Intended for healthcare professionals

Letters

Intersalt data

BMJ 1997; 315 doi: https://doi.org/10.1136/bmj.315.7106.484 (Published 23 August 1997) Cite this as: BMJ 1997;315:484

We received many letters commenting on the cluster of papers on salt that we published in May last year. We had several problems with the letters but have now resolved these and are publishing the letters now. These 10 letters and three others are available on the BMJ's website (http://www.bmj.com/).

Cross cultural studies such as Intersalt study cannot be used to infer causality

  1. James Le Fanu, General practitionera
  1. a Mawbey Brough Health Centre, London SW8 2UD
  2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
  3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
  4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
  5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
  6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
  7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
  8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
  9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
  10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
  11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
  12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
  13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
  14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
  15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
  16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

    Editor—The practical issue of whether moderate dietary salt restriction can lower blood pressure was resolved in the classic Glyncorrwg community study in south Wales, which found it to be ineffective.1 Clearly then the claims of the Intersalt investigators that their findings support “the reduction of salt intake to control adverse blood pressure levels” must be wrong—and the reasons are not hard to find.2

    Firstly, cross cultural studies such as the Intersalt study cannot be used to infer causality as they are based on the false assumption that the populations of diverse societies—irrespective of their genetic and cultural composition—have the same susceptibility to environmental factors. Indeed, it is this very difference in susceptibility that explains why it is not possible to show a cross cultural correlation between smoking and lung cancer. For both the countries of northern Europe and of the Mediterranean there is a clear dose-response relation (confirming causality), but the adverse effects of smoking are much more pronounced in the countries of northern Europe.3

    The multiplicity of confounding variables in cross cultural studies makes it almost inevitable that these studies will fail to reflect genuine cause and effect relations (such as smoking and lung cancer), while the associations they do identify are likely to be spurious. Hence the only epidemiological method for determining whether salt intake is implicated in raised blood pressure is to use within population studies, in which the problem of differing susceptibilities does not arise. Here the results of the Intersalt study are both trivial and contradictory: a fall of 1 mm Hg in diastolic pressure for every extra 100 mmol sodium excretion in 24 hours for men and women between the ages of 20 and 39, and a change of similar magnitude but in the reverse direction for those between the ages of 40 and 59.

    This leaves only the clinical studies of the effects of salt reduction on blood pressure to be considered. Malcolm Law's assertion, in his commentary, that these studies show that “dietary salt intake is a serious health hazard” is based on a reference to his own meta-analysis,4 whose methods, he fails to point out, have been subjected to critical scrutiny and conclusions rejected.5

    The important question that emerges from these papers is why the combined intellects of so many distinguished epidemiologists should maintain that the evidence incriminating salt in hypertension is so convincing when clearly it adds up to very little. Readers of the BMJ may rightly wonder how many of the other epidemiological discov-eries of recent years linking diet with disease are similarly insecure.

    References

    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.

    Data linking sodium intake to subsequent morbid and fatal outcomes must be studied

    1. Michael Alderman, President, American Society of Hypertensionb
    1. a Mawbey Brough Health Centre, London SW8 2UD
    2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
    3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
    4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
    5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
    6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
    7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
    8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
    9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
    10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
    11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
    12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
    13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
    14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
    15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
    16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

      Editor—Papers and editorials in the issue of 18 May 1996 addressed the relation of dietary sodium intake and blood pressure. The heart of the matter is whether a low sodium diet reduces blood pressure in populations. The BMJ printed a more and JAMA a less enthusiastic view of the magnitude of this relation.1 Several commentators in the BMJ then made the unjustified leap of converting what is a scientific controversy about a physiological mechanism into a public health issue.

      Public health recommendations must be based on proof of safety and benefit. Even if a low sodium diet could lower the blood pressure of most people (probably not true) and both the diet and the change in blood pressure could be sustained (not established), this alone would not justify a recommendation to reduce sodium intake.

