Monday, August 3, 2009

The Diet-Heart Hypothesis: Oxidized LDL, Part I

In my reading about lipoprotein particles (LDL, HDL, etc.) and how they associate with cardiac risk, I've come across three LDL-related markers that associate with risk: LDL cholesterol, LDL particle number, and LDL size/density. Is this a coincidence, or is there a reason for it?

The first marker, LDL cholesterol, is probably nothing more than a crude approximation of particle number. But LDL particle number and size/density are related to something else, that probably actually causes atherosclerosis rather than simply being associated with it: oxidized LDL (oxLDL).

oxLDL is formed when the lipids in LDL particles react with oxygen and break down. This happens specifically to the unsaturated fats in LDL, because saturated fats, by their chemical nature, are very resistant to oxidative damage. Polyunsaturated fats are much more susceptible to oxidative damage than saturated or monounsaturated fats. Linoleic acid (the omega-6 fatty acid found abundantly in industrial seed oils) is the main polyunsaturated fatty acid in LDL.

LDL is packaged with antioxidants in the liver, primarily vitamin E and coenzyme Q10 (CoQ10), to prevent its oxidation. However, the more time it spends in the blood, the more likely it is to exhaust its antioxidant store and become oxidized. Also, the smaller the LDL particle, the more likely it is to become trapped in the vessel wall and become oxidized there.

Oxidized LDL Correlates Tightly with Cardiac Risk

oxLDL has turned out to be a very sensitive marker of cardiac risk, surpassing traditional markers like LDL, HDL, and triglycerides in most studies to date. Since the discovery of sensitive assays that detect oxidized LDL drawn directly from patient blood, a number of studies have been published supporting its ability to detect atherosclerosis (plaque buildup in the arteries), heart attack risk and even the metabolic syndrome.

Holovet and colleagues published a study comparing the ability of oxLDL and a traditional risk factor assessment to detect coronary artery disease. The traditional method is called the Global Risk Factor Assessment Score (GRAS), and includes age, total cholesterol, HDL, blood pressure, diabetes and smoking status. It's similar to the commonly used Framingham risk score (which, interestingly enough, doesn't include LDL).

GRAS was able to correctly differentiate a healthy person from a person with coronary artery disease 49% of the time, while oxLDL was correct 82% of the time. Thus, oxLDL by itself was far more accurate than a whole battery of traditional cholesterol and cardiac markers. Coronary patients had more than twice the level of circulating oxLDL than the healthy comparison group.

In a large prospective study by Meisinger and colleagues, participants with high oxLDL had a 4.25 higher risk of heart attack than patients with lower oxLDL. oxLDL blew away all other blood lipid markers by nearly a factor of two. From the abstract:
Plasma oxLDL was the strongest predictor of CHD events compared with a conventional lipoprotein profile and other traditional risk factors for CHD.
Oxidized LDL Makes Sense

 Regular, non-oxidized LDL has few properties that would make it a suspect in atherosclerosis. It's just a little particle carrying cholesterol and fats from the liver to other organs. As soon as it oxidizes, however, it becomes pro-inflammatory, immunogenic, damaging to the vessel wall, and most importantly, capable of transforming immune cells called macrophages into foam cells, a major constituent of arterial plaque.

Researchers have been interested in the plaque-generating properties of oxLDL for over three decades, and quite a bit of data have accumulated. They've identified cellular receptors that allow macrophages to ingest oxLDL (CD36 and SR-A). These receptors are specific for oxLDL and do not recognize normal LDL to a significant degree. Mice whose macrophages lack either of these two receptors have the same amount of circulating LDL as normal mice, yet have 60 to 70 percent less atherosclerosis when fed a plaque-forming diet (1, 2). Shorter-term studies have not always been consistent however, suggesting that there are alternative mechanisms. I'll expand on this more later.

