evidenced base

Schizophrenia Research 166 (2015) 194–200

Contents lists available at ScienceDirect

Schizophrenia Research

journal homepage: www.elsevier.com/locate/schres

The risks and benefits of switching patients with schizophrenia or schizoaffective disorder from two to one antipsychotic medication: A randomized controlled trial

Robert J. Constantine a,⁎, Ross Andel a, Marie McPherson a, Rajiv Tandon b a Florida Mental Health Institute, University of South Florida, 13301 Bruce B. Downs Blvd, Tampa, FL, USA b Department of Psychiatry, University of Florida College of Medicine, Gainesville, FL, USA

⁎ Corresponding author.: Tel.: +1 813 245 3770. E-mail address: rconstantine@usf.edu (R.J. Constantin

http://dx.doi.org/10.1016/j.schres.2015.05.038 0920-9964/© 2015 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 18 March 2015 Received in revised form 24 May 2015 Accepted 27 May 2015 Available online 30 June 2015

Keywords: Schizophrenia Antipsychotic Treatment Polypharmacy Monotherapy Switch Clinical trial Randomized

Background: : Despite little evidence to support its use and practice guidelines discouraging the practice, antipsy- chotic polypharmacy is widely prevalent in schizophrenia. This randomized controlled trial studied the effects of switching patients stable on two antipsychotic medications to one antipsychotic medication. Method: : 104 adult outpatients with schizophrenia from 7 community mental health centers clinically stable on concurrent treatment with 2 antipsychotics were randomly assigned to stay on polypharmacy or to switch to an- tipsychotic monotherapy. Participants were followed for 1-year with assessments of symptoms and side effects occurring every 60 days (7 total assessments). We examined differences in time trajectories in symptoms (PANSS, CGI) and side effects (EPS, metabolic, other) as a function of group assignment (switch vs. stay) and time, using intention-to-treat analysis. Results: : Participants who switched to antipsychotic monotherapy experienced greater increases in symptoms than stay patients. These differences emerged in the second 6 months of the trial. All-cause discontinuation rates over the 1-year trial were higher in the switch-to-monotherapy group than in the stay-on-polypharmacy group (42% vs. 13%; p b 0.01). There were no differences in change over time in any of the side effect measures, except that stay patients experienced a greater decrease in Simpson Angus total scores than switch patients.

Conclusion: : Clinicians should be cautious in switching patients with chronic schizophrenia who are stable on 2 antipsychotics to one antipsychotic. Given the challenges in discontinuing antipsychotic polypharmacy, adequate trials of evidence-based treatments such as clozapine and long-acting injectable antipsychotics should be under- taken in inadequately responsive schizophrenia patients before moving to antipsychotic polypharmacy.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

The concurrent use of two or more antipsychotics in the treatment of schizophrenia is widely prevalent worldwide (Li et al., 2015; Park et al., 2014; Sun et al., 2014) and continues to increase (Gilmer et al., 2007; Mojtabai and Olfson, 2010) despite unfavorable risk/benefit data (Fleischhacker and Uchida, 2012; Gallego et al., 2012; Young et al., 2015) and notwithstanding expert guidelines recommending against the practice (Buchanan et al., 2010; National Collaborating Centre for Mental Health, 2009; Tandon et al., 2008). One explanation for the persis- tence of the practice is that good data on the risks and benefits of initiating antipsychotic polypharmacy are limited and their results are equivocal at best (Barnes and Paton, 2011; Correll et al., 2009; Hatta et al., 2014; Katona et al., 2014). Another is that in routine clinical settings, the deci- sion more often confronting clinicians involves the discontinuation of an- tipsychotic polypharmacy rather than its initiation (Tani et al., 2013;

e).

Tsutsumi et al., 2011). Since switching relatively stable patients on anti- psychotic monotherapy to another is associated with significant risk (Essock et al., 2006; Tandon et al., 2010), physicians are reluctant to switch patients relatively stable on two antipsychotic agents to antipsy- chotic monotherapy as it involves discontinuing one agent.

Clinical trial data regarding the risks and benefits of switching pa- tients with schizophrenia from two to one antipsychotic are limited to two non-randomized studies with first-generation antipsychotics (Godleski et al., 1989; Suzuki et al., 2004 and two randomized trials con- ducted primarily with second-generation antipsychotic medications (Essock et al., 2011; Hori et al., 2013). While all these studies concluded that a majority of patients could be successfully switched from antipsy- chotic polypharmacy to monotherapy, the two randomized trials noted higher discontinuation rates in the switch group (31% vs. 14% in Essock et al. and 15% vs. 0% in Hori et al.) because of worsening symptoms.

This study aims to build on the above work and reports on the risks and benefits of switching from antipsychotic polypharmacy to mono- therapy in 104 patients with schizophrenia or schizoaffective disorder over a 12-month period.

Fig. 1. Flow of patients into switch to monotherapy and stay on polypharmacy groups dur- ing study. Eligible consenting subjects (N = 104).

