This review was inspired by a recent paper in the Annals of Clinical and Laboratory Science by Uffe Ravnskov and Kilmer McCully.1 Ravnskov, a physician and independent researcher, is probably the leading critic worldwide of the cholesterol hypothesis. McCully, a pathologist now associated with the Veterans Affairs Healthcare System in Boston and the Department of Pathology, Harvard Medical School, is famous for proposing the connection between homocysteine and atherosclerosis which is now generally accepted but was rejected when first proposed.
Today it is widely believed that high cholesterol causes atherosclerosis and thus cardiovascular disease. The public is reminded of this view frequently through pharmaceutical TV advertising where an actor tells us that now that he has had a heart attack he regrets not paying more attention to his cholesterol and now trusts his heart to a lipid-lowering drug. Few dare challenge this conventional wisdom in spite of the fact that there are serious, even fatal defects in what is really only a hypothesis. Graphic TV advertising also attempts to convince the general public that lipid-lowering drugs halt or reverse atherosclerosis although as regards coronary plaque, the evidence is compelling that this is not the case and that there is no clinically significant evidence to support this assertion.2 The cholesterol hypothesis of cardiovascular disease (CVD) and in particular coronary heart disease (CHD) has evolved to become dogma. This has seriously impeded scientific research into alternative explanations which might lead to more successful therapy and has led to the tendency to grossly exaggerate the benefits of therapy based on the dogma, principally through the use of relative rather than absolute benefits and the failure to distinguish between true primary prevention and secondary prevention. The peer review process as applied to publications in medical journals and in the research grant approval process provides a high barrier for those critical of conventional wisdom or daring to think outside the box. The engine that drives the current conventional wisdom is in part the huge profits associated with lipid-lowering drugs.
This review will examine what is in fact a very old theory for the explanation of atherosclerosis, namely infections. This was a popular theory a hundred years ago but was supplanted after the Second World War when the cholesterol hypothesis came into vogue. Many consider the failure of trials with antibiotics to have sealed the fate of the infection hypothesis. In what follows, we will first present the Standard Model for the development of atherosclerosis, briefly outline its weaknesses and inconsistencies, and then examine the infection hypothesis as an alternative approach to explaining the relevant biological and clinical characteristics of CVD and CHD.
THE STANDARD MODEL
The first step, according to the Standard Model, a term used in particle physics to designate a prevailing paradigm, is dysfunction or damage to the innermost layer of cells in the artery. This single cell layer is called the endothelium and constitutes the surface over which the circulating blood flows. Damage or dysfunction is viewed as caused by elevated cholesterol, elevated homocysteine, or other toxic factors in the blood. Descriptions of the Standard Model explicitly state that high levels of circulating LDL trigger endothelial cell dysfunction in the artery.3 The damage is then hypothesized to allow the migration of cholesterol, including LDL, and blood cells (monocytes) into the arterial wall. The next event is the oxidation of the LDL which leads to an immune reaction and this modified LDL is converted by monocyte derived macrophages into lipid-laden foam cells to form the early stage of plaque formation. This is similar to the innate immune system reaction, and is postulated to be triggered by oxidized LDL. This process it viewed as taking place just under the endothelial layer below what is called the basement membrane in what is called the intima.
The final act involves only some of the plaques which are described as vulnerable, which translates into ease or susceptibility of rupture, and the dumping or leaking of the vulnerable plaque contents into the circulation resulting in either a clot-induced reduced blood flow and unstable angina or total blockage and a heart attack. Many plaques are not vulnerable, in part due to a fibrous cap which protects them for erosion and rupture. The ruptured vulnerable plaques are called "culprit" plaques and can be found and studied in detail at autopsy.
