select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to study looking at statin intolerance

A study published the European Heart Journal looks at the prevalence of statin intolerance.


Prof Sir Nilesh Samani, Medical Director at the British Heart Foundation, said:

“Decades of evidence have proven that statins save lives. This latest analysis, showing that the risk of side effects from statins are less than previously thought, should provide reassurance to those who are recommended this medicine to reduce their risk of a heart attack or stroke. Furthermore, other studies have shown that side effects commonly attributed to statins are often not from the drug itself. 

“Having said this, the decision to take a statin to reduce the risk of heart and circulatory disease should always be a joint one between a GP and patient after a discussion about the likely benefit.”


Prof Peter Sever, Professor of Clinical Pharmacology and Therapeutics, Imperial College London, said:

“This publication on statin related side effects adds little to our existing knowledge, the pooled analysis from a large number of studies highlights the fact that statin associated side effects occur far less frequently than commonly reported. However, by merging data from blinded randomised trials with observational cohort data derived from individuals who know they are taking a statin, the true incidence of adverse effects caused by a statin has not been clearly separated from the much larger incidence of side effects associated with statin use which is caused by misconceptions of patients about the side effects of statins and the nocebo effect. The latter is an adverse response to a drug which is not due its pharmacological effect, but is a response resulting from the expectation of some untoward effect occurring that is influenced by prior information obtained through media or other sources. The frequency of side effects of statins published in this report was 9% overall – a figure derived from randomised trials (5%) and observational studies (17%). Clinical practice often quotes a figure of around 20 % for side effects of statins and previous comprehensive analyses of randomised trial around 1%.

“The importance for clinicians and patients is to realise that commonly reported symptoms, such as muscle aches and pains and lethargy, are not due to the chemistry of the drug. These ‘nocebo’ symptoms may be psychological in origin but they are no less real than pharmacological symptoms in how they affect quality of life. However, it’s important to note that as they are not directly caused by the drug, they should not override the decision to prescribe and take statins on account of their proven benefit in reducing death and disability from heart attacks, strokes and other cardiovascular conditions.”


Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“In statistical terms, I think this study is competent, and the researchers have put together data from an impressive numbers of previous studies involving over 4 million patients in all. Given that many previous studies have demonstrated that taking statins in appropriate cases can reduce illness and deaths from cardiovascular disease (heart disease, strokes, etc.), there is clearly an issue if patients stop taking statins after they are prescribed. Many patients do discontinue statin treatment, very often because of side effects (actual or perceived) of the statin treatment, so it’s important to know how common side effects are, and previous studies have varied a lot in their estimates of the prevalence of adverse effects.

“Though the overall study is competent and pretty comprehensive, there is one thing in the press release that, I think, could easily be misinterpreted (though technically it’s not incorrect). I’ll describe that towards the end of my comment, but first I need to give some more general context.

“These researchers have produced estimates of the overall frequency of intolerance to statins, by combining the findings of 176 different previous studies, using standard statistical methods to carry out the analysis (which is called a meta-analysis). The studies that they combined were of two kinds. Roughly two thirds were randomised clinical trials (RCTs) where patients received treatment either with statins or with some other pills, which would often have been a placebo (a pill that looks exactly like the statin pill but contains no active ingredients) or some other drug that the researchers running the RCT wanted to compare with the statin. In RCTs, patients are allocated the treatment (statin or the comparator drug or placebo) at random. The other third of studies were observational cohort studies, where the researchers record whether the participants take statins or not, but do not choose whether the participants take statins or allocate the treatments at random. In both types of study, the patients are followed up for a period of time, and the researchers record (among many other things, typically) whether they experience adverse effects that are described as statin intolerance.

“Both types of study have advantages and disadvantages. But an issue with observational cohort studies is that the participants will know whether or not they are taking a statin. If, for whatever reason, they have a conscious or even unconscious perception that statins may have harmful or painful side effects, that might make them perceive such effects and attribute them to the statins even if they are really caused by something else, and not by the effects of the statins in their bodies. “Such effects, not caused by the drug itself but by the act of taking it, are called nocebo or drucebo effects. (‘Drucebo’ is a relatively new term, from ‘drug’ and ‘-cebo’ as in placebo or nocebo.) They can arise in some RCTs too, because in some RCTs the patients do actually know whether or not they are taking the statin under study, whereas in others they do not (because the non-statin pills look identical to the statin pills).

