11th October 2018
I have tended to avoid talking about blood pressure, because I am not entirely sure what I think about it as a cause of CVD. However, since more people now take blood pressure lowering medication than any other type of medication in the world, including statins, it wold be remiss of me not to at least mention it.
Another problem is that, whilst blood pressure may seem a very simple subject. Either it is high, or it is not, nothing could be further from the truth. It is immensely complicated, and fragments rapidly into thousands of different strands, looping and whirling in front of you.
For example, let us take an apparently simple question, what is a high blood pressure? Well, with almost every passing year, this changes. The experts and the guideline writers get together on a regular basis and decide that well, hey ho, we thought 140/90 was high, turns out we are wrong. It is 130/85 – or whatever. By the way, the definition of a ‘normal’ blood pressure always goes down – never up. On current trends we should hit 0/0mmHg by the year 2067. What happens after that is hard to say.
I suppose a question that may seem reasonable to ask is the following; is average normal. Not, not even slightly. In a similar fashion to blood cholesterol levels, average and normal do not even remotely match up. Last time I looked eighty-five per cent of the population in almost all Western countries had high cholesterol levels. I would suspect another eighty-five per cent have high blood pressure.
In fact, if you wish to stretch logic to its very boundaries it is possible to propose that 99.9% of the population has a high blood pressure level. How can this make any sort of sense, you may ask? Well, in a moment of ennui I opened the American College of Cardiology/American Heart Association (ACC/AHA) risk calculator. You can find it here. http://www.cvriskcalculator.com/
I put in all my risk factors, kept them all the same, apart from my blood pressure which I started moving up and down, as you do when there is nothing good on the TV. What I found was that, as I reduced my blood pressure on the calculator, my CV risk went down, and down, until I got to a systolic (upper figure) of 90mmHg. You cannot make your blood pressure go any lower than this on the calculator.
The reason why you cannot get your blood pressure below 90mmHg is that if you go below this figure, you will be diagnosed with hypotension. Hypo = low. So, we have the weird situation whereby at 90mmHg your blood pressure is perfect. Above this, your risk of CVD goes up, below this the pressure is dangerously low and should be raised.
Therefore, at exactly 90mmHg your blood pressure is ‘normal’. At any other pressure it is abnormal – in that it increases the risk of death. Which is the definition of any ‘abnormal’ clinical test. It must be said that this constitutes a pretty narrow range. A doctor should be trying to keep your systolic blood pressure between 90mmHg and 90mmHg. And good luck with that. A very delicate titration indeed.
Clearly this is nuts, and it is not based on any clinical data whatsoever. There has never been a study whereby the systolic blood pressure has been lowered to 90mmHg. Nor will there ever be one done. This, I can guarantee. The mortality rate would be catastrophic.
So, how is this figure arrived at?
It comes from a mathematical smoothing technique whereby you get all the points on a graph, then draw the ‘best fit’ line through them all. I include an example here, which is where someone (Zoe Harcombe actually) looked at the cholesterol levels and rate of death from CVD in every country in the world (these are the dots). As you can see everything is rather scattered. [Data taken from the World Health Organisation].
However, there is a trend here, and that trend can be worked out. In this case, you feed in all the data points and a formula works out the underlying association between cholesterol and CVD death. As you can see in this case, as cholesterol goes up – CVD deaths go down. If you half close your eyes (which gets rid of the outlying points) the association between the dots and the line seems clearer.
(Or maybe that is just me).
What does this prove? Well, it proves nothing very much, for certain. What it almost certainly disproves, however, is an association between raised cholesterol and CVD.
When it comes to blood pressure it cannot be denied that. as the blood pressure rises, the risk of CVD also rises. However, the association is non-linear. By this I mean that, if your blood pressure goes from 100mgHg to 110mmHg, the risk of CVD does rise (but not by a statistically significant amount]. It only rises very, very slightly.
From 110mmHg to 120mmHg another very slight rise
From 120mmHg to 130mmHg another very slight rise
It is only when you get to about 160mmHg that the risk suddenly starts to go up sharply. From then on, things become rapidly worse. So, if your systolic blood pressure is above 160mmHg you should probably do something about it. However, what is the risk for a systolic blood pressure below this? Here we must rely on mathematics.
A paper that I have mentioned a few times, because I think it is a belter is from the European Heart Journal – from the year 2000 (believe me nothing of significance has happened since then). It is entitled ‘There is non-linear relationship between mortality and blood pressure.’
