How valid are clinical signs of dehydration in infants?

Our objective was to determine the ability of several clinical signs of dehydration to distinguish among degrees of dehydration in infants with acute diarrhea. The design was a prospective cohort study in a pediatric referral hospital in Cairo, Egypt. Infant boys, 3-18 months old, with a history of acute diarrhea (5 or more watery stools per day for no more than 7 days) were eligible, except those with frank protein-energy malnutrition, serious nongastrointestinal illness, or being exclusively breast-fed. Several clinical signs of dehydration were assessed upon study entry. Subjects were then rehydrated with an oral rehydration solution and fed a standardized diet until diarrhea ceased (no watery or loose stools for 16 h). The main outcome measure was percent body weight gain at rehydration and at resolution of illness. Data from 135 subjects were available for analysis. Average (SD) rehydration phase duration was 5.2 (2.1) h, and average (SD) duration of illness was 54.5 (38) h. Multiple regression analysis selected prolonged skinfold, altered neurologic status, sunken eyes, and dry oral mucosa as the clinical signs that correlated best with percent dehydration (R2 for model 0.244, p < 0.001). Mean weight gain for the two assessment systems was 3.6-3.9% for mild, 4.9-5.3% for moderate, and 9.5-9.8% for severe dehydration. The most valid clinical signs of dehydration include prolonged skinfold, altered neurologic status, sunken eyes, and dry oral mucosa. Children with clinical signs of mild or moderate dehydration have fluid deficits on the order of 3 or 5% body weight, respectively.