icon-folder.gif   Conference Reports for NATAP  
 
 
 
 
Interpretation of the evidence for the efficacy and safety of statin therapy
 
 
  Download the PDF here
 
Download the PDF here
 
Summary
 
This Review is intended to help clinicians, patients, and the public make informed decisions about statin therapy for the prevention of heart attacks and strokes. It explains how the evidence that is available from randomised controlled trials yields reliable information about both the efficacy and safety of statin therapy. In addition, it discusses how claims that statins commonly cause adverse effects reflect a failure to recognise the limitations of other sources of evidence about the effects of treatment. Large-scale evidence from randomised trials shows that statin therapy reduces the risk of major vascular events (ie, coronary deaths or myocardial infarctions, strokes, and coronary revascularisation procedures) by about one-quarter for each mmol/L reduction in LDL cholesterol during each year (after the first) that it continues to be taken. The absolute benefits of statin therapy depend on an individual's absolute risk of occlusive vascular events and the absolute reduction in LDL cholesterol that is achieved. For example, lowering LDL cholesterol by 2 mmol/L (77 mg/dL) with an effective low-cost statin regimen (eg, atorvastatin 40 mg daily, costing about 2 per month) for 5 years in 10 000 patients would typically prevent major vascular events from occurring in about 1000 patients (ie, 10% absolute benefit) with pre-existing occlusive vascular disease (secondary prevention) and in 500 patients (ie, 5% absolute benefit) who are at increased risk but have not yet had a vascular event (primary prevention). Statin therapy has been shown to reduce vascular disease risk during each year it continues to be taken, so larger absolute benefits would accrue with more prolonged therapy, and these benefits persist long term. The only serious adverse events that have been shown to be caused by long-term statin therapy-ie, adverse effects of the statin-are myopathy (defined as muscle pain or weakness combined with large increases in blood concentrations of creatine kinase), new-onset diabetes mellitus, and, probably, haemorrhagic stroke. Typically, treatment of 10 000 patients for 5 years with an effective regimen (eg, atorvastatin 40 mg daily) would cause about 5 cases of myopathy (one of which might progress, if the statin therapy is not stopped, to the more severe condition of rhabdomyolysis), 50-100 new cases of diabetes, and 5-10 haemorrhagic strokes. However, any adverse impact of these side-effects on major vascular events has already been taken into account in the estimates of the absolute benefits. Statin therapy may cause symptomatic adverse events (eg, muscle pain or weakness) in up to about 50-100 patients (ie, 0⋅5-1⋅0% absolute harm) per 10 000 treated for 5 years. However, placebo-controlled randomised trials have shown definitively that almost all of the symptomatic adverse events that are attributed to statin therapy in routine practice are not actually caused by it (ie, they represent misattribution). The large-scale evidence available from randomised trials also indicates that it is unlikely that large absolute excesses in other serious adverse events still await discovery. Consequently, any further findings that emerge about the effects of statin therapy would not be expected to alter materially the balance of benefits and harms. It is, therefore, of concern that exaggerated claims about side-effect rates with statin therapy may be responsible for its under-use among individuals at increased risk of cardiovascular events. For, whereas the rare cases of myopathy and any muscle-related symptoms that are attributed to statin therapy generally resolve rapidly when treatment is stopped, the heart attacks or strokes that may occur if statin therapy is stopped unnecessarily can be devastating.
 
Consequently, the proportional reduction in the risk of major vascular events per mmol/L was about one-quarter in the trials of statin versus no statin (after an initial delay) and of more versus less intensive therapy. Based on the combined findings from these two sets of trials, it can be estimated that reducing LDL cholesterol concentrations by 2 mmol/L would reduce the risk of major vascular events by about 45% (derived as [1⋅0-(0⋅75 x 0⋅75)] x 100) during each year treatment is continued. In principle, even larger reductions in LDL cholesterol would be expected to produce even larger risk reductions (eg, 60-70% with 3-4 mmol/L LDL cholesterol reductions); however, this is likely only to be clinically relevant in limited circumstances (eg, for individuals with familial hypercholesterolaemia who have very high LDL cholesterol concentration).
 
Kidney-related outcomes
 
In light of the increased incidence of diabetes with statin therapy, it is appropriate to consider whether there are any excesses of microvascular complications related to the kidney. In a meta-analysis274 of 57 randomised controlled trials involving a total of about 140 000 patients treated for at least 6 months, statin therapy slowed the rate of decline of the estimated glomerular filtration rate (eGFR) by 0⋅41 mL/min per 1⋅73 m2 per year (95% CI 0⋅11-0⋅70). In addition, compared with control, statin therapy produced a smaller standardised mean difference in change in albuminuria or proteinuria of 0⋅65 standard deviations (95% CI 0⋅37-0⋅94) among about 5000 patients in 29 trials that had reported such data. Despite these beneficial effects, statin therapy did not appear to have an effect on progression to end-stage renal disease in randomised trials: 1261 (13⋅5%) cases on statin versus 1282 (13⋅6%) cases on control (odds ratio 0⋅98, 95% CI 0⋅90-1⋅07).
 
Consequently, as with differences in the rates of other outcomes that have been associated with statin use in observational studies, the evidence from randomised controlled trials does not provide support for an adverse effect of statin therapy on the kidney (except perhaps in the perioperative setting) and, instead, indicates that it may slow the progression of renal impairment (although the clinical significance of the small effect that has been observed is uncertain). If, however, statin therapy is not stopped when statin-related myopathy occurs, this may lead to renal failure, so doctors and patients should be alert to the possibility of this rare complication (while also being careful not to attribute muscle symptoms to statin therapy without confirmatory evidence and so not stop the statin unnecessarily).
 
Based on the combined findings from these two sets of trials, it can be estimated that reducing LDL cholesterol concentrations by 2 mmol/L [45 mg] would reduce the risk of major vascular events by about 45% (derived as [1⋅0-(0⋅75 x 0⋅75)] x 100) during each year treatment is continued. In principle, even larger reductions in LDL cholesterol would be expected to produce even larger risk reductions (eg, 60-70% with 3-4 mmol/L LDL cholesterol reductions); however, this is likely only to be clinically relevant in limited circumstances (eg, for individuals with familial hypercholesterolaemia who have very high LDL cholesterol concentration).
 
In summary, given the 0⋅3% absolute excess of muscle problems based on more than 10 000 reported cases in meta-analyses of randomised trials during 3-5 years of treatment (appendix),250 the excess rate of symptomatic muscle pain and other muscle-related problems due to statin therapy would appear to be no more than about 10-20 cases yearly per 10 000 treated individuals, with only about one of those cases associated with substantial elevations in creatine kinase concentrations (ie, myopathy) and requiring statin therapy to be stopped.
 
