Do Statins Cause the Symptoms People Report?

The question

A patient starts a statin. Within weeks their thighs ache, they feel unusually tired, and they cannot concentrate. They stop the drug, the symptoms ease, and they conclude that statins do not agree with them. Their experience is real. The ache is real, the fatigue is real, and the timing is real.

The best blinded evidence says the drug probably did not cause it.

Both statements can be true at once. That is the puzzle this brief takes seriously. The question is not whether people feel worse while taking statins. Many do. The question is what fraction of that symptom burden is caused by the drug molecule, and what fraction is caused by the act of taking a pill, by prior expectation, by ordinary background symptoms, and by the fact that people notice their bodies more carefully when they have just been told they are at risk of a heart attack.

This is a bounded evidence synthesis, not medical advice. It asks how much of the reported harm is pharmacological, how much is nocebo or drucebo — symptoms produced by expectation and by the ritual of taking a tablet rather than by the active ingredient — and what that implies for how the drugs are described, prescribed, and stopped.

Why the puzzle matters

Statins are among the most prescribed drugs in the world. Muscle symptoms are the most common reason people stop them. A meta-analysis of 176 studies covering more than four million patients estimated that about 9 percent of statin users meet a definition of statin intolerance, with estimates ranging from roughly 5 percent in randomized trials to 17 percent in cohort studies [7]. In a real-world Italian cohort of 16,717 patients, 9.6 percent reported statin-associated muscle symptoms [5]. In the PRIMO study of 7,924 patients on high-dose statins, 10.49 percent reported muscle symptoms [6].

Those numbers are not trivial. If 9 percent of the tens of millions of people taking statins stop because of symptoms, the population-level consequences are large. Observational studies consistently associate statin discontinuation with more heart attacks and higher cardiovascular mortality. The causal interpretation of those associations is contestable, because people who stop treatment may differ from those who continue in ways that records do not capture. But the direction is consistent and biologically plausible: statins prevent vascular events, so stopping them forgoes prevention.

At the same time, the blinded evidence is unusually clear. In the largest individual-participant meta-analysis on the question, the average excess of muscle pain caused by statins was small, concentrated in the first year, and absent afterward [1]. The contrast between a 9 percent reporting rate and a near-null average causal effect is the central anomaly. It is not a contradiction. It is a lesson in what different designs measure.

A report is a behavior. A cause is a counterfactual. Open-label studies count reports. Blinded trials estimate what would have happened without the drug. When the two diverge, the divergence is the finding.

What the blinded trials find

The Cholesterol Treatment Trialists' Collaboration analyzed individual participant data from 19 double-blind placebo-controlled trials with 123,940 participants, plus four trials comparing more intensive with less intensive statin regimens [1]. The population was typical of statin trials: mean age 63, 28 percent women, nearly half with existing vascular disease, 18.5 percent with diabetes, median follow-up 4.3 years.

During the trials, 27.1 percent of participants allocated to statins reported muscle pain or weakness, compared with 26.6 percent allocated to placebo. The rate ratio was 1.03, with a 95 percent confidence interval of 1.01 to 1.06. In the first year, the relative increase was 7 percent (RR 1.07, 95 percent CI 1.04 to 1.10), corresponding to an absolute excess of 11 events per 1,000 person-years (95 percent CI 6 to 16). After the first year, the excess disappeared: RR 0.99 (95 percent CI 0.96 to 1.02), or zero extra events per 1,000 person-years (95 percent CI −2 to 1).

The attributable fraction follows arithmetic. If the rate ratio in year one is 1.07, then (1.07 − 1.00) / 1.07 ≈ 0.065, or about one in fifteen. In plain terms, for every fifteen people who reported new muscle pain or weakness in the first year of a moderate-intensity statin, roughly one was reporting something the drug actually caused. The other fourteen would have reported something similar without it. The investigators put the same point more bluntly: more than 90 percent of muscle-symptom reports among people allocated statins were not due to the statin.

