Journal of Affectice Disorders, 24 ( 1992) 1- 10 0 1992 Elsevier Science Publishers B.V. AEI rights reserved 01650327/92/$05.00
JAD 00857
luoxetine: no association w in obsessive-compulsive Charles M. Beasley, Jr., Janet H. Potvin, Daniel N. Masica, David E. Wheadon, Bruce E. Dornseif and Law-a A. Genduso Division of Clinical Neurosciences, Lilly Research Laboratoms, Eli Lilly and Conzpany, Indianapolis, IN 46285, USA
(Received 12 August 1991) (Revision received 20 September 1991) (Accepted 25 September 1991)
Summary
Since (a) obsessive-compulsive disorder (QCD) may involve serotonergic neural transmission abnormalities also thought to be related to regulation of suicide and aggression, (b) comorbidity between OCD and depression is substantial, and (c) depression is a major risk factor for suicide, a comprehensive analysis of clinical trial data was undrrtaken to assess the potential association of fluoxetine, a serotonin uptake inhibitor, and suicidality (suicidai acts and ideation). Pooled data from clinical trials comparing fluoxetine (n = 266) and placebo (n = 89) in patients with DSM-IIIR OCD were analyzed retrospectively. No suicidal acts occurred during placebo lead-in or double-blind therapy. Mean Hamilton Depression Scale item 3 (suicide item) scores improved statistically significantly with fluoxetine compared with placebo. Worsening in suicidal ideation was statistically significantly more frequent with placebo than with fluoxetine. Emergence of substantial suicidal ideation (change in baseline item 3 score of 0 or 1 to 3 or 41 was numerically greater with placebo than with fluoxetine (3.6%. vs. 1.7%; not statistically significant). The incidence of suicidality in fluoxetine-treated patients with OCD was low, compared favorably with rates in corresponding placebo-treated patients, and was well within the range of estimates in previous studies of patients with OCD. These controlled clinical trial results suggest no undue risk of suicidality in. patients with OCD treated with fluoxetine.
Key words: Fluoxetine; Suicidality; Obsessive-compulsive
disorder; Depression; Serotonin
Introduction .4ddress for correspondence: Dr. Charles M. Reasley, Jr., Eli Lilly and Company. Lilly Corporate Center 2128,.Indianapolis, 1N 46285, USA. Tel.: (317) 276-4994; Fax: (317) 276-6696.
Fluoxetine hydrochloride, a highly potent and selective serotonin uptake inhibitor, has been
demonstrated to be an effective antidepressant in large placebo-controlled trials (Stark and Hardison. 1985; Wernicke et al., 1987, 1988) and has been suggested to be effective in the treatment of obsessive-compulsive ciisorder (OCD) in preliminary trials (tenike et al., 1989; Levine et al., 1989; Kim and Dysken, 1990; Landry and Chouinard, 1990; Pigott et al., 1990). Fluoxetine inhibits the uptake of serotonin without affecting other neurotransmitter uptake systems (Lemberger et al., 1978a,b; Michelson and Pfenning, 1984) and is essentially devoid of affinity for multiple neurotransmitter receptors at concentrations that inhibit serotonin uptake (Richelson and Nelson, 1984; Wander et al., 1986). The desmethylated metabolite, norfluoxetinc, is also selective for inhibition of serotonin uptake (Fuller et al., 1978). Fluoxetine appears to result in increased net serotonergic neural function as a result of desensitization of both presynaptic (Blier et al., 1988) and somatic (Weiner et al., 1989) inhibitory autoreceptors. OCD is characterized by persistent, intrusive thoughts and images (obsessions) and repetitive, ritualistic behaviors the individual feels driven to perform (compulsions) (American Psychiatric Association, 1987); the obsessions sometimes include thoughts of aggression and violent behavior (homicidal, suicidal) (Perse, 1988). Although the etiology of OCD is unknown (Perse, 1988) it is hypothesized that abnormal regulation of brain scrotonergic function is related to at least some aspects of the disorder (Tinsel et al., 1984, 1985; Turner et al., 1985; Charney et al., 1988; Landry and Chouinard, i990). Extensive animal literature supports the role of central serotonin (S-E-IT) in the regulation of aggressive impulses (Cocarro, 1989). Animal and clinical studies have shown that bidchemical indices suggestive of deficient serotonergic function are associated with the expression of aggressive and/or suicidal behaviors (Valzelli, 1984; Charney et al., 1988; Cocarro, 1989). Specifically, low concentrations of CSF 5-HlAA have been associated with violent suicide, interpersonal aggressiveness, and criminality (Charney et al., 1988; Cocarro, 1989). Increases in postsynaptic serotonin receptors (5-HT,) also have been associated with suicide (Cocarro, 1989).
