Concordance between methods for assessment of sodium intake of participants Brazilian Longitudinal Study of Adult Health (ELSA-Brasil)

INTRODUCTION:  Food Frequency Questionnaires (FFQ) have been used to estimate population´s food consumption in longitudinal studies, but they present little accuracy to assess sodium (Na) intake while golden standard is sodium urinary excretion. We aimed investigate the correlation between FFQ´s estimated intake of Na and intake measured by urinary excretion in subjects of ELSA-Brasil study.

METHODS:  Semiquantitative FFQ (114 items) and 12-h nocturnal urinary excretion were used.  FFQ´s nutrient intake was adjusted for energy consumption. Results from both methods sodium´s average intake were compared and agreement was assessed using correlation and tertiles classification of individuals according to the distribution of consumption. The sum of percentages of “exact” and “adjacent” agreements is presented.

RESULTS:  12,596 participants were evaluated. Mean age was 52±9 years, 45.9% were male. The average consumption of sodium and salt was 4.3±2.1g/day, 10.9±5.4g/day for urinary excretion, 4.5±1.7g/day, 11.4±4.2g/day for crude FFQ and 4.3±0.6g/day, 10.8±1.7g/day for energy adjusted FFQ. Significant difference was found when comparing the unadjusted intakes between the two methods (p<0.001); but when adjusted for energy no difference for sodium (p=0.528) and salt (p=0.275) intake was observed. Significant positive correlation between the methods was found for crude sodium (r=0.19) and adjusted (r=0.10). The "exact + adjacent" agreement between the methods was 81.1% for crude sodium intake, and 80.8% for adjusted. 

CONCLUSIONS: Although weak, correlation between methods was significant and agreement between them was observed. The average intake of Na and salt showed no significant differences when adjusted.