Atypical antipsychotic drugs and diabetes mellitus in the US Food and Drug Administration Adverse Event database: a systematic Bayesian signal detection analysis.

BACKGROUND: Prior literature suggests that the risk of diabetes-related adverse events (DRAEs) differs between atypical antipsychotics. The present study evaluated the potential association between atypical antipsychotics or haloperidol and diabetes using data from the FDA AERS database.
METHODS: Analysis of AERS data was conducted for clozapine, risperidone, olanzapine, quetiapine, ziprasidone, aripiprazole or haloperidol with 24 DRAEs from the Medical Dictionary for Regulatory Activities using a Multi-item Gamma Poisson Shrinker (MGPS) data-mining algorithm. Using MGPS, adjusted reporting ratios (Empiric Bayes Geometric Mean or EBGM) and 90% confidence intervals (CIs; EB05-EB95) were calculated to estimate the degree of drug-event association relative to all drugs and events. Logistic regression odds ratios and 90% CIs (LR05-LR95) were calculated for diabetes mellitus events.
RESULTS: All six atypicals had an EB05 >/= 2 for at least one DRAE. The most common event was diabetes mellitus (2,784 cases). Adjusted reporting ratios (CIs) for diabetes mellitus were: olanzapine 9.6 (9.2-10.0; 1306 cases); risperidone 3.8 (3.5-4.1; 447 cases); quetiapine 3.5 (3.2-3.9; 283 cases); clozapine 3.1 (2.9-3.3; 464 cases); ziprasidone 2.4 (2.0-2.9; 74 cases); aripiprazole 2.4 (1.9-2.9; 71 cases); haloperidol 2.0 (1.7-2.3; 139 cases). Logistic regression odds ratios agreed with adjusted reporting ratios.
CONCLUSIONS: In the AERS database, lower associations with DRAEs were seen for haloperidol, aripiprazole and ziprasidone, and higher associations were seen for olanzapine, risperidone, clozapine and quetiapine. Our findings support differential risk of diabetes across atypical antipsychotics, reinforcing the need for metabolic monitoring of patients taking antipsychotics.