Abstract

Purpose
Traumatic brain injury (TBI) presents competing risks of venous thromboembolism (VTE) and intracranial haemorrhage progression, rendering pharmacological thromboprophylaxis (PTP) timing uniquely contentious. The dilemma is most acute in severe TBI, where both risks are greatest. This systematic review and meta-analysis (PROSPERO: CRD420261388300) evaluates early versus late PTP on VTE, intracranial haemorrhage progression, and mortality in adults with severe TBI, applying a strict severity definition (Glasgow Coma Scale (GCS) 8 or less in 100% of the analysed cohort) to confirm and refine findings from prior mixed-severity syntheses.

Methods
MEDLINE, EMBASE, and CENTRAL were searched from inception to 1 May 2026. Studies enrolling adults with severe TBI (GCS 8 or less throughout analysed cohort) were eligible. Primary outcomes were VTE, intracranial haemorrhage progression, and in-hospital mortality. Definitions of VTE, and timing thresholds for early and late PTP were those used by the investigators in the individual studies. Risk of bias and certainty of evidence were assessed using ROBINS-I and GRADE. Random-effects meta-analysis generated pooled odds ratios (ORs).

Results
Seven reports representing six unique cohorts (6,825 patients) were included. Definitions used for early and late PTP timing varied across studies. Earlier PTP was associated with lower odds of VTE (pooled OR 0.47, 95% CI 0.23–0.99; I 2 = 30.6%; p = 0.046). The association was of borderline statistical significance: it did not persist on leave-one-out removal of either large contributing cohort, and was non-significant when re-estimated as a risk ratio (RR 0.50, 95% CI 0.25–1.02; p = 0.057). No difference in in-hospital mortality was observed (pooled OR 1.15, 95% CI 0.90–1.46; p = 0.253). Evidence regarding haemorrhage progression was derived from a single cohort with three total events (OR 0.55, 95% CI 0.03–11.52; p = 0.701); assessment relied on qualitative binary radiological review, which is insensitive to small-to-moderate progression. Among patients receiving PTP, pooled VTE prevalence was 10.1% (95% CI 5.9–15.3%), varying nearly fourfold (5.5%-20.7%) according to surveillance intensity. Three cohorts were judged at moderate and three at serious risk of bias; certainty of evidence was very low for all primary outcomes.

Conclusions
In adults with severe TBI (defined by GCS 8 or less in 100% of the analysed cohort), earlier PTP was associated with lower VTE risk of borderline statistical significance, and no difference in in-hospital mortality. The direction was consistent with prior mixed-severity syntheses, but the estimate was fragile under sensitivity analyses. The safety of earlier PTP with respect to intracranial haemorrhage progression remains unresolved, as only one cohort contributed this outcome using qualitative assessment. Adequately powered randomised evidence with quantitative haemorrhage assessment is warranted.