Evaluating a novel formula for noninvasive estimation of arterial carbon dioxide during post-resuscitation care.

Abstract

Controlling arterial carbon dioxide is paramount in mechanically ventilated patients, and an accurate and continuous noninvasive monitoring method would optimize management in dynamic situations. In this study, we validated and further refined formulas for estimating partial pressure of carbon dioxide with respiratory gas and pulse oximetry data in mechanically ventilated cardiac arrest patients. A total of 4741 data sets were collected retrospectively from 233 resuscitated patients undergoing therapeutic hypothermia. The original formula used to analyze the data is PaCO₂ -est1 = PETCO₂  + k[(PIO₂  - PETCO₂ ) - PaO₂ ]. To achieve better accuracy, we further modified the formula to PaCO₂ -est2 = k₁ *PETCO₂  + k₂ *(PIO₂  - PETCO₂ ) + k₃ *(100-SpO₂ ). The coefficients were determined by identifying the minimal difference between the measured and calculated arterial carbon dioxide values in a development set. The accuracy of these two methods was compared with the estimation of the partial pressure of carbon dioxide using end-tidal carbon dioxide. With PaCO₂ -est1, the mean difference between the partial pressure of carbon dioxide, and the estimated carbon dioxide was 0.08 kPa (SE ±0.003); with PaCO₂ -est2 the difference was 0.036 kPa (SE ±0.009). The mean difference between the partial pressure of carbon dioxide and end-tidal carbon dioxide was 0.72 kPa (SE ±0.01). In a mixed linear model, there was a significant difference between the estimation using end-tidal carbon dioxide and PaCO₂ -est1 (P < .001) and PaCO₂ -est2 (P < .001) respectively. This novel formula appears to provide an accurate, continuous, and noninvasive estimation of arterial carbon dioxide.