195R.J. Constantine et al. / Schizophrenia Research 166 (2015) 194–200

2. Methods

2.1. Study participants

Between November 2011 and November 2012, seven sites including 6 not-for-profit community mental health providers and one psychoso- cial rehabilitation center recruited patients with schizophrenia or schizoaffective disorder (DSM-IV-TR, American Psychiatric Association, 2000) who had been receiving 2 antipsychotic medications concurrent- ly for at least 90 days. Recruits had to have been stable on this medica- tion regimen, as indicated by the lack of a psychiatric hospitalization or emergency room visit in the previous 90 days, and the treating physician's certification that there were no plans to change the antipsy- chotic regimen. Diagnosis was confirmed through review of medical re- cords and use of the Schizophrenia Checklist (Astrachan et al., 1972). All patients were between 18 and 64 years old, enrolled in Florida's Medicaid program, and with a stable residence and/or case manager who could keep in touch with the patient during the study period. Exclusion criteria included incarceration, legal incompetence, co-occurring developmental disability, pregnancy, a general medical condition that in the opinion of the treating physician made it unsafe for the patient to participate in the study, or having had 2 or more unsuccessful trials of antipsychotic mono- therapy or clozapine of sufficient durations and doses as defined by the Florida Medicaid Drug Therapy Management Program (2011) guidelines for adults within the prior 3 years.

The study protocol was approved by the University of South Florida Institutional Review Board and its implementation was overseen by a Data and Safety Monitoring Board. Additionally, each site received re- view and approval of the protocol either from its internal IRB or from the Western Institutional Review Board. The ability of recruits to con- sent to participate in the study was initially assessed by their treating physician. Potential subjects then received a thorough description of the study and had to correctly answer at least 7 of 10 questions to indi- cate clear understanding of the study and the consent process. Partici- pants were paid a stipend at baseline and at each assessment.

After confirming that patients were eligible to participate in the study and obtaining informed consent, the research coordinator at each site ap- plied a site-specific random assignment protocol. Randomization was to a switch or a stay condition. Participants were then informed of their re- search status and scheduled for baseline interviews. Switch participants were required to switch from the two antipsychotics they were currently receiving to one of these two within 60 days of baseline assessments. Physicians were free to choose which of the two antipsychotics to con- tinue and at what dose, except that participants currently on treatment with an injectable antipsychotic or those receiving clozapine were re- quired to remain on these medications. Of the 98 study participants who were randomized and completed baseline interviews (Fig. 1), 46 were receiving either clozapine or a long-acting injectable antipsychotic at baseline with an equal number randomized to the switch and stay groups. Stay participants were required to remain on the two antipsy- chotic medications they were currently receiving but treating clinicians had flexibility with dosing. Physicians were free to augment treatment of both switch and stay participants with psychotherapeutic medications other than antipsychotics.

The protocol required participants to remain in their assigned re- search status (switch or stay) for 360 days, unless a change was clearly needed in response to a participant's clinical condition. Participants who had to change their assigned status remained in the study and were followed through the 360-day period. While treatment was open- label, baseline and subsequent assessments were conducted by “inde- pendent assessors” blinded to the research status of participants.

2.2. Baseline and follow-up measures

The primary outcome measure was severity of symptoms assessed by the Positive and Negative Syndrome Scale (PANSS, Kay et al.,

1987). A secondary outcome measure was all-cause treatment discon- tinuation. Additional measures included the Clinical Global Impressions CGI-S (Guy, 1976), the Abnormal Involuntary Movement Scale (AIMS, Guy, 1976), the Simpson Angus Scale (SAS, Simpson and Angus, 1970) and the Barnes Akathisia Scale (Barnes, 1989) as well as body mass index (BMI) and blood pressure. Participants were assessed using these instruments at baseline and at 60, 120, 180, 240, 300, and 360 days for a total of 7 assessments. In addition, hemoglobin A1c and fasting lipids (Marder et al., 2004; Tandon and Halbreich, 2003) were obtained at baseline, and at 180 and 360 days following baseline measurements.

Table 1 Participant characteristics at baseline.

Switch (n = 43) Stay (n = 47) p-value

Age, mean (SD) 43.9 (9.6) 47.0 (11.1) .157 Sex, % male 48.8 44.7 .693 Education, % .426

b12 years 34.9 21.3 12 years 48.8 53.2 N 12 years 16.3 25.5

Race/ethnicity, % .288 White 53.5 63.8 Black 39.5 34.0 Hispanic or other 7.0 2.1

Marital status, % .230 Single or widowed 72.1 59.6 Divorced 23.3 25.5 Married 4.7 14.9

Living arrangement, % .681 Congregate setting 20.9 31.9 With spouse/family 37.2 34.0 Alone 32.6 27.7 With a friend 9.3 6.4

Schizoaffective disorder (% yes) 60.5 53.2 .487 Total AP dose, mean (SD) 32.9 (10.7) 41.2 (18.5) .011 Length of current AP episode, % .865

Less than 1 year 18.6 19.2 1–2 years 41.9 34.0 3–6 years 23.2 29.8 More than 6 years 16.3 17.0

Prior MHI admission, % 28.6 46.8 .077 History of substance abuse, % 44.2 40.4 .718 Augmentation, % yes 79.1 72.3 .458 Anticholinergics, % yes 65.1 55.3 .343 FACT team patient, % yes 18.6 25.5 .430 Client of center (years), mean (SD) 10.6 (6.9) 12.7 (8.6) .214

Notes. AP = antipsychotic; FACT = Florida Assertive Community Treatment.

