Yves here. To add some anecdata to KLG’s discussion about how statins became the magic bullet for lowering cholesterol, seen as a key way to reduce heart disease. First, for those concerned about dietary sources of cholesterol, it appears that humans make cholesterol from carbohydrates, and not fats. The ketogenic Atkins diet (which I don’t regard as healthy on a long-term basis but is a useful expedient for weight loss, since your body makes blood sugar way less efficiently when in a ketogenic state), where users get to eat lots of meat and eggs and fats but skimp on carbs, reduces blood cholesterol levels. But is that actually such a great idea? In women, the total cholesterol level correlated with the lowest all-cause mortality is 270. Similarly, and again for women, low LDL levels are correlated with higher stroke risk. By KLG, who has held research and academic positions in three US medical schools since 1995 and is currently Professor of Biochemistry and Associate Dean. He has performed and directed research on protein structure, function, and evolution; cell adhesion and motility; the mechanism of viral fusion proteins; and assembly of the vertebrate heart. He has served on national review panels of both public and private funding agencies, and his research and that of his students has been funded by the American Heart Association, American Cancer Society, and National Institutes of Health
The first post on the Diet-Heart Hypothesis covered fat and carbohydrates (starch, sugar), leaving cholesterol for today. But this fascinating compound was present at the beginning. A simple version of the first presentation of the Diet-Heart Hypothesis by Ancel Keys can be summarized as: (1) cholesterol level predicts heart disease, (2) saturated fat in the diet increases serum cholesterol levels, and that (3) monounsaturated fats protect against heart disease.
The history of cholesterol and heart disease is covered well by Joseph L. Goldstein and Michael S. Brown in this review from Cell in 2015. Atherosclerotic plaques were first described in the late 19th century, and early in the 20th century cholesterol was identified as the primary constituent of these lesions. In 1913 Nikolaj Anitschkow fed pure cholesterol to rabbits, who then developed plaques.
In my reading most descriptions of this work fail to note that the natural diet of a rabbit contains no cholesterol[1], which indicates the cholesterol that makes up as much as 50% by weight of the cell membrane in a rabbit (and all animals) is synthesized de novo from precursors. It is not clear that these rabbits were studied to determine if they eventually died of heart disease caused by atherosclerosis. Heart attacks proper were not recognized until the development of electrocardiography by James Herrick in 1919. Familial hypercholesterolemia (FH) was described and an autosomal dominant genetic disease[2] in 1938. Those with FH have elevated serum cholesterol and are susceptible heart attacks in middle age. Brown and Goldstein note the Seven-Country Study of Ancel Keys, fully supports (with a nod to a few quibbles at the margin) the notion that dietary fat leads to an increase in serum cholesterol which leads to coronary heart disease (CHD).
The complex, energy intensive biosynthesis of cholesterol[3] was elucidated primarily by Konrad Block and Feodor Lynen in the 1950s, with the identification of HMG-CoA as the first intermediate committed to the cholesterol pathway. The enzyme HMG-CoA reductase converts HMG-CoA to mevalonate, starting the cascade of reactions leading to cholesterol. In the 1970s naturally occurring inhibitors of HMG-CoA reductase were isolated from the penicillin mold. Now known as statins, the original and derivatives of these natural products were used to reduce serum cholesterol levels in experimental animals by inhibiting HMG-CoA reductase, which led to the upregulation of LDLR and subsequent increased uptake of dietary cholesterol into cells. It was a natural and logical step to extend these results to the lowering of serum cholesterol in patients with high cholesterol.
Lovastatin (Merck: Mevacor) was the first statin approved for use in humans in 1987 because it lowers plasma cholesterol/LDL and is tolerated well by most people.
The big question was then whether statins reduce the incidence in heart attacks. The Scandinavian Simvastatin Survival Study (4S) was one of the first comprehensive clinical trials of statins and showed this to be the case, and statins were destined to be blockbusters all around. 4S (1994) enrolled 4,444 patients with CHD/heart attack to test whether Zocor (Merck) lessened the likelihood of future cardiac events in this population. At the end of the trial, which lasted 5+ years, 8% of the patients in the statin group had died versus 12% of the patients who received the placebo. Over the course of the study, the statin group had a 91.3% probability of surviving, while those in the placebo group had an 87.6% chance of surviving.
