Derivation and Internal Validation of a Score to Predict Dehydration Severity in Patients over 5 Years with Acute Diarrhea

Abstract

ABSTRACT. Diarrheal disease accounts for more than one million deaths annually in patients over 5 years of age. Although most patients can be managed with oral rehydration solution, patients with severe dehydration require resuscitation with intravenous fluids. Scoring systems to assess dehydration have been empirically derived and validated in children under 5 years, but none have been validated for patients over 5 years. In this study, a prospective cohort of 2,172 patients over 5 years presenting with acute diarrhea to International Centre for Diarrhoeal Disease Research, Dhaka Hospital, Bangladesh, were assessed for clinical signs of dehydration. The percent difference between presentation and posthydration stable weight determined severe (≥ 9%), some (3–9%), or no (< 3%) dehydration. An ordinal regression model was derived using clinical signs and demographics and was then converted to a 13-point score to predict none (score of 0–3), some (4–6), or severe (7–13) dehydration. The Novel, Innovative Research for Understanding Dehydration in Adults and Kids (NIRUDAK) Score developed by our team included age, sex, sunken eyes, radial pulse, respiration depth, skin turgor, and vomiting episodes in 24 hours. Accuracy of the NIRUDAK Score for predicting severe dehydration, as measured by the area under the receiver operating characteristic curve, was 0.76 (95% confidence interval = 0.73–0.78), with a sensitivity of 0.78 and a specificity of 0.61. Reliability was also robust, with an Inter-Class Correlation Coefficient of 0.88 (95% confidence interval = 0.84–0.91). This study represents the first empirically derived and internally validated scoring system for assessing dehydration in children ≥ 5 years and adults with acute diarrhea in a resource-limited setting.