Dose matters, but less than the reporting pattern. More intensive regimens — 40 to 80 milligrams of atorvastatin or 20 to 40 milligrams of rosuvastatin daily — produced a larger relative increase than moderate-intensity regimens (RR 1.08 versus 1.03), and a small excess persisted after the first year (RR 1.05, 95 percent CI 0.99 to 1.12), though the interval crossed one. Median creatine kinase rose by about 0.02 times the upper limit of normal.

Serious muscle injury is real but rare. The same analysis estimates roughly one extra case of myopathy per 10,000 person-years and two to three cases of rhabdomyolysis per 100,000 person-years [1]. Those risks are why high-dose statins require attention to interactions, renal function, and symptoms such as severe weakness or dark urine. They do not explain the typical ache.

The year-one concentration is important. It fits a genuine early pharmacological effect, but also heightened attention: people are most watchful in the first months after starting a drug they have been told they need. The data cannot fully separate those explanations. They can say that a new muscle symptom after years of tolerated treatment has a very low prior probability of being drug-caused.

The trial that separated the drug from the pill

The cleanest experiment on this question did not just compare statin with placebo. It compared statin, placebo, and nothing at all.

SAMSON enrolled 60 patients who had stopped statins because of side effects that began within two weeks of starting [2]. Each received twelve one-month bottles in random order: four with atorvastatin 20 milligrams, four with placebo, four empty. Symptoms were recorded daily on a 0-to-100 smartphone scale. The design allows three comparisons: statin versus placebo isolates the drug, placebo versus no tablet isolates the effect of taking a pill, and statin versus no tablet captures the total burden of treatment.

The results were unambiguous. Mean symptom intensity was 8.0 during no-tablet months (95 percent CI 4.7 to 11.3), 15.4 during placebo months (95 percent CI 12.1 to 18.7), and 16.3 during statin months (95 percent CI 13.0 to 19.6). Placebo and statin were statistically indistinguishable (P = 0.39). The nocebo ratio — the share of the statin-associated symptom increment that placebo also produced — was 0.90. In other words, about 90 percent of the symptom burden elicited by the statin challenge was also elicited by a tablet with no active drug. The drug-specific increment was under one point on a 100-point scale.

A secondary analysis of the same trial examined the diagnostic intuitions clinicians use [3]. Neither the intensity of symptoms at the start of a treatment period nor the degree of relief when stopping distinguished statin from placebo. Stopping was no more frequent on statin than on placebo (P = 0.173). The pattern that patients and doctors treat as proof of causation — "I started it and it began; I stopped it and it went away" — appeared identically with inert tablets.

StatinWISE replicated the core result in a larger, primary-care-based series of n-of-1 trials [4]. Two hundred participants who had recently stopped or were considering stopping statins because of muscle symptoms were randomized to six two-month periods of atorvastatin 20 milligrams or placebo. Among the 151 participants with at least one statin and one placebo period, mean muscle symptom scores were 1.68 with statin and 1.79 with placebo on a 0-to-10 scale, a mean difference of −0.11 (95 percent CI −0.36 to 0.14; P = 0.40). Withdrawals for intolerable muscle symptoms were 18 (9 percent) during statin periods and 13 (7 percent) during placebo periods. Two-thirds of completers reported restarting long-term statins.

Neither trial is perfect. SAMSON was small, enrolled only people whose symptoms began within two weeks, and demanded a year of adherence. StatinWISE lacked a no-tablet arm, so it cannot separate drug from pill-taking. But the designs converge, and together they explain the trial-clinic gap better than any single mechanism.

The gap between trials and clinics

If the causal effect is so small, why do so many people report symptoms?

Part of the answer is that different studies measure different things. The 9 percent figure from the meta-analysis of 176 studies is a prevalence of reported intolerance, not a causal estimate [7]. Within that same meta-analysis, the prevalence was 4.9 percent in randomized controlled trials and 17 percent in cohort studies. The design, not the biology, moved the number by a factor of three.

