Assessing Dehydration Employing End-Tidal Carbon Dioxide in Children With Vomiting and Diarrhea

OBJECTIVE: Serum bicarbonate reflects dehydration severity in children with gastroenteritis. Previous work in children receiving intravenous rehydration has correlated end-tidal carbon dioxide (EtCO2) with serum bicarbonate. We evaluated whether EtCO2 predicts weight change in children with vomiting and/or diarrhea. METHODS: A prospective cohort study was conducted. Eligible children were 3 months to 10 years old and presented for emergency department (ED) care because of vomiting and/or diarrhea. End-tidal carbon dioxide measurements were performed after triage. The diagnostic standard was weight change determined from serial measurements after symptom resolution. A receiver operating characteristic curve was constructed to identify a cut-point to predict 5% or more dehydration. RESULTS: In total, 195 children were enrolled. Among the 169 (87%) with EtCO2 measurements, the median (interquartile range [IQR]) was 30.4 (27.8 to 33.1). One hundred fifty-eight had repeat weights performed at home; the median (IQR) weight change from ED presentation to well weight was +0.06 (-0.14 to +0.30) or +0.72% (-1.2% to +2.1%). Sixteen percent (25/158) had 3% or more and 4% (6/158) had 5% or more weight gain (ie, percent dehydration). One hundred sixteen (60%) completed home follow-up and had acceptable EtCO2 recordings. Receiver operating curve analysis revealed an area under the curve of 0.34 (95% confidence interval, 0.06 to 0.62) for EtCO2 as a predictor of 5% or more dehydration. CONCLUSIONS: The limited accuracy of EtCO2 measurement to predict 5% or more dehydration precludes its use as a tool to assess dehydration severity in children. End-tidal carbon dioxide monitoring does not have the ability to identify those children with 5% or more dehydration in a cohort of children with vomiting and/or diarrhea presenting for ED care.