Association of lifesaving versus non-lifesaving extracranial surgery with long-term outcome after severe traumatic brain injury: a prospective CENTER-TBI cohort analysis

Background: Patients with traumatic brain injury (TBI) frequently undergo extracranial surgery (ES). These procedures may be lifesaving (LS), such as haemorrhage or damage-control interventions, or non-lifesaving (NLS), including definitive fracture fixation or wound closure. Whether ES independently affects neurological recovery remains unclear. Methods: We analysed adult ICU patients with polytrauma and structural TBI (ISS ≥ 16, AIS Brain ≥ 3, and at least one extracranial injury as AIS ≥ 3) enrolled in the prospective multicentre CENTER-TBI cohort across 65 European sites. Predictors of extracranial surgery (ES) were evaluated using a cause-specific Cox proportional hazards model with competing risks, considering death before surgery as a competing event. LS and NLS categories were assigned by local investigators based on urgency. Associations between ES category (LS, NLS, or no-ES) and six-month unfavourable neurological outcome (GOSE ≤ 4) were assessed using multivariable logistic regression adjusted for IMPACT extended prognostic variables. Sensitivity analyses excluding patients who died within 48 h and secondary analysis in ICU survivors evaluated six-month outcome while minimizing immortal time bias from later NLS procedures. Results: Of 870 patients, 369 (42.4%) underwent ES: 168 (19.3%) LS and 201 (23.1%) NLS. LS patients were younger, had higher ISS, greater haemodynamic instability, more severe shock, and higher transfusion requirements compared with NLS and no-ES groups. Unadjusted six-month unfavourable outcome rates did not differ significantly (48.9% no-ES, 51.8% LS, 42.3% NLS; p = 0.2). LS remained associated with worse neurological outcome after adjustment (OR 1.86, 95% CI 1.15-3.01; p = 0.011), whereas NLS was not. Results remained consistent across sensitivity analyses, the multiple imputation model, and the secondary analysis. Conclusions: Emergency/damage-control extracranial surgery operationally classified as LS identified a phenotype of patients with greater extracranial injury burden and physiological instability, associated with worse long-term neurological outcome. This association should be interpreted as reflecting injury severity and treatment indication rather than a direct detrimental effect of surgery itself. NLS procedures were not independently associated with unfavourable recovery. Trial registration: ClinicalTrials.gov Identifiers: NCT02210221.