jollant 2004 genes y riesgo suicida

8/14/2019 Jollant 2004 Genes y Riesgo Suicida

1/6

Serotonin Transporter Gene May Be Involved in Short-Term Risk of Subsequent Suicide AttemptsPhilippe Courtet, Marie-Christine Picot, Frank Bellivier, Stephane Torres, Fabrice Jollant,Cecile Michelon, Didier Castelnau, Bernard Astruc, Catherine Buresi, and Alain Malafosse Background: In the first year following a suicide attempt, patients are at high risk for reattempt and for completed suicide. We aim to determine the predictive value of two serotonin-related genes, the tryptophan hydroxylase (TPH) and serotonin transporter (5-HTTLPR) genes that have been involved in the susceptibility to suicidal behavior. Methods: After a one-year follow-up study of 103 patients hospitalized after a suicide attempt, patients have been genotyped for both the A218C TPH and the functional S/L 5-HTTLPR polymorphisms. Results: Patients who reattempted suicide during the follow-up period had significantly higher frequencies of the S allele and the SS genotype. The odds ratio for the SS genotype vs. the LL genotype was 6.5 (95% CI [1.1835.84]). No difference was observed for TPH gene. Patients carrying the SS genotype were more impulsive. However, multivariate analysis suggested an independent effect of both the SS genotype and impulsivity on the risk of repeated suicide attempts. Conclusions: These results suggest that the 5-HTTLPR SS genotype is associated with further suicide attempts among patients who have previously attempted suicide. Biol Psychiatry 2004;55:4651 2004 Society of Biological Psychiatry

Key Words: Serotonin transporter, tryptophan hydroxylase, sui-cidal behavior, repetition, polymorphism, association study,impulsivity

T he prevention of suicide and the reduction of nonfatalsuicidal behavior is part of the World Health Organiza-tions Health-for-All Policy (World Health Organization1999). Thus, suicide prevention strategies should focus in priority on high-risk subjects, which means that we need to improve theirrecognition. A previous suicide attempt is the best predictor of afuture suicide or suicide attempt (Beck et al 1975; Beck and Steer1989; Leon et al 1990; Hawton et al 1998), and subsequentattempts are generally more severe (Malone et al 1995). Indeed,in a recent systematic literature review, Owens et al (2002)

estimated that the median rates of nonfatal repetition were about16% at 1 year and 23% in studies lasting longer than 4 years. Therates of repeated suicide attempts were even higher in severalstudies, reaching 40% during follow-up periods of 3 to 8 years(Rygnestad 1988; Brauns and Berzewski 1988; Johnsson-Fridellet al 1996). For subsequent suicide, the median rates were about2% at 1-year follow-up, increasing to 7% in the studies lastingover 9 years. Patients who attempted suicide are well-knownconsumers of psychiatric resources and are at a high risk of actually committing suicide (Rissmiller et al 1994).

The lack of information concerning which preventive strate-gies are effective is probably related at least partly to the difficulty involved in identifying the subjects who should benefit fromthem. In a large-scale survey in the United Kingdom, where 24%

of all suicide cases had been in contact with the mental healthservices within the 12 months before their death, most people who committed suicide had been estimated by the clinicians tobe at no or low risk at the time of the final service contact(Appleby et al 1999). Despite the substantial risks of eventualsuicide in subjects who have already attempted to commitsuicide, clinicians lack robust predictors that can be used toquantify this risk. Although further efforts may identify clinicalprofiles of more value as predictors, a search for relevantbiological measures is clearly warranted. A large amount of datafrom studies in biological psychiatry led Mann (1998) to suggestthat the serotonin (5-HT) dysfunction is related to the diathesisfor suicidal behavior. Since the seminal studies by Asberg et al(1976), follow-up studies have suggested that low cerebrospinalfluid (CSF) 5-hydroxyindole acetic acid (5-HIAA) levels are apredictor of future suicide attempts and suicide completion andare associated with a substantial increase in short-term suiciderisk in suicide attempters (Traskman et al 1981; Roy et al 1989;Nordstrom et al 1994). Roy (1999) reported that the low affinity of the platelets serotonin uptake protein for serotonin has apredictive value for subsequent suicidal behavior during a 5-yearfollow-up period.

Impulsive aggression is closely related to lifetime suicidalbehavior and is also associated with a serotonergic dysfunction(Mann et al 1999; Placidi et al 2001). People who make repetitivesuicide attempts often have dramatic or impulsive personality traits (Paris 1996), and impulsivity is the component of borderlinepersonality disorder that best correlates with repetitive suicideattempts (Brodsky et al 1997).

