Do Statins Help People Who Have Never Had a Heart Attack?
Yes, but the size of the benefit is easy to overstate and worth seeing in per-year terms. Across the trials, about 1 in 250 avoided a death, 1 in 123 avoided a heart attack and 1 in 263 avoided a stroke. Those trials ran anywhere from six months to six years. Spread across a five-year course, the heart attack benefit works out to roughly 1 person in 600 per year of treatment. And those figures come from a population already at increased risk that was screened for statin tolerance before randomization, so a genuinely low-risk person should expect less than that.
Dr. Kumar’s Take
The honest headline is that statins work in primary prevention, with a caveat that belongs attached to that sentence every time it is said. The US Preventive Services Task Force pooled 19 trials covering 71,344 participants and found lower all-cause mortality, lower cardiovascular mortality, fewer strokes and fewer heart attacks. Anyone telling you statins do nothing for people without heart disease is arguing against a large and consistent body of randomized evidence.
The caveat: that benefit was measured in people already at increased cardiovascular risk, in trials that had screened out those who could not tolerate the drug, and it amounts to roughly one heart attack avoided per 600 people per year of treatment.
The size of that benefit is where most reporting goes wrong, in both directions. Statin trials are usually described in relative terms, and relative reductions sound far larger than what any individual gets. A third fewer heart attacks is a real finding, but in these trials it meant the heart attack rate fell by 0.81 percentage points. Put another way, about 123 people took a statin for roughly five years so that one of them avoided a heart attack. I use the absolute numbers throughout this piece for that reason.
The interesting part is the sentence most summaries skip. The USPSTF review states that absolute benefits were higher in subgroups at higher baseline risk, and its conclusion is explicitly about adults at increased cardiovascular risk without prior events. So the pooled absolute numbers describe a population already carrying elevated risk. Relative reductions hold up across subgroups, but a 36% reduction applied to a small starting risk yields a small absolute gain. That is arithmetic, not a criticism of the drug, and it is why the same evidence supports a confident yes for one patient and a genuine discussion for another.
The harms side is where I part company with the trial data, and for a structural reason rather than a hunch. Many statin trials used a run-in period: everyone took the drug for several weeks before randomization, and people who developed side effects or did not stick with it were dropped before the trial proper began. That design produces a randomized population pre-selected for tolerating statins. Whatever it tells you about efficacy, it cannot tell you the side effect rate in a general population, because the people most prone to side effects were removed first. The practice is documented well enough to have been studied in its own right.
The funding pattern compounds it. The great majority of statin trials were sponsored by the companies selling the drug, and industry-sponsored trials report favorable results more often than independently funded ones. That does not make the findings wrong. It does mean the reported benefits are best read as a best case and the reported harms as a floor.
Then there is what the trials were not designed to catch. Statins work by inhibiting HMG-CoA reductase, and that enzyme sits upstream of more than cholesterol. The same mevalonate pathway produces coenzyme Q10, which mitochondria need, and menaquinone-4, the tissue form of vitamin K2. Blocking the top of a pathway to lower one product also lowers the others. That is pharmacology, not speculation, and it gives a plausible mechanism for the muscle symptoms, the fatigue and the metabolic complaints patients actually report. What it does not yet have is outcome-trial confirmation, and I will not pretend otherwise.
Where this leaves me in clinic: if you have established cardiovascular disease or high calculated risk, this is straightforward and you should be on one. If your ten-year risk is genuinely low, you are being offered something like a 1-in-123 chance of avoiding a heart attack over five years, measured in a population already at higher risk than you and screened for tolerance before the trial started. Against that sit diabetes risk, muscle symptoms and a pathway argument that has not been settled. I do not think a number that small automatically wins that trade, and I think the patient should be the one making it.
