Letters

Long term effects of advice to reduce dietary salt

Front cover was highly misleading

Critical faculties should always be exercised

Salt needs to be reduced in manufacturing and processing food

Authors' reply

Front cover was highly misleading

EDITORThat small reductions in salt intake (2 g/day) have a small but significant effect on blood pressure is hardly surprising.¹ Nevertheless, in populations this would have a large effect on reducing strokes, heart attacks, and heart failure.

Hooper et al do not ask why reducing salt intake in the long term is so difficult. They claim that the interventions used were intensive, but most studies gave no details about what advice was offered. Furthermore, 75% of salt intake comes from processed food.² This needs to be avoided or contain less salt. None of the studies provided reduced salt foods.

Interpreting the study by Hooper et al is not helped by the editor writing the front cover of the BMJ, who seems to have read a different paper and misinterpreted the important positive findings. The confusion is increased by the authors' press release,³ which rightly blames the difficulty in reducing salt intake squarely on the food industry.

This confusion is compounded by errors in the meta-analysis. For example, the 18 month TOPH trial (phase I) was included as an intervention trial over "60 months," but salt intake was reduced for only 18 months, after which all participants returned to their normal diet. References were misquoted, and the correspondence following these papers was ignored. The totality of evidence for reducing salt is stronger than for any other non-pharmacological treatment.

Ninety five per cent of the population are at risk of developing cardiovascular disease,⁴ and 40% die from it. There are no controlled trials showing a reduction in mortality on stopping smoking, reducing fat intake alone (without fish oil supplements), reducing salt intake, losing weight, increasing fruit and vegetable consumption, or increasing exercise. For most of these factors no attempt has been made to conduct long term trials, owing to the innate difficulty of conducting and funding such trials and, now, the ethics of randomly putting a group of people on a high salt diet for the rest of their lives. The question that Hooper et al need to consider is what strength of evidence is needed to give dietary and lifestyle advice to try to prevent cardiovascular disease.

The study indicates the importance of reducing salt intake in the population, even by small amounts, particularly in treating high blood pressure.⁵ The BMJ should publish a retraction of its misleading front cover and read the authors' press release.

Graham A MacGregor, professor of cardiovascular medicine. 
Feng J He, cardiovascular research fellow. 
Blood Pressure Unit, St George's Hospital Medical School, London SW17 0RE

Competing interests: None declared.

1.	Hooper L, Bartlett C, Davey Smith G, Ebrahim S. Systematic review of long term effects of advice to reduce dietary salt in adults. BMJ 2002; 325: 628-632[Abstract/Free Full Text]. (21 September.)
2.	Nestle M. Food politicsHow the food industry influences nutrition and health. In: London: University of California Press, 2002.
3.	University of Bristol. New ways of reducing salt intake needed to make a long-term impact on blood pressure. Media release, 20 September 2002. http://bris.ac.uk/Depts/Info-Office/news/archive/salt.htm (accessed 3 Dec 2002).
4.	Beaglehole R. Global cardiovascular disease prevention: time to get serious. Lancet 2001; 358: 661-663[CrossRef][ISI][Medline].
5.	MacGregor GA, Markandu ND, Singer DRJ, Cappuccio FP, Shore AC, Sagnella GA. Moderate sodium restriction with angiotensin converting enzyme inhibitor in essential hypertension: a double blind study. BMJ 1987; 294: 531-534[ISI][Medline].

Critical faculties should always be exercised

EDITORThe paper by Hooper et al on the long term effects of advice to reduce salt intake in adults adds nothing new to the literature.¹ Substantial evidence accumulated over several decades shows that reducing salt intake lowers blood pressure.²

It has also been clear for many years that advice targeted at individuals will not produce substantial and sustained reductions in salt intake as most salt in the diet is added by the food industry to processed food such as bread, cooked meat, and breakfast cereals.² Data on mortality and cardiovascular events from sodium restriction trials are indeed limited, an important issue that has been highlighted repeatedly in the literature in recent years.

The discussion section of the paper by Hooper et al has elements of spin worthy of tabloid journalism, with selective and uncritical citation of relevant papers and a lack of context. The arguments seem largely based on a simplistic, individually based model of health promotion. Only cursory reference is made to the fact that dietary salt restriction is a population health issue that needs to be tackled in populations, by both regulation and collaborative work with the food industry.

The authors raise the spectre of possible harm from sodium restriction, raising the possibility of adverse effects on cardiovascular disease and all cause mortality. This speculation, which goes well beyond the clinical trial data, is largely based on two papers by Alderman et al that are widely regarded as methodologically flawed and have been extensively criticised in correspondence and reviews.^2-4 Hooper et al do not cite the paper by Tuomilehto et al, which links higher dietary salt intake with increased risk of cardiovascular events and increased mortality.⁵ Given that the current high dietary salt intake among children and adults can largely be attributed to salt added to processed food at concentrations well in excess of physiological requirements, the notion that efforts to achieve modest reductions in salt intake will have adverse effects on health is implausible to say the least. Meta-analysis is a powerful tool, but it does not absolve its practitioners from the need to exercise their critical faculties.

