Cell-free DNA is associated with central venous catheter-associated deep venous thrombosis in critically ill children

Abstract Introduction: Neutrophil extracellular traps (NETs) are webs of histone-coated DNA released from activated neutrophils. Increased NETosis is associated with a rise in plasma cell-free DNA (cfDNA) and has been linked with an increased risk of thrombosis in adults. There are limited data assessing NETs in children. Central venous catheters (CVC) are the most important risk factor for deep venous thrombosis (DVT) in pediatrics. In children, CVC-associated DVT often occur in the setting of inflammation. While cfDNA can originate from other sources besides NETs, we used this as a preliminary marker of NETosis in this study. We investigated the association of CVC-associated DVT with cfDNA in a secondary analysis of samples from 2 studies of critically ill children with CVCs. CRETE was a multicenter, phase 2b randomized trial of prophylactic enoxaparin compared to no anticoagulation for primary prevention of CVC-associated DVT. ATLAS was a multicenter prospective cohort study of critically ill children with a CVC who were at a high risk of bleeding. Methods: Whole blood was collected in citrated tubes from participants in ATLAS and CRETE within 24 hours of CVC insertion and centrifuged to obtain platelet-poor plasma. Demographic and clinical variables were collected in each study. Severity of illness was assessed using the Pediatric Index of Mortality 2 score. Plasma cfDNA levels were quantified via SYTOX™ Green (ThermoFisher Scientific) fluorescent plate assay. CVC-associated DVT was determined by surveillance ultrasounds per the original study protocols. Clinically relevant bleeding was defined as per the International Society on Thrombosis and Haemostasis. Bivariate associations were assessed with Wilcoxon rank-sum test and multivariable associations with logistic regression. Results: A total of 73 children from CRETE (n=41) and ATLAS (n=32) with available plasma samples were included in this analysis. Of these, 18 (25%) were <1 year of age, 39 (53%) were 1-13 years of age, and 16 (22%) were >13 years of age. Prophylactic anticoagulation was administered in 21 participants (29%). CVC-associated DVT developed in 28 (39%) children with an ultrasound available.The median cfDNA for the entire cohort was 0.82 µg/mL (IQR: 0.49, 2.19 µg/mL). The median cfDNA was higher for those with CVC-associated DVT (0.88 µg/mL: IQR: 0.79, 1.39 µg/mL) than in those without CVC-associated DVT (0.69 µg/mL; IQR: 0.61, 0.89 µg/mL), p=0.009. Increased cfDNA was associated with CVC-associated DVT after controlling for age group, prophylactic anticoagulation, and severity of illness (odds ratio: 4.8, 95% CI: 1.3-18.7, p=0.02). Clinically relevant bleeding occurred in 17 (23%) children. We did not find any association between clinically relevant bleeding and cfDNA. Median cfDNA levels for children with and without clinically relevant bleeding were 0.78 µg/mL (IQR: 0.61, 0.92 µg/mL) and 0.82 µg/mL (IQR: 0.64, 1.17 µg/mL; p=0.37), respectively. Conclusions: cfDNA levels on the day of CVC insertion are associated with CVC-associated DVT in critically ill children. Ongoing analysis of other markers of NETosis will further investigate the association between NETs and pediatric CVC-associated DVT, the role of inflammation in the development of DVT in children, and whether targeted therapies to block NETs would reduce thrombotic risk.