Four RCTs (838 patients) directly compared the drop-out rates of olanzapine and risperidone, while two RCTs (716 patients) directly compared the response rates of olanzapine and risperidone. Data from five RCTs of olanzapine (928 patients) and three studies of risperidone (290 patients) were included in the assessment of maintenance of response.
Direct comparison of drop-out rates with olanzapine and risperidone.
The risk of drop-out was significantly greater with risperidone than with olanzapine; the OR (fixed-effect model) was 1.50 (95% CI: 1.13, 2.00, P=0.006). No statistically significant heterogeneity was detected (P=0.80). The results were similar when using the log rank test (P=0.006). The NNT to avoid one drop-out with olazapine compared with risperidone was 11.
The results were also similar when using the censored approach (Cox analysis); the RR was 1.35 (95% CI: 1.08, 1.69, P=0.008). No statistically significant differences between the studies were detected (P for variable study =0.91)
Direct comparison of response rates with olanzapine and risperidone.
The response rates of risperidone and olanzapine were similar; the OR (fixed-effect model) was 1.03 (95% CI: 0.76, 1.39). No statistically significant heterogeneity was detected (P>0.90).
Maintenance of response.
Kaplan Meier curves showed a slightly worse pattern of maintenance of response for risperidone compared with olanzapine (the curves were presented). Studies of risperidone were of a longer duration than studies of olanzapine: 365 to 800 days with risperidone versus 196 to 365 days with olanzapine.