Association between peri-procedural red blood cell transfusion and clinical outcomes in transcatheter aortic valve replacement: a systematic review and meta-analysis

Résumé

Background Peri-procedural red blood cell (RBC) transfusion is common after transcatheter aortic valve replacement (TAVR), but its relation to clinical outcomes is uncertain. This systematic review and meta-analysis assessed that association. Methods PubMed, Embase, ScienceDirect, and ClinicalTrials.gov were searched to April 2026 for observational comparisons of transfused and non-transfused TAVR patients. After removal of overlapping TRITAVI populations and one study at critical risk of bias, 11 studies entered the primary analysis. Binary outcomes were pooled as risk ratios (RR) with random-effects models (restricted maximum likelihood, Hartung–Knapp; 0.5 continuity correction for zero-event arms). Continuous outcomes were pooled as mean differences. Adjusted odds ratios were pooled separately. Prediction intervals were reported when at least three studies informed an estimate. Risk of bias (ROBINS-I), leave-one-out analyses, Egger's test, and GRADE were applied. Results Transfusion was associated with higher 30-day mortality (RR 3.42, 95% CI 1.27–9.18; I 2 = 86.5%; prediction interval 0.28–41.03), 1-year mortality (RR 2.33, 95% CI 1.58–3.45; I 2 = 79.9%; prediction interval 0.81–6.73), stroke (RR 1.59, 95% CI 1.25–2.03; I 2 = 19.4%; prediction interval 1.00–2.53), life-threatening bleeding (RR 9.28, 95% CI 2.56–33.68; I 2 = 66.2%; prediction interval 0.98–87.45), major vascular complications (RR 4.07, 95% CI 1.85–8.96; I 2 = 30.8%; prediction interval 2.11–7.85), and stage 2–3 acute kidney injury (RR 2.19, 95% CI 1.39–3.44; I 2 = 86.3%; prediction interval 0.69–6.91). In-hospital mortality and non-fatal myocardial infarction were not significant. Adjusted estimates pointed toward harm but were underpowered. Egger's test suggested asymmetry for stage 2–3 acute kidney injury, length of stay, and life-threatening bleeding. Leave-one-out analyses were stable for 1-year mortality, stroke, stage 2–3 acute kidney injury, and length of stay, but 30-day mortality and life-threatening bleeding shifted after omission of a single study. Conclusion Peri-procedural RBC transfusion in TAVR is associated with worse outcomes, most robustly for stroke, major vascular complications, and 1-year mortality. The life-threatening bleeding signal is partly definitional, because VARC criteria include transfusion volume. Findings support bleeding avoidance and restrictive thresholds; trials are needed to establish causality. Systematic Review Registration identifier CRD420261405548.