Hospitalization risks in the treatment of bipolar disorder: comparison of antipsychotic medications


  • FG and R-HW have been consultants for AstraZeneca Pharmaceuticals LP, Janssen, Schering Plough, Novartis, Aventis, and Merck. KR is an employee of AstraZeneca Pharmaceuticals LP. JFG has received research grants from and/or has served as a consultant for Abbott Laboratories, AstraZeneca, Bristol Myers Squibb, Eli Lilly & Co., GlaxoSmithKline, Janssen Pharmaceutica, and Pfizer. This study received financial support from AstraZeneca Pharmaceuticals LP.

Frank Gianfrancesco, PhD, HECON Associates, 9833 Whetstone Drive, Montgomery Village, MD 20886, USA. Fax: +1 301 990 7641; e-mail:


Objectives:  This study compared the relative risk for hospitalization of patients with bipolar and manic disorders receiving atypical and typical antipsychotics.

Methods:  This retrospective study was based on administrative claims data extracted from the PharMetrics database during 1999 through 2003. Comparisons were made among atypical antipsychotics (risperidone, olanzapine, quetiapine or ziprasidone), as well as between each of these versus a combined group of the leading typical antipsychotics. Relative risk for hospitalization was estimated with Cox proportional regression, which adjusted for differences in patient characteristics.

Results:  Risperidone and olanzapine demonstrated higher risks for hospitalization than quetiapine [hazard ratio (HR) 1.19, p < 0.05 for both], translating into higher annual mental health inpatient charges of $260 per patient. Risperidone and olanzapine also showed higher estimated risks than ziprasidone, which approached the p < 0.05 threshold. Differences between each of the atypicals and the combined typicals were not significant. Patients with putative rapid cycling had a threefold greater risk for hospitalization than other patients with bipolar disorder. In these patients, comparisons among atypical antipsychotics showed that risperidone had a significantly higher hospitalization risk than olanzapine (HR 3.31, p < 0.05), resulting in higher annual mental health inpatient charges of $4,930 per patient.

Conclusions:  In the treatment of bipolar and manic disorders, risperidone and olanzapine were associated with a higher risk for hospitalization than quetiapine, and possibly ziprasidone. In the treatment of putative rapid cyclers, olanzapine was associated with a lower risk for hospitalization than risperidone.