tratamiento de la sepsis
TRANSCRIPT
TESIS DOCTORAL
TRATAMIENTO DE LA SEPSIS:
CONTROL DEL FOCO DE INFECCIÓN
MARÍA LUISA MARTÍNEZ GONZÁLEZ
DIRECTORES DEL TRABAJO DE TESIS
DR. ANTONI ARTIGAS RAVENTÓS DR. RICARD FERRER ROCA
TUTOR DEL TRABAJO DE TESIS
DR. FERRAN SEGURA
PROGRAMA DE DOCTORADO EN MEDICINA
DEPARTAMENTO DE MEDICINA
UNIVERSIDAD AUTÓNOMA DE BARCELONA
AÑO 2019
A mis padres
Al abuelo de Lalín
AGRADECIMIENTOS
Después de tantos años, de este largo camino, ver que todo puede cambiar de la forma más
inesperada pero que a pesar de todo nada me frena, que la pasión que siento por mi
profesión hace que siempre quiera mejorar. Con la tesis finalizo una etapa bonita que me
ha aportado cosas buenas y otras no tanto, pero lo más importante, con un final mejor en
todos los sentidos. Esta tesis me sirve también, y no menos importante, para poder
agradecer a tanta gente que me ha ayudado por el camino.
En primer lugar, querría agradecer a mis directores de tesis que me dieron la oportunidad y
confianza de trabajar con ellos en este gran proyecto.
Al Dr. Antoni Artigas, por todo el apoyo y confianza que me ha demostrado siempre.
Gracias por confiar en mí. Porque sin tu constancia, ejemplo, optimismo, ayuda e
insistencia, esta tesis, y otras muchas cosas, no hubieran llegado a su fin.
Al Dr. Ricard Ferrer, aún me acuerdo del día, hace ya unos cuantos años, en que tenía que
escoger entre dos caminos para hacer la tesis, y con muy buen criterio un compañero del
Taulí me dijo que siguiera investigando contigo. Que desde la “distancia” me enseñarías
mucho. No se equivocó. Gracias por todos estos años. Espero que éste no sea el último de
nuestros proyectos.
A Gemma Gomà, porque sin tu soporte y ayuda en el largo camino de los datos esta tesis
no se hubiera podido finalizar. A David Suárez, gracias por tu paciencia y ayuda. A Miguel
Ángel Díaz, alguien que he conocido recientemente pero que ha puesto el broche final a la
tesis.
A todo el grupo Edusepsis, porque sin ellos esto no hubiera sido posible.
A mis nuevos y no tan nuevos compañeros del HGC, mi nueva familia con la que disfruto
y aprendo cada día. Gracias por la paciencia y el apoyo durante estos meses. A Sandra y
Roser, porque sois unas compañeras estupendas. A Aitor (señor RAE ) y Marián, ha sido
un placer conoceros. Sin saberlo sois un gran apoyo. A Olga, ¡qué haría sin ti y tus rápidas
soluciones para todo! Gracias por el ordenador . A Jorge, mi mejor compañero de
guardia y ahora ya de mañanas. Por muchas más juntos.
A Mariló, gracias por todas las facilidades que me has dado para acabar la tesis, por
apoyarme y por tu energía, que se transmite.
A mi etapa en el H. Parc Taulí, por todo lo que aprendí y los buenos momentos que pasé.
Al Dr. Jordi Vallés y al Dr. Melcior Martínez, dos médicos excelentes que me mostraron el
respeto por el paciente y a anticiparse a los acontecimientos, algo que no sale en los libros.
Al Dr. Paco Baigorri, mi tutor y compañero de viaje durante muchos años.
A Sebas y Masip, Raquel y Caroline, porque fuisteis unos R-grandes inmejorables, siempre
apoyando y enseñando con paciencia. A Guillem, mi co-R y “hermano”. Por los buenos y
divertidos momentos que pasamos juntos. A mis resis pequeñas, tampoco me olvido de los
momentos compartidos con vosotras.
A Chelo, mi gran “mamá”, por todo lo vivido. A Sandra y Bego y a muchos otros
compañeros de residencia y del SEM. A Berta Cisteró, por los buenos años vividos
compartiendo el código sepsis.
A mis grandes compañeras de ratas y más, Raquel y Eva. Gracias por toda vuestra ayuda.
A toda mi familia y amigos, mis tíos, mis primos, mis abuelos. A Marta e Ione. A Laura,
por nuestra primera sesión en Castellón, el principio de todo jaja. A Ana, Xis y Mane. A
Guillem por estar siempre y demostrar que eres un buen amigo. A Xavi y Maikel; a Albert
mi cardiólogo de referencia, al resto de Gallenistas.
A Almu, mi hermana, amiga y alma gemela. Por tener la suerte de habernos conocido
siendo niñas.
A mis padres, por apoyarme siempre incluso cuando me decían que mejor no estudiar
medicina. No me equivoqué al escogerla, no lo cambiaría por nada. A mis hermanos, a
Iciar, por su apoyo investigador y a Alberto. A Andrea, una hermana más.
A Pablo, gracias por estar siempre a mi lado. Por apoyarme en todo y por recordarme
siempre que los límites me los pongo yo y solo yo. A Ana, por ser la niña más bonita,
dulce y especial que conozco. Gracias por hacerme feliz.
Y para acabar, esta tesis se la quiero dedicar a Ana Navas. Gracias Ana, por todo, porque
sobran las palabras. Porque eres el mejor ejemplo que he tenido como médico. Por
enseñarme tanto. Por demostrarme que mereces mucho la pena. No cambies. Ah! Te paso
el relevo ;-)
LISTADO DE ABREVIACIONES
ABISS. Antibiotic Intervention in Severe Sepsis.
APACHE II. Acute Physiology and Chronic Health Evaluation II score.
ATB. Antibiótico.
AUC. Área bajo la curva.
CDRE. Cuaderno electrónico de recogida de datos.
DAMPs. Danger-associated molecular patterns.
DE. Desviación estándar.
GTEIS. Grupo de trabajo de enfermedades infecciosas y sepsis.
IC. Intervalo de confianza.
IQR. Rango intercuartil.
ITU. Infección del tracto urinario.
NA. Noradrenalina.
NS. No significativa.
PAMPs. Pathogen-associated molecular patterns.
PCR. Proteína C reactiva.
PCT. Procalcitonina.
SEMICYUC. Sociedad Española de Medicina Intensiva, Crítica y Unidades Coronarias.
SRIS. Síndrome de respuesta inflamatoria sistémica.
SOFA. Sequential Organ Failure Assessment score.
SSC. Surviving Sepsis Campaign.
SNC. Sistema nervioso central.
UCI. Unidad de Cuidados Intensivos.
OR. Odds ratio.
VM. Ventilación mecánica.
RESUMEN
INTRODUCCIÓN
La sepsis, disfunción orgánica que pone en peligro la vida debido a la falta de regulación
de la respuesta del huésped a la infección, es una de las causas más frecuentes de ingreso
hospitalario y una de las primeras causas de mortalidad con una morbilidad asociada muy
importante. Dado que estamos delante de un síndrome infeccioso los dos pilares del
tratamiento de la sepsis son, por tanto, la administración del antibiótico de manera precoz y
adecuada y el drenaje percutáneo o quirúrgico del foco infeccioso si está presente.
OBJETIVOS
Objetivo principal:
Evaluar el impacto de una intervención múltiple de traslación del conocimiento dirigida a
mejorar el tratamiento antimicrobiano empírico de la sepsis y el shock séptico en la
mortalidad.
Objetivos secundarios:
1. Evaluar la frecuencia del control del foco y su implicación en el manejo de los
pacientes con sepsis y shock séptico.
2. Analizar el tiempo del control del foco y su impacto en la mortalidad.
MATERIAL Y MÉTODOS
Estudio prospectivo, multicéntrico con un diseño antes/después de un Programa
Educacional realizado en 99 Unidades de Cuidados Intensivos (UCI) distribuidas por toda
España. Se incluyeron consecutivamente todos los pacientes que ingresaban en UCI con
diagnóstico de sepsis y shock séptico en los períodos abril-junio 2011, abril-junio 2012 y
enero-abril 2013. El programa educativo o intervención se realizó durante un período de 3
meses (enero-marzo 2012). Se registraron datos clínicos y en relación con los tratamientos
administrados. El desenlace principal fue la mortalidad hospitalaria.
RESULTADOS GENERALES
Se incluyeron un total de 3663 pacientes con sepsis o shock séptico: 63.3% hombres, 64
(SD: 15.1) años, 70.1% médicos, 62.4% infecciones comunitarias, APACHE II medio 21.8
(16-27). Completaron los períodos pre-intervención y post-intervención 2628 pacientes
(1352 en la fase pre-intervención y 1276 en la post-intervención). El tiempo medio desde
el inicio de la sepsis hasta la administración del primer antibiótico fue significativamente
más corto en el período post-intervención comparado con el pre-intervención (2.0 (2.7) vs.
2.5 (3.6) horas; p=0.002). La proporción de pacientes que recibieron el tratamiento
antibiótico empírico inapropiado disminuyó del 8.9% en el grupo pre-intervención hasta el
6.5% en el post-intervención (p=0.024). La mortalidad hospitalaria global fue del 29.9%,
no encontrándose diferencias entre los dos períodos.
El 32% de los pacientes precisaron alguna técnica del control del foco; este grupo de
pacientes presentaba más edad, mayor proporción de shock y un mayor número de fallos
orgánicos en las primeras 6 horas desde el inicio de la sepsis comparado con los pacientes
que no precisaban control del foco. Adicionalmente, estos pacientes recibieron un peor
tratamiento inicial. El análisis multivariado mostró que los pacientes que precisaron control
del foco presentaron una mortalidad menor tanto de UCI como hospitalaria (OR 0.809 [IC
95%, 0.658–0.994]; p=0.044). Realizar el control del foco en las primeras 12 horas desde
el diagnóstico de la sepsis no se asoció a menor mortalidad.
CONCLUSIONES
Un programa educacional multicéntrico realizado a nivel nacional enfocado en la
importancia del manejo infeccioso de los pacientes con sepsis se asocia a una reducción en
el tiempo de administración del tratamiento antibiótico y también a un aumento de la
adecuación; y que presentar un foco infeccioso drenable se asocia a menor mortalidad.
Palabras clave: sepsis, shock séptico, antibiótico empírico, desescalamiento, control del
foco, mortalidad hospitalaria.
ABSTRACT
INTRODUCTION
Sepsis, a life-threatening organ dysfunction due to dysregulated host response to infection,
is common and its incidence seems to be increasing. Sepsis is associated with high
morbidity and mortality. Infection control is the cornerstone of treatment, and early
appropriate antimicrobial therapy and source control are considered essential aspects of
sepsis management.
OBJECTIVES
Main objective:
We aimed to evaluate the impact of a multifaceted educational intervention to improve
antibiotic treatment in patients with sepsis and septic shock.
Secondary objectives:
1. We aimed to assess the epidemiology of the need for source control and its role
in the management of these patients.
2. We hypothesized that delays in source control after onset of sepsis or septic
shock would worsen outcome.
PATIENTS AND METHODS
We prospectively studied all consecutive patients with sepsis/septic shock admitted to 99
intensive care units (ICUs) throughout Spain in two 4-month periods (before and
immediately after the 3-month intervention). We compared process-of-care variables
(resuscitation bundle and time-to-initiation, appropriateness, and de-escalation of empirical
antibiotic treatment) and outcome variables between the two cohorts. Also, we recorded
the need and time for source control technique. The primary outcome was hospital
mortality.
RESULTS
We included 3663 patients (men 63.3%, age 64 (SD: 15.1) years; APACHE II, 21.9 (16-
27). A total of 2628 patients were included during the pre-intervention (n=1352) and post-
intervention periods (n=1276). In the post-intervention cohort, the mean (SD) time from
sepsis onset to empirical antibiotic therapy was lower (2.0 (2.7) vs. 2.5 (3.6) h; p=0.002),
the proportion of inappropriate empirical treatments was lower (6.5% vs. 8.9%; p=0.024),
and the proportion of patients in whom antibiotic treatment was de-escalated was higher
(20.1% vs. 16.3%; p=0.004); the expected reduction in mortality did not reach statistical
significance (29.4% in the post-intervention cohort vs. 30.5% in the pre-intervention
cohort; p=0.544).
A total of 1173 patients (32%) underwent source control. Compared with patients who did
not require source control, patients who underwent source control were older, with a
greater prevalence of shock and major organ dysfunction. In addition, compliance with the
resuscitation bundle was worse in those undergoing source control. In patients who
underwent source control, crude ICU mortality was lower (21.2% vs 25.1%; p=0.010);
after adjustment for confounding factors, hospital mortality was also lower (odds ratio,
0.809 [95% CI, 0.658–0.994]; p=0.044). In this observational database analysis, source
control after 12 hours was not associated with higher mortality.
CONCLUSIONS
The ABISS intervention reduced the time to antibiotic administration and the proportion of
patients in whom antibiotic treatment was de-escalated, thus demonstrating that despite
advances in sepsis treatment in recent years, educational interventions can still improve the
delivery of care. Despite greater severity and worse compliance with resuscitation bundles,
mortality was lower in septic patients who underwent source control than in those who did
not. The time to source control could not be linked to survival in this observational
database.
Key words: sepsis, septic shock, timing of antibiotics, de-escalation, source control,
hospital mortality.
ÍNDICE
1.- PRESENTACIÓN: ............................................................................................................... 17
2.- INTRODUCCIÓN: ............................................................................................................... 19
3.- HIPÓTESIS: ...................................................................................................................... 29
4.- OBJETIVOS GENERALES: .................................................................................................. 31
5.- METODOLOGÍA GENERAL: ............................................................................................... 33
6.- RESULTADOS: .................................................................................................................. 39
7.- DISCUSIÓN GENERAL: ...................................................................................................... 51
8.- CONCLUSIONES GENERALES: ........................................................................................... 61
9.- LÍNEAS DE FUTURO: ......................................................................................................... 63
10.- BIBLIOGRAFÍA MÁS RELEVANTE .................................................................................... 65
11.- ANEXO 1: PUBLICACIONES DE LA TESIS DOCTORAL ....................................................... 75
12.- ANEXO 2: MATERIAL SUPLEMENTARIO ELECTRÓNICO ................................................... 97
13.- ANEXO 3: DATOS NO PUBLICADOS ............................................................................... 127
14.- ANEXO 4: MATERIAL EDUCATIVO ................................................................................ 131
15.- ANEXO 5: BECAS Y PREMIOS ........................................................................................ 135
16.- ANEXO 6: OTRAS PUBLICACIONES EN RELACIÓN A LA TESIS DOCTORAL ..................... 137
16
17
1.- PRESENTACIÓN:
La presente tesis doctoral está estructurada de acuerdo a las directrices del Marco
Regulador del Doctorado de la Escuela de Posgrado y del Departamento de Medicina de la
Universidad Autónoma de Barcelona (RD 1393/2007), modificada por el RD 861/2010 y
se presenta como compendio de publicaciones, tal y como ha aceptado la Comisión
Académica del programa de Doctorado de Medicina a fecha de 18 de julio de 2018.
Los estudios que conforman esta tesis doctoral pertenecen a una misma línea de
investigación dirigida a evaluar la importancia del manejo infeccioso adecuado y precoz en
pacientes con sepsis y shock séptico. En los diferentes apartados de esta tesis se exponen
los aspectos generales más destacados de los trabajos, y los aspectos más específicos están
detallados en las publicaciones que se encuentran en los anexos. La doctoranda ha
publicado estos dos trabajos en revistas internacionales de medicina crítica y con elevado
factor de impacto (revistas de primer cuartil):
1) María Luisa Martínez, Ricard Ferrer, Eva Torrents, Raquel Guillamat-Prats, Gemma
Gomà, David Suárez, Luis Álvarez-Rocha, Juan Carlos Pozo Laderas, Ignacio Martín-
Loeches, Mitchell M. Levy, Antonio Artigas, for the Edusepsis Study Group. Impact of
Source Control in Patients With Severe Sepsis and Septic Shock. Critical Care Medicine
2017 Jan;45(1):11-19. doi: 10.1097/CCM.0000000000002011. IP 7.05.
2) Ricard Ferrer*, María Luisa Martínez*, Gemma Gomà, David Suárez, Luis Álvarez-
Rocha, María Victoria de la Torre, Gumersindo González, Rafael Zaragoza, Marcio
Borges, Jesús Blanco, Eduardo Palencia Herrejón, Antonio Artigas and for the ABISS-
Edusepsis Study group. Improved empirical antibiotic treatment of sepsis after an
educational intervention: the ABISS-Edusepsis study. Crit Care. 2018 Jun;22(1):167. doi:
10.1186/s13054-018-2091-0. IP 6.425.
* Han contribuido por igual.
18
Todos los coautores han aprobado el uso, por parte de la doctoranda, de los estudios
presentados como trabajo de tesis doctoral.
Durante la realización de los estudios que conforman esta tesis la doctoranda ha
participado en otros trabajos vinculados a la presente tesis doctoral que se adjuntan en el
Anexo 6.
19
2.- INTRODUCCIÓN:
La sepsis, un síndrome complejo, está definida como una disfunción grave de órganos que
pone en peligro la vida debida a una respuesta anormal del organismo a la infección.
Cuando además se asocian alteraciones circulatorias, celulares y metabólicas que
condicionan un mayor riesgo de muerte, hablamos de shock séptico (1).
La sepsis es un problema importante de salud pública que afecta a millones de personas en
todo el mundo y representa una de las primeras causas de muerte siendo incluso su
mortalidad más elevada que la del infarto agudo de miocardio o el accidente
cerebrovascular (2). Cada año más de 19 millones de personas desarrollan sepsis (3).
Aproximadamente 14 millones sobreviven al alta hospitalaria pero con un pronóstico
posterior muy variable: la mitad se recupera, un tercio muere durante el siguiente año, y
una sexta parte presentará importantes secuelas físicas y/o psíquicas (4). Una modesta
disfunción orgánica en la sepsis se asocia a un 10% de mortalidad hospitalaria,
incrementándose de manera significativa a mayor número de disfunciones orgánicas siendo
reportadas incluso mortalidades de entre el 40% y el 80% para el shock séptico (1,5).
Además, su incidencia va en aumento, reflejando probablemente el envejecimiento
poblacional con más comorbilidades e incluso un mejor reconocimiento de la patología (1).
En Cataluña, por ejemplo, la incidencia de la sepsis es de 212.7 casos por cada 100.000
habitantes/año con una mortalidad hospitalaria del 21.6%. Como muestra la Figura 1,
existe un aumento de la incidencia anual del 7.3% y una reducción anual de la mortalidad
hospitalaria del 3.4% (6). Estos datos son similares a los publicados en varios países
diferentes, principalmente en EUA, Europa, Australia y Nueva Zelanda (7–11).
20
Figura 1. Número de casos, tasas de mortalidad y tasas de incidencia hospitalaria por sepsis en
Cataluña (2008-2012) (6). La incidencia de la sepsis aumentó de 12.809 casos a 20,228 casos en el período
de estudio de 5 años (media 16.460 casos por año), representando el 1.3 y el 2.1% (p <.0001) de los ingresos
hospitalarios y un aumento promedio anual de 6%. Sin embargo, la mortalidad hospitalaria disminuyó de
23.7 a 19.7% (p <.0001) para una reducción relativa anual de 3.4%.
FISIOPATOLOGÍA
La sepsis se produce por una respuesta inmunológica anómala del huésped a un patógeno
que puede amplificarse significativamente por factores endógenos que incluyen la
activación temprana de respuestas pro-inflamatorias y anti-inflamatorias, junto con
modificaciones importantes en vías no inmunológicas como las cardiovasculares,
neuronales, autonómicas, hormonales, bioenergéticas, metabólicas y de coagulación, las
cuales tienen importancia pronóstica. Lo que diferencia la sepsis de la infección es una
respuesta aberrante o anómala del huésped con la presencia de disfunción orgánica (1). La
teoría fisiopatológica actual expone que el trigger infeccioso desencadena una respuesta
inmune con liberación de citoquinas que se propaga independientemente del
desencadenante infeccioso subyacente (12). Es decir, serían los PAMPs (pathogen-
associated molecular patterns) quienes desencadenarían la respuesta inmune y los DAMPs
(danger-associated molecular patterns), factores liberados por el propio huésped, quienes
la mantendrían (13). Por tanto, según esta teoría, no importa lo rápido que elimines el
microorganismo responsable ya que es la respuesta inflamatoria la que mantiene la sepsis.
21
Recientemente, Kumar (14) propone una nueva teoría para explicar la fisiopatología de la
sepsis donde la carga bacteriana es el principal impulsor de la disfunción orgánica y el
shock séptico y, por tanto, la rápida eliminación de los patógenos sería fundamental para
cambiar la evolución de estos pacientes. Este paradigma incorpora el concepto de shock
irreversible y sugiere que el mejor enfoque para el tratamiento es minimizar el tiempo en el
que está presente una cantidad de microorganismos suficiente para generar shock (Figura
2). Por lo tanto, el tratamiento antimicrobiano potente, adecuado y precoz así como el
adecuado control del foco infeccioso serían los componentes clave en el tratamiento de la
sepsis.
Figura 2 (14). 2A: Teoría fisiopatológica de la sepsis propuesta por Kumar (ver texto para mayor
información). 2B y 2C: El tratamiento antimicrobiano precoz, adecuado y potente conseguiría cambiar la
evolución de la sepsis
2A
2C
2B
22
CAMPAÑA SOBREVIVIR A LA SEPSIS (SURVIVING SEPSIS CAMPAIGN)
En los últimos años se ha avanzado de manera significativa en el conocimiento de la
sepsis, no sólo a nivel epidemiológico sino también desde el punto de vista fisiopatológico
y de tratamiento. Estos avances terapéuticos, así como la elevada incidencia y mortalidad
de la sepsis, llevaron en el año 2002, en colaboración conjunta de la Society of Critical
Care Medicine (SCCM) y la European Society of Intensive Care Medicine (ESICM), al
desarrollo de la Surviving Sepsis Campaign (SSC): una campaña de esfuerzo mundial
dirigida a mejorar el tratamiento de los pacientes con sepsis y así disminuir la
morbimortalidad. Esta campaña ha ido avanzando en diferentes fases y ha consistido
principalmente en la creación de unas guías internacionales de tratamiento basadas en la
evidencia científica, la implementación de un programa de mejora del cumplimiento de las
recomendaciones terapéuticas, y la recogida, el análisis y posterior publicación de datos de
más de 30.000 registros de pacientes con sepsis y shock séptico de todo el mundo
(http://www.survivingsepsis.org).
Las primeras guías de la SSC (International Guidelines for Management of Sepsis and
Septic Shock) se publicaron en el 2004 (15) y desde entonces han sido actualizadas en
función de la nueva evidencia científica: en el 2008, 2012 y la última versión del 2016
(16). La “sepsis bundle” ha sido el elemento principal de traslación del conocimiento de
estas guías. Un bundle o paquete de medidas de tratamiento es un grupo de intervenciones
con las que se consiguen mejores resultados si se aplican de forma conjunta que si se hace
por separado. Los elementos individuales que constituyen el bundle se seleccionan si se
considera que tienen suficiente evidencia científica como para incorporarlos a la práctica
clínica habitual. El estudio de la presente tesis doctoral se ha basado en las guías de la SSC
del 2008 (8), donde la bundle de tratamiento para la sepsis se conocía como “Bundle de
Resucitación”, e incluía:
1. Determinación de lactato en plasma.
2. Obtención de hemocultivos previos a la administración de antibióticos.
3. Administración de antibióticos de amplio espectro en menos de 3 horas en
pacientes de urgencias y en menos de 1 hora en pacientes ingresados.
23
4. En presencia de hipotensión:
1. Iniciar administración de cargas de volumen (20 ml/kg de cristaloides o
equivalente en coloides).
2. Fármacos vasoactivos para mantener una Presión arterial media (PAM) ≥
65 mmHg.
5. En presencia de hipotensión refractaria a volumen o lactacidemia > 36 mg/dl:
1. Conseguir una Presión Venosa Central (PVC) ≥ 8 mmHg.
2. Conseguir una Saturación venosa central (SvcO2) ≥ 65%.
Describe 7 acciones que deben ser realizadas en las primeras 6 horas desde el diagnóstico
de la sepsis. Además, las guías incluyen otra serie de recomendaciones a realizar dentro de
las primeras 24 horas (control del foco infeccioso, administración de corticoides
sustitutivos, ventilación mecánica protectiva, control estricto de la glicemia, entre otras).
La SSC, en colaboración con el Institute for Healthcare Improvement, ha llevado a cabo,
no sólo el desarrollo de las guías, sino también una intervención múltiple de mejora de la
calidad del tratamiento de la sepsis con un registro internacional de pacientes con sepsis o
shock séptico. Tal y como muestra el análisis de los 30.000 pacientes incluidos durante los
primeros 7.5 años de la campaña, la participación en el registro y en las actividades de la
SSC se asoció a un incremento progresivo en el cumplimiento de las recomendaciones de
tratamiento de la sepsis con una caída progresiva en la mortalidad (17). Son muchos los
estudios que demuestran que seguir las recomendaciones de la SSC se asocia a mejores
resultados (18–20). La notable evidencia en la literatura que demuestra la asociación entre
el cumplimiento de las bundles y la mejora en la supervivencia de los pacientes con sepsis
y shock séptico ha llevado a las diferentes organizaciones médicas de los EUA a adoptar
las recomendaciones de la SSC como una medida pública obligatoria (21–23). La
importante relación entre la aplicación de las bundles y la supervivencia ha sido
confirmada recientemente en una publicación de los resultados de esta medida adoptada, la
New York State Initiative (24).
24
La conclusión más importante que se ha obtenido gracias a la amplia evidencia científica
es que podemos afirmar que la sepsis es una emergencia médica, un síndrome tiempo
dependiente que requiere un manejo estandarizado, secuencial y multidisciplinar con la
necesidad de coordinación de recursos hospitalarios e interhospitalarios. Al igual que con
el politraumatismo, el infarto agudo de miocardio y el accidente cerebrovascular, la rápida
identificación y el manejo inmediato apropiado en las primeras horas desde el inicio de la
sepsis puede influir en los resultados.
Toda esta evidencia científica ha llevado a la SSC a actualizar recientemente (2018), no las
guías, pero sí la bundle de tratamiento inicial acortando incluso el tiempo a realizar estas
medidas a la primera hora desde el diagnóstico de la sepsis, lo que se conoce como “Hour-
1 bundle” (Figura 3) (21):
Figura 3. Hour-1 Bundle Surviving Sepsis Campaign (21)
25
SEPSIS-3
Dado que estamos, por tanto, delante de una patología tiempo-dependiente, no sólo han
sido revisadas las guías de tratamiento de la sepsis, sino que también se han generado
nuevas definiciones con el fin de hacer más uniforme la selección de pacientes y tratar de
reconocer con mayor rapidez la sepsis. Estas nuevas definiciones, conocidas como Sepsis-
3, han sido publicadas recientemente en The Third International Consensus Definitions for
Sepsis and Septic Shock (1). La primera definición se generó en el año 1991 (Sepsis-1) (25)
cuando se propusieron los criterios de "síndrome de respuesta inflamatoria sistémica"
(SRIS), que han sido de gran utilidad pero son muy inespecíficos, y la segunda revisión
data del 2001 (Sepsis-2) y aporta pocos cambios a la primera (26). En esta tercera
actualización, principalmente, desaparece el concepto de “sepsis grave”, término
redundante, pasando a utilizarse únicamente los términos sepsis y shock séptico. Por tanto,
los criterios de SRIS dejan de estar incluidos en la definición de sepsis, considerándose una
respuesta inflamatoria a la infección pero no asociada a gravedad.
La sepsis ahora se define como "una disfunción orgánica potencialmente mortal causada
por una respuesta anómala del huésped a la infección". Para identificar la disfunción
orgánica, establecen un aumento de 2 ó más de 2 en la puntuación de la escala SOFA
(Sequential Organ Failure Assessment score).
Hablamos de shock séptico cuando además se asocian alteraciones en el metabolismo
circulatorio y celular identificadas por la necesidad de tratamiento vasopresor para
mantener una presión arterial media de 65 mmHg y un nivel de lactato en suero superior a
2 mmol/L después de una reanimación con líquidos adecuada.
ESTUDIO EDUSEPSIS
El proyecto internacional de la SSC en España tuvo un formato y un diseño específico
(Estudio Edusepsis). El estudio Edusepsis, coordinado por nosotros mismos, contó con la
participación de 77 Unidades de Cuidados Intensivos (UCIs) españolas donde se
implementó una intervención múltiple para la mejora del cumplimiento de las
recomendaciones terapéuticas mediante las bundles. Se demostró que la intervención
26
realizada fue capaz de mejorar el cumplimento de las recomendaciones terapéuticas y
también de reducir la mortalidad (27).
Sin embargo, no todos los efectos de la intervención fueron sostenidos; por ejemplo, el uso
precoz de antibióticos disminuyó en el seguimiento a largo plazo. Esto es especialmente
relevante si se considera el resultado del análisis secundario del estudio Edusepsis (28)
donde se evaluó el impacto individual de cada componente de la bundle de resucitación
sobre la mortalidad; se encontró que la administración de antibiótico de amplio espectro en
la primera hora en los pacientes con sepsis reducía la mortalidad significativamente y que
el efecto beneficioso desaparecía progresivamente en función del retraso antibiótico.
TRATAMIENTO ANTIMICROBIANO EMPÍRICO PRECOZ Y ADECUADO EN LA SEPSIS
El efecto beneficioso de la administración precoz de antibióticos en la sepsis no es un
concepto nuevo. De hecho, ya a principios del 1900, Ehrlich describió por primera vez el
concepto de “hit hard and fast” (“golpear fuerte y rápido") (29). Más recientemente,
Kumar, demostró en más de 2000 pacientes con shock séptico que recibir el tratamiento
antibiótico en la primera hora desde el diagnóstico se asociaba a una supervivencia del
79.9% y que cada hora de retraso, dentro de las primeras 6 horas, se asociaba con una
disminución de la supervivencia del 7.6% (30,31). Un metaanálisis posterior (18) muestra
que el tratamiento mediante bundles reduce la mortalidad y se asocia a un incremento en el
uso precoz de antibióticos de amplio espectro y los resultados de la SSC internacional,
confirman, en más de 17000 pacientes, que la administración tardía de antibióticos se
asocia a una mayor mortalidad hospitalaria (32).
Por tanto, los clínicos deberían prescribir rápidamente el tratamiento antibiótico de amplio
espectro en la sepsis en función de los hallazgos clínicos, sin esperar a los resultados
microbiológicos. Además el tratamiento prescrito debe administrarse con la máxima
celeridad, por lo que se debe facilitar desde el punto de vista organizativo que este objetivo
pueda cumplirse (disponer de un stock de antibióticos adecuado, personal de enfermería
suficiente, etc). En el estudio Edusepsis el porcentaje de cumplimiento de la administración
precoz de antibióticos fue: basal: 65%, post-intervención: 71% y al año de la intervención:
56%, existiendo un amplio margen de mejora. Una limitación del estudio Edusepsis fue
que no se recogió la adecuación del tratamiento antibiótico. Múltiples estudios previos han
27
identificado peores resultados con la administración del tratamiento antibiótico inadecuado
(31,33–35). Por lo tanto, el efecto beneficioso de la precocidad del tratamiento antibiótico
en la sepsis puede magnificarse administrando el antibiótico adecuado. Son necesarias
también intervenciones que incrementen la adecuación del tratamiento antibiótico empírico
en la sepsis. Las guías de la SSC también recomiendan reevaluar el tratamiento antibiótico
posteriormente para determinar si es posible la reducción del espectro o desescalamiento
(16). La no reducción del tratamiento antibiótico se asocia con peor pronóstico (36,37) y
podría conducir al desarrollo de resistencias microbianas (38,39).
Dada la rápida desaparición del efecto de la intervención inicial, el amplio margen de
mejora que existe en la precocidad del tratamiento antibiótico y en la adecuación de las
pautas empíricas y los potenciales beneficios sobre la mortalidad de la sepsis,
consideramos de gran interés hacer un esfuerzo de traslación del conocimiento focalizado
en la antibioticoterapia empírica precoz. Para ello diseñamos una intervención múltiple
dirigida a mejorar la precocidad y adecuación del tratamiento antibiótico empírico en la
sepsis. Este proyecto, llamado Estudio ABISS-Edusepsis, es la base principal de la
presente tesis doctoral y pretende evaluar el impacto de esta intervención y analizar su
eficiencia.
CONTROL DEL FOCO INFECCIOSO
El término “control del foco” data desde la antigüedad y engloba todas aquellas medidas
físicas que son utilizadas para controlar y eliminar el origen de la infección y modificar los
factores locales que promueven el crecimiento de los microorganismos e influyen en la
función normal de las defensas del huésped, incluyendo también el proceso de restaurar el
estado anatómico y funcional previo a la infección. Los cuatro principios en los que se basa
el control del foco son: drenaje, desbridamiento, retirada de dispositivos y restauración de
la anatomía y funcionalidad del área dañada (40).
El origen de la infección en la sepsis es muy variable dando lugar a diferentes posibles
focos susceptibles de ser controlados: colangitis, colecistitis, abscesos intraabdominales,
perforación gastrointestinal, isquemia intestinal, pielonefritis obstructiva, infección
necrotizante de tejidos blandos, artritis séptica, empiema, e infecciones de dispositivos
implantados, entre otros. No todos los pacientes van a presentar un foco susceptible de ser
28
drenado, pero, tal y como recomiendan las guías de la SSC (16), en todo paciente con
sepsis o shock séptico se debe evaluar la presencia de un foco de infección susceptible de
ser erradicado mediante algún tipo de intervención.
Existe un amplio espectro de medidas para controlar el foco (percutáneas y quirúrgicas) y
su utilidad en cada paciente dependerá no sólo del sitio y naturaleza de la infección, sino
también del estado previo del enfermo y de la disponibilidad de los recursos tanto técnicos
como humanos. En la situación ideal, el mejor método sería aquel que consiguiese eliminar
totalmente el foco de infección con el mínimo de traumatismo para el paciente. La decisión
de qué intervención aplicar es compleja, habiéndose de plantear de forma individual para
cada paciente y haciendo, por tanto, necesario un enfoque multidisciplinar de estos
enfermos.
La última actualización de las guías de la SSC del 2016, además, recomienda que el
control del foco se lleve a cabo lo antes médica y logísticamente posible desde que se
realiza el diagnóstico de sepsis (“as soon as posible”) (16). Esta recomendación ha ido
modificándose a lo largo de los años, desde la recomendación del 2008 donde se proponía
controlar el foco en menos de 6 horas a la recomendación de las guías del 2012 en que se
amplía a las primeras 12 horas desde la presentación de la sepsis. La falta de evidencia
científica debido a la dificultad para realizar apropiados estudios controlados y
randomizados, algo éticamente difícil de llevar a cabo, hace que no exista una
recomendación más precisa en cuanto al tiempo de realizar el control del foco. La mayoría
de trabajos publicados son pequeños, observacionales y realizados en focos concretos, y no
siempre en pacientes con sepsis así como tampoco en el global de los pacientes con sepsis
o shock séptico.
Aunque parece lógico que el control del foco representa una medida principal en el
tratamiento de la sepsis, ha recibido menos atención que otros tratamientos recomendados
por la SSC. Además, el impacto de llevar a cabo una técnica del control del foco en el
global de pacientes con sepsis no ha sido evaluado hasta el momento. Es por ello, que nos
planteamos un análisis secundario del estudio ABISS-Edusepsis con el objetivo de evaluar
la epidemiología de la necesidad del control del foco en los pacientes con sepsis y shock
séptico, su papel en el manejo global de estos pacientes y analizar si un retraso en el
control del foco se asocia a peor evolución.
29
3.- HIPÓTESIS:
Dada la importancia del control de la infección en el manejo de los pacientes con sepsis y
shock séptico, el trabajo de la presente tesis doctoral pretende, mediante un programa
educativo del manejo de la sepsis y el shock séptico a escala nacional, mejorar el
tratamiento antimicrobiano de estos pacientes reduciendo el tiempo de administración,
aumentando la proporción de pacientes que reciben el tratamiento antimicrobiano empírico
adecuado, favoreciendo el desescalamiento, y por tanto, conseguir un descenso en la
mortalidad.
Adicionalmente, hipotetizamos que la necesidad del control del foco infeccioso, mediante
técnica percutánea o quirúrgica, es frecuente en los pacientes con sepsis y shock séptico,
tiene implicaciones pronósticas y en el manejo global, y que un retraso en su aplicación se
asocia a un aumento de la mortalidad.
30
31
4.- OBJETIVOS GENERALES:
OBJETIVO PRINCIPAL:
1. Evaluar el impacto de una intervención múltiple de traslación del conocimiento dirigida
a mejorar el tratamiento antimicrobiano empírico de la sepsis y el shock séptico en la
mortalidad.
Este objetivo se desarrolla en la publicación #1 del anexo 1.
OBJETIVOS SECUNDARIOS:
1. Evaluar la frecuencia del control del foco y su implicación en el manejo de los pacientes
con sepsis y shock séptico.
2. Analizar el tiempo del control del foco y su impacto en la mortalidad.
Estos objetivos se desarrollan en la publicación #2 del anexo 1.
32
33
5.- METODOLOGÍA GENERAL:
A través del grupo de estudio Edusepsis y el grupo de trabajo sobre enfermedades
infecciosas y sepsis (Grupo GTEIS) de la Sociedad Española de Medicina Intensiva,
Crítica y Unidades Coronarias (SEMICYUC), invitamos a 115 centros de toda España a
participar en una intervención educativa nacional para mejorar el control de la infección en
la sepsis: el estudio ABISS-Edusepsis. Participaron un total de 99 unidades de cuidados
intensivos médico-quirúrgicas distribuidas por toda España. El estudio fue aprobado por el
Comité de Ética de Investigación Clínica de cada centro participante. Se garantizó el
anonimato de los pacientes.
DISEÑO DEL ESTUDIO
Estudio prospectivo con un diseño antes/después de un programa educacional. Se comparó
una fase llamada pre-intervención donde se incluían de manera consecutiva todos los
pacientes con sepsis y shock séptico que ingresaban en las UCIs participantes (Abril –
Julio 2011), con la fase post-intervención, período donde se incluían de manera
consecutiva todos los pacientes con sepsis y shock séptico que ingresaban en las UCIs
participantes durante 4 meses después del programa educativo (Abril – Julio 2012). El
programa educativo o intervención se realizó durante un período de 3 meses (Enero –
Marzo 2012). Se llevó a cabo un tercer período de recogida de datos de 4 meses para el
seguimiento a largo plazo (Enero – Abril 2013).
Todas las admisiones en las UCIs procedentes del departamento de urgencias, planta de
hospitalización o quirúrgicas y todos los pacientes ingresados en las UCIs, se examinaron
diariamente para detectar la presencia de sepsis o shock séptico.
Debido a que el estudio se llevó a cabo entre los años 2011-2013, las definiciones
empleadas para la inclusión de pacientes fueron las de la conferencia internacional de
consenso del 2003 ("Sepsis 2") (26), y no las definiciones del Tercer Consenso
Internacional del 2016 ("Sepsis 3"), recientemente publicadas (1). En resumen: la sepsis
grave se definió como la sepsis asociada con al menos una disfunción orgánica aguda. El
shock séptico se definió como insuficiencia circulatoria aguda sostenida (presión arterial
sistólica < 90 mm Hg, presión arterial media < 65 mm Hg o una reducción de la presión
34
arterial sistólica < 40 mm Hg desde el inicio) a pesar de la reanimación adecuada con
volumen.
Se consideró inicio de la sepsis (hora cero) el momento en que se realiza una primera
valoración compatible con sepsis (anotación médica o de enfermería), o en su defecto, el
momento en que se inician las medidas de resucitación y tratamiento. En el caso de que el
paciente llegase al hospital con criterios de sepsis o shock séptico, se recogía como hora
cero la hora de triaje.
RECOGIDA DE DATOS
Se utilizó un Cuaderno electrónico de Recogida de Datos (CRDe) para registrar los datos
clínicos (datos demográficos, comorbilidad, tipo y origen de la infección, escalas de
gravedad) y en relación con los tratamientos administrados incluidos en las bundles de
resucitación de la SSC (8) y su tiempo de administración.
Además se recogieron datos del tratamiento antibiótico empírico administrado, tiempo de
administración, adecuación del tratamiento y desescalamiento. Se consideró el tratamiento
como adecuado cuando al menos se hubiera incluido en el tratamiento antibiótico empírico
un fármaco efectivo en función de la susceptibilidad en el antibiograma del
microorganismo de un cultivo significativo. Se evaluó el desescalamiento del tratamiento
antibiótico a las 72 horas, definido como una reducción del espectro antibiótico en función
de la respuesta clínica, el resultado de los cultivos y la susceptibilidad de los patógenos
identificados.
Los pacientes fueron seguidos hasta la muerte o el alta hospitalaria. El desenlace principal
fue la mortalidad hospitalaria. Los desenlaces secundarios incluyeron días de ventilación
mecánica, días de soporte vasoactivo, días de estancia en UCI y hospital y mortalidad en
UCI.
INTERVENCIÓN
Durante los 3 meses que duró la fase intervención se implementó un programa educacional
múltiple que principalmente consistió en:
- Se distribuyó a cada centro el soporte para las sesiones clínicas que incluía: identificación
precoz de pacientes con sepsis, sistemática de toma de muestras microbiológicas,
35
algoritmos de tratamiento antibiótico precoz y adecuado al foco sospechado de sepsis,
importancia del control quirúrgico o percutáneo del foco y casos clínicos ilustrativos. Se
organizaron sesiones interactivas que incluían a todo el personal médico y de enfermería
de las áreas de urgencias, UCI y plantas de hospitalización que habitualmente atienden
pacientes sépticos.
- Se suministraron los algoritmos de soporte para la prescripción del tratamiento antibiótico
en formato de bolsillo y póster (se permitió adaptar en función del patrón de resistencias
antimicrobianas del centro) que se distribuyó a todos los asistentes a las sesiones
interactivas (ver Anexo 4).
- Para ayudar en la prescripción de antibióticos, se facilitó un sistema electrónico de apoyo
a la decisión clínica (http://www.es.dgai-abx.de).
- Se colocaron recordatorios de la importancia de la administración precoz de antibióticos
en las salas de trabajo y áreas de preparación de medicación.
- Se evaluaron los stocks de antibióticos de amplio espectro de las Urgencias y UCI de los
centros participantes. Se recomendó incorporar stock de antibióticos de amplio espectro en
los centros que no dispusieran.
- Los asistentes a las sesiones recibían, durante la fase educacional, mensajes de texto
semanales al teléfono móvil y al correo electrónico con recordatorios reiterando los puntos
más importantes de la educación. Adicionalmente, se desarrolló un videojuego para poner
en práctica mediante simulación clínica el manejo del paciente con sepsis
(http://www.edusepsis.org/en/training.html).
- Se registró la intervención realizada por cada centro: número de sesiones realizadas,
número de asistentes, difusión de las auditorías semanales. Se suministró información
semanalmente sobre los tiempos de administración de antibióticos en cada centro y en el
global del estudio (audit-feedback) haciéndose difusión de esta información a todo el
personal asistencial posteriormente.
36
ANÁLISIS ESTADÍSTICO
Para evaluar la efectividad de la intervención, se comparó el cumplimiento de los
diferentes procesos y resultados de los pacientes incluidos en el período pre-intervención
frente al período post-intervención. Se realizaron las mismas comparaciones entre los
pacientes incluidos en el período post-intervención y en el período largo plazo.
Se realizó un análisis de regresión lineal múltiple para determinar la asociación entre la
intervención y el tiempo de administración del tratamiento antimicrobiano. Además, como
análisis de sensibilidad, realizamos un modelo de regresión segmentado para estimar el
efecto de la intervención en reducir el tiempo de administración del antibiótico.
Así mismo, llevamos a cabo un modelo de regresión logística múltiple por pasos para
evaluar el impacto de la intervención en la mortalidad hospitalaria. Las variables incluidas
en el modelo de regresión logística fueron aquellas con una relación en el análisis
univariado (p ≤ 0.1) o con una posible relación plausible con el resultado. El modelo final
incluyó intervención, edad, sexo, comorbilidades, APACHE II score, SOFA score, tipo de
infección y origen de la sepsis como variables independientes.
Para el análisis de los datos se utilizó la versión 15.0 de SPSS (Chicago, IL, USA). La
significación estadística se estableció para valores de p < 0.05.
5.1- METODOLOGÍA OBJETIVOS SECUNDARIOS:
Análisis secundario del estudio ABISS-Edusepsis donde se incluyeron todos los pacientes
con sepsis y shock séptico de las 99 UCIs participantes en los diferentes períodos del
estudio.
Se recogió, además de los datos expuestos en la metodología general, si el paciente había
precisado o no un técnica percutánea o quirúrgica para el control del foco. La decisión de
llevar a cabo el control del foco se realizó de manera individual por cada clínico tratante
basándose en los protocolos de cada centro. Se registró también el tiempo desde el inicio
de la sepsis o el shock séptico hasta el procedimiento. Se dividieron a los pacientes en
37
aquellos que recibieron un control del foco precoz (< 12 horas) versus tardío (≥ 12 horas)
de acuerdo a las guías de la SSC del 2012 (41).
ANÁLISIS ESTADÍSTICO
Para evaluar el impacto del control del foco en los pacientes con sepsis se realizó primero
un análisis comparativo de los pacientes que habían precisado control del foco frente a los
que no. Así mismo, se llevó a cabo un análisis de regresión lineal múltiple con mortalidad
hospitalaria y otro con mortalidad en UCI como variables dependientes y necesidad de
control del foco y otros factores de confusión como variables independientes.
Para el grupo de pacientes que sí precisaron alguna técnica de control del foco,
comparamos los pacientes que lo habían recibido de manera precoz (< 12 horas) frente a
tardío (≥ 12 horas). Además, se estudiaron otros puntos de corte (datos no publicados, se
adjuntan en el Anexo 3): < 4 horas frente a ≥ 4 horas, y otro modelo donde comparamos <
4 horas, 4-8 horas, 8-12 horas, 12-24 horas y ≥ 24 horas. Para determinar el tiempo del
control del foco que mejor se asocia a mortalidad, se analizó el tiempo del control del foco
como variable continua y se llevó a cabo un análisis de ROC con cálculo del área bajo la
curva (AUC).
Realizamos también un análisis de regresión lineal múltiple para evaluar el impacto del
tiempo del control del foco (< 12 horas vs ≥ 12 horas) en la mortalidad hospitalaria. Para
comprender mejor la importancia del tiempo del control del foco, aplicamos el mismo
modelo de regresión para infecciones abdominales, urológicas y de piel y tejidos blandos.
Adicionalmente, como análisis de sensibilidad, dado que la sepsis de origen urinario se
asocia a menor mortalidad (42) siendo por tanto un potencial factor de confusión, se
realizó un análisis secundario donde excluimos a este grupo de pacientes tanto para el
análisis global como para evaluar la relación del tiempo del control del foco con la
mortalidad (datos no publicados, se adjuntan en el Anexo 3).
38
39
6.- RESULTADOS:
6.1- RESULTADOS OBJETIVO PRINCIPAL:
Completaron los períodos pre-intervención y post-intervención 77 de las 99 UCIs
participantes, reclutándose un total de 2628 pacientes (1352 en la fase pre-intervención y
1276 en la fase post-intervención). El seguimiento a largo plazo se pudo hacer en 50 de las
77 UCIs (n = 830).
Comparado con los pacientes de la fase pre-intervención, los pacientes del grupo post-
intervención presentaron un APACHE II discretamente menor y una mayor proporción de
infecciones de origen comunitario, sin encontrar otras diferencias epidemiológicas
significativas (Tabla 1, Publicación 1).
Después de la intervención se produjo una mejora en el cumplimiento de las
recomendaciones terapéuticas incluidas en las bundles de resucitación de la SSC (Figura
4).
Figura 4. Cumplimiento (%) de las recomendaciones del bundle de las primeras 6 horas
en los períodos pre-intervención y post-intervención. ns, no significativo; * p < 0.05;
*** p < 0.005
100
60
50
40
20
Lactato Hemocultivos Antibiótico Fluidos +
vasopresores
Pre-intervención
Post-intervención
***
*
*
ns
0
%
40
El tiempo medio desde el inicio de la sepsis hasta la administración del primer antibiótico
empírico fue significativamente más corto en el período post-intervención comparado con
el período pre-intervención (2.0 (2.7) vs. 2.5 (3.6) horas; p = 0.002). Incluso, después de
ajustar por características demográficas, severidad, tipo y origen de infección, el tiempo
medio de antibiótico fue significativamente menor en el período post-intervención (-0.45
(IC 95%: -0.75 a -1.56); p = 0.003) mostrando una asociación entre la intervención y el
tiempo de antibiótico (Tabla 1, Material suplementario de la Publicación 1).
La Figura 5 muestra el tiempo medio de antibiótico desde el inicio de la sepsis por mes en
los períodos pre-intervención (mes 1 al mes 4) y post-intervención (mes 1 al mes 4). El
análisis de regresión segmentada mostró un cambio significativo en el tiempo de
antibiótico (-0.92 (IC 95%: -1.51 a -0.33); p = 0.010), indicando un cambio abrupto efecto
de la intervención educacional (ver también Tabla 2, Material suplementario de la
Publicación 1).
Figura 5. Análisis de regresión segmentada de datos. Tiempo medio desde el inicio de la sepsis hasta la
administración del tratamiento antimicrobiano antes y después del programa educacional. Se excluyeron los
pacientes con antibiótico previo al inicio de la sepsis (n = 858)
Tie
mp
o m
ed
io d
el
pri
mer
an
tib
ióti
co (
h)
Inicio del período post-intervención
3.0
2.5
2.0
1.5
1 2 3 4 1 2 3 4
Meses antes y después de la intervención
41
La proporción de pacientes que recibió el tratamiento antibiótico empírico inapropiado
disminuyó del 8.9% en el grupo pre-intervención hasta el 6.5% en el grupo post-
intervención (p = 0.024) (Figura 6A) y la proporción de pacientes a quienes se desescaló el
tratamiento antibiótico fue mayor en el grupo post-intervención (16.3% vs 20.1% en el
grupo post-intervención; p = 0.004) (Figura 6B).
Figura 6A. Evaluación del tratamiento antibiótico a las 72h. * p < 0.05
Figura 6B. Cambio del tratamiento antibiótico a las 72h. *** p < 0.005
Pre-intervención
intervención
Post-intervención
intervención
Tratamiento
inapropiado
60
*
40
20
0
Tratamiento
apropiado Cultivos
negativos
Cultivos
no realizados
Éxitus
< 72h
Reducción
del espectro No cambio Cambio por
mala
evolución
clínica
Cambio por
microorganismo
no cubierto
Otros
60
40
20
0
Pre-intervención
intervención
Post-intervención
intervención
***
%
%
42
No se obtuvieron diferencias estadísticamente significativas en los días de ventilación
mecánica (VM) y días de tratamiento vasopresor, así como tampoco en los días de estancia
en UCI y hospitalaria entre los períodos pre y post-intervención. La mortalidad hospitalaria
global fue del 29.9% no encontrándose diferencias entre los dos períodos (Tabla 1).
Tabla 1. Días de ventilación mecánica y vasopresores, días de estancia en UCI y hospital,
mortalidad en UCI y hospital, en los períodos pre-intervención y post-intervención
Grupo pre-
intervención
(n = 1352)
Grupo post-
intervención
(n = 1276)
P
Días de VM, media (DE) 6.9 (14.4) 6.6 (12.4) 0.577
Días de vasopresores, media (DE) 4.0 (8.0) 4.3 (7.0) 0.369
Días de estancia en UCI, media (DE) 12.0 (17.0) 11.5 (14.9) 0.443
Días de estancia hospitalaria, media (DE) 30.0 (29.7) 28.4 (28.9) 0.161
Mortalidad, n (%)
UCI 332 (24.6) 301 (23.6) 0.562
Hospital
412 (30.5) 375 (29.4) 0.544
VM, ventilación mecánica; DE, desviación estándar; UCI, unidad de cuidados intensivos.
El análisis de regresión logística no mostró relación entre la intervención y la mortalidad
hospitalaria (Tabla 2).
43
Tabla 2. Análisis multivariado de factores de riesgo para mortalidad hospitalaria
OR IC 95% P
Grupo intervención
1.08
0.89-1.31
0.419
Edada 1.02 1.01-1.03 <0.001
Sexob 0.82 0.67-1.01 0.057
SOFAa 1.11 1.07-1.15 <0.001
APACHE IIa 1.08 1.06-1.10 <0.001
CHARLSONa 1.06 1.02-1.11 0.005
Tipo de infecciónc
Nosocomial 2.03 1.61-2.56 <0.001
UCI 2.49 1.61-3.87 <0.001
Asociada a la asistencia sanitaria 1.33 0.99-1.79 0.058
Origen de la sepsisd
Abdominal 0.79 0.63-1.00 0.051
Urinaria 0.25 0.18-0.35 <0.001
Meningitis 1.11 0.60-2.05 0.742
Piel y tejidos blandos 0.69 0.47-1.03 0.072
Bacteriemia por catéter 0.54 0.27-1.09 0.084
Otras infecciones 1.00 0.66-1.52 0.999
OR, odds ratio; IC, intérvalo de confianza; SOFA, Sequential Organ Failure Assessment score; APACHE II, Acute
Physiology and Chronic Health Evaluation II score; UCI, Unidad de Cuidados Intensivos. aPor cada punto de aumento. bComparado con sexo masculino. cComparado con infección comunitaria. dComparado
con infección respiratoria.
44
Seguimiento a largo plazo
En las 50 UCIs que hicieron el seguimiento a largo plazo, no hubo diferencias destacables
entre los pacientes del grupo post-intervención y el grupo largo plazo, a excepción de una
mayor proporción de neumonía en el grupo de largo plazo (Tabla 3, Material
suplementario de la Publicación 1). El porcentaje de pacientes que cumplieron las medidas
de resucitación se mantuvo estable con respecto al período post-intervención. El tiempo
medio de administración del antibiótico aumentó discretamente pero sin diferencia
estadísticamente significativa respecto a la post-intervención, manteniéndose estable
también el desescalamiento antibiótico a las 72 horas (Tabla 4, Material suplementario de
la Publicación 1). No se encontraron diferencias en la mortalidad (Tabla 5, Material
suplementario de la Publicación 1).
6.2- RESULTADOS OBJETIVOS SECUNDARIOS:
Control del foco vs no control del foco
Durante los 3 períodos del estudio ABISS-Edusepsis se incluyeron un total de 3663
pacientes con diagnóstico de sepsis o shock séptico; el 32% (1173) precisaron alguna
técnica del control del foco, percutánea (25.4%) o quirúrgica (74.6%). Las técnicas
utilizadas para el control del foco se adjuntan en el Anexo 2 (Apéndice S2, Material
suplementario de la Publicación 2). Cuando comparamos los pacientes que precisaron
alguna técnica del control del foco frente a los que no (la decisión de controlar el foco era
llevada a cabo por el equipo tratante del paciente), encontramos diferencias significativas
tanto demográficas como clínicas y de tratamiento recibido. En primer lugar, los pacientes
que precisaron control del foco presentaban más edad, mayor proporción de shock y un
mayor número de fallos orgánicos en las primeras 6 horas desde el inicio de la sepsis.
Además, presentaban una mayor proporción de bacteriemia y mayores niveles en sangre de
proteína C reactiva (PCR) y procalcitonina (PCT) en las primeras 24h. El origen
nosocomial también fue más frecuente en los pacientes con control del foco. Las
diferencias más importantes se resumen en la Tabla 3 (para más información ver Tabla 1,
Publicación 2).
45
Tabla 3. Características clínicas y demográficas para el global de pacientes, el grupo de pacientes
que precisó control del foco y el grupo que no precisó control del foco
Global
n = 3663
NO Control foco
n = 2490 (68%)
Control foco
n = 1173 (32%) p
Edad (años), media (DE) 64 (15.1) 62.8 (15.2) 66.7 (14.6) < 0.001
Shock, n (%) 2497 (68.2) 1630 (65.5) 867 (73.9) < 0.001
Charlson, media (DE) 2.6 (2.3) 2.6 (2.3) 2.7 (2.2) 0.531
Nosocomial, n (%) 780 (21.3) 443 (17.8) 337 (28.7) < 0.001
PCR (mg/dl)a, media (DE) 24.2 (13.7) 23.6 (13.9) 25.5 (12.9) < 0.001
PCT (ng/dl)b, media (DE) 26.2 (37.6) 24.1 (34.7) 31.2 (43) 0.001
Bacteriemia, n (%) 1211 (40.1) 821 (37.9) 390 (45.5) < 0.001
ATB apropiado, n (%) 1911 (51.9) 1231 (49.4) 670 (57.1) < 0.001
Fallos orgánicos < 6 horas (%)
Hemodinámico 3019 (82.4) 1994 (80.1) 1025 (87.4) < 0.001
Renal 2068 (56.5) 1351 (54.3) 717 (61.1) < 0.001
Hepático 606 (16.5) 386 (15.5) 220 (18.8) 0.013
Coagulación 1143 (31.2) 749 (30.1) 394 (33.6) 0.032
Hiperlactatemia 1630 (44.5) 1056 (42.4) 574 (48.9) < 0.001
an = 2763 pacientes. bn = 2084 pacientes. DE, desviación estándar; PCR, proteína C reactiva; PCT, procalcitonina;
ATB, antibiótico.
El origen de la infección también difirió entre un grupo y otro, siendo el abdominal más
frecuente en los pacientes con control del foco (n = 788; 67.2%) y el respiratorio en los
pacientes que no requirieron control del foco (n = 1189; 47.8%) (Figura 7).
46
Figura 7. Origen de la infección (%) en los pacientes que precisaron control del foco comparado con
los pacientes que no precisaron control del foco. ITU, infección del tracto urinario.
Los pacientes que precisaron alguna técnica del control del foco recibieron
significativamente un peor tratamiento inicial (Figura 8): menor medición de lactato en
plasma, menor obtención de hemocultivos y administración más tardía del tratamiento
antibiótico empírico inicial.
Figura 8. Cumplimiento (%) de las bundle de resucitación de la SSC en los pacientes que precisaron
control del foco comparado con los pacientes que no precisaron control del foco. NA, noradrenalina; PVC, presión venosa central; SvcO2, saturación venosa central de oxígeno.
47
El grupo de pacientes que requirió control del foco precisó más días de tratamiento
vasopresor y presentó una estancia hospitalaria significativamente mayor (32.5 vs 27.4
días; p < 0.001). La mortalidad en UCI fue menor en los pacientes que requirieron control
del foco (21.2% frente a 25.1%; p = 0.010), siendo la mortalidad hospitalaria similar en los
dos grupos (Tabla 3, Publicación 2).
El análisis multivariado ajustado para factores de confusión mostró que los pacientes que
precisaron control del foco presentaron una mortalidad menor tanto de UCI (Tabla S1,
Material suplementario de la Publicación 1) como hospitalaria (OR 0.809 [IC 95%, 0.658–
0.994]; p = 0.044) (Figura 9).
0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2 2,2
Odds ratio
Figura 9. Análisis multivariado de factores de riesgo para mortalidad hospitalaria para el global de
pacientes (n = 3663). *Neumonía como referencia.
OR, odds ratio; IC, intérvalo de confianza; APACHE II, Acute Physiology and Chronic Health Evaluation II
score; NA, noradrenalina; ATB, antibiótico; ITU, infección del tracto urinario.
