Statins are overprescribed. The absolute risk numbers explain why.

The one number that changes the conversation

In a 2022 meta-analysis published in JAMA Internal Medicine, Byrne and colleagues went back through 21 statin trials and did something that the original press releases mostly did not: they reported absolute risk reductions alongside the familiar relative ones. The relative-risk picture was the one the statin literature has reported for thirty years — roughly 25-30% reductions in major cardiovascular events. The absolute-risk picture was much smaller. For all-cause mortality, the absolute reduction was about 0.8%. For myocardial infarction, 1.3%. For stroke, 0.4%. Over approximately four to five years of treatment.

Those are the headline numbers from the recalculation. They are not anti-statin numbers. They are pro-statin numbers — for the patients who actually need them. They are also, for low-risk primary-prevention patients, the numbers that make the argument that we are prescribing this drug to a population whose individual probability of benefit is much smaller than the public conversation implies.

Both things can be true. The drug works. The drug is also being prescribed to people whose absolute benefit is, statistically, near the noise floor. The honest read is that secondary prevention is one of cardiology's strongest evidence bases, and primary prevention in low-risk patients is one of cardiology's softest. Most of the people currently on statins in North America are in the second group.

Relative risk reduction versus absolute risk reduction

The distinction matters enough to spell out. Imagine a five-year trial of a drug. In the placebo group, 4 out of 100 people have a heart attack. In the drug group, 3 out of 100 do. The relative risk reduction is 25% (from 4 down to 3 is a 25% drop). The absolute risk reduction is 1 percentage point (from 4% to 3%). The number needed to treat (NNT) — how many patients you have to put on the drug for five years to prevent one heart attack — is 100.

Both numbers describe the same trial. They communicate very different things to a patient. "This drug cuts your heart attack risk by 25%" sounds enormous. "This drug means that if 100 people like you take it for five years, one of them avoids a heart attack who otherwise would have had one" sounds like a real decision with real trade-offs.

The cardiology literature has historically led with the first framing. The patient-decision literature increasingly argues for the second. The first is technically correct. The second is the one a patient needs to make an informed choice about whether to be on the medication for the rest of their life.

The 2022 JAMA Internal Medicine recalculation

The Byrne et al. 2022 meta-analysis pulled together 21 randomized statin trials covering both primary and secondary prevention. Their headline finding — quoted from the published abstract numbers — was that pooled absolute risk reductions were 0.8% for all-cause mortality, 1.3% for myocardial infarction, and 0.4% for stroke. Pooled relative risk reductions, for the same outcomes, were 9%, 29%, and 14%.

The authors made two further points. First, the relationship between LDL-C reduction and absolute clinical benefit was inconsistent across trials, which complicates the cleanest version of the "lower LDL equals proportional benefit" framing. Second, the magnitude of benefit varied substantially by baseline cardiovascular risk — exactly the pattern you would expect if the drug works better in patients with more risk to reduce.

That second point is the bridge to the prescribing critique. The drug works best in high-risk patients. The drug works least in low-risk patients. The current prescribing guidelines have lowered the threshold for statin initiation in primary prevention to the point where a substantial share of new prescriptions are to patients in the low-benefit zone. The absolute risk reduction for those patients is real but small. The lifetime cost — including side effect exposure, daily medication adherence, and the framing of one's own health as drug-dependent — is the same as for the high-benefit patients.

The drug works. The drug is also being prescribed to people whose absolute benefit is, statistically, near the noise floor. Both things are true. That is the part the press releases never spell out.

Secondary prevention: where the evidence is strong

Secondary prevention means using a statin in a patient who has already had a cardiovascular event — a heart attack, a stroke, a coronary revascularization, documented atherosclerotic coronary or peripheral arterial disease. In this population, the baseline event rate is high, the risk of a second event is elevated for years afterward, and the absolute benefit of statin therapy is correspondingly larger.

The Cholesterol Treatment Trialists' (CTT) meta-analyses have shown that in secondary prevention, intensive statin therapy produces meaningful absolute reductions in major vascular events. The number needed to treat to prevent one major event over five years is in the range of 20-25 for patients with established disease, versus more than 100 for low-risk primary prevention patients. The benefit-risk math in secondary prevention is not close — it is a strongly favorable trade.

This is the population where the "statins save lives" framing is most defensible. If you have had a heart attack, the cardiology consensus that you should be on a high-intensity statin is not a marketing position. It is supported by the absolute risk data, by the mechanistic plausibility, and by the consistent pattern across two decades of trials. The argument in this article is not about this group.

Primary prevention: where the math gets uncomfortable

Primary prevention means using a statin in a patient who has not yet had a cardiovascular event. The 2018 ACC/AHA cholesterol guideline and the 2022 USPSTF update both moved toward broader use of statins in primary prevention based on 10-year ASCVD risk calculations. The USPSTF currently recommends statin initiation in adults aged 40-75 with one or more cardiovascular risk factors and a 10-year ASCVD risk of 10% or greater, with selective use in the 7.5-10% range.

