Home Grand Rounds Cholesterol, Statins, and What the Numbers Actually Say

Cholesterol, Statins, and What the Numbers Actually Say

Sep 25, 2026

143

Introduction

I recently had a medical check-up wherein my doctor strongly urged me to take a statin drug because my 167 milligrams per deciliter low-density lipoprotein cholesterol (LDL) meant I might suffer a heart attack any moment. With all due respect, I courteously declined.

“Well, I’m going to have to record that in your record,” he scowled.

“Fine. Of course,” I replied. I was unconcerned for reasons shared below. As a retired oral surgeon and medical professional educator I added, “I don’t believe in taking statin drugs in my case.”

The doctor arrogantly replied, “That’s your problem. You have a belief. You don’t have knowledge of the truth. I don’t have a belief,” he bragged. “I have the scientific proof, the facts. You need to be on a statin, or you will likely drop dead.”

Background

Measuring and treating blood cholesterol is a huge industry, not simply diligent patient care. Statin pills alone brought in more than 15 billion dollars worldwide in 2025, and the broader class of cholesterol-lowering drugs—statins plus newer shots and tablets—around 32 to 35 billion. (1) (2) (3) (4) Laboratory work sits on top of that: market reports put cholesterol and lipid testing as high as 20 billion a year. (5) (6) Those cholesterol-drug sales are smaller than the diabetes-drug market (antidiabetic medicines near 96 billion dollars in 2025) and far smaller than cancer-drug spending (on the order of 190 to 290 billion dollars in 2025, depending on the scorekeeper). (7) (8) (9) Volume, not price per pill, is what makes statins a business: they are among the most written prescriptions on earth, and new heart-society rules that widen who “qualifies” expand that volume further.

Incentives are given physicians for prescribing drugs. Direct pay-per-prescription kickbacks are illegal in the United States under federal anti-kickback law, but quieter money is paid by manufacturers for meals, speaking fees, consulting, and training as listed in the federal Open Payments database. Those payments are associated with more brand-name statin prescribing. In Massachusetts Medicare data, doctors who took industry money wrote more brand-name statins than colleagues who did not; every additional 1,000 dollars in payments tracked a small rise in the brand-name share, and “educational training” payments tracked a larger jump. A national study found that even a single industry meal worth less than 20 dollars was linked to higher prescribing of the promoted brand-name statin (rosuvastatin) over other statins. That is influence is not “doctors are unpaid and untouched.” (10) (11) (12)

This article looks at two questions that official heart guidelines usually treat as one. First: when people take statin drugs that lower blood cholesterol, do fewer heart attacks, strokes, and deaths show up in controlled experiments? Second: is the common blood test called low-density lipoprotein (LDL) cholesterol the main reason arteries clog, and therefore a good reason to put tens of millions of people who have never had a heart attack on a daily drug?

The answers: People who already have heart disease, studies show a modest drop in heart attacks. In people who only have a risk score and a lab number, the benefit is much smaller than the advertisements and press releases suggest. How you tell the story—percent drop versus actual lives changed—matters more than most patients are ever told.*

What the experiments measure, in plain language

To generate business, and presumably save lives, albeit generally disregarding damaging ‘side effects,’ drug companies and medical societies prefer “relative” numbers. This is an important distinction: If a heart attack rate falls from 3 in 100 people to 2 in 100 people, that can be called a 33 percent reduction. The number that matters to a person sitting in a clinic is the “absolute” change: one fewer event among 100 people treated. The “number needed to treat” is how many people must take the drug for a set number of years to prevent one event.

A 2022 review in JAMA Internal Medicine added up 21 randomized experiments that compared a statin with a dummy pill or no statin. Across those studies the relative drop in death from any cause was 9 percent. The absolute drop was much lower–only 0.8 percentage points. The relative drop in heart attack was 29 percent; the absolute drop was 1.3 percentage points. The relative drop in stroke was 14 percent; the absolute drop was 0.4 percentage points. (13) From these numbers, a reasonably intelligent person can discern how the system is rigged to confuse and manipulate patients.

The same paper split the results by who was enrolled. In people who had not yet had a heart attack or stroke, the absolute drops were 0.6 percentage points for death, 0.7 for heart attack, and 0.3 for stroke. In people who already had vascular disease, the absolute drops were 0.9, 2.2, and 0.7 percentage points. That is the central fact. The drug looks strongest where the starting risk is already high.

From these numbers, a reasonably intelligent person can discern how the system is rigged to confuse and manipulate patients.