It has been suggested that the obsessions related to aggression in patients with OCD may be considered a disinhibition of normally suppressed aggressive thoughts, whereas other obsessions and many compulsive rituals may be viewed as related to failure of habituation (Charney et al., 1988). The effectiveness of serotonin uptake inhibitors in treating OCD then may be related to their ability to increase serotonergic function (Blier et al., 1988; Weiner ct al., 1989) arrd thereby reduce aggression directed at external objects and oneself (Kostowski et al., 1984; Molina et al., 1987). A relationship between OCD and depression is well recognized (Gittelson, 1966a-d; Goodwin et al., 1969; Rachman and Hodgson, 1980; Perse, 1988). Rachman and Hodgson (1980) reviewed data and reported findings indicating that 45% of patients with OCD were depressed at the onset of the disorder and many patients free of depression at onset developed subsequent depressions. The Epidemiologic Catchment Area study estimated that 31.7% of respondents with OCD also had a history of a major depressive episode (Karno et al., 1988). Depression is a major risk factor for suicide (Hirschfeld and Davidson, 1988; Black and Winokur, 1990; Buda and Tsuang, 1990). It has been reported that 15% of depressed patients die by suicide (Guze and Robins, 1970): 20-40% of depressed patients will display suicidal behavior (Brent et al., 1988); and perhaps as many as 80% of depressed patients will experience suicidal ideation (Roose and Glassman, 1988). Since (a) the pathophysiology of OCD may involve abnormalities in serotonergic neural transmission also thought to be related to regulation of suicide and aggression, (b) substantial comorbidity exists between OCD and depression, and (c) depression is a major risk factor for suicide, it might be hypothesized that OCD patients would be at increased risk of suicide. However, suicide has been reported to occur infrequently in patients with OCD (Goodwin et al., 1969; Coryell, 1981; Black and Winokur, 1990). Although comparative analysis has failed to establish a specific association (Fava and Rosenbaum, 1991), recent case reports have noted emergence or a worrpning of suicidality in some patient=. with symptoms of severe refractory depression, panic disorder, and/or QCD treated
3
with fluoxetine (Papp and Gorman, 1990; Teicher et al., 1990a,b; Masand et al., 1991). Therefore, it is of interest to examine potential relationships between fluoxetine therapy and suicidality during treatment of OCD. A comprehensive analysis of double-blind, placebo-controlled clinical trial data was thus undertaken to assess the potential associatioi, uetween fluoxetine and suicidality in patients with OCD.
for 1 week before receiving the higher dose. For the primary assessment of suicidality, data were analyzed for the 13-week, double-blind therapy phase of the trials. Data were also reviewed for the 24-week double-blind extension. Patients A total of 355 randomized patients who participated in the studies after giving written informed consent were included in the suicidality analyses. A total of 266 patients were treated with fluoxetine during the double-blind phase; 89 were treated with placebo. Patients were male and female outpatients, aged 14-70 years, who met DSM-IIIR (American Psychiatric Association, 1987‘) criteria for OCD. Patients had OCD of at least 1 year’s duration and of at least moderate severity. The diagnosis was made based on routine psychiatric interview, Depression, if present, was secondary to OCD. Pregnant women, lactating women, and women of childbearing potential not using medically accepted means of contraception were excluded. Patients with coexisting axis I disorders (other than a secondary depressive mood disorder), psychotic features, history of alcohol or drug abuse, serious suicidai risk, unstable serious medical illness, or history of seizures were also excluded. Additional exclusion criteria were use of other psychotropic drugs or tryptophan. Characteristics of the patient population are summarized in Table 1. A total of 76 patients continued therapy in the double-blind responder extension (six placebo, 70
d data in the US lnvestigational New ND) fluoxetine clinical trial data base were analyzed from two multicenter, doubleblincl, Lontrolled clinical trials with identical decomparing fluoxetine and placebo in patients OGD. The studies consisted of a l-week, single-blind placebo screening phase and a 13week, double-blind phase, followed by two concurrent 24-week extensions: a double-blind extension In patients who responded to therapy by exhibiting at least a 25% decrease in Yale-Brown Obsessive Compulsive Scale (Goodman et al., 1989a,b) total score at the end of the double-blind phase (responder extension) and an open-label extension in patients who did not respond to double-blind therapy (nonresponder extension). Patients were randomly allocated to therapy with placebo or fixed daily doses of fluoxetine (20 mg, 40 mg, and 60 mg). Patients randomized to fluoxetine 60 mg/day received a dose of 40 mg/day
TABLE
1
SUMMARY
OF PATIENT
CHARACTERISTICS
AT BASELINE Fluoxetine
Placebo
Characteristic
(n = 89)
No. (%
bwhite
20 mg
40 mg
60 mg
(n=87)
(n=891
tn = 901
81 (91.01
82 (94.3)
79 (88.81
80 (88.9)
No. (u/cl female
45 (50.61
46 (52.91
49 (55.11
56 (62.2)
No. (%l obsessive and compulsive
82 (92.11
84 (96.6)
87 (92.1)
87 (96.7)
Age (years) ’
37.2
+ 12.3
38.0
of:12.7
37.0
+ 10.8
35.4
+11.3
24.3
+
5.7
23.6
f
5.7
23.5
+
5.6
24.4
+
5.2
11.6 k
3.9
11.6 f
3.4
11.4 +
3.9
11.8 +
3.5
5.9
9.6 +
5.2
9.4 f
5.9
0.60
0.37&
0.65
0.39+
0.67
Yale-Brown
obsessive-compulsive scale total score a
Yale-Brown Hamilton
obsessive-compulsive scale obsession subscore a Depression Scale 17.item total score a
Hamilton
Depression
Scale item 3 score a
’ Mean +_standard deviation.
