A Method for Estimating 24 h Urinary Sodium and Potassium Excretion by Spot Urine Specimen in Stroke Patients.

Abstract

Spot urine specimens have been used to estimate 24 h urinary sodium (Na) excretion (24UNaV) and potassium (K) excretion (24UKV). However, the validity is limited for 24UNaV and unknown for 24UKV in stroke patients, using the existing formulas. Herein, we developed and validated a new formula for 24UNaV and 24UKV by spot urine specimens in stroke patients. Spot and 24 h urine samples were collected from 970 stroke patients. The models of 24UNaV and 24UKV were developed using stepwise multivariate linear regression in 689 patients. The performance of different formulas was internally validated in 281 patients at the population and individual levels. The obtained new formulas were: (1) estimated 24UNaV (mmol/day): −0.191 × Age + 4.349 × BMI + 0.229 × SpotNa + 1.744 × SpotNa/Spot creatinine (Cr) + 41.492 (for male); −1.030 × Age + 2.011 × BMI + 0.143 × SpotNa + 1.035 × SpotNa/SpotCr + 147.159 (for female); and (2) estimated 24UKV (mmol/day): −0.052 × Age + 0.410 × BMI + 0.031 × SpotK + 33.280 × Ln (spotK/spot Cr) − 5.789 × Ln (spotNa/spot Cr) − 1.035 (for male); −0.235 × Age + 0.530 × BMI + 0.040 × SpotK + 30.990 × Ln (spot K/spot Cr) − 7.837 × Ln (spotNa/spotCr) + 4.318 (for female). The new formula obtained the lowest mean bias (5.17 mmol/day for 24UNaV and 0.85 mmol/day for 24UKV) and highest proportion at the cutoff under the ±30% level for the estimation of 24UNaV (59.43%) and 24UKV (70.11%). The new formula provides a meaningful exploration to estimate 24UNaV and 24UKV in stroke patients by using spot urine specimens.