select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
A meta analysis published in The Lancet looks at side effects of statins listed on product labels.
Prof Ian Douglas, Professor of Pharmacoepidemiology, London School of Hygiene & Tropical Medicine (LSHTM), said:
“Statins are a very effective class of medication which we know have prevented enormous numbers of potentially devastating cardiovascular problems over the last 40 years. They have also been subject to flawed claims of both extraordinary, unexpected benefits as well as harms over the years, highlighting the difficulty of reliably studying medication effects. As with all effective medications, statins can cause harms, which this study has set out to better understand.
“This very well conducted study uses data from 19 randomised trials to systematically evaluate the evidence for each adverse reaction currently included in statin prescribing information. They found that the majority of listed reactions didn’t occur more frequently amongst statin users compared with placebo. In many ways this is reassuring, and indeed, I suspect some of the adverse effects listed for statins are not actually caused by statins. Potential adverse effects listed in patient information leaflets are not only derived from randomised trials and non-randomised observational studies. They are also informed by individual case reports of suspected adverse reactions. Despite the inherent limitations of such case reports, they can also bring uniquely useful insights around details such as what happens if a patient is rechallenged with a drug. In many instances it is very difficult to be certain about a causal association, especially when more formal studies fail to find evidence of such a link. It is also almost never stated in the prescribing information where the evidence for adverse reactions comes from which can make them difficult to interpret.”
“Some of the outcomes investigated in this study are stated to be very rare in the product information. This means our best guess is that they may occur less frequently than one per 10,000 users, and in general the evidence for these effects comes from case reports. Even in the dataset for this study with ~80,000 statin users, we may simply fail to see enough cases to be able to be able to confidently conclude there is no causal association, which the authors also acknowledge. The authors chose to make an intention-to-treat comparison in their analysis. This is a valid approach, but it may underestimate on-treatment effects as people in the trial could discontinue their treatment but still be counted in the treated group during time when they couldn’t have been at risk of an adverse drug reaction.
“On the whole, I agree with the conclusion that the potential side effects of statins should be reviewed in detail. Once a potential adverse reaction is added to a label, it is incredibly rare for it to be removed, even if good contradictory evidence emerges and I think this is a failing of our current systems. It doesn’t help patients or clinicians make the best decisions. But I would be less certain that we now have conclusive evidence that all the adverse effects studied here are not caused by statins.”
Comments from our friends at the Spanish SMC:
Carlos Guijarro Herráiz, former president of the Spanish Society of Arteriosclerosis (SEA) and doctor at the Internal Medicine Unit of the Alcorcón University Hospital Foundation, said:
“A meta-analysis that includes individual patient data is an extraordinary collaboration that provides a higher quality of information than most meta-analyses, which use aggregated data from individual studies. In this regard, we are grateful for the collaboration of the Cholesterol Treatment Trialists’ Collaboration, which has previously provided us with detailed information on the protective effects of statins in cardiovascular prevention and on some side effects (myopathy, diabetes mellitus), whose importance is far outweighed by the unquestionable benefit of lipid-lowering treatment in patients with high cardiovascular risk.
“Several placebo-controlled studies report that most of the alterations attributed to statins are based on the “nocebo effect”, which is the mirror image of the placebo (subjective harm/benefit to the patient independent of the actual pharmacological effect of the treatment). In other words, the apparent discomfort or side effects of statins have no real pathophysiological or pharmacological basis. Paradoxically, the detailed description of possible side effects in the information contained on statin packaging (technical data sheet) potentially reinforces the “nocebo” effect. Who does not experience some discomfort over the years, whether or not they are taking medication?
“Once it has been confirmed that there is no objective basis in clinical trials for many of the disorders attributed to statins, a note of caution should be added. Clinical trials are not real life; the use of statins is widespread, including in patients who are usually excluded from pivotal clinical trials: the elderly, ethnic minorities, patients with concomitant diseases and polypharmacy. Therefore, individual assessment of each patient with their doctor is essential.
“Fortunately, we can reassure patients by pointing out that the primary (protective) and secondary (harmful) effects of statins have been extensively evaluated, and that hundreds of thousands of people currently take them and have taken them for decades with unquestionable cardiovascular protective effects.
“If, after explaining all this, the patient considers that any of their discomfort is attributable to statins, we can agree with the patient to briefly suspend treatment (in most cases disproving the relationship with the drug) and, if the patient continues to consider that there are problems with the treatment, fortunately we have lipid-lowering treatments used specifically in statin-intolerant patients or with completely different mechanisms of action (ezetimibe, iPCK9), which have demonstrated cardiovascular benefits and can provide effective and safe alternatives for reducing cardiovascular complications by controlling hypercholesterolaemia.
“Unfavourable press reports about statins are not irrelevant, as they are associated with a decline in their use and an increase in cardiovascular disease, so providing reassuring news has significant and undeniable public health value.”
José Luis López Sendón, Cardiologist at La Paz Hospital and researcher at IdiPAZ (Madrid), said:
“The article is of unquestionable scientific quality. Randomised, double-blind clinical trials were selected, comparing a statin with a placebo or a statin at different doses, with more than 1,000 cases in each study and a follow-up of more than two years. This is not a conventional meta-analysis; the authors used individual data from the patients included in each study. In total, individual data from 123,924 cases included in 19 studies with an average follow-up of 4.5 years were analysed. The analysis methodology was correct.
