RESUMO
INTRODUCTION: Capnography is used to monitor the endtidal carbon dioxide tension (EtCO2) in exhaled gas. Sidestream capnography has great potential to monitor mechanically ventilated pediatric patients, given the continuous sampling from the endotracheal tube into a gas sensor. However, hemodynamic and respiratory impairments may reduce reliability and validity of sidestream capnography to monitor arterial carbon dioxide tension (PaCO2) in critically ill, mechanically ventilated children. METHODS: In 47 mechanically ventilated pediatric patients (aged 0-14 years, median age 17.2 months), a total of 341 consecutive measurements of PaCO2, EtCO2, respiratory and hemodynamic parameters were performed, and capnogram shape was determined. Validity was assessed with the Bland-Altman limit of agreement (loa), mixed models were used to adjust for variation between patients, and potential confounders were considered with multivariable analyses. RESULTS: EtCO2 (mean 4.50 ± 0.96 kPa) underestimated PaCO2 (mean 5.37 ± 0.87) considerably, resulting in a loa of 0.87 kPa [95% confidence interval (95% CI) -1.03;2.77] and 42.2% percentage error. The association improved significantly b = 0.54 [95 %CI = 0.45;0.64, P < 0.001] when corrected for individual differences. The association between EtCO2 and PaCO2 was not influenced by any of the potential confounders. CONCLUSIONS: Sidestream capnography in mechanically ventilated infants and children seems moderately reliable and valid when corrected for individual differences. Therefore, it could only be used with caution for trend estimation in the individual patient.