It is important to estimate the sodium intake of primary hypertension patients through 24-hour urine sodium excretion, but the collection and testing processes are complicated. It is not clear whether the existing equation for estimating 24-hour urinary sodium excretion from spot urine sample is valid in Chinese primary hypertension patients.

To evaluate the validity of the Kawasaki, INTERSALT, and Tanaka equations using spot urine samples to estimate 24-hour urinary sodium excretion among Chinese primary hypertension patients.

This study retrospectively selected hospitalized primary hypertension patients from Department of Cardiology Peking University International Hospital from March 2018 to March 2021. General data were collected. The 24-hour urinary sodium extraction in the 24-hour urine sample was measured. The 24-hour urinary sodium extraction from spot urine sample was estimated. Paired t-test was used to measure the significance of difference between the 24-hour urine sodium excretion from spot urine samples predicted using Kawasaki, INTERSALT or Tanaka equation and measured 24-hour urinary sodium excretion from 24-hour urine collections, and the correlation and agreement between them were analyzed using the Spearman correlation coefficient and Bland–Altman plots, respectively. Relative deviation was calculated. The accuracy of the three equations were also evaluated in patients with different levels of salt intake, diabetes or proteinuria.

One hundred and ninety-six patients were finally included for analysis. The average 24-hour urinary sodium excretion level was (165.04±78.53) mmol/d, which was equivalent to daily NaCl intake of (9.65±4.59) g/d. Female patients were older, had higher ratios of microalbuminuria and proteinuria, while lower BMI, diastolic blood pressure on admission, 24-hour average diastolic blood pressure compared with male patients (P<0.05) . Moreover, female patients were less likely to be smokers, and had lower spot urine sodium, as while as lower 24-hour urinary sodium excretion (P<0.05) . The 24-hour urinary sodium excretion estimated by Kawasaki equation was significantly higher than the measured value (t_paired=-8.008, P<0.001) , and the Bland-Altman plots showed that 94.90% spots were in 95%CI. The 24-hour urinary sodium excretion estimated by INTERSALT equation was significantly lower than the measured value (t_paired=4.167, P<0.001) . The Bland-Altman plots showed that 94.39% spots were in 95%CI. The 24-hour urinary sodium excretion estimated by Tanaka equation was higher than the measured value (t_paired=-0.547, P>0.05) without significant difference, and the Bland-Altman plots showed that 96.43% spots were in 95%CI. The relative deviation of Kawasaki, INTERSALT and Tanaka equations was -31.82%, 12.94 % and -1.90%, respectively. After dividing patients according to salt intake, the moderate correlation between predicted and actual values of each equation was no longer significant. There was no significant difference in the predictive accuracy of the equations between patients with and without diabetes or proteinuria.

The accuracy and consistency of the Kawasaki, INTERSALT and Tanaka equations in estimating 24-hour urinary sodium levels by spot urine in primary hypertension patients were poor. Tanaka's overall estimated value is the closest to the measured mean value. No comorbidity (such as diabetes, proteinuria) that might affect urinary sodium excretion was found to have a significant effect on the accuracy of the estimation equation.