To analyze the predictive value of a single 24-hour urine sample to accurately estimate real salt intake, we compared true salt intake with measured 24-hour sodium excretion in the urine. Because current computerized models often calculate the projected effect of a 3-gram reduction in salt intake on cardiovascular outcome, we tested the accuracy of UNaV to correctly estimate real salt intake within a 3-gram (50 mmol) range. Accuracy of each individual UNaV for correct assessment of daily salt intake was performed by definition of true positives of salt intake. We investigated the difference between recorded sodium intake and UNaV with Bland-Altman plots (Online Supplement S2). We considered a ±25 mmol (corresponding to ±1.5 gram salt) deviation of the mean difference between the recorded sodium intake and renal sodium excretion as true positive urine sample (correct prediction of salt intake). UNaV samples, which were outside this range, were considered as true negative (misclassification of salt intake). To test the effect of salt intake on UNaV measures, we conducted multilevel modeling using linear mixed models. We tested a random-intercept versus a random intercept-slope model and selected the best-fit model. A p-value < 0.05 was considered statistically significant. Data analysis was performed with IBM/SPSS software (Version 20.0, IBM Corporation, Armonk, USA) and R (Version 3.1.1 R Foundation for Statistical Computing, Vienna, Austria), using the packages “lme4” and nlme”.