The level of serotonergic system activity is probably a genet-ically determined trait, and one mechanism by which geneticfactors may affect the risk of suicide (Brent et al 1996) is by reducing serotonergic function (Mann 1998). Personality traitsrelated to impulsive aggression are probably caused by geneticfactors (Coccaro et al 1994, 1997) and involved in the transmis-sion of suicidal behavior in families (Brent et al 1996). Thus,impulsivity may be an intermediate phenotype associated withthe serotonergic genes involved in vulnerability to suicidalbehavior and to repetitive suicide attempts.

We showed that two major serotonin-related candidate genes,one encoding tryptophan hydroxylase (TPH), the rate-limitingenzyme of serotonin synthesis, and the other encoding the

From theDepartment of Psychological Medicine andPsychiatry (PC, FJ, DC,BA),Lapeyronie Hospital, InstitutNational de la Sante et de la RechercheMedicale E 0361(PC, ST, FJ, AM),La Colombiere Hospital, Department of Biostatistics (M-CP, CM), Arnaud de Villeneuve Hospital, and UniversityDepartment of Psychiatry (ST), La Colombiere Hospital, Montpellier,France; University Department of Psychiatry(FB), Hopital Henri Mondor,Assistance Publique-Hopitaux de Paris, and INSERM U513 (FB), Facultede Medecine de Creteil, Creteil France; and Department of Psychiatry(CB, AM), Geneva University Hospital, Geneva, Switzerland.

Address reprint requests to Dr. PhilippeCourtet, Department of Psycholog-ical Medicine and Psychiatry, Lapeyronie Hospital, Centre HospitalierUniversitaire Montpellier, 34295 Cedex 5, France.

ReceivedApril 30,2003; revised July 21,2003; accepted July 29,2003.

BIOL PSYCHIATRY 2004;55:46510006-3223/04/$30.00doi:10.1016/j.biopsych.2003.07.004 2004 Society of Biological Psychiatry


2/6

functional polymorphism in the serotonin transporter gene pro-moter region (5-HTTLPR), are associated with suicidal behavior(Abbar et al 2001; Courtet et al 2001). In both studies, wheresuicide attempters have been compared with subjects withoutpersonal or family history of suicidal behavior, we reported adominant effect of both the TPH A218 and 5-HTTLPR S alleles.For the TPH gene, using bearers of the C allele as the referencegroup, the odds ratios (ORs) for the A allele and the AA genotype

were 1.54 [95% confidence interval (CI) (1.19 1.98)] and 2.49[95% CI (1.43 4.33)], respectively (Abbar et al 2001). For the5-HTTLPR, using bearers of the L allele as the reference group,the ORs for the S allele and the SS genotype were 1.72 [95% CI(1.09 2.71)] and 3.63 [95% CI (1.27 10.40)], respectively (Courtetet al 2001).

Recently, it has been shown that TPH is encoded by twodifferent genes, TPH2 being 150 times more expressed in brain-stem than TPH1 (Walther et al 2003); however, the scientificrationale for testing the implication of TPH1 in reattempt isprevious studies demonstrating the TPH1 gene polymorphism isa genetic vulnerability factor for suicidal behavior (Bellivier et al,in press).

No genetic follow-up studies of cohorts of high-risk subjects

(i.e., individuals who have previously attempted suicide) havebeen carried out. To determine the predictive value of thesepolymorphisms, we carried out a 1-year follow-up study of patients who had previously attempted suicide. The new sampleof suicide attempters was independent from those included inour prior studies (Abbar et al 2001; Courtet et al 2001). We alsolooked at whether impulsivity in suicide attempters was associ-ated with the investigated genes in relation to the risk of futuresuicide attempts.

Methods and Materials

SampleThe follow-up study included 103 consecutive patients who

were hospitalized in our unit after a suicide attempt. Writteninformed consent was obtained from all subjects. DSM-IV axis Ipsychiatric disorders were diagnosed by experienced psychia-trists (PC, ST) following psychiatric evaluations and by using theMini International Neuropsychiatric Interview version 5.0.0(Sheehan et al 1998). A suicide attempt was defined as inten-tional self-harm that was not self-mutilatory in nature andrequired medical evaluation and treatment in an emergency orintensive unit (Mann 1998). Suicide attempts were classified as violent according to the criteria proposed by Asberg et al (1976).Serious attempts were violent or nonviolent attempts that justi-fied admission to an intensive unit. Lifetime impulsivity history was rated using the French version of the Barratt Impulsivity Scale (BIS) (Bayle et al 2000).

After the index admission, all patients were discharged backto the care of their referring psychiatrist. After 1 year, all patientsand current psychiatrists were interviewed to determine if therehad been any suicidal behavior since the index admission. At 1 year, 20 patients had made a subsequent suicide attempt.

To minimize population heterogeneity, patients were all of West European Caucasian origin. The study was approved by theinstitutional ethics committee.

