31st August 2018
One of the greatest problems in researching possible causal factors for any disease in humans, is that if you want to do clinical studies, you run straight into a major ethical issue ‘first do no harm’.
For example, if you believe that vaping causes heart disease, it would be extremely difficult to get the go-ahead to find ten thousand people, ask them to start vaping, and see what happens. To confirm your hypothesis, you need a significant number of people to drop dead.
Which is why almost all clinical trials, at least on humans, are designed to study interventions that are supposed to make people better. Of course, ethically this is a good thing, but it does make it extremely difficult to prove causality – for sure. Which is probably why Bradford Hill, in one of his famous nine criteria for causation stated the following ‘Experiment’. Occasionally it is possible to appeal to experimental evidence. (More on Bradford Hill later)
In effect, when we try to study causal ‘risk factors’ for disease, we are normally forced to rely on epidemiological, or observational evidence. We can look at risk factors in populations, but we can’t touch. At least we can’t touch, if we are trying to create the disease we are studying. This is why you get so much conflicting advice on, just to pluck a topic from thin air, diet.
However, there are times when you get a chance to look at a causal agent in action, as with smoking. On other occasions, the window opens by accident. A drug is being used to treat condition X, and you find it triggers disease Y. This recently happened with Proton Pump Inhibitors (PPIs) such as omeprazole.
PPIs were recently found to interfere with NO synthesis1. As readers of this blog know, NO is vitally important for endothelial cell health, preventing blood clots, and endothelial progenitor cell (EPC) production in the bone marrow. Knowing this, you would expect that PPIs would increase the risk of CVD. At least you would expect that, if you believe that lowering NO is likely to cause CVD.
As some of you know, I wrote a blog on this very issue and, yes, people prescribed PPIs have a significantly increased risk of CVD – almost a doubling of risk. Not that this has had the slightest impact on long-term omeprazole prescribing, anywhere.
Of course, you can argue that the data on PPIs did not come from an interventional clinical study, specifically designed to prove that PPIs cause CVD. You are never going to get one of those. However, in a world of imperfect evidence, this is the next best thing to experimental evidence. A drug that should, theoretically, cause CVD, causes CVD.
Moving beyond PPIS, there is another class of drug which could have a far greater impact on CVD. Before I get to it, I should remind everyone that the hypothesis I am outlining in this blog is that CVD is caused by three interlinked processes:
- Endothelial damage
- Clot formation at the site of damage
- Repair of the clot/damage.
These things are going on, all the time, in everybody. Atherosclerotic plaque growth – and potentially fatal blood clots – occur when damage > repair. Greater damage is caused by such things as: PPIs, or smoking, or air pollution, or raised blood glucose levels, or lead poisoning, or high blood pressure, or vitamin C deficiency, or sickle cell disease – and suchlike. However, you can also tip the balance towards plaque formation in the opposite way, by impairing the repair systems. Ensuring that: repair < damage.
One of the most important repair systems in the body consists of white blood cells, primarily monocytes and macrophages. These latch onto, engulf, and clear up the debris left by any assault of the body, including blood clots.
It is mainly the macrophages that do the heavy lifting. They destroy and digest any ‘alien’ material in the body. They start by firing a super-oxide burst at any junk in the body, which could be bacteria, or broken-down cells remnants, or what is left of blood clots. They engulf the ‘oxidised’ material, then they transport themselves to the nearest lymph nodes, where everything in them (and the macrophage itself) is broken down and, eventually excreted by the kidneys. [Or they get stuck, turn into foam cells, and die].
The other critical part of the repair system, following endothelial damage, are the Endothelial Progenitor Cells (EPCs) themselves. I have mentioned them many times in this blog. They are synthesized in the bone marrow. They cover areas of damage in blood vessels, and then mature and re-grow into a new layer of endothelial cells.
However, EPCs have another repair ‘trick’ up their sleeves. Because they are not mature cells, they can travel down other developmental pathways. Which means that they do not necessarily become mature endothelial cells, they can also transform into monocytes which, in turn, can further mature into macrophages.
