Early red blood cell transfusion burden, hemoglobin at first transfusion, and in-hospital mortality in patients supported with veno-arterial ECMO: a multicenter cohort study

Abstract

Graphical abstract http://www.w3.org/1999/xlink float portrait fx1.jpg fig0015 anchor portrait graphical abs0005 Highlights • The 24-h landmark cohort included 1,708 adults from six VA-ECMO centers. lsti0005 • Day-1 RBC dose showed a graded mortality association (RR 1.08 per 2 units). lsti0010 • Exploratory analyses linked lower first-transfusion Hb to higher mortality. lsti0015 • Most sensitivity analyses yielded directionally similar estimates. lsti0020 • Findings are prognostic and do not establish an optimal transfusion threshold. lsti0025 simple lis0005 author-highlights abs0010 Background Red blood cell (RBC) transfusion is common during venoarterial extracorporeal membrane oxygenation (VA-ECMO), but the prognostic relevance of early cumulative transfusion burden and hemoglobin (Hb) at transfusion initiation remains uncertain. We examined associations of ECMO day-1 RBC dose and, among transfused patients, first-transfusion Hb and transfusion-pattern phenotype with in-hospital mortality. Methods Adults aged ≥18 years initiating peripheral VA-ECMO at six Korean centers between January 2018 and March 2025 were analyzed. The primary analysis used a 24 -h landmark cohort of patients alive and still receiving ECMO beyond 24 h. Day-1 RBC dose was categorized as 0, 1–2, 3–4, or ≥5 units and modeled per 2-unit increase. Adjusted risk ratios (RRs) were estimated using modified Poisson regression. Secondary analyses assessed first-transfusion Hb and consistent- versus variable threshold phenotypes among transfused complete cases. Results Among 2,020 included patients, 1,708 entered the 24-h landmark cohort, and 1,140 complete cases comprised the adjusted primary analysis. Compared with no day-1 transfusion, adjusted RRs for in-hospital mortality were 1.20 (95% confidence interval [CI], 1.02–1.40) for 1–2 units, 1.37 (95% CI, 1.17–1.59) for 3–4 units, and 1.48 (95% CI, 1.28–1.71) for ≥5 units. Each 2-unit increase was associated with higher mortality (adjusted RR, 1.08; 95% CI, 1.05–1.11), with concordant findings over ECMO days 1–2 (adjusted RR, 1.06; 95% CI, 1.04–1.08). Among 1,043 transfused complete cases with qualifying pretransfusion Hb, each 1 g/dL decrease in first-transfusion Hb was associated with higher mortality (adjusted RR, 1.15; 95% CI, 1.10–1.20). The association was evident in the consistent-threshold phenotype (adjusted RR, 1.16; 95% CI, 1.10–1.22) but not in the variable-threshold phenotype (adjusted RR, 1.02; 95% CI, 0.92–1.12). Conclusions In adults receiving peripheral VA-ECMO, early RBC transfusion burden showed a graded association with in-hospital mortality, whereas lower first-transfusion Hb was a prognostic marker chiefly within the consistent-threshold phenotype. These findings support dose- and threshold-aware prognostic assessment but do not establish causality or define an optimal transfusion threshold abs0015