Longevity

The Signal and the Hype

Longevity science is having a moment. Billions in funding, celebrity scientists, and a wave of self-described longevity practitioners have created a climate where speculation routinely outpaces evidence. The popular discourse mixes genuinely robust findings (exercise extends healthspan; smoking shortens it) with speculative extrapolations from mouse models and short-term human studies.

The key distinction worth maintaining: lifespan (how long you live) and healthspan (how many years you live in good health) are related but different outcomes. Most longevity interventions that work in humans extend healthspan more reliably than they extend maximal lifespan. Compression of morbidity — dying later and faster rather than earlier and slower — is the more tractable goal for most people.

Individual variation in aging rate is substantial. Biological aging markers between individuals of the same chronological age can differ by 10–20 years. This makes personal measurement unusually important in this domain.

Exercise: The Most Replicated Longevity Intervention

No pharmacological intervention comes close to exercise for longevity outcomes in human data. A landmark 2018 study by Kokkinos et al. (n = 122,007, median 10-year follow-up) found that the most fit individuals had a 45% lower all-cause mortality hazard ratio than the least fit. There was no plateau — every increment of cardiorespiratory fitness was associated with reduced mortality.

VO2 max is the single strongest predictor of all-cause mortality in large prospective studies, stronger than blood pressure, cholesterol, or diabetes status. Peter Attia's reanalysis of NHANES data shows that individuals in the bottom quartile of VO2 max have a roughly 5x higher mortality risk than those in the top quartile — an effect size larger than smoking.

The minimum effective dose for longevity benefit appears to be around 150 minutes per week of moderate-intensity aerobic exercise, with additional benefit up to roughly 450 minutes per week. Above that, the evidence for further incremental mortality benefit weakens, though there's no strong evidence of harm in recreational athletes.

Resistance training predicts longevity through a separate mechanism: muscle mass preservation. Sarcopenia (age-related muscle loss) is one of the strongest predictors of functional decline and all-cause mortality in older adults. Grip strength — a proxy for overall muscle strength — predicts mortality in large cohort studies independently of other health variables (HR ~0.70 per SD increase in strength, Leong et al., Lancet 2015).

Diet: What Survives Scrutiny

The diet-longevity literature is contaminated by confounding, short intervention windows, and misaligned incentives. The following findings are more robust than average:

Caloric restriction (CR) extends lifespan in nearly every organism studied, from yeast to primates. The CALERIE-2 trial — the only multi-year CR RCT in humans — found 12% caloric restriction over 2 years produced favorable changes in metabolic markers, inflammatory markers, and thymic function. Lifespan extension in humans from CR alone remains unproven but biologically plausible.

Protein intake in older adults: The conventional "0.8 g/kg/day" RDA is calculated to prevent deficiency, not optimize muscle preservation. Multiple studies show that 1.6–2.2 g/kg/day supports muscle protein synthesis in older adults, with higher thresholds needed post-exercise. Under-eating protein in older age accelerates sarcopenia.

Mediterranean and MIND diets show the most consistent epidemiological associations with reduced dementia and cardiovascular risk — better than single nutrient approaches. A 2018 RCT (PREDIMED-Plus) found Mediterranean diet with olive oil or nuts reduced major cardiovascular events by 30% versus low-fat diet (n = 7,447, median follow-up 4.8 years).

Ultra-processed food consumption is associated with mortality in multiple large prospective studies, with hazard ratios of 1.14–1.26 per 10% increase in ultra-processed food proportion of diet (NOVA classification). These are observational studies, but the biological mechanisms (displacement of whole foods, additives, ultra-palatability driving overconsumption) are plausible.

Sleep and Biological Aging

The relationship between sleep and aging operates at the cellular level. A 2021 study by Carroll et al. found that self-reported short sleep (< 6 hours) was associated with accelerated epigenetic aging by 1.5–3 years on DNAm clocks in the UK Biobank cohort. The relationship is dose-dependent and holds after controlling for confounders.

