OBJECTIVES: Thiazolidinedione (TZD) was recommended not for patients with moderate and serious heart failures due to the adverse reaction of body fluid retention. Our study investigated the use of TZD by type 2 diabetic patients after they were discharged from hospitalizations for heart failure. METHODS: The cohort observation was based on claims database of Taiwan’s National Health Insurance (NHI) for 2002∼2004. There were4774 diabetic patients who were hospitalized due to heart failure during 2003: 379 TZD users and 4395 non-TZD users. A total of 2692 non-TZD users who used sulfonylurea after discharge were selected as the control group. Cox proportional hazard models were estimated to compared the outcomes of death and readmission to hospitals of TZD users and sulfonylurea users,one year after the time patients were discharged. RESULTS:There were 7.9% of diabetic patients who were hospitalized for heart failure and received TZD treatments. There was no significant difference on readmission over one year after discharge between the two groups. However, among those who used TZDor sulfonylurea less than 60 days, those who used TZD had lower HR of readmission than those who used sulfonylurea (HR0.728, 95% CI 0.551∼0.962). Overall, the HR for death among patients receiving TZD was significant lower than the control group of sulfonylurea users (HR 0.092, 95% CI = 0.022–0.380).CONCLUSIONS: Based on the precautions stated in TZD package inserts, 7.9% of type 2 diabetic patients with heart failure received potentially inappropriate prescriptions. Short term use of TZD significantly reduced hazards ratio in readmission, though long term use of both TZD and sulfonylurea showed no difference on readmission, and the use of TZD significantly reduced the hazard of death.
