[Skip to Navigation]
September 14, 2011

Urinary Sodium Excretion and Cardiovascular Disease Mortality

Author Affiliations

Author Affiliation: Brigham and Women's Hospital, Boston, Massachusetts (ncook@rics.bwh.harvard.edu).

JAMA. 2011;306(10):1083-1087. doi:10.1001/jama.2011.1295

To the Editor: A recent study1 found an inverse association of urinary sodium excretion and subsequent CVD, in contrast to a wide body of evidence showing the opposite.2 The study has several analytic limitations.

First, the posited mechanism by which sodium affects CVD is through blood pressure. The authors, however, controlled for baseline blood pressure, thus removing the intermediate effect of sodium on CVD. A well-known principle of epidemiology is to avoid controlling for factors in the causal pathway. In their analysis of CVD outcomes, the authors included an additional analysis unadjusted for blood pressure, but this still controlled for antihypertensive medication. They inappropriately controlled for baseline blood pressure in analyses of incident hypertension, reversing the crude positive association.