Bringing all this together, if you find a drug that throws a spanner into EPC production – and thus macrophage development – whilst damaging NO synthesis and interfering with the growth of new endothelium, you will have found a drug that is almost perfectly designed to increase CVD risk.
And, yes, there is a class of drug that does exactly that and they are also, believe it or not, prescribed to humans. They are called vascular endothelial growth factor inhibitors. (VEGF-inhibitors). At one time is was thought that vascular endothelial growth factor (VEGF) was only active in the developing foetus, helping to stimulate EPCs, new endothelium growth, and driving the development of the entire vascular (blood vessel) system.
But it is now clear that VEGF still has a role in adults. It has a critical role in maintaining and helping to repair and re-grow the endothelium. Knowing this, you would expect that a drug specifically designed to inhibit VEGF could do some pretty serious damage to the cardiovascular system.
I have mentioned this class of drugs before, a few times, but I think it is worth highlighting them once more, as they provide almost perfect proof of the ‘three interlinked process’ hypothesis.
The most widely used VEGF-inhibitor is Avastin, the generic name is Bevacizumab. The mab at the end means it is a monoclonal antibody. It is an anti-cancer drug. Avastin works by inhibiting angiogenesis (‘angio’ = blood vessels, ‘genesis’ = new). Many cancers, as they grow, stimulate new blood vessel growth, which provides the tumour with the nutrients it needs. Cut the blood vessel production and the tumour shrivels and dies. This works. Avastin is an effective anti-cancer drug, and it is widely used.
Avastin is also used in macular degeneration where, in many cases, the growth of excess new blood vessels at the back of the eye (under the macula) is the problem, causing inexorably progressive blindness. With macular degeneration, Avastin is injected directly into the eyeball. (Yes, I know…ouch).
Avastin does not, as far as I can establish, seriously damage already existing endothelium – although I would imagine you would find that it does, if you looked hard enough. However, it seriously damages repair systems once the endothelium has been damaged. Therefore, it tips the scales heavily towards repair < damage. This effect has been directly studied in animals.
‘Systemic VEGF inhibition disrupts endothelial homeostasis and accelerates atherogenesis, suggesting that these events contribute to the clinical cardiovascular adverse events of VEGF-inhibiting therapies.2
That animal study was followed four years later, by a detailed review of all the clinical trials on Avastin, and the impact on ‘cardiovascular events.’. Some trials only went on for a few weeks, some were longer, lasting more than two years.
In this paper, the cardiovascular events themselves were listed in the strangest way I have ever come across. For example, we have, ‘arterial adverse events’ including arterial hypertension… is there any other sort?
‘Arterial adverse events’ were then further subdivided into one of the following: myocardial ischemia or infarction, cerebral infarction, cerebrovascular accident, cerebral ischemia, ischemic stroke, and peripheral or visceral arterial thrombotic events. Basically, it boils down to heart attacks and/or strokes – with a couple of other things thrown in.
Because I did not want to edit the results, I have listed them below, exactly as described in the paper. A risk of 2.40 means a two-point four times increase in the risk of something happening. This number could also be expressed as a 140% increase in risk.
The number 12.39 represents a twelve point three nine times increase in risk. Which can also be expressed as a one thousand, one hundred, and thirty-nine per cent (1,139%) increase in risk [These are relative risks].
INCREASE IN CARDIOVASCULAR EVENTS WITH AVASTIN3
Arterial adverse events 2.40 (1.64–3.52), P<0.001
Cardiac ischemia (heart attack) 5.16 (0.91–29.33), P=0.06
Cerebral ischemia (stroke) 12.39 (1.62–94.49), P=0.02
Venous adverse events 1.37 (1.11–1.68), P=0.03
Bleeding 2.96 (2.46–3.56), P<0.001
Arterial hypertension 4.81 (3.10–7.46), P=0.001
If you read the paper in more detail you will note that the longer the trials went on for, the greater the increased risk of an arterial adverse event.
At this point I think it is time to introduce you to the full set of Bradford Hills cannons/criteria for causation. Bradford Hill was a famous epidemiologist who worked with Richard Doll to ‘prove’ that smoking causes lung cancer. Within a certain arcane world, Bradford Hill’s cannons for causation are revered. I have listed them out below, having copied this version from Wikipedia.
