Estimated urinary osmolality based on combined urinalysis parameters: a critical evaluation.

Background Urinary conductivity allows a coarse prediction of urinary osmolality in most cases but is insensitive to the osmolal contribution of uncharged particles and the presence of roentgen contrast media. Urinary osmolality can be estimated on the recently introduced Sysmex UF-5000 urine analyzer using conductivity. In this study, we evaluated the analytical performance of this research parameter. Secondly, we aimed to improve the manufacturer's algorithm for estimating urinary osmolality, based on standard urinalysis parameters (creatinine, glucose, relative density). Methods The analytical performance was determined and a prediction model to estimate urinary osmolality based on urinalysis parameters was developed. We further developed and validated a prediction model using another set of routine urine samples. In addition, the influence of roentgen contrast media on urinary osmolality was studied. Results The within-run and between imprecision for osmolality and conductivity measured on the Sysmex UF-5000 ranged from 1.1% to 4.9% and 0.7% to 4.8%, respectively. Multiple regression analysis revealed urinary creatinine, conductivity and relative density to be the strongest predictors to estimate urinary osmolality. A mean difference of 1.3 mOsm/kg between measured and predicted osmolality demonstrated that the predictive performance of our model was favorable. An excellent correlation between the relative density and % contrast media was demonstrated. Conclusions Urinary osmolality is an important parameter for assessing specimen dilution in urinalysis. Urinary conductivity, along with relative density and urinary creatinine allows a coarse prediction of urinary osmolality and is insensitive to the osmolal contribution of uncharged particles and the presence of roentgen contrast media.