PROSISA shows what happens when reported symptoms are tested. In the Italian cohort, 9.6 percent reported muscle symptoms [5]. Of those, 82.2 percent underwent some form of dechallenge or rechallenge — stopping the drug, switching it, or restarting it. Symptoms reappeared in 38.4 percent of the rechallenged patients, which corresponds to 3.01 percent of the entire cohort. Even without blinding, a structured rechallenge cut the apparent prevalence by roughly two-thirds. The remaining 3 percent is an upper bound on the pharmacologically intolerant group, not a lower bound, because rechallenge without blinding still runs through expectation.

The placebo arm of trials in people who believe they are statin-intolerant makes the same point from the other direction. A systematic review of eight blinded trials enrolling 8,095 patients with a history of statin intolerance found that 21.34 percent reported muscle adverse events while receiving placebo, and 6.12 percent discontinued because of them [16].

The International Lipid Expert Panel estimates that complete statin intolerance affects only 3 to 5 percent of treated patients, and that more than half of all muscle pain or weakness attributed to statins may be nocebo or drucebo [9]. The panel cites a systematic review of trials with both open-label and blinded phases estimating that 38 to 78 percent of statin-associated muscle symptom intolerance could be attributed to expectation alone. That range is wide and comes from a secondary synthesis, so it should be treated as an order of magnitude rather than a precise quantity. But it is consistent with the primary n-of-1 data.

Several ordinary mechanisms generate the gap. Background muscle aches are common in the age groups that take statins. People start statins after a health scare, when attention to bodily sensation is high. Symptoms fluctuate, so the first month of a new drug coincides with heightened surveillance and regression to the mean. Media coverage and labels prime people to expect specific symptoms. And taking a daily tablet is itself a reminder that one is a patient, which can produce real symptoms without active pharmacology.

None of this makes the symptoms imaginary. The placebo arm of SAMSON generated a real increase in symptom intensity, from 8.0 to 15.4 on a 100-point scale. The symptoms are experienced. What is mistaken is the inference from experience to molecule.

What statins do cause

The corrective to "statins cause everything people report" is not "statins cause nothing." The same blinded evidence identifies real harms, and they deserve the same precision as the nocebo findings.

A small muscle excess. The CTT meta-analysis found a genuine, dose-related increase in muscle pain, concentrated in the first year [1]. STOMP, a six-month randomized trial of atorvastatin 80 milligrams in 420 healthy statin-naive volunteers, found myalgia in 19 participants on atorvastatin versus 10 on placebo (P = 0.05), and an average creatine kinase increase of 20.8 units per liter [8]. Muscle strength and exercise capacity did not change significantly. This is a real but modest signal: high-dose therapy can produce mild muscle injury and symptoms even in people who have never taken a statin.

Diabetes. A collaborative meta-analysis of 13 trials with 91,140 participants found a 9 percent relative increase in incident diabetes (odds ratio 1.09, 95 percent CI 1.02 to 1.17) [11]. The absolute risk was small: treating 255 patients for four years produced one extra case (95 percent CI 150 to 852). Most excess cases occur in people whose glucose was already near the diagnostic threshold. The cardiovascular benefit in the same trials was larger.

Liver enzymes, oedema, and urinary changes. The 2026 label-wide meta-analysis found abnormal liver transaminases about 41 percent more common on statins (RR 1.41, 95 percent CI 1.26 to 1.57), and other liver function test abnormalities about 26 percent more common (RR 1.26, 95 percent CI 1.12 to 1.41) [10]. The absolute annual excess for combined liver function test abnormality was 0.13 percent. It also found small excesses of oedema (RR 1.07, 95 percent CI 1.02 to 1.12) and urinary composition alteration (RR 1.18, 95 percent CI 1.04 to 1.33). These are modest and of uncertain clinical significance, not liver failure.