This outline omits many details and variations but represents the essential features of the Standard Model. It is noteworthy that this model involves a process which starts at the inner wall of the artery, the endothelium, and requires a mechanism that involves the penetration of the endothelium by monocytes, the precursors of macrophages which accomplish what is called phagocytosis, part of the end-game in the immune reaction that destroys undesirable and dangerous substances. As well, the penetration of other blood components including cholesterol is postulated. The Standard Model ignores the fact that the artery itself with its smooth muscle core (the media) has an external blood supply called the vasa vasorum. The Standard Model seems to view everything that takes place in the artery as using the endothelium as the entrance and exit even though the vascularization of the artery is via the vasa vasorum as is the vascularization of arterial plaque. This model also views elevated LDL cholesterol as the driving force behind atherosclerosis and advocates of this hypothesis point to the risk reductions for acute events achieved by cholesterol-lowering drugs in selected individuals, even though the absolute risk reductions are small and the drugs used have many non-lipid lowering actions that may account for the majority of benefits. Furthermore, this view ignores the compelling evidence that there is no correlation between cholesterol or LDL and plaque prevalence or progression. It would appear that various aspects of the pathophysiology are merely selected to fit the prevailing dogma which of necessity must have circulating cholesterol as a major actor since that is what is measured in screening, presents the weak correlations with CVD in selected individuals and is the target for therapy. Historically, there are numerous examples of dogma that turned out to be wrong but had a strong, even official following. Contradictory evidence was ignored since a priori it had to be wrong.
A significant portion of the evidence for this hypothesis rests on rodent and rabbit models of atherosclerosis, but we differ from rodents and rabbits in aspects relevant to atherosclerosis. It should be well known that rabbits did not evolve on a diet rich in cholesterol, but are fed diets high in cholesterol for the purposes of studying atherosclerosis. Rabbits presumably never were carnivores or scavenged for eggs and one can question their use in modeling human atherosclerosis. The Standard Model also is partly justified by mechanistic inferences based on what is found in dissected plaques, a practice potentially associated with the risk of being misled.
Ravnskov and McCully1 summarize and document the major conflicts of the Standard Model with clinical, epidemiological, pathological and experimental observations.
To this list can be added the compelling evidence derived from non-invasive imaging studies that that there is no association between total or LDL cholesterol and the presence or progression of coronary plaque, both calcified and non-calcified, a result consistent with a number of early autopsy studies. Furthermore, lowering LDL with drug therapy has no impact on the progression of coronary atherosclerosis.2 However, the lipid-lowering trials, mostly involving individuals a high risk of or with established heart disease, have had a huge impact in terms of reinforcing the almost universal belief in the Standard Model even though the absolute benefits were small. Furthermore, although disputed by the experts, lowering event rates by lowering LDL levels does not prove that LDL is a causative factor since the drugs used have a very large number of non-lipid lowering effects, more of which are being discovered every month, and these so-called pleiotropic effects would correlate with lipid lowering since both are, in general, dose dependent.
In addition to these problems with the Standard Model, there is the recent observation from a study that over 50% of a large group of patients hospitalised for a heart attack had normal or low LDL levels (= 105 mg/dL), and incidentally low levels were associated with decreased 3-year survival.4 There are many other inconsistencies and contradictions not included in the above list.5-7 Finally, the small risk of CHD associated with cholesterol levels seen in younger men may have its origin in psychological stress and exaggerated blood pressure response to aggravation with cholesterol acting only as a marker.8 Obviously, something important is being overlooked.
These problems with the Standard Model have incidentally been described in the medical literature and books and discussed for several decades but ignored by the proponents. Thus the question remains, what is the mechanism where an artery starts out as normal and ends up with vulnerable plaque which rupture and cause cardiovascular morbidity and mortality, while this never happens in veins but may when a vein is transplanted to serve the function of an artery? While mainstream medicine seems comfortable with the Standard Model and it forms the basis of much research and is the foundation upon which officially approved therapy rests, the science seems far from "settled", to borrow a term from the climate warming debate. Among the alternative hypotheses, the one involving infections as driving atherosclerosis appears to merit serious consideration, and the recent paper cited above should stimulate increased interest. Devout believers in the Standard Model would probably describe this renewed interest as an unjustified exhumation.
