Kurt G. Harris MD

PāNu means paleonutrition. The "paleo" here signifies "old" and not necessarily paleolithic. The PāNu approach to nutrition is grounded on clinical medicine and basic sciences disciplined by knowledge of evolutionary biology and paleoanthropology. The best evidence from multiple disciplines supports eating a pastoral (animal-based) diet rather than a grain-based agricultural one, while avoiding what I call the neolithic agents of disease - wheat, excess fructose and excess linoleic acid.

Support PāNu

PāNu is ad-free, completely independent and has no outside sponsorship. If you value PāNu, now you can support it. Read this for more information.

In addition to buying from the book list, you can also support PāNu by making all of your Amazon purchases for any item through the Amazon Portal below

Amazon Portal

« Guest Post – Professor Gumby - essay 001 | Main | The China Study - Polish a turd and find a diamond? »

Statins and the Cholesterol Hypothesis – Part I

Reader Stephen is a young man I have corresponded with a few times on the subject of his diagnosis of FH (heterozygous Familial Hypercholesterolemia) and what he should do about it. His TC runs about 467 mg/dl and his LDL about 333 mg/dl. He has a CAC (calcium score) of 16, which is very high for a 24 year old man. This would be at about the 50th percentile for a 50 year old male on the SAD.

Stephen has of course given me permission to discuss his case here.

This is his most recent email to me:

Hi Dr. Harris,

It's been awhile since I last emailed you about any recent information regarding my heart scan. I also forgot to thank you for your last response and really appreciate the help you've given me. I recently had more lab work done and I am waiting to hear back for the results, although not much has changed since our last correspondence. However, now that I'm back in Dallas for my summer break I was able to get a proper heart scan at the Cooper Clinic. My CAC score was 16. I attended the Metabolism Society Symposium conference in Seattle a few months back and spoke with Jimmy Moore and a few others about my situation and they recommended I send Dr. Davis an email to get his advice on the matter. I recently heard back from him telling me basically that he strongly believes I go on statins while maintaining my current diet. Jimmy asked me to keep him posted about Dr. Davis' recommendations, so I emailed him, and he is now interested in posting a blog about it and asking others for suggestions since he is strongly against statin use. Anyways, I just wanted to update you, since you mentioned an interest in eventually making a post about this as well. If you're interested in having a look at the scan I can send another copy over to you as well as the results of the NMR profile I drew blood for a few days ago. If not, I completely understand as well, you have a lot on your plate as is. Let me know what you'd prefer and I'll do whatever works best for you. I look forward to hearing back from you. 

Thanks again, 

Stephen U.

By now you already know that I personally would avoid statins under any circumstances as they only work at all on heart attack risk via their effects on inflammation. Taking a sledgehammer to the entire cholesterol machinery has all kinds of negative knock-on effects, including promotion of cancer and interference with muscle and liver function, in addition to the accidental side benefit of reducing inflammation. It’s a crude approach, it has side effects that make all cause mortality over long periods of time likely to be worse, and the only group with any demonstrated benefit is men with established disease. The vast majority of those taking statins have no scientific basis for taking them. The rest, including those who have had an MI already, could likely accomplish far greater improvements in health with non-drug dietary measures like improved glucoregulatory control, and avoidance of the Neolithic agents of disease.

Let me summarize what we know about statins

1)   A few older trials show an all-cause mortality benefit to statins in secondary prevention IN MEN (not women). The relative risk reduction is at most about 30% and the maximum absolute risk reduction is about 1% per year. (This means you still have about 70% of the relative risk of dying you had before the drug. Is it possible dietary changes could decrease your risk more substantially? I think so.) Secondary prevention means you have already had a heart attack or been proven to have coronary artery disease (CAD). If you only have high TC or LDL, this does not include you, and this does not include the majority of those now taking statins.

2)   Other trials, especially more recent ones, are less likely to show a benefit, even in secondary prevention

3)   For primary prevention, there is no demonstrated mortality benefit to taking statins.

4)   When there are decreased deaths from cardiac events in primary prevention, there are more deaths by other causes. You may well be trading your heart attack for cancer. Because of the long lead time for cancer deaths, there are good theoretical reasons to believe that over periods of time - much longer than a few years (the length of the typical drug trial) all cause mortality could easily be higher when taking a statin for primary prevention.