196 R.J. Constantine et al. / Schizophrenia Research 166 (2015) 194–200

2.3. Antipsychotic equivalent doses

Data on the antipsychotic doses received by switch and stay partici- pants were collected at baseline and at each of the 60-day follow-up as- sessment periods. All antipsychotic regimens were converted into olanzapine equivalent daily doses using the methodology described by Gardner et al. (2010).

2.4. Rater training

Independent assessors had to be masters-level trained clinicians. As- sessors were trained by a certified PANSS trainer prior to the baseline measurement and on 2 additional occasions during the study period. The trainer additionally provided on-site consultation and training in the context of actual participant assessments. At the completion of each group training session, assessors viewed audio-visual recordings of patient interviews and were asked to assign PANSS and CGIs scores. These scores were compared to a “gold standard score” assigned to the patients by a group of psychiatric experts. Assessors who deviated 2 or more points on the 7-point PANSS rating scale on 3 or more of the 30 PANSS items received remediation until their score fell within ac- ceptable levels. If acceptable scoring was not achieved after attempts at remediation, the assessor was terminated from the study. Assessors were trained on the side-effect scales with patient volunteers used by the trainer to demonstrate proper assessment, and used by assessors to practice proper scoring. Patients involved in this training were not participants in the clinical trial.

Assessors were informed about the importance of remaining blinded to the participant's research status. Research coordinators at each site, who were not blinded to the research status of the partici- pants, scheduled all assessments and conducted all communication with participants.

2.5. Data analysis

We examined differences in trajectories of change in the outcome measures as a function of group assignment. To properly account for the longitudinal structure of the data, we used mixed effects models in SAS (Littell et al., 2006) procedure MIXED. This procedure allowed us to evaluate (a) baseline differences in those who switched vs. stayed (i.e., the main effect of condition), (b) change over time in each outcome (i.e., the overall effect of time), and (c) differences in change over time as a function of whether the participant was assigned to stay on two or switch to one antipsychotic (i.e., the condition-by-time interaction). Mixed effects models are a variant of repeated measures analysis of var- iance with a few advanced features that allow for better representation of the data, full utilization of data including the use of participants with some values missing, and accounting for random effects (e.g., different starting values across participants, which can affect how a participant's scores change over time). Baseline PANSS Total scores were controlled for in the analysis. Since baseline differences in antipsychotic dose were observed, the results for baseline differences as well as differences in trajectories of change were adjusted for differences in baseline anti- psychotic dose. Therefore, each model included the following factors: main effects of baseline antipsychotic dose and PANSS Total score, inter- actions of baseline antipsychotic dose by time and of baseline PANSS Total score by time (all covariates), main effect for time, main effect for condition, and interaction of condition by time. Using these methods, we compared the stay (two antipsychotics) vs. switch (anti- psychotic monotherapy) groups.

Preliminary analyses run at midpoint of the study indicated no lon- gitudinal differences in PANSS Total scores for the switch vs. stay groups, whereas final results pointed to significant differences. There- fore, we applied a two-slope model to the analysis of change in PANSS Total score in addition to the conventional simple one slope model. In this model, the centering time was the midpoint (assessment 4) with

one slope (time-by-condition interaction) being estimated up to assess- ment 4 and one slope from assessment 4 through 7.

3. Results

Of the 104 participants who consented, 98 completed baseline as- sessments (49 switch and 49 stay). Of these 98 patients, 8 dropped out after baseline assessments but before the first scheduled follow-up as- sessment at 60 days; 6 of these were from the switch group and 2 were from the stay group (Fig. 1). A total of 90 participants who complet- ed 2 or more assessments (43 switch and 47 stay) constituted the final sample. The eight dropouts did not differ from the 90 participants with respect to any of the characteristics presented in Table 1 other than con- comitant anticholinergic treatment, with only 2 (25%) of the dropouts having such therapy compared to 70% of the participants. Participants randomized to switch to one antipsychotic and those who stayed on two antipsychotics were similar on all characteristics with the exception of olanzapine equivalent daily doses at baseline, with stay participants receiving about 8 mg/day more than participants who switched (Table 1). Similarly, there were no group differences in any outcome measures at baseline (Table 2). There were, however, significant differ- ences in length of follow-up, with switch participants having data for an average of 6.2 assessments (SD = 1.4) and stay participants for an av- erage of 6.8 assessments (SD = 0.7) (p = .016).

Although there were significant baseline differences in antipsychotic dose (Estimate = 12.5, SE = 3.5, p b .001), the changes in doses over time were not significant (p = .553), and there were also no changes in dose as a function of group assignment over time (p = .558).

A total of 12 participants, 10 switch (23%) and 2 stay (4%), changed their assigned condition during the study period but remained in the study. Twelve additional participants did not complete all seven re- quired assessments. Eight of these were switch participants who com- pleted an average of 3.2 assessments and 4 were stay participants

Table 2 Baseline values on outcome measures, mean (SD).

