Suicide attempts in a prospective cohort of patients with schizophrenia treated with sertindole or risperidone

European Neuropsychopharmacology (2010) 20, 829–838

www.elsevier.com/locate/euroneuro

Suicide attempts in a prospective cohort of patients with schizophrenia treated with sertindole or risperidone M.A. Crocq a,⁎, D. Naber b , M.H. Lader c , F. Thibaut d , M. Drici e , B. Everitt c , G.C. Hall f , C. Le Jeunne g , A. Mittoux h , J. Peuskens i , S. Priori j , M. Sturkenboom k , S.H.L. Thomas l , P. Tanghøj h , M. Toumi m , R. Mann n , N.D. Moore o a

Centre Hospitalier, BP29, 68250 Rouffach, France Universitätsklinikum Hamburg-Eppendorf, Klinik für Psychiatrie und Psychotherapie, Martinistrasse 52, 20246 Hamburg, Germany c Institute of Psychiatry, King's College London, Denmark Hill, London SE5 8AF, United Kingdom d University Hospital Ch. Nicolle, INSERM U 614, Faculty of Medicine, Rouen, France e Pharmacologie-Toxicologie, Hôpital Pasteur, BP 69, 30 Avenue de la Voie Romaine, 06002 Nice Cedex 1, France f Grimsdyke House, Herts, United Kingdom g Groupe Hospitalier Hotel-Dieu - La Collegiale, Paris, France h H. Lundbeck A/S, Ottiliavej 9, Valby, Copenhagen, DK 2500, Denmark i University Psychiatric Center K.U. Leuven, Belgium j Molecular Cardiology, University of Pavia, Via Ferrata 8, Pavia 27100, Italy k Dep. Medical Informatics and Epidemiology, Erasmus University Medical Center, PO BOX 2040 3000CA, Rotterdam, The Netherlands l Institute of Cellular Medicine, Wolfson Unit of Clinical Pharmacology, Newcastle University, Newcastle, United Kingdom m Université Claude Bernard Lyon, Bat. Nautibus, 43 Bd. du 11 Novembre 1918, 69622 Villeurbane Cedex, France n 42 Hazleton Way, Waterlooville, Hampshire PO8 9BT, United Kingdom o Département de Pharmacologie, CHU de Bordeaux-Pellegrin, France b

Received 27 April 2010; received in revised form 14 August 2010; accepted 9 September 2010

KEYWORDS Schizophrenia; Suicide; Suicide attempt; Risperidone; Sertindole

Abstract The incidence of suicide attempts (fatal and non-fatal) was analysed in a prospective cohort of patients with schizophrenia randomly assigned to sertindole (4905 patients) or risperidone (4904 patients) in a parallel-group open-label study with blinded classification of outcomes (the sertindole cohort prospective study — SCoP). The total exposure was 6978 and 7975 patient-years in the sertindole and risperidone groups, respectively. Suicide mortality in the study was low

⁎ Corresponding author. Tel.: +33 3 89 32 01 15; fax: +33 3 89 78 72 00. E-mail address: [email protected] (M.A. Crocq). 0924-977X/$ - see front matter © 2010 Elsevier B.V. and ECNP. All rights reserved. doi:10.1016/j.euroneuro.2010.09.001

830

M.A. Crocq et al. (0.21 and 0.28 per 100 patients per year with sertindole and risperidone, respectively). The majority (84%) of suicide attempts occurred within the first year of treatment. Cox's proportional hazards model analysis of the time to the first suicide attempt, reported by treating psychiatrists and blindly reviewed by an independent expert group according to the Columbia Classification Algorithm of Suicide Assessment (both defining suicide attempts by association of suicidal act and intent to die), showed a lower risk of suicide attempt for sertindole-treated patients than for risperidone-treated patients. The effect was statistically significant with both evaluation methods during the first year of randomized treatment (hazard ratios [95% CI]: 0.5 [0.31–0.82], p = 0.006; and 0.57 [0.35–0.92], p = 0.02, respectively). With classification by an independent safety committee using a broader definition including all incidences of intentional self-harm, also those without clear suicidal intent, the results were not significant. A history of previous suicide attempts was significantly associated with attempted suicides in both treatment groups. © 2010 Elsevier B.V. and ECNP. All rights reserved.

1. Introduction 1.1. Suicide and schizophrenia Suicide is substantially more frequent in individuals with schizophrenia than in the general population. According to earlier reviews (Miles 1977; Caldwell and Gottesman, 1990), the lifetime risk of completed suicide in schizophrenia reached 10%. Recently, that estimate was reduced to 4.9% by a meta-analysis (Palmer et al., 2005), which covered 61 studies with a total of 48,176 patients, and also confirmed that suicides are concentrated early in the illness course. The risk is particularly high in the first two years of the illness and decreases with time since first hospital admission (Heilä et al., 2005). Suicide attempt rates at the time of first hospitalization are highest in patients between age 16 and 20 and decrease in older patients (Levine et al., 2010). Clinical contact may exert a protective influence, as suggested by the fact that outpatients with schizophrenia commit suicide at a higher rate than inpatients (Helgason, 1990). The first weeks (Qin and Nordentoft, 2005) or months (Rossau and Mortensen, 1997) after hospital discharge are a critical period. Other risk factors include high premorbid socioeconomic class and high IQ (Siris, 2001), better cognitive functioning (Nangle et al., 2006), greater insight into the illness (Schennach-Wolff et al., 2010), living alone, social fragmentation (Evans et al., 2004), and duration of untreated psychosis longer than one year (Altamura et al., 2003). The association with clinical features is more controversial. Suicide has been associated with severity of hallucinations and delusional thinking, command auditory hallucinations (Heila et al., 1997), thought disorder (Bakst et al., 2010), post-psychotic depression occurring after acute psychotic episodes (Meltzer and Fatemi, 1995) and severity of depressive symptoms (Bakst et al., 2010). Several studies indicate that patients with schizophrenia tend to utilize violent and lethal methods of suicide, in particular jumping from heights. According to one study, fewer patients with schizophrenia plan the suicide, compared to those without the disorder (Kreyenbuhl et al., 2002). Lack of planning would suggest that suicide in schizophrenia is impulsive and occurs when an opportunity becomes available. However, this is debated and a possible bias is that schizophrenic patients tend not to verbalize their intentions. In the general population, completed suicide rates are consistently higher in males than

in females (Bray and Gunnell, 2006). This is also the case in patients with schizophrenia (Roy, 1982, Breier and Astrachan, 1984), although the gender difference is significantly less pronounced. Contrary to completed suicides, suicide attempts are observed at the same rate in males and females with schizophrenia (Harkavy-Friedman et al., 2001). A study in three groups of 160 inpatients each, with unipolar depression, schizophrenia, or opioid dependency showed that a positive family history of suicide was associated with a higher risk for suicide attempt, with high-lethality methods, with repeated attempts, and with a number of attempts, while the interaction between family history and diagnostic group was not significant (Trémeau et al., 2005). Finally, one of the most important predictors of suicide in schizophrenia is a history of a recent suicide attempt (Pompili et al., 2007; Bakst et al., 2010; Reutfors et al., 2009).

