Sleep

Validation of transcutaneous CO2 measurement in awake children undergoing sleep studies.

TL;DR

PtcCO2 measurements provide a reliable estimation of PcapCO2 in awake, hemodynamically stable children undergoing sleep studies, with a mean bias of -0.59 mmHg and limits of agreement from -5.56 to +4.38 mmHg.

Key Findings

Transcutaneous CO2 measurement showed strong positive correlation with arterialized capillary CO2 pressure in awake children undergoing sleep studies.

  • Pearson correlation coefficient r = 0.703; p < 0.0001
  • Study included 335 pediatric records analyzed retrospectively
  • Median age of participants was 12.2 years (IQR 8.8–15.3), with 60% being male
  • Measurements were taken prior to sleep on the day of polysomnography while breathing room air
  • Each patient underwent both earlobe arterialized capillary blood gas measurement and PtcCO2 recording

Bland-Altman analysis revealed a small mean bias between PtcCO2 and PcapCO2 with defined limits of agreement.

  • Mean bias (ΔCO2) was -0.59 mmHg
  • Limits of agreement ranged from -5.56 to +4.38 mmHg
  • The negative bias indicates PtcCO2 readings were on average slightly lower than PcapCO2
  • The relatively wide limits of agreement suggest individual-level differences may occur

No significant difference in ΔCO2 was found between the two transcutaneous CO2 devices used (Sentec ResMed and Radiometer TCM4).

  • p = 0.41 for device comparison
  • Two devices used were Sentec ResMed and Radiometer TCM4
  • Device type was not a significant predictor of measurement difference

Mean pH was the sole independent predictor of the difference between PtcCO2 and PcapCO2 in multivariate regression analysis.

  • Each 0.01-unit increase in pH was associated with an approximately 0.31 mmHg larger ΔCO2 (β = 0.31; 95% CI = 0.18, 0.43)
  • Higher pH (indicating more alkalosis/hyperventilation) was associated with larger discrepancy between PtcCO2 and PcapCO2
  • Analysis included univariate and multivariate regression models to determine predictors of ΔCO2
  • Authors attributed temporary differences to hyperventilation during blood sampling, which would lower PcapCO2 without immediately affecting PtcCO2

PtcCO2 measurements were considered to provide reliable estimation of PcapCO2 in hemodynamically stable children, though transient discrepancies can occur due to hyperventilation during blood sampling.

  • The agreement between PtcCO2 and arterial CO2 pressure in hemodynamically stable children had not previously been assessed
  • The study was retrospective in design
  • Temporary differences were attributed to hyperventilation during blood sampling, elevating pH and lowering PcapCO2 relative to the continuously recorded PtcCO2
  • Findings support the use of PtcCO2 for continuous alveolar ventilation monitoring in pediatric sleep studies

What This Means

This research suggests that a non-invasive method of monitoring carbon dioxide (CO2) levels through the skin — called transcutaneous CO2 measurement (PtcCO2) — gives reliable readings that closely match CO2 levels measured from a small blood sample taken from the earlobe in children who are awake before a sleep study. Researchers analyzed records from 335 children (median age about 12 years) and found that the two measurements agreed well on average, differing by only about half a millimeter of mercury (mmHg). The findings also showed that two different monitoring devices produced similar results, suggesting the method is consistent across equipment. However, the study found that some children showed larger discrepancies between the two measurements, and this was linked to higher blood pH — a sign of hyperventilation (breathing too fast or deeply). This likely happened because children sometimes breathe differently when having blood drawn, which temporarily lowers the CO2 in their blood sample but doesn't immediately change the skin sensor reading. This means that in some cases, the skin sensor may appear to overestimate CO2 compared to the blood test, simply because the blood test captured a moment of anxious or exaggerated breathing. This research matters because sleep studies in children commonly use transcutaneous CO2 sensors to continuously track breathing quality overnight, but it was previously unknown how well these sensors matched actual blood CO2 levels in awake children. These findings provide reassurance that the technology is reasonably accurate, while also highlighting that transient breathing changes — such as those caused by anxiety during blood draws — can create temporary mismatches between sensor and blood readings.

Have a question about this study?

Citation

Mariani A, Dardenne J, Beydon M, Taytard J, Beydon N. (2026). Validation of transcutaneous CO2 measurement in awake children undergoing sleep studies.. Sleep medicine. https://doi.org/10.1016/j.sleep.2026.108816