(527) Diabetic Ketoacidosis and the Incidence of Hospital-Acquired Venous Thromboembolism in Children

Introduction: Hospitalization for diabetic ketoacidosis (DKA) may increase the risk of hospital-acquired venous thromboembolism (HA-VTE) secondary to severe dehydration, hyperosmolarity, and metabolic acidosis. We aimed to estimate HA-VTE frequency in this population and identify prothrombotic risk factors.

Methods: Using the Pediatric Health Information Systems registry, we performed a retrospective cohort study of critically ill patients aged 0 - 21 years treated for DKA between January 2017 - December 2023 within 48 United States (US) pediatric intensive care units (PICUs). Those with a known VTE present on admission were excluded. The primary outcome was the occurrence rate of HA-VTE, including deep venous thromboses, pulmonary emboli, and cerebral sinovenous thromboses. Clinical features were compared for those with and without a HA-VTE using Student’s t, Wilcoxon rank-sum, and Chi-square tests. An adjusted logistic regression model was employed to assess putative HA-VTE risk factors including concurrent cerebral edema, central venous catheterization (CVC), invasive mechanical ventilation (IMV), infection, and prolonged length of stay (LOS).

Results: 27,613 encounters were studied of which 93 (0.3%) developed a HA-VTE. Compared to those without HA-VTE, those with HA-VTE experienced a longer median LOS (10 [interquartile (IQR) range: 5-21] versus 2 [IQR: 2-3] days) and greater rates of cerebral edema (25.8% versus 6.6%), CVC (23.7% versus 1.1%), concurrent infection (72% versus 23.5%), and IMV (39.8% versus 1.4%) (all p< 0.001). In an adjusted model, factors independently associated with increased HA-VTE risk were CVC (adjusted odds ratio [aOR]: 3.04, 95% confidence interval [CI]: 1.49-6.19), infection (aOR: 4.61, 95%CI: 2.81-7.56), IMV (aOR: 9.24, 95%CI: 4.83-17.56), and increasing LOS (aOR: 1.05, 95%CI: 1.02-1.06) (all p< 0.01).

Conclusions: Among 48 US PICUs, the frequency of HA-VTE among critically ill children hospitalized for DKA was 0.3%, a rate substantially lower than prior estimates for children hospitalized for any critical illness (1-2.2%). Factors independently associated with HA-VTE were concurrent CVC, IMV, infection, and prolonged LOS. Following validation, these data should inform the enrollment strategies of upcoming DKA thromboprophylaxis trials.