9.0 0.27f
f
5.9
9.2
0.56
0.25h
+
active therapy). The patients who continued active therapy were approximately evenly distributed among treatment groups: fluoxetine 20 mg, n = 23; fluoxetine 40 mg, IZ= 21; fluoxetine 60 mg, n = 26. A total of 62 of the 76 patients completed the 34-week extension. Case definitions and screefzing metkods A suicidal act was defined as a behavior pur-
posefully undertaken by the patient in which the outcome wzs reasonably likely to be self-harm, and no explicit data suggested that suicide was not intended (Hirschfeld and Davidson, 1988). The definition was broad so as not to exclude behaviors that might be labeled suicidal gestures. The definition excluded the following behaviors: accidental overdose, attempt to increase efficacy by ingestion of excess medication, drug or medicine abuse, and self-harm where c!inical report text indicated the patient was nonsuicidal. For example, the taking of several aspirin tablets, regardless of the psychosocial context, would be classified as a suicidal act unless comments explicitly indicated that the excess ingestion had the intended purpose of treating a particularly severe headache or some other explicit nonsuicidal purpose. The threat of taking such an overdose, however, anuld not be considered a suicidal act but rather the verbal expression of suicidal ideation. The case definition targeted any event that occurred through the day following the last day study drug was taken during the double-blind therapy phase of the study in compliance with the protocol. The time limit was adopted because (1) postdiscontinua\ion data were not systematically collected; (2) an event occurring after discontinuation might have been influenced by terminating participation in an experimental protocol or withdrawal of treatment; and (3) additional treatment with other medications might have been initiated after discontinuation. Suicidal ideation was defined as including the foIlowir,g types of thinking: wishing he or she were dead, would like to kill self but would not, thinking about suicide with neutral intent, thinking might commit suicide but unsure, and thinking about suicide with planning and/or intent. The definition excluded dreams and thinking of iife as not worth living.
New or increasing suicidal ideation was identified by increases in Hamilton Depression Scale item 3 scores and reports of treatment-emergent adverse events (events that first occurred or worsened after randomization to double-blind aherspy) indicative of suicidal ideation. Emergence of substantial suicidal ideation was defined as a change in score on item 3 (suicide item) of the Hamilton Depression Scale (Hamilton, 1967) from 0 or 1 at baseline to a score of 3 or 4 during double-blind therapy. The Hamilton Depression Scale is used to systematically rate severity of depression in clinical trials. Item 3 rates the severity of suicidality as 0 = no thinking of suicide, 1 = feels life is not worth living, 2 = wishes he or she were dead (or having any thoughts of possible death to self), 3 = suicidal ideas or gestures, 4 = attempts at suicide. The comprehelwive incidetzce of emergent suicidal ideation was based on the combination of cases identified by analysis of change in item 3 scores described above and those identified as having suicidal ideation as a treatment-emergent adverse event. To screen for patients who might have experienced suicidality, the US IND fluoxetine OCD clinical trial data base was searched electronically for information obtained from clinical report forms (adverse events, reasons for study discontinuation, free-text clinical comments, and a score of 4 on item 3 of the Ha ilton Depression Scale). The Drug Experience Network (DEN) data base. which contains reports of serious adverse events (as defined by Food and Drug Administration criteria) occurring in clinical trials and all adverse events spontaneously reported to Eli Lilly and Company as part of postmarketing surveillance (Talbott et al., 19871, was also searched electronically to obtain additional information relevant to suicidality in the OCD clinical trials. Potential cases identified electronically were reviewed individually by two Lilly psychiatrists, while blinded to treatment assignment, to confirm that they represented instances of suicidality. Descriptive analyses
Rates of suicidal acts, emergence of substantial suicidal ideation, and the comprehensive incidence of emergent suicidal ideation were calcu-
5
lated for each treatment group. Rates of emergence of substantial suicidal ideation were based on patients with a baseline and at least one Hamilton Depression Scale item 3 measurement after random allocation to double-blind therapy. Rates of acts and the comprehensive incidence of emergent suicidal ideation were determined based on all randomized patients. Four analyses dealing exclusively with change in the Hamilton Depression Scale item 3 scores were done: the first examined the mean change from baseline to endpoint; the second compared the proportion of patients who demonstrated a worsening in their item 3 scores from baseline to the highest score during double-blind therapy; the third compared the proportion of patients who demonstrated a worsening in their item 3 scores from baseline to their last score in double-blind therapy (endpoint); and the fourth examined the proportions of patients who manifested the emergence of substantial suicidal ideation. Analyses of suicidahty during the double blind responder extension were based only on adverse event reports, reasons for discontinuation, and clinical comments since Hamilton Depression Scale item 3 data were :tot collected. Inferential analyses
Differences in mean change in Hamilton Depression Scale item 3 scores from baseline to endpoint between fluoxetine- and placebo-areated patients were assessed by Wilcoxon’s rank sum test, and among the four treatment groups by analysis of variance of rank-transformed differences since this procedure readily provided pairwise comparisons. Pearson’s chi-square test was used to compare rates of worsening of suicidal ideation between ffuoxetine- CIO-mg, 40-mg, and 60-mg groups pooled) and placebo-treated patients. In addition, the Mantel-Haenszel chisquare test (one degree of freedom) was used to test for a linear Jose-response relationship in the incidence of suicidal ideation among fluoxetine treatment groups. Pearson’s chi-square test was used to compare the comprehensive incidence rate of emergent suicidal ideation among treatment groups, Statistical analyses were performed using SAS procedures (SAS, 1989). Statistical significance was defined as P d 0.050 (two-tailed).