Control foco
Edad
APACHE-II
Charslon
ATB precoz
Fluidos/NA
ATB apropiado
Nosocomial
Abdominal*
ITU
Meningitis
Piel
Otros
OR; IC 95%
48
Tiempo del control del foco
El tiempo del control del foco desde el inicio de la sepsis se registró en 1090 pacientes de
los 1173 que precisaron alguna técnica del control del foco. En el 75.7% (n = 825) la
técnica se llevó a cabo en las primeras 12 horas desde el diagnóstico de la sepsis. Cuando
comparamos los pacientes con control precoz del foco (< 12 horas) frente a tardío (≥ 12
horas), no encontramos diferencias epidemiológicas significativas a excepción de una
mayor proporción de bacteriemia en el grupo tardío y una mayor proporción de origen
abdominal en el precoz (Tabla S2, Material suplementario de la Publicación 2). Los
pacientes con control del foco precoz recibieron una mejor reanimación hemodinámica
inicial (Tabla S3, Material suplementario de la Publicación 2). No se observaron
diferencias significativas entre el grupo precoz y el grupo tardío en días de estancia en
UCI/hospitalaria ni en mortalidad de UCI/hospitalaria (Tabla 4).
Tabla 4. Días de VM y vasopresores, días de estancia en UCI y hospital, mortalidad en UCI y
hospital en el grupo con control precoz del foco infeccioso frente al tardío.
Grupo
control del
foco
n = 1090
Control foco
< 12 h
n = 825
Control foco
≥ 12 h
n = 265
p
Días de VM, media (DE) 7.1 (13.1) 7.1 (12.9) 7.1 (13.9) 0.995
Días de vasopresores, media (DE) 4.8 (8.1) 4.6 (7.5) 5.4 (9.7) 0.168
Días de estancia en UCI, media (DE) 12.2 (15.3) 12.1 (15.2) 12.6 (15.4) 0.518
Días de estancia
hospitalaria, media (DE) 32.3 (31.3) 31.9 (29.7) 31.6 (28.5) 0.884
Mortalidad, n (%)
UCI 226 (20.7) 172 (20.8) 54 (20.4) 0.869
Hospital
299 (27.4) 228 (27.6) 71 (26.8) 0.789
VM, ventilación mecánica; UCI, unidad de cuidados intensivos; h, horas; DE, desviación estándar.
49
El análisis mediante curvas ROC para evaluar el tiempo del control del foco como
predictor de mortalidad mostró un AUC de 0.504 (p = ns) (Figuras 1, 2 y 3, Material
suplementario de la Publicación 2). Adicionalmente, se estudiaron otros puntos de corte de
tiempo del control del foco: < 4 horas frente a ≥ 4 horas, y otro modelo donde comparamos
< 4 horas, 4-8 horas, 8-12 horas, 12-24 horas y ≥ 24 horas. En ninguno de ellos
encontramos asociación entre el tiempo del control del foco y mortalidad (Tablas 1 y 2 del
Anexo 3, datos no publicados). El análisis de regresión múltiple, no mostró tampoco
asociación entre el control del foco tardío y mortalidad (Tabla S8, Material suplementario
de la Publicación 2).
Dado que la sepsis de origen abdominal, urinaria y piel y tejidos blandos precisa con
mayor frecuencia una técnica del control del foco, realizamos en estos subgrupos un
análisis univariado y multivariado ajustado por factores de confusión no encontrando
diferencias significativas en la mortalidad entre el control del foco precoz y el tardío
(Tablas S4 a S6 y Tablas S9 a S11, Material suplementario de la Publicación 2).
En el análisis de sensibilidad, donde excluimos los pacientes con sepsis de origen urinario
del global de pacientes, observamos que precisar alguna técnica del control del foco
mantiene asociación con una menor mortalidad en UCI (Tabla 4, Anexo 3, datos no
publicados) la cual persiste, tanto en UCI como hospitalaria, tras ajustar por múltiples
factores de confusión (Tabla 5 Anexo 3, datos no publicados). El análisis de regresión
logística multivariable en pacientes con control del foco pero excluyendo pacientes con
infecciones urinarias no mostró relación entre el control precoz del foco y menor
mortalidad (Tabla 6, Anexo 3, datos no publicados).
50
51
7.- DISCUSIÓN GENERAL:
Los resultados de la presente tesis doctoral apoyan la importancia del adecuado manejo
infeccioso en los pacientes con sepsis y shock séptico. La administración del tratamiento
antimicrobiano de manera precoz y apropiada y la búsqueda del origen de la infección y
posterior drenaje del foco si lo hay, son de vital importancia en estos pacientes.
Intervención múltiple dirigida a mejorar la antibioticoterapia empírica en la sepsis
Los resultados principales de nuestro trabajo de tesis doctoral demuestran que una
intervención educativa a gran escala mejora el uso general de antibióticos en la sepsis
mejorando la eficacia, al reducir el tiempo desde el inicio de la sepsis hasta la
administración del primer antibiótico y aumentando la proporción de pacientes que reciben
un tratamiento empírico apropiado y, mejorando la seguridad, al aumentar la proporción de
pacientes que reciben un adecuado desescalamiento antibiótico.
Estos resultados refuerzan la creciente literatura que muestra que las intervenciones
educativas pueden mejorar la calidad asistencial en diferentes contextos y condiciones
(32,43–46). Dado que las intervenciones múltiples parecen ser más efectivas que los
enfoques más limitados para cambiar un determinado comportamiento (47,48), incluimos
en nuestra intervención nuevas estrategias educativas con el objetivo de fortalecer y
reforzar el conocimiento. Estas estrategias consistían en recordatorios semanales mediante
mensaje de texto al teléfono móvil o vía mail al personal responsable del paciente séptico y
la posibilidad de participar en un juego educativo de simulación de manejo de pacientes
sépticos, ambas estrategias respaldadas por experiencias más limitadas (49–52).
Las guías internacionales de la SSC para el manejo de los pacientes con sepsis y shock
séptico recomiendan administrar el tratamiento antibiótico lo antes posible y dentro de la
primera hora desde el reconocimiento de la sepsis y, además, que éste sea adecuado y por
lo tanto de amplio espectro. Esta recomendación, incluso, se está introduciendo en
determinados lugares como medida obligatoria o de calidad (22,23,53). Ya en los años 80
se empezó a reconocer la importancia del antibiótico precoz en el tratamiento de la
meningitis y desde entonces son muchos los estudios que apoyan esta recomendación, no
52
sólo asociando el retraso antibiótico a un peor pronóstico en términos de mortalidad sino
también asociando este retraso con otros efectos adversos como son más días de estancia
hospitalaria o un mayor número de fallos orgánicos (54–57). El primer estudio en pacientes
con shock séptico que muestra una asociación lineal entre el retraso en el tratamiento
antibiótico y la mortalidad es un trabajo realizado por Kumar et al. (58) en 2731 pacientes
donde encuentran que la mortalidad hospitalaria disminuye un 7% por cada hora de retraso
en el inicio del tratamiento antibiótico desde el inicio de hipotensión. Los estudios más
recientes, realizados en un gran número de pacientes con sepsis o shock séptico reafirman
este concepto. Por ejemplo, Ferrer et al. (32) realizaron un análisis retrospectivo de los
datos de la SSC que incluyó a 17990 pacientes con sepsis y shock séptico de 165 UCIs
entre 2005 y 2010 confirmando que la administración tardía de antibióticos se asocia con
un aumento de la mortalidad hospitalaria. Más recientemente, en una gran muestra
multicéntrica de pacientes con sepsis, Liu et al. (59) encuentran una asociación lineal entre
el retraso en la administración del antibiótico y la mortalidad. En el análisis multivariante,
por cada hora de retraso la probabilidad de muerte aumentó desde la primera hora siendo el
impacto más marcado en los pacientes con shock (OR 1.14; IC 95% 1.06-1.23; aumento en
la mortalidad absoluta 1.8%; IC 95% 0.8-3%; P = 0.001).
En nuestro trabajo, a pesar de la mejora significativa en la reducción del tiempo de inicio
del tratamiento antibiótico, no obtuvimos una disminución significativa de la mortalidad.
El grupo MEDUSA recientemente ha publicado los datos de un estudio multicéntrico,
educacional y randomizado que incluyó a 4183 pacientes con sepsis y shock séptico (60).
Aunque el riesgo de muerte aumentó en un 2% por hora de retraso en el inicio del
tratamiento antimicrobiano, la intervención no pudo reducir ni el tiempo medio del
tratamiento antimicrobiano (1.5 vs. 2.0 h, p = 0.41) ni la mortalidad. Un reciente
metaanálisis, que incluyó a más de 16000 pacientes con sepsis y shock séptico, no encontró
un beneficio significativo de administrar el tratamiento antibiótico de manera precoz (61).
Una posible explicación de estos resultados contradictorios podría estar en relación con la
incertidumbre diagnóstica con respecto a la sepsis y el posible daño asociado con
antibióticos innecesarios. Reconocer la sepsis no es una tarea sencilla, y a pesar de que con
la última actualización de las definiciones (SEPSIS-3) se intenta detectar a los pacientes
más graves y de manera más uniforme (1), su definición es un concepto teórico y no existe
una prueba diagnóstica o un biomarcador específico para detectarla. La heterogeneidad
53
biológica y clínica en estos pacientes (edad, comorbilidades, otros factores como el
requerir una cirugía urgente, el tratamiento previo o el origen de infección) le añaden una
mayor complejidad. Esto puede llevar a infra e incluso supra-diagnosticar la sepsis (62).
Además, el proceso de atención inicial a la sepsis requiere que en un período muy corto de
tiempo se lleven a cabo varias medidas terapéuticas. Cumplir con la recomendación de las
guías puede aportar ventajas, pero también conlleva una serie de riesgos: tratamiento
antibiótico innecesario de pacientes sin sepsis, no toma de muestras microbiológicas,
efectos secundarios del tratamiento antibiótico y aumento de resistencias en el hospital y
en la comunidad por el uso excesivo de antibióticos de amplio espectro (63). Existe un
preocupante aumento de resistencia a los antimicrobianos considerándose ya una crisis
global (64). Además, los antibióticos en sí también causan daño, por ejemplo, lesión
orgánica, disfunción mitocondrial, impacto en el microbioma y el crecimiento excesivo de
hongos y Clostridium difficile (64–67). Seguir las recomendaciones de la SSC implica, por
tanto, adoptar una serie de medidas posteriormente. En primer lugar, esta medida
obligatoria en relación a la antibioticoterapia precoz, debe ir asociada a una evaluación
posterior. Es necesario el desescalamiento antibiótico en función de los resultados
microbiológicos e incluso, aun siendo éstos negativos o no disponer de ellos, habría que
valorarlo si la evolución del paciente es favorable e incluso retirarlos si se descarta
infección bacteriana (68,69). La constante innovación en las técnicas microbiológicas de
diagnóstico rápido debería también contribuir a identificar precozmente tratamientos
inadecuados. La evidencia científica apoya la importancia de la administración de
antibióticos de manera “agresiva” e inmediata en los pacientes más graves, con rápido
deterioro y en los que presentan shock (20,59,64,70). Sin embargo, si existe incertidumbre
diagnóstica, los clínicos deben tratar de medir la gravedad de la enfermedad y la
probabilidad de infección (62).
Nuestro trabajo de tesis doctoral no sólo consigue acortar el tiempo de antibiótico, sino
que, tras la educación, el tratamiento apropiado y el desescalamiento aumentan. Los
esfuerzos para mejorar el tratamiento de la sepsis, especialmente aquellos dirigidos a la
administración empírica de antibióticos, deben abarcar todos los niveles de atención. La
clave para mejorar los resultados en la sepsis es motivar a los profesionales a implementar
medidas basadas en evidencia y proporcionarles retroalimentación sobre el impacto de
estas medidas (71). Uno de los mayores beneficios de intervenciones como la nuestra es su
54
contribución a la formación de una cultura que fomenta el deseo de mejorar y un
compromiso continuo con la excelencia en la atención al paciente (72).
Impacto del control del foco quirúrgico o percutáneo en la sepsis
Los datos de nuestro segundo trabajo de tesis doctoral demuestran la importancia del
control del foco en pacientes con sepsis y shock séptico. Este estudio muestra, en más de
3000 pacientes con sepsis y shock séptico, que aquellos que precisan alguna técnica del
control del foco, a pesar de ser un grupo de enfermos más graves y con un peor manejo
inicial de acuerdo a las recomendaciones de la SSC, presentan una menor mortalidad.
En nuestro estudio, un tercio de los pacientes que ingresan en UCI con diagnóstico de
sepsis o shock séptico requiere alguna técnica del control del foco. Hasta la fecha, sólo hay
otro trabajo publicado que evalúe la necesidad del control del foco en el global de los
pacientes con sepsis. Se trata de un estudio multicéntrico realizado en Alemania durante el
mismo período de tiempo que nuestro estudio en el cual el 41.7% de los 1011 pacientes
con sepsis o shock séptico incluidos requirió alguna técnica del control del foco (73).
Nuestro trabajo aporta información interesante, no estudiada previamente, en cuanto a las
diferencias que presentan los pacientes sépticos que precisan o no control del foco. Los
pacientes que requieren control del foco presentan una mayor proporción de shock, mayor
disfunción multiorgánica (mayor proporción de disfunción renal, cardiovascular, hepática,
coagulopatía e hiperlactatemia), y mayores tasas de bacteriemia y de niveles de PCR y
PCT en las primeras 24 horas. Tanto la PCR como la PCT son biomarcadores que se
correlacionan con la respuesta inflamatoria e incluso con la gravedad de la sepsis y la
disfunción orgánica (74–77). Esta mayor respuesta inflamatoria en los pacientes sometidos
a control del foco, asociada a una mayor disfunción orgánica, se podría explicar por la
propia manipulación de un foco con una alta carga microbiana (42,78). Además, como
podemos observar en los resultados, se trata de dos poblaciones claramente diferentes en
cuanto el origen de la infección: los pacientes que precisan control del foco generalmente
tienen infecciones intra-abdominales, infecciones necrotizantes de tejidos blandos,
nefropatía obstructiva o infecciones asociadas a catéteres externos; y los pacientes que no
precisan control del foco generalmente tienen neumonía o infecciones del sistema nervioso
55
central (SNC). El foco anatómico de infección se asocia de manera independiente con la
mortalidad tal y como muestra el estudio realizado por Leligdowicz et al. (42) en más de
7000 pacientes con sepsis y shock séptico, donde la isquemia intestinal o la infección
diseminada son los focos con mayor mortalidad, focos con alta carga bacteriana. Pero no
solo eso, pueden existir diferencias en la naturaleza de los patógenos infectantes en cada
sitio anatómico así como también en las condiciones del tejido local y la capacidad de
reparación o la tolerancia funcional a la intervención. Es posible que las diferencias en los
resultados estén en parte relacionadas con estas características además de con el propio
procedimiento, percutáneo o quirúrgico, del control del foco.
Cuando comparamos ambos grupos, objetivamos también que los pacientes que precisaron
control del foco presentaban más edad y mayor proporción de infecciones de origen
nosocomial, factores asociados de forma independiente con la mortalidad (79).
Otra diferencia importante que encontramos entre estos dos grupos de pacientes es en
relación al manejo inicial que recomiendan las guías de la SSC: aunque los pacientes que
requieren control del foco reciben una antibioterapia más apropiada (el poder obtener una
muestra dirigida del tejido infectado permite una determinación mejor del
microorganismo), el tiempo hasta la administración del primer antibiótico fue más
prolongado en este grupo. Además, el cumplimiento de otras medidas de la bundle de
resucitación fue peor en los pacientes que requirieron control del foco. Una posible
explicación para este peor manejo inicial puede ser debida a que en estos pacientes se
prioriza el llevar a cabo el control del foco frente al resto de medidas recomendadas por la
SSC.
Por lo tanto, aunque los pacientes que requieren control del foco presentan un mayor riesgo
de muerte que los que no, su mortalidad es menor incluso después del ajuste por factores
de confusión. Estos hallazgos apoyan firmemente la importancia del control del foco y por
tanto eliminar la carga microbiana, con el fin de eliminar/controlar el trigger infeccioso que
haría que la disfunción orgánica o el shock progresen, tal y como propone Kumar en una
revisión reciente sobre la fisiopatología de la sepsis (14) y que se expone en la
introducción de la presente tesis doctoral. La falta de un foco drenable, es decir, el hecho
de no poder eliminar físicamente el foco infeccioso, parece estar asociado con un peor
resultado. Un estudio retrospectivo que revisó los hallazgos macroscópicos en las autopsias
de 235 pacientes quirúrgicos que murieron por sepsis o shock séptico en UCI, encontró en
56
aproximadamente el 80% de ellos la persistencia de un foco séptico no drenado o mal
drenado, lo que sugiere que la necesidad del control del foco puede estar poco reconocida
(80).
Tiempo del control del foco
Aunque el control del foco es esencial para el manejo exitoso de los pacientes con sepsis y
shock séptico, el momento ideal para llevarlo a cabo no está bien definido. Nuestro trabajo
no ha podido demostrar relación entre el control del foco tardío y mayor mortalidad. La
última actualización de las guías de la SSC recomienda que el control del foco se realice lo
antes posible desde que se presenta la sepsis (16). Esta recomendación ha ido cambiando
desde las guías 2008 donde se debía realizar el control del foco en las primeras 6 horas
desde el diagnóstico, a las guías del 2012 donde el margen era de hasta 12 horas. Las guías
principalmente se basan en estudios observacionales pequeños como por ejemplo un
estudio retrospectivo realizado en 106 pacientes con shock séptico e infecciones
necrotizantes de tejidos blandos en los que realizar la cirugía pasadas 14 horas se asoció de
forma independiente con la mortalidad hospitalaria (81). Otros estudios realizados también
en pacientes con infección de partes blandas muestran la importancia del control precoz
pero con tiempos que varían de las 2 a las 24 horas (82–84). Existen otros trabajos llevados
a cabo en sepsis de origen abdominal que muestran la relación entre el tiempo del control
del foco y peor evolución (85). Por ejemplo, en un estudio retrospectivo publicado
recientemente y realizado en 76 pacientes con sepsis secundaria a perforación intestinal,
cada hora de retraso de la cirugía se asoció a menor supervivencia (86). Pero no todos los
trabajos muestran estos resultados positivos; un estudio realizado en una gran población de
pacientes con peritonitis fecaloidea no encontró relación entre el tiempo de control del foco
y mortalidad (87). Solo hay otro estudio que evalúe el tiempo del control del foco en el
global de pacientes con sepsis y shock séptico (60): este reciente trabajo multicéntrico
concluye que los pacientes con control del foco quirúrgico posterior a las 6 horas desde el
inicio de la sepsis tuvieron una mayor mortalidad a los 28 días (35.6% vs. 27.9%, p <
0.001).
En nuestra población, los pacientes que recibieron control precoz del foco también
recibieron una mejor resucitación hemodinámica, lo que sugiere que estos pacientes
podrían haber estado más graves; sin embargo, no encontramos otras diferencias
57
significativas entre ambos grupos. Se podría argumentar que los pacientes que tuvieron
intervenciones posteriores posiblemente tengan una tasa de mortalidad más baja que la
intervención temprana si la práctica clínica es proceder de inmediato al control del foco
solo para el paciente más crítico. Parece que hay al menos tres razones potenciales por las
cuales los médicos retrasarían el control del foco necesario en pacientes con sepsis:
(1) que el foco de la infección no fuera clínicamente evidente en la evaluación inicial, y
que se hiciera más evidente con el tiempo durante un período en que el paciente se
encontraba en una condición más estable;
(2) otra posibilidad es que los médicos optaran electivamente por demorar la intervención
para estabilizar la hemodinámica y el estado fisiológico del paciente y realizar
posteriormente la intervención;
(3) finalmente, es posible que la intervención quirúrgica se retrasara específicamente para
permitir que el área de necrosis o la formación de abscesos se localice y se defina para así
conseguir una intervención quirúrgica más adecuada (como el caso de pancreatitis
necrotizante).
Por varias razones, es posible que los pacientes más enfermos se trataran con una
intervención temprana, mientras que los pacientes menos enfermos se trataran con una
intervención posterior. Esto sesgaría el estudio a favor de una mayor mortalidad en
pacientes con intervención temprana versus intervención tardía. Hasta que se realice un
estudio en el que exista una aleatorización formal de los pacientes para el control del foco
precoz versus tardío esto seguirá siendo una pregunta sin respuesta.
Nuestro trabajo apoya la importancia de controlar o drenar el foco infeccioso cuando está
presente. Cuándo llevarlo a cabo sigue siendo una pregunta difícil de responder. Cada
paciente presenta su propio conjunto único de diferentes circunstancias, comorbilidades,
estado funcional inmune, etc., que podrían favorecer un enfoque sobre otro. Es probable
que exista una gradación de la relación riesgo-beneficio de la intervención precoz a la
tardía. En estudios con animales, es posible demostrar un umbral de concentración de
carga bacteriana letal a pesar de los intentos terapéuticos de salvar la vida a estos animales
(14,58). Una situación similar probablemente existe en los humanos, pero la patogenicidad
del organismo infectante, las capacidades inmunitarias del huésped y las reservas
58
fisiológicas de cada paciente séptico son muy variables y, en última instancia,
determinarán el resultado (88).
Los resultados de nuestro segundo trabajo de tesis doctoral, no justifican que los clínicos
adopten un abordaje indiferente en esta población de pacientes. Más bien, nuestro estudio
apoya la idea de que la medicina y la cirugía personalizadas son el enfoque óptimo para
decidir en qué momento es mejor realizar el control del foco. Algunos pacientes
necesitarán una intervención precoz, mientras que en otros probablemente se pueda
posponer hasta que su estado clínico haya mejorado. Es difícil concebir una situación en la
que se pueda diseñar un ensayo clínico prospectivo, aleatorizado para determinar
rigurosamente el momento óptimo para realizar el control del foco en pacientes sépticos.
La evidencia actual indica que es obligatorio un examen cuidadoso del foco de infección
que pueda ser drenado en todos los pacientes sépticos. El momento de las intervenciones
de control del foco, en caso de que surja la necesidad, es una decisión desafiante, difícil,
individual, que debe tomarse en base a la totalidad de la evidencia para cada paciente.
Limitaciones
El diseño antes y después de nuestro estudio tiene debilidades inherentes al propio diseño
como la influencia de las tendencias seculares que puede ser difícil de separar de los
efectos de la intervención. No utilizar un grupo control hace que sea imposible asegurar
que los cambios observados sean debidos a la intervención. Sin embargo, durante los tres
períodos del estudio no se realizaron cambios importantes en el protocolo. Además, en
nuestro estudio realizamos un análisis de regresión segmentado, un método potente para
estimar los efectos de una intervención en series de tiempo interrumpido (89). Otra
limitación es que evaluamos el tratamiento antimicrobiano como apropiado sólo en base a
los resultados microbiológicos sin tener en cuenta la farmacocinética ni la
farmacodinámica tan importante en los pacientes críticos. En tercer lugar, el seguimiento a
largo plazo de nuestro estudio, realizado 6 meses después del período post-intervención,
podría considerase demasiado cercano para evaluar el impacto tardío de la intervención.
Sin embargo, en otro estudio publicado por nosotros (90) donde comparamos los pacientes
del grupo pre-intervención del estudio Edusepsis del 2005 (27) con los pacientes del grupo
59
pre-intervención del estudio ABISS-Edusepsis del 2011 identificamos un mejor manejo en
el grupo del 2011 asociado también a una menor mortalidad.
En cuanto al objetivo secundario de nuestro trabajo, una limitación a considerar es que
analizamos solo a los pacientes que requirieron ingreso en UCI, posiblemente
introduciendo un sesgo de selección donde los pacientes que recibieron alguna técnica de
control del foco muy temprana pudieron mejorar lo suficiente como para evitar el ingreso
en UCI. Otros factores importantes que pueden influir en la evolución de los pacientes con
sepsis son el tipo de microorganismo y el tipo de técnica del control del foco que se aplica,
algo que no hemos evaluado en nuestro estudio. Otra limitación a tener en cuenta es que no
evaluamos la adecuación o “éxito” de la técnica del control del foco y la necesidad de
nuevas intervenciones, factores que pueden influir de manera significativa en los resultados
(73,82,91).
60
61
8.- CONCLUSIONES GENERALES:
Las conclusiones principales obtenidas de la presente tesis doctoral son:
1. Un programa educacional realizado a nivel nacional enfocado en la importancia del
manejo infeccioso de los pacientes con sepsis se asocia a un mejor uso global del
tratamiento antimicrobiano mejorando:
la eficacia: al reducir el tiempo de administración del tratamiento antibiótico
y aumentar la proporción de pacientes que reciben un tratamiento empírico
adecuado y
la seguridad: al aumentar la proporción de pacientes que reciben un
desescalamiento antibiótico posterior.
2. No hemos podido demostrar una mejoría significativa en la mortalidad evidenciando que
aún queda un largo camino de mejora en el manejo de los pacientes con sepsis y shock
séptico.
3. Una tercera parte de los episodios de sepsis precisan alguna medida percutánea o
quirúrgica de control del foco, especialmente el origen abdominal y piel y tejidos blandos.
4. Los pacientes que presentan un foco drenable, a pesar de que presentan mayor gravedad
y reciben un peor cumplimiento de las medidas iniciales de tratamiento que recomiendan
las guías de la SSC, tienen un mejor pronóstico inicial.
5. No hemos podido demostrar en nuestra serie que el control precoz del foco reduzca la
morbimortalidad.
62
63
9.- LÍNEAS DE FUTURO:
Las futuras nuevas vías de trabajo que se podrían llevar a cabo son:
1) analizar el impacto biológico del tratamiento antibiótico adecuado y precoz y del control
del foco infeccioso;
2) determinar biomarcadores para el diagnóstico precoz y la respuesta terapéutica del
control de la infección;
3) analizar el impacto del diagnóstico microbiológico rápido en los pacientes con sepsis;
4) analizar in vitro e in vivo el efecto del tratamiento combinado de antibióticos y terapia
celular con células mesenquimales u otros tratamientos inmunomoduladores.
64
65
10.- BIBLIOGRAFÍA MÁS RELEVANTE
1. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et
al. The Third International Consensus Definitions for Sepsis and Septic Shock
(Sepsis-3). JAMA. 2016 Feb 23;315(8):801.
2. Hall MJ, Levant S, DeFrances CJ. Trends in inpatient hospital deaths: National
Hospital Discharge Survey, 2000-2010. NCHS Data Brief. 2013 Mar;(118):1–8.
3. Fleischmann C, Scherag A, Adhikari NKJ, Hartog CS, Tsaganos T, Schlattmann P,
et al. Assessment of Global Incidence and Mortality of Hospital-treated Sepsis.
Current Estimates and Limitations. Am J Respir Crit Care Med. 2016 Feb
1;193(3):259–72.
4. Prescott HC, Angus DC. Enhancing Recovery From Sepsis. JAMA. 2018 Jan
2;319(1):62.
5. Shankar-Hari M, Phillips GS, Levy ML, Seymour CW, Liu VX, Deutschman CS, et
al. Developing a New Definition and Assessing New Clinical Criteria for Septic
Shock. JAMA. 2016 Feb 23;315(8):775.
6. Yébenes JC, Ruiz-Rodriguez JC, Ferrer R, Clèries M, Bosch A, Lorencio C, et al.
Epidemiology of sepsis in Catalonia: analysis of incidence and outcomes in a
European setting. Ann Intensive Care. 2017 Dec 20;7(1):19.
7. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR.
Epidemiology of severe sepsis in the United States: Analysis of incidence, outcome,
and associated costs of care. Crit Care Med. 2001 Jul;29(7):1303–10.
8. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R, et al.
Surviving Sepsis Campaign: International guidelines for management of severe
sepsis and septic shock: 2008. Crit Care Med. 2008 Jan;36(1):296–327.
9. Stevenson EK, Rubenstein AR, Radin GT, Wiener RS, Walkey AJ. Two Decades of
Mortality Trends Among Patients With Severe Sepsis. Crit Care Med. 2014
Mar;42(3):625–31.
66
10. Kaukonen K-M, Bailey M, Suzuki S, Pilcher D, Bellomo R. Mortality Related to
Severe Sepsis and Septic Shock Among Critically Ill Patients in Australia and New
Zealand, 2000-2012. JAMA. 2014 Apr 2;311(13):1308.
11. Bouza C, López-Cuadrado T, Saz-Parkinson Z, Amate-Blanco JM. Epidemiology
and recent trends of severe sepsis in Spain: a nationwide population-based analysis
(2006-2011). BMC Infect Dis. 2014 Dec 21;14(1):3863.
12. Hotchkiss RS, Monneret G, Payen D. Sepsis-induced immunosuppression: from
cellular dysfunctions to immunotherapy. Nat Rev Immunol. 2013 Dec
15;13(12):862–74.
13. Wiersinga WJ, Leopold SJ, Cranendonk DR, van der Poll T. Host innate immune
responses to sepsis. Virulence. 2014 Jan 17;5(1):36–44.
14. Kumar A. An alternate pathophysiologic paradigm of sepsis and septic shock.
Virulence. 2014 Jan;5(1):80–97.
15. Dellinger RP, Carlet JM, Masur H, Gerlach H, Calandra T, Cohen J, et al. Surviving
Sepsis Campaign guidelines for management of severe sepsis and septic shock. Crit
Care Med. 2004 Mar;32(3):858–73.
16. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al.
Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and
Septic Shock: 2016. Intensive Care Med. 2017 Mar 18;43(3):304–77.
17. Levy MM, Rhodes A, Phillips GS, Townsend SR, Schorr CA, Beale R, et al.
Surviving Sepsis Campaign: association between performance metrics and outcomes
in a 7.5-year study. Crit Care Med. 2015 Jan;43(1):3–12.
18. Barochia A V., Cui X, Vitberg D, Suffredini AF, OʼGrady NP, Banks SM, et al.
Bundled care for septic shock: An analysis of clinical trials. Crit Care Med. 2010
Feb;38(2):668–78.
19. Levy MM, Dellinger RP, Townsend SR, Linde-Zwirble WT, Marshall JC, Bion J, et
al. The Surviving Sepsis Campaign: Results of an international guideline-based
performance improvement program targeting severe sepsis. Crit Care Med. 2010
Feb;38(2):367–74.
67
20. Seymour CW, Gesten F, Prescott HC, Friedrich ME, Iwashyna TJ, Phillips GS, et al.
Time to Treatment and Mortality during Mandated Emergency Care for Sepsis. N
Engl J Med. 2017 Jun 8;376(23):2235–44.
21. Levy MM, Evans LE, Rhodes A. The Surviving Sepsis Campaign Bundle: 2018
update. Intensive Care Med. 2018 Jun 19;44(6):925–8.
22. https://www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-sheets/2014-Fact-
sheets-items/2014-08-04-2.html. Accessed on 28 December 2017).
23. http://www.nytimes.com/2012/12/21/nyregion/one-boys-death-moves-state-to-
action-to-prevent-others.html. Accessed on 27 December 2017).
24. Levy MM, Gesten FC, Phillips GS, Terry KM, Seymour CW, Prescott HC, et al.
Mortality Changes Associated with Mandated Public Reporting for Sepsis: The
Results of the New York State Initiative. Am J Respir Crit Care Med. 2018 Sep
7;2776:rccm.201712-2545OC.
25. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al.
Definitions for Sepsis and Organ Failure and Guidelines for the Use of Innovative
Therapies in Sepsis. Chest. 1992 Jun;101(6):1644–55.
26. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, et al. 2001
SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference.
Intensive Care Med. 2003 Apr 28;29(4):530–8.
27. Ferrer R. Improvement in Process of Care and Outcome After a Multicenter Severe
Sepsis Educational Program in Spain. JAMA. 2008 May 21;299(19):2294.
28. Ferrer R, Artigas A, Suarez D, Palencia E, Levy MM, Arenzana A, et al.
Effectiveness of Treatments for Severe Sepsis. Am J Respir Crit Care Med. 2009
Nov;180(9):861–6.
29. Ehrlich P. Chemotherapeutics: Scientific principles, methods, and results. Address
in pathology to 17th International Congress of Medicine. Lancet. 1913;2:445–51.
30. Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, et al. Duration of
hypotension before initiation of effective antimicrobial therapy is the critical
68
determinant of survival in human septic shock. Crit Care Med. 2006
Jun;34(6):1589–96.
31. Kumar A, Ellis P, Arabi Y, Roberts D, Light B, Parrillo JE, et al. Initiation of
Inappropriate Antimicrobial Therapy Results in a Fivefold Reduction of Survival in
Human Septic Shock. Chest. 2009 Nov;136(5):1237–48.
32. Ferrer R, Martin-Loeches I, Phillips G, Osborn TM, Townsend S, Dellinger RP, et
al. Empiric Antibiotic Treatment Reduces Mortality in Severe Sepsis and Septic
Shock From the First Hour. Crit Care Med. 2014 Aug;42(8):1749–55.
33. Suberviola Cañas B, Jáuregui R, Ballesteros MÁ, Leizaola O, González-Castro A,
Castellanos-Ortega Á. Effects of antibiotic administration delay and inadequacy
upon the survival of septic shock patients. Med Intensiva (English Ed. 2015
Nov;39(8):459–66.
34. Ulldemolins M, Nuvials X, Palomar M, Masclans JR, Rello J. Appropriateness is
Critical. Crit Care Clin. 2011 Jan;27(1):35–51.
35. Vallés J, Rello J, Ochagavía A, Garnacho J, Alcalá MA. Community-acquired
bloodstream infection in critically ill adult patients: impact of shock and
inappropriate antibiotic therapy on survival. Chest. 2003 May;123(5):1615–24.
36. Garnacho-Montero J, Gutiérrez-Pizarraya A, Escoresca-Ortega A, Corcia-Palomo
Y, Fernández-Delgado E, Herrera-Melero I, et al. De-escalation of empirical therapy
is associated with lower mortality in patients with severe sepsis and septic shock.
Intensive Care Med. 2014 Jan 12;40(1):32–40.
37. Masterton RG. Antibiotic De-Escalation. Crit Care Clin. 2011 Jan;27(1):149–62.
38. Bochud P-Y, Bonten M, Marchetti O, Calandra T. Antimicrobial therapy for
patients with severe sepsis and septic shock: An evidence-based review. Crit Care
Med. 2004 Nov;32(Supplement):S495–512.
39. Dryden M, Johnson AP, Ashiru-Oredope D, Sharland M. Using antibiotics
responsibly: right drug, right time, right dose, right duration. J Antimicrob
Chemother. 2011 Nov 1;66(11):2441–3.
69
40. Jimenez MF, Marshall JC, International Sepsis Forum. Source control in the
management of sepsis. Intensive Care Med. 2001;27 Suppl 1:S49-62.
41. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al.
Surviving Sepsis Campaign 2012. Crit Care Med. 2013 Feb;41(2):580–637.
42. Leligdowicz A, Dodek PM, Norena M, Wong H, Kumar A, Kumar A. Association
between Source of Infection and Hospital Mortality in Patients Who Have Septic
Shock. Am J Respir Crit Care Med. 2014 May 15;189(10):1204–13.
43. Scales DC. A Multifaceted Intervention for Quality Improvement in a Network of
Intensive Care Units. JAMA. 2011 Jan 26;305(4):363.
44. Groth ME, Chan-Tompkins NH, Hegde GG, Shang J, Venkat A. Delays in the
administration of antimicrobials in the emergency department and the impact of an
educational intervention to improve this administration. Eur J Emerg Med. 2013
Oct;20(5):339–43.
45. Capelastegui A, España PP, Quintana JM, Gorordo I, Ortega M, Idoiaga I, et al.
Improvement of Process‐of‐Care and Outcomes after Implementing a Guideline for
the Management of Community‐Acquired Pneumonia: A Controlled
Before‐and‐After Design Study. Clin Infect Dis. 2004 Oct;39(7):955–63.
46. Herrán-Monge R, Muriel-Bombín A, García-García MM, Merino-García PA,
Cítores-González R, Fernández-Ratero JA, et al. Mortality Reduction and Long-
Term Compliance with Surviving Sepsis Campaign. SHOCK. 2016 Jun;45(6):598–
606.
47. Grimshaw JM, Shirran L, Thomas R, Mowatt G, Fraser C, Bero L, et al. Changing
provider behavior: an overview of systematic reviews of interventions. Med Care.
2001 Aug;39(8 Suppl 2):II2-45.
48. Needham DM. Patient safety, quality of care, and knowledge translation in the
intensive care unit. Respir Care. 2010 Jul;55(7):922–8.
49. Evans KH, Daines W, Tsui J, Strehlow M, Maggio P, Shieh L. Septris. Acad Med.
2015 Feb;90(2):180–4.
70
50. Akl EA, Sackett KM, Pretorius R, Bhoopathi PSS, Mustafa R, Schünemann H, et al.
Educational games for health professionals. In: Akl EA, editor. Cochrane Database
of Systematic Reviews. Chichester, UK: John Wiley & Sons, Ltd; 2008. p.
CD006411.
51. Weiner SG, Brown SF, Goetz JD, Webber CA. Weekly E-mail Reminders Influence
Emergency Physician Behavior: A Case Study Using the Joint Commission and
Centers for Medicare and Medicaid Services Pneumonia Guidelines. Acad Emerg
Med. 2009 Jul;16(7):626–31.
52. Sinuff T, Cook D, Giacomini M, Heyland D, Dodek P. Facilitating clinician
adherence to guidelines in the intensive care unit: A multicenter, qualitative study.
Crit Care Med. 2007 Sep;35(9):2083–9.
53. https://www.nice.org.uk/guidance/indevelopment/gid-qs10032. Last accessed 25th
March 2017.
54. Funk DJ, Kumar A. Antimicrobial Therapy for Life-threatening Infections: Speed is
Life. Crit Care Clin. 2011 Jan;27(1):53–76.
55. Zhang D, Micek ST, Kollef MH. Time to Appropriate Antibiotic Therapy Is an
Independent Determinant of Postinfection ICU and Hospital Lengths of Stay in
Patients With Sepsis. Crit Care Med. 2015 Oct;43(10):2133–40.
56. Garnacho-Montero J, Aldabo-Pallas T, Garnacho-Montero C, Cayuela A, Jiménez
R, Barroso S, et al. Timing of adequate antibiotic therapy is a greater determinant of
outcome than are TNF and IL-10 polymorphisms in patients with sepsis. Crit Care.
2006 May 9;10(4):R111.
57. Bagshaw SM, Lapinsky S, Dial S, Arabi Y, Dodek P, Wood G, et al. Acute kidney
injury in septic shock: clinical outcomes and impact of duration of hypotension prior
to initiation of antimicrobial therapy. Intensive Care Med. 2009 May 9;35(5):871–
81.
58. Kumar A, Haery C, Paladugu B, Kumar A, Symeoneides S, Taiberg L, et al. The
duration of hypotension before the initiation of antibiotic treatment is a critical
determinant of survival in a murine model of Escherichia coli septic shock:
71
association with serum lactate and inflammatory cytokine levels. J Infect Dis. 2006
Jan 15;193(2):251–8.
59. Liu VX, Fielding-Singh V, Greene JD, Baker JM, Iwashyna TJ, Bhattacharya J, et
al. The Timing of Early Antibiotics and Hospital Mortality in Sepsis. Am J Respir
Crit Care Med. 2017 Oct;196(7):856–63.
60. Bloos F, Rüddel H, Thomas-Rüddel D, Schwarzkopf D, Pausch C, Harbarth S, et al.
Effect of a multifaceted educational intervention for anti-infectious measures on
sepsis mortality: a cluster randomized trial. Intensive Care Med. 2017 Nov
2;43(11):1602–12.
61. Sterling SA, Miller WR, Pryor J, Puskarich MA, Jones AE. The Impact of Timing
of Antibiotics on Outcomes in Severe Sepsis and Septic Shock. Crit Care Med. 2015
Sep;43(9):1907–15.
62. Klompas M, Calandra T, Singer M. Antibiotics for Sepsis—Finding the
Equilibrium. JAMA. 2018 Sep 14;3099(14):1–2.
63. Piacentini E, Ferrer R. Early and adequate empirical antibiotic treatment in sepsis
saves lives, but how should it be provided? Med Intensiva (English Ed. 2015
Nov;39(8):457–8.
64. Singer M. Antibiotics for Sepsis: Does Each Hour Really Count, or Is It Incestuous
Amplification? Am J Respir Crit Care Med. 2017 Oct 1;196(7):800–2.
65. Singh R, Sripada L, Singh R. Side effects of antibiotics during bacterial infection:
Mitochondria, the main target in host cell. Mitochondrion. 2014 May;16:50–4.
66. Alverdy JC, Krezalek MA. Collapse of the Microbiome, Emergence of the
Pathobiome, and the Immunopathology of Sepsis. Crit Care Med. 2017
Feb;45(2):337–47.
67. Wright J, Paauw DS. Complications of Antibiotic Therapy. Med Clin North Am.
2013 Jul;97(4):667–79.
68. Garnacho-Montero J, Gutiérrez-Pizarraya A, Escoresca-Ortega A, Fernández-
Delgado E, López-Sánchez JM. Adequate antibiotic therapy prior to ICU admission
72
in patients with severe sepsis and septic shock reduces hospital mortality. Crit Care.
2015 Dec 27;19(1):302.
69. Leone M, Bechis C, Baumstarck K, Lefrant J-Y, Albanèse J, Jaber S, et al. De-
escalation versus continuation of empirical antimicrobial treatment in severe sepsis:
a multicenter non-blinded randomized noninferiority trial. Intensive Care Med. 2014
Oct 5;40(10):1399–408.
70. Timsit J-F, Ruppe E, Ferrer R. Focus on sepsis: new concepts and findings in sepsis
care. Intensive Care Med. 2018 Oct 10;7–9.
71. Levy MM, Pronovost PJ, Dellinger RP, Townsend S, Resar RK, Clemmer TP, et al.
Sepsis change bundles: Converting guidelines into meaningful change in behavior
and clinical outcome. Crit Care Med. 2004 Nov;32(Supplement):S595–7.
72. Lagu T. Translating Knowledge into Practice in Critical Care Settings. Crit Care
Med. 2013 Nov;41(11):2659–60.
73. Bloos F, Thomas-Rüddel D, Rüddel H, Engel C, Schwarzkopf D, Marshall JC, et al.
Impact of compliance with infection management guidelines on outcome in patients
with severe sepsis: a prospective observational multi-center study. Crit Care. 2014
Mar 3;18(2):R42.
74. Garnacho-Montero J, Huici-Moreno MJ, Gutiérrez-Pizarraya A, López I, Márquez-
Vácaro J, Macher H, et al. Prognostic and diagnostic value of eosinopenia, C-
reactive protein, procalcitonin, and circulating cell-free DNA in critically ill patients
admitted with suspicion of sepsis. Crit Care. 2014;18(3):R116.
75. Wacker C, Prkno A, Brunkhorst FM, Schlattmann P. Procalcitonin as a diagnostic
marker for sepsis: a systematic review and meta-analysis. Lancet Infect Dis. 2013
May;13(5):426–35.
76. Clec’h C, Ferriere F, Karoubi P, Fosse JP, Cupa M, Hoang P, et al. Diagnostic and
prognostic value of procalcitonin in patients with septic shock. Crit Care Med. 2004
May;32(5):1166–9.
77. Arora S, Singh P, Singh PM, Trikha A. Procalcitonin Levels in Survivors and
Nonsurvivors of Sepsis. Shock. 2015 Mar;43(3):212–21.
73
78. Marshall JC, al Naqbi A. Principles of Source Control in the Management of Sepsis.
Crit Care Clin. 2009 Oct;25(4):753–68.
79. Blanco J, Muriel-Bombín A, Sagredo V, Taboada F, Gandía F, Tamayo L, et al.
Incidence, organ dysfunction and mortality in severe sepsis: a Spanish multicentre
study. Crit Care. 2008;12(6):R158.
80. Torgersen C, Moser P, Luckner G, Mayr V, Jochberger S, Hasibeder WR, et al.
Macroscopic Postmortem Findings in 235 Surgical Intensive Care Patients with
Sepsis. Anesth Analg. 2009 Jun;108(6):1841–7.
81. Boyer A, Vargas F, Coste F, Saubusse E, Castaing Y, Gbikpi-Benissan G, et al.
Influence of surgical treatment timing on mortality from necrotizing soft tissue
infections requiring intensive care management. Intensive Care Med. 2009 May
20;35(5):847–53.
82. Kobayashi L, Konstantinidis A, Shackelford S, Chan LS, Talving P, Inaba K, et al.
Necrotizing Soft Tissue Infections: Delayed Surgical Treatment Is Associated With
Increased Number of Surgical Debridements and Morbidity. J Trauma Inj Infect Crit
Care. 2011 Nov;71(5):1400–5.
83. Wong C-H, Chang H-C, Pasupathy S, Khin L-W, Tan J-L, Low C-O. Necrotizing
fasciitis: clinical presentation, microbiology, and determinants of mortality. J Bone
Joint Surg Am. 2003 Aug;85–A(8):1454–60.
84. Moss RL, Musemeche CA, Kosloske AM. Necrotizing fasciitis in children: Prompt
recognition and aggressive therapy improve survival. J Pediatr Surg. 1996
Aug;31(8):1142–6.
85. Buck DL, Vester-Andersen M, Møller MH. Surgical delay is a critical determinant
of survival in perforated peptic ulcer. Br J Surg. 2013 Jul;100(8):1045–9.
86. Hecker A, Schneck E, Röhrig R, Roller F, Hecker B, Holler J, et al. The impact of
early surgical intervention in free intestinal perforation: a time-to-intervention pilot
study. World J Emerg Surg. 2015 Dec 6;10(1):54.
87. Tridente A, Clarke GM, Walden A, McKechnie S, Hutton P, Mills GH, et al.
Patients with faecal peritonitis admitted to European intensive care units: an
74
epidemiological survey of the GenOSept cohort. Intensive Care Med. 2014 Feb
4;40(2):202–10.
88. Opal SM. Source Control in Sepsis Urgent or Not So Fast? Crit Care Med. 2017
Jan;45(1):130–2.
89. Wagner a. K, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regression
analysis of interrupted time series studies in medication use research. J Clin Pharm
Ther. 2002 Aug;27(4):299–309.
90. Sánchez B, Ferrer R, Suarez D, Romay E, Piacentini E, Gomà G, et al. Declining
mortality due to severe sepsis and septic shock in Spanish intensive care units: A
two-cohort study in 2005 and 2011. Med Intensiva. 2017 Jan;41(1):28–37.
91. Koperna T, Schulz F. Relaparotomy in Peritonitis: Prognosis and Treatment of
Patients with Persisting Intraabdominal Infection. World J Surg. 2000 Jan
14;24(1):32–7.
75
11.- ANEXO 1: PUBLICACIONES DE LA TESIS DOCTORAL
PUBLICACIÓN 1
Improved empirical antibiotic treatment of sepsis after an educational intervention:
the ABISS-Edusepsis study. Ricard Ferrer*, María Luisa Martínez*, Gemma Gomà,
David Suárez, Luis Álvarez-Rocha, María Victoria de la Torre, Gumersindo González,
Rafael Zaragoza9, Marcio Borges, Jesús Blanco, Eduardo Palencia Herrejón, Antonio
Artigas and for the ABISS-Edusepsis Study group. Crit Care 2018 Jun 22;22(1):167.
* Han contribuido por igual.
Factor de impacto de la revista: 6.425 según ISI Journal Citation Reports 2017.
PUBLICACIÓN 2
Impact of Source Control in Patients With Severe Sepsis and Septic Shock. María
Luisa Martínez; Ricard Ferrer; Eva Torrents;Raquel Guillamat-Prats; Gemma Gomà;
David Suárez; Luis Álvarez-Rocha; Juan Carlos Pozo Laderas; Ignacio Martín-Loeches;
Mitchell M. Levy; Antonio Artigas; for the Edusepsis Study Group. Critical Care Medicine
2017 Jan;45(1):11-19.
Factor de impacto de la revista: 7.05 según ISI Journal Citation Reports 2016.
76
RESEARCH Open Access
Improved empirical antibiotic treatment ofsepsis after an educational intervention: theABISS-Edusepsis studyRicard Ferrer1,2*†, María Luisa Martínez3†, Gemma Gomà4, David Suárez5, Luis Álvarez-Rocha6,María Victoria de la Torre7, Gumersindo González8, Rafael Zaragoza9, Marcio Borges10, Jesús Blanco2,11,Eduardo Palencia Herrejón12, Antonio Artigas2,3,4 and for the ABISS-Edusepsis Study group
Abstract
Background: Early appropriate antibiotic treatment is essential in sepsis. We aimed to evaluate the impact of amultifaceted educational intervention to improve antibiotic treatment. We hypothesized that the intervention wouldhasten and improve the appropriateness of empirical antibiotic administration, favor de-escalation, and decrease mortality.
Methods: We prospectively studied all consecutive patients with sepsis/septic shock admitted to 72 intensive care units(ICUs) throughout Spain in two 4-month periods (before and immediately after the 3-month intervention). We comparedprocess-of-care variables (resuscitation bundle and time-to-initiation, appropriateness, and de-escalation of empiricalantibiotic treatment) and outcome variables between the two cohorts. The primary outcome was hospital mortality. Weanalyzed the intervention’s long-term impact in a subset of 50 ICUs.
Results: We included 2628 patients (age 64.1 ± 15.2 years; men 64.0%; Acute Physiology and Chronic Health Evaluation(APACHE) II, 22.0 ± 8.1): 1352 in the preintervention cohort and 1276 in the postintervention cohort. In thepostintervention cohort, the mean (SD) time from sepsis onset to empirical antibiotic therapy was lower (2.0(2.7) vs. 2.5 (3.6) h; p = 0.002), the proportion of inappropriate empirical treatments was lower (6.5% vs. 8.9%;p = 0.024), and the proportion of patients in whom antibiotic treatment was de-escalated was higher (20.1%vs. 16.3%; p = 0.004); the expected reduction in mortality did not reach statistical significance (29.4% in thepostintervention cohort vs. 30.5% in the preintervention cohort; p = 0.544). Gains observed after theintervention were maintained in the long-term follow-up period.
Conclusions: Despite advances in sepsis treatment, educational interventions can still improve the delivery ofcare; further improvements might also improve outcomes.
Keywords: Sepsis, Septic shock, Quality improvement, Timing of antibiotics, De-escalation, Hospital mortality
BackgroundSepsis, a life-threatening organ dysfunction due to dysregu-lated host response to infection, is common and its inci-dence seems to be increasing [1, 2]. About a quarter ofpatients with sepsis go on to develop septic shock [3, 4].Sepsis is associated with high morbidity and mortality
[1, 2, 5]. The Surviving Sepsis Campaign (SSC), aninternational effort to optimize treatment for sepsis throughevidence-based guidelines, has resulted in sustained con-tinuous quality improvement associated with decreasedmortality [6]. In Spain, a nationwide effort based on theSSC, the EDUSEPSIS intervention, succeeded in improvingprocess-of-care and outcome measures [7].Infection control is the cornerstone of treatment, and
both the timeliness and appropriateness of empirical anti-biotic treatment are considered essential aspects of sepsismanagement [8, 9]. The SSC guidelines recommendadministering empirical broad-spectrum antibiotics within
* Correspondence: [email protected]†Ricard Ferrer and María Luisa Martínez contributed equally to this work.1Intensive Care Department, Shock, Organ Dysfunction and ResuscitationResearch Group, Vall d’Hebron Research Institute, Vall d’Hebron UniversityHospital, Autonomous University of Barcelona, Barcelona, Spain2CIBER Enfermedades Respiratorias, Madrid, SpainFull list of author information is available at the end of the article
© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Ferrer et al. Critical Care (2018) 22:167 https://doi.org/10.1186/s13054-018-2091-0
1 h of identification of sepsis [10]. This recommendationis supported by a secondary analysis of the Edusepsisstudy, which showed that early administration ofbroad-spectrum empirical antibiotics was the componentof the SSC bundles most strongly associated withincreased survival [11]. Several studies have identifiedworse outcomes associated with delays in administeringappropriate antibiotic treatment [1, 2, 12–15] and withinappropriate antibiotic therapy [12, 13, 16]. The SSCguidelines also recommend reassessing antibiotic treat-ment to determine whether de-escalation is possible, andfailure to de-escalate might be associated with worseoutcomes [8, 16, 17] and could lead to development ofresistant microbes [8, 18].Given the importance of infection control in sepsis
management, we designed a multifaceted educationalintervention to improve antimicrobial therapy inpatients with sepsis: the Antibiotic Intervention inSevere Sepsis (ABISS) study. We hypothesized that theintervention would decrease the time to the administra-tion of empirical antibiotics, increase the proportion ofpatients receiving appropriate empirical antibiotics, favorde-escalation, and decrease mortality.
MethodsThrough the Edusepsis study group and the Spanish Soci-ety of Critical Care Medicine’s working group on infectiousdisease and sepsis, we invited 115 centers throughout Spainto participate in a national educational intervention to im-prove infection control: the ABISS-Edusepsis study. A totalof 72 medical-surgical intensive care units (ICUs) locatedthroughout Spain took part. The study was approved byour institutional review board (reference 2,011,521) and theethics committees at each participating center approved thestudy protocol and waived the need for informed consentbecause the intervention was a quality improvement pro-gram and patients’ anonymity was guaranteed.
Study designWe designed a before-and-after study to compare a prein-tervention cohort consisting of all consecutive patientswith severe sepsis or septic shock admitted to the partici-pating ICUs in the 4-month period before the educationalprogram began (April–July 2011) against a postinterven-tion cohort in the 4-month period immediately after theintervention (April–July 2012). The intervention tookplace over a 3-month period (January–March 2012) dur-ing which no patient data were collected. Furthermore, toassess the long-term impact of the intervention, we ana-lyzed a third cohort 6 months after the postinterventionperiod (January–April 2013).The study sites, study design, data collection, and
quality-control measures are detailed in Additional file 1.Severe sepsis was defined as sepsis associated with organ
dysfunction unexplained by other causes. Septic shockwas defined as sepsis associated with systolic blood pres-sure < 90 mmHg, mean arterial pressure < 65 mmHg, or areduction in systolic blood pressure > 40 mmHg frombaseline despite adequate volume resuscitation [19]. Theonset of sepsis (T0) was determined according to thepatient’s location within the hospital when sepsis was diag-nosed: we used the time of triage for patients diagnosed inthe emergency department and searched the clinical docu-mentation for clues indicating the time of diagnosis forpatients diagnosed in the wards or the ICU (seeAdditional file 1).
InterventionBetween January and March 2012, we implemented amultifaceted educational program to train physicians andnursing staff in the emergency department, medical andsurgical wards, and ICU in sepsis care, with specialemphasis on antimicrobial management. Training includedprimary sepsis care focused on timing and strategy ofempiric measures against infection. The interventionconsisted of educational outreach, periodic reminders,auditing and feedback, and a videogame. The educationaloutreach included interactive educational sessions in whichthe local leader gave a 30-min slide presentation based onthe SSC guidelines recommendations focused on theimportance of (a) infection control in sepsis, (b) the timeli-ness and appropriateness of empirical antibiotic adminis-tration, and (c) de-escalation of antibiotic treatment. Eachcenter was provided with pocket guides and posters withrecommendations from the Spanish Society of CriticalCare Medicine and Coronary Units. To facilitate antibioticprescription, researchers preferentially used their localguideline or an electronic clinical decision support system(www.es.dgai-abx.de). Attendees received weekly email andcellphone text reminders reiterating the most importantpoints from the educational outreach program. Addition-ally, a videogame was developed to provide staff attendingseptic patients (http://www.edusepsis.org/en/training.html)with opportunities to put the guidelines into practice withsimulations. The educational intervention is detailed inAdditional file 2.
Process-of-care and outcome measurementsWe recorded demographic characteristics (age and sex),Charlson Comorbidity Score, diagnosis on admission(medical, emergency surgery, elective surgery), source ofsepsis, type of infection (community-acquired,healthcare-related, hospital-acquired, or ICU-acquired),Sequential Organ Failure Assessment (SOFA) scale onadmission, and worst Acute Physiology and ChronicHealth Evaluation (APACHE) II score during the first24 h in the ICU.
Ferrer et al. Critical Care (2018) 22:167 Page 2 of 10
We recorded the following process-of-care variablesrelated to empirical antibiotic treatment: drugs adminis-tered, time initiated, appropriateness, and de-escalation.Time to first antibiotic administration is reported as thedifference between onset of sepsis and first antibiotic ad-ministration. Empirical treatment was classified as ap-propriate (when ≥ 1 of the drugs administered wasconsidered effective based on the susceptibility in theantibiogram of the causative microorganism isolated incultures), inappropriate (when the above criterion wasnot met), or indeterminate (when no causative micro-organism was isolated in cultures or no cultures weretaken). Antibiotic strategies once culture results wereavailable were classified as “de-escalation” (switch to orinterruption of a drug class resulting in a less broadspectrum of coverage, “no change” (empirical therapymaintained without modification), or “change” (due topoor clinical evolution, uncovered microorganism, pos-sible toxicity, or other reasons).We also recorded time from T0 to other acts and tar-
gets prescribed in the SSC guidelines: measuring serumlactate, obtaining blood cultures, and administeringfluids and/or vasopressors in patients with hypotensionand/or lactate > 4 mmol/L [10].Patients were followed up until death or hospital dis-
charge. The primary outcome variable was hospital mor-tality. Secondary outcome measures included days onmechanical ventilation, days on vasopressors, hospitaland ICU lengths of stay, and ICU mortality.
Statistical analysisDescriptive statistics included frequencies and percent-ages for categorical variables and means and standarddeviations (SD) for continuous variables. To comparecategorical variables, we used the chi-squared (χ2) testor Fisher’s exact test as appropriate. To compare con-tinuous variables, we used Student’s t test or theMann-Whitney U test as appropriate.To assess the effectiveness of the intervention, we com-
pared the values of the process and outcome variablesrecorded in the preintervention cohort against thoserecorded in the postintervention cohort. To assesslong-term effectiveness, we compared the values recordedin the long-term follow-up period against those recorded inthe postintervention period in the same subset of hospitals.We used multivariate linear regression to determine
the association between the intervention and time to an-tibiotics after adjusting for possible confounders. More-over, as a sensitivity analysis, we performed segmentedregression analysis to estimate the size of the effect ofthe intervention for reducing time to antibiotic.We used multivariate stepwise logistic regression to
assess the impact of the intervention on outcome (hospitalmortality). Variables entered in the logistic regression
model were those with a relationship in the univariateanalysis (p ≤ 0.1) or with a potential plausible relationshipwith the outcome. The final model included the interven-tion, age, sex, comorbidities, APACHE II, SOFA, type ofinfection, and source of sepsis as independent variables.Statistical tests were two-tailed. We used SPSS version15.0 (SPSS, Chicago, IL, USA) for all analyses.