5)   When statins do work, they work by accident via their effects on inflammation. The side effects may be related to the lowering of LDL, but the benefits are not. Trials have failed to show the linear relationship between LDL lowering and cardiac end-points that one would expect if the effect were due to the LDL level.

I do not believe in any of the versions of the lipid hypothesis, ranging from Ancel Keys' original idea that cholesterol or dietary fat clogs the arteries, to the currently fashionable one that “small, dense” LDL particles are like microscopic rodents that are designed to burrow under the intima of your blood vessels and kill you.

Neither cholesterol nor any of the lipoproteins nor LP(a) is a "cause" of CAD (coronary artery disease). There is no evidence that “fixing” these numbers is of benefit other than by accident and there is plenty of evidence that you can kill people by trying to do so. 

HDL, particle numbers, particle sizes, LP(a) are all parameters that are more or less associated with CAD. If they respond positively to changes in diet, then they are just covariant with decreased risk of CAD or MI due to the changes you made in your diet. They are not necessarily, and not usually the direct mediators of the decreased risk.

They may track the positive changes you make in your diet, but they are not causing heart attacks any more than shoe size causes height!

You cannot decrease your stature by amputating your toes. Believing any of the lipid hypotheses or the cholesterol hypothesis is the intellectual equivalent of amputating toes to decrease height because shoe size causes stature. 

Correlation is not causation. Causation can cause a correlation to occur, but proof of correlation is not sufficient to prove causation. None of the lipid hypotheses are biologically plausible, and all have failed to be proven despite decades of research and billions in expenditure. 

Say you observe that the neighborhoods that have the most numbers of police on patrol have more crime. A neighborhood in downtown Milwaukee has more than 8 times the per capita police presence than in Sturgeon Bay 120 miles to the north. The crime rate in downtown Milwaukee is also more than an order of magnitude higher. Say the calculated correlation coefficient is .85 – with 0 being no correlation and 1.0 perfect correlation. Is it reasonable to propose we reduce the number of police in Milwaukee in order to effect a lower crime rate? There is a very high (and statistically significant) correlation, but if we think of mechanisms, and how police presence relates to crime, we would probably think that the police are there in response to the crime and are not likely causing it. We might go a step further and say that it might be dangerous to reduce the number of police, as for all we know the city has put them there for good reason, and the crime differential between our two towns might become even greater without them.

In the same way, although high HDL indeed correlates with lower risk of MI (heart attack or myocardial infarction) and CAD (coronary artery disease), when we understand the biology of HDL production, we might be wary of approaches that attempt to increase HDL as if HDL is the agent protecting your heart. It might well matter how we alter HDL. It might be that HDL is high in response to whatever lowers heart attack risk.

It might be (In fact I think it is) that a diet high in saturated fat protects against atherosclerosis and also affects the numerical value of HDL (most of you know it does because you see it happen to you), such that the association of HDL and lower MI risk is because they are caused by the same thing (diet) instead of because one “causes” the other.   

Consider the particle du jour of LP(a). Measured Lipoprotein (a) has a correlation with MI risk. It’s not really that high as a risk factor (not even remotely close to glycated hemoglobin or calcium score), but there is a positive correlation. In an individual human, Lipoprotein (a) could be high as an adaptive response to having lots of oxidized LDL and thereby may indicate you have an atherogenic diet, like the crime ridden neighborhood that has a substantial police presence. Or you could be just born with high levels of LP (a). In the same way a somewhat paranoid and wealthy community has lots of policemen on patrol even though the crime rate is low, you might have patrol cars of LP (a) cruising around even though you do not have intimal damage occurring.

Sidenote: I admit I have a bias towards this interpretation. My Lp(a) level is 85 yet my CIMT* measurement would be normal for someone 15 years younger and my CAC score is zero. Lp(a) levels can be highly heritable, and longevity and freedom from atherosclerosis runs in my family. As Peter would say, some of us may just be born with more sticking plasters, and some of us make more sticking plasters because we have more damage. It should be noted that the best way to decrease LP (a) is increasing saturated fat intake, but also that although there is positive correlation of LP(a) with disease on a population level, there is not yet a shred of evidence that any intervention to modify your LP(a) level modifies your risk. 

