AimThe harmful effect of hypocapnia on the neonatal brain emphasizes the importance of monitoring arterial carbon dioxide tension (PaCO2). Transcutaneous monitoring of carbon dioxide (tcPCO2) reduces the need for arterial blood sampling. Drawbacks are high electrode temperature causing risks of skin burning. The aim was to determine the accuracy and precision of tcPCO2 at reduced electrode temperature.MethodsForty newborns (GA 24.9-41.7) were included. Two tc-monitors were applied (TCM4, Radiometer, Copenhagen). Arterial blood gas sampling and monitoring of tcPCO2-level at different electrode temperatures was done simultaneously (39°C, 40°C, 41 °C, 42°C, 44°C). Difference of PaCO2-tcPCO2 was expressed as a percentage of the mean.ResultsMean PaCO2 was 5.8kPa [3,2; 7.9]. Bias (PaCO2 -tcPCO2) increased from 5% at 44°C to 17% at 39°C, but did not differ significantly between 41°C and 40°C. The precision of the tcPCO2 at each temperature ranged from +7-10%. After correction for the temperature-dependent over-reading, we found increasing PaCO2 — tcPCO2 difference with increasing PaCO2, approx. 2% pr. kPa increase of CO2. Only mild transient erythema was observed.ConclusionA lower electrode temperature in tcPCO2-monitoring increases systematic overreading of the tc-electrode. However, in very preterm babies, monitoring at 40°C or 41°C is possible provided a bias correction of 12-15% is applied.