Key Takeaways
- ✔ About 1 in 250 avoided a death (absolute difference 0.40 percentage points)
- ✔ About 1 in 123 avoided a heart attack (absolute difference 0.81 percentage points)
- ✔ About 1 in 263 avoided a stroke (absolute difference 0.38 percentage points)
- ✔ About 1 in 72 avoided any major cardiovascular event (absolute difference 1.39 points)
- ✔ Per year of treatment that is roughly 1 in 600 for a heart attack, 1 in 1,250 for death
- ✔ Absolute benefit rises with baseline risk, so these figures do not transfer to low-risk people
- ✔ Trial harm rates are probably an undercount, because run-in periods removed people who could not tolerate the drug before randomization
What the Benefit Actually Is
The USPSTF evidence review pooled 19 trials with 71,344 participants, mean ages 51 to 66. Against placebo or no statin:
- Death from any cause: 0.40 percentage points fewer, about 1 in 250 (relative reduction 14%, RR 0.86, 95% CI 0.80 to 0.93)
- Cardiovascular death: 0.43 points fewer, about 1 in 233 (relative reduction 31%, RR 0.69, 95% CI 0.54 to 0.88)
- Heart attack: 0.81 points fewer, about 1 in 123 (relative reduction 36%, RR 0.64, 95% CI 0.57 to 0.71)
- Stroke: 0.38 points fewer, about 1 in 263 (relative reduction 29%, RR 0.71, 95% CI 0.62 to 0.82)
- Any major cardiovascular event: 1.39 points fewer, about 1 in 72 (relative reduction 30%, RR 0.70, 95% CI 0.63 to 0.78)
Read the left-hand numbers first. The relative reductions are in brackets because they are what gets quoted, not because they are what you should weigh.
Per year of treatment. The review does not state a single follow-up length, and the trials in the updated version ran from six months to six years. Taking five years as a working figure, the numbers above come out at roughly 1 in 1,250 per year for death, 1 in 600 per year for a heart attack, 1 in 1,300 per year for stroke, and 1 in 360 per year for any major cardiovascular event. That arithmetic assumes risk is spread evenly across the years, which is an approximation rather than a finding. I include it because a yearly figure is closer to how anyone actually experiences taking a daily pill, and because it makes the scale of the benefit legible in a way a five-year total does not.
The Cholesterol Treatment Trialists Collaboration, pooling individual data from 27 trials, found the same direction in participants without a history of vascular disease. Its figures are proportional rather than absolute, at 15% lower vascular mortality and 9% lower all-cause mortality per 1.0 mmol/L of LDL reduction, with the proportional reductions similar across baseline risk categories. What that converts to for any individual again depends on where their risk starts.
An older meta-analysis of 11 trials covering 65,229 participants and 244,000 person-years reported no mortality reduction in high-risk primary prevention. It is the smaller and earlier analysis, and the two larger syntheses that came after it point the other way.
Why the Reported Harms Are Probably Too Low
Three features of how statin trials were built push the harm figures down, and they need stating before anyone quotes a reassuring safety number.
Run-in periods removed the intolerant. In many statin trials, participants took the drug for a period before randomization, and those who developed side effects or stopped taking it were excluded before the trial began. The population that then got randomized was, by construction, one that had already demonstrated it tolerated statins. This design has been examined specifically for its effect on statin efficacy and safety estimates. It is the most likely reason pooled trials show no excess of myalgias while clinical practice and dedicated trials find muscle symptoms in roughly one in twenty.
Industry funding is close to universal. Most statin trials were sponsored by manufacturers, and industry-sponsored trials report favorable results for the sponsor’s drug more often than independently funded ones do. Raw trial data have also largely remained with the sponsors rather than being released for independent reanalysis.
Some harms were never systematically collected. Trials tend to record the quantifiable and serious, such as rhabdomyolysis and liver enzymes. Fatigue, brain fog and mild but persistent muscle aching are the complaints that bring patients to my office, and they are the ones least likely to appear in a trial’s adverse event table unless someone thought to ask.
The Pathway Argument
Statins inhibit HMG-CoA reductase, the rate-limiting enzyme of the mevalonate pathway. That pathway does not exist to make cholesterol alone. It also produces coenzyme Q10, which mitochondria require for electron transport, and menaquinone-4, the tissue form of vitamin K2 involved in directing calcium into bone rather than arteries.
Inhibiting the top of a shared pathway to reduce one output reduces the others. That is the mechanism, and it is not controversial as chemistry. What remains unsettled is how much it matters clinically. The most direct evidence comes from supplementation trials: a meta-analysis of 12 randomized trials covering 575 patients found that coenzyme Q10 supplementation improved statin-associated muscle symptoms, which is at least consistent with depletion contributing to them.
I would apply the same skepticism here that I applied to the benefit data. A plausible mechanism plus a supplementation signal is a reason to take patient complaints seriously and to investigate further. It is not proof that statins cause cardiomyopathy or cognitive impairment. On cognition specifically, a large analysis in adults 65 and older found statin use was not associated with incident dementia, mild cognitive impairment, or decline in cognitive domains, and I am not going to leave that out because it cuts the other way.
