Dietary Salt: Medical Science Corrects a Long-Standing Error



The Federal Center for Disease Control (CDC) commissioned a review of the health benefits of reducing salt intake, and a draft of the final report is available on line.  The take-home message is that salt, in the quantities consumed by most Americans, is no longer considered a substantial health hazzard.  The average American eats about 1½ tsp. per day.  What the CDC study reported explicitly is that there is no benefit, and may be a danger, from reducing our salt intake below 1 tsp per day  But even above 1 tsp, the evidence is tenuous and inconsistent.  It may be that we’re better off with more salt than less, up to 2 or even 3 tsp per day. How did it happen that such standard medical advice drifted astray, then went un-corrected for so long?

This review by the National Academies Institute of Medicine (IOM), commissioned by CDC, considered dozens of studies, from cross-cultural (less reliable) to prospective, randomized with control (most reliable).  Most studies showed no relationship between salt intake and any health outcome.  Some seemed to indicate that more salt had a beneficial effect.  Gina Kolata of the NYTimes published an excellent summary of the report a few weeks ago, and I won’t try to improve on her article.

Here’s an excerpt:

One 2008 study the committee examined, for example, randomly assigned 232 Italian patients with aggressively treated moderate to severe congestive heart failure to consume either 2,760 or 1,840 milligrams of sodium a day, but otherwise to consume the same diet. Those consuming the lower level of sodium had more than three times the number of hospital readmissions — 30 as compared with 9 in the higher-salt group — and more than twice as many deaths — 15 as compared with 6 in the higher-salt group.

Another study, published in 2011, followed 28,800 subjects with high blood pressure ages 55 and older for 4.7 years and analyzed their sodium consumption by urinalysis. The researchers reported that the risks of heart attacks, strokes, congestive heart failure and death from heart disease increased significantly for those consuming more than 7,000 milligrams of sodium a day and for those consuming fewer than 3,000 milligrams of sodium a day.

To translate this last study into teaspoons: the finding was that anything between 1½ and 3 tsp of salt per day is just fine, and there were adverse effects from eating more than that or less than that.  Most Americans who are not consciously restricting salt fall in this range (1½ to 3 tsp).  People who are on low-salt diets for medical reasons are getting as little as ½ tsp, and they’re well into the range where dearth of salt is harming them.  The worst impact of low salt is on insulin sensitivity. Loss of insulin sensitivity is a big risk factor for all the diseases of old age.

Comments and Context

Perspiration is the biggest source of variability in individual requirement for salt.  Given that our need for salt varies widely with exercise and with temperature, it’s surprising that most of the studies proceeded blindly with a “one-size-fits-all” model.  A few tested for salt excreted in the urine, which is a better measure than the amount consumed.

The IOM finding was heralded by a great number of alternative web health sources which have told us for years that salt has been falsely demonized (Dr MercolaDr WeillAdele Hite)

How did the medical community fall into such a large error, lasting so long?  The medical research community gets a lot of things right, and when they make major errors, you can usually see the trail of money leading them astray (Vioxx, hormone replacement therapy…).  But there was no Pharma Giant making hay from people who ate less salt, so how did this happen?  I have a personal theory:

Blood Pressure and Clogged Arteries – Chicken and Egg

The central problem (I believe) is that the causal association between salt and heart disease was and still is based on faulty logic.  The error is in the connection through blood pressure as an intermediate variable.  The new review does attempt to look for direct evidence of a link between salt and CVD (cardio-vascular disease), independent of blood pressure, but the IOM review also echoes the faulty reasoning, perpetuating the old error.

As Kolata writes: “Until about 2006, almost all studies on salt and health outcomes relied on the well-known fact that blood pressure can drop slightly when people eat less salt. From that, and from other studies linking blood pressure to risks of heart attacks and strokes, researchers created models showing how many lives could be saved if people ate less salt.

Here’s the logical fallacy:  High blood pressure is associated with increased risk of heart attack.  Salt causes blood pressure to increase.  Therefore, (the community consensus concludes), salt increases the risk of heart attack.  The problem with this thinking is an example of the old warning, “correlation is not necessarily an indication of causation.”  In this case, there may be no causation at all.  Here’s why:

Atherosclerosis is the most common precursor to heart disease.  In atherosclerosis, arteries lose their elasticity and also become clogged with inflamed fatty deposits. The deposits can break off and block blood flow.  If this happens in the brain, the result is a stroke, and if it is in a coronary artery, the result is a heart attack.  So we understand why atherosclerosis implies a big risk of heart attack and stroke.

