Objective: Urine output, specific gravity, and ketones (urinary indices) are commonly used as an objective means to assess for dehydration and gastroenteritis severity; however, their utility has not been established. The study was designed to evaluate the accuracy of urinary indices as diagnostic tests to identify acute dehydration.
Methods: We completed a prospective cohort study in the Emergency Department of an urban pediatric hospital. Seventy-nine subjects ages 3 months to 36 months with gastroenteritis, clinically suspected moderate dehydration, and the need for intravenous rehydration were enrolled in the trial. Urine specific gravity and urine ketone levels were determined with bedside calorimetric (dipstick) testing, and urine output during rehydration and observation was measured by commonly used techniques. An internally validated, weight-based criterion standard for the percent dehydration on enrollment was used to identify the cohort of dehydrated subjects. Correlation statistics were calculated for urine output, specific gravity, and ketones. In addition, multilevel tables were created to determine the sensitivity, specificity, and likelihood ratio at varying test cutoff values to detect 3% and 5% dehydration.
Results: Urine specific gravity (r = -0.06, P = 0.64), urine ketones (r = 0.08, P = 0.52), and urine output during rehydration (r = 0.01, P = 0.96) did not correlate with the initial degree of dehydration present. Clinically useful cutoff values for urine specific gravity and ketones to increase or decrease the likelihood of dehydration at the time of enrollment could not be identified.
Conclusions: Urinary indices are not useful diagnostic tests to identify the presence of dehydration during the initial assessment of children with gastroenteritis.