June 13, 2012

Cardiovascular Disease

The straight dope on cholesterol – Part VII

In this post we’ll address the following concept: Does “HDL” matter after all?

Read Time 15 minutes

Previously, in Part I, Part II, Part III, Part IV, Part V ,and Part VI of this series, we addressed these 8 concepts:

     #1What is cholesterol?

     #2What is the relationship between the cholesterol we eat and the cholesterol in our body?

     #3Is cholesterol bad?

     #4 How does cholesterol move around our body?

     #5 How do we measure cholesterol?

     #6How does cholesterol actually cause problems?

     #7Does the size of an LDL particle matter?

     #8 Why is it necessary to measure LDL-P, instead of just LDL-C?

(No so) Quick refresher on take-away points from previous posts, should you need it:

  1. Cholesterol is “just” another fancy organic molecule in our body but with an interesting distinction: we eat it, we make it, we store it, and we excrete it – all in different amounts.
  2. The pool of cholesterol in our body is essential for life.  No cholesterol = no life.
  3. Cholesterol exists in 2 formsunesterified or “free” (UC) and esterified (CE) – and the form determines if we can absorb it or not, or store it or not (among other things).
  4. Much of the cholesterol we eat is in the form of CE. It is not absorbed and is excreted by our gut (i.e., leaves our body in stool). The reason this occurs is that CE not only has to be de-esterified, but it competes for absorption with the vastly larger amounts of UC supplied by the biliary route.
  5. Re-absorption of the cholesterol we synthesize in our body (i.e., endogenous produced cholesterol) is the dominant source of the cholesterol in our body. That is, most of the cholesterol in our body was made by our body.
  6. The process of regulating cholesterol is very complex and multifaceted with multiple layers of control.  I’ve only touched on the absorption side, but the synthesis side is also complex and highly regulated. You will discover that synthesis and absorption are very interrelated.
  7. Eating cholesterol has very little impact on the cholesterol levels in your body. This is a fact, not my opinion.  Anyone who tells you different is, at best, ignorant of this topic.  At worst, they are a deliberate charlatan. Years ago the Canadian Guidelines removed the limitation of dietary cholesterol. The rest of the world, especially the United States, needs to catch up.  To see an important reference on this topic, please look here.
  8. Cholesterol and triglycerides are not soluble in plasma (i.e., they can’t dissolve in water) and are therefore said to be hydrophobic.
  9. To be carried anywhere in our body, say from your liver to your coronary artery, they need to be carried by a special protein-wrapped transport vessel called a lipoprotein.
  10. As these “ships” called lipoproteins leave the liver they undergo a process of maturation where they shed much of their triglyceride “cargo” in the form of free fatty acid, and doing so makes them smaller and richer in cholesterol.
  11. Special proteins, apoproteins, play an important role in moving lipoproteins around the body and facilitating their interactions with other cells.  The most important of these are the apoB class, residing on VLDL, IDL, and LDL particles, and the apoA-I class, residing for the most part on the HDL particles.
  12. Cholesterol transport in plasma occurs in both directions, from the liver and small intestine towards the periphery and back to the liver and small intestine (the “gut”).
  13. The major function of the apoB-containing particles is to traffic energy (triglycerides) to muscles and phospholipids to all cells. Their cholesterol is trafficked back to the liver. The apoA-I containing particles traffic cholesterol to steroidogenic tissues, adipocytes (a storage organ for cholesterol ester) and ultimately back to the liver, gut, or steroidogenic tissue.
  14. All lipoproteins are part of the human lipid transportation system and work harmoniously together to efficiently traffic lipids. As you are probably starting to appreciate, the trafficking pattern is highly complex and the lipoproteins constantly exchange their core and surface lipids.
  15. The measurement of cholesterol has undergone a dramatic evolution over the past 70 years with technology at the heart of the advance.
  16. Currently, most people in the United States (and the world for that matter) undergo a “standard” lipid panel, which only directly measures TC, TG, and HDL-C.  LDL-C is measured or most often estimated.
  17. More advanced cholesterol measuring tests do exist to directly measure LDL-C (though none are standardized), along with the cholesterol content of other lipoproteins (e.g., VLDL, IDL) or lipoprotein subparticles.
  18. The most frequently used and guideline-recommended test that can count the number of LDL particles is either apolipoprotein B or LDL-P NMR, which is part of the NMR LipoProfile.  NMR can also measure the size of LDL and other lipoprotein particles, which is valuable for predicting insulin resistance in drug naïve patients, before changes are noted in glucose or insulin levels.
  19. The progression from a completely normal artery to a “clogged” or atherosclerotic one follows a very clear path: an apoB containing particle gets past the endothelial layer into the subendothelial space, the particle and its cholesterol content is retained, immune cells arrive, an inflammatory response ensues “fixing” the apoB containing particles in place AND making more space for more of them.
  20. While inflammation plays a key role in this process, it’s the penetration of the endothelium and retention within the endothelium that drive the process.
  21. The most common apoB containing lipoprotein in this process is certainly the LDL particle. However, Lp(a) and apoB containing lipoproteins play a role also, especially in the insulin resistant person.
  22. If you want to stop atherosclerosis, you must lower the LDL particle number. Period.
  23. At first glance it would seem that patients with smaller LDL particles are at greater risk for atherosclerosis than patients with large LDL particles, all things equal.
  24. “A particle is a particle is a particle.”  If you don’t know the number, you don’t know the risk.
  25. With respect to laboratory medicine, two markers that have a high correlation with a given outcome are concordant – they equally predict the same outcome. However, when the two tests do not correlate with each other they are said to be discordant.
  26. LDL-P (or apoB) is the best predictor of adverse cardiac events, which has been documented repeatedly in every major cardiovascular risk study.
  27. LDL-C is only a good predictor of adverse cardiac events when it is concordant with LDL-P; otherwise it is a poor predictor of risk.
  28. There is no way of determining which individual patient may have discordant LDL-C and LDL-P without measuring both markers.
  29. Discordance between LDL-C and LDL-P is even greater in populations with metabolic syndrome, including patients with diabetes.  Given the ubiquity of these conditions in the U.S. population, and the special risk such patients carry for cardiovascular disease, it is difficult to justify use of LDL-C, HDL-C, and TG alone for risk stratification in all but the most select patients.
  30. To address this question, however, one must look at changes in cardiovascular events or direct markers of atherosclerosis (e.g., IMT) while holding LDL-P constant and then again holding LDL size constant.  Only when you do this can you see that the relationship between size and event vanishes.  The only thing that matters is the number of LDL particles – large, small, or mixed.


Concept #9 – Does “HDL” matter after all?  

Last week was the largest annual meeting of the National Lipid Association (NLA) in Phoenix, AZ.  The timing of the meeting could not have been better, given the huge buzz going around on the topic of “HDL.”  (If you’re wondering why I’m putting HDL in quotes, I’ll address it shortly.)

What buzz, you ask? Many folks, including our beloved health columnists at The New York Times, are talking about the death of the HDL hypothesis – namely, the notion that HDL is the “good cholesterol.”

Technically, this “buzz” started about 6 years ago when Pfizer made headlines with a drug in their pipeline called torcetrapib.  Torcetrapib was one of the most eagerly anticipated drugs ever, certainly in my lifetime, as it had been shown to significantly raise plasma levels of HDL-C.   You’ll recall from part II of this series, HDL particles play an important role in carrying cholesterol from the subendothelial space back to the liver via a process called reverse cholesterol transport (RCT).  Furthermore, many studies and epidemiologic analyses have shown that people with high plasma levels of HDL-C have a lower incidence of coronary artery disease.

In the case of torcetrapib, there was an even more compelling reason to be optimistic.  Torcetrapib blocked the protein cholesterylester transfer protein, or CETP, which facilitates the collection and one-to-one exchange of triglycerides and cholesterol esters between lipoproteins.  Most (but not all) people with a mutation or dysfunction of this protein were known to have high levels of HDL-C and lower risk of heart disease. Optimism was very high that a drug like torcetrapib, which could mimic this effect and create a state of more HDL-C and less LDL-C, would be the biggest blockbuster drug ever.

The past month or so has seen this discussion intensify, which I’ll quickly try to cover below.


The data


After several smaller clinical trials showed that patients taking torcetrapib experienced both an increase in HDL-C and a reduction in LDL-C, a large clinical trial pitting atorvastatin (Lipitor) against atorvastatin + torcetrapib was underway.  This trial was to be the jewel in the crown of Pfizer.  It was already known that Lipitor reduced coronary artery disease (and reduced LDL-C, though this may have been a bystander effect and real reduction in mortality may be better attributed to the reduction in LDL-P).

I still remember exactly where I was standing, on the corner of Kerney St. and California St. in the heart of San Francisco’s financial district, on that December day back in 2006 when it was announced the trial had been halted because of increased mortality in the group receiving torcetrapib.  In other words, adding torcetrapib actually made things worse.  I was shocked.

Many reasons were offered for this, including the notion that torcetrapib was, indeed, helpful, but because of unanticipated side-effects, (raising blood pressure in some patients and altering electrolyte balance in others), the net impact was harmful.  Some even suggested that the drug could be useful in the “right” patients (e.g., those with low HDL-C, but normal blood pressure). Furthermore, in two subsequent studies looking at carotid IMT (thickening of the carotid arteries) and intravascular ultrasound, there was no reduction in atherosclerosis.

