Confusion over new salt research

Friday May 6 2011

“Salt is GOOD for you,” according to claims in the Daily Mail. The newspaper challenged conventional health advice by suggesting that “eating more could even lower the chances of heart disease”.

However, these claims are somewhat unjustified as they are based on a study that actually looked at a one-off measure of salt in people’s urine rather than in their diet. The research looked at 3,700 people’s urinary salt levels and then followed them for nearly eight years to look at their risk of high blood pressure, cardiovascular disease (CVD) and related deaths.

Among the main results the researchers observed 84 CVD-related deaths. Surprisingly, they found that there were 50 CVD-related deaths in the third of participants with the lowest salt levels, and just 10 deaths in those passing the most salt. This would initially seem to challenge the conventional wisdom that salt raises blood pressure and, therefore, the risk of heart problems. However, this study is not straightforward to interpret, particularly as the single urinary sodium measure analysed is not necessarily a direct indicator of how much salt a person eats. For example, it may indicate how hydrated a person is or how well their kidneys are filtering sodium.

The limitations of this study mean that, on its own, it does not challenge the accepted association between salt intake, blood pressure and related disease, and certainly does not suggest that eating more salt is good for you.

Where did the story come from?

The study was carried out by investigators from the European Project on Genes in Hypertension (EPOGH), a research project based in Belgium and supported by various European study and research grants. The study was published in the peer-reviewed Journal of the American Medical Association.

The Daily Mail ’s headline implying that eating salt is good for you is a rather simplistic conclusion from this complex study, and the study cannot be interpreted in this way. Crucially, it should be remembered that a single measure of someone’s urinary salt excretion does not necessarily equate to the level of salt they consume. Health recommendations are unlikely to change based on this study alone.

What kind of research was this?

This study aimed to assess whether participants’ blood pressure (BP) and cardiovascular health outcomes could be predicted using 24-hour urinary sodium (salt) excretion measures. This was done by measuring the level of salt passed in the participants’ urine in a 24-hour period. The researchers looked at two study cohorts of healthy, middle-aged people who were involved in either the Flemish Study on Environment, Genes and Health Outcomes (FLEMENGHO, 1985-2004), or in the European Project on Genes in Hypertension (EPOGH, 1999-2001).

What did the research involve?

The members of both cohorts were randomly sampled from the general population of Belgium (average age 38-40), with this study recruiting 3,681 participants who were free of cardiovascular disease. At the start of the study the participants had their urinary sodium excretion levels measured, as well as their blood pressure and body measurements. Various health and lifestyle factors were also assessed.

During a follow-up period of, on average, 7.9 years, the researchers identified any diseases and causes of death among the participants using medical databases, death certificates and hospital and medical records. They specifically looked at fatal and non-fatal cardiovascular events such as heart attacks and strokes, and looked at the association between sodium excretion levels at the start of the study, development of high blood pressure and CVD, and CVD-related mortality. They calculated risk according to three tertiles of sodium excretion by splitting the participants into three groups based on their urinary salt levels.

What were the basic results?

At the start of the study there were 3,681 participants but over the follow-up 219 people died. After exclusion of those who were seriously ill, those who moved out of the study area and those who did not wish to attend for further assessment, the researchers were left with a total of 2,856 people who were able to attend for re-assessment.

Of the 2,856 participants, 2,096 had normal blood pressure at the start of the study, allowing the researchers to assess whether starting salt level predicted development of high blood pressure in this group. A total of 1,499 participants had assessments of their blood pressure and urinary sodium excretion at both the study start and follow-up, allowing the researchers to assess how changes in sodium level reflected changes in blood pressure within this group.

Of the 3,681 people in the study, 232 experienced a fatal or non-fatal CVD event, such as a heart attack, over the 7.9 years.

There were 84 cardiovascular deaths, which were distributed according to tertile of salt excretion:

  • low tertile (mean urinary sodium 107mmol): 50 deaths
  • medium tertile (mean urinary sodium 168mmol): 24 deaths
  • highest tertile (mean urinary sodium 260mmol): 10 deaths

With adjustment for potential confounders, this meant that those in the lowest group had a higher risk of CVD mortality (hazard ratio 1.56, 95% confidence interval I 1.02 to 2.36) compared with the overall risk calculated for the cohort as a whole.

Among the 2,096 participants with normal blood pressure at the start of the study, baseline salt excretion level was not associated with risk of developing high blood pressure.

From data on the 1,499 participants with assessments at both study start and end of follow-up, the researchers calculated that a 100mmol increase in sodium excretion was associated with 1.71mm Hg increase in systolic blood pressure (which is the top figure in a two-figure blood pressure reading which reflects the arterial pressure when the heart contracts and pumps blood into the arteries).  There was no change in diastolic blood pressure (the bottom figure, which reflects the arterial pressure at the point when the heart relaxes and fills with blood).

How did the researchers interpret the results?

The researchers conclude that in their population-based cohort, an increase in sodium excretion is associated with an increase in systolic blood pressure but not diastolic pressure. However, this association was not linked to an increased risk of CVD complications, with them finding the unexpected result that lower sodium excretion was associated with higher CVD mortality.


This study aimed to look at whether 24-hour urinary sodium excretion is predictive of blood pressure and CVD outcomes and has found some conflicting results that are quite difficult to interpret.

The traditionally held theory is that higher salt intake and higher salt levels in the body increase blood pressure, which would therefore be expected to increase a person’s risk of developing CVD, or dying from CVD. However, some of the results of this particular study do not seem to reflect this accepted mechanism, with lower urinary salt levels (the proxy measure of salt intake used in this study) being associated with higher death risk, and higher levels being associated with lower death risk. However, they conversely found in a smaller sample of people who had their blood pressure and salt excretion measured at both study start and end of follow-up that an increase in salt excretion over time was associated with a small systolic blood pressure increase. This is consistent with current understanding about the association between salt levels and blood pressure.

These results are puzzling, and should be interpreted cautiously for a number of reasons. Crucially, this study looked at urinary salt levels over a single 24-hour period at the start of the study, which creates a number of potential problems:

  • Salt excretion does not necessarily equate to dietary salt intake and, therefore, those passing less salt should not be considered to be eating less salt on the basis of a single measurement. A single measure may be influenced by how well hydrated a person is or how well their kidneys are functioning.
  • Although some people had a second salt measure taken at the end of the study, the participants did not have salt measurements taken during the course of the 7.9-year follow-up period. This means we cannot tell if these measurements reflected their levels throughout the study period or in everyday life.

There are also some other limitations to the study:

  • Although the study population was large there were only 84 cardiovascular deaths that occurred. This is largely because the participants were fairly young (average 38-40 years) and free from CVD at the study start, so you would not expect many deaths in this young cohort over eight years. Having small numbers of deaths in each of the three salt excretion groups increases the risk of calculating inaccurate risk associations between salt intake and risk of death.
  • As the researchers note, their results seemed to vary between the FLEMENGHO and EPOGH cohort study members, which means that further research in other population groups would be beneficial.
  • As the researchers also say, their results were mainly applicable to white Europeans and should not be generalised to other ethnic groups.  

Further research into the association between salt intake, blood pressure and related mortality is likely before any change to general health recommendations is considered.

Analysis by Bazian
Edited by NHS Choices