Associations between the number of antipsychotics prescribed and metabolic parameters in Japanese patients with schizophrenia

Abstract Aim There is little evidence on the effects of antipsychotic polypharmacy on metabolic parameters in patients with schizophrenia. Thus, this cross‐sectional study explored the associations between the number of antipsychotics prescribed and metabolic parameters in Japanese patients with schizophrenia. Methods We obtained metabolic parameter data from 19,675 patients with schizophrenia. Of these, 1380 (7.0%), 8422 (42.8%), 6326 (32.2%), and 3547 (18.0%) were treated with none, one, two, and three or more antipsychotics, respectively. We compared eight metabolic parameters among the four groups using univariate analyses. We then performed multiple regression analysis to assess the effect of the number of antipsychotics prescribed on metabolic parameters after controlling for the effects of age, sex, type of care (outpatient/inpatient), chlorpromazine‐equivalent dose, and antipsychotic type (aripiprazole, olanzapine, and risperidone). Results There were significant differences in body mass index (BMI), systolic and diastolic blood pressure (dBP), total cholesterol, low‐density lipoprotein cholesterol, and triglycerides among the four groups. The multiple regression analysis showed that the number of antipsychotics prescribed was significantly correlated with BMI and dBP (standardized regression coefficient = 0.031 and 0.026, respectively). Conclusion Our results suggested that the number of antipsychotics prescribed adversely affects BMI and dBP. Clinicians should avoid inappropriate antipsychotic polypharmacy, especially polypharmacy involving three or more antipsychotics.


INTRODUCTION
Although antipsychotic polypharmacy is frequently used in clinical practice, the treatment guidelines for schizophrenia recommend antipsychotic monotherapy. 1,2In Japan, approximately half of all patients with schizophrenia receive antipsychotic polypharmacy. 3,4The benefits and risks of antipsychotic polypharmacy are controversial. 1,2For example, a recent meta-analysis reported that mortality risk does not differ between antipsychotic polypharmacy and monotherapy, 5 whereas antipsychotic polypharmacy may be associated with adverse drug events. 6tabolic syndromes and metabolic abnormalities are major side-effects of antipsychotic use. 7However, there is little evidence on the effects of antipsychotic polypharmacy on metabolic syndromes and metabolic parameters in patients with schizophrenia. 80][11][12][13] Using univariate analysis, Correll et al. 10 found that patients treated with antipsychotic polypharmacy had higher rates of metabolic syndromes than those undergoing antipsychotic monotherapy.However, this result became non-significant after controlling for relevant variables using multiple regression.Therefore, examining the effects of antipsychotic polypharmacy and relevant variables (e.g., age and sex) on metabolic parameters using both univariate and multivariate analyses is valuable.
This cross-sectional study aimed to assess the associations between the number of antipsychotics prescribed and metabolic parameters in Japanese patients with schizophrenia.

Participants
This study was approved by the Institutional Review Board of the Japan Psychiatric Hospital Association.All participants provided written informed consent.

This study was part of the Joint Project on Antipsychotic
Treatment and Physical Risk established by the Japan Psychiatric Hospital Association and the Japanese Society of Clinical Neuropsychopharmacology. [14][15][16][17] We conducted a questionnaire survey across 520 outpatient and 247 inpatient facilities belonging to the Japan Psychiatric Hospitals Association between January 2012 and July 2014.5][16][17]

Measurements
We compiled a brief questionnaire that encompassed age, sex, type of care (outpatient/inpatient), prescribed dose of antipsychotics, height and weight, systolic (sBP) and diastolic blood pressure (dBP), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglycerides (TG), and fasting blood sugar (FBS) level.Body mass index (BMI) was calculated as the ratio of body weight to height (kg/m 2 ).We calculated chlorpromazine (CP)-equivalent doses using established conversion formulas. 20

