A Retrospective Comparison of Transcutaneous and End-Tidal Carbon Dioxide Monitoring in Pediatric Polysomnography.

Background Accurate monitoring of carbon dioxide (CO₂) is important in pediatric polysomnography (PSG), particularly for identifying nocturnal hypoventilation, a condition with significant clinical consequences in children, especially those with neuromuscular or pulmonary disorders. Two primary modalities are available for non-invasive CO₂ estimation: end-tidal CO₂ (EtCO₂) and transcutaneous CO₂ (TcCO₂). While both are widely used, their diagnostic equivalence remains unclear. This study aimed to address the clinical question of whether TcCO₂ and EtCO₂ measurements in pediatric PSG can replace each other.  Methods A single-center retrospective study was conducted by including pediatric patients who underwent PSG with concurrent EtCO₂ and TcCO₂ monitoring. For each patient, mean, maximum, and minimum CO₂ values, as well as sleep duration with values exceeding 50 mmHg, were extracted and analyzed between the two modalities. The diagnostic rates of nocturnal hypoventilation as defined by the American Academy of Sleep Medicine (AASM) Manual were also compared. Results Among 127 children (median age 11 years, interquartile range age eight to 14), based on the Bland-Altman analysis, a positive mean bias was observed in all three comparisons, suggesting that EtCO₂ measurements were systematically lower than TcCO₂. Wide 95% limits of agreement in all TcCO₂ and EtCO₂ measurements were observed. A weakly positive correlation but with marginal statistical significance was noted between mean TcCO₂ and EtCO₂ values (rs=0.173, p=0.052). Maximum TcCO₂ and EtCO₂ values were also only weakly correlated (rs=0.288, p=0.001). The median of mean TcCO₂ was significantly higher than that of mean EtCO₂: 42.95 mmHg vs. 39.00 mmHg, respectively (Z = -8.06, p < 0.001). The median of maximum TcCO₂ was also significantly higher than that of maximum EtCO₂ values: 47.40 mmHg vs. 44.00 mmHg, respectively (Z = -7.50, p < 0.001). However, no statistically significant difference was observed for the corresponding minimum values (36.60 mmHg vs. 35.00 mmHg; Z = -1.94, p = 0.052). Ten out of 127 (7.9%) patients were diagnosed with hypoventilation by TcCO₂ measurement, none of whom could be diagnosed based on EtCO₂ measurement. Subgroup analysis by age, sex, obstructive sleep apnea (OSA) severity, obesity, craniofacial abnormalities, tonsillar hypertrophy, and syndromic diagnosis yielded no significantly different findings. Conclusions Our study suggested that EtCO₂ and TcCO₂ measurements in pediatric PSG may not be interchangeable, given the observed suboptimal agreement between the two modalities, i.e., the tendency for TcCO₂ values to be higher than EtCO₂, and the weak and mostly statistically insignificant correlation observed for CO₂ measurements. However, in the absence of arterial blood gas (ABG) as a reference standard, it was not possible to determine which modality more accurately reflects true arterial CO₂ levels or was more reliable for identifying nocturnal hypoventilation.