Correlation and prediction of arterial partial pressure of carbon dioxide from venous umbilical blood gases

Background. Arterial partial pressure of carbon dioxide (pCO(2)) samples are lower in children and higher in fetuses when compared with venous samples. The correlation and prediction of pCO(2) from umbilical venous (UVBG) to umbilical arterial blood gas (UABG) dyad in neonates are identified. Methods. A prospective study was performed from July 2018 to December 2019. Two dependent tests and a multivariate regression model were used to analyze the comparison and correlation tests. Results. A total of 116 paired UABG and UVBG samples were obtained. The medians (interquartile ranges, IQR) were as follows: gestational age of 34 (29-37) weeks, birth weight of 2122 (1146-2839) g, and postnatal age of 2.3 (1.4-10.8) h. The median (IQR) pCO(2(UABG)) and pCO(2(UVBG)) measurements were 40.2 (33.5-45.8) and 40.4 (34.7-46.8) mmHg, respectively (rho = 0.75, p < 0.001). The median of the differences (IQR) in pCO(2(UABG)) and pCO(2(UVBG)) was -0.9 (-4.7 to 2.3) mmHg, (p = 0.06). The equation to predict pCO(2(UABG) )was 0.9 x pCO(2(UVBG)) + 4, as derived from simple linear regression. The best model for predicting pCO(2(UABG)) was 0.9 x pCO(2(UVBG)) - 0.7 x venous base excess + 0.6 x 5-min Apgar score + 6.1 x meconium aspiration syndrome - 7.7 x patent ductus arteriosus - 6.5 (adjusted r(2) = 0.74). Conclusions. pCO(2(UVBG)) correlates with and can predict pCO(2(UABG)). Therefore, pCO(2(UVBG)) can be applied to pCO(2(UABG)) in neonates for whom UAC insertion is unsuccessful or to avoid an arterial puncture.