Statins in Absolute Terms: Mortality Benefit, Number Needed to Treat, and the Limits of Statin Therapy

Introduction: The Relative-Risk Problem

Statins are among the most extensively studied preventive medications in clinical medicine. Their central biological effect is LDL-C reduction, and larger absolute LDL-C reductions generally produce larger proportional reductions in major vascular events.^[1] That evidence does not mean that every person receives the same absolute benefit.

A 20% relative reduction can represent a major clinical gain in a patient whose untreated event risk is 25%, but a small gain in a patient whose untreated risk is 2%. The first patient may avoid about 5 events per 100 treated over the relevant period; the second may avoid about 0.4 events per 100. The relative effect can be similar while the NNT differs by more than an order of magnitude.

This is why statements such as “statins reduce cardiovascular risk by about one fifth” are incomplete. A useful treatment estimate must answer four questions: one fifth of what baseline risk, for which outcome, over what period, and compared with what alternative?

What NNT Measures, and What It Does Not

The number needed to treat (NNT) expresses an absolute treatment benefit: how many people would need to receive a treatment rather than a specified comparator for one fewer person, on average, to experience a particular adverse outcome over a defined period. Its interpretation requires the population, comparator, endpoint, and follow-up duration.

Relative Risk Reduction, Absolute Risk Reduction, and NNT

Relative risk reduction (RRR) describes the proportional decrease in risk compared with the control group. Absolute risk reduction (ARR) describes the difference between the groups’ risks. NNT is the reciprocal of ARR. These are complementary descriptions of the same treatment effect.

Consider a hypothetical five-year trial in which the outcome occurs in 10% of control participants and 8% of treated participants:

ARR must be expressed as a proportion when calculating NNT: 2 percentage points becomes 0.02, not 2. Noninteger NNTs are conventionally rounded upward.

A 20% relative reduction does not mean that 20 of every 100 treated people avoid an event. Here, the expected difference is 20 fewer people experiencing the outcome per 1,000 treated over five years: 100 with control versus 80 with treatment.

Baseline Risk Determines the Absolute Benefit

The same relative benefit can produce very different NNTs. Assuming a 20% RRR for the same endpoint over five years, reducing baseline risk from 2% to 1.6% produces an ARR of 0.4 percentage points and an NNT of 250. Reducing risk from 20% to 16% produces an ARR of 4 percentage points and an NNT of 25.

An NNT is therefore not a fixed property of a statin or the statin class. It depends on baseline cardiovascular risk, achieved LDL-C reduction, adherence, background therapy, comparator, endpoint, and duration. A trial comparing intensive with standard statin therapy estimates the incremental benefit of intensification, not the total benefit of statin treatment versus no statin.

What “One Additional Person Benefits” Means

An NNT of 50 describes an expected difference between groups. It does not guarantee exactly one prevented event in every group of 50, identify who benefits, or establish that the other 49 receive no benefit of any kind. Conversely, improvement in a laboratory value does not prove that everyone benefits clinically.

NNT also cannot distinguish permanent prevention from postponement beyond follow-up. An event absent during five years might occur later; NNT alone does not quantify event-free time or life-years gained.

The endpoint must remain attached to the number. An NNT of 78 for a composite cardiovascular outcome does not mean that treating 78 people prevents one death. The qualifying event could be a nonfatal myocardial infarction, stroke, revascularization, or another prespecified component. A first-event composite does not establish equal benefit across components or describe the total number of first and recurrent events prevented.

Duration and Statistical Uncertainty Matter

“NNT 50” is incomplete without specifying whether it applies to one, five, or ten years. NNTs from different follow-up periods should not be ranked directly or scaled linearly without supporting evidence.

For time-to-event trials, calculations should use appropriately estimated cumulative event risks at a specified time, accounting for censoring. A hazard ratio alone is insufficient to calculate NNT.

NNT is also an estimate with uncertainty. Because it is the reciprocal of ARR, small absolute differences can produce wide, asymmetric confidence intervals. For example, an ARR of 2 percentage points gives an NNT of 50; an ARR confidence interval of 0.5 to 3.5 percentage points corresponds to an NNT interval of approximately 29 to 200. When the ARR interval crosses zero, the NNT interval passes through infinity and includes possibilities of benefit and harm.

