The effectiveness of clozapine is clearly superior to other antipsychotics in the treatment of refractory schizophrenia. Clozapine leads to various side effects, and therefore many patients are forced to discontinue. In this study, we analyzed the registry database of all cases in Japan to identify risk factors for discontinuation of clozapine.
The Clozaril patient monitoring service® (CPMS) database from July 31, 2009 to January 26, 2020 was acquired. We defined the following exclusion criteria: patients who had ever taken clozapine by a non-CPMS method, such as an individual import or clinical trial, patients who did not receive clozapine after being enrolled in CPMS, and patients with initial doses other than 12.5 mg (outside the current protocol). Therefore, all patients in this study are new users. Multivariate Cox regression analysis was used to investigate independent risk factors associated with time to discontinuation of clozapine.
We identified 8,263 patients as the study population. Clozapine discontinuation was significantly associated with age 40 and older [hazard ratio (HR)=1.66, p<0.001], intolerance to olanzapine (HR=1.31, p=0.018), previous treatment with clozapine (HR=1.30, p=0.001), and leukocyte counts <6,000/mm3 (HR=1.24, p<0.001). The Kaplan-Meier curves for clozapine discontinuation by age group revealed that older age at the time of clozapine introduction tended to have lower continuation rates.
Careful administration is important because patients with these factors have a high risk of discontinuation. In addition, the initiation of clozapine during the younger period was more effective and more tolerated.
Schizophrenia continues to be an important disease in psychiatry. Schizophrenia is consistently high in terms of disability-adjusted life years [
The most effective treatment for refractory schizophrenia is continuous administration of clozapine [
A small study previously reported that in African-Americans (in a study on racial differences) [
Treatment-resistant schizophrenia, which is indicated for clozapine in Japan, is defined by two criteria. The first is poor response, with limited response after an administration of sufficient doses of two or more antipsychotics for 4 weeks or longer. It is at least 600 mg/day of chlorpromazine equivalent and contains one or more second-generation antipsychotics. The second is poor tolerability, the dose of antipsychotics could not be increased sufficiently due to side effects. In Japan, clozapine induction is performed in the hospital, and it requires inpatient treatment up to 18 weeks after the administration. Regular blood test at least once a week up to 26 weeks is also demanded. After 26 weeks, it is needed follow-up with a blood test at least once every two weeks.
Data were obtained from the CPMS center as requested by the Japanese Society of Neuropsychopharmacology. The CPMS data consisted of registered data for all patients who had received clozapine in Japan between July 31, 2009 and January 26, 2020. The CPMS data we acquired were anonymized, and each patient was given a unique identification code, which included patient registration date, registered medical institution, sex, date of birth, test date, leukocyte count, neutrophil count, blood glucose level, hemoglobin A1c (HbA1c), and prescription date. If a patient had previously discontinued clozapine for any reason and then readministered clozapine, the patient was newly enrolled and treated as a separate case. Due to personal information protection considerations, patient consent requirements were waived.
To describe the risk factors for clozapine discontinuation using the CPMS database, we defined the following exclusion criteria: patients who have ever taken clozapine by a non-CPMS method, such as an individual import or clinical trial, patients who did not receive clozapine after being enrolled in CPMS, and patients with initial doses other than 12.5 mg (outside the current protocol). Therefore, all patients in this study were new users with clozapine administration that was compatible with the standard protocol.
Patient age was defined as the age at the first prescription of clozapine. The clozapine treatment period was defined as the period from the first prescription date to the last prescription date plus the number of prescription days. Patients who did not register their records in CPMS for discontinuation of clozapine by the observation end date, January 26, 2020, were clozapine continued cases, that is, it was called a clozapine censored case. Neutropenia and leukopenia were defined as medical conditions with neutrophil counts of less than 1,500/mm3 and leukocyte counts of less than 3,000/mm3, respectively. In Japan, if the laboratory values fell into these ranges, the clozapine prescription must be immediately suspended according to CPMS criteria. Furthermore, agranulocytosis was defined as a neutrophil count of less than 500/mm3.
