Development of Heart Failure in Medicaid Patients with Type 2 Diabetes Treated with Pioglitazone, Rosiglitazone, or Metformin

BACKGROUND: Medicaid covers a high-risk population typically underrepresented in clinical trial data and largely absent in observational studies of real-world cardiovascular risks associated with thiazolidinediones (TZDs), such as pioglitazone and rosiglitazone, which are used to manage type 2 diabetes. In November 2013, the FDA removed prescribing restrictions for rosiglitazone in light of new evidence that rosiglitazone did not increase the risk of heart attack compared with standard type 2 diabetes medications. Further investigation is needed to elucidate whether the risk of heart failure (HF) associated with TZDs may be exacerbated in the Medicaid population. OBJECTIVE: To determine the relative risk of incident HF in patients initiating rosiglitazone, pioglitazone, and metformin therapy in a Medicaid population. METHODS: We retrospectively examined claims data for patients with type 2 diabetes enrolled in Maryland State Medicaid and managed care or fee-for-service programs between July 2005 and June 2010. Patients initiated on metformin, pioglitazone, or rosiglitazone treatments were extracted for analysis. Relative risks of incident HF after initiating treatment were compared using survival analysis, adjusting for switching or adding antidiabetic therapies during follow-up and other baseline risk factors for HF. RESULTS: Of 6,271 patients meeting inclusion criteria, 88% were started on metformin; 7% were started on pioglitazone; and 5% were started on rosiglitazone. Patients who initiated rosiglitazone had higher risk of HF than patients who initiated metformin using either univariate (HR=1.81, 95% CI=1.37-2.39), multivariate (HR=1.57, 95% CI=1.15-2.15), or propensity score-matched (HR= 1.79, 95% CI =1.16-2.76) analysis. There was no significant difference in risk between patients who initiated pioglitazone and metformin therapy. CONCLUSIONS: Compared with metformin, there may be higher risk of developing HF in Medicaid patients started on rosiglitazone but not pioglitazone. While pioglitazone was associated with a lower risk of developing HF compared with rosiglitazone, health care professionals should continue to work closely with their patients to determine the treatment options most appropriate.