The Cholesterol Connection

The Cholesterol Connection

Although I didn’t mention it by name in my recent 4-part blog series on dietary fat, the so-called “Lipid Hypothesis” is central to any discussion of the relationship between fat and heart disease.  The idea behind the Lipid Hypothesis is a pretty simple one:  eating saturated fat leads to increased cholesterol in the blood, which in turn causes cardiovascular disease.  The Lipid Hypothesis is behind all, or nearly all, dietary recommendations, and almost everything you read about diet in the popular press toes this party line. 

I ended the last blog post with a statement that flies in the face of the Lipid Hypothesis, namely that eating saturated fat does not cause cardiovascular disease.   But what about cholesterol—does eating fat, any fat, cause an increase in cholesterol?   And do elevated cholesterol levels cause heart disease?  Part of the reason that there is so much confusion about a healthy diet in general is that the story is highly nuanced, so an answer to any question concerning nutrition requires more than a flat yes or no.

But in answer to the first question, here is what the data says:  eating some kinds of fat does result in higher cholesterol levels, but the increase is negligible and it depends on the type of fat.  The answer to the second question is what this blog post is all about.

Cutting right to the chase, I can say yes, the data supports the hypothesis that increased cholesterol in blood plasma is correlated with increased cardiovascular disease.  The science here is very good and as a result, there is near unanimity (or “consensus”) of opinion that the connection between cholesterol and heart disease is more than a mere correlation—there is causation as well.  To add fuel to the fire, high cholesterol is also correlated with increased mortality. 

But there are always nonbelievers, in this case “The International Network of Cholesterol Skeptics.”  So, I, of course, went to their website, expecting to find scientific literature dispelling, or at least chipping away at, the “cholesterol to cardiovascular disease” connection.  And what did I find?  A bunch of YouTube videos and press releases ranting about—my last four blogs.  Well, they hadn’t read my blogs, but the points they made were essentially the same.  That is, they were attacking the anti-fat dogma.  They also railed about the negative health effects of statin drugs, but I didn’t find anything on their website about cholesterol and cardiovascular disease per se except the following:

For decades, enormous human and financial resources have been wasted on the cholesterol campaign, more promising research areas have been neglected, producers and manufacturers of animal food all over the world have suffered economically, and millions of healthy people have been frightened and badgered into eating a tedious and flavorless diet or into taking potentially dangerous drugs for the rest of their lives. As the scientific evidence in support of the cholesterol campaign is non-existent, we consider it important to stop it as soon as possible.
   The International Network of Cholesterol Skeptics (THINCS) is a steadily growing group of scientists, physicians, other academicians and science writers from various countries. Members of this group represent different views about the causation of atherosclerosis and cardiovascular disease, some of them are in conflict with others, but this is a normal part of science. What we all oppose is that animal fat and high cholesterol play a role. The aim with this website is to inform our colleagues and the public that this idea is not supported by scientific evidence; in fact, for many years a huge number of scientific studies have directly contradicted it.   

Now, the majority of this diatribe is most certainly false.  Except for the part about the public being badgered into eating flavorless food.  I mean, who would want skim milk in their latte rather than Half & Half?  On the contrary, the scientific evidence for the connection between blood cholesterol levels and cardiovascular disease is overwhelming, especially in people whose cholesterol is very high.  (And although I have no poll data, I suspect the group of folks supporting this Network is not growing very fast.  I count 98 people on their “member” list.)  Certainly the dietary fat/saturated fat connection is starting to fall apart, and I’ll grudgingly concede that they may have a point concerning the negatives associated with a carbohydrate-based diet.  But but on balance the rest of their assertions are not supported by the most recent evidence.

So let’s look at some of the data.  The cholesterol story is very interesting, and some elements are surprising. 

First, there is a genetic disease called “familial hypercholesterolemia,” (note the prefix “hyper,” meaning “more”) or “FH” for short.  People with a gene causing this rare condition have cholesterol levels in excess of 350mg/dL (“normal” is currently defined as less than 200 mg/dL).  Now, geneticists love genes that cause medical conditions, because it enables them to target specific causes and biochemical pathways.  And in this case the most common gene implicated in FH (there are at least three) is one that causes a defect in the liver’s removal of LDL (the “bad” cholesterol) from the blood.  So LDL builds up, the patients develop plaque, and left untreated, a lot of them die prematurely from cardiovascular disease.

In a study published in 2008, researchers surveyed 3,382 British people in their 40’s.  Before receiving treatment for cardiovascular disease, 28% of the men and 20% of the women had already had a heart attack, by-pass surgery, or angioplasty, which is well above the norm for any population.  In 1992, the test subjects started treatment with a class of drugs called “statins” that interferes with the formation of cholesterol.   In those without any signs of cardiovascular disease prior to treatment, mortality from the disease dropped 48% and in those who already had cardiovascular disease, mortality dropped 25%.

