Liberal vs. restrictive transfusion strategies for acute brain injury: a systematic review and frequentist-Bayesian meta-analysis

Abstract

Purpose

To determine whether a liberal transfusion strategy (≥ 9 g/dL) improves neurological outcomes in adults with acute brain injury (ABI).

Method

We systematically searched MEDLINE, EMBASE, the Cochrane Library, and trial registries for randomized controlled trials comparing liberal (≥ 9 g/dL) vs. restrictive (≥ 7 g/dL) transfusion in adults with ABI (traumatic brain injury, subarachnoid hemorrhage, intracranial hemorrhage) and Glasgow Coma Scale ≤ 13. Frequentist, Bayesian, and trial sequential analyses were used. The primary outcome was favorable neurological status at 180 days.

Results

Four randomized controlled trials (N = 1853; 922 liberal, 931 restrictive) were included. The pooled frequentist risk ratio (RR) for favorable neurological outcome was 0.84 (95% CI 0.65–1.09; I² = 58%). In a pre-specified sensitivity analysis including only low-risk-of-bias trials, the results suggested a potential benefit in favor of the liberal strategy (RR 0.74 [95% CI 0.63–0.87]) with no heterogeneity (I² = 0%). Subgroup analyses for patients with traumatic brain injury or stratified by initial Glasgow coma scale were consistent with the main findings. Bayesian analyses showed that the estimated treatment effect depended on the assumptions and priors used, with an unfavorable prior derived from one trial with distinct protocol appearing less likely than neutral or favorable priors. Trial sequential analysis indicated that current evidence is insufficient to confirm a definitive effect. Secondary outcomes did not differ significantly between groups.

Conclusions

This review did not provide definitive evidence of a neurological benefit from liberal transfusion strategies in acute brain injury. Both frequentist and Bayesian analyses highlight the influence of a single trial on the overall effect estimate and heterogeneity. However, sensitivity analyses excluding this trial and focusing on studies with low risk of bias suggested that liberal transfusion strategies could improve neurological outcomes. Future research should focus on identifying patient subgroups most likely to benefit, guiding a more individualized approach.