      For such advice to be responsibly given there must be evidence that the change will improve and not impair health. While the advantage of a lower blood pressure, at any level, is well established,2 it is not true that every method to lower blood pressure would necessarily improve health. Some techniques to lower blood pressure, like giving short acting calcium antagonists,3 may not be safe.

      All interventions aimed at enhancing or extending life by manipulating a single mechanism inevitably produce a variety of effects, some of which may not be advantageous. Extrapolation from mechanistic thinking demands evidence that the sum total of all the effects of the intervention—and not just one, such as lowering blood pressure—will help and not harm; and particularly here since the target is the whole population.

      A low sodium intake produces many effects, not all of which are salutary.4 The integrated impact of these effects remains to be established. The scanty evidence directly linking sodium intake to morbidity and mortality is not encouraging.5

      Unfortunately, we simply do not know whether a universal change in sodium consumption will cause benefit or harm. Insufficient evidence—for good or ill—is not a sturdy basis for making health policy. Gratuitous exhortation, reflecting the hopes of even the most well meaning authorities, is no substitute for data. Toward this end, a good start would be to collect and analyse further observational data linking sodium intake to subsequent morbid and fatal outcomes.

      References

      1. 1.
      2. 2.
      3. 3.
      4. 4.
      5. 5.

      Conclusions drawn in paper “revisiting” Intersalt data are of questionable validity

      1. Alexander Macnair, Private consultant in preventive medicine*c
      1. a Mawbey Brough Health Centre, London SW8 2UD
      2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
      3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
      4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
      5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
      6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
      7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
      8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
      9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
      10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
      11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
      12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
      13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
      14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
      15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
      16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

        Editor—Despite editorial assertions to the contrary,1 the conclusions drawn in the paper by Paul Elliott and colleagues “revisiting” the data from the Intersalt study are of questionable validity.2 The original report showed that 24 hour urinary sodium excretion was not significantly related to the prevalence of hypertension.3 This second reanalysis depends on two statistical manoeuvres to draw the opposite conclusion. The first manoeuvre is to correct for regression dilution bias. The authors have assumed that the differences between true mean blood pressures and the measured blood pressures were greater than those allowed for in the first analysis. Since they have no means of knowing what the true blood pressures were at the time of the Intersalt measurements, there is no valid basis for their assumption.

        Secondly, the effect of body mass index is removed from the multiple regression analysis. This was not an allowance for an assumed error but the withholding of an accurately measured variable which in the original report showed a strong and significant independent relation with blood pressure. The reason given for this startling decision is a supposed correlation between body weight and sodium excretion. There surely is no logical basis for believing that fat people eat more salt than thin people, and the exclusion of body mass index from the multiple regression analysis is without justification. Without all this statistical prestidigitation, the third visit to the Intersalt study still does not stand up.

        As an ad hoc adviser to several food manufacturing interests, including some that add salt to their products, I take exception to the suggestion that my views are necessarily more “commercial” than those of the academics. Surely the science speaks for itself.

        References

        1. 1.
        2. 2.
        3. 3.

        Epidemiological studies should be designed to reduce correction needed for measurement error to a minimum

        1. N E Day, Directord
        1. a Mawbey Brough Health Centre, London SW8 2UD
        2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
        3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
        4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
        5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
        6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
        7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
        8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
        9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
        10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
        11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
        12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
        13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
        14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
        15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
        16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

          Editor—George Davey Smith and Andrew N Phillips's critique1 of the most recent Intersalt paper,2 and the response of the Intersalt investigators in their commentary,1 raises an important issue. Measurement error can bias estimates of effect in epidemiology, and some correction is needed to remove the bias. The correction made, however, needs careful justification. The main problem relates to the error structure of the two variables one is attempting to associate and the degree of correlation between these measurement errors. The critical issue is the quantitative effect due to correlation between measurement errors.

