(1169) Performance of Noninvasive CO2 Monitoring in Critically Ill Adults

Introduction: Blood gas sampling is considered the gold standard for CO2 analysis in adult patients despite limitations of sampling risk, specimen processing delay, and a resultant single discrete measurement. Continuous transcutaneous CO2 (tcCO2) monitoring has been available for decades with devices marketed to neonatal and pediatric populations to limit blood draws but remains largely unused in adult ICUs. In addition to knowledge or financially based limitations, critically ill adult patients are commonly obese, hypothermic, or on vasopressors with potential impact on device accuracy. Correlations between tcCO2, etCO2, and paCO2 have been demonstrated in selected care settings but as technology continues to advance there is potential for tcCO2 monitoring in additional populations. This study assesses the accuracy and limitations of continuous transcutaneous carbon dioxide monitoring in real-world conditions among adult, critically ill, mechanically ventilated patients.

Methods: A retrospective chart review was conducted following a departmental feasibility assessment of the Sentec Digital Monitor System (SDMS). SDMS data was compared to correlating ABG results using paired t-test and sign tests. Linear regression modeling was performed for variables of concern (age, sex, weight, temperature, pressor use) using STATA12.

Results: 25 observations among 20 subjects were analyzed. tcCO2 monitoring demonstrated no difference with ABG sampling by paired t-testing (p=0.5) or sign rank testing (p=0.5). SLR revealed no relationship between tcCO2 and ABG results for age, sex, or use of pressors. Weight and temperature were significantly associated with pCO2 correlation between modalities (p=0.005 & p=0.046 respectively).

Conclusions: tcCO2 performed accurately as compared to invasive blood gas sampling in real-world conditions among critically ill patients requiring invasive mechanical ventilation. This correlation was maintained despite several confounding factors. Given the potential benefits of non-invasive monitoring systems (continuous measurement, absence of vascular injury risk, improved staff/resource demand), consideration for broader adoption is appropriate.