Switch (n = 43) Stay (n = 47) p-value

Symptom measures PANSS Total 71.3 (18.6) 66.6 (19.7) .246 PANSS Positive Symptoms 17.1 (4.6) 15.6 (5.3) .159 PANSS Negative Symptoms 19.0 (6.8) 17.6 (6.7) .303 PANS General Psychopathology 35.2 (9.8) 33.5 (9.9) .398 CGI Severity 3.7 (0.9) 3.4 (1.1) .249 CGI Improvement . . .

Adherence measures Medication Adherence 1 1.7 (0.5) 1.8 (0.4) .235 Medication Adherence 2 1.8 (0.4) 1.8 (0.4) .847 Medication Adherence 3 1.9 (0.3) 2.0 (0.2) .206 Medication Adherence 4 1.8 (0.4) 1.9 (0.3) .640

Side effect measures AIMS Total 1.5 (2.4) 1.9 (2.3) .447 AIMS Incapacitation 0.2 (0.5) 0.2 (0.4) .952 Barnes Total 0.7 (1.4) 0.9 (1.3) .448 Simpson Angus 2.0 (2.1) 2.9 (3.5) .184 Body mass index 33.6 (7.4) 31.3 (7.3) .149 Hemoglobin A1C 10.0 (23.7) 6.1 (1.7) .291 Total cholesterol 174.0 (38.4) 183.3 (43.8) .294 Triglycerides 165.1 (121.5) 149.9 (93.0) .508

Note. PANSS = Positive and Negative Syndrome Scale; CGI = Clinical Global Impressions Scale; BMI = body mass index.

197R.J. Constantine et al. / Schizophrenia Research 166 (2015) 194–200

who completed an average of 4.3 assessments. Thus, all-cause discon- tinuation was 42% in the switch group (18 out of 43) in comparison to 13% in the stay condition (6 of 47), representing a significant difference (chi-square = 9.72; df = 1; p b 0.01).

Main analyses designed to assess differences in trajectories of change in outcome measures across the study period are presented in Table 3. In this table, the intercept value refers to the baseline value on each outcome measure. The overall trajectory of change was

Table 3 Mixed effects models of change in outcome scores in relation to switch/stay condition assignm

Intercept Ti

Symptom measures PANSS Total 69.62 (6.42)⁎⁎ 1. PANSS Positive Symptoms 16.44 (1.86)⁎⁎ 0. PANSS Negative Symptoms 18.65 (2.19)⁎⁎ 0. PANS General Psychopathology 34.53 (3.37)⁎⁎ 0. CGI Severity 3.80 (0.34)⁎⁎ 0. CGI Improvement . .

Adherence measures Medication Adherence 1 1.64 (0.16)⁎⁎

Medication Adherence 2 1.82 (0.12)⁎⁎

Medication Adherence 3 1.98 (0.10)⁎⁎ − Medication Adherence 4 1.80 (0.10)⁎⁎

Side effect measures AIMS Total 1.76 (0.81)⁎ − AIMS Positive 0.19 (0.08)⁎ − AIMS Severity 0.54 (0.23)⁎ − AIMS Incapacitation 0.48 (0.14)⁎⁎ − Barnes Total 0.96 (0.44)⁎ − Barnes Global 0.50 (0.25)⁎ − Simpson Angus 1.89 (1.04) Duke Health Physical Health 49.63 (8.45)⁎⁎ − Duke General Health 5.92 (0.60)⁎⁎ − Body mass index 36.81 (2.91)⁎⁎ − Hemoglobin A1C 6.89 (0.68)⁎⁎ − Total cholesterol 180.82 (15.93)⁎⁎

Triglycerides 191.28 (42.48)⁎⁎ −

Note. Estimate (standard error of measurement) are reported. Both intercept and slope were ad CGI = Clinical Global Impressions Scale; Est. = Estimate; SE = standard error of measuremen ⁎ p b .05. ⁎⁎ p b .01.

relatively flat for most assessed outcome measures with a few exceptions. For the PANSS Total score, there was an overall increase in scores that averaged 1.46 points per assessment interval. This effect approached (but did not cross) the threshold of statistical significance (p = .050). A similar effect was observed for the PANSS subscales.

Finally, the condition by time interaction yielded a significant effect for the PANSS Total score, indicating that the overall increase in PANSS Total score observed as a time effect could be attributed to a substantial- ly greater increase in PANSS Total scores in the switch group compared to the stay group. Specifically, this effect was quantified as an increase in PANSS Total scores of 0.85 points per measurement occasion greater in the switch group than in the stay group. Therefore, over the seven mea- surement occasions, the model estimated that the groups would differ by about additional 5 points at the end of the study. The same results were observed for the PANSS Positive Symptoms subscale and for the General Psychopathology subscale but not for the Negative Symptom subscale. There was no difference between the two subgroups with re- gard to any adverse effect measure with the exception of the Simpson Angus scale, where there was a greater decline in scores for the stay group (Table 3).

Since stay participants received higher olanzapine equivalent daily doses than switch patients at baseline, and throughout the course of the study, we evaluated the effect of dose on these observed differences. The patterns of change remained even after we controlled for baseline dose and for dose over the entire study period, and after a sensitivity analysis in which patients on equivalent doses greater than 50 mg at baseline (11 stay and 1 switch) were eliminated.

