Measurement of mean population Na and K intakes typically uses laboratory-based assays, which can add significant logistical burden and costs. A valid field-based measurement method would be a significant advance. In the current study, we used 24 h urine samples to compare estimates of Na, K and Na:K ratio based upon assays done using the field-based Horiba twin meter v. laboratory-based methods.

The performance of the Horiba twin meter was determined by comparing field-based estimates of mean Na and K against those obtained using laboratory-based methods. The reported 95 % limits of agreement of Bland–Altman plots were calculated based on a regression approach for non-uniform differences.

The 24 h urine samples were collected as part of an ongoing study being done in rural China.

One hundred and sixty-six complete 24 h urine samples were qualified for estimating 24 h urinary Na and K excretion.

Mean Na and K excretion were estimated as 170·4 and 37·4 mmol/d, respectively, using the meter-based assays; and 193·4 and 43·8 mmol/d, respectively, using the laboratory-based assays. There was excellent relative reliability (intraclass correlation coefficient) for both Na (0·986) and K (0·986). Bland–Altman plots showed moderate-to-good agreement between the two methods.

Na and K intake estimations were moderately underestimated using assays based upon the Horiba twin meter. Compared with standard laboratory-based methods, the portable device was more practical and convenient.