Validity of a portable urine refractometer: the effects of sample freezing

The use of portable urine osmometers is widespread, but no studies have assessed the validity of this measurement technique. Furthermore, it is unclear what effect freezing has on osmolality. One-hundred participants of mean (±SD) age 25.1 ± 7.6 years, height 1.77 ± 0.1 m and weight 77.1 ± 10.8 kg provided single urine samples that were analysed using freeze point depression (FPD) and refractometry (RI). Samples were then frozen at -80°C (n = 81) and thawed prior to re-analysis. Differences between methods and freezing were determined using Wilcoxon's signed rank test. Relationships between measurements were assessed using intraclass correlation coefficients (ICC) and typical error of estimate (TE). Osmolality was lower (P = 0.001) using RI (634.2 ± 339.8 mOsm · kgH2O(-1)) compared with FPD (656.7 ± 334.1 mOsm · kgH2O(-1)) but the TE was trivial (0.17). Freezing significantly reduced mean osmolality using FPD (656.7 ± 341.1 to 606.5 ± 333.4 mOsm · kgH2O(-1); P < 0.001), but samples were still highly related following freezing (ICC, r = 0.979, P < 0.001, CI = 0.993-0.997; TE = 0.15; and r=0.995, P < 0.001, CI = 0.967-0.986; TE = 0.07 for RI and FPD respectively). Despite mean differences between methods and as a result of freezing, such differences are physiologically trivial. Therefore, the use of RI appears to be a valid measurement tool to determine urine osmolality.