OBJECTIVES Prior analyses showed that AF patients treated with initial IV amiodarone had significantly higher adjusted average resource use compared with other therapies. This analysis assesses whether these results can be explained by presence of classic AF covariate (AFC) diagnoses (heart failure; peripheral vascular disease). METHODS We used 2004-05 discharges from the Premier Perspective database, including patients with primary AF diagnoses and evidence of initial therapy with electric conversion (EC) or an IV antiarrhythmic [amiodarone (AM), ibutilide (IB) or procainamide (PR)]. Patients were classified by AFC status. Inpatient costs and LOS were adjusted for clinical, demographic and hospital factors. RESULTS We stratified 10048 discharges into groups with AFC (28%) and without. Results showed that patients with AFC had higher costs (p<.0001) and longer LOS overall (p<.0001) and were 34% more likely to be treated with initial AM (p<.0001) vs. patients without, thus explaining some of the differences observed earlier. However, resource use with initial AM remained significantly higher than with any other initial therapy regardless of patients' AFC status. Among those without AFC, adjusted average costs were $440-$1988 higher (p<.0001) and LOS was 0.4-1 day longer (p<.0001) with initial AM than with any other initial therapy. Among patients with AFC, initial AM was $1602- $3266 higher (p<.0001) and LOS was 1.1-1.5 days longer (p<.005) than with any other initial therapy. CONCLUSIONS There are significant inpatient cost and LOS differences among AF patients depending on initial therapy and presence of AFC. Patients with AFC had higher adjusted average costs and LOS than those without and were more likely to be treated with initial AM. However, adjusted average costs and LOS were highest among patients treated with initial AM regardless of AFC status. Further research should explore whether factors such as time to conversion affect resource use among these patients.