HbA1c Stratification Identifies High-Risk ICU Patients: Mortality and Complications in a Retrospective Critical Care...

IntroductionGlycemic variability is common in critical illness and can obscure the prognostic significance of premorbid glycemic status. Hemoglobin A1c (HbA1c) reflects chronic glycemia and may improve risk stratification beyond acute glucose measurements. We examined whether HbA1c-defined glycemic phenotypes were associated with in-hospital complications and mortality among critically ill adults.MethodsWe conducted a retrospective cohort study using MIMIC-IV (v3.1), a de-identified database of ICU admissions at Beth Israel Deaconess Medical Center (2008-2019). Adults (≥18 years) with a documented HbA1c during the index hospitalization and first ICU admission were included (N = 13,552). Patients were classified into eight mutually exclusive groups based on diabetes diagnosis status and HbA1c: healthy (≤5.7% without diabetes), undiagnosed prediabetes (5.8%-6.4%), undiagnosed diabetes (≥6.5%), and five diagnosed-diabetes strata (<6.0%, 6.0%-6.9%, 7.0%-7.9%, 8.0%-8.9%, ≥9.0%). Multivariable logistic regression adjusted for age and sex compared each group with healthy controls for composite complications and in-hospital mortality.ResultsHbA1c was available in 20.7% of first ICU admissions; 40.7% of included patients had diagnosed diabetes and 2.6% had undiagnosed diabetes. Compared with healthy controls, odds of any complication were higher in undiagnosed prediabetes (OR 1.28, 95% CI 1.11-1.48; p < .001) and several diagnosed-diabetes strata (OR range 1.31-1.90; all p ≤ .006). Renal complications showed the most consistent elevation across dysglycemic groups, including very poor glycemic control (OR 2.22, 95% CI 1.95-2.53; p < .001). Sepsis-related complications were increased across all diabetes-range groups; undiagnosed diabetes had OR 2.08 (95% CI 1.53-2.83; p < .001). In-hospital mortality was highest in undiagnosed diabetes (OR 2.33, 95% CI 1.73-3.14; p < .001).ConclusionAmong ICU patients in whom HbA1c was measured, HbA1c-defined dysglycemia, particularly undiagnosed diabetes, was associated with substantially higher ICU morbidity and mortality. Because HbA1c testing was selective and provided limited incremental prognostic value beyond established severity scores, these findings should be considered hypothesis-generating and require prospective validation before supporting routine HbA1c screening.