Statistical analysis
Participants were assigned to four groups according to the number of antipsychotics prescribed: none, one, two, and three or more.We compared age, sex, type of care, CP-equivalent dose, type of antipsychotics prescribed (aripiprazole, olanzapine, and risperidone), and eight metabolic parameters (BMI, sBP, dBP, TC, HDL-C, LDL-C, TG, and FBS) among the four groups using analyses of variance (ANOVAs) and χ 2 tests.The antipsychotic type has been shown to have an effect on metabolic parameters. 10,21,22Olanzapine, risperidone, and aripiprazole have the poorest, moderate, and safest metabolic side-effect profiles, respectively. 21These were the three most commonly prescribed antipsychotics in our study.Therefore, we included aripiprazole, olanzapine, and risperidone in the analyses.
For the univariate analyses, the significance level was set at p < 0.003 based on Bonferroni correction for 15 tests.When significant differences were found among the four groups in the ANOVA, we performed post hoc tests to identify significant differences between groups with Bonferroni correction.
We then performed multiple regression analysis to assess the effects of potential predictors (age, sex, type of care, number of antipsychotics prescribed, CP-equivalent dose, and antipsychotic type) on the eight metabolic parameters.Men and women were coded as 1 and 2, respectively.Outpatients and inpatients were coded as 1 and 2, respectively.The prescription of 0, 1, 2, and 3 or more antipsychotics was coded as 0, 1, 2, and 3, respectively.Not taking and taking aripiprazole, olanzapine, and risperidone were coded as 0 and 1, respectively.For the multiple regression analysis, the significance level was set at p < 0.006 based on Bonferroni correction for eight tests.
We used SPSS for Windows Version 19.0 (IBM Japan) for all statistical analyses.

RESULTS
The study participants comprised 19,675 patients with schizophrenia.
Of these, 1380 (7.0%), 8422 (42.8%), 6326 (32.2%), and 3547 (18.0%) were treated with none (Group 0), one (Group 1), two (Group 2), and three or more (Group 3) antipsychotics, respectively (Table 1).There were significant differences in age, sex, type of care, CP-equivalent dose, antipsychotic type (aripiprazole, olanzapine, and risperidone), and six of the eight metabolic parameters (i.e., BMI, sBP, dBP, TC, LDL-C, and TG) among the four groups.Age was highest in Group 0, followed by Groups 1, 2, and 3.In contrast, CP-equivalent dose was highest in Group 3, followed by Groups 2, 1, and 0. BMI was significantly higher in Groups 1, 2, and 3 than in Group 0. BMI was significantly higher in Group 2 than in Group 1.The sBP was significantly higher in Groups 0 and 1 than in Group 3. The dBP was significantly higher in Group 3 than in Group 1. TC was significantly higher in Groups 1 and 2 than in Group 3. LDL-C and TG were significantly higher in Group 1 than in Group 3.

