Optimal hemoglobin threshold for blood transfusions in sepsis and septic shock: a retrospective analysis

Abstract

Transfusions of red blood cells (RBCs) are crucial for improving tissue oxygenation in anemic patients with sepsis. Nevertheless, the debate continues over the ideal hemoglobin level for transfusions. This study aimed to assess the impact of different hemoglobin levels on the outcomes of patients with sepsis who received transfusions. This retrospective analysis included adult patients with sepsis treated in the general medical ward and intensive care unit at a University affiliate hospital. Patients needing RBC transfusions were included. The primary outcome was the 28-day mortality rate. From March 2018 to January 2022, 806 patients were studied. Of these, 480 (59.6%) were transfused at hemoglobin levels of 7–9 g/dL (“liberal group”), while 326 (40.4%) received RBC transfusions when their hemoglobin was < 7 g/dL (“restrictive group”). Mean hemoglobin levels at transfusion were 8.1 ± 0.8 g/dL and 6.3 ± 0.8 g/dL for each group, respectively(P < 0.001). On day 28, the liberal group had a mortality rate of 51.2% (246 patients), compared to 59.2% (193 patients) in the restrictive group (Odds ratio [OR] 0.88, 95% confidence interval [CI] 0.79–0.98, P = 0.031). Adjusted comparisons showed 46.8% mortality in the liberal group (141/301patients) versus 59.3% in the restrictive group (178/300patients) at 28 days (OR 0.78, 95% CI 0.66–0.92, P = 0.002). Multivariate analysis revealed transfusion at hemoglobin 7–9 g/dL as an independent variable linked to lower 28-day mortality (OR 0.70, 95% CI 0.49–0.99, P = 0.042). Other factors correlated with 28-day mortality were platelet counts ≤ 150 × 103/µL, albumin ≤ 2.5 g/dL, shock, mechanical ventilation, and renal replacement therapy. This retrospective study suggests that RBC transfusion at hemoglobin levels of 7–9 g/dL associates with lower 28-day mortality in sepsis patients compared to transfusion at hemoglobin levels below 7 g/dL. Clinical trial registration The study was registered with the Thai Clinical Trials Registry (identification number TCTR20231003003). (https://www.thaiclinicaltrials.org/show/TCTR20231003003).