Outcome beats surrogate
A lower lab value, more NAD+, or a younger “biological age” can be interesting. It is not evidence that people lived longer. Prefer all-cause mortality, disability, falls, or disease events.
Quick reference
The Leverage Ladder
Bar length is an intentionally rough visual composite of human outcome evidence—not a personal percentage or an additive calculator.
Quit smoking / never startA
Cardiorespiratory fitnessA
Regular activityA
Blood pressure / metabolic riskA
Resistance trainingB
Sleep, diet, social connectionB
Alcohol reduction / medicationsC / A*
Most “anti-aging” supplementsD
*Prescription therapies can have Grade-A evidence for a defined disease or risk group; they are not a universal longevity stack.
Read claims before you optimize them
Relative ≠ absolute
A 20% relative reduction means very different absolute benefit at 1% versus 20% baseline risk. Ask: compared with whom, over how long, and how many events?
Cohorts can reverse
People who quit because they are ill, or who take supplements because they are health-conscious, can make observational comparisons misleading. Randomization helps, but is scarce for lifespan.
Grades are evidence, not virtue
A: strong human outcome evidence; B: strong association or practical trial evidence; C: mixed/indication-specific; D: mechanism, animals, or null data.
Working knowledge
The actions worth doing first
| Action | Human evidence & concrete threshold | Who gains most / practical next step |
|---|---|---|
| Stop smoking | In a U.S. cohort, stopping at ages 25–34, 35–44, or 45–54 gained about 10, 9, or 6 years versus continuing smoking (Jha 2013). | Every current smoker; use evidence-based cessation support rather than “cutting down” indefinitely. |
| Build aerobic capacity | In 122,007 treadmill-tested adults, low versus elite fitness carried adjusted mortality HR 5.04; selection bias means this is not a promised 5× reversible effect. | People with low baseline fitness; begin with tolerable walking/cycling and progress. See running or cycling. |
| Do any regular activity | 661,137 adults: even <7.5 MET-hours/week was associated with 20% lower mortality than none; 1–2× the guideline range with 31% lower risk. | Inactive people. Start below the 150-min/week guideline; consistency beats an ideal plan never begun. |
| Strength train | 2022 meta-analysis: any resistance training associated with 15% lower all-cause mortality; the model’s largest reduction was around 60 min/week. | Older adults and anyone losing strength/function. See strength-training programming. |
| Measure waist, BP, glycemia | Use a tape, validated cuff, and appropriately ordered labs. A waist under half of height is a simple screening rule; an isolated measure is not a diagnosis. | Those with central weight gain or family risk. Use the biomarker scorecard then the blood-test decoder. |
| Treat hypertension when indicated | Blood pressure is a causal risk factor, unlike many “aging” scores. Targets and medication effects vary by person, measurement method, comorbidity, and harms. | Anyone with repeatedly elevated home readings; bring a validated log to a clinician, not one anxious reading. |
| Eat a durable pattern | Prioritize minimally processed plants, protein adequacy, and fiber-rich foods over single “longevity foods.” Dietary studies have confounding; adherence is the intervention. | People choosing between sustainable patterns. Avoid replacing a diet with powders or olive-oil theater. |
| Sleep enough—and diagnose disorders | Adults generally need 7–9 hours; duration has a U-shaped observational relationship with outcomes. Severe daytime sleepiness, snoring, or apneas deserve assessment. | People sacrificing sleep or using sedatives as sleep hygiene. See sleep optimization. |
| Maintain social connection | Isolation and loneliness predict worse outcomes, but the causal “dose” is not a weekly friend quota. Build recurring obligations and real support. | People whose health plan has no people in it. Choose one repeatable group, call, or shared activity. |
| Use prevention, not biohacks | Vaccination, screening, dental care, seat belts, and addressing substance use have direct risk pathways. The right test depends on age, sex, history, and jurisdiction. | People with an “optimization” budget: make routine care boringly complete first. |
| Alcohol: less is safer | Claims that moderate drinking is protective are vulnerable to abstainer and healthy-user bias. Do not start drinking for heart health. | People weighing a claimed benefit. If drinking creates harm or loss of control, seek support rather than setting a longevity quota. |
| Use indicated medicines | Statins, antihypertensives, diabetes therapies, and other drugs can have hard outcome evidence for specific indications—not as generic anti-aging prescriptions. | People with diagnosed risk. Compare absolute benefits/harms in NNT medical interventions. |
| Protect mobility and falls | Strength, balance, vision correction, medication review, and home hazards are more actionable than a “young” wearable score. | Especially after 65 or after a fall; include balance work and a home safety review. |
| Manage depression and chronic stress | They affect sleep, activity, adherence, substance use, relationships, and suicide risk. No supplement substitutes for evidence-based mental-health care. | Anyone whose plan fails because life is unmanageable: treat the bottleneck, not the self-discipline myth. |
| Measure function, not only labs | VO2 max estimate, grip, gait speed, waist and BP are low-cost tracking tools; their trends guide behavior better than a one-off “age” score. | Use the biomarker reference to choose a small dashboard. |
Edge & advanced
Myths, half-truths, and uncertainty
“Metabolism crashes after 30.”
