11. March 2004

PROVE-IT AND BEYOND 
(originally published in RedFlagsDaily)

A Massive Nail In The Coffin Of The Cholesterol, Or LDL, Hypothesis?  

For those who follow my ramblings, I felt the need for a very rapid follow-up to my column on the PROVE-IT study, which has been hailed as proof that the more you lower LDL levels the greater the protection against heart disease, and death. 

A very similar study to PROVE–IT was done, looking at the effect of pravastatin and atorvastatin on the growth of the atherosclerotic plaque.   This was the REVERSAL study, published late last year, and a review of this trial was sent to me today. So I thought I should share it with you. 

In this study, as with the PROVE-IT study, atorvastatin reduced the LDL to a greater degree than pravastatin – on average. And the progression rate of atherosclerosis was reduced more by atorvastatin than pravastatin. 

This, again, would seem to prove that greater LDL lowering provides greater protection against CHD, and death. However, the investigators decided to look at the data in more detail in a ‘post-hoc' analysis. 

One of the things they looked at was the progres

sion of atherosclerosis in those patients taking pravastatin who achieved the greatest lowering of their LDL level. Down to 88mg/dl – similar to the level achieved on atorvastatin. 

Dr Steve Nissen, lead investigator from the Cleveland Clinic noted:

‘Surprisingly, despite attaining a low LDL level on pravastatin, these patients showed highly significant progression for percent atheroma volume and percent obstructive volume… At any LDL level, progression was less on atorvastatin than on pravastatin. When I started this study, I believed that any reduction in progression would just be due to lower LDL levels, but now I'm not so sure. This analysis suggests that it may be more than just LDL it seems to be the drug as well…. Yes, this is a post-hoc analysis and should be considered hypothesis generating,' but I would say it is a robust finding.' 

What does this mean? It means that atorvastatin protects against CHD more effectively than pravastatin. It also means that the degree of protection has absolutely nothing to do with the reduction in LDL levels. The protection is provided by a direct drug effect.

Which is, or should be, a massive nail in the coffin of the cholesterol, or LDL, hypothesis. Of course, it won't be. 

If I were to summarise all of the statin trials up till now it would be as follows:

·    Statins reduce the risk of heart disease

·    They reduce the risk of heart disease regardless of the starting level of LDL, and regardless of how much they lower the LDL

·    Statins protect in conditions where a raised LDL is not a risk factor e.g. stroke

·    Statins reduce the overall risk of death in a small group of very high risk men e.g. those who have already had an MI, or who have diabetes

·    Statins do no reduce the overall risk of death in women – at all, regardless of their level of risk.

Conclusion: Statins work through direct effects on atherosclerotic plaque development. This effect has nothing to do with lowering LDL levels.

March 10, 2004

PROVE IT – PROVED WHAT?  
(originally published in RedFlagsDaily)

New Study Data Has Researchers And “Rent-A-Quote” Doctors Pushing For More Intensive Cholesterol-Lowering Treatment. This Is Massive Hype. 

 ‘The implications of this turning point – that is, of the new era of intensive statin therapy – are profound. Even today, only a fraction of the patients who should be treated with a statin are actually receiving such therapy…. More than 200 million people worldwide meet the criteria for treatment, but fewer than 25 million take statins.'

- Dr Eric Topol Cleveland Clinic 

At last, to massive cheers from all Pfizer shareholders everywhere, it has been proved that the more you lower the LDL level (bad cholesterol) the greater the protection against heart disease, and death, and world poverty and… sorry, got a little carried away in the emotion of the moment. 

For years it has been known that statins protect against heart disease, and in a select groups of very high risk patients, it has been confirmed that statins may even reduce the death rate. But no-one has ever managed to demonstrate that the protection was related to the degree of LDL lowering. A fact that may surprise you, but it's true. 

Now, however, in the ever so aptly named PR avastatin O r Ator V astatin E valuation and I nfection s T udy PROVE-IT , (talk about a tortuous anagram) all remaining doubts have been swept away in an avalanche of data. Results have been produced that will, according to Dr Eric ‘rent-a-quote' Topol, ‘Herald a shake-up in the field of cardiovascular prevention.' Hmmm. I beg to differ. 

Let's start by looking a little more closely at the PROVE-IT study. What it was and what it showed. 

In PROVE-IT the investigators took 4,162 patients who had been in hospital following an MI, or unstable angina (almost, but not quite an MI). They then split the group in half and gave one half pravastatin (made by BMS), and the other half atorvastatin (made by Pfizer). As expected LDL level, or ‘bad cholesterol' level, was reduced to a greater extent in the atorvastatin group. 

LDL in the treated pravastatin group: average 95 mg/dl           (range 79 – 113)

LDL in the treated atorvastatin group: average 62 mg/dl          (range 50 – 79) 

In short, in the atorvastatin group there was a thirty two percent greater reduction in LDL levels, and there was also a sixteen percent greater reduction in – well – almost everything you can think of: all cause mortality, MI, unstable angina, hospital readmission, interventional procedures. You name it, it was all quite wonderful. 