To further explore the significant longitudinal effect of group assign- ment on PANSS Total score, we separately assessed change in scores early and later in the study within one model by using the study midpoint (assessment 4) as the pivot point. We found a significant condition-by- time interaction for the latter interval (Estimate = −1.96, SE = 0.92, p = .034), but not for the former interval (Estimate = 0.23, SE = 0.89, p = .796). During the latter half of the study, there was a greater increase

ent.

me Condition Time × condition

46 (0.75) −3.97 (3.70) −0.85 (0.42)⁎

43 (0.22) −1.40 (1.07) −0.28 (0.13)⁎

42 (0.30) −1.45 (1.26) −0.13 (0.17) 63 (0.40) −1.13 (1.94) −0.44 (0.23) 01 (0.05) −0.26 (0.19) −0.01 (0.03)

. .

0.01 (0.02) 0.08 (0.09) −0.01 (0.01) 0.04 (0.02) 0.11 (0.07) −0.02 (0.01) 0.04 (0.02)⁎ 0.05 (0.06) 0.01 (0.01) 0.01 (0.02) 0.05 (0.06) 0.00 (0.01)

0.07 (0.11) 0.05 (0.47) −0.09 (0.06) 0.02 (0.02) −0.04 (0.05) 0.00 (0.01) 0.03 (0.04) 0.09 (0.13) −0.03 (0.02) 0.05 (0.03)⁎ −0.09 (0.08) 0.01 (0.02) 0.10 (0.08) 0.26 (0.25) −0.06 (0.05) 0.06 (0.05) 0.15 (0.14) −0.04 (0.03) 0.12 (0.13) 0.88 (0.59) −0.18 (0.07)⁎

1.43 (1.15) 1.74 (4.87) 0.63 (0.64) 0.10 (0.07) 0.35 (0.35) 0.05 (0.04) 0.05 (0.09) −1.98 (1.68) 0.04 (0.05) 0.05 (0.12) −0.71 (0.39) 0.09 (0.07) 0.35 (2.22) 14.29 (9.18) −1.02 (1.25) 7.79 (5.89) −14.09 (24.48) 3.96 (3.31)

justed for antipsychotic dose at baseline. PANSS = Positive and Negative Syndrome Scale; t.

Fig. 2. A two-slope mixed effects model showing change in the PANSS Total (A), PANSS Negative (B), and PANSS General (C) scores as a function of being assigned to the switch vs. stay condition.

198 R.J. Constantine et al. / Schizophrenia Research 166 (2015) 194–200

in PANSS Total scores of 3.24 points per measurement occasion in the switch group compared to the stay group (Fig. 2).

4. Discussion

In only the third randomized clinical trial of its kind in which partic- ipants with schizophrenia on 2 antipsychotics were randomly assigned to switch to one or to stay on both, we observed that symptoms of schizophrenia increased significantly over time for the participants assigned to the switch condition compared to those assigned to the stay condition. While 81% of switch participants who completed 2 or more assessments did not switch back to antipsychotic polypharmacy, they paid a price in terms of increased positive symptoms. Through the course of the 1-year study, all-cause treatment discontinuation was also significantly higher among switch patients than stay partici- pants (42% vs. 13%).

With respect to side effects, we did not observe condition by time in- teractions for any of the side effect measures included in the study ex- cept for the Simpson Angus Scale scores which declined significantly

Table 4 Comparison of randomized controlled trials of switching from antipsychotic polypharmacy to

Essock et al., 2011 Ho

Number of investigator sites 19 1 Sample size 127 39 Study design Randomized Ra

Open-label Op Blinded assessment Bli

Diagnosis Schizophrenia and schizoaffective disorder

Sc

Study duration 6 months 6 m Baseline average PANSS Total score 71 67 Average baseline BMI 31.6 No Antipsychotic agents allowed All oral antipsychotics other than

clozapine Al th

Injectables excluded Inj

Baseline average haloperidol equivalent daily dose

6.7 mg 12

All-cause discontinuation rates in switch vs. stay

31 % vs. 14 % 15 p b 0.05 p =

Change in PANSS Total in switch vs. stay groups over study

No significant difference between groups over study

No be

Change in Side-effects in switch vs. stay groups over study

Greater reduction in BMI in switch-to-monotherapy vs. stay-on-2 antipsychotics

No

Difference in cognitive function between groups at end of study

No data provided Im sw

more over time in the stay group than in the switch group despite the consistently higher antipsychotic dose received by stay patients. Simi- larly, there were no significant differences between groups in changes over time in BMI, blood pressure, or laboratory values including hemo- globin A1c, triglycerides, and cholesterol.

Our findings differ in two important ways from those of Essock et al. (2011) and Hori et al. (2013) in that those two studies noted no differ- ence in symptom severity over time between switch and stay groups while observing side-effect advantages for patients switched to antipsy- chotic monotherapy in comparison to those who remained on two anti- psychotic agents. Some important differences between the design of our study and that of those two studies need to be considered (Table 4).

First, our sample may have been weighted towards greater chronic- ity and treatment-refractoriness than those of Essock et al. and Hori et al. despite the similarity in baseline PANSS scores. Our sample was se- lected from community mental health centers in contrast to the other two samples that were derived primarily from academic medical cen- ters. The average duration of antipsychotic polypharmacy episode in our sample at study enrollment was 2.5 years. The relative chronicity of our sample may help explain the lack of differential change over time in the side effects experienced by switch vs. stay participants as the side effect burden of both groups may have been well established before study entry.