Meta-analysis can also reduce the impact of selective emphasis on effects observed in particular trials that may overestimate the real effects3, 20, 46 (eg, the excess of diabetes cases with statin therapy first noticed in the JUPITER trial48 was found to be smaller in other statin trials49) or may not even be real (eg, the small excesses of incident cancer cases in the CARE trial50 and the PROSPER trial51 were not confirmed by the much larger numbers of cases in the other statin trials52, 53).
 
---------------------
 
The large volume of data available was beneficial for this meta-analysis. To our knowledge, the current study represents the largest systematic review of statin administration on kidney disease progression and the first meta-analysis that evaluates the effect of statin treatment on kidney failure events. Several prior overviews have evaluated effects of statins on kidney disease outcomes in patients with cardiovascular risk or CKD. Meta-analyses that included 26 trials with 39,704 participants reported significant benefits of statins, with differences in eGFR decline of 1.22 (95% CI, 0.44-2.00) mL/min per year.17 A recent review that included 41 studies with a total of 88,523 patients found that statin therapy reduced the slope of eGFR decline.96 However, both studies focused on effects of proteinuria reduction or eGFR decline and not on clinically relevant renal benefits, such as kidney failure events. Additionally, these studies cannot exclude the possibility that statin administration increases creatinine excretion and influences serum creatinine level. The findings of these studies therefore have limited application in clinical practice. Our study, which examined nearly double the number of participants in prior reviews, found a nonsignificant beneficial effect in composite kidney failure events and ESRD. This finding aligns with SHARP, a large trial.14
 
Pdf attached above
 
-------------------------
 
Known adverse effects of statin therapy>
 
⋅The only adverse events that have been reliably shown to be caused by statin therapy are myopathy (defined as muscle pain or weakness combined with large increases in creatine kinase blood concentrations) and new-onset diabetes mellitus, along with a probable increase in strokes due to bleeding (ie, haemorrhagic strokes).
 
⋅Typically, treatment of 10 000 patients for 5 years with a standard statin regimen (such as atorvastatin 40 mg daily) would be expected to cause about 5 cases of myopathy, 50-100 new cases of diabetes, and 5-10 haemorrhagic strokes.
 
⋅Despite reports based largely on non-randomised observational studies, there is good evidence that statin therapy does not cause adverse effects on other health outcomes (chiefly muscle pain and weakness) that have been claimed prevent a large proportion of patients from continuing it long term (so-called "statin intolerance").
 
⋅Large-scale evidence from randomised trials rules out excesses of muscle pain and weakness with statin therapy of more than about 10-20 cases annually per 10 000 treated patients, with only about one of those cases being associated with large creatine kinase elevations (ie, myopathy) and  requiring statin discontinuation.
 
⋅Absolute excesses of adverse events that are caused by statin therapy are not more than about 100-200 per 10 000 patients (ie, 1-2%) treated for 5 years, and it is unlikely that large adverse effects on serious adverse events await discovery.
 
⋅The harmful effects of statin therapy can usually be reversed without any residual effects by stopping it, whereas the harmful effects of heart attacks or strokes that occur because statin therapy has not been used can be devastating.
 
Potential benefits and harms of lowering LDL cholesterol concentrations
 
Associations between LDL cholesterol and vascular disease

 
By contrast with observational studies of treatment, observational epidemiological studies are valuable for the assessment of causal risk factors. In particular, such studies have shown that there is a continuous positive association between blood concentrations of LDL cholesterol and the rates of coronary heart disease events in different populations, without any suggestion within the range that has been studied of a threshold below which a lower concentration is not associated with a lower risk.156, 157 The absolute difference in coronary disease risk associated with a given absolute difference in LDL cholesterol is greater at higher concentrations (figure 2A), which helps to explain the emphasis in previous treatment guidelines on individuals with hypercholesterolaemia. However, if risk is plotted on a logarithmic scale, then the proportional difference in risk associated with a given absolute difference in LDL cholesterol concentration is similar throughout the range (figure 2B).
 
Observational studies can provide evidence about associations of risk factors with health outcomes, but they do not necessarily suffice to confirm the causal nature of such associations. In the case of LDL cholesterol, several additional sources of evidence have helped to show that the continuous association with atherosclerotic disease is causal. These include experimental studies of atherosclerosis in animals, monogenic and polygenic associations in human beings, and randomised trials of LDL cholesterol-lowering therapy (which also assess the extent of risk reversibility and its timescale).
 
Lower concentrations of cholesterol have been associated in observational studies with higher rates of all-cause mortality, particularly in older people.164, 165, 166 However, such associations can be shown not to be causal. For example, using the Mendelian randomisation approach, lower genetically determined LDL cholesterol concentrations are associated with lower all-cause mortality even among individuals aged older than 90 years.167 It appears that pre-existing disease causes lower cholesterol concentrations (so-called "reverse causality");168 spurious associations can often be reduced in analyses of observational epidemiological studies of risk factors by censoring the first few years of follow-up.
 
Experimental studies in animals have shown that diets that raise LDL cholesterol concentrations increase the extent of atherosclerosis in the arterial wall, and that lowering LDL cholesterol concentrations with either diet or drugs (including statins) can reduce atherosclerosis.169, 170

0104171

0104172

Introduction
 
Used appropriately, modern medical therapies have the potential to prevent a large proportion of the burden of cardiovascular disease. However, their appropriate use relies on the availability of robust data on safety and efficacy, as well as on a sound understanding of the interpretation and application of such evidence.
 
Randomised controlled trials of adequate size are needed to be confident that any moderate benefits and any moderate harms of a treatment have been assessed sufficiently reliably.1, 2, 3, 4 In certain circumstances, available evidence from randomised trials about the effects of a treatment may be limited (perhaps because it is deemed not possible or too difficult to do adequate trials).2 However, the particular context that this Review addresses is the appropriate interpretation of evidence about the safety and efficacy of a treatment when randomised trials of it have been conducted in large numbers of many different types of patient (as is the case for statin therapy), as well as the additional value of information from observational studies based on cohorts, health-care databases, or other sources.3, 4, 5 Not only have the limitations of observational studies4, 6, 7, 8, 9 often been underestimated when attributing adverse effects to treatment (such as misleading claims that statins cause side-effects in one-fifth of patients10, 11, 12), but also the strengths of randomised trials with masked treatment allocation and systematic ascertainment of many different types of adverse event have been under-estimated for the reliable assessment of the safety and efficacy of treatment.3, 9, 13, 14, 15
 
This Review first considers the generic strengths and limitations of randomised trials and observational studies for assessing the effects of treatment, and then considers the specific evidence that is available on the efficacy and safety of statin therapy. It concludes by considering the public health implications of the failure to recognise the full benefits of using statin therapy and of the exaggerated claims that have been made about the rates of side-effects.
 