THE MICROBIAL HYPOTHESIS
The notion that bacteria and viruses are the main cause of atherosclerosis has a long history going back over a century. Early observations included finding arterial lesions in patients who died from typhoid fever and so-called hardened arteries in many who survived. The duration of an infectious disease and the degree of atherosclerosis found at autopsy appeared correlated, and the famous physician Sir William Osler likened what is now called a vulnerable plaque to a pustule (one common manifestation of acne) which is like a boil. it was generally thought early in the 20th century that arterial plaques were the result of irritation associated with infections or toxins.1 Studies cited by Ravnskov and McCully1 as well as a recent review9 and other studies10-13 have over the years provided the following circumstantial evidence:
The above provides an indication of the body of evidence which should direct attention to infections as a major actor in both the formation and progression of arthrosclerosis and acute cardiac and other vascular events.
WHAT MIGHT BE THE MECHANISM?
If it is assumed that infectious disease plays a major role in the etiology of atherosclerosis, just how does it do it? The Standard Model focuses on the inner arterial wall and its injury and penetration as events leading to the formation of plaque and acute events. A major point made by Ravnskov and McCully1 is that this model ignores the vasa vasorum, the microcirculation system serving the arteries, which provides another route into the main part of the artery (the media). In addition, they point out that the Standard Model ignores the remarkable role played by LDL particles in the immune system.
Ravnskov and McCully1 propose the following hypothesis for the sequence of events leading to plaque formation. This hypothesis is based on the aftermath of infection, the primary trigger. They start with the unappreciated role that LDL plays in the immune system. Ravnskov and McCully tabulate 12 studies which demonstrate that LDL binds microbes and microbial products, the most common microbial product being lipopolysaccharide. These LDL complexes stimulate the immune system to attack and destroy them through what is called phagocytosis, a process carried out by macrophages derived from monocytes in the circulation. Furthermore, there is evidence indicating these complexes of LDL and both microbes and their products lead to aggregation of the lipoprotein particles which can result in large particles that have the potential to block the microvasculature. Elevated homocysteine may also increase aggregation of LDL because of complex formation with homocysteine or molecules generated by the reaction of homocysteine with other proteins. In addition, they cite other lipid complexes that infections can promote.
While LDL does not initiate an immune response, its various complexes described above as well as the resultant aggregates do and if this takes place in the arterial wall or in the vasa vasorum, multiple end results are possible including the various immune and inflammatory responses that are through to lead eventually to arterial plaque, both fibrous (capped) and vulnerable. Aggregates may block the microvasculature (ischemia) resulting in oxygen deprivation and arterial tissue damage. The stage is set for macrophages generated from monocytes in the circulating blood to mount a campaign which results in the generation of oxidized LDL and a variety of inflammatory products. The nature of the plaque formed will depend on the details of the immune reaction, but the essential point is that this starts from the introduction of immune active molecules and complexes traceable back to infections which enter the artery via the vasa vasorum rather than through an injured endothelium as required by the Standard Model. Once a plaque develops the penetration of the endothelium by monocytes adds to the pool of potential macrophage precursors. If the immune system manages the above problems such that there is no permanent change, then no plaque forms, but when the damage is such that the changes are permanent, boil-like plaque can result which can be either innocuous aside from stenosis (partially or even totally blocking arterial flow) or a vulnerable plaque can rupture resulting in an acute event, the nature of which depends on the extent of blockage caused by the resultant clot.
According to this hypothesis, LDL does not enter the artery through the endothelium but via the capillary system of the vasa vasorum. Oxidized LDL is the result of the action of macrophages during the "digestion" of the complexes involving LDL (phagocytosis), not oxidation after the LDL enters trough the endothelium. Oxidized LDL can then further stimulate immune response and creature a vicious circle, but its oxidation is not a primary event. This may be why the use of antioxidants in prevention studies fails to produce significant results. When the aggregates involving LDL complexes obstruct the vasa vasorum circulation, local ischemia can occur within the arterial wall along with microbial growth and inflammation. The internal elastic lamina separating the media from the intima is breached as the plaque develops and pushes out the endothelial cell layer.
