Withdrawal of Life-Sustaining Treatment and the Severity-Matched Mortality Difference Between Abusive and Unintentional Pediatric Traumatic Brain Injury: A Propensity-Matched Cohort Study of the ACS TQIP.

PURPOSE: Abusive head trauma carries higher mortality than unintentional traumatic brain injury (TBI), but whether this persists at comparable presenting severity, and how much reflects withdrawal of life-sustaining treatment (WLST), is unclear. We compared severity-matched in-hospital mortality between abusive and unintentional severe pediatric TBI and quantified the contribution of WLST.

METHODS

Retrospective cohort study of the American College of Surgeons Trauma Quality Improvement Program registry (2019-2024) including children with severe TBI (Glasgow Coma Scale ≤8) from abusive or unintentional blunt mechanisms. Abusive cases were matched 1:1 to unintentional controls on 43 indicators of demographics, presenting severity, and hospital factors. Mortality was analyzed with conditional logistic regression; WLST's contribution was assessed by restriction to patients without WLST, counterfactual decomposition, and multiple imputation retaining all abusive cases.

RESULTS

Crude mortality was 40.5% (abusive) vs 17.2% (unintentional). In 395 matched pairs, abusive injury was associated with higher mortality (odds ratio [OR] 1.39, 95% CI 1.01-1.90) and more frequent WLST (19.0% vs 14.2%). The association was attenuated among patients without WLST (OR 1.11, 0.71-1.71), but this underpowered finding was not reproduced after multiple imputation (1,009 pairs: mortality OR 1.58, 1.21-2.07; without WLST 1.55, 1.03-2.34). WLST mediated one-third of the excess mortality (33%, 7-60%) and preceded half of the excess deaths; the excess was confined to deaths after hospital day 1 (OR 1.65, 1.22-2.23).

CONCLUSION

At comparable presenting severity, abusive TBI carries a modest excess mortality emerging after the first hospital day; roughly one-third runs through more frequent WLST and the remainder persists among children who continue treatment.

LEVEL OF EVIDENCE: III - retrospective comparative study (propensity-matched cohort).