Economic assessment of troglitazone as an adjunct to sulfonylurea therapy in the treatment of type 2 diabetes

Objective: To assess the economic efficiency of adding troglitazone to sulfonylurea therapy to improve glycemic control. Background: Despite the high prevalence of type 2 diabetes, existing treatment strategies often fail. New oral agents give a wider segment of the population with type 2 diabetes hope of achieving near-normal blood-glucose levels. Troglitazone, a novel chemical entity, is one promising new agent. Methods: We conducted an economic analysis based on glycemic-control data from a randomized clinical trial comparing troglitazone with placebo, each added to glyburide. A patient simulation model was used to translate these data to long-term outcomes associated with diabetes. Patients had poorly controlled type 2 diabetes mellitus despite glyburide therapy. Risk functions of developing and progressing through nephropathy, retinopathy, neuropathy, hypoglycemia, and macrovascular disease were developed from the Diabetes Control and Complications Trial and large epidemiologic studies. Cost estimates were based on data from 5 states, all payor databases, surveys, and literature. The main outcomes of the model were cost-consequences, number of patients developing each type of complication, mean time to development of the complication, cost per life-year gained (LYG), and cost per quality-adjusted life-year. Results: The model predicts that for every 1000 patients treated with troglitazone, the improved glycemic control could mean that 95 to 140 fewer patients would experience one of the most severe diabetic complications leg, blindness, end-stage renal disease, amputation), which may increase life expectancy by 2.0 years. These benefits are obtained at an additional $2100 per LYG (undiscounted). The ratio remains <$50,000 per LYG for most variations in input. Conclusions: The clinical trial demonstrated that troglitazone + glyburide improves glycemic control compared with glyburide alone. Based on these results, the model estimates fewer diabetic complications at a cost well below accepted cost-effective thresholds.