Laboratory MethodsBlood samples (15 mL) were collected into ethylenedia-

minetetraacetic acid (EDTA)-treated tubes, and genomic DNA was extracted from white blood cells (DNA extraction kit from

Amersham-Pharmacia Biotech, Du bendorf, Switzerland). Poly-morphisms in the promoter region of the serotonin transporter(5-HTT) gene and intron 7 of the TPH gene were genotyped aspreviously described (Abbar et al 2001; Courtet et al 2001).

Statistical AnalysisMeans and standard deviations were used to describe quan-

titative variables, and proportions were used to describe categor-

ical variables. The distributions of continuous variables (age,scores) were tested with the Shapiro-Wilk test and were notnormal. Therefore, the Mann Whitney rank sum test andKruskall-Wallis test were used to compare the means of morethan two groups. For categorical variables, the percentagesaccording to relapse or genotype were compared with the 2

analysis (2-tailed) or Fisher exact test if 2 was not valid. Theglobal score of the BIS was transformed in dichotomous vari-ables, using a receiver operating characteristic (ROC) curve todetermine the cutoff point that gave the best number of correctly predicted event responses (reattempts). We obtained a thresholdof 66. Odds ratios with 95% confidence intervals were calculatedby Woolf s method. The Armitage linearity tendency test wasused to identify dose effects of susceptibility alleles (Armitage

and Berry 1987). Logistic regression models were used to study the multivariate relationships of potential predictors with reat-tempt status and their relative importance. Predictors in thesemodels were selected a priori on the basis of previous researchand on the results of the univariate analysis. Variables wereselected in a stepwise manner. Age at inclusion in years wasincluded in the model as a continuous variable. The otherdichotomous variables entered in the model were gender anddepressive disorder. The categories with the most favorableoutcome were taken as the reference level. All of the othercategories were included in the model as dummy variables. Oddsratios and their confidence intervals were calculated. The to-enter and to-exit were set, respectively, at .20 and .10. Toassess the predictive ability of the model, we calculated the

concordance rate between the predicted and observed re-sponses. The goodness-of-fit of the logistic regression model wasassessed using the likelihood ratio test and the Hosmer andLemeshow test (Hosmer and Lemeshow 1989).

The SAS version 6.12/UNIX (PROC FREQ, PROC UNIVARI- ATE, PROC NPAR1WAY, PROC LOGISTIC) software was usedfor statistical analysis (SAS Institute 1989).

Results

The clinical data concerning the 103 suicide attempters attheir index admission are shown in Table 1. Twenty-seven (26%)of the 103 patients admitted to our unit after a suicide attemptdropped out of the study during the 1-year follow-up period,mostly after the third month and without any further attemptsduring this period. These patients did not differ from the 76remaining patients for the suicidal history (number and severity of attempts) and for DSM-IV Axis I diagnosis ( Table 1) .

Seventy-six patients completed the follow-up study: 20 pa-tients reattempted suicide and 56 did not reattempt suicideduring the follow-up period. The gender distribution, the psy-chiatric diagnoses, and the suicidal behavior history were similarin the two groups on index admission: there were no differencesfor the rates of repeaters, violent attempters, and serious attempt-ers; for the age at onset; or for the family history of suicidalbehavior ( Table 1) .

Concerning the 5 HTTLPR polymorphism, the frequency of the S allele and the short/short (SS) genotype were significantly

P. Courtet et al BIOL PSYCHIATRY 2004;55:46 51 47

www.elsevier.com/locate/biopsych


3/6

higher in the group of patients who reattempted suicide duringthe follow-up period than in individuals who did not (72.5% vs.48%, 2 7.01, df 1, p .008; 55% vs. 20%, 2 9.08, df 2, p .01, respectively) ( Table 2) .

When using subjects bearing the L allele and the long/long(LL) genotype as the reference group, the odds ratios for the Sallele and the SS genotype were 2.83 [95% CI (1.29 6.22)] and 6.5[95% CI (1.18 35.84)], respectively.

This suggests that patients who carry the S allele or the SSgenotype are at high risk of short-term subsequent suicideattempts; however, most subjects from both groups had a history of previous suicide attempts at index admission. We investigated whether the S allele or the SS genotype were also associated witha higher risk of repeated suicide attempts before and/or duringthe follow-up. For this, we compared the 17 subjects who hadattempted suicide only once with the 59 patients who hadattempted suicide at least twice. The SS genotype frequencies were higher in patients who made several suicide attempts (31%)than in patients who had only made one suicide attempt (24%)and than in controls with no history of suicide attempts wepublished previously (16%) (Courtet et al 2001). The frequency of the SS genotype increased with the repetition of suicide

attempts (Armitage linearity tendency test: 5.59, df 1, p .01).No difference was found for intron 7 of the TPH gene ( Table3).

