OBJECTIVES: Certain combination anti-hypertensives vary in their levels of potentially adverse metabolic effects in patients with comorbid diabetes. We examined the characteristics of patients initiating combination therapy and the effects of treatment choice on direct medical costs, using integrated medical and pharmacy claims data. METHODS: Patients with prior evidence of diabetes who were newly-treated for hypertension from January 1998 to March 2002 and continuously enrolled for at least 6 months before and after therapy initiation were selected. Patients were stratified by initial fixed combination treatment (trandolapril/verapamil [TV] vs. benazepril/amlodipine [BA]). One-year costs of care were examined following initiation of therapy; costs were categorized as cardiovascular-related and all other-related care. 95% confidence intervals for cost differences were calculated using nonparametric bootstrapping techniques. RESULTS: The mean age of the sample (n = 174) was 53 years; 47% were female. 22%, 6%, 6%, and 3% of patients had comorbid diagnoses of hyperlipidemia, cardiac arrhythmias, other ischemic heart disease, and myocardial infarction respectively during pre-treatment. Patients in the TV group had lower cardiovascular-related cost as compared to the BA group ($2311 vs. $2570, mean difference: -$259, 95% CI [-$2730, $1438]). Differences in cardiovascular-related cost were manifested primarily in the cardiovascular-related inpatient cost ($615 vs. $1,209 for TV and BA, respectively, mean difference:-$594, 95% CI [-$2900, $823]). All other-related costs were considerably lower in the TV group ($5006 vs. $6404, mean difference: -$1397, 95% CI [-$9491, $3669]). Patients in the TV group ($7,317) had lower overall costs as compared to the BA group ($8,974; mean difference -$1,656, 95% CI [-$12,657,$4,735]). CONCLUSIONS: Use of fixed combination trandolapril/verapamil therapy among patients with diabetes is associated with reduced direct medical costs in comparison to combination benazepril/amlodipine.