Severity and Clinical Outcomes Among ICU Patients With Diabetic Ketoacidosis: An Analysis of the Japanese Intensive Care Patient Database.

Abstract

Importance: Diabetic ketoacidosis (DKA) is an acute metabolic emergency characterized by hyperglycemia, ketosis, and acidosis due to insulin deficiency. While ICU admission is frequently considered for DKA, admission thresholds and resource utilization vary across countries and institutions. Current data linking admission severity to clinical outcomes among ICU-admitted patients with DKA remain limited.

Objectives: To describe clinical outcomes and ICU length of stay (LOS) stratified by arterial pH recorded during ICU day 1 among ICU-admitted patients with DKA in Japan.

Design, setting, and participants: Retrospective, multicenter observational study of adult patients admitted with a diagnosis of DKA to 95 ICUs across Japan between fiscal years 2015 and 2022. Patients were stratified by arterial pH recorded during ICU day 1 into severe (pH < 7.00), moderate (7.00 ≤ pH < 7.25), and mild (pH ≥ 7.25) groups.

Main outcomes and measures: The primary outcome was the in-hospital mortality rate. The secondary outcomes were ICU mortality, hospital LOS, and ICU LOS.

Results: Among 866 eligible patients (mild, n = 449; moderate, n = 347; severe, n = 70), in-hospital mortality was low regardless of ICU day 1 pH severity (mild: 4.5%, moderate: 3.2%, and severe: 5.7%). ICU mortality was also low (mild: 1.6% [7/449], moderate: 1.4% [5/347], and severe: 2.9% [2/70]). Hospital LOS was similar across groups (mild: 17.0 d, interquartile range [IQR], 10.0-26.0 d; moderate: 15.0 d, IQR, 9.0-24.0 d; and severe: 15.0 d, IQR, 10.0-22.0 d), whereas ICU LOS tended to increase with severity (mild: 44.3 hr, IQR, 24.6-72.9 hr; moderate: 48.3 hr, IQR, 27.9-76.8 hr; and severe: 57.5 hr, IQR, 31.1-115.0 hr).

Conclusions and relevance: In this nationwide ICU registry, in-hospital mortality was low across arterial pH recorded during ICU day 1-defined severity strata, while ICU LOS tended to be longer in strata with more severe DKA. These findings are descriptive; clinically meaningful mortality differences cannot be excluded due to the small number of events, and further studies are needed to evaluate prognostic models and triage strategies.