Cognition. This is where public belief and trial evidence diverge most. In 2012 the U.S. Food and Drug Administration added a label warning about rare, generally non-serious, reversible cognitive effects such as memory loss and confusion, based on post-marketing reports [15]. Onset ranged from one day to years; resolution had a median of about three weeks. The agency's review did not suggest that statins cause fixed or progressive dementia, and the data did not suggest that cognitive change was common or clinically significant [15]. A meta-analysis of 25 randomized trials with 46,836 participants found no adverse effect on cognitive tests in cognitively normal participants (standardized mean difference 0.01, 95 percent CI −0.01 to 0.03) or in Alzheimer's disease (standardized mean difference −0.05, 95 percent CI −0.19 to 0.10) [12].

The 2026 label-wide meta-analysis went further. Of 66 further undesirable outcomes listed in statin labels, only four were significant after false-discovery-rate control: liver transaminases, other liver function abnormalities, urinary composition alteration, and oedema. Cognitive impairment, depression, sleep disturbance, and peripheral neuropathy were not supported [10]. STAREE, a double-blind trial of atorvastatin 40 milligrams in 9,971 adults aged 70 or older followed for a median of 5.9 years, found no benefit on the composite of death, dementia, or persistent disability (HR 0.94, 95 percent CI 0.84 to 1.05), but a 30 percent reduction in major cardiovascular events (HR 0.70, 95 percent CI 0.61 to 0.82); serious adverse events were equally common, though musculoskeletal, hepatobiliary, and diabetes-related events were more frequent with atorvastatin [13].

The honest formulation: statins do not commonly cause clinically significant cognitive impairment, and the trial evidence does not support a causal syndrome. That does not rule out a rare reversible idiosyncratic reaction. Rare reports leave room for a susceptible minority, but the base rate is low enough that a new cognitive complaint should prompt a search for other causes before the drug is blamed.

The costs of misattribution

The nocebo gap is not merely an academic curiosity. It has consequences at the population level.

A Danish nationwide cohort followed 674,900 people who started statins between 1995 and 2010 [14]. Early discontinuation rose from 6 percent in 1995 to 18 percent in 2010. Each negative statin-related news story was associated with a 9 percent increase in the odds of early discontinuation (odds ratio 1.09, 95 percent CI 1.06 to 1.12); positive stories were associated with an 8 percent decrease (odds ratio 0.92, 95 percent CI 0.90 to 0.94). People who discontinued early had higher rates of myocardial infarction (HR 1.26, 95 percent CI 1.21 to 1.30) and cardiovascular death (HR 1.18, 95 percent CI 1.14 to 1.23).

That study is observational, and its causal interpretation is limited. People who stop statins may be sicker, less engaged with preventive care, or different in ways the registers do not measure. The healthy-adherer bias is real: people who take preventive medicines consistently also tend to do other health-promoting things. Five studies within a 52-study review of more than four million participants reported all-cause mortality hazard ratios of 1.36 to 3.65 after discontinuation [17], but the same confounding applies. These studies establish a consistent association, not a precise causal effect.

Still, the direction is hard to dismiss. Statins prevent vascular events in people at meaningful risk. If a substantial share of discontinuation is driven by symptoms the drug did not cause, then the nocebo effect is not benign. It converts a manageable risk into an unmanaged one.

The label problem

The 2026 label-wide meta-analysis raises a policy question harder than the statistics. The authors argue that labels list adverse effects that blinded trials do not support, and that labels should be revised so patients and doctors can make better-informed decisions [10]. If 62 of 66 labeled outcomes show no significant excess in blinded trials, the label is not a neutral summary of evidence; it shapes belief.

