Skip to content

Statins? A public-health triumph or the "cholesterol hoax"

Statins as first-line therapy are neither the unqualified public-health triumph implied by ever-widening prevention guidelines nor the "cholesterol hoax" promoted by internet contrarians.

Statins as first-line therapy are neither the unqualified public-health triumph implied by ever-widening prevention guidelines nor the "cholesterol hoax" promoted by internet contrarians.

Stripped of both pharmaceutical optimism and populist denial, current evidence splits sharply by baseline risk: for secondary prevention and high-risk patients, statins rest on some of the strongest causal and randomized evidence in modern medicine; for broad, low-risk primary prevention, the case relies heavily on relative-risk framing, composite endpoints, and multi-decade extrapolations that outrun direct trial data.

In Short, there is strong randomized evidence supporting statins for secondary prevention and high-risk patients, while highly questionable use for broad primary prevention in low-risk groups due to reliance on relative risks and long-term extrapolations.


Interactive Widget

Statin 5-Year Risk-Benefit Trade-off

Compare how a constant relative risk reduction translates into absolute cardiovascular benefit versus excess new-onset diabetes across patient risk strata.

Relative Risk Reduction

33.8%

-1.75 mmol/L LDL-C

Absolute Risk Reduction

1.69%

17 fewer CV / 1,000

5-Year NNT

59

to prevent 1 MVE

CV Prevented : DM

0.9 : 1

20 excess DM / 1,000

5-Year Outcomes per 1,000 Patients

Borderline Primary · High-Intensity Statin

5.0%
135 mg/dL
5.5%

Model grounded in CTT Collaboration 2012 (21% MVE RRR per 1.0 mmol/L LDL-C reduction) and CTT Collaboration 2024 (new-onset diabetes RR 1.10 moderate, 1.36 high intensity).

Why Serious Experts Refuse to Abandon Statins

Any rigorous critique must confront an empirical anchor: multi-pathway genetic and pharmacological convergence.
Loss-of-function mutations in the LDL receptor (LDLR, causing familial hypercholesterolemia) produce severe, premature atherosclerosis even in lean, non-smoking, normotensive individuals.

Conversely, loss-of-function variants in PCSK9, HMGCR, and NPC1L1 lower circulating LDL-cholesterol (LDL-C) from birth and dramatically reduce lifetime coronary disease.

Crucially, this is not unique to statins: non-statin drugs that upregulate hepatic LDL clearance through distinct mechanisms—ezetimibe, PCSK9 monoclonal antibodies, and bempedoic acid—reduce cardiovascular events along a remarkably similar proportional curve per unit of LDL-C lowered.

Add over 170,000 randomized patients across 27 trials and pennies-per-day generic pricing, and the consensus reason statins sit at first-line is clear: they target a proven causal pathway at trivial direct financial cost.

Separating Observation from Inference

The friction begins when the field moves from "statins lower vascular events" to "millions of asymptomatic, low-risk adults should take a statin for 30 years."

  • Direct observations: In the Cholesterol Treatment Trialists' (CTT) individual participant data meta-analysis of 27 trials ($N \approx 174,000$, median follow-up ~5 years), each 1.0 mmol/L (39 mg/dL) drop in LDL-C reduced major vascular events by 21% relative (95% CI 19%–23%), major coronary events by 24%, coronary revascularizations by 24%, and strokes by 15%. Yet when Byrne et al. (2022, JAMA Internal Medicine) stripped out softer revascularization procedures across 21 trials ($N \approx 140,000$, ~4.4 years average follow-up) and reported absolute effects on hard individual outcomes, the pooled reductions were modest: 0.8% for all-cause mortality (9% relative), 1.3% for myocardial infarction (29% relative; NNT ~77), and 0.4% for stroke (14% relative).
  • The elderly primary-prevention readout: In August 2026, the long-awaited Australian STAREE trial ($N = 9,971$, mean age 74.7, independent adults aged 70+ without prior cardiovascular disease, diabetes, or dementia) reported that atorvastatin 40 mg over 5.9 years cut major cardiovascular events by 30% relative (HR 0.70; ~23 fewer events per 1,000 treated, approximate NNT ~44). However, it produced no statistically significant improvement in its co-primary endpoint of disability-free survival (death, dementia, or persistent physical disability: 12.8% vs. 13.6%; HR 0.94, 95% CI 0.84–1.05).
  • Model-dependent inferences: Modern guidelines—including the 2026 ACC/AHA Dyslipidemia Guideline introducing the PREVENT-ASCVD equations—recommend lipid-lowering therapy at intermediate 10-year risk (5% to under 10%), call it reasonable at borderline risk (3% to under 5%), and even allow consideration below 3% 10-year risk when 30-year risk reaches 10%. That policy depends on two unverified extrapolations: first, that the 4- to 6-year proportional risk reduction seen in trials compounds cleanly over 20 to 30 years of real-world adherence; second, that the absolute excess of statin-induced new-onset diabetes—a dose-dependent relative increase of 10% on low/moderate-intensity and 36% on high-intensity statins in CTT's 2024 Lancet Diabetes & Endocrinology meta-analysis—carries negligible downstream microvascular and metabolic cost over decades compared with averted vascular events.