We then investigated whether impulsivity was associated withthe risk of subsequent suicide attempts and the 5-HTTLPR SSgenotype. Impulsivity tended to be slightly higher in reattempt-ers than in patients who did not reattempt: 65.35 12.59 versus60.43 10.50 (analysis of variance [ANOVA], F 2.74, p .10). When compared with the data reported in the validation study of the French version of the Barratt Impulsivity Scale (Bayle et al2000), patients in the present study showed similar impulsivity scores to those of multiple suicide attempters (63.3 15.6) andhigher impulsivity scores than subjects from the general popula-tion (55.8 17.5). As we hypothesized that a high level of impulsivity might be associated with a higher risk of reattempt insuicide attempters, we sought a threshold for the impulsivity score associated with better sensitivity and specificity. The cutoff was obtained for a score of 66 on the Barratt Impulsivity Scale.The risk of reattempt was higher in patients with an impulsivity score above 66 than in those with lower scores (40% vs. 20%respectively, p .058).

Impulsivity score was significantly higher in patients carryingthe SS genotype than in patients carrying the short/long (SL) andLL genotypes (65.58, 63.53, and 57.24, respectively; ANOVA, F 3.79, p .026). Classical suicidal behavior risk factors (gender,

age, history of major depressive disorder, history of substance oralcohol abuse or dependence, and history of multiple attempts atinclusion), 5 HTTLPR genotype, and impulsivity scores were

Table 1. Characteristics and Lifetime Psychiatric Diagnoses of Suicide Attempters at Index Admission

CharacteristicRepeaters(n 20)

Nonrepeaters(n 56)

Lost duringFollow-Up(n 27)

Women, No (%) 17 (85%) 47 (84%) 21 (78%)Age, y (mean SD) 34.05 11.71 37.89 14.42 37.63 14.68History of Suicide Attempts (SA) at Index Admission

At least 1 violent SA, no (%) 5 (25%) 12 (21%) 4 (15%)At least 1 serious SA, no (%) 6 (30%) 20 (36%) 6 (22%)Age at onset 20 years 7 (35%) 20 (36%) 7 (26%)At least 2 previous SA, no (%) 12 (60%) 39 (70%) 14 (52%)Number of SA (mean SD) 5.00 7.64 3.61 6.66 2.33 2.08

Family History (rst-degree relatives) of SuicidalBehavior, No (%)

8 (40%) 23 (41%) 10 (37%)

Lifetime Psychiatric Diagnoses at Index AdmissionMDD 16 (80%) 33 (59%) 21 (78%)Bipolar disorders 4 (20%) 8 (14%) 3 (11%)Eating disorders 2 (10%) 6 (11%) 0 (0%)Alcohol and substance abuse/dependence 5 (25%) 11 (20%) 2 (7%)GAD 5 (25%) 8 (14%) 3 (11)Agoraphobia 2 (10%) 2 (4%) 1 (4%)Social phobia 3 (15%) 3 (5%) 1 (4%)Panic disorder 4 (20%) 3 (5%) 1 (4%)

Suicide attempts were classied as violent according to the criteria proposed by Asberg et al (1976). Seriousattempts were violent or nonviolent attempts that justied admission to an intensive unit.

SA, suicide attempts; MDD, major depressive disorder; GAD, general anxiety disorder.

Table 2. Population Association Between 5-HTTLPR and Repeated Suicide Attempt at 1 Year

Genotypes Alleles

SS (%) SL (%) LL (%) S (%) L (%)

Reattempt during Follow-Up(n 20)

11 (55) 7 (35) 2 (10) 29 (72.5) 11 (27.5)

No Reattempt during Follow-Up(n 56)

11 (20) 32 (57) 13 (23) 54 (48) 58 (52)

5-HTTLPR, serotonin transporter gene; SS, short/short; SL, short/long; LL, long/long; S, short; L, long.

48 BIOL PSYCHIATRY 2004;55:46 51 P. Courtet et al



4/6

included in a logistic regression. The logistic regression results(Table 4) confirmed that the most important prognostic factorsfor relapse were the SS genotype [adjusted OR 5.4; 95% CI(1.6 20.8)], a high impulsivity score [adjusted OR 3.9; 95% CI(1.15 14.4)], and a depressive disorder [adjusted OR 5.3; 95%CI (1.4 26.7)]. Age of inclusion also had a prognostic value afteradjustment for these three variables. The adjusted odds ratio forevery 10-year decrease in age was 1.6 [95% CI (1.0 2.9)]. Theconcordance rate between the predicted and observed responses was good (79.5%). The Hosmer and Lemeshow goodness-of-fittest was not significant ( p .80).

