Table 1_Association of delayed enteral nutrition with hospital-acquired pneumonia and 30-day mortality following craniotomy for severe traumatic brain injury: a retrospective cohort study.docx

Sensitivity analyses included forced inclusion of tracheostomy and comparison of the primary model (pre-HAP ventilation duration) with the total ventilation model. A 48-h landmark analysis was also performed. The primary outcome was postoperative HAP; secondary outcomes were ICU stay and 30-day mortality.</p>Results<p>Postoperative HAP incidence was 11.5%.

After adjustment, delayed enteral nutrition (≥48 h) was independently associated with HAP (OR = 4.46, 95% CI: 2.52–8.19, p < 0.001). Other associated factors included age, admission GCS, mechanical ventilation duration, and serum albumin. The model showed good discrimination (AUC = 0.795) and calibration (Hosmer–Lemeshow p = 0.829).

In the time-dependent Cox model, HAP was associated with higher 30-day mortality (HR = 4.31, 95% CI: 1.52–12.25, p = 0.006). The primary model and total ventilation model yielded nearly identical results. In the landmark analysis, mechanical ventilation at 48 h was associated with subsequent HAP (OR = 2.19, 95% CI: 1.02–5.30, p = 0.06).

Forced inclusion of tracheostomy did not materially alter core results.</p>Conclusion<p>In severe TBI patients undergoing craniotomy, delayed enteral nutrition was independently associated with increased HAP risk, and HAP was associated with higher 30-day mortality. These findings highlight the potential importance of early enteral nutrition, but causal inference is limited by the observational design; prospective studies are needed.</p>