Objectives: To assess the diagnostic accuracy of commonly used signs and symptoms of low-intake dehydration in older care home residents.
Design: Prospective diagnostic accuracy study.
Setting: 56 care homes offering residential, nursing, and/or dementia care to older adults in Norfolk and Suffolk, United Kingdom.
Participants: 188 consecutively recruited care home residents aged ≥65 years, without cardiac or renal failure and not receiving palliative care. Overall, 66% were female, the mean age was 85.7 years (standard deviation 7.8), and the median Mini-Mental State Examination MMSE score was 23 (interquartile range 18-26).
Index tests: Over 2 hours, participants underwent double-blind assessment of 49 signs and symptoms of dehydration and measurement of serum osmolality from a venous blood sample. Signs and symptoms included skin turgor; mouth, skin, and axillary dryness; capillary refill; sunken eyes; blood pressure on resting and after standing; body temperature; pulse rate; and self-reported feelings of thirst and well-being.
Reference standard: Serum osmolality, with current dehydration defined as >300 mOsm/kg, and impending dehydration ≥295 mOsm/kg.
Outcome measures: For dichotomous tests, we aimed for sensitivity and specificity >70%, and for continuous tests, an area under the curve in receiver operating characteristic plots of >0.7.
Results: Although 20% of residents had current low-intake dehydration and a further 28% impending dehydration, none of the commonly used clinical signs and symptoms usefully discriminated between participants with or without low-intake dehydration at either cut-off.
Conclusions/implications: This study consolidates evidence that commonly used signs and symptoms of dehydration lack even basic levels of diagnostic accuracy in older adults, implying that many who are dehydrated are not being identified, thus compromising their health and well-being. We suggest that these tests be withdrawn from practice and replaced with a 2-stage screening process that includes serum osmolarity, calculated from sodium, potassium, urea, and glucose levels (assessed routinely using the Khajuria and Krahn equation), followed by serum osmolality measurement for those identified as high risk (calculated serum osmolarity >295 mmol/L).
Keywords: Aged; dehydration; diagnosis; drinking; osmolar concentration; sensitivity and specificity.
Crown Copyright © 2019. Published by Elsevier Inc. All rights reserved.