Diet-Heart Hypothesis: R.I.P.
This is another nail in the coffin of the Diet-Heart Hypothesis.
Researchers in New Zealand re-examined all the high-quality scientific studies looking at the effect of dietary fats on coronary heart disease, the leading cause of death in Western societies. Atherosclerosis in the heart arteries can cause chest pain (angina), heart attacks, heart failure, death, or the need to overcome a clogged artery with angioplasty or bypass surgery.
Here are the major findings of the meta-analysis:
- total fat intake is not associated with coronary heart disease
- higher trans fatty acid intake is linked to higher heart deaths and events
- saturated fatty acid consumption is not linked to heart deaths or events
- data on heart disease and polyunsaturated fatty acids are "inconsistent and unreliable"
- monunsaturated fatty acid intake was not associated with heart disease
- higher omega-3 fatty acid consumption (from fish oils or actual fish) is linked to lower risk of heart disease, although the data are not quite as strong as the authors would prefer
With the exception of trans fats and omega-3 fats, the authors write, "The available evidence from cohort and randomized controlled trials is unsatisfactory and unreliable to make judgment about and substantiate the effects of dietary fat on risk of CHD [coronary heart disease]."
But they have made a judgement: Total and saturated fats are not related to heart disease.
In an interesting post-script, the authors mention "Expert Consultation," which sounds like an oversight panel. This committee seems to have insisted on modifcation to the article to the effect that "replacing saturated fat with polyunsaturated fat reduces CHD risk." Reading between the lines, I sense the authors had to swallow hard before adding that.
-Steve Parker, M.D.
Reference: Skeaff, C. Murray and Miller, Jody. Dietary fat and coronary heart disease: Summary of evidence from prospective cohort and randomised controlled trials. Annals of Nutrition and Metabolism, 55 (2009): 173-201. Available free online September 15, 2009.
 older posts |
home | newer posts  |
| Finally, Some GOOD News: Heart Disease Hospitalizations Down | | The Heart Health Blog | | ICD: Women, Think Twice Before Accepting This Heart Hardware |
Posted by: Monica Reinagel | Sep 22, 2009 9:01:36 AM
Steve, maybe the Fat-Heart Disease Hypothesis is on the ropes but I think the Diet-Heart Disease Hypothesis is still very much in play. The focus is simply shifting from fat to refined carbohydrates.
Posted by: Celeste | Sep 22, 2009 10:29:53 AM
I think it goes further than just refined carbohydrates (WHITE SUGAR AND WHITE FLOUR)... it is the whole processed food diet in general that is causing heart disease (and a plethora of other chronic illnesses).
Remember Dr. Harvey Wiley??? NO??? Google his name and read about the research he did 100 years ago with relation to preservatives in our food supply. Processed food contains huge amounts of preservatives and produces a whole laundry list of diseases.... this is what is killing us all.
Posted by: Joseph Gentzel | Sep 22, 2009 10:49:51 AM
This really should not be that surprising although it certainly goes against the grain of popular thought.
Many fats are now being reclassified as "good" fats. Olive oil has always been the star, but there was a day not so long ago. We see now that duck and goose fat are very good for us, supported for centuries by the high fat diet from these of the Gascony people in SW France. They happen to be one of the longest lived people on earth. This is specific to this region in France, so it’s not just the wine.
Posted by: Kurt | Sep 22, 2009 10:55:31 AM
Again, I have no problem believing this, but IF eating a lower fat diet lowers blood cholesterol, as it seemed to do for me (HDL decreased from 155 to 95), then it would seem to follow that cholesterol levels may also not be related to heart disease.
Posted by: Steve Parker, M.D. | Sep 22, 2009 12:52:57 PM
Monica-
You're right: the term "diet-heart hypothesis" can be misleading. But it's the predominant term used in the medical literature to describe the idea that total fat, saturated fat, and dietary cholesterol cause heart attacks and atherosclerosis. Many casual readers here won't know that.
The carbohydrate theory of heart disease is discussed by Dr. Frank Hu, of Harvard Medical School, in the July 3, 2007 issue of Journal of the American College of Cardiology:
http://content.onlinejacc.org/cgi/content/full/50/1/22
Kurt-
I still think that cholesterol levels - specifically, high LDL cholesterol and low HDL cholesterol - are significantly related to increased heart disease. Total cholesterol is less important. There are cholesterol skeptics out there, no doubt.
