PaCO2, transcutaneous PCO2 (PtcCO2), and end-tidal PCO2 (PetCO2) measurements were studied in 12 critically ill neonates. PtcCO2 was measured using a combination CO2/O2 sensor during the routine care of these patients. End-tidal sidestream sampling was performed during blood gas measurement as dictated by the patient's clinical condition. There was a linear correlation between PtcCO2 and PaCO2 (n = 51, r = .71, slope = 0.90). PetCO2 and PaCO2 did not correlate as well (n = 51, r = .52, slope = 0.42). Acidosis negatively affected the correlation between PtcCO2 and PaCO2. When pH was greater than 7.30, r = .75 and slope = 1.28 (n = 38), whereas when pH was less than 7.30, r = .62 and slope = 0.73 (n = 13). The presence or absence of a metabolic acidosis did not have a significant effect on the slopes obtained. PtcCO2 monitoring using combined sensors is a useful and practical means of monitoring in the neonatal ICU, although acidosis affects the ability to correlate transcutaneous and arterial values. End-tidal sidestream measurements are not as clinically useful because they vary due to different ventilation/perfusion relationships in the sick neonate.