It is worth quoting the first two paragraphs in full. Sorry, for those not of a scientific bent:
‘Stamler stated that the relationship of systolic blood pressure (SBP) to risk of death is continuous, graded, and strong, and there is no evidence of a threshold…’ The formulation of this ‘lower is better’ principle, in terms of the linear logistic model (often referred to simply as the linear model) is the paradigm for the relationship of all cardiovascular risks to blood pressure and form the foundation for the guidelines for hypertension [and still does].
But it is often forgotten that when a study reports a linear (or any other) relationship between two variables it is not the data itself, but the model used to interpret the data, that is yielding the relationship. Almost universally, studies that report a linear relationship of risk to blood pressure do so via the linear models, such as the Cox model, or the linear logistic model.
Formally that model can be applied to any bivariate data and, independently of the data will always show that there is a linear relationship between the two variables. Before one can have confidence that the stated linearity correctly reflects the behaviour of the data, and is not just an artefact of the model, it is necessary to carefully examine the data in relation to the proposed model. At a minimum, it must be demonstrated that the model actually ‘fits’ the data and that it does not ‘smooth away’ important features of the data.’ 1
To paraphrase, your carefully constructed mathematical model may well be bollocks. In fact, it most probably is.
The statisticians who wrote this paper went back to the Framingham study – from whence all guidelines on blood pressure have since flowed, in all countries, everywhere – and found that the data ‘statistically rejected the model.’ They made the following statement ‘the paradigm MUST be false.’
I hate to say it, but the first person to recognise that the linear model ‘in terms of the relationship of overall and coronary heart disease death to blood pressure was unjustified.,’ was Ancel Keys. I am not sure what to make of that, as Keys is my number one medical historical villain. Still, he wasn’t stupid.
Anyway, the response to the European Heart Journal paper was…. Complete silence. Nothing. No counter arguments were proposed, nothing. “First they ignore you, then they laugh at you, then they fight you, then you win.” Gandhi. In this case we never got beyond ‘first they ignore you.’ Oh well, such is life.
None of this means that blood pressure has no role to play in CVD – or vice-versa – I just wanted to make it clear that that the whole area has become such a mess that it is very difficult to see through the forest of bias. What is a fact here? Frankly, sometimes, I have no real idea.
So, where does this leave us regarding blood pressure and CVD? It leaves us with only a few certainties. First, and most important, the only blood vessels in the body that normally develop atherosclerosis are the larger arteries. These blood vessels have a high blood pressure in them. Let us say around 120/70mmHg.
[120mmHg is equivalent to a column of water about three metres high. The units used to measure blood pressure are mmHg i.e., millimetres of mercury. Mercury was used to measure blood pressure, because it is many times denser than water. To measure blood pressure using a water sphygmomanometer would need a device more than three metres tall.]
On the other hand, the larger arteries in the lungs (pulmonary arteries) have an average blood pressure of around 20/6mmHg. In normal circumstances they never develop atherosclerosis. The veins have a blood pressure of about 6mmHg. It does not go up and down, because the pressure in the veins is unaffected by the pumping of the heart. The veins never develop atherosclerosis.
This makes it clear that the blood pressure, and turbulent blood flow, needs to reach a certain level before atherosclerosis can start. I sometimes liken this to a river flowing down a mountainside, with rushing and roaring and white water foaming and raging. That would be an artery. When the river reaches the plain below, the speed of water flow drops, the river widens and meanders. That would be a vein.
I think it is pretty clear that the lining of an artery is put under far more biomechanical stress than the lining of a vein – or a pulmonary artery. Which is why atherosclerotic plaques develop in [systemic] arteries, and nowhere else.
This idea is further supported by the fact that it is perfectly possible to get atherosclerotic plaques to develop pulmonary arteries, and veins. But only if you significantly raise the blood pressure. There is a condition known as pulmonary arterial hypertension (high blood pressure in lungs). There are many causes of this, and I am not going through them all here.
Let’s just say that people who suffer from pulmonary hypertension can, and do, develop atherosclerosis in the lungs. It should be pointed out that the pressure still gets nowhere near that in the rest of the body, perhaps 50/20mmHg, or suchlike. However, the blood vessels in the lungs were never designed to cope with high(er) blood pressure, and so damage will occur at a lower level.
When it comes to veins, if you take a vein from the leg, and use it as a coronary artery bypass graft (CABG), it will very rapidly develop atherosclerosis. Both of which prove that there is nothing inherently different about arteries and veins that normally protects veins and pulmonary arteries. It is all due to pressure.
Low pressure – no atherosclerosis
High pressure – atherosclerosis
So, surely the lower you get the blood pressure the better? Maybe, maybe not. There are many, many, other issues to be taken into account here – some of which I will discuss in the next blog.
1: Port S, et al: ‘There is a non-linear relationship between mortality and blood pressure.’ Eur Heart J, Vol 21, issue 20 October 2000