I have been somewhat silent over the last two or three weeks on this blog. The word ‘swamped’ springs to mind. The main swamping thing (alongside work and suchlike) that I have been doing is to analyse the Lancet paper which claimed that, basically, statins cause no adverse events. Professor Peter Sever (corresponding author), followed up the publication of this paper with statements such as:

‘While statins do have some potentially serious side effects, including a slightly raised risk of developing type II diabetes and, very rarely, a potentially fatal muscle condition known as rhabdomyolysis, Sever said that the Medicines and Healthcare Products Regulatory Agency (MHRA) should remove warnings of side-effects including muscle pain and weakness, sleep disturbance, erectile dysfunction and problems with cognitive function” (https://www.theguardian.com/society/2017/may/02/statin-side-effects-down-to-negative-expectations-not-the-drugs-nocebo).

In an interview with UK national newspaper, The Daily Telegraph, Peter Sever went on to say that:

‘There are people out there who are dying because they’re not taking statins, and the numbers are large, the numbers are tens of thousands, if not hundreds of thousands.

He said it was a “tragedy” akin to the MMR scandal that high risk patients had been deterred from taking drugs which could save their lives. Urging patients not to “gamble” with the risk of heart attacks and strokes, he said “bad science” had misled the public, deterring many from taking life-saving medication” (http://www.telegraph.co.uk/news/2017/05/03/statins-myth-thousands-dying-warnings-non-existent-side-effects/).

And so on and so forth. This paper, as you may expect, has been picked up with great enthusiasm by the mainstream medical media, and other doctors. Here is a Dr John Mandrola writing a Commentary in Medscape.

‘The frequency of muscle symptoms with statins is hotly debated. Randomized controlled trials (RCTs) in which patients don’t know whether they are taking the statin or a placebo report nearly identical rates of muscle-related adverse events. Observational studies, however, report higher rates of statin muscle complaints.

As a practicing doctor, I have always felt the truth lies closer to the observational data. A study published recently in the Lancet suggests I may be wrong. This new study, which has impeccable methods, suggests statin muscle complaints stem not from human muscles but from the human brain. In the Lancet paper, researchers took advantage of two distinct parts of the primary prevention ASCOT-LLA trial.

In the first part of ASCOT-LLA, more than 10,000 people were randomized to either atorvastatin 10 mg daily or placebo in a double-blinded fashion. After completion of the blinded phase of ASCOT-LLA, study participants were invited to take part in a nonblinded and nonrandomized extension study in which they could take atorvastatin open label.

The results turned on whether people knew they were on the statin. In the double-blinded phase of the trial, muscle symptoms occurred at the same rate—2.0% per year in both the statin and placebo groups. In the second phase of the trial, when people knew they were on the statin, side effects occurred at a higher rate (1.3% per year) in the statin group vs the placebo group (1.0% per year). This difference reached statistical significance (hazard ratio 1.41, CI 1.10–1.79; P=0.006).

These are remarkable observations, which are hard to dispute. In an accompanying editorial, two Spanish authors emphasized the obvious strengths of this paper: these were the same patients in both phases, and there was no run-in period in which patients intolerant to statins were excluded’ (http://www.medscape.com/viewarticle/879762_print).

So, this is a slam dunk. Right?

Well, I have taking a pretty forensic look at the Lancet Paper. It has the snappy title. ‘Adverse events associated with unblinded, but not with blinded, statin therapy in the Angle-Scandinavian Cardiac Outcomes Trial – Lipid Lowering Arm (ASCOT-LLA); a randomised double-blind placebo-controlled trial and its non-randomised non-blind extension phase.’ May 2^nd 2017’.

You may not be surprised to know that Professor Sir Rory Collins was a co-author.

I believe it may have a weakness – or two – or three – or … you get the picture. However, if you are going to attempt to argue against such a paper, or pick holes in it, you need to study it with extreme care, to make sure that you have your facts absolutely right.

