Objective: This study was designed to determine the degree to which second-generation antipsychotics were used off-label for patients with bipolar disorder (BPD) and to identify patient characteristics which were most associated with off-label second-generation antipsychotic use. Methods: The multi-state (PHARMetrics) medical claims database was used to evaluate105,771 adult patients with a diagnosis of bipolar disorder. The study period was between 1998-2002. The off-label use of a second-generation antipsychotic medication was defined as a patient either receiving olanzapine before March 2000 or any second-generation antipsychotic, other than olanzapine, during the entire study period. Olanzapine was differentiated from the others in its class because it was approved for use in BPD in 2000. Multivariate logistic regression analysis was used to assess the risk of receiving a drug off-label. Results: Sixty-three percent of the patients were female, and the mean age was 40.3 years. Eleven percent of patients were on lithium, 25% received other anticonvulsants, 34% were treated with antidepressants, and 10.5% (7.1% off-label) took second-generation antipsychotics. A higher risk of off-label use was associated with psychiatry specialist prescribers (Odds Ratio=1.52, 95% Confidence Interval 1.44-1.59) and certain comorbidities such as substance abuse (OR=1.51, 1.38-1.66), anxiety disorder (OR=1.20, 1.14-1.26), diabetes mellitus (OR=1.26, 1.16-1.37), cerebral vascular disease (OR=1.26, 1.10-1.45), and hypertension (OR=1.12, 1.05-1.20). Conclusion: The off-label use of second-generation antipsychotics in treating BPD was fairly common from 1998-2002 and their use was associated with some key clinical factors. Our results add to a growing literature that relationship links the use of newer antipsychotics with diabetes, hypertension, obesity, and cerebral-vascular disease.