|
|
| |||||||||||||||||||||||||||||||||||||||||||||||
Background Drug-induced psychosis is a difficult problem to manage in patients with Parkinson's disease. Multiple open-label studies have reported that treatment with clozapine at low doses ameliorates psychosis without worsening parkinsonism.
Methods We conducted a randomized, double-blind, placebo-controlled trial of low doses of clozapine (6.25 to 50 mg per day) in 60 patients at six sites over a period of 14 months. The patients (mean age, 72 years) had idiopathic Parkinson's disease and drug-induced psychosis of at least four weeks' duration. All the patients continued to receive fixed doses of antiparkinsonian drugs during the four weeks of the trial. Blood counts were monitored weekly in all the patients.
Results The mean dose of clozapine was 24.7 mg per day. The patients in the clozapine group had significantly more improvement than those in the placebo group in all three of the measures used to determine the severity of psychosis. The mean (±SE) scores on the Clinical Global Impression Scale improved by 1.6±0.3 points for the patients receiving clozapine, as compared with 0.5±0.2 point for those receiving placebo (P<0.001). The score on the Brief Psychiatric Rating Scale improved by 9.3±1.5 points for the patients receiving clozapine, as compared with 2.6±1.3 points for those receiving placebo (P=0.002). The score on the Scale for the Assessment of Positive Symptoms improved by 11.8±2.0 points for the patients receiving clozapine, as compared with 3.8±1.9 points for those receiving placebo (P=0.01). Seven patients treated with clozapine had an improvement of at least three points on the seven-point Clinical Global Impression Scale, as compared with only one patient given placebo. Clozapine treatment improved tremor and had no deleterious effect on the severity of parkinsonism. In one patient, clozapine was discontinued because of leukopenia.
Conclusions Clozapine, at daily doses of 50 mg or less, is safe and significantly improves drug-induced psychosis without worsening parkinsonism.
The introduction of the "atypical" antipsychotic drugs — those with few or no extrapyramidal side effects12 — raised the possibility of treating drug-induced psychosis without reducing motor function in patients with Parkinson's disease. Clozapine is virtually free of extrapyramidal side effects.13 It causes potentially fatal agranulocytosis14 in 1 to 2 percent of those exposed, however, and the Food and Drug Administration has mandated weekly blood-count monitoring of patients taking this drug. Such monitoring has made fatal agranulocytosis a rare event.15 Since the first report of the successful use of clozapine in treating drug-induced psychosis in patients with Parkinson's disease,16 there have been confirmatory reports of its effectiveness, but all such reports were from open-label studies.17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39 These studies, involving a total of more than 400 patients, reported significant benefit in approximately 85 percent of patients using 1/10 or less of the usual dosage of clozapine prescribed for schizophrenia. Open-label studies of the newer atypical antipsychotic drugs risperidone and olanzapine in patients with Parkinson's disease have yielded conflicting results,40,41,42,43,44,45 with clear worsening of motor signs in some patients.40,41,43,44 We report the results of a multicenter, randomized, placebo-controlled, double-blind study of clozapine for the treatment of drug-induced psychosis in patients with Parkinson's disease.
Methods
Patients
Patients were candidates for the study if they had idiopathic Parkinson's disease and psychosis induced by antiparkinsonian drugs. Parkinson's disease was diagnosed when at least three of the four cardinal features were present (tremor at rest, rigidity, hypokinesia, and problems with posture and balance typical of Parkinson's disease) and there were no alternative explanations for the findings. Psychosis was defined as a major psychiatric illness in which reality testing was impaired, typically by the presence of hallucinations or delusions, leading to substantial communication and social problems.46 Neither age nor the severity of motor dysfunction was a factor in the selection of patients. Special attempts were made to recruit women and members of minority groups. The protocol was approved by the institutional review board at each participating site.