The authors discuss in some detail the evidence for each aspect of this mechanism and as well, show the mechanism to be consistent with the pathological findings when both stable, fibrous and vulnerable plaque are examined in great detail at autopsy.
THE FAILED ANTIBIOTIC TRIALS
The standard argument against the above infection based hypothesis involves clinical trials using antibiotics. It was in fact these studies that led to the dismissal of the theory a number of years ago. Ravnskov and McCully argue that these unsuccessful trials involved mostly the use of a single antibiotic known to be effective against Chlamydia pneumoniae, one of the organisms found in atherosclerotic plaques. However, over 50 different microbial species have been identified in these plaques, but not a single one in normal arterial tissue. Patients typically have on average 12 different microbial species in the plaque. Thus it is highly unlikely that a single antibiotic would be effective. In addition, antibiotics are not effective against viral infections. Also, the trials regarded as falsifying the infection hypothesis were of relatively short duration. There is also the study mentioned above which found a strong association between the number of positive antibody reactions to different pathogens and the degree of atherosclerosis and cardiovascular mortality. This strongly argues against putting much weight on single-antibiotic short-term studies. Thus it would appear that the rejection of the infection hypothesis and its retreat into obscurity were not based on satisfactory evidence but rather on studies that suffered from a flawed design.
IS CHOLESTEROL ACTUALLY PROTECTIVE?
Recognition that LDL cholesterol takes part in the immune process suggests that it has been unjustly demonized. Ravnskov and McCully cite a number of studies providing considerable evidence that cholesterol is protective against infectious diseases. Cholesterol levels have been found to be inversely associated with total mortality in the elderly. In addition there is an inverse association between cholesterol levels and mortality from respiratory and gastrointestinal diseases, most of which have an infectious origin. This inverse association is also found for post-operative mortality from abdominal infections, the risk of being admitted to hospital because of an infectious disease, and the risk of contracting HIV and AIDS. Lipid lowering drugs such as the statins have a large number of non-lipid lowering actions. Some of these may neutralize the risks of infectious disease associated with an artificially induced low a cholesterol level.
It is also noteworthy that high cholesterol predicts longevity rather than increased mortality in older individuals, even those with genetically-related very high cholesterol.14 Also in many observational studies mortality was found to increase at low serum cholesterol levels. The mortality issue was discussed at length in a three-part Research Review in the November 2007 issue of International Health News, which concluded in the February 2008 issue.
THE IRON CONNECTION
Most readers of this Newsletter are aware of the advice to men to avoid multivitamins with iron and that for women the normal reduction of iron stores during menstruation is probably protective in the context of iron-related disorders. The hypothesis that elevated body iron stores were a risk factor for coronary heart disease (CHD) was suggested in 1981 by Jerome Sullivan. It was based in part on the low incidence of CHD in premenopausal women as compared to age-matched men. This was attributed to monthly blood loss and supported by the disappearance of this gender difference in postmenopausal women. Today the iron-heart hypothesis is viewed as valid and testable. One of the most significant obstacles to seriously considering this theory was that individuals genetically predisposed to iron overload (hemochromatosis) do not experience increased risk of atherosclerosis. In the past few years the recognition of the role of the hormone hepcidin in the regulation of iron balance and recycling has brought the iron-heart hypothesis to a new level of credibility and removed several strong objections. In addition, the iron-heart hypothesis points to the macrophage (the central player in the immune reaction) as an important player and brings us back to the connection with infectious disease where the macrophage releases inflammatory substances during phagocytosis.