“The main objective of the study was to analyse potential side effects, especially those included in the product information lists. The hypothesis was that the lists of side effect warnings indicated in the “package insert” documentation are not based on scientific evidence, but rather on isolated observations or those described in general follow-up observational studies.
“The evidence provided is compelling. Only four of the 66 side effects listed by the various regulatory agencies showed a statistically higher incidence in the statin group compared to the placebo group:
Muscle discomfort.
New diagnosis of diabetes.
Abnormal urine tests (0.03% excess per year compared to placebo (not significant in the text of the results, but it does appear in Figure 1).
Abnormal liver function tests (0.13% excess per year compared to placebo).
“The figures are excellent in terms of information, but they can be difficult to understand if you do not know anything about statistics. In any case:
Side effects are more common with high doses of statins.
The incidence of excess side effects compared to placebo is very low.
“The study does not seek to compare with previous evidence, but rather with the “popular” perception of safety (among many doctors, patients, and the general population). And, above all, with the warnings from regulatory agencies.
“All studies have limitations. In this study, they are the usual ones found in clinical trials and are not significant enough to cast doubt on the results and conclusions. In this case:
All statins are mixed together, and the side effects may be different.
“For protection, people included in clinical trials are selected using inclusion criteria that exclude potential contraindications and comorbidities (other associated diseases), i.e., it is a population with a lower risk of complications, and it is not known what happened to patients who did not qualify for inclusion in the studies. Those who are excluded fare worse and may have more complications.
“In all clinical trials, patients receive more medication than the study drug or placebo. They receive multiple medications for their diseases. Some of the medications may have similar side effects or additional benefits. To exclude the influence of other concomitant diseases and the potential effect of other medications, special statistical analyses are necessary to determine more accurately whether the effect of the study drug is independent of other factors (other diseases and medications). With such a high number of cases, it is unlikely that this type of analysis would yield different results, and it is not usually performed in the first evaluation.
“The analysis was done on an intention-to-treat basis, but there are patients on statins who stop taking the medication and patients assigned to placebo who start treatment with statins.
“The patients included in the clinical trials analysed were selected based on safety criteria. Regulatory agencies consider all available information, including study data, safety observations, publications of isolated cases with significant complications, etc. Therefore, it is understandable that regulatory agencies are more restrictive and go beyond the safety conclusions demonstrated in clinical trials. In any case, agencies must change the information provided to patients (in package inserts) or they will be misleading consumers and healthcare professionals.
“The results of this work should be considered a very important contribution to the safety of statins, which patients, doctors and regulatory agencies should be aware of and consider. But this study does not end with this publication. The database built with data from all the studies is immense and will provide more information: benefits, subgroups, etc.
“My congratulations to the group of researchers who carried out the work, starting with the most difficult part: agreeing to pool the individual patient data from each study and obtaining authorisation to do so from the owners of the databases for each study, generally the commercial companies sponsoring the clinical trials. Bravo!”
—-
Prof Stephen Evans, Emeritus Professor of Pharmacoepidemiology, The London School of Hygiene and Tropical Medicine, said:
Background:
“To obtain approval by regulatory authorities, pharmaceutical companies wishing to introduce new treatments to treat diseases, usually have to conduct randomised trials in patients with the disease, comparing patients getting the new treatment with patients getting a dummy pill (placebo). Both groups get the current standard of care. The typical number of patients studied in such randomised trials is less than 1500 patients. The consequence is that adverse effects that occur relatively rarely will not be detected in those trials. [n.b. an “adverse event” is one which may or may not be caused by the medicine, while an “adverse effect” is caused by the medicine.]
“The regulatory authorities monitor reports of individual patients with adverse events and new studies, usually not randomised. Experience is gained from the usage of the newly introduced medicine in many thousands or millions of patients. The authorities and the companies will notice many new adverse events that occur in patients. They attempt to ensure that when new adverse events are added as adverse effects to the official information about the medicine, there is some good evidence that the adverse event is caused by the medicine, rather than occurring coincidentally. This is not a simple process since many adverse events occur because of the disease being treated and may also occur in patients who are not being treated. A precautionary approach tends to be taken, and this is true especially for treatments intended to prevent disease rather than treat it. This results in many adverse events being added for which the evidence for their inclusion is not strong.
“Once an adverse event has been included in the official information, it is more difficult to remove it when better information arises that shows the adverse event is NOT caused by the medicine.”
This paper:
“This paper analyses the many large randomised trials of statins used for preventing cardiovascular disease. Most treatments do not have such large amounts of data based on randomised studies where whether an effect is causal or not can be assessed reliably. The official information has had many adverse events added to the list of adverse effects, possibly when they were not caused by the statin. With the enormous numbers included in randomised trials of statins, adverse events that occur rarely may be shown to occur at similar rates in those on the new treatment and in those on placebo. Such events are strong candidates for being removed from the official information as being caused by statins.