Before entering the study, each patient was treated as deemed most appropriate by the enrolling investigator but without an antipsychotic drug. Patients were excluded if they had taken clozapine or a neuroleptic drug within three months before the start of the study. The doses of antiparkinsonian medications were generally reduced to the lowest level that was considered tolerable for motor function. To prevent the enrollment of patients whose improvement was the result of an adjustment of their antiparkinsonian medication, all the patients had to have been taking a fixed dose of medication for at least seven days before enrollment. The criteria for study entry were psychosis severe enough to have warranted treatment with a standard neuroleptic drug if the patient had not had Parkinson's disease and a score of 3 (indicating mild psychosis) or higher on the Clinical Global Impression Scale47 for the severity of psychosis.
To be included in the study, each patient had to have a documented history of psychosis of at least four weeks' duration before enrollment and had to have a reliable care giver who could accurately report the patient's daily level of function, accompany the patient to each visit, and administer the study drug.
Criteria for exclusion were a history of leukopenia; the presence of any systemic factor that might contribute to a behavioral disorder; therapy with any dopamine-blocking drug within the three months before the study began; therapy with neuroleptic drugs administered in depot form within the year before the study; a change in antidepressant or anxiolytic drugs within the month before the study; previous therapy with clozapine for the treatment of psychosis; the presence of symptomatic orthostatic hypotension, uncontrolled seizures, uncontrolled angina, the acquired immunodeficiency syndrome or another illness that would make the use of clozapine potentially hazardous, or narrow-angle glaucoma; myocardial infarction during the three months before the study; treatment with chemotherapeutic drugs that lower white-cell counts; an inability to tolerate a fixed dose of antiparkinsonian drugs for one month; an inability to tolerate the current level of psychosis for one month; or the presence of dementia severe enough to preclude assessment on the psychiatric-test battery. Women of childbearing potential who were not using reliable forms of contraception were also excluded.
Design of the Study
The study was designed to determine whether clozapine, administered at low doses, is an effective treatment for drug-induced psychosis in patients with Parkinson's disease and to determine its effect on motor function in such patients. The primary outcome measures were the scores on the Clinical Global Impression Scale47 for psychosis and the Unified Parkinson's Disease Rating Scale for parkinsonism.48
If the screening laboratory tests revealed no clinically significant abnormality, the patient or his or her legal guardian was asked to give written informed consent. The patient was then enrolled in the Clozaril Patient Monitoring System administered by Sandoz Pharmaceuticals (now Novartis Pharmaceuticals, Hanover, N.J.) and enrolled in the study. The random assignment of the patients to the clozapine group or the placebo group was stratified according to the site and the age of the patient (under 70 years vs. 70 or older).
The double-blind evaluation lasted four weeks. At base line, a physical examination was conducted and the Mini–Mental State Examination49 was administered. Three measures of psychosis were used: the Clinical Global Impression Scale,47 which rates the severity of psychotic symptoms on a scale from 0 (no symptoms) to 7 (the most severe symptoms); the Scale for the Assessment of Positive Symptoms, in which the score for the severity of symptoms ranges from 0 (no symptoms) to 17650; and the Brief Psychiatric Rating Scale, in which scores can range from 18 to 126, with higher scores indicating more severe psychosis.51 A modified version of the Brief Psychiatric Rating Scale was also used in which the four items measuring tension, mannerisms, posturing, and motor retardation that we considered unreliable because of confounding with motor signs of Parkinson's disease were omitted.
All the patients were evaluated on the psychiatric scales by the site investigator (a neurologist or a psychiatrist). All the investigators were trained in the administration of the tests to patients with Parkinson's disease at a teaching session in which videotaped interviews were used, but no formal study of interrater reliability was performed. The severity of parkinsonism was assessed with the use of sections II and III of the Unified Parkinson's Disease Rating Scale (possible range of scores, 0 to 160),48 including the Step–Seconds Test, the Hoehn–Yahr staging system (stage I indicates newly diagnosed disease, and stages IV and V advanced disease),52 and the Schwab and England Activities of Daily Living scale (range of scores, 0 to 100, with lower scores indicating more severe dysfunction).53 The scores for tremor on the Unified Parkinson's Disease Rating Scale were analyzed in isolation because of reports that clozapine decreases tremor in Parkinson's disease.54,55,56,57,58 Adverse events and changes in sleep patterns were evaluated in detailed questionnaires.
