Early predictors of weight gain risk during treatment with olanzapine: analysis of pooled data from 58 clinical trials

This analysis evaluated the usefulness of different predictors in identifying patient risk of substantial weight gain (SWG) during olanzapine treatment. Data were from 58 studies with 3826 patients diagnosed with schizophrenia, schizophrenia spectrum disorders, bipolar mania, bipolar depression, or borderline personality disorder. The primary definition for SWG was gaining/=12% of baseline weight by endpoint (30 weeks +/-5 weeks); other definitions of SWG were also examined. Potential predictors of SWG included baseline patient characteristics, weight change, and percent weight change at Weeks 1, 2, 3, and 4 after olanzapine initiation. To facilitate model building and validation, the data set was randomly partitioned into training (N = 1912), validation (N = 1149), and test (N = 765) sets and 2 complementary analytic techniques were used: logistic regression with stepwise variable selection followed by receiver operating characteristic analysis for evaluation of resulting candidate models and decision trees. Approximately 24% of patients gained/=12% of their initial weight, about 30% gained/=10%, and 45% gained/=7% or/=5 kg by the 30-week endpoint. Baseline covariates significantly and positively associated with probability of SWG were lower baseline body mass index, younger age, female sex, United States residency, and African ethnicity. Early weight changes substantially improved the prediction of the risk for longer-term SWG. These results confirm that cut-offs for weight gain during the first 4 weeks of treatment may be useful in evaluating SWG risk for an individual patient.