Another line of evidence comes from the ability of LDL-borne antioxidants to prevent atherosclerosis in animal models. The powerful synthetic antioxidant probucol greatly reduces atherosclerosis in a number of animal models. It also reduces the extremely high cholesterol rodents and herbivorous animals get when they eat a high-cholesterol "atherogenic diet", but several studies have concluded that the majority of probucol's effect is due to its antioxidant ability rather than its ability to reduce cholesterol (ref).

Vitamin E and CoQ10 are two other LDL-borne antioxidants that can reduce atherosclerosis in animal models, particularly in combination with one another. Vitamin E alone is not as effective, and in some studies totally ineffective, which is one possible explanation for the equivocal results of vitamin E cardiovascular trials in humans. The most effective combination of antioxidants is probably the one provided by a nutrient-dense diet.

In Summary

Multiple lines of evidence suggest that oxidized LDL plays a dominant role in atherosclerosis. Not only is it associated with cardiovascular risk, there's also a large body of evidence suggesting it actually directly contributes to it. 


Ankur said...

How does oxLDL compare with CRP measurement as a predictor of cardio vascular disease, specifically plaque formation ?

TedHutchinson said...

Activation of calpain-1 in human carotid artery atherosclerotic lesions
This may explain the route.

Effects of equol on oxidized low-density lipoprotein-induced apoptosis in endothelial cells.
Interesting 50% of adults do not produce equol after eating soy.

Robert Andrew Brown said...

Another excellent informative and thought provoking post Stephan, and no doubt Omega 3 and 6 will appear at some point (-: if we have been reading the same things.

Ankur I do not have the ref to hand but I recall an article saying the prediction rate was 95% plus for ox ldl. I have only seen the abstract of the trial.

Do you have any figures for CRP?

Found it

Inphidel said...


What causes LDL particles to become small?

TedHutchinson said...

Dr Davis's explanation

J said...

great post as usual Stephan,
looking forward to the conclusion to this one.

thanks for all your work.


homertobias said...

My take on ox LDL:
It makes sense in explaining the MECHANISM of atheroma formation as it relates to dietary fat intake. I think this is why Stephan is bringing it up.
It is a mediocre marker of CAD risk. Please don't rush out and ask your doctor for an ox LDL test. It doesn't perform that much better than a traditional TC/HDLC ratio and is probably measuring the same phenonema in a slightly different way. In other words if your TC/HDLC is bad then your ox LDL is bad and vice versa.

What is interesting is that ox LDL DOES NOT correlate with HSCRP. Here we are measuring two different phenonema. And it looks like combining two measures TC/HDLC (OR LDL particle number OR ox LDL) AND HSCRP enhances predictability of future CV events.

homertobias said...


I always read your references, thank you. This time both of your references link to the first one.

Brock Cusick said...

Pastured butter and avocados? Looks like I'm in the clear then! :-)

Stephan, your posts continue to be the most informationally dense and informative blogging on the net. Thank you so much for doing this. I would be so much poorer without your contributions.

Did you find a link between Omega-6 consumption and Omega-6 concentrations in LDL? Because that would explain the link between their consumption and heart risk. I'm also interested in Inphidel's question about why LDL becomes small.

Stephan Guyenet said...


From what I can tell, oxLDL is equivalent to or better than hsCRP as a cardiovascular risk marker.


Low-fat diets, genetics and probably other factors.


You have to look at studies that were done since the development of sensitive assays for oxLDL. All but one that I've looked at indicated that oxLDL is superior to conventional cholesterol markers and equivalent to or better than hsCRP. In this study, they only broke participants up into tertiles but the upper tertile still had a 4.25-fold higher risk than the lower tertile. Imagine how big the RR would have been if they had broken it up into quintiles.

oxLDL does have some relationship with other lipid markers, but it is still an "independent" predictor of CHD events. The composition of the LDL particle, not just the amount of circulating LDL, determines the susceptibility to oxidation. So diet can influence oxLDL independently of LDL particle number.

