On bias in medical research; reflections on present salt-cholesterol controversies

Abstract

A common form of biased reporting is to avoid quoting evidence that contradicts the authors’ interpretation of scientific data. Against this background the recommendations by the Institute of Medicine of the National Academies of USA (IOM) regarding a population-based policy of salt restriction have been reviewed. It is suggested that the recommendations may not be valid since some important evidence has not been included in the IOM analysis.

Key words::

• Salt
• Salt restriction
• biased reporting
• cholesterol
• medical controversies

When it comes to prevention, and/or long-term handling, of such complex disorders of civilisation as primary (essential) hypertension (PH) and the metabolic syndrome (MS), biased opinions and premature conclusions begin to offer serious problems, as scrutinised in several recent surveys (1–4).

For example, despite multifactorial, and partly overlapping, natures of PH and MS (5–7), there is presently an unusually strong drive to consider salt (NaCl) and cholesterol as the real villains, worthy of being pursued at any cost. However, on closer scrutiny of present information, the weighed evidence for these assumed villain-ships is meagre or lacking – with reservation for some rare and special cases, or situations (4). Rather, both NaCl and cholesterol are of vital importance for a variety of biological processes (4–7) and should, in the great majority of cases, be considered as innocent co-travellers to other really serious contributors to both PH and MS (4–7).

Nevertheless the assumed two villains continue to be pursued with almost evangelic passions – and often in ways reflecting disrespect for opposing results and interpretations, or for influences of different nature. Thus, enormous resources are presently concentrated on what we eat, as well as to an intense search for drugs which, in this respect, are assumed to offer advantages. This, in turn, implies gigantic economical transfers between this and other fields of biomedical research, and between food-and-drug industries and in these topics involved research establishments – per se increasing risks for bias and misdirected information (1–4).

For example, the intense concentration on the intake issue is too often combined with a remarkable neglect for two other major health threats in today's mentally-hectic though physically-sedentary life. After all the present Brave New World is biologically artificial in so many respects, when one considers mankind's genetically determined mental and physical constitution. Along with many thousands of generations Homo sapiens has become selected the hard way to suit a hunter-gatherer small-group existence. Further, supported by a multivore constitution, roaming the globe in search of food and shelter, our species has managed to thrive on the strangest of diets (6,7). However, these biological facts, deeply rooted in our genetic framework, have at least two major consequences for present-day generations, though too often disregarded in the present search for assumed, and more easily caught villains in what we eat, and perhaps drink.

First, modern, mentally hectic life offers almost daily elements of arousal, mental strain and/or frustration – though no longer coped with by means of appropriate physical actions (which in primitive life are the natural solution). This, in turn, leads to more or less intense, and more prolonged, engagements of the age-old defence reaction and defeat reaction, designed to cope with various physical challenges – though not necessarily with present-day mental ones. Their neuro-endocrine expressions are dominated by the sympatho-adreno-medullary axis and the hypophyseal-adrenoglucocorticoid axis, respectively, though often more or less mixed along with shifting mental states.

It is, however, in ordinary life easily forgotten that their respective neuroendocrine influences affect all bodily functions, from mental state down to molecular-genetic levels. Further, that such influences often affect health and well-being far more than dietary factors. One of many examples, mental stress can elevate plasma cholesterol levels much more than what, for example, a prompt consumption eight eggs would do (4). However, one meets with a technical problem in this context, because the naturally phasic and rapidly undulating neuro-endocrine influences are far more difficult to adequately document than are, for example daily salt intakes or plasma cholesterol levels. Consequently they usually are more or less neglected in mass screenings of diet-influences – and conclusions often suffer.

Secondly, in most mass screenings of dietary influences, the many biological effects of physical exercise are difficult to control, simply because this factor differs considerably, not only between individuals but also with season, weather, and time or space available for exercise. However, exercise-related influences on health more often than not exceed by far those related to diet per se. They affect not only metabolism and organ function-and-structure, but also – via, for example central endorphin release – mental state, sympathetic activity, immune functions, etc, (4,6,7). It is easy to understand that bias and misinterpretations of results can be serious, indeed, when such powerful though highly variable influences come as unforeseen disturbers in dietary studies.

Thirdly, as in dietary and salt-restriction studies the resulting changes are usually fairly marginal, placebo effects, (i.e. psycho-physiologically induced endorphin effects – see above) must be particularly considered: They may have substantial effects on quite a few of the parameters used to detect, for example, the influence of fat-reduced diet, or of salt restriction. More often than not participants in such studies are aware of whether they are targets or controls, and placebo effects alone can, for example, often lower blood pressure more than a salt-restricted diet can.

