The superiority of point of care ultrasound in localizing central venous line tip position over time

The primary objective was to study agreement between X-rays and point of care ultrasound (POC-US) in determining central venous line (CVL) tip position. The secondary objective was to examine malposition rates over time using POC-US. Fifty-six neonates were enrolled who had a CVL placed. Initial X-rays and POC-US were obtained. POC-US was performed daily thereafter for the total of 6 days. US video clips were acquired in four standard echocardiographic views: subcostal, four-chamber, and short- and long-axis parasternal views. Gwet’s agreement coefficient (AC1) for agreement measured inter-rater reliability of X-rays and POC-US (correct position/malposition). A generalized linear mixed model for binary clustered data estimated malposition rate over time. All analyses were conducted using SAS version 9.4 and Agree Stat. The study included 108 “pairs” of X-rays and POC-US images. Agreement coefficient (AC1), with respect to correct position/malposition of CVL tip, was high AC1 = 0.872 (UVC-AC1 = 0.814, PICC-AC1 = 0.94). Among birth weight (BW) < 1000 g, 1000–1499 g, and BW > 1500 g, AC1 values were 0.922, 0.774, and 0.873, respectively. CVL tip malposition rate decreased over time.

Conclusions: Agreement between POC-US and X-rays for CVL tip position was high, with the highest in BW < 1000 g. The data suggest that POC-US can be used for initial confirmation and follow up of CVL tip position.