Objective. We sought to obtain representative data on the risk factors, diagnosis, current management, and short-term outcome of neonatal thrombosis. Research Design. A case registry was established at McMaster University. Standardized questionnaires were mailed to collaborators at participating centers every 4 to 6 months. Setting. Eighty-five level III and modified level II neonatal units in North America, Europe, and Australia were invited to join the registry. Patients. Eligible infants were born between January 1990 and June 1993. Large-vessel thrombosis was diagnosed during the first month of life or up to 44 weeks post-conception after premature birth. The clinical impression of thrombotic vessel obstruction was confirmed using at least one imaging technique. Results. Physicians in 64 centers expressed their willingness to participate. A total of 97 cases (excluding stroke) were registered from 29 centers. Spontaneous renal venous thrombosis (n = 21) was diagnosed at a median age of 2 days. The other venous (n = 39), arterial (n = 33), and mixed (n = 4) thromboses presented later; 89% of them were associated with an intravascular catheter and 29% with systemic infection. Doppler ultrasonography was the definitive diagnostic test in 68% of cases; contrast angiography was performed infrequently (14%). A third of all patients (but 62% of infants with renal venous thrombosis) received supportive therapy only. Thrombolytic agents were prescribed for 28% of catheter-associated venous thromboses and 30% of all arterial thromboses. The remainder of the patients were given heparin. Most patients (82%) survived to hospital discharge. Mortality rates were highest among infants with aortic thrombosis or central venous line-associated thrombosis affecting the right atrium or the superior vena cava (33%). Conclusions. Neonatal thrombosis is diagnosed fairly rarely. With the exception of spontaneous renal venous thrombosis, almost all cases are associated with indwelling catheters. Doppler ultrasound techniques are the most popular means of confirming the diagnosis in virtually all centers. Treatment varies greatly among different centers, probably because of the lack of scientific evidence about the optimum management of affected infants.