I think it's interesting you say that CRP isn't related to oxLDL, do you have a reference for that? As far as I can tell, atherosclerosis is a combination of oxLDL plus inflammation.

Stephan Guyenet said...


Dietary linoleic acid increases LDL oxidation. I'll have to look into how it affects LDL composition, but I have to assume it increases the proportion of linoleic acid in LDL. There's data for lipoprotein composition in the Kitavans that I can draw from.

TedHutchinson said...

@ homertobias
Effects of Equol on Oxidized Low-Density Lipoprotein-Induced Apoptosis in Endothelial CellsSorry I forgot to check the link.

Dave said...

Found this via a quick Googling, might be relevant to some of the questions above:

TedHutchinson said...

Clickable version of Dave's link
Whole-body insulin sensitivity, low-density lipoprotein (LDL) particle size, and oxidized LDL in overweight, nondiabetic men

Dave said...

Thanks Ted.

Robert Andrew Brown said...

Stephan said

"Dietary linoleic acid increases LDL oxidation. I'll have to look into how it affects LDL composition, but I have to assume it increases the proportion of linoleic acid in LDL. There's data for lipoprotein composition in the Kitavans that I can draw from."

Linoleic acid content does appear to vary with intake.

The Kitavans had a low Omega 6 intake maybe 1/2% and still quite a high amount of Omega 6 LA in LDL which suggests that a Omega 6 is an important constituent of LDL. The body seems to seek to maintain Omega 6 at a certain minimum level.

The amount of Omega 6 content will also depend on the proportional amount of Omega 9 and other fats.

The amount of Omega 3 in LDL will also increase quite significantly with a relatively modest increase in intake.

Author Omega Six The Devils Fat


(Revised version off to the editor shortly)

sverlyn said...

OT, but if you're looking for grass fed beef, avoid Slankers. I ordered from him and it arrived defrosted and stank to high heaven- I had to toss the whole thing
He refused to accomodate me at all for the poor packaging, so I contacted my CC and disputed the charge. IN hios response to the CC eompany he lied saying that I never contacted him- I have phone records to prove the opposite

Extremely unethical

homertobias said...


When triglycerides and VLDL particles are plentiful they interact with LDL and cause the formation of small dense LDL. Eating refined carbohydrates, high fructose corn syrup, wheat, fructose,omega 6 PUFA's all increase triglycerides and VLDL particles and produce small dense particles.
OxLDL and HSCRP measure two different things. My take is that oxLDL makes a mess in the artery wall and HSCRP makes the mess unstable so that it erupts causing a CV event.
Your link is incomplete and right now I don't have the time to comb through CIRCULATION.
RE references that HSCRP and oxLDL are unrelated markers and do not co-vary: For one REREAD Your MENSINGER reference. It is right in there. No relationship between the two. Also in the Mensinger article is the conclusion that although oxLDL is a great marker, it really is just a fancy more expensive alternative to TC/HDL measurement and is NOT independent.
You picked that reference.
References you might like:
CRP and other emerging biomarkers to optimize risk stratification, S. Tsimikas,JACC 2006. This compares hscrp,oxldl,and other candidates of which there are many(LPPLA2, MMP9,soluable adhesion molecule,apoClll,etc,etc) I can only get the abstract and am jealous of your university connections.
Also of course Paul Ridler, JACC 2007 49:2129-2138 "CRP & prediction of CV events amoung those at intermediate risk." Cites 20 large randomized trials showing that HSCRP is an INDEPENDENT risk factor for CV events. Just look past the statin references.

Natalia said...

Hi, Stephan.
Long time reader, first time responder. Like your blog a lot.

I've read somewhere that CRP helps to load cholesterol from brocken sells on to HDL to be taken back to liver. Isn't it just a scavenger and innocent marker of inflamation?
Sorry, English is not my first language.

Stephan Guyenet said...


In the Meisinger paper, the RR for oxLDL was 4.25 between upper and lower tertiles. The RR for TC/HDL was 2.32. They were not equivalent.