Potential benefits and harms of lowering LDL cholesterol concentrations
 
Associations between LDL cholesterol and vascular disease

 
By contrast with observational studies of treatment, observational epidemiological studies are valuable for the assessment of causal risk factors. In particular, such studies have shown that there is a continuous positive association between blood concentrations of LDL cholesterol and the rates of coronary heart disease events in different populations, without any suggestion within the range that has been studied of a threshold below which a lower concentration is not associated with a lower risk.156, 157 The absolute difference in coronary disease risk associated with a given absolute difference in LDL cholesterol is greater at higher concentrations (figure 2A), which helps to explain the emphasis in previous treatment guidelines on individuals with hypercholesterolaemia. However, if risk is plotted on a logarithmic scale, then the proportional difference in risk associated with a given absolute difference in LDL cholesterol concentration is similar throughout the range (figure 2B).

0104173

Consequently, with a treatment that acts through lowering LDL cholesterol, the proportional reduction in cardiovascular disease risk per mmol/L reduction in LDL cholesterol should be expected to be similar irrespective of the starting cholesterol concentrations (rather than, as has been suggested for statin therapy,10 being evidence that the effects are not related to cholesterol lowering). Moreover, the absolute reduction in vascular risk per mmol/L reduction in LDL cholesterol would also be expected to be similar for individuals who have similar levels of risk but present with different cholesterol concentrations. The results of randomised controlled trials of statin therapy support these epidemiological expectations (as described later),29, 30, 31, 32, 33, 160, 161 and treatment guidelines now tend to focus on an individual's risk of having atherosclerosis-related events as well as on their LDL cholesterol concentration.162, 163 Lower concentrations of cholesterol have been associated in observational studies with higher rates of all-cause mortality, particularly in older people.164, 165, 166 However, such associations can be shown not to be causal. For example, using the Mendelian randomisation approach, lower genetically determined LDL cholesterol concentrations are associated with lower all-cause mortality even among individuals aged older than 90 years.167 It appears that pre-existing disease causes lower cholesterol concentrations (so-called "reverse causality");168 spurious associations can often be reduced in analyses of observational epidemiological studies of risk factors by censoring the first few years of follow-up.
 
Causal relationship between LDL cholesterol and vascular disease
 
Observational studies can provide evidence about associations of risk factors with health outcomes, but they do not necessarily suffice to confirm the causal nature of such associations. In the case of LDL cholesterol, several additional sources of evidence have helped to show that the continuous association with atherosclerotic disease is causal. These include experimental studies of atherosclerosis in animals, monogenic and polygenic associations in human beings, and randomised trials of LDL cholesterol-lowering therapy (which also assess the extent of risk reversibility and its timescale).
 
Experimental studies in animals have shown that diets that raise LDL cholesterol concentrations increase the extent of atherosclerosis in the arterial wall, and that lowering LDL cholesterol concentrations with either diet or drugs (including statins) can reduce atherosclerosis.169, 170 Genetic disorders in human beings (in particular, LDL receptor mutations) that cause large elevations of LDL cholesterol concentrations are associated with substantially elevated rates of atherosclerotic disease.171, 172 Moreover, these disorders (ie, familial hypercholesterolaemia) provide compelling evidence of dose effects, whereby individuals in European and North American populations who inherit the abnormal genetic variant from both parents typically have LDL cholesterol concentrations greater than 13 mmol/L and coronary events before the age of 20 years,171 whereas those who inherit the abnormal variant from one parent typically have concentrations greater than 8 mmol/L and events in early middle-age.172 In addition, several common genetic variants have been identified that cause much smaller increases in LDL cholesterol concentration and these are associated with correspondingly smaller increases in the risk of coronary events, providing further evidence in support of a causal association.173
 
Proven beneficial effects of lowering LDL cholesterol concentration with statin therapy
 
In the pre-statin era, meta-analyses of randomised controlled trials of cholesterol-lowering diets, drugs, and ileal bypass surgery showed that, within a few years of reducing blood cholesterol concentrations, rates of non-fatal myocardial infarction and coronary death are reduced.174 In addition, the randomised trials that involved larger and more prolonged cholesterol reductions yielded larger reductions in the rates of coronary events. However, it was suggested that these beneficial effects might be offset by excesses in non-coronary deaths and cancers, which generated uncertainty about the overall benefits of lowering cholesterol.175, 176, 177 The development of statins, which can lower LDL cholesterol concentration to a greater extent than any of the previously available treatments, provided an opportunity to obtain clear evidence about the beneficial effects of LDL cholesterol lowering on atherosclerotic events and deaths, as well as to determine whether it produces adverse effects on other causes of major morbidity and mortality.178 For, although it may not always be possible to distinguish between adverse effects caused by lowering LDL cholesterol concentration and those due to off-target effects of statins (such as myopathy), reliable evidence of a lack of adverse effects with statin therapy should be generalisable about the safety of lowering LDL cholesterol concentrations per se.
 
Effects of statin therapy on LDL cholesterol concentrations
 
During the past 20 years, the increasingly widespread use of statin therapy among individuals who are known to have occlusive vascular disease or are considered to be at increased risk of cardiovascular events for other reasons (eg, having high cholesterol concentrations or other risk factors, such as older age, hypertension, or diabetes) has been associated with downward shifts in the distributions of LDL and total cholesterol concentrations in many populations.179, 180 In addition, because of the tendency for statin therapy to be prescribed more commonly to individuals with elevated LDL cholesterol concentrations, the proportions with high concentrations have been preferentially reduced.179, 180 Representative data from population-based studies conducted before evidence of beneficial effects of statin therapy on fatal and non-fatal vascular events emerged from large randomised trials indicate that average LDL cholesterol concentrations in European and North American populations among people in middle and old age are about 4 mmol/L in the absence of statin therapy.181, 182
 
