⚠ Draft — not for publication. Every citation and caveat is under clinician + legal review before this page goes live.
Longevity Program Join the Waitlist

Peak You · The Evidence

The Evidence

We measure the biological markers most strongly tied to how long — and how well — people live, then modify them on a clinical cadence and re-measure. This page is the peer-reviewed evidence behind each lever, and an honest account of what it does and doesn't prove.

Drawn from 55 peer-reviewed studies — meta-analyses, randomized trials, and large cohorts. Every citation on this page links to its source on PubMed. Each claim is tagged by how strong its evidence actually is.

How to read this page

Longevity claims live or die on one distinction: association is not causation. Much of the evidence that a marker predicts lifespan is observational — people with better values tend to live longer, which is not the same as proving that changing your value changes your fate.

Where randomized trials or genetic (Mendelian-randomization) data establish cause and effect, we say so. Where the evidence is only associational — however large and consistent — we say that too. Every lever below carries a badge for exactly how strong its evidence is, and every caveat sits on the same card as the claim.

Tier 1 · Causal

Randomized trials or genetic evidence show that intervening on this lever changes hard outcomes.

Tier 2 · Strong association

Large, consistent, dose-graded cohorts. Prognostic and plausible — but the marker predicts, it isn't proven to be the causal lever.

Tier 3 · Real, with caveats

Benefit is genuine but timing-, population-, or domain-specific, with meaningful trade-offs.

Tier 4 · Emerging

Predictive or surrogate-level evidence. The outcome benefit of measuring or moving it is not yet established.

The causal bedrock

Tier 1 · Proven in trials

Lipids — ApoB, LDL & Lp(a)

Causal · RCT + genetic

Lowering the particles that carry cholesterol lowers cardiovascular events — one of the most firmly established cause-and-effect chains in medicine.

For every ~39 mg/dL (1.0 mmol/L) reduction in LDL, major vascular events fell ~22% (RR 0.78, 95% CI 0.76–0.80) across 26 randomized trials and 170,000 people — with no floor below which benefit stopped. Genetic studies show LDL/ApoB particles are causal, not merely correlated.

  • Cholesterol Treatment Trialists' Collaboration. Individual-data meta-analysis, 26 RCTs, N≈170,000 · Lancet 2010. PMID 21067804 ↗
  • Ference BA, et al. (EAS Consensus). Genetic + epidemiologic + clinical triangulation · Eur Heart J 2017. PMID 28444290 ↗

Honest readRobustly causal — established by randomized trials and genetics, not observation alone. Caveat: dedicated Lp(a)-lowering outcome trials were still ongoing, so a treatment benefit of lowering Lp(a) specifically is not yet proven.

At Peak YouWe measure ApoB and Lp(a) at baseline (not just standard LDL), design a clinician-led protocol, and re-test to confirm your particle burden is actually falling.

Blood pressure

Causal · RCT

Lower blood pressure causes fewer heart attacks, strokes, and deaths — down to levels once considered "normal."

Intensive control (target <120) cut all-cause mortality 27% (HR 0.73) in SPRINT. Across 123 trials and 613,000 people, each 10 mmHg drop in systolic pressure reduced major cardiovascular events ~20% and mortality ~13%.

Honest readThe strongest tier of evidence. The trade-off is real: intensive targets raised low-blood-pressure episodes and kidney stress — which is why targets are individualized, not one-size-fits-all.

At Peak YouBaseline plus home blood-pressure monitoring, a clinician protocol with a target balanced to your risk, and tracked deltas — every sustained 5–10 mmHg drop is a causally-backed reduction in risk.

GLP-1 medications

Causal · RCT

In higher-risk people with obesity, semaglutide reduced major cardiovascular events — the first time a weight medication proved it in people without diabetes.

In SELECT (N=17,604), semaglutide cut major cardiovascular events 20% (HR 0.80, 95% CI 0.72–0.90). Tirzepatide produced mean weight loss up to −20.9% in a 72-week trial.

