Improved outcomes following a switch to olanzapine treatment from risperidone treatment in a 1-year naturalistic study of schizophrenia patients in Japan

Authors


Naohiro Nakahara, BSc, Eli Lilly Japan K.K., Sannomiya Plaza Building, 7-1-5 Isogami-dori, Chuo-ku, Kobe 651-0086, Japan. Email: nakahara_naohiro@lilly.com

Abstract

Aims:  This study assessed clinical and functional outcomes following a switch from risperidone to olanzapine in a 1-year naturalistic study of schizophrenia patients in Japan.

Methods:  We used data from a large 1-year prospective, multicenter, observational non-interventional study of individuals who were initiated on olanzapine for the treatment of schizophrenia in Japan. Current analyses focused on patients who were switched at study entry from risperidone to olanzapine (n = 258). Repeated measures analysis was employed to assess outcomes on validated measures.

Results:  At study entry, 45% were inpatients and 55% outpatients. Participants were in their early 40s with mean illness duration of 14 years. Approximately half were male. Most were switched from risperidone to olanzapine due to poor medication efficacy (67.8%) rather than medication intolerability (29.1%). Most patients (67.8%) completed the 1-year study. Patients experienced clinically and statistically significant (P < 0.05) improvements in global symptom severity, positive, negative, depressive, and cognitive symptoms, health-related quality of life, and paid work rates. Most patients (59.2%) demonstrated treatment response to olanzapine and 43.4% experienced symptom remission. Mean weight gain was 2.19 kg, with one-third of patients (33.3%) experiencing clinically significant weight gain (≥7%).

Conclusions:  In this 1-year naturalistic study, inpatients and outpatients who were switched from risperidone to olanzapine experienced clinically and statistically significant improvements in their clinical and functional outcomes. One-third of all patients experienced clinically significant weight gain. Current findings highlight the favorable benefit-to-risk profile of switching to olanzapine therapy following treatment failure on risperidone among patients with schizophrenia in Japan.

SCHIZOPHRENIA IS A chronic and debilitating mental illness that negatively impacts many aspects of patients' lives. Antipsychotic medications are the cornerstone of treatment for both reducing acute symptoms and for preventing future relapses.1–3 Long-term treatment with antipsychotics in usual care settings is a dynamic process that often involves changes in medications,4,5 due primarily to medication efficacy,4,6 but also to patient preferences, medication intolerability, and non-adherence.

Antipsychotics appear to differ in their pharmacology, efficacy, safety, and tolerability.7,8 The process of identifying an antipsychotic medication with the proper balance of effectiveness and tolerability for a specific patient at a given point in time can be challenging. Some individuals with schizophrenia who do not respond to one atypical antipsychotic, will respond to a different one,9 but relatively little information is available to clinicians regarding which antipsychotic to switch to after one fails. Clinicians appear to consider the reasons the previous antipsychotic failed when selecting the next antipsychotic.10

In Japan, risperidone is the most commonly used atypical antipsychotic to treat schizophrenia followed by olanzapine.11,12 Given the common use of these two medications, when a patient fails to respond adequately to risperidone, olanzapine may often be the next choice. However, little is known about the treatment outcomes for patients in usual care who fail risperidone therapy and are switched to olanzapine. Patients who fail treatment with risperidone may represent a more difficult-to-treat subpopulation.

Much of the effectiveness research for atypical antipsychotics has been completed in Europe and the USA; less is known about the outcomes for patients with schizophrenia treated in usual care in Japan. In Japan, antipsychotics tend to be used in higher doses13 and in combination,13 particularly the combination of a typical and an atypical agent.11 Furthermore, Japan has the highest level of psychiatric inpatient beds per capita.14 Many patients who are treated in the hospitals could possibly be treated in the community,14 but community-based psychiatric services to handle this load of patients are not fully developed.11 Understanding differences in outcomes for inpatient and outpatients may be of particular importance in Japan.

To better understand the outcomes for patients with schizophrenia who switch from risperidone to olanzapine in usual care in Japan, we analyzed data for a subset of patients from a larger observational study.15 The objectives of this analysis were two-fold: (i) to assess clinical, functional, and safety-related outcomes following a switch from risperidone to olanzapine in a 1-year naturalistic study of schizophrenia patients in Japan; and (ii) to compare treatment outcomes between inpatients and outpatients who were switched from risperidone to olanzapine in this 1-year naturalistic study.