The proportional reductions in LDL cholesterol achieved with statin therapy are not materially affected by the starting LDL cholesterol concentration or by other patient characteristics (such as age, sex, vascular risk, genetic markers).31, 183 Different statins have different potencies, with the newer agents (eg, atorvastatin and rosuvastatin) able to produce larger reductions in LDL cholesterol per mg of drug than the older agents (eg, simvastatin and pravastatin; table 3).160, 163 Irrespective of the statin used, each doubling of the dose produces an extra reduction of about 6 percentage points in LDL cholesterol (eg, 43% vs 49% reductions with atorvastatin 20 mg vs 40 mg daily). The American College of Cardiology/American Heart Association 2013 Blood Cholesterol Guideline classified statin regimens as being of low intensity (eg, <30% LDL cholesterol reduction with simvastatin 10 mg daily), moderate intensity (eg, 30% to <50% reduction with simvastatin 20-40 mg, atorvastatin 10-20 mg, or rosuvastatin 5-10 mg daily), or high intensity (eg, ≥50% reduction with atorvastatin 40-80 mg or rosuvastatin 20-40 mg daily).162 Use of high-intensity statin therapy would be expected to reduce LDL cholesterol by at least 2 mmol/L in individuals who present with concentrations of 4 mmol/L or more (ie, about half of the population in the absence of statin therapy181, 182), but by only about 1 mmol/L in those presenting with concentrations of 2 mmol/L. Consequently, since the proportional reductions in rates of vascular events with statin therapy are related to the absolute reductions in LDL cholesterol that are achieved, intensive statin therapy should be focused on patients at higher risk of vascular events rather than just on those with high cholesterol concentrations.162, 163, 186

0104174

Effects of statin therapy on LDL cholesterol concentrations
 
During the past 20 years, the increasingly widespread use of statin therapy among individuals who are known to have occlusive vascular disease or are considered to be at increased risk of cardiovascular events for other reasons (eg, having high cholesterol concentrations or other risk factors, such as older age, hypertension, or diabetes) has been associated with downward shifts in the distributions of LDL and total cholesterol concentrations in many populations.179, 180 In addition, because of the tendency for statin therapy to be prescribed more commonly to individuals with elevated LDL cholesterol concentrations, the proportions with high concentrations have been preferentially reduced.179, 180 Representative data from population-based studies conducted before evidence of beneficial effects of statin therapy on fatal and non-fatal vascular events emerged from large randomised trials indicate that average LDL cholesterol concentrations in European and North American populations among people in middle and old age are about 4 mmol/L in the absence of statin therapy.181, 182
 
The proportional reductions in LDL cholesterol achieved with statin therapy are not materially affected by the starting LDL cholesterol concentration or by other patient characteristics (such as age, sex, vascular risk, genetic markers).31, 183 Different statins have different potencies, with the newer agents (eg, atorvastatin and rosuvastatin) able to produce larger reductions in LDL cholesterol per mg of drug than the older agents (eg, simvastatin and pravastatin; table 3).160, 163 Irrespective of the statin used, each doubling of the dose produces an extra reduction of about 6 percentage points in LDL cholesterol (eg, 43% vs 49% reductions with atorvastatin 20 mg vs 40 mg daily). The American College of Cardiology/American Heart Association 2013 Blood Cholesterol Guideline classified statin regimens as being of low intensity (eg, <30% LDL cholesterol reduction with simvastatin 10 mg daily), moderate intensity (eg, 30% to <50% reduction with simvastatin 20-40 mg, atorvastatin 10-20 mg, or rosuvastatin 5-10 mg daily), or high intensity (eg, ≥50% reduction with atorvastatin 40-80 mg or rosuvastatin 20-40 mg daily).162 Use of high-intensity statin therapy would be expected to reduce LDL cholesterol by at least 2 mmol/L in individuals who present with concentrations of 4 mmol/L or more (ie, about half of the population in the absence of statin therapy181, 182), but by only about 1 mmol/L in those presenting with concentrations of 2 mmol/L. Consequently, since the proportional reductions in rates of vascular events with statin therapy are related to the absolute reductions in LDL cholesterol that are achieved, intensive statin therapy should be focused on patients at higher risk of vascular events rather than just on those with high cholesterol concentrations.162, 163, 186
 
Reductions in rates of major vascular events (panel 3)
 
The prespecified purpose of the CTT meta-analyses was to assess the effects of lowering LDL cholesterol on atherosclerotic events in different types of patient more reliably than would be possible in any of the separate randomised trials and (given previous concerns about cholesterol-lowering therapy) to determine whether there were adverse effects on non-vascular causes of death and site-specific cancers.79 Consequently, data were sought for each of the eligible trials about the baseline characteristics of each patient and about myocardial infarctions, strokes, coronary revascularisations, cancers, and causes of death that occurred during the scheduled treatment period (but not any other adverse events, which is the subject of an ongoing project187). Follow-up of outcomes in the trials was reported to be about 99% complete. It was pre-specified that results of the meta-analyses would be presented as risk reductions per mmol/L reduction in LDL cholesterol.29, 79
 
Panel 3
 
Proven beneficial effects of statin therapy

 
⋅Effective low-cost statin regimens (eg, generic atorvastatin 40 mg daily costs about 2 per month) reduce LDL cholesterol by more than 50% (ie, at least 2 mmol/L in individuals presenting with LDL cholesterol concentrations of ≥4 mmol/L).
 
⋅Large-scale evidence from randomised trials shows that each 1 mmol/L reduction in LDL cholesterol with statin therapy produces a proportional reduction of about 25% in the rate of major vascular events (coronary deaths, myocardial infarctions, strokes, and coronary revascularisations) during each year (after the first) that it continues to be taken. Consequently, lowering LDL cholesterol by 2 mmol/L reduces risk by about 45%.
 
⋅Lowering LDL cholesterol by 2 mmol/L with an effective statin regimen for about 5 years in 10 000 patients would typically prevent major vascular events in about 1000 (10%) patients at high risk of heart attacks and strokes (eg, secondary prevention) and 500 (5%) patients at lower risk (eg, primary prevention).
 
⋅Despite reports based largely on non-randomised observational studies, there is not good evidence that statin therapy produces beneficial effects on other health outcomes (eg, cancer, infections, respiratory disease, arrhythmias).
 