Primary Prevention in Absolute Terms

Primary prevention includes people without a prior myocardial infarction, ischemic stroke, symptomatic peripheral artery disease, or other established atherosclerotic cardiovascular disease (ASCVD).^[6,8] This is a heterogeneous population. A younger nonsmoker with mildly elevated LDL-C and normal blood pressure has a very different short-term risk from an older adult with diabetes, chronic kidney disease, hypertension, and coronary calcium.^[6,8]

Pooled primary-prevention outcomes

The evidence review supporting the 2022 US Preventive Services Task Force recommendation pooled randomized trials in adults at increased cardiovascular risk but without prior cardiovascular events.^[2] The absolute effects were consistently larger for nonfatal or composite cardiovascular outcomes than for mortality.^[2]

Outcome	Pooled relative risk	Absolute risk difference	Approximate NNT and follow-up
All-cause mortality	0.92	0.35% lower	286 over 1-6 years[2]
Fatal or nonfatal myocardial infarction	0.67	0.85% lower	118 over 2-6 years[2]
Fatal or nonfatal stroke	0.78	0.39% lower	256 over 1-6 years[2]
Composite cardiovascular outcome	0.72	1.28% lower	78 over 1-6 years[2]
Cardiovascular mortality	0.91; 95% CI crossed 1.0	0.13% lower	No reliable benefit NNT from the main RR analysis[2]

Table interpretation: The NNTs reproduce the pooled analysis, which used full-precision estimates; displayed absolute differences are rounded. Trial durations, populations, statin regimens, and composite definitions varied.^[2] These values should not be applied as individualized predictions.

The pooled all-cause mortality result was statistically significant in the main analysis, but it was small in absolute terms.^[2] In sensitivity analyses restricted to trials that were not stopped early or had at least 3 years of follow-up, the mortality estimate was attenuated and no longer statistically significant.^[2] The fairest summary is therefore not that primary-prevention statins “do not affect mortality,” nor that they produce a large survival gain. The evidence supports a modest pooled all-cause mortality benefit with uncertainty about its robustness across trial selections, while prevention of myocardial infarction and composite vascular events is more consistent.^[2]

Low-risk populations

The Cholesterol Treatment Trialists’ Collaboration examined people whose 5-year risk of a major vascular event was below 10%.^[7] For each 1.0 mmol/L reduction in LDL-C, statin therapy prevented approximately 11 major vascular events per 1,000 people over 5 years.^[7] That corresponds to an NNT of about 91 for one major vascular event, not one death.^[7]

This estimate illustrates two points. First, a proportional effect can remain clinically real at lower risk.^[7] Second, when baseline risk is low, many people must take treatment for several years for one person to avoid the selected event.^[7] Whether that tradeoff is worthwhile depends on treatment burden, safety, cost, competing priorities, and the patient’s values.

A trial-level example: HOPE-3

HOPE-3 enrolled 12,705 people at intermediate cardiovascular risk without known cardiovascular disease and compared rosuvastatin 10 mg daily with placebo for a median of 5.6 years.^[9] The first co-primary outcome occurred in 3.7% versus 4.8%, an ARR of 1.1 percentage points and an NNT of approximately 91.^[9] The broader co-primary outcome occurred in 4.4% versus 5.7%, an ARR of 1.3 percentage points and an NNT of approximately 77.^[9] These are composite-event NNTs, not mortality NNTs.^[9]

How current recommendations use absolute risk

The 2026 ACC/AHA multisociety dyslipidemia guideline replaced the 2018 cholesterol guideline and recommends the PREVENT-ASCVD equations for 10- and 30-year risk assessment in primary prevention among adults aged 30 to 79 years.^[6] It states that LDL-lowering therapy can be considered at a 10-year PREVENT-ASCVD risk of 3% to less than 5% and should be considered at 5% to less than 10% after clinician-patient discussion.^[6] Selective use of risk-enhancing factors and coronary artery calcium can refine uncertain decisions.^[6]

The 2022 USPSTF framework applies to adults aged 40 to 75 years with at least one major cardiovascular risk factor.^[8] It recommends initiating a statin when estimated 10-year CVD risk is at least 10% and selectively offering one when risk is 7.5% to less than 10%.^[8] It found insufficient evidence to recommend for or against initiating a statin for primary prevention at age 76 years or older.^[8]

These frameworks use different risk models, thresholds, and recommendation methods.^[6,8] Neither converts a risk estimate into a guaranteed personal outcome. Risk equations are decision aids, and the predicted endpoint may not exactly match the endpoint used in a trial-derived NNT.