Results Suicidak-y at baseline
_AnaIysis of Hamilton Depression Scale item 3 scores at baseline indicated that passive or active suicidal thinking (score > 2 on Ham&on Depression Scale item 3) was present in 8.2% (29 of 355) of patients. One patient, who had passive suicidal ideation identified as an event at study entry (visit l), was randomly allocated to the fluoxetine 40”mg group. The ideation had ceased by the first visit after randomization (visit 3). Suicidal acts No suicidal acts, either fatal or nonfatal, oc-
curred during the single-blind placebo lead-in phase, the double-blind therapy phase, or the double-b!ind responder extension. Suicidal ideation
Hamilton Depression Scale item 3 data were analyzed for 349 of the 355 patients in the trial who had item 3 measurements for at least two visits: one immediately before treatment (baseline, visit 2) and at least one after beginning double-blind therapy. Six patients were excluded from this analysis because they did not have at least one Hamilton Depression Scale item 3 measurement following random allocation to doubleblind therapy. Mean change item 3
in Hamilton Depression Scale
Mean change in Hamilton Depression Scale item 3 scores from i aseline to endpoint is summarized in Table 2. A statistically significant (P = 0.003) difference in the mean improvement in the Hamilton Depression Scale item 3 mean score was observed between fluoxetine and placebo. All three fluoxetine dosage groups demonstrated improvement from baseline to endpoint. Statistically significant worsening of the idean score from baseline to endpoint was observed with placebo (Wilcoxon signed rank test). Worsening in Hamilton Depression Scale item 3
The proportion of patients who experienced worsening in suicidal ideation, as assessed by an increase in Hamilton Depression Scale item 3
TABLE 2 MEAN CHANGE IN HAMILTON BLIND THERAPY a
DEPRESSION
SCALE
ITEM 3 SCORES
FROM
BASELINE ---
-----Fluoxetine
Placebo
Variable -I Scale item 3 score
Mean baseline Hamilton Depression Mean change in Hamilton Depression
TO END OF DQUBLE-
Scale item 3 score h
0.27 (n = 891 0.06 ’ (n = 881
All
20 mg
40 mg
60 mg
0.34 (n = 2661 -0.15 (n = 2611
0.25 (n = 87) - 0.08 (n = 86)
0.37 (‘1 = 89) -0.17 Cn = 86)
0.39 (n = 90) - 0.20 (n = 891
a Negative change indicates improvement. b Six patients did not have at least one Hamilton Depression Scale item 3 score measurement following assignment to double-blind therapy. ’ Overa!l treatment comparison, placebo vs. alt fluoxetine, P = 0.003 (Wilcoxon rank sum test); placebo vs. individual fluoxetine treatment groups, P = 0.005 (analysis of variance of rank-transformed data).
from baseline to the highest score at any subsequent visit, was statktically significantly higher for placebo (29.6%) than for fluoxetine (19.2%: P = 0.041, see Table 3). Comparison of worsening in Hamilton Depression Scale item 3 scores among individual treatment groups showed a statistically significant (P = 0.014) linear decrease from the placebo group to the fluoxetine 6C-mg group in the proportions of patients whose suicidal ideation worsened (Table 3). These data did not indicate clustering of worsening at higher doses. Worsening in Hamilton Depression Scale item 3 bcores from baseline to endpoint (Table 4)
occurred more than three times more frequently in placebo-treated patients (14.8%) than in fluoxetine-treated patients (4.6%), again a statistically significant (P = 0.001) difference. Comparison of baseline-to-endpoint change among individual treatment groups also showed a statistically significant (P = 0.012) decrease from the placebo group to the fhroxetine groups in the proportion of patients who demonstrated a worsening in their suicidal ideation (Table 4). Emergertce of suicidal ideation
baseline, 238 of the 266 fluoxetine-treated patients and 83 of the 89 placebo-treated patients At
TABLE 3 NUMBER AND PERCENTAGE OF PATIENTS EXPERIENCING WORSENING OR NO CHANGEJMPROVEMENT IN HAMILTON DEPRESSION SCALE ITEM 3 SCORES FROM BASELINE TO THE HIGHEST SURSEQUENT SCORE
TABLE 4
Measure
Measure
Worse
Placebo a.h Fluoxetine
No change/ improved
26 (29.6%) 62 (70.5%)
Total
88
All
20mg
40mg
6Ctmg
50 i19.2%1 211 (80.8%) 261
20 (23.3%) 66 (76.7%) 86
I7 I19.8%) 69 (80.2%1 86
13 (14.6%) 76 (85.4%) 89
_
a Overall treatment comparison, placebo vs. all fluoxetine, P = 0.041 (Pearson chi-square). h Test for linear association between level of response and treatment, P = 0.014 !blantel-Haenszel chi-square).
