[Skip to Navigation]
September 14, 2011

Urinary Sodium Excretion and Cardiovascular Disease Mortality

Author Affiliations

Author Affiliations: Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, Louisiana (jhe@tulane.edu).

JAMA. 2011;306(10):1083-1087. doi:10.1001/jama.2011.1292

To the Editor: Low sodium excretion was associated with higher CVD mortality in a prospective study of 3681 participants from Europe.1 The results should be interpreted with caution considering several flaws in study design, analysis, and interpretation.

A major source of bias in this study was misclassification of urinary sodium excretion. A single 24-hour urine sample, which was used for all the analyses, cannot represent usual sodium intake at the individual level. Nor can it be used to accurately classify individuals into categories as the authors did in analyzing 3 levels of sodium excretion with CVD mortality. It has been well documented that differential misclassification bias occurs when a continuous exposure variable with random measurement error is classified into multiple categories.2 Although the authors tried to exclude participants with inaccurate urine collection, it would have been helpful to examine the association of CVD mortality with sodium/creatinine ratio because this ratio is less affected by incompleteness of urine sample collection.