-Steve
Posted by: Dave | Sep 23, 2009 8:59:56 AM
@Steve,
The question asked by cholesterol skeptics is one of root cause: does high cholesterol play an independent role in the development of heart disease (making it a treatment target) or is it a symptom of some other cause? There is certainly an association between high LDL/low HDL and heart disease, as there is with many other metabolic markers. The key question is one of causality, i.e. if I lower LDL independent of other metabolic factors, does that significantly reduce disease? Or is high LDL a symptom of a deeper metabolic problem? And there's the closely related question of whether lowering cholesterol has a beneficial effect on total disease incidence and mortality, see e.g.
http://www.canibaisereis.com/2009/09/19/low-cholesterol-certainly-not-healthy/
The issue with LDL in particular may be deeper than just total amount, because the inflammatory reaction now believed to be at the root of formation of atherosclerotic plaques is driven by oxidized or sugar-damaged (glycated) LDL. This links several observations, i.e. if your LDL is chronically high, it's probably hanging around longer in the blood which gives it a better chance of being damaged. If your LDL particles are small, they have greater surface area, again increasing the potential for damage. Uncontrolled blood sugar increases the potential for glycation.
I would guess the keys to avoiding heart disease lay in maintaining metabolic normality (e.g. glucose tolerance) and keeping inflammatory reactions in check via appropriate omega-6/3 ratio and Vitamin D repletion.
Posted by: Lars | Sep 30, 2009 2:37:16 AM
Well . . . try this study, which found just about the opposite conclusion.
Am J Clin Nutr. 2009 May;89(5):1425-32.
Major types of dietary fat and risk of coronary heart disease: a pooled analysis of 11 cohort studies.
Jakobsen MU, O'Reilly EJ, Heitmann BL . . . Willett WC, Ascherio A
"CONCLUSION: The associations suggest
that replacing SFAs with PUFAs rather than MUFAs or carbohydrates prevents CHD over a wide range of intakes."
The evidence for a protective effect of n-6 when replacing saturated fats is, frankly, overwhelming.
There's much more evidence to quote of course.
Posted by: Lars | Sep 30, 2009 3:04:07 AM
No, Dave does not make a lot of good points.
1. LDL particle size is not related to CVD, only particle number.
2. There is no evidence that n-6/n-3 ratio is significant at all; only total n-3 intake.
Look it up.
Posted by: greg | Sep 30, 2009 4:16:43 AM
Steve - (I hope you are still reading posts here) I am a bit confused. Did we read the same paper? You wrote:
"But [the authors] have made a judgement: Total and saturated fats are not related to heart disease."
It seems to me pretty clear that the authors have made no such judgment. The authors write in the discussion:
"The null results of the observational studies of dietary lipids and CHD do not negate the importance of the underlying associations [between fat consumed and variation in the incidence of CHD], but reflect the combined effects of limitations of dietary assessment methods, inadequate numbers of participants studied and the prolonged follow-up of individuals."
They continue: "Hence, the null results are very likely to result from regression dilution bias and confounding of one nutrient by another."
In other words, contrary to the implications of your post, the authors attribute the failure to reject the null hypothesis to the presence of measurement error not the absence of a relationship.
The authors go on to point out: "By contrast, CHD risk is moderately strongly related to dietary patterns, such as vegetarian or Mediterranean diet, which are less influenced by mis-classification."
And if there was any doubt in the authors views: "The null results probably reflect the unreliability of the evidence on dietary fats from cohort studies that differs markedly from the reliability of ecological studies or metabolic ward studies of diet and cholesterol."
I also think you misinterpret the Post-Script. I do not believe the Expert Consultation referenced an referee insisting on any changes. Rather, this background paper is part of the FAO/WHOs Expert Consultation project. The paper "Fats and Fatty Acids in Human Nutrition: Introduction" explains the roles and methods of the Expert Consultation. I'd take the post-script at face value: an alternative meta-study by Jakobesen, O'Reilly, et. al. had "a significant advance in quality" over the present study and found that there was a statistically significant decrease in CHD risk when PUFA replaces SFA in a diet. Hence the Expert Consultation relied on that finding.
Steve, I'd love to see another post from you on this paper, perhaps the paper by Jakobesen, O'Reilly, et. al., and perhaps the paper by Elmadfa & Kornseiner "Fats and Fatty Acid Requirements for Adults" and put all this into perspective for us.
Unless I am very mistaken, your first post on this paper seems extremely misleading to your readers. This website seems to be a hot-bed for low-carbers and that is well and good and everyone is entitled to their opinions, but I would hate for the appearance to emerge that they are being pandered to and as a result all your readers are not getting a complete and balanced analysis of new literature. After all, you and the rest of the NutritionData team are the experts - the rest of us here are just interested layman - and it is access to your expertise that brings us to this site.
Posted by: greg | Sep 30, 2009 4:54:06 PM
@Dave
you wrote: "And there's the closely related question of whether lowering cholesterol has a beneficial effect on total disease incidence and mortality, see e.g."
I am curious, are you aware of a peer-reviewed study which raises such a question?
"Cardiovascular Benefit of Magnitude of Low-Density Lipoprotein Cholesterol Reduction. A Comparison of Subgroups by Age." Rahilly-Tierney CR, et. al. explicitly finds that lowering LDL reduces cardiovascular risk.
