Why Medicine Might Be Wrong About Salt, Fat & BMI


Salt and fat kill you early, and your BMI tells you how early. That has been the wisdom for years, but wisdoms have an expiry date, too. Particularly medical wisdoms. Recent research says those three are probably well beyond their use-by date. [tweet this]. 

We live in interesting times. Admitted, my view of times is myopic, it's focused on the biomedical. So, I'm obviously not referring to Greece teetering on the economic brink. In biomedicine our Greeks are the cherished wisdoms about salt, fat and BMI. Similarity 1: They are not doing so well. Similarity 2: Their balance sheet screams bankruptcy. Similarity 3: Our authorities won't kick them out for fear of a domino effect.

Actually, medical history is full of interesting times. Remember, when a young doctor suggested that simply washing hands between dissecting cadavers and helping women give birth would seriously reduce the regular 1-in-five death rate from childbed fever? 
Of course, you won't remember this: the place was the Vienna General Hospital in Austria, the year was 1849 and the young doctor's name was Ignaz Semmelweis. While he didn't publish his observations, one of his students did, in the grand old dame of British medical journals, the Lancet [1]. At that time infection was known per se, but it was believed to work like this: a "peculiar morbid atmospheric influence which extends beyond the range of personal communication". Semmelweis believed in washing hands. And he had the numbers to prove his belief. After introducing a hand-washing rule in his department, childbed deaths dropped by 75%. The response of Semmelweis' peers and superiors to his challenging notion is today known as the Semmelweis reflex. Unlike Semmelweis it is very much alive. It is the reflex-like rejection of new insights because they disagree with entrenched beliefs. 

Semmelweis' observations were one of the initial steps in the development of the germ theory of disease. Three more names are attached to its development, Louis Pasteur, Robert Koch and Joseph Lister. The latter was the first to apply this new theory to surgical procedures. So, the next time you gargle with Listerine, you might want to say a silent thank you to all the men and women of science who had the guts to challenge those pompous idiots who, in true Semmelweis-reflex mode, did everything to discredit the new insights. And they usually are quite successful. Who wants to argue with a praetorian guard of honorable old professors.  In Semmelweis' case, they got him barred from medical practice, they publicly ridiculed him and ultimately drove him to insanity. All the while women continued to die in childbed. Unsurprisingly, because medical textbooks continued to teach the old views on childbed fever until the 1890s. But once germ theory, and with it hygiene, was adopted into medicine and daily life, the mortality landscape changed dramatically. Infectious diseases disappeared from the pole position of the death tables, and the 1900s witnessed the emergence of their replacements: cardiovascular disease (CVD) in all its flavors, from hypertension and atherosclerosis to heart attack, stroke and heart failure.  

Recently, potential triggers for those Semmelweis reflexes have been coming out of research, though probably not as dramatic in consequence as the infection theory. 

Salt is no evil

I have written about the potentially flawed obsession with trying to get everybody to reduce salt intake in order to reduce blood pressure and stroke in the entire population.  The latest nail in this obsession's coffin is Alderman and Cohen's review of 23 observational studies and 7 randomized controlled trials (RCT), all investigating the effects of salt intake on parameters of health [2]. The observational studies accumulated data on 360,000 participants and recorded 26,000 disease events. While 7 studies showed a direct relationship between increasing sodium and increasing CVD events, another 6 studies demonstrated the opposite: a clear inverse relationship. Two studies showed a J-shaped relationship, in which low and high intakes increased disease risk, whereas 8 studies didn't show any relationship or only inconsistent results. 

Only 9 of the 23 studies measured salt intake objectively, that is by measuring participants' urinary sodium excretion. I mention this because the other 14 studies relied on self-reported sodium intake, which is prone to recall errors. Of those 9 studies with objective measurements, only 1 showed a direct association, whereas in three studies higher intake was associated with fewer disease events. Two studies showed a J-shaped relationship and one didn't show any relation. 

Is it now wise to simply balance those scores, like in football, and let the highest scorer win the tournament? In our case the team "inverse relation" would win hands down over its competitors "direct relation" "J-shaped relation" and "no relation". Should we draw our conclusions in this way? No, science does not work this way. And it's not the way the authors chose. When they looked a little closer at the sodium intake ranges in each of the RCT's they found that the seemingly conflicting results of the observational studies could be reconciled easily. 

It turns out that it all depends on the baseline intake from which you increase or decrease your salt consumption. The safe range being actually quite large, between 2.5-6.0 g/day. Go below or above that and you will face some increased risk. Interestingly, this range is way in excess of the current authoritative recommendation of "less than 2.0g/day". What is also rarely mentioned by those "authoritative guidelines" are some other side effects of lowering salt intake. In several of those RCTs salt reduction came with an increase in blood cholesterol, insulin resistance, adrenaline secretion and sympathetic nerve activity. None of those effects is beneficial to health. That's curious because all, except for cholesterol, tend to raise blood pressure. So the net effect of salt reduction, or increase, in you is always a composite of all those biochemical responses to sodium intake change.  

Want to bet whether these insights will trigger the Semmelweis reflex in some of those who have built their career on maligning salt? On to the next subject:

Fat isn't so bad either

Khaw and colleagues wanted to know the answer to an old question: does the fat in your diet give you heart disease [3]? The belief that dietary fat and heart disease march lock-step is so ingrained that you are forgiven to wonder why anybody would spend time and effort on this question. Well, that's because a recent meta-analysis of 21 studies, following 347,000 people for 5 to 23 years, could NOT find any association between the two [4].

