Patterns and Predictors of Use of Warfarin and Other Common Long-Term Medications in Patients with Atrial Fibrillation

Abstract

No study has been published that compared persistence and discontinuation of warfarin with other long-term medications in patients with atrial fibrillation (AF). The objective of this study was to evaluate persistence and discontinuation patterns in AF patients taking warfarin and other common long-term medications and identify predictors of persistence and discontinuation in this population. Patients with warfarin prescription within 3 months after AF hospitalization discharge and 12-month data before and after the first prescription were evaluated using administrative claims data (1 January 2005 - 31 December 2007). For comparison, persistence patterns for other long-term medications for treatment of hypertension, hyperglycemia, heart disease, and dyslipidemia, including once- (od) and twice- (bid) daily medications, were evaluated. Non-persistence was defined as the presence of a ≥60-day gap in medication use. Permanent discontinuation was defined as no use of the medication for ≥90 days until the end of the follow-up period. Multivariate analysis was conducted to identify predictors of warfarin non-persistence and discontinuation. 28 384 patients with AF were identified; 16 036 (56.5%) had a warfarin prescription following AF hospitalization. 53.5% of warfarin users were persistent for at least 1 year, similar to other long-term medications commonly prescribed to the AF population (ranging between 45.2% and 61.3%). 42.6% of warfarin users permanently discontinued warfarin within 1 year, also consistent with the discontinuation rate of 32.9-52.0% of other long-term medications. Residence in the South and West regions of the US, history of cardiac dysrhythmias, and warfarin cost-sharing significantly decreased the likelihood of warfarin persistence and increased the likelihood of discontinuation, while older age, history of ischemic stroke, and warfarin use before hospitalization significantly increased warfarin persistence and decreased the likelihood of discontinuation. Adherence of od and bid medications was similar. Persistence and discontinuation with warfarin in patients with AF is consistent with other long-term medications. Identifying factors associated with non-persistence and discontinuation with long-term medications can help in developing targeted adherence programs.