Transcutaneous versus blood carbon dioxide monitoring during acute noninvasive ventilation in the emergency department - a retrospective analysis.

QUESTIONS UNDER STUDY Transcutaneous measurement of carbon dioxide (PtCO2) has been suggested as an alternative to invasively obtained PaCO2 for the monitoring of patients with hypercapnic respiratory failure during noninvasive ventilation (NIV). Current data on monitoring in hypoxaemic respiratory failure are scarce and show conflicting results in hypercapnic patients in the emergency department. METHODS AND SETTING We performed a retrospective comparison of real-time PtCO2 (SenTec Digital Monitor) and arterial/venous carbon dioxide tension (PaCO2/PvCO2) measurements in patients with severe hypoxaemic and/or hypercapnic respiratory failure during NIV. Agreement between PtCO2 and PaCO2/PvCO2 was the primary endpoint. Bland-Altman analysis and linear regression were used. RESULTS 102 patients had at least one matched measurement of PtCO2 and PaCO2/PvCO2. For patients with arterial blood gas analysis, the mean difference was 0.46 kPa at baseline (95% confidence interval [CI] 0.23 to 0.60, limits of agreement 95% CI -0.54 to 1.45) and 0.12 kPa after NIV (95% CI -0.04 to 0.29, limits of agreement 95% CI: -0.61 to 0.86). The linear regression analysis found a correlation R2 of 0.88 (p <0.001) at baseline and an R2 of 0.99 (p <0.001) after initiating NIV. For patients with venous blood gas analysis, the mean difference was 0.64 kPa at baseline (95% CI 0.04 to 1.24, limits of agreement 95% CI -0.72 to 2) and 0.80 kPa after NIV (95% CI 0.51 to 1.10, limits of agreement 95% CI 0.29 to 1.32), R2 0.78 (p <0.001) at baseline and R2 0.91 (p <0.001) after initiating NIV. A PaCO2/PvCO2 >8 kPa was associated with a lesser degree of agreement between the levels of PtCO2 and PaCO2/PvCO2 (p <0.001). CONCLUSION Transcutaneous PCO2 monitoring shows a good concordance with PaCO2 and is a reliable, feasible, patient-friendly and safe alternative to repeated blood gas analysis for patients with severe hypoxaemic and/or hypercapnic respiratory failure receiving emergency NIV in the emergency department. An initial blood gas analysis to evaluate the respiratory and metabolic state and to rule out a significant discrepancy compared with the transcutaneous measurement is recommended.