The Cholesterol Wars: The Cholesterol Skeptics vs the Preponderance of Evidence
227 pages. New York, NY: Academic Press; 2007. $69.95. ISBN 978-0123739797
I consider The Cholesterol Wars by Daniel Steinberg to be required reading for every serious student of atherosclerosis and lipidology. It is written in an authoritative, scholarly, yet readable style. Steinberg has carefully reviewed the material for which he previously laid the groundwork with a series of articles in the Journal of Lipid Research. His subject is one about which he knows a great deal and in which he himself played a major role over the last 5 decades. I remember first hearing him speak about the subject matter of the book during the 2001 Drugs Affecting Lipid Metabolism meeting during a dinner at the Russian Tea Room in New York the night before 9/11. Everyone present that night remembers both the scintillating talk as well as where they were on that particular occasion.
I will make a few comments, observations, and an occasional critical remark that is not intended to detract from the quality of this publication. First, Steinberg begins by discussing the early cholesterol-feeding experiments of Nikolai Anitschkow and rightly points out that they followed studies on the toxicity of feeding large amounts of protein to rabbits. I believe that part of the failure to accept the cholesterol hypothesis on the basis of Anitschkow’s work is related to the extreme conditions of the experiment and to the very high levels of cholesterol obtained in the blood, which raised doubts about its relevance to humans. We now know with the benefit of hindsight that one can develop atherosclerosis even with lesser degrees of cholesterol elevation, but many also suspected that the cholesterol-feeding experiments might have induced an inflammatory response that was separate and distinct from the atherosclerotic process.
As its title implies, the book is more cholesterol-centric than lipoprotein-centric, although Steinberg rightly credits the work of J.L. Oncley, John Gofman, and all the members of the lipidology community who worked and trained with Don Frederickson, during his days at the National Institutes of Health. Steinberg does not discuss the possible role of dietary cholesterol in coronary heart disease (CHD) apart from its effect on serum cholesterol levels, which I still consider to be an unresolved question. Incidentally, a clarification is required on Frederickson’s career: He was a research fellow at the Massachusetts General Hospital under Ivan Frantz, but he first did an internship and fellowship on thyroid disease at the Peter Bent Brigham Hospital with George Thorn, who later became president of the Howard Hughes Medical Institute before Fredrickson.
Taking a historical perspective, Steinberg traces the lipoproteins from their discovery by Macheboeuf in 1929, through the work of Oncley and E.J. Cohn with electrophoresis, to studies by Gofman, Frank Lindgren, Alex Nichols, and colleagues with the analytic ultracentrifuge. Steinberg refers to a 4-center study led by Gofman in the 1950s on the value of measuring lipoproteins versus serum cholesterol as a predictor of CHD.1 The unfortunate thing about the Gofman study is that 2 opinions were published, and the view that serum cholesterol was as important as serum lipoproteins in predicting CHD became the accepted dogma for a long time. I believe that this prevailing view hindered the development of the field by perhaps a decade. In my opinion, it was not until Frederickson’s phenotyping of cholesterol and lipoprotein disorders, done in collaboration with Robert Lees and Robert Levy, that the field became more accessible to the general physician and began to develop at a greatly accelerated rate. The work by Frederickson and his colleagues helped to popularize the view that measuring total cholesterol has relatively little value compared with measuring the lipoproteins. These classic experiments were made possible by the earlier work of Lees and Fred Hatch at the Massachusetts General Hospital, who discovered that in order to obtain good separation of lipoproteins by paper electrophoresis, one had to add albumin to the buffer.
Steinberg devotes a chapter to the pivotal discovery of the low-density lipoprotein (LDL) receptor by Mike Brown and Joe Goldstein, which is what really tied together cholesterol and lipoprotein metabolism. The discovery of the LDL receptor, which does not recognize cholesterol itself but rather the specific protein in LDL, provided a way to conceptualize the regulation of serum lipoproteins in the body, and it helped to establish the link between high levels of LDL cholesterol and CHD. Coupled with the further discovery of the statins by Akira Endo, it then became possible to demonstrate that the inhibition of cholesterol biosynthesis has a clear effect on serum cholesterol levels and lowers the risk of CHD. These 2 discoveries helped to confirm the lipid hypothesis.