METHODS

Data source

Data for this post-hoc analysis come from a large (n = 1949) multicenter, naturalistic, 1-year, post-marketing surveillance study in Japan.15 Patients were eligible for the parent study if they were diagnosed with schizophrenia based on DSM-IV criteria and initiated treatment with olanzapine. Because this was a naturalistic, observational, and non-interventional study, all treatment decisions were left to the discretion of the treating clinician. Patients may have initiated their first antipsychotic treatment with olanzapine, switched from another antipsychotic to olanzapine, or augmented their current antipsychotic treatment with olanzapine. The study enrollment began in November 2003 and was completed in July 2004. Data were collected at the baseline, 3-month, 6-month, and 12-month visits. When patients discontinued olanzapine, they discontinued participation in the study.

The study procedures were approved by the internal review boards at each of the participating institutions. Informed consent was obtained based on the rules at each participating medical facility.

Measures

In this observational study, the procedures were designed to capture only a limited amount of information. Cumbersome or invasive measures were not included. All of the outcome measures were collected at the baseline, 3-month, 6-month, and 12-month visits.

The Clinical Global Impression – Schizophrenia (CGI-SCH) is a clinician-rated measure consisting of a global severity rating and four additional symptom cluster ratings: positive, negative, cognitive, and depressive symptoms. All ratings are made on an anchored scale ranging from no symptoms (0) to severe symptoms (6).16 The concurrent validity of the CGH-SCH subscales with the corresponding subscales from the more rigorous Positive and Negative Syndrome Scale (PANSS)17 have been found to be moderate to high depending on the subscale. Correlation coefficients ranged from 0.86 for positive symptoms to 0.61 for depressive symptoms, with the remaining scales varying from 0.75 to 0.80. Inter-rater reliability has been found to be high (interclass correlation coefficients ranging from 0.73 to 0.82 for all but the depressive subscale [0.64]).16

The European Quality of Life – 5 Dimensions (EQ-5D) is a generic measure of health-related quality of life that contains five dimensions (Level of Movement, Control of Environment, Normal Activities, Pain/Discomfort, and Anxiety/Depression) and a Visual Analog Scale (VAS) of overall health. The ratings on the five dimensions are used to generate health states that have been assigned values (utilities) ranging from death (0) to perfect health (1). The VAS is a single rating from 0 to 100 that patients make of their overall health.18

In addition, the study collected a broad array of other information on treatment, functioning, and adverse events. Measures of the treatment intervention include the number of days hospitalized for schizophrenia during each study period and details on concomitant medication use. Medication information included drug name, dose, route of administration, start and stop dates, and therapeutic category (antipsychotic, anticholinergic, antidepressant, anxiolytic/hypnotic, mood stabilizer, or other). Beyond the EQ-5D, measures of functioning included employment status (including working for pay) and number of social activities in the past 4 weeks (0, 1, 2, 3, 4, or 5+). In addition to spontaneously reported adverse events, the study included ratings for dystonia/akathisia/parkinsonism, tardive dyskinesia, decreased libido, amenorrhea/other menstrual dysfunction, erectile/sexual dysfunction, gynecomastia, and galactorrhea. There was a specific inquiry at baseline into the presence of any of the following medical complications: hypertension, hyperlipidemia, hepatic dysfunction, renal dysfunction, or other. Finally, bodyweight in kg was measured at each visit.

Procedures

This analysis was restricted to individuals who had been treated with, and discontinued, risperidone prior to initiating olanzapine at the study start (n = 258). Based on the setting at the time of olanzapine initiation, patients were grouped into two cohorts: inpatients or outpatients.

Symptomatic response and remission were defined based on the CGI-SCH using previously published definitions. Response was defined as a 2-point decrease in the CGI-SCH global severity rating when the baseline rating was between 4 and 6, or a 1-point decrease when the baseline rating was between 1 and 3.19 Symptomatic remission was defined as achieving a score ≤ 2 (mild symptoms) on CGI-SCH positive, negative, cognitive, and global severity scores.20

Patients' body mass indices (BMI) were categorized based on the World Health Organization recommendations for Asians.21 The following classification was used: underweight (<18.5), normal (≥18.5, ≤23), overweight (>23, ≤30), and obese (>30).

