A Double-blind, Placebo-Controlled Study of Risperidone Addition in Serotonin Reuptake Inhibitor–Refractory Obsessive-compulsive Disorder FREE

SEROTONIN REUPTAKE inhibitors (SRIs) are the most effective monotherapy currently available for obsessive-compulsive disorder (OCD).¹ To date, 2 major approaches have been taken in the development of pharmacological treatments for the SRI-refractory patient. Controlled studies of the first approach, adding drugs that further enhance serotonin (5-hydroxytryptamine [5-HT]) function, such as lithium,² buspirone,³ and clonazepam,⁴ have yielded discouraging results. The second approach has involved adding low-dose dopamine antagonists, such as pimozide⁵ and haloperidol,⁶ to the treatment of SRI-refractory patients. This strategy has been shown to be effective for reducing OCD symptoms primarily in patients who have a comorbid chronic tic disorder, such as Tourette's disorder^5- 6 or schizotypal personality disorder (SPD).⁵ Because of the potential for significant acute and long-term adverse effects with conventional dopamine antagonists, however, safer treatments are needed for patients who may require long-term pharmacotherapy.

Multiple open-label reports have suggested that adding risperidone, an atypical antipsychotic and potent 5-HT_2A/D₂ antagonist, may be effective for reducing OCD symptoms in patients unimproved after adequate SRI monotherapy.^7- 11 This double-blind, placebo-controlled study was designed to determine the short-term efficacy and tolerability of potent SRIs in combination with risperidone in treating OCD patients refractory to SRIs alone. It was hypothesized that OCD patients with a concurrent chronic tic disorder or SPD would be preferentially responsive. It was also hypothesized that the acute extrapyramidal and other adverse effects associated with risperidone would be only minimally greater than those of placebo.

Seventy patients gave written informed consent for participation in this study, which had been approved by the Human Investigation Committee of Yale University School of Medicine, New Haven, Conn. A semistructured interview modeled after the Schedule for Affective Disorders and Schizophrenia was administered to ascertain that all patients met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV)¹² criteria for OCD. Only patients with OCD symptoms of at least 1 year's duration and of at least moderate severity on the Clinical Global Impression Scale (CGI)¹³ were included. Patients were categorized as "depressed" (secondary to OCD) if they met DSM-IV criteria for major depressive disorder or received a baseline score on the modified 19-item Hamilton Rating Scale for Depression (HAM-D)¹⁴ of greater than 20. Relevant portions of the Schedule for Tourette's Syndrome and Other Behavioral Syndromes¹⁵ modified for DSM-IV were used to diagnose tic disorders. Patients with chronic tics (motor and/or vocal) were included, provided that OCD symptoms, rather than tics, constituted the chief complaint. The Schedule for Schizotypal Personalities¹⁶ and DSM-IV criteria were used to make diagnoses of SPD. Each subject and each subject's available first-degree relatives were also asked to report on each of their first-degree relatives' psychiatric symptoms using a semistructured format. All diagnostic instruments were administered by the same experienced clinician (S.W.) who had attained satisfactory interrater reliability with other diagnosticians within our research group. Final diagnoses for patients were made independently following the best-estimate diagnostic procedures developed by Leckman et al.¹⁷

Patients were healthy based on results of physical examination, electrocardiogram, and screening tests of blood and urine. Females had negative results on serum human chorionic gonadotropin testing. Patients were free of psychotropic medications for at least 4 weeks before starting the study.

Seventy patients received 12 weeks of open-label SRI monotherapy. The choice of SRI was based on each patient's prior SRI treatment. The SRI therapy was started at a dose of 50 mg/d (clomipramine, fluvoxamine, sertraline) or 20 mg/d (fluoxetine, paroxetine) and increased to a maximum tolerated daily dose of clomipramine (250 mg/d), fluvoxamine (300 mg/d), sertraline (200 mg/d), fluoxetine (80 mg/d), and paroxetine (60 mg/d). Clomipramine and fluvoxamine were increased by 50 mg every 4 to 5 days; sertraline was increased by 50 mg/d each week and fluoxetine and paroxetine were increased by 20 mg/d each week. Thus, patients received the maximum tolerated dose of the SRI for at least 8 weeks before being considered for the risperidone-addition phase.