So let’s pause here and consider the facts.  It is clear that persons with highly-elevated cholesterol have above-normal rates of cardiovascular disease.  It is also clear that when you interfere with cholesterol production in these persons, their rate of cardiovascular disease is greatly diminished.  And in fact, this is all pretty much scientific dogma today—I don’t think any scientists are disputing it.

But what about people without pre-existing cardiovascular disease?  

A large meta-analysis published in 2009 surveyed 68 long-term studies (2.79 million person years) in Europe and North America involving 302,430 people who DID NOT have cardiovascular disease at the start of the study.  The authors looked at the relationship among total cholesterol, LDL cholesterol (“bad” cholesterol), HDL (“good” cholesterol), and triglycerides. What they found was that high total cholesterol, high LDL, and low HDL were significantly associated with cardiovascular disease.  Interestingly, triglycerides (a measure of “fat” in your blood) showed a poor association.  Let me say that another way:  there was no relationship between triglyceride levels and cardiovascular disease.  Further, the incidence of ischemic stroke (the interruption of blood flow to the brain caused, for example, by a clot blocking an artery), showed a modest association with these blood factors, but the incidence of hemorrhagic stroke (rupture of a blood vessel) did not.

(And surprisingly, fasting before a cholesterol test was irrelevant.  In other words, it didn’t matter whether or not the patients had eaten prior to the test or not.)

I also looked at an earlier 2007 meta-analysis based on 61 studies, mostly in western Europe and North  America but also a few from China and Japan.  These studies involved almost 900,000 adult patients who DID NOT have previous cardiovascular disease.  The meta-analysis showed that total cholesterol, HDL, and LDL were linearly associated with heart disease mortality.  High HDL was “good” (low association with mortality) and high LDL was “bad” (high association with mortality), but the ratio of total cholesterol to HDL was the best predictor.  And as in the 2009 study, there was only a modest relationship between levels of these blood factors and ischemic stroke.

Okay, so these studies give us evidence for correlation when what we really need is evidence for causation.   Clearly, the best evidence for a causal relationship between high cholesterol and death would a finding of decreased mortality after specifically lowering cholesterol levels—without changing anything else.  Fortunately we have a way of doing just that:  by administering “statins.”

In 1971 statins were discovered in Penicillium sp. by Japanese biochemist Akira Endo (and they have since been found in other fungi as well).  At that time it was already known that cholesterol is manufactured in the liver through a very complicated series of biochemical steps called the “Mevalonate Pathway”, which is central to the formation of hormones and their precursors as well as such diverse chemicals as the rubber (in the rubber tree) and latex (in the sap from milkweeds).  What’s important here is that all biochemistry students had to memorize the Mevalonate Pathway, including Mr. Endo (and me).

Anyway, Akira Endo worked with fungi, and with his knowledge of the Mevalonate Pathway, he hypothesized that fungi produce a chemical that inhibits cholesterol formation in the parasitic organisms that prey on fungi.  The chemical that he found was compactin, now known as the very first “statin.”  In 1987, sixteen years after Endo’s discovery, Merck started marketing lovastatin, the first of this new class of drugs to reach the marketplace.  (What an interesting story this is!   I wish I knew more about it, as it is an excellent example of an obscure area of science leading to one of the most significant breakthroughs in history.) 

So statins have been used commercially since 1980.  What do we know about their efficacy and safety?

A meta-analysis published in 2011 reviewed 76 studies that compared statin use with placebo in 170,255 patients who were, on average, 60 years old.  The studies, which lasted an average of 2.7 years, included both men (74%) and women (26%), some of who already had coronary heart disease.  Here is what they found: statins reduced all mortalities by 10%, all cardiovascular deaths by 20%, and deaths from heart attack by 18%.  (There was also a reduction in deaths from stroke, but it was insignificant.)  Statin use resulted in a 26% reduction in non-fatal heart attacks and had no impact on cancer, but there was a TINY BUT STATISICALLY SIGNIFICANT RELATIONSHIP BETWEEN STATIN USE AND THE INCIDENCE OF DIABETES—an increase of 3.8% for statin users compared to 3.5% for controls.  But for purposes of the cholesterol hypothesis, the most relevant finding was that for every 10% reduction in LDL, there was a 2% reduction in cardiovascular mortality.  The bottom line is that using statins to reduce LDL cholesterol results in fewer deaths from heart disease.  

But here is the kicker:  because statins have shown such dramatic effects in cardiovascular disease over the past few decades, physicians have started prescribing statins to people at low risk.  And in Great Britain statins are available over-the-counter—at low dosages to be sure, but nevertheless available without a prescription.  Even if you can get them only in 10 mg pills, what’s to prevent someone from taking 10 pills, or more?

As a result, good scientists have started to wonder about the effect of statins on people at low risk.  Several studies have looked at this issue, but here are two that reach interesting, though OPPOSITE, conclusions.

A 2010 meta-analysis analyzed 11 studies and 65,229 people, ages 51-75, without cardiovascular disease but with LDL levels considered borderline high (an average of 138 mg/dL).  The subjects were divided into two groups—one received statin treatment and the other did not.  After 3.7 years, the statin group’s LDL averaged 94 mg/dL, while the placebo control group averaged 134 mg/dL.  HOWEVER, there was no difference between the two groups with regard to mortality.