1.2. The impact of first- and second-generation antipsychotic drugs on suicidal behavior The literature on the impact of typical or first-generation antipsychotics (FGAs) on suicidal behavior is contradictory (Mamo, 2007; Aguilar and Siris, 2007). This may be due to methodological shortcomings, such as the retrospective design of most studies, lack of treatment randomization, and confounding factors such as comorbidity and polypharmacy (Meltzer, 2001). Earlier studies reported that FGAs might actually increase suicidal ideation (Beisser and Blanchette, 1961), possibly because of antipsychotic induced akathisia and dysphoria. A study in a cohort of 3474 patients with schizophrenia followed over 4 years found that treatment with thioxanthenes was associated with a significant excess of suicide mortality, compared with phenothiazines, butyrophenones, or benzamides (Montout et al., 2002). In contrast, more recent studies have indicated that adherence to any antipsychotic treatment is likely to reduce suicide risk substantially (Haukka et al., 2008). In a register linkage study among first episode patients, the suicide risk of those not currently using any antipsychotic drug was about 37-fold compared to current users (Tiihonen et al., 2006). Further, a prospective case–control study found more self-injurious behavior following discontinuation of depot antipsychotics at 18 months when compared with patients who were maintained on the medication (Johnson et al., 1983). A negative attitude towards treatment (De Hert et al., 2001) and poor

Suicide attempts in SCoP adherence to medication (Pompili et al., 2009) are risk factors for suicide. Studies with atypical or second-generation antipsychotics (SGAs) yielded robust evidence for clozapine, which has been shown to have a substantial effect on attempted (Spivak et al., 1997, Meltzer and Okayli, 1995) and completed suicide (Tiihonen et al., 2009, Walker et al., 1997). A randomized, prospective 2-year study (InterSePT) (Meltzer et al., 2003) compared the effect of clozapine versus olanzapine in 980 patients who were considered at high risk for suicide. Suicidal behavior was significantly less frequent in patients treated with clozapine compared with olanzapine (hazard ratio [95% CI]: 0.76 [0.58–0.97], p = 0.03). However, few of these highrisk patients died of suicide (5 clozapine- and 3 olanzapinetreated patients) and the study could not demonstrate a specific preventative effect of clozapine on completed suicide. Other reports indicate an improvement in suicidality in schizophrenia patients treated with other SGAs, such as olanzapine, risperidone, and sertindole (Kerwin and Bolonna, 2004). A retrospective comparison of 378 patients with schizophrenia who attempted suicide with the same number of patients who did not (Barak et al., 2004), suggested that risperidone and olanzapine may both confer protection from suicide. Risperidone had a larger effect-size than olanzapine, although the difference was not statistically significant. Clearly, retrospective studies have limitations but there are few prospective randomized studies comparing the effect of SGAs on suicide in schizophrenia. This report analyses the incidence of attempted suicides (fatal and non-fatal) in a population of patients with schizophrenia who took part in the sertindole cohort prospective (SCoP) study, which focused on cardiac safety (Thomas et al., 2010) (Clinical Trial Registration: Safety study of sertindole versus risperidone under normal conditions of use, NCT 00856583 [EudraCT 2004-000213-19] [http://clinicaltrials. gov/ct2/results?term=00856583]). The primary results from the SCoP study have recently been published (Thomas et al., 2010); here we present a more detailed analysis of suicide attempts (fatal and non-fatal). With 9809 patients enrolled, the SCoP study is one of the largest prospective, randomized studies ever conducted in patients with schizophrenia. The size of the study allowed a relevant evaluation of the risk of suicide attempts in patients treated with sertindole or risperidone.

2. Experimental procedures 2.1. Study design The overall objective of the study was to compare the safety of sertindole with that of risperidone under normal conditions of use. The design of the SCoP study is described in more detail elsewhere (Peuskens et al., 2008). Briefly, this was a multinational, multi-centre, parallel-group, random-allocation, open-label study with blinded classification of outcomes. Six hundred and nine investigators took part in 30 European (most of Eastern and Western Europe, and Turkey) and 8 South and East Asian countries (Hong Kong, India, Malaysia, the Philippines, Singapore, South Korea, Taiwan, and Thailand). The study included patients with schizophrenia (category F20 in ICD-10) (World Health Organization, 1992), who were at least 18 years of age, for whom a change in antipsychotic treatment was indicated. The patients were randomized (1:1) to treatment with sertindole (4905 pts.) or risperidone (4904 pts.). Investigators were not blind to the allocated

831 treatment and managed their patients according to their usual clinical practice. The titration and maintenance dosages were determined by the investigator, in accordance with the national Summary of Product Characteristics (SPC). Sertindole and risperidone tablets were administered orally. Study assessments were performed monthly during the first 3 months of treatment and on a quarterly basis thereafter. The study was carried out between July 2002 and February 2008. Treatment duration for an individual patient was not limited. Thus, many patients were treated over several years (up to 63.5 months). The mean number of days of treatment was 489.6 and 564.0 in the sertindole and risperidone groups, respectively.

2.2. Assessment of suicide attempts Firstly, suicide attempts (fatal and non-fatal) were reported by the treating psychiatrists (investigators) as serious adverse events (SAE) as predefined in the study plan. Adverse events were coded according to Version 10.1 of the Medical Dictionary for Regulatory Activities (MedDRA). MedDRA is a clinically validated international adverse event classification dictionary used by regulatory authorities and the biopharmaceutical industry (Bousquet et al., 2005). MedDRA terms are organized into five hierarchical levels: system organ class, high level group terms, high level terms (HLTs), preferred terms (PTs) and low level terms. This study analysed the events coded with the two following PTs: Completed suicide, and Suicide attempt. Therefore, the six other PTs gathered under the same HLT – Suicidal and self-injurious behavior – (i.e., Intentional self-injury, Self-injurious behavior, Self-mutilation, Self-injurious ideation, Suicidal behavior, and Suicidal ideation) were not considered. A suicide attempt required the conjunction of an observable suicidal act, and intent to commit suicide. For all cases where a suicide, suicide attempt, intentional overdose, poisoning, or self-injury was suspected, the treating psychiatrist was asked to confirm whether it was a suicide or suicide attempt, if not stated clearly in the initial SAE report. Secondly, as predefined in the study protocol, an Independent Safety Committee (ISC), including members with expertise in cardiology, epidemiology, pharmacovigilance, psychiatry, and statistics, reviewed blinded case reports, and recorded completed suicide and suicidal behavior. Completed suicide was determined by the ISC on the basis of the death certificate, post-mortem examination if applicable, and other information from investigators. The ISC working procedures specified that an episode of serious selfharm or intentional overdose or poisoning was considered a suicide attempt, even if the patient expressed no overt suicidal intention, and regardless of the investigator's assessment. Thus, the ISC could consider that events such as suicidal ideation and suicidal tendency, as well as self-injuries and overdoses represented suicide attempts, even if not confirmed as such by the treating psychiatrist. Thirdly, in a post-hoc analysis requested by the FDA, suicide attempts were blindly reviewed by an independent expert group at Columbia University (New York), in accordance with the Columbia Classification Algorithm of Suicide Assessment (C-CASA) (Posner et al., 2007). The analysis included the following three C-CASA categories: (i) completed suicide (self-injurious behavior that resulted in fatality and was associated with at least some intent to die), (ii) suicide attempt (potentially self-injurious behavior, associated with at least some intent to die), and (iii) preparatory act toward imminent suicidal behavior (the individual takes steps to injure him- or herself, but is stopped by self or others).

