A four-year retrospective study ending on 31st December 2022 was performed on ventilated polytrauma patients admitted to a Level 1 Trauma Centre ICU. The institutional trauma registry was queried for: polytrauma patients (defined as Abbreviated Injury Scale (AIS) > 2 in two or more body regions), older than 16 years and admitted to the ICU on mechanical ventilation. Patients who did not receive EN during ICU admission, those who were discharged or died within 48 h from admission were excluded. Clinical data beyond the trauma registry was collected from the electronic Record for Intensive Care (eRIC) and patient files.

Collected patient characteristics included: age, sex, mechanism of injury, Injury Severity Score (ISS) and AIS of all body regions. The worst vital signs in the emergency department were collected (heart rate, systolic blood pressure, and Glasgow Coma Scale). Laboratory tests included: on arrival serum lactate and haemoglobin concentrations. Blood products transfused within 24 h of injury were also recorded.

Nutritional data included: hours from ICU admission to EN, feeding route, dietician input, time to nutritional goal, and interruptions to EN (longer than 6 h). EN interruptions were categorised by causation: ileus, airway event (e.g. extubation planning or attempt), ICU diagnostics or procedures (e.g. transoesphageal echocardiogram, endoscopy), procedures outside the ICU (e.g. imaging, surgery), accidental NGT removal, and undetermined (reason was not identified). Surgical procedures were categorized into craniotomy/craniectomy, laparotomy with and without bowel resection, damage control laparotomy, orthopedic procedures, and others. Disposition from hospital (home, nursing facility, acute care facility, rehabilitation facility, or death), as well as complications during ICU admission, (ventilator-acquired pneumonia (VAP), sepsis, ileus and vomitus) were also collected. Severe traumatic brain injury (TBI) was defined as AIS Head ≥ 4 and initial GCS ≤ 9. Severe abdominal trauma was defined as AIS abdomen ≥ 4.

Primary outcome was time to EN initiation. Secondary outcomes included: EN interruptions, dietician involvement, time to reach nutritional goal, ICU and hospital LOS, mortality, MOF, complications.

Subgroup analysis was performed between patients who received early EN (< 24 h from admission to ICU) and late EN (> 24 h).

Normality of distribution was determined with Shapiro-Wilk test. While categorical variables were tested with Chi-Square, continuous variables were tested with Wilcoxon-Mann–Whitney test.

A logistic regression model was used to calculate the odds of developing multi organ failure (MOF) given a change in time to EN. Furthermore, as ISS is not a true continuous variable, the appropriateness of using linear regression model was assessed using model residual diagnostics and the assumption was deemed appropriate. The association between ISS as a continuous variable and time to EN was tested with a linear regression model. Statistical significance was set at p-values < 0.05.