To determine the associations of relative hypoglycemia and hemoglobin A1c-adjusted time in blood glucose (BG) band (HA-TIB) with mortality in critically ill patients.
Retrospective cohort investigation.
University-affiliated adult medical-surgical ICU.
Three thousand six hundred fifty-five patients with at least four BG tests and hemoglobin A1c (HbA1c) level admitted between September 14, 2014, and November 30, 2019.
MEASUREMENTS AND MAIN RESULTS:
Patients were stratified for HbA1c bands of <6.5%; 6.5–7.9%; greater than or equal to 8.0% with optimal affiliated glucose target ranges of 70–140, 140–180, and 180–250 mg/dL, respectively. HA-TIB, a new glycemic metric, defined the HbA1c-adjusted time in band. Relative hypoglycemia was defined as BG 70–110 mg/dL for patients with HbA1c ≥ 8.0%. Further stratification included diabetes status-no diabetes (NO-DM, n = 2,616) and preadmission treatment with or without insulin (DM-INS, n = 352; DM-No-INS, n = 687, respectively). Severity-adjusted mortality was calculated as the observed:expected mortality ratio (O:EMR), using the Acute Physiology and Chronic Health Evaluation IV prediction of mortality. Among NO-DM, mortality and O:EMR, decreased with higher TIB 70–140 mg/dL (p < 0.0001) and were lowest with TIB 90–100%. O:EMR was lower for HA-TIB greater than or equal to 50% than less than 50% and among all DM-No-INS but for DM-INS only those with HbA1 greater than or equal to 8.0%.Among all patients with hba1c greater than or equal to 8.0% And no bg less than 70 mg/dl, mortality was 18.0% For patients with relative hypoglycemia (bg, 70–110 mg/dl) (p < 0.0001) And was 0.0%, 12.9%, 13.0%, And 34.8% For patients with 0, 0.1–2.9, 3.0–11.9, And greater than or equal to 12.0 Hours of relative hypoglycemia (p < 0.0001).
These findings have considerable bearing on interpretation of previous trials of intensive insulin therapy in the critically ill. Moreover, they suggest that BG values in the 70–110 range may be deleterious for patients with HbA1c greater than or equal to 8.0% and that the appropriate target for BG should be individualized to HbA1c levels. These conclusions need to be tested in randomized trials.