Clozapine (Clozaril) was obtained in bulk from Sandoz Pharmaceuticals. It was ground into powder form and repackaged in 6.25-mg capsules along with matching placebo at the University of Rhode Island School of Pharmacy. At the time of study entry, the patients received a 10-day supply of the assigned medication and started taking one capsule each night. On day 3 or 4, the study coordinator or investigator spoke with the patients' care givers to determine, on the basis of an unstructured conversation carried out over the telephone, the response to treatment. The dose could then be increased or decreased. If there were adverse events before the scheduled phone call, the care givers were asked to contact the study coordinator or investigator. The patients were evaluated weekly in the office, where vital signs (pulse, blood pressure, and temperature), weight, adverse events, and compliance (measured by pill counts) were assessed. Telephone calls to the care givers were made mid-week. The investigators assessed the level of psychosis weekly.
Possible daily doses were as follows: 6.25, 12.5, 18.75, 25, 37.5, and 50 mg. All daily doses started at 6.25 mg and could be raised one level depending on the patient's clinical response; if the patient's daily dose had been increased from the initial 6.25-mg level, it could also be lowered one level. The dose could be changed once between weekly office visits or at the visit. A dose could be increased even if the patient had not been able to tolerate that dose on a previous occasion. The dosage reached at the beginning of the final week was the maximal dose; it could not be increased further but could be decreased, if necessary, because of side effects. Thus, at the final assessment, when all base-line measures were repeated, the patient had been receiving a stable or declining dose of study medicine for at least seven days.
During the study, all the patients had a complete blood count performed weekly. The results were reported to a monitor who was not aware of the patients' treatment assignments. Monitoring guidelines were the same as those for all patients in the Clozaril Patient Monitoring System. Since a request for an additional complete blood count if the white-cell count declined could reveal the treatment assignment to the participants, additional blood tests were also randomly performed. Thus, some patients had additional blood drawn even if the standard weekly complete blood count was normal. The results were available only to the monitor, who remained unaware of the treatment assignments. The study participants were not told whether the additional blood test was triggered by a decline in the white-cell count or was randomly ordered. At the end of the fourth week, all base-line tests were repeated.
Safety
A safety monitoring committee, composed of two physicians and a statistician who did not otherwise participate in the study, met quarterly. This committee reviewed all laboratory data and reports of adverse events; its members had full access to treatment assignments.
Statistical Analysis
Data were sent to the Clinical Trials Coordination Center (Rochester, N.Y.). Forms were reviewed for completeness and accuracy before the data were entered into the data base. The protocol required a sample of 60 subjects in order to have the power to demonstrate that clozapine did not worsen motor function in patients with Parkinson's disease. Unpublished data on drug-induced psychosis in patients with Parkinson's disease that were obtained with use of the Brief Psychiatric Rating Scale indicated that fewer patients would be required in order to demonstrate that clozapine improved drug-induced psychosis. Assuming a standard deviation of six units for the change from base line in the scores on the Unified Parkinson's Disease Rating Scale (items II and III) and a dropout rate of 10 percent, we estimated that a sample of 30 patients per treatment group would give the study a 90 percent power to detect a worsening of five points on the Unified Parkinson's Disease Rating Scale in the group assigned to clozapine (P=0.05 by a two-sided test),59 a decline thought to be clinically significant.
The base-line characteristics of the patients in the two groups were compared by chi-square tests or t-tests, as appropriate. The effect of treatment on each outcome measure was assessed by analysis of covariance, with adjustments made for the base-line value and the investigator. Consistency with respect to possible investigator bias was examined by the inclusion of terms for investigator–treatment interactions.
