Measurement of urinary concentration: a critical appraisal of methodologies

The measurement of urine concentration provides information concerning the kidney's ability to appropriately respond to variations in fluid homeostasi s. It also assists in the interpretation of other tests performed on the sa me urine specimen. The gold standard of estimating urinary concentration is the measurement of its osmolality; however, this procedure is not readily available to the practicing physician. Therefore, urine concentration is us ually determined by measurement of its specific gravity (SG), which provide s a fair estimate of urine osmolality. Over thee years numerous tests have been developed to measure urine SG in a simple, quick, reliable and easily available method. These tests measure SG either directly (e.g., gravimetry) or by indirect methods (e.g., refractometry and reagent strip). All these tests have certain limitations based on their underlying physical principle s. Specific gravity as measured by refractometry is influenced by proteinur ia, such that for each 10 g/l protein the SG increases by 0.003. SG is also influenced by glucosuria such that it increases by approximately 0.002 per 10 g/l glucose when compared with urinary osmolality. Unlike osmolality, w hich is only affected by the number of particles, refractometry is affected by number, mass and chemical structure of the dissolved particles; hence l arge molecules like radiographic contrast or mannitol will increase SG rela tive to osmolality. The reagent strip is minimally affected by glucose, man nitol or radiographic contrast. However, it is affected by urinary pH such that only urine in the pH range of SG by reagent strip is based on the ioni c strength of the urine and thus is significantly affected by the ionic com position of the urine and by proteins which have an electric charge in solu tion. In our experience, SG measured by the refractometer is consistently m ore accurate than the reagent strip. For the clinician who is interpreting urine SG results, it is important to be aware of these limitations and unde rstand the reasons for possible potential errors of each particular method.