Statistical methods

Baseline comparisons between the inpatient and outpatient groups were completed with t-tests for continuous variables and χ2-tests for categorical variables. Mixed Models for Repeated Measures (MMRM) with baseline covariates for age, sex, duration of illness, and presence of any medical complication, were used to assess changes over time in continuous variables. Comparisons of categorical variables between baseline and post-baseline time periods were completed using McNemar's test with missing observations imputed using the last observation carried forward method. Comparisons of bodyweight between inpatients and outpatients were completed using t-tests at each visit. Survival analysis was conducted for time to treatment discontinuation using log–rank tests. Survival curves were constructed using unadjusted Kaplan–Meier estimates. A sensitivity analysis was completed using the primary symptoms measure (CGI-SCH) with the 52.7% of patients who were treated with olanzapine monotherapy throughout the study. This sensitivity analysis was conducted to confirm the results were not due to the use of concomitant antipsychotic medications. All analyses were completed using sas version 9.1.3 (sas Institute, Cary, NC, USA).

RESULTS

Baseline characteristics

The patients were on average 42.0 years old, half (50.8%) were female, and they had been diagnosed with schizophrenia on average 14 years previously. The inpatients (45%) generally had a more chronic and severe symptom profile and more medical complications (see Table 1). The most common reason (67% of patients) for discontinuing risperidone was insufficient efficacy rather than medication intolerability (29.1%). Clinicians reported discontinuing risperidone immediately before initiating olanzapine for most patients (80.6%, 208/258). For patients where the switch was not recorded as immediate, only nine patients (3.5%, 9/258) had a gap greater than 5 days between discontinuing risperidone and initiating olanzapine. The average final daily dose for risperidone prior to entering the study was 4.2 ± 2.7 mg, including 3.5 ± 2.4 mg for outpatients and 5.0 ± 2.8 mg for inpatients (P < .001). Prior to initiating olanzapine, 57.0% (147/258) of the patients were treated with risperidone monotherapy, including 61.3% (87/142) of the outpatients and 51.7% (60/116) of the inpatients (P = 0.12).

Table 1.  Baseline characteristics
CharacteristicAll PatientsOutpatientsInpatientsP
(n = 258)(n = 142)(n = 116)
  1. The P-value reflects the comparison of outpatients to inpatients.

  2. n = 185. n = 254. §n = 252. n = 220.

  3. CGI-SCH, Clinical Global Impression – Schizophrenia; BMI, body mass index; EQ-5D VAS, European Quality of Life – 5 Dimensions – Visual Analog Scale.

Age (years)mean ± SD42.0 ± 15.536.3 ± 13.348.8 ± 15.2<0.001
Female(%)50.851.450.00.82
Duration of illness (years)mean ± SD14.1 ± 13.89.1 ± 9.619.5 ± 15.6<0.001
Any medical complications(%)32.925.442.20.004
Working for pay(%)9.415.12.6<0.001
CGI-SCH overallmean ± SD3.3 ± 1.03.0 ± 0.93.7 ± 1.0<0.001
CGI-SCH positivemean ± SD2.9 ± 1.42.5 ± 1.43.4 ± 1.4<0.001
CGI-SCH negativemean ± SD3.1 ± 1.32.9 ± 1.23.3 ± 1.30.002
CGI-SCH cognitivemean ± SD2.9 ± 1.32.6 ± 1.33.2 ± 1.2<0.001
CGI-SCH depressivemean ± SD1.9 ± 1.42.0 ± 1.41.7 ± 1.50.18
EQ-5D VASmean ± SD§48.1 ± 22.349.3 ± 19.546.6 ± 25.10.33
EQ-5D Utility Scoremean ± SD0.68 ± 0.190.71 ± 0.150.66 ± 0.220.03
BMImean ± SD21.8 ± 4.122.1 ± 3.721.5 ± 4.50.28
Reason for switch from risperidone (%)   0.29
 Insufficient efficacy67.868.367.2 
 Medication intolerability29.127.531.0 
 Patient request7.89.26.0 
 Non-compliance4.33.55.2 

Treatment patterns

Most patients completed the 1-year study (67.8%) with 16.7% discontinuing by the 3-month visit and 26.0% discontinuing by the 6-month visit. Inpatients and outpatients had similar discontinuation rates. Average daily dose for olanzapine during the study was 11.7 ± 6.1 mg, including 9.9 ± 5.7 mg for outpatients and 13.9 ± 5.8 mg for inpatients (P < 0.001).

