Homocysteine, an amino acid thought to play a role in the pathogenesis of atherosclerosis, is now forgotten or ignored by many physicians. Professional opinion and guidelines in the United States almost universally suggest not managing homocysteine levels in all except the most extreme cases. However, in forming these recommendations, experts relied heavily on a series of inadequate trials while simultaneously undervaluing a compelling body of literature to the contrary.
Homocysteine levels can be measured on a typical blood sample for ten or twenty dollars. Treatment for mildly or moderately elevated homocysteine levels is straight-forward. It usually consists of nutritional improvements and targeted supplementation, neither of which are associated with harm when they are implemented in a measured fashion.
Guidelines and expert opinions have prematurely dismissed homocysteine when it appears that it may be a modifiable risk factor for the primary prevention of cardiovascular and cerebrovascular disease. While this may remain a controversial issue for some time, the preponderance of the evidence is in favor of managing homocysteine on an individualized basis in order to lower the risk of heart attacks and strokes.
A Brief History of Cardiovascular Disease Risk Assessment and Treatment Controversies
Since the 1950s, physicians and scientists have sought to understand atherosclerotic cardiovascular disease (ASCVD) with the goal of preventing and treating heart attacks and strokes. The field of cardiovascular medicine may be the most controversial of the last hundred years. It seems risk factors and treatments are perpetually controversial. Assuredly, some of this arises from business incentives and bad actors, but most of it is simply because cardiovascular disease and lipidology are complex.
Lipid Theory of Disease Much of modern cardiovascular medicine is based on the lipid theory of disease. That is, atherosclerosis (plaques in the arteries) form over time because cholesterol from the blood stream is deposited in the arterial wall. We give drugs to reduce blood levels of cholesterol and we use mechanical interventions to treat narrowed or obstructed arteries.
Two of the most notable and mainstream treatments are statins and percutaneous coronary intervention (PCI), also known as a coronary catheterization. Statins like atorvastatin and rosuvastatin are two of the most popular drugs used to prevent coronary and cerebrovascular disease. These drugs have the ability to lower Apolipoprotein B (ApoB) bearing lipoproteins. ApoB lipoproteins include chylomicrons, VLDL, IDL, LDL, Lp(a), and Remnants. Each lipoprotein carries lipids including cholesterol. On a typical lipid panel, the target of therapy for most physicians is the LDL-cholesterol (LDL-C), which is the average concentration of cholesterol contributed to a fixed volume of blood by the ApoB lipoproteins. In most cases, this is almost entirely LDL-C. I’ve written in more detail on ApoB and lipoproteins here.
So, by lowering LDL-C with statins, we are usually lowering LDL as well. Statins can lower LDL-C by 20-60% depending on the drug, dose, and the individual.
Statins are thought to be responsible for significant improvement in the morbidity and mortality of ASCVD and are based on the lipid hypothesis of ASCVD. This hypothesis says that atherosclerosis can only develop when ApoB particles like LDL are retained within the wall of an artery. Therefore, lowering LDL-cholesterol (LDL-C) will decrease the rate at which atherosclerosis develops. Statins are also thought to stabilize plaques, decreasing the likelihood of plaque rupture and myocardial infarction (MI). In patients at low risk for ASCVD, meta-analyses demonstrate that statins may decrease all-cause mortality by 10% and decrease the risk of stroke and MI by roughly 20%.
In patients who have had a prior MI or are at high risk for MI, statin therapy has been demonstrated to reduce the 5 year risk of major vascular events by 22% and all-cause mortality by 10%.
Despite this well documented success, statins are still the topic of controversy. There are cases of patients with LDL-C levels 4 or 5 times the normal level who have lived into old age with no sign of atherosclerosis or cardiovascular disease. We also know that many patients who have heart attacks have completely normal blood levels of LDL-C. These examples are worthy of investigation, but they don’t negate the overwhelming evidence that many patients benefit from statins.
The other treatment which has transformed the field is percutaneous coronary intervention (PCI), in which a cardiologist deploys a stent or balloon across a narrowed or obstructed portion of a coronary artery. This procedure is responsible for saving the lives of patients who otherwise would likely have died from heart attacks. Even in non-fatal MI, opening an acutely obstructed artery can prevent myocardial death, thus limiting the damage done by an MI and preserving myocardial function. However, PCI is not without controversy. A landmark trial in 2017 failed to find a benefit for PCI in patients with stable angina when compared to maximal medical management alone. Again, there is seemingly conflicting evidence about the efficacy of PCI, but the truth is that some patients benefit from this procedure and others don’t. This treatment, like all others, must be used judiciously.
