Restrictive and liberal transfusion strategies in extracorporeal membrane oxygenation: A retrospective observational study

• Published in: Transfusion (Philadelphia, Pa.) Pa.), 2023-02, Vol.63 (2), p.294-304
• Main Authors: Ng, Pauline Yeung, Chan, Ho Ching Victor, Ip, April, Ling, Lowell, Chan, Kai Man, Leung, Kit Hung Anne, Chan, King Chung Kenny, So, Dominic, Shum, Hoi Ping, Ngai, Chun Wai, Chan, Wai Ming, Sin, Wai Ching
• Format: Article
• Language: English
• Subjects: Adult; Blood Transfusion; Clinical trials; ECMO; Erythrocytes; Extracorporeal Membrane Oxygenation; Hemoglobin; Hemoglobins - analysis; Heterogeneity; Humans; intensive care unit; Membranes; Mortality; Observational studies; Oxygenation; Prospective Studies; red blood cell transfusion; Retrospective Studies; Risk reduction; Strategy; Subgroups; Transfusion
• ISSN: 0041-1132; 1537-2995
• DOI: 10.1111/trf.17221

Background To compare the outcomes of patients requiring extracorporeal membrane oxygenation (ECMO) support who had a restrictive transfusion strategy with those who had a liberal strategy. Study Design and Methods We retrospectively reviewed all adult patients from 2010 to 2019 who received a minimum of one packed red blood cell (pRBC) during ECMO. Hemoglobin values before each transfusion were retrieved. Restrictive transfusion strategy was defined as a transfusion threshold ≤8.5 g/dl in all transfusion episodes for a single patient, while liberal transfusion strategy was defined as a transfusion threshold >8.5 g/dl in any transfusion episode. Results The analysis included 763 patients, with 138 (18.1%) patients in the restrictive and 625 (81.9%) in the liberal transfusion strategy group. The median hemoglobin level, taking into account all measured hemoglobin values, during ECMO support was 8.3 and 9.9 g/dl, and the average units of pRBC received per day were 0.7 (0.3–1.8) and 1.2 (0.6–2.3), respectively. There were no significant differences in intensive care unit (ICU) mortality (adjusted odds ratio (OR), 0.86; 95% CI 0.56–1.30; p = .47), hospital mortality (adjusted OR, 0.79; 95% CI 0.52–1.21; p = .28), and 90‐day mortality (adjusted OR, 0.84; 95% CI 0.55–1.28; p = .42) between the two groups. Among subgroup analyses, a restrictive transfusion strategy was associated with decreased risk of ICU mortality in patients on veno‐venous ECMO (adjusted OR, 0.36; 95% CI 0.17–0.73; p = .005). There was no heterogeneity on outcomes across patients stratified by age, APACHE IV score, or need for large volume transfusion. Discussion Our data suggested it may be safe to adopt a restrictive red cell transfusion threshold of 8.5 g/dl in patients on ECMO, and highlighted the need for prospective trials in this heavily‐transfused population.