Operative versus conservative management of distal humerus fractures in patients aged 65 years and older: a target-trial-informed retrospective cohort study of five-year mortality.

INTRODUCTION: Older patients with distal humerus fractures die more often after conservative than after operative treatment in unadjusted comparisons; whether this reflects treatment or the selection of healthier patients for surgery is unresolved. We estimated the association between operative versus conservative management and five-year all-cause mortality in patients aged ≥ 65 years. MATERIALS AND METHODS: Retrospective comparative cohort study at two tertiary trauma centres (2012-2025). Cases were ascertained by automated screening of codes and free text, then adjudication of all uncertain records against source notes and injury radiographs, masked to treatment and vital status. Patients were classified by management initiated within 14 days as operative (n = 75) or conservative (n = 78); vital status was registry-linked. The primary estimand was the five-year restricted mean survival time (RMST) difference; principal adjusted estimates were inverse-probability-weighted (IPW) RMST and survival differences, requiring no proportional-hazards assumption. Cox, window and landmark analyses are supportive. RESULTS: Among 153 patients (52 deaths), conservatively managed patients were older (median 83 vs. 73 years, p < 0.001). Unadjusted five-year mortality was 41.9% versus 18.4%: RMST difference + 11.2 months (95% CI + 5.6 to + 16.8), log-rank p = 0.006. After weighting for age, sex, comorbidity and centre this shrank to + 3.4 months (95% CI - 1.5 to + 8.6; adjusted five-year survival difference + 1.4%, - 14.7 to + 18.5), and to + 1.0 month (- 2.9 to + 5.5) within the covariate-overlap cohort (65-84 years). Comorbidity (HR 1.35 per Charlson point) and age (HR 1.72 per decade) predicted mortality in both arms. Sensitivity analyses were concordant. CONCLUSIONS: The large unadjusted survival advantage of operative fixation shrank substantially after adjustment, and the remaining estimates are too imprecise to support any claim of survival benefit or harm. Because functional outcomes, the principal indication for surgery, were not measured, the study cannot inform individual treatment choice. Crude survival differences here should not be read as treatment effects. LEVEL OF EVIDENCE: Level III, Retrospective comparative cohort study.