  • Lincoff AM, et al. (SELECT). CV-outcomes RCT, N=17,604 · N Engl J Med 2023. PMID 37952131 ↗
  • Jastreboff AM, et al. (SURMOUNT-1). RCT, N=2,539 · N Engl J Med 2022. PMID 35658024 ↗

Honest readThe weight-loss trials prove weight loss, not longevity. Only SELECT shows a reduction in cardiovascular events, and only in a high-risk population — we don't generalize it to healthy, low-risk people. Both are industry-funded.

At Peak YouA clinician-supervised GLP-1 protocol for eligible members, with weight and cardiometabolic deltas tracked over the membership — never a default, always indicated.

Targeted cancer screening

Causal · RCT (in the right groups)

In the populations it was proven in, screening catches cancer early enough to lower cancer death.

Low-dose CT cut lung-cancer death 20% in heavy smokers (NLST, N=53,454). Stool-based screening cut colorectal-cancer death up to 32% over 30 years.

Honest readThe disease-specific benefit is real, but the all-cause signal is small (lung) or absent (colorectal), and screening carries overdiagnosis and false-positive harms. Whole-body MRI in healthy people has no such outcome evidence — we don't present it as proven.

At Peak YouAn evidence-tiered diagnostics program that leads with the screens shown in trials to save lives in the right people — matched to your history, not sold as a panel to everyone.

Strong predictors of how long you live

Tier 2 · Strong association

Cardiorespiratory fitness (VO₂max)

Strong association + AHA consensus

Fitness is one of the single strongest predictors of mortality medicine has — stronger than most traditional risk factors.

People in the lowest fitness group had ~5× the all-cause mortality of the fittest (HR 5.04), with no observed ceiling to the benefit. Each 1-MET higher capacity tracked 13% lower mortality. The American Heart Association calls fitness a "clinical vital sign."

Honest readObservational — illness can lower fitness (reverse causation) — but the consistent dose-response, temporal ordering, and supporting exercise trials make it one of medicine's most trusted prognostic markers.

At Peak YouVO₂max is a headline baseline and re-test delta; low fitness is a clinician-actionable flag that drives a structured aerobic protocol.

Muscle, strength & resistance training

Strong association

Strength predicts survival, and building it is tied to lower mortality.

Each 5-kg lower grip strength tracked 16% higher all-cause mortality across 139,691 people in 17 countries — outperforming systolic blood pressure as a predictor. Any regular muscle-strengthening activity was linked to 15% lower mortality, with most of the benefit by ~30–60 min/week.

Honest readGrip strength is a prognostic biomarker of whole-body health — squeezing harder isn't itself the lever. The training data are observational, but the mechanism (strength preserves function) is well supported.

At Peak YouGrip strength and DEXA lean-mass as low-cost baselines and deltas, plus a clinician-prescribed resistance-training protocol (~2 sessions/week) with adherence tracking.

Coronary artery calcium (CAC)

Strong prognostic marker

A quick scan shows the actual plaque in your arteries — and reclassifies your risk beyond any calculator.

Each doubling of calcium score raised cardiovascular risk ~14% independent of standard risk factors; a score of zero marks durably low risk (the "power of zero"), while everyone with a score ≥100 exceeded the treatment threshold.

Honest readA powerful prediction tool — but no trial has proven that treating by CAC score changes outcomes. Its value is identifying who most benefits from the causally-proven levers above.

At Peak YouA CAC scan to stratify risk — zero to reassure and de-intensify, elevated to trigger the clinician lipid/BP protocol — paired with ApoB so treatment matches true arterial burden.

Visceral fat & waist

Strong association (with a caveat)

Where your body stores fat predicts mortality better than BMI does.

The highest vs lowest waist measurements carried ~1.5–2× the all-cause mortality across 359,000 Europeans; a waist under half your height is a simple, validated screen.

Honest readObservational, and honestly, one large collaboration found waist adds little extra prediction once blood pressure, lipids, and glucose are known (ERFC, Lancet 2011). It's a useful screen, not a proven causal lever — "associated with," not "causes."