There are other metrics and treatments that have come and gone from the medical and public zeitgeist. For instance, it was long believed that dietary cholesterol was the primary culprit in causing elevated LDL-C and cardiovascular disease. In recent years, it’s been recognized that dietary intake of cholesterol has a limited effect on LDL-C and ASCVD in most otherwise healthy patients.
However, there are exceptions. Some patients absorb more dietary cholesterol than others. These patients might benefit from a low-cholesterol diet.
Likewise, there was a time when experts believed that high-density-lipoprotein cholesterol (HDL-C) indicated protection from ASCVD. High-density-lipoproteins (HDL) play an essential role in reverse cholesterol transport. It was thought that having high HDL-C might be as important or more important than having low LDL-C and that increasing HDL-C might prevent CVD. This idea came from the famous Framingham study which was an epidemiological study that established the traditional risk factors for ASCVD.
HDL and CETP
More recently, Nir Barzilai’s group found that a significant proportion of centenarians have high HDL-C and/or large HDL particle size. In some cases this was due to low levels of serum cholesterol-ester-transferase-protein (CETP). CETP is a protein that transfers cholesteryl-esters between lipoprotein particles and thus facilitates the indirect pathway of reverse cholesterol transport. Low levels of CETP are thought to protect these individuals from CVD by increasing HDL-C and HDL size and thus increasing reverse cholesterol transport.
In the early 2000s, drugs were developed to block CETP. These drugs worked in the sense that they increased HDL-C, but did not decrease mortality or CVD, at least not in the initial trials. In one of the trials, mortality rate actually increased, and in another, there was no improvement in atheroma volume in 24 months. Many experts believe that the failure of these drugs was not related to an inaccurate theory of HDL-C, but rather to unrelated toxicity of the drugs—they may increase blood pressure—which could offset the benefit conferred by increasing HDL size or HDL-C. For a brief yet thorough discussion of this subject, I recommend this NEJM editorial.
In any event, the failure of these trials cast doubt on the HDL-C theory of disease, as has subsequent research that has revealed the heterogeneity and complex functioning of HDL particles. The bottom line with HDL is that, while there is clearly a correlation between low HDL-C and increased CVD, and that high HDL-C is correlated with lower incidence of CVD on the population level, it’s not clear on the individual level that high HDL-C is protective and that low HDL-C is a risk factor.
HDL-C is a poor measure, because HDL is not just one thing. HDL is a blanket term that refers to a myriad of complex particles that have different sizes, features, and functions which vary to some degree from person to person. Measuring HDL-C and trying to conclude something about cardiovascular disease risk is like finding out that there are 2,000 people traveling on your local highway right now, and trying to deduce from that piece of data alone which vehicles they’re driving, where they’re going, and what they’re doing at their destinations. A head-count just isn’t enough information to draw these kinds of conclusions.
As a result of the complexity of HDL, HDL-C is not currently used as a target of treatment or prevention. But the fact remains that high HDL-C is strongly associated with a lower incidence of cardiovascular disease and vice versa.
The history of cardiovascular medicine is, and continues to be, reliably controversial. It’s necessary to consider and apply each piece of evidence contextually. Failing to do so can result in the dismissal of useful diagnostics and treatments, and thus a lost opportunity for patients.
Another cardiovascular risk factor that received massive attention within the medical profession in the 1980s through the early 2000s is an amino acid called homocysteine. Homocysteine is an amino acid that comes from the metabolism of another amino acid, methionine, which comes from dietary protein.
Source: Moll and Varga 2015
The homocysteine theory of atherosclerosis states that homocysteine can build up in the bloodstream in patients who have certain enzyme insufficiencies and/or nutritional deficiencies. This increase in homocysteine can negatively affect blood vessels, the clotting system, lipids, and can cause inflammation. As a result, it is thought to contribute to atherosclerosis and thrombosis, and thus heart attacks and strokes.
Dr. Kilmer McCully and Homocystinuria
Medical schools teach students about homocystinuria. It’s a disease of inborn metabolic error in which a child is born with a genetic deficiency in one of the several enzymes which are involved in the metabolism of methionine, homocysteine, and cysteine. This disease, if severe, can be fatal within the first decade of life by causing multi-organ failure. The metabolism of homocysteine requires multiple B-vitamins, methyl-donors like choline and betaine, and active enzymes in order to function properly.
Source: By Guillaume Pelletier – Own work, CC BY-SA 3.0
In 1968, a physician scientist named Dr. Kilmer McCully became aware of a case of homocystinuria in which an eight year old died from a stroke. McCully reviewed the case and found wide spread atherosclerosis.