If there is anything in cardiology that is beyond question and rests on a rock solid foundation of evidence, it is the fact that there is a very large gender difference in the risk of CVD which decreases in the postmenopausal years. In fact, as Sullivan points out15 not only are healthy premenopausal women protected from CVD, so are premenopausal women with familial hypercholesterolemia who turn out to have the same rates as women without this genetic disorder. Actually this is another argument against the cholesterol hypothesis. The conventional wisdom views this gender difference as a hormonal phenomenon, but this view is not sustainable in the light of a number of studies and randomized trials involving hormone replacement therapy. Indeed, already in 1963 there was evidence from studies of women who had both ovaries removed and exhibited essentially identical prevalence of atherosclerotic heart disease as controls. That result has never been challenged. Sullivan suggests that these observations point to a uterine rather than ovarian (hormonal) function that is responsible for premenopausal CVD protection.15 He points out that as part of the Framingham research, natural menopause, premenopausal simple hysterectomy, or premenopausal hysterectomy with ovary removal all were associated with similar increases in heart disease risk. He also cites additional supporting studies. However, the observational Nurses' Health Study provided evidence that contradicts this view in that a simple hysterectomy was not found protective. The reason for this discrepancy is not clear, but the weight of evidence appears to point to a protective function of the intact uterus with its monthly cycle. The iron-heart hypothesis provides a simple explanation. The healthy range of serum ferritin, a measure of iron load, is commonly given as 80-90 mg/L. Menstruating women typically have values below 30 mg/L.15
The current understanding of the role of hepcidin in iron metabolism and regulation has had a great impact on the iron-heart hypothesis. This small protein (25 amino acid peptide) appears to have first been isolated and purified in 2000-2001, which in itself is quite remarkable considering the important role it plays in human iron biochemistry. Hepcidin has the following properties:16
These functions and properties make clear the connection between iron status, infection and inflammation. The critical aspect in terms of the infection hypothesis of CVD is hepcidin's role in regulating the release of iron from macrophages and thus its role in the iron content of atherosclerotic plaques which contain both macrophages and the remnants of macrophages which have undergone cell death. As Sullivan discusses in a recent paper,17 there is growing evidence that iron depletion protects against atherosclerosis and atherosclerotic plaque have been found to have roughly ten times the iron content of the healthy arterial wall. Inflammation may increase levels of hepcidin which in turn promotes iron retention in macrophages within the atherosclerotic plaque and in extreme cases can promote the bursting of the macrophage releasing reactive iron into the interior of the plaque. On the other hand, low levels of hepcidin are associated with hemochromatosis and this would reduce the macrophage iron levels, given that this protein regulates the iron uptake and loss in this immune system cell. This of course would explain the paradox of the lack of association between hemochromatosis and arterial atherosclerosis and remove one of the major obstacles to the iron-heart hypothesis.
HOW IT ALL FITS TOGETHER
According to Ravnskov and McCully, their hypothesis explains a number of clinical and pathological observations. These include why all the classic features of infectious disease are common in a significant number of individuals presenting with a heart attack, and why elevated CRP in individuals with atherosclerosis is a marker of increased risk of heart attack. Furthermore, fatty streaks, viewed in the Standard Model as plaque precursors, appear in newborns and children and frequently disappear, and this can be viewed as a normal and reversible response to infections.
The hypothesis explains why atherosclerosis is related to elevated blood pressure and the associated hydrodynamic pressure because lipoprotein complexes generated in response to infections are more easily trapped in the arterial vasa vasorum due to pressure differentials. Pulmonary hypertension appears to produce the same effect with the development of plaques in pulmonary arteries when conditions arise that lead this type of hypertension.
The current view of the anatomy of the vasa vasorum is that these vessels are so-called end-arteries, supplying the smooth muscle core of the artery where blood flow and patency are compressed by the pressure derived from the arterial circulation. When congested with the products of an immune response, end-arteries can burst, depositing elevated concentrations of toxins directly into the arterial tissue, initiating or enhancing the immune response and producing local inflammation.