The intention-to-treat analysis called for in the protocol was modified to account for one patient who was randomly assigned to receive placebo but actually received clozapine. When this error was discovered midway through the trial, we opted to keep giving the patient the active drug. Very similar results with respect to efficacy were obtained regardless of whether this patient was excluded from or included in the group to which she had been randomly assigned. The analyses reported here are based on the treatment the patients received, not on the intention-to-treat analysis, and include data only for those who completed evaluations at base line and four weeks.
Results
Base-Line Characteristics
Sixty patients were enrolled between April 1995 and October 1996, with the numbers recruited varying from 9 to 12 patients per site. All had had hallucinations or delusions. There were some significant imbalances at base line between the groups in the intention-to-treat analysis (the patients receiving clozapine had slightly less severe psychosis than those receiving placebo), but not between the groups in the analysis based on the treatment the patients actually received (Table 1). Fifty-four patients completed the trial.
|
|
Three patients receiving placebo and three receiving clozapine withdrew from the study. The psychiatric condition of two of the three patients receiving placebo worsened. One patient required psychiatric hospitalization, and the other discarded her medications, declaring herself "cured." The third patient was hospitalized for pneumonia. Of the three patients in the clozapine group who withdrew from the study, one discontinued the drug because of leukopenia (white-cell count, 2900 per cubic millimeter), one because of myocardial infarction, and one because of sedation.
There were no significant changes in the mean neutrophil or white-cell counts in either group. There were no significant differences between groups in changes in orthostatic blood pressure, but there was a small but statistically significant increase from base line, of 3.9 beats per minute, in the mean resting heart rate of the patients receiving clozapine as compared with no increase in that of the patients receiving placebo (P=0.046). Weight increased by 0.7 kg (1.6 lb) in the patients receiving clozapine and 0.1 kg (0.27 lb) in those receiving placebo (P=0.005). There were no significant differences between the groups in the severity of drooling, memory impairment, constipation, orthostatic lightheadedness, urinary hesitancy, confusion, headaches, fatigue, or daytime sedation.
Analyses were repeated with the use of the intention-to-treat method, the last-observation-carried-forward method, or both methods; the results of all analyses were similar.
Efficacy
The mean daily dose of clozapine prescribed at the end of the study was 24.7 mg (range, 6.25 to 50). The mean daily dose of placebo was equivalent to 35.2 mg (range, 6.25 to 50).
Improvements were seen in all measures of psychosis in the clozapine group, including the Clinical Global Impression Scale, the Brief Psychiatric Rating Scale (original and modified versions), and the Scale for the Assessment of Positive Symptoms (Table 3). In 13 patients in the clozapine group but in only 3 patients in the placebo group the severity of psychosis decreased to a level that, if present at base line, would have made them ineligible for the study. Table 4 shows the distribution of changes in the scores on the Clinical Global Impression Scale. Hallucinations (almost entirely visual) improved by 1.9 points for the patients receiving clozapine and 0.7 point for those receiving placebo (P=0.002) on the Brief Psychiatric Rating Scale. There were no significant differences between the groups with respect to any of the five items on our sleep questionnaire.
|
|
There were no statistically significant investigator–treatment interactions in any of the outcome measures, indicating consistency of results among investigators. Compliance, as measured by pill counts, was 96 percent in the placebo group and 98 percent in the clozapine group.
Discussion
In this double-blind study, clozapine ameliorated psychosis despite the fact that the patients continued taking antiparkinsonian drugs. Moreover, clozapine did not worsen motor function and actually decreased tremor.
