Social and Metabolic Characteristics Associated With Multiple DKA Admissions at a Large County Hospital

• Published in: Journal of the Endocrine Society 2024-01, Vol.8 (3), p.bvad173
• Main Authors: Peedikayil, Josh, Reddy, Shrenika, Nair, Rohit, Gunasekaran, Uma, Nelson, Carolyn, Shakoor, Musa, Ahmad, Zahid
• Format: Article
• Language: English
• Subjects: Analysis; Electronic records; Homelessness; Marijuana; Medical records; Medical research; Medicine, Experimental; Social aspects; Substance abuse; Type 1 diabetes; Type 2 diabetes
• ISSN: 2472-1972; 2472-1972
• DOI: 10.1210/jendso/bvad173

Abstract Context Diabetic ketoacidosis (DKA) is a preventable, deadly, and costly complication of type 1 diabetes mellitus (T1DM). Some individuals with T1DM have recurrent DKA admissions. Objective We sought to characterize social factors that differ between patients with single vs multiple DKA admissions at an urban, safety-net hospital. Methods We queried the electronic health records for T1DM patients admitted for DKA from 2019 to 2021. Admission laboratory values, demographic information, and detailed social histories were collected and analyzed statistically, including logistical regression. Results A total of 243 patients were admitted for DKA, 64 of whom had multiple DKA admissions. There was no significant difference between the groups in their admission laboratory values, hospital length of stay, health-care payer status, history of homelessness, current employment, living alone, independence of activities of daily living, and barriers to discharge. T1DM patients with multiple DKA admissions had greater rates of substance use disorder (33.0% vs 60.9%; P < .001), especially with cannabis (6.7% vs 25.0%; P < .001), tobacco (26.3% vs 46.3%; P = .002), and psychoactive substance use (1.1% vs 6.3%; P = .043). Regression models of substance use showed increased risk with any substance use (odds ratio [CI] 3.17 [1.78-5.73]; P < .001) and cannabis (3.70 [1.55-8.83]; P = .003). Conclusion We identified substance use as a possible predictor of T1DM patients at risk for multiple DKA admissions. Our findings identify a group of T1DM patients for whom interventions may help to decrease recurrence of DKA episodes within similar community hospital populations.