Estimating salt intake in a Caucasian population: can spot urine substitute 24-hour urine samples?

Abstract

Background: A simple and valid alternative for 24-hour urine collection to estimate populational 24-hour urinary sodium excretion would be desirable for monitoring sodium intake in populations.

Aim: To assess the validity of the predicted 24-hour urinary sodium excretion using spot urine and two different prediction methods in a Danish population.

Methods: Overall, 473 Danish individuals provided a para-aminobenzoic acid-validated complete 24-hour urine collection and a spot urine sample. Data were collected in the DanThyr study (248 women aged 25-30 years and 60-65 years) and the Inter99 study (102 men and 113 women aged 30-60 years), respectively. The measured 24-hour urine sodium excretion was compared with the predicted 24-hour sodium excretion from a causal urine specimen, using both the Tanaka prediction method and a prediction model developed in a Danish population.

Results: The measured 24-hour sodium excretion (median, 5th to 95th percentile) was men 195 (110 to 360) and women 139 (61 to 258), whereas the predicted 24-hour sodium excretion for the Tanaka model was men 171 (117 to 222) and women 153 (92 to 228) and for the Danish model was men 207 (146 to 258); women 134 (103 to 163). The Spearman correlation between predicted and measured 24-hour sodium excretion was 0.39 and 0.49 for the Tanaka and the Danish model, respectively. For both prediction models, the proportion of individuals classified in the same or adjacent quintile was 74% for men and 64% for women.

Conclusions: Both prediction models gave a reasonable classification of individuals according to their sodium excretion. However, the median daily sodium intake was estimated more precisely by the Danish model, especially among men.

Keywords: Monitoring; prediction model; sodium; validation.