The hospitalization rates decreased at each visit during the study. Those who were inpatients when initiating olanzapine spent 85.5%, 67.2%, and 62.3% of the days hospitalized for schizophrenia prior to the 3-month, 6-month, and 12-month visits, respectively. For outpatients, the respective percent of days hospitalized for schizophrenia was 5.4%, 2.5%, and 0.9%.

Combination antipsychotic treatment was common: only about half of the patients (52.7%) were treated with olanzapine monotherapy throughout the study, including 60.6% of the outpatients and 43.1% of the inpatients (P = 0.005). Antipsychotics combined with olanzapine were more likely to be typical antipsychotics (31.4%) than atypical antipsychotics (9.7%). Anticholinergic use (28.9%) did not differ between outpatients (26.9%) and inpatients (31.1%, P = 0.48), nor did antidepressants use (10.7%, including 14.3% for outpatients and 6.6% for inpatients, P = 0.06). However, the inpatients were significantly more likely to be treated with anxiolytics/hypnotics (68.4% for all patients, 82.1% for inpatients, and 56.3% for outpatients, P < 0.001) and mood stabilizers (13.3% for all patients, 22.6% for inpatients, and 5.0% for outpatients, P < 0.001). Other psychotropic and non-psychotropic medications (41.3%) were also more likely to be used by inpatients (57.5%) than outpatients (26.9%, P < 0.001).

Effectiveness outcomes

Most patients (67.8%) continued with treatment olanzapine through to the end of the 1-year study period. There were no differences between inpatients and outpatients in time to all-cause discontinuation (P = 0.98; see Fig. 1).

Figure 1.

Time to all-cause discontinuation of olanzapine for all patients, inpatients, and outpatients. There was no significant difference in time to all-cause discontinuation between the inpatient and outpatient cohorts (P = 0.98). (inline image) All Patients; (inline image) Inpatients; (inline image) Outpatients.

On all of the CGI-SCH subscales, the patients improved from baseline to each post-baseline visit after initiating treatment with olanzapine (P < 0.001). The inpatients had statistically significant (P < 0.05) higher scores on the CGI-SCH global severity and positive subscales at all visits except the 6-month visit on the positive subscale (see Fig. 2). There were no significant time-by-hospitalization interactions, indicating that the inpatients and outpatients had similar patterns of response over time.

Figure 2.

Clinical Global Impression – Schizophrenia (CGI-SCH) scores over the 1-year study period. The CGI-SCH subscale scores changed significantly from study entry to post-baseline visits (P < 0.001) for all patients, outpatients and inpatients. The Mixed Models for Repeated Measures models were adjusted for age, sex, duration of illness, and presence of any medical complication and differ from the unadjusted baseline comparisons in Table 1 for the negative and cognitive subscales. Asterisks indicate differences between inpatients and outpatients at each visit (P < 0.05). (inline image) All Patients; (inline image) Inpatients; (inline image) Outpatients.

The response rate was 59.2% and did not differ between the inpatients (57.0%) and outpatients (61.2%, P = 0.52). However the remission rate was higher for the outpatients (51.5%) than inpatients (35.6%, P = 0.02).

Patient functioning improved after initiating treatment with olanzapine. On the EQ-5D VAS, a broad rating of health-related quality of life, patients improved from 48.5 at baseline to 68.5 at the 12-month visit (P < 0.001). Inpatients (50.1–69.3, P < 0.001) and outpatients (47.0–67.7, P < 0.001) improved similarly from baseline to the 12-month visit on the EQ-5D VAS. Likewise, the utility score on the EQ-5D did not differ between inpatients and outpatients at any point in time, but improved from 0.68 (0.66 for inpatients and 0.71 for outpatients) at baseline to 0.83 (0.81 for inpatients and 0.84 for outpatients) at the 12-month visit for all patients (P < 0.001). Over the 1-year study period, the percent of outpatients working for pay increased from 14.8% to 21.3% (P = 0.02), the percent of inpatients working for pay increased from 2.8% to 6.5% (P = 0.05), and the percent of all patients working for pay increased from 9.2% to 14.4% (P = 0.003). The proportion of patients engaging in five or more social activities in the previous 4 weeks increased from 11.1% at baseline to 20.6% at the 12-month visit (P = 0.001); for outpatients the change was from 17.2% to 27.0% (P = 0.01), whereas for inpatients the change was from 3.8% to 13.1% (P = 0.007).

