Abstract / Summary
Abstract Background Among the major global health burdens, traumatic brain injury (TBI) stands out as a significant source of death and long-term disability, characterized by intricate pathophysiological mechanisms. As an emerging inflammatory biomarker, the leuko-glycemic index (LGI) has demonstrated associations with unfavorable clinical outcomes across multiple disease states. The objective of this investigation was to evaluate how LGI levels relate to clinical outcomes among TBI patients admitted to the intensive care unit (ICU). Methods Clinical information from TBI patients was retrospectively retrieved from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database for this cohort analysis. Calculation of the LGI involved multiplying white blood cell (WBC) count (× 10 9 /L) by the blood glucose level (mmol/L). To mitigate the influence of extreme values, the product was natural logarithm–transformed (Log_LGI). Patients were categorized into quartiles (Q1–Q4) according to Log_LGI. All-cause mortality at 30 days served as the primary endpoint, with 90-day and 365-day mortality as secondary endpoints. Survival differences among quartile groups were evaluated through Kaplan–Meier analysis. The relationship between Log_LGI and clinical outcomes was examined through Cox regression and restricted cubic spline (RCS) modeling. Results The final cohort comprised 2,640 TBI patients admitted to the ICU, spanning a broad range of injury severity, including mild (Glasgow Coma Scale [GCS] score 13–15; 46.02%), moderate (GCS score 9–12; 14.39%), and severe (GCS score 3–8; 39.58%). Elevated Log_LGI values showed significant associations with heightened mortality risk in Cox regression analyses. Q4 demonstrated substantially higher 30-day mortality risk than Q1 (hazard ratio [HR] = 2.66, 95% confidence interval [CI]: 1.92–3.69; P < 0.001). Q3 patients similarly showed elevated 30-day mortality risk relative to Q1 (HR = 1.60, 95% CI: 1.14–2.25; P = 0.007). A nonlinear relationship between Log_LGI and 30-day mortality was identified through RCS modeling (P for nonlinearity < 0.001), with an inflection point at 3.861. Above the threshold of 3.861, every 1-unit rise in Log_LGI was associated with a 2.4-fold higher risk of 30-day mortality ( P < 0.001). Conclusions These results support the LGI as an independent prognostic marker for all-cause mortality in TBI patients admitted to the ICU. This index may serve as a potential biomarker for assessing short-term and intermediate- to long-term mortality risk in this patient population. Clinical trial number. Not applicable.