Then you need to look at all other associated papers around the entire ASCOT study. For example, I have been amusing myself, or not, by studying ‘Rationale, design, methods and baseline demography of participants in the Anglo-Scandinavian Cardiac Outcomes Trial’…. And a few other papers as well. I have also been speaking to some very bright people who understand exactly how clinical studies are done, how adverse events are reported and recorded. It is an arcane and opaque world indeed.

You need to try to understand comments such as this, in the paper:

Procedures

‘After randomisation, study participants were scheduled to be seen at 6 weeks and 3 months and then at 6 monthly intervals thereafter during both the blinding randomised and non-blinding randomised phases of the ASCOT-LLA (until the ASCOT-BPLA completed – yes this was two trials in one). At each study visit all adverse events (AEs) reported by participants were recorded by the study team in the case report form. Specific questions relating to any putative AEs were not asked at these visits.

Reports of AEs by the study participants were initially recorded verbatim and subsequently classified with use of the Medical Dictionary for Regulatory Activities into 26 separate system organ class (SOC) groups, 2288 unique preferred terms, and 5109 separated low-level terms…..’

Now, I defy anyone to make sense of that. [I had no idea what the word putative meant in this context. Having looked it up, I am none the wiser]. Either adverse reports were initially recorded onto a case report form, or comments were recorded verbatim and subsequently classified…. You can do one, or the other, not both. As for attempting to reclassify verbatim reports, in several different languages, fifteen years later…. Hmmmmm.

However, whilst trying to get my head around that, my interest was piqued by those involved in this data analysis. It turns out that the lead author, Ajay Gupta, was provided with financial support from the ‘Foundation for Circulatory Health’. I had never heart of this ‘charity’ before. So I tried to find out how it was funded – always tricky. You can usually find out who provides the dosh, but not how much.

Looking at their accounts, the foundation for Circulatory Health seems to be funded largely (almost entirely?) by the pharmaceutical industry. Companies which include, guess who, Pfizer, who funded the initial ASCOT study and who also funded the recent Lancet Nocebo paper.

Supporters (of the Foundation for Circulatory Health (http://www.ffch.org/supporters.html):

• Pfizer
• Sanofi-Aventis
• Menarini
• Novartis
• Medtronic
• Boston Scientific
• Pulsecor
• Patients attending the Hypertension and Cardiology Clinics

Digging further it then turned out that that Peter Sever and Neil Poulter (key authors on the ‘nocebo’ paper) are also directors of the Foundation for Circulatory Health, which Funded Dr Gupta to work on the Nocebo paper – supported by Pfizer. Well, who’d a thunk? [Well, me actually].

Neil Poulter is a very well-known researcher in CV medicine, well known to those who keep track of such things. His name turns up all over the place. Here was his declaration of interest statement in the Lancet paper:

‘Neil Poulter’s institution (Imperial College London) held a grant for the conduct of the Anglo-Scandinavian Cardiac Outcomes Trial in the UK and Ireland and he has also received a speaker’s honoraria from Pfizer outside the submitted work. He is also a recipient of the National Institute for Health Research Senior Investigator Award to Imperial College Healthcare NHS Trust.’

Sounds quite reasonable(ish) and above board. However, compare this with a conflict of interest statement from 2008: ‘Poulter disclosed receiving ad hoc payments to appear on advisory boards/deliver lectures for “all the major pharmaceutical companies that produce major agents in hypertension and CV medicine” and receiving grant income from Pfizer and Servier’(http://www.medscape.com/viewarticle/790044?t=1#vp_2).

Perhaps he just forgot that he had received money from all the major pharmaceutical companies that produce major agents in hypertension and CV medicine. Must be hard to keep track of what you have previously disclosed. Is there a time limit on conflicts of interest?

For now, I shall continue to dig. I shall continue to analyse the paper. Watch this space. It is all rather time consuming, but it may turn out rather well in the end. Although, I suppose, that rather depends on which side you are on in this debate.

