RESUMO
BACKGROUND: Urine osmolarity (OsmU) is the gold standard for the evaluation of the kidney's urine concentration capacity; nevertheless, urinary density (UD) is often used as a surrogate for its estimation. OBJECTIVE: The objective of this study was to analyze the accuracy of UD in estimating OsmU. MATERIALS AND METHODS: A transversal study including patients with simultaneous determination of UD measured with refractometry and OsmU measured by osmometer (OsmUm). We multiplied the last two digits of the UD by 35, 30, 32, 33.5, and 40 to estimate OsmU; the estimates were considered precise if the value was ± 30 mOsm/kg from the OsmUm. A Bland-Altman analysis was conducted. RESULTS: Among 205 patients, there was no difference between OsmUm and the estimated form when using a factor of 33.5 (p = 0.578). When analyzing by the absence or presence of proteinuria and/or glycosuria, there were no differences when using the factors 35 (p = 0.844) and 32 with adjusted UD (p = 0.898). In the linear correlation analysis, values for Pearson's r = 0.788 and r2 = 0.621 were obtained (p < 0.001). The areas under the curve obtained by the receiver operating characteristics curves to estimate urine osmolarity values < 100 and > 600 mOsm/kg were > 0.90. CONCLUSION: The estimation of the OsmU from UD showed adequate performance. If an osmometer is unavailable, we recommend using the factor 35 for clean samples and 32 with adjusted UD for samples with proteinuria and/or glycosuria.