          In the Intersalt study the outcome, blood pressure, is measured at the same time as urinary sodium excretion. Both measures are taken to represent the long term average value for that individual. Both will be measured with considerable error. Correlation between the two errors may not be negligible, due, for example, to parallel seasonal effects. For the interpopulation association, no problem should arise. For the intrapopulation association, however, the correction for measurement error will depend on the between error correlation.

          The Intersalt authors assume the correlation to be zero, and because urinary sodium has a large measurement error the correction they make is large. For systolic blood pressure, 1.6 mm Hg/100 μmol sodium intake/day is “corrected” to 4.3 mm Hg/100 μmol/day. This value is considered to be consistent with the between population regression coefficient of 7.1. (These values refer to the estimates adjusted for age and sex, over all age groups. Similar consideration would apply to the multivariate estimate.) If the two errors are correlated, however, the correction is overdone. Approximately, for values of the correlation of 0.1, 0.3, and 0.5 the proper corrected values should be 4.0, 3.5, and 2.9 respectively. If one considers a correlation of 0.5—for example, with parallel seasonal effects—to be credible then the agreement claimed by the Intersalt investigators between regression coefficients based on within and between population comparisons is substantially overstated. If one believes a value of 0.1 to be the largest likely to occur then the Intersalt conclusion is approximately correct. Given the dearth of information on what the correct value is likely to be, the large correction made in the Intersalt study seems to be inadequately justified. More generally,

          • corrections for measurement error are not free of assumption

          • when the correction is large the assumptions must be clearly stated and evidence presented in their support, perhaps with a sensitivity analysis

          • large corrections presented without such supporting evidence are unsound.

          Statistical complexity should not be used to conceal inadequacies of the data. Generating accurate quantitative data on the variance-covariance structure of measurement error is difficult. The clear message is that epidemiological studies should be designed to reduce to a minimum the correction needed for measurement error. This can be achieved by improving the measurement instruments, taking repeat measures, and choosing study populations to maximise the between individual variance. None of these are achieved by simply increasing the study size.

          References

          1. 1.
          2. 2.

          Correction for regression dilution bias in Intersalt study was misleading

          1. George Davey Smith, Professor of clinical epidemiologye,
          2. Andrew N Phillips, Reader in epidemiology and biostatisticsf
          1. a Mawbey Brough Health Centre, London SW8 2UD
          2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
          3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
          4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
          5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
          6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
          7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
          8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
          9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
          10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
          11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
          12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
          13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
          14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
          15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
          16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

            Editor—We were surprised by the response of Alan R Dyer and colleagues in their commentary on our article on correction for regression dilution bias in the Intersalt study.1 They fail to address the substantive points adequately while restating evidence on the association between dietary salt and blood pressure, which was not the concern of our article.

            A random 8% of participants in the Intersalt study had repeat 24 hour urine collections and blood pressure measurements taken on average three weeks after the initial measurement. The repeat 24 hour urine collections were used to estimate the degree to which single 24 hour collections were inadequate indicators of usual sodium excretion. The reliability coefficient calculated from these repeat measures was used to greatly inflate the estimated strength of association between sodium excretion and blood pressure. However, if changes in urinary sodium excretion and blood pressure between the first and the repeat measurements tend to coincide then the association between sodium excretion and blood pressure would not have been underestimated to the degree to which the correction method assumes. We cited papers that show that changes in sodium excretion and blood pressure do indeed coincide, for measurements taken one to two months apart. As N E Day shows [letter above], if the correlations between changes in sodium excretion and changes in blood pressure are even of moderate strength then the correction method used by the Intersalt team is highly misleading.