The investigators said "Plasma oxLDL was the strongest predictor of CHD events compared with a conventional lipoprotein profile and other traditional risk factors for CHD." It was a stronger predictor than anything else they measured, including CRP.

You said that oxLDL is not an independent predictor of cardiac events. That is not what Meisinger et al. found, and it's not what the majority of the papers on oxLDL have found. They stated that it was an independent predictor after adjustment for the other lipid values they measured.

But in any case, too much emphasis is placed on whether or not something is independent. That only matters from a clinical diagnosis perspective, not from a mechanistic perspective. Even if oxLDL weren't independent of TC/HDL, that would probably suggest TC/HDL is just a marker of oxLDL, not the other way around. oxLDL is the marker with the mechanism behind it.

Stephan Guyenet said...


By the way, the paper you cited seems to be free full text:

C-Reactive Protein and Other Emerging Blood Biomarkers to Optimize Risk Stratification of Vulnerable Patients

Drs. Cynthia and David said...

Thanks for another great post. I came across an article that was originally available in draft form as full text. I like the first three sentences: "Atherosclerosis was once thought to be a simple lipid storage disease that caused pathology by arterial obstruction. The pathology leads to arterial obstruction, but not simply by accumulation of lipid. Atherosclerotic lesions are foci of vessel wall inflammation (1-3)."

I can send you the full text if you want it. Anyway, the article is about innate immunity in the form of Toll-like Receptors residing on endothelial cells that have become stimulated to become proinflammatory. Apparently the cells are sensitive to blood flow, and in locations where they sense turbulence, they are stimulated and express TLRs on their surfaces. Certain ligands bind to the TLRs and stimulate the endothelial cells into further proinflammatory condition. Oxidized LDL is one culprit. Bacterial endotoxin is another very potent stimulator.

The article is pretty interesting, but is mainly about mouse models, with and without LDL receptors, with and without TLRs (you can't make atherosclerotic mice without TLRs). The infectious disease theory of atherosclerosis may fit in here too, though it's not clear from what I've read whether they're talking about proinflammatory microbial products such as endotoxin or actual microbes invading the artery walls. Anyway, it's all fascinating.

Thanks again.


Bris said...

Vitamin E and a-Lipoic Acid Supplementation Increase Bleeding Tendency via an Intrinsic Coagulation Pathway
Clinical and Applied Thrombosis/Hemostasis, Vol. 12, No. 2, 169-173 (2006)
DOI: 10.1177/107602960601200204

"Dietary supplementation of vitamin E and a-lipoic acid increases bleeding tendency via inhibition of the intrinsic coagulation pathway with no change in markers of lipid peroxidation. Such supplementation could benefit patients with cardiovascular disease who exhibit elevated levels of coagulation and oxidative stress."

Stephan Guyenet said...


CRP is an inflammatory mediator released by the liver. It prevents cholesterol from exiting macrophages in cell culture.

homertobias said...

Mensinger article:
"However oxLDL did not significantly increase the prediction of a coronary event. After inclusion of oxLDL in a model containing CRP,TC/HDLC ratio and other CV disease risk factors the additional improvement in risk prediction was rather low."
See also:
Is plasma oxidized LDL and independent predictor of CAD amoung US men & women? JACC 2006 Vol 48 5 pp973-982 T Wu:
"Our results suggest that circulating oxLDL measured by AB4E6 is not an independent predictor of CHD. It is less predictive than apoB or TC/HDLC ratio."
Which brings up another point. OxLDL is a heterogeneous group of compounds measured by different antibody elisa assays. The problem becomes which one are you going to use......all of them?
I think we are going at this from two different perspectives. You are explaining the mechanism of the origins of cardiovascular disease and relating it to dietary fat consumption subtypes. Oxldl works well here. I am thinking about how much the lab bill will cost. And believe me, if my oxLDL level doesn't kill me, my lab bill will give me a heart attack.

Stephan Guyenet said...