Discussion

This is the first follow-up cohort study of at-risk subjects toinvestigate whether two serotonergic genes involved in vulner-ability to suicidal behavior are associated with an increased riskof subsequent suicidal behavior. Our results suggest that the Sallele of 5-HTTLPR is involved in the propensity to repeatedsuicidal behavior. At the functional level, the 5-HTTLPR S allele isassociated with reduced serotonergic activity (Lesch et al 1996;Reist et al 2001) and with a lower level of CSF 5-HIAA inmonkeys with deleterious rearing experiences (Bennett et al2002). Thus, our results are concordant with the findings of early biochemical studies, suggesting a correlation between low levelsof 5-HT transmission and the risk of future suicide attempts (Roy et al 1989; Roy 1999).

The lack of association between the short-term risk of reat-tempt and the TPH gene suggest that the serotonergic genesencoding for TPH and the serotonin transporter have differenteffects on susceptibility to suicidal behavior. It may also reflect alack of power in our study.

Our results may help to refine the phenotype associated withgenetic susceptibility to suicidal behavior, as well as the pheno-type associated with the 5-HTTLPR. Previous studies suggestedthat the 5-HTTLPR was associated with violent suicidal behavior(Bondy et al 2000; Courtet et al 2001) but not with nonviolentsuicidal behavior (Courtet et al 2003). In a new sample of patientsincluded in the present study, despite the small size of each

subgroup, we observed that the frequencies of the S allele and SSgenotype in violent suicidal attempters and in nonviolent at-tempters were similar to the frequencies reported previously inthese groups (data not shown) (Courtet et al 2001, 2003). Theheterogeneity of suicide attempters according to the characteris-tics of the suicidal acts (violence, repetition) may explain discor-dances in the results of association studies. Indeed, whenconsidering our whole sample, we observed a trend for anassociation between suicide attempt and the SS genotype ( p .07) and a marginal association with the S allele ( p .05). Thistrend may be explained by the high frequency of multiple suicide

attempters (73% of whom 22% have a history of violent suicideattempts) in the population studied.The suicidal phenotype may include different phenomena

and types, each with distinct biology bases. Thus, refining thephenotype may be useful to reduce the amount of etiologic andgenetic heterogeneity and thus help us to identify susceptibility genes. Taken together with the previously published associationstudies, our data suggest that the low-activity allele of the5-HTTLPR is associated with a psychobiological trait involvedboth in the risk of violent suicidal behavior and in the propensity to repeated suicidal behavior.

Our data support the hypothesis that such a trait is related toimpulsivity. We observed an association between the 5-HTTLPR and impulsivity; however, multivariate analysis showed that both5-HTTLPR and impulsivity significantly influenced the risk of reattempt. We observed that six of the eight patients with boththe SS genotype and a high impulsivity score (above the definedthreshold of 66) reattempted suicide ( p .001). About 50% of patients carrying the SS genotype, 40% of patients with animpulsivity score above 66, and 75% of patients with bothreattempted suicide. This is consistent with the notion that othergenetic and/or developmental factors interact with genes in theconstitution of psychobiological traits and that environmentaltriggers interplay to increase the risk of subsequent suicideattempts. Indeed, impulsivity seems to be associated preferen-tially with low lethality suicidal behavior (Oquendo et al 2003),as was the case in most of our patients. Nevertheless, we cannotexclude the possibility that some impulsive patients may attemptsuicide using a violent means. This would explain the role of theavailability of the suicidal means as a well-known risk factor.

A large number of association studies have shown that the5-HTTLPR is involved in susceptibility to affective disorders, inthe pharmacological response to antidepressants, in the risk of antidepressant-induced mania in bipolar patients, and in suscep-tibility to rapid cycling, suggesting that the associated trait isrelated to mood instability (Bellivier et al 2002). Together withthese data, our results suggest that the serotonin transporter geneis associated with an instability pattern involved in a propensity to recurrence of mood states, as well as suicidal acts. One may hypothesize that the 5-HTTLPR S allele is associated with acommon phenotype characterized by an affective and behavioral

Table 3. Population Association Between Intron 7 of TPH and Repeated Suicide Attempt at 1 Year

Genotypes Alleles

AA (%) AC (%) CC (%) A (%) C (%)

Reattempt during Follow-Up(n 20)

3 (15) 9 (45) 8 (40) 15 (38) 25 (62)

No Reattempt during Follow-Up(n 56)

12 (21) 26 (46) 18 (32) 50 (45) 62 (55)

TPH, tryptophan hydroxylase.

Table 4. Multivariate Analysis (Logistic Regression) of Prognostic Factorsof Subsequent Suicide Attempt

VariablesAdjusted

ORs 95% CI p (Wald 2 )

Genotype (SS vs. SL or LL) 5.4 1.6 20.8 .009BIS Score ( 66 vs. 66) 3.9 1.114.4 .03Age of Inclusion (OR for

10-Year decrease)1.6 1.0 2.9 .05

Depressive Disorder (withvs. without)

5.3 1.4 26.7 .02

OR, odds ratio; CI, condenceinterval; SS, short/short; SL, short/long; LL,long/long; BIS, Barratt Impulsivity Scale.




