Background: Acute diarrhea is frequent, costly because of the number of hospital admissions required, and sometimes serious, even fatal to children in France. The clinical diagnosis of dehydration is difficult, but essential to determine management.
Objective: To summarize the published data on the value of clinical history, clinical signs and laboratory results for diagnosing dehydration during acute diarrhea in young (1 month-5 years) non-malnourished children.
Methods: Four databases (Medline, INIST, Ovid, and Cochrane) were searched through November 2006, with the key words "dehydration" subcategories "diagnosis, or etiology, or history", "diarrhea" subcategory "diagnosis", and age limits "infant or preschool child". We selected the articles and reviews that included as an endpoint for dehydration "weight gain > 5% after recovery" (the gold standard).
Results: Thirteen studies were selected. No single clinical history item, clinical sign or laboratory value was sufficient to discriminate between children with and without dehydration. The reproducibility of clinical signs varied substantially between studies. Persistent skin folds and signs of vasoconstriction contributed the most information, with good specificity but sensitivity < 50%. The combination of at least 3 clinical signs was most discriminative for dehydration. No dehydration scale has been validated. None of the studies selected had a very high level of proof (level 1 and 2); neither signs nor scores have been validated internally or externally because of the low number of subjects.
Conclusion: The diagnosis of dehydration due to acute diarrhea in young children depends on the number of signs present, since no individual element of clinical history, clinical picture or laboratory tests distinguished dehydration. Other studies are necessary.