The Kaplan-Meier event-free curve was used to calculate the event-free probability of clozapine discontinuation. A log-rank test was performed to verify the differences between each group. Multivariate Cox regression analysis was used to investigate independent risk factors associated with time to event. The event was the discontinuation of clozapine. Variance inflation factors (VIFs) of 10 or more were considered evidence of the existence of multicollinearity. The proportional hazards assumption was evaluated using the Kaplan-Meier curve. A trend test for survival data was performed for trends by age group in time to event. For the trend test, we assigned the median age within each level in the age group as the variable, that is, 18 for “≤20,” 27 for “21–30,” 36 for “31–40,” 45 for “41–50,” 55 for “51–60,” 64 for “61–70,” and 73 for “≥71.” A Cox regression model was fitted to assess the trend of age group assigned these values. The Cochran-Armitage trend test was performed to clarify the trend of incidence of neutropenia/leukopenia and agranulocytosis in clozapine recipients by age group. All p-values reported were two-sided, and the significance level was set at p<0.05. Statistical analysis was performed using R version 3.6.2 (R Development Core Team, Vienna, Austria).
The total number of patients enrolled in the CPMS database was 9,635. Of these, 1,196 patients were enrolled in duplicate after being transferred to another hospital, so they were excluded after consolidating the data from the hospitals before and after. Of these 8,439 patients, 22 were excluded who had individually imported clozapine prior to CPMS enrollment. Forty-eight patients had participated in the clinical trial and were excluded. Sixty-nine patients who did not receive clozapine after CPMS enrollment were excluded. Finally, there were 37 patients with an initial dose other than 12.5 mg, who were excluded because it was outside the scope of the current protocol. As a result, the number of patients was 8,263 (
The clinical background of the patients is shown in
The Kaplan-Meier curves for clozapine discontinuation events are shown in
Seven variables with p<0.2 in the log-rank test were included in the multivariate analysis. Multivariate Cox regression analysis revealed four independent risk factors for clozapine discontinuation (
The Kaplan-Meier curves for clozapine discontinuation by age group revealed that older age at the time of clozapine introduction tended to have lower continuation rates (
The incidence of agranulocytosis tended to be higher in older age groups, although the incidence of neutropenia/leukopenia tended to be the similar increasing across ages (
This study is one of the nationwide surveys of clozapine prescription, with the largest sample size and nearly a single ethnicity. We focused on the risk factors for clozapine discontinuation and performed an analysis. The analysis revealed that being the group who was 40 years and older, who had poorly tolerated olanzapine, who had readministered clozapine, and with leukocyte counts <6,000/mm3 were independent risk factors for clozapine discontinuation. Among these, the effect of older age had the highest risk of discontinuation, and other risk factors were similar. Furthermore, it was found that the rate of agranulocytosis increased significantly in patients aged 40 years and older, which probably contributed to the higher rate of clozapine discontinuation. Increasing the rate of agranulocytosis at higher age was consistent with studies conducted in the UK [
Among the risk factors identified, the most influential risk factor was the age of clozapine introduction of 40 years or older. The result that the discontinuation rate of clozapine increased with the delay of the induction age was in agreement with the previous study [
Since clozapine has a similar chemical skeleton to olanzapine, its pharmacological properties are also similar. It is known that side effects such as weight gain and blood glucose increases caused by olanzapine can also occur with clozapine [