If irritants directly attacked the endothelium as the Standard Model requires, then atherosclerosis should be just as common in veins, and if elevated LDL were the driving force then atherosclerosis should be a more generalized disease. The hypothesis addresses these issues. Veins function under a different pressure environment, and LDL functions as part of the immune system rather than driving atherosclerosis by invading through the endothelium and then being oxidized to start the process.
There is consistency between the infection hypothesis and other established risk factors. Mental stress increases the risk of infections, partly through its impact on the hormone cortisol. Anger increases homocysteine levels. Smoking predisposes to a number of infectious diseases and also causes an increase in homocysteine levels. Diabetes is a major risk factor for heart disease and diabetics have a higher prevalence of infections than do healthy individuals.
This hypothesis also raises questions regarding the mechanistic role of inflammation since inflammation is a necessary step in the normal healing process but is implicated in the etiology of heart disease. In this context, the connection between inflammation and homocysteine may be part of the answer. Studies suggest that homocysteine in an enhancer of inflammatory activation and as well plays a role in the autoimmune triggering process. Thus its connects inflammatory factors to the acceleration of atherosclerosis.18 The well known connection between autoimmune diseases such as rheumatoid arthritis and the risk of heart disease is also consistent with the hypothesis since elevated homocysteine is associated with these diseases,18 and according to the theory of Ravnskov and McCully, complexes of homocysteine with LDL and other proteins are postulated to participate in the immune triggering and vasa vasorum blockage that eventually leads to plaque formation.
Ravnskov and McCully discuss modifiable risk factors associated with this hypothesis. From these one can outline preventive actions.
Other suggested actions that have appeared in the literature include hand-washing, avoidance of individuals with infectious diseases, vaccinations and diet. It can also be argued that it is a good idea to stay away from hospitals unless absolutely necessary since they offer exposure to antibiotic resistant infections.
The infectious disease hypothesis refocuses attention away from cholesterol to the immune system. Thus a prevention program would also involve maintaining a healthy and effective immune system. The hypothesis also considers a dysfunctional immune system as promoting unstable plaque formation. Naturally, infections can probably never be eliminated and during a lifetime there are bound to be a few serious infections. Perhaps this is why most older individuals have some atherosclerosis. It is noteworthy that optimum vitamin D status appears related to good immune response and a low risk of, for example, of influenza, and the risk of CVD is inversely related to vitamin D status. Furthermore, the gut plays an important role in immunity (see Chapter 3 of Gut and Psychology Syndrome by Dr. Natasha Campbell-McBride for a brief introduction), and in the age of excessive antibiotic use and poor dietary patterns, the dysfunctional gut appears to be common and its importance unappreciated.
In his new book Fat and Cholesterol are Good for You!, Uffe Ravnskov has a chapter on avoiding heart attacks that follows a chapter describing the hypothesis discussed above and presented in the paper by Ravnskov and McCully. He points out that the really relevant question involves avoiding premature death from heart attack or stroke. Once one has reached a ripe old age he points out that it may be better to succumb to a heart attack than die of cancer or problems related to Alzheimer's disease. Aside from actions discussed above, Ravnskov suggests eating a diet low in refined grains and sugars, avoiding ingesting the hundreds of chemicals mixed in with so-called foods and offered at the supermarket, exercising, and paying attention to the huge impact an individual's long-chain omega-3 status (EPA + DHA) has on the risk of sudden death due to "electrical problems". He also recommends that anyone with unstable angina or a victim of a heart attack or stroke should be checked out with a blood culture as soon as possible after admission and any infection targeted immediately. This approach to increasing the chances of avoiding a second acute event may be more effective than the conventional pos-heart attack protocol, or may offer considerable benefit when added to the standard protocol.
The preventive strategy outlined above would appear to involve no risks and should provide benefits independent of whether or not the infection hypothesis is correct. It is almost certain that this new view which points to infection as driving atherosclerosis and adverse vascular events will be resisted or ignored by mainstream medicine, but it is hoped that for those who are willing to think outside the box, benefits may derive from a renewed interest in the immune system as potential a major player in CVD.