The effect of liberal and restrictive transfusion practices on survival in oncology patients with sepsis in the intensive care unit

Red blood cell (RBC) transfusion is a commonly utilized supportive therapy in intensive care units (ICUs). However, the optimal transfusion practice in oncologic patients with sepsis remains controversial. This study aimed to compare the clinical outcomes of liberal versus restrictive transfusion practices in oncologic patients with sepsis admitted to the ICU. This retrospective, non-randomized study included oncologic patients with sepsis admitted to the ICU between 2014 and 2023. Patients were classified as having restrictive or liberal transfusion practice based on serial hemoglobin (Hb) measurements and corresponding decisions to administer or withhold RBC transfusion, reflecting lower (< 7–8 g/dL) or higher (≤ 9–10 g/dL) Hb thresholds, respectively. Among patients who did not receive RBC transfusion, withholding transfusion at Hb levels of 7.1–8.9 g/dL and 9.1–10.0 g/dL was classified as restrictive and liberal transfusion practice, respectively; thus, actual receipt of RBC transfusion was not required for group assignment. Patients with mixed or otherwise unclassifiable transfusion practices were excluded. As treatment decisions were determined by treating physicians, the approach may introduce selection bias. Demographic characteristics, clinical scores, primary tumor type, laboratory parameters, organ dysfunction, infection characteristics, and mortality outcomes were compared. Descriptive statistics were calculated. Propensity score analysis with stabilized inverse probability of treatment weighting (IPTW) was used to minimize confounding and selection bias. Logistic regression models were used to evaluate associations between transfusion practice and binary outcomes, and Cox proportional hazards models were used to assess survival. A total of 658 patients were included. Baseline severity scores were higher in the restrictive transfusion group, as reflected by significantly greater Acute Physiology and Chronic Health Evaluation (APACHE) II and Sequential Organ Failure Assessment (SOFA) scores. There was a marked imbalance in primary tumor type’s distribution between groups, with lung cancer more frequent in the liberal group and gastrointestinal malignancies predominating in the restrictive group. Overall organ dysfunction rates were similar between groups; however, acute kidney injury occurred more frequently in the restrictive group. No differences were observed in duration of mechanical ventilation, use of vasopressors, or requirement for renal replacement therapy. Unadjusted analysis suggested higher 90-day mortality in the restrictive group, but this association was not significant after IPTW adjustment. No difference in 28-day mortality was observed. The restrictive transfusion group was associated with an increased risk of renal injury compared to the liberal transfusion group, and this association remained significant after clinical adjustment. In critically ill patients with sepsis and malignancy, restrictive and liberal transfusion practices were not associated with significant differences in mortality after IPTW adjustment. The observed association between restrictive transfusion practice and renal injury should be interpreted cautiously, as it does not establish a causal relationship. Prospective studies evaluating individualized transfusion practices are warranted.