OBJECTIVES:  Among NVAF patients, compare all-cause cost and healthcare resource utilization (HCRU) between dabigatran and apixaban, and dabigatran and rivaroxaban. METHODS: Adult NVAF patients newly diagnosed and treated with dabigatran, apixaban or rivaroxaban during 10/01/2010-12/31/2014 were identified using MarketScan claims databases. Patients were continuously enrolled for 12-months prior to index date (first NOAC claim) and were followed from index until medication switch, discontinuation, inpatient death or end of continuous enrollment or study period. Dabigatran patients were matched to apixaban and rivaroxaban patients and two 1:1 propensity score matched groups were created. Per-patient-per-month (pppm) HCRU and costs were reported. Generalized linear models compared total, inpatient, outpatient, and pharmacy costs within each group. RESULTS:  A total of 26,646 and 8,857 dabigatran patients were matched to rivaroxaban and apixaban patients, respectively (mean age 68 years; 37% female in each matched group). Compared with rivaroxaban, dabigatran patients had significantly lower mean all-cause, pppm total healthcare, inpatient, and pharmacy costs ($4,147 vs. $4,559; $1,484 vs. $1,812; and $630 vs. $644, respectively, all P<0.05) and significantly fewer hospitalizations (0.06 vs. 0.07), outpatient visits (4.81 vs. 4.95), and pharmacy claims (4.77 vs. 4.91) (all P<0.01). Compared with apixaban, dabigatran patients had similar mean, all-cause, pppm total healthcare, inpatient, outpatient, and pharmacy costs ($3,849 vs. $3,807; $1,225 vs. $1,198; $1,953 vs. $1,930 and $671 vs. $678, respectively, all P>0.05) and, similar hospitalizations (0.05 vs. 0.05, p=0.097) but significantly higher outpatient visits (4.68 vs. 4.24) and pharmacy claims (4.79 vs. 4.57), (both P<0.01). Multivariate analyses demonstrated dabigatran had significantly lower total, inpatient, and outpatient costs than rivaroxaban and similar costs (across all categories) as apixaban patients. CONCLUSIONS: Dabigatran had significantly lower total costs and HCRU than rivaroxaban patients. Total costs were similar between dabigatran and apixaban. Dabigatran had more outpatient and pharmacy HCRU than apixaban but similar hospitalizations, the main driver of costs.