Of course when they say a sixteen percent reduction, they actually mean a sixteen percent reduction in relative risk. Which, as you will hopefully all know by now, may mean something – or nothing? 

However, I shall be fair. Just focusing on all-cause mortality which, I think, is the most important figure to choose. The absolute reduction in the rate of death from taking atorvastatin, rather than pravastatin, was one percent, a decrease from 3.2% to 2.2%, over twenty four months. Or, to put it another way, a 0.5% absolute risk reduction per year. 

Maybe not as mind-boggling as you would hope, but in line with previous studies of this type. So what are the problems with this study? Apart from massive over-hype.

I suppose the most basic problem with the study is the old ‘two variables' conundrum. It is true that those with the greatest LDL lowering were protected against death.

However, as you may have noticed, those who were protected not only had a greater degree of LDL lowering, THEY WERE ALSO ON A DIFFERENT DRUG! which is rather important, yet seems to have been swept aside on a wave of hype. 

In reality, if you really want to prove that the more you lower the LDL level, the greater the protection, then you must use the same drug. This achieves an absolutely critical requirement of any scientific experiment, which is to remove all possible uncontrolled variables. 

Ideally, in an experiment of this type, you should not only use the same drug, you should also use the same dose of the drug; then group patients by how much the LDL was lowered, then see if the protection against death is related to the degree of LDL lowering. This would remove the possibility that a higher dose of the same drug protects against CHD through other non-lipid effects. Which is kind of important as, over the last few years, it has been found that statins do a great deal more than lower LDL. 

‘In addition to their lipid-modulating properties, statins have a large number of beneficial cardiovascular effects that have emerged over time and that were not anticipated during drug development…. By acting on the vessel wall, statins may prevent lesion initiation and repair injuries, enhance myocardial perfusion, slow lesion progression, and prevent coronary occlusion. They may also directly reduce myocardial damage, favor myocardial repair, and protect against immune injury.' Davignon J. Curr Atheroscler Rep. 2004 Jan 

Presently, for the PROVE-IT study to mean anything, requires that atorvastatin and pravastatin are identical in all actions, only differing in the amount that they lower LDL. If they do have other direct, and drug specific effects, which could explain the difference in protection from heart disease, then all the PROVE-IT proved was that atorvastatin provides more protection from heart disease than pravastatin – and this may have nothing to do with the impact on LDL. 

Why on Earth didn't the investigators choose to use different doses of atorvastatin to lower the LDL by different amounts? Or, even better, use the same dose of atorvastatin then group patients by the degree of LDL lowering achieved? What stopped them from doing either of these things? 

There is no practical reason why this couldn't have been done, and from a scientific perspective it would have been infinitely preferable. In fact, it would have prevented the study from being a completely meaningless load of rubbish. 

As this study presently stands, because they used different drugs, anyone can make the case that the benefits seen in the patients on atorvastatin had nothing to do with greater LDL lowering; they were purely due to direct drug effects of atorvastatin. And it is impossible for the authors to argue that this is not the case. 

In addition, there is some very powerful evidence out there that directly contradicts the hypothesis that the degree of LDL lowering, and the protection against death are connected. This evidence comes from across the ocean, and is provided by another study which – perhaps to no-one's great surprise, attracted very little attention at all.

A couple of years ago a much larger study than PROVE-IT finished in Japan. The Japan Lipid Intervention Trial (J-LIT) Matsuzaki M et al Circ J 2002;66: 1087 - 1095 .   In this trial over thirty thousand patients with raised LDL levels were put on simvastatin at the single dose of 5mg daily, and then patients were grouped by the degree of LDL lowering achieved – the ideal study if you want to remove all variables. 

As you would expect, although everyone took the same dose of the drug, not everyone's LDL level responded the same way. Some patients had no reduction in LDL levels (I would suggest that these were the patients who didn't bother taking the drug). Some had a moderate fall in LDL, and some had very large LDL reductions. 

This allowed for a nice, simple, single variable analysis to be done. The drug was the same in all patients, the drug dose was the same in all patients, and the patients were all chosen to provide a homogeneous sample with a similar starting level of LDL. The only thing that was different was the amount by which the LDL fell. Nice and simple. And what, gentle readers, do you think it showed. 

It showed that there is no correlation whatsoever between the amount of LDL lowering, and death rate. None. This, please remember, in a study that had ten times as many patients, lasted almost three times as long and – perhaps most importantly – used the same drug, at the same dose, in all patients. So it actually means something. 

In contrast what did PROVE-IT really prove? It proved that atorvastatin protects against heart disease and death better than pravastatin. What it most certainly did not prove is that the more you lower the LDL level the greater the protection.

J-LIT is the only study done so far that has looked scientifically, rather than marketingly, at the correlation between LDL lowering and protection against death, and it PROVED-THAT there is no connection at all.

  More essays by Malcolm Kendrick