Author's response to reviews Title: Predictors of psychiatric hospitalization during 6 months of maintenance treatment with olanzapine long-acting injection: post hoc analysis of a randomized, double-blind study Authors:

Background: Hospitalization is a costly and distressing event associated with relapse during schizophrenia treatment. No information is available on the predictors of hospitalization during maintenance treatment with olanzapine long-acting injection (olanzapine-LAI) or how the risk of hospitalization differs between olanzapine-LAI and oral olanzapine. This study aimed to identify the predictors of psychiatric hospitalization during maintenance treatment with olanzapine-LAI and assessed four parameters: hospitalization prevalence, incidence rate, duration, and the time to first hospitalization. Olanzapine-LAI was also compared with a sub-therapeutic dose of olanzapine-LAI and with oral olanzapine. Methods: This was a post hoc exploratory analysis of data from a randomized, doubleblind study comparing the safety and efficacy of olanzapine-LAI (pooled active depot groups: 405 mg/4 weeks, 300 mg/2 weeks, and 150 mg/2 weeks) with oral olanzapine and sub-therapeutic olanzapine-LAI (45 mg/4 weeks) during 6 months’ maintenance treatment of clinically stable schizophrenia outpatients (n = 1064). The four hospitalization parameters were analyzed for each treatment group. Within the olanzapine-LAI group, patients with and without hospitalization were compared on baseline characteristics. Logistic regression and Cox’s proportional hazards models were used to identify the best predictors of hospitalization. Comparisons between the treatment groups employed descriptive statistics, the Kaplan–Meier estimator and Cox’s proportional hazards models. Results: Hospitalization was best predicted by suicide threats in the 12 months before baseline and by prior hospitalization. Compared with sub-therapeutic olanzapine-LAI, olanzapine-LAI was associated with a significantly lower hospitalization rate (5.2% versus 11.1%, p < 0.01), a lower mean number of hospitalizations (0.1 versus 0.2, p = 0.01), a shorter mean duration of hospitalization (1.5 days versus 2.9 days, p < 0.01), and a similar median time to first hospitalization (35 versus 60 days, p = 0.48). Olanzapine-LAI did not differ significantly from oral olanzapine on the studied hospitalization parameters. Conclusions: In clinically stable schizophrenia outpatients receiving olanzapine-LAI maintenance treatment, psychiatric hospitalization was best predicted by a history of suicide threats and prior psychiatric hospitalization. Olanzapine-LAI was associated with a significantly lower incidence of psychiatric hospitalization and shorter duration of hospitalization compared with sub-therapeutic olanzapine-LAI. Olanzapine-LAI did not differ significantly from oral olanzapine on hospitalization parameters. 4. In many parts of the manuscript, hospitalization and relapse were discussed in a parallel manner without sufficient attention to their difference. The authors need to organize the discussion by focusing their difference. Reply: As suggested, we have better organized the Background and Discussion sections by differentiating the relapse and hospitalization. The Discussion section now states: “The risk of hospitalization – as assessed here using four hospitalization parameters – however, was not found to differ significantly between the oral and LAI formulations of olanzapine. This finding is likely driven by the study design, a randomized clinical trial (RCT) of stable, mildly ill and mostly adherent outpatients. In this sample, the hospitalization rate was relatively low, which lowers the power of finding significant findings. Although non-adherent patients are typically the ones chosen for depot formulations in clinical practice, they are often reluctant to enroll in RCTs. This interpretation seems consistent with a recent study that found differences in effectiveness when comparing results from RCTs and observational studies, with observational studies finding larger medication differences [20]. A large review of RCTs comparing oral with LAI treatment found significant differences in relapse rates between RCTs of depot and oral antipsychotics, but did not find differences in hospitalization rates or other outcome parameters [18]. As mentioned above, naturalistic studies have shown differences in hospitalization rates between oral and depot formulations. For example, Tiihonen and colleagues conducted retrospective analyses of Finnish databases and found depot therapy to be associated with a significantly lower risk of hospital admission compared with oral formulations of the same compounds [11,12]. In another similar study in Hungary, depot antipsychotics were found to have a lower hospitalization rate than many other oral antipsychotics [19]. Similar findings were observed in the Schizophrenia Outpatients Health Outcomes (SOHO) Study [35], a large pan-European, naturalistic, prospective, observational study of schizophrenia patients. That analysis of SOHO focused on non-adherent patients who were initiated on typical antipsychotics in oral or depot formulations (n = 431) and found that patients initiated on depot formulations had a significantly lower rate of hospitalization and a lower mean number of hospitalizations following 6 months of treatment. The potential confounding impact of clinical trials versus naturalistic practice settings will require further research to clarify the relative advantage of depot versus oral atypical antipsychotics in reducing the risk of hospitalization among patients with schizophrenia. Observational studies may be better suited to study the impact of depot therapy on treatment outcomes among non-adherent patients because these patients tend not to participate in RCTs.” New references added: 19. Bitter I, Katona L, Zámbori J, Takács P, Fehér L, Diels J, Bacskai M, Lang Z, Gyáni G, Czobor P: Comparative effectiveness of depot and oral second generation antipsychotic drugs in schizophrenia: A nationwide study in Hungary. Eur Neuropsychopharmacol 2013 Mar 7. [Epub ahead of print] 20. Kirson NY, Weiden PJ, Yermakov S, Huang W, Samuelson T, Offord SJ, Greenberg PE, Wong BJ: Efficacy and effectiveness of depot versus oral antipsychotics in schizophrenia: synthesizing results across different research designs. J Clin Psychiatry 2013, 74:568–575. Discretionary Revisions: 5. The authors included solely Olanzapine LAI in the title, and explain the finding of Olanzapine LAI with more words than that of oral Olanzapine in Conclusion section though it could not be determined which was more excellent in terms of the findings of this study. . Is it a neutral and balanced attitude? Reply: The primary objective of this study was to provide novel information about olanzapine-LAI, aiming to identify the predictors of psychiatric hospitalization in the maintenance treatment of schizophrenia with olanzapine-LAI and compare olanzapineLAI with a sub-therapeutic dose of olanzapine-LAI on hospitalzation parameters. A secondary objective was to compare olanzapine-LAI with oral olanzapine on the hospitalization parameters. The title of the manuscript is consistent with the study’s primary objective. Please also note that the title of the manuscript is already quite long (“Predictors of psychiatric hospitalization during 6 months of maintenance treatment with olanzapine long-acting injection: post hoc analysis of a randomized, double-blind study”). We believe that the addition of information about the secondary comparison between olanzapine-LAI and oral olanzapine to the title would have made it much too long and cumbersome for the reader.