Of the many trials since 4S, few seem to have produced the same level of protection. Moreover, those who refer to this trial in the biomedical literature, or otherwise use it as positive evidence of the efficacy of statins, do not always note that this trial included patients at very high risk of heart attack, so this demonstration of secondary prevention cannot be used to make conclusions about primary prevention of CHD/heart attack. While it is not clear that any subsequent well designed trial has produced results this strong, meta-analyses of many trials in the tradition of Evidence-Based Medicine can be made to support the hypothesis.
In a test of statins for primary prevention the AFCAPS/TexCAPS trial (1998) enrolled 5,608 men and 997 women with average plasma cholesterol levels. Of those in the statin arm of the trial, 3.3% had heart attacks, and 5.6% of those taking the placebo had heart attacks. It apparently remained unmentioned by the authors that 80 patients in the statin group died, while 77 in the placebo group died. This can be explained because 63 subjects in the statin group died of non-cardiovascular causes versus 52 in the placebo group.
Nevertheless, the conclusion was confirmation that reducing plasma LDL cholesterol is beneficial. In another primary prevention test of statins, the ASCOT-LLA trial (2003) enrolled 10,305 subjects with high blood pressure or other risk factors for cardiovascular disease (CVD) but with average/normal plasma cholesterol levels. These were potential patients who probably would be prescribed statins today. 1.9% of the subjects in the statin group had a heart attack or died from heart disease, while 3.0% of the placebo group died of CVD/CHD. Thus, the absolute risk of death from heart disease was 1.1% higher in the placebo group. This was reported as a 37% reduction in relative risk associated with statins, however, which is arithmetically correct.[4] But what is heard by the public is naturally, “If you have normal cholesterol plus one or more risk factors, you are 37% less likely to die of a heart attack if you take this drug!” This trial was halted early because of its positive results. There was no significant difference in all-cause mortality between the statin and placebo groups.
If a drug works, then there should be a dose-response effect, i.e., a higher dose, if well tolerated with no untoward side effects, should work better or faster than a lower dose. The Treating to New Targets (TNT, 2006) initiative addressed this hypothesis.
In this study of secondary prevention of CHD, either 10 mg or 80 mg doses of atorvastatin (Lipitor) were used to treat patients with CHD and metabolic syndrome. At the end of 5 years 13% of 10 mg statin group had died, while 9.5% of 80 mg statin group had died. The absolute risk reduction was 3.5%, which is statistically significant, and the relative risk reduction was approximately 27% (3.5/13). All-cause mortality in the two groups was 6.3% versus 6.2%, which is not statistically significant.
But this may be clinically relevant, in that the higher dose of Lipitor decreased the incidence/risk of CHD in these at-risk patients, but that decrease may have been counterbalanced by a higher risk of death from other causes? This study was a post-hoc analysis of the original TNT initiative, “Intensive Lipid Lowering with Atorvastatin in Patients with Stable Coronary Disease.” The original included 10,001 patients with CHD and the average, i.e., “healthy,” plasma cholesterol level of 130 mg/dL. Based on the way I read the results, the plasma cholesterol of the 10 mg group decreased to 101 mg/dL, while the high dose group cholesterol decreased to 77mg/dL. 10.9% of the 10 mg statin group died compared to 8.7% of the 80 mg group (absolute risk reduction of 2.2%; relative risk reduction of 20.1%, 2.2/10.9). Thus, the higher dose of statin reduced plasma cholesterol more effectively, and fewer people died in the high-dose group.