NUMBER AND PERCENTAGE OF PATIENTS EXPERIENCiNG WORSENING OR NO CHANGE/IMPROVEMENT IN HAMILTON DEPRESSION SCALE ITEM 3 SCORES FROM BASELINE TO ENDPOINT Placebo a.b Fluoxetine All Worse No change/ improved Total
13 (14.8%) 75 (85.2%1 88
20mg
12 4 (4.6%) (3.7%) 249 P2 (95.4%) (95.3%) 261 86
4Omg
6Omg
4 (4.7%) 82 (95.3%! 86
4 (4.5%) 85 (95.5%) 89
a Overall treatment comparison, placebo vs. all fluoxetine, P = O.GOl (Pearson chi-square). b Test for linear association between level of response and treatment, P = 0.012 (Mantel-Haenszel chi-square).
who had a baseline and at least one Hamilton Depression Scale measurement after randomization had little or no thinking about suicide (Hamilton Depression Scale item 3 scores of 0 or 1). Of these, seven patients (four fluoxetine, three placebo) had emergence of substantial suicidal ideation during double-blind treatment as defined by a change in Hamilton Depression Scale item 3 scores from 0 or 1 to 3 or 4. One flucxetine-treated patient was receiving a 20-mg dose; two a 40-mg dose; and one a 60-mg dose. The patient receiving the fluoxetine 60-mg dose also had suicidal ideation noted as a treatment-emergent adverse event, as described below. Emergence of substantial suicidal ideation in patients without ideation at baseline (Hamilton Depressio7 Scale item 3 score = 0 or I; fiuoxetine n = 238, placebo it = 83) occurred at a numerically higher rate in placebo-treated patients than in fluoxetine-treated patients (3.6% vs. 1.7%), but this difference was not statistically significant (P = 0.299; chi-square test). Suicidal ideation was noled as a treatmentemergent adverse event for three patients (all in the fluoxetine 6C-mg group) during the d_rubleblind phase of the trial. No instances of suicidal ideatton were reported as treatment-emergent events with either fluoxetinr: o; placebo during the double-blind responder extension. The comprehensive incidence rate of emergent suicidal ideation (combin;ng cases identified by review of treatment-emergent adverse events and those identified by analysis of change in Hamilton Depression Scale item 3 and considering all randomized patients) was 2.3% for fluoxetine (6 of 266 patients) and 3.4% for placebo (3 af 89 patients). Among the fluoxetine-treated patients, two were identified by review of cl11cal report form data (adverse events), three by change in Hamilton Depression Scale item 3 score, and one by both case screening methods. Among placebo-treated patients, all cases were identified by change in Hamilton Depression Scale item 3 score. While the comprehensive incidence of treatment-emergent suicidal ideation occurred at a numerically lower rate in flusxetine-treated than in placebo-treated patients, this difference was not statistically significant (P = 0.562; chi-square test).