Secondary Prevention: Why Absolute Benefit Is Usually Larger

Secondary prevention refers to patients with established ASCVD, such as prior myocardial infarction, ischemic stroke of atherosclerotic origin, symptomatic peripheral artery disease, or previous coronary revascularization.^[6] Their untreated recurrent-event risk is much higher than the first-event risk in most primary-prevention populations.^[1-7] A similar relative risk reduction therefore yields a larger ARR and a lower NNT.^[1,7]

Trial and population	Outcome and follow-up	Control vs statin	ARR and NNT
4S: established coronary heart disease with elevated cholesterol[3]	All-cause mortality; median 5.4 years	11.5% vs 8.2%	ARR 3.3%; NNT 31
LIPID: prior myocardial infarction or unstable angina[4]	All-cause mortality; mean 6.1 years	14.1% vs 11.0%	ARR 3.1%; NNT 33
Heart Protection Study: high-risk population, many with vascular disease or diabetes[5]	All-cause mortality; about 5 years	14.7% vs 12.9%	ARR 1.8%; NNT 56
Heart Protection Study: same population[5]	First major vascular event; about 5 years	25.2% vs 19.8%	ARR 5.4%; NNT 19

The 4S, LIPID, and Heart Protection Study results demonstrate why the phrase “statin NNT” is misleading. The mortality NNTs ranged from approximately 31 to 56, while the NNT for a first major vascular event in the Heart Protection Study was approximately 19.^[3-5] The endpoint matters as much as the population.

These landmark trials predated many elements of contemporary preventive care, so their historical control-event rates should not be treated as fixed current predictions.^[3-6] The trials remain powerful demonstrations of causality, but their NNTs are not timeless forecasts.

Current ACC/AHA guidance retains statin therapy as the foundation of secondary prevention and recommends LDL-C goals below 55 mg/dL for patients at very high risk and below 70 mg/dL for other patients with ASCVD, with nonstatin therapy added when maximally tolerated statin treatment does not achieve the relevant goal.^[6]

The Same Relative Effect Can Produce Very Different NNTs

The following simplified calculation applies the approximately 21% proportional reduction in major vascular events per 1.0 mmol/L LDL-C reduction reported in lower-risk trial data.^[7] It is a mathematical illustration, not a patient-specific prediction.

Untreated 5-year event risk	Expected treated risk at RR 0.79	ARR	Approximate NNT over 5 years
5%	3.95%	1.05%	96
10%	7.9%	2.1%	48
20%	15.8%	4.2%	24

The model assumes the same endpoint, follow-up, proportional effect, and 1.0 mmol/L LDL-C reduction.^[7] Real-world estimates vary because actual LDL-C lowering, adherence, competing events, baseline treatment, and risk-model calibration differ.

Intensification: Incremental Benefit Is Not the Same as Initial Benefit

A trial comparing high-intensity with lower-intensity statin therapy asks a different question from a placebo-controlled trial. Its NNT estimates the extra benefit of intensification on top of active treatment, not the total benefit of receiving a statin.

In TNT, 10,001 patients with stable coronary disease received atorvastatin 80 mg or 10 mg daily.^[10] Over a median of 4.9 years, the primary major cardiovascular outcome occurred in 8.7% versus 10.9%, an ARR of 2.2 percentage points and an NNT of approximately 46 for one additional major cardiovascular event prevented by the higher dose.^[10] All-cause mortality was not significantly reduced.^[10]

This distinction is clinically important. Escalating intensity can prevent additional nonfatal events without producing a detectable mortality difference within the trial’s duration.^[10] A nonsignificant mortality result does not erase the morbidity benefit, but it prevents clinicians from presenting the intensification NNT as a survival NNT.

Residual Risk: What Statins Cannot Do

Statins lower vascular risk; they do not eliminate atherosclerosis or all of its consequences.^[1,5,6] In the Heart Protection Study, 19.8% of statin-treated participants still experienced a first major vascular event over about 5 years.^[5] Residual risk reflects persistent lipid-related risk and nonlipid factors that require separate assessment and treatment.^[6]

For this reason, statin therapy should not displace blood pressure control, smoking cessation, physical activity, nutrition, diabetes treatment, weight management, antithrombotic therapy when indicated, or nonstatin LDL-lowering therapy when risk and achieved lipid levels justify it.^[6] The 2026 guideline’s return to LDL-C and non-HDL-C goals reflects the need to assess achieved response rather than treating a prescription as the endpoint of care.^[6]

Another limit is adherence. Trial efficacy assumes substantial exposure to assigned treatment. Missed doses, discontinuation, inadequate dose intensity, and a smaller-than-expected LDL-C response reduce realized benefit.^[1,6] An NNT calculated from a well-conducted trial can therefore be more favorable than an effectiveness estimate in a poorly adherent population.

