Unclear results for salt reduction study

The Daily Express has today claimed that “salt is safe to eat”, and that, after years of lecturing, the “health fascists” have been proved wrong.

This news is based on a systematic review that combined data from seven earlier studies looking at how reduced-salt diets affected the risk of cardiovascular disease (CVD), blood pressure and death. The authors of this combined study wanted to examine this because, although there is evidence that reducing salt leads to a reduction in blood pressure, no reviews had directly looked at these outcomes.

Although the studies provided data on nearly 6,500 participants, the researchers found that the number of participants who died or developed outcomes such as heart attacks was low, making it hard to estimate the impact of salt reduction in a meaningful way. The researchers concluded that there is not enough evidence to say with certainty that advising a reduced-salt diet is linked to reductions in death and CVD events, but crucially they have not said that salt reduction has no effect, as the coverage in the Express might suggest.

The review’s authors have called for further high-quality, long-term research to determine conclusively the effects of salt reduction. The NHS’s current advice, that adults should limit their salt intake to 6g a day, remains unchanged.

Where did the story come from?

The study was carried out by researchers from the University of Exeter, the University of Bristol, Florida Atlantic University, the University of East Anglia and the London School of Hygiene and Tropical Medicine, as part of the Cochrane Collaboration. The study was funded by the UK National Institute for Health Research, as part of the Cochrane Collaboration.

The study was published in the peer-reviewed American Journal of Hypertension.

The findings of this study were overstated by media outlets, with the Daily Express commenting that a pool of 6,489 people from across seven studies was large enough to draw conclusions from. However, although the number of people included in the studies was relatively large, the researchers say that the number of events seen during these trials was too small to detect a reduction in risk with sufficient certainty. In email bulletins from the Cochrane Library, one of the reviewers, Professor Rod Taylor, estimates that data from at least 18,000 individuals would be required before any clear health benefits could be revealed.

What kind of research was this?

This was a systematic review and meta-analysis of randomised controlled trials that had examined the effects that salt-restricted diets have on outcomes including death, heart attacks, blood pressure, angina and stroke.

When conducting a systematic review, researchers pull together and analyse all appropriate studies relating to a particular topic, regardless of their findings. A meta-analysis, which pools the findings of included trials and analyses the data as an aggregate, is used to increase the ability (or “power”) of the study to detect an effect. This ability is improved as the number of participants studied increases. In this instance, the published study included nearly 6,500 participants.

What did the research involve?

The researchers conducted a literature search and included studies for analysis if:

The design was a randomised control trial with a follow-up period of at least six months.
The intervention being tested was a reduced-salt diet or advised reduced-salt diet.
The outcomes of interest included death due to cardiovascular disease (CVD) or any cause of CVD events, such as heart attack, angina, stroke or heart failure, or hospitalisation due to CVD. Data on blood pressure and estimated salt intake were also analysed where available.

The researchers assessed the quality (risk of bias) in each of the selected studies, and analysed the data by pooling the results in several different ways.

Researchers analysed the data in three separate groups, depending on the characteristics of the population under study at the start of the study:

people with normal blood pressure
people with high blood pressure
a mixed population of the two.

The only trial that included a mixed group was carried out in people diagnosed with heart failure who had been hospitalised within the previous month.

Across the seven studies, the treatments offered to the intervention and control groups varied. The interventions included group counselling sessions, advice and information leaflets, dietary counselling and specially prepared meals in which regular table salt was swapped for low-sodium salt. Across the control groups, some received no active behavioural intervention or advice, while others received specially prepared meals containing regular table salt.

What were the basic results?

The analysis included seven trials comprising a total of approximately 6,500 participants. These trials had formed the basis of 39 separate published research papers.

A total of 665 deaths from all causes (including 98 deaths from CVD) and 293 CVD events were reported across the included trials. Data was analysed separately for trials of people with normal blood pressure, high blood pressure and a mix of the two.

