A study of dietary sodium intake and mortality presented Tuesday at Scientific Sessions showed no association between reduced sodium intake and increased risk at the low end of the sodium spectrum. Earlier observational studies had raised the possibility of a J-curve association, but the analysis presented Tuesday showed a linear association between increased dietary sodium and increased risk of future mortality over 20 years at all levels of sodium intake.
“We had a healthy group of individuals and showed that there is a straight line relationship for dietary sodium intake with both cardiovascular disease incidence and now total mortality,” said lead author Nancy R. Cook, ScD, professor of medicine at Harvard Medical School and professor of epidemiology at Harvard T.H. Chan School of Public Health in Boston. “All the evidence points to lower sodium being better.”
The AHA recommends that most adults consume no more than 1,500 mg of sodium per day.
“That is very difficult to achieve, given that most of our dietary sodium is found in prepared foods,” Cook said. “Not many people have sodium intake that low. But it seems it would be beneficial for total mortality as well as cardiovascular disease.”
Cook presented findings from a post-trial follow-up of the two phases of the Trials of Hypertension Prevention (TOHP), randomized clinical trials of various blood pressure control strategies in the 1980s and 1990s. A total of 744 phase I participants and 2,382 phase II participants were randomized to either 18 months or 36 months of dietary sodium restriction or control and compared to 3,021 usual-care participants.
Participants were between 30 and 54 years of age at baseline. None had frank cardiovascular disease and none had hypertension. The participants had what was called at the time “high normal” blood pressure and might now be assessed as prehypertensive, Cook said.
Because TOHP was a randomized clinical trial, all of the participants had multiple clinical assessments, including between three and six 24-hour urine specimens during the trial phases to assess their true sodium intake. More than 3,000 participants were not assigned to a sodium-reduction intervention and were classified by average sodium intake for the analysis of average sodium levels. The participants were then followed for between 23 and 26 years for cardiovascular disease and for all-cause mortality in the post-trial period.
“This study utilized the breadth and depth of clinical measurements collected during the two phase I and II clinical trial periods,” said Donna K. Arnett, PhD, MSPH, chair and professor of epidemiology at the University of Alabama Birmingham School of Public Health and a past AHA president. “Rather than relying on self-reported sodium intake, these participants had multiple 24-hour urine samples, the gold standard of sodium intake. They were randomized to a wide range by sodium intake and they were followed for up to 26 years. There is no evidence of harm and, most importantly, no evidence of any J-shaped relationship within this very well-done study.”
Earlier studies assessed TOHP participants for cardiovascular disease 10 to 15 years following the trial, Cook said. In addition to statistically significant reductions in cardiovascular disease associated with lower sodium intake, there was a trend for lower overall mortality associated with reduced dietary sodium.
The current analysis is based on more than 20 years of mortality follow-up through the end of 2013. Active intervention to reduce sodium intake was associated with a hazard ratio of 0.85 with a nonsignificant p value of 0.19. Participants were also classified by dietary sodium intake: less than 2,300 mg/day; 2,300 to 3,599 mg/day; 3,600 to 4,799 mg/day; and 4,800 mg/day or higher.
“Of particular interest to us was the bottom of the curve, to see whether there was any evidence of an uptick, a J shape, or if the dietary sodium-mortality relationship remained linear,” Cook said. “We saw a very linear relationship throughout. Using the best measure of sodium available, 24-hour urine, we saw no evidence of a J-shaped curve.”