Adults Aged 75 Years or Older: Treatment and Deprescribing Are Different Questions

Randomized evidence remains less direct for primary prevention after age 75 than for secondary prevention.^[11] An individual-participant meta-analysis of 28 trials found that statins or more intensive statin regimens reduced major vascular events by about 21% per 1.0 mmol/L LDL-C reduction across age groups.^[11] However, among participants older than 75, much of the evidence came from people with established vascular disease, and the primary-prevention evidence was less extensive.^[11]

The 2026 ACC/AHA guideline states that LDL-lowering pharmacotherapy can be considered after age 75 together with lifestyle intervention and individualized risk assessment.^[6] The decision should distinguish starting therapy, continuing a well-tolerated therapy, and deprescribing therapy in the setting of limited life expectancy, frailty, polypharmacy, or changed goals of care.^[6,12]

The 2026 SAGA/SITE trial

SAGA/SITE directly studied discontinuation rather than initiation.^[12] The pragmatic, open-label French trial enrolled approximately 1,160 adults aged 75 years or older who had been taking a statin for primary prevention and had no prior ASCVD.^[12] At 3 years, all-cause mortality was 7.2% among participants assigned to discontinue and 7.9% among those assigned to continue.^[12] Discontinuation met the prespecified noninferiority criterion, which allowed an absolute mortality difference of up to 5 percentage points.^[12] Major cardiovascular events and quality-of-life measures did not differ significantly, while LDL-C rose by about 50% after discontinuation.^[12]

The finding is clinically relevant but narrow. The trial was open-label, enrolled fewer participants than originally planned, observed lower mortality than anticipated, used a comparatively wide noninferiority margin, and was not powered to exclude important differences in individual cardiovascular outcomes.^[12] Follow-up was 3 years, participants had already tolerated statin therapy, and secondary-prevention patients were excluded.^[12]

SAGA/SITE therefore supports a documented, individualized deprescribing conversation for selected older adults receiving a statin solely for primary prevention.^[12] It does not establish that statins are ineffective after age 75, does not address long-term outcomes beyond 3 years, and does not justify routine discontinuation in patients with prior myocardial infarction, stroke, peripheral artery disease, or other ASCVD.^[6,11,12]

Harms in Absolute Terms

Benefit discussions are incomplete without harms, but harms also require an absolute denominator and a valid comparator.

Muscle symptoms

An individual-participant meta-analysis of large double-blind trials found a 7% relative increase in muscle pain or weakness during the first year of statin therapy, corresponding to approximately 11 excess reports per 1,000 person-years.^[13] There was little or no excess after the first year, and more than 90% of muscle symptoms reported by statin-assigned participants were not caused by the statin.^[13]

This evidence should not be used to dismiss symptoms. It means that muscle pain is common in the population and that temporal association alone does not establish causation.^[13] Clinicians should evaluate severity, timing, interacting medications, other medical causes, and the feasibility of rechallenge, dose adjustment, or a different statin.^[13] Severe weakness, marked creatine kinase elevation, or dark urine requires urgent evaluation because serious muscle injury, although uncommon, can occur.^[13]

New-onset diabetes and glycemia

A 2024 individual-participant meta-analysis found a dose-dependent increase in diabetes diagnoses.^[14] Compared with placebo, low- or moderate-intensity statin therapy increased new-onset diabetes by about 10% proportionally, with annual rates of 1.3% versus 1.2%.^[14] High-intensity therapy increased diagnoses by about 36% proportionally, with annual rates of 4.8% versus 3.5% in the contributing trials.^[14]

Those absolute rates should not be converted into a universal number needed to harm because baseline glycemia and the frequency of hemoglobin A1c testing differed across trials.^[14] Approximately 62% of statin-attributable new diabetes diagnoses occurred among participants in the highest quarter of baseline glycemia.^[14] In practice, the excess risk is concentrated near the diagnostic threshold and must be weighed against cardiovascular benefit, which is usually larger in patients whose metabolic risk also raises ASCVD risk.^[14]