            Dyer and colleagues state that “to the best of our knowledge, there are no data showing physiological day to day parallel fluctuations of sodium or blood pressure.” They miss the point that fluctuations in sodium excretion and blood pressure do not need to coincide on a day to day basis to invalidate their method. If the fluctuations correspond over a three week period, which was the interval in their validation study, then this will render their correction method fallacious. What is most odd about their response is that the Intersalt investigators do indeed have the data that can test this point, as they measured both blood pressure and sodium excretion in their validation study. How can they then state that no such data exist? We ask that they report the correlations between the changes in sodium excretion and the changes in blood pressure over the three week period to allow other investigators to evaluate whether one of the central assumptions of their correction method is met.

            The Intersalt investigators seem not to appreciate the difficulties of correcting for measurement imprecision in the presence of confounders. Body mass index is only a proxy for the various aspects of body composition that confound the association between sodium excretion and blood pressure. The mathematical construct of weight divided by the square of height is clearly not a perfect measure of these. The appropriate reliability coefficient is that between body mass index and the relevant aspects of body composition, not between body mass index measured on two occasions. The importance of measurement error in confounders is, fortunately, well recognised by other investigators.2

            All of the conclusions of our paper remain unchallenged by the response by Dyer and colleagues; indeed, they are strengthened by the demonstration that a research team that uses these methods to essentially triple the magnitude of their association between sodium excretion and blood pressure does so without a clear appreciation of the implications of these correction methods. The use of epidemiology to inform public health policy would be furthered better by improving study design3 than by using obfuscating and potentially erroneous statistical “corrections,” the assumptions of which have not been adequately tested.

            References

            1. 1.
            2. 2.
            3. 3.

            Slow decremental change in dietary sodium load in whole populations is needed

            1. Graham Watt, Professorg,
            2. Julian Tudor Hart, Professorh
            1. a Mawbey Brough Health Centre, London SW8 2UD
            2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
            3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
            4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
            5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
            6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
            7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
            8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
            9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
            10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
            11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
            12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
            13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
            14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
            15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
            16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

              Editor—As authors of papers whose conclusions could be misused by the commercial salt interests,1 2 3 4 we support the professional consensus in favour of a decremental reduction in dietary sodium load in whole populations everywhere. In our experience, individuals who endured abrupt reductions in sodium load from about 150 mmol to about 60 mmol found their new diet almost intolerable. Our research team and their families shared the experience of our subjects for the first week and agreed with them entirely. An important, though unfortunately unmeasured, consequence was that all subjects added more fat to their food in an attempt to make it taste of something, so that an abrupt reduction in sodium intake can in these circumstances increase rather than reduce overall cardiovascular risk. Doctors who rely on instructing their patients to abjure salt forthwith are careful never to verify compliance by measuring sodium outputs.

              Graham A MacGregor and Peter S Sever emphasise the delay of about one month before taste adapts to a new dietary sodium intake, a period that probably varies between individuals.5 Given time for this adaption to take place, large reductions in dietary sodium may not be difficult; without it, a huge range of foods—all soups, breads, and most cheeses, as well as obvious items like Marmite, bacon, and kippers—disappear from one's diet. And when people have adapted and they can no longer stand the taste of what remains “normal” for everyone else, they can no longer find a reasonable choice of prepared foods, eat out in a restaurant, or dine with their friends.

              We need a slow, decremental, across the board approach, strictly regulated so that competing food manufacturers don't cheat; successive targets should be based on evidence from continuing market research on how, or even whether, changes are perceived by consumers (nobody seems to have noticed the 20% reduction in sodium in bread). Without regulation, sodium load will continue to rise, as food manufacturers move toward North American levels of taste deception.

              References

              1. 1.
              2. 2.
              3. 3.
              4. 4.
              5. 5.

              Science demands data sharing

              1. Keith Rennolls, Professor of applied statisticsi
              1. a Mawbey Brough Health Centre, London SW8 2UD
              2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
              3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
              4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
              5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
              6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
              7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
              8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
              9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
              10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
              11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
              12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
              13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
              14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
              15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
              16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

                Editor—Research indicating the influence of dietary salt on blood pressure is convincing.1 However, the size and scientific methodology of the Intersalt study give particular weight to its conclusions.2 That the Salt Institute should put a counter case is not surprising3 since the scientific debate takes place in a market economy.