The point is that just by observing more police patrols or more Lp(a), even in the context of a true correlation between crime and police and Lp(a) and MI, you cannot necessarily tell why the police are there or why the Lp(a) is elevated. More importantly, the idea that reducing the number of police will reduce crime or that pharmacologically altering Lp(a) will reduce risk may be not only false, but dangerously false. Some neighborhoods need their police, and some humans may need their Lp(a), just like some humans may have more heart attacks when we “help” them by pharmacologically jacking up their HDL. 

This is in fact precisely the story of torcetrapib. This drug cost Pfizer over 800 million dollars to develop. Torcetrapib is an inhibitor of CETP - cholesterol ester transfer protein. CETP facilitates the transfer of cholesterol from HLD to VLDL or LDL. Now, if you believe version 3.0 of the lipid hypothesis, there are these things called “good” and “bad” cholesterol. This is cardiologist shorthand for lipoprotein-cholesterol complexes that cause or prevent heart disease. HDL is the good one and we want it to be high, so it can hoover up the cholesterol from arterial plaque (a ridiculous idea that as Malcolm Kendrick pointed out defies physical chemistry – Ornish’s “garbage trucks”) and shrink them. LDL is the bad one and we want it to be low, as it is constantly fighting to transport cholesterol to the arterial plaque.

Cardiologists and Pfizer actually believe that LDL is trying to kill you and HDL is trying to save you at the same time. They figured, why not throw our pharmacological weight into the fight? If we inhibit CETP, the level of HDL will rise, and the level of LDL will fall. They actually combined torcetrapib with an extant statin (atorvastatin or Lipitor) to really, really get the LDL down as well as the HDL up.

So how did it work? Well, it worked spectacularly. HDL levels soared and LDL levels went down. I mean, we are talking HDL to LDL ratios that epidemiologists and cardiologists would say should reduce the risk of heart attack to zero. There was a glitch, though. Although the HDL and LDL went exactly the way they wanted, the risk of death in the group that got the torcetrapib was 60% higher. The trial was halted early. 

Sidenote: We could belabor the police metaphor by saying pharmacologic elevation of HDL with torcetrapib was a bit like recruiting Alex’s droogs into the police in “Clockwork Orange”. The crime rate would go up with more police recruited from among those who are themselves criminals. 

So do you think it is plausible that LDL causes heart disease and HDL saves us from it? Or is it more likely that these laboratory numbers are merely epiphenomena to the real primary process?

Is it really plausible that our bodies evolved to simultaneously create two substances that are like some perverse yin and yang, with good HDL trying to save us from the LDL that is furiously trying to clog our arteries with cholesterol?

Or is this all a bunch of biologically implausible nonsense? 

Taking all the evidence in its totality, I’ve see no compelling use for statins for anyone if they have truly optimized their diet, and very little even if they haven’t. The benefits of statins are unrelated to their effects on “cholesterol” or lipoprotein levels, are small and are accidental.

The risk/benefit ratio and effect size of dietary modifications are likely to be superior to any of the current statin drugs.

I do believe there are dietary ways that we can minimize the risk of atherosclerosis and other inflammatory processes, minimize the risk of acute plaque rupture and thrombosis if we do have atherosclerosis, and minimize the risk of a fatal arrhythmia if these upstream steps fail 

I don’t believe once you have taken these steps, that pharmacology targeted to “cholesterol” or lipoproteins has anything to add.

I suggest:

1)   Avoidance of excess PUFA in the diet that leads to oxidized LDL and possible endothelial damage

2)   Avoidance of all the causes of leaky gut that may lead to the suite of inflammatory processes known as the metabolic syndrome – avoid excess PUFA (esp. linoleic acid), wheat, and fructose

3)   Plenty of saturated fat intake

4)   If your glucoregulation is impaired, reduce carbohydrate consumption to whatever is necessary to minimize glycation and endothelial damage. Minimize glycated hemoglobin (Hemoglobin A1c) to the degree possible

5)   Replacement of essential micronutrients that may be deficient on the SAD. Magnesium, sunshine, pastured butter for Vit K2.