Thirty-four of the 70 patients did not enter the risperidone-addition phase. Of these, 23 responded to SRI alone, 7 had treatment-limiting adverse effects to the SRI, and 4 were noncompliant with treatment.

Thirty-six patients were refractory to SRI monotherapy and entered the 6-week controlled risperidone-addition phase. Criteria for refractoriness to the SRI included (1) less than 35% improvement on the Yale-Brown Obsessive Compulsive Scale (Y-BOCS)^18- 19 or a Y-BOCS score of greater than 16, (2) no better than "minimally improved" on the CGI global improvement item, and (3) consensus of the treating clinician and 2 of the authors (C.J.M. and C.N.E.) that the patient's condition was unimproved (based on direct clinical interview and review of behavioral ratings). Only patients who met all 3 criteria were considered refractory.

The SRI dose was maintained during the risperidone-addition phase. Assignment to risperidone (1 mg per capsule) or placebo was done randomly from a computer-generated list, with patients, treating staff (including prescribing physicians), and the rater blind to assignment. Patients were given risperidone, 1 mg/d for 7 days, with the dosage then increased by 1 mg every week, to a maximum of 6 mg/d, as tolerated. No other drugs were administered and patients were not instructed in formal behavior therapy techniques. The clinical characteristics of each treatment group are summarized in Table 1.

After completing the controlled phase of the study, placebo-treated patients received open-label risperidone addition to SRI for 6 weeks.

Patients were assessed weekly by the same trained blind rater (S.W.). Obsessive-compulsive symptoms were measured with the Y-BOCS. Depression was rated with a 19-item version of the HAM-D, which excludes an item for rating OCD symptoms. Anxiety was measured with the Hamilton Rating Scale for Anxiety (HAM-A).²⁰ The frequency dimension of the Yale Global Tic Severity Scale²¹ (rated from 0 = none to 5 = always) was used to measure change in tic frequency for those patients with comorbid tics. A clinician rating of global improvement was made with the CGI (7 = "very much worse," 4 = "no change," 1 = "very much improved"), compared with the prerandomization baseline.

Criteria for response to the SRI-risperidone combination included (1) 35% or greater improvement on the Y-BOCS from the beginning of the risperidone-addition trial and a final Y-BOCS score of 16 or less, (2) a final CGI rating of "much improved" or "very much improved," and (3) consensus of the treating clinician and 2 of the primary investigators (C.J.M. and C.N.E.) that the patient's condition was improved (based on direct clinical interview and behavioral ratings). Patients who met all 3 criteria were classified as "marked" responders, those who met two thirds of the criteria were "partial" responders, and those who met less than 2 of the criteria were nonresponders.

Blood samples were collected before and after the 6-week double-blind trial to determine SRI blood levels. Blood levels of SRIs were determined by high-performance liquid chromatography with fluorescence detection by a commercial laboratory. A commercially available assay for fluvoxamine was not available at the time of the study.

Sitting and standing blood pressure and pulse, temperature, and respiratory rate were recorded at baseline and at the end of weeks 1 through 6 of the controlled trial. Each patient was examined for extrapyramidal (abnormal gait, ataxia, dystonia, hyperkinesia, hypertonia, hypokinesia, involuntary muscle contractions, oculogyric crisis, and tremor) and other adverse effects (agitation, blurry vision, constipation, coughing, diaphoresis, diarrhea, dizziness, dry mouth, dyspepsia, enuresis, gynecomastia, headache, increased appetite, insomnia, lightheadedness, menstrual pattern changes [females], muscle stiffness, nausea, palpitations, restlessness, rhinitis, sedation, sialorrhea, tinnitus, urinary urgency, and vomiting) at baseline and at the end of weeks 1 through 6 of the controlled study.

Thirty-three of the 36 refractory patients completed the study and were included in the efficacy analyses. One patient (risperidone group) dropped out prior to completing the first week of treatment owing to intolerable insomnia. Two others (1 receiving risperidone, 1 receiving placebo) were noncompliant with the protocol and dropped out prior to completing the first week of treatment. Only baseline ratings were recorded for these 3 patients. Baseline ratings were obtained after 12 weeks of SRI treatment. Student t tests were calculated to determine baseline differences in rating scale measures, age, age of onset and duration of OCD, and number of prior adequate medication and SRI trials. The χ² test with Yates correction was used to determine differences in sex distribution, treatment setting, and previous behavior therapy between groups.