ResultsPatients characteristicsA total of 2628 patients (mean age 64.1 (15.2) years;mean APACHE II score 22.0 (8.1); 64% male) wereincluded during the preintervention (n = 1352) and post-intervention periods (n = 1276). Table 1 reports thedemographic and clinical characteristics of the patientsin these two cohorts. Compared to patients in the prein-tervention cohort, those in the postintervention cohorthad slightly less severe illness and had a greater propor-tion of community-acquired infections. There were nodifferences in age, sex, or SOFA scores. Diagnosis on ad-mission was mainly medically or surgically urgent inboth cohorts and the main sources of sepsis in both pe-riods were pneumonia and acute abdominal infections.Compliance with three of the six tasks in the 6-h
resuscitation bundle (lactate measurement, blood cul-tures before antibiotics, and early administration ofbroad spectrum antibiotics) improved significantly afterthe intervention (see Fig. 1). Compared to the preinter-vention cohort, the mean time from sepsis onset toempirical antibiotic therapy was shorter in the postinter-vention cohort (2.0 (2.7) vs. 2.5 (3.6) hours; p = 0.002).The proportion of patients receiving inappropriate em-pirical antibiotic treatment decreased from 8.9% in thepreintervention cohort to 6.5% in the postinterventioncohort (p = 0.024) and the proportion of patients inwhom antibiotic treatment was de-escalated was higherin the postintervention cohort (20.1% vs. 16.3% in thepreintervention cohort; p = 0.004) (see Fig. 1).After adjustment for severity of illness, type and source
of infection, and demographic characteristics, mean timeto first antibiotic after sepsis onset was significantly lowerin the postintervention group (− 0.45 (95% CI − 0.75 to −1.56); p = 0.003) (see Additional file 3: Table S1).Figure 2 shows the time series of mean time to first anti-
biotic after sepsis onset per month in the preinterventionand postintervention periods. The segmented regressionanalysis found a significant change in level (− 0.92 (95% CI− 1.51 to − 0.33); p = 0.010), indicating an abrupt interven-tion effect (see Additional file 4: Table S2).Table 2 reports the outcome variables. No significant
differences were observed in any of the outcome variables.Overall hospital mortality (29.9%) did not differ betweencohorts. Multivariable logistic regression (Table 3) to
Ferrer et al. Critical Care (2018) 22:167 Page 3 of 10
Table 1 Demographic and clinical characteristics of patients
Patient Characteristic Preintervention cohort (n = 1352) Postintervention cohort (n = 1276) P
General data
Age (years), mean (SD) 64.3 (15.3) 63.8 (15.1) 0.400
Sex (male), n (%) 858 (63.5) 824 (64.6) 0.552
APACHE II, mean (SD) 22.5 (8.1) 21.4 (8.0) 0.001
SOFA, mean (SD) 8.7 (3.5) 8.5 (3.5) 0.102
Charlson, mean (SD) 2.7 (2.3) 2.7 (2.3) 0.391
Source of sepsis, n (%) 0.018
Pneumonia 454 (33.6) 403 (31.6)
Acute abdominal infection 452 (33.4) 431 (33.8)
Urinary tract infection 229 (16.9) 209 (16.4)
Soft-tissue infection 82 (6.1) 98 (7.7)
Meningitis 26 (1.9) 43 (3.4)
Catheter-related bacteremia 30 (2.2) 19 (1.5)
Other infections 79 (5.8) 73 (5.7)
Type of infection, n (%) 0.013
Community 802 (59.3) 835 (65.4)
Nosocomial 302 (22.3) 249 (19.5)
ICU 65 (4.8) 51 (4)
Healthcare-related 183 (13.5) 141 (11.1)
Diagnosis on admission, n (%) 0.817
Medical 950 (70.3) 891 (69.8)
Urgent surgical 318 (23.5) 311 (24.4)
Non-urgent surgical 84 (6.2) 74 (5.8)
Abbreviations: APACHE II Acute Physiology and Chronic Health Evaluation II, SOFA Sequential Organ Failure Assessment, ICU intensive care unit, SDstandard deviations
1.5
2.0
2.5
3.0
Months before and after the intervention
Tim
e to
fisr
t ant
ibio
tic (h
)
1 2 3 4 1 2 3 4
Start of postintervention period
Fig. 1 Mean time to first antibiotic after sepsis onset before and after the educational intervention, excluding patients with previous antibiotictreatment (n = 858)
Ferrer et al. Critical Care (2018) 22:167 Page 4 of 10
adjust for possible confounders showed no relationshipbetween intervention and hospital mortality.
Long-term follow upFifty centers participated in a 4-month follow-up study tomeasure the long-term effects of the intervention (n = 830patients). Compared to patients from the postintervention
cohort, those in the long-term cohort had lower meanCharlson score and there was a higher proportion of pa-tients with pneumonia (see Additional file 5: Table S3).The percentage of patients in whom care complied
with resuscitation measures was stable with respect tothe postintervention cohort. Time from sepsis onset toempirical antibiotic therapy increased slightly but not
Reduct
ion o
f the s
pectru
m
No chan
ge of a
ntibio
tic
Change d
ue to p
oor clin
ical e
volu
tion
Change d
ue to u
ncove
red m
icroorg
anism
Other
0
20
40
60
Change of antibiotic at 72 h
%
***
***
******
***
Preintervention
Postintervention
Inap
propria
te tr
eatm
ent
Appropria
te tr
eatm
ent
Negat
ive cu
lture
s
Culture
s not d
one
Patien
t died
in le
ss th
an 72
h0
20
40
60
Evaluation of antibiotic treatment at 72 h
%
*
*
*
**
Lacta
te m
easu
red
Blood cu
lture
s bef
ore an
tibio
tics
Broad
spec
trum
antib
iotic
s
Fluid
s and va
sopre
ssors
0
20
40
60
80
100
%
Sepsis Resuscitation Bundle 6 h
***
*
*ns
Fig. 2 Compliance with process-of-care measures in the preintervention vs. postintervention cohort. a The proportion of adherence to theresuscitation bundle in the preintervention vs. postintervention cohort. b Evaluation of treatment at 72 h. c Change of antibiotic at 72 h. NS,not statistically significant; *p < 0.05; ***p < 0.005
Table 2 Outcome measurements in the preintervention vs. postintervention cohort
Outcome measurements Preintervention cohort (n = 1352) Postintervention cohort (n = 1276) P
Duration of MV, days mean (SD) 6.9 (14.4) 6.6 (12.4) 0.577
Duration of vasopressors, days mean (SD) 4.0 (8.0) 4.3 (7.0) 0.369
ICU stay, days mean (SD) 12.0 (17.0) 11.5 (14.9) 0.443
Hospital stay, days mean (SD) 30.0 (29.7) 28.4 (28.9) 0.161
Mortality, n (%)
ICU 332 (24.6) 301 (23.6) 0.562
Hospital 412 (30.5) 375 (29.4) 0.544
Abbreviations: ICU intensive care unit, MV mechanical ventilation, SD standard deviations
Ferrer et al. Critical Care (2018) 22:167 Page 5 of 10
significantly in the long-term cohort and the proportionof patients in whom antibiotic treatment wasde-escalated remained unchanged (see Additional file 6:Table S4).No significant differences were observed in the out-
come variables (see Additional file 7: Table S5). Multi-variable logistic regression showed no relationshipbetween the intervention and hospital mortality (seeAdditional file 8: Table S6).
DiscussionThis large-scale multifaceted educational interventionimproved the overall use of antibiotics in sepsis, improv-ing efficacy by lowering the time from sepsis onset toantibiotic treatment and increasing the proportion of pa-tients who received appropriate empirical treatment andalso improving safety by increasing the proportion of pa-tients who received appropriate de-escalation. Import-antly, gains observed after the intervention weremaintained in the long-term follow-up period.These results lend strength to a growing body of litera-
ture showing that educational interventions can improvethe process of care in different contexts and conditions[15, 20–24]. However, Ramsay et al. [25] reviewed the ef-fectiveness of interventions to improve antibiotic prescrib-ing in hospital inpatients and concluded that most studies
supporting these interventions had fundamental flaws indesign and/or execution, pointing out that segmented re-gression analyses are recommendable when analyzing theeffects of interventions on process measures. In our data,these analyses estimating intervention effects in inter-rupted time series studies showed a significant change inlevel, indicating an abrupt intervention effect, and thusconfirming the results of our multivariate linear regressionand strengthening the conclusion that the intervention re-duced the time to first antibiotic.Since multifaceted interventions appear to be more ef-
fective than more limited approaches to changing behavior[26, 27], we aimed to strengthen the intervention by includ-ing various approaches to transferring and reinforcingknowledge. Two of the approaches we used, educationaloutreach and auditing and feedback, are well-establishedapproaches to knowledge translation; the other two, weeklyreminders and an educational game to increase awarenessand improve adherence to guidelines, are supported bymore limited experiences [28–31].Despite significant improvement in antibiotic treat-
ment, no significant decrease in mortality was observed.One factor that probably contributed to our not identify-ing an impact on mortality is that our study focusedmainly on improving antimicrobial treatment. It is un-likely that a limited intervention at a single point in timewould have a profound impact on survival. Seymour etal. [32] in a study with more than 49,000 emergency de-partment patients with sepsis showed a linear associ-ation between time to antibiotic and mortality. Thisstudy was done after the implementation of a statewidemandate requiring protocolized sepsis care rather thanafter educational intervention, and compliance with the3-h bundle was very high.The decrease in mortality associated with the reduction
in time to antibiotic observed in the present study wascomparable to forecasts based on previous research. Inthe Seymour et al. study [32], the odds of death increasedby 4% for every 1-h delay in receiving antibiotics, andKumar et al. [14] found mortality decreased 7% per hourof reduction in time to antibiotics. We observed ahalf-hour reduction in time to antibiotic. Interestingly, theobserved mortality reduction was practically the same asin the Kumar et al. study: 7.2% per hour (3.6% perhalf-hour, from 30.5% mortality in the preintervention to29.4% in the postintervention period).Unfortunately, how-ever, our study was underpowered to detect this differencein mortality; to achieve a statistically significant result witha type I error rate of 5% and 80% power would haverequired the inclusion of 50,000 patients.Although delaying antibiotic administration in patients
with sepsis is inadvisable, the evidence supporting the mor-tality benefits of early antibiotic administration is inconclu-sive. Ferrer et al. [15], in a study with more than 17,000
Table 3 Multivariate analysis of risk factors for hospital mortality
Factors OR 95% CI P
Interventional cohort 1.08 0.89–1.31 0.419
Agea 1.02 1.01–1.03 < 0.001
Sexb 0.82 0.67–1.01 0.057
SOFAa 1.11 1.07–1.15 < 0.001
APACHE IIa 1.08 1.06–1.10 < 0.001
Charlsona 1.06 1.02–1.11 0.005
Type of infectionc
Nosocomial 2.03 1.61–2.56 < 0.001
ICU 2.49 1.61–3.87 < 0.001
Healthcare-related 1.33 0.99–1.79 0.058
Source of sepsisd
Acute abdominal infection 0.79 0.63–1.00 0.051
Urinary tract infection 0.25 0.18–0.35 < 0.001
Meningitis 1.11 0.60–2.05 0.742
Soft-tissue infection 0.69 0.47–1.03 0.072
Catheter-related bacteremia 0.54 0.27–1.09 0.084
Other infections 1.00 0.66–1.52 0.999
Abbreviations: SOFA Sequential Organ Failure Assessment, APACHE II AcutePhysiology and Chronic Health Evaluation II, ICU intensive care unitaPer each point of increasebCompared with male sexcCompared to community-acquired infection.dCompared to pneumonia
Ferrer et al. Critical Care (2018) 22:167 Page 6 of 10
patients, confirmed that delayed antibiotic administrationis associated with increased hospital mortality. More re-cently, in a large multicenter sample of patients with sepsis,Liu et al. [33] found a linear association between delays inantibiotic administration and mortality; patients with septicshock received the greatest benefit from early administra-tion. Whiles et al. [34] similarly reported an 8% increase inthe chance of developing septic shock for each hour ofdelay in antibiotic administration. Therefore, the mortalitybenefits of early antibiotic administration are probably es-pecially important in the most critically ill patients.A recent meta-analysis of six studies including more
than 16,000 patients found no significant mortality benefitof administering antibiotics within 3 h of emergency de-partment triage or within 1 h of recognition of shock, al-though the inclusion of studies with small samples andheterogeneity among studies may limit the conclusions[35]. In a cluster randomized trial to evaluate the effect ofa multifaceted educational intervention for anti-infectiousmeasures on sepsis mortality, the MEDUSA study group[36] found no association between the intervention andimpact on time to empiric antibiotic or mortality; how-ever, the authors concluded the intervention was insuffi-cient, inconsistent, and mainly applied only in the ICU.Our results are similar to those of the pediatric substudyof the ABISS-Edupsesis project [37], a multifaceted educa-tional intervention in children with sepsis and septicshock, which found decreased time to antibiotic adminis-tration but not decreased mortality after the intervention,probably due to the small sample size.Singer [38] questions the importance of earlier anti-
biotic treatment mainly because the risk of increasingantimicrobial resistance. Our intervention increased theproportion of patients receiving appropriatede-escalation from 16.3% to 20.1%, and these improve-ments were maintained in the long-term follow-upperiod. In a recent prospective study, about 35% of pa-tients with sepsis received appropriate de-escalation andde-escalation was associated with lower mortality [39]. Amulticenter non-blinded randomized non-inferiority trialin 1116 patients with sepsis found that de-escalation didnot worsen patient outcomes [40].Efforts to improve the treatment of sepsis, especially
those targeting empirical antibiotic administration,need to encompass all levels of care. We targeted allprofessionals caring for septic patients. For knowledgetransfer to benefit patients, it is often necessary to re-organize how care is delivered [41]. The key to im-proving outcomes in sepsis is motivating professionalsto implement evidence-based measures and providingthem with feedback about the impact of these mea-sures [42]. To this end, it is important to monitorprocess-of-care variables and outcome variables. Oneof the greatest benefits of interventions like ours is
their contribution to shaping a culture that fostersthe desire to improve, and an ongoing commitmentto excellence in patient care [41].
LimitationsThe before-and-after design of our study has inherentweaknesses. The influence of secular trends can be difficultto separate from the effects of the intervention in studiesemploying before-and-after designs. In these cases, expertsrecommend using a stepped-wedge design. However, noother major changes in protocol were instituted in the rela-tively short gaps between the three periods. Not using acontrol group makes it effectively impossible to ensure thatthe changes observed after the intervention would not havehappened anyway. Although using centers where nothingwas done to improve antibiotic therapy for sepsis as a con-trol group might have enabled us to sort out the effects ofa possible secular trend, we considered this approachmight be unethical. Another argument against uncon-trolled before-and-after studies is that it is impossible toensure that the intervention site is representative andchange is merely an expression of regression to the mean[25]; however, the large number of centers participating inour study safeguard against this. We also performed a seg-mented regression analysis, a powerful method for estimat-ing intervention effects in interrupted time series [43].Although our intervention employed a broad multifa-
ceted approach, other measures such as real-time auto-mated alerting to remind clinicians were not included,and this may be partly responsible for our failure to finda strong effect on outcome. The long-term follow-upanalysis (6 months after the postintervention period)might not be late enough to assess the long-term impactof the intervention. Our earlier study found that someimprovements were maintained after 1 year [7], and thevery long-term impact of the interventions was recentlyconfirmed in another study between 2005 and 2011 thatidentified dramatically decreased mortality related to se-vere sepsis/septic shock [44].Despite these limitations, our study has noteworthy
strengths. The large number of ICUs that participated en-abled us to prospectively enroll and follow large numbersof patients with sepsis in each data collection period andincreases the likelihood that our results can be applied inother contexts. Our strict quality control helped ensure ahomogeneous database and the validity of our data.
ConclusionSepsis is a time-dependent condition in which early em-pirical antibiotic treatment can improve survival. Boththe time from onset to administration of antibiotics andantibiotic de-escalation are modifiable factors worthy ofour attention. The ABISS intervention reduced the timeto antibiotic administration and the proportion of
Ferrer et al. Critical Care (2018) 22:167 Page 7 of 10
patients in whom antibiotic treatment was de-escalated,thus demonstrating that despite advances in sepsis treat-ment in recent years, educational interventions can stillimprove the delivery of care. Further improvementsmight also improve outcomes.
Additional files
Additional file 1: Appendix 1. Appendix describing in detail the studydesign, the approach to data collection, and the quality-control measuresto ensure data reliability. (DOC 27 kb)
Additional file 2: Appendix 2. Appendix describing in detail theeducational intervention. (DOC 25 kb)
Additional file 3: Table S1. Multivariate linear regression for time toantibiotic. (DOC 38 kb)
Additional file 4: Table S2. Segmented regression model for time inhours to first antibiotic. (DOC 29 kb)
Additional file 5: Table S3. Demographic and clinical characteristics ofpatients in the long-term cohort. (DOC 48 kb)
Additional file 6: Table S4. Compliance with process-of-care measure-ments in the long-term cohort. (DOC 44 kb)
Additional file 7: Table S5. Outcome measurements in the long-termcohort. (DOC 31 kb)
Additional file 8: Table S6. Multivariate analysis of factors associatedwith mortality in the long-term cohort. (DOC 37 kb)
AbbreviationsABISS: Antibiotic Intervention in Severe Sepsis; APACHE: Acute Physiologyand Chronic Health Evaluation; CI: Confidence interval; ICU: Intensive careunits; IQR: Interquartile range; SD: Standard deviation; SOFA: SequentialOrgan Failure Assessment; SSC: Surviving Sepsis Campaign; T0: Onset ofsepsis
AcknowledgementsThe authors gratefully acknowledge John Giba for his assistance in thepreparation of the manuscript.The ABISS Edusepsis Study Group included Gemma Gomà, María LuisaMartínez, Antonio Artigas (Hospital de Sabadell, Consorci HospitalariUniversitari Parc Taulí), María del Mar Cruz (Hospital Virgen de la Salud,Toledo), Sandra Barbadillo (Hospital Universitario General de Cataluña),Francisco Fernández (Centro Clínico Delfos), Alberto PensadoCastiñeiras;Mª Teresa Rey Rilo, Luis Alvarez Rocha (Complexo HospitalarioUniversitario de A Coruña), Belén Jiménez Bartolomé (Hospital ClínicoUniversitario Lozano Blesa Zaragoza), Juan Diego Jiménez Delgado(Hospital Comarcal Don Benito-Villanueva, Extremadura), Demetrio Car-riedo Ule, Ana María Domínguez Berro, Francisco Javier Díaz Domínguez(Complejo Asistencial Universitario de León), Juan Machado Casas (Com-plejo Hospitalario de Jaén), Clara Laplaza Santos; Manuel García-Montesi-nos; Enrique Maraví Poma (Complejo Hospitalario de Navarra-Pamplona),Víctor López Ciudad; Pablo Vidal Cortés (Complejo Hospitalario de Ou-rense), Miguel Martínez Barrio, Mª Jesús López Pueyo (Hospital GeneralYagüe, Burgos), María Jesús López Cambra (Hospital general de Segovia),Pau Torrabadella, Álvaro Salcedo, Claudio Durán (Hospital UniversitariGermans Trias i Pujol), Iratxe Seijas (Hospital de Cruces de Bilbao), TeresaRecio Gómez (Hospital San Pedro de Alcántara, Cáceres), Ángel Arenzana(Hospital Virgen de la Macarena, Sevilla), Izaskun Azkarate (Hospital deDonostia), Sandra Rodríguez Bolaño (Hospital de Baza. Granada), PabloOlivares García (Hospital Gregorio Marañón, Madrid), Jordi Solé Violán(Hospital Universitario de Gran Canaria Dr. Negrín), Gerardo Aguilar Agui-lar (Hospital Clínic Universitari Valencia), Ángel Rodríguez Rencinas, MartaPaz Pérez, Elena Pérez Losada (Hospital de Salamanca), Fernando Martí-nez Sagasti (Hospital Clínico San Carlos, Madrid), José Luis García Allut(Hospital Clínico Universitario de Santiago de Compostela), FernandoDíez Gutiérrez, Francisco Gandía, Amanda Francisco Amador (HospitalClínico Universitario de Valladolid), Ramón Vegas Pinto (Hospital deAntequera, Málaga), Pilar Martínez Trivez (Hospital de Barbastro Huesca),
Nieves García Vázquez (Hospital Universitario de La Princesa), Luis Zapata,Paula Vera (Hospital de la Santa Creu i Sant Pau), Eduardo Antón (Hos-pital de Manacor), Juan Carlos Yébenes (Hospital de Mataró), María delas Olas Cerezo Arias (Hospital de Mérida), Francisco García delgado(Hospital de Montilla), Javier Fierro Rosón, Josefa Peinado Rodríguez(Hospital de Poniente), María Álvarez (Hospital de Terrassa), Paco ÁlvarezLerma (Hospital del Mar), Francisco Valenzuela (Hospital de Jerez), Patri-cia Albert de la Cruz (Hospital del Sureste), Rafael Blancas Gómez-Casero(Hospital del Tajo, Madrid), Montserrat Sisón Heredia (Hospital Dr. JoséMolina Orosa, Las Palmas), Perico Olaechea, Celia Sañudo (Hospital deGaldakao-Usansolo), José Manuel Gutiérrez Rubio (Complejo HospitalarioUniversitario de Albacete), Roberto Reig Valero (Hospital General de Cas-tellón), Hasania Abdel-Hadi Álvarez (Hospital General de Ciudad Real),Leandro Fajardo Feo (Hospital General de Fuerteventura), Pau Garro(Hospital General de Granollers), Francisco Navarro Pellejero (HospitalGeneral de la Defensa en Zaragoza), Ana Esther Trujillo Alonso (Hospitalgeneral de La Palma), Rosa Catalán (Hospital General de Vic), AssumptaRovira, Nicolás Rico (Hospital General Hospitalet de LLobregat), JoséManuel Allegue Gallego (Hospital General Universitario Santa Lucía Carta-gena), José Córdoba Alonso, Dolores Ocaña (Hospital La Inmaculada deHuercal-Overa. Almería), Juan Mora Ordóñez, Manuel Salido Mota (Hos-pital Regional Universitario Carlos Haya Málaga), Mª José Tolón Herrera,Paloma Dorado (Hospital Royo Villanova de Zaragoza), Arantxa LanderAzcona (Hospital San Jorge Huesca), Diego Mendoza (Hospital Sant JoanDespí Moisès Broggi), Francisca Prieto (Hospital Sta. Bárbara Puertollano),Mª Carmen Ramagge Martín (Hospital de La Línea), José Ignacio Ayes-tarán Rota (Hospital Son Espases), Marcio Borges (Hospital de Son Llat-zer), Enrique Piacentini, Ricard Ferrer (Hospital Mútua de Terrassa), JosepMaria Sirvent, Cristina Murcia, Gina Rognoni (Hospital Universitari Dr.Josep Trueta de Girona), José Antonio Gonzalo, Diego Parra Ruiz, NataliaBretón Díez, José Ignacio Argüelles Antuña (Hospital Universitario Centralde Asturias), Leonardo Lorente Ramos (Hospital Universitario de Canar-ias), Helena Yáñez (Hospital Universitario de Guadalajara), Ana Loza (Hos-pital Universitario de Valme), Borja Suberbiola (Hospital UniversitarioMarqués de Valdecilla), Domingo Ruiz de la Cuesta Martín (Hospital Uni-versitario, Miguel Servet Zaragoza), María del Mar Martín Velasco (Hos-pital Universitario Nuestra Señora de Candelaria), Antonio PontesMoreno, Rafael León López, Juan Carlos Pozo (Hospital UniversitarioReina Sofía de Córdoba), Luis Tamayo Lomas, Jesús Blanco, Arturo Muriel,José Ángel Berezo (Hospital Universitario Río Hortega), Paula Ramírez,Miguel Ángel Chiveli Monleón (Hospital Universitario y Politécnico La Fede Valencia), Juan Carlos Ruiz Rodríguez, Jesús Caballero, Adolf Ruiz, Ale-jandra García, Jordi Riera, Javier Sarrapio, Francesc Sanpedro, José CarlosMartin, Tatiana Acero (Hospital Universitari Vall Hebrón), Ana Carolina Ca-ballero, Silvia María Cortés Díaz (Hospital Virgen de la Concha, Zamora),M. Victoria de la Torre, Begoña Mora Ordóñez (Hospital Virgen de laVictoria), José Garnacho Montero (Hospital Virgen del Rocío), EduardoPalencia Herrejón, Begoña Bueno García (Hospital Infanta Leonor,Madrid), Gumersindo González-Díaz, Andrés Carrillo (Hospital MoralesMeseguer), Pedro Jesús Domínguez García (Hospital Juan Ramón Jiménez), RuthJorge García (Hospital Nuestra Señora de Gracia Zaragoza), Almudena Simón (Hos-pital Nuestra Señora del Prado), José Carlos Torralba Allué (Hospital GeneralObispo Polanco Teruel), Teresa Recio Gómez (Hospital San Pedro de Alcántara,Cáceres), Ricardo Díaz Abad (Hospital Severo Ochoa, Madrid), Mar Gobernado(Hospital Santa Bárbara, Soria), Francisco Guerrero Gómez (Hospital Torrecárdenas),José Castaño Pérez (Hospital Virgen de las Nieves), Fernando Bueno Andrés (Hos-pital Virgen del Puerto, Plasencia), Elena Bustamante Munguira, Gaspar Tuero (Hos-pital de Can Misses), José Francisco Olea Parejo (Hospital Lucus Augusti-Lugo),Miguel Soto, Susana Sancho Chinesta, Rafa Zaragoza (Hospital Universitario Dr.Peset de Valencia), Carmen Fernández González (Complejo Hospitalario de FerrolArquitecto Marcide), Manuel Castellano (Hospital Alto Guadalquivir), José MaríaBonell (Hospital Clínica Palma Planas), Mª Jesús Broch Porcar (Hospital de Sa-gunto), Néstor Bacelar (Clínica Corachan), Isabel Cremades (Hospital Reina Sofíade Murcia), and Miguel Valdivia and Pedro Galdós (Hospital Puerta del Hierro).
FundingResearch grant Instituto de Salud Carlos III (FIS 10/01497), CM12/00066.
Availability of data and materialsThe datasets used and/or analyzed during the current study are availablefrom the corresponding author on reasonable request.
Ferrer et al. Critical Care (2018) 22:167 Page 8 of 10
Authors’ contributionsRF and MLM had full access to all of the data in the study and takeresponsibility for the integrity of the data and the accuracy of the dataanalysis. GG and MLM: data monitoring. RF and AA: study concept anddesign. RF, ML M, LAR, MVT, GG, RZ, MB, JB, EPH, and AA: acquisition of data.RF and AA: general coordination. LAR, MVT, GG, RZ, MB, JB, and EPH: areacoordinators. RF, MLM, and AA: drafting of the manuscript and analysis andinterpretation of data. All authors: critical revision of the manuscript forimportant intellectual content. RF and DS: statistical analysis. All authors readand approved the final manuscript.
Ethics approval and consent to participateThe study was approved by our institutional review board (reference2,011,521) and the ethics committees at each participating center approvedthe study protocol and waived the need for informed consent because theintervention was a quality improvement program and patients’ anonymitywas guaranteed.
Consent for publicationNot applicable.
Competing interestsAA received funding for himself from Grifols, Lilly Foundation, and Fisher &Paykel. Other authors declare that they have no competing interests.
Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.
Author details1Intensive Care Department, Shock, Organ Dysfunction and ResuscitationResearch Group, Vall d’Hebron Research Institute, Vall d’Hebron UniversityHospital, Autonomous University of Barcelona, Barcelona, Spain. 2CIBEREnfermedades Respiratorias, Madrid, Spain. 3Intensive Care Department,Hospital Universitario General de Catalunya, Autonomous University ofBarcelona, Sant Cugat del Vallés, Spain. 4Intensive Care Department,Corporación Sanitaria Universitaria Parc Taulí, Autonomous University ofBarcelona, Sabadell, Spain. 5Epidemiology and Assessment Unit, FundacióParc Taulí, Autonomous University of Barcelona, Sabadell, Spain. 6IntensiveCare Department, Hospital Universitario de la Coruña, A Coruña, Spain.7Intensive Care Department, Hospital Universitario Virgen de la Victoria,Málaga, Spain. 8Intensive Care Department, Hospital General UniversitarioMorales Meseguer, Murcia, Spain. 9Intensive Care Department, HospitalUniversitario Doctor Peset, Valencia, Spain. 10Intensive Care Department,Hospital Son Llatzer, Palma de Mallorca, Spain. 11Intensive Care Department,Hospital Universitario Rio Hortega, Valladolid, Spain. 12Intensive Caredepartment, Hospital Universitario “Infanta Leonor”, Madrid, Spain.
Received: 10 January 2018 Accepted: 8 June 2018
References1. Martin GS. Sepsis, severe sepsis and septic shock: changes in incidence,
pathogens and outcomes. Expert Rev Anti-Infect Ther. 2012;10:701–6.2. Gaieski DF, Edwards JM, Kallan MJ, Carr BG. Benchmarking the incidence
and mortality of severe sepsis in the United States. Crit Care Med. 2013;41(5):1167–74.
3. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis inthe United States from 1979 through 2000. N Engl J Med. 2003;348:1546–54.
4. Brun-Buisson C. The epidemiology of the systemic inflammatory response.Intensive Care Med. 2000;26(Suppl1):S64–74.
5. Mayr FB, Yende S, Angus DC. Epidemiology of severe sepsis. Virulence. 2014;5:4–11.
6. Levy MM, Dellinger RP, Townsend SR, et al. The surviving Sepsis campaign:results of an international guideline-based performance improvementprogram targeting severe sepsis. Intensive Care Med. 2010;36:222–31.
7. Ferrer R, Artigas A, Levy MM, et al. Improvement in process of care andoutcome after a multicenter severe sepsis educational program in Spain.JAMA. 2008;299:2294–303.
8. Bochud P-Y, Bonten M, Marchetti O, Calandra T. Antimicrobial therapy forpatients with severe sepsis and septic shock: an evidence-based review. CritCare Med. 2004;32:S495–512.
9. Siddiqui S, Razzak J. Early versus late pre-intensive care unit admission broadspectrum antibiotics for severe sepsis in adults. Cochrane Database Syst Rev.2010;10:CD007081. https://doi.org/10.1002/14651858.CD007081.pub2.
10. Dellinger RP, Levy MM, Rhodes A, et al. Surviving Sepsis campaign:international guidelines for management of severe sepsis and septic shock.Crit Care Med. 2013;41:580–637.
11. Ferrer R, Artigas A, Suarez D, et al. Effectiveness of treatments for severesepsis: a prospective, multicenter, observational study. Am J Respir Crit CareMed. 2009;180:861–6.
12. Suberviola Cañas B, Jáuregui R, Ballesteros MÁ, Leizaola O, González-CastroA, Castellanos-Ortega Á. Effects of antibiotic administration delay andinadequacy upon the survival of septic shock patients. Med Int. 2015;39(8):459–66.
13. Kumar G, Ellis P, Arabi Y, et al. Initiation of inappropriate microbial therapyresults in a fivefold reduction of survival in human septic shock. Chest. 2009;136:1237–48.
14. Kumar A, Roberts D, Wood KE, et al. Duration of hypotension beforeinitiation of effective antimicrobial therapy is the critical determinant ofsurvival in human septic shock. Crit Care Med. 2006;34:1589–96.
15. Ferrer R, Martin-Loeches I, Phillips G, et al. Empiric antibiotic treatmentreduces mortality in severe sepsis and septic shock from the first hour:results from a guideline-based performance improvement program. CritCare Med. 2014;42:1749–55.
16. Ulldemolins M, Nuvials X, Palomar M, et al. Appropriateness is critical. CritCare Clin. 2011;27:35–51.
17. Masterton RG. Antibiotic de-escalation. Crit Care Clin. 2011;27:149–62.18. Dryden M, Johnson AP, Ashiru-Oredope D, Sharland M. Using antibiotics
responsibly: right drug, right time, right dose, right duration. J AntimicrobChemother. 2011;66:2441–3.
19. Levy MM, Fink MP, et al. For the international Sepsis definitions conference.2001 SCCM/ESICM/ACCP/ATS/SIS international Sepsis definitions conference.Intensive Care Med. 2003;29:530–8.
20. Scales DC, Dainty K, Hales B, et al. A multifaceted intervention for qualityimprovement in a network of intensive care units: a cluster randomizedtrial. JAMA. 2011;305:363.
21. Groth ME, Chan-Tompkins NH, Hegde GG, et al. Delays in the administrationof antimicrobials in the emergency department and the impact of aneducational intervention to improve this administration. Eur J Emerg Med.2013;20:339–43.
22. Houck PM, Bratzler DW. Administration of first hospital antibiotics forcommunity-acquired pneumonia: does timeliness affect outcomes? CurrOpin Infect Dis. 2005;18:151–6.
23. Capelastegui A, España PP, Quintana JM, et al. Improvement of process-of-care and outcomes after implementing a guideline for the management ofcommunity-acquired pneumonia: a controlled before-and-after designstudy. Clin Infect Dis. 2004;39:955–63.
24. Herran-Monge R, Muriel-Bombín A, García-García MM, Merino-García PA,Cítores-Gonzalez R, Fernandez-Ratero JA, Albala N, Carriedo D, Moradillo-Gonzalez S, Alvarez-Martínez B, Macías S, María J, Calavia F, Tarancon C, VillarJ, Blanco J, On behalf of the GRECIA network. Mortality reduction and long-term compliance with surviving sepsis campaign: a nationwide multicenterstudy. Shock. 2016;45:598–606.
25. Ramsay C, Brown E, Hartman G, et al. Room for improvement: a systematicreview of the quality of evaluations of interventions to improve hospitalantibiotic prescribing. J Antimicrob Chemother. 2003;52:764–71.
26. Grimshaw JM, Shirran L, Thomas R, et al. Changing provider behavior: anoverview of systematic reviews of interventions. Med Care. 2001;39(8 Suppl2):II2–45.27.
27. Needham DM. Patient safety, quality of care, and knowledge translation inthe intensive care unit. Respir Care. 2010;55:922–8.
28. Evans KH, Daines W, Tsui J, et al. Septris: a novel, mobile, online, simulationgame that improves sepsis recognition and management. Acad Med. 2015;90:180–4.
29. Akl EA, Kairouz VF, Sackett KM, et al. Educational games for healthprofessionals. Cochrane Database Syst Rev. 2013;1:CD006411. https://doi.org/10.1002/14651858.CD006411.pub3.
30. Weiner SG, Brown SF, Goetz JD, Webber CA. Weekly e-mail remindersinfluence emergency physician behavior: a case study using the joint
Ferrer et al. Critical Care (2018) 22:167 Page 9 of 10
commission and Centers for Medicare and Medicaid Services pneumoniaguidelines. Acad Emerg Med. 2009;16:626–31.
31. Sinuff T, Cook D, Giacomini M, et al. Facilitating clinician adherence toguidelines in the intensive care unit: a multicenter, qualitative study. CritCare Med. 2007;35:2083–9.
32. Seymour CW, Gestin F, Prescott HC, et al. Time to treatment and mortalityduring mandated emergency care for sepsis. N Engl J Med. 2017;376(23):2235–44.
33. Liu VX, Fielding-Singh V, Greene JD, et al. The timing of early antibiotics andhospital mortality in sepsis. Am J Respir Crit Care Med. 2017;196(7):856–63.
34. Whiles BB, Deis AS, Simpson SQ. Increased time to initial antimicrobialadministration is associated with progression to septic shock in severesepsis patients. Crit Care Med. 2017;45(4):623–9.
35. Sterling SA, Miller WR, Pryor J, et al. The impact of timing of antibiotics onoutcomes in severe sepsis and septic shock: a systematic review and meta-analysis. Crit Care Med. 2015;43:1907–15.
36. Bloos F, Rüddel H, Thomas-Rüddel D, et al. Effect of a multifacetededucational intervention for anti-infectious measures on sepsis mortality: acluster randomized trial. Intensive Care Med. 2017; https://doi.org/10.1007/s00134-017-4782-4. [Epub ahead of print]
37. Esteban E, Belda S, García-Soler P, et al. A multifaceted educational interventionshortened time to antibiotic administration in children with severe sepsis andseptic shock: ABISS Edusepsis pediatric study. Intensive Care Med. 2017; https://doi.org/10.1007/s00134-017-4721-4. [Epub ahead of print]
38. Singer M. Antibiotics for sepsis: does each hour really count, or is itincestuous amplification? Am J Respir Crit Care Med. 2017;196(7):800–2.
39. Garnacho-Montero J, Gutiérrez-Pizarraya A, Escoresca-Ortega A, et al. De-escalation of empirical therapy is associated with lower mortality in patientswith severe sepsis and septic shock. Intensive Care Med. 2014;40(1):32–40.
40. Leone M, Bourgoin A, Cambon S, Dubuc M, Albanèse J, Martin C. Empiricalantimicrobial therapy of septic shock patients: adequacy and impact on theoutcome. Crit Care Med. 2002;31:462–7.
41. Lagu T. Translating knowledge into practice in critical care settings. Crit CareMed. 2013;41:2659–60.
42. Levy MM, Pronovost PJ, Dellinger RP, et al. Sepsis change bundles:converting guidelines into meaningful change in behavior and clinicaloutcome. Crit Care Med. 2004;32:S595–7.
43. Wagner AK, Soumerai SB, Zhang F, et al. Segmented regression analysis ofinterrupted time series studies in medication use research. J Clin PharmTher. 2002;27:299–309.
44. Sánchez B, et al. Declining mortality due to severe sepsis and septic shockin Spanish intensive care units: a two-cohort study in 2005 and 2011. MedInt. 2017;41(1):28–37.
Ferrer et al. Critical Care (2018) 22:167 Page 10 of 10
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Critical Care Medicine www.ccmjournal.org 1
Objectives: Time to clearance of pathogens is probably critical to out-come in septic shock. Current guidelines recommend intervention for source control within 12 hours after diagnosis. We aimed to deter-mine the epidemiology of source control in the management of sepsis and to analyze the impact of timing to source control on mortality.
Design: Prospective observational analysis of the Antibiotic Inter-vention in Severe Sepsis study, a Spanish national multicenter educational intervention to improve antibiotherapy in sepsis.Setting: Ninety-nine medical-surgical ICUs in Spain.Patients: We enrolled 3,663 patients with severe sepsis or septic shock during three 4-month periods between 2011 and 2013.Interventions: Source control and hospital mortality.Measurements and Main Results: A total of 1,173 patients (32%) underwent source control, predominantly for abdominal, urinary, and soft-tissue infections. Compared with patients who did not require source control, patients who underwent source control were older, with a greater prevalence of shock, major organ dys-function, bacteremia, inflammatory markers, and lactic acidemia. In addition, compliance with the resuscitation bundle was worse in those undergoing source control. In patients who underwent source control, crude ICU mortality was lower (21.2% vs 25.1%; p = 0.010); after adjustment for confounding factors, hospital mortal-ity was also lower (odds ratio, 0.809 [95% CI, 0.658–0.994]; p = 0.044). In this observational database analysis, source control after 12 hours was not associated with higher mortality (27.6% vs 26.8%; p = 0.789).Conclusions: Despite greater severity and worse compliance with resuscitation bundles, mortality was lower in septic patients who underwent source control than in those who did not. The time to source control could not be linked to survival in this observational database. (Crit Care Med 2016; XX:00–00)Key Words: critical care; infection control; mortality; sepsis; septic shock; severe sepsis
Sepsis is an inflammatory response to severe infection with organ dysfunction (1). Infection initiates cytokine release, leading to a global inflammatory cascade. Under
the recent hypothesis that bacterial load is the primary driver of septic organ dysfunction, the rapid clearance of pathogens is the central determinant of outcome in septic shock (2), and early appropriate antimicrobial therapy and source control are key to sepsis management.
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
DOI: 10.1097/CCM.0000000000002011
1Intensive Care Department, Sabadell Hospital, Instituto Universitario Parc Taulí, Autonomous University of Barcelona, Sabadell, Spain.
2Intensive Care Department, Vall d’Hebron University Hospital, Autono-mous University of Barcelona, Barcelona, Spain.
3CIBER-Enfermedades Respiratorias, Barcelona, Spain.4Epidemiology and Assessment Unit, Fundació Parc Taulí, Autonomous University of Barcelona, Sabadell, Spain.
5Intensive Care Department, A Coruña University Hospital, A Coruña, Spain.
6Intensive Care Department, Reina Sofía de Córdoba University Hospital, Córdoba, Spain.
7Multidisciplinary Intensive Care Research Organization (MICRO), St James’s University Hospital, Trinity Centre for Health Sciences, Dublin, Ireland.
8Medical Intensive Care Unit, Rhode Island Hospital, Brown University School of Medicine, Providence, RI.
A list of Edusepsis Study Group is provided in the ACKNOWLEDGMENT section.
Supplemental digital content is available for this article. Direct URL cita-tions appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal).
Supported, in part, by research grant from Instituto de Salud Carlos III (FIS 10/01497) (CM12/00066).
Dr. Martinez’s institution received grant support (research grant from Insti-tuto de Salud Carlos III [FIS 10/01497] [partial payment]). Dr. Ferrer served as a board member for Laboratorios Ferrer, Abot, and Grifols and lectured for Grifols. His institution received grant support from Instituto Carlos III. Dr. Gomà’s institution received grant support from Gemma Goma. Dr. Álva-rez-Rocha lectured for Pfizer. Dr. Levy’s institution received grant support from ImmuneExpress (study on sepsis on the medical wards). Dr. Artigas’ institution received grant support research grant from Instituto de Salud Carlos III [FIS 10/1497] [partial payment]). The remaining authors have disclosed that they do not have any potential conflicts of interest.
For information regarding this article, E-mail: [email protected]
Impact of Source Control in Patients With Severe Sepsis and Septic Shock
María Luisa Martínez, MD1; Ricard Ferrer, MD, PhD2,3; Eva Torrents, MD1;
Raquel Guillamat-Prats, PhD3; Gemma Gomà, RN1; David Suárez, MSc, PhD4;
Luis Álvarez-Rocha, MD5; Juan Carlos Pozo Laderas, MD, PhD6; Ignacio Martín-Loeches, MD, PhD7;
Mitchell M. Levy, MD, FCCP, FCCM8; Antonio Artigas, MD, PhD1,3; for the Edusepsis Study Group
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Martínez et al
2 www.ccmjournal.org XXX2016•VolumeXX•NumberXXX
Source control comprises “all physical measures to eliminate sources of infection, to control contamination, and to restore anatomy and function” (3). It includes draining infected flu-ids, debriding infected soft tissues, removing infected devices or foreign bodies, and correcting anatomic derangement caus-ing microbial contamination. Source control’s effectiveness depends on the infection site, the patient’s premorbid state, and the resources available (4).
The Surviving Sepsis Campaign (SSC) recommends all patients with severe sepsis or septic shock be evaluated as soon as possible for specific infection sites amenable to source con-trol and undergo source control within 12 hours after diag-nosis (grade 1C, SSC 2012) (5). However, source control has received less attention than other treatments in the SSC. In addition, although studies show that early compliance with the SSC bundles is associated with lower mortality (6–8), evi-dence regarding the impact of the timing of source control in patients with severe sepsis or septic shock is lacking.
The impact of source control in septic patients is not fully understood. We aimed to assess the epidemiology of the need for source control and its role in the management of patients with severe sepsis or septic shock. We hypothesized that delays in source control after onset of severe sepsis or septic shock would worsen outcome.
MATERIALS AND METHODS
Study DesignWe conducted a prospective secondary analysis of the Anti-biotic Intervention in Severe Sepsis study, a Spanish national multicenter educational intervention in 99 medical-surgical ICUs homogeneously distributed throughout Spain. All con-secutive adults with severe sepsis or septic shock admitted to participating ICUs during three 4-month periods (April to July 2011, April to July 2012, and January to April 2013) were eli-gible for the study.
Each center’s ethics committee approved the study and waived the informed consent requirement due to the study’s observational and anonymous nature.
The study design, data collection, and quality control mea-sures are in detailed in Appendix S1 (Supplemental Digital Content 1, http://links.lww.com/CCM/C27). Severe sepsis was defined as sepsis associated with organ dysfunction unex-plained by other causes. Septic shock was defined as sepsis associated with systolic blood pressure less than 90 mm Hg, mean arterial pressure less than 65 mm Hg, or a reduction in systolic blood pressure more than 40 mm Hg from baseline despite adequate volume resuscitation (9).
Process of Care and Outcome MeasurementsWe recorded demographic characteristics (age and sex), Charl-son comorbidity score, presence of shock, diagnosis at admis-sion (medical, emergency surgery, and elective surgery), site of infection, type of infection (community acquired, healthcare related, hospital acquired, or ICU acquired), organ dysfunction at sepsis presentation, worst value of inflammatory markers
(C-reactive protein and procalcitonin) in the first 24 hours of sepsis onset, presence of bacteremia, need for source control, and worst Acute Physiology and Chronic Health Evaluation II score during the first 24 hours in the ICU.
Attending physicians decided when specific percutaneous or surgical source control were necessary. Appendix S2 (Supplemental Digital Content 2, http://links.lww.com/CCM/C28) describes the source control techniques in detail. We recorded time from onset of severe sepsis or septic shock to source control and divided patients into those who received early (< 12 hr) and late (≥ 12 hr) source control. We also recorded time from onset to other acts and targets prescribed in the SSC guidelines (10): measuring serum lactate, obtaining blood cultures, administering broad-spectrum antibiotics, administering fluids and/or vasopressors in patients with hypotension and/or lactate more than 4 mmol/L (36 mg/dL), and achieving central venous pressure greater than or equal to 8 mm Hg and central venous oxygen saturation greater than or equal to 70%. We also recorded the appropriateness of antibiotic therapy, defined as the administration of an antimicrobial agent with in vitro microbiologic activity against the isolated pathogen. To facilitate antibiotic prescription, researchers used preferentially their local guideline or an electronic clinical decision support sys-tem (http://www.es.dgai-abx.de).
Patients were followed up until death or hospital dis-charge. The primary outcome variable was hospital mortality. Secondary outcome measures included days of mechanical ventilation, days of vasopressors, hospital and ICU lengths of stay, and ICU mortality.
Statistical AnalysisDescriptive statistics included frequencies and percentages for categorical variables and means, sd, CIs, medians, and inter-quartile ranges for continuous variables. To compare contin-uous variables, we used Student t test or the Mann-Whitney U test as appropriate. To analyze categorical variables, we used the chi-square test or Fisher exact test as appropriate.
To assess the impact of source control, we used multi-variate logistic regression with ICU and hospital mortality as the dependent variables and source control, age, sex, Acute Physiology and Chronic Health Evaluation II score, the pres-ence of shock, Charlson comorbidity score, patient location at sepsis diagnosis, site of infection, appropriateness of antibiotic therapy, and compliance with early antibiotic administration and resuscitation with fluids and/or vasopressors as indepen-dent variables.
In the group who underwent source control, we did a multi-variable analysis to assess the impact of time to source control (< 12 vs ≥ 12 hr) on hospital mortality, including the same variables as in the previous model. To better understand the importance of time to source control, we applied the same regression model for abdominal, urologic, and skin and soft-tissue infections.
We also examined time to source control as a continuous variable. To determine the best cutoff time to source control to discriminate mortality, we plotted a receiver operating characteristic curve (ROC) and calculated the area under the curve (AUC).
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigation
Critical Care Medicine www.ccmjournal.org 3
Statistical tests were two tailed with significance defined as p value less than 0.05. We used SPSS version 15.0 (SPSS, Chicago, IL) for all analyses.
RESULTS
Patients Undergoing Source Control Versus Those Who Did NotA total of 3,663 patients met the criteria for severe sepsis or septic shock during the study periods; 1,173 (32%) of these underwent surgical or percutaneous procedures for source control. Table 1 summarizes patients’ demographics, preex-isting medical conditions, and baseline clinical, physiologic, and laboratory variables. Compared with patients who did not require source control, patients undergoing source control were older, and a higher proportion had shock. Coagulopathy, hyperlactatemia, and heart, kidney, and liver failures were more common in patients who required source control. Patients who required source control had more bacteremic episodes and higher levels of C-reactive protein and procalcitonin. The most common source of sepsis was abdominal infection in patients undergoing source control (n = 788; 67.2%) and respira-tory infection in patients who did not require source control (n = 1,189; 47.8%). A greater proportion of patients in the source control group were admitted after urgent surgery and had hospital-acquired infections.
Compliance with three of the six tasks in the 6-hour resus-citation bundle (lactate measurement, blood cultures before antibiotics, and early administration of broad spectrum anti-biotics) was worse in patients requiring source control than in those who did not. Overall compliance with all elements in the resuscitation bundle was also lower in patients requiring source control (Table 2).
Table 3 reports the outcome data for patients who required source control versus those who did not. Patients who required source control needed more days of vasopressor treatment, but no differences were observed in days of mechanical ventilation. The hospital stay was longer in patients who required source control (32.5 vs 27.4 d in those that did not; p < 0.001); the ICU stay was similar in the two groups. ICU mortality was lower in patients requiring source control (21.2% vs 25.1%; p = 0.010), but hospital mortality was similar in the two groups. After adjusting for possible confounders, ICU mortality remained lower in patients who underwent source control (Table S1, Supplemental Digital Content 3, http://links.lww.com/CCM/C29), and hospital mortality was also lower in patients under-going source control (odds ratio, 0.809 [95% CI, 0.658–0.994]; p = 0.044) (Table 4).
Timing of Source ControlA total of 1,173 patients (32%) underwent procedures for source control, and time was recorded in 1,090 of these; thus, 83 patients were excluded because time to source control was unknown. Median time to source control was 4.6 hours (1–11.5 hr). Interventions for source control were done within 12 hours of sepsis onset in 825 patients (75.7%). No significant
differences in demographic or clinical characteristics were found between patients who underwent source control within 12 hours of onset and those who underwent source control later (Table S2, Supplemental Digital Content 4, http://links.lww.com/CCM/C30). Compliance with the items in the 6-hour resuscitation bundle was better in patients undergoing source control within 12 hours of onset than in those undergoing source control later, except blood cultures before antibiotics and early administration of broad-spectrum antibiotics, where no differences were observed (Table S3, Supplemental Digital Content 5, http://links.lww.com/CCM/C31).
Table 5 reports the outcome data for patients who under-went source control within 12 hours of onset versus those who underwent source control later. No significant differences between the two groups were observed in hospital stay, ICU stay, hospital mortality, or ICU mortality.
ROC curves analyzing time to source control as a continu-ous variable failed to identify a point of maximum sensitivity and specificity to predict the optimum time for source control; we observed no relationship between time to source control and mortality in the group of patients who underwent source control (AUC = 0.504, nonsignificant [ns]) (Fig. S1, Supplemental Digital Content 6, http://links.lww.com/CCM/C32; legend, Supplemental Digital Content 17, http://links.lww.com/CCM/C43) or in the subgroups of patients who underwent percutaneous source con-trol (AUC = 537, ns) (Fig. S2, Supplemental Digital Content 7, http://links.lww.com/CCM/C33; legend, Supplemental Digital Content 17, http://links.lww.com/CCM/C43) or surgical source control (AUC = 523, ns) (Fig. S3, Supplemental Digital Content 8, http://links.lww.com/CCM/C34; legend, Supplemental Digital Content 17, http://links.lww.com/CCM/C43).
When we analyzed the outcomes for patients who needed source control in the subgroups with abdominal, urinary, and skin and soft-tissue infections, we found no significant differ-ences between the less than 12 hour and greater than or equal to 12 hour groups, except longer duration of vasopressors in patients with skin and soft-tissue infections (early source control: 4.3 d vs late source control: 8.6 d; p < 0.001) (Table S4, Supplemental Digital Content 9, http://links.lww.com/CCM/C35; Table S5, Supplemental Digital Content 10, http://links.lww.com/CCM/C36; and Table S6, Supplemental Digital Content 11, http://links.lww.com/CCM/C37).
Univariate analysis found no difference in time to source control between survivors and nonsurvivors (Table S7, Supplemental Digital Content 12, http://links.lww.com/CCM/C38). Multivariable logistic regression adjusting for pos-sible confounders showed no relationship between time to source control less than 12 hours and hospital mortality in the group of patients who underwent source control (Table S8, Supplemental Digital Content 13, http://links.lww.com/CCM/C39) or in the subgroups of patients with abdominal, urinary, and skin and soft-tissue infections (Table S9, Supplemental Digital Content 14, http://links.lww.com/CCM/C40; Table S10, Supplemental Digital Content 15, http://links.lww.com/CCM/C41; and Table 11, Supplemental Digital Content 16, http://links.lww.com/CCM/C42).
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Martínez et al
4 www.ccmjournal.org XXX2016•VolumeXX•NumberXXX
TAbLE 1. Demographic and Clinical Characteristics of Patients
Patient CharacteristicAll Patients,
n = 3,663
Patients Not Requiring
Source Control, n = 2,490 (68%)
Patients Requiring Source
Control, n = 1,173 (32%) p
General data
Age (yr), mean (sd) 64 (15.1) 62.8 (15.2) 66.7 (14.6) < 0.001
Sex (male), n (%) 2,319 (63.3) 1,621 (65.1) 698 (59.5) 0.001
Acute Physiology and Chronic Health Evaluation II, mean (sd)
21.8 (8.01) 22.03 (8.2) 21.3 (7.6) 0.010
Shock, n (%) 2,497 (68.2) 1,630 (65.5) 867 (73.9) < 0.001
Charlson comorbidity score, mean (sd) 2.6 (2.3) 2.6 (2.3) 2.7 (2.2) 0.531
C-reactive protein (mg/dL), mean (sd)a 24.2 (13.7) 23.6 (13.9) 25.5 (12.9) < 0.001
Procalcitonin (ng/mL), mean (sd)b 26.2 (37.6) 24.1 (34.7) 31.2 (43) 0.001
Bacteremia, n (%) 1,211 (40.1) 821 (37.9) 390 (45.5) < 0.001
Appropriate antibiotic therapy, n (%) 1,911 (51.9) 1,231 (49.4) 670 (57.1) < 0.001
Organ failure at sepsis presentation, n (%)
No. of organ failures (sd) 2.98 (1.4) 2.98 (1.4) 2.98 (1.4) 0.914
Cardiovascular 3,019 (82.4) 1,994 (80.1) 1,025 (87.4) < 0.001
Respiratory 1,602 (43.7) 1,275 (51.2) 327 (27.9) < 0.001
Renal 2,068 (56.5) 1,351 (54.3) 717 (61.1) < 0.001
Hyperbilirubinemia 606 (16.5) 386 (15.5) 220 (18.8) 0.013
Thrombocytopenia 856 (23.4) 620 (24.9) 236 (20.1) 0.001
Coagulation 1,143 (31.2) 749 (30.1) 394 (33.6) 0.032
Hyperlactatemia 1,630 (44.5) 1,056 (42.4) 574 (48.9) < 0.001
Site of infection, n (%)
Abdominal 1,234 (33.7) 446 (17.9) 788 (67.2) < 0.001
Respiratory 1,232 (33.6) 1,189 (47.8) 43 (3.7)
Urologic 606 (16.5) 459 (18.4) 147 (12.5)
Skin and/or soft tissue 258 (7.0) 140 (5.6) 118 (10.1)
Central nervous system 87 (2.4) 82 (3.3) 5 (0.4)
Other 246 (6.7) 174 (7.0) 72 (6.1)
Type of infection (acquisition site), n (%)
Community acquired 2,285 (62.4) 1,607 (64.5) 678 (57.8) < 0.001
Healthcare related 438 (12) 318 (12.8) 120 (10.2)
Hospital acquired 780 (21.3) 443 (17.8) 337 (28.7)
ICU acquired 160 (4.4) 122 (4.9) 38 (3.2)
Diagnosis at admission, n (%)
Medical 2,567 (70.1) 2,256 (90.6) 311 (26.5) < 0.001
Surgical 205 (5.6) 109 (4.4) 96 (8.2)
Urgent surgical 891 (24.3) 125 (5) 766 (65.3)a n = 2,763 patients.b n = 2,084 patients.
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigation
Critical Care Medicine www.ccmjournal.org 5
DISCUSSIONThis prospective observational study in patients with severe sepsis and septic shock found significantly lower mortality, even after adjustment for confounding factors, in patients who underwent measures for source control than in those who did not, despite the greater risk of death in patients who under-went source control. These findings underline the importance of source control in the management of patients with severe sepsis or septic shock.
In this large population of patients with severe sepsis or septic shock, one third of all patients underwent source con-trol. Another recent prospective observational multicenter study including 1,011 patients with severe sepsis or septic shock reported that 41.7% underwent source control (11). The characteristics of the patients and most importantly the site of infection in this study were similar to those in our study (12).
Compared with patients who did not require source con-trol, a higher proportion of patients who underwent source control presented shock and major organ dysfunction, prob-ably because of higher bacterial load as suggested by the higher rates of bacteremia and higher levels of procalcitonin and C-reactive protein biomarkers that correlate with the inflam-matory response. Serum procalcitonin increases with the severity of sepsis and organ dysfunction (13–16). A greater
proportion of patients in the group who underwent source control had undergone surgery, and manipulation of a focus with a high microbial load might explain the higher rates of bacteremia and more severe inflammatory response (3). In addition, patients who underwent source control were older and a greater proportion had abdominal and nosocomial infec-tions, factors independently associated with mortality (12, 17). Although patients who required source control received more appropriate antibiotherapy, the time to first antibiotic admin-istration was longer in this group, and delays in antibiotic administration over the first 6 hours after sepsis identification are associated with increased mortality (7). Furthermore, com-pliance with two other tasks in the resuscitation bundle (blood cultures before antibiotic administration and lactate measure-ment) was worse in patients who required source control. It is unclear why there should be differences in compliance based on the use of source control, but delays might be because of prioritizing source control when needed.
Thus, although patients who underwent source control were at a greater risk than those who did not, they had lower mortality even after adjustment for confounding factors. These findings strongly support the importance of source control and lend weight to the new paradigm proposed by Kumar et al (2) to explain the pathophysiology of sepsis, where microbiologic
TAbLE 2. Compliance With Sepsis Resuscitation bundle in Source Control Group Versus Nonsource Control Group
Sepsis Resuscitation bundle, 6 h, n (%)
All Patients, n = 3,663
Patients Not Requiring Source Control,
n = 2,490
Patients Requiring Source Control,
n = 1,173 p
All resuscitation measures 379 (10.3) 279 (11.2) 100 (8.5) 0.013
Measure lactate 2,709 (74.0) 1,876 (75.3) 833 (71.0) 0.005
Blood cultures before antibiotics 1,906 (52.0) 1,376 (55.3) 530 (45.2) < 0.001
Early broad-spectrum antibiotics 2,550 (69.6) 1,763 (70.8) 787 (67.1) 0.023
Fluids and vasopressors 2,152 (58.7) 1,452 (58.3) 700 (59.7) 0.434
Central venous pressure, ≥ 8 mm Hg 1,609 (43.9) 1,113 (44.7) 496 (42.3) 0.170
Central venous oxygen saturation, ≥ 70% 1,236 (33.7) 849 (34.1) 387 (33.0) 0.510
TAbLE 3. Outcome Measurements in Source Control Group Versus Nonsource Control Group
Outcome MeasurementsAll Patients,
n = 3,663
Patients Not Requiring Source Control,
n = 2,490
Patients Requiring Source Control,
n = 1,173 p
Duration of mechanical ventilation, d, mean (sd) 6.88 (13.2) 6.78 (13.0) 7.11 (13.6) 0.480
Duration of vasopressors, d, mean (sd) 4.26 (7.2) 4.01 (6.6) 4.8 (8.4) 0.002
ICU stay, d, mean (sd) 11.8 (15.4) 11.6 (15.03) 12.3 (16.02) 0.202
Hospital stay, d, mean (sd) 29.04 (28.6) 27.4 (27.8) 32.5 (30.1) < 0.001
Mortality, n (%)
ICU 875 (23.9) 626 (25.1) 249 (21.2) 0.010
Hospital 1,088 (29.7) 756 (30.4) 332 (28.3) 0.203
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Martínez et al
6 www.ccmjournal.org XXX2016•VolumeXX•NumberXXX
load is the main driver of septic shock and rapid clearance of pathogens is central to outcome. This paradigm incorporates the concept of irreversible shock and suggests that the best approach to treatment is to minimize the time when a num-ber of microorganisms sufficient to generate shock are pres-ent. Thus, early potent antimicrobial therapy and adequate source control are key components in sepsis management. Both crude and adjusted ICU mortality rates were lower in
patients with source control; however, the decrease in hospital mortality was not evident until we adjusted for confounding factors. The hospital stay was longer in patients who required source control. One reason for this difference could be that abdominal infections were the most common source of sepsis in the source control group, and the definitive management of abdominal infections often requires more than one source control intervention, increasing morbidity, hospital stays, and
TAbLE 4. Multivariate Analysis of Risk Factors for Hospital Mortality in All Patients (n = 3,663)
Factors OR 95% CI p
Source controla 0.809 0.658–0.994 0.044
Ageb 1.017 1.011–1.023 < 0.001
Sex, femalec 1.131 0.955–1.340 0.154
Acute Physiology and Chronic Health Evaluation IIb 1.102 1.090–1.115 < 0.001
Septic shockd 1.338 1.108–1.616 0.002
Charlson comorbidity scoreb 1.067 1.029–1.106 < 0.001
Early broad-spectrum antibiotics 0.804 0.672–0.963 0.018
Fluids and vasopressors 0.863 0.733–1.014 0.074
Appropriate antibiotic therapye 0.710 0.538–0.938 0.016
Nosocomial acquired infectionf 1.971 1.610–2.414 < 0.001
Site of infectiong
Abdominal 0.952 0.758–1.196 0.671
Urologic 0.289 0.215–0.388 < 0.001
Central nervous system 1.179 0.683–2.036 0.554
Skin and soft-tissue 0.935 0.667–1.311 0.696
Others 1.073 0.779–1.478 0.664a Compared with patients not requiring control.b Per each point increase.c Compared with male sex.d Compared with severe sepsis.e Compared with inappropriate antibiotic therapy.f Compared with the emergency department.g Compared with respiratory infection.
