Preventive effect of aripiprazole once monthly on rehospitalization for bipolar disorder: A multicenter 1‐year retrospective mirror image study

Abstract Aim We conducted a 1‐year retrospective mirror‐image study to investigate the effect of aripiprazole once monthly (AOM) on rehospitalization for bipolar disorder. Methods Participants were recruited from psychiatric emergency and acute care hospitals in western Japan. We included 39 participants with bipolar disorder who had been administered AOM for at least 1 year with no missing medical records during the observational period. The primary outcomes were rehospitalization rate, number of rehospitalizations, total hospitalization days, and time to rehospitalization in the context of overall psychiatric readmissions. The significance level was set at p < 0.05. Results AOM significantly reduced the rehospitalization rate from 23/39 (59%) to 7/39 (18%) (p = 0.001). The number of rehospitalizations decreased significantly from a mean of 0.85 per person‐year to 0.41 per person‐year (p = 0.048). The total hospitalization days significantly decreased from a mean of 34.9 days to 14.4 days (p = 0.008). AOM significantly prolonged the time to rehospitalization (p < 0.001). Conclusion This study found that AOM reduces overall psychiatric rehospitalization for bipolar disorder based on data from 1 year before and after AOM administration in the real‐world setting. Future studies should examine the robustness and persistence of the rehospitalization preventive effect of AOM with larger sample sizes and longer observation periods beyond 1 year.

quality of life. 6 Hence, bipolar disorder ranks 28th in years lived with disability and 67th in disability-adjusted life years. 7 Prevention of recurrence is clinically essential to avoid poor outcomes in patients with bipolar disorder.
Maintenance treatment with medication is crucial for preventing the recurrence of bipolar disorder. Treatment guidelines in various countries recommend maintenance treatment with mood stabilizers and antipsychotics. [8][9][10] However, approximately 50% of patients with bipolar disorder have poor adherence. 11 In a prospective 21month observational study in the treatment of bipolar disorder, the non-adherent group had more recurrences, rehospitalizations, and suicide attempts than those of the adherent group. 12 Using longacting injectable (LAI) other than oral medication improves medication adherence in bipolar disorder, as well as therapeutic alliances, shared decision-making, and psychoeducation. 11 A database study showed that the proportion of days covered (PDC) increased by 5% and the medication discontinuation rate decreased by 19% in LAIinitiated patients with bipolar disorder compared with those in oral antipsychotic-initiated patients with bipolar disorder. 13 Once-monthly long-acting injection of aripiprazole, aripiprazole once monthly (AOM) has recently been approved of maintenance treatment of bipolar disorder. AOM prolonged the time to recurrence of any mood episode in a 52-week, double-blinded, randomized controlled trial of patients with bipolar disorder type I in the maintenance phase (hazard ratio: 0.45). 14 Another 52-week, openlabel study of AOM investigated patient-oriented outcomes, and more than 70% of patients with bipolar disorder reported being extremely or very satisfied with AOM treatment, while more than 65% reported either no side effect or much less side effects compared with previous medications. 15 Owing to its effectiveness and associated patient satisfaction, AOM has been recommended as a maintenance treatment for bipolar disorder by several pharmacotherapeutic guidelines. 9,16 Previous studies have shown that LAIs effectively reduce rehospitalizations for bipolar disorder in the real-world setting. In a large Finnish cohort study, compared with identical oral antipsychotics, first-generation antipsychotic LAIs and second-generation antipsychotic LAIs (SGA-LAIs) had an adjusted hazard ratio of 0.70 for psychiatric rehospitalization on bipolar disorder. 17 One-year, retrospective, mirror-image studies have reported that both risperidone LAI and paliperidone LAI reduce rehospitalization in patients with bipolar disorder after their use. 18,19 Moreover, 1-year, retrospective, mirror-image studies conducted in an overall population of patients with bipolar disorder using three SGA-LAIs-risperidone LAI, paliperidone LAI, and AOM-showed a statistically significant decrease in rehospitalization rates, 20 the number of rehospitalizations, 21 and total hospitalization days post-treatment. [20][21][22] However, previous studies have not verified whether AOM reduces rehospitalization for bipolar disorder after its use. A large Finnish cohort study included only risperidone LAI and olanzapine LAI but not AOM among SGA-LAIs. 17 In addition, 1-year, retrospective, mirror image studies conducted in Turkey and Italy included a mixed group of patients with bipolar disorder using multiple SGA-LAIs, such as risperidone LAI, paliperidone LAI, and AOM. [20][21][22] These studies did not examine the effect of AOM alone on rehospitalization for bipolar disorder by selecting only AOM users. [20][21][22][23] Here, we conducted a 1-year, retrospective, mirror-image study to determine whether AOM reduces rehospitalization rates, number of rehospitalizations, and total hospitalization days and prolongs the time to rehospitalization in patients with bipolar disorder after its use. Our study examined the clinical impact of AOM by evaluating various types of readmissions, including overall psychiatric readmissions, readmissions specifically related to manic or depressive episodes, and both voluntary and involuntary readmissions. By analyzing these specific aspects, we aimed to provide a comprehensive understanding of the clinical benefits and effectiveness of AOM.