This was a big strike against the HDL hypothesis and work on torcetrapib was immediately halted.


Niacin has long been known to raise HDL-C and has actually been used therapeutically for this reason for many years.  The AIM-HIGH trial (Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides – you can’t have trials in medicine without catchy names!) sought to test this.  The trial randomly assigned over 3,000 patients with known and persistent, but stable and well treated cardiovascular risk, to one of two treatments:

  1. Simvastatin (40-80 mg/day), +/- ezetimibe (10 mg/day) as necessary to maintain LDL-C below 70 mg/dL + placebo (a tiny dose of crystalline niacin to cause flushing);
  2. As above, but instead of a placebo, patients were given 1,500 to 2,000 mg/day of extended-release niacin.

Both arms of the study had their LDL-C < 70 mg/dL, non-HDL-C < 100 md/dL and apoB < 80 mg/dL, but despite the statin or statin + ezetimibe treatment still had low HDL-C.  So, if niacin raised HDL-C and reduced events, the HDL raising hypothesis would be proven.

Simvastatin, as its name suggests, is a statin which primarily works by blocking HMG-CoA reductuse, an enzyme necessary to synthesize endogenous cholesterol.  Ezetimibe works on the other end of problem, by blocking the NPC1L1 transporter on gut enterocytes and hepatocytes at the hepatobiliary junction (for a quick refresher, go back to part I of this series and look at the second figure – ezetimibe blocks the “ticket taker” in the bar).

After two years the niacin group, as expected, had experienced a significant increase in plasma HDL-C (along with some other benefits like a greater reduction in plasma triglycerides).  However, there was no improvement in patient survival.  The trial was futile and the data and safety board halted the trial.  In other words, for patients with cardiac risk and LDL-C levels at goal with medication niacin, despite raising HDL-C and lowering TG, did nothing to improve survival. This was another strike against the HDL hypothesis.


By 2008, as the AIM-HIGH trial was well under way, another pharma giant, Roche, was well into clinical trials with another drug that blocked CETP.  This drug, a cousin of torcetrapib called dalcetrapib, albeit a weaker CETP-inhibitor, appeared to do all the “right” stuff (i.e., it increased HDL-C) without the “wrong” stuff (i.e., it did not appear to adversely affect blood pressure). It did nothing to LDL-C or apoB.

This study, called dal-OUTCOMES, was similar to the other trials in that patients were randomized to either standard of care plus placebo or standard of care plus escalating doses of dalcetrapib.   A report of smaller safety studies (called dal-Vessel and Dal-Plaque) was published a few months ago in the American Heart Journal, and shortly after Roche halted the phase 3 clinical trial.  Once again, patients on the treatment arm did experience a significant increase in HDL-C, but failed to appreciate any clinical benefit.  Another futile trial.

Currently, two additional CETP inhibitors, evacetrapib (manufactured by Lilly) and anacetrapib (manufactured by Merck) are being evaluated. They are much more potent CETP inhibitors and, unlike dalcetrapib, also reduce apoB and LDL-C and Lp(a). Both Lilly and Merck are very optimistic that their variants will be successful where Pfizer’s and Roche’s were not, for a number of reasons including greater anti-CETP potency.

Nevertheless, this was yet another strike against the HDL hypothesis because the drug only raised HDL-C and did nothing to apoB. If simply raising HDL-C without attacking apoB is a viable therapeutic strategy, the trial should have worked. We have been told for years (by erroneous extrapolation from epidemiologic data) that a 1% rise in HDL-C would translate into a 3% reduction in coronary artery disease.  These trials would suggest otherwise.

Mendelian randomization

On May 17 of this year a large group in Europe (hence the spelling of randomization) published a paper in The Lancet, titled, “Plasma HDL cholesterol and risk of myocardial infarction: a mendelian randomisation study.”  Mendelian randomization, as its name sort of suggests, is a method of using known genetic differences in large populations to try to “sort out” large pools of epidemiologic data.

In the case of this study, pooled data from tens of studies where patients were known to have myocardial infarction (heart attacks) were mapped against known genetic alterations called SNPs (single nucleotide polymorphisms, pronounced “snips”).  I’m not going to go into detail about the methodology because it would take 3 more blog posts., But, the reason for doing this analysis was to ferret out if having a high HDL-C was (only) correlated with better cardiovascular outcome, which has been the classic teaching, or if there was any causal relationship.  In other words, does having a high HDL-C cause you to have a lower risk of heart disease or is it a marker for something else?

This study found, consistent with the trials I’ve discussed above, that any genetic polymorphism that seems to raise HDL-C does not seem to protect from heart disease.  That is, patients with higher HDL-C due to a known genetic alteration did not seem to have protection from heart disease as a result of that gene. This suggests that people with high or low HDL-C who get coronary artery disease may well have something else at play.

Oh boy.  This seems like the last nail in the casket of the entire “HDL” hypothesis, as evidenced by all of the front page stories worldwide.


The rub: the difference between HDL-C and HDL-P

The reason I’ve been referring to high density lipoprotein as “HDL,” unless specifically referring to HDL-C, is that HDL-P and HDL-C are not the same thing.  Just as you are now intimately familiar with the notion that LDL-C and LDL-P are not the same thing, the same is true for “HDL” which simply stands for high density lipoprotein, and like LDL is not a lab assay. In fact, unpublished data from the MESA trial found that the correlation between HDL-C and HDL-P was only 0.73, which is far from “good enough” to say HDL-C is a perfect proxy for HDL-P.

HDL-C, measured in mg/dL (or mmol/L outside of the U.S.), is the mass of cholesterol carried by HDL particles in a specified volume (typically measured as X mg of cholesterol per dL of plasma). HDL-P is something entirely different.  It’s the number of HDL particles (minus unlipidated apoA-I and prebeta-HDLs: at most 5% of HDL particles) contained in a specified volume (typically measured as Y micromole of particles per liter).

As you can see in the figure below (courtesy of Jim Otvos’ presentation at the NLA meeting 2 weeks ago), the larger an HDL particle, the more cholesterol it carries. So, an equal number of large versus small HDL particles (equal HDL-P) can carry very different amounts of cholesterol (different HDL-C).  Of course, it’s never this simple because HDL particles, like their LDL counterparts, don’t just carry cholesterol.  They carry triglycerides, too. Keep in mind, HDL core CE/TG ratio is about 10:1 or greater – if the large HDL carries TG, it will not be carrying very much cholesterol.

HDL sizes


So, the important point is that HDL-C is not the same as HDL-P (which is also not the same as apoAI, as HDL particles can carry more than one apoAI).


But there’s something else going on here.  If you look at the figure below, from the Framingham cohort, you’ll note something interesting.  As HDL-C rises, it does so not in a uniform or “across the board” fashion.  A rise in HDL-C seems to disproportionately result from an increase in large HDL particles.  In other words, as HDL-C rises, it doesn’t necessarily mean HDL-P is rising at all, and certainly not as much.


HDL-C changes with HDL-P

As you can see, for increases in HDL-C at low levels (i.e., below 40 mg/dL) the increase in small particles seems to account for much of the increase in total HDL-P, While for increases over 40 mg/dL, the increase in large particles seems to account for the increase in HDL-C. Also note that as HDL-C rises above 45 mg/dL, there is almost no further increase in total HDL-P – the rise in HDL-C is driven by enlargement of the HDL particle – more cholesterol per particle – not the drop in small HDL-P.  This reveals to us that the small HDL particles are being lipidated.


Is there a reason to favor small HDL particles over large ones?

In the 2011 article, “Biological activities of HDL subpopulations and their relevance to cardiovascular disease,” published in Trends in Molecular Medicine, the authors describe in great detail some of protective mechanisms imparted by HDL particles.

Large HDL particles may be less protective and even dysfunctional in certain pathological states, whereas small to medium-sized HDL particles seem to confer greater protection through the following mechanisms:

  • Greater antioxidant activity
  • Greater anti-inflammatory activity
  • Greater cholesterol efflux capacity
  • Greater anti-thrombotic properties

In other words, particle for particle, it seems a small HDL particle may be better at transporting cholesterol from the subendothelial space (technically, they acquire cholesterol from cholesterol-laden macrophages or foam cells in the subedothelial space) elsewhere, better at reducing inflammation, better at preventing clotting, and better at mitigating the problems caused by oxidative free radicals.

Of course, reality is complicated.  If there was no maturation from small to large HDL particles (i.e., the dynamic remodeling of HDL), the system would be faulty. So, the truth is that all HDL sizes are required and that HDL particles are in a constant dynamic state (or “flux”) of lipidating and delipidating, and the real truth is no particular HDL size can be said to be the best. If the little HDLs do not enlarge, the ApoA-I mediated lipid trafficking system is broken.


The truth about the old (and overly simplistic) term called reverse cholesterol transport (RCT)

HDL particles traffic cholesterol and proteins and last in plasma on average for 5 days. They are in a constant state of acquiring cholesterol (lipidation) and delivering cholesterol (delipidation). There are membrane receptors on cells that can export cholesterol to HDL particles (sterol efflux transporters) or extract cholesterol or cholesterol ester from HDL particles (sterol influx transporters).