But labels are not purely scientific documents. They are legal instruments, communication tools, and sometimes the basis for liability. Listing a rare reported effect can be a way of disclosing uncertainty rather than claiming proven causation. Removing it could reduce nocebo symptoms, but could also suppress reporting of genuine rare harms and undermine trust when a patient experiences something that used to be listed.

The better fix is hierarchy rather than deletion: distinguish established causal effects, uncertain signals, and rare post-marketing reports, with plain-language explanations. That preserves informed consent and evidential honesty. A rare true harm and a common nocebo symptom can produce the same sentence in a leaflet; only the study design can tell them apart. If public conversation treated that distinction as important, the 9 percent reporting rate and the much smaller causal rate would not feel like a contradiction.

What the trials cannot see

The blinded evidence is strong, but it has boundaries.

First, trials exclude people who cannot tolerate the drug. A run-in period or a clinician's decision not to enroll can remove the people who would have the most symptoms. The 3 to 5 percent with complete intolerance are underrepresented in the averages.

Second, adverse-event ascertainment varies. In the CTT meta-analysis, first-year muscle-symptom reporting ranged from 1.0 percent to 60.5 percent across trials, reflecting how actively symptoms were solicited [1]. The reassuring finding is that the rate ratio did not depend on the absolute reporting rate, but a trial that never asks about a symptom will not record it.

Third, follow-up is short relative to the diseases statins prevent. A median of 4.3 to 5.9 years is reasonable for muscle symptoms but not long enough to settle rare cognitive effects. STAREE's null dementia result is informative, not definitive.

Fourth, averages cannot tell an individual what caused a particular symptom. The right clinical tool is a blinded or structured n-of-1 trial, not a population estimate applied with certainty.

Finally, the nocebo estimate is design-dependent, and reporting bias cuts both ways: manufacturer-sponsored trials may under-report subjective symptoms, while independent investigators may select symptomatic patients. The CTT analysis, using individual participant data and prespecified outcomes, is the least vulnerable, but not immune.

How confident is the evidence?

Calibrated judgments, with reasoning:

• About 95 percent confident that, in blinded trials, moderate-intensity statins produce a small average excess of muscle symptoms, on the order of 11 extra events per 1,000 person-years in the first year and little or none afterward. This rests on a large individual-participant meta-analysis with consistent results across trials and a dose-response pattern.
• About 90 percent confident that more than half of real-world muscle-symptom reports attributed to statins are not caused by the drug molecule. The n-of-1 trials, the placebo-arm data in statin-intolerant patients, and the unblinded rechallenge data converge.
• About 85 percent confident that a genuine small causal muscle excess exists and is concentrated in the first year and at higher doses. STOMP's myalgia signal and the CTT dose-response support this, though the mechanisms are not fully characterized.
• About 85 percent confident that statins do not commonly cause clinically significant cognitive impairment. The RCT meta-analysis, the label-wide meta-analysis, and STAREE all point the same way. The residual uncertainty is about rare idiosyncratic reactions, not a common syndrome.
• About 80 percent confident that the population-level cardiovascular cost of nocebo-driven discontinuation is real and non-trivial. The Danish cohort and the discontinuation literature are consistent, but confounding by health behavior cannot be eliminated.
• About 60 percent confident that a substantial fraction of people diagnosed as statin-intolerant would not reproduce their symptoms under a properly blinded rechallenge. The trials are small and selected, and the estimate depends heavily on the definition of intolerance.

These reflect design strength, consistency, and known biases; they should move if better data arrive.

What would change the answer

A large blinded trial that systematically solicited symptoms and found a persistent excess after the first year would challenge the "early and small" model. The CTT analysis looked and found none, but a trial enriched for people with prior symptoms could detect a subgroup effect that averages hide.

A validated biomarker or genotype predicting statin-associated muscle symptoms would change the interpretation. If a subgroup has a measurable pharmacological vulnerability — for example, a transporter variant that raises muscle exposure — their symptoms would be drug-caused even if the population average is null. SLCO1B1 variants are plausible but not yet a routine diagnostic.