At Peak YouA DEXA / visceral-fat and waist baseline as an evidence-backed marker beyond BMI, with deltas tracked over time.

Real, with real caveats

Tier 3 · Genuine but bounded

Protein & nutrition

Real · pattern proven, dose nuanced

The right eating pattern lowers cardiovascular risk — and protein needs change with age.

A Mediterranean diet cut major cardiovascular events ~30% in a randomized trial (HR ≈0.70). Protein's relationship with mortality is age-dependent — a concern in mid-life that reverses in older adults, who need more protein to preserve muscle.

Honest readThe randomized evidence proves the Mediterranean pattern cuts events — not a specific protein dose. The protein/age findings are observational and hypothesis-generating. The takeaway is a nuanced, age-tiered plan, not "more protein."

At Peak YouAge- and goal-personalized clinician-set protein targets within a Mediterranean-pattern frame, tracked against DEXA lean-mass and metabolic response.

Menopause hormone therapy (HRT)

Real · timing- & person-dependent

Started near menopause, hormone therapy relieves symptoms and protects bone — and, over 18 years, did not change overall mortality either way.

Long-term all-cause mortality was neutral (HR 0.99) across the Women's Health Initiative trials. But the same program, begun on average a decade past menopause, showed real absolute harms — more breast cancer, stroke, and clots.

Honest readThe "timing hypothesis" (near-menopause initiation may be net-favorable) is plausible but not settled. Critically, HRT is not FDA-approved to prevent heart disease or extend life — it's approved for symptoms and bone, where the benefit is strongest.

At Peak YouHormone labs and symptom staging, then clinician-supervised, individualized HRT weighed against personal risk — framed honestly as symptom and bone care, with ongoing monitoring. Not marketed as longevity.

Testosterone therapy (TRT)

Real · for confirmed low-T only

In men with properly diagnosed low testosterone, replacement improves symptoms with acceptable cardiovascular safety.

In the FDA-mandated safety trial (N=5,246), testosterone did not raise major cardiac events (HR 0.96, non-inferior). Benefits were domain-specific — sexual function improved reliably; vitality and physical-function effects were weak or absent.

  • Lincoff AM, et al. (TRAVERSE). RCT, N=5,246 · N Engl J Med 2023. PMID 37326322 ↗
  • Snyder PJ, et al. (Testosterone Trials). RCTs, N=790 · N Engl J Med 2016. PMID 26886521 ↗

Honest readThe safety trial ran ~33 months and flagged higher atrial fibrillation, clots, and kidney injury. No trial shows TRT extends lifespan; the FDA approves it only for hypogonadism from a defined medical cause — not age-related decline alone.

At Peak YouTwo morning fasting testosterone labs plus LH/SHBG/hematocrit/PSA to confirm true low-T before treatment, then supervised TRT with structured monitoring — framed as symptom care, not longevity.

Sleep

Strong, consistent association

Both too little and too much sleep track higher mortality, cardiovascular disease, and dementia.

Across ~1.38M people, short sleep carried 12% higher and long sleep 30% higher mortality (a U-shape). Persistent short sleep in mid-life tracked ~30% higher dementia risk over 25 years.

Honest readObservational, and the long-sleep arm is heavily confounded — long sleep often marks underlying illness rather than causing harm. Treat this as risk-marker optimization, not proof that changing your hours changes your lifespan.

At Peak YouWearable sleep tracking (duration, regularity, quality), with chronically short or long sleep flagged for clinical review of the driver — apnea, mood, cardiometabolic risk.

Emerging — promising, not proven

Tier 4 · Investigational

Biological age & epigenetic clocks

Emerging · predictive, not proven to move

DNA-methylation "clocks" predict mortality well. Whether lowering your clock reading adds years is not yet proven.

A one-year-older reading on a validated clock tracked 4.5% higher mortality risk. The one randomized human trial found caloric restriction slowed one clock by ~2–3% — and did not move two others.