Those are my general recommendations for everyone.

So what if you have FH? Are the recommendations different?

More discussion in the next post. 


*CIMT is carotid intima/media thickness - easily measured noninvasively with ultrasound and a more sensitive measure of early atherosclerosis than CAC (calcium score) in men under 40 and women under 50.

Reader Comments (9)

Fantastic summary of the issues. Should be on the front page of every newspaper in the world. If only.

But we all know that'll never happen. Big Pharma isn't going to give up $25 billion a year in statin profits anytime soon. And as long as the majority of doctors get their information about drugs from drug rep sales brochures (a documented fact), they'll continue handing out statins like candy.

As Chris Rock said, "the money isn't in the cure, the money's in the medicine."

July 21, 2010 | Registered CommenterChris Kresser, L.Ac

I was surprised by how much the addition of a modest amount (200-400 mg) of magnesium changed my total cholesterol (from 240 to 200 mg/dL, both measurements while on a paleo diet).

Regarding HDL, I have to wonder if the type of saturated fat makes a difference. Vastly increasing my saturated fat intake (mostly from heavy cream) made no difference to my HDL (about 50 mg/dL). Perhaps a lauric acid rich fat source like coconut milk would have had more impact on HDL. Assuming it would, the question is does this equate to better health?

July 21, 2010 | Registered CommenterStephen

My answer is that it is the saturated fat that is healthy, and the PUFAs and sucrose you have displaced with it that are unhealthy. The HDL is just a marker. Remember that the Kitavans described by Lindeberg have "low" HDL and little evidence of coronary disease. It can't be the HDL itself. Low HDL with a wheat and sucrose and excess PUFA free diet seems to be no problem.

July 21, 2010 | Registered CommenterKurt G. Harris MD

I've had three patients over the years become psychotic on statins. The issues resolved when the statin was removed. Numerous "foggy brains" and worsening of depression. So many women on statins, everyone with diabetes... and yet such a side effect load. To see the critique of the JUPITER trial published in the Archives of Internal Medicine was heartening, to say the least.

July 21, 2010 | Registered CommenterEmily Deans MD

Comments on Jupiter coming up soon. WOMEN on statins? Where are the pink ribbons when you need them?

July 21, 2010 | Registered CommenterKurt G. Harris MD

I saw Stephen's question on Jimmy Moore's site too, and pointed out a paper on FH that I think is telling. See http://www.ncbi.nlm.nih.gov/pmc/articles/PMC31037/pdf/1019.pdf Basically, families carrying the FH gene did not show shorter lifespan or greater mortality during the 19th and early 20th centuries, not until after about 1915, and increasingly as time went on after that period. It's consistent with dietary changes such as increased sucrose consumption, carbohydrates in general, and perhaps increasing vegetable oils and trans fats. The authors only speculated that it was due to some environmental trigger.

I saw Dr. de Logeril point out that the recent statin trials have not shown efficacy and he believes that this may be due to the changes in rules regarding how the trials are to be conducted, i.e. the earlier positive results were fabricated in some way (by cherry picking patients? or excluding non-responders?). It's not clear what changes are responsible for the change. Any ideas?

July 21, 2010 | Registered CommenterCynthia M

FH to be discussed in more detail next post...

July 21, 2010 | Registered CommenterKurt G. Harris MD

Same experience as Emily. Female patient post-MI complaining of brain-fog, depression, low energy, peripheral neuropathy, weakness... all resolved as soon as removed from rosuvastatin. Now feels a million bucks. Funnily enough, cholesterol perfectly 'normal' (lab marked total cholesterol as high - was due to high HDL).

60y.o. Coeliac who was only diagnosed as such 10 years ago & who is gluten-free but still consumes large amounts of carbs, n-6 PUFA, low SFA, low serum vit D... Doctor joined the dots. Had MI, total cholesterol was high, therefore TC drove the MI and she should be on statin, aspirin, & fluoxetine for the rest of her (what was turning out to be a miserable) life.

July 21, 2010 | Registered CommenterJamie S

I got a fever . . . and the only prescription . . . is more PaNu!

July 23, 2010 | Registered CommenterTom G
Comments for this entry have been disabled. Additional comments may not be added to this entry at this time.