Is a Benefit That Small Worth the Metabolic Cost?
This is the question the guidelines do not ask, and I think it is the right one for a low-risk patient.
On one side: roughly one heart attack avoided per 600 people per year of treatment, measured in a population at higher risk than the person asking, and in trials that had already removed those who could not tolerate the drug.
On the other side: a 9% relative increase in incident diabetes in the 13-trial meta-analysis, rising to 25% in one high-intensity trial. Muscle symptoms in roughly one in twenty when a trial goes looking for them. A mevalonate pathway that supplies coenzyme Q10 and vitamin K2 as well as cholesterol, inhibited daily and indefinitely. Fatigue and cognitive complaints that trials were not built to capture.
Diabetes is not a minor trade. It carries its own cardiovascular risk, which means a drug prescribed to lower cardiovascular risk can give some of that benefit back through a different door. Nobody has run the trial that nets those two against each other in a low-risk population, and until someone does, the honest answer is that the trade is unquantified rather than settled.
For a patient with established disease, none of this changes the recommendation; the benefit is large enough to absorb the cost. For a healthy patient with a low ten-year risk, I do not think a 1-in-600-per-year benefit automatically outweighs a daily intervention in a pathway that does several jobs. That is a judgement, and it is theirs to make with the numbers in front of them.
Why Baseline Risk Changes the Answer
Relative benefit is roughly constant. Absolute benefit is not. The USPSTF review says so directly: relative benefits appeared consistent across demographic and clinical subgroups, including people without marked hyperlipidemia, while absolute benefits were higher in subgroups at greater baseline risk.
Work it through. A third fewer heart attacks sounds decisive until you attach it to a starting number. Applied to someone with a 20% ten-year risk it prevents a lot of events. Applied to someone with a 2% ten-year risk it prevents very few, while the pill, the cost, the monitoring and the side effect risk are unchanged. This is why guidelines key treatment to calculated risk rather than to a cholesterol value, and why “should I take a statin” has no single answer.
It also means the pooled absolute figures above should not be read as your personal odds. The trials enrolled people already at increased risk, which is the population the USPSTF conclusion names. For someone genuinely low risk, the true absolute benefit sits below these numbers.
The Harms, Read From the Trials
This is where the popular account and the randomized evidence diverge most.
In the USPSTF pooled analysis, statins were not associated with increased serious adverse events (RR 0.99, 95% CI 0.94 to 1.04), not associated with increased myalgias (RR 0.96, 95% CI 0.79 to 1.16), and not associated with increased liver harms (RR 1.10, 95% CI 0.90 to 1.35).
Diabetes is genuinely contested. The USPSTF pooled analysis found no significant increase (RR 1.05, 95% CI 0.91 to 1.20), though with meaningful heterogeneity between trials, and one trial found high-intensity statins associated with 25% higher risk (RR 1.25, 95% CI 1.05 to 1.49). A separate collaborative meta-analysis of 13 trials covering 91,140 participants found a 9% relative increase (OR 1.09, 95% CI 1.02 to 1.17), which worked out to one extra case of diabetes for every 255 patients treated for four years.
Muscle symptoms are contested in the other direction. Randomized trials show no excess of myalgias, but the STOMP trial, designed specifically to measure statin effects on skeletal muscle, found myalgia in 4.8% of participants and commonly raised creatine kinase. STOMP ran six months.
Observational data adds a different picture again. A cohort of 2,004,692 patients in England and Wales found statin use associated with increased risks of moderate or serious liver dysfunction, acute renal failure, moderate or serious myopathy, and cataract, highest in the first year, with cataract risk returning to normal within a year of stopping. Observational data cannot separate the drug from the people who take it, so I weigh it below the randomized comparisons.
What to Do With This
- Ask for your ten-year risk, not just your cholesterol: the calculated risk is what determines how much a statin can do for you
- Ask for absolute numbers: a 36% reduction means something different at 20% baseline risk than at 2%
- Do not stop a statin on your own: if you have established heart disease or high risk, the evidence for staying on it is strong
- Report muscle symptoms rather than quitting silently: they are usually reversible, and switching agents often resolves them
- Raise diabetes risk if you already have risk factors for it: the signal is strongest with high-intensity statins
- Treat lifestyle as additive, not alternative: diet, exercise and not smoking act on the same outcomes and combine with treatment
Limitations To Keep In Mind
The trials pooled by the USPSTF enrolled adults at increased cardiovascular risk, so they do not directly answer the question for people at genuinely low risk, and the absolute figures overstate what a low-risk person should expect. Trial follow-up is measured in years while statin therapy is often lifelong, so both benefits and harms beyond that window are extrapolation. Run-in periods and near-universal industry sponsorship both push reported harms down and reported benefits up, which is why I treat the safety figures as a floor rather than an estimate. The pathway argument for coenzyme Q10 and vitamin K2 depletion is mechanistically sound but has not been tested against hard outcomes. And the cognitive question, where I would have expected the mechanism to show up, has a large null result against it.