  • Levine ME, et al. (PhenoAge). Biomarker + multi-cohort validation · Aging 2018. PMID 29676998 ↗
  • Waziry R, et al. (CALERIE). RCT, N=197 · Nat Aging 2023. PMID 37118425 ↗

Honest readClocks are robustly predictive, but prediction is not proof: it has not been shown that lowering a clock reading extends life. We treat "clock reversal" as a research signal, not an outcome — a moving number is not added years.

At Peak YouA biological-age baseline and repeat-measured trajectory dashboard — explicitly a tracking tool, not a proven lever.

Continuous glucose monitoring in non-diabetics

Emerging · personalization, not outcome-proven

A CGM reveals your personal glucose responses. Whether wearing one improves outcomes in healthy people is not established.

People considered normoglycemic by standard labs spent ~15% of the time in a prediabetic glucose range. But a 2025 peer-reviewed review found that of the non-diabetic CGM studies, only one even measured a clinical outcome.

Honest readHard-outcome evidence in healthy people is thin. CGM here is hypothesis-personalizing — a way to see your own responses — not proven to prevent disease or extend life.

At Peak YouCGM offered as a personalization and self-experiment tool against reference ranges — never marketed as a proven longevity intervention.

Multi-cancer early detection (blood tests)

Emerging · immature outcome data

Blood tests that screen for many cancers at once are promising — and still unproven for saving lives.

A leading test was highly specific (99.5%) but caught only 16.8% of the earliest-stage (most curable) cancers; in real-world use, a positive result was a true cancer only 38% of the time.

  • Klein EA, et al. (Galleri/CCGA). Validation, N=4,077 · Ann Oncol 2021. PMID 34176681 ↗
  • Schrag D, et al. (PATHFINDER). Prospective cohort, N=6,621 · Lancet 2023. PMID 37805216 ↗

Honest readCurrent evidence is analytic performance and feasibility — not proof it reduces mortality. Early-stage sensitivity is modest and false positives are common; definitive outcome trials are ongoing.

At Peak YouOffered only as a clearly-labeled emerging adjunct, with honest counseling on false positives — never as a substitute for the screens proven in trials.

Why continuous, measured care

The membership thesis

Measure → modify → re-measure, on a cadence

Mechanism proven in RCTs

The evidence doesn't just support individual levers — it supports the loop itself: sustained, measured, feedback-driven care beats the once-a-year physical.

Sustained, multifactorial risk-factor management cut all-cause mortality 46% in a landmark trial (HR 0.54). A long-term relationship with one clinician (>15 years) tracked 25% lower mortality across 4.5M people; patient-measured, feedback-driven titration lowered blood pressure more than clinic visits alone.

Honest readThe continuity-and-mortality link is observational (more-engaged patients differ). But the mechanism — measure, titrate, re-measure — is proven in randomized trials. That mechanism is exactly the Peak You loop.

At Peak YouA persistent clinical relationship plus longitudinal self-measurement and active titration — the same feedback loop that moved outcomes in these trials, run continuously instead of once a year.

What the evidence does not show

The most important section on this page.

  • No one has run a trial on "the Peak You program." The evidence above supports individual levers; the bundled membership has not itself been shown to extend lifespan. We build on proven levers — we don't claim the bundle is proven.
  • Moving a marker is not the same as living longer. Improving VO₂max, ApoB, a glucose curve, or a biological-age estimate is a reasonable, monitored proxy — not a guarantee of added years.
  • Medicine is humbling. In one major randomized trial, even an intensive lifestyle intervention did not reduce cardiovascular events (Look AHEAD, N Engl J Med 2013). We measure and adjust — we don't promise outcomes.
  • Population statistics are not personal promises. A hazard ratio from a study of thousands is not a prediction about you. Individual results vary.
  • The emerging tools are labeled emerging — CGM in healthy people, epigenetic-clock reversal, and multi-cancer blood tests are promising research, not proven interventions.
  • Hormone therapy targets symptoms, not longevity — HRT and TRT help the right patients feel and function better; neither is proven to extend life.