"Step 0" Assumptions and Instrument Blind Spots

Even accepting the trial numbers at face value, the "statins-first for cholesterol" framework embeds a major Step 0 assumption and inherits four structural blind spots in how trials and guidelines measure risk and harm.


First, Step 0 treats circulating cholesterol mass as the origin of the disease rather than a necessary substrate. Atherogenesis requires ApoB-containing lipoproteins to cross a dysfunctional endothelium, become retained in the arterial intima, and trigger a chronic immune-inflammatory cascade. Starting clinical intervention at "inhibit hepatic HMG-CoA reductase" does not address why the endothelium became permeable or insulin-resistant in the first place. It also conflates historical precedence with mechanistic optimality: statins remain first-line largely because cheap generic patents and legacy trial volume make them unbeatable on cost-effectiveness, not because broad mevalonate-pathway inhibition (which also depletes downstream intermediates like CoQ10 and impairs glycemic homeostasis) is inherently cleaner than targeted ApoB clearance.

Second, the evidence pipeline systematically tilts toward smoother efficacy and quieter harms:

  • LDL-C mass vs. ApoB particle count: Standard lipid panels measure cholesterol mass inside LDL rather than the actual number of atherogenic particles (ApoB). In insulin resistance and metabolic syndrome, patients routinely carry elevated numbers of small, cholesterol-depleted ApoB particles alongside a "normal" LDL-C—meaning cholesterol-centric screening misses high-risk biology while flagging low-risk patients with large, buoyant particles.
  • Composite endpoint elasticity: Pooling hard events (death, MI, stroke) with physician-driven procedures (coronary revascularization) inflates event rates and stabilizes statistical significance. Even STAREE had to add coronary revascularization to its primary cardiovascular composite after a blinded interim review found fewer hard events than expected.
  • Run-in filtering and raw data opacity: Several foundational statin trials used active run-in phases that excluded patients who experienced early intolerance or poor adherence before randomization, structurally depressing reported adverse-event rates compared with routine clinical practice. Meanwhile, the Oxford-based CTT Collaboration holds the individual participant data for most major industry-sponsored trials under closed sharing agreements—and even while publishing its February 2026 Lancet meta-analysis ($N = 154,664$ across 23 trials) concluding that most labeled side effects (cognitive decline, depression, neuropathy, sleep or sexual dysfunction) lack causal RCT backing and urging label revisions, the group reiterated that raw participant data will not be released for independent third-party audit.
  • The relative-risk illusion in low-risk strata: In CTT's own 2012 low-risk analysis, participants with under 5% 5-year vascular risk showed an impressive 38% relative risk reduction—yet the underlying annual event rate moved only from 0.56% to 0.38% per year. Presenting constant relative percentages across wildly different denominators hides the point at which clinical benefit approaches statistical noise for an individual patient.

Serious Insider Alternatives (and What Would Falsify Them)

Inside preventive cardiology and clinical epidemiology—well clear of fringe claims that LDL does not enter plaques—three rigorous counter-models challenge the blanket "statins-first by population risk score" consensus:

Existing evidence decisively rules out both caricatures: lowering ApoB-containing lipoproteins unmistakably reduces atherosclerotic events, and in high-risk or post-MI patients, statins remain an unusually high-value intervention. Where consensus quietly overreaches is at the low-risk frontier—treating a constant ~21%–30% relative risk reduction as self-evident justification for medicating healthy adults for decades when hard-outcome absolute gains shrink toward 1% over five years, disability-free survival in older primary-prevention adults did not budge in STAREE, and independent researchers still cannot audit the raw individual-patient trial data behind the official safety meta-analyses.

Explore the Denominator Effect: Relative vs. Absolute Trade-Offs

Because the entire primary-prevention debate turns on how a fixed relative risk reduction behaves across different baseline risk strata—and how quickly absolute benefit shrinks relative to statin-induced dysglycemia—adjust the patient profile below to compare projected 5-year absolute risk reduction (ARR), Number Needed to Treat (NNT), and estimated excess new-onset diabetes using CTT Collaboration effect sizes.

Key insight: Switch between the Low-Risk Primary preset and Secondary Prevention while keeping LDL-C and statin intensity fixed. Notice that the relative risk reduction barely changes, yet the clinical reality flips completely: in secondary prevention, prevented cardiovascular events tower over excess diabetes cases, whereas in low-risk primary prevention on high-intensity therapy, the number of excess diabetes diagnoses per 1,000 patients can rival or exceed the number of major vascular events prevented over a 5-year window.

Eduardo Bergel and Gemini Neon

T333T.com Research

The Symbiont

Comments

Latest