RESUMO
BACKGROUND: The authors investigated the accuracy of transcutaneous capnometry (TcPCO(2)) in estimating arterial blood carbon dioxide partial pressure (PaCO(2)) during spontaneous breathing in patients admitted to our surgical intensive care unit (ICU). METHODS: Serial TcPCO(2) and PaCO(2) measurements were taken in stable patients undergoing postoperative monitoring after major abdominal, vascular, or thoracic surgery. Patients were enrolled 12 hours after extubation. Exclusion criteria were pulmonary dysfunction, hemodynamic instability, or anemia. Linear regression, mixed models, and Bland-Altman analyses were used to compare accuracy and correlation between the two variables. Data are presented as means (95% confidence intervals). RESULTS: PaCO(2) values ranged between 26 mmHg and 52 mmHg. Mean values for TcPCO(2) and PaCO(2) were 35.3 (33.8-36.8) mmHg and 39.2 (37.6-40.7) (P<0.001). Bland-Altman analysis showed a bias of -3.8 (-5.1-2.5) mmHg with upper and lower limits of agreement of 5.6 (3.4-7.9) mmHg and -13.3 (-15.6 -11.1) mmHg, respectively. Thirty-one (55.4%) measurements disagreed by 3 mmHg or more. Linear mixed model analysis with adjustment for repeated measurements showed low correlation (r=0.63; P<0.0005). There were small but significant differences between patients in the correlation coefficient (P=0.04). CONCLUSION: TcPCO(2) showed moderate bias with wide limits of agreement when compared to PaCO(2) in our patients. The regression model showed low correlation of the measurements when adjusting for repeated measurements and between-subject variances. TcPCO(2) may not provide an accurate quantitative estimation of PaCO(2) in stable, spontaneously breathing, nonintubated patients in the early postoperative period.