In total in the CTT meta-analyses, there were about 25 000 major vascular events (defined as the composite of coronary deaths or non-fatal myocardial infarctions, strokes of any type, and coronary revascularisation procedures) during an average of about 5 years of scheduled study treatment. The proportional reductions in these major vascular event rates were related to the absolute reductions in LDL cholesterol that were achieved (figure 3). Overall, in the trials of routine statin therapy versus no routine use, there was a 20% proportional reduction in the major vascular event rate per mmol/L LDL cholesterol reduction (figure 4). The proportional risk reduction was smaller during the first year after starting treatment, whereas it was 24% (ie, a risk ratio of 0⋅76) on average during each subsequent year that allocation to statin therapy was continued (p<0⋅0001 for difference between effects in first vs later years). In the trials of more versus less intensive statin regimens, the average 0⋅5 mmol/L further reduction in LDL cholesterol yielded a 15% further proportional reduction in the rate of major vascular events (figure 3), corresponding to a 28% reduction (ie, a risk ratio of 0⋅72) per mmol/L further LDL cholesterol reduction during each year of treatment (with no apparent delay after increasing the intensity of statin therapy).31
 
Consequently, the proportional reduction in the risk of major vascular events per mmol/L was about one-quarter in the trials of statin versus no statin (after an initial delay) and of more versus less intensive therapy. Based on the combined findings from these two sets of trials, it can be estimated that reducing LDL cholesterol concentrations by 2 mmol/L would reduce the risk of major vascular events by about 45% (derived as [1⋅0-(0⋅75 x 0⋅75)] x 100) during each year treatment is continued. In principle, even larger reductions in LDL cholesterol would be expected to produce even larger risk reductions (eg, 60-70% with 3-4 mmol/L LDL cholesterol reductions); however, this is likely only to be clinically relevant in limited circumstances (eg, for individuals with familial hypercholesterolaemia who have very high LDL cholesterol concentration).
 
In these meta-analyses, statin therapy produced similar proportional reductions per mmol/L LDL cholesterol reduction in the risks of each of the main components of the composite outcome of major vascular events (ie, myocardial infarctions and coronary deaths; strokes of any type; or coronary revascularisations).32 The proportional reductions in major vascular events were also similar among different types of patient.29, 30, 31, 32, 33, 34 For example, as would be expected from the log-linear associations in observational epidemiological studies between coronary disease risk and cholesterol concentration (figure 2B), the proportional reductions in risk per mmol/L reduction were about the same irrespective of the concentrations of cholesterol at presentation (figure 1). The proportional risk reductions appeared to be smaller among individuals aged older than 75 years who were included in these trials, but they had a higher prevalence of severe heart failure and end-stage renal disease (conditions associated with non-atherosclerotic vascular outcomes not much influenced by lowering LDL cholesterol).188 Moreover, since the absolute risks of major vascular events were higher among older individuals, the absolute benefits were of similar size to those among younger individuals. The proportional risk reductions also appeared to be slightly smaller among the women included in these trials. However, this apparent difference could be accounted for largely by differences in non-sex-related characteristics, and the relative effects were similar for men and women at equivalent risk of cardiovascular events.33 The risks of major vascular events were reduced in secondary prevention as well as in primary prevention (including among individuals with diabetes or hypertension),29, 31 but the proportional reductions were somewhat larger among lower-risk individuals. This finding is consistent with results from Mendelian randomisation studies,189 which indicate that genetically determined exposure to lower LDL cholesterol concentrations before atherosclerosis has developed may produce larger risk reductions.173
 
In general, the absolute benefits of using statin therapy depend on an individual's absolute risk of atherosclerotic events and the absolute reduction in LDL cholesterol that can be achieved. For example, 5 years of treatment with a statin regimen that lowers LDL cholesterol by 2 mmol/L would be expected to prevent major vascular events in about 1000 (10%) higher-risk patients per 10 000 treated and in about 500 (5%) lower-risk patients per 10 000 treated (figure 5; which also provides estimates of the absolute benefits with smaller LDL cholesterol reductions).32 The continued follow-up of patients beyond the end of the trials has found that the benefits of statin therapy persist (and may even become larger)98, 99, 100, 101, 102, 103, 104, 105, 106 for many years after the differences in statin use between the randomised groups have ceased. However, of more relevance for a treatment that is intended to be continued for life once it has been started, the meta-analyses show that statin therapy reduces the risk of major vascular events during each year that it is continued (figure 4). Consequently, even larger absolute benefits would be expected with statin therapy that is continued for longer than the average of about 5 years in these randomised trials.
 
Reductions in coronary mortality
 
Overall in the CTT meta-analyses, there was a statistically robust 12% proportional reduction in vascular mortality per mmol/L LDL cholesterol reduction (figure 6), attributable chiefly to a 20% proportional reduction in coronary deaths (with, as was seen for major vascular events, a greater proportional effect after the first year of treatment), along with an 8% reduction in other cardiac deaths (some of which, such as those due to arrhythmias or heart failure, may not be due to atherosclerotic causes and so not amenable to LDL cholesterol-lowering therapy) and little effect on death due to all types of stroke combined.31, 32 Both for the aggregate of all vascular deaths and for coronary and non-coronary causes considered separately, the proportional reductions in risk per mmol/L LDL cholesterol reduction appear to be similar in patients with and without pre-existing vascular disease, and in those who present at different levels of baseline vascular risk, as well as in other subgroups that have been considered.29, 30, 31, 32, 33, 34
 
As discussed above, when there is compelling evidence of an effect of a treatment on a particular outcome (ie, vascular mortality) and this is supported by the effects on related outcomes (ie, the even more statistically robust reductions in non-fatal major vascular events with statin therapy), then the appropriate question to ask is whether there is good evidence that the treatment does not reduce that outcome in different circumstances (rather than whether there is direct evidence of benefit in every circumstance).3, 13, 71 Even in the aggregate of all of the trials in the CTT meta-analyses, too few vascular deaths occurred among lower-risk participants for reliable direct assessment of the effects of statin therapy in such individuals considered in isolation (as has been proposed by some commentators11, 12, 190). However, the proportional risk reduction was statistically compatible with the reduction observed in higher-risk patients (trend p=0⋅7) and it was supported by the clear reduction in major vascular events among lower-risk patients.32 Similarly, although there were too few women in these trials to assess the effects on vascular mortality directly (which has been the basis of assertions that statin therapy is not beneficial for women191, 192, 193, 194, 195), the proportional reductions were similar among women and men (interaction p=0⋅8) and were reinforced by definite reductions in major vascular events among women.33
 
Consequently, it is reasonable to conclude that statin therapy produces proportional reductions of at least 20% in coronary mortality per mmol/L LDL cholesterol reduction among people at different levels of occlusive vascular risk irrespective of their sex and, assuming that the proportions of vascular deaths due to coronary and non-coronary causes are similar, of 12% in deaths from all vascular causes. The availability of additional evidence from large trials (such as the HOPE-3 trial in primary prevention196 and the ongoing STAREE trial in people aged older than 70 years197) will provide more direct evidence about the effects in particular circumstances.
 