Tolerability outcomes

Patients who completed the study gained an average of 2.2 kg over the 1-year period (see Fig. 3). Overall, one in three patients (33.3% of 216 patients, 33.0% of 109 outpatients, and 33.6% of 107 inpatients; P = 0.92) experienced clinically significant weight gain defined as a 7% or greater increase over baseline. In terms of BMI category, 5.2% of olanzapine users (6.1% of outpatients and 4.3% of inpatients) decreased a category, 76.0% (77.6% of outpatients and 74.5% of inpatients) remained in the same category, and 18.8% (16.3% of outpatients and 21.3% of inpatients) increased a category. No cases of new-onset diabetes were reported.

Figure 3.

Observed bodyweight over the 1-year study period. Asterisks indicate differences between inpatients and outpatients at each visit (P < 0.05). (inline image) All Patients; (inline image) Inpatients; (inline image) Outpatients.

New onsets of other measured adverse events were relatively infrequent. There were 11 cases (4.3%) of new-onset dystonia/akathisia/parkinsonism, including seven outpatients (5.0%) and four inpatients (3.5%, P = 0.54); three cases (1.3%) of new-onset decreased libido, including one outpatient (0.8%) and two inpatients (2.0%, P = 0.42); two cases (1.9%) of new-onset amenorrhea/menstrual dysfunction, including one outpatient (1.5%) and one inpatient (2.5%, P = 0.72); and a single case of new-onset erectile/sexual dysfunction reported by an inpatient (1.1%, P = 0.24). There were no cases of new-onset tardive dyskinesia, gynecomastia, or galactorrhea.

Sensitivity analysis

To confirm the symptom improvements were not due to the effects of concomitant antipsychotic medications, the analysis for the CGI-SCH scales was repeated using only the 136 patients who were treated with olanzapine monotherapy throughout the study. When used as monotherapy, the average daily dose of olanzapine was 10.8 mg (outpatients: 9.8 mg; inpatients: 12.5 mg). In this sensitivity analysis, the significant changes from baseline to each post-baseline for inpatients and outpatients on the CGI-SCH global severity and each of the subscales (positive, negative, cognitive and depressive symptoms) were confirmed (P < 0.05). However, the differences between inpatients and outpatients on the CGI-SCH global severity and positive symptoms scales were no longer significant, with the exception of the baseline score on the CGI-SCH global severity.

DISCUSSION

Among Japanese inpatients and outpatients with schizophrenia who had failed risperidone therapy, the switch to olanzapine was successful for most patients, as 67.8% continued treatment with olanzapine for the full 1-year study period. Time to all-cause discontinuation has been thought of as a measure of overall effectiveness, incorporating efficacy, safety, and tolerability.22 Consistent with this view, patients showed significant improvements in symptom severity as measured by the global severity, positive symptoms, negative symptoms, cognitive symptoms, and depressive symptoms scales on the CGI-SCH. Patients' functional levels improved as measured broadly by the EQ-5D, as well as by increases in the percent of patients working for pay and the percent of patients engaging in social activities. One in three patients experienced clinically significant weight gain.

The inpatients appeared to be a unique subgroup of patients. Although the level of improvement was similar for both inpatients and outpatients, the inpatients were treated with higher doses, were more likely to be treated with antipsychotic combination therapy, had more severe symptoms throughout the study, were less likely to achieve symptomatic remission, and had lower rates of paid work and social activities. There were no significant differences between inpatients and outpatients in time to all-cause discontinuation. Our findings replicate previous research showing that inpatients have more severe symptoms,23,24 are treated with higher doses of antipsychotics,24 and have higher needs for care.23