2.3. Statistical analysis To avoid adjunctive effects of polypharmacy the “only randomized treatment period” (ORT), when patients received only sertindole or risperidone as antipsychotic medication, was a key period for the reporting and analysis of events. The ORT was defined as the period

832

M.A. Crocq et al.

from the date of prescription of randomized treatment until randomized treatment was stopped plus 1 day (to include events that occurred the day after the last dose of randomized treatment). The “whole randomized treatment period” (i.e. including any addon antipsychotic therapy) and the 30 days thereafter (WRT + 30) constituted another key period for reporting and analysis of events as this period was conservatively prespecified as the primary analysis period in the study protocol. Since the focus of treatment in reducing suicidality in schizophrenia should be on preventing suicidal attempts (completed suicide represents a fatal suicide attempt), the rates of suicide attempt (both fatal and non-fatal) were compared between treatments. This was done using Cox's proportional hazards model analysis of the estimated time to the first suicide attempt, with a variable for the treatment group, and including the baseline covariates: age, sex, duration of schizophrenia (b5 years, 5 to 10 years, N10 years prior to study entry), and time since last suicide attempt (no attempt, b1 year, 1 to 5 years, N5 years prior to study entry). In addition, study accrual time was examined for prognostic information to adjust for possible drift in patient inclusion. In this model, the suicide attempt ratio of the hazard in the sertindole group compared to that in the risperidone group was estimated and the null hypothesis of no difference (suicide attempt ratio = 1) was tested against the one-sided alternative of fewer suicide attempts in the sertindole group (suicide attempt ratio b1). However, all p-values presented here are two-sided.

3. Results 3.1. Patients The demographic characteristics of both treatment groups did not differ significantly (Table 1). The mean age at baseline was similar in both groups; ages ranged from 18 to 85 years. Slightly more patients in the sertindole group than in the risperidone group were older than 65, and 75 (99 and 84 patients, and 14 and 9, respectively). Men comprised a slight majority of both groups

Table 1

Patient characteristics.

Patients treated [Europe/Asia] Total exposure (patient-years) a Exposure per patient (days) (median [mean]) b Gender (% male) Mean age (years) [range] Age groups (%) (b 35/35–65/≥65) Last antipsychotic treatment (%) (typical/ before study inclusion atypical/both) (%) (monopharmacy/ polypharmacy) Primary reason for prescription Adverse drug reaction of investigational medicinal Lack of efficacy product (%) Non or poor compliance Patient's choice Other Duration of schizophrenia (%) (b 5 years/5–10 years/N10 years) Time since last suicide attempt (%) (no attempt/b1 year/1–5 years/N5 years) a b

(55%) and their age distribution was skewed towards a younger age than for women. This is consistent with the younger age at onset of schizophrenia in males. The proportion of men was slightly larger in Asia (60%) than in Europe (54%), particularly in the sertindole group (61% men in Asia versus 53% men in Europe). The mean age at baseline was slightly younger in Asia (36 years) than in Europe (39 years). At baseline, approximately two-thirds of the patients in each treatment group had been diagnosed with schizophrenia for 5 years or more prior to study entry. Twelve percent of patients in each treatment group had a history of previous suicide attempts. There were no significant differences between treatment groups in the number of previous suicide attempts, nor in the time since the last suicide attempt. Treatment groups did not differ either within each region, and there were no regional differences in disease duration or history of suicide attempts. The proportions of patients who had received a FGA, a SGA, or both were similar in both treatment groups. The proportion of patients who had received a SGA was smaller in Asia (26%) than in Europe (41%). The reasons for enrolment into the study were similar in both treatment groups: more than half of the patients in each treatment group were included because of the lack of efficacy of the previous antipsychotic (52% in the sertindole group and 53% in the risperidone group). At weeks 12, 24, and 48, the median doses of risperidone and sertindole were 4 mg and 12 mg, respectively, and approximately 82% of the patients in the sertindole group and 91% of the patients in the risperidone group received doses within the recommended dose range (i.e., 12–20 mg/day for sertindole, and 2–8 mg/day for risperidone), though with a preference for lower doses (12 mg/day sertindole and 2 and 4 mg/day risperidone). This may reflect the common clinical practice of using the lowest effective dose for each individual patient. Only a few patients received an additional antipsychotic as an add-on therapy (424 [8.6%] in the risperidone group, and 361 [7.3%] in the sertindole group), mostly for lack of efficacy.

WRT period + 30 days. WRT period.

Sertindole

Risperidone

Total

4905 [3545/1360] 6978 360.0 [489.6] 55.2 38.4 [18.0–84.6] 43.8/54.2/1.7 53.3/36.8/9.9

4904 [3543/1361] 7975 475.5 [564.0] 55.4 38.3 [18.0–81.6] 44.3/54.0/1.7 53.5/36.4/10.1

9809 [7088/2721] 14,953 415.0 [526.8] 55.3 38.3 44.0/54.1/1.9 53.4/36.6/10.0

77.1/22.9

76.7/23.3

76.9/23.1

21.8 52.6 3.4 19.9 2.1 30.33/26.23/43.43

22.8 51.8 3.2 20.2 1.8 30.48/26.53/42.99

22.3 51.8 3.2 20.2 1.8 30.41/26.38/43.21

87.58/2.50/4.62/5.30 87.68/2.39/4.46/5.48 87.63/2.44/4.54/5.39

Suicide attempts in SCoP

833

3.2. Suicide attempts (fatal and non-fatal) 3.2.1. Reports from the treating psychiatrists (MedDRA) Cox's proportional hazards model analysis of the estimated time to the first suicide attempt (fatal or non-fatal) showed a significantly lower risk for sertindole-treated than for risperidone-treated patients during the ORT and WRT + 30 periods (Fig. 1). During antipsychotic monotherapy (ORT), the estimated risk of a patient attempting suicide with a fatal or nonfatal outcome was almost 40% lower with sertindole (HR [95% CI]: 0.61 [0.39–0.94]) and 50% lower during the first year of treatment (HR [95% CI]: 0.50 [0.31–0.82]). The Kaplan–Meier plot of the estimated probability of not attempting suicide is shown in Fig. 2. As can be deduced from the plot, the majority of events occurred within the first year of treatment at a reducing rate (34% within the first 3 months, 52% within the first 6 months, and 84% within the first year).

