Lancet Paper Adds To Evidence That Reducing Salt To Very Low Levels May Be Dangerous

A new paper from a very large ongoing observational study offers additional and more powerful evidence that dramatic reductions in salt consumption may not be beneficial and might even prove harmful. The finding supports growing criticism that current guideline recommendations to dramatically lower salt intake in the general population may be misguided.

The study also suggests that too much salt in the diet is not a widespread global problem, though some populations which consume very high levels of salt would likely benefit from curbing salt.

The new paper in the Lancet is from the Prospective Urban Rural Epidemiology (PURE) study, an ambitious ongoing observational study. The current paper reports blood pressure and cardiovascular community outcomes in 94,000 adults from 21 countries based on estimated sodium and potassium intake calculated from a single fasting morning urine specimen. Study subjects have now been followed for an average of 8 years.

Worldwide, sodium intake is slightly less than 4 g/day, though there are wide geographical and cultural variations. The World Health Organization (WHO) guideline recommends that sodium consumption be lower than 2 g/day (equivalent to 5 g/day of salt, or 1 teaspoon), though the American Heart Association goes further and recommends an upper threshold of 1.5 g/day. These guidelines are rooted in the well established relationship between sodium intake and blood pressure. Reducing sodium intake, which leads to lower blood pressure, will inevitably result in important reductions in cardiovascular events, low salt advocates contend.

But a growing number of skeptics disagree. Andrew Mente (McMaster University), first author of the Lancet paper, said that when it comes to major reductions in salt ” there is little evidence in terms of improved health outcomes.” Further, he pointed out there is no evidence “that free-living individuals ever achieve such a low level” as recommended in the guidelines.

The new results from PURE result offer no support for very low sodium levels recommended in the guidelines but they do support more moderate reductions from very high levels. An increased cardiovascular risk was only observed in communities where the average intake was greater than 5 g/day. But only about 5% of people in developed countries have sodium intake over 5 g/day. The study found no increase in risk in the communities with sodium levels between 3-5 g/day.

By contrast, rather than finding a benefit the researchers found an increase in cardiovascular risk in communities with the lowest levels of sodium. This finding supports the perspective of the guideline critics, who propose that lowering sodium, an essential nutrient, below a certain threshold may be dangerous.

Most of the increased risk associated with very high sodium levels was observed in China, where 80% of the communities in the PURE study had an average sodium intake greater than 5 g/day. Most other communities in the study outside of China had sodium intake between 3-5 g/day.  There were no communities in the study that had  sodium intake below 3 g/day.

One important caveat to the high sodium findings is that even in communities with high sodium intake over 5 g/day there was no increase in risk if the community also consumed sufficient amounts of potassium.

“We found all major cardiovascular problems, including death, decreased in communities and countries where there is an increased consumption of potassium which is found in foods such as fruits, vegetables, dairy foods, potatoes and nuts and beans,” said Mente. Mente acknowledged that the potassium finding might not be causal. High potassium intake might be just a marker of a healthy lifestyle.

Mente emphasized that he is not saying that blood pressure isn’t a risk factor, and he acknowledges that high sodium levels help explain the association of blood pressure with stroke. But he speculates that at very low levels of sodium “you get activation of the renin angiotensin system and that explains the U shaped curve.”

An earlier 2014 report from PURE focused on data from individual participants. Critics pointed out that because of variability and reliability of the single urine specimen method the individual data was open to question. The new paper presents data from the separate communities in the study, including 255 with cardiovascular outcomes data. The PURE investigators say that this approach helps smooth over the uncertainty in the individual analysis. Mente said the community analysis also helps minimize the issue of reverse causality, since there may be a few individuals “who reduce salt intake because they are sicker… this problem is substantially minimized in the community analysis.”

Martin O’Donnell, a co-author of the paper, said that the findings “demonstrate that community-level interventions to reduce sodium intake should target communities with high sodium consumption, and should be embedded within approaches to improve overall dietary quality. There is no convincing evidence that people with moderate (average) sodium intake need to reduce their sodium intake, for prevention of heart disease and stroke.”
Dan Jones (University of Mississippi) has been a supporter of low sodium guidelines. More recently he helped lead a consensus among salt guideline advocates and critics to propose a randomized clinical trial. “This observational study  examining the relationship between dietary sodium and cardiovascular disease is not the study design capable of offering new evidence in the relationship. As recommended by the United States Institute of Medicine in 2013, an event based clinical trial is needed. Nothing in this report should change recommendations of science groups around the world to maintain daily sodium intake to 2300 mg per day or less.”

