Assistant Professor of Clinical Pediatrics Children's Hospital Los Angeles Los Angeles, California, United States
Background: Central venous catheters (CVCs) are frequently used in NICU patients and are critical for survival. Adverse events such as thromboses and subsequent vessel death can be associated with CVC use. Catheter characteristics, placement factors, and patient characteristics may influence risk for thrombosis, however the prevalence of and risk factors for CVC-related thrombosis in the NICU population are not well understood. Objective: To retrospectively examine the 1) prevalence of CVC-related thrombosis and 2) risk factors for CVC-related thrombosis in patients cared for in a level 4 NICU at a single center. Design/Methods: We used natural language processing to create a database of every CVC placed in NICU patients at our institution from 1/2017-2/2023. Data was obtained from clinical notes and daily nursing documentation. We collected data related to catheter characteristics such as type of CVC (PICC, temporary tunneled, permanent tunneled, etc.), number of lumens, catheter material, catheter size, as well as placement factors (vessel used, vessel size, ultrasound guidance, number of attempts, location of placement, proceduralist), and patient factors (age, weight). We used patient-level pharmacy and radiology data to identify the outcome of thrombosis. Results: 3,292 CVCs were placed in 1,782 unique NICU patients. 6.4% of CVCs were associated with thrombosis. PICCs were the most common CVC (77.0%), followed by umbilical (12.2%), temporary tunneled (7.7%), and permanent tunneled catheters (2.7%). 53.7% of CVCs had a single lumen and 46.3% had a double lumen. The most common catheter material was polyurethane (40.5%), followed by silicone (9.3%), and endexo polyurethane (9.2%). The most common sites of CVC placement were upper extremity (29.0%), femoral (21.5%), umbilical (18.1%), saphenous (18.1%), and right internal jugular (7.6%) veins. The majority of CVCs were placed at the NICU bedside and without ultrasound guidance. Mean dwell time was 20.9 days (+/- 25.7 days). Preliminary analyses indicate that lower weight at time of placement and longer dwell time (27.4 vs. 19.3 days, p < 0.001) was associated with thrombosis. Analyses of how catheter characteristics (material, size, lumens) and placement factors (ultrasound guidance, vessel size, attempts) are associated with thrombosis are ongoing.
Conclusion(s): Based on single center data, catheter-related thrombosis is relatively common in NICU patients and may be related to catheter and patient factors. Using natural language processing we created a database that provides granularity in analyzing these factors which is not possible using discharge databases.
