Risk Factors for Umbilical Venous Catheter-Associated Thrombosis in Very Low Birth Weight Infants.

Abstract Background: Thrombosis in neonates is a rare but serious occurrence that is usually associated with central catheterization. Among acquired risk factors, thrombocytosis has often been thought to play a role in neonatal thrombosis, but little evidence exists to support this impression. Objectives:To investigate the effects of platelet count on catheter-related thrombosis in neonates.To investigate the effects of being small for gestational age (SGA) on catheter-related thrombosis in neonates. We hypothesized that neonates with catheter-related thrombosis would have relative thrombocytosis and would be SGA. Methods: The present retrospective study was performed using data from a randomized trial of duration of umbilical venous catheters (UVC) placement among infants <1250 g birth weight (Butler-O’Hara, Pediatrics2006;118:e25–e35). In this study, all subjects had UVC that were left in place for up to 28 days. All subjects were screened biweekly for thrombi with echocardiograms. Twenty-two cases of UVC-associated thrombosis were identified in this sample. The remaining study sample (n=188) served as controls. Data on thrombosis, platelets, gestational age, birth weight, hematocrit, serum sodium (as a measure of dehydration), duration of catheter placement, study group assignment and demographic factors were collected using database and record review. Results: Among the total subjects (n=210), 112 (53%) were males and 126 (60%) were Caucasians, with mean gestational age of 27.7 ± 2.1 wks (SD) and mean birth weight of 923 ± 195g. Bivariate analysis revealed significant association of thrombosis with hematocrit >55% in the first wk (OR, 5.4; 95% CI, 2.0–14.6; p=0.0003), being small for gestational age (OR, 2.9; 95% CI, 1.2–7.4; p=0.02), lower platelet counts in the first wk (193 ± 57 x 103/uL in infants with thrombus vs. 238 ± 70 x 103/uL in infants without thrombus, p=0.005) and gestational age (27.8 ± 2.5 wks in infants with thrombus vs. 27.6 ± 2.0 wks in infants without thrombus, p=0.02). In multivariate logistic regression analysis, only higher hematocrit was independently associated with thrombus (OR, 3.9; 95% CI 1.3–12.6; p=0.02). There was a trend towards an independent negative association between platelets and thrombosis (OR, 0.93 per 10 x 103/uL platelet rise; 95% CI, 0.85–1.02; p=0.12). Conclusion: This study demonstrates a significant, independent association of elevated hematocrit and development of UVC-associated thrombosis. We did not observe an increased risk of thrombosis with increased platelet count. Careful monitoring for catheter-associated thrombosis is suggested for neonates with hematocrit >55% in the first wk of life.