| Study design
This 1-year, multicenter, retrospective, mirror-image study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines. 24 The study examined whether AOM introduction reduces rehospitalization in patients with bipolar disorder. Figure 1 shows the observation period of the study. The index date was the date of the first AOM administration. If AOM was introduced in an outpatient setting, the years before and after the index date were defined as the pre-and post-treatment periods, respectively ( Figure 1A). If AOM was introduced in an inpatient setting, the period up to 1 year before the index date was defined as the pre-treatment period, whereas the year after discharge from the hospitalization in which AOM was introduced was defined as the post-treatment period ( Figure 1B); this was done because rehospitalizations cannot occur if patients have not first been discharged.

| Setting of the study
The survey period was from January 1, 2023 to March 31, 2023. Six psychiatric emergency and acute care hospitals in Western Japan participated in this study. The information required for the study was retrospectively collected from the medical records of the participating hospitals.

| Eligibility criteria
We selected participants who (1) had been prescribed AOM between May 1, 2015, and December 31, 2022; (2) were diagnosed with bipolar disorder according to the International Classification of Diseases, 10th Revision (ICD-10) 25 ; (3) received AOM every consecutive month for more than 1 year; (4) had no missing medical data from both the pre-and post-treatment periods; and (5) did not opt out of participating in the study. We excluded participants (1) diagnosed with schizophrenia, schizoaffective disorder, or delusional disorder according to the ICD-10 and (2) whose administration interval of AOM was longer than 2 months.

| Definitions of primary and secondary outcomes
The following four outcomes were established: rehospitalization rate, number of rehospitalizations, total hospitalization days, and time to rehospitalization. The proportion of patients readmitted at a participating psychiatric hospital at least once during the pre-or post-treatment period was defined as the rehospitalization rate. The total number of psychiatric readmissions during each period was defined as the number of rehospitalizations. The total length of stay in the psychiatric hospital during each period was defined as the total hospitalization days. Finally, the time from the beginning of either the pre-or post-treatment period to the first readmission was defined as the time to rehospitalization.
Primary outcomes included the rehospitalization rate, number of rehospitalizations, total hospitalization days, and time to rehospitalization in the context of overall psychiatric readmissions.
In contrast, the secondary outcomes specifically examined these same measures in the context of manic or depressive readmis-

| Definitions of the participant characteristics
Sociodemographic, clinical, and therapeutic data were collected.
Sociodemographic variables were age, sex, education, and employment status at the time of AOM introduction. Clinical variables included age at the onset of bipolar disorder, duration of illness, and number of psychiatric hospitalizations from the onset of bipolar disorder to AOM introduction. Therapeutic variables were the annual prescription days of aripiprazole oral formulation during the year before the first AOM administration; maintenance dose of AOM after continuation of its administration for a year; and concurrent use of oral antipsychotics, mood stabilizers, and antidepressants after a year of AOM administration.

| Consideration of biases
A common case-reporting form was used at the participating hospitals to ensure that the information was collected comprehensively as per the uniform definitions in the research protocol. An individual double-checked the data collected at each participating hospital.

| Statistical methods
We used the R version 4.1.0. software (R Foundation for Statistical Computing) for statistical analyses. In descriptive statistics, normally distributed continuous variables are expressed as means and standard deviations, whereas non-normally distributed data are expressed F I G U R E 1 Definitions of observation periods for the mirror-image study. (A) Observation periods for aripiprazole once monthly introduction in the outpatient setting; (B) Observation periods for aripiprazole once monthly introduction in the inpatient setting. as medians and interquartile ranges. Meanwhile, categorical variables are expressed as numbers and percentages.
The following statistical analyses were conducted on the primary and secondary outcomes. McNemar's test was performed on the difference in the rehospitalization rates between the pre-and post-treatment periods, and its effect size was evaluated using Cohen's g. The number of rehospitalizations and total hospitalization days were determined using the Shapiro-Wilk test for nonnormality. The differences between each period were examined using the Wilcoxon signed-rank test, and non-parametric effect sizes were evaluated using Cliff's delta. Cohen were also provided. Furthermore, we conducted the same analysis on the subgroup of AOM users who were not concurrently receiving any concomitant antipsychotics, mood stabilizers, or antidepressants. This analysis aimed to evaluate the treatment effects of AOM by excluding the potential additional effects of concomitant medications.

