Dealing with dehydration in hospitalized oldest persons: accuracy of the calculated serum osmolarity

Background Assessment of hydration status is complex and difficult to detect in older persons. Different methods have been developed to determine hydration status in clinical settings, but their diagnostic accuracy remains questionable. Aims The aim of this study was to determine and compare the diagnostic accuracy of all methods routinely used in acute settings to detect dehydration in a cohort of hospitalized oldest-old persons, using as primary reference standard blood urea nitrogen (BUN) to creatinine ratio. Methods This retrospective study was conducted on 59 oldest-old subjects at hospital admission in an acute setting, with complete physical, biochemical, bioelectrical impedance analysis (BIA) and ultrasound assessment, including inferior vena cava diameters. Results Fifty-nine (45 women/14 men) subjects, with a mean age of 87.4 +/- 5.9 years, were studied. Based on the value of the BUN/creatinine ratio, the whole population was divided into hyperhydrated (n = 10), normohydrated (n = 42), and dehydrated (n = 7) groups. Among parameters indicating the hydration status, serum sodium levels (p < 0.0001), serum chloride levels (p = 0.010), calculated plasma osmolarity (p < 0.0001), and fat mass (FM) (p = 0.030) differed significantly among groups. A ROC analysis showed that the highest and most significant value for dehydration detection was the calculated plasma osmolarity (AUC: 0.820, p = 0.013), which significantly correlated with clinical parameters including heart rate (r = 0.300; p = 0.021), capillary refill (r = 0.379; p = 0.013) and systolic blood pressure (r = - 0.261; p = 0.046). Discussion The measurement of calculated serum osmolarity is simple and inexpensive and may quickly provide high sensitivity and specificity indication of dehydration in hospitalized oldest-old persons.