The vast majority of lipidation occurs (in order): 1) at the liver, 2) the small intestine, 3) adipocytes and 4) peripheral cells, including plaque if present. The liver and intestine account for 95% of this process. The amount of cholesterol pulled out of arteries (called macrophage reverse cholesterol transport) is critical to disease prevention but is so small it has no effect on serum HDL levels. Even in patients with extensive plaque, the cholesterol in that plaque is about 0.5% of total body cholesterol. HDL particles circulate for several days as a ready reserve of cholesterol: almost no cell in humans require a delivery of cholesterol as cells synthesize all they need.  However, steroidogenic hormone producing tissues (e.g., adrenal cortex and gonads) do require cholesterol and the HDL particle is the primary delivery truck.

If, as is the case in a medical emergency, the adrenal gland must rapidly make a lot of cortisone, the HDL particles are there with the needed cholesterol.  This explains the low HDL-C typically seen in patients with severe infections (e.g., sepsis) and severe inflammatory conditions (e.g., Rheumatoid Arthritis).

Sooner or later HDL particles must be delipidated, and this takes place at: 1) the adrenal cortex or gonads 2) the liver, 3) adipocytes, 4) the small intestine (TICE or transintestinal cholesterol efflux) or give its cholesterol to an apoB particle (90% of which are LDLs) to return to the liver.  A HDL particle delivering cholesterol to the liver or intestine is called direct reverse cholesterol transport (RCT), whereas a HDL particle transferring its cholesterol to an apoB particle which returns it to the liver is indirect RCT. Hence, total RCT = direct RCT + indirect RCT.

The punch line: a serum HDL-C level has no known relationship to this complex process of RCT. The last thing a HDL does is lose its cholesterol.  The old concept that a drug or lifestyle that raises HDL-C is improving the RCT process is wrong; it may or may not be affecting that dynamic process. Instead of calling this RCT, it would be more appropriately called apoA-I trafficking of cholesterol.


Why do drugs that specifically raise HDL-C seem to be of little value?

As I’ve argued before, while statins are efficacious at preventing heart disease, it’s sort of by “luck” as far as most prescribing physicians are concerned. Most doctors use cholesterol lowing medication to lower LDL-C, not LDL-P.  Since there is an overlap (i.e., since the levels of LDL-P and LDL-C are concordant) in many patients, this misplaced use of statins seems to work “ok.”  I, and many others far more knowledgeable, would argue that if statins and other drugs were used to lower LDL-P (and apoB), instead of LDL-C, their efficacy would be even greater.  The same is true for dietary intervention.

Interestingly, (and I would have never known this had Jim Otvos not graciously spent a hour on the phone with me two weeks ago giving me a nuanced HDL tutorial), a study that went completely unnoticed by the press in 2010, published in Circulation, actually did a similar analysis to the Lancet paper, except that the authors looked at HDL-P instead of HDL-C as the biomarker and looked at the impact of phospholipid transfer protein (PLTP) on HDL metabolism.  In this study, though not the explicit goal, the authors found that an increase in the number of HDL particles and smaller HDL particles decreased the risk of cardiovascular disease.   The key point, of course, is that the total number of HDL particles rose, and it was driven by increased small HDL-P. The exact same thing was seen in the VA-HIT trial: the cardiovascular benefit of the treatment (fibrate) was related to the rise in total HDL-P which was driven by the fibrates’ ability to raise small HDL-P.

It seems the problem with the “HDL hypothesis” is that it’s using the wrong marker of HDL.  By looking at HDL-C instead of HDL-P, these investigators may have missed the point.  Just like LDL, it’s all about the particles.



  1. HDL-C and HDL-P are not measuring the same thing, just as LDL-C and LDL-P are not.
  2. Secondary to the total HDL-P, all things equal it seems smaller HDL particles are more protective than large ones.
  3. As HDL-C levels rise, most often it is driven by a disproportionate rise in HDL size, not HDL-P.
  4. In the trials which were designed to prove that a drug that raised HDL-C would provide a reduction in cardiovascular events, no benefit occurred:  estrogen studies (HERS, WHI), fibrate studies (FIELD, ACCORD), niacin studies, and CETP inhibition studies (dalcetrapib and torcetrapib).  But, this says nothing of what happens when you raise HDL-P.
  5. Don’t believe the hype: HDL is important, and more HDL particles are better than few. But, raising HDL-C with a drug isn’t going to fix the problem. Making this even more complex is that HDL functionality is likely as important, or even more important, than HDL-P, but no such tests exist to “measure” this.


Two apolipoprotein A1 chains (magenta ribbons) complexed with cholesterol (orange balls) and phospholipids, after PDB 3K2S by Ayacop [Public domain], via Wikimedia Commons

Disclaimer: This blog is for general informational purposes only and does not constitute the practice of medicine, nursing or other professional health care services, including the giving of medical advice, and no doctor/patient relationship is formed. The use of information on this blog or materials linked from this blog is at the user's own risk. The content of this blog is not intended to be a substitute for professional medical advice, diagnosis, or treatment. Users should not disregard, or delay in obtaining, medical advice for any medical condition they may have, and should seek the assistance of their health care professionals for any such conditions.


  1. Sorry, still dont get it. You are what you eat! If you enjoy a bit deep fried meal your plasma colesterol jumps up as does the lipid content in general come to that? I’d love to see the data proving the contrary, Thanks

    • Overly simplistic view of a much more complex process. Go back to part I of this series if you want to understand the interaction between ingested cholesterol and serum cholesterol.

  2. I don’t have access to an NMR test so I am looking at ApoB to estimate particle number. How can I translate from ApoB to LDL-P? The ApoB result is measured in mg/dL.

    The graph of LDL-P distribution showing percentiles (like the one from the MESA trial) is useful so my end goal is to approximate where an ApoB result would sit on this graph.

    Thanks. I’ve really enjoyed this series of posts.

  3. Hi Peter,

    I haven’t seen any commentary by you about the role homocysteine plays in all this. Dr. Kilmer McCully’s pioneering work in 1968 on homocysteine was prior to the growing domination of the cholesterol theory. As the cholesterol theory gained momentum, he was demonized, stripped of tenure and ultimately fired and black-listed. In the 1990s, international studies confirmed his earlier work and he began to be given credibility again.

    Essentially, as I understand it, his work indicates that elevated homocysteine levels cause the inflammation that is necessary for atherosclerosis to develop by enabling cholesterol to enter the arterial walls.

    His “cure” for atherosclerosis is to avoid the all too common dietary deficiency of B6, B12 and folic acid by dietary means if possible and supplement them if not possible due to veganism or some other reason that limits access to animal products.



    P. S. Some of his papers over the years:

    1. McCully K S, “Vascular pathology of homocysteinemia: implications for the pathogenesis of arteriosclerosis”, Am. J. Pathol. (1969), 56: pp. 111–128.
    2. McCully K S, “Chemical pathology of homocysteine”, I. Atherogenesis Ann. Clin. Lab. Sci. (1993), 23: pp. 477–493.
    3. McCully K S, “Homocysteine and vascular disease”, Nature Med. (1996), 2: pp. 386–389.
    4. McCully K S, “Homocysteine, folate, vitamin B6 and cardiovascular disease”, J.Am. Med.Assoc. (1998), 279: pp. 392–393.
    5. McCully K S,“Atherosclerosis, serum cholesterol and the homocysteine theory: a study of 194 consecutive autopsies”, J.Am. Med. Sci. (1990), 299: pp. 217–221.
    6. McCully K S,“Keeping the Young-Elderly Healthy. Homocysteine, vitamins, and vascular disease prevention”, Am. J. Clin. Nutr. Nov. 2007 Vol. 86, No. 5, 1563S-8S https://www.ajcn.org/cgi/content/full/86/5/1563S.

  4. I’m still 100% convinced its all about avoiding inflammation. Your Cholesterol just doesn’t get under the “cracks” unless the cracks show-up first – you need a reason the cholesterol shows up and gets “stuck”. Avoid the “cracks” and avoid the problems. How’s that for a simple conclusion 🙂

    Once you have inflammation – your cholesterol numbers and size doesn’t matter – you may postpone the inevitable, but by how much? What if inflammation triggers “bad cholesterol” to be produced in the first place?

    • If this were true, how would it explain the association between apoB and CHD? If it were only about inflammation, as you hypothesize, any association with apoB would vanish or be so significantly diluted that we could never see it. Yet we see it loud and clear?

    • This debate between inflammation v. cholesterol as a cause of heart disease is a prevalent theme in the health blogosphere. The “cholesterol deniers” champion the inflammation hypothesis while most of the medical/pharmaceutical community goes with cholesterol. Unfortunately, nobody seems to be pointing out that there is a much more sophisticated history of competing theories in the medical research literature- the “Response to Retention” v. the “Response to Injury” hypotheses which pretty much parallel the blogosphere debate. It is just not possible to review the extensive literature here but we can say, albeit grossly oversimplified, that the Retention hypothesis corresponds pretty much to the “LDL particles get stuck in the arterial walls” idea while the Injury hypothesis mirrors the “inflammation damages the arterial wall” concept. As often the case, there may not be a stark dichotomy between the two explanations but the most important observation I can make is in reference to Peter’s “Take Away Points” 19-20:

      “The progression from a completely normal artery to a “clogged” or atherosclerotic one follows a very clear path: an apoB containing particle gets past the endothelial layer into the subendothelial space, the particle and its cholesterol content is retained, immune cells arrive, an inflammatory response ensues “fixing” the apoB containing particles in place AND making more space for more of them….While inflammation plays a key role in this process, it’s the penetration of the endothelium and retention within the endothelium that drive the process.”