Discussion
NO suicidal acts occurred during the placebo lead-in phase, the double-blind therapy phase, or the double-blind responder extension with either fluoxetine or placebo. During the double-blind therapy phase, suicidal ideation as measured hy change in item 3 of the Hamilton Depression Scale improved in fluoxetine-treated patients, while worsening was observed in placebo-treated patients and the difference was statistically significant. Worsening of suicidal ideation as indicated by an increase in Hamilton Depression Scale item 3 score tram baseline to the highest score and from baseline to endpoint was statistically significantly more frequent with placebo than with fluaxetine. Although patients with serious suicidal risk were excluded from the trials, passive or active suicidal thinking was observed in 8.2% of the patients at baseline. During the tria!s, emergence of substantial suicidal ideation (Hamilton Depression Scale item 3 score increase from 0 or 1 to 3 or %.!was observed in 1.7% of flnaxetinetreated patients and 3.6% of placebo-treated patients, a nonsignificant difference. The rates of suicidal acts (Oc,j observed in the fluoxetine OCD comrolled clinical trials are lower than those previously reported during longer-term evaluation in patients with OCD. Gittleson (1966a-d) retrospectively evaluated 398 patients hospitalized for severe depression. The group of patients Gittleson studied included 52 patients who had frank obsessional neuroses prior to the depression. Using current diagnostic categories, these patients would likely be characterized as having OCD with secondary depression. In these patients with probable primary OCD, the overall rate of suicidal attempts was 13.5% (7 Of 52) compared with a rate of 29.8% (103 of 346) in those with primary depressive illness. The rate of suicidal ideation was S6% in bath groups. Among those who had suicidal ideation, the rate of suicidal acts was 27% (7 of 26) in the group with probable primary OCD compared with a rate of .52% (103 of I97) in those with primary depressive illness. Patients, regardless of primary diagnosis, who had obsessions during depression had a statisti-
tally significantly lower rate of suicidal acts than those without obsessions (5.7%, 7 of 124 VS. 37.6%, 95 of 253 with one fatality), bul the effect was not sustained in patients whose obsessions converted to delusions (38,1%, 8 of 21). In the follow-up period, age of the patients with obsessions attempted suicide (0.8%, 1 of 124) compared with 18 of the patients without obsessions (7.1%, 18 of 253, eight fatalities). Rosenberg (1968) studied a group of patients whose diagnoses closely resembled today’s categories of OCD (n = 144) and panic disorder (n = 144). A subset of patients in each group had cor.current moderate to severe depression (OCD, 49 of 144; panic disorder, 37 of 144). Most of those in the OCD subgroup had received treatment for the obsessional symptoms for some years prior to the onset of depression. In the OCD group, nine patients had made suicide attempts, with no known fatalities (18.4%, 9 of 49); in the panic disorder group, 21 patients had made attempts, with three fatalities (56.8%, 21 of 37). The author concluded that although depression is a common complication of obsessional neurosis, the risk of suicide is small. Goodwiil et al. (1969), who reviewed long-term follow-up studies in patients with OCD, found that in most studies, less than 1% of patients with OCD committed suicide, despite the frequency of suicidal content in the obsessional thinking. Coryell (19811, who retrospectively reviewed University of Iowa Psychiatric Hospital records, identified 44 patients with primary DSM-III OCD with onset before 40 years of age. He also identified a control group of 44 patients with primary unipolar depression, matched for sex, age, and year of admission. On follow-up information was located for 39 patients in each group. Thirteen of the OCD patients had died (11 death certificates available) and 17 of the depressed patients had died (16 death certificates available). Of the 11 deaths in OCD patients for whom information was available, there were no suicides. Of the 16 deaths in the depressed patients for whom information was available, there were four suicides. In addition, one depressed patient had died by drowning and one of exhaustion due to mania. These data indicate that suicide is higher in depression than in OCD.
In conclusion, the overali incidence of suicidality in patients with OCD treated with fluoxetine in US IND controlled clinical trials was low, compared favorably with rates in corresponding placebo-treated patients, and was well within the range of estimates in previous studies of patients with OCD. The exclusion of comorbid axis I disorders, especially depressive mood disorders, substance abuse disorders, and schizophrenia, may have been a factor contributing to the low incidence of suicidality with both placebo and fluoxetine. As discussed above, noncomorbid OCD has been found by others to be associated with a relatively low incidence of suicide. Depression, substance abuse, and schizophrenia, however, have been associated with relatively higher incidences of suicide (Black and Winokur, 1990; Buda and Tsuang, 1990; Flavin et al., 1990). These results suggest no undue risk of suicidality in patients with OCD treated with fluoxetine. References American
Psychiatric Association
tistical
Manual
American Black, D.W.
of Mental
Diagnostic
and Winokur, In:
G. (1990)
S.J. Blumenthal
Suicide over
the Life
and Treatment P.. Chaput,
revised.
DC.
Suicide and psychiatric
and
D.J.
Kupfer
(Eds.),
Cycle: Risk Factors, Assessment,
of Suicidal Patients. American
Press, Washington,
and Sta-
3rd edn.
Psychiatric Press, Washington,
diagnosis.
Blier,
(1987)
Disorders,
Psychiatric
DC, pp. 135-153.
Y. and deMontigny,
C. (1988)
Long-term
S-HI’ reuptake blockade, but not monoamine oxidase inhibition, decreases the function of !erminal tors: an electrophysiological Pharmacol. Brent, D.A., The
337, 246-254.
Kupfer, D.J., Bromet. E.J. and Dew, M.A.
assessment and treatment
suicide.In: A.J. Update:
5-HT autorecep-
study in the rat brain. Arch.
of patients
(1988)
at risk
for
Frances and R.E. Hales (Eds.), Psychiatry
American
Psychiatric
view, Vol. 7. American
Annual
Re-
Psychiatric Press, Washington,
Association
DC,
pp. 353-385. Buda,
M.
and Tsuang,
suicide: implications
M.T.
(1990)
The
epidemiology
of
for clinical practice. In: S.J. Blumen-
thal and D.J. Kupfer (Eds.). Suicide Over the Life Cycle: Risk Factors, Assessment, and Treatment tients. American
of Suicidal Pa-
Psychiatric F’ress, Washington,
DC,
pp.
17-37. Charney,
D.S.,
Rasmussen,
Goo&nan, S.A.
W.K..
Price,
and Heninger,
L.H.,
G.R.
