■ site map
BLOG: March 2010
Hypertension and salt war
It is silent, but deadly: often without warning signs, high blood pressure (hypertension) kills well over 50,000 Americans each year. Moreover, it is among causative factors in another 300,000 deaths, or so. In short, it is an enemy with long and terrifying record. Two major governmental health organizations - Centers for Disease Control (CDC) and Institute of Medicine (IOM) - decided that it warrants declaring war on salt, which is for a long time linked to high blood pressure.
More so considering that hypertension-related deaths are rising at an alarming rate - according to the American Heart Association, nearly 20% from 1996 to 2006.
A few recent studies landed support to this anti-hypertension uprising, all coming up with some impressive figures of lives and money saved with even moderate reduction in salt intake.
Here's the battle plan.
In its report from February this year (A Population-Based Policy and Systems Change Approach to Prevent and Control Hypertension), IOM identifies the main targets. Through proposed active collaboration with state and local health organizations, IOM wants people to know about the main risk factors for hypertension:
▶ too much salt and too little potassium
▶ overweight/obesity, and
▶ too little physical activity
IOM also wants better implementation of the JNC (Joint National Committee) guidelines in the treatment of high blood pressure by physicians
Well, not much new under the sun. The first three factors primarily target prevention, and the last one treatment quality, although all four are inevitably interconnected.
While it is high salt intake that usually gets all the attention, IOM makes important point of the excess salt is likely to become health problem only
when accompanied with low potassium intake.
Knowing that we get potassium mainly from plant foods takes us back to the sore point of American diet: too much meat and processed foods, too few vegetables. It is not getting any better in this respect; in fact, it's getting worse.
And so do, according to recent studies, the rates of overweight, obese and physically inactive Americans. That doesn't leave much of a "choice" to the rates of hypertension but to go up, and points to the main culprits of this and many other health problems:
unhealthy diet and lifestyle.
Even with hypertension being only a small fraction of the price that Americans are paying for it, it is affecting some 73 millions of Americans with hypertension, with about one in every six adult deaths related to it.
But it is all too known how hard it is to change lifestyle habits of the population. So the overall focus is shifting toward seemingly easier target: reduction in salt intake. This course of action was particularly encouraged by the recent claimed UK success in reducing salt intake of the general population by an estimated 9.5% over the past five years.
While there is no data yet of any health benefits directly resulting from it - and the actual reduction could be significantly smaller (the estimate is based on the salt content reduction in processed foods, a figure relying on the data supplied by manufacturers, also not taking into account possible compensatory increase in table salt use), it prompted wave of analytical studies. Assuming that similar reduction in salt intake is possible in the U.S. as well, these studies are coming up with very optimistic, impressive figures vs. resulting large-scale health benefit.
For instance, a study published this month concludes that 9.5% decrease in the mean U.S. population salt intake - which amounts to less than 1g of salt a day - would avert
over half a
over the lifetime of 40 to 85y adults alive today (Population Strategies to Decrease Sodium Intake and the Burden of Cardiovascular Disease, Smith-Spangler et al, 2010). In addition, it would save $32.1 billion in medical costs.
The study goes so far as to consider tax on salt, which would require 1,600 times increase in the price of industrial food-grade salt to result in somewhat lower (6%) decrease in the mean population salt intake, and nearly proportionally lower benefits.
What doesn't seem to be adding up at the first sight is that less than 1g of "extra" salt a day can cause so much damage. As many of us know, salt is chemically sodium chloride, i.e. its molecule is composed of sodium and chloride atoms. With the atomic weight of sodium being 22.99, and that of chloride 35.45, one mole (6.02x1023) of their atoms weights 22.99g and 35.45g, respectively. This implies that sodium makes 39.3% of the weight of salt molecule. Hence, 1g of salt contains slightly less than 0.4g of sodium.
As a vital nutrient, a major body electrolyte, sodium is
crucial for proper cellular functioning and very survival.