5/6

instability. Other studies in nonclinical populations indicated thatthis polymorphism may be associated with anxiety-related traitsand with a greater neuronal activity of the amygdala in responseto fearful stimuli (Lesch et al 1996; Hariri et al 2002). These resultssuggest that genes coding for key proteins of the serotoninergicsystem are associated with traits involved in several psychiatricdisorders and/or acting at different stages of the suicidal process,from the negative thoughts to the suicidal act itself (Marusic and

Farmer 2001). It is possible that the variants of the serotonintransporter gene have differential effects, depending on theirexpression levels in specific brain areas (prefrontal cortex,amygdala) and in combination with other genetic and nonge-netic factors, leading to different patterns of emotional or behav-ioral instability.

The serotonin transporter is the initial target for specificserotonin reuptake inhibitors (SSRI). In a recent controlled trial, Verkes et al (1998) reported that paroxetine, an SSRI, effectively reduced the recurrence of suicide attempts in subjects withpersonality disorders, independently of its antidepressant effect;however, the treatment only reduced the recurrence of suicideattempts in patients who had made less than five suicide attemptsat inclusion. Specific pharmacogenomic studies would be usefulto investigate whether patients carrying the S allele are lessresponsive to SSRI treatments, leading to a lack of preventiveeffect for subsequent suicide attempts.

The main limitation in our prospective study is the number of patients lost during the 1-year follow-up period, which wasabout one quarter (27/103). The failure for outpatients to comply with aftercare is a well-documented problem among attemptedsuicide patients hospitalized after a suicide attempt (Isacsson andRich 2001). In clinical trials of psychosocial interventions after asuicide attempt, as few as 40% to 50% of patients could becontacted directly after 1 year (Hawton et al 1981; Van Heeringenet al 1995; van der Sande et al 1997). Although we cannot be surethat the patients that dropped out made further suicide attemptsor committed suicide, some data suggest that this is not the case.First, the patients that dropped out tended to present, while notsignificantly different, fewer psychiatric conditions and a lesssevere suicidal history than the patients who completed thefollow-up, suggesting that they would present a lower risk of reattempt. Furthermore, Rossi et al (2002) suggested that patients who drop out of care have fewer psychiatric conditions and thattheir global functioning is better. Second, as suicide attemptersfrom our geographic area are mainly admitted to our university hospital, it is unlikely that most of these patients reattemptedsuicide without us finding out. Lastly, it is unlikely that asignificant number of the 27 patients committed suicide based onthe estimates of the number of completed suicides at 1 year, which is about 2% of suicide attempters (Owens et al 2002).Interestingly, we did not find any difference for the serotonintransporter genotypes and allele frequencies between the pa-tients that dropped out and the patients who completed the study (genotypes: 2 2.52, p .28; alleles: 2 2.30, p .13).Moreover, the patients that dropped out had significantly lowerS allele and SS genotype frequencies than patients who reat-tempted suicide (genotypes: 2 8.93, p .01; alleles: 2

8.30, p .003).In summary, we carried out a 1-year follow-up study of

suicidal behavior among suicide attempters. Our results suggestthat the serotonin transporter gene, but not the TPH gene, isassociated with a propensity toward repeated suicidal behavior.Our results also suggest that among patients who have previ-ously attempted suicide, this genotype may be a predictive

marker of an increased risk of further suicidal behavior. Theserotonin transporter gene may be associated with impulsivity.Before drawing definitive conclusions, replication is required inlarger samples, especially with a higher number of males. If confirmed, the SS genotype may be predictive of an increasedrisk of further suicidal behavior in suicide attempters, andimpulsivity may be a psychobiological trait associated with theserotonin transporter gene. The identification of polymorphisms

in serotonin-related candidate genes may provide a more pow-erful tool for the prediction and prevention of suicidal behaviorin patients who have already attempted to commit suicide.

This study was supported by Grant No. 31-53849.98 and Grant No. 32-66793.01 from the Swiss National Foundation to Dr. A. Malafosse, and grants from the Clinical Research Unit of the Centre Hospitalier Universitaire of Montpellier (Programme Hospitalier de Recherche Clinique 7653) and from the Fondation pour la Recherche Me dicale to Dr. P. Courtet.

Abbar M, Courtet P, Bellivier F, Leboyer M, Boulenger JP, Castelhau D, et al(2001): Suicide attempts and the tryptophan hydroxylase gene. Mol Psychiatry 6:268273.

Appleby L, Shaw J, Amos T, McDonnell R, Harris C, McCann K, et al (1999):Suicide within 12 months of contact with mental health services. Na-tional clinical survey. BMJ 318:12351239.