Testing with the Shapiro-Wilks statistic showed that Y-BOCS scores could be approximated by the normal distribution at all assessment time points in both the risperidone and placebo groups. Two-way analysis of variance (ANOVA) with repeated measures was calculated for the 6 weeks of risperidone or placebo addition to assess the main effects of drug, time, and drug × time interactions for the behavioral ratings. Analysis of covariance (ANCOVA), using baseline score as a covariate, was used to determine when significant risperidone-placebo differences took place on the CGI and Y-BOCS.

The χ² test with Yates correction was used to compare the 2 groups for rate of responders (marked or partial) vs nonresponders, and to determine if response was related to sex or treatment setting. The relationship between response vs nonresponse to risperidone and the presence or absence of a comorbid chronic tic disorder or SPD was determined by the Fisher exact test. Paired t tests were used to determine change in behavioral ratings (baseline to end of week 6) for patients given open-label risperidone. Fourteen of these 16 patients completed the 6-week open-label risperidone phase. One patient was noncompliant and the other moved to another country prior to entering this phase of the study. Data are presented as mean ± SD unless otherwise indicated, and are reported as significant when P<.05 (2 tailed).

Twenty patients were randomized to risperidone and 16 to placebo. The 2 groups did not differ significantly in treatment setting, age, sex distribution (although the power may be too low for this particular χ² test), age at onset of OCD, duration of OCD, comorbid secondary major depression, chronic tic disorders, or SPD, number of previous medication trials, number of previous SRI trials, previous behavior therapy, or baseline Y-BOCS, HAM-D, and HAM-A scores (Table 1 and Table 2). The particular SRI used to treat each patient appears in Table 2. There was no significant difference in dosage between patients randomized to risperidone (2.2 ± 0.7 mg/d) and placebo (2.6 ± 0.8 mg/d) (Table 2).

Ratings on the CGI global improvement item showed risperidone addition superior to placebo (F_1,31 = 7.65, P<.001) (Figure 1). The ANCOVA demonstrated that risperidone had significant superiority beginning at week 3 and continuing at weeks 4, 5, and 6 (Figure 1). The ANOVA demonstrated a significant degree of global improvement over time in the risperidone group (F_1,17 = 12.86, P<.001) but not in the placebo group.

Nine (50%) of 18 risperidone-treated patients responded (4 marked response, 5 partial response) (Figure 2). None of 15 placebo-treated patients were categorized as responders (χ² = 8.0, P<.005) (Figure 2).

Seven (50%) of 14 patients who received 6 weeks of open-label risperidone addition following the double-blind placebo phase were categorized as responders (4 marked response, 3 partial response).

As measured by reduction in total Y-BOCS scores, risperidone addition was superior to placebo (Figure 3). The ANOVA revealed significant interaction effects of drug × time (F_1,31 = 6.59, P<.001) for the risperidone- vs placebo-addition phase. In comparing the effects of risperidone and placebo on OCD symptoms, ANCOVA showed that risperidone was significantly better than placebo beginning at week 5 and continuing at week 6 (Figure 3). The ANOVA showed a highly significant decrease in total Y-BOCS scores from baseline in the risperidone group (31.8%; 27.4 ± 5.4 to 18.7 ± 8.3) (F_1,17 = 14.61, P<.001). In responders, Y-BOCS scores decreased by 51.6%, from a baseline score of 27.5 ± 3.8 to a final score of 13.3 ± 6.0, following 6 weeks of risperidone. In contrast, ANOVA showed no significant change in Y-BOCS scores for any week following baseline in the placebo group (27.6 ± 3.7 to 25.0 ± 4.4) (Figure 3).

As measured by reduction in total HAM-D scores, risperidone addition was superior to placebo for reducing depressive symptoms. The ANOVA revealed significant interaction effects of drug × time (F_1,31 = 5.29, P<.001) for the risperidone- vs placebo-addition phase. The ANOVA showed a significant decrease in HAM-D scores from baseline (11.7 ± 7.8 to 7.6 ± 4.9) with risperidone (F_1,17 = 4.16, P = .002). No significant change in HAM-D scores occurred for any week following baseline in the placebo group (12.2 ± 5.0 to 15.2 ± 8.7).