TAbLE 5. Outcome Measurements in the Source Control Group
Outcome Measurements
All Patients Receiving
Source Control, n = 1,090
Patients Receiving Source
Control < 12 hr, n = 825
Patients Receiving Source
Control ≥ 12 hr, n = 265 p
Duration of mechanical ventilation, d, mean (sd) 7.1 (13.1) 7.1 (12.9) 7.1 (13.9) 0.995
Duration of vasopressors, d, mean (sd) 4.8 (8.1) 4.6 (7.5) 5.4 (9.7) 0.168
ICU stay, d, mean (SD) 12.2 (15.3) 12.1 (15.2) 12.6 (15.4) 0.518
Hospital stay, days mean (SD) 32.3 (31.3) 31.9 (29.7) 31.6 (28.5) 0.884
Mortality, n (%)
ICU 226 (20.7) 172 (20.8) 54 (20.4) 0.869
Hospital 299 (27.4) 228 (27.6) 71 (26.8) 0.789
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigation
Critical Care Medicine www.ccmjournal.org 7
hospital mortality (18, 19). By contrast, in patients who did not require source control, the predominant source of sepsis was pneumonia, a less drainable focus. The lack of a drainable focus seems to be associated with worse outcome. One retro-spective study that reviewed macroscopic findings in autopsies of 235 surgical ICU patients who died of sepsis or septic shock found a septic focus in approximately 80%, suggesting that the need for source control may be under recognized (20).
Although source control is essential to the successful man-agement of severe sepsis and septic shock, in this observational database, we could not demonstrate that source control was time dependent. The SSC guidelines recommend that source control be undertaken within 12 hours of diagnosis (5), up from 6 hours in the previous guidelines. This increase was based on a recent retrospective study in 106 patients with sep-tic shock and necrotizing soft-tissue infections where a delay of surgery more than 14 hours was independently associated with hospital mortality, but that study did not analyze other cut-off times (21). Although it is reasonable to assume that rapid source control is essential to maximize survival in severely septic patients with acute physiologic deterioration, scant evi-dence supports this approach. Only one study in patients with severe sepsis and septic shock showed a reduction (16%) in 28-day mortality when source control was performed within the first 6 hours, and this study only analyzed 234 of the 488 patients who needed source control (11). Several studies demonstrate the importance of early source control in necro-tizing infections, but the definition of early source control var-ied between 2 and 24 hours (22–24). Another study in a large population of patients with fecal peritonitis found that early source control was not associated with better outcome (25). In our population, patients who received early source control also received better early resuscitation, suggesting that these patients might have been sicker; however, we found no sig-nificant differences in baseline characteristics between patients who received early source control and those who received late source control. Yet, despite better early management, the mor-tality for patients receiving early source control was similar to those receiving late source control. The most likely explanation is that the clinical team considered source control more urgent in patients who underwent earlier source control and that the multivariate analysis failed to measure this confounder. There are at least three reasons for delaying source control in severely septic patients: 1) small foci of infection might not be clini-cally evident at first; 2) physicians aware of the need for source control might delay intervention in apparently stable patients to enable nonemergency source control; or 3) surgical inter-vention might be deferred to allow necrosis to define itself anatomically to optimize intervention (e.g., in necrotizing pancreatitis) (26, 27). Determining the impact of early versus late source control would require formal randomization and prospective trials in more homogenous populations of patients and specific sources of infection (28).
This study has several limitations. We analyzed only patients who required admission to the ICU, possibly intro-ducing a selection bias where patients with very early source
control improve enough to avoid ICU admission. Although we adjusted for a number of predisposing patient factors, there may be other confounding factors that we did not measure. In addition, despite adherence to SSC guidelines as far as pos-sible, many aspects of source control that we could not control (e.g., adequate preprocedure resuscitation) affect the outcome. Furthermore, we did not evaluate the type or the success of the source control measure; the specific source control technique and the technical success of the intervention can influence outcomes. A recent study in 44 ICUs found inadequate source control in 13.3% of patients with severe sepsis and septic shock (11), but percentages could be higher in necrotizing soft-tissue or abdominal infections. One study found that 64% of patients with necrotizing soft-tissue infections required more than one debridement (29), and another found that inadequate debride-ment was associated with increased mortality (22). Others found that failure to control the septic source in abdominal infections and the method used for source control affected outcomes (18, 30). Finally, this secondary analysis of an obser-vational study has the weaknesses inherent to observational studies.
Our study has also strengths. We prospectively enrolled a large cohort of patients with severe sepsis and septic shock in a short time and monitored them until death or hospital dis-charge, resulting in a homogeneous database with high-quality control measures to ensure validity. Furthermore, the large number of ICUs participating means that the results can be extrapolated.
CONCLUSIONSOne third of patients with sepsis admitted to the ICU needed source control, especially those with abdominal and soft-tissue infections. Although patients who underwent source control were more severe and received worse initial resuscitation, their outcomes were better than those who did not undergo source control; these findings underline the importance of source control in the management of patients with severe sepsis or septic shock. We failed to demonstrate lower mortality for early source control versus late source control, but there is no ratio-nale to defer source control in severe patients. Well-designed clinical trials including all patients with severe sepsis and septic shock, not just those admitted to the ICU, should examine the effects of early versus later source control in specific infectious foci. Educational and quality control programs are required to identify and control infectious foci in patients with severe sep-sis and septic shock.
ACKNOWLEDGMENTWe gratefully acknowledge John Giba for his assistance in the preparation of the article. The Edusepsis Study Group included María Luisa Martínez, Antonio Artigas (Hospital de Sabadell, Consorci Hospitalari Parc Taulí), María del Mar Cruz (Hospital Virgen de la Salud, Toledo), Sandra Barbadillo (Hospital Gen-eral de Cataluña), Francisco Fernández (Centro Clínico Delfos), Alberto Pensado Castiñeiras; María Teresa Rey Rilo, Luis Alvarez
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Martínez et al
8 www.ccmjournal.org XXX2016•VolumeXX•NumberXXX
Rocha (Complexo Hospitalario Universitario de A Coruña), Belén Jiménez Bartolomé (Hospital Clínico Universitario Loz-ano Blesa Zaragoza), Juan Diego Jiménez Delgado (Hospital Comarcal Don Benito-Villanueva, Extremadura), Demetrio Carriedo Ule, Ana María Domínguez Berro, Francisco Javier Díaz Domínguez (Complejo Asistencial Universitario de León), Juan Machado Casas (Complejo Hospitalario de Jaén), Clara Laplaza Santos; Manuel García-Montesinos; Enrique Maraví Poma (Complejo Hospitalario de Navarra - Pamplona), Víc-tor López Ciudad; Pablo Vidal Cortés (Complejo Hospitalario de Ourense), Miguel Martínez Barrio, María Jesús López Pueyo (Hospital General Yagüe, Burgos), María Jesús López Cambra (Hospital general de Segovia), Pau Torrabadella, Álvaro Salcedo, Claudio Durán (Hospital Universitari Germans Trias i Pujol), Iratxe Seijas (Hospital de Cruces de Bilbao), Teresa Recio Gómez (Hospital San Pedro de Alcántara, Cáceres), Ángel Arenzana (Hospital Virgen de la Macarena, Sevilla), Izaskun Azkarate (Hospital de Donostia), Sandra Rodríguez Bolaño (Hospital de Baza. Granada), Eduardo Palencia Herrejón (Hospital Gregorio Marañón Madrid), Jordi Solé Violán (Hospital Universitario de Gran Canaria Dr Negrín), Gerardo Aguilar Aguilar (Hospital Clínic Universitari Valencia), Ángel Rodríguez Rencinas, Marta Paz Pérez, Elena Pérez Losada (Hospital de Salamanca), Fer-nando Martínez Sagasti (Hospital Clínico San Carlos, Madrid), José Luis García Allut (Hospital Clínico Universitario de San-tiago de Compostela), Fernando Díez Gutiérrez, Francisco Gandía, Amanda Francisco Amador (Hospital Clínico Universi-tario de Valladolid), Ramón Vegas Pinto (Hospital de Antequera, Málaga), Pilar Martínez Trivez (Hospital de Barbastro Huesca), Nieves García Vázquez (Hospital Universitario de La Princesa), Luis Zapata, Paula Vera (Hospital de la Santa Creu i Sant Pau), Eduardo Antón (Hospital de Manacor), Juan Carlos Yébenes (Hospital de Mataró), María de las Olas Cerezo Arias (Hospital de Mérida), Francisco García delgado (Hospital de Montilla), Javier Fierro Rosón, Josefa Peinado Rodríguez (Hospital de Poni-ente), María Álvarez (Hospital de Terrassa), Paco Álvarez Lerma (Hospital del Mar), Francisco Valenzuela (Hospital de Jerez), Patricia Albert de la Cruz (Hospital del Sureste), Rafael Blancas Gómez-Casero (Hospital del Tajo, Madrid), Montserrat Sisón Heredia (Hospital Dr. José Molina Orosa, Las Palmas), Perico Olaechea, Celia Sañudo (Hospital de Galdakao-Usansolo), José Manuel Gutiérrez Rubio (Complejo Hospitalario Universitario de Albacete), Roberto Reig Valero (Hospital General de Castel-lón), Hasania Abdel-Hadi Álvarez (Hospital General de Ciudad Real), Leandro Fajardo Feo (Hospital General de Fuerteven-tura), Pau Garro (Hospital General de Granollers), Francisco Navarro Pellejero (Hospital General de la Defensa en Zaragoza), Ana Esther Trujillo Alonso (Hospital general de La Palma), Rosa Catalán (Hospital General de Vic), Assumpta Rovira, Nicolás Rico (Hospital General Hospitalet de LLobregat), José Manuel Allegue Gallego (Hospital General Universitario Santa Lucía Cartagena), José Córdoba Alonso, Dolores Ocaña (Hos-pital La Inmaculada de Huercal-Overa. Almería), Juan Mora Ordóñez, Manuel Salido Mota (Hospital Regional Universi-tario Carlos Haya Málaga), María José Tolón Herrera, Paloma Dorado (Hospital Royo Villanova de Zaragoza), Arantxa Lander
Azcona (Hospital San Jorge Huesca), Diego Mendoza (Hospi-tal Sant Joan Despí Moisès Broggi), Francisca Prieto (Hospi-tal Sta. Bárbara Puertollano), María Carmen Ramagge Martín (Hospital de La Línea), José Ignacio Ayestarán Rota (Hospital Son Espases), Marcio Borges (Hospital de Son Llatzer), Enrique Piacentini, Ricard Ferrer (Hospital Mútua de Terrassa), Josep Maria Sirvent, Cristina Murcia, Gina Rognoni (Hospital Univer-sitari Dr. Josep Trueta de Girona), José Antonio Gonzalo, Diego Parra Ruiz, Natalia Bretón Díez, José Ignacio Argüelles Antuña (Hospital Universitario Central de Asturias), Leonardo Lorente Ramos (Hospital Uni Canarias), Helena Yáñez (Hospital Uni-versitario de Guadalajara), Ana Loza (Hospital Universitario de Valme), Borja Suberbiola (Hospital Universitario Marqués de Valdecilla), Domingo Ruiz de la Cuesta Martín (Hospital Universitario, Miguel Servet Zaragoza), María del Mar Martín Velasco (Hospital Universitario Nuestra Señora de Candelaria), Antonio Pontes Moreno, Rafael León López, Juan Carlos Pozo (Hospital Universitario Reina Sofía de Córdoba), Luis Tamayo Lomas, Jesús Blanco, Arturo Muriel, José Ángel Berezo (Hos-pital Universitario Río Hortega), Paula Ramírez, Miguel Ángel Chiveli Monleón (Hospital Universitario y Politécnico La Fe de Valencia), Juan Carlos Ruiz Rodríguez, Jesús Caballero, Adolf Ruiz, Alejandra García, Jordi Riera, Javier Sarrapio, Francesc Sanpedro, José Carlos Martin, Tatiana Acero (Hospital Univer-sitari Vall Hebrón), Ana Carolina Caballero, Silvia María Cortés Díaz (Hospital Virgen de la Concha, Zamora), María Victoria de la Torre, Begoña Mora Ordóñez (Hospital Virgen de la Victo-ria), José Garnacho Montero (Hospital Virgen del Rocío), Edu-ardo Palencia Herrejón (Hospital Infanta Leonor), Gumersindo González-Díaz; Andrés Carrillo (Hospital Morales Meseguer), Pedro Jesús Domínguez García (Hospital Juan Ramón Jiménez), Ruth Jorge García (Hospital Nuestra Señora de Gracia Zara-goza), Almudena Simón (Hospital Nuestra Señora del Prado), José Carlos Torralba Allué (Hospital General Obispo Polanco Teruel), Teresa Recio Gómez (Hospital San Pedro de Alcántara, Cáceres), Ricardo Díaz Abad (Hospital Severo Ochoa, Madrid), Mar Gobernado (Hospital Santa Bárbara, Soria), Francisco Guerrero Gómez (Hospital Torrecárdenas), José Castaño Pérez (Hospital Virgen de las Nieves), Fernando Bueno Andrés (Hos-pital Virgen del Puerto, Plasencia), Elena Bustamante Munguira, Gaspar Tuero (Hospital de Can Misses), José Francisco Olea Parejo (Hospital Lucus Augusti-Lugo), Miguel Soto, Susana San-cho Chinesta, Rafa Zaragoza (Hospital Universitario Dr. Peset de Valencia), Carmen Fernández González (Complejo Hospi-talario de Ferrol Arquitecto Marcide), Manuel Castellano (Hos-pital Alto Guadalquivir), José María Bonell (Hospital Clínica Palma Planas), María Jesús Broch Porcar (Hospital de Sagunto), Néstor Bacelar (Clínica Corachan), Isabel Cremades (Hospital Reina Sofía de Murcia), Miguel Valdivia, Pedro Galdós (Hospital Puerta del Hierro).
REFERENCES 1. Vincent JL, Opal SM, Marshall JC, et al: Sepsis definitions: Time for
change. Lancet 2013; 381:774–775 2. Kumar A: An alternate pathophysiologic paradigm of sepsis and sep-
tic shock: Implications for optimizing antimicrobial therapy. Virulence 2014; 5:80–97
Copyright © 2016 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigation
Critical Care Medicine www.ccmjournal.org 9
3. Marshall JC, al Naqbi A: Principles of source control in the manage-ment of sepsis. Crit Care Clin 2009; 25:753–68, viii
4. Marshall JC, Maier RV, Jimenez M, et al: Source control in the man-agement of severe sepsis and septic shock: An evidence-based review. Crit Care Med 2004; 32:S513–S526
5. Dellinger RP, Levy MM, Rhodes A, et al; Surviving Sepsis Campaign Guidelines Committee including the Pediatric Subgroup: Surviving sepsis campaign: International guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med 2013; 41:580–637
6. Ferrer R, Artigas A, Levy MM, et al; Edusepsis Study Group: Improve-ment in process of care and outcome after a multicenter severe sep-sis educational program in Spain. JAMA 2008; 299:2294–2303
7. Ferrer R, Martin-Loeches I, Phillips G, et al: Empiric antibiotic treat-ment reduces mortality in severe sepsis and septic shock from the first hour: Results from a guideline-based performance improvement program. Crit Care Med 2014; 42:1749–1755
8. Barochia AV, Cui X, Vitberg D, et al: Bundled care for septic shock: An analysis of clinical trials. Crit Care Med 2010; 38:668–678
9. Levy MM, Fink MP, Marshall JC, et al: 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med 2003; 31:1250–1256
10. Dellinger RP, Levy MM, Carlet JM, et al; International Surviving Sep-sis Campaign Guidelines Committee; American Association of Crit-ical-Care Nurses; American College of Chest Physicians; American College of Emergency Physicians; Canadian Critical Care Society; European Society of Clinical Microbiology and Infectious Diseases; European Society of Intensive Care Medicine; European Respiratory Society; International Sepsis Forum; Japanese Association for Acute Medicine; Japanese Society of Intensive Care Medicine; Society of Critical Care Medicine; Society of Hospital Medicine; Surgical Infec-tion Society; World Federation of Societies of Intensive and Critical Care Medicine: Surviving Sepsis Campaign: International guidelines for management of severe sepsis and septic shock: 2008. Crit Care Med 2008; 36:296–327
11. Bloos F, Thomas-Rüddel D, Rüddel H, et al; MEDUSA Study Group: Impact of compliance with infection management guidelines on out-come in patients with severe sepsis: A prospective observational multi-center study. Crit Care 2014; 18:R42
12. Leligdowicz A, Dodek PM, Norena M, et al; Co-operative Antimicro-bial Therapy of Septic Shock Database Research Group: Association between source of infection and hospital mortality in patients who have septic shock. Am J Respir Crit Care Med 2014; 189:1204–1213
13. Garnacho-Montero J, Huici-Moreno MJ, Gutiérrez-Pizarraya A, et al: Prognostic and diagnostic value of eosinopenia, C-reactive protein, procalcitonin, and circulating cell-free DNA in critically ill patients admitted with suspicion of sepsis. Crit Care 2014; 18:R116
14. Wacker C, Prkno A, Brunkhorst FM, et al: Procalcitonin as a diagnos-tic marker for sepsis: A systematic review and meta-analysis. Lancet Infect Dis 2013; 13:426–435
15. Clec’h C, Ferriere F, Karoubi P, et al: Diagnostic and prognostic value of procalcitonin in patients with septic shock. Crit Care Med 2004; 32:1166–1169
16. Arora S, Singh P, Singh PM, et al: Procalcitonin levels in survivors and nonsurvivors of sepsis: Systematic review and meta-analysis. Shock 2015; 43:212–221
17. Blanco J, Muriel-Bombín A, Sagredo V, et al; Grupo de Estudios y Análisis en Cuidados Intensivos: Incidence, organ dysfunction and mortality in severe sepsis: A Spanish multicentre study. Crit Care 2008; 12:R158
18. Koperna T, Schulz F: Relaparotomy in peritonitis: Prognosis and treat-ment of patients with persisting intraabdominal infection. World J Surg 2000; 24:32–37
19. Marshall JC: Intra-abdominal infections. Microbes Infect 2004; 6:1015–1025
20. Torgersen C, Moser P, Luckner G, et al: Macroscopic postmortem findings in 235 surgical intensive care patients with sepsis. Anesth Analg 2009; 108:1841–1847
21. Boyer A, Vargas F, Coste F, et al: Influence of surgical treatment tim-ing on mortality from necrotizing soft tissue infections requiring inten-sive care management. Intensive Care Med 2009; 35:847–853
22. Kobayashi L, Konstantinidis A, Shackelford S, et al: Necrotizing soft tissue infections: Delayed surgical treatment is associated with increased number of surgical debridements and morbidity. J Trauma 2011; 71:1400–1405
23. Wong CH, Chang HC, Pasupathy S, et al: Necrotizing fasciitis: Clini-cal presentation, microbiology, and determinants of mortality. J Bone Joint Surg Am 2003; 85-A:1454–1460
24. Moss RL, Musemeche CA, Kosloske AM: Necrotizing fasciitis in chil-dren: Prompt recognition and aggressive therapy improve survival. J Pediatr Surg 1996; 31:1142–1146
25. Tridente A, Clarke GM, Walden A, et al; GenOSept Investigators: Patients with faecal peritonitis admitted to European intensive care units: An epidemiological survey of the GenOSept cohort. Intensive Care Med 2014; 40:202–210
26. Mier J, León EL, Castillo A, et al: Early versus late necrosectomy in severe necrotizing pancreatitis. Am J Surg 1997; 173:71–75
27. Hartwig W, Maksan SM, Foitzik T, et al: Reduction in mortality with delayed surgical therapy of severe pancreatitis. J Gastrointest Surg 2002; 6:481–487
28. Opal SM, Dellinger RP, Vincent JL, et al: The next generation of sepsis clinical trial designs: What is next after the demise of recombinant human activated protein C?*. Crit Care Med 2014; 42:1714–1721
29. Bosshardt TL, Henderson VJ, Organ CH Jr: Necrotizing soft-tissue infections. Arch Surg 1996; 131:846–852
30. Mulier S, Penninckx F, Verwaest C, et al: Factors affecting mortal-ity in generalized postoperative peritonitis: Multivariate analysis in 96 patients. World J Surg 2003; 27:379–384
96
97
12.- ANEXO 2: MATERIAL SUPLEMENTARIO ELECTRÓNICO
Material Suplementario Electrónico en relación a la publicación #1 del trabajo de
tesis doctoral
(Improved empirical antibiotic treatment of sepsis after an educational intervention: the
ABISS-Edusepsis study. Ricard Ferrer, María Luisa Martínez, Gemma Gomà, David
Suárez, Luis Álvarez-Rocha, María Victoria de la Torre, Gumersindo González, Rafael
Zaragoza9, Marcio Borges, Jesús Blanco, Eduardo Palencia Herrejón, Antonio Artigas and
for the ABISS-Edusepsis Study group. Crit Care 2018 Jun 22;22(1):167).
Additional file 1: Appendix describing in detail the study design, the approach to data
collection, and the quality-control measures to ensure data reliability.
Study Sites
The steering committee of the ABISS-Edusepsis study defined the project’s
purpose, timeline, interventions, and design. Through the Spanish Society of Intensive
Care, all Spanish ICUs for adults (115 units) were invited to participate. No fees were
provided for participation. ICUs were not asked to provide reasons for not participating.
Seventy-two medical-surgical Spanish ICUs homogeneously distributed around the
country were included in the study. All ICUs were closed units with a critical care
specialist on hand 24 hours per day, 365 days per year. The general coordinating center
was located at the Critical Care Center of the Hospital Universitario Mútua de Terrassa,
Barcelona. Each ICU belonged to a geographical area coordinated by an area coordinator
and at least 1 physician was designated principal investigator in each center.
98
Patients
All ICU admissions from the emergency department or medical or surgical wards
and all ICU patients were actively screened daily for severe sepsis or septic shock. Patients
who received initial infection control measures for sepsis in another hospital were
excluded.
Severe sepsis was defined as sepsis associated with at least one acute organ
dysfunction: (1) respiratory dysfunction (bilateral pulmonary infiltrates with PaO2/FIO2
ratio < 300 mmHg); (2) renal dysfunction (urine output < 0.5 mL/kg per hour for ≥2 hours
or creatinine increase > 0.5 mg/dL or creatinine level > 2.0 mg/dL); (3) coagulation
abnormalities (international normalized ratio >1.5 or partial thromboplastin time > 60
seconds); (4) thrombocytopenia (platelets <100 x 103/μL); (5) hyperbilirubinemia (total
plasma bilirubin >2.0 mg/dL); (6) hypoperfusion (lactate >3 mmol/l); or (7) hypotension
(systolic blood pressure < 90 mmHg, mean arterial pressure < 65 mm Hg, or a reduction in
systolic blood pressure > 40 mm Hg from baseline measurements). Septic shock was
defined as sustained acute circulatory failure (systolic blood pressure < 90 mm Hg, mean
arterial pressure < 65 mm Hg, or a reduction in systolic blood pressure < 40 mm Hg from
baseline) despite adequate volume resuscitation (19).
The onset of sepsis (T0) was determined according to the patient’s location within
the hospital when sepsis was diagnosed. In patients diagnosed with sepsis in the emergency
department, T0 was defined as the time of triage. For patients admitted to the ICU from the
medical or surgical wards or other non-emergency department units, T0 was determined by
searching the clinical documentation for the time of diagnosis of severe sepsis. This might
include, for example, a physician’s note or timed and dated orders, a timed and dated note
of a nurse’s discussion of severe sepsis with a physician, or timed records initiating referral
99
to the ICU for severe sepsis. If no time and date could be found by searching the chart, the
default time of presentation was the time of admission to the ICU. Lastly, for patients who
developed sepsis after admission to the ICU, the time of presentation was again determined
on the basis of the clinical documentation.
Data Collection and Quality Control
Data were collected prospectively daily using an electronic database (EDICS,
www.edics.org). Before data collection started, all investigators received detailed
information explaining the aim of the study, instructions for data collection, and definitions
for various items. The database included automated filters to check for incorrect entries
and alert researchers to possible errors. For quality assurance purposes, two research nurses
and a physician with experience in sepsis trials checked data for completeness, accuracy,
and uniformity. Errors or blank fields generated queries that were returned to each center
for correction.
100
Additional file 2: Appendix describing in detail the educational intervention.
To standardize the educational program, the general coordinating center organized
meetings with the area coordinators and the principal investigator of the participating
centers before and after each study period. Before the implementation of the educational
program, the importance of the study was explained to the hospital manager at each
participating center to ensure full institutional support. The general coordinating center
provided specific material for this meeting with information regarding the epidemiology,
morbidity, mortality, and costs of sepsis.
The principal investigator acted as local champion, charged with creation of a local
multidisciplinary team with representatives of all pertinent stakeholders, including
physicians and nurses from the ICU and infectious diseases, emergency, and internal
medicine departments.
Between January and March 2012, the local multidisciplinary team at each hospital
implemented a homogeneous predefined multifaceted educational program based on the
SSC guidelines with special emphasis on antimicrobial management. The educational
program was implemented in the emergency department, medical and surgical wards, and
the ICU. The intervention consisted of educational outreach, periodic reminders, auditing
and feedback, and a videogame. The educational outreach included interactive educational
sessions in which the local leader gave a 30-minute slide presentation based on the SSC
guidelines recommendations focused on the importance of infection control in sepsis. Each
center was provided with pocket guides and posters with recommendations from the
Spanish Society of Critical Care Medicine and Coronary Units. Posters were displayed in
prominent places in the emergency departments, medical and surgical wards, and ICUs.
Pocket versions were distributed to all participants in the educational program sessions. To
101
facilitate antibiotic prescription, researchers used preferentially their local guideline or an
electronic clinical decision support system (www.es.dgai-abx.de). Attendees received
weekly email and cellphone text reminders reiterating the most important points from the
educational outreach program. Additionally, a videogame was developed to provide staff
attending septic patients with an opportunity to practice applying the guidelines in a
simulated environment and to receive feedback about their performance
(http://www.edusepsis.org/en/training.html).
Each center’s performance was audited and compared with the performance of the
overall group, and local leaders received weekly feedback about their center’s performance
and distributed these results to all staff. The general coordinating center maintained
continuous contact with the principal investigator at each center through a mailing list.
Moreover, after the educational program, a survey was distributed to all principal
investigators to check that all participating centers had completed the educational program,
to know the number and duration of lectures, and to know the principal investigator’s
subjective evaluation of the main endpoints of the educational program, which were
institutional support, creation of a multidisciplinary team, improvement in knowledge, and
improvement in hospital processes.
102
Additional file 3: Table 1. Multivariate linear regression for time to antibiotic
Factors Coefficient 95% CI p
Post-intervention cohorta -0.45 -0.75 to -1.56 0.003
Ageb 0.01 -0.006 to 0.02 0.380
Sexc -0.06 -0.38 to -0.25 0.694
SOFAb 0.02 -0.04 to 0.08 0.427
APACHE IIb -0.01 0.03 to 0.22 0.739
CHARLSONb -0.23 -0.09 to 0.04 0.466
Type of infectiond
Nosocomial 0.92 0.47 to 1.37 <0.001
ICU 3.40 2.57 to 4.23 <0.001
Healthcare related -0.19 -0.27 to 0.66 0.415
Source of sepsise
Acute abdominal infection 0.31 -0.07 to 0.69 0.11
Urinary tract infection -0.23 -0.68 to 0.21 0.303
Meningitis 0.09 -0.80 to 0.98 0.841
Soft-tissue infection 0.03 -0.59 to 0.65 0.926
Catheter-related bacteremia -0.81 -2.01 to 0.39 0.185
Other infections -0.02 -0.75 to 0.71 0.954
Excluding patients with previous antibiotics (n = 858).
Abbreviations: SOFA, Sequential Organ Failure Assessment; APACHE II, Acute Physiology and Chronic Health Evaluation II; ICU,
Intensive Care Unit. aCompared with pre-intervention cohort. bPer each point of increase cCompared with female sex. dCompared to community-acquired
infection. eCompared to pneumonia
103
Additional file 4: Table 2. Segmented regression model for time to first antibiotic in hoursa
Coefficient 95% CI p
Intercept 2.25 1.87-2.63 < 0.001
Change in level
(Post vs. pre-
intervention)
-0.92 -1.51 - -0.33 0.010
Trend 0.11 -0.03-0.23 0.092
aExcluding patients with previous antibiotics (n = 858).
Interaction between trend and intervention was not significant (p=0.288) and therefore was not included in the final
model.
Additional file 5: Table 3. Demographic and clinical characteristics of patients in the long-term
cohort
104
Patient Characteristic
Preintervention
Cohort
(n= 1352)
Postintervention
Cohort
(n= 1276)
Long-term
Cohort
(n= 830)
p
General data
Age (years), mean (SD) 64.3 (15.3) 63.8 (15.1) 63.5 (15.1) 0.689
Sex (male), n (%) 858 (63.5) 824 (64.6) 513 (61.8) 0.197
APACHE-II, mean (SD) 22.5 (8.1) 21.4 (8.0) 21.3 (7.8) 0.591
SOFA, mean (SD) 8.7 (3.5) 8.5 (3.5) 8.7 (3.3) 0.180
Charlson, mean (SD) 2.7 (2.3) 2.7 (2.3) 2.4 (2.2) 0.007
Source of sepsis, n (%) 0.004
Pneumonia 454 (33.6) 403 (31.6) 320 (38.6)
Acute abdominal infection 452 (33.4) 431 (33.8) 270 (32.5)
Urinary tract infection 229 (16.9) 209 (16.4) 131 (15.8)
Soft-tissue infection 82 (6.1) 98 (7.7) 18 (2.2)
Meningitis 26 (1.9) 43 (3.4) 51 (6.1)
Catheter-related bacteremia 30 (2.2) 19 (1.5) 14 (1.7)
Other infections 79 (5.8) 73 (5.7) 26 (3.1)
Type of infection, n (%) 0.317
Community 802 (59.3) 835 (65.4) 521 (62.8)
Nosocomial 302 (22.3) 249 (19.5) 190 (22.9)
ICU 65 (4.8) 51 (4) 33 (4)
Health care related 183 (13.5) 141 (11.1) 86 (10.4)
Diagnosis on admission, n (%) 0.061
Medical 950 (70.3) 891 (69.8) 604 (72.8)
Urgent surgical 318 (23.5) 311 (24.4) 196 (23.6)
Non-urgent surgical 84 (6.2) 74 (5.8) 30 (3.6)
105
Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II; SOFA, Sequential Organ Failure
Assessment; ICU, intensive care unit.
Additional file 6: Table 4. Compliance with process-of-care measurements in the long-term
cohort
106
aExcluding patients with previous antibiotics (n = 761)
Process-of-care
Preintervention
Cohort
(n= 1352)
Postintervention
Cohort
(n= 1276)
Long-
term
cohort
(n= 830)
p
Sepsis Resuscitation Bundle 6 hours, n (%)
Lactate measured 964 (71.3) 986 (77.3) 635 (76.5) 0.683
Blood cultures before antibiotics 656 (48.5) 687 (53.8) 492 (59.3) 0.014
Broad spectrum antibiotics 903 (66.8) 915 (71.7) 587 (70.7) 0.625
Fluids and vasopressors 815 (60.3) 744 (58.3) 476 (57.3) 0.663
Time to antibiotics, mean (SD)
hoursa 2.5 (3.6) 2.0 (2.7) 2.2 (3.7) 0.306
Evaluation of treatment at 72 hours, n (%) 0.002
Inappropriate treatment 120 (8.9) 83 (6.5) 72 (8.7)
Appropriate treatment 678 (50.1) 654 (51.3) 467 (56.3)
Negative cultures 438 (32.4) 417 (32.7) 244 (29.4)
Cultures not done 71 (5.3) 56 (4.4) 23 (2.8)
Patient died in less than 72
hours 45 (3.3) 66 (5.2) 24 (2.9)
Change of antibiotic at 72 hours, n (%) 0.232
Reduction of the spectrum 220 (16.3) 257 (20.1) 181 (21.8)
No change of antibiotic 762 (56.4) 695 (54.5) 437 (52.7)
Change due to poor clinical
evolution 143 (10.6) 97 (7.6) 76 (9.2)
Change due to uncovered
microorganism 82 (6.1) 65 (5.1) 50 (6)
Other 145 (10.7) 162 (12.7) 86 (10.4)
107
Additional file 7: Table 5. Outcome measurements in the long-term cohort
Outcome
measurements
Preintervention
cohort
(n=1352)
Postintervention
cohort
(n= 1276)
Long-term
cohort
(n= 830)
p
Duration of MV, days
mean (SD) 6.9 (14.4) 6.6 (12.4) 6.9 (11.3) 0.623
Duration of
vasopressors, days mean
(SD)
4.0 (8.0) 4.3 (7.0) 4.6 (6.3) 0.289
ICU stay, days
mean (SD) 12.0 (17.0) 11.5 (14.9) 11.5 (12.3) 0.986
Hospital stay, days mean
(SD) 30.0 (29.7) 28.4 (28.9) 27.2 (23.6) 0.319
Mortality, n (%)
ICU 332 (24.6) 301 (23.6) 174 (21.0) 0.159
Hospital 412 (30.5) 375 (29.4) 219 (26.4) 0.134
Abbreviations: ICU, intensive care unit; MV, mechanical ventilation; SD, standard deviation.
108
Additional file 8: Table 6. Multivariate analysis of factors associated with mortality in the long-
term cohort
Factors OR 95% CI P
Interventional cohort 0.83 0.67-1.03 0.091
Agea 1.02 0.64-1.01 <0.001
Sexb 0.81 0.75-1.17 0.063
SOFAa 1.12 1.06-1.16 <0.001
APACHE IIa 1.07 1.05-1.09 <0.001
CHARLSONa 1.09 1.04-1.14 <0.001
Type of infectionc
Nosocomial 2.41 1.86-3.13 <0.001
ICU 2.81 1.72-4.61 <0.001
Healthcare related 1.12 0.83-1.69 0.351
Source of sepsisd
Acute abdominal infection 0.74 0.57-0.96 0.023
Urinary tract infection 0.32 0.22-0.47 <0.001
Meningitis 0.86 0.44-1.70 0.670
Soft-tissue infection 1.09 0.72-1.65 0.697
Catheter-related bacteremia 0.69 0.31-1.55 0.365
Other infections 0.91 0.56-1.50 0.716
Abbreviations: SOFA, Sequential Organ Failure Assessment; APACHE II, Acute Physiology and Chronic Health
Evaluation II; ICU, Intensive Care Unit. aPer each point of increase. bCompared with male sex. cCompared to community-acquired infection. dCompared to
pneumonia.
109
Material Suplementario Electrónico en relación a la publicación #2 del trabajo de
tesis doctoral
(Impact of Source Control in Patients With Severe Sepsis and Septic Shock. María Luisa
Martínez; Ricard Ferrer; Eva Torrents;Raquel Guillamat-Prats; Gemma Gomà; David
Suárez; Luis Álvarez-Rocha; Juan Carlos Pozo Laderas; Ignacio Martín-Loeches; Mitchell
M. Levy; Antonio Artigas; for the Edusepsis Study Group. Critical Care Medicine 2017
Jan;45(1):11-19).
Appendix S1. Appendix describing in detail the study design, the approach to data
collection and the quality-control measures to ensure data reliability.
Patients
All ICU admissions from the emergency department or medical or surgical wards
and all ICU patients were actively screened daily for the presence of severe sepsis or septic
shock. When the onset of severe sepsis could not be determined, patients were not included
in the study.
Severe sepsis was defined as sepsis associated with at least one acute organ
dysfunction: (1) respiratory dysfunction, bilateral pulmonary infiltrates with a ratio of
PaO2/FIO2 of less than 300 mm Hg; (2) renal dysfunction, urine output of less than 0.5
mL/kg per hour for at least 2 hours or creatinine increase > 0.5 mg/dL or creatinine level of
greater than 2.0 mg/dL; (3) coagulation abnormalities, international normalized ratio (INR)
greater than 1.5 or a partial thromboplastin time greater than 60 seconds; (4)
thrombocytopenia, platelet count of less than 100 x 103/μL; (5) hyperbilirubinemia, total
plasma bilirubin level of greater than 2.0 mg/dL; (6) hypoperfusion, lactate level greater
than 3 mmol/l or (7) arterial hypotension, systolic blood pressure of less than 90 mm Hg,
mean arterial pressure of less than 65 mm Hg, or a reduction in systolic blood pressure of
greater than 40 mm Hg from baseline measurements. Septic shock was defined as
110
sustained acute circulatory failure (systolic blood pressure < 90 mm Hg, mean arterial
pressure < 65 mm Hg, or a reduction in systolic blood pressure < 40 mm Hg from baseline)
despite adequate volume resuscitation [9].
The onset of sepsis (time zero) was determined according to the patient’s location
within the hospital when sepsis was diagnosed. In patients diagnosed with sepsis in the
emergency department, time zero was defined as the time of triage. For patients admitted
to the ICU from the medical or surgical wards or other non-emergency department units,
time zero was determined by searching the clinical documentation for the time of diagnosis
of severe sepsis. This might include, for example, a physician’s note or timed and dated
orders, a timed and dated note of a nurse’s discussion of severe sepsis with a physician, or
timed records initiating referral to the ICU for severe sepsis. If no time and date could be
found by searching the chart, the default time of presentation was the time of admission to
the ICU. Lastly, for patients who developed severe sepsis after admission to the ICU, the
time of presentation was again determined on the basis of the clinical documentation.
Data Collection and Quality Control
Data were collected prospectively daily using an electronic database EDICS,
www.edics.org. Detailed instructions explaining the aim of the study, instructions for the
data collection, and definitions for various items were made available to all investigators
before data collection started. The database included alarms that checked for errors:
automated filters which check for incorrect entries. For quality assurance purposes, two
research nurses and a physician with experience in sepsis trials checked data for
completeness, accuracy, and uniformity. Errors or blank fields generated queries that were
returned to each center for correction.
111
Appendix S2. Appendix describing the specific source control techniques separated in
percutaneous versus surgical source control.
Surgical source control: Appendectomy, Gastric surgery, Gastrectomy, Esophagectomy,
Gastric ulcer suture, Duodenal ulcer suture, Liver surgery, Small intestinal resection,
Ileostomy, Partial colectomy, Total colectomy, Colostomy, Right hemicolectomy, Left
hemicolectomy, Sigmoidectomy, Partial pancreatectomy, Total pancreatectomy,
Cholecystectomy, Nephrectomy, Bronchus and lung surgery, Intervention on mediastinum,
Abscess debridement, Debridement of neck and mediastinum, Fasciotomy, Skin-soft tissue
intervention, Hysterectomy, Other obstetric surgery, Valvular heart surgery, Cranial
surgery, Others.
Percutaneous source control: Abdominal percutaneous drainage, Biliary drainage,
Nephrostomy, Ureteral catheterization, Venous catheter removal, Chest drain tube.
112
Table S1. Multivariate analysis of risk factors for ICU mortality in all patients (n = 3,663)
Factors OR 95% CI P
Source controla 0.779 0.625-0.972 0.027
Ageb 1.009 1.002-1.015 0.008
Sex, womanc 1.253 1.048-1.499 0.013
APACHE IIb 1.103 1.090-1.116 < 0.001
Septic shockd 1.475 1.201-1.810 < 0.001
CHARLSONb 1.065 1.026-1.106 0.001
Early broad-spectrum antibiotics 0.839 0.694-1.014 0.070
Fluids and vasopressors 0.799 0.674-0.949 0.010
Appropriate antibiotic therapy e 0.639 0.480-0.850 0.002
Nosocomial acquired infectionf 1.698 1.373-2.100 < 0.001
Site of infectiong
Abdominal 0.852 0.670-1.083 0.190
Urologic 0.257 0.185-0.357 < 0.001
Central nervous system 1.135 0.647-1.992 0.659
Skin and soft-tissue 0.897 0.629-1.279 0.549
Others 0.932 0.669-1.300 0.680
Abbreviations: APACHE II. Acute Physiology and Chronic Health Evaluation II. a Compared with patients not requiring control. b Per each point increase. c Compared with male sex. d Compared with
severe sepsis. e Compared with inappropriate antibiotic therapy. f Compared with the emergency department. g Compared
with respiratory infection.
Table S2. Demographic and clinical characteristics in early and late source control patients
113
Global
n= 1,090
Source Control
< 12h
n= 825
Source
Control ≥ 12h
n= 265
p
General data
Age (years), mean (SD) 67.0 (14.3) 67.1 (14.3) 67.0 (14.2) 0.937
Sex (male), n (%) 651 (59.7) 498 (60.4) 153 (57.7) 0.448
APACHE-II, mean (SD) 21.3 (7.6) 21.4 (7.8) 21.0 (6.7) 0.538
Shock, n (%) 808 (74.1) 619 (75) 189 (71.3) 0.230
Charlson, n (SD) 2.6 (2.3) 2.6 (2.3) 2.6 (2.1) 0.827
CRP (mg/dl), mean (SD)a 26.0 (12.9) 25.6 (12.5) 27.4 (13.9) 0.082
PCT (ng/ml), mean (SD)b 30.8 (43.5) 31.1 (44.8) 30.0 (39.9) 0.796
Bacteremia, n (%) 350 (44.1) 246 (41.3) 104 (52.3) 0.007
Appropriate antibiotic
therapy, n (%) 613 (56.2) 470 (57.0) 143 (54.0) 0.145
Organ failure at sepsis presentation, n (%)
Number of organ failures
(SD) 3.01 (1.4) 3.01 (1.4) 3.0 (1.3) 0.942
Cardiovascular 956 (87.7) 731 (88.6) 225 (84.9) 0.110
Respiratory 300 (27.5) 230 (27.9) 70 (26.4) 0.643
Renal 669 (61.4) 505 (61.2) 164 (61.9) 0.844
Hyperbilirubinemia 214 (19.6) 159 (19.3) 55 (20.8) 0.597
Thrombocytopenia 219 (20.1) 152 (18.4) 67 (25.3) 0.015
Coagulation 376 (34.5) 284 (34.4) 92 (34.7) 0.930
Hyperlactatemia 546 (50.1) 423 (51.3) 123 (46.4) 0.169
Site of infection, n (%) 0.023
Abdominal 753 (69.1) 592 (71.8) 161 (60.8)
114
Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II; CRP, C-reactive protein; ICU,
intensive care unit; PCT, Procalcitonin; SD, standard deviations. a n = 857 patients. b n = 581 patients
Respiratory 43 (3.9) 28 (3.4) 15 (5.7)
Urologic 147 (13.5) 102 (12.4) 45 (17.0)
Skin and/or soft-tissue 113 (10.4) 81 (9.8) 32 (12.1)
Central nervous system 4 (0.4) 2 (0.2) 2 (0.8)
Other 30 (2.8) 20 (2.4) 10 (3.8)
Infection acquisition site, n (%) 0.357
Community 656 (60.2) 487 (59.0) 169 (63.8)
Healthcare-related 110 (10.1) 83 (10.1) 27 (10.2)
Nosocomial 296 (27.2) 235 (28.5) 61 (23.0)
ICU 28 (2.6) 20 (2.4) 8 (3.0)
Diagnosis on admission, n (%) < 0.001
Medical 275 (25.2) 158 (19.2) 117 (44.2)
Surgical 83 (7.6) 62 (7.5) 21 (7.9)
Urgent Surgical 732 (67.2) 605 (73.3) 127 (47.9)
115
Table S3. Compliance with sepsis resuscitation bundle in the early versus late source
control group
Sepsis resuscitation bundle 6h, n (%)
Global
n = 1,090
Source
Control < 12h
n = 825
Source
Control ≥ 12h
n = 265
p
All resuscitation measures 91 (8.3) 68 (8.2) 23 (8.7) 0.823
Measure lactate 769 (70.6) 616 (74.7) 153 (57.7) < 0.001
Blood cultures before
antibiotics 486 (44.6) 356 (43.2) 130 (49.1) 0.092
Early broad-spectrum
antibiotics 739 (67.8) 569 (69.0) 170 (64.2) 0.144
Fluids and vasopressors 650 (59.6) 528 (64.0) 122 (46.0) < 0.001
Central venous pressure ≥
8 mm Hg 461 (42.3) 370 (44.8) 91 (34.3) 0.003
Central venous oxygen
saturation ≥ 70% 352 (32.3) 283 (34.3) 69 (26.0) 0.012
116
Table S4. Outcome measurements according to time to source control in abdominal
foci (n = 753)
Source
Control < 12h
592 (78.6%)
Source Control
≥ 12h
161 (21.4%)
P
Duration of MV, days
mean (SD) 7.4 (13.0) 7.3 (13.8) 0.936
Duration of
vasopressors, days
mean (SD)
4.9 (8.2) 5.03 (6.3) 0.817
ICU stay, days mean
(SD) 12.2 (15.4) 13.4 (15.6) 0.404
Hospital stay, days
mean (SD) 32.4 (27.6) 32.8 (30.4) 0.829
Mortality, n (%)
ICU 141 (23.8) 37 (23) 0.825
Hospital 186 (31.4) 50 (31.1) 0.930
Abbreviations: ICU, intensive care unit; MV, mechanical ventilation; SD, standard deviations.
117
Table S5. Outcome measurements according to time to source control in urinary
focus (n = 147)
Source
Control < 12h
102 (69.4%)
Source Control
≥ 12h
45 (30.6%)
P
Duration of MM, days
mean (SD) 2.8 (8.6) 4.3 (17.7) 0.489
Duration of
vasopressors, days mean
(SD)
2.5 (3.2) 5.5 (18.9) 0.130
ICU stay, days Mean
(SD) 6.3 (9.1) 8.7 (18.8) 0.302
Hospital stay, days
Mean (SD) 17.7 (14.7) 21.3 (21.0) 0.232
Mortality, n (%)
ICU 6 (5.9) 4 (8.9) 0.505
Hospital 8 (7.8) 5 (11.1) 0.520
Abbreviations: ICU, intensive care unit; MV, mechanical ventilation; SD, standard deviations.
118
Table S6. Outcome measurements according to time to source control in
skin and soft-tissue foci (n = 113)
Source Control
< 12h
80 (70.8%)
Source Control
≥ 12h
33 (29.2%)
P
Duration of MV, days
mean (SD) 8.1 (11.2) 10.6 (11.9) 0.296
Duration of vasopressors,
days mean (SD) 4.5 (5.0) 7.6 (7.1) 0.010
ICU stay, days
Mean (SD) 13.9 (14.3) 16.1 (13.1) 0.440
Hospital stay, days mean
(SD) 40.3 (41.3) 42.3 (31.5) 0.801
Mortality, n (%)
ICU 12 (15.0) 8 (24.2) 0.242
Hospital 15 (18.8) 9 (27.3) 0.314
Abbreviations: ICU, intensive care unit; MV, mechanical ventilation; SD, standard deviations.
119
Table S7. Univariate analysis of factors associated with in-hospital mortality
Global
n = 3,663
Survivors
n = 2,575
(70.3%)
Non survivors
n = 1,088
(29.7%)
p
General data
Age (years), mean (SD) 64 (15.1) 62.4 (15.6) 67.8 (13.3) < 0.001
Sex (male), n (%) 2,319 (63.3) 1,621 (63) 698 (64.2) 0.490
APACHE-II, mean (SD) 21.8 (8.01) 19.9 (7.2) 26.3 (8.1) < 0.001
SOFA, mean (SD) 8.6 (3.4) 7.9 (3.2) 10.3 (3.4) < 0.001
Shock, n (%) 2,497 (68.2) 1,643 (63.8) 854 (78.5) < 0.001
Charlson, mean (SD) 2.6 (2.3) 2.4 (2.2) 3.2 (2.4) < 0.001
Infection acquisition site, n (%)
Nosocomial 780 (21.3) 442 (17.2) 338 (31.1) < 0.001
Site of infection, n (%) < 0.001
Abdominal 1,234 (33.7) 827 (32.1) 407 (37.4)
Respiratory 1,232 (33.6) 828 (32.2) 404 (37.1)
Urinary 606 (16.5) 529 (29.5) 77 (7.1)
Skin and soft-tissue 258 (7.0) 182 (7.1) 76 (7.0)
Central nervous system 87 (2.4) 61 (2.4) 26 (2.4)
Other 246 (6.7) 148 (5.7) 98 (9.0)
Organ dysfunction criteria at sepsis presentation, n (%)
Number of organ failures
(SD) 2.98 (1.39) 2.8 (1.3) 3.5 (1.5) < 0.001
Hemodynamic 3,019 (82.4) 2,073 (80.5) 946 (86.9) < 0.001
Respiratory 1,602 (43.7) 1,016 (39.5) 586 (53.9) < 0.001
Renal 2,068 (56.5) 1,386 (53.8) 682 (62.7) < 0.001
120
Abbreviations: APACHE II. Acute Physiology and Chronic Health Evaluation II; ICU. intensive care unit; IQR.
interquartile range; SD. standard deviations; MV. mechanical ventilation.
Hyperbilirubinemia 606 (16.5) 371 (14.4) 235 (21.6) < 0.001
Thrombocytopenia 856 (23.4) 540 (21) 316 (29) < 0.001
Coagulation 1,143 (31.2) 730 (28.3) 413 (38) < 0.001
Hyperlactatemia 1630 (44.5) 1,043 (40.5) 587 (54) < 0.001
Sepsis resuscitation bundle 6h, n (%)
All resuscitation measures 379 (10.3) 324 (12.6) 55 (5.1) < 0.001
Measure lactate 2,709 (74) 1,920 (74.6) 789 (72.5) 0.198
Blood cultures before
antibiotics 1,906 (52.0) 1,396 (54.2) 510 (46.9) < 0.001
Early broad-spectrum
antibiotics 2,550 (69.6) 1,882 (73.1) 668 (61.4) < 0.001
Fluids and vasopressors 2,152 (58.7) 1,551 (60.2) 601 (55.2) 0.005
Central venous pressure ≥
8 mm Hg 1,609 (43.9) 1,197 (46.5) 412 (37.9) < 0.001
Central venous oxygen
saturation ≥ 70% 1,236 (33.7) 961 (37.3) 275 (25.3) < 0.001
Outcome measurements
Duration of MV, days
mean (SD) 6.88 (13.2) 5.4 (11.2) 13.5 (17.9) < 0.001
Duration of vasopressors,
days mean (SD) 4.26 (7.2) 3.2 (5.3) 6.8 (9.9) < 0.001
ICU stay, days mean (SD) 12.2 (15.3) 11.1 (14.1) 13.5 (17.9) < 0.001
Hospital stay, days mean
(SD) 32.3 (31.3) 31.2 (29.2) 23.8 (26.5) < 0.001
Time to source control, h
median (IQR) 4.6 (10.5) 4.5 (10.5) 4.9 (10.5) 0.465
121
Table S8. Multivariate analysis of risk factors for hospital mortality in source control
group (n=1,090)
Factors OR 95% CI P
Time to source control ≥ 12ha 1.082 0.756-1.548 0.668
Ageb 1.027 1.014-1.040 < 0.001
Sex, womanc 1.272 0.928-1.743 0.134
APACHE IIb 1.102 1.078-1.127 < 0.001
Septic shockd 1.577 1.072-2.320 0.021
CHARLSONb 1.108 1.037-1.184 0.002
Early broad-spectrum antibiotics 0.744 0.534-1.036 0.080
Fluids and vasopressors 1.035 0.756-1.417 0.831
Appropriate antibiotic therapy e 0.803 0.484-1.333 0.396
Nosocomial acquired infectionf 1.772 1.242-2.528 0.002
Site of infectiong
Abdominal 0.682 0.325-1.428 0.310
Urologic 0.156 0.061-0.401 < 0.001
Central nervous system 5.445 0.393-75.471 0.206
Skin and soft-tissue 0.609 0.256-1.453 0.264
Others 0.805 0.256-2.536 0.711
Abbreviations: APACHE II. Acute Physiology and Chronic Health Evaluation II. a Compared with time to source control < 12h. b Per each point increase. c Compared with male sex. dCompared with
severe sepsis. eCompared with inappropriate antibiotic therapy. f Compared with the emergency department. g Compared
with respiratory infection.
122
Table S9. Multivariate analysis of risk factors for hospital mortality in abdominal
foci (n=753)
Factors OR 95% CI P
Time to source control ≥ 12ha 0.997 0.664-1.499 0.990
APACHE IIb 1.109 1.082-1.136 <0.001
CHARLSONb 1.123 1.044-1.207 0.002
Early broad-spectrum antibiotics 0.604 0.430-0.850 0.004
Fluids and vasopressors 1.093 0.773-1.546 0.616
Abbreviations: APACHE II. Acute Physiology and Chronic Health Evaluation II. aCompared with time to source control < 12h. bPer each point increase.
Table S10. Multivariate analysis of risk factors for hospital mortality in urinary
foci (n=147)
Factors OR 95% CI P
Time to source control ≥ 12ha 2.432 0.614-9.627 0.205
APACHE IIb 1.130 1.028-1.242 0.011
CHARLSONb 1.411 1.081-1.841 0.011
Early broad-spectrum antibiotics 3.618 0.379-34.549 0.264
Fluids and vasopressors 0.572 0.154-2.115 0.402
Abbreviations: APACHE II. Acute Physiology and Chronic Health Evaluation II. aCompared with time to source control < 12h. bPer each point increase.
123
Table S11. Multivariate analysis of risk factors for hospital mortality in skin and
soft-tissue foci (n=113)
Factors OR 95% CI P
Time to source control ≥ 12ha 2.226 0.553-8.954 0.260
APACHE IIb 1.130 1.035-1.234 0.006
CHARLSONb 1.061 0.782-1.438 0.704
Early broad-spectrum antibiotics 2.344 0.202-27.269 0.496
Fluids and vasopressors 0.355 0.080-1.578 0.174
Abbreviations: APACHE II. Acute Physiology and Chronic Health Evaluation II. aCompared with time to source control < 12h. bPer each point increase.
124
Figure 1. ROC - Mortality by time to source control. We analyzed time of source
control as a continuous variable, and no relationship was observed between time to source
control and hospital mortality. The ROC curve did not identify a point of maximum
sensitivity and specificity to predict the optimal time for source control intervention (AUC
0.504, ns)
125
Figure 2. ROC - Mortality by time to source control in percutaneous group. The ROC
curve did not identify a point of maximum sensitivity and specificity to predict the optimal
time for source control intervention (AUC 0.523, ns).
AUC 0.523
p ns
126
Figure 3. ROC - Mortality by time to source control in surgical group. The ROC curve
did not identify a point of maximum sensitivity and specificity to predict the optimal time
for source control intervention (AUC 0.537, ns).
AUC 0.537
p ns
127
13.- ANEXO 3: DATOS NO PUBLICADOS
Datos no publicados en relación a la publicación #2 del trabajo de tesis doctoral
(Impact of Source Control in Patients With Severe Sepsis and Septic Shock. María Luisa
Martínez; Ricard Ferrer; Eva Torrents;Raquel Guillamat-Prats; Gemma Gomà; David
Suárez; Luis Álvarez-Rocha; Juan Carlos Pozo Laderas; Ignacio Martín-Loeches; Mitchell
M. Levy; Antonio Artigas; for the Edusepsis Study Group. Critical Care Medicine 2017
Jan;45(1):11-19).
Table 1. Outcome measurements according to time to source control (n = 1,090)
< 4h
n = 484
4-8h
n = 233
8-12h
n = 108
12-24h
n = 194
> 24h
n = 71 p
Duration of MV, days mean (SD)
6.7 (12.3) 8.6 (15.4) 5.8 (8.2) 7.8 (15.6) 4.7 (6.4) 0.109
Duration of
vasopressors, days
mean (SD) 4.8 (8.4) 4.3 (6.5) 4.3 (5.4) 5.7 (19.9) 4.2 (4.4) 0.414
ICU stay, days
mean (SD) 11.7 (14.5) 13.5 (17.9) 10.6 (11.2) 13.5 (17.3) 10.2 (7.7) 0.201
Hospital stay, days
mean (SD) 34.3 (34.7) 30.1 (30.3) 29.2 (23.2) 33.6 (31.4) 26.2 (17.4) 0.119
Mortality, n (%)
ICU 97 (20.0) 49 (21.0) 25 (23.1) 39 (20.1) 16 (22.5) 0.948
Hospital 129 (26.7) 65 (27.9) 33 (30.6) 54 (27.8) 18 (25.4) 0.928
Abbreviations: ICU, intensive care unit; MV, mechanical ventilation; SD, standard deviations.
128
Table 2. Outcome measurements according to time to source control (n = 1,090)
< 4h
n = 484 ≥ 4h
n = 606 p
Duration of MV, days mean
(SD) 6.7 (12.3) 7.4 (13.7) 0.361
Duration of vasopressors, days
mean (SD) 4.8 (8.4) 4.7 (7.9) 0.930
ICU stay, days mean (SD) 11.7 (14.5) 12.6 (15.8) 0.344
Hospital stay, days mean (SD) 34.3 (34.7) 30.6 (28.3) 0.049
Mortality, n (%)
ICU 97 (20.0) 129 (21.3) 0.614
Hospital 129 (26.7) 170 (28.1) 0.607
Abbreviations: ICU, intensive care unit; MV, mechanical ventilation; SD, standard deviations.
129
Table 3. Outcome measurements in group with source control versus group without
source control, excluding urinary focus (n = 3,057)
Without source
control
n = 2,036
With source
control
n = 1,026
p
Duration of MV, days mean
(SD) 7.9 (13.9) 7.7 (13.7) 0.618
Duration of vasopressors, days mean (SD)
4.3 (7.1) 5.0 (7.9) 0.030
ICU stay, days mean (SD)
13.1 (18.1)
13.1 (16.3)
0.943
Hospital stay, days mean
(SD)
29.2 (29.9)
35.0 (32.9)
< 0.001
Mortality, n (%)
ICU 581 (28.6) 239 (23.3) 0.002
Hospital 692 (34.1) 319 (31.1) 0.098
Abbreviations: ICU, intensive care unit; MV, mechanical ventilation; SD, standard deviations.
Table 4. Multivariate analysis of risk factors for ICU mortality in all patients excluding
those with a urinary focus (n = 3,057)
Factors OR 95% CI P
Source controla 0.724 0.596-0.878 0.001
Ageb 1.009 1.002-1.015 0.009
Sexc 1.213 1.007-1.460 0.042
APACHE IId 1.101 1.087-1.114 < 0.001
Septic shocke 1.428 1.157-1.762 0.001
CHARLSONf 1.044 1.004-1.086 0.033
Early broad-spectrum antibiotics 0.812 0.668-0.987 0.036
Fluids and vasopressors 0.813 0.681-0.972 0.023
Appropiate antibiotic therapyg 0.669 0.496-0.902 0.008
Nosocomial acquired infectionh 1.664 1.339-2.068 < 0.001
Abbreviations: APACHE II. Acute Physiology and Chronic Health Evaluation II. a Compared with non-source control. b Per each point increase. c Compared with male sex. d Per each point increase. e Compared with severe sepsis. f Per each point increase. g Compared with inappropiate antibiotic therapy. h Compared with the emergency department.