The United States Preventive Services Task Force reviewed 22 experiments in people without prior heart attacks or strokes. Statin use was linked to an 8 percent relative drop in death from any cause, which came to an absolute drop of 0.35 percentage points—about one death avoided for every 286 people treated over one to six years. Heart attacks fell by an absolute 0.85 percentage points (about one for every 118 people). Strokes fell by an absolute 0.39 percentage points (about one for every 256 people). Death specifically from heart disease was no longer statistically clear once two large older-adult data sets were included. When the reviewers looked only at longer experiments that were not stopped early, even the all-cause death finding lost statistical clarity. (14) (15)

A 2026 experiment called STAREE, published in the New England Journal of Medicine, enrolled community-dwelling adults age 70 and older who did not already have heart disease, diabetes, or dementia. Half received atorvastatin 40 milligrams; half received a dummy pill. After a median of 5.9 years, major heart events were lower on the drug (10.9 versus 15.5 events per 1,000 person-years). A broader measure that patients actually care about—death from any cause, dementia, or lasting physical disability—did not improve. (16)

The Cholesterol Treatment Trialists Collaboration, the group whose summaries most guidelines quote, reports about 22 percent fewer “major vascular events” for every 39 milligrams per deciliter drop in low-density lipoprotein cholesterol. That figure is a relative change in a mixed bundle of events that includes procedures such as stents, not a proof that every extra point of cholesterol lowering adds years of life.

How to read a headline

Take the widely advertised Anglo-Scandinavian Cardiac Outcomes Trial lipid arm. Advocates said atorvastatin cut fatal coronary disease and nonfatal heart attack by 36 percent. The raw counts were about 3.0 percent of people on the dummy pill versus 1.9 percent on the drug—an absolute difference of 1.1 percentage points. There was no clear saving of lives overall in that trial, and no benefit in women. Relative language turned a small absolute change into a campaign number. (17)

Where the “cholesterol causes death” story weakens

If high low-density lipoprotein cholesterol were a reliable killer across the lifespan, older people with higher levels should die sooner. A 2016 systematic review in BMJ Open by Uffe Ravnskov and colleagues gathered 19 follow-up studies in 30 groups totaling 68,094 people age 60 and older. In 16 of the groups that recorded death from any cause—covering 92 percent of those participants—higher low-density lipoprotein cholesterol went with lower death rates, and in 14 of those groups the finding was statistically significant. Death from heart disease was highest in the lowest cholesterol group in two cohorts and showed no link in seven. (18) (19)

That review does not erase plaque biology in a middle-aged man with inherited sky-high cholesterol. It does undercut the habit of treating an older person’s lab slip as if it were a death sentence.

Ravnskov, Michel de Lorgeril, and coauthors have also argued that influential summaries leave out experiments in which more cholesterol lowering did not produce fewer deaths, and that when those experiments are put back on the graph the “lower cholesterol, longer life” slope flattens or reverses. Readers can disagree about which studies belong in a pool. They cannot honestly say that death tracks cholesterol lowering as tightly as heart-attack composites do. (20)

A 2023 review of 60 randomized experiments including about 324,000 people, discussed in clinical commentary on that paper, found lipid-lowering drugs associated with roughly 1.33 fewer deaths per 1,000 people per year. That works out to treating on the order of 754 people for about four years to postpone one death from any cause, and more than 1,000 people to postpone one death attributed to the heart. When the authors asked whether more cholesterol lowering bought more survival, the line was essentially flat. (21)

Drugs that lower low-density lipoprotein cholesterol by means other than statins have often cut procedure-heavy event lists more cleanly than they have cut deaths. That pattern fits a drug that changes what gets counted in a hospital more reliably than it changes how long a person lives.

Who paid for the evidence, and who writes the rules

Almost all of the large statin experiments were paid for by the companies that sell the pills. Reviews of medical research in general find that industry-paid trials report more favorable results and friendlier conclusions than independently paid trials. The large United States government-leaning experiment called ALLHAT-LLT did not show that pravastatin reduced death or coronary heart disease in a primary-prevention population. That is a main reason the Preventive Services Task Force death estimates shrank when those data were added. (22)

The Cholesterol Treatment Trialists Collaboration at Oxford holds the combined individual-person records from the major trials. The Collaboration says it does not take industry grants for its own meta-analyses. The original experiments were industry-funded. Access to the raw person-level files is restricted by contracts with trial leaders and companies; outside scientists cannot freely re-check the counts. Reporting around a 2014 British Medical Journal dispute put more than 200 million pounds from Merck into Oxford’s related Clinical Trial Service Unit over prior decades. Closed data plus industry-origin trials is a structural conflict, even if today’s Collaboration grant line is listed as charitable. (23) (24) (25)

Guideline committees have been entangled with the same companies. Eight of nine authors of the 2004 United States cholesterol-treatment update had financial ties to statin makers. Of 15 authors of the 2013 American College of Cardiology / American Heart Association cholesterol guideline, six—including a co-chair—had recent industry ties. A review of earlier American College of Cardiology / American Heart Association guidelines found financial relationships in 56 percent of committee members. The societies now ban drug-company employees from writing panels and publish disclosure tables. They do not ban a professional world whose talks, research budgets, and advisory fees run through makers of cholesterol drugs. (26) (27) (28) (29)