DISCUSSION
The rate of antipsychotic polypharmacy is considered to be unnecessarily high in Japan. 2 In this study, 9873 of 19,675 (50.2%) patients with schizophrenia had been prescribed multiple antipsychotics.
Notably, 3547 (18.0%) patients were treated with three or more antipsychotics.Moreover, recent studies conducted in Japan have reported that approximately half of all patients with schizophrenia are undergoing antipsychotic polypharmacy. 3,4Despite decreases in the rate of antipsychotic polypharmacy in Japan, rates remain the highest among Asian countries. 4Training and education programs for Japanese psychiatrists (e.g., the Effectiveness of Guidelines for Dissemination and Education Psychiatric Treatment Project 3 ) may improve and minimize the rate of antipsychotic polypharmacy.
Our univariate analyses revealed significant differences in six metabolic parameters (i.e., BMI, sBP, dBP, TC, LDL-C, and TG) as well as potential predictors (i.e., age, sex, type of care, CP-equivalent dose, and antipsychotic type) among the four prescribed antipsychotic number groups.After controlling for the effects of these potential predictors, the associations between the number of antipsychotics prescribed and BMI and dBP remained significant using multiple regression analysis.
A recent study revealed that patients with schizophrenia are genetically protected against abnormalities in several metabolic parameters (i.e., BMI, sBP, HDL-C, TG, and blood sugar), although the negative correlation between dBP and schizophrenia was not significant. 23These findings suggest that non-genetic factors play an important role in the high prevalence of metabolic abnormalities in patients with schizophrenia.Our results indicate that the number of antipsychotics prescribed, rather than CP-equivalent dose, is a risk factor for obesity and hypertension in patients with schizophrenia.Notably, a higher CP-equivalent dose does not necessarily reflect antipsychotic polypharmacy.The higher the number of antipsychotics prescribed, the greater the likelihood that patients' body weight and blood pressure will be affected because metabolic side-effect profiles differ among antipsychotics. 21Indeed, the type of antipsychotics prescribed was associated with BMI and dBP, and aripiprazole was negatively correlated with BMI.Notably, meta-analyses have demonstrated that aripiprazole augmentation of antipsychotics is associated with lower body weight and BMI. 24,25Of the 2121 patients prescribed aripiprazole in the current study, 1208 (57.0%) received antipsychotic polypharmacy.The negative correlation between aripiprazole and BMI may reflect the protective effects of aripiprazole against antipsychotic-induced weight gain.In contrast, olanzapine has been documented as the most BMI-increasing antipsychotic. 21wever, we found a negative correlation between olanzapine and BMI.This may be because, in real-world practice, clinicians avoid prescribing olanzapine to patients with a high BMI.In addition, although olanzapine was correlated with dBP, risperidone was negatively correlated with dBP.A recent meta-analysis reported that olanzapine and risperidone did not affect dBP. 26However, the number of studies on the effects of olanzapine and risperidone were seven and two, respectively, and the I 2 values for olanzapine and risperidone were 53.0% and 61.3%, respectively, indicating high heterogeneity across studies.Thus, to draw definitive conclusions about the effect of different antipsychotics on blood pressure, further prospective studies are needed.0][11][12][13] Our sample size (n = 19,675) was substantially larger than that of previous studies (n = 118-543), [9][10][11][12][13] which enabled the detection of a small yet statistically significant adverse effect of the number of antipsychotics prescribed on BMI and dBP using both univariate and multivariate analyses.The differences in BMI and dBP among the groups were small, and these parameters were within normal limits for all four groups.Therefore, the clinical significance of our results should be interpreted with caution.

T A B L E 1 Number of antipsychotics prescribed and metabolic parameters
Our study had several limitations.First, we used a cross-sectional design; therefore, we were unable to examine causal relationships.
Second, antipsychotics were not randomly chosen.Thus, our results may have been influenced by selection bias.Third, we did not adjust for other potential confounding factors, such as duration of illness, concomitant use of other psychotropics (e.g., mood stabilizers), smoking status, and symptom severity. 21,28Fourth, the data on physical illnesses and the use of medications to treat physical illnesses were not available.Approximately 7% of inpatients did not receive antipsychotics despite being hospitalized; some of these patients may not have been prescribed antipsychotics because of a physical illness.There were also differences in age among the four groups, of which those who were not treated with antipsychotics were the oldest.These patients were more likely to have physical illnesses and take more medications than the other groups.
Therefore, we cannot exclude the possibility that our results were influenced by physical illnesses and the use of medications to treat physical illnesses.
In summary, our results suggest that the number of antipsychotics prescribed adversely affects BMI and dBP in patients with schizophrenia.However, large-scale observational studies have reported that antipsychotic polypharmacy is more strongly associated with lower rehospitalization and mortality than antipsychotic monotherapy. 29,30Therefore, the benefits and risks of antipsychotic polypharmacy remain controversial.b Reference is outpatient.
c Reference is not taking the antipsychotic.
Data are expressed as means ± standard deviations.The level of significance was set at an overall p < 0.003 based on Bonferroni correction for 15 tests.We applied Bonferroni correction to post-hoc tests to identify significant differences between groups.Abbreviations: BMI, body mass index; CP, chlorpromazine; dBP, diastolic blood pressure; FBS, fasting blood sugar; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; sBP, systolic blood pressure; TC, total cholesterol; TG, triglycerides.aCalculated using an analysis of variance.
Briefly, patients were diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders, fourth edition, text revision 18 criteria, or the International Statistical Classification of Diseases and Related Health Problems, Version 10 19 criteria.We excluded individuals who were under 20 years old.
The level of significance was set at p < 0.006 based on Bonferroni correction for eight tests.Abbreviations: BMI, body mass index; CP, chlorpromazine; dBP, diastolic blood pressure; FBS, fasting blood sugar; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; sBP, systolic blood pressure; SRC, standardized regression coefficient; TC, total cholesterol; TG, triglycerides.