Proven adverse effects of statin therapy (panel 4)
 
The only excesses of adverse events that have been reliably demonstrated to be caused by statin therapy are myopathy and diabetes mellitus, along with a probable excess of haemorrhagic stroke. These excesses are larger in certain circumstances, but the absolute risks remain small by comparison with the absolute benefits.
 
Panel 4
 
Known adverse effects of statin therapy

 
⋅The only adverse events that have been reliably shown to be caused by statin therapy are myopathy (defined as muscle pain or weakness combined with large increases in creatine kinase blood concentrations) and new-onset diabetes mellitus, along with a probable increase in strokes due to bleeding (ie, haemorrhagic strokes).
 
⋅Typically, treatment of 10 000 patients for 5 years with a standard statin regimen (such as atorvastatin 40 mg daily) would be expected to cause about 5 cases of myopathy, 50-100 new cases of diabetes, and 5-10 haemorrhagic strokes.
 
⋅Despite reports based largely on non-randomised observational studies, there is good evidence that statin therapy does not cause adverse effects on other health outcomes (chiefly muscle pain and weakness) that have been claimed prevent a large proportion of patients from continuing it long term (so-called "statin intolerance").
 
⋅Large-scale evidence from randomised trials rules out excesses of muscle pain and weakness with statin therapy of more than about 10-20 cases annually per 10 000 treated patients, with only about one of those cases being associated with large creatine kinase elevations (ie, myopathy) and requiring statin discontinuation.
 
⋅Absolute excesses of adverse events that are caused by statin therapy are not more than about 100-200 per 10 000 patients (ie, 1-2%) treated for 5 years, and it is unlikely that large adverse effects on serious adverse events await discovery.
 
⋅The harmful effects of statin therapy can usually be reversed without any residual effects by stopping it, whereas the harmful effects of heart attacks or strokes that occur because statin therapy has not been used can be devastating.
 
Increases in rates of myopathy
 
Myopathy (sometimes referred to as myositis) is typically defined as muscle pain, tenderness, or weakness that is accompanied by substantial increases in blood creatine kinase concentrations (eg, greater than ten times the laboratory upper limit of normal).145, 216 Rhabdomyolysis is a severe form of myopathy involving muscle breakdown (usually identified by even larger increases in creatine kinase concentrations), with myoglobin released into the circulation and, in some cases, leading to acute renal failure or worsened renal function.145 Myopathy is rare in normal circumstances. Approved statin regimens have been associated both in observational studies and in randomised trials with large relative risks for myopathy,145, 152, 217 but typically with small absolute excesses (about 1 case per 10 000 people treated per year) and even smaller excesses in the incidence of rhabdomyolysis (about 2-3 cases per 100 000 treated per year).31, 218 It usually resolves rapidly when statin therapy is stopped.145
 
The underlying mechanisms for statin-related myopathy are not well understood. The risk of myopathy is dose related and it appears to depend on the levels of the statin in the circulation (as indicated by its association with a SLCO1B1 gene variant that reduces the transport of all statins from the blood into the liver).217, 219, 220 Cerivastatin was withdrawn from use because the myopathy rate observed in post-marketing surveillance with approved doses was much higher than with other statins.221 In the SEARCH randomised trial,83 simvastatin 80 mg daily produced a more than ten-fold higher rate (at least 1 case of myopathy per 1000 patients treated yearly) than 20 mg daily (or 40 mg daily in HPS;222 about one case per 10 000 yearly), so the high-dose regimen is no longer recommended routinely.223 The rates of reports of myopathy in regulatory databases are also higher with higher doses of atorvastatin, although such spontaneous reports may be biased and the absolute risks are still small even with the highest approved dose.217 The rate of myopathy can be increased substantially when statins are used in combination with other drugs that affect their metabolism (in particular, inhibitors of cytochrome P450 or the P-glycoprotein, such as ciclosporin and azole antifungals) and in certain types of patient (eg, people of Asian origin and those who have functional variation in the SLCO1B1 gene).145, 178, 218, 224 More moderate increases (eg, risk ratios of about 1⋅5 to 2) in the rate of myopathy are also seen in other circumstances (eg, in combination with certain antihypertensive drugs and in women, people aged older than 80 years, and those with diabetes).219
 
Despite this causal association with myopathy, the evidence from randomised controlled trials indicates that statin therapy has little effect on less severe muscle pain (ie, myalgia) or weakness, although such symptoms are commonly attributed to statins in routine practice. Indeed, an excess of muscle-related symptoms has generally only been reported in trials when it occurs in combination with increased creatine kinase concentrations, with bigger relative risks reported with larger creatine kinase increases. For example, in the Heart Protection Study of simvastatin 40 mg daily versus placebo, the relative risk for any myalgia irrespective of increased creatine kinase concentrations was 0⋅99 (95% CI 0⋅95-1⋅03), whereas it was 1⋅7 (0⋅9-3⋅1) for myalgia in patients with a creatine kinase concentration more than four times the upper limit of normal, and 2⋅5 (0⋅8-8⋅0) for those with an increase of more than ten times the upper limit of normal.38, 222 This result provides another illustration of the value of using specific outcomes to detect treatment effects, rather than composites of outcomes that are affected by treatment and those that are not.
 
Increases in rates of diabetes
 
In the JUPITER randomised trial among 17 802 patients without a history of vascular disease, concentrations of glycated haemoglobin were slightly higher after about 2 years among the patients allocated rosuvastatin 20 mg daily than among those allocated placebo (5⋅9% vs 5⋅8%; p=0⋅001).48, 225 There was also a small excess of newly diagnosed diabetes (3⋅0% vs 2⋅4%; p=0⋅01), which corresponds to a 25% (95% CI 5-49) proportional increase. In subsequent meta-analyses of the available results from the randomised trials, standard statin dose regimens were associated with a proportional increase of about 10% in reported diabetes, and more intensive statin regimens (as used in JUPITER) with about a 10% further increase.49, 226 This excess of diabetes diagnoses appeared soon after the start of statin therapy, chiefly among patients who had risk factors for diabetes (eg, elevated body-mass index or HbA1c, or impaired fasting glucose), and did not appear to get larger as treatment continued.225, 227, 228 Prior to these reports from randomised trials, statin therapy had not been associated with increased diabetes incidence in observational studies, although several reports of such associations have been published subsequently.229, 230
 
