Moderate and severe diabetic ketoacidosis at type 1 diabetes onset in children over two decades: A population-based study of prevalence and long-term glycemic outcomes.

Abstract

To investigate in a population-based pediatric cohort: prevalence of moderate-severe diabetic ketoacidosis (DKA) at type 1 diabetes (T1D) diagnosis over two decades and its association with long-term glycemic control. Children <16 years diagnosed with T1D in Western Australia 2000-2019 were included and followed up for ≤14 years. Moderate-severe DKA at diagnosis was defined as serum pH < 7.2 or bicarbonate<10 mmol/L with hyperglycemia and ketosis. HbA1c was measured ~3-monthly. Trend in prevalence of moderate-severe DKA at diagnosis was investigated using a logistic regression model adjusting for sex, age, socioeconomic status, and area of residence. Long-term glycemic control associated with DKA at diagnosis was investigated using linear mixed models adjusting for the same variables and also for visit frequency, CGM and pump use. Moderate-severe DKA occurred in 534 of 2111 (25.3%) participants. Odds of presenting with moderate-severe DKA increased by 4.1% (95% CI: 2.3, 5.9; p< 0.001) per year. Patients with moderate-severe DKA at diagnosis had higher HbA1c levels than other patients initially; the groups were similar between 2 and 6 years duration; from 7 years HbA1c levels tracked higher in the group with moderate-severe DKA at diagnosis with significant differences at 8 and 12 years (p< 0.05). The increasing prevalence of DKA at diagnosis of pediatric T1D is concerning and highlights the need for early detection programs. Unlike a similar US study, this study did not find a consistent, clinically significant relationship between DKA at diagnosis and long-term HbA1c, raising important questions about the influence of other factors on long-term glycemic outcomes.