Body's sodium content averages about 1.4g per kg of body weight; it is just not plausible, from the evolutionary point of view, that human body in general is so vulnerable to such small variations in its intake level. The proposed 9.5% reduction in daily sodium intake would amount to less than 0.4g, or less than 1/200 of body's average sodium content; how likely it is that this minute amount poses an obstacle to body's very sophisticated and capable salt-regulatory mechanisms?
From this standpoint, the estimated enormous benefits by Smith-Spangler et al. seem to be overly optimistic. It is hard to imagine how as little as 1.25 mm Hg decrease in mean systolic blood pressure (MSBP) in the hypertensive population, which is what the authors think would result from 9.5% reduction in salt intake, would do so much good.
How did they come up with those figures?
These estimates were based on another recent study (Potential Societal Savings From Reduced Sodium Consumption in the U.S. Adult Population, Palar and Sturm, 2009), which itself depended on dose-response relationships estimated in the meta study by He and MacGregor (Effect of longer term modest salt reduction on blood pressure, 2004). From the estimated decrease in mean blood pressure is then statistically derived how many people it would switch from the "hypertensive" (over 140/90 mm Hg) to "elevated" blood pressure classification.
Despite the authors characterizing this method as "conservative", it is rather obvious that it
creates statistical fiction,
overestimating the benefit of only slightly lowered blood pressure. There wouldn't be fewer hypertensive individuals in any meaningful sense. Statistically, we might say that X people with systolic blood pressure just over 140 mm Hg would get just below that mark. What would actually happen - assuming the estimated figure of average blood pressure reduction are accurate - is exactly what the figures imply: average blood pressure within hypertensive population would go slightly lower, and the
health effect of it can only be commensurate.
In addition, He and MacGregor estimates are based on mainly small, short-term studies (with their "longer-term" classification applying to trials lasting four weeks, or more), averaging about 100 participants (3,022 total). Those are nether sufficiently long trials, nor sufficiently large samples to reliably represent the effect, or the population.
Beside, taking a closer look at how the 1.25 mm Hg MSBP reduction figure is arrived at reveals that it is almost certainly unrealistically high. Smith-Spangler et al. starts with assuming 3.9g average daily sodium intake (i.e. 9.75g of salt) in the U.S. which is some 15% higher than the more often quoted 3.4g. Since 3.9g of sodium equals 0.17 moles or 170 mmol (millimole, 1/1000 of a mole), it implies 9.5% reduction as 16mmol of sodium.
Now, the direct result obtained from trials analyzed in He and MacGregor was 5.06 mm Hg decrease in MSBP for 78 mmol lower mean sodium excretion (which is a rough indicator of sodium intake). This was extrapolated to 7.2 mm Hg per 100 mmol reduction in sodium intake, which implies mildly exponential change in pressure vs. intake, that approximately fits B~S1.4, where S the ratio of sodium intake reduction change vs. base 78 mmol figure, and B is the ratio of MSBP change vs. corresponding 5.06 mm Hg.
According to it, sodium intake reduction of 16 mmol in the Smith-Spangler et al. model (thus S=16/100=0.16) would only result in 0.4 mm Hg (from B~0.161.4x5.06) lower MSBP - a hardly measurable change.
Instead, Smith-Spangler et al, inconsistent with their source, obtained their MSBP drop estimate by assuming it proportional to the reduction in sodium intake vs. extrapolated value for 100 mmol intake reduction. That gives S=(16/100)=0.16 which, multiplied with 7.2 mm Hg gives 1.15 - a number close to their declared MSBP reduction benefit of 1.25 mm Hg. Should they use the base value of 5.06 mm HG drop with 78 mmol sodium intake reduction, the corresponding drop for 16 mmol reduction would have been 1 mm Hg. But
the only MSBP drop value
consistent with their own source
If this figure-sloppiness seems inappropriate for a study intended to inform the general public and its major institutions about important health matters, wait - there's more to come.