Related Studies and Research
Statins and Primary Prevention: The Debate: the controversy surrounding statins for primary prevention in low-risk populations.
Rosuvastatin, CRP, and the JUPITER Trial: the JUPITER findings on rosuvastatin and C-reactive protein in primary prevention.
Statins and Neuromuscular Side Effects: potential neuromuscular complications associated with statin use.
HOPE-3 Trial: Rosuvastatin in Primary Prevention: results in people at intermediate cardiovascular risk.
Frequently Asked Questions
Do statins help if I have never had a heart attack?
Yes, with a caveat. In 19 pooled trials covering 71,344 participants without prior cardiovascular disease, about 1 in 250 avoided a death, 1 in 123 avoided a heart attack and 1 in 263 avoided a stroke. Spread over five years that is roughly 1 in 600 per year for a heart attack, and it was measured in people already at increased risk.
Then why do people say statins do not work for healthy people?
Because absolute benefit depends on baseline risk. The trials enrolled people already at increased risk, and the USPSTF states that absolute benefits were higher in higher-risk subgroups. For someone at genuinely low risk, the same proportional reduction produces a much smaller absolute gain than the 1-in-123 figure above.
Do statins cause muscle pain?
Randomized trials show no excess of myalgias compared with placebo (RR 0.96). A trial designed specifically to measure muscle effects found myalgia in 4.8% over six months. Report symptoms to your physician; they are usually reversible.
Do statins cause diabetes?
The evidence is mixed. Pooled primary-prevention trials found no significant increase (RR 1.05), while a 13-trial meta-analysis found a 9% relative increase, about one extra case per 255 people treated for four years. High-intensity statins carry the clearest signal.
Should I take one?
That depends on your calculated ten-year risk. High risk makes it straightforward. Low risk makes it a decision worth making with your physician using your own numbers.
Conclusion
Statins reduce death and cardiovascular events in people who have never had a heart attack. Stated in absolute terms, roughly 1 in 250 avoided a death and 1 in 123 avoided a heart attack over about five years, in trials that enrolled people already at increased risk. For someone genuinely low risk the true figure is smaller than that.
Against a benefit that size, the harms deserve more weight than the trial tables give them. Run-in periods removed people who could not tolerate the drug before randomization, most trials were funded by the manufacturers, and the symptoms patients actually report were often not collected systematically. The mevalonate pathway that statins block also produces coenzyme Q10 and vitamin K2, which is a real mechanism for the metabolic and muscular complaints even though it has not been tested against hard outcomes.
So the honest answer depends on where you start. High risk makes this straightforward and I recommend it without hesitation. Low risk makes it a genuine trade between a small, well-measured benefit and a set of harms that are probably underestimated, and that is a decision for the patient rather than a guideline.
Read the full study here
Sources: Chou et al, USPSTF evidence report, JAMA 2016;316:2008-2024 (PMID 27838722) · Cholesterol Treatment Trialists Collaboration, Lancet 2012;380:581-590 (PMID 22607822) · Ray et al, Arch Intern Med 2010;170:1024-1031 (PMID 20585067) · Parker et al, STOMP trial, Circulation 2013;127:96-103 (PMID 23183941) · Sattar et al, Lancet 2010;375:735-742 (PMID 20167359) · Hippisley-Cox & Coupland, BMJ 2010;340:c2197 (PMID 20488911) · Redberg & Katz, JAMA 2016;316:1979-1981 (PMID 27838702) · Fralick et al, run-in studies and statin trials, J Gen Intern Med 2018;33:792-794 (PMID 29520746) · Qu et al, CoQ10 and statin myopathy, J Am Heart Assoc 2018;7:e009835 (PMID 30371340) · Zhou et al, statins and cognitive decline in older adults, J Am Coll Cardiol 2021;77:3145-3156 (PMID 34167639)
