Patient characteristics, vital signs and tcpCO2 dynamics significantly influence tcpCO2 accuracy in extremely preterm infants: an observational study

Purpose The estimation of the partial pressure of carbon dioxide (pCO 2 ) is of interest in extremely premature infants. Neonatal intensive care units use transcutaneous pCO 2 monitoring (tcpCO 2 ) as a trend parameter and continuous estimate of pCO 2 . Our objective was to identify candidate influence factors on the accuracy of tcpCO 2 in extremely premature infants in their first week of life. Methods In a retrospective cohort study, linear mixed-effects models were used to analyze the influence of patient characteristics, vital signs and tcpCO 2 dynamics on the difference of tcpCO 2 and capillary pCO 2 as reference (ΔpCO 2 ). Candidate parameters identified in univariate analyses were incorporated into multivariable models to confirm their effect on ΔpCO 2 . Results 500 measurements from 29 extremely preterm infants were analyzed. Postnatal age, mean fraction of inspired oxygen, tcpCO 2 , capillary pCO 2 , last tcpCO 2 and last ΔpCO 2 were positively associated with ΔpCO 2 , while gestational age and hemoglobin were negatively associated with ΔpCO 2 . Conclusion These candidate parameters may be used to enhance the non-invasive estimation of pCO 2 from tcpCO 2 in the future. Trial registration German Clinical Trials Register (DRKS00033228, first registered 29 August 2024, https://drks.de/search/en/trial/DRKS00033228/details )