“Another issue with observational cohort studies is that the patients who do or do not take the statins will differ in many ways other than whether they take the statins, and any difference in reporting possible side effects between the statin-takers and the non-takers could be due to these other differences rather than the statins themselves. Statistical adjustments can be made to allow for this to some extent, and they were generally made in the studies that fed into this new research, but one can never tell whether everything appropriate has been adjusted for. This issue would not generally arise with RCTs, because treatments are allocated at random.

“The researchers in the new study found high levels of heterogeneity in the studies that they combined. That is, the estimates of the prevalence of statin intolerance between the studies differed a lot from one study to another, much more than would be expected just from statistical variability. That could well be due, as the researchers say, to differences in the importance of nocebo/drucebo effects between different studies, or perhaps because different cohort studies have made different adjustments for other factors. But it could also be due to different studies involving groups of patients with different characteristics, perhaps because they were carried out in different populations with different demographic characteristics or different patterns of previous illness. That does make it inappropriate to say that the overall prevalence of statins intolerance would be at the level estimated in this analysis, whatever type of patients were involved – but the researchers don’t claim that. They report an overall average figure – an estimated prevalence of 9.1% on their broadest measure of intolerance. But they also investigated which patient characteristics might mean that their risk of having statin intolerance was different from the average.

“The groups where they did find evidence of a different risk level are listed in the press release. For example the press releases says that the increased risk of statin intolerance in women was 48%. Really my main reason for commenting on this study is that that 48% figure could be very misleading. It’s not wrong – it’s just that it needs context, I think. It does NOT, for example, mean that the risk of statin intolerance in women was 48%, in the same way that they found the overall risk to be about 9%. It also does not mean that 48% of the people who had statin intolerance were women.

“What it does mean is as follows. Overall, the incidence of statin intolerance was 9.1%. This means that, if you had a group of 100 people starting on statins, who on average were broadly like the 4 million included in this research, about 9 of them would turn out to be intolerant of the statin they were taking. That’s an absolute risk measure. But the 48% figure for women is a relative measure – it means that the risk of intolerance in women is 48% higher than the risk of intolerance in men, but that doesn’t help much unless you know the risk of intolerance in men. It can be worked out from various information in the research report that, if you had a group of 100 men starting on statins, who on average were broadly like the men in this research, then about 7 or 8 of them would have statin intolerance. If you had a group of 100 women who were broadly like the men in other characteristics, roughly about 11 of them would have statin intolerance. So the prevalence of statin intolerance for women is higher than for men, but it isn’t as much higher as the 48% relative risk (actually relative odds) might lead you to believe.”


Prof Riyaz Patel, Professor of Cardiology, UCL & Consultant Cardiologist, Barts Health NHS Trust, said:

“This is a substantial analysis of a large number of studies including over 4M patients from around the world that looked at reported side effects from statins. The authors conclude that the prevalence of statin intolerance is lower than usually reported.

“The study summarises the effects of multiple studies to give an estimate of what the overall effect may be. This approach is usually quite reliable but where studies are very different in quality or methodology, results may need to be interpreted with a little caution. For example, to gather a large sample size the authors combined two different types of studies called trials and cohorts.  The former is usually much more stringent in who is included in the studies and what data is captured, the latter less so and usually based on observing what happens to people when they take statins without intervening or scrutinising symptoms.  In this analysis only 200K or so patients were from trials and 4M from cohorts. This doesn’t make the analysis unreliable but needs to be considered especially as the overall effect suggested that statin side effects were much lower in trials and higher in cohorts indicating differences in the two settings. It is possible more people reported side effects in the cohorts which may not truly be statin side effects and people who were less likely to report side effects and underwent enrolment screening procedures were participating in trials.

“To add context, based on what we know about statin side effects to date, the results are likely to be broadly valid, and indicate that we should not overestimate statin side effects or be too quick to stop statins without due consideration. Clinically, it should be stated that some patients do experience real side effects and we do our best to help them with alternative therapies as with any other medicine. However for the vast majority of people experiencing statin side effects, we can usually work with the patient to understand the symptoms, use proven strategies to manage these and ensure they do not miss out on the well-established benefits of statins. This is especially important for people who have already had a heart attack or stroke where statin therapy is really important in preventing further events.”



“Prevalence of stain intolerance”, by Ibadete Bytyci et al. was published in the European Heart Journal at 00:05 UK time on Wednesday 16th February.




Declared interests

Prof Peter Sever: “Previous research grants and consultancy from Pfizer and Amgen”

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  My quote above is in my capacity as an independent professional statistician.”

Prof Riyaz Patel: “Have in the past received honoraria and consulting fees from drug companies manufacturing new cholesterol lowering drugs. Currently work with NICE as a topic advisor for CVD prevention.”

None others received.

in this section

filter RoundUps by year

search by tag