Built on the evidence — not on hype.

Peak You is a clinician-led longevity membership built around the levers above, measured and modified on a clinical cadence. Founding members are forming now.

Join the Waitlist

References

Every study cited on this page, linked to its source. Compiled from 55 peer-reviewed references; each was verified against its primary source. Full bibliography available on request.

  1. Cholesterol Treatment Trialists' Collaboration. Lancet 2010;376:1670–1681. PMID 21067804
  2. Ference BA, et al. Eur Heart J 2017;38:2459–2472. PMID 28444290
  3. SPRINT Research Group. N Engl J Med 2015;373:2103–2116. PMID 26551272
  4. Ettehad D, et al. Lancet 2016;387:957–967. PMID 26724178
  5. Lincoff AM, et al. (SELECT). N Engl J Med 2023;389:2221–2232. PMID 37952131
  6. Jastreboff AM, et al. (SURMOUNT-1). N Engl J Med 2022;387:205–216. PMID 35658024
  7. Aberle DR, et al. (NLST). N Engl J Med 2011;365:395–409. PMID 21714641
  8. Shaukat A, et al. N Engl J Med 2013;369:1106–1114. PMID 24047060
  9. Mandsager K, et al. JAMA Netw Open 2018;1:e183605. PMID 30646252
  10. Kodama S, et al. JAMA 2009;301:2024–2035. PMID 19454641
  11. Leong DP, et al. (PURE). Lancet 2015;386:266–273. PMID 25982160
  12. Momma H, et al. Br J Sports Med 2022;56:755–763. PMID 35228201
  13. Detrano R, et al. (MESA). N Engl J Med 2008;358:1336–1345. PMID 18367736
  14. Blaha MJ, et al. Atherosclerosis 2020;294:72–79. PMID 31784032
  15. Pischon T, et al. (EPIC). N Engl J Med 2008;359:2105–2120. PMID 19005195
  16. Ashwell M, et al. Obes Rev 2012;13:275–286. PMID 22106927
  17. Emerging Risk Factors Collaboration. Lancet 2011;377:1085–1095. PMID 21397319
  18. Estruch R, et al. (PREDIMED). N Engl J Med 2018;378:e34. PMID 29897866
  19. Levine ME, et al. Cell Metab 2014;19:407–417. PMID 24606898
  20. Manson JE, et al. (WHI mortality). JAMA 2017;318:927–938. PMID 28898378
  21. Rossouw JE, et al. (WHI). JAMA 2002;288:321–333. PMID 12117397
  22. Lincoff AM, et al. (TRAVERSE). N Engl J Med 2023;389:107–117. PMID 37326322
  23. Snyder PJ, et al. (TTrials). N Engl J Med 2016;374:611–624. PMID 26886521
  24. Cappuccio FP, et al. Sleep 2010;33:585–592. PMID 20469800
  25. Sabia S, et al. Nat Commun 2021;12:2289. PMID 33879784
  26. Levine ME, et al. (PhenoAge). Aging 2018;10:573–591. PMID 29676998
  27. Waziry R, et al. (CALERIE). Nat Aging 2023;3:248–257. PMID 37118425
  28. Hall H, et al. PLoS Biol 2018;16:e2005143. PMID 30040822
  29. Wilczek MP, et al. Sensors 2025;25:187. PMID 39796978
  30. Klein EA, et al. (Galleri). Ann Oncol 2021;32:1167–1177. PMID 34176681
  31. Schrag D, et al. (PATHFINDER). Lancet 2023;402:1251–1260. PMID 37805216
  32. Gæde P, et al. (Steno-2). N Engl J Med 2008;358:580–591. PMID 18256393
  33. Sandvik H, et al. Br J Gen Pract 2022;72:e84–e90. PMID 34607797
  34. McManus RJ, et al. (TASMINH4). Lancet 2018;391:949–959. PMID 29499873
  35. Look AHEAD Research Group. N Engl J Med 2013;369:145–154. PMID 23796131