A mechanism explaining why symptoms appear in the first year and then vanish — adaptation, mitochondrial conditioning, or a transient immune response — would strengthen the causal reading of the early excess. Better prospective cognitive data, with objective testing before and after initiation, would do more than another meta-analysis of the same trials.

What follows for practice and policy

The implications are not "believe the trials and ignore the patient," nor "believe the patient and abandon the drug."

Take the symptom seriously. The fact that most reports are not drug-caused does not make the experience less real, and dismissing it damages trust. Explain the base rate: muscle aches are common, and the first year is when both drug effects and attention are highest.

Use a structured assessment. Check for alternative causes — thyroid disease, vitamin D deficiency, electrolyte abnormalities, inflammatory myopathy, exertion, and drug interactions. Consider a temporary discontinuation and a blinded or open rechallenge. The n-of-1 design is the most informative test, though not practical for every patient.

Consider dose and statin choice. High-intensity therapy carries a larger muscle signal. Switching statins, lowering the dose, or alternate-day dosing can preserve most of the benefit. Non-statin lipid-lowering drugs remain an option when statins truly cannot be used.

Do not miss the rare serious case. Severe pain, weakness, dark urine, or a marked creatine kinase rise requires prompt evaluation. The nocebo explanation applies to the typical ache, not every ache. For cognition, investigate sleep, mood, interactions, thyroid, B12, and vascular risk before stopping the statin.

For policy, make labels and public communication distinguish established from reported effects. That is harder than defending or deleting the current labels, but it is what the evidence implies.

Bottom line

Statins cause a small, mostly first-year excess of muscle pain, a modest increase in diabetes and liver enzymes, and rare serious muscle injury. They do not appear to commonly cause clinically significant cognitive impairment. Most people who report muscle symptoms on a statin are not experiencing a pharmacological effect of the drug; they are experiencing the real consequences of taking a tablet, expecting harm, and noticing ordinary symptoms. That distinction is not a reason to dismiss them. It is a reason to test the attribution rather than assume it, and to recognize that a culture that treats reported side effects as proven ones produces real cardiovascular harm.