“Patients can then be reassured that a number of adverse events, that some believe are caused by statins, are very much more likely to be coincidental. Patients can obtain the benefits, for which there is convincing evidence, without being concerned about these adverse events.
“Most treatments do not have such large randomised trials where unusual causal effects can be studied. It is the large size of the randomised trials of statins that allow for this.”
Dr Emily Herrett, Associate Professor at the London School of Hygiene & Tropical Medicine, said:
“Statins are a key pillar in prevention of heart attacks and strokes, and this study gives us further important evidence about their safety.
“This press release accurately reflects the science. The findings come from a very large and carefully conducted review of randomised trials, which is the most reliable way we have to assess drug side effects. The study design is particularly strong because it compares people taking statins with people taking placebo tablets, meaning any differences in symptoms can be attributed to the drug rather than expectations or background health problems.
“The conclusions are well supported by the data. Across tens of thousands of participants followed for several years, rates of most symptoms were the same whether people were taking statins or not. This strongly suggests that statins are not the cause of most symptoms listed in package leaflets. These findings fit well with previous work from this group and others, which has already shown that most muscle symptoms attributed to statins are not actually caused by the drug. This study adds further high-quality evidence across a much wider range of symptoms.
“As with all medicines, statins can have side effects, and the study clearly identifies the small number that are genuinely associated with treatment, such as mild changes in liver blood tests. Importantly, serious harms remain very rare.
“This evidence should reassure both patients and clinicians. Many people stop statins because of perceived side effects or fear of side effects, but this study shows that symptoms occurring during statin use would probably have occurred even if the patient was not taking statins.
“Anyone who develops new symptoms while taking statins should talk to their GP, but it’s important to know that the chances of these symptoms being caused by the statin itself are very low.”
Prof Riyaz Patel, Professor of Cardiology & Consultant Cardiologist, University College London (UCL), said:
“This is an important study, from a highly regarded academic group, looking to carefully examine whether there is any basis to many of the possible adverse effects listed on standard medication inserts for statins.
“By going back to the original data from 19 trials with over 120K participants, the authors evaluated the frequency of all the possible adverse events reported – these usually have to be meticulously documented during any clinic trial.
“The authors found that for most of the potential side effects listed in medicines inserts, there was no evidence of a link to statins, as they were just as common in the placebo group as the statin group.
“Beyond muscle effects and diabetes risk which the same group previously documented, they did find an effect on increasing liver enzymes modestly, which is well known (hence the NICE recommendation to check liver enzymes after 3 months) and rarely if ever leads to any harm. Another effect they found was on urinary protein levels in some people, but this does not change with higher doses suggesting it is an uncertain effect.
“The study is significant as it helps both doctors and patients feel more confident that there are few true side effects from statins. The ones they did uncover fit with what we already know after 30+ years of statin use so this adds further validity to and confidence in the results.
“There are some limitations which the authors acknowledge, such as the duration of follow up being relatively short during the trials, which limits understanding of potential concerns that could take many years to develop. However, when interpreted in the context of all other studies on statins, and 30+ years of clinical experience, these are unlikely to be common or significant.
“Overall, this is a really important paper. One which should be made available to anyone who takes or prescribes a statin as it will hopefully allay many concerns.”
Prof Kausik Ray, Professor of Public Heath, Imperial College London, said:
“This is the most robust blinded trial data that to date. Trials are the true test of causality and free from confounding that plagues observational data and anecdotes.
“Warning labels often come from post marketing observations with no control arm. From a legal perspective, it makes sense but from a biological perspective it doesn’t. The true test comes from blinded trials testing whether a treatment causes more harm than an inert comparator (placebo), when tested in double blind trials.
“These data are very important to provide reassurance and need to be widely disseminated as they are clinically relevant and actionable as they provide a reliable source of information for patients and help us with shared decision making. Statins save lives, but they are often underutilised, and with patient access to unrealisable opinions or non-randomised data these, data are timely and critical.”
‘Assessment of adverse effects attributed to statin therapy in product labels: a meta-analysis of double-blind randomised controlled trials’ by Reith et al. was published in The Lancet at 23:30 UK time on Thursday 5th February.
Declared interests
Carlos Guijarro Herráiz: No conflicts of interest.
José Luis López Sendón: No conflicts of interest.
Prof Ian Douglas: ID is a professor of pharmacoepidemiology at the London School of Hygiene & Tropical Medicine. He has shares in and research grants from GSK. He has also studied statins in large non-randomised studies and also found them to be associated with very few adverse effects.
Prof Stephen Evans:
Dr Emily Herrett: No conflicts of interest to declare.
Prof Riyaz Patel: “In 2022/2023 advised NICE on the latest NICE guidance on lipid management. I run the Lipids and CVD risk service at Barts Health NHS Trust. I have a private practice at One Welbeck where I treat patients with cholesterol concerns. I have in the last two years advised drug companies on study design and research projects for experimental drugs and received honoraria for my consulting services.”
Prof Kausik Ray: “Currently none related to statins, but previously consulted for Pfizer, AstraZeneca and Kowa. I am involved with the development on novel lipid modification therapies currently (not statins).”