Epidemiology and statistics are most important tools in biomedical research, but when several complex variables are at hand – and particularly if one of them is of special interest to investigators – great harm can ensue. Not for nothing, Mark Twain once had some salty remarks about complicated statistical problems – and events discussed in references (1–4) certainly put emphasis to his classic comment.

There is, however, in this context yet another form of bias in biomedical research, which is not always so easy to see, but which can have just as serious consequences as what is exemplified above: Not seldom, studies or reviews show a remarkable ability to avoid quoting reports, in which results and/or conclusions do not concur with supposedly accepted truths. Of course, in the present avalanche of scientific publications anyone can miss, or misunderstand, per se important findings – but if deliberate and put in system, it certainly collides with, for example, Karl Popper's ‘rules in science’ as outlined in his autobiography ‘Unended Quest’.

An example from the ‘Salt Arena’ will here be given and it is, in a way particularly remarkable as it emerges from the Institute of Medicine of the US National Academies, in a recent 220 page book (8) – though undoubtedly the best of intentions dominate thinking and conclusions. A committee and a selected panel of authors here summarise in six major chapters plus appendices their views concerning what is the proper salt intake for the US population. When one screens the several hundreds of references, however, there is striking lack of mention of some, in this context, really important data, studies or reviews, which do not consider present-day 8–12 g NaCl daily-intake as any real health threat. Rather, these not mentioned studies – further discussed below – indicate that this level of intake is genetically-determined, and suited for handling risks at both ends of the intake spectrum. In fact, hardly any type of intake regulation is so well-guarded and so powerful as that controlling salt intake and NaCl metabolism in higher organisms (9). After all, NaCl is the electrolyte skeleton of the plasma and interstitial-fluid volumes, and the Na⁺ and the Cl² ions are key elements in the function of cell membranes.

There is, however, in (8) no mention of the JAMA article by Graudal et al. (10), where a meta-analysis of no less than 58 careful studies is presented, which describe the extent of blood pressure reduction induced by substantial restrictions of NaCl intake. In normotensives mean arterial pressure (MAP) was reduced by a tiny 0.6 mmHg, and in hypertensives with a humble 3 mmHg. Further, these meagre MAP effects were associated with an accentuated sympathetic activity and a mobilisation of the renin-angiotensin-aldosteron axis already in the ‘resting-steady-state’. Thus – as is also evident from the diagrams of Laragh's group in 1972 (11) – such mobilisations occur already when daily NaCl intake is lowered to 6–8 g. But no mention of these data can be traced from the reference list in (8).

Nor is it mentioned that Korner's group in Melbourne (12) observed how three 40-min exercise-bouts per week led to fairly sustained 6–7 mmHg MAP reductions in normotensives, while in hypertensives the MAP reductions matched those induced by conventional anti-hypertensive drugs. To this comes, of course, the many other advantages of exercise, described above – and which no drug can offer… On top of this, exercise increases appetite, and thereby food (and salt) intake, but nevertheless the exercise-induced MAP reduction was some 10 times larger than that caused by salt restriction alone in normotensives, according to Graudal et al. (10).

In addition, no mention is made in (8) of the data and considerations presented, and referred to, by Alderman et al. (13,14) or by McCarron et al. (15). Both groups also mention the indeed meagre evidence for the views in (8) that a reduction of NaCl intake to some 6 g daily should prevent, or substantially reduce the occurrence of hypertension. As stressed by McCarron et al. (15), biologically determined regulatory processes, safeguarding homeostasis, should not be overruled by ever so devoted committees – or still less by political decrees. In fact, the nowadays common 8–12 g daily salt intake in man coincides closely with what, for example, omnivore rats or herbivore sheep choose at free access to salt, when corrections are made for size and metabolic rate (16).

Not even in patients with imminent cardiac failure, where anyone could believe that salt restriction would be helpful, this common belief has been put in doubt. In a recent clinical study from Italy (17), it was observed how patients with cardiac failure had clinically better outcome when on ordinary salt intake than when put on low sodium diet. The authors state already in the paper title: “….is sodium an old enemy or a new friend?” But no mention in (8) of these interesting findings. It is instead concluded that daily NaCl intakes in adults should be lowered to 6 g – according to some in the panel to 3–4 g. But little mention about how this could be achieved, and whether it in the end would have much effect, considering the contrary odds outlined above, concerning observations and data by Graudal et al., Alderman et al., McCarron et al., etc.

