Statins: the arguments you've heard, and what the evidence actually shows

Patients ask me about statins constantly, and the question has changed in the last few years. It used to be "is this drug right for me." Now it is "should I be on this drug at all." A patient joins a video visit with a printout, or a podcast clip queued on their phone, or a list of arguments they have collected from social media, and they want to know whether the foundation of preventive cardiology is sound.

The origin of the confusion

The first generation of statin trials, in the 1990s and early 2000s, was remarkable. The Scandinavian Simvastatin Survival Study (4S), CARE, LIPID, HPS, JUPITER. These were large, well-conducted, placebo-controlled trials, and they showed that lowering LDL with a statin reduced heart attacks and cardiovascular death. The effect was big enough that cardiology was reshaped around it. By the 2010s, statins had become the most-prescribed drug class in the United States.

What followed was a backlash, and the backlash had real points underneath it. Some of the early enthusiasm was overstated. Side effects, particularly muscle complaints, were under-acknowledged in early trial reporting. The threshold for prescribing crept downward in ways that put patients on statins who would not have qualified a decade earlier. Industry funding of guideline panels was a real issue and was, in some places, slow to be addressed. And the lipid hypothesis itself was sometimes communicated as if LDL were the only number that mattered, when in fact ApoB and Lp(a) and the rest of the lipid panel carry independent information.

The contrarian wave grabbed those legitimate criticisms and extrapolated them into a broader claim: that statins are overprescribed, that LDL does not cause heart disease, that the trials cannot be trusted, that lifestyle and metabolic health are the real story. Some of those arguments have a kernel of truth. Most do not survive contact with the actual trial data. The problem is that the trial data is not what social media is built to convey. Five sentences of confident skepticism reach a million views before a careful response can put on its shoes.

Three things that are settled

Separate the science into three layered claims. The first two are not really in dispute among cardiologists who have read the trials.

One: lower LDL means fewer heart attacks. This is the most consistent finding in cardiovascular medicine. The Cholesterol Treatment Trialists' Collaboration is a meta-analysis of every major statin trial ever conducted, pooled at the individual-patient level. The line is essentially linear: for every 1 mmol/L (about 39 mg/dL) of LDL reduction sustained over a few years, the risk of a major vascular event falls by about 22%. That dose-response holds across primary prevention, secondary prevention, high-risk patients, low-risk patients, men, women, diabetics, elderly. The slope is the slope.

Two: it is the LDL reduction that does the work, not the statin itself. This is the part that should settle the "is it really the LDL" debate, and it is the part many patients have not heard. Ezetimibe is not a statin; it works by blocking cholesterol absorption in the gut. PCSK9 inhibitors are not statins; they are monoclonal antibodies. Bempedoic acid is not a statin; it works upstream in cholesterol synthesis. Inclisiran is not a statin; it is a small interfering RNA. All four of these drugs lower LDL by different mechanisms, and all four have been tested in cardiovascular outcome trials (IMPROVE-IT, FOURIER, ODYSSEY OUTCOMES, CLEAR Outcomes). Every one of them lowered cardiovascular events proportionally to how much LDL they lowered. The trials slot directly onto the CTT line. It is the LDL, by whatever mechanism.

Three: cause is not confounded. The cleanest test of causality we have in any field of medicine is a Mendelian randomization study. People are born with gene variants that lifelong raise or lower their LDL. The variants are assigned at conception and cannot be confounded by diet, exercise, smoking, income, or the placebo effect. People with lifelong low-LDL variants have proportionally low rates of heart attacks. People with lifelong high-LDL variants (including familial hypercholesterolemia, the extreme version) have proportionally high rates of heart attacks. Multiple gene variants, replicated across populations, converge on the same answer. The "LDL causes atherosclerosis" claim is not a hypothesis. It is a finding.

The arguments with a kernel of truth

Most of the social-media arguments you have heard are not wrong because they are facile. They are wrong because they take a real piece of nuance and over-extend it.

"LDL is too simple. ApoB and particle count matter more." Half right, in a way that has already been incorporated into modern lipidology. ApoB is in fact a better marker than LDL-C, because it counts every atherogenic particle (LDL, IDL, VLDL remnants, Lp(a)) rather than estimating cholesterol content. Modern preventive cardiology orders ApoB routinely. The right conclusion is "measure ApoB, not LDL-C alone," not "lipids are irrelevant." The Mendelian randomization argument above runs through ApoB-containing lipoproteins specifically. Causality is intact.

