Association of all glycemic spectrum with the risk of sepsis and mortality: two cohort studies in the UK and USA

Background: Sepsis causes substantial morbidity and mortality, underscoring the urgent need for standardized risk biomarkers. Prior studies of glycated hemoglobin (HbA1c) and sepsis are inconsistent and largely confined to diabetic populations. We examined the full glycemic spectrum in individuals with and without diagnosed diabetes, assessing sepsis incidence in the general population as well as short-term and intermediate-term all-cause mortality among septic patients to evaluate HbA1c for risk stratification. Materials and Methods: In the UK Biobank, HbA1c was categorized as low-normal (≤34 mmol/mol), normal (35–41 mmol/mol), pre-diabetes (42–47 mmol/mol), undiagnosed diabetes (≥48 mmol/mol), and diagnosed diabetes. HbA1c categories were further stratified into percentile groups. Cox proportional hazards models were used to estimate hazard ratios and 95% confidence intervals for associations of HbA1c with incident sepsis and all-cause mortality at 28, 90, and 365 days after sepsis. We additionally analyzed 2749 septic patients with diagnosed diabetes in the MIMIC-IV database to examine associations between HbA1c and all-cause mortality. Results: Individuals with pre-diabetes, undiagnosed diabetes, and diagnosed diabetes had incremental risks for each sepsis outcome compared to those with normal HbA1c. Percentile analyses showed increased sepsis risks at the 0–5th percentile within the low-normal category, the 95–100th percentile within the normal category, and at HbA1c ≥48 mmol/mol among those with diagnosed diabetes. Associations for diagnosed diabetes were more pronounced in participants aged <65 years compared with those aged ≥65 years. Sensitivity analyses yielded results consistent with the primary analyses. In the ICU cohort of septic patients with diabetes, both HbA1c ≤34 mmol/mol and HbA1c ≥48 mmol/mol were associated with higher all-cause mortality at 28, 90, and 365 days. Conclusion: HbA1c could serve as a risk-stratification biomarker for sepsis susceptibility and sepsis prognosis, supporting population-level stratified management and informing clinical decision-making for hospitalized septic patients with diabetes.