Mingkun Yang, Yanwen Lu, Yijing Liu, Yifeng Chen, Xindian Pan, Xiaofeng Zhang, Weihang Hu, Jing Yan
Four distinct blood glucose trajectories observed during the first 72 h of ICU admission among critically ill patients with non-traumatic ICH were independently associated with 28-day and longer-term mortality. These findings may support early risk stratification and inform future prospective studies.
BACKGROUND: Dysglycemia is frequently observed in patients with non-traumatic intracranial hemorrhage (ICH) and has been linked to adverse clinical outcomes. However, the prognostic impact of longitudinal blood glucose trajectories in critically ill non-traumatic intracranial hemorrhage remains incompletely understood.
METHODS: We conducted a retrospective cohort study using the MIMIC-IV database (United States, 2008-2019), including 2,260 critically ill non-traumatic ICH patients. Latent class mixed modeling (LCMM) was employed to identify distinct longitudinal blood glucose trajectory subtypes, utilizing blood glucose data collected during the initial 72 h of ICU admission. Associations between trajectory groups and mortality were assessed using multivariable Cox proportional hazards regression.
RESULTS: Four glucose trajectories were identified: Trajectory 1 (persistent low), Trajectory 2 (moderate stable), Trajectory 3 (progressive decline), and Trajectory 4 (progressive increase). Significant differences in baseline characteristics were observed across trajectories. Even after adjusting for baseline blood glucose and other potential confounders, the 28-day mortality risk remained significantly elevated in Trajectories 2-4 compared to Trajectory 1, with hazard ratios of 1.71 (95% CI: 1.33-2.20), 1.87 (95% CI: 1.25-2.80), and 2.41 (95% CI: 1.64-3.54), respectively. Consistent findings were observed for 180-day and 1-year mortality.
CONCLUSION: Four distinct blood glucose trajectories observed during the first 72 h of ICU admission among critically ill patients with non-traumatic ICH were independently associated with 28-day and longer-term mortality. These findings may support early risk stratification and inform future prospective studies.