Some of you will have noted that researchers have now decided that statins do not have any side effects at all. To be pedantic, the correct term is not side-effects, it is drug related adverse events. A side effect can be positive, or negative.

In order to prove that statins cause no adverse events, a paper was published in the Lancet entitled: ‘Adverse events associated with unblinded, but not with blinded, statin therapy in the Anglo-Scandinavian Cardiac Outcomes Trial—Lipid-Lowering Arm (ASCOT-LLA): a randomised double-blind placebo-controlled trial and its non-randomised non-blind extension phase.’

A virtually impenetrable title which could mean almost anything. But the key message can be found here:

‘These analyses illustrate the so-called nocebo effect, with an excess rate of muscle-related AE reports only when patients and their doctors were aware that statin therapy was being used and not when its use was blinded. These results will help assure both physicians and patients that most AEs associated with statins are not causally related to use of the drug and should help counter the adverse effect on public health of exaggerated claims about statin-related side-effects.’

Funding: Pfizer, Servier Research Group, and Leo Laboratories

Statement by authors in original ASCOT study [The Lancet vol 361 April 5th 2003. Pp1149-1158] ‘The Anglo-Scandinavian Outcomes Trial (ASCOT) is an independent, investigator-initiated and investigator-led multicentre, randomised trial designed to compare two antihypertensive treatment strategies for the prevention of CHD events…’

Funding of the original ASCOT study: Pfizer, Servier Research Group and Leo Laboratories

The ASCOT study was published over fifteen years ago.

There was a lot of noise about this study on the radio, newspaper and television. At least there was in the UK. Professor Peter Sever, one of the authors, and a key investigator, stated on the radio, that the inserts warning of drug related adverse effects should be removed from the packaging, as they simply encourage patients to believe that they are suffering from adverse effects. He also stated that statins caused muscle problems in less than one in ten thousand patients.

I tend to disagree with him. I was asked to be interviewed on various radio stations, including BBC radio Scotland, and to write a newspaper article for the Scotsman newspaper. It went as follows:

The Great Statin Con

Yesterday, I was asked to appear on various programmes to discuss a study ‘proving’ that statins cause no side-effects at all. Or, at most, they may cause muscle pains in around one in ten thousand people, no more. At the same time, statins save thousands of lives a year. Therefore, everyone should take them, and patients should ignore scaremongering doctors – such as me I suppose – who state that side-effects are common, and potentially serious.

On the radio, Professor Peter Sever, the lead author of the study, suggested that the leaflets warning of side effects should be removed, because once a patient reads that there may be side effects, they will be far more likely to suffer from them, and report them. The so-called ‘nocebo’ effect. The opposite of the placebo effect, whereby people taking medicines think they will get better, or that their pain will be reduced.

There is no doubt that the nocebo effect is real, although the placebo effect is also real, so do these two effects not just cancel each other out? This is a difficult area of medicine, disentangling what is real, from what is imagined.

However, I watched my father in law become unable to walk, whilst taking statins. We were pushing him around in a wheelchair until, eventually, he agreed to stop his statins. At which point he became able to walk a good distance again, and even climb stairs again. A ‘nocebo’ effect? All in the mind? No, of course not.

I had a patient with such severe abdominal pains that she was going to undergo an investigative laparotomy to establish what was causing them. No investigations had revealed anything. I suggested she stop the statins and the pains were completely gone in two days. All in the mind? I have spoken to many other GPs who have reported seeing side effects in many patients.

I suppose if you are trying to push statins as hard as possible, and you built your academic reputation on running trials on statins, you will naturally want to push them as hard as possible. Some ‘experts’ have even suggested putting statins in the water supply.

But this latest report pushes things to a completely ridiculous point. Can I, as a GP, simply tell patients reporting side-effect that ‘you do not have a side effect, they do not exist, it is simply in your mind.’ No, this would be completely ridiculous, and a total denial of your job, which is to listen to what patients tell you. Not to take a horribly, I know best, paternalistic position.

