Background/AimsWeight gain is an undesirable side effect of treatment with 2nd/3rd generation antipsychotic drugs which may have genetic determinants. A number of candidate genes have been analyzed but no genome-wide association studies (GWAS) have yet been reported. The purpose of this study was to determine the rate of extreme weight gain (EWG) among adult users of 2nd/3rd generation antipsychotic medications for future GWAS.MethodsA standardized dataset was extracted on antipsychotic medications users at Group Health and Geisinger Health System from 2004–2010 that included demographics, enrollment, vitals, and pharmacy data. Electronic health record search algorithms were used to identify adult subjects with body weight increases of 15%+ within one year of the initiation of new 2nd/3rd generation antipsychotic drugs. To adjudicate the association of EWG with drug treatment, SAS was used to plot and visually inspect a single graph for each episode that encompassed all weight measurements, antipsychotic orders/fills, pregnancy- and cancer-related visits, and steroid and insulin orders/fills over time. The rate of EWG was then compared among the common types of antipsychotic medications.ResultsThere were 7128 episodes of antipsychotic use that qualified for analysis (5251 patients). The most common medication types were quetiapine (n=2543), risperidone (n=1817), aripiprazole (n=1315), olanzapine (n=971), and ziprasidone (n=413). The least common medications were clozapine (n=37), olanzapine/fluoxetine (n=23), and paliperidone (n=9). EWG was identified for 275 of the 7128 episodes (3.86 per 100 episodes). EWG was significantly different among common medication types (p-value<0.0001). EWG was highest for olanzapine (5.46 per 100 episodes, 95% CI=[4.77, 6.24]), followed by quetiapine (3.93 per 100 episodes, 95% CI=[3.25, 4.75]), risperidone (3.58 per 100 episodes, 95% CI=[3.14, 4.07]), and aripiprazole (3.50 per 100 episodes, 95% CI=[2.65, 4.62]). For ziprasidone, EWG was 2–3 fold lower (1.69 per 100 episodes, 95% CI=[1.46, 1.96]). EWG was not significantly different between Geisinger and Group Health (4.43 per 100 episodes versus 3.65 per 100 episodes, p=0.157).DiscussionRates of EWG differed significantly by type of 2nd/3rd generation antipsychotic medication, which was consistent in two geographically diverse populations. Future GWAS may help determine if genetic determinants of EWG are drug specific.