The relationship between minute ventilation and end tidal CO2 in intubated and spontaneously breathing patients undergoing procedural sedation

Background Monitoring respiratory status using end tidal CO2 (EtCO2), which reliably reflects arterial PaCO2 in intubated patients under general anesthesia, has often proven both inaccurate and inadequate when monitoring non-intubated and spontaneously breathing patients. This is particularly important in patients undergoing procedural sedation (e.g., endoscopy, colonoscopy). This can be undertaken in the operating theater, but is also often delivered outside the operating room by non-anesthesia providers. In this study we evaluated the ability for conventional EtCO2 monitoring to reflect changes in ventilation in non-intubated surgical patients undergoing monitored anesthesia care and compared and contrasted these findings to both intubated patients under general anesthesia and spontaneously breathing volunteers. Methods Minute Ventilation (MV), tidal volume (TV), and respiratory rate (RR) were continuously collected from an impedance-based Respiratory Volume Monitor (RVM) simultaneously with capnography data in 160 patients from three patient groups: non-intubated surgical patients managed using spinal anesthesia and Procedural Sedation (n = 58); intubated surgical patients under General Anesthesia (n = 54); and spontaneously breathing Awake Volunteers (n = 48). EtCO2 instrument sensitivity was calculated for each patient as the slope of a Deming regression between corresponding measurements of EtCO2 and MV and expressed as angle from the x-axis (θ). All data are presented as mean ± SD unless otherwise indicated. Results While, as expected, EtCO2 and MV measurements were negatively correlated in most patients, we found gross systematic differences across the three cohorts. In the General Anesthesia patients, small changes in MV resulted in large changes in EtCO2 (high sensitivity, θ = -83.6 ± 9.9°). In contrast, in the Awake Volunteers patients, large changes in MV resulted in insignificant changes in EtCO2 (low sensitivity, θ = -24.7 ± 19.7°, p < 0.0001 vs General Anesthesia). In the Procedural Sedation patients, EtCO2 sensitivity showed a bimodal distribution, with an approximately even split between patients showing high EtCO2 instrument sensitivity, similar to those under General Anesthesia, and patients with low EtCO2 instrument sensitivity, similar to the Awake Volunteers. Conclusions When monitoring non-intubated patients undergoing procedural sedation, EtCO2 often provides inadequate instrument sensitivity when detecting changes in ventilation. This suggests that augmenting standard patient care with EtCO2 monitoring is a less than optimal solution for detecting changes in respiratory status in non-intubated patients. Instead, adding direct monitoring of MV with an RVM may be preferable for continuous assessment of adequacy of ventilation in non-intubated patients.