How well do the 56,183 atrial fibrillation patients enrolled in the Contemporary Novel Oral Anticoagulant (NOAC) trials reflect the real world?

Purpose: Compare baseline data of atrial fibrillation (AF) patients in Novel oral anti-coagulant (NOAC) trials compared to Real World (RW) AF. Methods: AF patients in 4 randomized trials (RCTs) (AVERROES ARISTOTLE, RE-LY and ROCKET AF) were compared to 5 RW sources by baseline characteristics. RW Data sources: UK General Practice Research Database, the Swedish Hospital Discharge Register, Christiana Care Health System and claims data from I3 InVision Data Mart and PharMetrics in the US. Results: Descriptive results representing over 56,000 AF patients in RCTs and 550,000 AF patients in the RW show that the RW populations have a higher proportion of > 75 (range 70 to 77 vs 69 to 73 RCTs) and slightly more females in RW ( 41 to 48% RW vs. 35 to 42% RCT). More RCT AF patients had a higher use of beta blockers ~ 60% in RCTs to 17 to 30% in RW. RCT patients had more prior strokes, hypertension and statin use. Conclusions: Females and > 75 years are under-represented in the RCTs. I3 and PharMetrics under-represent >65 so differences in the RW may be greater than represented here. Substantially more beta blockers in the RCTs. Drug use rates are lower in the RW cohorts despite fairly similar conditions, which may be an indicator of less adequate treatment in the RW. In the absence of evidence that baseline differences do not affect the effectiveness of NOACs, adjustments need to be made to assess the benefits and potential risks of NOACs in the RW.