Genetic variants that reduce the activity of HMG-CoA reductase (which is analogous to inhibiting this enzyme with a statin) have been associated with an increased incidence of diabetes.231 Likewise, individuals with familial hypercholesterolaemia-in whom the numbers and function of LDL receptors on cell surfaces are reduced (by contrast with the increase in receptors produced by statins)-had been diagnosed with diabetes less frequently than were their unaffected relatives.232 These genetic experiments of nature provide support for the association of statin therapy with an excess of diabetes being causal. The mechanism is not known: it could be directly related to LDL cholesterol lowering,233 but it has also been hypothesised that increasing the numbers of LDL receptors (eg, with treatments like statins and PCSK9 inhibitors) might cause diabetes by enabling more cholesterol to enter and damage pancreatic cells.232
 
However, the clinical relevance of this excess of diabetes is less clear; in particular, the cardiovascular benefits of statin therapy are substantial despite any increase in diabetes-related morbidity. The underlying incidence of new-onset diabetes in the primary prevention trials was about 1% per year,49 so the absolute excess with statin therapy was about 10-20 per 10 000 per year (with this range reflecting the intensity of the statin regimen). If it is assumed that this statin-related diabetes is associated with as much as a doubling of cardiovascular risk (as is the case for spontaneously occurring diabetes234) then it might result in major vascular events among about 5-10 of 10 000 individuals with an underlying 5-year risk of 5-10% (eg, primary prevention) who are treated for 5 years. However, despite this potential adverse impact, lowering LDL cholesterol by 1-2 mmol/L with statin therapy prevents major vascular events among about 150-300 per 10 000 such individuals who are treated for 5 years (figure 5). The absolute benefits are even larger among higher-risk patients (including those who already have diabetes; Figure 1, Figure 5)30, 31, 32 and, again despite any adverse impact of the diabetes excess, increase while statin therapy continues to be taken (figure 4). There is also no good evidence of an excess of microvascular complications related to diabetes with statin therapy (as described below).
 
Probable increases in rates of haemorrhagic stroke
 
In observational studies, blood cholesterol concentrations have been negatively associated with rates of haemorrhagic stroke, particularly at low concentrations of cholesterol in people with high blood pressure.157, 235, 236 In the randomised SPARCL trial among 4731 patients with prior cerebrovascular disease, allocation to atorvastatin 80 mg daily produced a definite reduction in ischaemic stroke (218 [9⋅2%] vs 274 [11⋅6%]; p=0⋅008), but there was also a possible increase in haemorrhagic stroke (55 [2⋅3%] vs 33 [1⋅4%]; p=0⋅02).39 When these results were combined with those from the other trials included in the CTT meta-analysis, there was a 21% (95% CI 5-41; p=0⋅01) proportional increase in the incidence of haemorrhagic stroke per mmol/L reduction in LDL cholesterol.31
 
In European and North American populations, this would typically translate into an absolute excess of about 5-10 haemorrhagic strokes per 10 000 patients in whom LDL cholesterol is reduced by 1-2 mmol/L for 5 years with statin therapy. The absolute excess would be expected to be bigger in individuals with pre-existing cerebrovascular disease39 and in populations (such as in Asia) where the underlying rates of haemorrhagic stroke are higher.237 However, statin therapy has been found to reduce the overall risk of stroke in many different settings (including in people who have already had a stroke39 or have hypertension238) irrespective of the underlying risk of vascular disease.32 For example, the increase in haemorrhagic stroke is outweighed by the reduction in the risk of ischaemic stroke, as well as in other occlusive vascular events and deaths, even among individuals with a 5-year risk of major vascular events below 10%.
 
Other adverse events that have been attributed to statin therapy
 
It has been claimed that statin therapy causes increased rates of other types of adverse health outcome, as well as symptomatic side-effects (chiefly muscle pain and weakness) that prevent a large proportion of patients from continuing to take statin therapy long term, often now referred to as "statin intolerance".10, 11, 12, 22, 64, 73, 74, 75 These claims have been chiefly based on reports to regulatory authorities of adverse events that have been attributed to a statin and on non-randomised observational studies based on health-care databases. However, they are not supported by the evidence from randomised controlled trials: in particular, statin therapy has been found to be no less well tolerated than placebo.53, 65, 239, 240, 241, 242
 
As is discussed above, the potential biases inherent in studies without both randomly assigned control groups and masked ascertainment of outcomes limit their ability to demonstrate causal associations (except for large effects on rare outcomes). This is particularly the case for symptomatic adverse events that are attributed to statin use, especially if such reports have been prompted by guidance from clinicians to their patients or from patient information leaflets and other sources.146, 147, 148, 243, 244 By contrast, the inclusion of large numbers of different patient types in randomised controlled trials of prolonged statin therapy with different eligibility criteria provides unbiased evidence about adverse effects of treatment that are relevant to routine clinical practice.
 
Muscle-related outcomes (other than myopathy)
 
The adverse events most commonly attributed to statin therapy relate to muscle pain (ie, myalgia) or other muscle-related symptoms. For example, based on the NHANES database, it was reported that 23% of 671 statin users who did not have arthritis recalled having episodes of musculoskeletal pain (not muscle pain specifically) during the previous month compared with 18% of 4499 individuals who were not taking a statin.245 After statistical adjustment for the recorded differences (which were substantial) between the characteristics of the patients using and not using a statin, a prevalence ratio of 1⋅33 (95% CI 1⋅06-1⋅67; p=0⋅02) was reported with statin use. In another observational study91 of statin use based on health-care data, musculoskeletal pain was reported by 73⋅4% of 6967 statin users compared with 71⋅6% of 6967 non-users during a median of 4⋅7 years, yielding an odds ratio of 1⋅09 (95% CI 1⋅02-1⋅18; p=0⋅02) after attempting to match patients with propensity scores based on recorded characteristics (which, again, differed substantially).
 
Both of these reports discussed the inability of such non-randomised studies to assess causality because of the potential for residual differences between patients who had used statins and those who had not (despite statistical adjustment for recorded characteristics). They also mentioned the potential for ascertainment bias due to patients who were taking statins being examined more frequently. However, neither report commented on the inherent lack of masking of treatment in such studies and the consequent potential for bias due to patients prescribed statins being advised by their doctors that they may cause muscle pain (whereas such advice is, of course, not given to patients not prescribed a statin).246 In addition, the result for NHANES excluded the 3058 individuals with arthritis in whom statin use was not associated with any excess of musculoskeletal pain (prevalence ratio 0⋅96, 95% CI 0⋅81-1⋅15).245 Such data-dependent selection of which results to emphasise introduces yet another potential source of bias into this assessment of the effects of statin therapy.3
 
In general, the data available for observational studies based on health-care records do not derive from a systematic approach to seeking and recording information about symptoms or about the use of statin therapy. The PRIMO survey tried to overcome this limitation by systematically seeking information about the muscle symptoms that were reported.247 Among 7924 patients with hyperlipidaemia receiving high-dose statin therapy, 10⋅5% reported muscle symptoms at a median of about 1 month after starting it. However, those patients were required to give informed consent, which presumably involved advising them that statins can cause muscle problems and that the aim was to assess this outcome specifically, increasing the likelihood of prompting reports of muscle symptoms. In any case, since there was no control group in that study, it is not able to provide any useful information as to whether statins cause an increase in such symptoms.
 