                Accepted scientific practice demands that results should be reproducible by independent scientists. Financial considerations mean that it is not realistic to reproduce data comparable to those of the Intersalt study, but the conflict between the Intersalt Cooperative Research Group and the Salt Institute focuses on the reproducibility and validity of the statistical analysis of the Intersalt data.

                While it is reasonable for medical researchers to retain their data until they have published their primary analysis, the Intersalt Steering and Editorial Committee4 claims the data as “the confidential property of local investigators” and so forces dissenters into a limited framework of scientific discussion.

                Though I would not wish to support the analyses of Richard L Hanneman,3 I find I have as many doubts about the statistical analysis of the Intersalt group and the commentaries on Hanneman's analysis.3 Some of my qualms are:

                • the criss crossing of population trends mentioned by Hanneman, and described by Malcolm Law in his commentary as “bizarre” and “implausible,” is seen to occur if the trend lines are plotted

                • the claim by J Stamler and colleagues that the use of the intercept is statistically invalid3 is incorrect, and the biological argument used is spurious

                • the expected effect of a difference of 100 mmol in 24 hour sodium excretion (for example, 70 v 170 mmol) over an age range of 30 years (for example, 25 v 55) is problematic, since the magnitude of the standard error of this measure depends critically on the end values of the intervals that are chosen

                • has the Intersalt group taken into account the possible inhomogeneity of the variances about the regressions? Has the prediction theory been correctly based on the errors in regressor variables model?

                For the medical-scientific community to achieve a shared understanding it is necessary that confirmatory secondary analyses be carried out by other researchers. For these reasons, and the powerful ethical arguments made in Tony Delamothe's editorial,4 the sharing of medical research data is long overdue. A good model is provided by the Economic and Social Research Council's data archive, which ensures use only by bona fide researchers and that the primary researchers are appropriately acknowledged.

                References

                1. 1.
                2. 2.
                3. 3.
                4. 4.

                Reply for Intersalt Steering and Editorial Committee

                1. Paul Elliott, Professorj,
                2. Jeremiah Stamler, Professor emeritusk,
                3. Alan R Dyer, Professork,
                4. Rose Stamler, Professor emeritusk,
                5. Hugo Kesteloot, Professorl,
                6. Michael Marmot, Professorm
                1. a Mawbey Brough Health Centre, London SW8 2UD
                2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
                3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
                4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
                5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
                6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
                7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
                8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
                9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
                10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
                11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
                12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
                13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
                14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
                15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
                16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

                  Editor—James Le Fanu selects one trial on salt and blood pressure, ignoring others.1 He also cites only two (among 36) published Intersalt coefficients for individuals, whereas all 18 coefficients for sodium and systolic pressure and nine for sodium and diastolic pressure are significant.2

                  Michael Alderman cites a 1996 overview of trials but fails to note (a) its report of significant falls in the systolic pressure of non-hypertensive and hypertensive individuals with reduced sodium intake, (b) several criticisms of this overview, and (c) findings of larger reductions in blood pressure in other overviews.1 Contrary to what Alderman says, recommendations for moderate salt reduction in populations are studied judgments by expert groups, including two reports that he co-signed. When discussing safety, Alderman cites his two papers but fails to note critiques of them and inability to replicate his findings.3 4

                  Contrary to Alexander Macnair's comments, the 1988 analysis across 52 Intersalt population samples (prior hypothesis) showed that median urinary sodium excretion was significantly related to hypertension (P<0.01).

