Impact of rosiglitazone therapy on the lipid profile, glycemic control, and medication costs among type 2 diabetes patients.

OBJECTIVE: To investigate the impact of rosiglitazone therapy on lipid profiles, glycemic control, and costs associated with cholesterol-lowering and diabetic medications among Type 2 diabetes mellitus (T2DM) patients in a standard practice setting.

METHOD: This retrospective cohort study was conducted using data from the General Practice Research Database during 1999-2006. T2DM patients were classified based on the addition of rosiglitazone versus either metformin or a sulfonylurea ('comparison group') to pre-existing glucose lowering agents. After propensity score matching to control for differences in baseline patient characteristics, 1450 matched pairs were identified. The mean changes in total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), glycosylated hemoglobin A1C (A1C), and daily medication costs were calculated. To investigate the incremental costs for lipid-lowering medications, a two-part model was utilized.

RESULTS: The mean changes in TC and A1C for the rosiglitazone and metformin/sulfonylurea groups were 9 vs -10 mg/dL for TC, -2 vs -9 mg/dL for LDL-C, and -0.8% vs. -1.2% for A1C, respectively. The mean changes in daily medication costs of glucose- and lipid-lowering drugs were $3.95 for rosiglitazone patients and $0.27 for metformin/sulfonylurea patients. For patients with positive incremental lipid-lowering costs, rosiglitazone use was significantly associated with costs eight times greater than metformin/sulfonylureas. Generalizability of the study is limited due to cost estimates using the national formulary and potential selection bias.

CONCLUSIONS: Addition of rosiglitazone to an existing antidiabetic medication regimen improved glycemic control to a lesser extent than metformin/sulfonylurea, and also deteriorated patients' lipid profiles, leading to significantly greater daily costs. Economic evaluations of alternative therapies should consider such costs to estimate the full impact of different therapeutic approaches in diabetes.