3.2.2. Assessment with C-CASA classification Following blinded C-CASA classification, Cox's proportional hazards model analysis of the estimated time to the first

suicide attempt (fatal or non-fatal) yielded results very similar to those from the psychiatrists' reports (Fig. 1). However, the reduced risk of attempting suicide with sertindole as compared with risperidone, in the ORT and WRT + 30 periods, was significant only during the first year of treatment.

3.2.3. Assessment by the ISC According to the ISC classification, which also included episodes of serious self-harm even if the patient expressed no overt suicidal intention, more patients exhibited suicidal behavior than reported by the treating psychiatrist or classified according to C-CASA. Cox analysis of the time to the first suicide attempt showed lower risks for sertindoletreated patients, but this did not reach statistical significance (Table 2).

3.2.4. Predictive factors for suicide attempts Study accrual time had a significant effect on the time to attempting suicide (p = 0.006) (Table 3), reflecting a decreasing hazard for patients who were enrolled later in the study. Neither age nor gender was predictive of the hazard function.

All patients (n = 9,809) Period ORT

Classification

a

MedDRA C-CASA st

ORT 1 year

b

c

WRT + 30

e

f

st

d

Hazard ratio (95% CI)

p-value

34 (0.54)

55 (0.76)

0.61 (0.39-0.94)

0.0249

37 (0.58)

55 (0.76)

0.67 (0.44-1.02)

0.0630

MedDRA

26 (0.78)

49 (1.34)

0.50 (0.31-0.82)

0.0058

C-CASA

28 (0.84)

47 (1.28)

0.57 (0.35-0.92)

0.0206

66 (0.83)

0.66 (0.45-0.98)

0.0373 0.1094

MedDRA

WRT + 30 1 year

g Sertindole Risperidone (N/100PYE) (N/100PYE)

43 (0.62)

C-CASA

48 (0.69)

67 (0.84)

0.74 (0.51-1.07)

MedDRA

33 (0.91)

58 (1.48)

0.56 (0.36-0.86)

0.0086

C-CASA

36 (1.00)

64 (1.64)

0.62 (0.41-0.95)

0.0287

Hazard ratio (95% CI)

p-value

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Hazard ratio (95%CI)

h

High-risk patients (n = 683) Period

ORT

Classification

a

MedDRA C-CASA

ORT 1 year b st

g

h

MedDRA

c

WRT + 30

Sertindole Risperidone (N/100PYE) (N/100PYE) 13 (3.00)

21 (5.15)

0.55 (0.27-1.10)

0.0886

15 (3.49)

21 (5.15)

0.63 (0.33-1.23)

0.1779

12 (5.22)

20 (9.22)

0.53 (0.26-1.08)

0.0812

C-CASA

14 (6.11)

20 (9.22)

0.62 (0.31-1.23)

0.1678

MedDRA

16 (3.38)

27 (5.71)

0.56 (0.30-1.04)

0.0656

C-CASA

18 (3.81)

27 (5.71)

0.63 (0.35-1.15)

0.1309

WRT + 30 1 year MedDRA

14 (5.60)

24 (9.84)

0.54 (0.28-1.05)

0.0695

C-CASA

16 (6.43)

24 (9.84)

0.62 (0.33-1.17)

0.1417

st

d

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Hazard ratio (95% CI)

Figure 1 Estimated hazard ratios (sertindole versus risperidone) for suicide attempts (fatal plus non-fatal) all patients and high-risk patients reported by the treating psychiatrists coded to MedDRA or classified blindly according to C-CASA. a ORT: Only Randomized Period. b ORT 1 year: Only Randomized Period during first year of treatment c WRT + 30: Whole Randomized Period+ 30 days. d WRT + 30 1 year: Whole Randomized Period+ 30 days during first year of treatment. e MedDRA: the Medical Dictionary for Regulatory Activities (used by treating psychiatrists to code adverse events). f C-CASA: Columbia Classification Algorithm of Suicide Assessment. g Two sertindole patients with non-fatal suicide attempts are not included in the model because information on some of the prognostic variables was missing. h High-risk patients: those with a history of at least one suicide attempt during the five years prior to study entry.

834

M.A. Crocq et al.

estimated attempt free probability

1

0.995

0.99

sertindole risperidone

0.985

0.98 Day

0

180

360

540

720

900

1080

Sertindole Cumulative events

0

18

33

36

38

41

41

4905

3560

2635

1907

1406

952

719

0

37

58

64

64

64

65

4904

3925

3039

2286

1713

1163

880

At risk (N) Risperidone Cumulative events At risk (N)

Figure 2 period.

Kaplan–Meier plot of suicide attempts (fatal and non-fatal) as reported by the treating psychiatrists during the WRT + 30 day

Patients who had previously attempted suicide had a significantly higher risk of attempting suicide again than patients who had never attempted suicide. The Cox analysis showed no interaction between treatment and the history of suicide attempts. Patients with disease duration of 5 to 10 years tended to have a relatively higher risk of attempting suicide than patients with shorter or longer disease duration. Region of origin was not a significant predictor of the time to suicide attempt.

3.2.5. High-risk group characteristics High-risk patients were those with a history of at least one suicide attempt during the five years prior to study entry. This high-risk group represented 7% of the total population but accounted for nearly 50% of the suicide attempts during the study. Compared to the whole population, these patients were slightly younger (mean age 35.1 years compared to 38.3 years) than the whole group and fewer had schizophrenia for longer than 10 years (29.3% compared to 42.3%). In addition, 35.0% of the high-risk patients had attempted suicide during the year

Table 2

ORT ORT 1 year b WRT + 30 c WRT + 30 1 year d a b c d

3.3. Fatal suicide attempts (completed suicide) The treating psychiatrists reported 34 suicides (22 men and 12 women): 13 (0.20 per 100 patient-years of exposure) in the sertindole group, and 21 (0.28 per 100 PYE) in the risperidone group. The ISC reported 35 suicides (24 men and 11 women): 14 (0.21 per 100 patient-years of exposure) in the sertindole group and 21 (0.28 per 100 PYE) in the risperidone group. The discrepancy in the number of completed suicides between the treating psychiatrists' reports and the ISC classification is due to the fact that (i) 2 cases reported as overdose (1 risperidone and 1 sertindole) were classified as suicide by the ISC and (ii) 1 case reported as suicide (risperidone) but not documented as such was considered a cardiac death by the ISC.