It was revealed that the clozapine readministration group was also a high-risk group with early discontinuation, although they represented a small group of 5.3%. It is presumed that the readministration group had previously discontinued clozapine use due to some side effects or poor adherence. Clozapine rechallenge can be problematic and should be done with caution. In particular, in the case of interruption due to neutropenia or QT prolongation syndrome, it is recommended to increase the dose in small amounts. Based on the results of this study, rechallenge to patients with leukocyte counts <6,000/mm3 should be done with caution. However, clozapine rechallenge is not recommended for patients with agranulocytosis or myocarditis [
This study has some inevitable limitations because this report is based solely on an analysis of CPMS data. First, the exact adherence of clozapine is unknown. Second, CPMS data use anonymous identification codes to protect the patient’s personal information, so detailed information such as medical records, including age at onset, subclassification of schizophrenia, family history, symptomatic records scored by Positive and Negative symptoms scale (PANSS) and combination data related to drug metabolism or genetic data, are not included in the analysis. Third, there is a large variability in the number of individuals previously on other drugs, which can affect the p-value of the analysis. Furthermore, no information is available on concomitant medications. For example, it is speculated that some doctors prescribe lithium for granulocyte depletion. Therefore, white blood cells and neutrophils are increased in patients receiving lithium, although it could not be considered adequately in this study because of the limitation of analyzed data source. One of the strengths of this study is that CPMS is a nationwide database of all cases in Japan and therefore directly reflects treatment with clozapine in one country. In addition, it can be said that off-label use of antipsychotics is not mixed because it is certain that the target patients for clozapine use are treatment-resistant schizophrenia patients.
In conclusion, we identified independent risk factors for clozapine discontinuation: aged 40 and older, intolerance to olanzapine, previous treatment with clozapine, and leukocyte counts <6,000/mm3. Careful administration is important because patients with these factors have a high risk of discontinuation. In addition, a detailed examination of age revealed that introduction of clozapine during younger periods was more effective and better tolerated.
The authors thank the Japanese Society of Neuropsychopharmacology for their cooperation in data acquisition.
Dr. Toyoda has received speaker’s honoraria from Eisai. Dr. Yamauchi has received speaker’s honoraria from Meiji Seika Pharma, Otsuka. Dr. Kinoshita has received speaker’s honoraria from Meiji Seika Pharma, Otsuka, Sumitomo Dainippon Pharma, EA Pharma and Eisai. Dr. Inada has received speaker’s honoraria from Eisai, Eli Lilly, Janssen, Meiji Seika Pharma, Mitsubishi Tanabe Pharma, Mochida, MSD, Novartis, Otsuka, Shionogi, Sumitomo Dainippon Pharma, and Yoshitomiyakuhin, and research grants from Ministry of Health, Labor and Welfare, Mitsubishi Tanabe Pharma, MSD, and National Center of Neurology and Psychiatry. Dr. Kanazawa has received speaker’s honoraria from Janssen, Meiji Seika Pharma, Otsuka, Shionogi, Sumitomo Dainippon Pharma, and Takeda Pharmaceutical Company, research grants from Ministry of Health, Labor and Welfare, and Health and Labor Sciences Research Grant (20GC1017 Head: Takefumi Ueno). The other authors declare no conflicts of interest.
Conceptualization: Katsunori Toyoda, Takeo Hata, Ken Inada, Tetsufumi Kanazawa. Data curation: Katsunori Toyoda, Takeo Hata. Formal analysis:Katsunori Toyoda, Takeo Hata, Tetsufumi Kanazawa. Funding acquisition: Tetsufumi Kanazawa. Investigation: Katsunori Toyoda, Takeo Hata, Tetsufumi Kanazawa. Methodology: Katsunori Toyoda, Takeo Hata. Project administration: Tetsufumi Kanazawa. Resources: Tetsufumi Kanazawa. Software: Takeo Hata. Supervision: Tetsufumi Kanazawa. Validation: Katsunori Toyoda, Takeo Hata, Tetsufumi Kanazawa. Visualization: Takeo Hata. Writing— original draft: Katsunori Toyoda. Writing—review & editing: all authors.
Flow diagram describing the construction of the study population. CPMS: Clozaril patient monitoring service®.