Second, half the patients in our sample were receiving either cloza- pine or a long-acting injectable antipsychotic as one of their 2 antipsy- chotics at baseline. This is in contrast to the other two samples that excluded patients on these agents.

Third, we followed participants for 360 days rather than 180 days as in the other two studies. We observed that from baseline to 180-day follow-ups, the mean total PANSS score was relatively stable, a 6- month finding that is consistent with that of both the Essock et al. and Hori et al. studies. The differences in symptom control between the stay vs. switch groups only emerged in the second 6 months of our trial. This is similar to the findings of Covell et al. (2012), who observed no difference between stayers and switchers in the first 6 months but did observe significantly greater all-cause discontinuation among switchers than stayers in the second 6 months of a 1-year study of the effectiveness of switching from a first-generation long-acting antipsy- chotic agent to a second-generation agent in stable patients with schizophrenia.

monotherapy in schizophrenia.

ri et al., 2013 Present study

7 104

ndomized Randomized en-label Open-label nded assessment Blinded assessment hizophrenia Schizophrenia and schizoaffective

disorder onths 1 year

69 t provided 32.2 l oral antipsychotics other an clozapine

All oral antipsychotics including clozapine

ectables excluded Long-acting injectable antipsychotics allowed

.5 mg 37.0 mg

% vs. 0 % 42 % vs. 13 % 0.08 p b 0.01

significant difference tween groups over study

Significantly greater increase in PANSS Total in switch vs. stay

ne noted No difference in BMI, lipids, HbA1C, EPS, or TD between groups

provement in cognitive function in the itch group but not the stay group

No data provided

199R.J. Constantine et al. / Schizophrenia Research 166 (2015) 194–200

These findings suggest that the answer to the question of the effec- tiveness of switching patients with schizophrenia from 2 concomitant antipsychotics to a single agent may depend on the characteristics of the patients, the time period observed, and the specific antipsychotic agents involved. For chronically ill patients who in the judgment of their treating physicians are stable on a 2 antipsychotic regimen, the risks of switching to one may outweigh the benefits. For such patients, clinicians must exercise caution if they consider transitioning them from antipsychotic polypharmacy to monotherapy and closely monitor them during the process (Tandon et al., 2006).

It should be noted that our results do not provide information on the risks and benefits of initiating antipsychotic polypharmacy in patients with schizophrenia. They do suggest, however, that since it may be dif- ficult to get patients off antipsychotic polypharmacy once it has become a sustained treatment strategy, trials of clozapine and injectable anti- psychotics should precede consideration of antipsychotic polypharmacy (Brissos et al., 2014; Constantine et al., 2010, 2013; Goren et al., 2013; Kales et al., 1999; Tandon, 2011). Of note, a recent analysis of outcomes of Medicaid beneficiaries with schizophrenia noted that those on cloza- pine monotherapy had better outcomes than those on antipsychotic polypharmacy (Velligan et al., 2015).

5. Limitations

The results of our study should be interpreted with caution for sev- eral reasons. First, we did not measure antipsychotic blood levels as an indicator of antipsychotic adherence but instead relied on participant responses to the Medication Adherence Questionnaire (Morisky et al., 1986 ). Second, the switch and stay groups differed significantly on baseline antipsychotic dose with stay participants receiving an average of 8 mg. more olanzapine equivalents per day, although this difference did not appear to explain our results. Additionally, the study relied on olanzapine equivalent dose formulas based on expert consensus (Gardner et al., 2010) and these equivalencies may not be entirely accu- rate. Third, our sample did not include patients exhibiting violent be- havior or those whose symptoms precluded them from legitimately consenting to participate in the study.

6. Conclusion

Our study provides some guidance for clinicians regarding how to proceed with schizophrenia patients who are currently on 2 antipsy- chotic medications. For stable patients with long histories of antipsy- chotic treatment, the risks of switching may exceed the benefits and should be approached with great caution. The symptoms as well as the side effects of patients who are switched from 2 to one antipsychotic should be carefully monitored and measured over at least 12 months with the dosing of antipsychotic agents adjusted as needed.

Funding source This study was funded by the Florida Agency for Health Care Administration. They had

no role in the design of the study, the analysis of the results, or development of the manuscript.

Contributors Robert Constantine and Rajiv Tandon designed the study. Robert Constantine and

Marie McPherson coordinated the conduct of the study. Ross Andel provided statistical consultation throughout the study and conducted data analysis. All authors participated in the development of the manuscript and have approved the final manuscript.

Conflict of Interest The authors have no relevant conflicts of interest.

Acknowledgement We wish to thank Christina Guenther for her invaluable assistance in executing the

study Research sites and Principal Investigators included Lakeview Center (Edward Mobley,

M.D.), Directions for Living (James Zenel, M.D., and Sangita Desai, M.D.), Peace River Cen- ter (Jorge Dorta-Duque, M.D.), Boley Behavioral Health (Randy Hemsath, M.D.), Meridian

Behavioral Health (Fernando Castro, M.D.), Henderson Mental Health (Bhagerathy Sahasranaman, M.D.), and Lifestream Behavioral Health (Thomas Valente, M.D.).