"Furthermore, we investigated whether oxLDL would predict
future CHD events independent of CRP and TC/HDL-C.
Analyses were restricted to 320 persons (81 cases and 239
controls) because of missing data for CRP. When assessed in
a separate model (Figure 1), CRP was a powerful predictor in
multivariable analysis (HR for men in the upper tertile
compared with the lower tertile was 2.64; 95% CI, 1.23 to
5.66; P=0.013). The corresponding value for oxLDL was
3.02 (95% CI, 1.33 to 6.86; P=0.008; Figure 2). When
oxLDL and CRP were simultaneously assessed in the same
model, both parameters still predicted future CHD events,
even after multivariable adjustment."

The only thing oxLDL was not "independent" of was a model including every single other marker they measured, including TC/HDL, CRP, hypertension, weight, physical activity, etc. oxLDL was definitely independent of TC/HDL, and other lipid markers, as stated above.

oxLDL was the most predictive single marker they measured.

oxLDL has been a consistent independent predictor of cardiac events, with the exception of the Wu et al. paper you cited. That is the only exception I'm aware of that looked at cardiac events in a prospective fashion.

toddhargrove said...

Anyone want to comment on this new study:

"High Cholesterol Linked to Alzheimer's"

Bris said...

Anyone want to comment on this new study:

"High Cholesterol Linked to Alzheimer's"

Translation:unhealthy people who just happen to have high total cholesterol at greater risk.

Ellen said...

Maybe it's been discussed already but how would you ask for an oxidized LDL test? What's it called since it's not included in the lipid panel.

TedHutchinson said...

full text PDF of Midlife Serum Cholesterol and Increased
Risk of Alzheimer’s and Vascular
Dementia Three Decades Later

Bris said...

The Kuopio Ischaemic Heart Disease Risk Factor Study

"In a Cox proportional hazards model adjusting for age, examination years and all other predictive coronary disease risk factors, blood donors (middle-aged males had a 88% reduced risk (relative hazard = 0.12, 95% confidence interval 0.02–0.86, p = 0.035) of acute myocardial infarction, compared with non-blood donors. "

88% less risk of an infarct by donating blood!

Bris said...

Heres the paper - I forgot to give the details.
American Journal of Epidemiology Vol. 148, No. 5: 445-451

Donation of Blood Is Associated with Reduced Risk of Myocardial Infarction
The Kuopio Ischaemic Heart Disease Risk Factor Study

TedHutchinson said...

Donation of Blood Is Associated with Reduced Risk of Myocardial Infarction
The Kuopio Ischaemic Heart Disease Risk Factor Study

Full free text pdf link.

TedHutchinson said...

In the discussion they point to
Lowering of body iron stores by blood letting and oxidation resistance of serum lipoproteinsThese observations indicate that the reduction of body iron stores by venesection can increase the oxidation resistance of serum VLDL/LDL in regularly smoking men.

Dave said...

"High Cholesterol Linked to Alzheimer's"

Correlation is not causation.

I love the fantastically stupid comment by Rachel Whitmer in the WebMD article. Great example of the lack of critical thinking amongst health-care professionals.

Lynn M. said...


These were the quotes I saw attributed to Dr. Whitmer in the WebMD article:
"People tend to think of the brain and the heart as totally separate, but they are not," study co-author Rachel A. Whitmer, PhD of Kaiser Permanente Division of Research in Oakland, Calif., tells WebMD. "We are learning that what is good for the heart is also good for the brain -- and that midlife is not too soon to be thinking about risk factors for dementia."

"But it is safe to assume that most people whose total cholesterol was high had high levels of bad cholesterol because about two-thirds of total cholesterol reflects LDL, Whitmer says."

I guess my critical thinking is on the blink, because I don't see what is fantastically stupid about those comments. I'm not saying I agree with the conclusions of the study,and I do understand that correlation is not causation. I thought I had seen the idea that what is good for the heart is good for the brain substantiated elsewhere. Is that idea wrong or unproven?