It has been asserted that the rates of muscle-related symptoms caused by statins may be underestimated in randomised trials because of the exclusion of patients at risk of these problems (such as those with a history of muscle problems or creatine kinase elevations with statin therapy) and a perceived lack of systematic questioning and standardised definitions.11, 12, 22, 23, 24, 74, 248 However, as discussed above, few patients would have been exposed to statin therapy prior to recruitment into many of the large clinical outcome trials and use of a pre-randomisation placebo run-in phase in about half of the trials (appendix) would tend to increase the sensitivity of the subsequent randomised comparisons to detect any effects.80 The inclusion of large numbers of different types of patients in different randomised trials with different eligibility criteria also makes the evidence about any side-effects of statin therapy far more widely generalisable to routine practice than is often asserted.11, 12, 22, 23, 24, 76, 77, 78
 
In addition, use of masked control groups ensures that health outcomes are ascertained in the same way in the different treatment groups within any particular trial.17, 18, 246 Consequently, even though different randomised trials of statin therapy did not always use the same methods to identify or classify muscle symptoms (and may even have failed to detect some relevant events; table 1), each within-trial masked comparison should still provide a reliable assessment of the effects of statin therapy on muscle-related problems (and, indeed, on other adverse events).19 Moreover, even though some of the trials did not seek information about muscle-related problems, this would not introduce bias into the assessment of the effects of statin therapy based on the trials that did record them. In principle, the failure of some trials that did record such outcomes to publish their results does have the potential to introduce bias. However, muscle-related problems are common, and the large numbers of such outcomes that have been reported from many different trials (appendix) makes it unlikely that material bias exists in the published literature. Consequently, the general lack of differences between the randomised treatment groups in the rates of the different muscle-related outcomes recorded in the large masked trials that are eligible for the CTT meta-analysis (some of which assessed such symptoms particularly carefully; appendix) provides strong evidence against statin therapy causing much effect on muscle-related symptoms. In the JUPITER and HOPE-3 trials (of rosuvastatin 20 mg and 10 mg daily, respectively), there were small excesses in some muscle-related outcomes.48, 196, 249 However, no excesses of muscle-related outcomes were observed among the large numbers of patients in the other large randomised masked trials of long-term statin therapy. Nor were there excesses in those trials that sought information about the severity of any muscle symptoms or about stopping study treatment because of muscle symptoms.246
 
The STOMP trial248 was specifically designed to assess the effects of statin therapy on several prespecified muscle-related measures. Compared with 236 patients allocated placebo, there were no apparent effects on muscle strength or endurance, aerobic performance, or physical activity among 232 statin-naive patients randomly allocated atorvastatin 80 mg daily for 6 months. Cases of unexplained muscle pain (23 [9⋅9%] vs 14 [5⋅9%]; p=0⋅1) and the subset of those cases defined as myalgia (19 [8⋅2%] vs 10 [4⋅2%]; p=0⋅08) were reported more commonly among patients allocated atorvastatin, but these differences in the prespecified intention-to-treat comparisons were compatible with chance. In a meta-analysis of 26 masked trials (including STOMP) that involved at least 6 months of statin therapy,250 there was little difference in the reported rates of muscle problems during an average treatment duration of 3 years: 12⋅7% among 59 237 participants allocated statin versus 12⋅4% among 54 458 allocated placebo; an absolute excess of 0⋅3% (95% CI 0-0⋅7; p=0⋅06) or, alternatively, a range from zero to 20 cases per 10 000 years of treatment. Similarly, combination of the results in the large placebo-controlled trials that were eligible for the CTT meta-analyses (appendix) yields similar results: 5162 (11⋅7%) cases allocated statin therapy versus 5015 (11⋅4%) allocated placebo during an average of 5 years of treatment (p=0⋅10). Moreover, the difference is even smaller in the numbers of cases of muscle problems that resulted in study treatment being stopped: 201 (0⋅63%) versus 183 (0⋅58%); p=0⋅37.
 
Crossover trials, in which active and placebo treatment are allocated in a random sequence to each patient, may be particularly sensitive for detecting adverse effects that emerge rapidly after treatment starts and resolve soon after stopping treatment. No differences in myalgia or other pain measures were observed in a randomised re-challenge trial with three statin-placebo paired crossover comparisons among 8 patients with prior statin-related myalgia (with or without creatine kinase elevations), and 5 of the patients resumed statin therapy.251 In another trial, 86 patients were assigned simvastatin 40 mg daily (combined with amlodipine, losartan, and hydrochlorothiazide) or a matching placebo in a random sequence; muscle aching was reported more commonly on the active polypill (9 vs 1 cases), but it was not considered sufficiently troublesome to stop treatment.252 Among 492 patients with a history of not tolerating two or more statin regimens who were randomised to receive atorvastatin 20 mg daily then placebo or placebo then atorvastatin,253 muscle-related symptoms were reported by 43% of the patients when on atorvastatin but not on placebo versus 27% of them when on placebo but not on atorvastatin, yielding a risk ratio of 1⋅5 (although it has been suggested that this trial may not have been properly masked246). In a similar crossover trial among 131 patients with a history of muscle complaints who were randomised to simvastatin 20 mg daily then placebo or placebo then simvastatin, muscle pain was reported by 36% of the patients when on simvastatin but not on placebo versus 29% of them when on placebo but not on simvastatin.254, 255 These results indicate that, even among highly selected patients who have repeatedly attributed intolerable symptoms to statin therapy, some of the reported muscle-related intolerance may be due to the statin but most of it is not.
 
In summary, given the 0⋅3% absolute excess of muscle problems based on more than 10 000 reported cases in meta-analyses of randomised trials during 3-5 years of treatment (appendix),250 the excess rate of symptomatic muscle pain and other muscle-related problems due to statin therapy would appear to be no more than about 10-20 cases yearly per 10 000 treated individuals, with only about one of those cases associated with substantial elevations in creatine kinase concentrations (ie, myopathy) and requiring statin therapy to be stopped.