      This clearly comes down hard on the side of the “Response to Retention” hypothesis and neatly, perhaps too neatly, meshes with LDL-P as the critical variable. My objection is that this is presented as if it was settled fact instead of the subject of ongoing debate and research. The “association between apoB and CHD” is NOT conclusive evidence on the side of Retention. Since LDL cholesterol has immune functions, immune signaling is a possible explanation for the association, as noted by Lorraine in the above comments, and could also support the Injury hypothesis (Some interesting research in this area was done in Russia in the 1990’s showing that cholesterol in circulating immune complexes was a much more reliable marker of coronary atherosclerosis than standard measures).

      Also ignored in the series is the role of serum oxidized cholesterol which has also been shown to be a superior marker for atherosclerosis over standard measures as well as having the ability to move back and forth between vessel wall and the circulation.

      All of this cries out for far more discussion and debate than has been given here.

  5. Eating cholesterol has very little impact on the cholesterol levels in your body.
    Yes, but can the same be said about saturated fat. Dr. Krauss appears to think otherwise and suggests caution with regard to red meat. I hope you discuss this as it appears that diet via sat fat may affect cholesterol levels and number of particles. Thanks

  6. Were I near by I would absolutely attend! Living in New York City makes the logistics a little tough! Come out this way and I will be there for sure! Hope you address the question in a subsequent cholesterol post. Am guessing it is on the minds of many who have cut carbs( low or vlc) only to find sky rocketing LDL-C and LDL-P

  7. ? Peter–So, is this work simply wrong from heart.org??

    June 20, 2012
    View the most recent content at: https://www.theheart.org
    Most recent news
    Measuring apolipoproteins does not help risk prediction

    • From Dayspring, Dall, and others:

      All of these analyses are flawed with respect to examining the atherogenic lipoprotein variables in patients in which cholesterol measurements and lipoprotein concentration measurements are not also examined in the patients where the variables are discordant. These measures (cholesterol concentrations and apolipoprotein B) correlate highly – yet as we all know and these studies never address it, is in the many patients where the measures are discordant, apoB and LDL-P are the proven better variables to measure both risk prediction and therapeutic goals. It is also unfortunate that this study provided no LDL-P analysis. Thus these analyses might be of some interest to epidemiologists who look at entire populations and have virtually no value to practicing clinicians who treat people one at a time.

      It is difficult to make the case for apoA-I by itself in routine screening as it is not the most accurate way of quantifying total HDL-P. Thus I have little use for that measure by itself other than the investigation of patients with significant hypoalpha or hperalphalipoproteinemia. However both AMORIS (which this analysis conveniently left out) and INTERHEART (two gigantic trials) revealed that the best risk predictor was the apoB/apoA-I ratio. So in drug naive patients the ratio (which requires apoA-I measurement) is validated. No ratio is likely valid in patients on lipid modulating meds as drugs do not effect apoB and apoA-I equally nor do apoB and apoA-I have equal predictive abilities.

      With respect to inflammation markers: their appropriate use (as discussed in the recent NLA biomarker statement) is to be used not in place of lipid or lipoprotein concentrations but afterwards to better fine tune risk which several studies have shown they do. Their elevation, based on current knowledge, should lead the clinician to get more aggressive in attaining lipid and lipoprotein goals of therapy not per se any (now nonexistent) inflammatory goals of therapy. However current studies do suggest but further studies will be needed to show if it is important to also normalize at least some inflammatory markers.

      The paper actual states: “The addition of the combination apolipoprotein B and A-I, lipoprotein(a), or lipoprotein-associated phospholipaseA2to risk scores containing total cholesterol and HDL-C provided slight improvement in CVD prediction.” When you apply that slight improvement to 350 million Americans you are talking about millions of persons who would indeed benefit

      Interestingly, last year the NLA reviewed all of this data and more and came to the conclusion that apoB, LDL-P, Lp-PLA2 and Lp(a) were indeed useful in almost all folks who have greater than a 5% ten year Framingham Risk score (virtually every adult > 30-40 years old).

    • Peter,

      You indicate this response is from “Dayspring, Dall, and others” so I don’t know who I am actually responding to. I will continue nevertheless.

      I see two parts to this rebuttal to the JAMA article:


      first, that it failed to take “discordance” into account and second, that other studies show that apoB/apoA-I ratio is “the best risk predictor.” I will deal with the risk predictor argument first as stated here:

      “both AMORIS (which this analysis conveniently left out) and INTERHEART (two gigantic trials) revealed that the best risk predictor was the apoB/apoA-I ratio. ”

      In fact, neither study says that at all. INTERHEART:


      looked ONLY at apoB/apoA-I and for reasons of practicality as they wrote:

      “Because apolipoprotein concentrations are not affected by the fasting status of the individual (unlike calculated LDL), we used the ApoB/ApoA1 ratio as an index of abnormal lipids in the current analysis. 10”

      The study did not look at or compare any other predictors nor was the study designed to do that.

      As for AMORIS:

      “High apolipoprotein B, low apolipoprotein AI, and improvement in the prediction of fatal myocardial infarction (AMORIS study): a prospective study” (only abstract available online”

      Lets first deal with the rather underhanded swipe at the JAMA article authors, suggesting that there was something sinister in the omission of AMORIS. However, as they clearly state:

      “Data from the Apolipoprotein Related Mortality Risk Study (AMORIS) could not be incorporated into these current analyses because it did not measure baseline levels of HDL-C, blood pressure, smoking status, body mass index, or diabetes (eTable 5).17”

      Hardly “mysterious.” In any case AMORIS only looked at ApoB/ApoA1 compared to TC, TG, and LDL leaving out TC/HDL or Non-HDL cholesterol so it was hardly a comprehensive analysis.

      So, it does not seem that these two studies in any way speak to the superiority of ApoB/ApoA1 over the traditional markers. (As for the lament that LDL-P was not included in the JAMA study, I have spoken my piece on that already earlier.)

      You seem to want to salvage something for your case by arguing that claiming:

      “The paper actual states: “The addition of the combination apolipoprotein B and A-I, lipoprotein(a), or lipoprotein-associated phospholipaseA2to risk scores containing total cholesterol and HDL-C provided slight improvement in CVD prediction.” When you apply that slight improvement to 350 million Americans you are talking about millions of persons who would indeed benefit.”

      Aside from the silliness of applying the numbers to every, single American (what happened to your statement that not everybody needs an NMR???), unfortunately you didn’t include their analysis here:

      …we showed that each of the lipid-related markers studied herein slightly increased CVD prediction when using measures (eg, the C index and D measure) that are independent of clinical risk categories. Second, we found that none of these markers significantly improved reclassification of participants across the clinical risk cutoff levels that are currently used to inform treatment decisions. Third, we modeled a scenario assuming targeted lipid-related marker assessment in people judged as being at intermediate risk (10%- <20% 10-year predicted CVD risk) after initial screening by conventional risk factors alone. If such targeted measurement were to be coupled with allocation of statins per US Adult Treatment Panel III guidelines,24 then our data suggest that it could help prevent 1 extra CVD outcome over 10 years for approximately every 4500 people additionally screened with a combination of apolipoprotein B and A-I, …..

      Even if you accept the statin numbers, screening 4500 people to prevent one outcome, and on the basis of epidemiological evidence, is……well……madness. Thats why we have reclassification analysis, to bring some sanity to medical screening.

      Lets turn to the second argument" which is this assertion about discordance:

      "…in the many patients where the measures are discordant, apoB and LDL-P are the proven better variables to measure both risk prediction and therapeutic goals.

      I have no doubt that such measures are superior risk predictors to LDL-C but that begs the more important question. Since we already know, as I have commented in an earlier post, that discordance is closely associated with Metabolic Syndrome/Insulin Resistance, the real question is whether or not "discordance" is a better risk predictor than the usual set of markers used to diagnose MetSyn. Another way to ask that that question would be whether or not "discordance, in the absence of MetSyn/IR, adds any useful predictive information. I see no evidence for that here.

      Similarly, for those already with MetSyn/IR, we also need to ask if there is any evidence that the "therapeutic goal" (I suspect meaning medication) of reducing discordance has any value over any other treatment for the condition. I will wait for the treatment part of the series for an answer on this but, in the mean time, I am seeing a disturbing pattern here of studies not actually saying what they are purported to say. I will extend to you the benefit of the doubt since as you say, this posts taking a great deal of time to write and you are doing it in your spare time.

      • It’s a fair point you make, Hasan, about the filter of MetSyn/IR to predict discordance, and ultimately, the need for NMR. Something to consider is the following: Is there a harm in knowing someone’s LDL-P (whether concordant or discordant)? Assume, of course, that in the MetSyn/IR patients there will be much greater discordance, but as we saw in MESA, even in non-MetSyn/IR discordance can be present, albeit to a lesser degree. What I’m asking is, do we lose anything by having this information? Furthermore, when coupled with other clinical information (e.g., calcium score, Lp-PLA2, hs-CRP, LP-IR score), can a good clinician make a better decision with better targets of therapy?

