Evaluation of Prognostic Factors for Prolonged Intensive Care Unit Hospitalization After Trauma: Roles of Hemorrhagic Burden and Admission Laboratory Abnormalities

Abstract

Background and Objectives: Prolonged intensive care unit (ICU) hospitalization after trauma reflects sustained critical-care needs, but early prognostic assessment remains challenging. The primary objective was to identify factors associated with prolonged ICU hospitalization (≥7 days) using admission characteristics and treatment information available during the first 24 h. The secondary objective was to explore whether these associations differed according to a composite score of admission laboratory abnormalities (CE score). Materials and Methods: This retrospective cohort study included 2638 patients admitted between 1 January 2006 and 31 December 2021 from a prospectively maintained Level I trauma registry. Prolonged ICU hospitalization was defined as ICU length of stay (LOS) ≥7 days. Multivariable regression evaluated independent associations between early clinical factors and prolonged ICU hospitalization. Exploratory interaction analyses evaluated effect modification by the CE score, a composite of admission laboratory abnormalities. In-hospital deaths were retained in the primary cohort. The potential influence of in-hospital death on ICU LOS was evaluated in an additional survivor-restricted sensitivity analysis. Sensitivity analyses used ICU LOS ≥ 10 days and log-transformed ICU LOS. Results: Overall, 757/2638 patients (28.7%) had an ICU LOS ≥ 7 days. Older age (adjusted OR [aOR] 1.01 per year, 95% CI 1.01–1.02; p < 0.001), higher Injury Severity Score (aOR 1.06 per point, 95% CI 1.05–1.08; p < 0.001), massive transfusion (aOR 1.52, 95% CI 1.09–2.12; p = 0.013), and damage control surgery (aOR 2.35, 95% CI 1.60–3.46; p < 0.001) were independently associated with prolonged ICU hospitalization, whereas admission lactate, systolic blood pressure, and angioembolization were not. CE interactions were observed for lactate, SBP, and massive transfusion (likelihood-ratio p < 0.001, p = 0.006, and p = 0.037, respectively). Using the ≥10-day threshold, associations persisted for massive transfusion (aOR 1.56, 95% CI 1.10–2.22; p = 0.012) and DCS (aOR 1.88, 95% CI 1.27–2.79; p = 0.002). Among 688 hospital survivors in the additional analysis, none of the three CE interactions was statistically significant (p = 0.208, p = 0.572, and p = 0.574, respectively). Conclusions: Older age, greater injury severity, massive transfusion, and DCS were associated with prolonged ICU hospitalization after trauma. Exploratory CE interactions were not statistically significant in the survivor-restricted analysis and require further investigation. These findings do not establish the clinical utility of the CE score.