Trials of those with normal blood pressure showed:

no evidence for a reduction in the number of deaths (from any cause) in the intervention group
no evidence of reduction in CVD events in the intervention group
weak evidence for an average reduction in systolic blood pressure (maximum pressure exerted with each heart beat) of 1.1 mm Hg in the intervention group compared to control group participants
evidence for an average reduction in diastolic blood pressure (pressure exerted while the heart is at rest) of 0.80 mm Hg in the intervention group compared to control group participants
evidence for a reduction in estimated salt intake (as measured by changes in urinary sodium levels) of 34.19 mmol a day in the intervention group compared to control group participants, suggesting compliance with the interventions

Trials of those with high blood pressure showed:

no evidence for a reduction in the number of deaths (due to any cause) in the intervention group
no evidence for a reduction in the number of CVD deaths in the intervention group
no evidence of reduction in CVD events in the intervention group
evidence for an average reduction in systolic blood pressure (pressure exerted while the heart is beating) of 4.1 mm Hg in the intervention group compared to the control group
no evidence for a reduction in diastolic blood pressure (pressure exerted while the heart is at rest) in the intervention group compared to the control group evidence for a reduction in estimated salt intake (as measured by changes in urinary sodium levels) of 39.11 mmol a day in the intervention group compared to control group participants

The single trial looking at people with heart failure, who had either normal or high blood pressure, showed:

evidence for an increase in the number of deaths (from any cause) in the intervention group
evidence for an average reduction in systolic blood pressure (pressure exerted while the heart is beating) of 4.0 mm Hg in the intervention group compared to the control group
no evidence for a reduction in diastolic blood pressure (pressure exerted while the heart is at rest) in the intervention group compared to the control group
evidence that estimated salt intake (as measured by changes in urinary sodium levels) reduced by 27.00 mmol a day in the intervention group compared to control group participants

How did the researchers interpret the results?

Researchers concluded that, within the pooled research they analysed, there was no strong evidence that a restricted salt diet reduced death due to any cause, or reduced the number of CVD events in people with normal or high blood pressure. They also say that there was an increase in death due to any cause in people with heart failure who were on a restricted-salt diet.

The researchers say that they found consistent evidence for a reduction in urinary sodium levels, indicating that the participants continued to comply with the salt-restricted diet. However, they do say that this adherence is likely to decline over time.

The researchers also say that the observed reductions in blood pressure (between 1 and 4 mm Hg) would, according to current risk models, generally predict a reduction in death due to CVD by 5% to 20%. However, only two trials included data on CVD death, and the number of events reported was small. This increased the uncertainty surrounding the observed effect.

Researchers say that the small number of events observed across the seven trials is a limitation of their review. More observed events would be needed in order to reduce the uncertainty surrounding the estimated effects of salt reduction. They also say that since most of the participants knew whether they were on a reduced salt or regular salt diet, other lifestyle and dietary behaviours may have simultaneously been taken up by the control group, thus weakening the observed effect.

Conclusion

Performing a systematic review and meta-analysis can increase the statistical power of research on a topic, helping to detect differences in outcomes. However, although the number of individuals pooled in the analysis was large (approximately 6,500 people), the number of events which occurred was relatively small, reducing certainty as to what impact the interventions may have had. One of the study authors has estimated that at least 18,000 participants would need to be analysed in order to identify any effects clearly.

Points to note are that:

The reduced-sodium interventions did actually produce the expected positive effect, but the researchers could place little confidence in those measurements due to the relatively small number of events observed.
The researchers were not able to detect a clear effect, which is very different from detecting that no effect exists: the review did not find that salt reduction was not beneficial; it found that there was not enough evidence to determine conclusively the effect of salt reduction on death and CVD events.
The pooled studies varied quite widely in terms of their size, length of follow-up and intervention methods. Although the researchers have taken steps to analyse how well these differing studies can be pooled, these variations are likely to have had an effect on the accuracy of the results.
The studies followed participants for between 6 and 36 months – a relatively short time to observe outcomes such as CVD risk, which is partly determined by health choices made over several years. Although two of the studies also performed additional analyses several years after their original research, five of the studies did not.
These researchers were focusing on dietary advice and interventions that reduce salt intake in individuals. There is a well-described relationship between population levels of salt and cardiovascular disease. Small changes in salt at a population level are likely to have an effect on blood pressure and the link between blood pressure and cardiovascular disease is strong. This suggests that although the effect is small, large numbers of people might benefit from population approaches to reducing salt.

As the authors of this review have noted, further research into the long-term effects of a sustained low-salt diet would reduce the uncertainty surrounding the effects of dietary interventions to reduce salt intake.