McCully later became aware of additional cases. In each case, he found different in-born errors of metabolism, but a common thread of elevated homocysteine levels and atherosclerosis. This observation caused McCully to generate the hypothesis that homocysteine played a central role in the development of atherosclerosis. He spent the rest of his career researching this hypothesis. His research first led him back in time.
In 1953, an experiment revealed that monkeys formed hypercholesterolemia and atherosclerosis in response to a high-cholesterol diet only when they were also deprived of sulfur-containing amino acids like methionine and cysteine. Feeding them either of those amino acids would correct the dyslipidemia. Homocysteine levels were elevated in these monkeys before the correction, and low afterwards. The levels of homocysteine seemed to correlate with atherosclerosis and dyslipidemia. McCully took this as further evidence of an underlying relationship between homocysteine and atherosclerosis. Additional experiments in rabbits, baboons, and pigs demonstrated the ability of homocysteine to induce atherosclerosis.
In the 80s and 90s, scientists conducted observational and prospective trials in which correlations between elevated homocysteine levels and coronary artery disease were consistently found across various populations.
For instance, in this 1992 study of 14,916 male physicians from ages 40-84, 271 patients developed myocardial infarction (heart attack or MI for short) during a five year follow-up period. The study controlled for age and smoking to eliminate these as potential confounding variables. In the group that developed an MI, the average homocysteine level was 11.1 versus 10.5 in the group that did not have an MI. In addition, a graded effect of homocysteine was discovered, meaning that higher levels of homocysteine predicted higher risk for MI. Prospective studies like this one do not prove causation, only correlation, but this and others like it were evidence enough to cause physicians in the 90s to begin measuring homocysteine levels and trying to lower them.
The treatments that typically lower homocysteine levels are a combination of B-vitamins—6, 9, and 12, and methyl-donors like choline and betaine. These can be obtained in a healthy diet but can also be taken as supplements. Up until the early 2000s, there were no studies of the effects of homocysteine lowering on humans.
But now, we have eight double-blind placebo-controlled randomized trials that have attempted to answer the question—does lowering homocysteine with B vitamins decrease the risk of ASCVD? Each of the studies used slightly different approaches and had differing follow-up times ranging from as few as two to as many as seven years. All of the studies have been in patients over the age of forty and all patients in these studies have either known CAD, history of MI, or risk factors for ASCVD. These are not young and healthy patients.
None of the eight studies demonstrated an improvement in outcomes or in disease by the administration of B vitamins, even when homocysteine levels were successfully lowered.
Here are the possible reasons that these studies did not find improvements in ASCVD when homocysteine was lowered:
- Homocysteine does not have any causal relationship with ASCVD.
- The levels were not lowered enough.
- The follow-up times were too short.
- These patients already had disease, and lowering homocysteine cannot reverse or arrest disease, it can only prevent disease.
- The patients had only mildly elevated homocysteine levels and so this was not the driver of most of the pathology for the patients in these studies.
- The treatment did help, but it caused unintended consequences that resulted in no total improvement in disease or mortality.
- Since these patients were already at risk, they were on other disease modifying treatments that minimized their risk such that improving homocysteine had only a marginal and non-significant effect on top of these other treatments.
Nobody knows which if any of these are correct, but a combination of numbers three, four, five, and seven are probable. Given the large body of literature linking homocysteine and heart disease, number one is improbable. The failure of these studies is that they are looking at a high-risk population who is already being treated for ASCVD with other risk modifying therapies and who don’t have particularly high levels of homocysteine in the first place. In other words, it may be the case that lowering homocysteine in patients who are already diseased is unhelpful, and yet it might be helpful in patients who have yet to develop disease.
Primary Prevention Trials?
To date and to my knowledge, there are no homocysteine trials that used younger patients without ASCVD but who have moderately elevated homocysteine levels. What would happen in this population if they were evaluated for B-vitamin deficiencies, metabolic causes of elevated homocysteine, and were appropriately treated? Would they have better long-term health and lower risk of lifetime ASCVD? This question has not been answered.
It’s possible that homocysteine has no relationship to ASCVD, but I think at this point it is the hypothesis that must be disproven given the abundance of observational data in its favor and the lack of evidence that measuring or treating patients with elevations is dangerous.