As determined by reduction in total HAM-A scores, risperidone addition was also significantly better than placebo for reducing symptoms of anxiety. The ANOVA revealed significant interaction effects of drug × time (F_1,31 = 3.78, P = .003) for the controlled addition phase. The ANOVA demonstrated a significant decrease in HAM-A scores from baseline (13.2 ± 5.4 to 9.1 ± 4.3) with risperidone (F_1,17 = 3.52, P = .007) but not placebo (13.5 ± 6.0 to 15.4 ± 6.9).

For the 14 patients who received 6 weeks of open-label risperidone addition, the drug resulted in significant reductions in total Y-BOCS (24.8 ± 4.5 to 18.3 ± 6.2, t = 3.2, P = .007) and HAM-A (15.4 ± 7.9 to 11.4 ± 6.3, t = 2.3, P = .04) scores, but not HAM-D (13.4 ± 6.3 to 10.9 ± 7.1) scores.

Three of 5 patients with a comorbid chronic tic disorder (Tourette's disorder, n = 2; chronic motor tic disorder, n = 1) were responders (1 marked response, 2 partial response) to the double-blind addition of risperidone, whereas 6 of 13 patients without tics (no known family history of tics) responded (Fisher exact test, P = 1.0). Three of 5 patients with comorbid SPD were responders (1 marked response, 2 partial response) to risperidone, whereas 6 of 13 patients without SPD responded (Fisher exact test, P = 1.0).

For the 14 patients treated with 6 weeks of open-label risperidone addition, 2 of 4 patients with a comorbid chronic tic disorder (Tourette's disorder, n = 2) were responders (both marked response), whereas 5 of 10 patients without tics (no known family history of tics) responded (Fisher exact test, P = 1.0). The 1 patient with comorbid SPD was a nonresponder.

For the 5 patients with comorbid chronic tics who completed the double-blind risperidone addition phase, a significant reduction in tic frequency was determined (1.8 ± 0.4 to 1.0 ± 1.2, t = 4.0, P = .02) with the Yale Global Tic Severity Scale.

The SRI/risperidone treatment was generally well-tolerated. During the controlled phase, 18 of 20 risperidone-treated patients experienced at least 1 adverse effect. These included constipation (n = 1), diaphoresis (n = 1), dry mouth (n = 5), increased appetite (n = 6), insomnia (n = 1), lightheadedness (n = 1), restlessness (n = 6), sedation (n = 17), tinnitus (n = 2), and urinary urgency (n = 1). No clinically significant changes in blood pressure, heart rate, respiratory rate, or temperature were recorded, and no acute extrapyramidal effects, seizures, or cardiac events occurred. Anticholinergic agents or β-blockers were not required for any patient for risperidone-related extrapyramidal symptoms. The most prominent adverse effect was initial, mild transient sedation. Fifteen of 16 patients experienced at least 1 adverse effect while receiving placebo. These included blurry vision (n = 2), diaphoresis (n = 4), diarrhea (n = 1), dry mouth (n = 5), headache (n = 5), increased appetite (n = 3), insomnia (n = 1), lightheadedness (n = 4), muscle stiffness (n = 1), palpitations (n = 1), restlessness (n = 6), sedation (n = 8), and tinnitus (n = 1).

Complete SRI blood level data were available for 4 risperidone-treated patients (clomipramine [n = 2], fluoxetine [n = 2]) and 7 placebo-treated patients (clomipramine [n = 4], fluoxetine [n = 2], sertraline [n = 1]). In the risperidone group, clomipramine plus desmethylclomipramine levels changed from 450 ng/mL to 348 ng/mL and 514 ng/mL to 333 ng/mL, respectively; and fluoxetine plus norfluoxetine levels changed from 880 µg/L to 1049 µg/L and 841 µg/L to 845 µg/L, respectively. In the placebo group, clomipramine plus desmethylclomipramine levels changed from 461 ng/mL to 434 ng/mL, 558 ng/mL to 514 ng/mL, 424 ng/mL to 496 ng/mL, and 484 ng/mL to 297 ng/mL, respectively; fluoxetine plus norfluoxetine levels changed from 293 µg/L to 408 µg/L and 799 µg/L to 805 µg/L, respectively; and sertraline levels changed from 150 ng/mL to 308 ng/mL.