                  N E Day, George Davey Smith and Andrew N Phillips, and we agree that uncorrected coefficients underestimate the size of the association of sodium with blood pressure in individuals. We agree with Day that “corrections for measurement error are not free of assumptions”; hence we presented uncorrected and multivariate corrected estimates with and without body mass index2 and gave detailed methods and extensive sensitivity analyses.5

                  Davey Smith and Phillips call for the Intersalt correlations between change in sodium excretion and change in blood pressure for the 8% of people in whom repeat measurements were made. They were 0.15 for systolic pressure and approximately zero for diastolic pressure. These correlations are insufficient to make the corrections Davey Smith and Phillips discuss; estimates of within-person covariance and covariance reliability between sodium and blood pressure are needed. In the Intersalt study, obtaining such estimates validly was not possible due to 52 small population samples (200 men and women in eight age-sex groups); some extreme differences between first and repeat blood pressures; and a fall in mean blood pressure of 3.3/1.6 mm Hg over the three weeks between measurements,5 invalidating the usual assumption of a constant mean.

                  Other studies tell us whether correlated fluctuations exist over a period of around three weeks. Recent analyses from the trials of hypertension prevention found—as Intersalt assumed—little or no short term covariation of sodium and blood pressure (N Cook, fourth international conference on preventive cardiology, Montreal, 2 July 1997). Such data, casting further doubt on assertions from Davey Smith and Phillips, indicate that the Intersalt corrected estimates—a sodium intake of 100 mmol less/day influences systolic pressure of individuals by -3 to -6 mm Hg and diastolic pressure by 0 to -3 mm Hg—are scientifically sound, although likely still to be underestimates.

                  In the context of extensive concordant data from all research disciplines supporting the conclusion that high dietary salt intake has an important role in causing population-wide adverse blood pressure levels, the Intersalt data—both cross population and within population—have profound meaning for public policy, as has been widely recognised by unbiased expert groups.

                  We agree with Graham Watt and Julian Tudor Hart that food manufacturers should gradually reduce salt in their products, with clear food labelling.

                  A response to Keith Rennolls's letter and an extended version of this letter are available on the BMJ website.

                  Footnotes

                  • On behalf of the Intersalt Cooperative Research Group

                  References

                  1. 1.
                  2. 2.
                  3. 3.
                  4. 4.
                  5. 5.

                  Collaborative efforts must be made to reduce sodium in diet

                  1. Richard H Grimm, Directorn
                  1. a Mawbey Brough Health Centre, London SW8 2UD
                  2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
                  3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
                  4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
                  5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
                  6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
                  7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
                  8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
                  9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
                  10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
                  11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
                  12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
                  13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
                  14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
                  15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
                  16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

                    Editor—The article by Richard L Hanneman was entertaining in its demonstration of the lengths to which the food industry, specifically the Salt Institute, will go to obfuscate and confuse the question of salt's relation to blood pressure.1 The criticisms in the article related to the Intersalt study2 have been extensively covered, and it would be redundant to review these again. However, the important and extensive area of evidence from clinical trials has largely been ignored by Hanneman and the Salt Institute.

                    The Shapiro Center for Evidence-Based Medicine has been working in clinical trials of sodium reduction and blood pressure for 20 years. Earlier trials on sodium reduction combined with other lifestyle changes (weight loss, alcohol reduction) showed the usefulness of sodium reduction in removing and minimising the need for blood pressure drugs in hypertensive patients who had previously been treated for years with such drugs.3 More recently, the centre has conducted double blind crossover trials of sodium supplementation in white and black normotensive subjects, which showed that reducing sodium intake and giving back sodium in levels similar to average daily sodium intake will significantly influence blood pressure.4 Other studies have confirmed these results.5 Dietary sodium intake is clearly a major determinant of blood pressure based on epidemiological data (that is, data from the Intersalt study) confirmed by prospective, randomised, placebo controlled trials.

                    Let's stop debating what for some time has been ridiculously obvious; let's move on to working collaboratively to reduce the sodium in the diet (which is ingested largely in the form of processed foods and fast foods) so that the population at large will benefit.