Armitage P, Berry G (1987): Statistical Methods in Medical Research . Oxford,England: Blackwell Scientic Publications, 371 407.

Asberg M, Traskman L, ThorenP (1976): 5-HIAA in thecerebrospinaluid. Abiochemical suicide predictor? Arch Gen Psychiatry 33:11931197.

Bayle FJ, Bourdel MC, Caci H, Gorwood P, Chignon JM, Ades J, et al (2000):Factor analysis of French translation of the Barratt impulsivity scale(BIS-10).Can J Psychiatry 45:156165.

Beck AT, Beck R, Kovacs M (1975): Classication of suicidal behaviors: I.Quantifying intent and medical lethality. Am J Psychiatry 132:285287.

Beck AT, Steer RA (1989): Clinical predictors of eventual suicide: A 5- to10-year prospective study of suicide attempters. J Affect Disord 17:203209.

Bellivier F, Chaste P, Malafosse A (in press): Association between the TPH

gene A218 polymorphism and suicidal behavior: A meta-analysis. Am J Med Genet . Published online March 14, 2003.Bellivier F, Roy I, Leboyer M (2002): Serotonin transporter gene polymor-

phisms and affective disorder-related phenotypes. Curr Opin Psychiatry 15:4958.

Bennett AJ,Lesch KP,Heils A, Long JC,Lorenz JG,ShoafSE, et al (2002): Earlyexperience and serotonin transporter gene variation interact to inu-ence primate CNS function. Mol Psychiatry 7:118122.

Bondy B, Erfurth A, Jonge S, Kruger M, Meyer H (2000): Possible associationof theshort alleleof the serotonintransporter promoter gene polymor-phism (5-HTTLPR) with violent suicide. Mol Psychiatry 5:193195.

BraunsML, BerzewskiH (1988): Follow-upstudyof patients with attemptedsuicide. IntJ SocPsychiatry 34:285291.

Brent DA, Bridge J, Johnson BA, Connolly J (1996): Suicidal behavior runs infamilies. A controlled family study of adolescent suicide victims. ArchGen Psychiatry 53:11451152.

Brodsky BS, Malone KM, Ellis SP, Dulit RA, Mann JJ (1997): Characteristics of borderline personality disorder associated with suicidal behavior. Am J Psychiatry 154:17151719.

Coccaro EF, BergemanCS, KavoussiRJ,Seroczynski AD (1997):Heritability of aggression and irritability: A twin study of the Buss-Durkee aggressionscalesin adult male subjects. Biol Psychiatry 41:273284.

Coccaro EF, Silverman JM, Klar HM, Horvath TB, Siever LJ (1994): Familialcorrelates of reduced central serotonergic system function in patientswithpersonality disorders. Arch Gen Psychiatry 51:318324.

Courtet P, Baud P, Abbar M, Boulenger JP, Castelnau D, Mouthon D, et al(2001): Association between violent suicidalbehavior and the low activ-ityallele of theserotonin transporter gene. Mol Psychiatry 6:338341.

Courtet P, BuresiC, Abbar M, Baud P, BoulengerJP, CastelnauD, et al (2003):No association between non-violentsuicidalbehavior andthe serotonintransporter promoter polymorphism. Am J Med Genet 116(suppl 1):7276.

50 BIOL PSYCHIATRY 2004;55:46 51 P. Courtet et al



6/6

Hariri AR,MattayVS,TessitoreA, KolachanaB,Fera F,GoldmanD, etal (2002):Serotonin transporter genetic variation and the response of the humanamygdala. Science 297:400403.

HawtonK, Arensman E, TownsendE, Bremner S, Feldman E, Goldney R, etal (1998): Deliberate self harm: Systematic review of efcacy of psy-chosocial and pharmacological treatments in preventing repetition.BMJ 317:441447.

Hawton K, Bancroft J, Catalan J, Kingston B, Stedeford A, Welch N (1981):Domiciliary and out-patient treatment of self-poisoning patients bymedical and non-medical staff. Psychol Med 11:169177.

Hosmer DW,Lemeshow S (1989): Applied LogisticRegression. NewYork: JohnWiley and Sons, Inc, 431.

Isacsson G, Rich CL (2001): Management of patients who deliberately harmthemselves. BMJ 322:213215.

Johnsson-Fridell E, Ojehagen A, Traskman-Bendz L (1996): A 5-year fol-low-up study of suicide attempts. Acta Psychiatr Scand 93:151157.

Leon AC, Friedman RA, Sweeney JA, Brown RP, Mann JJ (1990): Statisticalissuesin theidenticationof risk factors for suicidal behavior:The appli-cation of survival analysis. Psychiatry Res 31:99108.

Lesch KP, Bengel D, Heils A, Sabol SZ, Greenberg BD, Petri S, et al (1996):Association of anxiety-related traits with a polymorphism in the seroto-nin transporter gene regulatory region. Science 274:15271531.