    • ” Is there a harm in knowing someone’s LDL-P (whether concordant or discordant)? ”

      If you mean that knowledge in the hands of a physician the harm, aside from cost, will depend on any ensuing treatment either based on that number or because that number was added to the mix. If you mean that knowledge in the hands of blog readers, well, thats another story and you have to measure the advantage of having more informed patients v. the reality that such knowledge can end up in added expense, anxiety, and possibly treatment if the patient is insistent enough. That takes us back full circle.

      “Furthermore, when coupled with other clinical information (e.g., calcium score, Lp-PLA2, hs-CRP, LP-IR score), can a good clinician make a better decision with better targets of therapy?”

      If the targets really are better which has yet to be established but if the truth be known I kind of hate the notion of “targets.” I would rather talk about the health of people.

      ” …as we saw in MESA, even in non-MetSyn/IR discordance can be present, albeit to a lesser degree.”

      You have yet to convince me that discordance in the absence of MetSyn/IR or other risk indictors is a risk marker.

      Don’t get me wrong, I am NOT trying to minimize the significance of MetSyn/IR. In fact, I consider it to be the hallmark of our time. If this series is heading towards “lifestyle change” as an approach to this problem, I am with you. If its headed toward adding even more people to the statin pool who have yet to participate because their LDL was not high enough, I am against you. I will anticipate the response that these two approaches are not mutually exclusive but, in reality, I believe they are for all intents and purposes but thats a discussion for another day.

    • “Is there a harm in knowing someone’s LDL-P (whether concordant or discordant)?”

      Yes, there could be substantial harm, and this is one of the main reasons I think Hasan is raising questions here.

      I suspect that Hasan has been influenced by some of the same people I have, such as Dr. Gilbert Welch of Dartmouth and Dr. Nortin Hadler of UNC. They have written extensively about the harms that can come looking too hard for signs of illness in apparently well people (i.e., screening).

      When even seemingly no-brainer tests to detect cancer early (e.g., PSA test, colonoscopy, mammography) are subjected to large, long-term clinical trials, it becomes clear that the harms are very real and may very well outweigh any benefits. In these cases, the physical harms are such very serious things as incontinence, impotence, loss of breasts, radiation-induced cancer, bowel perforation, or death, from the tests themselves or the further invasive tests and interventions that inevitably follow in large numbers. On top of this are the often excruciating psychological harms of health scares, endless testing, and becoming permanently labeled as, and feeling like, a sick person instead of a well one.

      The sought-after benefits, such as preventing or successfully treating cancers that otherwise would have killed, have turned out to be so elusive that even massive, long-term trials have had great difficulty detecting them. If they exist, which does not even seem entirely clear, they have to be extremely small.

      As Dr. Welch says, even if the chance of being seriously injured by agreeing to screening is very small, so is the chance of being helped, so it is far from clear that we should all be good boys and girls and agree to the standard of care in these matters. It is a personal choice, and also a social one given the enormous costs. Welch and Hadler write extensively about how to think about these choices. I myself have declined all such tests after looking into them. I am quite happy to forgo the minute possibility that I will be saved in return for avoiding the far more likely possibility of getting sucked into the medical mill unnecessarily, with perhaps dire consequences. I have seen it happen.

      How do we know the same won’t turn out to be the case for LDL-P screening? Statin drugs, for example, are certainly not benign, and their long-term risks are not even fully understood yet. I would guess that LDL-P will ultimately fare no better than PSA, colonoscopy, standard lipid screening, and all the rest, when applied to healthy populations. There are fundamental underlying reasons for this rooted in Bayesian statistics, the natural history of most illness, and the limitations of medical interventions. Testing and treating the truly sick is one thing, often able to produce benefit in excess of harm. Testing, and then inevitably treating, the well, e.g. in search for “risk factors,” is another matter entirely.

      Peter says, “When you apply that slight improvement to 350 million Americans you are talking about millions of persons who would indeed benefit.” But what of any slight, or maybe not so slight, *harms* of the testing and treatment, when also multiplied over 350 million people?

      Maybe if applied to a small subset of clearly sick people, LDL-P could be useful, just as all those other tests can be in some circumstances. But to assume away the harms of generalized population testing could be to repeat enormous mistakes of the past that will take many years to unwind. Notice the tantrums and hysteria that result when even when staid mainstream bodies such as the U.S. Preventive Task Force recommend the slightest relaxing of heavily-promoted prostate or breast cancer screening protocols.

      • All fair points, Bill, but keep a few things in mind. The overzealous interventions that have caused so much harm with unnecessary colonoscopy, PSA, and mammography are VERY different from what I think you’re suggesting (i.e., unnecessary prescription of a drug to reduce apoB/LDL-P/LDL-C/pick-your-target). This is a night and day difference. If folks were suggesting that elevated LDL-P means a trip to the cath lab, I’d be in 100% agreement with you. That’s crazy. It’s like doing a biopsy on every man with an elevated PSA, or like doing a full colonoscopy on everyone over the age of 30. But it doesn’t. It means (in the hands of a good doctor) more screening tests to determine how great your risk is and at the very least careful monitoring for progression of this, or other markers.
        I realize you, and others, may reject the idea that a cluster of biomarkers is of value (e.g., elevated apoB, elevated LpPLA2, elevated MPO), but at some point we have to pay attention even if a person in clinically asymptomatic.

        Do you know what the most common presentation of myocardial infarction is in the U.S.? Not chest pain. Not shortness of breath. It’s sudden death. So the single most common presentation of heart disease (this was true 10 years and may not be true today) was death. If you had looked at the person the day before they died, what would you have seen? Would it have been bad enough to justify treating? What about a year before? What about 5 years before? What would have been enough to treat? How high should the calcium score get before a treatment was suggested? When is and LDL particle screaming loud enough to justify treatment?

        These are not easy questions. If they were I’m sure we wouldn’t be doing this back and forth and back and forth and back and forth and back and forth … that I’m enjoying so much.

        But keep in mind the difference between the mindless algorithmic application of “standards” and a thoughtful, individual approach to patient care. I’m not here to tell anyone to take drug X or get test Y, but I refuse to accept that information is power. With this power comes responsibility. If someone can’t handle that, fine. Perhaps the information is not right.

        We shop around like crazy to find the best car mechanic or hair dresser (well, I don’t, but I’m sure some people do). Why are we not applying this same standard to our doctors? Maybe if people had better doctors who understood the nuances of this, fewer patients would have unnecessary colonoscopies, breast biopsies, prostate biopsies, and be prescribed fewer drugs.

    • I wanted you to see the website–but here is the complete article to save your time.

      heartwire home
      Measuring apolipoproteins does not help risk prediction
      June 19, 2012 Sue Hughes

      Cambridge, UK – Measuring newer lipid biomarkers with the aim of honing risk prediction for heart disease doesn’t actually appear to help very much in this regard, according to a the largest study yet to look at this issue [1].

      The study, published in the June 20, 2012 issue of the Journal of the American Medical Association, found that measuring a combination of apolipoprotein B (apoB) and apoA1, lipoprotein (a) (Lp[a]), or lipoprotein-associated phospholipase A2 (Lp-PLA2) gave worse predictions of risk than current lipid measures—total and HDL cholesterol. In addition, the study showed that measuring these alternative biomarkers added little information when added to conventional risk factors.

      Coauthor of the current study, Dr Emanuele Di Angelantonio (University of Cambridge, UK), commented to heartwire: “These are the best data yet on these apolipoprotein biomarkers, and our conclusion is that it is probably not worth measuring them routinely for screening purposes.”

      Di Angelantonio explained that there have been previous studies on these apolipoproteins in smaller data sets, but this is the first study to look at all these markers on top of conventional risk factors in a large primary-prevention population.

      The researchers combined data on 165 000 individuals from 37 cohort studies where these apolipoproteins were measured at baseline and the patients followed for an average of 10 years. They conducted two analyses—one on the use of the new biomarkers instead of total and HDL cholesterol, and the other on using the alternative biomarkers in addition to total and HDL cholesterol.

      “We found that use of any of the apolipoproteins instead of current markers gave a worse prediction. And use of any of them on top of current markers improved risk prediction but only very slightly. It is probably too small an effect to be worth doing.”

      The study measured improvement in risk prediction by looking at changes in the C statistic. These were improved by very small amounts with the different apolipoproteins.
      Improvement in C statistic with apolipoproteins on top of conventional factors

      Improvement in C statistic

      The researchers also estimated how many individuals would be reclassified as high risk with these new measures. They calculated that for 100 000 adults aged 40 years or older, 15 436 would be initially classified at intermediate risk using conventional risk factors alone and that additional testing with a combination of apoB and apoA1 would reclassify 1.1% of people to high risk and, therefore, eligible for statin treatment under current guidelines. Corresponding numbers for Lp(a) and Lp-PLA2 were 4.1% and 2.7%, respectively.

      They further estimate that this would lead to the prevention of one extra cardiovascular outcome over 10 years for approximately every 4500 people additionally screened for a combination of apoB and apoA1, about 800 people screened for Lp(a), or about 1000 people screened for Lp-PLA2.