But as noted in the final sentence of the Results summary, “There was no difference between the two treatment groups in overall mortality.” The entire Conclusion summary: “Intensive lipid-lowering therapy with 80 mg of atorvastatin per day in patients with stable CHD provides significant clinical benefit beyond that afforded by treatment with 10 mg of atorvastatin per day. This occurred with a greater incidence of elevated aminotransferase levels.” I missed/could not find the absolute levels of alanine aminotransferase in the two populations. The increase may not be clinically significant, but an increase in plasma liver enzymes is a marker of liver damage. I will leave it to the reader to parse the meaning of the quotations in bold font. But in keeping with my ongoing reading in EBM, I am now going to make two long lists, for which I apologize. The primary authors of this paper are from the following distinguished academic and medical institutions:
From the State University of New York Health Science Center, Brooklyn (J.C.L.); the University of Texas Southwestern Medical Center, Dallas (S.M.G.); San Francisco General Hospital, San Francisco (D.D.W.); Pfizer, Groton, Conn. (C.S.); the Heart Research Institute, Sydney (P.B.); Institut Pasteur, Lille, France (J.-C.F.); Weill Medical College of Cornell University, New York (A.M.G.); Universitätsklinikum Eppendorf, Hamburg, Germany (H.G.); Academic Medical Center, University of Amsterdam, Amsterdam (J.J.P.K.); the University of Glasgow, Glasgow, United Kingdom (J.S.); and Emory University School of Medicine, Atlanta (N.K.W.) (pdf, p. 1425)
The study was “Funded by Pfizer.”[5] Dr. LaRosa reports having received consulting fees from Pfizer, Merck, Bristol-Myers Squibb, and AstraZeneca and lecture fees from Pfizer; Dr. Grundy lecture fees from Merck, Pfizer, Kos Pharmaceutical, Abbott, and AstraZeneca and grant support from Kos Pharmaceutical and Merck; and Dr. Waters consulting fees from AstraZeneca and Pfizer; lecture fees from Merck, Pfizer, and Novartis; and grant support from Merck and Johnson & Johnson. Dr. Shear is an employee of Pfizer and owns stock in that company. Dr. Barter reports having received consulting fees from Pfizer, AstraZeneca, and Sanofi-Aventis; lecture fees from Pfizer, AstraZeneca, FournierPharma, and Sanofi-Aventis; and grant support from Pfizer; and Dr. Fruchart consulting fees from Pfizer and Fournier and lecture fees from Merck, Fournier, Pierre Fabrie, and AstraZeneca. Dr. Gotto reports having received consulting fees from AstraZeneca, Bristol-Myers Squibb, Merck, ScheringPlough, Pfizer, Novartis, and Reliant and lecture fees from AstraZeneca, Merck, ScheringPlough, Pfizer, and Reliant and having testified before the Food and Drug Administration on behalf of Johnson & Johnson–Merck. Dr. Greten reports having received consulting and lecture fees from Pfizer, Merck, and ScheringPlough; Dr. Kastelein consulting fees, lecture fees, and grant support from Pfizer, Merck, ScheringPlough, AstraZeneca, Bristol-Myers Squibb, and Sankyo; Dr. Shepherd consulting fees from AstraZeneca, GlaxoSmithKline, Merck, ScheringPlough and Oxford Biosensors and lecture fees from AstraZeneca, Merck, and ScheringPlough; and Dr. Wenger consulting fees from Eli Lilly, Merck, Bristol-Myers Squibb, Pfizer, and Kos Pharmaceuticals; lecture fees from Eli Lilly, Pfizer, Novartis, Merck, Bristol-Myers Squibb, and Kos Pharmaceuticals; and grant support from Eli Lilly, Novartis, Bristol-Myers Squibb, and AstraZeneca. (pdf, p. 1434; emphasis added in both paragraphs)
From where does the evidence come? This particular paper has been cited 2,444 times in the biomedical literature, so Big Pharma’s return on investment has been substantial. This approach is bound up in the nature of EBM, covered in the book reviewed previously here, where among other things, the nature of grants from Big Pharma is discussed. Current (2022) statin guidelines largely remain the same as they were 8 years ago and 30 years ago (Statin Use for the Primary Prevention of Cardiovascular Disease in Adults. US Preventive Services Task Force Recommendation Statement, here and here,) Some commentary on the updated guidelines remains on point, noting among other things that statins are cheap (depends on the definition) and a major problem remains that not enough people are taking them. Also, see here.
This conventional wisdom is based on the biochemistry and physiology of atherosclerosis, abstracted from how people live in a world where food might be a food-like substance, especially what is available to working people in and out of food deserts.
Still, a reasonable citizen might ask, “If heart disease can be prevented by statins, where is the proof after 40 years and hundreds of billions of dollars spent on research, clinical trials, and statins as prescription drugs, and why are we still arguing about this?”
Going back to the 4S Study of 1994, about 88 of 100 people on average in the placebo group died, while 91 of 100 died in the statin group. Statistically significant, yes, and much more than that for the three whose deaths may have been due to non-treatment with statins. But if higher plasma cholesterol is the culprit, the level of protection perhaps should be higher? Statistical arguments can prove the hypothesis, even when the underlying pathophysiology is unknown. When Richard Doll and Bradford Hill completed their survey on smoking and health among 41,000 British physicians from 1951 to 1954, 789 deaths were reported and 36 of these were attributed to lung cancer. All 36 of these physicians were smokers.[6] No statistical analysis required.