Woods,
(1988)
S.W.,
Serotonin
function in obsessive-compulsive disorder: a comparison of the effects of tryptophan and nn-chlorophenylpiperazine patients 177-185.
and healthy subjects. Arrh.
Gen.
in
Psychiatry 45,
9
Zoccaro, E.F. (1989) Central serotonin and impulsive aggression. Br. J. Psychiatry 155 (Suppl. 81, 52-62. Coryell, W. (1981) Obsessive-compulsive disorder and primary unipolar depression: comparisons of background, family history, course, and mortality. J. Nerv. Ment. Dis. 169, 220-224. Fava, M. and Rosenbaum, J.F. (1991) Suicidality and fluoxetine: is there a relationship? J. Clin. Psychiatry 52 108-I 11. Flavin, D.K., Franklin Jr., J.E. and Frances, R.J. (1990) Substance at%e and suicidal behavior. In: S.J. Blumenthal and D.J. Kupfer (Eds.), Suicide Over the Life Cycle: Risk Factors, Assessment, and Treatment of Suicidal Patients. American Psychiatric Press, Washington, DC, pp. 177-204. Fuller, R.W., Snoddy. H.D., Perry, K.W.. Bymaster, F.P. and Wow, D.T. (1978) Importance of duration of drug action in the antagonism of p-chloroamphetamine depletion of brain serotonin - comparison of fluoxetine and chlorimipramine. Biochem. Pharmacol. 27, 193-198. Gittleson, N.L. (1966a) The effect of obsessions on depressive psychosis. Br. J. Psychiatry ! 12, 2X-259. Gittleson, N.L. (1966b) The phenomenalogy of obsessions in depressive psychosis. Br. J. Psychiatry 112, 261-264. Gittleson, N.L. (I966c) Depressive psychosis in the obsessional neurotic. Br. J. Psychiatry 112, 883-887. Gittleson, N.L. (1966d) The relationship between obsessions and suicidal attempts in depressive psychosis. Br. J. Psychiatry 112, 889-890. Goodman, W.K., Price, L.H., Rasmussen, S.A., Mazure, C., Fleischmann, R.L., Hill. C.L.. Heninger. G.R. and Charney, D.S. (1989a) The Yale-Brown obsessive-compulsive scale, I: development, use, and reliability. Arch. Gen. Psychiatry 46. 1006-1011. Goodman, W.K.. Price, L.H., Rasmussen, S.A., Mazure. C., Delgado. P., Heninger. C,.R. and Charney, D.S. (1989b) The Yale-Brown obsessive-compulsive scale, II: validity. Aich. Gen. Psychiatry 46, 1012-1016. Goodwin. D.W.. Guze, S.B. and Robins, E. (1969) Follow-up studies in obsessional neurosis. Arch. Gen. Psychiatry 20, 182 187. Guze, S.B. and Robins, E. (1970) Suicide and primary affective disorders. Br. J. Psychiatw 117, 437-438. Hamilton, M. (1967) Development of a rating scale for primary depressive illness. Br. J. Sot. Clin. Psychol. 6, 278296. Hirschfeld, R.M.A. and Davidson, L. 11988) Pisk f&ors for suicide. In: A.J. Frances and R.E. Hales (Ed-.), Psychiatry Update: American Psychiatric Association Annual Review, Vol. 7. American Psychiatric Press, Washington, DC, pp. 307-333. Insel, T.R., Mueller III, E.A., Gillrn, J.C., Siever, L.J. and Murphy, D.L. (1984) Biological m:qrkers in obsessive-compulsive and affective disorders. J. Psychiatr. Res. 18, 407423. Insel, T.R., Mueller, E.A., Aiterman, I., Linnoila, M. and Murphy, D.L. (1985) Obsessive-compulsive disorder and serotonin: is there a connection? Biol. Psychiatry 20, 1174-I 188. Jenike, M.A., Buttolph. L., Baer, L., Ricciardi, J. and Hol-
land, A. (1989) Open trial of fluoxetine in obsessive-cornpulsive disorder. Am. J. Psychiatry 146, 909-911. Karno. M., Gelding, J.M., Sorenson, S.B. and Burnax, MA. (1988)The epidemiology of obsessive-compulsive disorder in five US communities. Arch. Gen. Psychiatry 45, 10941099. Kim, S.W. and Dysken, M.W. (1990) Open fixed dose trial of fluoxetine in the lreatment of obsessive compulsive disorder. Drug Dev. Res. 19,315-319. Kostowsk, W., Valzelli, L., Kozak, W. and Bernasconi, S. (1984) Activity of desipramine, fluoxetine and nomifensinc on spontaneous and P-CPA-induced muricidal aggression. Pharmacol. Res. Commun. 16, 265-271. Landry. P. and Chouinard, G. (1990) Obsessive-compulsive disorder: pathogenesis and fluoxrtine treatment. Rev. Contemp. Pharmacother. 