Longitudinal Determination of Diabetes Complications and Other Clinical Variables as Risk Factors for Diabetic Ketoacidosis in Type 1 Diabetes

OBJECTIVE We aimed to determine whether diabetes complications, such as kidney disease that may impair acid-base buffering capacity, independently predict the risk of subsequent diabetic ketoacidosis (DKA). RESEARCH DESIGN AND METHODS We accessed previously collected 34-year data from the Diabetes Control and Complications Trial and Epidemiology of Diabetes Interventions and Complications study through public data access. Multivariable Cox proportional hazards models with time-varying exposures and covariates were used to examine the associations of macrovascular disease and early and late stages of neuropathy, nephropathy, and retinopathy, with subsequent DKA occurrence as the outcome. RESULTS Of 1,441 participants, 297 experienced 488 DKA events over follow-up. Major adverse cardiovascular events [hazard ratio (HR) 3.16, 95% CI 1.57–6.35, P = 0.001] and late-stage neuropathy, which comprised serious foot ulcer or amputation (HR 1.59, 95% CI 1.04–2.45, P = 0.03) were independently associated with higher DKA risk. Higher risk was also associated with shorter diabetes duration (HR 0.76, 95% CI 0.64–0.91, P = 0.002), female sex (HR 2.04, 95% CI 1.56–2.67, P < 0.001), current insulin pump use (HR 3.04, 95% CI 2.29–4.02, P < 0.001), higher time-updated HbA1c (per additional 1%: HR 1.39, 95% CI 1.29–1.50, P < 0.001), and higher current insulin dose (per 1 additional unit/kg/day: HR 2.32, 95% CI 1.62–3.33, P < 0.001). CONCLUSIONS A major cardiovascular event, foot ulcer, or amputation confers the greatest risk of future DKA independent of previously recognized risk factors, implying a need to target patients with these events for DKA prevention interventions, such as self-management skills for metabolic control, management of depression, and DKA education.