Detractors Have Valid Points
Before drawing such a conclusion, it is important to consider the points made by detractors of the homocysteine theory. One problem with the evidence comes from McCully’s description of the atherosclerosis in the children who died from homocystinuria. These lesions do not exactly resemble the lesions found in adult atherosclerosis. McCully described the former as follows:
"A branch of the left coronary artery (Fig 2) was narrowed by a proliferation of intimal and medial loose vacuolated fibrous tissue, associated with fraying and destruction of the internal elastic membrane.”
In adults, however, atheromas contain cholesterol crystals and foam cells which are lipid and cholesterol laden macrophages. These were not seen in McCully’s homocystinuria cases, which he explains away as a matter of time. These patients were children and simply didn’t have time to develop the advanced lesions of adults. This could be true, but, there is another possible explanation, which is that cysteine and cystine are required for proper collagen synthesis which is an important building block of connective tissue. In addition, the presence of homocysteine may inhibit proper collagen synthesis and function. With weakened connective tissue, patients with homocystinuria develop a myriad of complications including retinal detachment, MI, stroke, and increased risk of thrombosis, all of which can be linked to connective tissue dysfunction. So, it’s not clear that homocystinuria patients have the same pathophysiology of atherosclerosis as adults with mildly or even moderately elevated homocysteine levels. To give you a rough idea, patients with homocystinuria can have a homocysteine level 10-50 times higher than a healthy individual. Similar questions and problems exist with the animal studies which may not be generalizable to humans.
Despite these potential shortcomings, McCully and other proponents of the homocysteine theory of atherosclerosis maintain that the preponderance of the evidence is in favor of the idea that homocysteine plays a significant causative role in ASCVD. However, it remains far from proven whether or not they are right, and if so, exactly how it works.
Always Problems with Expert Opinion and Guidelines
In the meantime, we are faced with a decision, do we measure and manage homocysteine levels or not?
The easiest way to decide is to defer to expert opinion and guidelines. But, experts and writers of guidelines have a heavy burden. They know their recommendations will be followed by scores of physicians and will be responsible for the devotion of time and resources to follow recommendations. Further, guidelines are used as standard of care in medical malpractice cases. Thus, guidelines often do not push the cutting edge of medicine and instead err on the side of the safest recommendation that can be followed broadly and which place a lower burden on physicians to execute. Because of this, guidelines often don’t allow for innovative, individualized, or resource intensive approaches to medical care.
According to an article published in 2015 in the American Heart Association Journal on the “Cardiology Patient Page:”
“Elevated homocysteine levels are associated with an increased risk for cardiovascular disease (see Table 2). The higher the level, the higher the risk. Cardiovascular disease can lead to coronary artery disease, heart attacks, and strokes.”
—“Homocysteine and MTHFR Mutations | Circulation”, p. e6
Despite this acknowledgement, their recommendations are only to test in one very rare circumstance:
“Who Should Have Their Homocysteine Levels Tested? The only group of people in whom testing appears indicated are young individuals, such as <20 or 30 years of age, who have had an unexplained heart attack, stroke, DVT, or PE, who are being evaluated for the rare homocystinuria, particularly if additional physical abnormalities suggestive of homocystinuria are present.”
—“Homocysteine and MTHFR Mutations | Circulation”, p. e8
They rationalize this position by stating that, "data show that the risk is only mildly increased" and point to a 2010 statement by the AHA that said they do not consider high homocysteine levels in the blood to be a major risk factor for cardiovascular disease (I cannot find this statement). They also say that, "the lowering of homocysteine levels does not change the risk of future recurrent events.”
The paper goes through a number of situations and questions, and to each, the answer is the same: it is not recommended to test or treat homocysteine levels in any circumstance other than the one mentioned above.
The United States Task Force for Preventive Services made similar statements in 2009 including an acknowledgment that for each 5 micromol/L increase in homocysteine there may be an 18% increase in risk for coronary events. Despite this, they stop short of recommending testing or treating homocysteine levels on the grounds that "in several well-conducted trials, homocysteine therapy did not prevent CHD events in persons with known heart disease."
