Contemporary management pathways in liver trauma: predominant and successful use of non-operative strategies at a high-volume tertiary center.

BACKGROUND

Non-operative management (NOM) has become the default for haemodynamically stable liver trauma, yet pathway-level performance and bedside predictors of mortality warrant contemporary appraisal in University trauma centres.

METHODS

We conducted a single-centre retrospective cohort study of consecutive adults (≥ 18 years) with blunt or penetrating hepatic injury managed within a protocolised, physiology- and CT-guided pathway (1 Jan 2018–30 Apr 2025). The primary endpoint was in-hospital mortality. Secondary endpoints included NOM failure, liver-specific complications, and era-level outcomes (2018–2021 vs. 2022–2025). Independent predictors of mortality were assessed with multivariable logistic regression; discrimination was evaluated by ROC analysis.

RESULTS

Among 350 patients, NOM was initiated in 304 (86.9%); delayed conversion to laparotomy occurred in 15/304 (4.9%), including 5 after prior angioembolisation (IR in 32/350; 9.1%). Overall in-hospital mortality was 14.9% (52/350) and clustered in patients with physiological compromise and high global injury burden. Definitive management stratified patients into NOM and operative management cohorts, with lower mortality after definitive NOM than after operative management. In multivariable analysis, Shock Index (OR 2.71 per unit; p = 0.022), ISS (OR 1.05 per point; p < 0.001) and lactate (OR 1.17 per mmol/L; p = 0.031) independently increased the odds of death, whereas GCS was protective (OR 0.89 per point; p = 0.002); AAST high grade and era were not independently associated. Model discrimination was excellent (AUC 0.87; Youden-optimal sensitivity 0.92, specificity 0.74); Shock Index alone achieved AUC 0.75, while AAST grade provided modest separation (AUC 0.608). Secondary transfers (40.3%) incurred no survival penalty. Era comparison showed reduced operative exposure (21.5%→12.3%, OR 0.51, p = 0.024) and mortality (18.5%→10.3%, OR 0.51, p = 0.035) after formalising the pathway.

CONCLUSIONS

In a university trauma centre, a protocolised physiology- and CT-guided pathway delivered high NOM utilisation with low failure and era-level gains. Short-term mortality was driven predominantly by physiology and global injury severity rather than anatomical grade. Multicentre prospective and implementation studies are warranted to mitigate residual biases and validate pragmatic bedside risk composites.