Impact of dynamic blood glucose trajectories on mortality in critically ill non-traumatic intracranial hemorrhage patients: a retrospective cohort study.
Four distinct blood glucose trajectories identified during the first 72 hours of ICU admission were independently associated with 28-day and longer-term mortality in critically ill non-traumatic intracranial hemorrhage patients.
Key Findings
Results
Four distinct longitudinal blood glucose trajectory subtypes were identified in critically ill non-traumatic ICH patients during the first 72 hours of ICU admission.
Trajectories were identified using Latent Class Mixed Modeling (LCMM) applied to blood glucose data collected during the initial 72 hours of ICU admission.
The four trajectories were labeled: Trajectory 1 (persistent low), Trajectory 2 (moderate stable), Trajectory 3 (progressive decline), and Trajectory 4 (progressive increase).
The study included 2,260 critically ill non-traumatic ICH patients from the MIMIC-IV database (United States, 2008-2019).
Significant differences in baseline characteristics were observed across the four trajectory groups.
Results
Trajectory 2 (moderate stable) was associated with significantly elevated 28-day mortality risk compared to Trajectory 1 (persistent low).
The hazard ratio for 28-day mortality in Trajectory 2 versus Trajectory 1 was 1.71 (95% CI: 1.33–2.20).
This association remained significant after adjusting for baseline blood glucose and other potential confounders.
Consistent findings were observed for 180-day and 1-year mortality endpoints.
Results
Trajectory 3 (progressive decline) was associated with significantly elevated 28-day mortality risk compared to Trajectory 1 (persistent low).
The hazard ratio for 28-day mortality in Trajectory 3 versus Trajectory 1 was 1.87 (95% CI: 1.25–2.80).
This association was independent of baseline blood glucose and other potential confounders.
Consistent findings were observed for 180-day and 1-year mortality endpoints.
Results
Trajectory 4 (progressive increase) was associated with the highest 28-day mortality risk among all trajectory groups compared to Trajectory 1 (persistent low).
The hazard ratio for 28-day mortality in Trajectory 4 versus Trajectory 1 was 2.41 (95% CI: 1.64–3.54).
This was the largest hazard ratio among the three non-reference trajectory groups.
This association remained significant after adjusting for baseline blood glucose and other potential confounders.
Consistent findings were observed for 180-day and 1-year mortality endpoints.
Results
Blood glucose trajectories during the first 72 hours of ICU admission were independently associated with both short-term and longer-term mortality in critically ill non-traumatic ICH patients.
Associations were assessed using multivariable Cox proportional hazards regression.
Mortality outcomes examined included 28-day, 180-day, and 1-year mortality.
Elevated mortality risk was observed across Trajectories 2, 3, and 4 relative to Trajectory 1 at all three time points.
The authors note these findings 'may support early risk stratification and inform future prospective studies.'
The study design was retrospective, using data from a single database (MIMIC-IV), which limits causal inference.
What This Means
This research examined how blood sugar levels change over time in patients who are critically ill due to non-traumatic bleeding in the brain (such as from a ruptured blood vessel), and whether those patterns of change are linked to survival. Using data from over 2,200 patients in a large U.S. hospital database, the researchers used a statistical method to group patients into four categories based on how their blood glucose behaved during the first three days in the ICU: those with persistently low glucose, those with moderately stable glucose, those whose glucose progressively declined, and those whose glucose progressively increased. They then tracked whether patients survived over 28 days, 6 months, and 1 year.
The study found that patients with persistently low blood glucose (Trajectory 1) had the best survival outcomes, while all other groups had significantly higher risks of dying. The group with progressively increasing blood glucose had the worst outcomes, with more than twice the risk of 28-day death compared to the low-glucose group, even after accounting for other factors like how sick patients were at admission. Patients with progressively declining glucose and those with moderately stable glucose also faced substantially elevated mortality risks. These associations held up across all three time points measured.
This research suggests that tracking how blood sugar changes over time—not just measuring it once—may provide important information about which patients with brain bleeding are at greatest risk of dying. This could help clinicians identify high-risk patients earlier and may help guide the design of future studies looking at whether managing blood glucose trajectories can improve outcomes. The retrospective nature of the study means it cannot prove that the blood glucose patterns directly cause worse outcomes, but it highlights dynamic glucose monitoring as a potentially meaningful tool in this patient population.
Yang M, Lu Y, Liu Y, Chen Y, Pan X, Zhang X, et al.. (2026). Impact of dynamic blood glucose trajectories on mortality in critically ill non-traumatic intracranial hemorrhage patients: a retrospective cohort study.. BMC neurology. https://doi.org/10.1186/s12883-026-05351-0