      “Not completely closing the door on Lp(a)”

      Di Angelantonio commented to heartwire: “Of the three markers, Lp(a) came out best, but even here the benefits were still very small.” However, he added that they did not include data on genotype or isoform size, which may be important for Lp(a), “so we are not completely closing the door on Lp(a) assessment for the future.”

      Noting that measurement of Lp(a) to augment risk prediction in intermediate-risk individuals was recommended in the most recent European Atherosclerosis Society guidelines, Di Angelantonio suggested that these may need to be “carefully revised.”

      Grundy in agreement

      In an accompanying editorial [2], Dr Scott Grundy (University of Texas Southwestern Medical Center, Dallas) agrees that these apolipoproteins “do not add much to risk prediction over routine lipid measures.” He suggests that the main reason for this is that cardiovascular disease is multifactorial and that apolipoproteins are only one set of factors among many.

      He writes: “A common mistake in clinical practice is to overly emphasize one or another risk factor as a predictive factor. This is particularly problematic when risk factors are colinearly linked, as occurs between lipid and apolipoprotein risk factors.”

      He also points out that the predictive power of various biomarkers does not necessarily equate to the benefit achieved by reducing them. So epidemiological studies cannot be taken as the final word, and drug therapy that targets the risk factor might still prove more efficacious than would be predicted from epidemiological studies.

      Grundy adds that recommendations for statin use in primary prevention may need to be revisited now anyway, highlighted by the recent meta-analysis from the Oxford group that showed benefits of statins in much lower-risk individuals than those for whom treatment is currently advised. And he suggests that risk assessment may in the future move away from measuring many biomarkers and instead focus on subclinical atherosclerosis with imaging methods or simple risk projection based on age, sex, LDL levels, and perhaps another major risk factor.


      « Previous heartwire article
      Alarming increase in hypertension in US children
      Jun 19, 2012 13:30 EDT Next heartwire article »
      Which drug, when? New anticoagulants force new decisions in AF treatment
      Jun 20, 2012 10:30 EDT

      The Emerging Risk Factors Collaboration. Lipid-related markers and cardiovascular disease prediction. JAMA 2012; 307:2499-2506.
      Grundy S M. Use of emerging lipoprotein risk factors in assessment of cardiovascular risk. JAMA 2012; 307:2540-2542.

      Related links

      Statins benefit those at much lower CV risk
      [Lipid/Metabolic > Lipid/Metabolic; May 16, 2012]
      ApoA size independently associated with MI
      [heartwire > Medscape Medical News; Apr 24, 2012]
      Lp(a) does predict risk in blacks, new ARIC data show
      [Lipid/Metabolic > Lipid/Metabolic; Dec 01, 2011]
      European Atherosclerosis Society recommends screening for Lp(a)
      [Prevention > Prevention; Jun 23, 2010]
      High Lp(a) levels and small ApoA size predict CV risk in dialysis patients
      [Lipid/Metabolic > Lipid/Metabolic; May 03, 2005]

    • Hasan, comments inline:

      “If you mean that knowledge in the hands of blog readers, well, thats another story and you have to measure the advantage of having more informed patients v. the reality that such knowledge can end up in added expense, anxiety, and possibly treatment if the patient is insistent enough. That takes us back full circle.”

      >> The genie is out of the bottle. It is all out there now. I think that the highly detailed, fundamental foundation approach that Peter is taking here is the perfect compromise between a simplistic and standardized site (like the Mayo site, which is a great place for patients who don’t want to go into this level of detail) and the medical journals, which are far too numerous and full of jargon (the actual concepts aren’t that bad until you get into the organic chemistry, but the studies and stats are no big deal once you develop a decoder ring).

      >> In any case, the uninformed patient is thankfully a thing of the past. If you look at the bright side, this kind of in depth analysis helps to counter the four page statin ads you see everywhere.

      “Don’t get me wrong, I am NOT trying to minimize the significance of MetSyn/IR. In fact, I consider it to be the hallmark of our time. If this series is heading towards “lifestyle change” as an approach to this problem, I am with you. If its headed toward adding even more people to the statin pool who have yet to participate because their LDL was not high enough, I am against you. ”

      >> Besides the fact that this cholesterol series is just the latest topic of hopefully many, many more topics in nutrition, I think the new title of “Eating Academy” and NuSI, the “Nutrition Science Institute”, should provide some clues as to where this is (frustratingly slowly) headed…

    • Bill,

      I suspect that Hasan has been influenced by some of the same people I have, such as Dr. Gilbert Welch of Dartmouth and Dr. Nortin Hadler of UNC.” They have written extensively about the harms that can come looking too hard for signs of illness in apparently well people (i.e., screening).”

      Indeed and I would add to that list Dr. David H. Newman. As for the rest, beautifully written (much better than I) and I concur 100%!


      “The overzealous interventions that have caused so much harm with unnecessary colonoscopy, PSA, and mammography are VERY different from what I think you’re suggesting (i.e., unnecessary prescription of a drug to reduce apoB/LDL-P/LDL-C/pick-your-target). This is a night and day difference.”

      There is a difference but not “night and day.” Unecessary testing/treatment of any kind, statin medication included, incurs a variety of costs- monetary, psychological, and time as well as the side effects. These are not trivial by any means as Bill has noted. (This idea that more information is necessarily better is part of the reason why the U.S. is spending 16% of its GDP on medical care as opposed to 8% in Europe).

      “Do you know what the most common presentation of myocardial infarction is in the U.S.? Not chest pain. Not shortness of breath. It’s sudden death. So the single most common presentation of heart disease (this was true 10 years and may not be true today) was death. If you had looked at the person the day before they died, what would you have seen? Would it have been bad enough to justify treating? What about a year before? What about 5 years before? What would have been enough to treat? How high should the calcium score get before a treatment was suggested? When is and LDL particle screaming loud enough to justify treatment? These are not easy questions.

      Indeed they are not easy questions but they won’t be answered without proper research and evaluation which is what I have been asking for here.. Dramatizing the issue doesn’t negate the need for such evidence. I haven’t seen it yet.

      “We shop around like crazy to find the best car mechanic or hair dresser (well, I don’t, but I’m sure some people do). Why are we not applying this same standard to our doctors? Maybe if people had better doctors who understood the nuances of this, fewer patients would have unnecessary colonoscopies, breast biopsies, prostate biopsies, and be prescribed fewer drugs.”

      I am sorrry but this is always the fallback position when evidence doesn’t support a procedure or treatment. To whit, well maybe the research doesn’t support it but in the hands of the annointed who now how to use it, it is miraculously transformed into something wonderful. Ok, if the argument is now that the testing under discussion should be left to the hands of such trusted practioners, have them produce the evidence that what they are doing is working. Otherwise, all we have are testimonials. On that basis I can argue that reading tea leaves or animal entrails are valid in the hands of those who know what they are doing.

      • Hasan, it seems we agree on much more than we disagree on, but you continue to belabor this point. I want to make a few points for clarification (others, not yours).

        You accuse me of “dramatizing the issue,” which perhaps I have, but you close by suggesting that an exceptional doctor with exceptional judgment versus a mediocre doctor with mediocre judgement is akin to someone who is really good at “reading tea leaves of animal entrails” versus someone who is not. Really?

        I say the following not be derogatory, though I realize it will come across as condescending, but if you haven’t done so I suggest spending more time with physicians in training and trying to gain an understanding of why residency — the training period POST 8 years of college and medical school — is anywhere from 3 to 12 years. In residency, and beyond, physicians are not learning new facts. If that were the case, an internist could be replaced by Google or a computer algorithm. Instead they learn judgment. There are great generals and poor generals. There are great surgeons and poor surgeons. There are are great internists and poor internists. The difference isn’t knowledge of military history, “gifted” hands, or encyclopedic knowledge. The difference is judgment. It’s knowing how to incorporate all of the evidence that is knowable and applying it to a situation where the limits of that evidence may be exceeded. I understand your disdain for this. That’s fine. It’s your right to hold everything to that highest standard. Just be mindful of what you’re dong.

        There are no randomized controlled trials showing that cigarette smoking increases the risk of lung cancer. There are no randomized controlled trials showing that jumping out of airplane with a parachute is safer than without a parachute. There are no randomized controlled trials showing that seatbelts save lives. There are no randomized controlled trials demonstrating the off-label efficacy of certain drugs we commonly use. But we use our judgement when we decide whether or not to employ these “treatments.”

        Few people enjoy a discussion of the poor quality of science out there running amok more than I do, but be mindful that you don’t confuse this issues. Some things can never be tested. Some thing haven’t been tested, but can be. It’s good to know the difference. I can think of a hundred “dramatizations” to make this point, but I have a feeling more examples will only serve to ensure I never write another blog post and instead continue in this back-and-forth discussion with you.

        Hasan, you’ve made your point. Anyone who fails to understand it by now is probably unworthy of hearing it. I’d really like to keep working on other blog posts, but I can’t do both (blog and go back-and-forth with you). I’m happy to respectfully disagree with you on this issue and I trust we’ve given the readers more than enough substrate to make up their own minds.

