Abstract 1448: Thiazolidinediones and the Risk of Incident Congestive Heart Failure Among Patients With Type 2 Diabetes Mellitus: A Population-Based Nested Case-Control Study
Background: Recent clinical trials suggest that thiazolidinediones (TZDs) may increase the risk of congestive heart failure (CHF). However, the effect of TZDs on the risk of incident CHF in unselected populations has not been thoroughly investigated.
Methods: Using data from the United Kingdom’s General Practice Research Database, we conducted a case-control study within a population-based cohort of patients with type 2 diabetes. Cases were identified by a clinical diagnosis of incident CHF between January 1st, 2000 and December 31st, 2006 and were then classified as possible or probable cases using prescription data. Up to 10 controls per case were matched on age, physician practice, and calendar date. A drug exposure window of 90 days was used in the primary analysis, which compared patients prescribed TZDs to those with no prescriptions for anti-diabetic medications. Data were analyzed by conditional logistic regression.
Results: We identified 3,405 incident cases (2,632 probable and 773 possible) of CHF and 32,042 corresponding controls. TZDs were prescribed in 6.4% of cases and 6.1% of controls. We found no definitive evidence that TZDs are associated with an increased incident rate of CHF but are unable to exclude a small but clinically meaningful increase in the rate of CHF (Table⇓). Restriction to probable CHF cases produced similar results (adjusted rate ratio (RR)=1.15, 95% CI=0.89, 1.47). In a secondary sensitivity analysis using a 30-day exposure-window, similar results were observed (adjusted RR=1.28, 95% CI=1.01, 1.65). Compared with metformin monotherapy, TZD therapy was associated with an increased rate of CHF (RR=1.38, 95% CI=1.12, 1.71).
Conclusions: Given the totality of the evidence from this and previous studies, the probability of an increased risk for CHF with these agents remains high. However, any increase in CHF risk associated with TZDs may be lower than previously reported.