Abstract
<title>Abstract</title> <p> Background Prolonged intensive care unit (ICU) stay increases healthcare costs, complications, and adverse outcomes. This study aimed to identify admission determinants of prolonged ICU stay and compare neurological with non-neurological critically ill patients. Material and Methods This retrospective cohort study included 609 adults admitted to five multidisciplinary ICUs of a tertiary care hospital between October 2019 and October 2024. Admission comorbidities, Glasgow Coma Scale (GCS), APACHE II and SOFA scores, and laboratory variables were recorded. Prolonged ICU stay was defined as a length of stay ≥ 14 days. Multivariable logistic regression analyses were restricted to admission variables, excluding mechanical ventilation and tracheostomy as time-dependent variables. Results Median age was 71 (63–83) years, 43.8% of patients were women, median ICU stay was 19 (10–30) days, and 66.2% remained in the ICU for ≥ 14 days. Neurological patients had shorter ICU stays (15 vs 20 days, <italic>p</italic> = 0.002) and a lower prevalence of prolonged stay (57.8% vs 69.7%, <italic>p</italic> = 0.006) than non-neurological patients. Lower admission GCS (OR = 0.90, 95% CI 0.84–0.95) and absence of diabetes mellitus (OR = 0.64, 95% CI 0.43–0.96) independently predicted prolonged ICU stay, whereas higher SOFA score, medical admission, and neutrophil-to-lymphocyte ratio showed borderline associations. Conclusions Lower admission GCS and absence of diabetes mellitus were independent admission determinants of prolonged ICU stay. Similar predictors were observed in neurological and non-neurological patients, indicating that baseline physiological severity rather than admission diagnosis primarily influences prolonged ICU utilization. Early identification of high-risk patients may improve individualized management and ICU resource allocation. </p>