The results of the Coronary Primary Prevention Trial offered concrete evidence in favor of the lipid hypothesis, but we would never have gotten over the goal line without the statins. I happened to be president of the American Heart Association at the time that the Coronary Primary Prevention Trial results were announced, and the cardiology community, and even the American Heart Association, were quite skeptical. Doubts were raised about the statistical validity of the trial and the unlikely possibility of increased accidental deaths and suicides in individuals trying to lower their cholesterol.
As Steinberg recounts, after reports of toxicity with compactin, the Sankyo Company was evidently fearful of similar findings with lovastatin. Lovastatin was discovered almost simultaneously by Alfred Alberts and colleagues at Merck and by investigators at Sankyo. Although investigators at Sankyo obtained patents in most European countries and in Japan, they never developed or marketed the drug, even though they blocked Merck’s introduction of lovastatin where they had priority. Merck later developed simvastatin, which has an additional methyl group and turned out to be a more potent agent. Sankyo then isolated pravastatin, which they viewed as safer because of its greater water solubility, and comarketed the drug worldwide with Squibb, now Bristol-Myers Squibb. This led to approximately a decade of “cholesterol wars” between pharmaceutical companies and investigators over the potential link between the hydrophobicity of a statin and its toxicity. It was postulated that the more hydrophilic statins would have less access to peripheral tissues and the brain than the more hydrophobic ones. As far as I can determine, the relative hydrophobicity or hydrophilicity has little if any effect on the clinical application of these drugs. In any case, on the basis of data from the Heart Protection Study, the Scandinavian Simvastatin Survival Study (4S), the West of Scotland Study, Cholesterol and Recurrent Events (CARE) study, the Long-term Intervention in Pravastatin and Ischemic Disease (LIPID) trial, and the Air Force/Texas Coronary Atherosclerosis Prevention study, it is now established that statins have a class effect on CHD risk reduction. These trial data have helped, probably more than any other single factor, in establishing the validity of the lipid hypothesis. Atorvastatin went on to become the No. 1–selling drug in the history of the pharmaceutical world and simvastatin No. 3.
We’re left today trying to address residual risk. Steinberg describes the high-density lipoprotein–raising approach, although the jury is still out on cholesteryl ester transfer protein inhibitors. We have not yet lowered LDL to the point at which we see toxicity just on the basis of LDL reduction per se, nor have we identified a threshold below which LDL reduction is no longer beneficial. In fact, serum cholesterol levels have been shown to predict CHD risk even in a population like China, where cholesterol levels were extremely low in the 1980s and only 7% of deaths were due to coronary disease.2 Therefore, it appears as if no lower threshold really exists, although the fact that the relationship between CHD and LDL cholesterol is logarithmic or curvilinear implies that there must be some value of LDL below which further reduction obtains little benefit.3 Raising high-density lipoprotein, reducing triglycerides, and trying to attack inflammation seem to be promising future approaches.
Finally, this is a wonderful book for someone who spent virtually his entire career in the lipid field. It brought me great pleasure to read and is worth every penny of the $70 it cost. It obviously represents a labor of love for Dan Steinberg, someone who has contributed so much to this field and who has been a shining example to several generations of young scholars. Without leading academicians like him, the development of statins and the resolution of the cholesterol wars would never have happened.
Dr Gotto is a current consultant for KOWA Pharmaceuticals, Merck, Merck/Schering-Plough, and Reliant Pharmaceuticals, and serves on the board of directors for Aegerion Pharmaceuticals and Arisaph Pharmaceuticals. He serves on advisory boards for DuPont and Novartis.
Gofman JW, Hanig M, Jones HB, Lauffer MA, Lawry EY, Lewis LA, Mann GV, Moore FE, Olmsted F, Yeager JF, Andrus EC, Barach JH, Beams JW, Fertig JW, Page IH, Shannon JA, Stare FJ, White PD. Evaluation of serum lipoprotein and cholesterol measurements as predictors of clinical complications of atherosclerosis: report of a cooperative study of lipoproteins and atherosclerosis. Circulation. 1956; 14: 689–741.
Chen Z, Peto R, Collins R, MacMahon S, Lu J, Li W. Serum cholesterol concentration and coronary heart disease in population with low cholesterol concentrations. BMJ. 1991; 303: 276–282.