Completed suicide and self-injurious behavior as classified by the Independent Safety Committee.

Period a

prior to study entry. Cox analysis of the time to the first suicide attempt showed lower risks for sertindole-treated patients, but this was not statistically significant (Fig. 1), probably due to the smaller number of patients in the high-risk group.

Sertindole (rate/100PYE)

Risperidone (rate/100PYE)

Hazard ratio (95% CI)

p-value

54 43 68 53

64 53 77 64

0.87 0.81 0.92 0.83

0.4454 0.3112 0.6019 0.3151

(0.85) (1.29) (0.98) (1.47)

(0.89) (1.45) (0.97) (1.64)

ORT: Only Randomized Period. ORT 1 year: Only Randomized Period during first year of treatment. WRT + 30: Whole Randomized Period + 30 days. WRT + 30 1 year: Whole Randomized Period + 30 days during first year of treatment.

(0.60–1.25) (0.54–1.22) (0.66–1.28) (0.57–1.20)

Suicide attempts in SCoP

835

Table 3 Estimated Cox's proportional hazards model of suicide attempts (fatal and non-fatal) as reported by treating psychiatrists during the WRT + 30 day period. Variable

Hazard ratio

Treatment (sertindole) Age Sex (female) Duration of schizophrenia (5 to 10 years) Last suicide attempt: b 1 year prior to entry 1 to 5 years prior to entry N 5 years prior to entry Study accrual time a

95% CI

p-value a

0.66

0.45–0.98

0.0373

0.99 1.07 1.46

0.97–1.01 0.73–1.58 0.98–2.19

0.2890 0.7140 0.0613

14.83

8.95–24.56

b 0.0001

6.91

4.14–11.51

b 0.0001

2.89

1.46–5.70

0.0022

0.77

0.64–0.93

0.0056

Wald's χ . 2

Half of the events in each treatment group occurred within the first 6 months of treatment (7 in the sertindole group, and 11 in the risperidone group) and two-thirds within the first 9 months (10 in the sertindole group, and 15 in the risperidone group). The suicide methods were mostly violent and highly lethal (hanging/strangulation: n = 14 [41%]; jumping/falling: n = 14 [41%]; drug overdose: n = 3 [9%]; gunshot: n = 1 [3%]; cutting/stabbing: n = 1 [3%]; and information missing: n = 1 [3%]). Using Cox's proportional hazards model analysis of the time to suicide, as reported by the treating psychiatrists, in the WRT + 30 day period, the estimated hazard for suicide risk was lower (though not significantly) for sertindole- than for risperidone-treated patients (HR [95% CI]: 0.66 [0.33–1.32], p = 0.2432). Three variables had significant or near-significant effect on the hazard function. Firstly, patients who had attempted suicide within the 5 years prior to study entry had a significant and much higher risk of committing suicide than patients who had never attempted suicide (HR [95% CI]: 7.98 [3.93–16.20], p b 0.0001). Secondly, patients who had received the combination of several antipsychotics (polypharmacy) prior to study entry had a significantly higher risk of committing suicide than the others (HR [95% CI]: 2.09 [1.04–4.20], p = 0.0376). Thirdly, patients who were included into the study later had a near-significantly lower risk of committing suicide than those who were enrolled earlier (HR [95% CIs]: 0.76 [0.55–1.05], p = 0.0940) for 1 year increase in study accrual time), reflecting a decreasing hazard for patients who were enrolled later in the study (estimated approx. 20% decrease per year of study accrual), irrespective of treatment group. Age or gender had no significant effect on the hazard function, although the estimated hazard function was lower (though not significant) for women than for men (HR [95% CI]: 0.70 [0.34–1.42], p = 0.3061). Region, duration of schizophrenia, and the type of last antipsychotic medication (FGA, SGA, both) did not provide any prognostic information about the time to committing suicide. Age interacted significantly with treatment (p = 0.0331): for the

younger half of the patients who committed suicide, almost 40% of the suicides (6 of 16) occurred in the risperidone group, whereas for the older half of the patients who committed suicide, more than 80% of the suicides (15 of 18) occurred in the risperidone group. Results from the ISC classification were similar to those from the treating psychiatrists' reports.

4. Discussion The strengths of this study are its prospective design, the randomized treatment assignment, and the exceptionally large number of patients and years of exposure. Patient follow-up was highly successful and there were very few missing data. Cox's proportional hazards model analysis of the time to the first suicide attempt, as reported by the treating psychiatrists, showed a significantly lower risk of suicide attempt for sertindole-treated patients than for their risperidone-treated counterparts. When suicide attempts were blindly reviewed by an independent expert group in accordance with C-CASA, the results confirmed the lower risk of suicide attempts in the sertindole group observed by the treating psychiatrist. This effect was observed both with antipsychotic monotherapy (ORT) and including any add-on therapy (WRT) and 30 days thereafter. Classification by an independent safety committee, using a broader definition including suicidal behavior and suicidal ideation, suicidal tendency, and self-injuries and overdoses for which there was no clear suicidal intent, yielded results that were still in favor of sertindole, although not statistically significant. Results from psychiatrists' reports, C-CASA and independent safety committee analyses all indicated that the effect of sertindole on suicide attempt was even stronger during the first year of treatment. This is the period when over 80% of the suicide attempts occurred. Although the estimated risk of suicide was lower for the sample of sertindole-treated patients in the study than for the sample of risperidone-treated subjects, there was no evidence that this reflected a population difference. A history of attempted suicide was the most potent predictor of subsequent suicide. Also, patients who had received polypharmacy prior to study entry had a significantly higher risk of committing suicide. This may reflect the fact that these patients presented with more intractable and treatment-resistant symptoms at inclusion. This study may be limited by cultural differences in the reporting of suicide, even though the region of origin (Europe or Asia) was not a significant predictor of the time to suicide attempt. Overall, suicide mortality in the study was fairly low: approximately 0.21 and 0.28 per 100 patient-years of treatment with sertindole and risperidone, respectively. For comparison, pooling five of the largest studies cited in the literature as satisfying stringent quality criteria (Mortensen and Juel, 1993; Osby et al., 2000; Allgulander et al., 1992; Newman and Bland, 1991; Sernyak et al., 2001) shows average suicide mortality in schizophrenia of 0.49 per 100 patient-years, in a total population of 26,991 patients followed-up over a mean of 7.8 years. The low rate in the present study may be partially explained by the study conditions. Most patients were in a chronic phase of the disorder, where suicide mortality is lower than shortly after illness onset. The proportion of women, who have lower suicide mortality, was somewhat higher than in usual clinical trials for schizophrenia. Finally, as in all clinical trials,