        Lastly, the U.S. is currently spending 17.6% of GDP on healthcare based on the most recent CMS figures and annual healthcare report from McKinsey & Co. I would argue, having studied this problem for 6 years, that it has much less to do with ordering too many screening tests or generating too much information, and much more to do with 3 other factors:

        1) Physician behavior in response to a litigious climate — doctors order head CT scans for an obvious case of caffeine withdrawal to ensure they are never sued, for example;
        2) Americans are culturally very different from Europeans and even Canadians (I’m highly qualified to make this statement since I’ve lived half my life in each Country) — in the U.S. we will nearly always make the choice to operate on 90 year-old-patient with a ruptured AAA, though the probability of that person returning to “normal” is infinitesimally small. Other nations are far more strict about explicitly (e.g., UK) or implicitly (e.g., Canada) applying QALY (Quality-adjusted life years) standards to these decisions. We are not;
        3) When a significant percentage of the population is uninsured, as is the case here, the relationship between payers and providers becomes an elaborate game of cross-subsidization. The heavy subsidization, which varies by payer type, insurance type, and even venue of service, is an embarrassing charade that is actually at the heart of our healthcare spending epidemic.

        Of course, my preferred solution to reducing healthcare spending to less than 10% of GDP by 2025 is just to have less people be sick. But that’s another story called NuSI for another day. Be well, Hasan, and know that I, and surely others, have appreciated your insights and devotion to this discussion.

    • Peter, I appreciate your points that 1) statin drugs are not the same as cutting off breasts, irradiating prostates, etc., 2) waiting for CVD to present itself symptomatically can mean you are dead before any chance of treatment, and 3) tests in the hands of a thoughtful, skilled clinician are not the same as tests in the context of a pre-defined standard of care that automatically triggers a cascade of further tests and intervention upon positive findings.

      On the first point, perhaps it’s true that statin drugs are more benign than the interventions that follow other screening tests. But in both cases, if you are screening healthy populations, due to basic Bayesian statistics your test must virtually 100% specific to avoid “treating” dozens, hundreds, or thousands of people unnecessarily to perhaps help one who really needed it. And so far at least screening tests, and medical tests in general, just aren’t anywhere near that specific.

      So unless the follow-on tests and treatments are 100% benign—also certainly not the case for any screening test I’m aware of—inevitably healthy people are unnecessarily harmed by the sequelae of screening. Then we can ask, say, if I’d rather lose my erections or my memory. I enjoy both, so it’s a difficult choice for me. I need to know the risks relative to the benefits. And the track record of screening when big enough studies are done to find out is so poor that I’m skeptical about subjecting myself, or anyone else, to new ones. Would the benefit-risk ratio be better with LDL-P (or other alternative) CVD screening and follow-on statin therapy, etc? Maybe. Maybe not. History is not encouraging, and the Bayesian challenge is enormous, but I must concede it’s possible and I hope it’s true.

      On the second point, this is really the same argument used to justify the other screening tests, e.g. my own primary care doc warns me that if I wait until colon cancer shows itself clinically, the outcome can be extremely grim. But true though this is, it’s not enough information to justify screening me now for the potential illness. I also need to know how effective the screening is at saving the people who are really destined to get sick vs. how harmful it is for the vastly larger pool of people who can’t benefit because they’re not going to get colon cancer anyway—or if they do, at an advanced age when something else was going to get them anyway (Nortin Hadler in particular emphasizes this critical last point: no medical intervention can actually “save a life,” contrary to screening propaganda, since death inevitably comes one per customer; the real issue is how much the quantity and quality of life might be extended). And now we are back to the very poor record of screening in this regard and the need for big, long-term trials to test seemingly common sense ideas like finding and treating catastrophic illnesses early.

      On the third point, I could not agree with you more. I allow my very smart, flexible, holistically-oriented doc, whom I see for a full hour at a time, to perform tests on me that I would certainly refuse in a more conventional medical context. I know that he and I will carefully discuss results and mutually decide further steps, even if outside the standard of care. It’s a big difference.

      However, even here I feel the need to be very careful. For example, when I had hand problems that *could* have been signs of serious early autoimmune illness, I refused rheumatological screening tests from both the conventional doc who insisted I needed them and from my holistic doc, who offered them if I wanted. The “information” I might have gotten seemed more likely to harm me than help me. Here again, the tests have low specificity, so I would risk becoming inappropriately labelled with an awful diagnosis, or the possibility of one, and I would not, at that point, agree to the treatments that a rheumatologist might suggest anyway (e.g., immune suppressing drugs) due to their substantial toxicity and risks. Much better ways to diagnose and treat these illnesses would have to exist, or I would have to be much sicker, for those rheumatological screening tests to make sense to me, even in the hands of my wonderful holistic doc. (Unlike the conventional orthopedist, who refused to even see me again unless I did the rheumatological screening, he found my position reasonable—and it was a good call, as it turned out, since the hand problem now seems to be non-inflammatory, which was the likelihood all along.)

      So I can easily appreciate how the alternative lipid tests might be useful in, say, your hands, even if not as part of yet another mass screening program pushed on the entire population long before evidence to justify it exists. But still we need such evidence to know if those tests will really advantage us, to use Hadler’s formulation. I greatly look forward to seeing your posts on what is known about that when you get to treatment. These are decisions I myself will have to make soon, since I am one of the LCHF people with (calculated) LDL-C high enough to make my primary care doc quite stern.

      And as others have said, many thanks for your herculean efforts here.

      • All very well said, Bill. Thanks for taking the time to comment on this issue. I trust your comments and those of others are as valuable, if not more valuable, than my actual posts.

    • Peter and Hasan, thanks for your kind words, and I can say them back to you both with much more validity!

      Peter has made clear—very reasonably—that it’s now time to move on, so all I will say is that to me a lot of the differences in point of view here come down to the question of how to think and what to do about important matters under conditions of uncertainty. This is a very large and to me a very interesting topic, which goes far beyond medicine (climate change is an all too obvious example) and in fact touches every moment of life for all conscious beings. Perhaps there will be opportunities to explore this critical topic along the way.

    • So, a question that may have been answered previously (this place is going to need a wiki soon, please get on it in your spare time…): The standard statin guidelines are based on studies. Do these studies track weight gain or loss during the study? If one assumes a constant or increasing weight with elevated TC for a guideline, maybe the guidelines make sense in those cases (leaving LCHF out of it for a moment). But wouldn’t elevated TC during weight loss be expected as more lipids are transferred from fat cells to the liver, especially on a LCHF diet? If that makes sense, have these studies been done?

    • Not for primary prevention, Axel. Of all the ratios, in patients with known CVD, it appears LDL-P/HDL-P > TG/HDL-C > TC/HDL-C. But as you know, a ratio is never used as a target of therapy, rather as a marker of risk.

    • Hi Peter,

      I hope you cover studies which elicit what to do when LDL-P/HDL-P is discordant with TG/HDL-C.


  8. Diet and Response of Triglycerides:

    4 cups cooked/eaten cold beans and wheat,barley,lentils.rice,millet mixture
    1 can tuna
    1 cup water fried potatoes
    1 large lettuce salad with vinegar
    1 slice whole wheat bread

    A 1500 calorie – very low fat diet that lasted six months – 70 pound weight loss – male subject(me) 35-6′ 3″ – 270 pounds starting

    A very low fat diet – my Grandma always tried to sneak some oil in when she fried the potatoes I usually kept an eye on her –

    Blood lipid Test near end of this six month diet:

    TC 150
    LDL can’t remember – this was in 1985
    HDL can’t remember

    Triglycerides 30

    The low Triglyceride number is what my Doctor mentioned as perhaps being too Low –

    The point of this post is that a high carb/very low fat diet can produce a very low triglyceride number

    Since the triglycerides where so low – I would assume a low LDL -P (particle number) and HDL -P( particle number)

    The 4 cups of cooked beans/mixture – eaten cold – and the low fat diet is what produced it

    The draw-backs of this diet where: I could smell the bean nitrogen coming out thru the pores of my skin and the cooking of the food produced indigestible food into the lower colon – which produced leg, thigh and buttock acne –

    This was in 1985 – now I have to eat a Raw Vegetable Diet (producing perfectly clear skin) without ant Yeast Contaminated Bread – I eat Low Carb now for other reasons – but at least a Teaspoon of DRY Beans/Mixture And One Teaspoon Fermented Bean/Mixture is i in my everyday Diet –

    The Dry Beans are for Carb Control and the Fermented Beans are used to fix my back issues(which works 100% by the way)-(via possible Vitamin K3 production)

    I also believe Theoretically – that the Bean/Mixtures – lower LDL and HDL Particle Numbers

  9. Listening to the interview with Dr. Dayspring on the June 20, 2012 podcast of The Livin’ La Vida Low-Carb Show with Jimmy Moore. Dr. Dayspring generally recommended a low carb diet to improve LDL(P) throughout the podcast, but at the end, briefly said that a low fat diet may help some people, those who are not insulin resistant, but that is not most people. It seems it would be very important to know whether you are insulin resistant, and if you are not, then low carb may not be the right diet for you? First time I’ve heard this, and would like to learn more.

    • It remains to be seen. I think the point is this, if you have dyslipidemia and you are IR, the likelihood that fixing your IR (via diet) will also correct your dyslipidemia is higher.

  10. Peter,

    Is #7 the final post in this cholesterol series?