A 2011 meta-analysis analyzed 29 studies, 8 of which were also covered by the 2010 meta-analysis discussed in the previous paragraph.   The 80,711 participants were 62% male and 38% female with an average age of 58 years and had a LOW risk of cardiovascular disease—less than a 10% probability.  (Interestingly, 47% of the study participants had high blood pressure, indicating that high blood pressure does necessarily not place people at high risk.) The patients were divided into two groups, with one group receiving a placebo, and the other a statin.  Those who were given statins had significantly lower rates of all-cause mortality, non-fatal heart attacks, and non-fatal strokes.  The was NO DIFFERENCE IN RATES OF CANCER OR DIABETES.

And guess what?  The last study didn’t even report on cholesterol.  So why did I even bring them up when cholesterol is central to any discussion of the Lipid Hypothesis?  Because I think one can conclude that as far as many scientists are concerned, the relationship between statin use and cholesterol reduction is so strong and so well-known that it no longer needs to be pointed out.  Since the authors of the last meta-analyses didn’t need to re-invent that wheel, they were able to ignore cholesterol altogether and proceed directly to the bottom line—the effect of statins on cardiovascular disease, cancer, and diabetes. 

But did you notice that the 2010 study found that statins had no effect on all-cause mortality and the 2011 study found that they had a significant effect?  I don’t know how to interpret these results, except to say that is just how science goes.  It may be that the first study did not last long enough for the effects of statin use to become apparent, but it is impossible to tell.  Unfortunately the 2010 study did not report on events other than death (such as non-fatal strokes or heart attacks), and neither of them reported the extent to which the test subjects had significant plaque build-up.

I should also point out that another 2012 meta-analysis showed that statin use did NOT result in fewer deaths from deep-vein thrombosis or “DVT” (an obstructive blood clot in a deep vein).  This particular analysis was conducted because of an earlier study indicating that statins can affect blood clotting, but it is not immediately clear why this should be the case.   The classic risk factors for DVT are numerous and varied (obesity, leg trauma, pregnancy, pancreatic cancer, old age, etc.), but elevated cholesterol is not one of them.  Even so, there was nearly a statistically-significant relationship between statin use and a reduction in PVT mortality (only off by .03%).

So what does this all say about the Lipid Hypothesis?  It seems that the first half is false, and the second half is true.  In other words, ingesting saturated fat has no significant effect on cholesterol levels, but cholesterol levels do have a significant effect on cardiovascular disease.  (At least as of 2012, when the most recent meta-analysis was published.)

And, finally, statins overall seem to be efficacious and safe.  They reduce deaths from cardiovascular disease, and they apparently do not cause cancer.  Nevertheless, there are those who continue to rail about the danger of statins in spite of all the evidence to the contrary.  In fact there are some new (2012) books with titles such as: How Statin Drugs Really Lower Cholesterol: And Kill You One Cell at a Time, and The  Truth About Statins:  Risks and Alternatives to Cholesterol-Lowering Drugs and Poisoned!:  Recovery from Statin “Side Effects”.  I have not read these books, so I don’t know where the authors are getting their information.   However, I have no doubt that if you look hard enough, you can find data to support any position you might want to take.  This is evident from the meta-analyses that I scrutinized as a part of the research for my blog series on dietary fat—I occasionally found individual studies that reached the opposite conclusion from that of the overall analysis.  By focusing on those few outlier studies to the exclusion of the many studies that support the majority conclusion, it would be possible to cite legitimate data in support of a position that runs counter to the weight of the evidence.  But why would you want to do that?  Unless maybe you thought you could make some money as a whistle-blower by spreading biased or unfounded theories, it really makes no sense. 

It seems to me that one should strive to look at ALL studies, whether they support your position or not.  That is why meta-analyses are so valuable.  Unlike most books, they are peer-reviewed, which means that prior to publication, other scientists were asked for their opinions as to the quality of the underlying research.  You can bet that any scientist who thinks a paper is not scientifically sound will not hesitate to make that opinion known.   That’s the beauty of peer-review, which is central to all scientific publishing, but NOT a part of the publishing process for most books intended for a general audience.

I should also make it clear that I do NOT believe everyone should be taking statins.  Why take even the small risk that they pose unless you are trying to avoid the much greater risk of cardiovascular disease?  And the modest increase in diabetes suggests that if you are taking statins, you should make sure you get tested for that, too.  (And along those lines, I should also point out that among the participants in these statin studies, liver enzymes were elevated in some (a bad thing) and depressed in others (a good thing).  So if you’re taking statins, it would be a good idea to get your liver enzymes tested per your doctor’s orders in case you’re one of the people who has a negative reaction.  Ain’t no free lunch.)

The verdict on cholesterol and cardiovascular disease?  Case closed.  I’m going to enjoy my saturated fats.