Glymphatic clearance — the brain's waste-removal system that clears amyloid-beta and tau proteins — is primarily active during sleep, particularly slow-wave sleep. Sleep disruption accelerates amyloid accumulation in humans measurable by PET scan (Ju et al., 2017). The Alzheimer's-sleep connection is now one of the strongest mechanistic links in aging research.

Seven to eight hours remains the range with the lowest mortality hazard in large meta-analyses, with J-shaped curves showing increased risk both below 6 and above 9 hours.

Biomarkers and Biological Age

The most meaningful development in longevity research in the last decade is the emergence of validated biological age measures. These are not marketing claims — several are based on rigorous epigenetic analysis.

Horvath's DNAm clock (2013) and subsequent "second generation" clocks (PhenoAge, GrimAge) correlate with mortality, disease incidence, and functional status more strongly than chronological age. GrimAge specifically predicts time-to-death with a hazard ratio of ~1.74 per 5-year increment of acceleration — a larger effect than most clinical risk factors.

These clocks are now commercially available through companies like TruDiagnostic, Elysium Index, and others at $100–400 per test. The critical limitation: they are snapshots. Knowing your biological age once tells you little without knowing whether interventions are changing the trajectory.

Other trackable biomarkers with longevity relevance:

• Fasting insulin and HOMA-IR (insulin resistance — a key mediator of metabolic aging)
• HbA1c (3-month glucose average)
• High-sensitivity CRP (chronic low-grade inflammation)
• ApoB (superior to LDL for cardiovascular risk)
• DHEA-S (declines with age; proxy for adrenal function)
• IGF-1 (GH axis; high in youth, lower in aging)

What to Measure

• VO2 max: best measured via graded exercise test with metabolic cart; wearable estimates (Garmin, Polar) have ~10–15% error but track directional changes
• Grip strength: hand dynamometer ($30–100); track dominant and non-dominant; compare to age/sex norms (Leong et al. normative data is published)
• Fasting labs: insulin, glucose, HbA1c, ApoB, hsCRP, Lp(a) — annual panel; most useful tracked over years, not single measurements
• DNAm biological age: test every 1–2 years with a validated clock (GrimAge or DunedinPACE preferred); the delta between tests, not the single score, is what matters
• DEXA scan (annual or biannual): lean mass, fat mass, visceral fat — the most useful body composition snapshot available outside a research setting
• Resting HRV: daily wearable tracking; autonomic health is an aging proxy with longitudinal relevance

What to Experiment With

→ Zone 2 cardio (30–45 min, 4x/week) for 12 weeks → VO2 max estimate and resting HRV trend Tests the most robust longevity intervention in the literature. The dose here is achievable and evidence-based; the wearable-estimated VO2 max provides directional feedback within 8–12 weeks.

→ Protein target increase (1.6 g/kg/day) for 8 weeks → grip strength, DEXA lean mass (if available), and subjective energy Addresses the most common undernutrition pattern in adults over 40. Track dietary protein with a food log app (Cronometer works well) to ensure the intervention is actually implemented.

→ Consistent 7.5-hour sleep window (same bedtime, same wake time) for 4 weeks → HRV, subjective energy, and cognitive speed test (simple reaction time apps) Tests whether your current sleep is suboptimal for your biology. HRV and reaction time are the most sensitive daily proxies for sleep quality effects on biological systems.

→ Time-restricted eating (8-hour window, consistent timing) for 12 weeks → fasting insulin, hsCRP, and body weight Krista Varady's RCTs show significant improvements in cardiometabolic markers from TRE independent of caloric restriction in some studies; others don't replicate this. Testing your own metabolic response with pre/post labs is more informative than the average effect.

Acting Before Full Certainty

Longevity research is advancing faster than clinical guidelines update. The gap between what the evidence supports and what most doctors recommend is particularly wide here. The practical implication: well-characterized, low-risk interventions (Zone 2 exercise, resistance training, adequate protein, sleep optimization) have strong enough evidence to act on now, while higher-risk or more speculative interventions (senolytics, metformin for non-diabetics, aggressive caloric restriction) warrant more caution and closer personal monitoring.