What you may notice is that nothing in Hill’s list is black and white. He was wise enough to know that absolute proof in something as complex as disease causation, is very tricky. Very tricky indeed. There are often contradictions, and gaps, in the knowledge. However, with Avastin, every single one of his criteria are fulfilled.
Strength (effect size): A small association does not mean that there is not a causal effect, though the larger the association, the more likely it is causal. [Avastin can cause a 1,139% increase in stroke risk in less than two years]
Consistency (reproducibility): Consistent findings observed by different persons in different places with different samples strengthens the likelihood of an effect. [Every study on Avastin has shown the same thing, to a greater or lesser extent]
Specificity: Causation is likely if there is a very specific population at a specific site and disease with no other likely explanation. The more specific an association between a factor and an effect is, the bigger the probability of a causal relationship. [We have a very specific connection here, with no other obvious explanation]
Temporality: The effect has to occur after the cause (and if there is an expected delay between the cause and expected effect, then the effect must occur after that delay). [There is a clear delay with Avastin, the problems only occur after the drug is given]
Biological gradient: Greater exposure should generally lead to greater incidence of the effect. However, in some cases, the mere presence of the factor can trigger the effect. In other cases, an inverse proportion is observed: greater exposure leads to lower incidence. [With Avastin we have a clear biological gradient]
Plausibility: A plausible mechanism between cause and effect is helpful (but Hill noted that knowledge of the mechanism is limited by current knowledge). [The mechanism of endothelial damage is well identified, and plausible, with Avastin]
Coherence: Coherence between epidemiological and laboratory findings increases the likelihood of an effect. However, Hill noted that “… lack of such [laboratory] evidence cannot nullify the epidemiological effect on associations”. [Almost perfect coherence with Avastin]
Experiment: “Occasionally it is possible to appeal to experimental evidence”. [The experiment, albeit inadvertently, has been done]
Analogy: The effect of similar factors may be considered. [Other agents that interfere with NO, e.g. omeprazole, steroids, have the same effect]
Now, whilst I am reluctant to keep harping back to the LDL hypothesis, I think it is worth asking the question. Can the LDL hypothesis explain the increase in CVD with Avastin? Answer, no it cannot. Because Avastin has no impact on LDL.
Of course, as you might expect, Avastin does increase the blood pressure (BP). If you significantly lower NO synthesis, then the blood pressure will inevitably rise. So, the classical risk factors do have something to say about Avastin – if not a great deal.
ACE-Inhibitors, such as enalapril, or perindopril, are used to keep the BP down when people are prescribed Avastin. This works, primarily because ACE-inhibitors raise NO synthesis. [Although, to be frank, I do not know if anyone involved in treating the raised BP caused by Avastin has the faintest idea that is how they work, in this case].
Anyway, if you have a hypothesis that CVD is caused by three interlinked processes:
- Endothelial damage
- Clot formation at the site of damage
- Repair of the clot/damage.
Or, to be more accurate CVD is caused by any factor, or factors, that can
- Increase endothelial damage
- Create bigger and more difficult to shift blood clots
- Interfere with the repair systems.
Then, your attention is bound to turn to drugs that can do one of these three things. PPIs are one, VEGF-inhibitors are another. Whilst few things are absolute in human research, the evidence linking VEGF-inhibitors to a ‘three process’ hypothesis is, I believe, compelling.
It is certainly true to say the VEGF-inhibitors are sufficient, to cause CVD, by themselves. No need for any other risk factor to be present. Does this mean that they are THE cause of CVD? Of course not, but they are A cause of CVD, and their impact cannot be explained by any of the other traditional risk factors for CVD.
What does this mean? It means we have a black swan on our hands. The blackest of black swans. An agent, that is perfectly designed to create endothelial mayhem, causes CVD, with no explanation available within the LDL/cholesterol hypothesis.
Not only that, the data on VEGR-inhibitors fits every single one of Bradford Hills cannons for causation, and that is a rare thing indeed. You might even argue that VEGF-Inhibitors have allowed us a direct and uninterrupted view of the true ‘cause’ of CVD.