1, 61-74. Lemberger, L., Rowe, H., Carmichael, R., Crabtree, R., Horng, J.S., Bymaster, F. and Wong, D. (19788) Fluoxetine, a selective serotonin uptake inhibitor. Clin. Pharmacol. Ther. 23, 42 l-429. Lemberger, L., Cowe, H., Carmichael, R., Oldham, S., Horng, J.S.. Bymaster, F.P. and Wong, D.T. (1978b) Pharmacologic effects in man of a specific serotonin-reuptake inhibitor. Science 199, 436-437. Levine, R., Hoffman, J.S., Day Knepple, E. and Kenin, M. (1989) Long-term fluoxetine treatment of a large number of obsessive-compulsive patients. J. Clin. Psychopharmacol. 9, 281-283. Masand, P., Gupta, S. and Dewan, M. (1991) Suicide: ideation related to fluoxetim .reatment (letter). New Engl. J. Med. 324. 420. Molina. V., Ciesielski, L., Gobaille, S., Isel, F. and Mandel, P. (1987) Inhibition of moue-e killing behavior by serotoninmimetic drugs: effects of partial alterations of serotonin neurotransmission. Pharmacol. Biochem. Behrv. 27, 123131. Papp, L.A. and Gorman, J.M. (1990) Suicidal preoccupation during fluoxetine treatment (letter). Am. J. Psychiatry 147, 1380. Perse, T. (1988) Obsessive-compulsive disorder: a treatment review. J. Clin. Psychiatry 49, 48-55. Pigott, T.A., Pato, M.T., Bernstein, S.E., Grover, G.N.,‘Hill, J.L., Tolliver, T.J. and Murphy, D.L. (1990) Controlled comparisons of clomipramine and fluoxetine in the treatment of obsessive-compulsive disorder: behavioral and biological results. Arch. Gen. Psychiatry 47, 926-932. Rachman. S.J. and Hodgson, R.J. (1980) Obsessions, cOmPulsions. and depression. In: Obsessions and Compulsions. Prentice-Hall, Englewood Cliffs, NJ, pp. 69-86. Rich&on. E. and Nelson, A. (1984) Antagonism by antidepressants of neurotransmitter receptors of normal human brain in vitro. J. Pharmacol. Exp. Ther. 230, 94-102. Richejson, E. and Pfenning. M. (1984) Blockade by antidepressants and related compounds of biogenic amine uptake into rat brain synaptosomes: most antidepressants selectively block norepinephine uptake. Eur. J. Pharmacol. 104. 277-286. Roose, S.P. and Glassman, A.H. (1988) Delusional depres-
10 sion. In: A. Georgotas and R. Cancro (Eds.1, Depression and Mania. Elsevier. New York, pp. 76-85. Rosenberg, CM. (1968) Complications cf obsessional neurosis. Br. 3. Psychiatry 114. 477-478. SAS Institute Inc. (1989) SAS/STAT& User’s Guide, Version 6, 6th edn. Vol. 2. SAS Institute Inc., Cary, NC, pp. 1072-1121. Stark, P. and Hardison, C.D. (1985) A review of multicenter controlled studies of fluoxerrne vs. imipramine and placebo in outpatients with major depressive disorder. .I. Clin. Psychiatry 46, 53-58. Talbott, M.W., Hadley, P.A., Lister, W.C., Rader. M., Hizer, R.E., Pettinga, C.W. and Thompson, W.L. (1987) Adverse drug events: global collection, analysis, and reporting. J. Clin. Res. Drug Dev. 1, 53-73. Teicher. M.H., Glod, C. and Cole, J.O. (1990a) Emergence of intense suicidal preoccupation during fluoxetine treatment. Am. J. Psychiatry 147, 207-210. Teicher. M.H.. Glad, C. and Cole, J.O. (i’iio’o) Suicidal preoccupation during fluoxetine treatment. Dr. Teicher and associates reply (letter). Am. J. Psychiatry 147. 13801381.
Turner, S.M., Beidel, D.C. and Nathan, R.S. (1985) Biological factors in obsessive-compulsive disorders. Psychol. Bull. 97, 430-450. Valzelli, L. (1984) Reflections on experimental and human pathology of aggression. Prog. Neuropsychopharmacol. Biol. Psych atry 8, 31? -325. Wander, T., Nelson, A., Okazaki, H. and Richelson, E. (1986) Antagonism by antidepressants of serotonin S1 and S, receptors of normal human brain in vitro. Eur. J. Pharmacol. 132, 115-1~1. Welner, S.A.. DeMontigny, C.. Desroches, J., Desjardins, P. and Suranyi-Cadotte. B.E. (lY89) Autoradiographic quantificatton of serotomn,, receptors in rat brain following antidepressant drug treatment. Synapse 4, 347-352. Wernicke. J.F., Dunlop. S.R., Dornseif, B.E. and Zerbe, R.L. (1987) Fixed-dose fluoxetine therapy for depression. Psychopharmaco!. Bull. 23, 164-168. Wernicke. J.F.. Dunlop, S.R.. Dornueif, BE., Eosomworth, J.C. and Humbert, M. (1988) Low-dose fluoxetine therapy for depression. Psychopharmacol. Bull. 24. 183-188.