The second comment I see as either true or false, not stupid or smart. Is that statement wrong?

Dave said...

@Lynn M.

Heart and brain are all part of the same system, with lots of shared feedback and control mechanisms, common supply of oxygen, nutrients, etc. That anyone (particularly scientists, who ought to know better) would think of them as being completely separable reflects a lack of much thinking at all.

Second point is that it is implied that the common cause of heart disease and Alzheimer's is high cholesterol, as opposed to that being a symptom of a systemic disorder affecting both the blood and brain, such as hyperglycemia or hyperinsulinemia. Associative evidence carries little weight toward causality, unless you at least have a hypothetical mechanism by which excess serum cholesterol leads to Alzheimer's.

Mamatha said...

I've been following your blog for the last couple of weeks. I decided to delurk today and express my gratitude for all the great work you are doing.

Best wishes,

MangoManDan said...

Is oxidized cholesterol in food also a concern? I remember reading that it was present in powdered skim milk. It seemed strange to me that milk without fat could have any sort of cholesterol in it.

Stephan Guyenet said...


I appreciate your support.


I don't really know about the effects of oxidized cholesterol in food at this point.

karl said...

Totally missing from this is dietary sources of oxy-chol which we know is a cause of CAD.

Only in the west, where heart disease is increasing, do we eat food that has been stored for so long. We know that powered milk, powdered eggs are a source of dietary oxy-chol. We also know that refrigerated meats have more an more oxy-chol as time passes. We also know that fast cooking methods limit the amount of oxy-chol in our food.

Dietary cholesterol is not the devil - it is dietary oxy-chol that needs looking at. The few studies of food content are from Australia. Many questions about food storage and preparation in association with oxy-chol are not answered.

It is not clear what the relative risk of dietary-oxy-chol vs oxy-chol that forms in our body are.

TedHutchinson said...

I found this interesting reading today. First a laymans guide to the paper from Spacedoc's website
Infections and Vulnerable Plaques
and you'll find the full text of Ravnskov U, McCully KS. paper here on the nutrition and metabolism society website.
Vulnerable Plaque Formation from Obstruction of Vasa Vasorum by Homocysteinylated and Oxidized Lipoprotein Aggregates Complexed with Microbial Remnants and LDL Autoantibodies

karl said...

Have you found anyway to get one's blood tested for OXY-LDL?

My hunch is we should be looking at LDL-P and oxy-LDL

digger said...

I an not a chemist or have any scientific background so please excuse my laymans understanding.

However I have a grave concern as my daughter is considering undergoing treatment for fat reduction using Radio Frequency and Ultra Sound treatment which both rely on heat.
It is my understanding that the heat will oxidise both the fat and cholesterol and lead to blocking of the arteries. The amount of fat lost in each session by this liquifying process also gives me concern that the liver will not be able to process this effectively in bulk and it will migrate to other organs.

karl said...

Have you found anyway to get one's blood tested for OXY-LDL?


Vignani said...

At present no one knows what is Ox-LDL. Take my word-I was one of the co-discoverers of Ox-LDL. Before we could understand it better, tocopherol advocates did several shot gun approach studies and shot down the hypothesis prematurely. Now vitamin E advocates are distancing themselves from tocopherols!
Measurement of "ox-LDL" is real and could be an important predictor-but we have a long way to go. Research isn dwindling on the topic.

Alexanderson12 said...

Hey everyone, great discussion here, hopefully people are still viewing this post and commenting!

Can someone explain how and why Macrophages ingest oxidised LDL?

Many thanks,


Heart health quest said...

Dear Stephan,
Thank you for your very informative blog. I myself suffer from atherosclerosis and I am looking for new information that will allow me to decrease or even stop taking statins with all the side effects. I will look forward to your next post on diet and other ways to help.

tusker said...

This post is called "The Diet-Heart Hypothesis: Oxidized LDL, Part I". I can't find a Part II, does it exist?