Ivan J Perry, professor of public health. 
Department of Epidemiology and Public Health, Distillery House, University College Cork i.perry{at}ucc.ie

Competing interests: None declared.

Salt needs to be reduced in manufacturing and processing food

EDITORHooper et al in their meta-analysis of randomised trials of individual dietary advice to reduce salt intake conclude that such intervention will have little effect on health.¹ They do not satisfactorily distinguish whether salt reduction itself confers only a small benefit or a large one, but people do not materially reduce their salt intake. As a result readers may conclude from the paper that reducing salt intake is unimportant.

This is not so. Reducing the current average salt consumption in Britain by 3 g/day (about one third) would reduce average blood pressure by about 5 mm Hg systolic in people over 50 and thereby reduce the incidence of heart attack and strokes by about 15% and 22% respectively.² A reduction of 6 g/day would reduce blood pressure by about twice as much with a corresponding additional reduction in the incidence of heart attacks and stroke. Reducing salt intake generally would thus have a major impact in the prevention of cardiovascular disease.

The obstacle to prevention is that nearly all the salt we eat is hidden, added to many foods in manufacturing and processing. Only about 15% is discretionary in that an individual can alter his or her intake through their own cooking and addition at meals. It is not therefore surprising that trials of advising people to reduce salt intake have little effect.

When salt intake is reduced blood pressure falls. Trials that show this best were not included in the meta-analysis of Hooper et al. They were trials in which dietary advice was reinforced by the provision of low salt staple foods such as bread, a major contributor to hidden salt in the national diet.^3-5

While the effect of avoiding discretionary salt is small it is achievable and worth while. Unfortunately it will have been underestimated in the analysis of Hooper et al because the trial participants included people who had already taken steps to avoid using discretionary salt, thereby diluting the effect.

The analysis of these trials by Hooper et al and the conclusions drawn are uninformative other than confirming the observation that little is gained by individual dietary advice. The public health challenge is to reduce salt used in the manufacturing and processing of food. Over 10-15 years, salt intake could be reduced by two thirds. This would cause no untoward effects and confer substantial health benefits.

M R Law, professor. 
m.r.law{at}qmul.ac.uk

N J Wald, professor. 
Department of Environmental and Preventive Medicine, Wolfson Institute of Preventive Medicine, Barts and the London, Queen Mary's School of Medicine and Dentistry, Queen Mary, University of London, London EC1M 6BQ

Competing interests: None declared.

Authors' reply

EDITORWe asked, "What are the long term effects on health and blood pressure of advice to reduce dietary salt intake?" and not, as commentators seem to imagine, "Can salt reduction lower blood pressure?" or "What would be the effect of reducing salt in processed foods?" We showed that advice does reduce urinary sodium excretion by about a quarter and this produces a 1 mm Hg fall in systolic blood pressure at 13-60 months.

Contrary to MacGregor and He's assertion, interventions provided by four studies (including 3007 of the 3514 participants) were well documented and highly intensive. It was not an error to use the 60 month outcomes of the TOHP phase I trial: although its 18 month intervention period had ended, there was no indication that all participants had returned to their normal diet. The point of such intensive intervention is precisely to encourage lifelong dietary change, and the authors clearly felt this was the case as they followed up participants to 60 months.

Potential harms of a reduced sodium diet do need discussion. Raised concentrations of low density lipoprotein cholesterol were highlighted in Graudal et al's systematic review,¹ and the evidence on mortality from three large cohort studies should not be dismissed as inconvenient. We cited the paper by Tuomilehto et al, which showed protective effects of low salt diets, to give a balanced account of the debate and draw attention to inconsistencies in the evidence.

We excluded short duration trials of salt restriction because these are not relevant to the question we posed and some may not be generalisable. At least seven trials conducted by MacGregor's group have produced mean blood pressure reductions that are greater than the upper 95% confidence interval of the effects found in meta-analysis of over 50 trials of salt restriction.² The reasons for such wide divergence remain of interest and have not been adequately explained.³

Law and Wald's estimate of the effect of salt restriction on blood pressure is extremely optimistic compared with other systematic reviews and has been more often cited (table).⁴ Their meta-analysis, which included 78 studies of salt restriction, only 10 of which were randomised, uses its own methodology.⁴ Effects about an order of magnitude greater than those reported by other meta-analyses were found.

Reduced sodium foods may be helpful. Only one of the studies in our review provided low salt foods for its intervention group throughout and recorded large reductions in blood pressure, but, as antihypertensive drugs also altered during the study, interpretation is difficult.

Lee Hooper, lecturer in evidence based care and systematic review. 
MANDEC, University Dental Hospital of Manchester, Manchester M15 6FH lee.hooper{at}man.ac.uk

Christopher Bartlett, research associate in health services research. 
MRC Health Services Research Collaboration

George Davey Smith, professor of clinical epidemiology. 
Shah Ebrahim, professor in epidemiology of ageing. 
Department of Social Medicine, University of Bristol, Bristol BS8 2PR

Competing interests: LH owns 285 shares in West Indies Rum Distillery, Barbados.