Other label-attributed adverse effects

A 2026 individual-participant meta-analysis of 19 double-blind placebo-controlled trials evaluated adverse outcomes listed in statin product information.^[15] Beyond previously established muscle and glycemic effects, only four of 66 additional outcomes met the prespecified false-discovery criterion: elevated transaminases, other liver-test abnormalities, altered urine composition, and edema.^[15] The combined absolute annual excess of liver-test abnormalities was 0.13%.^[15]

The blinded trial data did not support causal relationships between statins and most other listed conditions, including cognitive impairment, depression, sleep disturbance, and peripheral neuropathy.^[15] This does not prove that every rare event is impossible. It indicates that large blinded randomized datasets do not show an excess for most commonly attributed outcomes.^[15]

Seven Limits of Statin Therapy That Should Be Stated Explicitly

1. A composite benefit is not a mortality benefit. The lowest NNT in a trial often belongs to a composite endpoint dominated by nonfatal events.^[2-5,9,10]
2. Relative benefit does not determine absolute benefit. Baseline risk is the main driver of ARR when proportional effects are similar.^[1,7]
3. Trial duration limits what can be claimed. A 5-year study cannot directly prove a 20-year mortality benefit, and a 3-year deprescribing study cannot exclude later divergence.^[2,12]
4. Historical NNTs are not automatically contemporary NNTs. Background therapy, diagnostic practices, and event rates change over time.^[3-6]
5. Statins leave residual risk. Many treated patients continue to experience vascular events and may need additional risk-factor treatment.^[5,6]
6. Tolerance and adherence modify effectiveness. A prescribed statin cannot produce full trial-level benefit when it is not taken or does not lower LDL-C as expected.^[1,6]
7. Statins do not replace individualized judgment. Frailty, life expectancy, polypharmacy, competing illness, reproductive considerations, patient goals, and prior ASCVD change the benefit-harm calculation.^[6,8,12]

A Practical Clinical Approach

1. Define the prevention category correctly

Prior ASCVD places the patient in secondary prevention even when the event occurred years ago and the patient is currently asymptomatic.^[6] The expected absolute benefit is generally larger than in primary prevention.^[1-7]

2. Specify the outcome and time horizon

Ask whether the discussion concerns death, myocardial infarction, stroke, revascularization, or a composite event. State the period over which the estimate applies. “NNT 50 for a major vascular event over 5 years” is clinically meaningful; “NNT 50” is not.

3. Estimate baseline risk with a current framework

For US primary prevention, the 2026 ACC/AHA guideline uses PREVENT-ASCVD in adults aged 30 to 79 years, followed by personalization with risk-enhancing factors and selective coronary calcium testing.^[6] The 2022 USPSTF thresholds remain a separate preventive-services framework for adults aged 40 to 75 years with one or more major risk factors.^[8]

4. Estimate the likely LDL-C reduction

The proportional reduction in major vascular events scales with the absolute LDL-C reduction achieved.^[1] A person whose LDL-C falls by 60 mg/dL should not be assumed to receive the same effect as one whose LDL-C falls by 15 mg/dL, even when both are described as “taking a statin.”

5. Convert risk to an approximate absolute benefit without false precision

A simplified estimate multiplies baseline event risk by the expected relative risk. The difference is the ARR, and 1 divided by ARR is the NNT. This calculation is only approximate because risk-score endpoints and trial endpoints may differ, proportional effects may not be identical, and adherence and competing risks are uncertain.

6. Discuss harms using absolute frequencies

Muscle symptoms attributable to statins are uncommon in blinded trials, diabetes risk is concentrated in susceptible patients and rises with intensity, and liver-test abnormalities are increased slightly.^[13-15] The discussion should neither minimize a patient’s symptoms nor present every temporally associated complaint as drug-caused.^[13-15]

7. Reassess response and residual risk

Confirm adherence, the achieved LDL-C reduction, tolerance, and whether the current LDL-C goal has been reached.^[6] In secondary prevention or high-risk primary prevention, consider additional LDL-lowering therapy when maximally tolerated statin treatment is insufficient.^[6] Continue to treat smoking, blood pressure, diabetes, inactivity, diet, and other modifiable risks.^[6]

