OBJECTIVE: To identify risk factors associated with venous and arterial thrombosis in sick neonates admitted to the neonatal intensive care unit. STUDY DESIGN: A case-control study was conducted at 2 centers between January 2010 and March 2014 using the Children's Hospital Neonatal Database dataset. Cases were neonates diagnosed with either arterial or venous thrombosis during their neonatal intensive care unit stay; controls were matched in a 1:4 ratio by gestational age and presence or absence of central access devices. Bivariable and conditional logistic regression analyses for venous and arterial thrombosis were performed separately. RESULTS: The overall incidence of neonatal thrombosis was 15.0 per 1000 admissions. A higher proportion of neonates with thrombosis had presence of central vascular access devices (75% vs 49%; P < .01) were of extremely preterm gestational age (22-27 weeks; 26% vs 15.0%; P <.05) and stayed >/=31 days in the neonatal intensive care unit (53% vs 32.9%; P <.01), when compared with neonates without thrombosis. A final group of 64 eligible patients with thrombosis and 4623 controls were analyzed. In a conditional multivariable logistic regression model, venous thrombosis was significantly associated with male sex (AOR, 2.12; 95% CI, 1.03-4.35; P = .04) and blood stream infection (AOR, 3.47; 95% CI, 1.30-9.24; P = .01). CONCLUSIONS: The incidence of thrombosis was higher in our neonatal population than in previous reports. After matching for central vascular access device and gestational age, male sex and blood stream infection represent independent risk factors of neonatal venous thrombosis. A larger cohort gleaned from multicenter data should be used to confirm the study results and to develop thrombosis prevention strategies.