The 2026 American College of Cardiology / American Heart Association blood-fat guideline widens who “qualifies.” Analyses of United States survey data estimate that about 87.5 million adults ages 30 to 79—roughly 57 percent of that age band—meet eligibility, including more than 93 percent of people ages 70 to 79, with about 21.5 to 25 million newly added. The newly added group has a lower average ten-year risk than the old target group. That is a large expansion of a daily drug market sitting on the small primary-prevention absolute benefits described above. (30) (31) (32)

Harms, as counted in experiments versus as felt in clinics

The Collaboration’s recent pooled look at trial safety forms found that most items listed on statin labels were not more common on the drug than on dummy pills, and that extra muscle symptoms in blinded experiments were small. That is what short, coded trial forms capture. (33)

What is not in serious dispute: a small rise in new diabetes; rare severe muscle injury that rises with dose; liver-test changes. What trial paperwork undercounts: fatigue, mental fog, and “I feel worse” among people who know they are taking the drug. Clinic rates of muscle complaints are far higher than the Collaboration’s excess. Some of that is expectation. Some of it is the drug. Both can be true.

What numbers might reasonably prompt a statin

There is no single cholesterol number that, by itself, means a statin is clearly worth it for every man or woman at every age. After a heart attack, stent, bypass, or documented artery disease, most trial evidence supports offering a statin even when the low-density lipoprotein cholesterol number is not extreme, because the starting risk of another event is already high. In people who have never had an event, the number that has long been treated as “high enough on its face” is roughly 190 milligrams per deciliter or higher of low-density lipoprotein cholesterol (or a similarly extreme inherited pattern). Below that, the decision is supposed to follow overall five- to ten-year risk—blood pressure, smoking, diabetes, family history—not the lab slip alone. In middle-aged men with several of those risks, many clinicians begin a serious discussion in the 160s and treat more readily as estimated risk climbs; that is a risk conversation, not proof that 160 is a magic harm threshold. A reading of 160 to 180, or a bit higher, is in the range where offices often start talking about a statin, but it is still a middling-to-high laboratory band, not an emergency number, and it does not by itself outweigh a low overall risk, especially in women. In middle-aged women the same numbers are weaker grounds for a strong push: several landmark experiments showed little or no clear saving of lives in women, so a high number plus low overall risk is a weak case for insisting on a daily drug.

In older adults the case for treating a number gets thinner, not thicker, even though current American heart-society rules would make almost everyone in their seventies “eligible.” Follow-up studies in people 60 and older often find that higher low-density lipoprotein cholesterol tracks with the same or longer survival, and the 2026 experiment in healthy people 70 and older cut later heart events without improving a combined measure of death, dementia, and lasting disability. A cholesterol of 130 or 150 in an otherwise well woman or man in their seventies is not, on the evidence, a strong reason to lean hard on a statin; neither is 160 to 180 in an older person whose only “problem” is the blood test. A history of prior heart disease still is.

How to hold the evidence without joining a camp

After a heart attack, bypass, stent, or documented artery disease, several experiments show fewer later heart attacks. The absolute benefit over about five years is typically in the low single percentage points. The effect on death from all causes is smaller and not the same in every experiment.

In a person whose only “disease” is a risk calculator and a cholesterol number, the relative headlines still look strong and the absolute change is thin. Hundreds of people may take the pill for years so that one event is delayed. In healthy people 70 and older, STAREE cut heart events without improving a combined measure of death, dementia, and disability.

In people over 60 followed in ordinary life, higher low-density lipoprotein cholesterol often tracks with equal or longer survival. That is a hard limit on treating an older adult for a lab value alone.

Statins do more than lower cholesterol. They affect inflammation and other pathways. If extra cholesterol lowering does not reliably add extra years of life, “how low did the number go” is the wrong scoreboard.

Medical societies and the Oxford pooling group are interested parties. So are the sharpest anti-statin writers. The useful habit is to insist on absolute numbers and open counts, not on either brand.

A working rule that respects the data

Judge the person by their actual event history and their five- to ten-year risk, not by a cholesterol number in isolation. Any claim of a “36 percent reduction” should arrive in the same sentence as the absolute change and the number of people who must be treated. Treat industry-funded composites and locked Oxford files as interested evidence, not as scripture. Treat the elderly follow-up studies that find no death penalty for higher cholesterol as a real constraint on medicating old age for a blood test.

__________________

* This is information for readers, not personal medical advice.

About the author

Dr. Leonard G. Horowitz, DMD, MA, MPH, DNM (hon), DMM (hon), received his doctorate in dental medicine from Tufts University; trained in periodontology and oral surgery at the University of Rochester; received his master’s degree in public health from Harvard University; and subsequently earned multiple awards as an author, filmmaker, energy medicine pioneer, and nutraceutical entrepreneur. He pioneered the fields of ‘frequency therapeutics’ and electroceuticals featuring especially the 528 frequencies of sound and light impacting epigenetics, stem cell differentiation, and healing dynamics; and was granted two honorary degrees, one in naturopathic medicine, and the other in missionary medicine, from the World Organization for Natural Medicine, wherein officials named him a “World Leading Intellectual.” His official website is DrLenHorowitz.com.