Since all of the previous studies had been of an observational nature, which does not allow for conclusions of causality, the researchers used data from a prospective trial which had investigated the correlation of diet with cancer outcome, the European Prospective Investigation into Cancer (EPIC)-study. They looked at the fats in the blood of 10,000 participants aged 40–79 years, and they followed them from 1993–1997 through 2011. During this period 2,424 participants were diagnosed with heart disease. From among the remaining 10,000 participants the researchers chose 4,930 disease-free controls for a comparative evaluation.  They checked whether, and how strongly, saturated and unsaturated fatty acids in the blood correlated with heart disease. 

Since the fatty acid composition in the blood mirrors dietary fatty acid intake, this is as close as you can get to a conclusion about how the intake of type of fat affects your risk of heart disease. Of course, you need to adjust for other risk factors, such as age, sex, BMI, smoking, physical activity, alcohol intake, diabetes, blood pressure, cholesterol and other known risk factors for heart disease. 

The researchers did those adjustments, and they found the saturated fatty acids to be only weakly related with heart disease. But that's not the surprising find. The real surprise was about the unsaturated fatty acids of the famous omega-3 and omega-6 persuasion. You have always heard how omega-3, the fish-oil variant is so good for your heart and the omega-6 is not. Well, listen to this: omega-6 turned out to be protective against heart disease, and omega-3 wasn't. That's contrary to our hitherto held beliefs that popping fish oil pills will make you a Methusalem, and that reducing omega-6 intake will do wonders against inflammation in your arteries. See a Semmelweis reflex on the horizon?

BMI is a useless crutch  

There is probably no other number which has become so much engrained in our medical psyche as the BMI. This relation of bodyweight over height squared is the human equivalent of a meat stamp: if it's below 25 you are the prime cut which every health insurer wants on his client list. Bring it above 30 and you are a fat and soon-to-be sick bum whom nobody wants to talk to, unless your name is John Candy. In our society where we determine the winner of a Formula-1 race with millisecond precision, we accept being stamped at-risk with the accuracy of 20/100 vision (20/20 being the normal, and 20/200 being the cut-off for legal blindness). 

The problem with BMI is, it doesn't differentiate between muscle and fat tissue, which makes a body-builder look fat, and bad, on the chart. BMI also doesn't tell you where your fat resides, on your buttocks or on your waist. The latter is certainly worse for your health than the former. Still BMI is THE number to judge you by. Maybe not any more. 

Krakauer and Krakauer have developed a new measure from looking at the numbers of the National Health and Nutrition Examination Survey (NHANES) 1999-2004, and correlating those numbers with the death statistics. They wanted to blend weight, height and body shape into a more informative indicator of disease risk. Which is why they called it A Body Shape Index, ABSI. They also wanted this number to be easy to calculate from parameters which everyone can measure. Which is why the ABSI only calls for waist circumference to be measured in addition to the BMI's parameters of height and weight. I won't bore you with the details of statistical development of this ABSI, but suffice it to say, it's been done beautifully and very thoroughly. Then, after adjusting for other known risk factors, such as age, smoking, diabetes, blood pressure and cholesterol, the authors correlated BMI, WC and ABSI with death. While WC and BMI didn't show any correlation ABSI was strongly correlated. That's surprising, given the relatively short follow-up period of 5 years. 

Of course, there area lot more questions to be answered before the ABSI will make it into medicine's hall of fame. The most important: Does lowering the ABSI translate into increasing health, or improving risk or extending lifespan? 

So, be prepared for quite some time to pass before your doctor tells you that your ABSI requires some serious attention. Once that happens, you might remember this post and look up the time that has passed between its publication and your encounter with the ABSI in a medical environment. You will then see that medical science grinds slowly. But ultimately it grinds. 
In this case, medical science might grind a little slower, because the ABSI isn't the brainchild of a biomedical or public health scientist. It's developer is an assistant professor in the Department of Civil Engineering of the City College of New York. 
Now, how can a civil engineer's index beat our beloved BMI when biomedicine's best brains have been laboring over a BMI replacement for years?
Great potential for a Semmelweis effect, don't you think?  [tweet this]. 


1. Routh CH: On the Causes of the Endemic Puerperal Fever of Vienna. Medico-chirurgical transactions 1849, 32:27-40.
2. Alderman MH, Cohen HW: Dietary Sodium Intake and Cardiovascular Mortality: Controversy Resolved? Am J Hypertens 2012, 25(7):727-734.
3. Khaw K-T, Friesen MD, Riboli E, Luben R, Wareham N: Plasma Phospholipid Fatty Acid Concentration and Incident Coronary Heart Disease in Men and Women: The EPIC-Norfolk Prospective Study. PLoS Med 2012, 9(7):e1001255.
4. Siri-Tarino PW, Sun Q, Hu FB, Krauss RM: Meta-analysis of prospective cohort studies evaluating the association of saturated fat with cardiovascular disease. Am J Clin Nutr 2010, 91(3):535-546.


Routh CH (1849). On the Causes of the Endemic Puerperal Fever of Vienna. Medico-chirurgical transactions, 32, 27-40 PMID: 20895917

Alderman MH, & Cohen HW (2012). Dietary sodium intake and cardiovascular mortality: controversy resolved? American journal of hypertension, 25 (7), 727-34 PMID: 22627176

Khaw KT, Friesen MD, Riboli E, Luben R, & Wareham N (2012). Plasma Phospholipid Fatty Acid Concentration and Incident Coronary Heart Disease in Men and Women: The EPIC-Norfolk Prospective Study. PLoS medicine, 9 (7) PMID: 22802735

Siri-Tarino PW, Sun Q, Hu FB, & Krauss RM (2010). Meta-analysis of prospective cohort studies evaluating the association of saturated fat with cardiovascular disease. The American journal of clinical nutrition, 91 (3), 535-46 PMID: 20071648