End-Tidal Carbon Dioxide Measurement during Pediatric Polysomnography: Signal Quality, Association with Apnea Severity, and Prediction of Neurobehavioral Outcomes

Abstract

Study objectives: To identify the role of end-tidal carbon dioxide (EtCO2) monitoring during polysomnography in evaluation of children with obstructive sleep apnea syndrome (OSAS), including the correlation of EtCO2 with other measures of OSAS and prediction of changes in cognition and behavior after adenotonsillectomy.

Design: Analysis of screening and endpoint data from the Childhood Adenotonsillectomy Trial, a randomized, controlled, multicenter study comparing early adenotonsillectomy (eAT) to watchful waiting/supportive care (WWSC) in children with OSAS.

Setting: Multisite clinical referral settings.

Participants: Children, ages 5.0 to 9.9 y with suspected sleep apnea.

Interventions: eAT or WWSC.

Measurements and results: Quality EtCO2 waveforms were present for ≥ 75% of total sleep time (TST) in 876 of 960 (91.3%) screening polysomnograms. Among the 322 children who were randomized, 55 (17%) met pediatric criteria for hypoventilation. The mean TST with EtCO2 > 50 mmHg was modestly correlated with apnea-hypopnea index (AHI) (r = 0.33; P < 0.0001) and with oxygen saturation ≤ 92% (r = 0.26; P < 0.0001). After adjusting for AHI, obesity, and other factors, EtCO2 > 50 mmHg was higher in African American children than others. The TST with EtCO2 > 50 mmHg decreased significantly more after eAT than WWSC. In adjusted analyses, baseline TST with EtCO2 > 50 mmHg did not predict postoperative changes in cognitive and behavioral measurements.

Conclusions: Among children with suspected obstructive sleep apnea syndrome, overnight end-tidal carbon dioxide (EtCO2) levels are weakly to modestly correlated with other polysomnographic indices and therefore provide independent information on hypoventilation. EtCO2 levels improve with adenotonsillectomy but are not as responsive as AHI and do not provide independent prediction of cognitive or behavioral response to surgery.

Clinical trial registration: Childhood Adenotonsillectomy Study for Children with OSAS (CHAT). ClinicalTrials.gov Identifier #NCT00560859.

Keywords: CO2; capnography; end-tidal; hypercapnia; hypoventilation; pediatric; polysomnogram; sleep apnea.

Figure 1

Recruitment through study completion. Data were analyzed from: screening PSGs for descriptive analyses, and randomized baseline PSGs and endpoint PSGs for longitudinal analyses. PSG, polysomnogram; EtCO₂, end-tidal carbon dioxide; AHI, apneahypopnea index; eAT, early adenotonsillectomy group; WWSC, watchful waiting with supportive care group.

Figure 2

Hypoventilation and AHI categories. Among screening PSGs, an increase in hypoventilation was observed with an increase in AHI severity. AHI, apnea hypopnea index; PSGs, polysomnograms.

Figure 3

End-tidal carbon dioxide (EtCO₂) variables by AHI severity levels. The peak EtCO₂ value was greater than 50 mmHg in 761 (86.9%) of screening polysomnograms. Test of linear trend of AHI severity P < 0.0001 for all categories. AHI, apnea-hypopnea index; TST, total sleep time.

Figure 4

Comparison of AHI and percentage TST end-tidal carbon dioxide > 50 mmHg before and after adenotonsillectomy. A total of 267 children had baseline and endpoint polysomnogram data available (n = 136 eAT group; n = 131 WWSC group). AHI, apnea hypopnea index; eAT, early adenotonsillectomy; TST, total sleep time; WWSC, watchful waiting with supportive care.

Figure 5

Change in hypercapnia in children with hypoventilation. Change in %total sleep time end-tidal carbon dioxide > 50 mmHg in randomized subjects who had hypoventilation on the baseline polysomnogram with follow-up data on endpoint polysomnogram (n = 25 eAT group; n = 15 WWSC group). eAT, early adenotonsillectomy; WWSC, watchful waiting with supportive care.