The Kaplan-Meier survival curves describe the time to discontinuation of clozapine by clinical background. The Kaplan-Meier eventfree curve was used to calculate the event-free probability of clozapine discontinuation. A log-rank test was performed to verify the differences between each group. *p<0.05.
The Kaplan-Meier survival curves describe the time to discontinuation of clozapine based on intolerance to previously used antipsychotics. The Kaplan-Meier event-free curve was used to calculate the event-free probability of clozapine discontinuation. A log-rank test was performed to verify the differences between each group. *p<0.05. SGA: second-generation antipsychotics.
Multivariate Cox regression analysis for independent risk factors for clozapine discontinuation. Seven variables with p<0.2 in the log-rank test were included in the multivariate analysis. Multivariate Cox regression analysis was used to investigate independent risk factors associated with time to event. The event was the discontinuation of clozapine. VIFs of 10 or more were considered evidence of the existence of multicollinearity. It was found that each VIF did not exceed 10. The proportional hazards assumption was evaluated using the Kaplan-Meier curve. *p<0.05. CI: confidence interval, HR: hazard ratio, VIF: variance inflation factor.
Trend of discontinuation rates of clozapine by age group. A trend test for survival data was performed for trends by age group in time to event. For the trend test, we assigned the median age within each level in the age group as the variable, that is, 18 for “≤20,” 27 for “21–30,” 36 for “31–40,” 45 for “41–50,” 55 for “51–60,” 64 for “61–70,” and 73 for “≥71.” A Cox regression model was fitted to assess the trend of age group assigned these values.
Trend of incidence of neutropenia/leukopenia and agranulocytosis in clozapine recipients by age group. The Cochran-Armitage trend test was performed to clarify the trend in the incidence of neutropenia/leukopenia and agranulocytosis in clozapine recipients by age group.
Clinical background of the patients at the registration with the Clozaril patient monitoring service®
Background (N=8,263) | |
---|---|
Sex, N (%) | |
Female | 3,793 (45.9) |
Male | 4,470 (54.1) |
Age, median (range), year |
40 (10–87) |
Age group, N (%) |
|
≤20 | 266 (3.2) |
21–30 | 1,457 (17.6) |
31–40 | 2,458 (29.7) |
41–50 | 2,273 (27.5) |
51–60 | 1,246 (15.1) |
61–70 | 494 (6.0) |
≥71 | 69 (0.8) |
Reason patient introduced clozapine use, N (%) | |
Intolerance to previously used antipsychotics |
537 (6.5) |
Aripiprazole |
212 (2.6) |
Blonanserin |
140 (1.7) |
Olanzapine |
240 (2.9) |
Paliperidone |
83 (1.0) |
Perospirone |
40 (0.5) |
Quetiapine |
91 (1.1) |
Risperidone |
257 (3.1) |
Other SGA |
45 (0.5) |
Poor response to previously used antipsychotics |
7,659 (92.7) |
Aripiprazole |
2,063 (25.0) |
Blonanserin |
1,357 (16.4) |
Olanzapine |
4,650 (56.3) |
Paliperidone |
1,220 (14.8) |
Perospirone |
257 (3.1) |
Quetiapine |
1,526 (18.5) |
Risperidone |
3,804 (46.0) |
Other SGA |
304 (3.7) |
FGA |
766 (9.3) |
Previously treated with clozapine, N (%) | |
Non-treated | 7,828 (94.7) |
Treated | 435 (5.3) |
Leukocyte, median (range), /mm3 | 6,000 (4,000–20,500) |
Neutrophil, median (range), /mm3 | 3,397 (2,000–16,572) |
Hemoglobin A1c, median (range), % | 5.3 (3.2–11.8) |
age is expressed as the age at introduction of clozapine,
as some patients met neither or both of the requirements, the sum is not 100%,
as all patients received more than two antipsychotics, the sum is more than the total.
FGA: first-generation antipsychotics, SGA: second-generation antipsychotics