130
Table 5. Multivariate analysis of risk factors for hospital mortality in source control group,
excluding urinary focus (n = 943)
Factors OR 95% CI P
Time to source control ≥ 12ha 1.040 0.719-1.504 0.836
Ageb 1.027 1.014-1.040 < 0.001
Sexc 1.242 0.900-1.714 0.188
APACHE IId 1.103 1.078-1.129 < 0.001
Septic shocke 1.501 1.014-2.220 0.042
CHARLSONf 0.787 1.014-1.161 0.018
Early broad-spectrum antibiotics 0.711 0.509-0.991 0.044
Fluids and vasopressors 1.036 0.751-1.429 0.830
Appropiate antibiotic therapyg 0.787 0.466-1.328 0.369
Nosocomial acquired infectionh 1.783 1.247-2.550 0.002
Abbreviations: APACHE II. Acute Physiology and Chronic Health Evaluation II. a Compared with time to source control < 12h. b Per each point increase. c Compared with male sex. d Per each point
increase. e Compared with severe sepsis. f Per each point increase. g Compared with inappropriate antibiotic therapy. h Compared with the emergency department.
131
14.- ANEXO 4: MATERIAL EDUCATIVO
Parte del material educativo que se utilizó durante la fase intervención.
132
133
134
135
15.- ANEXO 5: BECAS Y PREMIOS
Premios y becas a los trabajos de la presente tesis doctoral:
Beca FIS 10/01497. Evaluación de la efectividad y eficiencia de una intervención
múltiple dirigida a mejorar la antibióticoterapia empírica precoz en la sepsis grave.
2010-2013. Instituto de Salud Carlos III.
PREMIO Dr. Ignacio Sánchez Nicolay a la Mejor Comunicación Oral presentada
en el XLVII Congreso Nacional de la SEMICYUC. Santander 13 Junio 2012. ML.
Martínez, R. Ferrer, C. de Haro, G. Gomà, G. Gili, A. Artigas. "Análisis
descripitivo de los pacientes sépticos en las UCIs españolas. Fase preintervención
del estudio ABISS-Edusepsis."
Contrato de Formación en Investigación Río Hortega del ISC III CM12/00167.
2013-2015. Beca otorgada para realizar la presente tesis doctoral, proyecto ABISS-
Edusepsis.
Premio Dra. María Jesús López Pueyo al trabajo publicado “Impact of source
control in patients with severe sepsis and septic shock”. Fundación Burgos por la
Investigación de la Salud. Hospital Universitario de Burgos. 21 de Junio 2018.
136
137
16.- ANEXO 6: OTRAS PUBLICACIONES EN RELACIÓN A LA TESIS DOCTORAL
Publicaciones en las que ha participado la doctoranda en relación a la presente tesis
doctoral:
- Ana Navas, Ricard Ferrer, Maria Luisa Martínez, Gemma Gomà, Gisela Gili, Jordi
Masip, David Suárez and Antonio Artigas. Impact of hemoperfusion with polymyxin B
added to hemofiltration in patients with endotoxic shock: a case–control study. Annals of
Intensive Care 2018;8:121.
- María Luisa Martínez, Juan Carlos Ruiz-Rodríguez, Ricard Ferrer. Improving knowledge
about sepsis 3 definition in critically ill patients: new insights. J Emerg Crit Care Med
2018;2:39.
- Yébenes JC, Ruiz-Rodriguez JC, Ferrer R, Clèries M, Bosch A, Lorencio C, et al.
Epidemiology of sepsis in Catalonia: analysis of incidence and outcomes in a European
setting. Ann Intensive Care. 2017 Dec 20;7(1):19.
- Sánchez B, Ferrer R, Suarez D, Romay E, Piacentini E, Gomà G, Martínez ML, Artigas
A; Edusepsis Study Group. Declining mortality due to severe sepsis and septic shock in
Spanish intensive care units: A two-cohort study in 2005 and 2011. Med Intensiva 2017
Jan - Feb;41(1):28-37.
- Póvoa P, Salluh JIF, Martínez ML, Guillamat-Prats R, Gallup D, Al-Khalidi HR,
Thompson BT, Ranieri VM, Artigas A. Clinical impact of stress dose steroids in patients
with septic shock: insights from the PROWESS-Shock trial. Critical care 2015;19(1):193.
- Díaz-Martín A, Martínez-González ML, Ferrer R, Ortiz-Leyba C, Piacentini E, Lopez-
Pueyo MJ, Martín-Loeches I, Levy MM, Artigas A, Garnacho-Montero J; for the
Edusepsis Study Group. Antibiotic prescription patterns in the empiric therapy of severe
sepsis: combination of antimicrobials with different mechanisms of action reduces
mortality. Crit Care 2012 Nov 18;16(6):R223.
Navas et al. Ann. Intensive Care (2018) 8:121 https://doi.org/10.1186/s13613-018-0465-8
RESEARCH
Impact of hemoperfusion with polymyxin B added to hemofiltration in patients with endotoxic shock: a case–control studyAna Navas1* , Ricard Ferrer2,3, Maria Luisa Martínez4, Gemma Gomà1, Gisela Gili5, Jordi Masip6, David Suárez7 and Antonio Artigas1,2
Abstract
Background: Septic shock is a leading cause of death in critical patients. In patients with gram-negative septic shock, hemoperfusion with polymyxin B aims to remove endotoxins from plasma. We analyzed the clinical and bio-logical response to hemoperfusion in patients with septic shock and acute kidney injury.
Methods: This prospective case–control study in the medical–surgical intensive care unit of a university hospital included consecutive adults patients with septic shock and suspected gram-negative bacteria infection with elevated plasma endotoxin activity (EAA > 0.6 EU/ml) and acute kidney injury requiring continuous renal replacement therapy (CRRT). At onset of septic shock, half underwent CRRT plus hemoperfusion with polymyxin B for two hours a day dur-ing two consecutive days (hemoperfusion group) and half received only CRRT (control group). We measured clinical, physiological, and biological parameters (EAA, C-reactive protein, procalcitonin, and cytokines) daily during the first 5 days.
Results: We included 18 patients (male, 33%; mean age, 67.5; mean SOFA score, 11.3). Abdominal infections predom-inated (50% had peritonitis). At the beginning of CRRT, RIFLE classification was “failure” for 72% and “injury” for 28%. Baseline characteristics did not differ between groups. Patients in the hemoperfusion group required longer mechani-cal ventilation (12.4 vs. 9.4 days, p = 0.03) and CRRT (8.5 vs. 6 days, p = 0.01) than in the control group. Noradrenaline doses, lactate, procalcitonin, and C-reactive protein decreased in both groups. At day 5, EAA was significantly lower in the hemoperfusion group (0.58 EU/ml vs. 0.73 EU/ml in controls, p = 0.03). There were no significant differences between groups in other biomarkers or ICU mortality (33.3% in the treatment group vs. 44.4% in the control group, p = 0.5). No adverse effects of hemoperfusion were observed.
Conclusions: Hemoperfusion with polymyxin B added to CRRT resulted in faster decrease in endotoxin levels, but we observed no improvements in clinical, physiological, or biological parameters.
Keywords: Endotoxic shock, Hemoperfusion, Polymyxin B, Acute kidney injury, Hemofiltration
© The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
Open Access
*Correspondence: [email protected] 1 Critical Care Center, Parc Tauli Hospital Universitari, Institut d’Investigació i Innovació Parc Taulí I3PT, Universitat Autònoma de Barcelona, Sabadell, SpainFull list of author information is available at the end of the article
Page 2 of 9Navas et al. Ann. Intensive Care (2018) 8:121
BackgroundDespite continuous improvements in the care of critical patients, septic shock is common and remains a leading cause of death [1]. Mortality rates for septic shock range from 28% to 50%, depending on the type of causal micro-organism, source of infection, age, sex, comorbidities, severity of disease, and inflammatory response.
In septic shock caused by gram-negative bacteria, endotoxin activates the toll-like receptor 4 (TLR-4), gen-erating an inflammatory response including an increase in proinflammatory cytokines and a coagulation cascade with increased neutrophils, endothelial and systemic damage, vasodilation, and increased interstitial perme-ability resulting in secondary edema at the tissue level [2].
Initial management and treatment of septic patients has improved since the publication of the Surviving Sep-sis Campaign guidelines, which prioritize early diagno-sis, early broad-range antibiotic administration, source control, and hemodynamic management [3]. In sep-tic patients with acute kidney injury, continuous renal replacement techniques (CRRT) have been extensively studied with the rationale that extracting inflamma-tory mediators from patients’ plasma would be benefi-cial; however, attempts to correlate the dose or timing of CRRT with outcomes have failed [4, 5]. Most studies of CRRT in septic patients found decreased cytokines, but no improvement in survival [6, 7].
Hemoperfusion with polymyxin B (Toraymyxin®, Toray Industries) is a blood purification technique in which the patient’s plasma is filtered through a cartridge containing polyurethane and polystyrene-derivative fibers with poly-myxin B, an antibiotic that has a high affinity for endo-toxin. This adsorptive technique eliminates circulating endotoxin by covalent bonding (1:1) to polymyxin B [8]. Endotoxin concentrations higher than 500 pg/ml (> 0.6 EU/ml) in septic shock patients are associated with poor outcome [9–11].
Preclinical studies have shown that hemoperfu-sion with polymyxin B adsorbs endotoxin from circu-lating blood. Hemoperfusion with polymyxin B also decreases the major inflammatory cytokines and procal-citonin. However, various studies and meta-analyses have reported disparate results about the effects of hemoper-fusion with polymyxin B on clinical outcomes [12–30].
We aimed to analyze the clinical, physiological, and biological effects of hemoperfusion with polymyxin B in patients with endotoxic shock and acute kidney injury treated with CRRT.
MethodsStudy designThis prospective case–control study was conducted in a mixed ICU at a university hospital from January 1, 2008,
to May 31, 2012. The Ethics Committee of the Hospital de Sabadell approved the protocol (CIR09065/CEAAH 2407), and all patients provided written informed consent.
We prospectively included consecutive adult patients with acute (< 48 h) septic shock (suspected bloodstream infection and need for vasoactive drugs) with an abdomi-nal, biliary, or renal focus of infection, with acute kidney injury requiring CRRT (RIFLE score indicating injury or worse), and with elevated plasma endotoxin activity, defined as > 0.6 EU/ml on the Endotoxin Activity Assay (EAA™) (Spectral Diagnostics, Toronto, Canada) [9]. All patients received standard care according to the recom-mendations of the Surviving Sepsis Campaign [3]. All patients underwent hemofiltration with effluent flow rate 35 ml/kg/h through a double-lumen 13-F catheter and an AN-69 membrane through a PrismaFlex CRRT system (Baxter®); the system was anticoagulated with sodium heparin unless contraindicated, and bicarbonate-buff-ered solution was used as the replacement fluid. The first patients were assigned to the hemoperfusion group; in addition to hemofiltration, these patients underwent 2 h hemoperfusion with polymyxin B (Toraymyxin®, Toray Medical Co) on two consecutive days, starting within 24 h of ICU admission. The next consecutive patients who met the inclusion criteria were assigned to the con-trol group; these patients underwent hemofiltration with-out hemoperfusion.
VariablesWe recorded the following variables: demographic char-acteristics (age and sex), severity (APACHE II), organ failure (SOFA scores), source of infection (peritonitis, bil-iary, or urinary tract infection), first antibiotic adminis-tered, technique used for infection source control, blood cultures, vasoactive drugs (doses and duration), mechan-ical ventilation, duration of CRRT, central venous oxygen saturation, and Pa02/Fi02 ratio.
All patients were monitored with a pulse index con-tinuous cardiac output (PiCCO) system (Pulsion®), and the following hemodynamic variables were recorded: mean arterial pressure, vascular resistance, global end-diastolic volume, extravascular lung water index, pul-monary vascular permeability index, and global ejection fraction. Every 12 h during the first 2 days and then every 24 h during the next 3 days, blood samples were collected for standard biochemical analyses (pH, HCO3, serum creatinine, azotemia) and EAA determinations. Plasma samples were frozen for further analyses of various bio-markers (neutrophil gelatinase-associated lipocalin (NGAL), soluble urokinase-type plasminogen activator receptor (SuPAR), and cytokines (IL-6, IL-8, IL-10, TNF-α, IL-1β) (See supplementary material).
Page 3 of 9Navas et al. Ann. Intensive Care (2018) 8:121
Patients were followed up until death or hospital dis-charge. Primary outcomes were variables related to the biological, physiological, and clinical effects of hemop-erfusion with polymyxin B. Secondary outcomes were ICU and hospital lengths of stay and ICU and hospital mortality.
Statistical analysisDescriptive statistics included frequencies and percent-ages for categorical variables and means, standard devia-tions (SD), and confidence intervals (CI) for continuous variables. To compare categorical variables, we used the χ2 test or Fisher’s exact test, as appropriate. To compare continuous variables, we used Student’s t test or the Mann–Whitney U test, as appropriate. Statistical tests were two-tailed with significance defined as p < 0.05. We used SPSS, version 15.0 (SPSS, Chicago, Illinois), for all analyses.
ResultsA total of 18 patients were included: nine patients in the hemoperfusion group and nine in the control group. No adverse effects or coagulation of the circuit were observed in association with any of the 18 hemoperfusion treatments. Table 1 reports the demographic and base-line clinical characteristics of patients in the two groups and their overall outcomes. Demographic and baseline
clinical characteristics were similar in the two groups. At the beginning of CRRT, 72% were classified as Fail-ure according to RIFLE. The main source of sepsis was peritonitis, followed by biliary and urinary foci. Patients in the hemoperfusion group required longer mechani-cal ventilation (12.4 vs. 6.3 days, p = 0.03) and longer CRRT (8.5 vs. 3.5 days, p < 0.01). The mean ICU stay was 19.8 days, and the mean hospital stay was 36 days. Over-all ICU mortality was 38.9% (33.3% in the hemoperfusion group vs. 44.4% in control group, p = 0.5); all in-hospital deaths occurred in the ICU. Mortality on day 2 (at end treatment) was 0% in the hemoperfusion group and 33% in the control group.
Table 2 reports the source of sepsis, microbiology findings, and initial empiric antibiotic treatment in each patient. In accordance with the ICU’s protocol, all patients were treated with appropriate broad-spectrum antibiotics within 3 h of the diagnosis of septic shock, and antibiotic treatment was de-escalated according to the microbiology results. All patients underwent source control within 6 h of septic shock diagnosis. Stress-dose steroids (hydrocortisone, 100 mg every 8 h) were admin-istered to 94%. All but one of the patients underwent invasive mechanical ventilation.
Table 3 reports the evolution of respiratory, hemo-dynamic, and renal parameters in the two groups over the first 5 days. Uremia, creatinine, and doses of
Table 1 Demographic and clinical characteristics of patients
APACHE II Acute Physiology and Chronic Health Evaluation II, SD standard deviation, SOFA Sequential Organ Failure Assessment, IMV invasive mechanical ventilation, CRRT continuous renal replacement therapy, LOS length of stay, ICU intensive care unit
Global(n = 18)
Control patients(n = 9)
Toraymyxin treated patients(n = 9)
p value
Age: years, mean (SD) 67.5 (9.9) 66 (10) 69.1 (9.5) 0.52
Male: n (%) 6 (33.3) 2 (22) 4 (44) 0.31
Apache II: mean (SD) 20.67 (4.7) 21.2 (5.3) 20.1 (4.3) 0.63
SOFA baseline mean (SD) 11.3 (2.6) 11.6 (2.9) 11 (2.4) 0.6
Sepsis focus n (%)
Peritonitis 9 (50) 4 (44) 4 (44) 0.4
Biliary tract 4 (23) 2 (22) 3 (33)
Urinary tract 5 (27) 3 (33) 2 (22)
RIFLE score (%)
Injury 28 33 22 0.5
Failure 72 64 78
Vasoactive drugs (days), mean (SD) 4.9 (3.8) 4.5 (3.4) 5.2 (4.3) 0.72
IMV (days), mean (SD) 9.4 (8.8) 6.3 (8) 12.4 (8.8) 0.03
CRRT (days), mean (SD) 6 (6) 3.5 (1.9) 8.5 (7.6) 0.01
ICU LOS (days), mean (SD) 19.8 (16.8) 14.7 (16) 24.9 (17) 0.21
Hospital LOS (days), mean (SD) 36 (31) 32 (34) 39.5 (29) 0.64
ICU mortality (%) 38.9 44.4 33.3 0.5
Hospital mortality (%) 38.9 44.4 33.3 0.5
Page 4 of 9Navas et al. Ann. Intensive Care (2018) 8:121
noradrenaline decreased significantly compared to the baseline, but the decrease did not differ between groups. Changes in score SOFA between day 1 and day 5 did not differ between groups.
Table 4 reports the evolution of biological parameters in the two groups over the first 5 days. Lactate, C-reac-tive protein, procalcitonin, NGAL, and suPAR decreased significantly compared to the baseline, but did not differ between groups. Pro- and anti-inflammatory interleu-kins gradually decreased, but the decrease did not differ between groups.
Figure 1 shows the evolution of endotoxin plasma activ-ity in the two groups from day 1 to day 5. At baseline, both groups had similar, elevated EAA values. On day 3, EAA had decreased significantly in both groups with respect to the baseline value, but on day 5, the decrease was significant only in the hemoperfusion group.
DiscussionThis prospective study analyzed the clinical and biologi-cal effects of hemoperfusion with polymyxin B combined with CRRT in a homogenous population of patients with endotoxic septic shock and multiorgan failure who underwent invasive hemodynamic monitoring. We found that hemoperfusion with polymyxin decreased plasma endotoxin activity, but did not significantly improve clini-cal or biological parameters.
Many studies (mainly from Japan) have reported that hemoperfusion with polymyxin B decreased endotoxin levels [12–15, 17, 22]. Two large randomized trials, EUPHAS [18] and ABDOMIX [19], tested hemoperfu-sion with polymyxin B in critical patients, but did not analyze plasma endotoxin levels. In our study, where elevated plasma endotoxin was an inclusion criterion, patients who received hemoperfusion with polymyxin B had significantly decreased endotoxin levels at day 5 compared with the control group.
Unlike other studies, we found no improvement in mul-tiorgan dysfunction, mortality, or biomarkers in patients who underwent hemoperfusion with polymyxin B com-pared to the control group. Various observational studies showed diverse benefits from hemoperfusion treatment, including improvement in respiratory and hemodynamic parameters, as well as increased survival. In contrast to Vincent et al. [16], who found improved cardiac and renal function, we found no differences in the improvement in organ failure between the hemoperfusion and control groups.
In a systematic review, Cruz et al. [17] concluded that hemoperfusion treatment was associated with improve-ments in mean arterial pressure, use of vasoactive drugs, Pa02/Fi02 ratio, and mortality. By contrast, the ABDO-MIX study [19] and Iwagami et al.’s retrospective study [31] found no improvements. The EUPHRATES trial [21], a multicenter, placebo-controlled, blind trial in
Table 2 Type of infection, microbiology findings, and initial empiric antibiotic treatment
Sepsis focus Microbiology findings Empiric antibiotic treatment
Control patients
Patient 1 Urinary tract Negative Meropenem plus vancomycin plus caspofungin
Patient 2 Urinary tract Escherichia coli Piperacillin–tazobactam
Patient 3 Urinary tract Escherichia coli Meropenem
Patient 4 Biliary tract Escherichia coli/Enterococcus faecalis, Enterococcus faecium Meropenem plus vancomycin
Patient 5 Peritonitis Escherichia coli Meropenem plus vancomycin
Patient 6 Peritonitis Mixed flora Meropenem plus amikacin
Patient 7 Biliary tract Escherichia coli Meropenem
Patient 8 Peritonitis Negative Meropenem
Patient 9 Peritonitis Enterococcus faecium, Candida albicans, Candida tropicalis Meropenem plus vancomycin plus anidulafungin
Toraymyxin treated patients
Patient 1 Urinary tract Escherichia coli Meropenem plus vancomycin
Patient 2 Peritonitis Escherichia coli/Enterococcus faecium/Bacteroides fragilis Piperacillin–tazobactam plus fluconazole
Patient 3 Biliary tract Escherichia coli/Streptococcus anginosus Piperacillin–tazobactam plus fluconazole
Patient 4 Peritonitis Escherichia coli Piperacillin–tazobactam plus fluconazole
Patient 5 Biliary tract Escherichia coli Meropenem
Patient 6 Peritonitis Mixed flora Piperacillin–tazobactam
Patient 7 Peritonitis Negative Piperacillin–tazobactam
Patient 8 Urinary tract Escherichia coli Meropenem
Patient 9 Peritonitis Mixed flora Piperacillin–tazobactam
Page 5 of 9Navas et al. Ann. Intensive Care (2018) 8:121
Table 3 Clinical results
MAP mean arterial pressure, ISVR indexed systemic vascular resistances, EVLWi extravascular lung water index, PVPi pulmonary vascular permeability index, GEDVi global end-diastolic volume index
* p < 0.05
Patients in hemoperfusion group/control group
Day 1 (pretreatment)(n = 18)9/9
Day 2 (posttreatment)(n = 16)9/7
Day 3(n = 15)9/6
Day 4(n = 15)9/6
Day 5(n = 15)9/6
Respiratory data, mean (SD)
Pa02/Fi02 ratio
Hemoperfusion group 260 (75) 260 (110) 242 (92) 222 (61) 251 (78)
Control group 209 (117) 218 (114) 247 (45) 215 (37) 185 (54)
HC03 (mmHg)
Hemoperfusion group 17.4 (2.8) 22 (3) 25.4 (4.8) 28.6 (2.7) 28.6 (1.6)
Control group 16 (4.7) 24 (4) 26.7 (3.6) 28.6 (2.7) 29.7 (0.5)
pH
Hemoperfusion group 7.3 (0.05) 7.4 (0.08) 7.4 (0.08) 7.5 (0.04) 7.5 (0.05)
Control group 7.2 (0.1)* 7.3 (0.12) 7.4 (0.04) 7.4 (0.04) 7.4 (0.06)
Hemodynamic data, mean (SD)
MAP (mmHg)
Hemoperfusion group 74.7 (8.4) 76.6 (14) 78.5 (11.2) 80.2 (15) 78.3 (11.7)
Control group 78.8 (13.4) 83.3 (15) 90.3 (9.5) 87.8 (14.4) 78.7 (10)
Noradrenaline (µg/kg/min)
Hemoperfusion group 0.8 (0.5) 0.7 (0.7) 0.5 (0.6) 0.3 (0.7) 0.4 (0.8)
Control group 1.3 (0.7) 1.1 (1.2) 0.3 (0.6) 0.2 (0.4) 0.2 (0.3)
Dobutamine (µg/kg/min)
Hemoperfusion group 1.5 (4.6) 3.2 (5.4) 3.1 (4.6) 2 (4.5) 2.1 (4.5)
Control group 0.5 (1.6) –* 1.6 (4) 1.6 (4) 0.8 (2.0)
Central venous oxygen Saturat
Hemoperfusion group 72 (6.5) 62 (12) 61,4 (5.5) 63.5 (8.6) 64.1 (15)
Control group 66 (8) 61 (8.6) 65.2 (7.7) 57.3 (9.9) 58.4 (5.3)
Cardiac index (L/min/m2)
Hemoperfusion group 3.3 (0.76) 2.6 (0.6) 3.2 (0.9) 3.1 (0.9) 3.8 (1.2)
Control group 3.3 (1.18) 2.5 (0.6) 2.6 (0.7) 2.7 (0.8) 2.5 (0.5)
ISVR (dynes/cm−5/m2)
Hemoperfusion group 1113.6 (430) 1468.8 (751) 1115.4 (420) 1097 (522) 1261 (699)
Control group 1689.9 (881) 2745.5 (1719) 2528.5 (585)* 2427 (483)* 2600 (400)
EVLWi (ml/kg)
Hemoperfusion group 7.5 (3) 6.9 (2.7) 6.3 (1.8) 7 (2.7) 7.5 (4)
Control group 7.7 (3.5) 6.3 (1.8) 6.1 (1.7) 5.7 (1.9) –
PVPi
Hemoperfusion group 2.1 (0.6) 1.8 (0.7) 1.8 (0.57) 1.9 (0.7) 1.75 (0.75)
Control group 1.6 (0.1) 1.6 (0.2) 1.7 (0.3) 1.3 (0.2) –
GEDVi
Hemoperfusion group 950 (380 965 (324) 872 (277) 832 (196) 978 (408)
Control group 1000 (392) 761 (348) 1107 (532) 809 (99) –
Renal data, mean (SD)
Serum creatinine (mg/dl)
Hemoperfusion group 3.8 (2) 2.6 (1.2) 1.9 (0.9) 1.6 (0.6) 1.8 (0.8)
Control group 2.9 (1.3) 1.8 (1.2) 1.5 (0.9) 1.5 (1) 1.5 (0.9)
Azotemia (mg/dl)
Hemoperfusion group 119 (36) 84 (33) 67.5 (28) 67 (28) 79 (37)
Control group 106 (38) 62 (19) 62 (19) 66 (30) 82 (41)
Page 6 of 9Navas et al. Ann. Intensive Care (2018) 8:121
50 ICUs in the USA and Canada, randomized patients with septic shock and EAA > 0.6 to treatment with or without two sessions of hemoperfusion with polymyxin B. Although the initial results showed no reduction in overall mortality at day 28 in the treated group ver-sus the control group, a post hoc analysis limited to the 244 patients with EAA between 0.6 and 0.9 (115 treated patients vs. 129 controls) found a significant reduction in absolute mortality (26.1% in the treated group vs. 36.8% in controls) and a relative mortality reduction of 30%. Romaschin et al. [32] suggest that current polymyxin B
filters are probably ineffective in patients with EAA > 0.9. In our study, all patients had initial EAA between 0.6 and 0.9, and none had EAA > 0.9 at inclusion.
After the publication of these trials in 2017, two sys-tematic reviews were published. In the first, Chang et al. [33] found that polymyxin B hemoperfusion reduced mortality in selected patients with intermediate and high risk of disease severity. In the second, Fujii et al. [34] con-cluded that the treatment does not decrease mortality or the number of organ failures and should therefore not be used routinely.
Table 4 Biological results
CRP C-reactive protein, PCT procalcitonin, Adren adrenomedullin
* p < 0.05
Patients in hemoperfusion group/control group
Day 1(n = 18)9/9
Day 2 (posttreatment)(n = 16)9/7
Day 3(n = 15)9/6
Day 4(n = 15)9/6
Day 5(n = 15)9/6
Lactate (mg/dl), mean (SD)
Hemoperfusion group 56.7 (33) 41.3 (32.7) 35.2 (34) 26.6 (10.4) 21.3 (14.2)
Control group 56.4 (26.7) 37.7 (26.3) 19.8 (7.5) 17 (3.8) 17 (3.8)
CRP (mg/dl), mean (SD)
Hemoperfusion group 33.5 (10.7) 30.4 (14.9) 27.6 (19.5) 16.8 (9.7) 14.1 (7.8)
Control group 24.6 (10.3) 24.1 (10.3) 12.9 (5) 8.2 (5.7) 7.4 (9.3)
PCT (ng/ml), mean (SD)
Hemoperfusion group 83 (92) 43 (59) 33.2 (45.8) 19.9 (27.2) 17.7 (23)
Control group 71 (67) 40 (24) 27.9 (21.2) 12.7 (12.4) 6.1 (5.4)
Adren (nmol/l), mean (SD)
Hemoperfusion group 17.8 (4.8) 10.9 (3.7) 7.2 (2.4) 5.6 (2.1) 6.4 (1.8)
Control group 17.7 (5.8) 12.3 (3.3) 6.7 (1.8) 4.1 (2) 4.8 (3.6)
IL-6 (ng/ml), mean (SD)
Hemoperfusion group 1115 (777) 387 (548) 317 (400) 94.5 (120) 367.8 (577)
Control group 8302 (6830)* 3698 (6285) 634 (1292) 9.8 (4.3) 19.2 (14.2)
IL-8 (pg/ml), mean (SD)
Hemoperfusion group 566.8 (291) 271 (278) 144 (88) 141 (118) 150 (126)
Control group 2766 (3684) 1412 (3090) 781 (1254) 29.4 (8.6) 40.6 (57)
IL-10 (pg/ml), mean (SD)
Hemoperfusion group 427.1 (339) 178.5 (122) 283.8 (287) 162.7 (164) 131.7 (198)
Control group 3411 (8650) 335 (361) 231 (323) 54.4 (47.4) 77 (95)
IL-1β (pg/ml), mean (SD)
Hemoperfusion group 4.7 (1.8) 3.7 (0.93) 3.6 (0.6) 3.4 (0.4) 5.3 (3)
Control group 35.1 (90) 2.1 (1.5) 89 (195) 2.4 (1.5) 1.8 (1.5)
TNFα (pg/ml), mean (SD)
Hemoperfusion group 177.2 (111) 56.8 (26.7) 78.2 (108) 47.7 (36.5) 49.1 (51)
Control group 386 (493) 76.6 (72.8) 170 (355) 28.7 (20.7) 27.4 (8)
SuPAR (ng/ml), mean (SD)
Hemoperfusion group 23.1 (9.8) 24.5 (12) 24.9 (17.5) 21.8 (13.3) 16.1 (6.8)
Control group 34.2 (21.2) 24.8 (9.4) 23.8 (15.3) 18.2 (11.5) 15.3 (5.4)
NGAL (ng/ml), mean (SD)
Hemoperfusion group 2331 (1028) 1725.2 (645) 1261 (382) 987 (346) 749 (197)
Control group 2264 (1444) 1284 (1037) 783 (671) 574 (593) 600 (674)
Page 7 of 9Navas et al. Ann. Intensive Care (2018) 8:121
Our study selected patients who met stringent inclu-sion criteria (septic shock probably due to gram-nega-tive bacteria with elevated endotoxemia and multiorgan failure) and received homogeneous treatment, includ-ing CRRT. The mortality in our study (38.9%) is close to the range reported for similar patients in other studies. Microbiology studies confirmed that infections were pre-dominantly due to gram-negative microorganisms. All patients received appropriate empirical treatment with broad-spectrum antibiotics and early source control. The same procedures were carried out in all patients who received hemoperfusion with polymyxin B, and no com-plications of this treatment were observed. Our failure to find clinical improvements with hemoperfusion with polymyxin B is likely due to the small sample size limiting the statistical power. Furthermore, concomitant CRRT might mask the effect of hemoperfusion with polymyxin B.
A systematic review of 41 recent articles dealing with the removal of cytokines with extracorporeal techniques found that standard hemofiltration was generally poor at removing cytokines and that high cut-off hemofiltration techniques with large-pore filters were consistently bet-ter [35]. Although the technique used in our study was standard hemofiltration, we found a significant decrease in cytokines in both groups, probably derived from the global effect of the treatment administered. In 2009, Payen et al. [6], analyzing several cytokines (IL-6, IL-1ra, and MCP-1) in patients with septic shock randomized to early CRRT or no CRRT, found no differences between the two groups regarding the decrease in cytokines. Ana-lyzing cytokines in patients in the ABDOMIX study,
Coudroy et al. [36] found no differences between patients treated with hemoperfusion with polymyxin B and con-trols. Thus, it seems likely that the decrease seen in both groups in our study was due to the overall effects of treat-ment rather than to CRRT alone.
Although CRRT techniques have long been used in the treatment of septic shock, the best modality, dose, and time to start remain unclear. In our study, in which 72% of patients had baseline RIFLE scores of Failure, CRRT con-sisted of hemofiltration at a dose of 35 ml/kg/h and was started within the first 24 h of septic shock. Recently, the ELAIN [37] and AKIKI [5] studies found differences in survival in relation to whether CRRT was initiated early or late. Another recent study found very early onset was associated with poorer outcome due to incorrect dosing of antibiotics [38] and side effects. Studies testing high dialysis flows (> 50 ml/kg/h) in septic patients have failed to find improvements in outcomes [4, 6, 7], and the Kid-ney Disease: Improving Global Outcomes guidelines [39] recommend prescribing an effluent volume of 30–35 ml/kg/h to achieve a flow of 20–25 ml/kg/h. Within the over-all management of septic patients, it is difficult to discern the specific effects of CRRT on outcomes.
To our knowledge, no published studies have prospec-tively compared septic patients undergoing CRRT and polymyxin B hemoperfusion versus patients undergoing CRRT alone. CRRT might influence the effects of poly-myxin B hemoperfusion, thus making it difficult to find significant differences between the two groups. Moreo-ver, the reduction in mortality achieved with improve-ments in the overall management of septic patients also makes it more difficult to find relevant differences related to specific treatments.
Our study has some important limitations. Patients were not randomized to the hemoperfusion and control groups. Nevertheless, to minimize the selection bias, we included all patients consecutively according to strin-gent inclusion criteria and applied homogeneous treat-ment protocols; moreover, the demographic, clinical, and hemodynamic parameters were similar in the two groups. Another limitation is that the inclusion criterion requiring patients to need CRRT might have delayed the initiation of hemoperfusion. Starting hemoperfusion treatment alone would have allowed earlier initiation of hemoperfusion, as in the recently EUPHRATES study [21], where the mean time to starting hemoperfusion was 4 h. However, when our study was designed, our ethics committee deemed the evidence supporting early hemoperfusion insufficient. Nevertheless, our inclu-sion criterion for CRRT was a RIFLE score of Injury or worse, and 88% of patients started CRRT (with or with-out hemoperfusion) within 24 h of ICU admission,
Fig. 1 Endotoxin activity levels in the first 5 days. On day 3, EAA had decreased significantly in both groups with respect to the baseline value: hemoperfusion group 0.54 versus 0.78; p = 0.02 and control group 0.57 versus 0.77; p = 0.05. On day 5, we observe a significantly decreased EAA with respect to the baseline value in the hemoperfusion group 0.58 versus 0.78; p = 0.03
Page 8 of 9Navas et al. Ann. Intensive Care (2018) 8:121
which is similar to the inclusion criteria of the EUPHAS study [18]. Finally, the early death of three patients in the control group resulted in missing values that could have affected some of our results (mortality, length of mechanical ventilation and CRRT, values of endotoxin and cytokines).
Most studies published on the treatment with poly-myxin B do not analyze endotoxin activity levels, and those that do only do not present a control group to com-pare. In this study, we analyzed the EAA in both groups, which does not measure the absolute value of plasma endotoxins. Clearly, endotoxin levels decreased faster in patients who received hemoperfusion with polymyxin B than in those treated with CRRT alone. Since the filter used for CRRT (AN69) has a very low endotoxin adsorp-tion capacity, we can infer that the decrease in endotoxin in the hemoperfusion group was a consequence of the polymyxin B cartridge. On the other hand, although the hemoperfusion group had lower EAA values than the control group on day 3, the two values are similar. These findings are very similar to those of the recent EUPHRA-TES study, probably because the EAA is an inadequate reflection of the absolute endotoxin value.
Taken together with the results of recently published trials, our results suggest that further studies are nec-essary to clarify the efficacy of hemoperfusion with polymyxin B, especially in patients with elevated blood endotoxin level and multiorgan failure.
ConclusionsHemoperfusion with polymyxin B decreases blood endo-toxin levels, although we found no improvement in clini-cal and biological parameters. Further studies in larger samples of specific patient populations are necessary to assess the efficacy of polymyxin B hemoperfusion.
Additional file
Additional file 1. Biomarkers analysis methods.
AbbreviationsEAA: Endotoxin activity assay; CRRT : Continuous renal replacement therapy; CRP: C-reactive protein; PCT: Procalcitonin; AKI: Acute kidney injury; RIFLE: Risk, Injury, Failure, Loss of Kidney Function, and End-Stage Kidney Disease; APACHE: Acute Physiology and Chronic Health Evaluation; SOFA: Sequential Organ Failure Assessment; NGAL: Neutrophil gelatinase-associated lipocalin; SuPAR: Soluble urokinase-type plasminogen activator receptor.
Authors’ contributionsAll authors contributed to the elaboration of this manuscript. AN, RF, and AA designed and coordinated the study. AN, GGo, and GGi ordered and applied the toraymyxin treatment. MLM and JM included patients and collected data. DS performed the statistical analysis. AN, RF, and AA analyzed the results and drafted the manuscript. All authors read and approved the final manuscript.
Author details1 Critical Care Center, Parc Tauli Hospital Universitari, Institut d’Investigació i Innovació Parc Taulí I3PT, Universitat Autònoma de Barcelona, Sabadell, Spain. 2 CIBER Respiratory Diseases, Madrid, Spain. 3 Intensive Care Department, Vall d’Hebron University Hospital, Shock, Organ Dysfunction and Resuscitation Research Group, Vall d’Hebron Research Institute, Barcelona, Spain. 4 Depart-ment of Intensive Care, Hospital Universitari General de Catalunya, Barcelona, Spain. 5 Unitat de Suport a la Investigación Clínica, Vall d’Hebron Institut de Recerca, Barcelona, Spain. 6 Servei de Medicina Intensiva, Hospital Universitari Sant Joan de Reus, Tarragona, Spain. 7 Epidemiology and Assessment Unit, Fundació Parc Tauli, Universitat Autònoma de Barcelona, Sabadell, Spain.
AcknowledgementsThe author gratefully acknowledges all the nurses and the physicians of the ICU at the Hospital de Sabadell for their contributions in the care of the study patients.
Competing interestsThe authors declare that they have no competing interests.
Availability of data and materialsThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Consent for publicationNot applicable.
Ethics approval and consent to participateOur centers’ ethics committees approved the study protocol (CIR09065/CEAAH 2407), and all patients provided written informed consent.
FundingThis study was supported by an unrestricted, investigator-initiated, research grant from Toray Industries. Toray Industries had no role in the conception, design, or conduct of the study; collection, management, analysis, interpreta-tion, or presentation of the data; or preparation, review, or approval of the manuscript.
Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims in pub-lished maps and institutional affiliations.
Received: 12 August 2018 Accepted: 30 November 2018
References 1. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR.
Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 2001;29:1303–10.
2. Opal SM. The host response to endotoxin, antilipopolysaccharide strategies, and the management of severe sepsis. Int J Med Microbiol. 2007;297:365–77.
3. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al. Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock. Intensive Care Med. 2017;43:304–77.
4. Palevsky PM, Zhang JH, O’Connor TZ, Chertow GM, Crowley ST, Choud-hury D, et al. VA/NIH Acute Renal Failure Trial Network. Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med. 2008;359:7–20.
5. Gaudry S, Hajage D, Schortgen F, Martin-Lefevre L, Pons B, Boulet E, et al. Initiation strategies for renal-replacement therapy in the intensive care unit. N Engl J Med. 2016;375:122–33.
6. Payen D, Mateo J, Cavaillon JM, Fraisse F, Floriot C, Vicaut E. Impact of continuous venovenous hemofiltration on organ failure during the early phase of severe sepsis: a randomized controlled trial. Crit Care Med. 2009;37:803–10.
Page 9 of 9Navas et al. Ann. Intensive Care (2018) 8:121
7. Joannes-Boyau O, Honoré PM, Perez P, Bagshaw SM, Grand H, Canivet JL, et al. High-volume versus standard-volume haemofiltration for septic shock patients with acute kidney injury (IVOIRE study): a multicentre randomized controlled trial. Intensive Care Med. 2013;39:1535–46.
8. Rimmelé T, Kellum JA. Clinical review: blood purification for sepsis. Crit Care. 2011;15:205.
9. Romaschin AD, Harris DM, Ribeiro MB, Paice J, Foster DM, Walker PM, Marshall JC. A rapid assay of endotoxin in whole blood using autolo-gous neutrophil dependent chemiluminescence. J Immunol Methods. 1998;212:169–85.
10. Marshall JC, Walker PM, Foster DM, Harris D, Ribeiro M, Paice J, et al. Meas-urement of endotoxin activity in critically ill patients using whole blood neutrophil dependent chemiluminescence. Crit Care. 2002;6:342–8.
11. Marshall JC, Foster D, Vincent JL, Cook DJ, Cohen J, Dellinger RP, et al. Diagnostic and prognostic implications of endotoxemia in critical illness: results of the MEDIC study. J Infect Dis. 2004;190:527–34.
12. Aoki H, Kodama M, Tani T, Hanasawa K. Treatment of sepsis by extracor-poreal elimination of endotoxin using polymyxin B-immobilized fiber. Am J Surg. 1994;167:412–7.
13. Ronco C, Brendolan A, Dan M, Piccinni P, Bellomo R, De Nitti C, et al. Adsorption in sepsis. Kidney Int Suppl. 2000;76:S148–55.
14. Shoji H, Tani T, Hanasawa K, Kodama M. Extracorporeal endotoxin removal by polymyxin-B immobilized fiber cartridge: designing and antiendotoxin efficacy in the clinical application. Ther Apher. 1998;2:3–12.
15. Shoji H. Extracorporeal endotoxin removal for the treatment of sepsis: endotoxin adsorption cartridge (Toraymyxin). Ther Apher Dial. 2003;7:108–14.
16. Vincent JL, Laterre PF, Cohen J, Burchardi H, Bruining H, Lerma FA, et al. A pilot-controlled study of a polymyxin B-immobilized hemoperfusion cartridge in patients with severe sepsis secondary to intra-abdominal infection. Shock. 2005;23:400–5.
17. Cruz DN, Perazella MA, Bellomo R, de Cal M, Polanco N, Corradi V, et al. Effectiveness of polymyxin B-immobilized fiber column in sepsis; a systematic review. Crit Care. 2007;11:R47.
18. Cruz DN, Antonelli M, Fumagalli R, Foltran F, Brienza N, Donati A, et al. Early use of polymyxin B hemoperfusion in abdominal septic shock: the EUPHAS randomized controlled trial. JAMA. 2009;301:2445–52.
19. Payen DM, Guilhot J, Launey Y, Lukaszewicz AC, Kaaki M, Veber B, et al. Early use of polymyxin B hemoperfusion in patients with septic shock due to peritonitis: a multicenter randomized control trial. Intensive Care Med. 2015;41:975–84.
20. Zhou F, Peng Z, Murugan R, Kellum JA. Blood purification and mor-tality in sepsis: a meta-analysis of randomized trials. Crit Care Med. 2013;41:2209–20.
21. Dellinger RP, Bagshaw SM, Antonelli M, Foster D, Klein DJ, Marshall J, et al. Effect of targeted polymyxin B hemoperfusion on 28-day mortality in patients with septic shock and elevated endotoxin level. The EUPHRATES randomized clinical trial. JAMA. 2018;320(14):1455–63.
22. Esteban E, Ferrer R, Alsina L, Artigas A. Immunomodulation in sepsis: the role of endotoxin removal by polymyxin b-immobilized cartridge. Media-tors Inflamm. 2013. https ://doi.org/10.1155/2013/50753 9.
23. Tani T, Hanasawa K, Kodama M, Imaizumi H, Yonekawa M, Saito M, et al. Correlation between plasma endotoxin, plasma cytokines, and plas-minogen activator inhibitor-1 activities in septic patients. World J Surg. 2001;25:660–8.
24. Nakamura T, Matsuda T, Suzuki Y, Shoji H. Polymyxin B-immobilized fiber hemoperfusion in patients with sepsis. Dial Transplant. 2003;32:602–7.
25. Zagli G, Bonizzoli M, Spina R, Cianchi G, Pasquini A, Anichini V, et al. Effects of hemoperfusion with an immobilized polymyxin-B fiber column on cytokine plasma levels in patients with abdominal sepsis. Minerva Anest-esiol. 2010;76:405–12.
26. Shimizu T, Hanasawa K, Sato K, Umeki M, Koga N, Naganuma T, et al. The clinical significance of serum procalcitonin levels following direct hemop-erfusion with polymyxin B-immobilized fiber column in septic patients with colorectal perforation. Eur Surg Res. 2009;42:109–17.
27. Yaroustovsky M, Abramyan M, Krotenko N, Popov D, Plyushch M, Popok Z. Endotoxin adsorption using polymyxin B immobilized fiber cartridges in severe sepsis patients following cardiac surgery. Int J Artif Organs. 2014;37:299–307.
28. Yaroustovsky M, Abramyan M, Krotenko N, Popov D, Plyushch M, Rogals-kaya E, Nazarova H. Combined extracorporeal therapy for severe sepsis in patients after cardiac surgery. Blood Purif. 2014;37:39–46.
29. Bobek I, Gong D, De Cal M, Cruz D, Chionh CY, Haapio M, et al. Removal of neutrophil gelatinase-associated lipocalin by extracorporeal therapies. Hemodial Int. 2010;14:302–7.
30. Suberviola B, Castellanos-Ortega A, Ruiz Ruiz A, Lopez-Hoyos M, Santi-bañez M. Hospital mortality prognostication in sepsis using the new bio-markers suPAR and proADM in a single determination on ICU admission. Intensive Care Med. 2013;39:1945–52.
31. Iwagami M, Yasunaga H, Doi K, Horiguchi H, Fushimi K, Matsubara T, et al. Postoperative polymyxin b hemoperfusion and mortality in patients with abdominal septic shock: a propensity-matched analysis. Crit Care Med. 2014;42:1187–93.
32. Romaschin A, Obieuz-Foster C, Shoji H, Klein D. Novel insights into the direct removal of endotoxin by polymyxin B hemoperfusion. Blood Purif. 2017;44:193–7.
33. Chang T, Tu YK, Lee CT, Chao A, Huang CH, Wang MJ, Yeh YC. Effects of Polymyxin B hemoperfusion on mortality in patients with severe sepsis and septic shock: a systemic review, meta-analysis update, and disease severity subgroup meta-analysis. Crit Care Med. 2017;45:e858–64.
34. Fujii T, Ganeko R, Kataoka Y, Furukawa TA, Featherstone R, Doi K, et al. Poly-myxin B-immobilized hemoperfusion and mortality in critically ill adult patients with sepsis/septic shock: a systematic review with meta-analysis and trial sequential analysis. Intensive Care Med. 2018;44:167–78.
35. Atan R, Crosbie DC, Bellomo R. Techniques of extracorporeal cytokine removal: a systematic review of human studies. Ren Fail. 2013;35:1061–70.
36. Coudroy R, Payen D, Launey Y, Lukaszewicz AC, Kaaki M, Veber B, et al. Modulation by polymyxin-B hemoperfusion of inflammatory response related to severe peritonitis. Shock. 2017;47:93–9.
37. Zarbock A, Kellum JA, Schmidt C, Van Aken H, Wempe C, Pavenstädt H, et al. Effect of early vs delayed initiation of renal replacement therapy on mortality in critically ill patients with acute kidney injury: the ELAIN randomized clinical trial. JAMA. 2016;315:2190–9.
38. Bellomo R, Kellum J, Ronco C, Wald R, Martensson J, Maidel M, et al. Acute kidney injury in sepsis. Intensive Care Med. 2017. https ://doi.org/10.1007/s0013 4-017-4755-7.
39. Kellum J, Lameire N, Aspelin P, Barsoum R, Burdmann E, Goldstein S, et al. Kidney disease: improving global outcomes (KDIGO) Acute Kidney Injury Work Group. KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl. 2012;2:1–126.
Page 1 of 4
© Journal of Emergency and Critical Care Medicine. All rights reserved. J Emerg Crit Care Med 2018;2:39jeccm.amegroups.com
Sepsis continues to be an important global public health problem with persisting elevated mortality rates. The reported incidence of sepsis is increasing, but considerable international variation in incidence (6–27%) of sepsis has been reported (1-4). Variations in the definition of sepsis and septic shock can explain differences in mortality rates among septic patients (as high as 80%) (5-7). Since the time of the original sepsis definitions in 1991 (and refinement in 2001, also known as sepsis-2 definition), clinical outcomes from sepsis have improved (1,2,8) because of their application and the interventions associated with their use. Nevertheless, all recent multinational trials assessing different treatments have failed to improve survival (9-11). Defining sepsis is often difficult because of the wide variation in patient characteristics, clinical presentation, and the varied standard-of-care found across the world. As suggested first in 2013 by Vincent (12), and later in 2014 by Gattinoni (13), it is time to change the sepsis definitions and create a better classification of sepsis severity. Following this, sepsis definitions were updated in “The Third International Consensus Definitions for Sepsis and Septic Shock” (sepsis-3) (7). Sepsis is now defined as ‘a life-threatening organ dysfunction caused by a dysregulated host response to infection’. For identifying organ dysfunction, the authors stablished an increase in the Sequential [sepsis-related] Organ Failure Assessment (SOFA) score of 2
points, which is associated in international databases with an in-hospital mortality of more than 10% (14). Some patients with sepsis develop septic shock, a more severe stage characterized by circulatory and cellular metabolism abnormalities identified by the need of vasopressor therapy requirement to maintain a mean arterial pressure of 65 mmHg, and serum lactate level greater than 2 mmol/L (>18 mg/dL) after adequate fluid resuscitation (15). The new definition excludes the concept of systemic inflammatory response syndrome (SIRS) and introduces a new score named quick SOFA (q-SOFA) as tool for to identify infected patients with high risk of death.
In our opinion, the new sepsis definition is necessary since it provides uniformity in clinical practice as well as for epidemiological studies and future trials. In regular clinical practice, we continue considering the SIRS criteria as indicative of infection, and if they are present, we look for severity data using q-SOFA outside the ICU and the SOFA score inside the ICU. In fact, many of the studies that show that early treatment of sepsis decreases mortality are performed in patients with organ dysfunction (severe sepsis and septic shock) (16,17).
How this new sepsis definition is going to affect the epidemiology of sepsis remains to be seen. In this context, Shankar-Hari and colleagues (18), who participated prominently in the sepsis-3 definitions, analyzed the effect
Editorial
Improving knowledge about sepsis 3 definition in critically ill patients: new insights
María Luisa Martínez1, Juan Carlos Ruiz-Rodríguez2,3, Ricard Ferrer2,3
1Department of Intensive Care, Hospital Universitario General de Catalunya, Barcelona, Spain; 2Department of Intensive Care, Vall d’Hebron
University Hospital, Barcelona, Spain; 3Shock, Organ Dysfunction and Resuscitation Research Group, Vall d’Hebron Research Institute, Barcelona,
Spain
Correspondence to: Ricard Ferrer. Department of Intensive Care, Vall d’Hebron University Hospital, Passeig de la Vall d’Hebron, 119-129, 08035
Barcelona, Spain. Email: [email protected].
Comment on: Shankar-Hari M, Harrison DA, Rubenfeld GD, et al. Epidemiology of sepsis and septic shock in critical care units: comparison between
sepsis-2 and sepsis-3 populations using a national critical care database. Br J Anaesth 2017;119:626-36.
Provenance: This is a Guest Editorial commissioned by the Section Editor Biao Zhang, MD (Department of Critical Care Medicine, Suzhou
Integrated Chinese and Western Medicine Hospital, Suzhou, China).
Received: 08 April 2018; Accepted: 17 April 2018; Published: 30 April 2018.
doi: 10.21037/jeccm.2018.04.05
View this article at: http://dx.doi.org/10.21037/jeccm.2018.04.05
Journal of Emergency and Critical Care Medicine, 2018Page 2 of 4
© Journal of Emergency and Critical Care Medicine. All rights reserved. J Emerg Crit Care Med 2018;2:39jeccm.amegroups.com
that the new sepsis definition had on incidence, mortality, and another epidemiological variable by comparing sepsis-2 severe sepsis/septic shock and sepsis-3 sepsis/septic shock populations using a national ICU database of 654,918 consecutive admissions to 189 adult English ICUs (that covers 96% of the adult general ICUs). To define sepsis-2 severe sepsis, the authors defined a SOFA score of >1 for organ dysfunction, and to define sepsis-2 septic shock they used cardiovascular a SOFA score of >1 or a lactate level >4 mmol L−1. The authors compared the epidemiology of sepsis based on sepsis-2 severe sepsis/septic shock and sepsis-3 sepsis/septic shock between January 2011 and December 2015.
Along the 5-year period of study, 654,918 patients were admitted in the participating ICUs, classified according the definitions as sepsis-2 severe sepsis 197,724 (30.2%) cases and as sepsis-3 197,142 (30.1%) cases. Sepsis-2 severe sepsis and sepsis-3 sepsis definitions were overlapped in 92% cases; these included a similar age, comorbidities, illness severity scores, infection source, and even a similar ICU and hospital mortality. In addition, when the epidemiology of sepsis-3 sepsis in the ICU setting is compared with the previously described for sepsis-2 severe sepsis the results were equivalent (3,14). Shankar-Hari and colleagues conclude that the new definition (sepsis-3) was reliable detecting a similar amount than the previous (sepsis-2) classification, with similar rates of mortality. These results are expected as diagnostic criteria are similar and the authors used the same score (SOFA) to identify severe sepsis (sepsis-2) and sepsis (sepsis-3). Recently, Williams and colleagues in a prospective study with 8871 patients from the emergency department, also found that overall organ dysfunction according to both definitions estimated similar mortality risk [12.5% (95% CI, 10.8–14.2%) vs. 11.4% (95% CI, 10.1–12.8%)]. In contrast, in this study 29% of patients with identified using the new criteria did not meet the previous criteria (19). Some authors argued about the lack of correlation between the previous concept of severe sepsis and the new definition of sepsis (20): some clinical situations could be included by the new definition, such as organ failure without hypotension or hyperlactatemia.
The new definition excludes the concept of SIRS and does not include the concept of sepsis without organ dysfunction. This has generated controversy since some authors suggest that, ideally, patients at risk of sepsis should be identified before organ dysfunction is established (21-24). In this regard, Shankar-Hari and colleagues described that only 4.1% of sepsis-2 severe sepsis patients do
not meet the stricter criteria for sepsis-3 organ dysfunction and 4.0% of sepsis-3 patients were SIRS negative. In their analysis, as most patients with organ dysfunction also tend to have SIRS, discarding SIRS as the initial step for sepsis diagnosis (in patients in the first 24 h of ICU admission) does not alter the epidemiology of sepsis.
One important finding is that the proportion of patients with septic shock differs between sepsis-2 and sepsis-3 definitions. Among patients admitted with sepsis, there were 153,257 (77.5%) sepsis-2 septic shock and 39,262 (19.9%) sepsis-3 septic shock, being 0.01% negative for systemic inflammation criteria. The severity scores, lactate levels and hospital mortality were higher in sepsis-3 septic shock. Thus, the sepsis-3 septic shock definition selects a very critically ill subpopulation. Recently, Driessen and colleagues (25) prospectively analyzed a cohort of 632 ICU septic patients: 300 patients (48.4%) according to the new definition and 482 (76.3%) had septic shock according to the former criteria. Patients meeting the sepsis-3 septic shock criteria had a higher mortality than patients meeting the old septic shock definition (38.9% vs. 34.0%). The findings of these two recent studies support the objectives of the Task Force to select a very severe and homogeneous septic shock populations.
Shankar-Hari and colleagues also calculate trends and risk factors for adjusted and unadjusted hospital mortality using four logistic-regression models that include a large number of confusion factors as illness severity. They found an increase in sepsis incidence and an improvement in hospital mortality. In addition, age and comorbidity are factors that increase the incidence and mortality of Sepsis-3 sepsis and septic shock, similar to previous epidemiologic studies (26). In the regression models, the highest increment in predictive validity was for sepsis-3 septic shock, even when adjusting for severity.
In summary, Shankar-Hari and coworkers present us a well-designed study, using an observational high-quality database. They present one of the first direct comparisons of old and new sepsis epidemiology using in England. For adult ICU admissions with sepsis, the new sepsis-3 sepsis definition does not involve a big change in epidemiology data. This study confirms that new septic shock is a population with high risk of death. This can be interpreted as better predictive validity and argued as prognostic enrichment maybe resulting better patient selection for clinical trials, and therefore could be considered as a risk-stratification screening-tool. For designing clinical trials, will be important to assess the magnitude of mortality
Journal of Emergency and Critical Care Medicine, 2018 Page 3 of 4
© Journal of Emergency and Critical Care Medicine. All rights reserved. J Emerg Crit Care Med 2018;2:39jeccm.amegroups.com
risk reduction that is viable to be effectively reduced. This new definition will create different challenges for both clinicians and researchers. As we further explore a more uniform epidemiological description of sepsis, we will better understand it.
Acknowledgements
None.
Footnote
Conflicts of Interest: The authors have no conflicts of interest to declare.
References
1. Yébenes JC, Ruiz-Rodríguez JC, Ferrer R, et al. Epidemiology of sepsis in Catalonia: analysis of incidence and outcomes for patients in an european setting. Ann Intensive Care 2017;7:19.
2. Kaukonen KM, Bailey M, Suzuki S, et al: Mortality related to severe sepsis and septic shock among critically ill patients in Australia and New Zealand, 2000-2012. JAMA 2014;311:1308-16.
3. Gaieski DF, Edwards JM, Kallan MJ, et al. Benchmarking the incidence and mortality of severe sepsis in the United States. Crit Care Med 2013;41:1167-74.
4. Shankar-Hari M, Harrison DA, Rowan KM. Differences in impact of definitional elements on mortality precludes international comparisons of sepsis epidemiology-a cohort study illustrating the need for standardized reporting. Crit Care Med 2016;44:2223-30.
5. Vincent JL, Marshall JC, Namendys-Silva SA, et al. Assessment of the worldwide burden of critical illness: The intensive care over nations (ICON) audit. Lancet Respir Med 2014;2:380-6.
6. Fleischmann C, Scherag A, Adhikari NK, et al; International Forum of Acute Care Trialists: Assessment of global incidence and mortality of hospital-treated sepsis. Current estimates and limitations. Am J Respir Crit Care Med 2016;193:259-72.
7. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (sepsis-3). JAMA 2016;315:801-10.
8. Stevenson EK, Rubenstein AR, Radin GT, et al: Two decades of mortality trends among patients with severe sepsis: A comparative metanalysis. Crit Care Med
2014;42:625-31.9. Caironi P, Tognoni G, Masson S, et al. Albumin
replacement in patients with severe sepsis or septic shock. N Engl J Med 2014;370:1412-21.
10. Yealy DM, Kellum JA, Huang DT, et al. A randomized trial of protocol-based care for early septic shock. N Engl J Med 2014;370:1683-93.
11. ARISE Investigators, ANZICS Clinical Trials Group, Peake SL, et al. Goal-directed resuscitation for patients with early septic shock. N Engl J Med 2014;371:1496-506.
12. Vincent JL, Opal SM, Marshall JC, et al. Sepsis definitions: time for change. Lancet 2013;381:774-5.
13. Gattinoni L, Ranieri VM, Pesenti A. Sepsis: needs for defining severity. Intensive Care Med 2015;41:551-2.
14. Seymour CW, Liu VX, Iwashyna TJ, et al. Assessment of Clinical Criteria for Sepsis: For the Third International Consensus Definitions for Sepsis and Septic Shock (sepsis-3). JAMA 2016;315:762-74.
15. Shankar-Hari M, Phillips GS, Levy ML, et al. Developing a new definition and assessing new clinical criteria for septic shock: for the third international consensus definitions for sepsis and septic Shock (Sepsis-3). JAMA 2016;315:775-87.
16. Ferrer R, Martin-Loeches I, Phillips G, et al. Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guideline-based performance improvement program. Crit Care Med 2014;42:1749-55.
17. Barochia AV, Cui X, Vitberg D, et al. Bundled care for septic shock: an analysis of clinical trials. Crit Care Med 2010;38:668-78.
18. Shankar-Hari M, Harrison DA, Rubenfeld GD, et al. Epidemiology of sepsis and septic shock in critical care units: comparison between sepsis-2 and sepsis-3 populations using a national critical care database. Br J Anaesth 2017;119:626-36.
19. Williams JM, Greenslade JH, McKenzie JV, et al. Systemic inflammatory response syndrome, quick sequential organ function assessment, and organ dysfunction. Chest 2017;151:586-96.