"The NNT is bad in primary prevention." Real point, badly applied. In a 10-year low-risk patient with no other risk factors, the number-needed-to-treat to prevent one heart attack is genuinely large, and that is a fair argument against blanket statin use in young, low-risk people. The fix is risk stratification, which is exactly what current guidelines do. ASCVD risk calculators, coronary calcium scoring, family history, ApoB, Lp(a), and CT angiography are now standard tools for deciding which patient actually has enough atherosclerosis or enough risk to make the math work. The real answer to "is the NNT bad" is "for the wrong patient, yes. That is why we stopped treating everyone the same way."

"Side effects are under-reported." The historical version of this is true. The current version mostly is not. The SAMS-CoQ10 trial and the Wood et al. crossover trial (Lancet, 2020) put patients with reported statin intolerance on randomized blinded statin, placebo, and no-treatment periods. Most of the symptoms attributed to the statin appeared on placebo too. That is the classic nocebo signature: the expectation of side effects produces the side effects. Real myopathy exists, real rhabdomyolysis exists, and they are rare. Cognitive side effects have not held up in randomized trials. Type-2-diabetes risk is real, dose-dependent, modest, and concentrated in patients already on the path to diabetes; it does not erase the cardiovascular benefit. 1990s trials understated the nocebo problem; we understand it better now.

"Guidelines are industry-captured." Was a real problem; has been addressed. The Cholesterol Treatment Trialists' Collaboration is independent of industry funding. The 2018 and 2026 ACC/AHA guideline panels are required to disclose conflicts and recuse on relevant votes. Cynicism about the historical record is fair; extrapolating it forward to "therefore the current recommendations are corrupt" is not.

"Lean keto patients with very high LDL are different. They are safe." The data here is genuinely incomplete. The argument is that very-low-carb dieters who become lean and metabolically optimized but whose LDL rises substantially (sometimes to 250–400 mg/dL) are not at the same atherosclerotic risk as someone with LDL in that range from a Western diet. Imaging studies, including CT angiography and coronary calcium scoring, are starting to read out, and the early results are mixed. Some lean low-carb individuals with high LDL show progression of coronary plaque. Some show little. The Mendelian randomization argument still applies (high ApoB is a causal exposure regardless of why it is high), and the precautionary stance is still to lower LDL. The data is still coming in. That is the fairest read.

The modern toolkit floor

Lipid management in 2026 is not "everyone on rosuvastatin." It is a ladder. Lifestyle first, always: Mediterranean pattern, weight optimization, exercise, replacement of saturated fat with mono- and polyunsaturated fat (I wrote about that side of it in the saturated fat piece). For many patients that is enough, and we recheck a panel and move on. When lifestyle hits its floor, ezetimibe is often the next step, then bempedoic acid or a low-dose statin, then a full statin, then PCSK9 inhibitors or inclisiran for patients whose LDL or ApoB or Lp(a) still sits too high for their risk profile. The 2026 ACC/AHA guidelines formalize this ladder.

Each step has its own trial-grade evidence behind it. Each can be titrated. Side effects can be managed by switching agents, lowering doses, alternate-day dosing, and combination therapy that uses lower doses of each drug. A patient who could not tolerate atorvastatin 80 may do fine on rosuvastatin 5 plus ezetimibe. A patient who genuinely cannot tolerate any statin has bempedoic acid, ezetimibe, and PCSK9 inhibitors as real options.

The decision a cardiologist makes with you in 2026 is not "statin yes or no." It is "what is your absolute risk, what does your ApoB look like, what does your imaging show, and what is the best combination of tools to get you where you need to be." The modern toolkit is good enough that almost everyone can be brought to target with side effects that they can live with. The work is choosing which tools, in which order, at which doses.

My advice to patients

Get a current lipid panel that includes ApoB and Lp(a), not just LDL-C. Get a coronary calcium score if you are between 40 and 70 and uncertain about your risk. Bring those numbers, and a real read on your family history, to a cardiologist who will explain the ladder rather than push the first rung.

If your risk is genuinely low and your lifestyle is doing the work, you may not need any medication. If your risk is high and lifestyle has plateaued, the right question is not "will the cardiologist make me take a statin." It is "which combination of tools, and at what doses, gets my ApoB to where my arteries are actually safe."

Statins are one tool in a larger toolkit. The arguments against them on social media are mostly arguments against a 2005 version of the field that no longer exists. The arguments that are still live are the narrow ones: lean low-carb-high-LDL imaging data, fine-grained risk stratification at the population edges, individual response variation. Those are the ones worth talking through with someone who has read the trials.

You should not feel cornered by the social-media debate. The trials are clearer than the discourse around them, and the toolkit is bigger than it has ever been.