Evaluation of laboratory tests in dehydrated children with acute gastroenteritis

The aim of the present study was to demonstrate the utility of laboratory tests in predicting dehydration severity.

We evaluated retrospectively 168 dehydrated children aged 1-21 months with acute gastroenteritis. The degree of dehydration in each patient was calculated as the percentage change in weight from admission to after rehydration. The associations between degree of dehydration, urea, bicarbonate and serum sodium were examined.

Dehydration severity was significantly related to concentrations of urea (P < 0.00001) and bicarbonate (P = 0.01), but not to serum sodium or age. Children were divided into three groups on the basis of blood urea: (i) normal (< 6.7 mmol/L or < 40 mg/dL); (ii) intermediate (6.8-16.6 mmol/L); and (iii) high (>or=16.7 mmol/L or >or=100 mg/dL). Groups 1 and 3 had statistically significant associations with the degree of dehydration and combining the urea groups with plasma bicarbonate levels enhanced the accuracy of predictions. Group 1 had only mild to moderate dehydration, but if bicarbonate was 15 mmol/L or higher, there was a positive predictive value (PPV) of 93% for mild dehydration. Group 3 had moderate to severe dehydration, but if bicarbonate was 15 mmol/L or higher, moderate dehydration could be predicted (PPV 100%). Sixty per cent of mildly dehydrated children (32/53) had low bicarbonate, but only 7% (2/29) with bicarbonate 15 mmol/L or higher had severe dehydration. Serum sodium did not have any significant association with the degree of dehydration, or the levels of bicarbonate or urea.

The results suggest that serum urea and bicarbonate concentrations can be helpful in the estimation of fluid deficit independently from serum sodium concentration, and may be considered to be adjuncts to clinical evaluation in assessing the degree of dehydration.