The blog at that link you posted is non-sense. Regressing TC versus Life Expectency aggregated by country isn't revealing anything. A quick look reveals Cameron, Nigeria & Sierra Leon are the low TC/low LE data points. Iceland & Australia are the hi TC/hi LE data points. Add another variable - say per capita income - and those pretty looking pictures are going to go away. TC goes up with wealth: so does LE. That doesn't make TC good for you, rather that is a side effect of the wealthy Western diet. That blog isn't serious science.
@Steve, are you aware of any studies on the efficacy of lowering TC or LDL for health outcomes?
Posted by: Steve Parker, M.D. | Oct 1, 2009 7:23:50 PM
Hi, Greg -
It seems to be awfully hard to prove that dietary saturated fats cause coronary heart disease. Nearly all the studies that support that idea are based on associations, not causality. Then there are the other studies that find no such association. Maybe saturated fats do cause atherosclerosis, but the small magnitude of the effect makes it hard to discern.
It's a critically important question. Maybe my position is wrong. Few practicing physicians are even aware that their belief in the diet-heart hypothesis is contradicted by published scientific reports.
You do a good job presenting evidence contrary to mine. We must look at all sides. As Dr. George Bray wrote on this issue, "Read and decide for yourself." I spent about 80 hours on it. Few other people, physician or otherwise, will do that.
I agree there are several studies associating better cardiovascular health with a diet replacing some saturated fats with polyunsaturated fats. These leave open the possibility that PUFAs actively promote cardiovascular health, while saturated fats are NEUTRAL or simply not AS healthy as PUFAs.
Most physicians, myself included, believe that reasonable lowering of elevated LDL cholesterol levels is associated with better overall health outcomes, including fewer heart attacks and longer life. We focus much less on total cholesterol. Most people with total cholesterol of 350 mg/dl have sky high LDL cholesterol. Treatment targets the LDL. [Have I studied this extensively? No. And I have no references at my fingertips.]
Excessively low LDL (under 75-80 mg/dl, if memory serves me well) is associated with shorter lifespan among the general population.
I'm not sure if NutritionData blogs are a hotbed of low-carbers. They don't seem hesitant to speak up, however! That's fine. I'm a moderate-carber myself.
For my conception of the optimal diet for the general population, see this blog post:
http://blog.nutritiondata.com/heart_health_blog/2009/07/the-21st-century-dietheart-hypothesis.html
-Steve
Posted by: O Primitivo | Oct 3, 2009 3:09:57 PM
Hi Dr. Steve, I'm the author of the post http://www.canibaisereis.com/2009/09/19/low-cholesterol-certainly-not-healthy/ and I'm also on the low-carb/ paleolithic/ mediterranean diets. I agree with your opinion, we don't have strong evidence against saturated fats, and this is why the best paradigm should be the traditional diets/lifestyle of healthy people. I personally don't believe people should move away from their traditional diets just because it increases some obscure physiological parameters, specially in my case, because I'm from Portugal, my traditional diet is the Mediterranean diet. By the way, you can replace the word "Mediterranean" by "hyperlipid" and you'll now understand the main advantage of our diet. High fat intake is what actually makes it protective against heart disease diabetes. We really eat a LOT of olive oil!
Posted by: James | Oct 5, 2009 6:49:00 PM
Saturated Fatty Acids are, in my humble opinion, over victimised. It doesn't seem to make sense that they would be bad for you. When a person loses weight they burn body fat (SFAs) in place of food. It's pretty widely accepted that losing weight is heart-healthy and certainly doesn't increase one's chance of suffering from CHD.
I have a question though. Excuse me if this is very basic, I'm not exactly very well informed of the details. Cholesterol is produced to carry out a task or tasks, right? Should high cholesterol not increase your efficiency to complete those tasks? Correct me if I'm wrong, but it seems that it must be something other than cholesterol itself that causes the problems. Something in my head is saying triglycerides but to be completely honest, I don't know what a triglyceride is! I'd appreciate an answer if anyone were to be so kind :)
Thanks
Posted by: Steve Parker, M.D. | Oct 7, 2009 1:02:57 PM
James-
Triglycerides are fats, composed of three fatty acid chains linked to a single small glycerol molecule. Wish I had a graphic represetation to show you.
Here's an educational link at WebMD.com:
http://www.webmd.com/cholesterol-management/tc/high-triglycerides-overview
Hope that helps.
-Steve
Posted by: acide folique | Oct 31, 2009 1:06:35 AM
The low-fat "diet-heart hypothesis" has been controversial for nearly 100 years. The low-fat-high-carbohydrate diet, promulgated vigorously by the National Cholesterol Education Program, National Institutes of Health, and American Heart Association since the Lipid Research Clinics-Primary Prevention Program in 1984, and earlier by the U.S. Department of Agriculture food pyramid, may well have played an unintended role in the current epidemics of obesity, lipid abnormalities, type II diabetes, and metabolic syndromes.
