A. Ammar, H. Fardoun, E. Hart, E. Cagadas, B. Tatar, TH. Ng, KM. Janzen, J. Addison, et al. (2026). Closing the Gap: Evaluating the Safety and Effectiveness of Diabetic Ketoacidosis Nomograms. The Annals of pharmacotherapy, 60(6), 537-545. https://doi.org/10.1177/10600280251388640

PubMed 41235596

[Abstract]
BACKGROUND: Diabetic ketoacidosis (DKA) is a severe complication of diabetes mellitus requiring insulin. However, there is no consensus on insulin infusion titration strategies, resulting in varied nomograms and potentially suboptimal management. Limited comparative data on insulin nomograms underscore the need for a standardized, evidence-based approach.

OBJECTIVE: The purpose of this study is to evaluate the safety and effectiveness of 3 DKA nomograms.

METHODS: This multicenter, retrospective observational study included adult patients with confirmed DKA treated with IV insulin for ≥3 hours. Each site used a different variable-rate nomogram: variable rate weight-adjusted IV insulin infusion with no bolus (VR-W-NB), variable-rate IV insulin infusion with bolus (VR-B) or variable-rate IV insulin infusion with no bolus (VR-NB). The primary endpoint was the incidence of hypoglycemia (<70 mg/dL). Secondary outcomes included time to DKA resolution and recurrence.

RESULTS: A total of 350 patients were included. Hypoglycemia occurred in 36.2% of patients in the VR-W-NB cohort, 23.9% of patients in the VR-B cohort and 6.8% of patients in the VR-NB cohort (P < 0.01). After controlling for confounding factors, hypoglycemia remained the highest in the VR-W-NB (odds ratio [OR] = 6.2, 95% confidence interval [CI] = 2.5-15.3, P < 0.01) and VR-B (OR = 5, 95% CI = 2.1-12.1, P < 0.01) cohorts compared with the VR-NB cohort. Diabetic ketoacidosis resolution was significantly shorter in patients receiving VR-W-NB (median of 9.7 hours, interquartile range [IQR] = 5.6-16.9) compared with those receiving VR-B (12.7 hours, IQR = 9.2-25) and VR-NB (16.1 hours, IQR = 10-24.5). DKA recurrence was higher in those treated with VR-B (19.7%) compared with those treated with VR-W-NB (7%) and VR-NB (15.5%) (P = 0.02).

CONCLUSION AND RELEVANCE: The VR-W-NB was associated with a higher incidence of hypoglycemia and faster DKA resolution compared with VR-B and VR-NB. These findings support the need for prospective studies to define optimal insulin strategies that improve safety, effectiveness, and resource utilization.