20. Simpson SQ. New sepsis criteria: a change we should not make. Chest 2016;149:1117-8.
21. Farkas J. PulmCrit—Top ten problems with the new sepsis definition. February 29, 2016. Available online: https://emcrit.org/pulmcrit/ problems-sepsis-3-definition.
22. Carneiro AH, Povoa P, Gomes JA. Dear Sepsis-3, we are sorry to say that we don’t like you. Rev Bras Ter Intensiva 2017;29:4-8.
Journal of Emergency and Critical Care Medicine, 2018Page 4 of 4
© Journal of Emergency and Critical Care Medicine. All rights reserved. J Emerg Crit Care Med 2018;2:39jeccm.amegroups.com
23. Cortés-Puch I, Hartog CS. Opening the debate on the new sepsis definition change is not necessarily progress: revision of the sepsis definition should be based on new scientific insights. Am J Respir Crit Care Med 2016;194:16-8.
24. Machado FR, Salomao R, Pontes de Azevedo LC, et al. Why LASI did not endorse the new definitions of sepsis published today in JAMA. Available online: http://ilas.org.
br/assets/arquivos/upload/statement-en.pdf 25. Driessen RGH, van de Poll MCG, Mol MF, et al. The
influence of a change in septic shock definitions on intensive care epidemiology and outcome: comparison of sepsis-2 and sepsis-3 definitions. Infect Dis (Lond) 2018;50:207-13.
26. Mayr FB, Yende S, Angus DC. Epidemiology of severe sepsis. Virulence 2014;5:4-11.
doi: 10.21037/jeccm.2018.04.05Cite this article as: Martínez ML, Ruiz-Rodríguez JC, Ferrer R. Improving knowledge about sepsis 3 definition in critically ill patients: new insights. J Emerg Crit Care Med 2018;2:39.
Yébenes et al. Ann. Intensive Care (2017) 7:19 DOI 10.1186/s13613-017-0241-1
RESEARCH
Epidemiology of sepsis in Catalonia: analysis of incidence and outcomes in a European settingJuan Carlos Yébenes1,2,3,16*, Juan Carlos Ruiz‑Rodriguez4,5, Ricard Ferrer4,5,6, Montserrat Clèries7, Anna Bosch7, Carol Lorencio8, Alejandro Rodriguez2,9,10, Xavier Nuvials11,12, Ignacio Martin‑Loeches13, Antoni Artigas6,14,15 and SOCMIC (Catalonian Critical Care Society) Sepsis Working Group
Abstract
Background: Up‑to‑date identification of local trends in sepsis incidence and outcomes is of considerable public health importance. The aim of our study was to estimate annual incidence rates and in‑hospital mortality trends for hospitalized patients with sepsis in a European setting, while avoiding selection bias in relation to different complexity hospitals.
Methods: A large retrospective analysis of a 5‑year period (2008–2012) was conducted of hospital discharge records obtained from the Catalan Health System (CatSalut) Minimum Basic Data Set for Acute‑Care Hospitals (a mandatory population‑based register of admissions to all public and private acute‑care hospitals in Catalonia). Patients hospital‑ized with sepsis were detected on the basis of ICD‑9‑CM codes used to identify acute organ dysfunction and infec‑tious processes.
Results: Of 4,761,726 discharges from all acute‑care hospitals in Catalonia, 82,300 cases (1.72%) had sepsis diagnoses. Annual incidence was 212.7 per 100,000 inhabitants/year, rising from 167.2 in 2008 to 261.8 in 2012. Length of hos‑pital stay fell from 18.4 to 15.3 days (p < .00001), representing a relative reduction of 17%. Hospital mortality fell from 23.7 to 19.7% (p < .0001), representing a relative reduction of 16.9%. These differences were confirmed in the multi‑variate analysis (adjusted for age group, sex, comorbidities, ICU admission, emergency admission, organ dysfunction, number of organ failures, sepsis source and bacteraemia).
Conclusions: Sepsis incidence has risen in recent years, whereas mortality has fallen. Our findings confirm reports for other parts of the world, in the context of scarce administrative data on sepsis in Europe.
Keywords: Sepsis, Septic shock, Mortality, Epidemiology
© The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
BackgroundSepsis has been recently redefined as an infection that leads to organ dysfunction [1]. Although a hospitalized patient with sepsis is more likely to die than a patient with heart attack or stroke [2], sepsis is still not evaluated with the same urgency as other critical conditions. Sepsis mortality can be reduced considerably by adopting early
recognition protocols and using standardized emergency treatment, but fewer than 1 in 7 patients are actively intervened in this way [3]. Treatment ineffectiveness is often due to late sepsis diagnosis—mainly a failure by caregivers or healthcare professionals to suspect sepsis. The clinical symptoms and laboratory signs currently used for diagnostic purposes are not specific for sepsis, and there is a lack of reliable systems for timely detection of septic patients.
Proper detection of sepsis and its progression are essen-tial to patient management. Epidemiology case studies using administrative hospital data have reported both
Open Access
*Correspondence: [email protected] 16 Servei de Medicina Intensiva, Hospital de Mataró, Carretera de Cirera s/n, 08304 Mataró, SpainFull list of author information is available at the end of the article
Page 2 of 10Yébenes et al. Ann. Intensive Care (2017) 7:19
growing incidence and declining mortality rates associ-ated with severe sepsis in several different countries but mainly in the USA, Australia and New Zealand [4–11]. However, administrative data may be affected by changes in coding practices that distort incidence and mortal-ity estimates. Recently, Stevenson et al. [12] compared incidence and mortality trends in trial data with those observed in administrative data, observing that, since 1991, risk-standardized 28-day severe sepsis mortality has tended to decline in parallel for both methods; this which would indicate that trends in severe sepsis mortality as calculated from administrative data and International Classification of Diseases, 9th revision, Clinical Modifica-tion (ICD-9-CM) algorithms are likely to be accurate.
Catalonia, an autonomous region in Spain, has actively begun to advocate more vigorous efforts to decrease the sepsis burden through the development of an Inter-hos-pital Sepsis Emergency Code, operational between the Catalan Health Service (CatSalut) and seven local medi-cal societies [13]. Since the identification of trends in sep-sis outcomes is of considerable public health importance, the aim of our study was to estimate population and annual in-hospital incidence rates and in-hospital mor-tality trends for patients with sepsis between 2008 and 2012 in Catalonia, before implementation of the Inter-hospital Sepsis Emergency Code designed to coordinate and optimize care of patients with sepsis.
MethodsData sourcesA retrospective analysis was conducted of hospital dis-charge records from the Minimum Basic Data Set for the Catalan Health System (CatSalut) Acute-Care Hospitals (a mandatory population-based register of admissions to all public and private acute-care hospitals in Catalonia) enables evaluation and optimization of resource use, pro-vides support to and improves healthcare planning and facilitates procurement management and payments. To ensure data quality, the CMBD-HA input data are sys-tematically validated internally and periodically vali-dated externally. The data set contains demographic and clinical data for patient care episodes, including age, sex, length of stay (days), one primary diagnosis, up to nine secondary diagnoses, one primary procedure, up to seven secondary procedures and status on discharge (alive or dead). Official data from the register of insured persons maintained by CatSalut were used to estimate crude and specific hospitalization rates (universal coverage for 7,601,813 inhabitants in 2012).
PatientsSepsis, formerly severe sepsis [1] was defined by the pres-ence of infection and at least one organ dysfunction.
Patients hospitalized with sepsis were detected, according ICD-9-CM codes used to identify acute organ dysfunction and infectious processes following the Angus methodology [5], over a 5-year period (2008–2012). To avoid overlaps, we excluded patients who were transferred from one acute-care hospital to another during the same severe sepsis episode.
CodingDiagnoses and procedures were coded using the ICD-9-CM, whose codes to identify patients with sepsis were updated in 2000 to the following: 995.91 (sepsis), 995.92 (severe sepsis) and 785.52 (septic shock) (Supplemen-tary Appendix: Additional file 1). Although information was not available regarding the unit or department where patients were treated (intensive care unit (ICU), internal medicine unit, etc.), we indirectly deduced ICU admission from procedures typically used in intensive care manage-ment (Supplementary Appendix: Additional file 1). The Charlson comorbidity index with its 17 comorbid disease categories [13] was used to assess the presence of underly-ing comorbidities. The ICD-9-CM codes used to identify acute organ dysfunction and infectious processes are listed in Supplementary Appendix: Additional file 1.
Statistical analysisThe hospitalization rate was defined as the yearly number of admissions per 100,000 population (excluded were 1590 admissions from non-residents in Catalonia). Crude over-all and specific hospitalization rates by age and sex were calculated. Continuous variables and discrete variables were compared using analysis of variance and the Chi-square test, respectively. Multivariate logistic regression, adjusted for other significant variables, was used to analyse hospital mortality risk by year of admission for the study population and for the ICU and non-ICU patient groups; variables were entered one by one and retained when their significance was <.10 and were clinically plausible. For the regression analysis, each of the clinical attributes included (comorbidities, acute organ failure and infection) were treated as binary (dummy) variables indicating the presence or absence of these conditions; a single patient could there-fore account for more than one attribute. The area under the receiver operating characteristic curve (AUROC) was used to evaluate how well the multivariate logistic regres-sion model discriminated between patients with severe sepsis who were discharged alive versus those who died in-hospital [14]. Data analysis was performed using SPSS 18.0 software (SPSS Inc, Chicago, IL, USA).
ResultsIncidence and main features of severe sepsisOf 4,761,726 discharges from all acute-care hos-pitals during the study period, 82,300 (1.72%) had
Page 3 of 10Yébenes et al. Ann. Intensive Care (2017) 7:19
sepsis. Demographic characteristics and comorbidities for patients with sepsis are shown in Table 1.
Annual incidence in the population was 212.7 per 100,000 inhabitants/year, increasing from 167.2 in 2008 to 261.8 per 100,000 inhabitants/year in 2012 (Fig. 1). Sepsis was significantly associated with age (more fre-quent in older patients) and sex (208.3 cases for men ver-sus 156.0 cases for women per 100,000 inhabitants/year) (Fig. 2).
The most frequent origins of sepsis were urinary and respiratory tract infections, accounting for 37.2 and 32.5% of cases, respectively, followed by the abdomen (11%). Almost a quarter (24.7%) of cases presented bacte-raemia. Acute kidney injury was the most frequent organ failure (58.4%), followed by respiratory failure (20.5%) and central nervous system failure (19.7%). Cardiovas-cular dysfunction was reported in 4.9% of cases. Two or more acute organ failures were documented in 14% of cases (Table 1).
Hospital outcomesIn-hospital mortality over the study period was 21.6% (95% confidence interval (CI); 18.6–24.9). Hospital mor-tality was higher in older patients with higher Charlson comorbidity score, in patients with bacteraemia (39.3% in patients with positive versus 15.8% in patients with nega-tive blood cultures), and was also higher in patients with more organ failures (three or more, 63.4%). Respiratory and abdominal origins were associated with higher mor-tality (24 and 28%, respectively). Mean (SD) length of stay was 16.7 (19.5) days, with no clinically relevant differ-ences for patients who died in-hospital versus who were discharged alive (despite a value of p < .001) (Table 1 and Table 2).
Incidence and in‑hospital outcome trendsIncidence of sepsis increased in the five-year study period from 12,809 cases to 20,228 cases (mean 16,460 cases per year over the period), representing 1.3 and 2.1% (p < .0001) of hospital admissions, respectively (Fig. 2). Observed in the same period were an increase in mean age, from 69.1 to 72.8 years (p < .0001), and an increase in mean Charlson comorbidities score, from 4.9 to 5.3 (p < .001) (Fig. 3).
Length of hospital stay decreased from 18.4 to 15.3 days (p < .0001), representing a relative reduction of 17% (Fig. 3). Univariate analysis showed that hospital mortal-ity also decreased—from 23.7 to 19.7% (p < .0001)—for a relative reduction of 18.6% (Fig. 2). These differences were confirmed in the multivariate analysis adjusted for all significant variables (age group, sex, comorbidities, ICU admission, emergency admission, organ dysfunction, number of organ failures, infection source and presence
of bacteraemia) (Tables 1, 2). Differences between 2008 and all the ensuing years except 2009 were statistically significant. The logistic regression (reference year 2008) indicated a mortality odds ratio (OR) for patients with sepsis in 2012 of 0.772 (95% CI 0.727–0.820) (Table 3). The falling trend in the mortality OR was linear through-out the study period for all patients with sepsis, whether or not treated in the ICU. Values for the AUROC were calculated to evaluate how well the multivariate logistic regression model discriminated between patients dis-charged alive and discharged dead: 0.782 (95% CI 0.779–0.786) for all patients, 0.746 (0.741–0.752) for non-ICU patients and 0.749 (0.743–0.756) for ICU patients.
DiscussionMost epidemiological data on sepsis refers to the first decade of twenty-first century and almost exclusively refer to the USA. This is a large observational study of patients discharged from all national health system acute-care hospitals conducted in a European setting. We estimated the mean sepsis incidence to be 212.7 cases per 100,000 inhabitants/year and in-hospital mortality to be 21.6%. Incidence and mortality varied over time, with a yearly increase in incidence of 7.3%, a yearly relative reduction of 3.3% in length of stay and a yearly reduc-tion in in-hospital mortality of 3.4%. After adjustments for relevant clinical and epidemiological variables, the reduction in mortality remained statistically significant.
The estimated incidence of sepsis in our study was lower than reported in the USA and slightly higher than reported in smaller European studies [2, 6, 7]. Previ-ous studies conducted in Spain reported incidences of between 110 and 230 cases per 100,000 inhabitants/year [15, 16], versus the 212.7 cases observed in our study. Dif-ferences in calculated incidences may be related to struc-tural or functional organization [17] or may be due to discharge diagnosis coding. Nonetheless, we would like to emphasize the importance of using local data to moni-tor trends in activity and results over time. Moreover, the number of sepsis cases in our study increased yearly, a finding which is consistent with findings reported in other epidemiological studies [10–12].
Estimates of sepsis incidence and trends are also essen-tial to estimate the resources needed to care for these patients. Sepsis incidence is increasing compared to inci-dence for other leading causes of mortality such as acute myocardial infarction or ischaemic stroke. The CatSalut data on hospital admissions/year for severe sepsis (five-year mean, 16,460 cases) are close to the combined num-bers for acute coronary syndrome and ischaemic stroke admissions together, at 11,000 and 8000, respectively [18, 19]. However, incidence rates for acute coronary syn-drome and ischaemic stroke, unlike for sepsis, are stable
Page 4 of 10Yébenes et al. Ann. Intensive Care (2017) 7:19
[19–21]. In the USA, the percentage of septic patients with a fatal outcome increased from 14% in 2000 to 16% in 2010; in contrast, mortality for respiratory failure decreased from 25 to 17%, for heart attack from 10 to 8%, for cancer from 8 to 4% and for stroke from 6 to 5% [22].
Prospective versus retrospective analysis observed differences in incidence and source of sepsis. Prospec-tive monitoring is laborious, costly and complex and can also be affected by issues such as inclusion criteria or data sources [23, 24]. Although retrospective analy-ses from hospital discharges—as in our study—can also be affected by definitions, codes and analytical meth-ods, they serve an important function in analysing local
trends and outcomes. Gaieski [24] observed a 3.5-fold difference in estimates of absolute incidence using differ-ent database abstraction methods. Nonetheless, trends were similar irrespective of the methodology. Stevenson et al. [12] recently found that severe sepsis mortality was 10% higher for patients included in the control group of clinical trials compared to administrative data (collected according to Angus’ criteria) [5]; nonetheless, mortality trends were similar, irrespective of the data source—and were also similar to the 3% yearly reduction found in our study. Stevenson et al. consequently concluded that administrative data are useful in monitoring mortality trends in patients with severe sepsis.
Table 1 Profile of patients with severe sepsis in Catalonia 2008–2012
Total N = 82,300 Alive on discharge N = 64,511
Dead on discharge N = 17,789
p
Mean (SD) Mean (SD) Mean (SD)
Age (years) 71.2 (19.7) 70.6 (20.4) 73.3 (16.6) <.0001
Length of hospital stay (days) 16.7 (19.5) 16.8 (19.3) 16.3 (20.3) <.001
Charlson comorbidity index 5.1 (2.6) 5.0 (2.6) 5.6 (2.7) <.0001
% % % p
Age (years)
<15 3.2 3.7 1.4 <.0001
15–44 6.1 6.6 4.2
45–64 16.4 16.2 17.3
65–74 17.5 17.3 18.4
75–84 32.9 32.8 33.6
>84 23.8 23.5 25.0
Sex
Males 56.7 56.1 59.0 <.0001
Comorbidities
Chronic kidney disease 23.8 24.5 20.9 <.0001
COPD 22.4 22.9 20.6 <.0001
Cancer 14.5 12.5 21.5 <.0001
Metastasis 4.8 5.8 8.3 <.0001
Peripheral vascular disease 4.4 4.2 4.9 <.0001
Complicated diabetes 4.0 4.3 3.0 <.0001
Liver disease: mild 9.7 8.7 13.1 <.0001
Liver disease: moderate–severe 3.0 2.4 4.9 <.0001
Myocardial infarction 3.8 3.6 4.7 <.0001
Congestive heart failure 20.1 19.1 23.8 <.0001
Cerebrovascular disease 6.7 6.5 7.6 <.0001
AIDS/HIV infection 1.0 0.9 1.3 <.0001
Emergency stay (%) 89.4 89.8 87.8 <.0001
ICU admissions 28.2 22.3 49.8 <.0001
Admission year
2008 15.6 15.1 17.1 <.0001
2009 17.9 17.4 19.6
2010 19.7 19.6 20.0
2011 22.3 22.6 21.0
Page 5 of 10Yébenes et al. Ann. Intensive Care (2017) 7:19
Table 1 continued
% % % p
2012 24.6 25.2 22.3
Sepsis origins
Urinary 37.2 41.1 23.0 <.0001
Respiratory 32.5 31.4 36.4 <.0001
Abdominal 11.0 10.1 14.3 <.0001
Skin and soft tissues 4.1 4.2 3.9 .06
Endocarditis 1.6 1.5 2.0 <.0001
Device‑related 1.1 1.2 0.6 <.0001
CNS 0.9 0.9 1.2 <.0001
Others 37.9 41.0 26.5 <.0001
Bacteraemia 24.7 19.1 44.9 <.0001
Organ dysfunction
Kidney 58.4 56.9 63.7 <.0001
Lung 20.5 15.7 37.9 <.0001
CNS 19.7 21.6 12.7 <.0001
Haematologic 11.1 11.0 11.5 .054
Cardiovascular 4.9 5.2 3.6 <.0001
Liver 1.3 0.7 3.3 <.0001
Number of organ failures
1 86.3 89.7 72.9 <.0001
2 12.0 9.4 21.8
3 1.6 0.8 4.7
4 or more 0.2 0.1 0.6
Data are presented as mean and standard deviation or %
AIDS acquired immune deficiency syndrome, CNS central nervous system, COPD chronic obstructive pulmonary disease, HIV human immunodeficiency virus, ICU intensive care unit
Fig. 1 Number of cases, mortality rates and hospital incidence rates for sepsis in Catalonia (2008–2012). Incidence of sepsis increased from 12,809 cases to 20,228 cases in the 5‑year study period (mean 16,460 cases per year), representing 1.3 and 2.1% (p < .0001) of hospital admissions and an average yearly increase of 6%. However, hospital mortality decreased from 23.7 to 19.7% (p < .0001) for a yearly relative reduction of 3.4%
Page 6 of 10Yébenes et al. Ann. Intensive Care (2017) 7:19
Fig. 2 Age‑specific incidence and mortality rates for all cases of severe sepsis by sex in Catalonia (2008–2012). The dark line represents incidence (thicker for men and thinner for women) expressed as a thousand cases per 100,000 inhabitants. Age‑adjusted mortality is expressed as the number of deaths with respect to the number of cases grouped according to 5‑year age brackets
Table 2 Overall mortality by patient characteristics, sepsis origins, presence of bacteraemia and organ dysfunction by year of admission in Catalonia (2008–2012)
Data are presented as number of cases or %
CNS central nervous system, NS non-significant
Condition N Presence of condition (%) In‑hospital mortality (%)
2008 2009 2010 2011 2012 p
Mortality 23.7 23.6 22.0 20.4 19.7 <.0001
Source of sepsis
Urinary tract 30,600 33.6 35.0 37.1 39.8 38.8 <.0001 13.9
Respiratory tract 26,748 34.2 34.0 31.2 30.9 32.7 <.0001 24.2
Abdomen 9065 10.7 11.2 11.3 11.0 10.9 NS 28.0
Skin and soft tissues 3394 3.9 4.0 4.0 4.4 4.2 NS 20.3
Endocarditis 1326 1.8 1.8 1.9 1.4 1.4 <.0001 27.3
Device‑related 869 1.1 1.1 1.2 1.1 0.9 .016 11.6
CNS 775 1.2 1.1 1.1 0.7 0.7 <.0001 27.5
Other 31,149 37.4 37.5 39.2 38.1 37.1 .001 15.1
Bacteraemia 20,285 25.5 25.5 24.5 24.8 23.5 <.0001 39.3
Organ dysfunction
Kidney 48,072 50.8 53.5 57.1 61.9 64.7 <.0001 23.6
Respiratory 16,876 26.1 24.1 20.9 17.9 16.3 <.0001 40.0
CNS 16,177 20.5 19.6 19.1 18.9 20.3 <.0001 13.9
Haematologic 9158 11.6 12.3 12.2 11.2 9.1 <.0001 22.4
Cardiovascular 3988 5.3 5.4 4.7 4.6 4.4 <.0001 16.1
Liver 1058 1.0 1.3 1.3 1.3 1.4 .029 55.9
Number of organ failures
1 70,874 86.6 85.9 86.5 86.0 85.8 NS 18.5
2 9964 11.7 12.2 11.8 12.4 12.3 39.0
3 or more 1462 1.7 1.9 1.9 1.7 1.8 63.4
Page 7 of 10Yébenes et al. Ann. Intensive Care (2017) 7:19
Length of stay and mortality both decrease yearly dur-ing the study period—to a statistically significant degree according to the multivariate analysis adjusted for demo-graphic data, comorbidities, infection source and number
of organ failures. The external validity of our findings is supported by the fact that mortality in 2008 was the same as that reports by the PROWESS-SHOCK study placebo group [25]. Other recent large randomized clinical trials
Fig. 3 Trends for main characteristics and hospital stay for patients with sepsis in Catalonia (2008–2012)
Table 3 Univariate and multivariate analyses of in-hospital mortality by year of admission in Catalonia (2008–2012)
Data are presented as number of death or %. The multivariate analysis is adjusted by sex, age group, comorbidities, ICU admission, emergency admission, organ dysfunction, number of organ failures, septic origins and bacteraemia
CI confidence interval, ICU intensive care unit, OR odds ratio, NS non-significant
Univariate analysis Multivariate analysis
N % mortality p OR 95% CI for OR p
All patients N = 82,300
2008 12,809 23.7 <.0001 1 – –
2009 14,736 23.6 0.988 0.928–1.052 NS
2010 16,192 22.0 0.911 0.856–0.969 .003
2011 18,335 20.4 0.818 0.770–0.870 <.0001
2012 20,228 19.7 0.772 0.727–0.820 <.0001
Non‑ICU patients N = 59,064
2008 8551 15.8 <.0001 1 – –
2009 10,137 16.3 1.028 0.945–1.117 NS
2010 11,497 15.4 0.941 0.867–1.022 NS
2011 13,545 14.8 0.859 0.793–0.931 <.0001
2012 15,334 13.9 0.769 0.711–0.832 <.0001
ICU patients N = 23,236
2008 4258 39.5 .002 1 – –
2009 4599 39.6 0.938 0.854–1.031 NS
2010 4695 38.2 0.868 0.790–0.953 .003
2011 4790 36.0 0.757 0.686–0.832 <.0001
2012 4894 37.6 0.778 0.702–0.855 <.0001
Page 8 of 10Yébenes et al. Ann. Intensive Care (2017) 7:19
(RCTs) have reported mortality rates of 18–30% [26, 27], further confirming the likely validity of 2008 as our base-line year. Kaukonen et al. [11], who recently reported similar results for Australia and New Zealand, observed an annual absolute decrease of 1.3% in risk, from 24% in 2008 to 19% in 2012. In our study, multivariate analysis revealed a robust association between mortality and year of detection, as adjusted for confounding factors including sex, age group, comorbidities, ICU admission, emergency admission, organ dysfunction, number of organ failures, septic focus and the presence of bacteraemia. Although our study was not designed to address this issue, a poten-tial improvement in the management of sepsis could be suggested, explained in part as a consequence of training and increased clinical awareness [3, 28–30].
Our results suggest also that sepsis outcomes should be interpreted according to the year of data collection and the presence of comorbidities. Moreover, under-powered RCTs would be avoided if these effects were taken into account in estimating statistical power and sample size. Yearly reduction in crude mortality rates should be expected, bearing in mind that overestimated mortality rates may lead to underpowered studies which might, in turn, lead to potentially useful treatments being downgraded due to lack of evidence. Furthermore, excluding elderly patients and patients with comorbidi-ties from RCTs represents a form of selection bias; Kau-konen et al. [11], for instance, reported a 4.6% mortality rate for comorbidity-free patients and young patients (versus our rate of 21.6%). Another issue is that sepsis outcomes are too often viewed as binary: The patient dies (failure) or survives (success). Studies also tend to focus on in-hospital mortality and length of stay as an outcome measure for ICU patients, overlooking the fact that many patients admitted for sepsis die after dis-charge. There is an unmet need to improve knowledge regarding long-term effects in patients with sepsis [1–8], so other outcome indicators such as long-term morbid-ity and quality of life are likely to be included in future trials.
Just under a quarter (24.7%) of our patients presented with bacteraemia, associated with higher mortality. Patients with bacteraemia could represent a suitable pop-ulation to monitor prospectively in clinical practice, as bacteraemia, unlike sepsis, is easily identified retrospec-tively, is easily distinguished from other non-infectious diseases that cause organ dysfunction and is also easily stratified using the sepsis-related organ failure assess-ment (SOFA) instrument [31].
The main strengths of our study are the large cohort of patients included in a European setting, the fact of includ-ing 100% of admissions to both public and private hospi-tals of Catalonia Health System, the long period of data
collection and the use of a previously validated strategy. Our study has several limitations. The fact that cases of sepsis were identified indirectly using ICD-9-CM codes implies less accuracy in identifying cases and a poor clini-cal analysis compared to prospective methods. Urinary infections appear as the main focus of sepsis in our study. The relevance of each focus can be affected by population characteristics or methodology. It also can affect incidence or severity results. Case recruitment may also have been affected by coding, as reflected in our results for cardiovas-cular dysfunction. Hypotension was poorly documented on discharge. Incidence, as reported in our study, probably does not reflect clinical incidence. Cardiovascular dys-function incidence rates of 7.2–42% have been reported for epidemiological or retrospective studies, in contrast to rates of up 90% for prospective studies [4, 5, 9, 13, 15, 32, 33]. Our study cannot account for reasons for reduced mortality and shorter stays. Inclusion of a specific severity scores, such as SOFA, in the multivariate analysis would allow insights into whether mortality reduction is related to the inclusion of less severe patients. Unfortunately, our study design does not admit this conclusion. Given that the CMBD-HA does not specifically collect data about ICU admission, the category ‘ICU stay’ was deduced from procedures typically used in ICUs. We think that since this definition is highly specific but not sensitive, we cannot rule out the possibility that some septic patients with less severity were excluded from ICU admission.
ConclusionsSepsis incidence has risen continuously in recent years in Catalonia. While mortality and length of stay have fallen, despite increases in age and in comorbidities, our find-ings corroborate results reported for other parts of the world, in the context of scarce administrative data on sepsis in Europe.
AbbreviationsAIDS: acquired immune deficiency syndrome; CI: confidence interval; CMBD‑HA: Minimum Basic Data Set—Acute‑Care Hospitals; COPD: chronic obstruc‑tive pulmonary disease; ICD‑9‑CM: International Classification of Diseases, 9th revision, Clinical Modification; ICU: intensive care unit; NS: non‑significant; OR: odds ratio; RCT: randomized clinical trials; SD: standard deviation; SOFA: sepsis‑related organ failure assessment.
Authors’ contributionsJCY, JCR, RF, MC, AB and AA conceived and designed the work. Data acquisi‑tion and statistical analysis were performed by MC and AB; interpretation of the data for the work was carried out by JCY, JCR, RF, MC, AB and AA; all
Additional file
Additional file 1. ICD‑9‑MC codes used to identify infectious process, site of infection, acute organ dysfunction and procedures used in intensive care unit.
Page 9 of 10Yébenes et al. Ann. Intensive Care (2017) 7:19
authors were involved in draft redaction and/or revision for important intel‑lectual content; all authors read and approved the final manuscript.
Author details1 Critical Care Department, Hospital de Mataró, Mataró, Spain. 2 Grup de Recerca en Sepsis, Imflamació i Seguretat (AGAUR 2014‑SGR926), Barcelona, Spain. 3 Escola Superior de Ciencies de la Salut, Universitat Pompeu Fabra, Mataró, Spain. 4 Critical Care Department, Vall d’Hebron University Hospital, Barcelona, Spain. 5 Shock, Organ Dysfunction and Resuscitation (SODIR) Research Group, Vall d’ Hebron Research Institut, Universitat Autònoma de Bar‑celona, Barcelona, Spain. 6 CIBER Enfermedades Respiratorias, Madrid, Spain. 7 Divisió d’Anàlisi de la Demanda i l’Activitat, Servei Català de la Salut (CatSalut), Barcelona, Spain. 8 Critical Care Department, Hospital Universitari Dr. Josep Tru‑eta, Girona, Spain. 9 Critical Care Department, Hospital Universitari Joan XXIII, Tarragona, Spain. 10 IISPV/URV, Tarragona, Spain. 11 Critical Care Department, Hospital Universitari Arnau de Vilanova, Lleida, Spain. 12 Institut de Recerca Biomèdica (IRB), Lleida, Spain. 13 Multidisciplinary Intensive Care Research Organization (MICRO), St James’s University Hospital, Trinity Centre for Health Sciences, Dublin, Ireland. 14 Critical Care Center, Sabadell Hospital, Corporació Sanitaria Universitaria Parc Tauli, Sabadell, Spain. 15 Universitat Autonoma de Barcelona, Sabadell, Spain. 16 Servei de Medicina Intensiva, Hospital de Mataró, Carretera de Cirera s/n, 08304 Mataró, Spain.
AcknowledgementsGroup Authorship: SOCMIC (Catalan Critical Care Society) Sepsis Working Group Members: Abdo Taché, Servei de Medicina Intensiva, Hospital Universi‑tari Dr Josep Trueta de Girona; Antoni Margarit, Servei de Medicina Intensiva, Hospital Nostra Senyora Meritxell d’Andorra; Assumpta Ricart, Servei de Urgències, Hospital Universitari Vall d’ Hebron; Adolf Ruiz‑Sanmartin, Servei de Medicina Intensiva, Hospital Universitari Vall d’ Hebron; Begoña Balsera, Servei de Medicina Intensiva, Hospital Universitari Arnau de Vilanova de Lleida; Berta Cisteró, Servei de Urgències, Hospital Parc Taulí de Sabadell; Candelària de Haro, Servei de Medicina Intensiva, Hospital Parc Taulí de Sabadell; Concepció Rovira, Servei de Medicina Intensiva, Hospital Sant Joan de Reus; Eva Torrents, Hospital Parc Taulí de Sabadell; Francisco Álvarez‑Lerma, Servei de Medicina Intensiva, Hospital del Mar de Barcelona; Herbert Baquerizo, Servei de Urgèn‑cies, Hospital Sant Joan de Déu de Manresa; Joan Balcells, Servei de Pediatria, Hospital Universitari Vall d’Hebrón de Barcelona; José L. Echarte, Servei de Urgències, Hospital del Mar de Barcelona; José Luna, Servei de Medicina Inten‑siva, Hospital de Tortosa Verge de la Cinta; Josep M. Sirvent, Servei de Medicina Intensiva, Hospital Universitari Dr Josep Trueta de Girona; Juan Méndez, Servei de Medicina Intensiva, Hospital de Mataró; Lluís Zapata, Servei de Medicina Intensiva, Hospital de Sant Pau i Santa Creu de Barcelona; Lluïsa Bordejé, Servei de Medicina Intensiva, Hospital Universitari Germans Trias i Pujol de Badalona; Lourdes Jiménez, Servei de Medicina Intensiva, Hospital Universitari Arnau de Vilanova de Lleida; Maite Martínez‑Izquierdo, Servei de Urgències, Hospital del Mar de Barcelona; María L. Martínez, Hospital Parc Taulí de Sabadell; María P. Gracia‑Arnillas, Servei de Medicina Intensiva, Hospital del Mar de Barcelona; Mercedes Palomar, Servei de Medicina Intensiva Hospital Universitari Arnau de Vilanova de Lleida; Miguel Sánchez, Servei de Urgències, Hospital Clínic de Barcelona; Pablo Pujol, Servei de Medicina Intensiva, Hospital Universitari Dr Josep Trueta de Girona; Pau Garro, Servei de Medicina Intensiva, Hospital General de Granollers; Pau Torrabadella, Servei de Medicina Intensiva, Hospital Universitari Germans Trias i Pujol de Badalona; Paula Vera, Servei de Medicina Intensiva, Hospital de Sant Pau i Santa Creu de Barcelona; Roger Bisbal, Servei de Medicina Intensiva, Hospital de Mataró; Ruth Hernández, Servei de Urgèn‑cies, Hospital de Sant Joan Despí Moisès Broggi; Teresa M. Tomasa, Servei de Medicina Intensiva, Hospital Universitari Germans Trias i Pujol de Badalona; Víctor Pérez‑Claveria, Servei de Urgències, Hospital de Mataró.
Competing interestsThe authors declare that they have no competing interests.
Availability of data and materials sectionThe data set supporting the conclusions of this article could be available upon request from [email protected].
Ethics approval and consent to participateIn agreement with Spanish regulations concerning epidemiologic observa‑tional studies that do not modify existing diagnosis or therapeutic strategies, no ethics committee approval was required to conduct the study.
Funding The study was partially funded by the Mataró Sepsis Challenge (www.sepsis‑challenge.cat) through the Fundació Salut del Consorci Sanitari del Maresme.
Prior presentationThis study was reported in part as a communication at the 43rd Annual Congress of the Society of Critical Care Medicine (SCCM, January, 2014, San Francisco, California, USA), the 27th Annual Congress of the European Society of Intensive Care Medicine (ESICM, September 2014, Barcelona, Spain) and the 36th Annual Congress of the Catalan Society of Critical Care Medicine (SOCMIC, March, 2015, Barcelona, Spain).
Received: 1 July 2016 Accepted: 4 February 2017
References 1. Singer M, Deutschman C, Seymour C, Shankar‑Hari M, Annane D, Bauer
M, Bellomo R, et al. The third international consensus definitions for sepsis and septic shock (sepsis‑3). JAMA. 2016;315:801–10.
2. Hall MJ, Levant S, DeFrances CJ. Trends in inpatient hospital deaths: National Hospital Discharge Survey, 2000–2010. NCHS data brief, no 118. Hyattsville: National Center for Health Statistics; 2013.
3. Ferrer R, Artigas A, Levy MM, Edusepsis Study Group et al. Improvement in process of care and outcome after a multicenter severe sepsis educa‑tional program in Spain. JAMA. 2008;299:2294–303.
4. Martin G, Mannino D, Eaton S, et al. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med. 2003;348:1546–54.
5. Angus DC, Linde‑Swirble T, Lidicker J, et al. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 2001;29:1303–10.
6. Padkin A, Goldfrad C, Brady AR, et al. Epidemiology of severe sepsis occurring in the first 24 hrs in intensive care units in England, Wales, and Northern Ireland. Crit Care Med. 2003;31:2332–8.
7. Brun‑Buisson C, Meshaka P, Pinton P, EPISEPSIS Study Group et al. EPISEP‑SIS: a reappraisal of the epidemiology and outcome of severe sepsis in French intensive care units. Intensive Care Med. 2004;30(4):580–8.
8. Finfer S, Bellomo R, Lipman J, et al. Adult‑population incidence of severe sepsis in Australian and New Zealand intensive care units. Intensive Care Med. 2004;30(4):589–96.
9. Dombrovskiy VY, Martin AA, Sunderram J, et al. Rapid increase in hos‑pitalization and mortality rates for severe sepsis in the United States: a trend analysis from 1993 to 2003. Crit Care Med. 2007;35(5):1244–50.
10. Lagu T, Rothberg MB, Shieh MS, et al. Hospitalizations, costs, and out‑comes of severe sepsis in the United States 2003–2007. Crit Care Med. 2012;40(3):754–61.
11. Kaukonen KM, Bailey M, Suzuki S, et al. Mortality related to severe sepsis and septic shock among critically ill patients in Australia and New Zea‑land, 2000–2012. JAMA. 2014;311(13):1308–16.
12. Stevenson EK, Rubenstein AR, Radin GT, et al. Two decades of mortality trends among patients with severe sepsis: a comparative meta‑analysis. Crit Care Med. 2014;42:625–31.
13. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD‑9‑CM and ICD‑10 administrative data. Med Care. 2005;43:1130–9.
14. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982;143:29–36.
15. Iñigo J, Sendra JM, Díaz R, Bouza C, Sarría‑Santamera A. Epidemiology and costs of severe sepsis in Madrid. A hospital discharge study. Med Intensiva. 2006;30:197–203.
16. Andreu Ballester JC, Ballester F, González Sánchez A, et al. Epidemiology of sepsis in the Valencian Community (Spain), 1995–2004. Infect Control Hosp Epidemiol. 2008;29:630–4.
17. Levy MM, Artigas A, Phillips GS, et al. Outcomes of the Surviving Sepsis Campaign in intensive care units in the USA and Europe: a prospective cohort study. Lancet Infect Dis. 2012;12:919–24.
18. Dégano I, Elosua R, Marrugat J. Epidemiology of acute coronary syn‑dromes in Spain: estimation of the number of cases and trends from 2005 to 2049. Rev Esp Cardiol. 2013;66(6):472–81.
Page 10 of 10Yébenes et al. Ann. Intensive Care (2017) 7:19
19. Colls C, Abillera S, Garcia ALtes A, Gallofre M. Beneficis de l’organització de l’atenció sanitària a les persones amb ictus: mortalitat evitada i impacte econòmic. Edita:AQuAS, 1st ed. Barcelona. Observatori del Sistema de Salut de Catalunya; 2013.
20. Sans S, Puigdefabregas A, Paluzie G, Monterde D, et al. Increasing trends of acute myocardial infarction in Spain: the MONICA‑Catalonia Study. Eur Heart J. 2005;26:505–15.
21. Gil M, Martı H, Elosua R, et al. Análisis de la tendencia en la letalidad, incidencia y mortalidad por infarto de miocardio en Girona entre 1990 y 1999. Rev Esp Cardiol. 2007;60:349–56.
22. Hall MJ, Williams SN, DeFrances CJ et al. Inpatient care for septicemia or sepsis: a challenge for patients and hospitals. NCHS data brief 2011;62:1–8. CDC National Hospital Discharge Survey. http://www.cdc.gov/nchs/nhds.htm. Accessed 17 Oct 2014.
23. Klein Klouwenberg PM, Ong DS, Bonten MJ, et al. Classification of sepsis, severe sepsis and septic shock: the impact of minor variations in data capture and definition of SIRS criteria. Intensive Care Med. 2012;38:811–9.
24. Gaieski DF, Edwards JM, Kallan MJ, et al. Benchmarking the incidence and mortality of severe sepsis in the United States. Crit Care Med. 2013;41:1167–74.
25. Ranieri M, Thompson T, Barie P, et al. Drotrecogin alfa (activated) in adults with septic shock. N Engl J Med. 2012;366:2055–64.
26. Caironi P, Tognoni G, Masson S, et al. Albumin replacement in patients with severe sepsis or septic shock. N Engl J Med. 2014;370:1412–21.
27. The ProCESS Investigators. A randomized trial of protocol‑based care for early septic shock. N Engl J Med. 2014;370:1683–93.
28. Iwashyna TJ, Angus DC. Declining case fatality rates for severe sepsis: good data bring good news with ambiguous implications. JAMA. 2014;311(13):1295–7.
29. Levy MM, Dellinger RP, Townsend SR, Campaign SS, et al. The Surviving Sepsis Campaign: results of an international guideline‑based perfor‑mance improvement program targeting severe sepsis. Crit Care Med. 2010;38:367–74.
30. Suarez D, Ferrer R, Artigas A, Edusepsis Study Group et al. Cost‑effective‑ness of the Surviving Sepsis Campaign protocol for severe sepsis: a pro‑spective nation‑wide study in Spain. Intensive Care Med. 2011;37:444–52.
31. Routsi C, Pratikaki M, Sotiropoulou C, et al. Application of the sequential organ failure assessment (SOFA) score to bacteremic ICU patients. Infec‑tion. 2007;35(4):240–4.
32. Bouza C, Lopez‑Cuadrado T, Amate‑Blanco JM. Use of explicit ICD9‑CM codes to identify adult severe sepsis: impacts on epidemiological esti‑mates. Crit Care. 2016;20:313. doi:10.1186/s13054‑016‑1497‑9.
33. Esteban A, Frutos‑Vivar F, Ferguson ND, et al. Sepsis incidence and outcome: contrasting the intensive care unit with the hospital ward. Crit Care Med. 2007;35(5):1284–9.
Med Intensiva. 2017;41(1):28---37
www.elsevier.es/medintensiva
ORIGINAL ARTICLE
Declining mortality due to severe sepsis and septicshock in Spanish intensive care units: A two-cohortstudy in 2005 and 2011
B. Sáncheza, R. Ferrerb,c,f,∗, D. Suarezd, E. Romaya, E. Piacentinia, G. Gomàc,e,M.L. Martínezc,e, A. Artigasc,e, for the Edusepsis Study Group♦
a Intensive Care Department, Hospital Universitari Mútua Terrassa, PhD Programme, University of Barcelona, Barcelona, Spainb Intensive Care Department, Vall d’Hebron University Hospital, Autonomous University of Barcelona, Barcelona, Spainc CIBER Enfermedades Respiratorias, Spaind Epidemiology and Assessment Unit, Fundació Parc Tauli, Autonomous University of Barcelona, Sabadell, Spaine Critical Care Centre, Hospital de Sabadell, Corporació Sanitària Universitària Parc Taulí, Autonomous University of Barcelona,Sabadell, Spainf Shock, organ dysfunction and Resuscitation Research Group (SODIR), VHIR, Barcelona, Spain
Received 12 May 2016; accepted 7 September 2016Available online 28 October 2016
KEYWORDSSepsis;Guidelines;Mortality;Critical care
AbstractObjective: To analyze the evolution of sepsis-related mortality in Spanish Intensive Care Units(ICUs) following introduction of the Surviving Sepsis Campaign (SSC) guidelines and the rela-tionship with sepsis process-of-care.Design: A prospective cohort study was carried out, with the inclusion of all consecutive patientspresenting severe sepsis or septic shock admitted to 41 Spanish ICUs during two time periods:2005 (Edusepsis study pre-intervention group) and 2011 (ABISS-Edusepsis study pre-interventiongroup).Scope: Patients with severe sepsis or septic shock admitted to Spanish ICUs.Patients: All ICU admissions from the emergency department or wards and all ICU patients witha diagnosis of severe sepsis or septic shock. A total of 1348 patients were included: 630 in the2005 group and 718 in the 2011 group.Intervention: None.Primary endpoints: ICU mortality, 28-day mortality and Hospital mortality, hospital length ofstay, ICU length of stay and compliance with the resuscitation bundle.Results: Compliance with the resuscitation bundle was significantly greater in the 2011 group(5.7% vs. 9.9%; p = 0.005), and was associated to lower mortality (OR 0.602 [0.365---0.994];p = 0.048). The 2011 group had lower absolute in-hospital mortality (44.0% vs. 32.6%; p = 0.01),28-day mortality (36.5% vs. 23.0%; p = 0.01), and adjusted mortality (OR 0.64 [0.49---0.83],p = 0.001).
∗ Corresponding author.E-mail address: [email protected] (R. Ferrer).
♦ The list of researchers in Edusepsis Study Group is included in Appendix.
http://dx.doi.org/10.1016/j.medin.2016.09.0040210-5691/© 2016 Elsevier Espana, S.L.U. y SEMICYUC. All rights reserved.2173-5727
Declining mortality due to severe sepsis and septic shock in Spanish intensive care units 29
Conclusions: Mortality related to severe sepsis or septic shock in Spain decreased between twopatient cohorts in 2005 and 2011, and was attributable to earliness and improvement in sepsiscare.© 2016 Elsevier Espana, S.L.U. y SEMICYUC. All rights reserved.
PALABRAS CLAVESepsis;Guías;Mortalidad;Medicina intensiva
Disminución de la mortalidad de la sepsis grave y shock séptico en las ucis espanolas:un estudio de dos cohortes en 2005 y 2011
ResumenObjetivo: Analizar la evolución de la mortalidad relacionada con la sepsis en las unidades decuidados intensivos (UCI) espanolas desde la introducción de las directrices Surviving SepsisCampaing y la relación con el proceso de atención de la sepsis.Diseno: Estudio prospectivo de cohortes. Se incluyeron de manera consecutiva, todos lospacientes con sepsis grave o shock séptico ingresados en 41 UCI espanolas durante 2 perio-dos de tiempo: en 2005 (grupo pre-intervención en el estudio Edusepsis) y en 2011 (grupopre-intervención en el estudio ABISS-Edusepsis).Ámbito: Pacientes con sepsis grave o shock séptico ingresados en las UCI espanolas.Pacientes: Todos los ingresos en UCI procedentes de Urgencias o planta y todos los pacientesde UCI con diagnóstico de sepsis grave/shock séptico. Se incluyeron 1348 pacientes: 630 delgrupo de 2005 y 718 del grupo de 2011.Intervención: Ninguna.Variables de interés principal: Mortalidad en UCI, a 28 días y hospitalaria, estancia en la UCI yen el hospital y cumplimiento con el bundle de reanimación.Resultados: El cumplimiento del bundle de reanimación fue significativamente mayor en elgrupo de 2011 (5,7 frente a 9,9%, p = 0,005) y se asoció con una menor mortalidad (OR 0,602[0,365 a 0,994], p = 0,048). El grupo de 2011 tuvo una menor mortalidad absoluta hospita-laria (44,0 frente a 32,6%, p = 0,01), mortalidad a los 28 días (36,5 frente a 23,0%, p = 0,01) ymortalidad ajustada (OR 0,64 [0,49 a 0,83], p = 0,001).Conclusiones: La mortalidad relacionada con la sepsis grave y el shock séptico en Espana dis-minuyó entre las 2 cohortes de pacientes de 2005 y 2011, atribuible a la precocidad y las mejorasen la atención de la sepsis.© 2016 Elsevier Espana, S.L.U. y SEMICYUC. Todos los derechos reservados.
Introduction
Severe sepsis and septic shock are major healthcareproblems worldwide, with high mortality and increasingincidence. Before the start of the Surviving Sepsis Campaign(SSC) in 2002, in the USA, there were 300 cases of severesepsis per 100,000 population and 2.26 cases per 100hospital discharges; half of those received intensive care.The overall rate of sepsis mortality was 28.6%; mortalityincreased with age, from 10% in children to 38.4% in those>85 years old. At an average cost of $22,100 per case,the total annual cost was $16.7 billion. The incidence wasprojected to increase by 1.5% per year.1 An epidemiologicalstudy in Spain, that analyzed the 2006---2011 NationalHospital Discharge Registry, reported that overall incidenceper year of severe sepsis was 86.97 cases per 100,000population (increasing from 63.91 cases/100,000 popula-tion in 2006 to 105.51 cases/100,000 population in 2011)representing 1.1% of all hospitalisations and 54% of hospi-talisations with sepsis. The overall mortality rate during thestudy period was 37.1 cases per 100,000 population witha significant decrease in mortality rates with an overall
annual percent change of −3.24%,2 the incidence of severesepsis attended in the Spanish ICU was 12.4% with high ICUand hospital mortality rates (48.2 and 54.3% respectively),3
with treatment costing around 500 million euros annually.4
In the past decade many studies have demon-strated improved survival in septic patients with earlyadministration of appropriate antibiotics,5---8 lactate lev-els measurements,9 early goal-directed hemodynamicresuscitation,10 management with replacement doses ofcorticosteroids,11 glycemic control,12 drotrecogin alfa(activated) administration,13 and protective mechanicalventilation.14 These therapeutic advances were collectedin the first Surviving Sepsis Campaign (SSC) guidelines15,16
with the intention to reduce sepsis mortality by 25% infive years. The progressive implementation of these recom-mendations achieved a progressive fall in mortality.17,18 AnSpanish study showed that compliance with the resuscita-tion bundle is associated with improvement in survival inpatients with severe sepsis/septic shock,19 also, in 2010, ameta-analysis of all studies comparing outcomes in patientswho received bundled care vs. non protocolized care demon-strated a clear association between the use of bundles and
30 B. Sánchez et al.
lower mortality,20 even though most of the original recom-mendations in bundled care were changed after randomizedcontrolled trials failed to confirm the efficacy of specificsepsis treatments (most of which conform the managementbundle),21---26 as is reflected in the recently updated SSCguidelines.27,28
Two recent studies concluded that there has been asecular decrease in severe sepsis mortality. One studyanalyzed patients pooled from the control groups of 36randomized controlled trials investigating severe sepsisand compared the 28-day mortality with patients in theadministrative Nationwide Inpatient Sample database.29
The other study analyzed hospital mortality due to severesepsis from a large database of patients in Australian andNew Zealand intensive care units (ICU).30 Both studies found1---3% annual improvement in crude severe sepsis mortality.The mechanism underlying this decline is unclear, but isprobably related to improved processes of care.
The Edusepsis study evaluated the impact of a nation-wide quality improvement intervention in Spain based on theSSC guidelines, showing an improvement in compliance withtreatment recommendations accompanied by a reduction inmortality.31 However, not all the effects of the interven-tion were sustained; for example, early use of antibioticsdecreased in the long-term follow-up. This is especiallyrelevant considering an analysis of the impact of individ-ual components of the resuscitation bundle on mortalityin the Edusepsis study found that early empirical antibi-otic administration was the most important factor.4 As inother time-dependent pathologies, in sepsis the timelinessand appropriateness of treatments administered in the firsthour after the onset of disease can influence outcomes.For all these reasons, a new study was designed to focusspecifically on educational interventions about early admin-istration of empirical antibiotics in severe sepsis and septicshock (ABISS-Edusepsis study). Both the original Edusepsisstudy and the ABISS-Edusepsis study employed a ‘‘controlgroup’’ documented before the educational interventions.
The primary objective of the study was to analyze theevolution of sepsis-related mortality in Spanish ICUs sincethe introduction of the Surviving Sepsis Campaing (SSC)guidelines and the relationship with the improvement of sep-sis process-of-care. The secondary objective was to analyzethe evolution of sepsis process-of-care by using the sepsisresuscitation bundle.
Patients and methods
Design
We designed a cohort study to compare two groups ofpatients with severe sepsis or septic shock treated inSpanish ICUs during two time periods: the first group(data collected between November and December 2005)was the pre-intervention group in the Edusepsis study,and the second group (data collected between April andJune 2011) was the pre-intervention group in the ABISS-Edusepsis study (an ongoing study, in the data analysis phase.http://www.edusepsis.org/en/abiss-edusepsis.html). Onlydata from ICUs that participated in both studies wereincluded in the present study.
Patients and process-of-care and outcomemeasurements
We used the same inclusion-exclusion criteria and defini-tions of severe sepsis/septic shock, acute organ dysfunction,and onset of sepsis (time zero) as in the two Edusep-sis studies.5,31 Briefly, in both studies, all ICU admissionsfrom the emergency department or from wards and allICU patients were actively screened daily for severe sep-sis or septic shock. Time zero was determined accordingto the patient’s location within the hospital when sepsiswas diagnosed. Researchers recorded data related to tenitems (tasks or targets), grouped in the sepsis resuscita-tion bundle (6 items that should begin immediately andbe accomplished within 6 h of time zero: lactate measure-ment, fluids and vasopressors, blood extraction for cultures,administration of broad spectrum antibiotics, achievementof central venous pressure ≥8 mmHg, and achievement ofcentral venous oxygen saturation ≥70%).
Bundle compliance and clinical outcome
The primary outcome measure was hospital mortality andcompliance with the individual items of the resuscita-tion bundle in the established time frames. Compliancewas defined as evidence that bundle tasks were done andtargets were achieved within the indicated time frame.Secondary outcome measures included 28-day mortality,hospital length of stay, and ICU length of stay.
Statistical analysis
Descriptive statistics included frequencies and percentagesfor categorical variables and means, standard deviations,medians, and interquartile ranges for continuous variables.To compare categorical variables between the two studyperiods, we used chi-square analysis. To compare continuousvariables during the two study periods, we used Student’st-test or the Mann---Whitney test, as appropriate. We con-structed 3 multivariate logistic regression models, withhospital mortality as the dependent variable:
• Model I was constructed to assess the protective effectof bundles and includes as independent variables: resus-citation bundle, APACHE II score, age, patient locationat sepsis diagnosis, site of infection, and baseline acuteorgan dysfunctions.
• Model II was constructed to assess the difference in mor-tality between the two study periods adjusted by APACHEII score, age, patient location at sepsis diagnosis, site ofinfection and baseline acute organ dysfunctions.
• Model III was constructed to assess the potential role ofbundle compliance in the reduction of the adjusted mor-tality and includes the same independent variables thanModel II adding the resuscitation bundle.
Statistical tests were two-tailed and significance was setat 0.05. We used SPSS version 17.0 (SPSS, Chicago, IL, USA)for all analyses.
Declining mortality due to severe sepsis and septic shock in Spanish intensive care units 31
Table 1 Demographic and clinical characteristics of patients, by group.
2005 group(n = 630)
2011 group(n = 718)
p
APACHE II, mean (SD) 20.7 (7.2) 22.4 (7.9) <0.001Age, years, mean (SD) 62.1 (16.7) 64.9 (14.9) 0.231Sex, male n (%) 375 (59.5) 457 (63.6) 0.134Source n (%) <0.001
Emergency department 268 (42.5) 523 (72.8)Ward 278 (44.1) 158 (22)ICU 84 (13.3) 37 (5.2)
Site of infection, n (%) <0.001Pneumonia 243 (38.6) 218 (30.4)Acute abdominal infection 189 (30) 253 (35.2)Urinary tract infection 58 (9.2) 123 (17.2)Meningitis 7 (1.1) 11 (1.5)Soft-tissue infection 32 (5.1) 48 (6.7)Catheter 10 (1.6) 18 (2.5)≥2 sites of infection 10 (1.6) 0Other infections 81 (12.9) 47 (6.5)
Baseline acute organ dysfunctions n (%)Cardiovascular 521 (82.7) 617 (85.9) 0.141Respiratory 408 (64.8) 314 (43.7) <0.001Renal 462 (73.3) 415 (57.8) <0.001Hyperbilirubinemia 127 (20.2) 115 (16) 0.072Thrombocytopenia 150 (23.8) 155 (21.6) 0.334Coagulation 222 (35.2) 350 (48.7) <0.001Hyperlactatemia 223 (35.4) 339 (47.2) <0.001
Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II; SD, standard deviation; ICU, intensive care unit.
Ethics committee approval
Each participating centers’ research and ethical reviewboards approved the study and patients remained anony-mous. The need for informed consent was waived inview of the observational and anonymous nature of thestudy.
Results
Data from the 41 ICUs participating in both studies wereincluded. All ICUs were medical-surgical and most (86%)were in teaching hospitals training residents. No patientswere excluded.
Patient characteristics
In the two periods, 1348 patients fulfilled criteria for severesepsis or septic shock (630 patients in the 2005 groupand 718 in the 2011 group). Patients in the 2011 groupwere older and more severely ill. In 2005, sepsis was diag-nosed predominantly in the ward, whereas in 2011 sepsiswas diagnosed predominantly in the emergency depart-ment. Pneumonia was most common infection in 2005 whileacute abdominal infection was the predominant infection in2011. More than 80% of patients in both periods had sep-tic shock. Table 1 shows patient characteristics in the twoperiods.
Outcome indicators
Table 2 reports the outcome data. Patients in the 2011 grouphad lower hospital mortality (32.6% vs. 44.0%; p < 0.001) and28-day mortality (23.0% vs. 36.5%; p < 0.001). No differenceswere observed in the ICU stay in the surviving populations,but the mean hospital stay was higher in the 2011 group(28.1 ± 22.9 vs. 33.9 ± 34.2 days; p = 0.003).
Change in compliance with bundle items over time
Rates of compliance with resuscitation bundle itemsincreased between the two periods. Compliance with theresuscitation bundle increased from 5.7% in the 2005 groupto 9.9% in the 2011 group (p = 0.005). Fig. 1 shows compli-ance with the items in the resuscitation bundle. In 2005, theonly two items in the resuscitation bundle for which compli-ance was higher than 50% were blood extraction for culturesbefore antibiotic administration (54.8%) and early adminis-tration of broad spectrum antibiotics (68.3%); compliancewith these two items was similar in 2011. Compliance withthe other 4 items was lower than 50% in 2005 and improvedsignificantly in 2011.
Multivariate logistic regression
Table 3 (Model I) showed, after to adjust for possible con-founders, that compliance with the resuscitation bundleare both associated with lower mortality. In addition, the
32 B. Sánchez et al.
Table 2 Outcomes measurements by group.
Measurements 2005 group 2011 group p
Hospital mortality, (%) 277 (44.0) 234 (32.6) <0.00128-day mortality, (%) 230 (36.5) 165 (23.0) <0.001Hospital staya, days
Mean (SD) 28.1 (22.9) 33.9 (34.2) 0.003Median [interquartile range] 21.4 [13.7---35.8] 21 [13---42] 0.253
ICU staya, daysMean (SD) 12.3 (14.8) 11.2 (14.5) 0.316Median [interquartile range] 7.5 [4.6---14.7] 6 [3---13] 0.832
Abbreviations: SD, standard deviation; ICU, intensive care unit.a Deaths are excluded.
diagnosis of sepsis when the patient is located at ward orat ICU was independently associated with an increase inhospital mortality compared with sepsis identification inthe Emergency Department. Two sites of infection (urinary-tract and soft-tissue) were associated to lower mortalitythan pneumonia, and two baseline acute organ dysfunc-tions (respiratory and trombocytopenia) were associated tohigher mortality.
After adjusting for possible confounders (Table 4, ModelII), the 2011 cohort was independently associated withlower hospital mortality (OR 0.64 [0.49---0.83], p = 0.003).When we included in the previous model the compliancewith the resuscitation bundle (Table 4, Model III), the 2011cohort kept a significant lower adjusted mortality (OR 0.64[0.531---0.95], p = 0.021).
