We assessed the presence of immortal time bias (ITB) in observational studies evaluating the effectiveness of radiofrequency catheter ablation for atrial fibrillation (AF) on reduction of stroke. Eligible studies were classified based on presence or absence of ITB. Hazard ratios (HRs) were pooled using the random-effects model. Eight of 10 (80%) studies were subject to ITB. Pooling studies without ITB indicated no statistically significant reduction in incident strokes (HR 0.75; 95% confidence interval [CI], 0.49-1.02]. In conclusion, the pervasiveness of ITB in observational studies precludes definitive conclusions regarding an effect of AF ablation on strokes. Further studies designed to avoid ITB are warranted.
Copyright © 2020 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved.