The present study set out to determine whether morning spot urine samples can be used to monitor Na (and K) intake levels in South Africa, instead of the ‘gold standard’ 24 h urine sample.

Participants collected one 24 h and one spot urine sample for Na and K analysis, after which estimations using three different formulas (Kawasaki, Tanaka and INTERSALT) were calculated.

Between 2013 and 2015, urine samples were collected from different population groups in South Africa.

A total of 681 spot and 24 h urine samples were collected from white (n 259), black (n 315) and Indian (n 107) subgroups, mostly women.

The Kawasaki and the Tanaka formulas showed significantly higher (P≤0·001) estimated Na values than the measured 24 h excretion in the whole population (5677·79 and 4235·05 v. 3279·19 mg/d). The INTERSALT formula did not differ from the measured 24 h excretion for the whole population. The Kawasaki formula seemed to overestimate Na excretion in all subgroups tested and also showed the highest degree of bias (−2242 mg/d, 95 % CI−10 659, 6175) compared with the INTERSALT formula, which had the lowest bias (161 mg/d, 95 % CI−4038, 4360).

Estimations of Na excretion by the three formulas should be used with caution when reporting on Na intake levels. More research is needed to validate and develop a specific formula for the South African context with its different population groups. The WHO’s recommendation of using 24 h urine collection until more studies are carried out is still supported.