These results demonstrate that the addition of risperidone to the treatment of OCD patients refractory to SRI alone is significantly more effective than placebo for reducing OCD, depressive, and anxiety symptoms. Consistent with a previous study,²² risperidone was also effective for reducing tics in patients with comorbid tics. Contrary to our hypothesis and results from previous studies of haloperidol⁶ and pimozide⁵ addition in fluvoxamine-refractory OCD, no significant difference in treatment response was found between OCD patients with and without comorbid chronic tic disorder or SPD. The ability to detect significant differences between these groups, however, may have been limited by the relatively small sample sizes of patients with such comorbidity. The broader effects of risperidone across these diagnostic subtypes of OCD, compared with haloperidol and pimozide, may be due to the drug's potent 5-HT_2A antagonism, in addition to D₂ blockade, which is particularly relevant at the relatively low dosages required in this study.²³ Risperidone's affinity in vitro is 20 times higher for 5-HT_2A than for D₂ receptors; in vivo it occupies 5-HT_2A receptors at a dose 10 times lower than it does D₂ receptors.²³ Moreover, unlike haloperidol, risperidone has been shown to dose-dependently increase extracellular concentrations of 5-HT in rat frontal cortex.²⁴

The design of this study does not exclude the possibility that treatment response was due to risperidone alone. Based on a review of the literature on atypical antipsychotics,²⁵ and on our previous investigation of clozapine monotherapy in refractory OCD,²⁶ this would not be anticipated. Considering prior reports, one might have hypothesized that an atypical antipsychotic (eg, clozapine,²⁷ risperidone,²⁸ olanzapine²⁹) would induce or exacerbate OCD symptoms. However, as we have previously commented,³⁰ all of these descriptions of clinical induction or worsening of OCD symptoms occurred in patients diagnosed with a psychotic disorder, rather than primary OCD.

That treatment response was due to a longer duration of SRI administration is unlikely, as none of the 15 placebo-treated patients responded. Although the sample sizes were small, no significant differences in SRI blood levels were found between the risperidone- and placebo-treated groups. Also, blood levels did not seem to be related to treatment response. Thus, it is unlikely that risperidone exerted a therapeutic effect through a pharmacokinetic interaction.

A balance of full 5-HT_2A receptor occupancy and partial D₂ receptor occupancy has been hypothesized to underlie risperidone's lower tendency to cause extrapyramidal effects compared with typical neuroleptics.²³ In the present study, risperidone was generally well-tolerated, with no evidence that acute extrapyramidal effects differed between drug and placebo. The most common adverse effect was mild, transient sedation, which resolved early in the treatment course. This adverse effects profile is in contrast to our experience with haloperidol addition in fluvoxamine-refractory OCD.⁶ In that study, 29% of haloperidol-treated patients required adjunctive propranolol for akathisia despite being prophylactically treated with benztropine. No patient required propranolol or anticholinergic agents in the current investigation. While this study does not address the potential long-term adverse effects of risperidone, only 2 (0.2%) of 882 patients with schizophrenia exposed to more than 12 weeks of risperidone have developed tardive dyskinesia (written communication, Martin Brecher, MD, 1999). Nonetheless, because of reports of possible tardive dyskinesia with risperidone,³¹ continued close monitoring of the drug and longer-term follow-up studies are warranted. Additional research is necessary to replicate these findings and to investigate this treatment strategy in children and adolescents with SRI-refractory OCD.

Accepted for publication March 24, 2000.

This work was supported in part by grant PO1 MH49351 (J. F. Leckman, MD) from the US Public Health Service, Bethesda, Md; Young Investigator Awards from the National Alliance for Research in Schizophrenia and Depression, Chicago, Ill (Drs McDougle, Epperson, and Pelton); Theodore and Vada Stanley Foundation Research Awards Program, Arlington, Va (Drs McDougle and Price); and the State of Connecticut, Department of Mental Health and Addiction Services, Hartford.

Elizabeth Kyle, AS, prepared the manuscript, Sally J. Vegso, MS, performed the statistical analyses, and Elizabeth M. Ruff and Krista Guenin, BA, constructed the graphics.

Reprints: Christopher J. McDougle, MD, Indiana University School of Medicine, Section of Child and Adolescent Psychiatry, James Whitcomb Riley Hospital for Children, 702 Barnhill Dr, Room 3701, Indianapolis, IN 46202-5200 (e-mail: cmcdougl@iupui.edu).