Clinical Bottom Line

Clinical context	Typical absolute pattern	Interpretation
Low-risk primary prevention[2,7]	Small ARR over several years; higher NNT	Shared decision-making is central; event prevention is more evident than mortality benefit
Intermediate- or high-risk primary prevention[1,2,6,9]	Larger ARR than low-risk prevention	Benefit becomes more compelling as baseline risk and achieved LDL-C reduction rise
Established ASCVD[3-6]	Largest absolute event reduction; lower NNT	Statins are foundational, but residual risk often requires broader treatment
Higher- vs lower-intensity statin[10]	Incremental reduction in major events	The NNT describes added benefit, not the total benefit of statin therapy
Age 75 or older, primary prevention[6,11,12]	Evidence is less direct and competing risk matters	Do not use age alone to start, continue, or stop therapy; SAGA/SITE supports individualized deprescribing discussions

References

1. Cholesterol Treatment Trialists’ (CTT) Collaboration. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet. 2010;376(9753):1670-1681. DOI. PubMed.
2. Chou R, Cantor A, Dana T, Wagner J, Ahmed AY, Fu R, Ferencik M. Statin Use for the Primary Prevention of Cardiovascular Disease in Adults: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force. JAMA. 2022;328(8):754-771. DOI. PubMed.
3. Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet. 1994;344(8934):1383-1389. DOI. PubMed.
4. Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. N Engl J Med. 1998;339(19):1349-1357. DOI. PubMed.
5. Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20,536 high-risk individuals: a randomised placebo-controlled trial. Lancet. 2002;360(9326):7-22. DOI. PubMed.
6. Blumenthal RS, Morris PB, Gaudino M, et al. 2026 ACC/AHA/AACVPR/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline on the Management of Dyslipidemia: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2026;153(17):e1154-e1276. DOI. PubMed. AHA guideline resources. Accessed August 18, 2026.
7. Cholesterol Treatment Trialists’ (CTT) Collaborators. The effects of lowering LDL cholesterol with statin therapy in people at low risk of vascular disease: meta-analysis of individual data from 27 randomised trials. Lancet. 2012;380(9841):581-590. DOI. PubMed.
8. US Preventive Services Task Force. Statin Use for the Primary Prevention of Cardiovascular Disease in Adults: US Preventive Services Task Force Recommendation Statement. JAMA. 2022;328(8):746-753. DOI. PubMed. USPSTF recommendation. Accessed August 18, 2026.
9. Yusuf S, Bosch J, Dagenais G, et al; HOPE-3 Investigators. Cholesterol Lowering in Intermediate-Risk Persons without Cardiovascular Disease. N Engl J Med. 2016;374(21):2021-2031. DOI. PubMed.
10. LaRosa JC, Grundy SM, Waters DD, et al; Treating to New Targets Investigators. Intensive Lipid Lowering with Atorvastatin in Patients with Stable Coronary Disease. N Engl J Med. 2005;352(14):1425-1435. DOI. PubMed.
11. Cholesterol Treatment Trialists’ Collaboration. Efficacy and safety of statin therapy in older people: a meta-analysis of individual participant data from 28 randomised controlled trials. Lancet. 2019;393(10170):407-415. DOI. PubMed.
12. Bonnet F, et al. Discontinuation of statins for primary prevention of atherosclerotic cardiovascular disease in adults aged 75 years or older (SAGA/SITE): a multicentre, open-label, pragmatic, non-inferiority randomised trial. Lancet Healthy Longev. Published online August 11, 2026. DOI. PubMed.
13. Cholesterol Treatment Trialists’ Collaboration. Effect of statin therapy on muscle symptoms: an individual participant data meta-analysis of large-scale, randomised, double-blind trials. Lancet. 2022;400(10355):832-845. DOI. PubMed.
14. Cholesterol Treatment Trialists’ Collaboration. Effects of statin therapy on diagnoses of new-onset diabetes and worsening glycaemia in large-scale randomised blinded statin trials: an individual participant data meta-analysis. Lancet Diabetes Endocrinol. 2024;12(5):306-319. DOI. PubMed.
15. Reith C, Blackwell L, Emberson JR, et al; Cholesterol Treatment Trialists’ Collaboration. Assessment of adverse effects attributed to statin therapy in product labels: a meta-analysis of double-blind randomised controlled trials. Lancet. 2026;407(10529):689-703. DOI. PubMed.