| Explanatory analysis
We explored the factors associated with rehospitalization after AOM by comparing groups with and without rehospitalization after AOM introduction. We tested for differences between the two groups with respect to sociodemographic, clinical, and therapeutic variables. The Fisher's exact test was used to compare categorical variables, and the Mann-Whitney U test was used to compare continuous variables.
The significance level was set at a p-value of <0.05.

| Participant selection
In total, 618 patients from the participating hospitals underwent AOM treatment at least once. Among them, 512 had non-bipolar disorders, including schizophrenia. A total of 106 patients had bipolar F I G U R E 2 Selection process for study participants. ICD-10, International Classification of Diseases, 10th Revision. disorder and had used AOM at least once, of whom 34 and 72 received AOM in outpatient and inpatient settings, respectively. Sixtyseven patients were excluded because they had been receiving AOM for less than 1 year (n = 62) or had missing medical data (n = 5).
Thirty-nine patients were included in the final analysis ( Figure 2).

| Participant characteristics
The participant characteristics are summarized in Table 1. The average age of the study participants was 39.2 ± 15.6 years, and 30.8% were men. The mean age of onset of bipolar disorder was 28.5 ± 11.3 years, the mean duration of illness was 10.7 ± 8.

| Results of the secondary analysis
Rehospitalizations for mania exhibited significant decreases across all measures between the pre-and post-treatment periods, while rehospitalizations for depression did not show any significant changes (Tables S1 and S2) Note: The number of lifetime psychiatric hospitalizations was counted from the onset of bipolar disorder to the first AOM administration. The annual prescription days of oral aripiprazole were recorded during the year before the first AOM administration. The AOM maintenance dose and concurrent use of oral antipsychotics, mood stabilizers, and antidepressants were assessed a year after AOM administration.

| Results of the subgroup analysis
Twenty of 39 (51%) participants were AOM users not concurrently taking any antipsychotic, mood stabilizer, or antidepressant medications. This subgroup showed a significant reduction in rehospitalization across all measures between the pre-and post-treatment periods (Table S5), along with a significant improvement in the time to rehospitalization (p < 0.001).

| Results of the explanatory analysis
Seven and 32 patients were in the rehospitalization and nonrehospitalization groups, respectively, after introducing AOM. Six of seven (86%) patients in the rehospitalization group and nine of 32 (28%) in the non-rehospitalization group were on mood stabilizers after 1 year of AOM, with a significant difference between the two groups (p = 0.02). No significant differences were detected between the two groups for any other variables. Details are shown in Table S6.

| Summary of the findings
This study found that AOM significantly reduced rehospitalization rates, number of rehospitalizations, and total hospitalization days and significantly prolonged the time to rehospitalization for patients with bipolar disorder in a 1-year retrospective mirror-image study.
These findings remained consistent even in subgroup analyses that excluded concomitant medications with maintenance effects.
Using the same 1-year mirror image design, previous studies have shown that risperidone LAI, paliperidone LAI, and overall SGA-LAIs significantly reduced rehospitalization for bipolar disorder. [18][19][20][21][22] However, these previous studies did not specifically reveal the TA B L E 2 Rehospitalization rate, number of rehospitalizations, and total hospitalization days related to overall psychiatric readmissions before and after aripiprazole once monthly in patients with bipolar disorder.

F I G U R E 3
Time to rehospitalization before and after aripiprazole once monthly. Kaplan-Meier survival plots are shown comparing times to rehospitalization between the pre-and post-treatment periods. The red survival plot represents the pretreatment period. The green survival plot represents the posttreatment period. The log-rank test was used to identify statistical significance (p < 0.001).

| Interpretation of the primary analysis
The effect of AOM on rehospitalization for bipolar disorder shown in this study is consistent with the results of several previous studies. In the present study, the rehospitalization rate significantly decreased from 59% to 18% between the pre-and post-treatment periods with AOM introduction; likewise, the rehospitalization rate significantly decreased from 57% to 23% in the SGA-LAI subgroup of the STAR study (n = 52). 20 The number of rehospitalizations significantly decreased from 0.85 to 0.41 in the present study and from 0.9 to 0.1 in the Turkish study (n = 17). 21 The more significant reduction in the Turkish study may have been caused by differences in the participants' background characteristics between the two studies.
The total hospitalization days were reduced from 34 to 14 days in the present study and from 30 to 17 days in the SGA-LAI subgroup of the NOMIAC study (n = 49) 22 ; thus, both studies have similar findings.
The STAR 20 and NOMIAC studies, 22  In the present study, the time to rehospitalization was significantly prolonged between the pre-and post-treatment periods.
The survival curve for rehospitalization in the pre-treatment period showed a cumulative decrease after 300 days, whereas that in the post-treatment period remained flat beyond 300 days. Medication non-adherence is a potential reason for the increase in rehospitalization after 300 days in the pre-treatment period. In a study examining oral medication adherence among patients with bipolar disorder using Japanese health insurance data, the 1-year average PDC ranged from 0.51 to 0.61, which is below 0.80, the PDC that indicates good adherence. 28 Thus, non-adherence with oral medication during the pre-treatment period may result in insufficient therapeutic concentrations, leading to recurrence and rehospitalization. 12 Conversely, AOM prevented recurrence and rehospitalization in the post-treatment period because the formulation maintained sufficient therapeutic concentrations for 1 month. 29