    Thank you for your generosity with the information you’re packing into this web site. Quite a boon!

    Kind regards,

  11. Hi Peter,

    Thanks for a thought-provoking series.

    I’m curious if you plan to cover the role of ox-LDL as a risk factor. I’ve read multiple studies suggesting that it’s an independent predictor of CVD with equal or superior predictive value to any other single marker. However, I havent been able to find a study comparing ox-LDL to LDL-P or apoB:apoA (or either marker individually), or a study investigating ox-LDL that controlled for LDL-P or apoB. Are you (or is anyone else following this thread) aware of such a study?

    Up until very recently, this would have been a moot point because it has been impossible to test for ox-LDL outside of research settings. But there are now at least two labs (Shiel and Doctor’s Data) that are offering this test, and I’m sure more will follow (if there aren’t already more).


    • Ah, the joy of WordPress…I had just typed up a 30 minute response to this very important question and, voila, lost it! I won’t be able to repeat it, but here’s a shorter version of my thoughts.

      First off, great to have you involved in this discussion, Chris. I really love your work and the nuance you bring to it. It’s an honor to use your question as a broader response to a very important theme. Whether we’re discussing ox-LDL or another sophisticated marker, the problem we face is one of population homogeneity (or lack thereof). [For those not interested in ox-LDL (oxidized LDL particles), please read this response, as it addresses an important theme.]

      Most studies try to minimize the degree of heterogeneity somewhat — men vs. women, diabetic vs. non-diabetic, previous MI vs. not, etc. — but ultimately, even within such populations, vast differences (e.g., polymorphisms, apoE genetics) exist. Until we can do a better job of this, I opt for a more “personalized” approach of triangulation. Let me use IR as an example.

      Besides using a euglycemic clamp, which is too impractical outside of a research setting, what is the best way to assess someone’s IR? HOMA-IR? TG/HDL-L? OGTT? LP-IR? I propose there is no one “best” tool. There probably are best tools, if we knew which were most predictive in which sub-population, but a priori we don’t know.

      Instead, through constant clinical interaction, we start to see patterns. For example, LP-IR seems, in most, to precede elevation of HOMA-IR and TG/HDL-C, and elevation of HOMA-IR seems to precede TG/HDL-C.

      Guidelines, both for risk and treatment targets, are often helpful, but they need to be interpreted in the context of the patient. Someone with and LDL-P of 1,500 and normal Lp-PLA2 is probably at lower risk than an age-matched person at 1,200 with an elevated Lp-PLA2, as an example.

  12. Thanks for your reply, Peter.

    My current approach *with low risk individuals* (no IR, normal weight, physically active, Paleo/Primal diet) is to begin with TC:HDL, since many studies suggest that is roughly analogous in predictive value to apoB:apoA or LDL-P (especially if triglycerides are in the normal range), and also because these markers are cheap and readily available. If >3.5, I’ll re-test in a couple of weeks because I’m aware that the TC:HDL ratio can vary by as much as 0.8 per measurement without any dietary/lifestyle changes or therapeutic intervention.

    If again elevated, I’ll order the NMR to check LDL-P and apoB:apoA. I also run a few other markers like an iron panel + ferritin (iron overload is correlated with higher ox-LDL levels, and ferritin is an acute-phase reactant that can signal both iron overload and inflammation), CRP and homocysteine (more as a marker for methylation status). If LDL-P and apoB:apoA and the inflammatory markers are normal, that’s where I stop.

    If any of those markers are elevated, my next step is to investigate possible causes of poor LDL receptor function (leptin resistance, thyroid hypofunction, infection/inflammation, genetics), and treat those if present. I will also suggest additional dietary tweaks and some nutritional/botanical supplements to address the inflammation.

    What I’ve been wondering is whether ox-LDL adds anything to step 2 in that process. For example, if a patient has elevated LDL-P, but low ox-LDL, is their risk lower than someone with elevated LDL-P and ox-LDL? Probably, but hard to know how much lower. Likewise, if a patient has normal LDL-P but elevated ox-LDL, what does that say about their risk?

    The study I linked to indicates that ox-LDL is as good at predicting risk as apoB:apoA, but doesn’t add anything to the mix when particle number is controlled for. But that only speaks to risk at the population level, which speaks to your point. ox-LDL may still be a useful marker in the “individual triangulation” process you mentioned.

    Again, great series Peter. I look forward to the remaining articles.

  13. Hi Peter,
    I’m not sure if this point got covered and I missed it or maybe it needs to be made more obvious, but, if dietary cholesterol (CE) does NOT have a significant impact on the cholesterol (UC) in our plasma lipids, why then do we see a significant jump in the LDL-Ps or any other lipid markers for that matter in people who go on a LCHF diet? I hope this question makes sense. Thanks, Diana

    • Hi Diana

      I don’t see that your question has really been addressed yet. Dr. Attia has done a great job of personally validating a dietary approach that works for him, with a highly positive effect on his lipids and other markers. And it apparently works for many others. But some of us don’t see the same effects, in fact, the LCHF diet seems to worsen lipids. Dr. Dayspring referred to this in passing, in one of his commentaries on this site, something to the effect that low carb helps many, but not all people.

      Do you have any theories, conjectures on this?

  14. Peter,

    Are you (or is anyone else) aware of a study indicating that apoB/LDL-P adds predictive value when adjusted for *both* trigs and HDL?

    I found one study suggesting apoB:apoA-1 still had value after adjustment for TC:HDL, but although they did measure trigs, as far as I can tell they didn’t adjust for them. https://www.ncbi.nlm.nih.gov/pubmed/18676970

    In the one study I could find that did adjust for both trigs & HDL, apoB & LDL-P didn’t have additional predictive value. https://www.ncbi.nlm.nih.gov/pubmed/17276177


    • I do not believe this study has been done yet. Allan Sniderman, who in my estimation understands this as well as anyone, wrote a wonderful paper on this exact point in 2008. I have it, but it does not appear to be available for free: https://www.ingentaconnect.com/content/fm/flp/2008/00000003/00000003/art00006
      AMORIS and INTERHEART come close, but don’t adjust for both. There is little doubt that *on average* adjusting for HDL-C, non-HDL-C, and TG sufficient. Furthermore, in patients with MS, the rate of discordance is high enough to justify the look at apoB/LDL-P. The question that remains unanswered is what to do with the patient with, say, 0-2 of the 5 MS criteria (who, *on average*, should be fine with just the other test)? Clearly *some* will be discordant, even after adjustment. But how many? What to do about it?

  15. Thanks, Peter. Actually the study I linked to did control for both TC:HDL and triglycerides, and found that apoB:apoA-1 lost its additional predictive value.

    It seems to me that, as you pointed out, apoB:apoA is helpful when LDL-C and LDL-P are discordant. According to the Framingham data, that begins to happen as trigs >125.

    So one possible approach in clinical practice would be starting with a basic lipid panel, and then proceeding to NMR if TC:HDL >4 and trigs >125.

    Another possibility would be going ahead with the NMR regardless of trigs, but giving it more weight when trigs >125.

    Thanks for the link to the paper. I have full-text access to many journals, so I may be able to get the PDF. Looking forward to reading it.

  16. I really enjoyed reading your blog Dr Attia BUT I live in the UK where it is impossible to measure the particle sizes of LDL, never mind HDL ! I once asked a doctor, a “lipid specialist” no less, if I could have an LDL particle size test done privately – I was willing to pay for it – and he said no ! So what do people who can’t get particle sizes tested do ? I don’t know one person in the UK who has had this test done !

  17. Hi Peter – ApoB testing is not available either in the UK…well I just looked at a private laboratory and yes they do the test BUT they require a doctor to request it and that’s what bugs me that I can’t get a doctor to request it for me even though I’m willing to pay.

    Btw, I like your website, been looking through your blogs. I’ve been very low carbing for nearly six years now à la Paleo though I do add some milk to tea and coffee, but that’s the only dairy, I have no grains at all and the only fruit, like you, is berries. I’ve never been overweight but I turned to this way of eating after reading about the general health benefits of Paleo and low carb (I do have atypical Type 2 diabetes but I’m skinnny and not insulin resistant !). I eat tons – lots of meat, fish, eggs and nuts – I notice you don’t eat many veggies, I eat a lot of leafy green ones which I sauté in coconut oil to give me extra fats. My cholesterol profile is pretty good I think…although your post on HDL made me think twice ! My HDL is 3 which is 116 in American, and my triglycerides are only 0.5 or 44 in American – not that my doctor is impressed with my total cholesterol but he knows nothing !

  18. Anne NMR testing is available in the UK through: https://www.gdx.net/uk/product/123 This company was recommended by Liposcience:
    Name: Genova Diagnostics UK
    Address: 356 West Barnes Lane
    City: New Maiden Surrey State: UK Zip Code: KT36NB
    Office Phone: (800)522-4762

    • Hi James – I know and I know that the Doctors Laboratory in London can now do it BUT those labs all require that a doctor write the request form, a patient cannot self refer. I’ve asked a lipid specialist for a request for these tests, I’ve offered to pay for them knowing that the NHS would not pay for it and he said “no”.

Facebook icon Twitter icon Instagram icon Pinterest icon Google+ icon YouTube icon LinkedIn icon Contact icon