Here is where the absolute risk math matters most. A patient with a 10-year ASCVD risk of 10% has, by definition, a 90% probability of not having a cardiovascular event over those ten years even without treatment. Statin therapy reduces relative risk by roughly 25%. Applied to that 10% baseline, the absolute reduction is approximately 2.5 percentage points over ten years — meaning the patient's ten-year cardiovascular event probability moves from about 10% to about 7.5%.

That is a real benefit. It is also a smaller absolute benefit than the press-release relative-risk framing implies. The number needed to treat in this risk stratum is roughly 40 patients on statin therapy for ten years to prevent one cardiovascular event. The other 39 patients take the drug daily for a decade and do not derive a measurable clinical benefit from it.

For patients with a 10-year risk in the 5-7.5% range — where some prescribing decisions still happen, especially in patients with elevated LDL-C — the math gets harder. Absolute risk reduction in this stratum falls below 1.5 percentage points over ten years. NNTs climb above 65. The drug is still doing what the trials showed it does. Individual patients in this group are unlikely to be the ones who derive clinical benefit from being on it.

The side-effect conversation, honestly

The statin side-effect conversation has been distorted from both directions, and both distortions are real. The first is the under-reporting in the trial literature. The original statin trials were largely industry-sponsored, had run-in periods that excluded intolerant patients before randomization, and used adverse-event definitions that captured serious events well but were less sensitive to the muscle aches, cognitive complaints, and fatigue that show up in clinical practice. The trial-reported side effect rates underestimate the real-world experience.

The second distortion is in the other direction. The N-of-1 placebo-controlled crossover trials — patients on statin, then matched placebo, then back to statin, blinded — have consistently shown that a substantial fraction of patient-attributed statin side effects are not specifically caused by the drug. The SAMSON trial in 2020 was the cleanest of these: 90% of the symptom burden patients attributed to statins also appeared when they took identical placebo capsules. The drug causes real side effects in some patients. It is also blamed for symptoms it did not cause in many others.

The honest synthesis is that statins have a real but modest side-effect burden, that the burden is over-attributed in some patients and under-attributed in others, and that the most useful tool for any individual patient is a structured trial — on the drug, off the drug, with a careful symptom log — rather than either the trial-data dismissal or the internet-driven panic. For a patient where the absolute benefit is small to begin with, even a small side-effect burden tilts the math.

Who should be on a statin — and who probably should not

The patients where the absolute-risk math strongly favors statin therapy: anyone with established atherosclerotic cardiovascular disease, anyone with familial hypercholesterolemia, most adults with diabetes who fall in the moderate-to-high ASCVD risk range, and primary-prevention patients with 10-year ASCVD risk above approximately 15-20%. In these groups, the absolute benefit is large enough that the prescribing decision is not a close call.

The patients where the math is genuinely uncomfortable: primary-prevention patients with 10-year ASCVD risk in the 5-10% range, especially if the dominant driver is age or single elevated LDL-C without other risk factors. This is the group the current guidelines have moved most aggressively into statin therapy, and it is also the group where individual patient absolute benefit is smallest and where the side-effect and lifelong-medication trade-off deserves more weight in shared decision-making than it typically gets.

The piece that is almost always missing from the prescribing conversation in this lower-risk band is the absolute risk number itself. "Your 10-year cardiovascular risk is 8%. A statin would reduce that to about 6%. The number needed to treat for ten years to prevent one event in patients like you is approximately 50. Here are the things we know and do not know about side effects. What would you like to do?" That is the conversation the absolute-risk data supports. It is not the conversation most primary-prevention patients are having.

What this article is not saying

This is not "statins are bad." Statins work. The mechanism is real, the trial data is real, and for the right patients the benefit-risk math is a strongly favorable trade. The argument here is about prescribing thresholds and the communication of risk, not about whether the drug class works.

This is not "go off your statin." Anyone reading this who is currently on a statin should keep taking it until they have a structured conversation with the cardiologist who knows their actual risk profile. Discontinuation after a cardiovascular event has documented harm. Discontinuation in low-risk primary prevention is a reasonable conversation, but it is a conversation with a clinician, not a decision based on an article.

This is not anti-cholesterol-management. LDL-C is one of the most validated risk factors in modern cardiology, and lowering it in high-risk patients has clear benefit. The question is whether pharmacologic lowering in low-risk patients delivers enough absolute benefit to justify a lifetime medication commitment, when lifestyle changes, when made consistently, can produce similar absolute risk reductions without the medication exposure.

Finally, this is not a recommendation to ignore your cardiologist. It is a recommendation to ask one question the literature now supports asking: "What is my actual absolute risk reduction from this drug, in numbers, given my specific risk profile?" If your cardiologist can answer that question precisely, the prescribing decision is being made well. If the answer is some version of "statins reduce heart attacks by 25%," the conversation has not yet covered the part that matters most for you.