836 the increased amount of clinical attention and background social support in some milieus were likely protective factors. The strength of the evidence from the analyses of suicide attempt lies in both the beneficial hazard ratios observed and in the consistency of these observations in relevant groups (all patients and high-risk subgroup) and periods (ORT, ORT first year, WRT + 30, WRT + 30 first year) which, in spite of the reduced number of events in some of these smaller samples, point in the same direction, supporting a systematic effect of sertindole. The treating psychiatrists' reports and the ISC classification yielded similar figures for completed suicide, but this was not the case for attempted suicide. Completed suicide is a distinct and typically salient event which can be assessed with a fair degree of certainty, even though the incidence of suicide is hard to establish in a study population because of its rare occurrence. Suicidal behavior, a frequent manifestation, is a heterogeneous ensemble with a complex etiology. There are numerous causative factors for suicidality, self-harm and self-aggression, including biological abnormalities (e.g., serotonergic imbalance), psychiatric syndromes (e.g., delusions, depression), and psychological dimensions (e.g., impulsivity, low self-esteem, self-hatred, feeling uncomfortable with one's body, or the propensity to use it as a medium to express affects). The assessment of suicidality is influenced by clinical judgment and definitions. There is no scientific consensus on the instruments to evaluate suicidality. Recently, classification algorithm (C-CASA) with an anchored approach, was tested to categorize suicidality in pediatric antidepressant studies. In this study, the investigators' reports of suicide attempts as well as the C-CASA classification supposed the conjunction of both self-aggressive action and intent to die. In contrast, the ISC definition of suicidal behavior required neither an act (suicidal ideation was sufficient), nor a proven intent to die (self-aggression is not necessarily guided by the desire to die). Other studies of suicidal behavior in schizophrenia have also used a priori definitions. For instance, the InterSePT study prespecified two types of suicidal events. A type 1 event was the occurrence of a significant suicide attempt (including completed suicides), or hospitalization and increased levels of surveillance because of imminent suicide risk. A type 2 event was defined as ratings of “much worse” or “very much worse” on the Clinical Global Impression of Severity of Suicidality (CGI-SS). The MedDRA category of suicide attempt used here was a narrower concept than the definition of suicidal behavior used by the InterSePT study. The biological mechanisms underlying completed suicide and suicide attempts are likely to differ. In fact, it has been shown that medications may exert diverging effects on completed and attempted suicide in schizophrenia. For instance, in a study of suicidal patients with schizophrenia (Sernyak et al., 2001), antidepressants seemed to increase the risk of hospital treated suicide attempts, whereas they actually decreased the number of suicide deaths. The InterSePT and the SCoP studies differed in several important ways. In the latter, the main objective was not suicide but all cause mortality; patients were not selected for being at high risk for suicide; the study population was comprised of patients at a non-acute stage of the illness, treated under naturalistic conditions; finally, treating psychiatrists reporting suicide attempts were not blinded to treatment. Thus, the overall SCoP study results might

M.A. Crocq et al. offer a more accurate reflection of the majority of the patients with schizophrenia. We considered the possibility that the apparent beneficial results of sertindole on suicide attempts were biased by treatment duration, since the median number of days of exposure was smaller in the sertindole than in the risperidone group (Table 1). However this potential bias loses relevance under the conditions of Cox's proportional hazards model, which analyses the time to committing the first suicide attempt, and censors patients at the time they stopped taking randomized treatment. SGAs may differ in their ability to prevent suicide in schizophrenia. There have been previous reports that sertindole treatment may be more effective for reducing risk of suicide than other SGAs. For instance, Azorin et al. (2008) studied retrospectively 344 patients who had followed up to four alternating six-month periods of treatment with sertindole and other antipsychotics (period 1: non-sertindole treatment [24.6% clozapine, and 24.6% risperidone]; period 2: sertindole; period 3: non-sertindole treatment [42.1% olanzapine]; and period 4: sertindole). The percentage of self-harm attempts decreased from 22.8% to 3.5% during the first sertindole-treatment period (p b 0.001), and from 5.8% to 1.8% (p b 0.05) during the second sertindole period. The number of suicide attempts also tended to decrease during both sertindole-treatment periods. The distinct pharmacological profiles of various SGAs might account for differing anti-suicide properties. Both risperidone and sertindole are SGAs sharing serotonin–dopamine antagonist properties, but they differ in other aspects. Firstly, similar to FGAs, some SGAs might induce dysphoria through their D2 receptor occupancy and extrapyramidal side effects. It has been shown that higher D2 receptor occupancy in striatal, temporal, and insular regions is associated with negative subjective experience in patients taking risperidone or olanzapine (Mizrahi et al., 2007). Extrapyramidal symptoms, including akathisia, can occur with risperidone at higher doses; this may be a mechanism for increased suicidality. In contrast, sertindole appears efficacious at a low D2 receptor occupancy, comparable to that produced by clozapine. It was shown (Nyberg et al., 2002) that D2 striatal receptor occupancy was 52–68% in patients receiving sertindole 20 mg/day for 6–8 weeks, similar occupancy being found in the thalamus, and the temporal and frontal cortices. Secondly, SGAs may exert an antidepressant action, either directly through serotonergic functions or other systems, or indirectly through the improving of cognitive functioning. Sertindole is a 5-HT6 antagonist (like clozapine, olanzapine, loxapine, quetiapine, and zotepine). Serotonin 5-HT6 receptors might be linked to the release of brain derived neurotrophic factor (Schechter et al., 2005), and the putative antidepressant effect of 5-HT6 antagonists is the subject of investigation (Wesołowska and Nikiforuk, 2007; Svenningsson et al., 2007). Sertindole with its particular pharmacological profile might confer protection from attempted (fatal and non-fatal) suicide in patients with schizophrenia.

Role of the funding source The study was sponsored by H. Lundbeck A/S.

Suicide attempts in SCoP

Contributors Study concept and design: MH Lader, M Sturkenboom, R Mann, ND Moore, GC Hall, and M Toumi. Acquisition of data: All study investigators. Analysis and interpretation of data: All authors. Drafting the manuscript: MA Crocq. Critical revision of the manuscript: All authors. Statistical expertise: P Tanghøj, B Everitt. Study supervision: IMC Members, A Mittoux.

Conflict of interest The authors have served as consultants or were employees of the sponsor during the conduct of the clinical trial. Independent Management Committee: R Mann (Chairman), MD Drici, M Sturkenboom, D Naber, SHL Thomas, B Everitt, MA Crocq, and E Sylvälahti (March 2002–April 2006). Independent Safety Committee: ND Moore (Chairman), S Priori, MH Lader, GC Hall, A Charlesworth, C Le Jeunne, F Thibaut, P Auby, and G Heebol-Nielsen.

Acknowledgments The SCoP study investigators express their appreciation to the 9858 participants, without whom the trial could not have been done.