References

1. Cholesterol Treatment Trialists' (CTT) Collaboration. Effect of statin therapy on muscle symptoms: an individual participant data meta-analysis of large-scale, randomised, double-blind trials. The Lancet. 2022;400(10355):832–845. doi:10.1016/S0140-6736(22)01545-8. https://doi.org/10.1016/S0140-6736(22)01545-8
2. Wood FA, Howard JP, Finegold JA, et al. N-of-1 trial of a statin, placebo, or no treatment to assess side effects. New England Journal of Medicine. 2020;383(22):2182–2184. doi:10.1056/NEJMc2031173. https://doi.org/10.1056/NEJMc2031173
3. Howard JP, Wood FA, Finegold JA, et al. Side effect patterns in a crossover trial of statin, placebo, and no treatment. Journal of the American College of Cardiology. 2021;78(12):1210–1222. doi:10.1016/j.jacc.2021.07.022. https://doi.org/10.1016/j.jacc.2021.07.022
4. Herrett E, Williamson E, Brack K, et al. Statin treatment and muscle symptoms: series of randomised, placebo controlled n-of-1 trials. BMJ. 2021;372:n135. doi:10.1136/bmj.n135. https://doi.org/10.1136/bmj.n135
5. Casula M, Gazzotti M, Bonaiti F, et al. Reported muscle symptoms during statin treatment amongst Italian dyslipidaemic patients in the real-life setting: the PROSISA Study. Journal of Internal Medicine. 2021;290(1):116–128. doi:10.1111/joim.13219. https://doi.org/10.1111/joim.13219
6. Bruckert E, Hayem G, Dejager S, Yau C, Bégaud B. Mild to moderate muscular symptoms with high-dosage statin therapy in hyperlipidemic patients — the PRIMO study. Cardiovascular Drugs and Therapy. 2005;19(6):403–414. doi:10.1007/s10557-005-5686-z. https://doi.org/10.1007/s10557-005-5686-z
7. Bytyçi I, Penson PE, Mikhailidis DP, et al. Prevalence of statin intolerance: a meta-analysis. European Heart Journal. 2022;43(34):3213–3223. doi:10.1093/eurheartj/ehac015. https://doi.org/10.1093/eurheartj/ehac015
8. Parker BA, Capizzi JA, Grimaldi AS, et al. Effect of statins on skeletal muscle function. Circulation. 2013;127(1):96–103. doi:10.1161/CIRCULATIONAHA.112.136101. https://doi.org/10.1161/CIRCULATIONAHA.112.136101
9. Penson PE, Bruckert E, Marais D, et al. Step-by-step diagnosis and management of the nocebo/drucebo effect in statin-associated muscle symptoms patients: a position paper from the International Lipid Expert Panel (ILEP). Journal of Cachexia, Sarcopenia and Muscle. 2022;13(3):1596–1622. doi:10.1002/jcsm.12960. https://doi.org/10.1002/jcsm.12960
10. Cholesterol Treatment Trialists' (CTT) Collaboration. Assessment of adverse effects attributed to statin therapy in product labels: a meta-analysis of double-blind randomised controlled trials. The Lancet. 2026;407(10529):689–703. doi:10.1016/S0140-6736(25)01578-8. https://doi.org/10.1016/S0140-6736(25)01578-8
11. Sattar N, Preiss D, Murray HM, et al. Statins and risk of incident diabetes: a collaborative meta-analysis of randomised statin trials. The Lancet. 2010;375(9716):735–742. doi:10.1016/S0140-6736(09)61965-6. https://doi.org/10.1016/S0140-6736(09)61965-6
12. Ott BR, Daiello LA, Dahabreh IJ, et al. Do statins impair cognition? A systematic review and meta-analysis of randomized controlled trials. Journal of General Internal Medicine. 2015;30(3):348–358. doi:10.1007/s11606-014-3115-3. https://doi.org/10.1007/s11606-014-3115-3
13. Zoungas S, Wolfe R, Moran C, et al. Atorvastatin, cardiovascular events, and disability-free survival in older adults. New England Journal of Medicine. 2026;395(13):1259–1272. doi:10.1056/NEJMoa2607314. https://doi.org/10.1056/NEJMoa2607314
14. Nielsen SF, Nordestgaard BG. Negative statin-related news stories decrease statin persistence and increase myocardial infarction and cardiovascular mortality: a nationwide prospective cohort study. European Heart Journal. 2016;37(11):908–916. doi:10.1093/eurheartj/ehv641. https://doi.org/10.1093/eurheartj/ehv641
15. U.S. Food and Drug Administration. FDA Drug Safety Communication: Important safety label changes to cholesterol-lowering statin drugs. 2012. Archived at https://web.archive.org/web/20210304043232/https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-important-safety-label-changes-cholesterol-lowering-statin-drugs
16. Louro R, Gouveia e Melo R, Ruivo J, Almeida AG, Pinto FJ, Caldeira D. Incidence of muscle symptoms in placebo arm among statin-intolerant patients: a systematic review with meta-analysis. Expert Review of Cardiovascular Therapy. 2023;21(11):887–894. doi:10.1080/14779072.2023.2274502. https://doi.org/10.1080/14779072.2023.2274502
17. Ageeb SA, Abdelmoghith A, ElGeed H, Awaisu A, ElMansor A, Owusu YB. Prevalence, associated risk factors, and adverse cardiovascular outcomes of statins discontinuation: a systematic review. Pharmacoepidemiology and Drug Safety. 2024;33(8):e5879. doi:10.1002/pds.5879. https://doi.org/10.1002/pds.5879