The Canadian Guidelines which were published in 2000 and have not been updated since are more measured:
There is insufficient evidence to include or exclude screening of tHcy [total homocysteine] levels in any population. Screening may enable identification of patients at high risk for CAD so that other risk factors can be managed aggressively. However, laboratory testing for homocysteine is currently restricted to research centres. Moreover, testing is not yet covered by provincial health insurance, and therefore patients may be required to cover the cost. (grade C recommendation)
Although folic acid effectively lowers plasma tHcy levels, there is insufficient evidence to suggest that its use would prevent CAD events. Adherence to the recommended daily allowance of dietary sources of folate and vitamins B6 and B12 may prevent hyperhomocysteinemia due to vitamin deficiency. Once elevated tHcy levels are discovered, vitamin deficiency should be ruled out to allow specific treatment and prevention of complications, such as neurological sequelae due to vitamin B12 deficiency. Some authorities recommend limiting folic acid intake to 1 mg/d or adding higher doses of vitamin B12 (0.2–1 mg/d) because of the theoretical risk of unmasking occult vitamin B12 deficiency. (grade C recommendation)
Overall, the USPSTF and the physicians publishing in the AHA Journal seem to believe that there is a strong signal for harm associated with elevated homocysteine levels, but can’t find a justification to measure or manage this problem. In order to validly hold this position, I think these authors would have to address the shortcomings of the trials that tested the efficacy of lowering homocysteine levels. Only the Canadian guidelines mention a nuanced approach that consider B vitamin deficiencies. Even though the writers of the guidelines don’t believe testing or treating homocysteine is helpful, they also fail to address whether it is harmful. The guidelines would be more understandable if they acknowledged the weakness of their position and discussed harms of testing and treatment.
Is Lowering Homocysteine Harmful?
I haven’t seen any evidence suggesting that simply having a lower homocysteine level is harmful, but there are several studies that suggest potential harms of treatment with homocysteine lowering B vitamins.
Increased Progression of Atherosclerosis from B Vitamins?
This study found an increased rate of progression of atheromas in patients who were treated with homocysteine lowering B vitamins.
Increased Incidence of Cancer in Patients Supplementing B Vitamins?
A clinical trial conducted in Norway, published in JAMA in 2009 found that out of 6,837 patients who were randomized either to B vitamins or placebo, the group who received B vitamins for roughly three years ended up having 53 more participants who were diagnosed with cancer (341 vs. 288) either during the 39 month treatment period or 38 month follow-up period. Neither of these studies is sufficient to conclude that lowering homocysteine or supplementing with B vitamins causes an increase in atherosclerosis or death from cancer, but they do suggest that it’s possible there are ill effects from unnecessary or excessive B vitamin supplementation.
Acting with Imperfect Information
Evidence in Favor of Homocysteine Hypothesis of ASCVD
There is a large body of evidence supporting the idea that high homocysteine is related to ASCVD which includes the following:
- biochemical experiments suggest elevated homocysteine levels could cause problems with endothelial function, thrombosis, connective tissue synthesis, and even increased atherogenicity of homocysteinylated LDL
- observations of patients with homocystinuria who develop severe and premature atherosclerosis along with a myriad of other diseases
- animal studies in which inducing B vitamin deficiencies or methionine excess or deficiency can induce or modify atherosclerosis depending on the circumstances, and furthermore, that in some monkeys, dietary cholesterol alone cannot induce atherosclerosis, but in the presence of methionine deficiency, it can
- many observational studies demonstrating higher rates of cardiovascular and cerebrovascular disease as well as venous thromboembolism and pregnancy complications in patients with higher homocysteine levels
- prospective studies demonstrating an association between ASCVD and homocysteine levels with a graded increase in risk coinciding with higher homocysteine levels
However, there is also the following contradictory evidence:
- several observational trials that fail to find a correlation between homocysteine and ASCVD
- randomized blinded and placebo controlled trials that fail to show a reduction in MI or death by lowering homocysteine with B vitamins
- one study that suggests a possible increase in the rate of progression of atherosclerosis in groups supplementing B vitamins
- one study that suggests a possible link between cancer and supplementation with folic acid
- Multiple American professional and expert guidelines recommend against testing or treating homocysteine levels
- Canadian guidelines suggest measuring and treating homocysteine levels on a case by case basis
An Individualized, Longitudinal, Thoughtful, Iterative Approach
My position is almost always that more data is better. If I’m taking care of a patient over a long period of time, I want as much data about that person as I can get at any point in time. If I measure a homocysteine level today and it’s mildly elevated, I can discuss this result with my patient and decide whether to treat depending on their age, the degree of elevation, their other risk factors, desire and ability to take supplements or change their diet, and new evidence if and when it becomes available.
Whatever we decide now, having additional information in the future is always useful. I might be able to use past test results to detect a trend in the future. In patients who track their nutrition, supplementation, and exercise, it might be possible to use homocysteine levels as a measure of adequate nutrition.
So, my position is that it’s best to measure homocysteine, to assess its relationship with other health parameters, to discuss the full picture with your patient, and based on the circumstances, to decide whether or not to make changes to nutrition, supplements, medications, or some combination of the three.
This is the methodology of an individualized, longitudinal, thoughtful, and iterative approach to preventive care.