Several aspects of our study design require explanation. We excluded patients whose psychosis had lasted for less than four weeks so as to avoid confounding between improvement due to clozapine and that due to a delayed response to a reduction in the dose of antiparkinsonian medication. Because there are no validated instruments for measuring psychosis in Parkinson's disease, we chose four separate measures of psychotic symptoms. The Clinical Global Impression Scale was the primary tool we used to measure the severity of psychosis. The Brief Psychiatric Rating Scale rates both positive (e.g., presence of hallucinations) and negative (e.g., absence of emotion) symptoms and is valid for assessing the severity of schizophrenia. We also used the Scale for the Assessment of Positive Symptoms, on the basis of numerous reports of drug-induced psychosis consisting of positive symptoms.60,61,62,63,64 We used these measures in combination to buttress the reliability of our results.
The patients in this study were fairly typical of those encountered in clinical practice, although we excluded those with the most severe psychosis because of our concern that four weeks of placebo therapy might jeopardize the patients' safety. We defined improvement of psychosis according to three separate clinical measures, and our results demonstrated both clinically and statistically significant improvement. These results held true in the analysis of the patients according to actual treatment as well as in the intention-to-treat analysis. The improvements were greater than those reported in most trials of antipsychotic drugs in schizophrenia.65,66 The finding that parkinsonism did not worsen in our study confirms the results of open-label studies. The improvement in tremor is commensurate with that in previous studies.54,55,56,57,58 Our results contrast markedly with those of the one previous placebo-controlled trial, which had only six subjects.67 The most obvious explanation for the difference in outcomes is the fact that higher doses were used in that study.
The reduction in the number of white cells in the patients receiving clozapine, necessitating discontinuation of the drug in one patient, is within the range expected on the basis of much larger samples. The complete blood count returned to normal once clozapine was discontinued. In more than 400 patients with Parkinson's disease who were treated with clozapine, there have been only 3 cases of leukopenia, a percentage somewhat smaller than that found in our study.37 Our results covered a period of four weeks, and the greatest risk of agranulocytosis extends through the first four months of treatment. Thus, clozapine can be safely given to elderly patients with Parkinson's disease for at least four weeks, with weekly monitoring of the white-cell count.
Although weekly visits could have revealed adverse effects that might have made the investigator aware of the treatment assignments, the nature and frequency of side effects were similar in the clozapine and placebo groups. Although clozapine has anticholinergic activity in the brain,68 we found that it did not impair memory or cognitive function as measured on the Mini–Mental State Examination. There were no anticholinergic effects, such as urinary retention, dry mouth, blurred vision, or constipation. Somnolence was as frequent in the placebo group as in the clozapine group, and there were no differences between the groups in any of the measures of sleep patterns — namely, the amount of time spent sleeping, the amount of time needed to fall asleep, and the duration of naps during the day.
Other adverse events associated with clozapine in our study were weight gain and an increase in the pulse rate. To our knowledge, weight gain in patients with drug-induced psychosis has not been noted before, although it has occurred in patients with Parkinson's disease who were treated with higher doses of clozapine for tremor.55 Weight gain is not necessarily a problem, since patients with Parkinson's disease commonly lose weight during the course of their illness. Tachycardia was asymptomatic in the patients in our study and was not of clinical importance. Whether this could be a problem for patients with unstable angina needs to be considered. Fever did not occur in our patients.
