U-shaped curve for association between sodium intake and CV events

Urinary sodium excretion, blood pressure, cardiovascular disease, and mortality: a community-level prospective epidemiological cohort study

Literature - Mente A, O’Donnell M, Rangarajan S et al. - The Lancet 2018; 392: 496–506

Introduction and methods

Guidelines recommend a mean sodium intake of no more than 2 g/day for the general population, based on the assumption that this dietary measure may lead to reduced incidence of elevated blood pressure (BP), CV disease and mortality [1,2]. However, there are no data supporting a direct association between increased sodium intake and adverse CV events on the population level.

The Prospective Urban Rural Epidemiology (PURE) study includes detailed data on mean sodium intake, CV disease, and mortality in a large sample of individuals from the general population in different countries and communities [3]. This analysis of the PURE study evaluated the association between mean sodium intake and CV outcomes at community level, in participants without CV disease at baseline.

PURE includes 168,067 individuals aged 35–70 years. For this analysis, a morning fasting midstream urine sample was collected from every participant, to estimate the 24 h urinary sodium and potassium excretion, which was used as a surrogate for sodium and potassium intake. Moreover, BP was measured. Participants were followed at 3, 6 and 9 years to collect data on risk factors and outcomes.

Tests for linearity and deviation from linearity were used to determine the slope describing the relations between mean sodium and potassium intake (per 1 g increase in intake for each) and the adjusted mean CV event rates. To assess whether the associations at community level varied by mean sodium or potassium intake, subgroups were made by subdividing the communities into tertiles of intake and estimated associations within each were assessed (for sodium: lowest tertile <4.43 g/day, middle tertile 4.43–5.08 g/day, highest tertile >5.08 g/day).

Main results

  • Data from 95,767 participants in 369 communities were used for the BP analyses, and data from 82,544 participants in 255 communities were used for the CV analyses.
  • The median follow-up was 8.1 years (IQR: 5.8–9.4).
  • The mean sodium intake for the 369 communities was 4.77 g/day (range: 3.22–7.52), while in China it was higher (5.58 g/day). The mean potassium intake was 2.16 g/day (range: 1.25–3.11).
  • For each 1 g increase in estimated sodium intake, systolic BP (SBP) increased by 2.86 mm Hg (95%CI: 2.12–3.60; P<0.0001). In the highest sodium intake tertile association between SBP and sodium intake was positive and significant, but an inverse and non-significant association was observed with sodium intakes in the middle and lower tertiles (P<0.0001 for heterogeneity). The results for diastolic BP (DBP) were similar.
  • Potassium intake was not significantly associated with SBP or DBP.
  • There was a positive association for each 1 g increase in sodium intake and the mean overall CV event rate (0.66 events per 1000 years, 95%CI: 0.46-0.87; P<0.0001 and after multivariable adjustment 0.73, 95%CI: 0.53-0.93, P<0.0001).
  • In the lowest tertile of sodium intake the association with CV events was inverse and significant (change –1.00 events per 1000 years, 95%CI: –2.00 to –0.01; P=0.0497), in the middle tertile there was no association (change 0.24 events per 1000 years, 95%CI: –2.12 to 2.61; P=0.8391), whereas in the highest tertile the association was positive but non-significant (change 0.37 events per 1000 years, 95%CI: –0.03 to 0.78; P=0.0712).
  • Each 1 g increase in sodium intake was associated with stroke (0.97 events per 1,000 years; 95%CI: 0.81–1.13; P<0.0001; after multivariable adjustment 1.01; 95%CI: 0.86–1.17; P<0.0001), with a positive association found only among communities in the highest tertile of sodium intake (0.54 events per 1000 years, 95%CI: 0.12-0.96; P<0.0001) and no significant association found in the middle or lowest tertiles.
  • In China, a strong association was seen between sodium intake and stroke (mean sodium intake 5.58 g/day, 0.42 events per 1000 years, 95% CI: 0.16- 0.67, P=0.0020).
  • An inverse association was seen between increasing mean potassium intake and all major CV events after multivariable adjustment.


In a large population-based study, decreased sodium intake was associated with an increased CV risk and an increased sodium intake was associated with an increase in stroke. The authors suggest that a population-specific strategy for sodium reduction targeted at countries or communities with sodium intake greater than 5 g/day would be preferable to a population-wide strategy of sodium reduction to reduce CVD.

Editorial comment

In their editorial article, Messerli, Hofstetter, and Bangalore [4] characterize the Mente et al. data as ‘provocative’, and criticize their work because the findings suggest that sodium may have a cardio protective effect, and the analyses are based on a sole urinary sodium excretion measurement, without taking into account any potential confounders. They mention the study that was published 2 years ago by Mente et al. stating that salt restriction could be harmful [5]. This resulted in a reaction of the community of 'disbelief' that 'such bad science' should be published in The Lancet [6]. The authors conclude: ‘Arguments put forward about the relation between salt and cardiovascular disease reflect a stance that is more befitting of medieval ecclesiasts than modern day scientists. Judgmental terms like “bad science” might simply reflect that findings are at variance with preconceived notions of the beholder… Methodological issues notwithstanding, there is little doubt that with the notable exception of benefits of increasing potassium intake, the findings of Mente and colleagues will again give rise to controversy. But then, “controversy is the life-blood of science” and “should be welcomed, as it is only through controversy that science advances”.


1. WHO. Guideline: sodium intake for adults and children. Geneva: World Health Organization, 2012.

2. Whelton PK, Appel LJ, Sacco RL, et al. Sodium, blood pressure, and cardiovascular disease: further evidence supporting the American Heart Association sodium reduction recommendations. Circulation 2012; 126: 2880–89.

3. Yusuf S, Rangarajan S, Teo K, et al; PURE Investigators. Cardiovascular risk and events in 17 low-, middle-, and high-income countries. N Engl J Med 2014; 371: 818–27.

4. Franz H Messerli, Louis Hofstetter, Sripal Bangalore. Salt and heart disease: a second round of “bad science”? Lancet 2018; 392:456-8.

5. Mente A, O’Donnell M, Rangarajan S, et al. Associations of urinary sodium excretion with cardiovascular events in individuals with and without hypertension: a pooled analysis of data from four studies. Lancet 2016; 388: 465–75.

6. Johnston I. Lancet attacked for publishing study claiming low-salt diet could kill you. Independent May 21, 2016: 12.

Find this article online at The Lancet

Facebook Comments


We’re glad to see you’re enjoying PACE-CME…
but how about a more personalized experience?

Register for free