Comparison of Two Techniques of Superior Vena Cava Flow Measurement in Preterm Infants With Birth Weight <1,250 g in the Transitional Period—Prospective Observational Cohort Study

Objectives Superior Vena Cava (SVC) flow in neonates measured by the standard approach has been validated by different groups around the world. The modified SVC flow measurement technique was recently suggested. The aim of our study was to evaluate standard and modified technique of echocardiography SVC flow measurement in a cohort of extremely preterm neonates in the immediate postnatal period. Methods Prospective, observational cohort study in a level III neonatal centre. Infants with birth weight less than 1250g were eligible for enrolment. SVC flow was measured by echocardiography using standard and modified methods at 6, 18 and 36 hours of age. Our primary outcome was equivalency (using raw bounds of –20 to +20 ml/kg/min difference between the paired measurements), agreement and correlation between standard and modified methods of the SVC flow measurements. Results Thirty-nine infants were enrolled. The mean gestational age of the cohort was 27.4 (SD 2.1) weeks of postmenstrual age, the mean birth weight was 0.95 kg (SD 0.2). The measurements at 6 and 36 hours of age were equivalent as defined in the design of the study (p=0.003 and p=0.004 respectively; raw bounds -20 to 20 ml/kg/min). At six hours of age the mean difference (bias) between the measurements was –0.8 ml/kg/min with 95% limits of agreement -65.0 to 63.4 ml/kg/min. At 18 hours of age, the mean difference (bias) between the measurements was +9.5 ml/kg/min, with 95% limits of agreement -79.6. to 98.7 ml/kg/min. At 36 hours of age the mean difference (bias) between the measurements was –2.2 ml/kg/min with 95% limits of agreement -73.4 to 69.1 ml/kg/min. There was a weak, but statistically significant correlation between the standard and modified method at 6 hours of age (r=0.39, p=0.04). Conclusion Both SVC flow echocardiography measurement techniques yielded clinically equivalent results, however due to wide limits of agreement and poor correlation they do not seem to be interchangeable.