Smith-Spangler et al. mention that there are other data sources for estimating the effect of reduced sodium intake on blood pressure, but they wouldn't use them because they see them less appropriate for the purpose. What they don't say is that meta studies attempting to establish the specific relation between reduction in salt intake and blood pressure in hypertensive individuals have resulted in a range of values, and that the one they use is significantly higher than the rest of them, and
nearly double the lowest one.
While all meta studies agree that substantial (30% to 50%) reduction in salt intake results in measurably lower blood pressure in hypertensive individuals - at least for the duration of studies - their specific numbers (and interpretations) differ. They are split between lower benefit (Midgley et al. 1996, Graudal et al. 1998) and somewhat higher (Cutler et al. 1997, He and MacGregor, 2004). Their respective mean values of MSBP drop for 100 mmol sodium intake reduction are 3.7, 3.9, 4.9 and 7.2 mm Hg.
Crystal-Spangler et al. used this last, highest, nearly certainly over-inflated figure.
It is worth noting that authors of the two meta studies with higher MSBP drop figures are well known salt hypothesis proponents. Another estimate from the salt hypothesis camp, based on an overview of randomized controlled trials, putting the drop in MSBP for 70 mmol reduction in sodium intake at 2.9 mm Hg, and that for 100 mmol reduction at 4.5 mm Hg (Nutritional factors in blood pressure, Elliott, 1994), almost 40% lower than the highest figure.
When weighing figures of these meta studies, it is reasonable to assume is that the results of these meta studies generally overestimate the effect, due to publication bias (studies finding insignificant effect are more likely not to get published) and placebo effect. Not a consideration in Crystal-Spangler et al.
Other existing data sources don't give support to their high figure as well. Good illustration of the expected effect on general population (including both, hypertensives and normotensives), is furnished by the Japanese NIPPON DATA80. In it, as much as 257mmol differential in sodium intake between the lowest and highest quintal of the study corresponded to as little as 2.9mm Hg (systolic) in Japanese man; Japanese woman in the highest quintile of sodium intake had only 1.2mm Hg higher averaged systolic pressure than women in the lowest quintile (219mmol higher sodium intake differential).
If these real-life proportions would apply to Smith-Spangler et al. 17mmol sodium intake reduction, the benefit would have melted
down to 0.1-0.2 mm Hg MSBP reduction
at the level of population.
Opting for the study with by far the highest MSBP increase by Smith-Spangler et al. does not seem to have other rationale but to support the purpose of the study:
picturing salt as major culprit in hypertension,
in order to support the call for reduction in salt intake. More so considering that the two studies with the lowest estimates of MSBP drop with salt intake reduction include substantially larger number of trials: 56 and 114, respectively, compared to 31 trials in He and MacGregor.
Finally, since Smith-Spangler et al. base their benefits estimate on the fictional number of people that this small reduction in blood pressure would push just below 140/90 mm Hg threshold, the applied value for MSBP drop
should be smaller than the average,
since the drop in blood pressure is generally commensurate with its nominal height (roughly, average MSBP drop for normotensives is only about 1/3 of that for hypertensives), and these individuals are, in that respect, at the very bottom of the hypertensive population. Specifically, the MSBP drop for normotensives in their source study was about 40% of that for hypertensives, implying that the correct value applied for those just above the 140 mm Hg level would be about 70% of the averaged value for the hypertensives.
In this context, if the study was consistent with its own model, it should have used 70% of the base averaged MSBP drop, or about 3.6 mm Hg with 78 mmol sodium intake reduction. A proportional MSBP drop corresponding to 16mmol salt intake reduction is about 0.7 mm Hg.
We could go on and on...
Obviously, we are not talking about number sloppiness; it looks more like
the case of number fabrication.