Even more curious, in Table 2–5 (and Figure 2–6) in (8) the sodium intake 1971–1974 was a mere 2.2 g, while the prevalence of hypertension was 39.8%, but in 1988–1994 the sodium intake had increased to 3.6 g (+ 64%) while the prevalence of hypertension had become reduced to 25.5% (-36%). Whatever this indicates – (and to what extent the tabled data are reliable) – it hardly lends support to the suggestion (e.g. p. 118 in (8)), that daily sodium intake should be reduced to 2.3 g, and preferably, to even less (according to some mathematical calculations of debatable realism).

Moreover, almost nowhere in (8) – or, for that matter, in other low salt enterprises – is it seriously discussed whether the suggested low-salt intakes could imply any risks. In general, it is – for example from a physiological point of view – remarkable how such risks are rarely considered in these types of studies in man, and how seldom heart rates (HR) are mentioned, even though HR is more easily measured than blood pressure – and, often enough, offers important information in this context.

Actually, changes of resting HR during salt restriction tells a lot about, for example reflex counter-regulation, and about what happens with the long-term load on left cardiac ventricle and on systemic arteries. The reason is that these types of load well correspond to the product of heart rate and systolic (or mean) blood pressure (i.e. HR × SBP). More often than not this product increases during salt restriction, as the percent HR increase often exceeds the percent SBP decrease (16). In addition, the for left ventricular coronary blood supply so important diastolic period is shortened when HR increases Moreover an intensified sympathetic activity increases myocardial oxygen consumption. But no mention of such potential complications in salt restrictions, nor of Alderman et al.'s (13,14) observations of increased death rates when cardiac patients were put on salt restriction.

Further, little or no interest is paid to the – after all not unimportant ‘quality-of-life’ – not to mention that salt restriction opens up for other types of risk, than those described above. This is so particularly in elderly people where food (and hence salt) intake begins to decline, and so does also the efficiency of reflex cardiovascular adjustments. For example, how often may salt restriction invite to cardiovascular collapse at, ordinarily trivial, gastrointestinal infections, or merely by sweating on a hot day? Or, simply by fainting spells at sudden changes of body position, perhaps ending in fractures and prolonged invalidisation? However, this side of the coin is hardly ever mentioned by supporters of salt restriction, still less ascribed to its biological consequences. But every stroke, or related cardiovascular incident, is meticulously noted as a result of high salt intake. Again an element of bias easily comes in – but can be very difficult to detect.

The many problems inherent in the cholesterol issue are amply discussed in (4), but to sum-up those inherent in the NaCl issue, the statements below may be justified. It is easy, indeed, to understand the sincere wish of the US Institute of Medicine Committee (8) to explore all means to reduce strokes and other cardiovascular insults in elderly people, and – if possible – to prevent hypertension. Further, it can often be worthwhile in cardiovascular disorders to test whether salt restriction may have beneficial effects. But, at the present state of knowledge – weighing available odds for and against –it is, indeed, questionable to argue for generalised salt restriction, independent of individual health, age, sex, etc.

There is, after all, abundant evidence (10,13–16) that the intrinsic salt appetite in mammals – and hence in Homo sapiens as well – is firmly set so as to best handle risks at both ends of the spectrum. There is also strong evidence that the all-important NaCl balance is safe-guarded by a uniquely sophisticated, and powerful, control system (9). Moreover, when it comes to adult man, this ends-up in a daily NaCl intake of 8–12 g (10–16), where the inherent salt appetite promptly increases on salt loss and decreases on occasional intake excesses. This, of course, by no means denies that real excesses of any intake have their risks – sudden gross intakes of innocent H₂O can have serious consequences, indeed…

Such considerations, of course, do not deny that there are situations, and individuals, where salt restriction can be justified, but then on an individual basis and under proper control. Nor does it deny, for example that preprocessed food should not unnecessarily be salt-enriched, or that NaCl contents – like those of other ingredients – should be declared. The reason is simply that it is nearly impossible, for example salt-sensitives to eliminate NaCl – but it is only too easy to add it according to individual salt appetite levels. But such common-sense rules are quite another matter than campaigns to more-or-less forcefully lower NaCl intakes for whole populations…

Declaration of interest: The author reports no conflicts of interest. The author alone is responsible for the content and writing of the paper.