OBJECTIVES: Blood gas monitoring is necessary in treatment of critically ill neonates. Whereas SaO2 can be estimated by pulse oximetry, PaCO2 is still most often assessed from blood samples. AIM: To compare long time performance of an ear sensor for combined assessment of transcutaneous carbon dioxide (PtcCO2) and oxygen saturation (SpO2) (TOSCA Monitor; Radiometer, Switzerland) with a conventional PtcCO2 monitor (MicroGas 7650-500 rapid, Radiometer, Switzerland) in critically ill neonates. METHODS: Prospective, observational study. Twenty critically ill neonates were monitored for PtcCO2 and SpO2 using the Tosca and the MicroGas monitor for 24 h. TOSCA ear sensor was changed to the other ear lobe after 12 h and the MicroGas sensor four hourly on the trunk. Values obtained were compared with SaO2 and PaCO2 from arterial blood gas analysis using Bland-Altman analysis. Data are presented as median (range). RESULTS: Eighty-two paired measurements were obtained. Median age of the 20 patients was 4.5 days (1-26 days) and weight was 3.05 kg (0.98-3.95 kg). Bias and precision between PaCO2 and PtcCO2 were 0.14 and 1.45 kPa for the Tosca monitor and -0.08 and 1.2 kPa for the MicroGas monitor, respectively. The two biases were significantly different (P = 0.0036). SpO2 assessment by TOSCA was comparable to SaO2 values (bias 0.26% and precision 4.14%). CONCLUSION: The TOSCA monitor allows safe estimation of PtcCO2 and SaO2 in neonates. Measurements of PtcCO2 were less reliable with TOSCA compared with conventional monitoring but still allow assessing a trend of ventilation status in newborn patients.