At the end of the double-blind trial, all the patients were offered treatment for three months with clozapine, provided free of charge. Fifty-three patients were enrolled in this open-label extension of the study. During the three-month extension, the white-cell count dropped to 3000 per cubic millimeter in one patient, who was then withdrawn from the study, after which the count returned to normal. Six patients died, three of whom had been living in nursing homes. Two of these three had advance directives limiting intervention; one died of a stroke and the other of bronchitis. The third nursing home resident died in her sleep of unknown causes. Two other patients died of pneumonia. The sixth patient died of cardiac arrest shortly after the three-month extension of the study ended. No deaths were associated with a low white-cell count. The site investigators did not believe clozapine was related to death in any of the patients. Other drugs were changed (the doses were either increased or decreased, or the drugs were discontinued, or new drugs were started) during this extension phase. We believe the deaths were due to complications of disease in patients who had generally requested that no aggressive interventions should be used to save their lives. Such a high death rate has not been reported before in studies of clozapine, but increased mortality and hallucinations have been reported in patients with Parkinson's disease in nursing homes.69
In summary, we found that clozapine at doses ranging from 6.25 to 50 mg per day, which are far lower than the doses of 300 to 900 mg required for the treatment of schizophrenia, reduced the severity of drug-induced psychosis in patients with Parkinson's disease. Although the most feared adverse effect is agranulocytosis, the incidence of which is reported to increase with age, it did not occur in our study. Leukopenia developed in one patient during the four weeks of clozapine therapy, but it resolved after clozapine was stopped. The fact that the three-month extension of the study had an open-label design makes it difficult to interpret the relation between clozapine and morbidity, since other drugs were changed during this period, but continued study of the morbidity and mortality associated with clozapine is appropriate.
Supported in part by a grant (FD-R-001416-02) from the Orphan Drug Division of the Food and Drug Administration and in part by the Parkinson Study Group.
We are indebted to Mary Eddy for assistance in the preparation of the manuscript; to R.B. Daroff, M.D., and H.Y. Meltzer, M.D., for editorial help; to Michael Dreiss, M.S., of the Food and Drug Administration for financial and administrative advice; to the white-cell count monitors at the study sites: D. Agle, M.D., E. Molho, M.D., P. O'Dowd, M.D., and A. Payne, R.N.; and to Sandoz Pharmaceuticals for providing clozapine.
Source Information
Address reprint requests to Dr. Joseph H. Friedman at the Memorial Hospital of Rhode Island, 111 Brewster St., Pawtucket, RI 02860, or at joseph_friedman{at}mhri.org. Dr. Friedman assumes responsibility for the overall content and integrity of the manuscript.
References
The following persons participated in the study: Investigators — J. Friedman and M. Lannon, Memorial Hospital of Rhode Island, Pawtucket, and Brown University School of Medicine, Providence; C. Comella, Rush–Presbyterian–St. Luke's Medical Center and Rush School of Medicine, Chicago; S. Factor, Albany College of Medicine, Albany, N.Y.; R. Kurlan and I. Richard, University of Rochester, Rochester, N.Y.; M. Parsa, Case Western University School of Medicine, Cleveland; R. Pfeiffer, University of Tennessee Medical Center, Memphis; Study Coordinators — R. Davies, Memorial Hospital of Rhode Island, Pawtucket, and Brown University School of Medicine, Providence; K. Janko, Rush–Presbyterian–St. Luke's Medical Center, Chicago; D. Brown, Albany College of Medicine, Albany, N.Y.; I. Gardner and N. Pearson, University of Rochester, Rochester, N.Y.; K. Large, Case Western University School of Medicine, Cleveland; S. Rast, University of Tennessee Medical Center, Memphis; Steering Committee — D. Oakes, University of Rochester, Rochester, N.Y.; C. Goetz, Rush–Presbyterian–St. Luke's Medical Center, Chicago; G. Paulson, Ohio State University, Columbus; F. Marshall, K. Kieburtz, and A. Rudolph, University of Rochester, Rochester, N.Y.; Clinical Trials Coordination Center — K. Bourgeois, C. Casacelli, A. Freimuth, B. Guthrie, R. Pelusio, and A. Watts, University of Rochester, Rochester, N.Y.; Safety Monitoring Committee — P. Tariot and R. Raubertas, University of Rochester, Rochester, N.Y.; and T. Greenamyre, Emory University, Atlanta.
| |||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Clinical Course of Acute Gastric Mucosal Lesions Caused by Acute Infection with Helicobacter pylori
Moriai T., Hirahara N.
Extract |
Full Text
N Engl J Med 1999;
341:456-457, Aug 5, 1999.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