| Interpretation of the secondary analysis
The findings of this study demonstrate a significant reduction in manic rehospitalizations in real-world settings with the use of AOM.
However, statistical significance was not achieved for depressed rehospitalizations. These results are consistent with those of previous studies. For instance, a 52-week double-blind randomized controlled trial showed that AOM significantly reduced the recurrence of manic and mixed episodes but did not significantly reduce the recurrence of depressive episodes. 14 Similarly, a 100-week double-blind randomized controlled trial of oral aripiprazole revealed a significant reduction in manic episodes but not in depressive episodes. 30 Correspondingly, the 1-year retrospective mirror image study reported a significant decrease in hospitalization days for mania before and after the introduction of LAIs, including AOM. However, no significant difference was observed in the number of hospitalization days for depression. 22 Based on these collective findings, it appears AOM is effective in preventing manic episodes but may not be as effective in preventing depressive episodes.

| Interpretation of the explanatory analysis
The rehospitalization group following the introduction of AOM exhibited a significantly higher likelihood of utilizing concomitant mood stabilizers than the non-rehospitalization group. The majority (75%) of rehospitalizations after AOM were attributed to depression, indicating the concurrent use of mood stabilizers served as a treatment approach for recurrence/relapse of depression following AOM introduction. In cases where the initiation of mood stabilizers was delayed, the worsening of depression may have led to rehospitalization. A network meta-analysis suggested that the combination of oral aripiprazole with either valproate or lamotrigine was more effective in preventing recurrence/relapse of the depressed phase in bipolar disorder compared to AOM alone. 31 Therefore, early implementation of these mood stabilizers in combination with AOM introduction may help reduce rehospitalization for depression in patients with bipolar disorder experiencing a predominantly depressed phase. 32

| Limitations
The findings of this study should be interpreted with caution owing to its limitations. First, although the number of AOM users in this study was larger than that in previous studies, it was still lower at 39. Second, because the study participants were recruited from six psychiatric emergency and acute care hospitals, the results of this study are applicable to severely ill patients with bipolar disorder and may not be generalizable. Third, as this study did not include patients with bipolar disorder who discontinued AOM within 1 year, it may have excluded patients who had an unfavorable response to AOM within 1 year, leading to selection bias and overestimation of the treatment effect. Furthermore, although we diagnosed bipolar disorder based on medical record data, we lacked sufficient information to distinguish bipolar disorder types I and II. We determined the mood episodes at readmission (manic or depressive); however, we were unable to assess the severity of these episodes using established rating scales such as the Young Mania Rating Scale or Hamilton Depression Scale. Additionally, our study did not include the observation of recurrence/relapse events not requiring rehospitalization. We were also unable to adjust for risk factors for rehospitalization, such as substance abuse and personality disorders, 33,34 and protective factors for rehospitalizations, such as psychoeducation. 35,36

| CON CLUS ION
Our study found that AOM significantly reduced rehospitalization rates, number of rehospitalizations, and total hospitalization days and significantly prolonged the time to rehospitalization in the context of overall psychiatric readmissions among patients with bipolar disorder by comparing data from 1 year before and after AOM treatment. These findings suggest that AOM may help prevent rehospitalization as a maintenance treatment for bipolar disorder. Future studies should examine the robustness and persistence of the rehospitalization preventive effect of AOM with larger sample sizes and extended observation periods beyond 1 year.

AUTH O R CO NTR I B UTI O N S
Goto J. and Nabeshima Y. conceptualized and designed the study.
critically revised the manuscript for important intellectual content.
All authors approved the version to be published and agreed to be accountable for all aspects of this work.

ACK N OWLED G M ENTS
We thank Negishi T., Miyagawa H., Yamada M., Fukuda N., Uno K., Murata T., Nishimura H., Hiraki F., Murakami R., and Ogata H. for clinical data collection and fine preparation of the case reporting form. We also thank Editage (www. Edita ge.com) for the English language editing.

FU N D I N G I N FO R M ATI O N
This study did not receive any specific grants from funding agencies in the public, commercial, or non-profit sectors.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data of the present study cannot be made publicly available because the disclosure of personal data was not included in the research protocol and because the opt-out clearly states that personal data will not be disclosed. Yoshihiro Nabeshima https://orcid.org/0009-0004-3072-0399