Validation of Continuous Transcutaneous Carbon Dioxide Monitoring in Neonates Undergoing Therapeutic Hypothermia for Neonatal Encephalopathy.

Abstract

To assess the accuracy of transcutaneous carbon dioxide (tcPCO₂) monitoring compared with intermittently sampled blood gas PCO₂ in neonates undergoing therapeutic hypothermia (TH). Neonates ≥34 weeks of gestation undergoing TH were enrolled and subsequently monitored with continuous tcPCO₂. Agreement between tcPCO₂ and both temperature-corrected and measured (uncorrected) blood gas PCO₂ from arterial, venous, and capillary sources was assessed using Bland-Altman (BA) analyses that accounted for paired repeated measures. 130 paired tcPCO₂-PCO₂ measurements were analyzed for 53 neonates. The mean difference (bias) between tcPCO₂ and temperature-corrected PCO₂ from all sources was 6.6 mm Hg with 95% limits of agreement (LOA) from -7.4 to +20.6 mm Hg. Relative to measured PCO₂, the geometric mean ratio indicated a 4% overestimation of tcPCO₂, with proportional 95% LOA from -25% to +35%. Capillary samples showed significant (P<0.05) non-constant variance in agreement with increasing variability at extreme PCO₂ values, whereas arterial and venous samples demonstrated stable bias and variance across PCO₂ ranges. Using transcutaneous CO₂ monitoring in infants with neonatal encephalopathy receiving TH, there is moderate agreement with wide 95% LOA between tcPCO₂ and PCO₂ from all sources. Despite potential bias, tcPCO₂ may assist in identifying trends and avoiding extreme PCO₂ values during TH. https://clinicaltrials.gov/study/NCT04603547.