The correlation and level of agreement between end-tidal and blood gas pCO2 in children with respiratory distress: a retrospective analysis

BACKGROUND To investigate the correlation and level of agreement between end-tidal carbon dioxide (EtCO2) and blood gas pCO2 in non-intubated children with moderate to severe respiratory distress. METHODS Retrospective study of patients admitted to an intermediate care unit (InCU) at a tertiary care center over a 20-month period with moderate to severe respiratory distress secondary to asthma, bronchiolitis, or pneumonia. Patients with venous pCO2 (vpCO2) and EtCO2 measurements within 10 minutes of each other were eligible for inclusion. Patients with cardiac disease, chronic pulmonary disease, poor tissue perfusion, or metabolic abnormalities were excluded. RESULTS Eighty EtCO2-vpCO2 paired values were available from 62 patients. The mean +/- SD for EtCO2 and vpCO2 was 35.7 +/- 10.1 mmHg and 39.4 +/- 10.9 mmHg respectively. EtCO2 and vpCO2 values were highly correlated (r = 0.90, p < 0.0001). The correlations for asthma, bronchiolitis and pneumonia were 0.74 (p < 0.0001), 0.83 (p = 0.0002) and 0.98 (p < 0.0001) respectively. The mean bias +/- SD between EtCO2 and vpCO2 was -3.68 +/- 4.70 mmHg. The 95% level of agreement ranged from -12.88 to +5.53 mmHg. EtCO2 was found to be more accurate when vpCO2 was 35 mmHg or lower. CONCLUSION EtCO2 is correlated highly with vpCO2 in non-intubated pediatric patients with moderate to severe respiratory distress across respiratory illnesses. Although the level of agreement between the two methods precludes the overall replacement of blood gas evaluation, EtCO2 monitoring remains a useful, continuous, non-invasive measure in the management of non-intubated children with moderate to severe respiratory distress.