References

American Psychiatric Association, 2000. Diagnostic and Statistical Manual of Mental Disorders-4th Edition Text Revision (DSM-IV-TR). American Psychiatric Association, Washington D.C.

Astrachan, B.M., Harrow, M., Adler, D., Brauer, L., Schwartz, A., Schwartz, C., et al., 1972. A checklist for the diagnosis of schizophrenia. Br. J. Psychiatry 121, 529–539.

Barnes, T., 1989. A rating scale for drug-induced akathisia. Br. J. Psychiatry 154, 672–676. Barnes, T.R.E., Paton, C., 2011. Antipsychotic polypharmacy in schizophrenia. CNS Drugs

25, 383–399. Brissos, S., Veguilla, M.R., Taylor, D., Balanza-Martinez, V., 2014. The role of long-acting an-

tipsychotics in schizophrenia: a critical appraisal. Ther. Adv. Psychopharmacol. 4, 198–219.

Buchanan, R.W., Kreyenbuhl, J., Kelly, D.L., et al., 2010. The 2009 schizophrenia PORT psy- chopharmacological treatment recommendations and summary statements. Schizophr. Bull. 36, 71–93.

Constantine, R., Andel, R., Tandon, R., 2010. Trends in adult antipsychotic polypharmacy: progress and challenges in Florida's Medicaid. Community Ment. Health J. 46, 523–530.

Constantine, R., McPherson, M.A., Jones, M.E., Tandon, R., Becker, E.R., 2013. Improving psychotherapeutic medication prescribing in Florida: implementation of the Florida Medicaid Drug Therapy Management Program (MDTMP). Community Ment. Health J. 49, 33–44.

Correll, C.U., Rummel-Kluge, C., Corves, C., Kane, J.M., Leucht, S., 2009. Antipsychotic com- binations vs monotherapy in schizophrenia: a meta-analysis of randomized con- trolled trials. Schizophr. Bull. 35, 443–457.

Covell, N.H., McEvoy, J.P., Schooler, N.R., Stroup, T.S., Jackson, C.T., Rojas, I.A., Essock, S.M., 2012. Effectiveness of switching from long-acting injectable fluphenazine or haloper- idol decanoate to long-acting injectable risperidone microspheres: an open-label, randomized controlled trial. J. Clin. Psychiatry 73, 669–675.

Essock, S.M., Covell, N.H., Davis, S.M., Stroup, T.S., Rosenheck, R.A., Lieberman, J.A., 2006. Effectiveness of switching antipsychotic medications. Am. J. Psychiatry 163, 2090–2095.

Essock, S.M., Schooler, N.R., Stroup, T.S., McEvoy, J.P., Rojas, I., Jackson, C., et al., 2011. Ef- fectiveness of switching from antipsychotic polypharmacy to monotherapy. Am. J. Psychiatry 168, 702–708.

Fleischhacker, W.W., Uchida, H., 2012. Critical review of antipsychotic polypharmacy in the treatment of schizophrenia. Int. J. Neuropsychopharmacol. 2, 1–11.

Florida Drug Therapy Management Program for Behavioral Health, 2011. Florida Best Prac- tice Psychotherapeutic Medication Guidelines for Adults. medicaidmentalhealth.org.

Gallego, J.A., Nielsen, J., de Hert, M., Kane, J.M., Correll, C.U., 2012. Safety and tolerability of antipsychotic polypharmacy. Expert Opin. Drug Saf. 11, 527–542.

Gardner, D.M., Murphy, A.L., O'Donnell, H., Centorrino, F., Baldessarini, R.J., 2010. Interna- tional consensus study of antipsychotic dosing. Am. J. Psychiatry 167, 686–693.

Gilmer, T.P., Dolder, C.R., Folsom, D.P., Mastin, W., Jeste, D.V., 2007. Antipsychotic polypharmacy trends among Medicaid beneficiaries with schizophrenia in San Diego county. 1999-2004. Psychiatr. Serv. 58, 1007–1010.

Godleski, L.S., Kerler, R., Barber, J.W., Glick, J.I., Kellog, E., Vieweg, W.V.R., et al., 1989. Mul- tiple versus single antipsychotic drug treatment in chronic psychosis. J. Nerv. Ment. Dis. 177, 686–689.

Goren, J.L., Meterko, M., Williams, S., Young, G.J., Baker, E., Chou, C.-H., et al., 2013. Anti- psychotic prescribing pathways, polypharmacy, and clozapine use in treatment of schizophrenia. Psychiatr. Serv. 64, 527–533.

Guy, W., 1976. The ECDEU assessment manual for psychopharmacology – Revised. U.S. Department of Health, Education, and Welfare, Alcohol, Drug Abuse, Mental Health Administration, NIMH Psychopharmacology Research Branch, Rockville, MD.

Hatta, K., Otachi, T., Fujita, K., Morikawa, F., Ito, S., Tomiyama, H., et al., 2014. Antipsychotic switching versus augmentation among early non-responders to risperidone or olanzapine in acute-phase schizophrenia. Schizophr. Res. 158, 213–222.

