Table 1.
Summary of findings
| Outcome | No. of Studies | Design | Events/Patient‐Years | Relative Effect (95% CI) | No. of Patients | Quality of Evidence (GRADE) | Comments | |
|---|---|---|---|---|---|---|---|---|
| PxAF | Non‐pxAF | |||||||
| Stroke and/or SE | 18 | 11 RCTs, 7 registries | 941/62 553 | 3611/176 975 | 0.72 (0.65‐0.80) | 239 528 | + + + + | |
| Guideline definition | 14 | 7 RCTs, 7 registries | 883/60 108 | 3349/166 278 | 0.70 (0.63‐0.78) | 226 386 | + + + + | |
| Stroke only | 10 | 5 RCTs, 6 registries | 298/21 059 | 955/48 314 | 0.75 (0.62‐0.89 | 69 373 | + + + − | Risk of bias in individual studies |
| Guideline definition | 8 | 2 RCTs, 6 registries | 269/20 054 | 848/44 662 | 0.72 (0.59‐0.88 | 64 716 | + + + − | Risk of bias in individual studies |
| Major bleedings | 7 | RCTs | 1100/37 550 | 3420/124 434 | 1.04 (0.92‐1.18) | 161 984 | + + − − | Data only from RCTs, inconsistency |
Abbreviations: CI, confidence interval; GRADE, Grading of Recommendations Assessment, Development, and Evaluation; RCTs, randomized controlled trials; SE, systemic embolism.
Studies with assessed risk of bias showed very low weight in the main analysis, and downgrade was not justified.