MaloneKM, Haas GL,SweeneyJA, Mann JJ (1995): Major depression andtherisk of attempted suicide. J Affect Disord 34:173185.

Mann JJ (1998): The neurobiology of suicide. Nat Med 4:2530.Mann JJ, Waternaux C, Haas GL, Malone KM (1999): Toward a clinical model

of suicidal behaviorin psychiatric patients. Am J Psychiatry 156:181189.Marusic A, Farmer A (2001): Toward a new classication of risk factors for

suicide behavior. Crisis22:4346.Nordstrom PM,Samuelsson M, Asberg M, Traskman-BendzL, Aberg-Wistedt

A, NordinC, et al(1994): CSF5-HIAA predictssuicide risk after attemptedsuicide. Suicide Life Threat Behav 24:19.

Oquendo MA, Placidi GP, Malone KM, Campbell C, Keilp J, Brodsky B, et al(2003): Positron emission tomography of regional brain metabolic re-sponses to a serotonergic challenge and lethality of suicide attempts inmajor depression. Arch Gen Psychiatry 60:1422.

Owens D, Horrocks J, House A (2002): Fatal and non-fatal repetition of self-harm. Systematic review. Br J Psychiatry 181:193199.

Paris J (1996): Social Factors in the Personality Disorders. A Biopsychosocial Approach to Etiology and Treatment . Cambridge, England: CambridgeUniversity Press.

Placidi GP, Oquendo MA, Malone KM, Huang YY, Ellis SP, Mann JJ (2001):

Aggressivity, suicide attempts,and depression: Relationshipto cerebro-spinal uid monoamine metabolite levels. Biol Psychiatry 50:783791.

Reist C, Mazzanti C, Vu R, Tran D, Goldman D (2001): Serotonin transporterpromoter polymorphism is associated with attenuated prolactin re-sponse to fenuramine. Am J Med Genet 105:363368.

Rissmiller DJ, Steer R, Ranieri WF, Rissmiller F, Hogate P (1994): Factorscomplicating cost containment in the treatment of suicidal patients.HospCommunity Psychiatry 45:782788.

Rossi A, Amaddeo F, Bisof G, Ruggeri M, Thornicroft G, Tansella M (2002):Dropping out of care: Inappropriate terminations of contact with com-munity-based psychiatric services. Br J Psychiatry 181:331338.

Roy A (1999): Suicidal behavior in depression: Relationship to platelet sero-tonin transporter. Neuropsychobiology 39:7175.RoyA, De Jong J, Linnoila M (1989): Cerebrospinal uid monoaminemetab-

olites and suicidal behavior in depressed patients. A 5-year follow-upstudy. Arch Gen Psychiatry 46:609612.

Rygnestad T (1988): A prospective 5-year follow-up study of self-poisonedpatients. Acta Psychiatr Scand 77:328331.

SAS Institute, Inc (1989): SAS/STAT Users Guide,Version 6, 4thed, vol. 2. Cary,NC: SAS Institute, Inc.

Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E, et al(1998): The Mini-International Neuropsychiatric Interview (M.I.N.I.). Thedevelopment and validation of a structureddiagnostic psychiatric inter-view for DSM-IV andICD-10. J Clin Psychiatry 59(suppl 20):2233.

Traskman L, Asberg M, Bertilsson L, Sjostrand L (1981): Monoamine metab-olitesin CSF and suicidal behavior. Arch Gen Psychiatry 38:631636.

van der Sande R, van Rooijen L, Buskens E, Allart E, Hawton K, van der Graaf

Y, et al (1997): Intensive in-patient and community intervention versusroutine care afterattempted suicide. A randomised controlled interven-tion study. Br J Psychiatry 171:3541.

Van Heeringen C, Jannes S, Buylaert W, Henderick H, De Bacquer D, VanRemoortelJ (1995): Themanagementof non-compliance withreferraltoout-patient after-care among attempted suicide patients: A controlledintervention study. Psychol Med 25:963970.

Verkes RJ, Van der Mast RC, Hengeveld MW, Tuyl JP, Zwinderman AH, VanKempen GM (1998): Reduction by paroxetine of suicidal behavior inpatientswith repeated suicide attemptsbut not majordepression. Am J Psychiatry 155:543547.

Walther DJ,Peter JU,Bashammakh S,HortnaglH, VoltsM, Fink H,et al(2003):Synthesis of serotonin by a second tryptophan hydroxylase isoform.Science 299:76.

World Health Organization (1999): Health 21: The Health for all Framework for the WHO European Region. European Health for All Series No 6.

Copenhagen, Denmark: World Health Organization, Regional Ofcefor Europe.



jollant 2004 genes y riesgo suicida

Documents