Hori, H., Yoshimura, R., Katsuki, A., Sugita, A.-I., Atake, K., Nakamura, J., 2013. Switching to antipsychotic monotherapy can improve attention and processing speed, and social activity in chronic schizophrenia patients. J. Psychiatr. Res. 47, 1843–1848.

Kales, H.C., DeQuardo, J.R., Tandon, R., 1999. Combined electroconvulsive therapy and clo- zapine in treatment-refractory schizophrenia. Prog. Neuropsychopharmacol. Biol. Psychiatry 23, 547–556.

Katona, L., Czobor, P., Bitter, I., 2014. Real-world effectiveness of antipsychotic monother- apy vs. polypharmacy in schizophrenia: to switch or to combine? A nation-wide study in Hungary. Schizophr. Res. 152, 246–254.

Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261–276.

Li, Q., Xiang, Y.-T., Su, Y.-A., Shu, L., Yu, X., Chiu, H.F.K., et al., 2015. Antipsychotic polypharmacy in schizophrenia patients in China and its association with treatment satisfaction and quality of life. Findings of the third national survey on use of psycho- tropic medications in China. Aust. N. Z. J. Psychiatry 49, 129–136.

Littell, R.C., Milliken, G.A., Stroup, W.W., Wolfinger, R.D., Schabenberger, O., 2006. SAS for mixed models. Second edition. SAS Institute Inc., Cary, NC.

Marder, S.R., Essock, S.M., Miller, A.L., et al., 2004. Physical health monitoring of patients with schizophrenia. Am. J. Psychiatry 161, 1334–1349.

Mojtabai, R., Olfson, M., 2010. National trends in psychotropic medication polypharmacy in office-based psychiatry. Arch. Gen. Psychiatry 67, 26–36.

Morisky, D.E., Green, L.W., Levine, D.M., 1986. Concurrent and predictive validity of a self- reported measure of medication adherence. Med. Care 24, 67–74.

200 R.J. Constantine et al. / Schizophrenia Research 166 (2015) 194–200

National Collaborating Centre for Mental Health, 2009. Schizophrenia, 2009. Core Inter- ventions in the Treatment and Management of Schizophrenia in Adults in Primary and Secondary Care (Update). Clinical Guidelinep. 82.

Park, S.-C., Lee, M.-S., Kang, S.-G., Lee, S.-H., 2014. Patterns of antipsychotic prescription to aptients with schizophrenia in Korea: results from the Health Insurance Review and Assessment Service – National patient sample. J. Korean Med. Sci. 29, 719–728.

Simpson, G.M., Angus, J.W., 1970. A rating scale for extrapyramidal side effects. Acta Psychiatr. Scand. Suppl. 212, 11–19.

Sun, F.F., Stock, E.M., Copeland, L.A., Zeber, J.E., Ahmedani, B.K., Morisette, S.B., 2014. Polypharmacy with antipsychotic drugs in patients with schizophrenia: trends in multiple health care systems. Am. J. Health Syst. Pharm. 71, 728–738.

Suzuki, T., Uchida, H., Tanaka, K.F., Nomura, K., Takano, H., Tanabe, A., et al., 2004. Revising polypharmacy to a single antipsychotic regimen for patients with schizophrenia. Int. J. Neuropsychopharmacol. 7, 133–142.

Tandon, R., 2011. Antipsychotics in the treatment of schizophrenia. J. Clin. Psychiatry 72, 1–8.

Tandon, R., Halbreich, U., 2003. The second-generation antipsychotics: similar efficacy but different neuroendocrine side-effects. Psychoneuroendoc. 28 (Suppl.1), 1–7.

Tandon, R., Targum, S.D., Nasrallah, H.A., Ross, R., 2006. Strategies for maximizing clinical effectiveness in the treatment of schizophrenia. Psychiatr. Serv. 12, 348–363.

Tandon, R., Moller, H.-J., Belmaker, R.H., et al., 2008. World Psychiatry Association Pharmacopsychiatry Section statement on comparative effectiveness of antipsy- chotics in the treatment of schizophrenia. Schizophr. Res. 100, 20–38.

Tandon, R., Nasrallah, H.A., Keshavan, M.S., 2010. Schizophrenia, “Just the facts” 5. Treat- ment and prevention. Schizophr. Res. 122, 1–23.

Tani, H., Uchida, H., Suzuki, T., Fujii, Y., Mimura, M., 2013. Interventions to reduce antipsy- chotic polypharmacy: a systematic review. Schizophr. Res. 143, 215–220.

Tsutsumi, C., Uchida, H., Suzuki, T., Watanabe, K., Takeuchi, H., Nakajima, S., et al., 2011. The evolution of antipsychotic switch and polypharmacy in natural practice—a longi- tudinal perspective. Schizophr. Res. 130, 40–46.

Velligan, D.I., Carroll, C., Lage, M.J., Fairman, K., 2015. Outcomes of Medicaid beneficiaries with schizophrenia receiving clozapine only or antipsychotic combinations. Psychiatr. Serv. 2015. http://dx.doi.org/10.1176/appi.ps.201300085.

Young, S.L., Taylor, M., Lawrie, S.M., 2015. “First do no harm”. A systematic review of the prevalence and management of antipsychotic adverse effects. J. Psychopharmacol. 29, 351.