We evaluated accuracy of end-tidal carbon dioxide tension (PETco2) monitoring and measured the effect of temperature correction of blood gas values in children after cardiac surgery. Data from 49 consecutive mechanically ventilated children after cardiac surgery in the cardiac intensive care unit were prospectively collected. One patient was excluded from the study. Four arterial-end-tidal CO2 pairs in each patient were obtained. Both the arterial carbon dioxide tension (Paco2) values determined at a temperature of 37 degrees C and values corrected to body temperature (Patcco2) were compared with the PETco2 values. After the surgical correction 28 patients had biventricular, acyanotic (mean age 2.7 +/- 4.8 years) and 20 patients had a cyanotic lesion (mean age 1.0 +/- 1.7 years). The body temperature ranged from 35.2 degrees C to 38.9 degrees C. The Pa-PETco2 discrepancy was affected both by the type of cardiac lesion and by the temperature correction of Paco2 values. Correlation slopes of the Pa-PETco2 and Patc-PETco2 discrepancies were significantly different (p = 0.040) when the body temperature was higher or lower than 37 degrees C. In children, after cardiac surgery, end-tidal CO2 monitoring provided a clinically acceptable estimate of arterial CO2 value, which remained stabile in repeated measurements. End-tidal CO2 monitoring more accurately reflects temperature-corrected blood gas values.