It’s also true that atherosclerosis causes elevated blood pressure.  This is a simple mechanical effect.  When you have a narrower tube, it takes more pressure to drive the fluid through it.  Atherosclerosis causes high blood pressure and also causes risk of CVD.  And most people are subject to atherosclerosis as we age.  So it’s no mystery that high blood pressure should be correlated with CV risk.  But that doesn’t mean that high blood pressure causes CV risk.

Eating salt causes the body to retain water, and the swelling increases ambient pressure in surrounding tissues, and blood pressure increases in response.  So salt increases blood pressure and atherosclerosis increases blood pressure, but they work via two very different mechanisms.  There’s every reason to believe atherosclerosis causes CV risk, but no reason to think that salt should lead to CV risk, even though it increases blood pressure.

It’s clear from epidemiology that atherosclerosis is correlated with blood pressure.  It’s clear from simple physics that atherosclerosis causes high blood pressure. (A => B.) But maybe there’s a causal link in the other direction as well (B => A).  This is what doctors and medical researchers have long assumed: that high blood pressure causes a mechanical strain in the arteries, and that in some unexplained way the body’s response is atherosclerosis.  It’s a long-standing theory.  And if it is true that B causes A, then of course we’d have reason to think that salt, which raises blood pressure, would also increase severity of atherosclerosis.

This hypothesis (B =>A) was first confirmed in an experiment on rabbits in 1989.  The experiment seems questionable, with hindsight.  Perhaps the researchers were looking too hard for  evidence to support an important and popular theory, and they were fooled by confirmation bias.  This is the well-studied tendency of all of us – with and without scientific training – to easily take in new information that fits well with what we already know, but to ignore or cast aside information that doesn’t fit with our existing beliefs.  When experiments don’t turn out the way we expect, we take it as a failure, and often we try to “fix” the problems, until the experiment “works”.

The 1989 study was by a senior team of researchers at Boston University, headed by Aram Chobanian  (who subsequently served as University president). They operated on young rabbits, removing one kidney and constricting blood flow to the other with a tight-fitting band around the artery.  About half the operated rabbits developed hypertension, and the other half either died or were excluded from the study.  The remaining rabbits, all of whom had high blood pressure, developed atherosclerosis – the result they were looking for.  The Chobanian study made a big splash, and has been cited hundreds of times since 1989.

But could it be that the experiment came out the way it did for a different reason entirely?  (This is my own thought, and I haven’t seen it in print.  You’re the first to read it.)  Restricting kidney function can lead quite directly to atherosclerosis, independent of the blood pressure connection.  Atherosclerosis is known to be elevated in kidney dialysis patients.  A 2011 study comparing autopsies of heart patients and dialysis patients found that the arteries of the dialysis patients were even more clogged than those of the heart patients.  This is not B=>A but K=>A.  Maybe the idea that B=>A has been a long detour into the wilderness.

Surveying web sites that offer the best in standard medical advice, I’m surprised to see that all of them embrace the logic (B=>A) without any acknowledgment that (A=>B) is the expectation of elementary physics, and that evidence for (B=>A) is tenuous and indirect.  None of them have yet updated their advice on salt to comport with the newly-recognized reality.

Crystal Ball

Given what we know about confirmation bias, it was rather brave of the IOM to report conclusions that were at odds with medical advice of the last 50 years.  Officially, the recommendation to reduce daily intake below 1 tsp has been lifted, but some of the studies cited seem to show that our bodies need more than this to avoid impaired insulin sensitivity.  Now that IOM has made it respectable to take a different view of salt, we can expect a sea change in the medical establishment’s attitudes toward salt, as the force of confirmation bias has been shifted to the other side of the sc



  1. My counter to everyones anti-wikipedia argument is usually that peer review processes/academic journals are baiased and flawed while also being our loudest method of channeling the conclusions made from the academic world to our mainstream culture, and that in a post wikipedia world we wouldn’t create vague ridiculous rumors like “you should drink 8 glasses of water a day”, so it’s awesome to see our methods of conclusion making and their effects on health brought in to question on hplus. Thanks for the article.

Leave a Reply