In doubly labeled water data, expenditure adjusted for fat-free mass was broadly stable from ages 20–60, then declined in older age. Body composition and activity can change; the folk deadline is wrong.
“Blue Zones prove a secret diet.”
Exceptional-age records and demographic data have been criticized; that does not make walking, social life, or plant-forward food bad ideas. It means do not infer a causal supplement or menu from a branded longevity map.
“Antioxidant pills slow aging.”
Oxidative biology is not a license for high-dose pills. A Cochrane review found increased mortality signals for vitamin A, beta-carotene and vitamin E in lower-bias trials.
“Detoxes remove toxins.”
There is no general detox pathway a cleanse activates. Exposure control and medical evaluation for a specific toxin are real; juice, laxative, and sauna narratives are not substitutes.
“Resveratrol/NAD boosters are longevity drugs.”
Mechanism and short-term biomarker changes are not human lifespan trials. See the supplement evidence ledger before buying an expensive stack.
“Lifting is dangerous after 50.”
Unsupervised maximal effort can be dangerous at any age; progressive resistance training is a central way to preserve strength and function. Technique, load selection and conditions matter.
“More exercise is always better.”
Dose-response curves rise steeply from none and flatten; elite volumes are not a required longevity dose. Train for capacity, recovery, enjoyment and injury history.
“Genes decide the whole story.”
Family history matters and can guide prevention, but it is not a reason to skip the modifiable exposures above. A genetic risk result is not a personal forecast.
“One biological-age test tells your true age.”
Clocks can be research tools, but test-retest variation and interpretation limits make them poor steering wheels compared with function, BP, waist and clinically useful labs.
“Moderate drinking protects the heart.”
Modern genetic and bias-aware analyses weaken the simple protective J-curve story. Do not prescribe alcohol as a health intervention.
Common mistakes
Buying exotic before boring.
Fitness, tobacco, BP and sleep are not optional prerequisites for an NR order.
Fitness, tobacco, BP and sleep are not optional prerequisites for an NR order.
Optimizing a surrogate.
A number that moves is not automatically a causal target or a longer life.
A number that moves is not automatically a causal target or a longer life.
Adding effects together.
Observational hazard ratios overlap; a ladder is a priority list, not an actuarial calculator.
Observational hazard ratios overlap; a ladder is a priority list, not an actuarial calculator.
Copying an influencer.
Disclosure, access to medical supervision, and a n=1 outcome are not a trial.
Disclosure, access to medical supervision, and a n=1 outcome are not a trial.
Sources and related references
Primary papers: Jha et al., NEJM 2013 (smoking cessation); Mandsager et al., JAMA Network Open 2018 (fitness); Arem et al., JAMA Internal Medicine 2015 (activity); Shailendra et al., 2022 (resistance training); Pontzer et al., Science 2021 (energy expenditure); Bjelakovic et al., Cochrane (antioxidants). All accessed 2026-07-13.
Continue with the cluster
Biomarkers: which numbers predict outcomes · Supplements: the evidence ledger · Actual Risk Dashboard · Reassurance Tables