Hospital-Associated Venous Thromboembolism in a Pediatric Cardiac ICU: A Multivariable Predictive Algorithm to Identify Children at High Risk

Objectives: Critically ill children with cardiac disease are at significant risk for hospital-Associated venous thromboembolism, which is associated with increased morbidity, hospital length of stay, and cost. Currently, there are no widely accepted guidelines for prevention of hospital-Associated venous thromboembolism in pediatrics. We aimed to develop a predictive algorithm for identifying critically ill children with cardiac disease who are at increased risk for hospital-Associated venous thromboembolism as a first step to reducing hospital-Associated venous thromboembolism in this population. Design: This is a prospective observational single-center study. Setting: Tertiary care referral children's hospital cardiac ICU. Patients: Children less than or equal to18 years old admitted to the cardiac ICU who developed a hospital-Associated venous thromboembolism from December 2013 to June 2017 were included. Odds ratios and 95% CIs are reported for multivariable predictors. Measurements and Main Results: A total of 2,204 separate cardiac ICU encounters were evaluated with 56 hospital-Associated venous thromboembolisms identified in 52 unique patients, yielding an overall prevalence of 25 hospital-Associated venous thromboembolism per 1,000 cardiac ICU encounters. We were able to create a predictive algorithm with good internal validity that performs well at predicting hospital-Associated venous thromboembolism. The presence of a central venous catheter (odds ratio, 4.76; 95% CI, 2.0-11.1), sepsis (odds ratio, 3.5; 95% CI, 1.5-8.0), single ventricle disease (odds ratio, 2.2; 95% CI, 1.2-3.9), and extracorporeal membrane oxygenation support (odds ratio, 2.7; 95% CI, 1.2-5.7) were independent risk factors for hospital-Associated venous thromboembolism. Encounters with hospital-Associated venous thromboembolism were associated with a higher rate of stroke (17% vs 1.2%; p < 0.001). Conclusions: We developed a multivariable predictive algorithm to help identify children who may be at high risk of hospital-Associated venous thromboembolism in the pediatric cardiac ICU.