SAT-LB119 Diabetic Ketoacidosis With PUMP Failure: Analysis of the 2006-2016 U.S. Kid Inpatient Database

Background: Insulin pumps are valuable tools in diabetes management and their use has increased dramatically over the past decade. Unfortunately, insulin pump use has also been associated with diabetic ketoacidosis (DKA), relating to pump malfunctions that result in the disruption in insulin administration. Our objective was to examine the prevalence and characteristics of DKA admissions associated with pump failure among pediatric patients. Methods: We used the national Kids’ Inpatient Database to identify pediatric admissions with a primary diagnosis of DKA in years 2006, 2009, 2012, and 2016. We defined a DKA-pump failure admission as an admission with either a primary diagnosis of DKA plus a secondary diagnosis of pump failure/complication or conversely, a primary diagnosis of pump failure/complication with a secondary diagnosis of DKA. We used descriptive statistics and logistic regression to describe the annual trends and characteristics of children admitted for DKA with or without pump failure. Lastly, logistic regression was used to assess the impact of pump failure on length of stay and severity of illness during DKA admissions. Results: Our dataset included 166,583 DKA admissions, of which 2,291 (1.4%) were associated with a primary or secondary diagnosis of insulin pump failure. Between 2006 and 2016, the number of total DKA admissions increased by 58%. Admissions for DKA with pump failure increased from 387 to 665 admissions during this time. Among all children admitted with DKA, those with pump failure were primarily older (60% above age 12), mostly white (63%), female (57%), from urban areas (78%), and almost 2/3rds had private insurance (60%). Adjusted analyses revealed that compared to DKA admissions without pump failure, pump failure was associated with older age, white race, residing in a rural area, private insurance, and higher income. Pump failure admissions were more likely in western and southern hospitals, otherwise there were no significant differences with respect to hospital characteristics. Compared to DKA admissions without pump failure, DKA admissions associated with pump failure had a longer mean length of stay (2.6 vs 1.5 days) and were more likely to have a higher severity of illness category. Conclusion: In this national sample, DKA with pump failure was more often observed among white, privately insured and high income children; these patient characteristics likely reflect the population of youth with diabetes who are more likely prescribed pumps in the US. Admissions for DKA concurrent with insulin pump failure accounted for a minority of pediatric DKA admissions but these admissions were associated with longer lengths of stay and severity of illness. Pump failure has important implications for care and management of children with diabetes.