Discussion
We assessed whether sepsis-related mortality in SpanishICUs has decreased since the introduction of the Surviv-ing Sepsis Campaign (SSC) guidelines and whether decreasesare attributable to bundle compliances and other improve-ments in sepsis care. We found that compliance withthe 6-h and 24-h bundles improved and that 28-day and
hospital mortality decreased in this 6 years period, suggest-ing a sustained effect of the SSC moreover, this reduction inhospital mortality remained significant after adjustments.We also found that compliance with the resuscitation bun-dle are independent protective factors for mortality. Ourresults are consistent with recent clinical trials, where themortality due to septic shock was around 24---26%,32 andwith epidemiological studies that show a declining trendin severe sepsis mortality over time.2,33 We found that28-day mortality and hospital mortality decreased despitean increase in predicted mortality as evidenced by higherAPACHE II scores; these results corroborate those reportedin Stevensons et al.’s29 meta-analysis and Kaukonen et al.’s30
large epidemiological study. A recently published study ofa collaborative change intervention aimed at facilitatingadoption of SCC bundles in 218 hospitals over 7.5 years foundcompliance improved over time and increased compliance isassociated with decreased mortality.34
The dramatic decrease in sepsis mortality between 2011and 2005 probably is multifactorial. Increased bundled careis playing a role but the lack of change in adjusted mortalitybetween Model II and III is suggesting that other uncontrolledfactors are also important. Probably, implementing the SSCguidelines beginning in 2005 resulted in earlier diagnosis(reflected in the increased proportion of sepsis diagnosed
0%
25%
50%
75%
100%
Lactatemeasure
Fluids andvasopressors
CVP 8 mmHg ScvO2 70% Bloodcultures
Broad spectrumATB
Completebundle
*P<.05
*
*
*
*42.4%
74.7%
44.7%
64.2%
23.7%
47.6%
6.1%
34.8%
54.8%51.5%
66.7%
9.9%
68.3%
5.7%
*
2005 group 2011 grou p
Figure 1 Resuscitation bundle, accomplished within 6 h.Abbreviation: CVP: central venous pressure; ScvO2: central venous oxygen saturation; ATB: antibiotic.
Declining mortality due to severe sepsis and septic shock in Spanish intensive care units 33
Table 3 Multivariate analysis of factors associated withmortality, Model I; only significant variables are shown.
Variable OR (95% CI)Model I
p
Age 1.018 (1.009---1.027) <0.001APACHE-II 1.087 (1.066---1.109) <0.001Source
Warda 2.007 (1.522---2.646) <0.001ICUb 2.396 (1.541---3.726) <0.001
Site of infectionUTIc 0.287 (0.173---0.476) <0.001Soft-tissue infectiond 0.414 (0.221---0.776) 0.006
Baseline acute organ dysfunctionsRespiratory 1.438 (1.076---1.923) 0.014Thrombocytopenia 1.288 (1.016---1.896) 0.039
Treatment bundleResuscitation bundle 0.602 (0.365---0.994) 0.048
Abbreviations: APACHE II, Acute Physiology and Chronic HealthEvaluation II; OR, odd ratio; CI, confidence interval; ICU, inten-sive care unit; UTI, urinary tract infection.
a Patient located in a ward at sepsis diagnosis compared to inthe emergency department.
b Patient located in the ICU at sepsis diagnosis compared to inthe emergency department.
c Patient with UTI compared to pneumonia.d Patient with Soft-tissue infection compared to pneumonia.
in the emergency department), and lower treatment vari-ability both among clinicians and within clinicians betweenpatients.35 According to Kaukonen et al.30 perhaps there areother reasons we have not controlled in our study (bettersource control, more adequate empirical antibiotic therapy,earlier transfer to ICU, better overall management of the
septic patient, dissemination of protocols, greater exper-tise of health workers, increased sensitivity for this disease,greater involvement of health institutions, etc.), that alsoaffect the drop in mortality over time. In addition, in arecently published study36 was described a number of factorsassociated with in-hospital mortality among patients withsevere sepsis or septic shock (age, active cancer, diabetes,DNR status on ED arrival, lack of fever, hypoglycemia, andintubation) despite receipt of early protocolized resuscita-tion in the ED, providing insights into aspects of early sepsiscare that can be targets for future intervention.
Most process-of-care indicators improved over time,but compliance with two recommended tasks (acquiringblood cultures before antibiotic administration and broad-spectrum antibiotics administration before 3 h) did notchange. In the 2005 group, these were the two items in theresuscitation bundle with the highest compliance (54.8% and68.3%, respectively), improved compliance with these tasksshould have an important impact on mortality.9,13 Milleret al.35 underlined the importance of these early inter-ventions when they reported that compliance with earlyresuscitation bundle items was associated with a lowerprobability of being eligible for later resuscitation and main-tenance bundle items, probably reflecting lesser severitydue to the improvements brought about by the early treat-ments.
A recent multicenter cohort study conducted in Holland37
showed that the implementation of a national programsepsis resulted in improved adherence to sepsis bundlesin severe sepsis and septic shock patients and was asso-ciated with reduced adjusted in-hospital mortality onlyin participating ICUs, suggesting direct impact of sepsisscreening and application bundle on in-hospital mortal-ity. Our study demonstrated improvements in hemodynamicresuscitation over time. However, recent trials in patients
Table 4 Multivariate analysis of factors associated with mortality, Models II and III; only significant variables are shown.
Variable OR (95% CI)Model II
p OR (95% CI)Model III
p
2011 group 0.64 (0.49---0.83) 0.003 0.64 (0.531---0.95) 0.021Age 1.016 (1.008---1.025) <0.001 1.019 (1.01---1.027) <0.001APACHE-II 1.104 (1.084---1.124) <0.001 1.092 (1.07---1.114) <0.001Source
Warda 1.796 (1.362---2.367) <0.001 1.849 (1.39---2.458) <0.001ICUb 2.168 (1.4---3.357) <0.001 2.191 (1.397---3.334) 0.001
Site of infectionUTIc 0.277 (0.172---0.445) <0.001 0.295 (0.178---0.449) <0.001Soft-tissue infectiond 0.437 (0.234---0.817) 0.010 0.418 (0.223---0.784) 0.007
Baseline acute organ dysfunctionsRespiratory 1.378 (1.028---1.846) 0.032 1.356 (1.01---1.821) 0.043Thrombocytopenia 1.362 (0.998---1.859) 0.052 1.371 (1.003---1.874) 0.048
Treatment bundleResuscitation bundle NA 0.634 (0.378---1.032) 0.066
Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II; OR, odd ratio; CI, confidence interval; ICU, intensive careunit; UTI, urinary tract infection; NA, not applicable.
a Patient located in a ward at sepsis diagnosis compared to in the emergency department.b Patient located in the ICU at sepsis diagnosis compared to in the emergency department.c Patient with UTI compared to pneumonia.d Patient with Soft-tissue infection compared to pneumonia.
34 B. Sánchez et al.
with septic shock failed to demonstrate any benefit ofprotocolized resuscitation when compared with ‘‘usualcare’’.38---41 Although no consensus exists among cliniciansregarding optimal hemodynamic monitoring and to date nomethod has proven superior, the ‘‘usual care’’ in these trialsincludes early identification of septic patients, early antibi-otic treatment, and early volume resuscitation measures.The improvement in hemodynamic resuscitation betweenthe two periods in our study is probably due to earlier resus-citation more than greater protocolized resuscitation. Oneof the most important changes between the two periodswas the place where sepsis was diagnosed. The proportionof cases diagnosed in the emergency department increasedfrom 42.5% in 2005 to 72.3% in 2011, and the proportionof cases diagnosed in the wards decreased from 44.1% in2005 to 22% in 2011. These findings indicate earlier detec-tion of sepsis and hence earlier initiation of treatment.Importantly, in the above-mentioned studies38---41 comparing‘‘usual care’’ with protocolized resuscitation, all cases ofseptic shock benefited from early detection and initiationof treatment, so perhaps earlier diagnosis and treatmentrather than differences in how treatment is administered iswhat determines prognosis. We agree with Levy42 that thepriority should be to establish systems to identify and treatseptic patients early.
On the other side, there are several differences betweenthe 2 cohorts: the 2011 group had a higher rate of urinarytract infection and lower rate of respiratory dysfunction(associated to lower mortality), but also those patients wereolder and had higher APACHE II score (associated to highermortality); we cannot discard that, despite the adjustments,those differences in the case-mix also influences the differ-ence in mortality.
Our study shows several limitations, the participation inboth studies was entirely voluntary, and the hospitals thatparticipated are not necessarily representative of those thatdid not participate; therefore, our findings may not be gen-eralizable. The length of study periods, its nonrandomizeddesign and the lack of control group precludes establishinga causal connection between the improvements in process-of-care variables and outcomes. Thus, although we observeda better compliance with most of the resuscitation bundlein 2011, related with a decrease in mortality, these findingsdo not necessarily imply a causal relationship between thecompliance with sepsis bundles and outcomes.
Moreover, our study was limited to patients admitted tothe ICU, and we cannot know how possible improvementin process-of-care variables in other areas of the hospitalaffected outcomes.
Finally, the latest SSC guidelines reflect some changes inthe standard of care at the time of our study, such as theuse of hydroxyethyl starch or drotrecogin alfa (activated).Considering these changes actually strengthens conclusionsdrawn from our results.
In conclusion, the mortality related to severe sep-sis/septic shock in Spain, between two cohorts of patientsin 2005 and 2011, decreased dramatically attributableto earliness and improvements in sepsis care, includinghigher compliance with resuscitation bundle. Nevertheless,compliance with some important items of the resuscitationbundle have not improved enough; early administration ofeffective antimicrobials could further improve outcomes.
Authors’ contributions
Baltasar Sánchez: data analysis, drafted, translated and cor-rected the manuscript.
Ricard Ferrer: national coordinator, study design, dataanalysis, drafted, translated and corrected the manuscript.
David Suárez: database development, data analysis, cor-rected the manuscript.
Eduardo Romay: data collection, corrected themanuscript.
Enrique Piacentini: site coordinator, data collection andcorrected the manuscript.
Gemma Gomá: data collection and corrected themanuscript.
María Luisa Martínez: data collection and corrected themanuscript.
Antonio Artigas: project coordinator, drafted and cor-rected de manuscript.
Ethical responsibilities
Protection of people and animals. The authors declare thatin this research have not been performed experiments onhumans or animals.
Data confidentiality. The authors declare that they havefollowed the protocols of their workplace about publicationof patient data.
Right to privacy and informed consent. The authorsdeclare that in this paper does not appear patient data.
Funding
This work was supported by Instituto de Salud Carlos III:PI10/01497 and CM12/00066.
Conflicts of interest
The authors declare no conflict of interest.
Acknowledgment
We thank John Giba for help with English.
Appendix A. Appendix
Edusepsis Study Group: Ma Mar Cruz (Virgen de la SaludToledo-Acquaroni); Carmen Fernández González (Com-plejo Hospitalario de Ferrol. Arquitecto Marcide), PazMerino, Elena Bustamante (Hospital Can Misses); San-dra Barbadillo (Hospital General de Cataluna); María dela Cruz Martín, Joaquin Ramon (Centro Médico Delfos);Luis Alvarez Rocha (Complexo Hospitalario Universitariode A Coruna); Nestor Bacelar (Clínica Corachan); BelénJimenez Bartolomé(Hospital Clínico Universitario LozanoBlesa Zaragoza); Juan Diego Jiménez Delgado (Hospi-tal Comarcal Don Benito-Villanueva) Demetrio CarriedoUle, Ana María Dominguez Berrot, Francisco Javier DíazDominguez (Complejo Asistencial Universitario de León);
Declining mortality due to severe sepsis and septic shock in Spanish intensive care units 35
Juan Machado Casas(Complejo Hospitalario de Jaén);Clara Laplaza Santos; Manuel García-Montesinos; EnriqueMaraví Poma (Complejo Hospitalario de Navarra, Pamplona);Victor Lòpez Ciudad; Pablo Vidal cortes (Complejo Hos-pitalario de Ourense); Ana Navas, Gemma Gomà, MaríaLuisa Martínez, Antonio Artigas (Hospital de Sabadell, Con-sorci Hospitalari Parc Tauli); Manuel Castellano Hernandez;Rafael Domínguez (Hospital Alto Guadalquivir, Andújar);Miguel Martínez, Jose Antonio Fernández, Fernando Callejo,María Jesús López Pueyo (Hospital General Yagüe); Alec Tal-let (Hospital General de Segovia); Pau Torrabadella, AlvaroSalcedo, Claudio Durán (Hospital Universitari Germans Triasi Pujol); Iratxe Seijas (Hospital de Cruces); Teresa RecioGómez, Abilio Arrascaeta (San Pedro de Alcántara, Cáceres);Angel Arenaza, Ana Morillo, Daniel Del Toro, Tomá s Guz-man (Hospital Virgen de la Macarena); Pilar Marco, IzaskunAzkarate (Hospital de Donostia); Isabel Rodríguez (HospitalGeneral de Baza); Eduardo Palencia, Pablo García Olivares,Patricia Santa Teresa Zamarro (Hospital Gregorio MaranónMadrid); Eugenia Yuste (Hospital Universitario San Cecilio);Jordi Solé Violán (Hospital Dr Negrín. Las Palmas); JoséBlanquer, Mónica García (Hospital Clínico Valencia); JuanCarlos Ballesteros (Hospital Universitario de Salamanca);Antonio Blesa, Fernando Martínez, Alejandro Moneo (Hos-pital San Carlos); Carlos Pérez (Hospital Santiago Apóstol);Jose Ángel Berezo, Jesús Blanco (Hospital Río Hortega Val-ladolid); Francisco Javier Martín López (Hospital comarcalSanta Ana, Motril); Ramón Vegas Pinto (Hospital de Ante-quera); Pilar Martinez Trivez (Hospital de Barbastro Huesca);Antonio Reyes Garcia (Hospital de la Princesa de Madrid);Lluís Zapata, Paula Vera (Hospital de la Santa Creu i SantPau); Eduardo Antón (Hospital de Manacor); Juan CarlosYebenes (Hospital de Mataró); María de las Olas Cerezo Arias(Hospital de Mérida); Francisco García delgado (Hospitalde Montilla); Javier Fierro Rosón, Josefa Peinado Rodriguez(Hospital del Poniente, El Ejido); Ma Jesús Broch Porcar(Hospital de Sagunto); María Álvarez (Hospital de Terrassa);Francisco Álvarez, Ma Pilar Gracia Arnillas (Hospital del Mar);Francisco Valenzuela (Hospital de Jerez); Patricia Albertde la Cruz (Hospital del Sureste); Rafael Blancas Casero,Blanca López Matamala (Hospital del Tajo); Monserrat SisónHeredia (Hospital Dr. José Molina Orosa); Pedro Olaechea,Celia Sanudo (Hospital Galdakao-Usansolo); Jose ManuelGutierrez Rubio (Hospital General de Albacete); RobertoReig (Hospital General de Castellón); Alfonso Ambrós, JulianOrtega (Hospital General de Ciudad Real); Leandro FajardoFeo (Hospital General de Fuerteventura); Pau Garro (Hos-pital General de Granollers); Francisco Navarro Pellejero(Hospital General de la Defensa en Zaragoza); Ana TrujilloAlonso (Hospital general de La Palma); Rosa Catalán (Hospi-tal General de Vic); Assumpta Rovira, Nicolas Rico (HospitalGeneral Hospitalet de LLobregat); Jose Manuel AllegueGallego, Luis Herrera Para, Josefa Murcia Paya (HospitalGeneral Universitario Santa Lucía, Cartagena); José CórdobaAlonso, Dolores Ocana (Hospital La Inmaculada de Huercal-Overa); Jose Francisco Olea Parejo (Hospital Lucus Augusti,Lugo); Pedro Galdos Anuncibay (Hospital Puerta del Hierro);Manuel Salido Mota (Hospital Regional Universitario CarlosHaya); María Jesús Gómez (Hospital General UniversitarioReina Sofía de Murcia); Ana Isabel Ezpeleta Galindo, PalomaDorado (Hospital Royo Villanova, Zaragoza); Arantxa LanderAzcona, Rosario Elbaile (Hospital San Jorge, Huesca); Diego
Mendoza (Hospital Sant Joan Despí Moisès Broggi); FranciscaPrieto (Hospital de Sta. Bárbara, Puertollano’; Luis Vallejo(Hospital SAS La Línea); Jose Ignacio Ayestarán Rota (Hos-pital Son Espasses); Marcio Borges (Hospital Son Llatzer);Enrique Piacentini, Ricard Ferrer (Hospital Univerisari MútuaTerrassa); Josep Maria Sirvent, Sara Herranz Ulldemolins(Hospital Universitari Josep Trueta de Girona); FernandoIglesias Llaca, Lorena Forcelledo Espina, Francisco TaboadaCosta, José Antonio Gonzalo Guerra (Hospital UniversitarioCentral de Asturias); Leonardo Lorente Ramos (HospitalUniversitario Canarias. Tenerife); Helena Yanez (Hospitaluniversitario de Guadalajara); Ana Loza (Hospital Univer-sitario de Valme); Jose Miguel Soto, Constantino Tormo(Hospital Universitario Dr. Peset); Borja Suberbiola (Hospi-tal Universitario Marqués de Valdecilla); Domingo Ruiz dela Cuesta Martin, Ignacio Tomás Marsilla (Hospital Universi-tario Miguel Servet Zaragoza); Mar Martín Velasco (HospitalUniversitario Nuestra Senora de Candelaria); Rafael LeónLópez, Juan Carlos Pozo (Hospital Universitario Reina Sofiade Córdoba); Jose Ángel Berezo, Jesús Blanco (HospitalRío Hortega Valladolid); Paula Ramírez (Hospital Universi-tario y Politecnico la Fe); Juan Carlos Ruiz Rodriguez, JesusCaballero, Adolf Ruiz, Alejandra García, Jordi Riera, JavierSarrapio, (Hospital Universitari Vall d’ Hebron); CarolaGiménez Esparza (Hospital Vega Baja orihuela); Ana CarolinaCaballero (Hospital Virgen de la Concha, Zamora); María Vic-toria de la Torre, Cristina Salazar (Hospital Virgen Victoriade Málaga); Carlos Ortiz (Hospital Virgen del Rocio); EduardoPalencia Herrejón, Begona Bueno (Hospital Infanta Leonor,Madrid); Gumersindo González, Díaz; Andres Carrillo (Hospi-tal General Universitario Morales Meseguer, Murcia); ManuelRodríguez (Hospital Juan Ramon Jiménez); Raquel ValeroGracia (MAZ MATEPSS SUMA Intermutual Zaragoza); RuthJorge García (Hospital Nuestra Senora de Gracia Zaragoza);Manuel Quintana, Miguel Ángel Taberna (Hospital NuestraSra del Prado); José Carlos Torralba Allué (Hospital GeneralObispo Polanco Teruel); Isidro Prieto del Portillo (Hospi-tal Universitario Ramón y Cajal); José Luis Monzón, AdolfoCalvo Martínez (Hospital de Logrono); Ricardo Diaz Abad,Miguel Ángel Blasco Navalpotro, Frutos Del Nogal Sáez, JesúsRebollo Ferreiro, José Suarez Saiz (Hospital UniversitarioSevero Ochoa), Mar Gobernado, Ma José Fernandez Calavia(Hospital de Sta. Bárbara, Soria); Francisco José Guerrero,Felipe Canada, Milagros Balaguer, Isabel Mertín, CarmenLópez, Daniel Sánchez (Hospital Torrecárdenas); Jose MaríaBonell (USP H Clínica Palma Planas); José Castano (Hospi-tal Universitario Virgen de las Nieves); Hospital Virgen delPuerto, Plasencia).
References
1. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J,Pinsky MR. Epidemiology of severe sepsis in the United States:analysis of incidence, outcome, and associated costs of care.Crit Care Med. 2001;29:1303---10.
2. Bouza C, López-Cuadrado T, Saz-Parkinson Z, Amate-Blanco JM.Epidemiology and recent trends of severe sepsis in Spain: anationwide population-based analysis (2006---2011). BMC InfectDis. 2014;14:3863.
3. Blanco J, Muriel-Bombín A, Sagredo V, Taboada F, Gandía F,Tamayo L, et al., Grupo de Estudios y Análisis en CuidadosIntensivos. Incidence, organ dysfunction and mortality in severesepsis: a Spanish multicentre study. Crit Care. 2008;12:R158.
36 B. Sánchez et al.
4. Suarez D, Ferrer R, Artigas A, Azkarate I, Garnacho-MonteroJ, Gomà G, et al., Edusepsis Study Group. Cost-effectivenessof the Surviving Sepsis Campaign protocol for severe sepsis:a prospective nation-wide study in Spain. Intensive Care Med.2011;37:444---52.
5. Ferrer R, Artigas A, Suarez D, Palencia E, Levy MM, Arenzana A,et al., Edusepsis Study Group. Effectiveness of treatments forsevere sepsis: a prospective, multicenter, observational study.Am J Respir Crit Care Med. 2009;180:861---6.
6. Garnacho-Montero J, Garcia-Garmendia JL, Barrero-AlmodovarA, Jimenez-Jimenez FJ, Perez-Paredes C, Ortiz-Leyba C. Impactof adequate empirical antibiotic therapy on the outcome ofpatients admitted to the intensive care unit with sepsis. CritCare Med. 2003;3:2742---51.
7. Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S,et al. Duration of hypotension before initiation of effectiveantimicrobial therapy is the critical determinant of survival inhuman septic shock. Crit Care Med. 2006;34:1589---96.
8. Ferrer R, Martin-Loeches I, Phillips G, Osborn TM, Townsend S,Dellinger RP, et al. Empiric antibiotic treatment reduces mor-tality in severe sepsis and septic shock from the first hour:results from a guideline-based performance improvement pro-gram. Crit Care Med. 2014;42:1749---55.
9. Puskarich MA, Illich BM, Jones AE. Prognosis of emergencydepartment patients with suspected infection and intermediatelactate levels: a systematic review. J Crit Care. 2014;29:334---9.
10. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B,et al. Early goal-directed therapy in the treatment of severesepsis and septic shock. N Engl J Med. 2001;345:1368---77.
11. Annane D, Sebille V, Charpentier C, Bollaert PE, Francois B,Korach JM, et al. Effect of treatment with low doses of hydro-cortisone and fludrocortisone on mortality in patients withseptic shock. JAMA. 2002;288:862---71.
12. Van den Berghe G, Wouters P, Weekers F, Verwaest C,Bruyninckx F, Schetz M, et al. Intensive insulin therapy in thecritically ill patients. N Engl J Med. 2001;345:1359---67.
13. The Acute Respiratory Distress Syndrome Network. Ventilationwith lower tidal volumes as compared with traditional tidal vol-umes for acute lung injury and the acute respiratory distresssyndrome. N Engl J Med. 2000;342:1301---8.
14. Bernard GR1, Vincent JL, Laterre PF, LaRosa SP, DhainautJF, Lopez-Rodriguez A, et al., Recombinant human protein CWorldwide Evaluation in Severe Sepsis (PROWESS) study group.Efficacy and safety of recombinant human activated protein Cfor severe sepsis. N Engl J Med. 2001;344:699---709.
15. Dellinger RP, Carlet JM, Masur H, Gerlach H, Calandra T, CohenJ, et al. Surviving Sepsis Campaign guidelines for manage-ment of severe sepsis and septic shock. Intensive Care Med.2004;30:536---55.
16. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, JaeschkeR, et al. Surviving Sepsis Campaign: international guidelines formanagement of severe sepsis and septic shock. Crit Care Med.2008;36:296---327.
17. Levy MM, Dellinger RP, Townsend SR, Linde-Zwirble WT,Marshall JC, Bion J, et al. The Surviving Sepsis Campaign: resultsof an international guideline-based performance improve-ment program targeting severe sepsis. Intensive Care Med.2010;36:222---31.
18. Castellanos-Ortega A, Suberviola B, García-Astudillo LA,Holanda MS, Ortiz F, Llorca J, et al. Impact of the Surviv-ing Sepsis Campaign protocols on hospital length of stay andmortality in septic shock patients: results of a three-yearfollow-up quasi-experimental study. Crit Care Med. 2010;38:1036---43.
19. Castellanos-Ortega Á, Suberviola B, García-Astudillo LA, OrtizF, Llorca J, Delgado-Rodríguez M. Late compliance withthe sepsis resuscitation bundle: impact on mortality. Shock.2011;36:542---7.
20. Barochia AV, Cui X, Vitberg D, Suffredini AF, O’Grady NP, BanksSM, et al. Bundled care for septic shock: an analysis of clinicaltrials. Crit Care Med. 2010;38:668---78.
21. Guidet B, Martinet O, Boulain T, Philippart F, Poussel JF, MaizelJ, et al. Assessment of hemodynamic efficacy and safety of6% hydroxyethylstarch 130/0.4 vs. 0.9% NaCl fluid replacementin patients with severe sepsis: the CRYSTMAS study. Crit Care.2012;16:R94.
22. Perner A, Haase N, Guttormsen AB, Tenhunen J, Klemenzson G,Åneman A, et al., 6S Trial Group, Scandinavian Critical Care Tri-als Group. Hydroxyethyl starch 130/0.42 versus Ringer’s acetatein severe sepsis. N Engl J Med. 2012;367:124---34.
23. Myburgh JA, Finfer S, Bellomo R, Billot L, Cass A, Gattas D,et al., CHEST Investigators, Australian and New Zealand Inten-sive Care Society Clinical Trials Group. Hydroxyethyl starch orsaline for fluid resuscitation in intensive care. N Engl J Med.2012;367:1901---11.
24. Sprung CL, Annane D, Keh D, Moreno R, Singer M, Freivogel K,et al. Hydrocortisone therapy for patients with septic shock. NEngl J Med. 2008;358:111---24.
25. Finfer S, Chittock DR, Su SY, Blair D, Foster D, Dhingra V, et al.Intensive versus conventional glucose control in critically illpatients. N Engl J Med. 2009;360:1283---97.
26. Annane D, Timsit JF, Megarbane B, Martin C, Misset B, MourvillierB, et al., for The APROCCHSS Trial Investigators. Recombi-nant human activated protein C for adults with septic shock:a randomized controlled trial. Am J Respir Crit Care Med.2013;187:1091---7.
27. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H,Opal SM, et al., Surviving Sepsis Campaign Guidelines Com-mittee including The Pediatric Subgroup. Surviving SepsisCampaign: international guidelines for management of severesepsis and septic shock, 2012. Intensive Care Med. 2013;39:165---228.
28. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, OpalSM, et al., Surviving Sepsis Campaign Guidelines Committeeincluding The Pediatric Subgroup. Surviving Sepsis Campaign:international guidelines for management of severe sepsis andseptic shock, 2012. Crit Care Med. 2013;41:580---637.
29. Stevenson EK, Rubenstein AR, Radin GT, Wiener RS, WalkeyAJ. Two decades of mortality trends among patients withsevere sepsis: a comparative meta-analysis. Crit Care Med.2014;42:1---7.
30. Kaukonen KM, Bailey M, Suzuki S, Pilcher D, Bellomo R. Mortal-ity related to severe sepsis and septic shock among criticallyill patients in Australia and New Zealand, 2000---2012. JAMA.2014;311:1308---16.
31. Ferrer R, Artigas A, Levy MM, Blanco J, González-Díaz G,Garnacho-Montero J, et al., Edusepsis Study Group. Improve-ment in process of care and outcome after a multicenter severesepsis educational program in Spain. JAMA. 2008;299:2294---303.
32. Ranieri VM, Thompson BT, Barie PS, Dhainaut JF, Douglas IS,Finfer S, et al., PROWESS-SHOCK Study Group. Drotrecoginalfa (activated) in adults with septic shock. N Engl J Med.2012;366:2055---64.
33. Azkárate I, Choperena G, Salas E, Sebastián R, Lara G, ElóseguiI, et al. Epidemiology and prognostic factors in severe sep-sis/septic shock. Evolution over six years. Med Intensiva.2016;40:18---25.
34. Levy MM, Rhodes A, Phillips GS, Townsend SR, Schorr CA, BealeR, et al. Surviving Sepsis Campaign: association between per-formance metrics and outcomes in a 7.5-year study. IntensiveCare Med. 2014;40:1623---33.
35. Miller RR, Dong L, Nelson NC, Brown SM, Kuttler KG, ProbstDR, et al., Intermountain Healthcare Intensive Medicine Clin-ical Program. Multicenter implementation of a severe sepsisand septic shock treatment bundle. Am J Respir Crit Care Med.2013;188:77---82.
Declining mortality due to severe sepsis and septic shock in Spanish intensive care units 37
36. Drumheller BC, Agarwal A, Mikkelsen ME, Sante SC, Weber AL,Goyal M, et al. Risk factors for mortality despite early proto-colized resuscitation for severe sepsis and septic shock in theemergency department. J Crit Care. 2016;31:13---20.
37. Van Zanten AR, Brinkman S, Arbous MS, Abu-Hanna A, LevyMM, de Keizer NF, Netherlands Patient Safety Agency Sep-sis Expert Group. Guideline bundles adherence and mortalityin severe sepsis and septic shock. Crit Care Med. 2014;42:1890---8.
38. ARISE Investigators, ANZICS Clinical Trials GroupPeake SL,Delaney A, Bailey M, Bellomo R, Cameron PA, Cooper DJ, et al.Goal-directed resuscitation for patients with early septic shock.N Engl J Med. 2014;371:1496---506.
39. Yealy DM, Kellum JA, Huang DT, Barnato AE, Weissfeld LA,Pike F, et al., ProCESS Investigators. A randomized trial ofprotocol-based care for early septic shock. N Engl J Med.2014;370:1683---93.
40. Mouncey PR, Osborn TM, Power GS, Harrison DA, Sadique MZ,Grieve RD, et al., ProMISe Trial Investigators. Trial of early,goal-directed resuscitation for septic shock. N Engl J Med.2015;372:1301---11.
41. Sterling SA, Puskarich MA, Summers RL, Jones AE. The effect ofearly quantitative resuscitation on organ function in survivorsof septic shock. J Crit Care. 2015;30:261---3.
42. Levy MM. Early goal-directed therapy: what do we do now? CritCare. 2014;18:705.
Póvoa et al. Critical Care (2015) 19:193 DOI 10.1186/s13054-015-0921-x
RESEARCH Open Access
Clinical impact of stress dose steroids in patientswith septic shock: insights from the PROWESS-Shock trialPedro Póvoa1,2*, Jorge I F Salluh3,4, Maria L Martinez5,6, Raquel Guillamat-Prats5,6, Dianne Gallup7,Hussein R Al-Khalidi7, B Taylor Thompson8, V Marco Ranieri9 and Antonio Artigas5,6
Abstract
Introduction: The aim of our study was to evaluate the clinical impact of the administration of intravenous steroids,alone or in conjunction with drotrecogin-alfa (activated) (DrotAA), on the outcomes in septic shock patients.
Methods: We performed a sub-study of the PROWESS-Shock trial (septic shock patients who received fluids andvasopressors above a predefined threshold for at least 4 hours were randomized to receive either DrotAA orplacebo for 96 hours). A propensity score for the administration of intravenous steroids for septic shock at baselinewas constructed using multivariable logistic regression. Cox proportional hazards model using inverse probabilityof treatment weighting of the propensity score was used to estimate the effect of intravenous steroids, alone or inconjunction with DrotAA, on 28-day and 90-day all-cause mortality.
Results: A total of 1695 patients were enrolled of which 49.5% received intravenous steroids for treatment ofseptic shock at baseline (DrotAA + steroids N = 436; DrotAA + no steroids N = 414; placebo + steroids N = 403;placebo + no steroids N = 442). The propensity weighted risk of 28-day as well as 90-day mortality in those treatedvs. those not treated with steroids did not differ among those randomized to DrotAA vs. placebo (interactionp-value = 0.38 and p = 0.27, respectively) nor was a difference detected within each randomized treatment.Similarly, the course of vasopressor use and cardiovascular SOFA did not appear to be influenced by steroidtherapy. In patients with lung infection (N = 744), abdominal infection (N = 510), Gram-positive sepsis (N = 420)and Gram-negative sepsis (N = 461), the propensity weighted risk of 28-day as well as 90-day mortality in thosetreated vs. those not treated with steroids did not differ among those randomized to DrotAA vs. placebo nor wasa difference detected within each randomized treatment.
Conclusions: In the present study of septic shock patients, after adjustment for treatment selection bias, we wereunable to find noticeable positive impact from intravenous steroids for treatment of septic shock at baseline eitherin patients randomized for DrotAA or placebo.
Trial registration: Clinicaltrials.gov NCT00604214. Registered 24 January 2008.
* Correspondence: [email protected] Intensive Care Unit, São Francisco Xavier Hospital, CentroHospitalar de Lisboa Ocidental, Lisbon, Portugal2NOVA Medical School, CEDOC, New University of Lisbon, Lisbon, PortugalFull list of author information is available at the end of the article
© 2015 Póvoa et al.; licensee BioMed Central. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly credited. The Creative Commons Public DomainDedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article,unless otherwise stated.
Póvoa et al. Critical Care (2015) 19:193 Page 2 of 10
IntroductionSevere sepsis and septic shock are amongst the majorcauses of intensive care admissions [1,2] and despite therecent improvements in clinical outcomes, mortality ratesare still elevated, varying from 20 to 35% [3-5]. Improvedoutcome is mainly ascribed to earlier identification andimprovements in the process of care of sepsis rather thanspecific pharmacologic interventions [6-9].In recent years, the uses of corticosteroids and to a lesser
degree drotrecogin-alfa activated (DrotAA) have been thecornerstones of adjunctive pharmacologic therapy for severesepsis and septic shock [10-13]. However, the results of themore recent clinical trials have failed to demonstrate clinicalbenefits from either intervention [14-16]. In addition, sup-porters of the use of corticosteroids for septic shock claimthat the CORTICUS study results had limited external valid-ity due to the fact that it excluded patients whose cliniciansdecided to treat with corticosteroids. This a priori decisionpotentially biased the study by enroling patients with eitherlower severity of illness or those thought to receive lessbenefit [17,18]. Furthermore, although there are potentialsynergies in the concomitant use of corticosteroids and Dro-tAA, only one recent study evaluated this issue, and it waslimited by the discontinuation of the DrotAA arm when thedrug was withdrawn from the market [16].In the present study, we have evaluated the clinical im-
pact of corticosteroids alone or in conjunction with Dro-tAA in patients with septic shock by analyzing data fromthe PROWESS-Shock trial [15]. We hypothesized thatthe patients who received corticosteroids as part of usualcare will improve their outcomes after adjustment forbaseline imbalances.
Materials and methodsStudy design and settingPatients diagnosed with septic shock were randomlyassigned to receive either DrotAA (24 μg/kg/hour) or pla-cebo administered intravenously for 96 hours [15]. All de-tails of the PROWESS-Shock trial and its design have beenreported elsewhere (NCT00604214) [15]. The PROWESS-Shock trial was approved by the research ethics boards ofall participating institutions. Patients, next of kin, or surro-gate decision-makers gave written informed consent in ac-cordance with local requirements. The trial was conductedin accordance with the Declaration of Helsinki. The presentanalysis was proposed and approved by the Steering Com-mittee of the PROWESS-Shock trial, which in additionconsidered there was no need for further ethical approval.The concomitant use of steroids as adjunctive treatment
of septic shock, according to the recommendations at thetime [13], was at the discretion of the attending physicianand was not required by the study protocol. The questionused in the case report form to collect data on steroid usewas the following: “Was the subject treated with any
intravenous steroid therapy for septic shock during the pre-treatment period (before study drug infusion)?” No add-itional information on steroid adjunctive therapy use wasrecorded, namely type of steroid, dose, type of infusion(intermittent or continuous) or duration of therapy.For the present analysis we divided the two arms of the
trial, the DrotAA and the placebo arms, into four groupsaccording to the prescription of intravenous steroid ther-apy for septic shock. These were: 1) patients receiving ste-roids at baseline and randomized to receive DrotAA; 2)patients not receiving steroids at baseline and randomizedto receive DrotAA; 3) patients receiving steroids at base-line and randomized to receive placebo, and 4) patientsnot receiving steroids at baseline and randomized to re-ceive placebo. Study outcome was all-cause mortality at28 days and at 90 days. In addition, we assessed the courseof organ failure, in particular cardiovascular sequentialorgan failure assessment (SOFA) score, as well as the mor-tality due to secondary refractory septic shock.
Statistical analysisBaseline characteristics of patients were compared bytreatment strategy using the Kruskal-Wallis test for con-tinuous variables and Pearson Chi-square test for categor-ical variables. Kaplan-Meier estimates and survival curvesfor 28-day and 90-day all-cause mortality amongst all pa-tients were weighted by the inverse probability of receiv-ing steroid therapy at baseline, using a propensity score.Estimates of the hazard ratios (95% CI) of all-cause mor-
tality for all patients and for patients within each of the fol-lowing subgroups were displayed in a forest plot on thelog-odds scale in: 1) patients with lung infection; 2) patientswith abdominal infection; 3) patients with Gram-positivesepsis, and 4) patients with Gram-negative sepsis. The haz-ard ratio (95% CIs) estimates and the P-values for the inter-action terms of randomized treatment*baseline steroid usewere obtained using inverse probability-weighted Cox pro-portional hazards models [19].A propensity score (that is, probability of receiving
intravenous steroid therapy for septic shock at baseline)was calculated using a multivariable logistic regressionmodel after adjusting for clinically relevant patient char-acteristics at baseline. The following variables were se-lected to be included in the propensity model: age;gender; baseline acute physiology and chronic healthevaluation (APACHE) II score; baseline total SOFAscore; time between first vasopressor and study drug in-fusion (hours); number of baseline organ dysfunctions(1 to 5); baseline lactate concentration (mmol/L); IVfluids in 24 hours before start of vasopressors (mL);intravenous (IV) fluids from the start of vasopressor tothe start of study drug infusion (mL); baseline vasopres-sor score (a dimensionless variable calculated as follows:dopamine dose (mcg/kg/min) × 1) + (dobutamine dose
Póvoa et al. Critical Care (2015) 19:193 Page 3 of 10
(mcg/kg/min) × 1) + (epinephrine dose (mcg/kg/min) ×100) + (norepinephrine dose (mcg/kg/min) × 100) + (phenyl-ephrine dose (mcg/kg/min) × 100 + (vasopressin dose (mcg/kg/min) × 100) [20-22]; whether or not the patient had anabdominal infection, and whether or not the patient had alung infection. The adequacy of the propensity model wasassessed by checking the distribution of the propensityscores by treatment for a reasonable overlap and the pre-and post-inverse probability of treatment-weighting balanceof the covariates [23].Missing data were handled using the Markov chain
Monte Carlo full-imputation strategy with a single imput-ation. After imputation, continuous variables were evalu-ated for linearity and cubic splines utilized, if necessary.Descriptive statistics for change in vasopressor-free daysfrom day 1 to day 6 and change in cardiovascular SOFAfrom baseline to day 6 were provided. No statistical testingwas performed to detect differences in these outcomesacross treatment strategy because no information was ob-tained during the trial on when steroid treatment beganpost-baseline. P-values <0.05 were used to determine stat-istical significance. All analyses were performed using SASversion 9.2 (SAS Institute Inc., Cary, NC, USA).
ResultsIn the PROWESS-Shock trial, 49.5% of patients receivedintravenous steroids for treatment of septic shock atbaseline in the pretreatment period before DrotAA infu-sion (steroid use at baseline + DrotAA, n = 436; no ster-oid use at baseline + DrotAA, n = 414; steroid use atbaseline + placebo, n = 403; no steroid use at baseline +placebo, n = 442; total n = 1,695. There was a differenceof two patients from the enroled patients in thePROWESS-Shock trial, one patient was randomizedwithout prior consent and another without baselinesteroid usage information) [15].Table 1 presents the baseline characteristics of the pa-
tients at trial inclusion according to the four definedgroups. Patients who received intravenous steroids hada significantly higher APACHE II score, total SOFAscore, need for mechanical ventilation and incidence ofacute respiratory distress syndrome, and higher need forrenal replacement therapy than patients who did not re-ceive intravenous steroids in both arms of the trial. Inaddition, patients on steroids had significantly greaterneed of vasopressors, as indicated by a higher vasopres-sor score. Patients treated with steroids received a sig-nificantly lower volume of fluids in the 24 hours beforestart of vasopressors, but from the start of vasopressorsto the start of the study drug (either DrotAA or placebo)they received a significantly greater volume. Finally, themedian total amount of fluid received by each of thefour groups was comparable (DrotAA with and withoutsteroids, 8,130 and 8,238 mL, respectively; placebo with
and without steroids, 8,000 and 8,065 mL, respectively,P = 0.69).
Impact of steroids in septic shock mortalityThe propensity-weighted risk of 28-day and 90-day all-cause mortality in those treated with steroids versusthose not treated with steroids at baseline did not differamong those randomized to DrotAA versus placebo(interaction P-value = 0.38 and P = 0.27, respectively) norwas a difference detected within each randomized treat-ment (Table 2). Figure 1 presents the weighted Kaplan-Meier 90-day mortality according to randomized treat-ment (DrotAA versus placebo) and baseline steroid usefor septic shock treatment.In patients with lung infection (n = 744), patients
with abdominal infection (n = 510), patients with Gram-positive sepsis (n = 420) and patients with Gram-negativesepsis (n = 461), the propensity-weighted hazard of 28-dayas well as 90-day mortality in those treated with steroidsversus those not treated with steroids did not differ forthose randomized to DrotAA versus placebo nor was adifference detected within each randomized treatment(Figure 2).
Impact of steroids in the course of septic shockWe described the effect of intravenous steroids in theweaning from vasopressor as expressed by propensity-weighted vasopressor-free days from day 1 to day 6 aswell as propensity-weighted change of cardiovascularSOFA from baseline to day 6 (Table 3). Septic shock pa-tients randomized to DrotAA or placebo who receivedsteroids seemed to present a similar course of vasopres-sor use as those without steroid therapy. Likewise, thereappeared to be a similar decrease in cardiovascularSOFA between those treated with steroids and those nottreated with steroids, regardless of which randomizedtreatment was assigned, DrotAA or placebo. Finally, 90-day mortality (propensity-weighted) by refractory septicshock in patients randomized to DrotAA or placebowho received steroids was similar to those without ster-oid therapy (25.6, 28.8, 25.2 and 28.4%, respectively).
DiscussionWe found no benefits from the use of intravenous ste-roids for treatment of septic shock at baseline either inpatients randomized to DrotAA or placebo. In addition,we observed that intravenous steroids did not seem toinfluence the clinical course of septic shock, expressedby the cardiovascular SOFA, vasopressor-free days, anddeath from refractory shock.The role of steroids as an adjunctive therapy in the
treatment of septic shock has been a controversial issuefor many decades [24]. A large meta-analysis including17 randomized controlled trials (RCT) and 3 quasi-RCTs
Table
1Baselinech
aracteristicsbysteroidusefortrea
tmen
tof
septicsh
ockat
baselinean
drandom
ized
trea
tmen
tof
allpatients
DrotAAan
dsteroiduse
DrotAAan
dno
steroiduse
Placeb
oan
dsteroiduse
Placeb
oan
dno
steroiduse
P-valuea
Num
ber
436
414
403
442
Age
,years
0.1007
Med
ian(25th,75th
percen
tile)
64.4(52.5,74.2)
66.2(55.4,76.0)
66.2(54.5,76.6)
63.6(51.4,75.2)
Female,n/total(%)
190/436(43.6%
)169/414(40.8%
)185/403(45.9%
)194/442(43.9%
)0.5350
Prioror
preexistingcond
ition
s
Alcoh
olde
pend
ence
58/436
(13.3%
)59/414
(14.3%
)53/403
(13.2%
)61/442
(13.8%
)0.9604
Chron
icliver
disease
18/436
(4.1%)
11/414
(2.7%)
12/403
(3.0%)
17/442
(3.8%)
0.6109
Chron
icob
structiveairw
aysdisease
69/436
(15.8%
)59/414
(14.3%
)66/403
(16.4%
)65/442
(14.7%
)0.8084
Chron
icrenald
isease
49/436
(11.2%
)35/414
(8.5%)
30/403
(7.4%)
37/442
(8.4%)
0.2380
Con
gestivehe
artfailure
27/436
(6.2%)
21/414
(5.1%)
22/403
(5.5%)
23/442
(5.2%)
0.8837
Coron
aryartery
disease
57/436
(13.1%
)55/414
(13.3%
)37/403
(9.2%)
49/442
(11.1%
)0.2024
Diabe
tesmellitus
100/436(22.9%
)89/414
(21.5%
)90/403
(22.3%
)126/442(28.5%
)0.0618
Hypertension
190/436(43.6%
)191/414(46.1%
)201/403(49.9%
)201/442(45.5%
)0.3341
Immun
odeficiency
40/436
(9.2%)
11/414
(2.7%)
41/403
(10.2%
)18/442
(4.1%)
<0.0001
Malignancy(cancer)
81/436
(18.6%
)64/414
(15.5%
)83/403
(20.6%
)64/442
(14.5%
)0.0629
Pancreatitis
18/436
(4.1%)
12/414
(2.9%)
13/403
(3.2%)
12/442
(2.7%)
0.6509
Stroke
25/436
(5.7%)
23/414
(5.6%)
19/403
(4.7%)
25/442
(5.7%)
0.9048
Positivebloo
dcultu
re143/436(32.8%
)127/414(30.7%
)122/403(30.3%
)117/442(26.5%
)0.2262
Source
controlo
finfectio
nb138/151(91.4%
)137/152(90.1%
)133/149(89.3%
)131/146(89.7%
)0.9355
Num
berof
baselineorgandysfun
ctions
<0.0001
18/436(1.8%)
10/414
(2.4%)
6/403(1.5%)
17/442
(3.8%)
241/436
(9.4%)
74/414
(17.9%
)35/403
(8.7%)
77/442
(17.4%
)
3133/436(30.5%
)144/414(34.8%
)132/403(32.8%
)162/442(36.7%
)
4185/436(42.4%
)142/414(34.3%
)162/403(40.2%
)154/442(34.8%
)
569/436
(15.8%
)44/414
(10.6%
)68/403
(16.9%
)32/442
(7.2%)
Timefro
mstartof
vasopressorto
infusion
start,ho
urs
0.1390
Med
ian(25th,75th
percen
tile)
19.1(13.0,22.7)
17.0(13.0,21.5)
18.2(12.8,22.0)
18.0(11.6,22.2)
ApacheIIscore
<0.0001
Med
ian(25th,75th
percen
tile)
25.0(20.0,31.0)
24.0(19.0,30.0)
26.0(21.0,32.0)
23.0(18.0,29.0)
Recent
surgery
159/436(36.5%
)157/414(37.9%
)143/403(35.5%
)165/442(37.3%
)0.8968
Mechanicalven
tilation
379/436(86.9%
)316/414(76.3%
)361/403(89.6%
)339/442(76.7%
)<0.0001
Póvoa et al. Critical Care (2015) 19:193 Page 4 of 10
Table
1Baselinech
aracteristicsbysteroidusefortrea
tmen
tof
septicsh
ockat
baselinean
drandom
ized
trea
tmen
tof
allpatients
(Con
tinued)
Renalrep
lacemen
ttherapy
88/432
(20.4%
)32/413
(7.7%)
80/402
(19.9%
)27/442
(6.1%)
<0.0001
Acute
respiratory
distress
synd
rome
144/436(33.0%
)80/414
(19.3%
)123/403(30.5%
)114/442(25.8%
)<0.0001
Intraven
ousfluidsfro
mstartof
vasopressorto
startof
stud
ydrug
,mL
0.0093
Med
ian(25th,75th
percen
tile)
4858
(3210,7840)
4458
(2746,6865)
4635
(2904,7508)
4222
(2755,6460)
Intraven
ousfluidsin
24ho
ursbe
fore
startof
vasopressors,m
L<0.0001
Med
ian(25th,75th
percen
tile)
2902
(1850,4384)
3250
(2197,5100)
2850
(1600,4350)
3250
(2105,5000)
Vasopressorscore
<0.0001
Med
ian(25th,75th
percen
tile)
50.0(23.0,98.0)
27.0(12.2,46.4)
50.0(22.9,90.9)
21.9(12.6,44.4)
Totalseq
uentialo
rgan
failure
assessmen
t<0.0001
Med
ian(25th,75th
percen
tile)
11.0(9.0,13.0)
10.0(8.0,12.0)
11.0(9.0,13.0)
10.0(8.0,12.0)
Lactate,mmol/L
<0.0001
Med
ian(25th,75th
percen
tile)
3.0(1.9,4.8)
2.2(1.4,3.3)
3.1(2.0,5.1)
2.1(1.4,3.3)
a Kruskal-W
allis
P-valueforcontinuo
usvaria
bles,chi-squ
aretest
forcatego
rical
varia
bles.bSo
urce
controla
chieved/source
controln
ecessary.A
PACHE,acuteph
ysiology
andchroniche
alth
evalua
tion;
DrotAA,
drotrecogin-alfa
activ
ated
.
Póvoa et al. Critical Care (2015) 19:193 Page 5 of 10
Table 2 Survival analysis of 28-day and 90-day all-cause mortality (propensity-weighted), all patients, by steroid use atbaseline and randomized treatment
Mortality DrotAA andsteroid usea
DrotAA andno steroid usea
Hazardratiob
Placebo andsteroid usea
Placebo andno steroid usea
Hazardratiob
InteractionP-valueb
Day 28
Estimate 24.8% 29.5% 0.826 23.5% 23.3% 1.001 0.3764
95% CI 20.6, 29.7 25.1, 34.5 0.616, 1.107 19.2, 28.6 19.4, 27.8 0.735, 1.364
Day 90
Estimate 34.1% 37.9% 0.874 32.8% 30.6% 1.074 0.2743
95% CI 29.4, 39.3 33.1, 43.0 0.678, 1.127 27.9, 38.4 26.2, 35.4 0.822, 1.403aInverse proportional weighted Kaplan-Meier rate and associated 95% CI. bHazard ratios of steroid versus no steroid use and randomized treatment*steroid useinteraction P-value obtained using inverse proportional weighted Cox proportional hazards model. Observations weighted by the inverse probability of beingprescribed a steroid therapy at baseline using a propensity model including the following variables: age, gender, baseline acute physiology and chronic healthevaluation II score, baseline total sequential organ failure assessment, time between first vasopressor and study drug infusion (hours), number of baseline organdysfunctions (1 to 5), baseline lactate concentration (mmol/L), intravenous (IV) fluids in 24 hours before start of vasopressors (mL), IV fluids from the start ofvasopressor to the start of study drug infusion (mL), and baseline vasopressor score. DrotAA, drotrecogin-alfa activated.
Póvoa et al. Critical Care (2015) 19:193 Page 6 of 10
suggested some survival benefit of prolonged low-dosecorticosteroid therapy in septic shock patients [25].However, the analysis of the impact of low-dose cortico-steroids in septic shock mortality assessed in largeclinical registries showed little or no effect [7] or a sig-nificant increase in mortality [26], even after adjustingfor clinical severity. Similarly, a recent meta-analysisfound no statistically significant difference in mortality(relative risk 1.00, 95% CI, 0.84, 1.18) [27]. Recently, Del-linger and coworkers [8] found that hydrocortisonefailed to show any benefit on outcome (relative risk1.06) if the meta-analysis included only the six high-levelRCTs with low risk of bias [11,14,28-31] and excludedstudies with placebo mortality >60%.
Figure 1 Propensity-weighted Kaplan-Meier 90-day all-cause mortality accosteroid use for septic shock treatment. P = 0.27. DrotAA, drotrecogin-alfa ac
In the midst of these conflicting results, two recent obser-vational studies were published [32,33] that brought a littlelight to these issues [34,35]. The first study from Katsenoset al. [32], showed a potential mortality benefit from earlyinitiation of steroids (in the first 9 hours after vasopressors).However, these results are compromised by several limita-tions, namely the small and asymmetric sample size, thefact that the impact of steroid therapy was not adjusted forclinical severity nor organ dysfunction, and the very highmortality rate at 28 days (almost 70% in patients with latesteroid initiation) [34]. The study from Funk et al. [33] wasa large retrospective multicenter propensity-matched co-hort study that showed no benefit from low-dose cortico-steroids in septic shock patients either in 30-day mortality
rding to randomized treatment (DrotAA versus placebo) and baselinetivated.
Prowess-Shock: Weighted Hazard Ratio of Steroid vs. No Steroid Use Within TreatmentOutcome = 28-day Mortality
X-axis is on the log scale
P-valueInteractionHR (95% CI)Hazard RatioNumber of Patients (%)Subgroup
0.38
0.65
0.58
0.49
0.12
0.83 (0.62,1.11)
1.00 (0.74,1.36)
0.85 (0.50,1.45)
1.01 (0.59,1.72)
0.84 (0.55,1.28)
1.00 (0.65,1.54)
0.97 (0.53,1.77)
1.33 (0.68,2.58)
0.62 (0.36,1.06)
1.26 (0.62,2.55)
0.25 0.5 1 2 4
<--Steroid Better---- ---No steroid Better--->
850 (50)
845 (50)
264 (31)
246 (29)
368 (43)
376 (44)
218 (47)
202 (48)
245 (53)
216 (52)
All Patients
Abdominal Infection
Lung Infection
Gram Positive
Gram Negative
Xigris
Placebo
Xigris
Placebo
Xigris
Placebo
Xigris
Placebo
Xigris
Placebo
Prowess-Shock: Weighted Hazard Ratio of Steroid vs. No Steroid Use Within TreatmentOutcome = 90-day Mortality
X-axis is on the log scale
P-valueInteractionHR (95% CI)Hazard RatioNumber of Patients (%)Subgroup
0.27
0.75
0.56
0.18
0.12
0.87 (0.68,1.13)
1.07 (0.82,1.40)
0.78 (0.51,1.18)
0.86 (0.55,1.33)
1.01 (0.68,1.50)
1.19 (0.80,1.78)
1.06 (0.63,1.79)
1.84 (1.00,3.38)
0.65 (0.40,1.04)
1.19 (0.66,2.16)
0.25 0.5 1 2 4
<--Steroid Better---- ---No steroid Better--->
850 (50)
845 (50)
264 (31)
246 (29)
368 (43)
376 (44)
218 (47)
202 (48)
245 (53)
216 (52)
All Patients
Abdominal Infection
Lung Infection
Gram Positive
Gram Negative
Xigris
Placebo
Xigris
Placebo
Xigris
Placebo
Xigris
Placebo
Xigris
Placebo
Figure 2 Impact of steroids in septic shock mortality; 28-day and 90-day all-cause mortality (propensity-weighted) for PROWESS-Shock patientsand subgroups (abdominal infection, lung infection, Gram-positive infection and Gram-negative infection). D28, day 28; D90, day 90; HR, hazardratio; DrotAA, drotrecogin-alfa activated.
Póvoa et al. Critical Care (2015) 19:193 Page 7 of 10
or vasopressor dependence. However, in those with higherseverity, with APACHE II ≥30, there might be a benefit,whereas in the lower clinical-severity quartiles there might
be potential harm. Similarly, this study has also several limi-tations, in particular the very long period of patient inclu-sion (11 years) during which a marked change in mortality
Table 3 Course of septic shock (propensity-weighted) from baseline to day 6 by steroid use at baseline andrandomized treatment
DrotAA andsteroid use
DrotAA andno steroid use
Placebo andsteroid use
Placebo andno steroid use
Change in cardiovascular sequential organ failure assessment
Mean (SD) −2.8 (1.6) −2.7 (1.6) −2.8 (1.6) −2.9 (1.6)
Median (25th, 75th percentile) −4 (−4,-1) −4 (−4,-1) −4 (−4,-1) −4 (−4,-2)
Minimum, maximum −4, 1 −4, 1 −4, 1 −4, 1
Vasopressor-free days
Mean (SD) 2.5 (2.0) 2.7 (2.1) 2.4 (2.0) 2.9 (2.0)
Median (25th, 75th percentile) 3 (0,4) 3 (0,5) 3 (0,4) 4 (1,5)
Minimum, maximum 0, 6 0, 6 0, 6 0, 6
DrotAA, drotrecogin-alfa activated.
Póvoa et al. Critical Care (2015) 19:193 Page 8 of 10
was expectable [6,7,9], and the propensity score did not in-clude variables associated with shock severity, namely theSOFA score or the number and dose of vasopressors. Fi-nally, almost 16% of low-dose corticosteroids given to septicshock patients did not have APACHE II score recorded.Conversely, when the two largest RCTs [11,14] of low-
dose corticosteroids were analyzed, one (n = 300) suggestsa marked positive impact of steroids on mortality in septicshock only in the patients who did not respond to theshort corticotropin test, whereas the second (n = 499)found no beneficial effect irrespective of the response tothe short corticotropin test [14].However, these two RCTs are not totally comparable.
Septic shock patients in the positive trial had a higherSimplified Acute Physiology Score II at baseline, were un-responsive to vasopressors, were all under mechanicalventilation (compared to 86% in CORTICUS and 82% inPROWESS-Shock) and there was a much higher rate ofdeath at 28 days in the placebo group (61% compared with32% in the CORTICUS trial). The enrolment of patientsin the positive trial was allowed only within 8 hours afterfulfilling inclusion criteria, as compared with a 72-hourwindow in the negative trial. Therefore, some authors per-ceive that results of the negative trial represent therandomization of patients whose clinicians decided not totreat with corticosteroids, that is, those with less severeclinical presentation [17,18]. Taken together, these find-ings might suggest a potential benefit of steroids for themost severe cases at the earliest stages of septic shock.In the present analysis of PROWESS-Shock trial, we con-
firmed that steroid use in usual care was indeed reservedfor more critically ill individuals. However, we could notconfirm such benefit from this practice, even when steroidswere administered early in the course of shock, with orwithout concomitant DrotAA. The same was true in thedifferent subgroup analyses, namely of patients with lunginfection, abdominal infection, Gram-positive infection orGram-negative infection.
Both trials that assessed the efficacy of DrotAA [10,15]allowed the use of intravenous steroids at the discretionof the attending physician. In line with the original rec-ommendations of the Surviving Sepsis Campaign [12], aswell as the 2008 revision [13], the administration ofintravenous steroids for treatment of septic shock wasrecommended and as a result its prescription increasedfrom 36.0% in the original PROWESS trial to 49.5% inthe PROWESS-Shock trial.The potential synergies in the concomitant use of corti-
costeroids and DrotAA were evaluated in only one recentstudy, and this analysis was limited by the discontinuationof the DrotAA [16]. However, the authors found no signifi-cant interaction between corticosteroids and DrotAA (P =0.47). Similarly, in our analysis we were unable to find anysignificant interaction between these two drugs among thetotal patient group, or in the different subgroups, namely ofpatients with lung infection, abdominal infection, Gram-positive or Gram-negative sepsis.In addition, we were unable to demonstrate any significant
improvement in hemodynamic stability associated with theuse of corticosteroids [11,14]. Nonetheless we could notevaluate the response to the short corticotropin test, as itwas not routinely performed and if performed those datawere not collected in the PROWESS-Shock trial. However,in the past decade a significant amount of data have ques-tioned the validity of the results of such a test in this setting[36,37]. First, there is a great variability of cortisol measure-ments observed between different methods and laboratories[36]. Also, the relationship between total and free cortisollevels had also been shown to be poor [38]. Finally, it hasbeen shown in critically ill patients that cortisol productionwas 83% higher and cortisol clearance was 50% lower incomparison to matched controls. These factors account fora 3.5 times greater cortisol level in these patients [39].In our study there are also several limitations that need
to be acknowledged. The present study was not designedto stratify by the use of corticosteroids a priori and
Póvoa et al. Critical Care (2015) 19:193 Page 9 of 10
unmeasured confounders may have been missed or incom-pletely accounted for in our propensity adjustments.Enroled patients must have survived the initial resuscitationperiod to be randomized, which was on average 17 hoursfrom the onset of vasopressor use [15]. As a result only sep-tic shock patients that survived to that time point were ana-lyzed, excluding very sick septic shock patients. As a resultpatients with early refractory shock and early deaths duringthis period were not included. We were unable to analyzethe type, the dose of corticosteroid drug administered andthe duration of steroid therapy, as these data were not col-lected. Only the prescription of intravenous steroid therapyfor septic shock during the pretreatment period (beforestudy drug infusion) was recorded. In addition, data on eto-midate and fludrocortisone prescription was not collectedin the PROWESS-Shock database. Similarly the steroid-related complications namely myopathy, nosocomial infec-tions, and metabolic alteration were not fully available.However, if we consider endpoints such as all-cause mortal-ity and duration of mechanical ventilation as surrogates forthese complications, we did not find significant differencesamong the groups. Nonetheless, we acknowledge that thesesafety issues deserve in-depth analysis with specific and ro-bust data collection in future studies.The present study does have several strengths. Our ana-
lyses utilized data from a large multicenter and well-conducted RCT collected during a 25-month period andwith a 90-day follow up. In addition, the inverse probabil-ity of treatment weighting using the propensity score toperform our analysis balances measured covariates be-tween those prescribed steroids and those not prescribedsteroids [23].