References Aguilar, E.J., Siris, S.G., 2007. Do antipsychotic drugs influence suicidal behavior in schizophrenia? Psychopharmacol. Bull. 40, 128–142. Allgulander, C., Allebeck, P., Przybeck, T.R., Rice, J.P., 1992. Risk of suicide by psychiatric diagnosis in Stockholm County: a longitudinal study of 80,970 psychiatric inpatients. Eur. Arch. Psychiatry Clin. Neurosci. 241, 323–326. Altamura, A.C., Bassetti, R., Bignotti, S., Pioli, R., Mundo, E., 2003. Clinical variables related to suicide attempts in schizophrenic patients: a retrospective study. Schizophr. Res. 60, 47–55. Azorin, J.M., Murteira, S., Hansen, K., Toumi, M., 2008. Evaluation of patients on sertindole treatment after failure of other antipsychotics: a retrospective analysis. BMC Psychiatry 8, 16. Bakst, S., Rabinowitz, J., Bromet, E.J., 2010. Antecedents and patterns of suicide behavior in first-admission psychosis. Schizophr. Bull. 36, 880–889. Barak, Y., Mirecki, I., Knobler, H.Y., Natan, Z., Aizenberg, D., 2004. Suicidality and second-generation antipsychotics in schizophrenia patients: a case-controlled retrospective study during a 5year period. Psychopharmacology 175, 215–219. Beisser, A.R., Blanchette, J.E., 1961. A study of suicides in a mental hospital. Dis. Nerv. Syst. 22, 365–369. Bousquet, C., Lagier, G., Lillo-Le Louët, A., Le Beller, C., Venot, A., Jaulent, M.C., 2005. Appraisal of the MedDRA conceptual structure for describing and grouping adverse drug reactions. Drug Saf. 28, 19–34. Bray, I., Gunnell, D., 2006. Suicide rates, life satisfaction and happiness as markers for population mental health. Soc. Psychiatry Psychiatr. Epidemiol. 41, 333–337. Breier, A., Astrachan, B.M., 1984. Characterization of schizophrenic patients who commit suicide. Am. J. Psychiatry 141, 206–209. Caldwell, C.B., Gottesman, I.I., 1990. Schizophrenics kill themselves too: a review of risk factors for suicide. Schizophr. Bull. 16, 571–588.

837 De Hert, M., McKenzie, K., Peuskens, J., 2001. Risk factors for suicide in young people suffering from schizophrenia: a long-term follow-up study. Schizophr. Res. 47, 127–134. Evans, J., Middleton, N., Gunnell, D., 2004. Social fragmentation, severe mental illness and suicide. Soc. Psychiatry Psychiatr. Epidemiol. 39, 165–170. Harkavy-Friedman, J.M., Nelson, E.A., Venarde, D.F., 2001. Suicidal behavior in schizophrenia and schizoaffective disorder. Clin. Neurosci. Res. 1, 345–350. Haukka, J., Tiihonen, J., Härkänen, T., Lönnqvist, J., 2008. Association between medication and risk of suicide, attempted suicide and death in nationwide cohort of suicidal patients with schizophrenia. Pharmacoepidemiol. Drug Saf. 17, 686–696. Heila, H., Isometsa, E.T., Henriksson, M.M., Heikkinen, M.E., Marttunen, M.J., Lönnqvist, J.K., 1997. Suicide and schizophrenia: a nationwide psychological autopsy study on age- and sexspecific clinical characteristics of 92 suicide victims with schizophrenia. Am. J. Psychiatry 154, 1235–1242. Heilä, H., Haukka, J., Suvisaari, J., Lönnqvist, J., 2005. Mortality among patients with schizophrenia and reduced psychiatric hospital care. Psychol. Med. 35, 725–732. Helgason, L., 1990. Twenty years' follow-up of first psychiatric presentation: what could have been prevented? Acta Psychiatr. Scand. 81, 231–235. Johnson, D.A., Pasterski, G., Ludlow, J.M., Street, K., Taylor, R.D., 1983. The discontinuance of maintenance neuroleptic therapy in chronic schizophrenic patients: drug and social consequences. Acta Psychiatr. Scand. 67, 339–352. Kerwin, R.W., Bolonna, A.A., 2004. Is clozapine antisuicidal? Expert Rev. Neurother. 4, 187–190. Kreyenbuhl, J.A., Kelly, D.L., Conley, R.R., 2002. Circumstances of suicide among individuals with schizophrenia. Schizophr. Res. 58, 253–261. Levine, S.Z., Bakst, S., Rabinowitz, J., 2010. Suicide attempts at the time of first admission and during early course schizophrenia: a population based study. Psychiatry Res. 177, 55–59. Mamo, D.C., 2007. Managing suicidality in schizophrenia. Can. J. Psychiatry 52 (6 Suppl 1), 59S–70S. Meltzer, H.Y., 2001. Treatment of suicidality in schizophrenia. Ann. NY Acad. Sci. 932, 44–60. Meltzer, H.Y., Fatemi, H., 1995. Suicide in schizophrenia: the effect of clozapine. Clin. Neuropharmacol. 18 (Suppl. 3), S18–S24. Meltzer, H.Y., Okayli, G., 1995. Reduction of suicidality during clozapine treatment of neuroleptic-resistant schizophrenia: impact on risk–benefit assessment. Am. J. Psychiatry 152, 183–190. Meltzer, H.Y., Alphs, L., Green, A., Altamura, A.C., Anand, R., Bertoldi, A., Bourgeois, M., Chouinard, G., Islam, M.Z., Kane, J., Krishnan, R., Lindenmayer, J.-P., Potkin, S., 2003. Clozapine treatment for suicidality in schizophrenia. Arch. Gen. Psychiatry 60, 82–91. Miles, C.P., 1977. Conditions predisposing to suicide: a review. J. Nerv. Ment. Dis. 164, 231–246. Mizrahi, R., Rusjan, P., Agid, O., Graff, A., Mamo, D.C., Zipursky, R.B., Kapur, S., 2007. Adverse subjective experience with antipsychotics and its relationship to striatal and extrastriatal D2 receptors: a PET study in schizophrenia. Am. J. Psychiatry 164, 630–637. Montout, C., Casadebaig, F., Lagnaoui, R., Verdoux, H., Philippe, A., Begaud, B., Moore, N., 2002. Neuroleptics and mortality in schizophrenia: prospective analysis of deaths in a French cohort of schizophrenic patients. Schizophr. Res. 57, 147–156. Mortensen, P.B., Juel, K., 1993. Mortality and causes of death in first admitted schizophrenic patients. Br. J. Psychiatry 163, 183–189. Nangle, J.M., Clarke, S., Morris, D.W., Schwaiger, S., McGhee, K.A., Kenny, N., Murphy, K., Gill, M., Corvin, A., Donohoe, G., 2006. Neurocognition and suicidal behaviour in an Irish population with major psychotic disorders. Schizophr. Res. 85, 196–200.