                    References

                    1. 1.
                    2. 2.
                    3. 3.
                    4. 4.
                    5. 5.

                    Sodium contents of restaurant foods in United States are high

                    1. Bonnie F Liebman, Director of nutritiono,
                    2. Michael F Jacobson, Executive directoro
                    1. a Mawbey Brough Health Centre, London SW8 2UD
                    2. b Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461-1602, USA
                    3. c Cae'n y Bwlch Isaf, Talsamau, Gwynedd LL47 6YB
                    4. *Alexander Macnair acts as a specialist adviser to several food manufacturers that add salt to their products.
                    5. d MRC Biostatistics Unit, Institute of Public Health, Cambridge CB2 2SR
                    6. e University of Bristol, Department of Social Medicine, Bristol BS8 2PR
                    7. f University Department of Primary Care and Population Sciences, Royal Free Hospital and School of Medicine, London NW3 2PF
                    8. g University Department of General Practice, Woodside Health Centre, Glasgow G20 7LR
                    9. h International Section, Department of Primary Health Care, Royal Free Hospital Medical School, London NW3 2QU
                    10. i School of Computing and Mathematical Sciences, University of Greenwich, London SE18 9PF
                    11. j Department of Epidemiology and Public Health, Imperial College School of Medicine, St Mary's Hospital, London W2 1PG
                    12. k Department of Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611-4402, USA
                    13. l Department of Epidemiology, St Raphael University Hospital, Leuven, Belgium
                    14. m Department of Epidemiology and Public Health, University College London Medical School, London WC1E
                    15. n Shapiro Center for Evidence-Based Medicine, Hennepin County Medical Center, Minneapolis Medical Research Foundation, 914 South 8th Street, D-2, Minneapolis, MN 55404, USA
                    16. o Center for Science in the Public Interest, 1875 Connecticut Avenue NW, Washington DC, 20009-6728, USA

                      Editor—The many articles on salt and hypertension in the BMJ of 18 May 1996—including Fiona Godlee's insightful commentary on the salt industry1—were excellent. In the United States the salt and food industries continue to undermine advice by health authorities to reduce salt intake. For example, on 21 May 1996 JAMA issued a press release with the headline “Reducing salt in diet has little effect on blood pressure.” The release neglected to mention that the meta-analysis, funded by the Campbell Soup Company's Institute for Research and Technology, found a significant drop in systolic blood pressure before the authors excluded the trials in which subjects were fed diets controlled by institutions.2

                      The snowballing effect of such reports is insidious. Two weeks after JAMA's press release the New York Times ran an article declaring that, “contrary to previous advice, recent studies indicate that salt is not the villain it was once said to be.” 3 The article later stated: “doctors still caution people who suffer from hypertension—estimated to be 10 percent of the population—to cut way back on salt.” In fact, 25% of American adults (50-60% of adults aged 60 or older) have hypertension. The only “authority” quoted in the article was Richard Hanneman, president of the Salt Institute.

                      The United States has made some progress in reducing sodium levels in packaged foods. Our surveys indicate that these levels seem to be dropping at a rate of just over 1% a year.4 However, Americans now spend nearly half of their food dollars on—and obtain a third of their energy from—foods eaten out of the home. There, sodium levels remain undisclosed and disturbingly high. The 1 shows the sodium contents that our nutritional analyses of restaurant foods (composites of nine to 12 samples purchased at mid-priced restaurants in at least three cities) have yielded for various selected entrees and meals.

                      Table 1

                      Sodium content of various foods sold at restaurants in United States

                      View this table:

                      These findings indicate that many Americans are exceeding, in a single meal, the 2400 mg of sodium recommended for an entire day. Only with the continued efforts of health authorities will the sodium content of restaurant foods and processed foods fall to safer levels.

                      References

                      1. 1.
                      2. 2.
                      3. 3.
                      4. 4.
                      View Abstract