Comparison of methods for estimating discretionary salt intake among females of reproductive age in Oromia region, Ethiopia.

BACKGROUND

Accurate and precise estimates of discretionary salt intake are essential to the design of salt fortification and sodium reduction interventions; however, there is limited methodological guidance for low-resource field settings.

OBJECTIVES

The objective of this study was to compare measures of association and agreement between the weighed food record (WFR) and 4 other methods for estimating daily discretionary salt intake among females of reproductive age (FRA) in the Oromia region of Ethiopia.

METHODS

A total of 100 FRA were enrolled in a cross-sectional study that assessed discretionary salt intake using 5 approaches: 1) 1-d WFRs; 2) duplicate diet composites to assess the total sodium content of the diet and preparation of corresponding replicate diet composites excluding discretionary salt to measure the intrinsic sodium content of the diet in 40 participants; 3) 24-h urinary sodium excretion; 4) 1-d household salt disappearance; and 5) 1-wk household salt disappearance. WFR, 1-d household salt disappearance, and 24-h urine collections were repeated in a subset of 40 participants to estimate usual discretionary salt intake from observed discretionary salt intake. Correlation and Bland-Altman analyses were conducted to assess the association and agreement between the WFR and methods 2-5.

RESULTS

Estimates of the mean ± SD discretionary salt intake ranged from 6.8 ± 1.9 g/d (WFR usual intake estimate) to 10.0 ± 7.3 g/d (1-d household salt disappearance per household member). One-day salt disappearance per adult female equivalent and per household member was most strongly associated with the WFR method (r = 0.675, P < 0.001 and r = 0.645, P < 0.001, respectively). The difference in mean discretionary salt intake was the smallest between the replicate diet composite and the WFR method (-0.58 ± 4.21 g).

CONCLUSIONS

The WFR method, our predefined reference method, with repeated assessments in a subgroup of participants likely yields the most precise and accurate estimation of actual discretionary salt intake. If resource constraints preclude this method, estimating salt utilization per household member via a 1-d disappearance study following preliminary WFRs to allow for potential adjustment may be an appropriate alternative.