838 Newman, S.C., Bland, R.C., 1991. Mortality in a cohort of patients with schizophrenia: a record-linkage study. Can. J. Psychiatry 36, 239–245. Nyberg, S., Olsson, H., Nilsson, U., Maehlum, E., Halldin, C., Farde, L., 2002. Low striatal and extra-striatal D2 receptor occupancy during treatment with the atypical antipsychotic sertindole. Psychopharmacology 162, 37–41. Osby, U., Correia, N., Brandt, L., Ekbom, A., Sparen, P., 2000. Mortality and causes of death in schizophrenia in Stockholm County, Sweden. Schizophr. Res. 45, 21–28. Palmer, B.A., Pankratz, V.S., Bostwick, J.M., 2005. The lifetime risk of suicide in schizophrenia. A reexamination. Arch. Gen. Psychiatry 62, 247–253. Peuskens, J., Tanghø, j.P., Mittoux, A., 2008. Sertindole Cohort. The Sertindole Cohort Prospective (SCoP) study: rationale, design and methodology. Pharmacoepidemiol. Drug Saf. 17, 425–433. Pompili, M., Amador, X.F., Girardi, P., Harkavy-Friedman, J., Harrow, M., Kaplan, K., Krausz, M., Lester, D., Meltzer, H.Y., Modestin, J., Montross, L.P., Mortensen, P.B., Munk-Jørgensen, P., Nielsen, J., Nordentoft, M., Saarinen, P.I., Zisook, S., Wilson, S.T., Tatarelli, R., 2007. Suicide risk in schizophrenia: learning from the past to change the future. Ann. Gen. Psychiatry. 6, 10. Pompili, M., Lester, D., Grissini, A., Innamorati, M., Calandro, F., Iliceto, P., De Pisa, E., Tatarelli, R., Girardi, P., 2009. Completed suicide in schizophrenia: evidence from a case–control study. Schizophr. Res. 167, 251–257. Posner, K., Oquendo, M.A., Gould, M., Stanley, B., Davies, M., 2007. Columbia Classification Algorithm of Suicide Assessment (CCASA): classification of suicidal events in the FDA's pediatric suicidal risk analysis of antidepressants. Am. J. Psychiatry 164, 1035–1043. Qin, P.N., Nordentoft, M., 2005. Suicide risk in relation to psychiatric hospitalization: evidence based on longitudinal registers. Arch. Gen. Psychiatry 62, 427–432. Reutfors, J., Brandt, L., Jönsson, E.G., Ekbom, A., Sparén, P., Ösby, U., 2009. Risk factors for suicide in schizophrenia: findings from a Swedish population-based case–control study. Schizophr. Res. 108, 231–237. Rossau, C.D., Mortensen, P.B., 1997. Risk factors for suicide in patients with schizophrenia: nested case–control study. Br. J. Psychiatry 171, 355–359. Roy, A., 1982. Suicide in chronic schizophrenia. Br. J. Psychiatry 141, 171–177. Schechter, L.E., Ring, R.H., Beyer, C.E., Hughes, Z.A., Khawaja, X., Malberg, J.E., Rosenzweig-Lipson, S., 2005. Innovative approaches for the development of antidepressant drugs: current and future strategies. NeuroRx. 2, 590–611.

M.A. Crocq et al. Schennach-Wolff, R., Jäger, M., Seemüller, F., Obermeier, M., Schmauss, M., Laux, G., Pfeiffer, H., Naber, D., Schmidt, L.G., Gaebel, W., Klosterkötter, J., Heuser, I., Maier, W., Lemke, M.R., Rüther, E., Klingberg, S., Gastpar, M., Möller, H.J., Riedel, M., 2010. Outcome of patients with schizophrenia: results from a naturalistic study. Acta Psychiatr. Scand. 121, 359–370. Sernyak, M.J., Desai, R., Stolar, M., Rosenheck, R., 2001. Impact of clozapine on completed suicide. Am. J. Psychiatry 158, 931–937. Siris, S.G., 2001. Suicide and schizophrenia. J. Psychopharmacol. 15, 127–135. Spivak, B., Mester, R., Wittenberg, N., Maman, Z., Weizman, A., 1997. Reduction of aggressiveness and impulsiveness during clozapine treatment in chronic neuroleptic-resistant schizophrenic patients. Clin. Neuropharmacol. 20, 442–446. Svenningsson, P., Tzavara, E.T., Qi, H., Carruthers, R., Witkin, J.M., Nomikos, G.G., Greengard, P., 2007. Biochemical and behavioral evidence for antidepressant-like effects of 5-HT6 receptor stimulation. J. Neurosci. 27, 4201–4209. Thomas, S.H.L., Drici, M.D., Hall, G.C., Crocq, M.A., Everitt, B., Lader, M.H., Jeunne, C.L., Naber, D., Priori, S., Sturkenboom, M., Thibaut, F., Peuskens, J., Mittoux, A., Tanghøj, P., Toumi, M., Moore, N.D., Mann, R.D., 2010. Safety of sertindole versus risperidone in schizophrenia: principal results of the sertindole cohort prospective study (SCoP). Acta Psychiatr. Scand. April 8. [Epub ahead of print]. Tiihonen, J., Walhbeck, K., Lönnqvist, J., Klaukka, T., Ioannidis, J.P., Volavka, J., Haukka, J., 2006. Effectiveness of antipsychotic treatments in a nationwide cohort of patients in community care after first hospitalisation due to schizophrenia and schizoaffective disorder: observational follow-up study. Br. Med. J. 333, 224–229. Tiihonen, J., Lönnqvist, J., Wahlbeck, K., Klaukka, T., Niskanen, L., Tanskanen, A., Haukka, J., 2009. 11-year follow-up of mortality in patients with schizophrenia: a population-based cohort study (FIN11 study). Lancet 374, 620–627. Trémeau, F., Staner, L., Duval, F., Corrêa, H., Crocq, M.A., Darreye, A., Czobor, P., Dessoubrais, C., Macher, J.P., 2005. Suicide attempts and family history of suicide in three psychiatric populations. Suicide Life Threat Behav. 35, 702–713. Walker, A.M., Lanza, L.L., Arellano, F., Rothman, K.J., 1997. Mortality in current and former users of clozapine. Epidemiology 8, 671–677. Wesołowska, A., Nikiforuk, A., 2007. Effects of the brain-penetrant and selective 5-HT6 receptor antagonist SB-399885 in animal models of anxiety and depression. Neuropharmacology 52, 1274–1283. World Health Organization, 1992. The ICD-10 classification of mental and behavioural disorders. Clinical descriptions and diagnostic guidelines. World Health Organization, Geneva.