The improvements found in this study are consistent with the findings from head-to-head studies comparing olanzapine and risperidone therapy in the treatment of patients with schizophrenia in other world geographies. A Cochrane meta-analysis synthesized the efficacy of randomized controlled trials comparing atypical antipsychotics and found olanzapine to have greater efficacy than risperidone, although the estimated difference was only 1.9 Positive and Negative Syndrome Scale total score points.8 The meta-analysis did not examine tolerability issues, but other head-to-head comparisons of olanzapine and risperidone have found a lower incidence of extrapyramidal symptoms in patients treated with olanzapine,25,26 less hyperprolactemia and corresponding sexual dysfunction,26 but greater weight gain.4,26,27 In addition, an open label randomized study found greater improvements in occupational functioning for patients treated with olanzapine than risperidone.28 These findings are also similar to those of large observational studies of usual care in healthcare systems across many different geographic regions outside of Japan.19,29

However, because this was a switching study, our findings are not directly comparable to the head-to-head studies. Unlike randomized clinical trials, when patients fail a given medication in usual care, the next treatment is not chosen at random and appears to reflect the issues resulting in the previous medication's discontinuation.10 The switching methodology more closely reflects the clinical challenge of finding the right medication for each individual patient. Some patients will fail one antipsychotic, but then may respond to a different antipsychotic that is, on average, less effective.9 This study replicates a previous switching study in Japan, in which first-episode patients were switched to olanzapine after an insufficient response to risperidone and had significant reductions in total symptoms, anxiety/depressive symptoms, and excitatory symptoms that were accompanied by an average weight increase from 62.2 kg to 65.5 kg.30 Interestingly, another switching study in Japan examined patients who failed olanzapine therapy and then switched to risperidone. When treated with risperidone, patients who failed olanzapine also had significant reductions in total symptoms, positive symptoms, and excitatory symptoms that were accompanied by significant increases in prolactin levels.31 These findings are generally consistent with the notion that atypical antipsychotics are a heterogeneous class of medications and that schizophrenia is a heterogeneous mental illness. Our results highlight the importance of finding the right medication for the right patient at a given point in time.

Limitations

This study was designed to maximize generalizability for patients with schizophrenia treated in Japan at either inpatient or outpatient settings. As a result, design considerations that favored external validity were given precedence over those that favored internal validity. In order to minimize interference with usual clinical care, brief outcome measures were used instead of more robust measures, such as the PANSS. This was a single-arm study. Although participants improved after initiating olanzapine we cannot be certain of the degree of improvement that was due to the treatment rather than simply an artifact of time. In addition, consistent with other studies of usual care in Japan, antipsychotic polypharmacy was common in this study.11,13 We cannot, therefore, be certain as to whether the improvements are due solely to olanzapine and not other medication or psychosocial treatments. However, the sensitivity analysis of the CGI-SCH scales confirmed the statistically significant improvements in the subset of patients treated with olanzapine monotherapy. Only 13.2% of the patients (23.3% of inpatients and 4.9% of outpatients) were severely ill (CGI-SCH – Global Severity ≥ 5) at baseline, therefore the results may not generalize to severely ill patients.

Conclusions

In this 1-year naturalistic study of patients with schizophrenia in Japan, inpatients and outpatients who were switched from risperidone to olanzapine experienced clinically and statistically significant improvements in their clinical and functional outcomes. One-third of all patients experienced clinically significant weight gain. Current findings highlight the favorable benefit-to-risk profile of switching to olanzapine therapy following treatment failure on risperidone among schizophrenia patients in Japan.

ACKNOWLEDGMENTS

Funding for this study was provided by Eli Lilly and Company.

Wenyu Ye, Shinji Fujikoshi, Naohiro Nakahara, and Michihiro Takahashi are full-time employees of Eli Lilly Japan, K.K. Haya Ascher-Svanum is a full-time employee of Eli Lilly and Company. This funded research by the request of Eli Lilly Japan K.K. was undertaken under the agreement of the funded research condition of publishing to Article 4 clause 1 of the University of Tokushima School of Medicine. Tetsuro Ohmori, from the University of Tokushima School of Medicine, took part in the data analysis and manuscript preparation.

Technical writing support was provided by Michael Stensland of Agile Outcomes Research Inc, Rochester, Minnesota, USA and Susan Dennett of Strategic Health Outcomes Inc, Carmel, Indiana, USA.

Ancillary