But not everyone agrees with the current dogma! For example, in a commentary this month (October 2022) in JAMA Internal Medicine, Anand R. Habib, Michael H. Katz, and Rita F. Redberg ask if it is “time to curb our enthusiasm” for the use of statins for primary cardiovascular disease prevention: “The updated evidence synthesis found that statins yielded a smaller, but still statistically significant, reduction in all-cause mortality…(but) these recommendations should be considered in the context of a meta-analysis (2010) which enrolled only patients receiving high-risk primary prevention; this study showed no benefit on all-cause mortality with statins (and) the benefit for CVD mortality was not statistically significant…”
More importantly, and this is one of the very few such statements I have encountered in my reading of the primary literature, “There is a difference between statistically significant and clinically meaningful benefit (and) the purported benefits of statins in terms of relative risk reduction are fairly constant across baseline lipid levels and cardiovascular risk score categories for primary prevention. Therefore, the absolute benefit for those in lower-risk categories is likely small given that their baseline absolute risk is low, which the chance of adverse effects is constant across risk categories.” Adverse effects such as excess deaths in the statin branch of the trial, perhaps? It gets even better in this commentary, as noted in comments here last week:
The practice of medicine is an art as well as a science…In the US, about $25 billion is spent annually on statins (but) cardiovascular disease incidence and mortality are the upshot of myriad social determinants. Although statins lower LDL cholesterol in individuals, investments at the community level to construct a more salubrious environment that enables healthy eating and promotes physical activity are more likely to have widespread multiplicative and pleiotropic effects on improving quality of life. The 2022 USPSTF recommendations are an opportunity to pause and refocus efforts to meaningfully improve CVD outcomes for all, rather than extol the marginal, likely small, and uncertain absolute benefits of statins for the few in primary CVD prevention.(pdf, p. 1023)
At last! A multi-level, multi-system view from physicians with a platform of the “Diet-Health Hypothesis” instead of only the “Diet-Heart Hypothesis.”
Finally, my interest in this subject began as I prepared a presentation with the goal of modifying the medical curriculum on nutrition and the social determinants of health to place responsibility for the obesity epidemic where it belongs. Which is not on the people. My gloss on the history of the Diet-Heart Hypothesis was surprising to some but well received. One colleague did remind me that CVD/CHD patients would still be discharged from the hospital with an obligatory statin prescription and the advice to eat a low-cholesterol, low-fat diet and, above all, to take their medicine. Another colleague sent me references showing that statins are useful because of their anti-inflammation activities. True, but a healthy, balanced diet that includes fruits and vegetables is also an anti-inflammation diet. Not only that, research over the past 20 years indicates that obesity itself is a condition associated chronic inflammation.
But I also come back to the essential roles of cholesterol in normal biological membrane structure and function. Does it make sense to interfere with cholesterol absorption and synthesis when every cell in our bodies depends on cholesterol for their physical integrity? Then there is the role of cholesterol as the precursor for the synthesis of the steroid hormones – androgens, estrogens, corticosteroids – without which we would not exist. It does not make sense that artificially lowering cholesterol levels will have only benign effects.
Cholesterol appeared on the cover of Time magazine in March 1984: “Cholesterol – And Now the Bad News.” In 2015, the USDA Dietary Guidelines Committee stated:
Previously, the Dietary Guidelines for Americans recommended that cholesterol intake be limited to no more than 300 milligrams per day. The 2015 DGAC [Dietary Guidelines Advisory Committee] will not bring forward this recommendation because available evidence shows no appreciable relationship between consumption of dietary cholesterol and serum cholesterol, consistent with the conclusions of the AHA/ACC report. Cholesterol is not a nutrient of concern for overconsumption. (Original reference can be found here; AHA is American Heart Association and ACC is the American College of Cardiology).
This channeling of Emily Litella has not gotten nearly the attention it deserves.
To be fair, butter appeared on the cover of Time 30 years later in June 2014: “Scientists labeled fat the enemy. Why they were wrong.” The difference in influence of Time between 1984 and 2014 is large.