On the other hand, the benefits of statins have been hyped to an almost completely ridiculous degree. We are told that they reduce the risk of having a heart attack by 30%, which sounds highly impressive, if you, like almost everyone, including me, do not understand statistics.

The reality is, that unless you have had a previous heart attack, statins have no effect on overall mortality. To put that another way, they don’t save lives. They don’t even prevent heart attacks or strokes in women with no previous history of heart disease.

The statistic you really want to know about statins is the following. If you have had a heart attack, or stroke, and take a statin for five years, you will increase your life expectancy by 4.2 days. Balance that against a twenty per cent chance of having side effects, some of which are very unpleasant and long-lasting, and you can see why I am not a fan of statins.
Ends.

Currently I am sifting through the original ASCOT paper to find out exactly what they did study, and what they found, and suchlike. The problem with trying to get to grips with research like this is that there are figures, and more figures, and data and exclusion criteria, and things that are not fully explained. So, it is difficult to make any statement about this entire saga, without many hours of detailed research.

However, I can certainly comment on the key finding from the recent Lancet ‘nocebo’ paper. Key or not, it is the finding that they made the most noise about.

‘During the non-blinded non-randomised phase, muscle-related AEs (adverse events) were reported at a significantly higher rate by participants taking statins than by those who were not (161 [1·26% per annum] vs 124 [1·00% per annum]; 1·41 [1·10–1·79]; p=0·006).’

To translate 161 people (out of more than six thousand) complained of muscle pain whilst taking the statin, and 124 people taking a placebo complained of muscle pain. In total 37 more people complained of muscle pain on the statin. This is not, what I would call, a lot. It was an absolute increase, in the risk of reporting adverse effects, of 0.26%.

Compare and contrast this figure with the findings of the ‘Statin USAGE’ study. As far as I know, this was the largest study to look at why people take, then stop taking, statins:

‘The USAGE survey – “Understanding Statin use in Ama and Gaps in Education” – is the largest known cholesterol survey conducted in the U.S., involving more than 10,100 statin users. The USAGE survey explores patient perceptions, attitudes, behaviors and concerns about statins, the most commonly prescribed medications to treat high cholesterol.’ http://www.statinusage.com/Pages/about-survey-respondents.aspx

A number of things were found. The most important of which, is just how many people stopped taking their statins after one year. A pretty staggering 75%. Why did they stop?

‘More than six in ten respondents (62%) said they discontinued their statin due to side effects, with the secondary factor (17%) being medication cost. Only 12% of respondents cited lack of efficacy in cholesterol management as a reason for stopping their medication. On average, respondents who experienced side effects due to their statin stopped after trying two different statins.

Three out of ten respondents experienced side effects of muscle pain and/or weakness, and 34% stopped taking their statin because of these side effects without consulting with their doctor.’

So, on one hand, what the Lancet study found was that 0.26% extra patients reported muscle pain – when they knew they were on a statin. On the other hand, the Statin USAGE survey found that 30% of people experience muscle pain and/or weakness when on a statin. Now, try to get those two figures to match up.

You could argue that the nocebo effect can only account for 0.26% of adverse effects. Therefore, the other 29.74% (30% in the Statin USAGE study – 0.26% nocebo effects) represents the true rate of adverse effects. You could argue that randomised controlled clinical trials do not reflect the experience of taking medication in the real-world environment. You could say that you can believe one of these studies, but not both.

On the other hand, you could move sideways a bit, and wonder why researchers suddenly decided to ‘data dredge’ a twenty-year-old study – not set up to look at adverse effects as a primary end-point – to prove that statins do not have any adverse effects. You could then look at who funded that research and you could ask yourself why would a company currently being sued in the US for not highlighting the adverse effects of statins, decided to use a study to prove that statins do not have adverse effects.

Alternatively, you could ask people who have taken statins, whether they suffered adverse effects, and try to match the number who claim that they do, with the one in ten thousand figure of Professor Peter Sever. And good luck with that. It is hard, I find, not to think that ‘he who pays the piper calls the tune.’