select search filters
roundups & rapid reactions
factsheets & briefing notes
before the headlines
Fiona fox's blog

expert reaction to study looking at salt consumption and health risks

Scientists publishing in The Lancet examine the association between salt consumption and health risks.


Prof Tom Sanders, Professor emeritus of Nutrition and Dietetics, King’s College London, said:

“This is a large observational study in men and women whose average age was 50 years, where the incidence of cardiovascular disease is relatively low.  Salt intake is difficult to measure from diet records and this study used the level of salt in urine passed in the morning to estimate 24 hour levels.  Salt intake can be estimated by the urinary excretion of sodium over 24 hours.  There are limitations to this, particularly in very hot climates where large amounts of salt may be excreted in sweat.  Furthermore, a single morning sample has some limitations as it assumes that it reflects the intake the previous day.  Usually to measure salt intake, a 24 hour urine sample is required and some measure of completeness of urine collection, which was lacking from the present study.  The concentration of sodium in urine also depends on how much water has been consumed.  The countries studied varied markedly in climate, which can influence salt and fluid intake as well as blood pressure (a higher ambient temperature is associated with lower blood pressure).

“This study confirms that increasing sodium excretion is associated with increases in blood pressure and that high intakes of salt over 12 g per day (about 5g of sodium per day) are associated with an increased risk of cardiovascular disease, particularly stroke, but not lower intakes.  What is controversial in this paper is the suggestion that it may be harmful and so not be worth lowering intakes as low as that recommended by the WHO guidelines.  The increased risk of heart disease observed in those with lowest sodium excretion is unlikely to be a causal relationship as there are no known mechanisms that could explain this observation.

“Previous research has shown that the benefit of salt reduction is greatest in people with raised blood pressure and those over the age of 50 years.  In the UK and most developed economies, most adults over the age of 50 have raised blood pressure.  Intakes of salt in the UK have fallen over the past thirty years from over 12 g per day to 7-8 g per day, and this has been accompanied by a fall in average blood pressure of the population.  Japan used to have a very high prevalence of high blood pressure and high rates of stroke, and took action to cut salt intake in the 1970s and now has much lower rates.  Stroke remains a major killer in China which still has very high intakes of salt because of the extensive use of soy sauce, which typically contains about 20% salt by weight, as well as a lot of pickled foods.  Current guidelines for salt intake in the UK and Europe are to consume less than 6 g salt per day, which is slightly higher than the 5 grams (2 grams of sodium) recommended by WHO.

“Sodium is an essential nutrient but the requirement is very low at about 0.5g per day – except in very hot conditions where large amounts of sweat may lead to increased losses of the mineral.

“In ageing populations such as the UK, it remains sensible to advise people to restrict the addition of salt to food.”


Dr Gunter Kuhnle, Associate Professor in Nutrition and Health, University of Reading, said:

“This large observational study confirms what is already known about high salt intake: it results in higher blood pressure and an increased risk of cardio-vascular diseases.  It also confirms that very low salt intake can also affect health badly – which is the case for most nutrients.

“The main limitation of the study is how salt intake was measured: the method used in the study is not suitable to estimate the actual salt intake reliably.  It is therefore not possible to identify an optimal range of intake based on the data presented, or to assess whether the current WHO recommendation of 5 g/d is sufficient.

“The accurate assessment of salt intake in a free living population is notoriously difficult as there are many dietary sources.  A convenient and more reliable alternative is therefore to measure salt excretion in urine.  However, as salt excretion varies considerably during the day, a 24h urine sample is required for an accurate estimate.  There are methods to estimate 24h salt excretion from morning or spot urine samples, such as those used in this study, but they are known to be unreliable.  While the data obtained in the study are sufficient to categorise participants according to their salt intake, it is not possible to estimate reliably their actual intake.”


Prof Francesco Cappuccio, Chair of Cardiovascular Medicine and Epidemiology, University of Warwick, said:

“The latest publication in The Lancet from the PURE study does not add anything to the knowledge of the effects of salt on cardiovascular outcomes and, more importantly, does not provide any evidence that reducing population sodium consumption by a moderate amount cause harm.

“The PURE study, due to the numerous flaws highlighted in the last few years in international journals, is not fit to address any of the issues regarding salt consumption and cardiovascular outcomes.

“The assessment of exposure by spot urine collections and the use of the Kawasaki formula is flawed, leading to biased estimates, also shown in the PURE validation of Chinese data.  The Kawasaki formula leads to a systematic bias – this bias leads to an overestimate of CVD risk at lower levels of intake.  The effect of this bias ‘creates’ a J-shaped curve between sodium consumption and cardiovascular outcomes, as recently shown using data from the Trials Of Hypertension Prevention (TOHP) I and II (published in International Journal of Epidemiology a few weeks ago) – the TOHP trials data was not taken into consideration in the present study.  Several repeated 24h urine collections with sodium estimates do show a linear and graded relationship between sodium consumption and CVD risk, with no J shape at all.