And to his considerable credit, Ancel Keys, who is primarily responsible for the Diet-Heart Hypothesis (which was, lest we forget, eminently reasonable in theory) was himself quoted in The New York Times in 1980: “I’ve come to think that cholesterol is not as important as we used to think it was. Let’s reduce cholesterol by reasonable means, but let’s not get too excited about it.” (quoted in Gary Taubes, Good Calories, Bad Calories, p. 79)
Which brings us back to my title: “What if Medicine Were Taught Like a Science?” And no, this does not imply that each medical practitioner must be an historian of his or her profession. We do not have to be historians or anthropologists to understand and appreciate the history of Reconstruction as presented by Eric Foner or the meaning of anthropology according to David Graeber and David Wengrow.
But just as the pathology textbooks state that Alzheimer’s disease is caused by amyloid plaques and tau tangles, they also say, “The HMG-CoA reductase inhibitors, commonly known as ‘statins,’ are the most commonly prescribed lipid altering therapy by virtue of their potency, tolerability, and impact on reduction of cardiovascular events.” All mostly true, as far as the statement goes.[7] But it does not go nearly far enough.
“Evidence-based medicine” sounds a lot like “science-based medicine.” But that all depends on whose evidence and what medicine, used for whose purpose. This can only be understood if the science and practice of medicine are taught, not told, with the history attached, at least at the margins. In any case, the margins are where the most meaningful discoveries are made.
________
[1] Rabbits are herbivores. Plants contain phytosterols, which are the cholesterol-like component of cell membranes in plants. Without getting too far into the weeds, biological membranes are composed of inner and outer leaflets of phospholipids with cholesterol/phytosterol inserted between the phospholipids. Sterols are large planar (flat) molecules that maintain fluidity of biological membranes while also contributing to requisite local stiffness. OK, that was in the weeds, but it is also evidence that cholesterol has other functions than to cause heart attacks. Such as being the starting material for sex hormones. I have wondered, but not so much as go searching through the literature, what the result would have been if rats, which are omnivores with a physiology very similar to humans, had been used in these initial animal experiments instead of rabbits.
[2] Autosomal dominant: Presence of mutation on one of the paired 22 non-sex chromosomes leads to disease. In this case the gene encodes the LDL receptor (LDLR) which is responsible for removing LDL from the circulation (LDL: low-density lipoprotein, so-called “bad” cholesterol; cholesterol is not water soluble, so it is transported in the circulation in lipoprotein particles). A heterozygote for the mutation with half the normal complement of LDLR has markedly elevated serum cholesterol. Homozygous FH patients complete lack LDLR and cannot clear LDL/cholesterol from the circulation. Huntington disease is another autosomal dominant disease. Compare with autosomal recessive traits, which result in disease only if both chromosomes carry the mutated gene, e.g., Tay-Sachs disease. Most autosomal recessive traits involve enzymes, for which 50% of the normal amount is usually “good enough.”
[3] If you really want to go there, you can start here (registration required) or any good biochemistry textbook. I once or twice memorized the entire pathway (structures included, with a memory half-life of 60 minutes) so as to regurgitate on exams. The key, though, is HMG-CoA reductase and its actions in reducing plasma cholesterol/LDL.
[4] Absolute risk and relative risk are often conflated, especially in marketing. This is covered concisely in Chapter 13 of Fineman’s Nutrition in Crisis. For example, if on average 1 of 10,000 people treated with a drug do not die from disease X, while 2 of 10,000 people in the same population die if they do not take the drug, the absolute risk of death is 0.01% (0.02%-0.01%, or 1-in-10,000). The relative risk of not taking the drug is 50%, however (0.01/0.02). 50% is the much more impressive number. But still, the difference is 9,999 healthy people versus 9,998. Which is not to ignore the death of that one person, on average. But if the drug has off-target effects that might lead to death from other causes, the calculation necessarily changes.
[5] A history of Lipitor. Wikipedia, but the included references seem mostly correct and complete. “Further reading” in this entry is a good place to begin.
[6] Siddhartha Mukherjee, The Emperor of All Maladies: A Biography of Cancer, 2010. Part Four: Prevention is the Cure.
[7] L.S. Lilly, Pathophysiology of Heart Disease, Seventh Edition, LWW, 2020. The “same” passage in the 4th edition (2007) reads, “The HMG CoA reductase inhibitors, commonly known as statins, are the most effective drugs for reducing LDL cholesterol. By virtue of their potency, excellent tolerability, and mortality benefits, they are the most widely prescribed lipid-regulating drugs.” The 2020 version is toned down somewhat and is more accurate, but no medical student and very few academics will compare textbooks in such a close reading.