In fact, such conclusion agrees with a recent long-term meta study on the effects of salt intake reduction more than double that suggested in Crystal-Spangler et al. In it, the average systolic blood pressure reduction is lower than what the latter projects (1.1 vs. 1.25 mm Hg), there is no direct relation between sodium intake and blood pressure, and no significant effect on cardiovascular morbidity/mortality (Systematic review of long term effects of advice to reduce dietary salt in adults, Hooper et al. 2002)
So, there we have it. Just a little closer look at the study promoting big health and monetary benefits from a minor reduction in salt intake reveals significant biases, inconsistencies and statistical manipulation. By any reasonable measure, the benefit in lowering MSBP in hypertensive population by 9.5% reduction in sodium intake should be well below 1 mm Hg mean. Plainly put, negligible.
Granted, it would be hard to construct all the health and monetary benefits around a number starting with zero.
But Smith-Spangler et al. is not just exception. It is only the last in a series of similar statistical studies promoting the same idea of enormous health and monetary benefits achievable by small reductions in population's salt intake (Dall et al. 2009, Palar et al. 2009, Bibbins-Domingo et al. 2010, the above mentioned studies, and others). Why such a
wide, organized, persistent insistence on
If these officially endorsed projections are flawed, what is the "right" number? The answer seems to be: We don't really know. Worse yet, most of the evidence that we have points to the opposite: life expectancy declining toward the lower-end salt intake levels.
But the government must be having solid evidence that lowering intake of a vital nutrient for the entire population is safe, right? The bad news is,
there is no such evidence.
It may be hard to believe, but all that this grand anti-salt campaign is based on is a statistical indication that substantial change in salt intake produces small but measurable raise in blood pressure, and high blood pressure is linked to a plethora of diseases.
Notice the disconnect between the first and second part of the sentence: significantly higher salt intake is usually associated with only a small elevation in blood pressure, NOT high blood pressure. There is no evidence that salt causes high blood pressure on the level of population. There is also
no substantial evidence that it actually lowers life expectancy.
In fact, the evidence to the contrary is overwhelming.
Why would the government, and all those supporting the anti-salt move engage in such misadventure? Let's try to find the answers; what specific evidence supports this major move by CDC, IOM, NLHBI (National Lung, Heart and Blood Institute) and a number of associated health organizations?
The reality is
much more complex than
the oversimplified, exaggerated
Sodium is among minerals vital for proper cellular function. It is not surprising that its body level can - and often does - have significant effect on its functioning.
The problem is that the specifics of this effect vary individually, depending on individual metabolism, nutritional status, age, race, gender, state of health, lifestyle and environmental exposure. This simple detail seems to get more neglected as the salt hysteria builds. The Institute of Medicine suggests that even the DRI for sodium, 2.3 g/day, is not safe for the majority of population (those over 40y old, those with hypertension, and African-Americans), which should keep sodium intake below 1.5g/day.
For comparison, the vast majority of Earth's human populations average between 2.5g and 5g/day of sodium (corresponds to 2.5 times larger quantity of salt, 5-10g). Mass addiction, or
indication that the
optimum salt intake does gravitate
Not a few people may get impression that drastic salt intake reduction is their best bet health-wise. And that is not necessarily so. Just like with anything else, the need and tolerance for salt can vary significantly from one person to another. The only certain part is that everyone must satisfy individual needs for it in order to maintain optimum health. And it may not be provided by 1.5g of salt per day, or less.
By now, you may be beginning to suspect that there is more involved in this salt controversy than just science. In fact, it is more travesty than science, to the extent that is hard to believe. The story is long, but if you want to figure out what is really happening here and how much of a danger salt really is, keep reading.
We'll start with how the whole idea of salt inducing high blood pressure came about, and then turn to the results of more recent studies. Since thousands of studies and research papers have been published on this subject, the focus will be on those few that were pivotal for the development of this view and health policy based on it, as well as those most important for the fact sheet of this controversy.