“The use of ‘ecological’ associations, which relate average exposures with average outcomes, is immaterial to the interpretations offered by the authors and editorial, due to the lack of allowance for important confounders.

“Concerted actions globally, led by the World Health Organization and endorsed by multiple international health organizations should continue to reduce population salt consumption.”


Prof Rod Taylor, Chair of Health Services Research and Director of Exeter Clinical Trials Unit & NIHR Senior Investigator, University of Exeter Medical School, said:

“This interesting observational study conducted in a number of counties looks well conducted in terms of the methods used.  The authors adjusted their analysis for factors that could skew the results, such as age, sex, body-mass index, education, alcohol intake, and smoking status, and they analysed individuals rather than populations.  The findings suggest that harmful impacts of sodium consumption – the main component of salt in our diets – on stroke are limited to people who eat more than ~5g sodium per day.  This is well above current public health guidelines, which are to decrease sodium intake to 2 g per day. This group was mainly in Asia.  The findings imply that such current guidance may have limited public health benefit in other settings with typically lower sodium intakes.  The results of this study would probably have been even more convincing if it had been possible to repeatedly test participants’ urine over 24 hours.

“I think the editorial nicely sums up what I believe to be the appropriate interpretation of this study.  These findings were observational, so cannot claim cause and effect, and they were limited to a mostly Asian population, so we don’t know how widely applicable they might be.  Salt intake in this study was estimated from the one measurement of overnight fasting urine.  So we can’t know from this study alone whether changes in salt intake interventions would be beneficial.

“The Cochrane review of randomised controlled trials of dietary intervention (last updated in 2013) suggested that the existing evidence as a whole was not highly powered enough to confirm clinically important effects of dietary advice or salt substitution on cardiovascular mortality1.  This new study, and that previous Cochrane review, both support a call for a definitive randomised controlled trial in this area.  However, randomised controlled trials in this field remain ethically and practically very difficult to design and deliver.”



Prof Graham MacGregor, Professor of Cardiovascular Medicine, Queen Mary University of London, said:

“This publication is simply a repeat of a study issued by the authors in 2016, using a slightly modified study population and analysis, however the authors have not addressed any of the serious criticisms from the wider scientific community of their 2016 study in this subsequent publication.  These criticisms include the use of ill participants in the study, leading to reverse causality (i.e. those suffering with heart disease don’t eat much food, and consequently eat less salt, but it is the illness that leads to death rather than lower salt intake), and the use of spot urine measurements, which, as our paper in the International Journal of Epidemiology1 demonstrates, is an inaccurate measurement and leads to incorrect findings on the relationship between salt intake and health.  The totality of evidence shows that lowering salt intake leads to a fall in cardiovascular disease events, the commonest cause of death and disability worldwide.  In our view, papers of poor scientific quality should not be considered as part of the evidence base.”

1 He FJ, Campbell NRC, Ma Y, MacGregor GA, Cogswell ME, Cook NRC. ‘Errors in estimating usual sodium intake by the Kawasaki formula alter its relationship with mortality ─ Implications for public health’. Int J Epidemiol 2018


Prof Peter Sever, Professor of Clinical Pharmacology & Therapeutics, Imperial College London, said:

“There are serious methodological flaws in the study design.  Early morning or spot urines cannot be used as a reliable measure of an individual’s salt intake.  The marked variation between individuals in salt intake in most populations (probably except in China, where daily intake is consistently high) makes use of spot urines an unreliable determinant of an individual’s average salt intake.

“The study has the usual problems with observational studies, where confounding factors influence associations with morbidity and mortality.  A classic example is ill health influencing poor diet (so ill people have low cholesterol and low nutrient intake).  It’s not the diet that causes the disease!  This is an example of reverse causation.”


* ‘Urinary sodium excretion, blood pressure, cardiovascular disease, and mortality: a community-level prospective epidemiological cohort study’ by Andrew Mente et al. published in The Lancet on Thursday 9 August 2018.


Declared interests

Prof Tom Sanders: “Scientific governor of British Nutrition Foundation, Honorary Director of Nutrition HEART UK.”

Dr Gunter Kuhnle: “I have received funding from the EU as part of a FP7 programme to develop new processing methods for meat, and from Mars for flavanol related research.”

Prof Francesco Cappuccio: “President and Trustee of the President of the British and Irish Hypertension Society, Head of the WHO Collaborating Centre for Nutrition, member of CASH, WASH, TRuE Consortium – all unpaid.”

Prof Rod Taylor: “RT is an author on the Cochrane systematic review Reduced dietary salt for the prevention of cardiovascular disease. He has no other competing interests.”

Prof Graham MacGregor: “Graham is Chair of Blood Pressure UK (BPUK), Action on Salt and World Action on Salt and Health.  BPUK, Action on Salt and WASH are non-profit charitable organisations and Graham does not receive any financial support from any of these organisations.”

Prof Peter Sever: “I have no conflict of interest when commenting on this study.”


in this section

filter RoundUps by year

search by tag