ConclusionsIn the present retrospective analysis of the PROWESS-Shock trial database, we were unable to find a noticeablepositive impact from intravenous steroids for treatment ofseptic shock at baseline either in patients randomized toDrotAA or those randomized to placebo. These data couldsupport the premise that intravenous steroids should not besystematically used in patients with septic shock; howeverfurther research in a large RCT is warranted.
Key messages
� In the PROWESS-Shock trial, 49.5% of patientsreceived intravenous steroids for treatment of septicshock at baseline.
� Septic shock patients treated with intravenoussteroids at baseline had more organ dysfunction,higher APACHE II and SOFA scores, and neededmore vasopressors.
� After adjustment for treatment selection bias,intravenous steroids for treatment of septic shock
at baseline had no impact on 28-day and 90-daymortality, either in patients randomized fordrotrecogin-alfa (activated) or placebo.
� The course of septic shock, assessed by the numberof vasopressor-free days (propensity-weighted), wasalso similar in patients treated with and withoutintravenous steroids.
� No significant interaction between intravenoussteroids and drotrecogin-alfa (activated) was found.
AbbreviationsAPACHE: acute physiology and chronic health evaluation II score;DrotAA: drotrecogin-alfa (activated); RCT: randomized controlled trial;SOFA: sequential organ failure assessment score.
Competing interestsPedro Póvoa: payment for lectures from Astellas, Gilead and Astra Zeneca.Jorge IF Salluh: the author declares that he has no conflict of interest. MariaL Martinez: the author declares that she has no conflict of interest. RaquelGuillamat-Prats: the author declares that she has no conflict of interest.Dianne Gallup: the author declares that she has no conflict of interest.Hussein R Al-Khalidi: the author declares that he has no conflict of interest.B Taylor Thompson: consultancy fee and travel support from Eli Lilly;co-principal investigator of the Prowess-Shock study. V Marco Ranieri:consultancy fee and honorarium from Eli Lilly; co-principal investigator of theProwess-Shock study. Antonio Artigas: board membership of Ferrer Pharma,consultancy fee and honorarium from Rubio Lab, and payment for lecturesfrom Almirall, Astute, Grifols and Virogates.
Authors’ contributionsPP, JIFS and AA conceived the study, participated in its design andcoordination, participated in data analysis and drafted the manuscript. DGand HRA contributed to the study conception and design, carried out andsupervised data analysis and helped to draft the manuscript. MLM and RGPparticipated in study design and helped to draft the manuscript. BTT andVMR contributed to and evaluated the study design, participated in dataanalysis and helped to draft the manuscript. All authors read and approvedthe final manuscript.
AcknowledgementsEli Lilly funded the PROWESS-Shock trial [15]: the database of the trial was usedfor the present study. We thank the Steering Committee of the PROWESS-Shocktrial (funded by Eli Lilly) for accepting the proposal of the present substudy, theData Monitoring Committee and PROWESS-Shock investigators; all data areprovided in: Supplement to Ranieri VM et al. Drotrecogin alfa (activated) in adultswith septic shock. N Engl J Med 2012;366:2055–64 [15]. We thank Karen Pieper forher insights in the analyses of study data and assistance in the preparation of themanuscript. This work was presented in part at the XIX CBMI - Brazilian Congressof Intensive Care, 6 to 8 November 2014, Goiania, Brazil.
Author details1Polyvalent Intensive Care Unit, São Francisco Xavier Hospital, CentroHospitalar de Lisboa Ocidental, Lisbon, Portugal. 2NOVA Medical School,CEDOC, New University of Lisbon, Lisbon, Portugal. 3D’or Institute forResearch and Education, Rio de Janeiro, Brazil. 4Postgraduation Program,Instituto Nacional de Câncer, Rio de Janeiro, Brazil. 5Critical Care Center,Sabadell Hospital, Corporación Sanitaria Universitaria Parc Taulí, UniversitatAutonoma de Barcelona, Sabadell, Spain. 6CIBER de EnfermedadesRespiratorias (CIBERES), Madrid, Spain. 7Duke Clinical Research Institute,Durham, NC, USA. 8Pulmonary and Critical Care Unit, Department ofMedicine, Massachusetts General Hospital, Boston, USA. 9Dipartimento diAnestesiologia e Rianimazione, Azienda Ospedaliera Città della Salute e dellaScienza e di Torino_Molinette, Università di Torino, Torino, Italy.
Received: 16 February 2015 Accepted: 13 April 2015
Póvoa et al. Critical Care (2015) 19:193 Page 10 of 10
References1. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR.
Epidemiology of severe sepsis in the United States: analysis of incidence,outcome, and associated costs of care. Crit Care Med. 2001;29:1303–10.
2. Wang HE, Shapiro NI, Angus DC, Yealy DM. National estimates of severesepsis in United States emergency departments. Crit Care Med.2007;35:1928–36.
3. Asfar P, Meziani F, Hamel JF, Grelon F, Megarbane B, Anguel N, et al. Highversus low blood-pressure target in patients with septic shock. N Engl JMed. 2014;370:1583–93.
4. Caironi P, Tognoni G, Masson S, Fumagalli R, Pesenti A, Romero M, et al.Albumin replacement in patients with severe sepsis or septic shock. N EnglJ Med. 2014;370:1412–21.
5. Pro CI, Yealy DM, Kellum JA, Huang DT, Barnato AE, Weissfeld LA, et al. Arandomized trial of protocol-based care for early septic shock. N Engl JMed. 2014;370:1683–93.
6. Ferrer R, Artigas A, Levy MM, Blanco J, Gonzalez-Diaz G, Garnacho-Montero J,et al. Improvement in process of care and outcome after a multicenter severesepsis educational program in Spain. JAMA. 2008;299:2294–303.
7. Levy MM, Dellinger RP, Townsend SR, Linde-Zwirble WT, Marshall JC, Bion J, et al.The Surviving Sepsis Campaign: results of an international guideline-basedperformance improvement program targeting severe sepsis. Intensive Care Med.2010;36:222–31.
8. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al.Surviving sepsis campaign: international guidelines for management ofsevere sepsis and septic shock: 2012. Crit Care Med. 2013;41:580–637.
9. Kaukonen KM, Bailey M, Suzuki S, Pilcher D, Bellomo R. Mortality related tosevere sepsis and septic shock among critically ill patients in Australia andNew Zealand, 2000–2012. JAMA. 2014;311:1308–16.
10. Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez-RodriguezA, et al. Efficacy and safety of recombinant human activated protein C forsevere sepsis. N Engl J Med. 2001;344:699–709.
11. Annane D, Sebille V, Charpentier C, Bollaert PE, Francois B, Korach JM, et al.Effect of treatment with low doses of hydrocortisone and fludrocortisoneon mortality in patients with septic shock. JAMA. 2002;288:862–71.
12. Dellinger RP, Carlet JM, Masur H, Gerlach H, Calandra T, Cohen J, et al.Surviving Sepsis Campaign guidelines for management of severe sepsis andseptic shock. Crit Care Med. 2004;32:858–73.
13. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R, et al.Surviving Sepsis Campaign: international guidelines for management ofsevere sepsis and septic shock: 2008. Crit Care Med. 2008;36:296–327.
14. Sprung CL, Annane D, Keh D, Moreno R, Singer M, Freivogel K, et al.Hydrocortisone therapy for patients with septic shock. N Engl J Med.2008;358:111–24.
15. Ranieri VM, Thompson BT, Barie PS, Dhainaut JF, Douglas IS, Finfer S, et al.Drotrecogin alfa (activated) in adults with septic shock. N Engl J Med.2012;366:2055–64.
16. Annane D, Timsit JF, Megarbane B, Martin C, Misset B, Mourvillier B, et al.Recombinant human activated protein C for adults with septic shock: arandomized controlled trial. Am J Respir Crit Care Med. 2013;187:1091–7.
17. Mason PE, Al-Khafaji A, Milbrandt EB, Suffoletto BP, Huang DT. CORTICUS:the end of unconditional love for steroid use? Crit Care. 2009;13:309.
18. Toma A, Stone A, Green RS, Gray S. Steroids for patients in septic shock: theresults of the CORTICUS trial. CJEM. 2011;13:273–6.
19. Cole SR, Hernan MA. Adjusted survival curves with inverse probabilityweights. Comput Meth Programs Biomed. 2004;75:45–9.
20. Shore S, Nelson DP, Pearl JM, Manning PB, Wong H, Shanley TP, et al.Usefulness of corticosteroid therapy in decreasing epinephrinerequirements in critically ill infants with congenital heart disease. Am JCardiol. 2001;88:591–4.
21. Russell JA, Walley KR, Singer J, Gordon AC, Hebert PC, Cooper DJ, et al.Vasopressin versus norepinephrine infusion in patients with septic shock.N Engl J Med. 2008;358:877–87.
22. Cruz DN, Antonelli M, Fumagalli R, Foltran F, Brienza N, Donati A, et al. Earlyuse of polymyxin B hemoperfusion in abdominal septic shock: the EUPHASrandomized controlled trial. JAMA. 2009;301:2445–52.
23. Austin PC. The relative ability of different propensity score methods tobalance measured covariates between treated and untreated subjects inobservational studies. Med Decis Making. 2009;29:661–77.
24. Patel GP, Balk RA. Systemic steroids in severe sepsis and septic shock. Am JRespir Crit Care Med. 2012;185:133–9.
25. Annane D, Bellissant E, Bollaert PE, Briegel J, Confalonieri M, De Gaudio R,et al. Corticosteroids in the treatment of severe sepsis and septic shock inadults: a systematic review. JAMA. 2009;301:2362–75.
26. Beale R, Janes JM, Brunkhorst FM, Dobb G, Levy MM, Martin GS, et al. Globalutilization of low-dose corticosteroids in severe sepsis and septic shock: areport from the PROGRESS registry. Crit Care. 2010;14:R102.
27. Sligl WI, Milner Jr DA, Sundar S, Mphatswe W, Majumdar SR. Safety andefficacy of corticosteroids for the treatment of septic shock: A systematicreview and meta-analysis. Clin Infect Dis. 2009;49:93–101.
28. Bollaert PE, Charpentier C, Levy B, Debouverie M, Audibert G, Larcan A.Reversal of late septic shock with supraphysiologic doses of hydrocortisone.Crit Care Med. 1998;26:645–50.
29. Briegel J, Forst H, Haller M, Schelling G, Kilger E, Kuprat G, et al. Stress doses ofhydrocortisone reverse hyperdynamic septic shock: a prospective, randomized,double-blind, single-center study. Crit Care Med. 1999;27:723–32.
30. Oppert M, Schindler R, Husung C, Offermann K, Graf KJ, Boenisch O, et al.Low-dose hydrocortisone improves shock reversal and reduces cytokinelevels in early hyperdynamic septic shock. Crit Care Med. 2005;33:2457–64.
31. Yildiz O, Doganay M, Aygen B, Guven M, Kelestimur F, Tutuu A. Physiological-dosesteroid therapy in sepsis [ISRCTN36253388]. Crit Care. 2002;6:251–9.
32. Katsenos CS, Antonopoulou AN, Apostolidou EN, Ioakeimidou A, KalpakouGT, Papanikolaou MN, et al. Early administration of hydrocortisonereplacement after the advent of septic shock: impact on survival andimmune response. Crit Care Med. 2014;42:1651–7.
33. Funk D, Doucette S, Pisipati A, Dodek P, Marshall JC, Kumar A, et al.Low-dose corticosteroid treatment in septic shock: a propensity-matchingstudy. Crit Care Med. 2014;42:2333–41.
34. Greenberg SB, Coursin DB. Timing of corticosteroids in refractory septicshock: a key or wishful thinking? Crit Care Med. 2014;42:1733–5.
35. Allen KS, Kinasewitz GT. The pendulum of corticosteroids in sepsis swingsagain? Crit Care Med. 2014;42:2442–3.
36. Briegel J, Sprung CL, Annane D, Singer M, Keh D, Moreno R, et al.Multicenter comparison of cortisol as measured by different methods insamples of patients with septic shock. Intensive Care Med. 2009;35:2151–6.
37. Marik PE. Critical illness-related corticosteroid insufficiency. Chest.2009;135:181–93.
38. Hamrahian AH, Oseni TS, Arafah BM. Measurements of serum free cortisol incritically ill patients. N Engl J Med. 2004;350:1629–38.
39. Boonen E, Vervenne H, Meersseman P, Andrew R, Mortier L, Declercq PE,et al. Reduced cortisol metabolism during critical illness. N Engl J Med.2013;368:1477–88.
Submit your next manuscript to BioMed Centraland take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at www.biomedcentral.com/submit
RESEARCH Open Access
Antibiotic prescription patterns in the empirictherapy of severe sepsis: combination ofantimicrobials with different mechanisms ofaction reduces mortalityAna Díaz-Martín1,2,3*, María Luisa Martínez-González4, Ricard Ferrer5,6, Carlos Ortiz-Leyba1,2,3, Enrique Piacentini5,Maria Jesus Lopez-Pueyo7, Ignacio Martín-Loeches4,6, Mitchell M Levy8, Antoni Artigas4,6,José Garnacho-Montero1,2,3 and for the Edusepsis Study Group
Abstract
Introduction: Although early institution of adequate antimicrobial therapy is lifesaving in sepsis patients, optimalantimicrobial strategy has not been established. Moreover, the benefit of combination therapy over monotherapyremains to be determined. Our aims are to describe patterns of empiric antimicrobial therapy in severe sepsis,assessing the impact of combination therapy, including antimicrobials with different mechanisms of action, onmortality.
Methods: This is a Spanish national multicenter study, analyzing all patients admitted to ICUs who receivedantibiotics within the first 6 hours of diagnosis of severe sepsis or septic shock. Antibiotic-prescription patterns incommunity-acquired infections and nosocomial infections were analyzed separately and compared. We comparedthe impact on mortality of empiric antibiotic treatment, including antibiotics with different mechanisms of action,termed different-class combination therapy (DCCT), with that of monotherapy and any other combination therapypossibilities (non-DCCT).
Results: We included 1,372 patients, 1,022 (74.5%) of whom had community-acquired sepsis and 350 (25.5%) ofwhom had nosocomial sepsis. The most frequently prescribed antibiotic agents were b-lactams (902, 65.7%) andcarbapenems (345, 25.1%). DCCT was administered to 388 patients (28.3%), whereas non-DCCT was administered to984 (71.7%). The mortality rate was significantly lower in patients administered DCCTs than in those who wereadministered non-DCCTs (34% versus 40%; P = 0.042). The variables independently associated with mortality wereage, male sex, APACHE II score, and community origin of the infection. DCCT was a protective factor against in-hospital mortality (odds ratio (OR), 0.699; 95% confidence interval (CI), 0.522 to 0.936; P = 0.016), as was urologicfocus of infection (OR, 0.241; 95% CI, 0.102 to 0.569; P = 0.001).
Conclusions: b-Lactams, including carbapenems, are the most frequently prescribed antibiotics in empiric therapyin patients with severe sepsis and septic shock. Administering a combination of antimicrobials with differentmechanisms of action is associated with decreased mortality.
* Correspondence: [email protected] Care Unit, Critical Care and Emergency Department, Virgen delRocío University Hospital, Avda. Manuel Siurot s/n, Seville 41013, SpainFull list of author information is available at the end of the article
Díaz-Martín et al. Critical Care 2012, 16:R223http://ccforum.com/content/16/6/R223
© 2012 Dias-Martin et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.
IntroductionSepsis is a prevalent disorder and one of the main causesof death among hospitalized patients. Treating sepsis isassociated with high costs; however, despite advances inmedical practice, the mortality rate of sepsis has notdeclined in recent decades [1]. In Spain, the incidence ofsevere sepsis is 104 cases per 100,000 adult residents peryear, and related in-hospital mortality is 20.7%; the inci-dence of septic shock is 31 cases per 100,000 adult resi-dents per year, and related in-hospital mortality is 45.7%[2]. Sepsis present at intensive care unit (ICU) admissionand ICU-acquired sepsis clearly differ in the types ofpatients affected, the sources of infection, the microor-ganisms responsible, and the prognosis [3].Diverse studies have confirmed that the prompt institu-
tion of antimicrobial therapy active against the causativepathogen is lifesaving in patients with severe sepsis [4,5].The Surviving Sepsis Campaign strongly recommendsinitiating antibiotic therapy within the first hour of recog-nition of severe sepsis, after suitable samples have beenobtained for cultures [6].Nevertheless, although antibiotic therapy is the corner-
stone in the treatment of sepsis, the optimal antimicro-bial strategy has not been defined. Few data are availableabout antibiotic prescription patterns used most in severesepsis.Furthermore, the advantages and disadvantages of
combination therapy compared with monotherapy arecontroversial, and studies comparing the two approacheshave mainly been limited to bacteremia, pneumonia, orserious Pseudomonas aeruginosa infections [7-9]. Impor-tantly, a recent retrospective study concluded that certaincombinations of antimicrobials, including antimicrobialswith different targets, improve survival in patients withseptic shock [10].We present a secondary analysis of the Edusepsis study,
which enrolled all patients with severe sepsis and septicshock admitted to the participating ICUs during 2 monthsin 2005 and 4 months in 2006. Our aims are (a) todescribe the patterns of empiric antimicrobial therapy,analyzing the differences between community-acquiredand nosocomial infections; and (b) to compare the impacton mortality of combination therapy, including at leasttwo antimicrobials with different mechanisms of action,with that of monotherapy and other combinations ofantimicrobials.
Materials and methodsDesign of the studyWe conducted a secondary analysis of the Edusepsisstudy, a Spanish national multicenter before-and-afterstudy involving 77 ICUs [11]. In this study, carried outbetween November 2005 and 2007, data were collectedbefore and after a 2-month educational intervention
based on the Surviving Sepsis Campaign guidelines; thisapproach to improving treatment of severe sepsis is cost-effective [12]. Each participating centers’ research andethical-review boards approved the study, and patientsremained anonymous. The need for informed consentwas waived in view of the observational and anonymousnature of the study.The study included all patients in these ICUs with
severe sepsis or septic shock. The study design isdescribed in detail elsewhere [11]. In brief, severe sepsiswas defined as sepsis associated with organ dysfunctionunexplained by other causes. Septic shock was defined assepsis associated with systolic blood pressure <90 mmHg, mean arterial pressure <65 mm Hg, or a reduction insystolic blood pressure >40 mm Hg from baseline despiteadequate volume resuscitation. Patients in whom theonset of severe sepsis could not be determined wereexcluded from the analysis. The approach to data collec-tion and the quality-control measures to assure datareliability also are described elsewhere [11,12].
VariablesThe following variables were recorded: demographiccharacteristics (age and gender), types of patients (medi-cal, trauma, emergency surgery, elective surgery), sourcesof infection, location at sepsis acquisition (community-acquired or nosocomial infection), and baseline lactatelevel and organ dysfunction at sepsis diagnosis. Severityof illness was evaluated by the Acute Physiology andChronic Health Evaluation (APACHE) II score, consider-ing the worst reading in the first 24 hours in the ICU[13]. All patients were followed up until death or hospitaldischarge. The primary outcome variable was in-hospitalmortality.
Antimicrobial therapyThe antimicrobial therapy prescribed at the diagnosis ofsevere sepsis and the time from severe sepsis presenta-tion to antibiotic administration were recorded. To facili-tate subsequent analysis, antimicrobial agents weregrouped into eight antibiotic families: b-lactams (exceptcarbapenems), carbapenems, quinolones, macrolides,aminoglycosides, anti-gram-positive antibiotics (vanco-mycin, teicoplanin, and linezolid), antifungal agents, andother antimicrobial agents (including antiviral and tuber-culostatic agents). Data for community-acquired andnosocomial infections also were analyzed separately. Wealso compared the clinical characteristics of patients thatreceived different-class combination therapy (DCCT)with those of patients that received any other antimicro-bial therapy (non-DCCT).DCCT was defined as the concomitant use of two or
more antibiotics of different mechanistic classes, as recentlydefined by Kumar et al. [10], specifically b-lactams or
Díaz-Martín et al. Critical Care 2012, 16:R223http://ccforum.com/content/16/6/R223
Page 2 of 9
carbapenems with aminoglycosides, fluoroquinolones, ormacrolides/clindamycin. Monotherapy or any other combi-nation therapy was considered non-DCCT for this analysis.To assess the impact of DCCT on mortality, we analyzed
only patients who received the first dose of antimicrobialwithin the first 6 hours after severe sepsis presentation.
Statistical analysisDiscrete variables were expressed as frequencies (percen-tage), and continuous variables, as means and standarddeviations (SDs), unless stated otherwise; all statisticaltests were two-sided. Differences in categoric variableswere calculated by using c2 tests or Fisher Exact test, anddifferences in continuous variables were calculated byusing the Mann-Whitney U or Kruskal-Wallis test, asappropriate.Backward logistic regression was used to assess the fac-
tors independently associated with in-hospital mortality.To avoid spurious associations, variables entered in theregression models were those with a relation in univariate
analysis (P ≤ 0.05) or a plausible relation with the depen-dent variable. SPSS for Windows 20.0 (SPSS, Chicago, IL,USA) was used for all statistical analyses.
Results and discussionDescriptive analysisThe Edusepsis study included 2,796 patients with severesepsis or septic shock; we analyzed the 1,372 patientsthat received antibiotic therapy in the first 6 hours fromthe diagnosis of sepsis, of whom 1,022 (74.5%) had com-munity-acquired sepsis and 350 (25.5%) had nosocomialsepsis. Table 1 shows the study group’s main demo-graphics, APACHE II scores, levels of lactate, and diag-noses on admission.The most frequent sources of sepsis were pneumonia
(n = 502; 36.6%), followed by abdominal infection (n =390; 28.48%), urinary tract infection (n = 182; 13.3%),central nervous system infection (n = 50; 3.6%), skin orsoft-tissue infection (n = 54; 3.9%), and catheter-relatedinfection (n = 24; 1.7%).
Table 1 Demographic and clinical characteristics of the patients
Globaln = 1,372
Community-acquiredn = 1,022 (74.5%)
Nosocomialn = 350 (25.5%)
P
General data
Sex (male) 837 (61%) 623 (61%) 214 (61%) 0.999
Age (years) 62.24 ± 16.22 62.00 ± 16.65 62.93 ± 14.87 0.354
APACHE II 21.44 ± 7.54 21.09 ± 7.49 22.45 ± 7.61 0.004
Lactate (mM) 35.56 ± 26.94 36.76 ± 27.68 32.06 ± 24.37 0.020
Diagnosis on admission
Medical 893 (65.4%) 734 (72%) 159 (45.8%)
Trauma 25 (1.8%) 8 (0.8%) 17 (4.9%) <0.001
Emergency surgery 382 (28%) 256 (25.1%) 126 (36.3%)
Elective surgery 66 (4.86%) 21 (2.1%) 45 (13%)
Type of infection
Pneumonia 502 (36.6%) 362 (35.4%) 140 (40%)
Abdominal 390 (28.4%) 270 (26.4%) 120 (34.3%)
Urologic 182 (13.3%) 163 (15.9%) 19 (5.4%)
Meningitis 50 (3.6%) 47 (4.6%) 3 (0.9%) <0.001
SSTI 54 (3.9%) 46 (4.5%) 8 (2.3%)
Catheter 24 (1.7%) 6 (0.6%) 18 (5.1%)
Others 138 (3.1%) 108 (10.6%) 30 (8.6%)
More than one focus 32 (2.3%) 20 (2.0%) 12 (3.4%)
Organ failure
Hemodynamic 1,129 (82.3%) 845 (82.7%) 284 (81.1%) 0.517
Respiratory 880 (64.1%) 641 (62.7%) 239 (68.3%) 0.062
Renal ,1006 (73.3%) 754 (73.8%) 252 (72.0%) 0.529
Hepatic 238 (17.3%) 176 (17.2%) 62 (17.7%) 0.870
Hematologic 344 (25.1%) 266 (26.0%) 78 (22.3%) 0.175
Coagulation 502 (36.6%) 394 (38.6%) 108 (30.9%) 0.010
Mortality 526 (38.3%) 356 (34.8%) 170 (48.6%) <0.001
Díaz-Martín et al. Critical Care 2012, 16:R223http://ccforum.com/content/16/6/R223
Page 3 of 9
Antimicrobial treatments prescribedThe most frequently prescribed antibiotic agents were b-lactams (n = 902; 65.7%), carbapenems (n = 345; 25.1%),and quinolones (n = 282; 20.6%). Table 2 presents thedata for the entire group of patients who received empiricantibiotic therapy within 6 hours of admission, and forthe groups of patients with community-acquired (n =1,022; 74.5%) and nosocomial infections (n = 350; 25.5%).The distribution of the antibiotics prescribed for commu-
nity-acquired infections was similar to that for the overallgroup, with predominance of b-lactams (n = 708; 69.3%),quinolones (n = 241; 23.6%), and carbapenems (n = 218;21.3%), , whereas in the group with nosocomial infections,although b-lactams were also the most-used treatment(n = 194; 55.4%), carbapenems were second (n = 127;36.3%), followed by aminoglycosides (n = 69; 19.7%) andanti-gram-positive agents (n = 65; 18.6%). Macrolides andquinolones were more frequently used in community-acquired sepsis than in nosocomial sepsis (see Table 2).
DCCT and non-DCCT groupsDCCTs were administered to 388 patients (28.3%), andnon-DCCTs, to 984 (71.7%). Table 3 shows the demo-graphic characteristics, diagnosis at admission, incidenceof associated organ failure, and sources of infection ofpatients in the DCCT and non-DCCT groups. Sex distri-bution, age, APACHE II score, and lactate levels werevery similar in the two groups.Significant differences between the two groups were
found in diagnosis at admission and source of infection.In the DCCT group, the percentage of patients with med-ical diagnoses was higher (79.9% versus 59.6%; P < 0.001)and the percentage with emergency surgical diagnoseswas lower (15.2% versus 33%; P < 0.001). The most com-mon source of sepsis was pneumonia in the DCCT group(59% versus 27.7%; P < 0.001) and abdominal infection inthe non-DCCT group (14.4% versus 33.9%; P < 0.001).Although the median number of organ failures was the
same in both groups, significant differences were notedin the organ-failure distribution: respiratory failure was
more common in the DCCT group (74.5% versus 60.1%;P < 0.001) and renal failure was more common in thenon-DCCT group (68% versus 75.4%; P = 0.007).In the DCCT group, the most frequently used agents
were b-lactams (n = 320; 82.5%), followed by quinolones(n = 186; 47.9%), aminoglycosides (n = 158; 40.7%), andcarbapenems (n = 76; 19.6%) (Table 4). These agents wereused in the following combinations: (a) a b-lactam plus anaminoglycoside or a quinolone or a macrolide (n = 312;80.4%); the most common combination in this group wasa b-lactam plus a quinolone (n = 163; 52.2%); (b) a carba-penem plus an aminoglycoside or a quinolone or a macro-lide (n = 68; 17.5%); the most common combination inthis group was a carbapenem plus an aminoglycoside (n =46; 67.6%); and (c) a b-lactam plus a carbapenem (n = 8;2.1%), usually associated with an aminoglycoside (n = 6;75.0%) (data not shown in table). It is noteworthy thatDCCT consisted only of a b-lactam or carbapenem plus amacrolide and/or an aminoglycoside and/or a quinolonein 311 (80%) patients; thus, other antimicrobials (antifun-gals, anti-gram-positive agents, and so on) were also admi-nistered in DCCT in only 75 (20%) (data not shown).
Predictors of mortalityIn the univariate analysis, factors significantly associatedwith mortality were gender, age, APACHE II score, lac-tate levels, source of infection, and DCCT (Table 5).Mortality was significantly lower in the DCCT group(34.0% versus 40%; P = 0.042). In the multivariate analysis(Table 6), including the variables that were significantlyassociated with mortality in the univariate analysis,higher age (OR, 1.023; 95% CI, 1.014 to 1.032; P < 0.001),male sex (OR, 1.350; 95% CI, 1.041 to 1.750; P = 0.024),higher APACHE II score (OR, 1.099; 95% CI, 1.099 to1.141; P < 0.001), and community-acquired infection(OR, 1.487; 95% CI, 1.119 to 1.974; P = 0.006) was asso-ciated with higher mortality, whereas urologic focus ofinfection (OR, 0.241; 95% CI, 0.102 to 0.569; P = 0.001)and DCCT were associated with lower mortality (OR,0.699; 95% CI, 0.522 to 0.936; P = 0.016).
Table 2 Antibiotic distribution in the entire cohort and in patients with community-acquired and nosocomial sepsis
Antibiotics Globaln = 1,372
Community-acquiredn = 1,022 (74.5%)
Nosocomialn = 350 (25.5%)
P
b-lactams 902 (65.7%) 708 (69.3%) 194 (55.4%) <0.001
Carbapenems 345 (25.1%) 218 (21.3%) 127 (36.3%) <0.001
Quinolones 282 (20.6%) 241 (23.6%) 41 (11.7%) <0.001
Aminoglycosides 183 (13.3%) 114 (11.2%) 69 (19.7%) <0.001
Macrolides 60 (4.4%) 54 (5.3%) 6 (1.7%) 0.004
Anti-gram-positive 161 (11.7%) 96 (9.4%) 65 (18.6%) <0.001
Antifungals 38 (2.8%) 20 (2.0%) 18 (5.1%) 0.004
Others 151 (11.0%) 111 (10.9%) 40 (11.4%) 0.767
Díaz-Martín et al. Critical Care 2012, 16:R223http://ccforum.com/content/16/6/R223
Page 4 of 9
For the DCCT-combination treatments associated withreductions in mortality, the results of the analysis,excluding patients who died within the first 6 hours, weresimilar to the results including these patients; hence,no evidence of immortal bias was found in our results.
DiscussionThis secondary analysis of the Edusepsis study reveals inter-esting data about the patterns of antibiotic prescription
in patients with severe sepsis and septic shock and aboutthe characteristics of patients receiving combination ther-apy, including antimicrobials, with different mechanisms ofaction (DCCTs) versus those receiving either monotherapyor any other combinations of antimicrobials (non-DCCTs).Our study confirms the increased survival in patients admi-nistered DCCTs (b-lactams plus aminoglycosides, quino-lones, or macrolides/clindamycin) within the first 6 hoursof severe sepsis presentation. We excluded patients that
Table 3 Comparisons of patients treated with DCCT or non-DCCT
DCCT groupn = 388 (28.3%)
Non-DCCT groupn = 984 (71.7%)
P
General data
Sex (male) 247 (63.7%) 590 (60%) 0.219
Age (years) 60.88 ± 16.79 62.78 ± 15.96 0.057
APACHE II 21.35 ± 7.43 21.47 ± 7.58 0.790
Lactate (mM) 36.37 ± 26.99 35.22 ± 26.93 0.578
Diagnosis on admission
Medical 310 (79.9%) 583 (59.6%) <0.001
Trauma 3 (0.8%) 22 (2.2%) 0.074
Emergency surgery 59 (15.2%) 323 (33%) <0.001
Elective surgery 16 (4.1%) 50 (5.1%) 0.487
Type of infection
Abdominal 56 (14.4%) 334 (33.9%) <0.001
Urologic 44 (11.3%) 138 (14%) 0.216
Meningitis 5 (1.3%) 45 (4.6%) 0.002
Skin and/or soft-tissue 6 (1.5%) 48 (4.9%) 0.003
Catheter 4 (1%) 20 (2%) 0.256
Others 31 (8%) 107 (10.9%) 0.134
More than one focus 13 (3.4%) 19 (1.9%) 0.162
Organ failure
Number of organ failures 2.98 ± 1.26 2.98 ± 1.25 0.955
Hemodynamic 319 (82.2%) 810 (82.3%) 0.999
Respiratory 289 (74.5%) 591 (60.1%) <0.001
Renal 264 (68%) 742 (75.4%) 0.007
Hepatic 61 (15.7%) 177 (18%) 0.343
Hematologic 94 (24.2%) 250 (25.4%) 0.679
Coagulation 131 (33.8%) 371 (37.7%) 0.191
Mortality 132 (34%) 394 (40%) 0.042
Table 4 Antibiotic prescription in patients treated with DCCT or non-DCCT
Antibiotics Non-DCCT groupn = 984 (71.7%)
DCCT groupn = 388 (28.3%)
P
b-Lactams 582 (59.1%) 320 (82.5%) <0.001
Carbapenems 269 (27.3%) 76 (19.6%) 0.003
Quinolones 96 (9.8%) 186 (47.9%) <0.001
Aminoglycosides 25 (2.5%) 158 (40.7%) <0.001
Macrolides 7 (0.7%) 53 (13.7%) <0.001
Anti-gram-positive 120 (12.2%) 41 (10.6%) 0.456
Antifungals 21 (2.1%) 17 (4.4%) 0.028
Others 121 (12.3%) 30 (7.7%) 0.016
Díaz-Martín et al. Critical Care 2012, 16:R223http://ccforum.com/content/16/6/R223
Page 5 of 9
received antimicrobial therapy after 6 hours of severe sepsisdiagnosis from this analysis, because strong evidence indi-cates that early administration increases survival in patientswith severe sepsis or septic shock [4,5,10].Appropriate empiric antimicrobial therapy is crucial for
the survival of sepsis patients [4,5]. Formerly, multidrug-resistant pathogens were found almost exclusively innosocomial infections. However, community-acquiredinfections are now often caused by antibiotic-resistantbacteria (for example, extended-spectrum b-lactamase-producing Enterobacteriaceae, multidrug-resistant Pseu-domonas aeruginosa, or methicillin-resistant Staphylococ-cus aureus) [14,15]. This striking change in epidemiologymay explain why the initial therapy frequently includes acombination of different antimicrobial agents [16].
b-Lactams, including carbapenems, are the most com-monly used antibiotics in the critical care setting [17].Likewise, this antibiotic family constitutes the mainstayof empiric treatment in patients with severe sepsis orseptic shock, whether administered alone or in combina-tion with other antimicrobials. Carbapenems are morefrequently prescribed in patients with nosocomial sepsis,although it is worth mentioning that one in five patientswith community-acquired sepsis is treated empiricallywith a carbapenem. This may reflect the increase in mul-tidrug-resistant gram-negative pathogens in the commu-nity [14]. Carbapenems might have been analyzed inconjunction with the rest of b-lactams. However, wedecided to analyze them separately from other b-lactamsbecause of the broader-spectrum, major role in empiricantibiotic therapy and the widespread use in the ICU.Quinolones are used mainly in community-acquired
infections and in combination therapy [18]. The extendeduse of quinolones in combination therapy in patientswith severe community-acquired pneumonia may explainthe increasing rate of quinolone resistance among noso-comial gram-negative pathogens [18,19].Numerous studies have evaluated the likely superiority
of combination therapy in patients with diverse types ofinfections. A French multicenter study of critical patientswith acute peritonitis found no difference in the rate oftherapeutic failure or length of antibiotic treatment whenb-lactams were administered alone or in combinationwith aminoglycosides, concluding that aminoglycosidesshould be added only when an infection by Pseudomonasspp or Enterococcus spp is suspected [20]. Two rando-mized clinical trials found no benefits of combinationtherapy over monotherapy in patients with ventilator-
Table 5 Univariate analysis of factors associated with in-hospital mortality
Globaln = 1,372
Survivorsn = 846 (61.7%)
Nonsurvivorsn = 526 (38.3%)
P
General data
Sex (male) 837 (61.0%) 489 (57.8%) 348 (66.2%) 0.002
Age (years) 62.2 ± 16.2 59.80 ± 16.81 66.16 ± 14.39 <0.001
APACHE II 21.4 ± 7.5 19.20 ± 6.86 25.09 ± 7.17 <0.001
Lactate (mM) 35.6 ± 26.9 31.09 ± 22.54 43.09 ± 31.69 <0.001
Type of infection
Pneumonia 502 (36.6%) 289 (34.2%) 213 (40.5%)
Abdominal 390 (28.4%) 228 (27%) 162 (30.8%)
Urologic 181 (13.3%) 142 (16.8%) 40 (7.6%)
Meningitis 50 (3.6%) 40 (4.7%) 10 (1.9%) <0.001
Skin and soft-tissue 54 (3.9%) 37 (4.4%) 17 (3.2%)
Catheter 24 (1.7%) 15 (1.8%) 9 (1.7%)
Others 138 (10.1%) 78 (9.2%) 60 (11.4%)
More than one focus 32 (2.3%) 17 (2%) 15 (2.9%)
Community acquired 1,022 (74.5%) 666 (78.7%) 356 (67.7%) <0.001
DCCT 388 (28.3%) 256 (30.3%) 132 (25.1%) 0.042
Table 6 Multivariate analysis of risk factors for mortality
Factors OR CI (95%) P
Age (years) 1.023 (1.014-1.032) <0.001
Sex (male) 1.350 (1.041-1.750) 0.024
APACHE II 1.099 (1.099-1.141) <0.001
Community-acquired 1.487 (1.119-1.974) 0.006
DCCT 0.699 (0.522-0.936) 0.016
Focus of infection
Pneumonia 0.784 (0.358-1.718) 0.543
Abdominal 0.595 (0.269-1.317) 0.200
Urologic 0.241 (0.102-0.569) 0.001
Meningitis 0.357 (0.122-1.046) 0.060
Skin and soft-tissue 0.424 (0.157-1.141) 0.089
Catheter 0.441 (0.135-1.445) 0.177
Others 0.772 (0.330-1.806) 0.551
More than one focus 1
Díaz-Martín et al. Critical Care 2012, 16:R223http://ccforum.com/content/16/6/R223
Page 6 of 9
associated pneumonia [21,22]. Moreover, in one trial,monotherapy was associated with lower rates of thera-peutic failure, superinfection, and side effects [22].Conversely, diverse studies have demonstrated lower
mortality and length of stay in patients with pneumococcalbacteremia or with community-acquired pneumoniareceiving combination therapy, including a b-lactam plus amacrolide or a quinolone, than in those receiving mono-therapy [23-25]. In these studies, the benefits seem to berestricted to more-severe patients or to those in septicshock [18,23]. Conversely, a recent retrospective studyconcluded that, in bacteremia caused by gram-negativebacilli, combination therapy with b-lactams and fluoroqui-nolones was associated with a reduction in 28-day crudemortality only among less severely ill patients [7].Two meta-analyses of studies performed in patients with
gram-negative bacteremia or sepsis found no benefit ofcombination therapy over monotherapy, except when bac-teremia was caused by multidrug-resistant bacteria orPseudomonas spp [26,27]. Moreover, higher rates of sideeffects (mainly nephrotoxicity) were reported in the groupof patients treated with b-lactam antibiotics plus amino-glycosides. More recently, a meta-analytic/meta-regressionstudy that included 50 studies found that combinationantibiotic therapy improves survival, particularly in septicshock patients, but may be harmful to less severely illpatients [28].Nevertheless, few data are available about the impact
on the outcome of combination therapy in large cohortsof patients with severe sepsis or septic shock. A recentpropensity-matched analysis concluded that, in patientswith septic shock, the use of combination therapy withtwo or more antibiotics of different mechanistic classeswas associated with lower 28-day mortality, shorter ICUstay, and lower in-hospital mortality [10].Our results confirm that combination therapy, including
two or more antimicrobials with different mechanisms ofaction (b-lactams in combination with aminoglycosides,fluoroquinolones, or macrolides/clindamycin), adminis-tered within the first 6 hours of sepsis presentation is anindependent protective factor against in-hospital mortality.Interestingly, severity of illness measured by APACHE IIscore, basal lactate levels, and the presence of hemody-namic failure did not differ between patients receivingDCCTs and those receiving non-DCCTs.The choice of empiric antimicrobial therapy is based on
the clinical presentation of the infection, the characteristicsof the patient, the local ecology, and previous antibioticexposure. Reducing the antibiotic pressure and side effectsare the main reasons for choosing monotherapy. Conver-sely, the main reason for prescribing combination therapyfor critically ill sepsis patient is to broaden the antimicro-bial spectrum in an attempt to ensure the coverage of alllikely pathogens. Our results permit us to speculate that
the synergistic mechanisms of different antimicrobial com-binations, or the immunomodulatory effects described withmacrolides and quinolones, may be of clinical transcen-dence in patients with severe sepsis or septic shock [29-31].Our study has several limitations. First, a major limita-
tion in our study is the lack of microbiology data due tothe initial study design. Accordingly, no data are availableon antibiotic susceptibility, appropriateness of antimicro-bial therapy, or the presence of bacteremia. Appropriateantimicrobial therapy based on culture results was animportant determinant of survival in a large cohort ofpatients with severe sepsis [32].Second, because of the absence of microbiology data,
we cannot explore whether the positive impact on mor-tality observed with DCCTs is related to a synergisticeffect of two mechanistically distinct antibiotics or abroader range of coverage with two or more agents.Third, we did not evaluate source control and other
important measures included in the Surviving SepsisCampaign care bundles.Fourth, this is a secondary analysis of an observational
study. Nevertheless, properly designed observational stu-dies with the appropriate analytic approach can providevaluable information on treatment effectiveness [4].However, our study has also important strengths. We
prospectively enrolled a large cohort of ICU patients withsevere sepsis and septic shock in a short time and moni-tored them until death or hospital discharge, resulting ina homogeneous database with high quality-control mea-sures that assure data validity [4]. Finally, our conclusionsare strengthened by absence of immortal bias.
Conclusionsb-Lactams, including carbapenems, are the mainstay ofempiric therapy in patients with severe sepsis and septicshock. Carbapenems are more frequently prescribed inpatients with nosocomial sepsis, although up to one infive patients with community-acquired sepsis is treatedempirically with a carbapenem. Our study supports thehypothesis that early administration of antimicrobialswith different mechanistic targets is associated withdecreased in-hospital mortality. Our findings extendthose of the propensity-matched analysis in patients withseptic shock published by Kumar et al. [10] because wealso included patients with severe sepsis, underlining theurgent need for well-designed randomized controlledtrials to evaluate the clinical benefit of DCCTs in criti-cally ill sepsis patients.
Key messages• b-Lactams, including carbapenems, are the antibio-tics most usually used in the critical care setting.• Although carbapenems are more frequently pre-scribed in patients with nosocomial sepsis, one in five
Díaz-Martín et al. Critical Care 2012, 16:R223http://ccforum.com/content/16/6/R223
Page 7 of 9
patients with community-acquired sepsis is treatedempirically with a carbapenem.• Urologic focus of infection is associated with thelowest mortality rate in patients with severe sepsis orseptic shock.• In our series, the mortality rate was significantlylower in patients receiving DCCTs than in thosereceiving non-DCCTs.• A DCCT in empiric therapy is a protective factor formortality in patients with severe sepsis or septic shock.
AbbreviationsAPACHE II: Acute Physiology and Chronic Health Evaluation II; DCCT:different-class combination therapy; ICU: intensive care unit.
AcknowledgementsWe received statistical support from David Suarez, MSc from Unidad deEpidemiologia, Fundación Parc Tauli, Universidad Autónoma de Barcelona.Barcelona, Spain; and data monitoring from Gemma Gomà, research nursefrom Centro de Críticos, Hospital de Sabadell. Sabadell, Spain.The Edusepsis Study Group included Ana Navas, Ricard Ferrer, AntonioArtigas (Hospital de Sabadell, Consorci Hospitalari Parc Tauli); María Álvarez(Hospital de Terrassa); Josep Maria Sirvent, Sara Herranz Ulldemolins (HospitalUniversitari Josep Trueta de Girona); Pedro Galdós, Goiatz Balziscueta(Hospital General de Móstoles); Pilar Marco, Izaskun Azkarate (Hospital deDonostia); Rafael Sierra (Hospital Universitario Puerta del Mar); José JavierIzua (Hospital Virgen del Camino); José Castaño (Hospital Universitario Virgende las Nieves); Alfonso Ambrós, Julian Ortega (Hospital General de CiudadReal); Virgilio Corcoles (Complejo Hospitalario Universitario de Albacete); LuisTamayo (Hospital Río Carrión); Demetrio Carriedo, Milagros Llorente (Hospitalde León); Paz Merino, Elena Bustamante (Hospital Can Misses); EduardoPalencia, Pablo García Olivares, Patricia Santa Teresa Zamarro (HospitalGregorio Marañon Madrid); Carlos Pérez (Hospital Santiago Apóstol); AnaRenedo (Hospital Morales Messeguer); Silvestre Nicolás-Franco (HospitalRafael Méndez); María Salomé Sánchez (Hospital Vega Baja); Francisco JavierGil (Hospital Santa Maria del Rosell); María Jesús Gómez (Hospital GeneralUniversitario Reina Sofía de Murcia); Enrique Piacentini (Hospital Mútua deTerrassa); Ana Loza (Hospital Universitario de Valme); Jordi Ibáñez (HospitalSon Dureta); Silvia Rodríguez (Hospital de Manresa); Jose Ángel Berezo, JesúsBlanco (Hospital Río Hortega Valladolid); Angeles Gabán, Mª Jesús LópezCambra, Alec Tallet (Hospital General de Segovia); Miguel Martínez, JoseAntonio Fernández, Fernando Callejo, Mª Jesús López Pueyo (HospitalGeneral Yagüe); Francisco Gandía (Hospital Clínico Universitario deValladolid); Mª José Fernández (Hospital Santa Bárbara); Juan CarlosBallesteros (Hospital Universitario de Salamanca); María Teresa Antuña,Santiago Herrero (Hospital de Cabueñes); Manuel Valledor, Mª Jose Gutierrez(Hospital San Agustín); Carmen Pérez (Hospital Universitario Insular GranCanaria); Oscar Rodríguez (Hospital Clínico Universitario de Valencia); RafaelDominguez (Hospital Alto Guadalquivir); Josefa Peinado (Empresa PúblicaHospital de Poniente); María Victoria de la Torre, Cristina Salazar (HospitalVirgen Victoria de Málaga); Mª Cruz Martín, Joaquin Ramon (Centro MédicoDelfos); Fernando Iglesias Llaca, Lorena Forcelledo Espina, Francisco TaboadaCosta, José Antonio Gonzalo Guerra (Hospital Universitario Central deAsturias); Francisco José Guerrero, Felipe Cañada, Mª Milagros Balaguer,Isabel Mertín, Carmen López, Daniel Sánchez (Hospital Torrecardenas); JosepCosta, Calizaya (Hospital de Barcelona SCIAS); Angel Arenaza, Ana Mª Morillo,Daniel Del Toro, Tomás Guzman (Hospital Virgen de la Macarena); AntonioBlesa, Fernando Martínez, Alejandro Moneo (Hospital San Carlos); Mª JesúsBroch (Hospital de Sagunt); Jose Antonio Camacho (Hospital San Agustín deLinares); Francisco J. Garcia (Hospital de Montilla); Xosé Luis Pérez (HospitalUniversitario de Bellvitge); Nieves Garcia (Hospital Universitario La Princesa);Juan Carlos Ruiz, Jesús Caballero, Esther Francisco, Tania Requena, AdolfoRuiz, José Luis Bóveda (Hospital Universitari Vall Hebrón); José Miguel Soto,Constantino Tormo (Hospital Universitario Dr Peset); Rafael Blancas (HospitalLa Mancha-Centro); Manuel Quintana, Miguel Ángel Taberna (HospitalNuestra Sra del Prado); Jose Maria Añon, Juan B. Aranjo (Hospital Virgen dela Luz); Manuel Rodríguez (Hospital Juan Ramon Jiménez); José Maria Garcia
(Hospital La Serrania de Ronda); Mª Isabel Rodríguez (Hospital General deBaza); Mª Jesús Huertos (Hospital Universitario Puerto Real); Carlos Ortiz(Hospital Virgen del Rocio); Mª Eugenia Yuste (Hospital Universitario SanCecilio); Juan Francisco Machado (Hospital Santa Ana-Motril); Dolores Ocaña(Hospital La Inmaculada); Ramón Vegas (Hospital Valle de los Pedroches);and Luis Vallejo (Hospital SAS La Linea).
Author details1Intensive Care Unit, Critical Care and Emergency Department, Virgen delRocío University Hospital, Avda. Manuel Siurot s/n, Seville 41013, Spain.2Instituto de Biomedicina de Sevilla (IBIS), Virgen del Rocío UniversityHospital/CSIC/Seville University, 41013 Seville, Spain. 3Spanish Network forResearch in Infectious Disease (REIPI), Virgen del Rocío University Hospital,Avda. Manuel Siurot, s/n. 41013, Seville, Spain. 4Critical Care Center, SabadellHospital, Autonomous University of Barcelona, Corporació SanitariaUniversitària Parc Taulí, 08208 Sabadell, Spain. 5Intensive Care Unit, MútuaTerrassa University Hospital, University of Barcelona, Plaça Dr, Robert 5,Terrassa, 08221, Barcelona, Spain. 6CIBER-Enfermedades Respiratorias.7Intensive Care Unit, General Yagüe Hospital, Avda del Cid Campeador, 9609005 Burgos, Spain. 8Medical Intensive Care Unit, Rhode Island Hospital/Brown University, 593 Eddy St,, Providence, RI 02903, USA.
Authors’ contributionsAs principal investigator, ADM had full access to all data in the study andtakes responsibility for the integrity of the data and the accuracy of the dataanalysis. Study concept and design were performed by ADM, MLMG, RF, AA,JGM. RF completed the acquisition of data. Analysis and interpretation ofdata were performed by ADM and RF. Drafting of the manuscript wasexecuted by ADM and JGM. Critical revision of the manuscript for importantintellectual content was done by MLMG, COL, EP, MJLP, IML, ML, and AA. RFand AA carried out the Edusepsis General Coordination and Surviving SepsisCampaign coordination by ML. All authors critically revised and approvedthe manuscript.
Competing interestsThe authors declare that they have no competing interests.
Received: 30 July 2012 Revised: 9 October 2012Accepted: 18 October 2012 Published: 18 November 2012
References1. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR:
Epidemiology of severe sepsis in the United States: analysis ofincidence, outcome, and associated costs of care. Crit Care Med 2001,29:1303-1310.
2. Esteban A, Frutos-Vivar F, Ferguson ND, Penuelas O, Lorente JA, Gordo F,Honrubia T, Algora A, Bustos A, Garcia G, Díaz-Regañón IR, de Luna RR:Sepsis incidence and outcome: contrasting the intensive care unit withthe hospital ward. Crit Care Med 2007, 35:1284-1289.
3. Alberti C, Brun-Buisson C, Burchardi H, Martin C, Goodman S, Artigas A,Sicignano A, Palazzo M, Moreno R, Boulme R, Lepage E, Le Gall R:Epidemiology of sepsis and infection in ICU patients from aninternational multicentre cohort study. Intensive Care Med 2002,28:108-121.
4. Ferrer R, Artigas A, Suarez D, Palencia E, Levy MM, Arenzana A, Perez XL,Sirvent JM: Effectiveness of treatments for severe sepsis: a prospective,multicenter, observational study. Am J Respir Crit Care Med 2009, 180:861-866.
5. Garnacho-Montero J, Garcia-Garmendia JL, Barrero-Almodovar A, Jimenez-Jimenez FJ, Perez-Paredes C, Ortiz-Leyba C: Impact of adequate empiricalantibiotic therapy on the outcome of patients admitted to the intensivecare unit with sepsis. Crit Care Med 2003, 31:2742-2751.
6. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R, Reinhart K,Angus DC, Brun-Buisson C, Beale R, Calandra T, Dhainaut JF, Gerlach H,Harvey M, Marini JJ, Marshall J, Ranieri M, Ramsay G, Sevransky J,Thompson BT, Townsend S, Vender JS, Zimmerman JL, Vincent JL:Surviving Sepsis Campaign: international guidelines for management ofsevere sepsis and septic shock: 2008. Intensive Care Med 2008, 34:17-60.
7. Al-Hasan MN, Wilson JW, Lahr BD, Thomsen KM, Eckel-Passow JE, Vetter EA,Tleyjeh IM, Baddour LM: Beta-lactam and fluoroquinolone combinationantibiotic therapy for bacteremia caused by gram-negative bacilli.Antimicrob Agents Chemother 2009, 53:1386-1394.
Díaz-Martín et al. Critical Care 2012, 16:R223http://ccforum.com/content/16/6/R223
Page 8 of 9
8. Garnacho-Montero J, Sa-Borges M, Sole-Violan J, Barcenilla F, Escoresca-Ortega A, Ochoa M, Cayuela A, Rello J: Optimal management therapy forPseudomonas aeruginosa ventilator-associated pneumonia: anobservational, multicenter study comparing monotherapy withcombination antibiotic therapy. Crit Care Med 2007, 35:1888-1895.
9. Weiss K, Tillotson GS: The controversy of combination vs monotherapy inthe treatment of hospitalized community-acquired pneumonia. Chest2005, 128:940-946.
10. Kumar A, Zarychanski R, Light B, Parrillo J, Maki D, Simon D, Laporta D,Lapinsky S, Ellis P, Mirzanejad Y, Martinka G, Keenan S, Wood G, Arabi Y,Feinstein D, Kumar A, Dodek P, Kravetsky L, Doucette S: Early combinationantibiotic therapy yields improved survival compared with monotherapyin septic shock: a propensity-matched analysis. Crit Care Med 2010,38:1773-1785.
11. Ferrer R, Artigas A, Levy MM, Blanco J, Gonzalez-Diaz G, Garnacho-Montero J, Ibanez J, Palencia E, Quintana M, de la Torre-Prados MV:Improvement in process of care and outcome after a multicenter severesepsis educational program in Spain. JAMA 2008, 299:2294-2303.
12. Suarez D, Ferrer R, Artigas A, Azkarate I, Garnacho-Montero J, Goma G,Levy MM, Ruiz JC: Cost-effectiveness of the Surviving Sepsis Campaignprotocol for severe sepsis: a prospective nation-wide study in Spain.Intensive Care Med 2011, 37:444-452.
13. Knaus WA, Wagner DP, Draper EA, Zimmerman JE, Bergner M, Bastos PG,Sirio CA, Murphy DJ, Lotring T, Damiano A, Harrell FE: The APACHE IIIprognostic system. Risk prediction of hospital mortality for critically illhospitalized adults. Chest 1991, 100:1619-1636.
14. Angel Diaz M, Ramon Hernandez J, Martinez-Martinez L, Rodriguez-Bano J,Pascual A: [Extended-spectrum beta-lactamase-producing Escherichia coliand Klebsiella pneumoniae in Spanish hospitals: 2nd multicenter study(GEIH-BLEE project, 2006)]. Enferm Infecc Microbiol Clin 2009, 27:503-510.
15. Kock R, Becker K, Cookson B, van Gemert-Pijnen JE, Harbarth S, Kluytmans J,Mielke M, Peters G, Skov RL, Struelens MJ, Tacconelli E, Navarro Torné A,Witte W, Friedrich AW: Methicillin-resistant Staphylococcus aureus (MRSA):burden of disease and control challenges in Europe. Euro Surveill 2010,15:19688.
16. Kumar A: Optimizing antimicrobial therapy in sepsis and septic shock.Crit Care Clin 2009, 25:733-751, viii.
17. de With K, Meyer E, Steib-Bauert M, Schwab F, Daschner FD, Kern WV:Antibiotic use in two cohorts of German intensive care units. J HospInfect 2006, 64:231-237.
18. Rodriguez A, Mendia A, Sirvent JM, Barcenilla F, de la Torre-Prados MV,Sole-Violan J, Rello J: Combination antibiotic therapy improves survival inpatients with community-acquired pneumonia and shock. Crit Care Med2007, 35:1493-1498.
19. Rotschafer JC, Ullman MA, Sullivan CJ: Optimal use of fluoroquinolones inthe intensive care unit setting. Crit Care Clin 2011, 27:95-106.
20. Dupont H, Carbon C, Carlet J: Monotherapy with a broad-spectrum beta-lactam is as effective as its combination with an aminoglycoside intreatment of severe generalized peritonitis: a multicenter randomizedcontrolled trial: The Severe Generalized Peritonitis Study Group.Antimicrob Agents Chemother 2000, 44:2028-2033.
21. Damas P, Garweg C, Monchi M, Nys M, Canivet JL, Ledoux D, Preiser JC:Combination therapy versus monotherapy: a randomised pilot study onthe evolution of inflammatory parameters after ventilator associatedpneumonia [ISRCTN31976779]. Crit Care 2006, 10:R52.
22. Heyland DK, Dodek P, Muscedere J, Day A, Cook D: Randomized trial ofcombination versus monotherapy for the empiric treatment ofsuspected ventilator-associated pneumonia. Crit Care Med 2008,36:737-744.
23. Baddour LM, Yu VL, Klugman KP, Feldman C, Ortqvist A, Rello J, Morris AJ,Luna CM, Snydman DR, Ko WC, Chedid MB, Hui DS, Andremont A,Chiou CC: Combination antibiotic therapy lowers mortality amongseverely ill patients with pneumococcal bacteremia. Am J Respir Crit CareMed 2004, 170:440-444.
24. Lodise TP, Kwa A, Cosler L, Gupta R, Smith RP: Comparison of beta-lactamand macrolide combination therapy versus fluoroquinolonemonotherapy in hospitalized Veterans Affairs patients with community-acquired pneumonia. Antimicrob Agents Chemother 2007, 51:3977-3982.
25. Tessmer A, Welte T, Martus P, Schnoor M, Marre R, Suttorp N: Impact ofintravenous {beta}-lactam/macrolide versus {beta}-lactam monotherapy
on mortality in hospitalized patients with community-acquiredpneumonia. J Antimicrob Chemother 2009, 63:1025-1033.
26. Paul M, Benuri-Silbiger I, Soares-Weiser K, Leibovici L: Beta lactammonotherapy versus beta lactam-aminoglycoside combination therapyfor sepsis in immunocompetent patients: systematic review and meta-analysis of randomised trials. BMJ 2004, 328:668.
27. Safdar N, Handelsman J, Maki DG: Does combination antimicrobialtherapy reduce mortality in Gram-negative bacteraemia? A meta-analysis. Lancet Infect Dis 2004, 4:519-527.
28. Kumar A, Safdar N, Kethireddy S, Chateau D: A survival benefit ofcombination antibiotic therapy for serious infections associated withsepsis and septic shock is contingent only on the risk of death: a meta-analytic/meta-regression study. Crit Care Med 2010, 38:1651-1664.
29. Giamarellos-Bourboulis EJ, Pechere JC, Routsi C, Plachouras D, Kollias S,Raftogiannis M, Zervakis D, Baziaka F, Koronaios A, Antonopoulou A,Markaki V, Koutoukas P, Papadomichelakis E, Tsaganos T, Armaganidis A,Koussoulas V, Kotanidou A, Roussos C, Giamarellou H: Effect ofclarithromycin in patients with sepsis and ventilator-associatedpneumonia. Clin Infect Dis 2008, 46:1157-1164.
30. Metersky ML, Ma A, Houck PM, Bratzler DW: Antibiotics for bacteremicpneumonia: improved outcomes with macrolides but notfluoroquinolones. Chest 2007, 131:466-473.
31. Parnham MJ: Immunomodulatory effects of antimicrobials in the therapyof respiratory tract infections. Curr Opin Infect Dis 2005, 18:125-131.
32. Harbarth S, Garbino J, Pugin J, Romand JA, Lew D, Pittet D: Inappropriateinitial antimicrobial therapy and its effect on survival in a clinical trial ofimmunomodulating therapy for severe sepsis. Am J Med 2003,115:529-535.
doi:10.1186/cc11869Cite this article as: Díaz-Martín et al.: Antibiotic prescription patterns inthe empiric therapy of severe sepsis: combination of antimicrobialswith different mechanisms of action reduces mortality. Critical Care 201216:R223.
Submit your next manuscript to BioMed Centraland take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at www.biomedcentral.com/submit
Díaz-Martín et al. Critical Care 2012, 16:R223http://ccforum.com/content/16/6/R223
Page 9 of 9