Asked to elaborate, Jones said that “the study methodology doesn’t allow for certainty of any findings.  The relationship between sodium, potassium, and blood pressure is more direct than with cardiovascular disease.”

Jones said he supports increasing potassium in the diet, but not at the expense of efforts to lower sodium. “In my view, the existing evidence supports a diet both lower in sodium and higher in potassium than the typical modern diet.  I strongly disagree with the conclusion that limiting sodium in the diet should be limited to populations consuming above 5 g/day.”

In an accompanying comment Franz H Messerli and Louis Hofstetter (University Hospital, Bern) and Sripal Bangalore (NYU) express support for a randomized controlled trial, but simultaneously point to an absurd part of the proposal, noting that “such a trial has been proposed in a closely controlled environment, the federal prison population in the USA… The simple fact that a trial looking at salt restriction has to be done in the federal prison population indicates that curtailing salt intake is notoriously difficult. incentivizing people to enrich their diets with potassium through eating more fruit and vegetables is likely to need less persuasion.”

Mente summarized his take home messages: “Our study adds to growing evidence to suggest that, at moderate intake, sodium may have a beneficial role in cardiovascular health, but a potentially more harmful role when intake is very high or very low. This is the relationship we would expect for any essential nutrient and health. Our bodies need essential nutrients like sodium, but the question is how much. The recommendation to lower sodium consumption to 2g/day is based on short-term trials of sodium intake and blood pressure, and the assumption that any approach to reduce blood pressure will necessarily translate into a lower risk of cardiovascular disease with no unintended consequences. While low sodium intake does reduce blood pressure, at very low levels it may also have other effects, including adverse elevations of certain hormones associated with an increase in risk of death and cardiovascular diseases.”

Related reading:



  1. ‘Dan Jones … propose[s] a randomized clinical trial. “… an event based clinical trial is needed. Nothing in this report should change recommendations of science groups around the world to maintain daily sodium intake to 2300 mg per day or less.”

    Asked to elaborate, Jones said that “the study methodology doesn’t allow for certainty of any findings”‘

    So Jones accepts that the only conclusive research would be an RCT. Nevertheless he defends the attempt to coerce the population to limit its salt intake to a daily 2.3 g when the evidence for that limit is the same sort of uncertain observational evidence that he disparages in the current study.

    It would be daft to suggest that they try teaching logic in medical schools, because the lack of logic in his position would be clear to a reasonably bright twelve year old.

    It’s dispiriting to reflect on how much public health advice is little better than an attempt to bully the population based on shoddy evidence. I wonder what the corresponding death toll is.

  2. Andrew Mente is a fine person to run observational studies. He has actually heard of Bradford Hill. Unlike Harvard, presumably.

  3. It looks as if reducing weight by modest amounts isn’t much use either.

    “While rates of major cardiac events were almost identical in those taking the drug, compared with the control group, those on the pills saw “small but significant” improvements in several factors linked to heart disease – including blood pressure, heart rate and blood sugar levels, the study found.”

    The logical interpretation of those results is that being overweight does not cause “cardiac events”, nor probably does having an unreduced blood pressure, an unreduced blood sugar level nor a changed heart rate.

    Of course, the scientifically literate would like to know something beyond chatter about cardiac events: how about mortality? Did it get no mention in the press release?

  4. Headline: “New heart treatment is biggest breakthrough since statins, scientists say”

    Set aside the thought that that i setting the bar rather low. Just ask: is it any ruddy good? Maybe not.

    “Prof Martin Bennett, a cardiologist from Cambridge who was not involved in the study, said the trial results were an important advance in understanding why heart attacks happen. But, he said, he had concerns about the side effects, the high cost of the drug and the fact that death rates were not better in those given the drug.”

    No mortality improvement? Is this a pattern in cardioworld?

Speak Your Mind