One number tracks how much oxygen your body can turn into work. It predicts how long you'll live better than almost anything else we can measure, it falls about 10% a decade if you let it, and it responds to training at every age. This page is how to raise it.
Drop in your numbers. If you don't know your VO2max, use the estimate tab — it's within ~10% for most people.
"VO2max is the strongest predictor of mortality you can train" is a big claim, so here is what holds it up — and where it wobbles.
What survives the caveats: VO2max is the strongest measurable correlate of all-cause mortality in the largest datasets we have, the gradient is steep and continuous with no plateau, and it responds to training in months. Effect size × modifiability × speed is why it earns the top slot.
The maximum rate your body can take in, deliver, and use oxygen, in milliliters per kilogram per minute. The physiology is a supply chain:
VO2 = cardiac output × (arterial O2 − venous O2)
Cardiac output is heart rate × stroke volume. The arteriovenous difference is how much oxygen the muscle strips out of each liter of blood.
Genetics set the ceiling. The HERITAGE Family Study found trainability was roughly half heritable: the average person gained ~15–20% after 20 weeks, but individuals ranged from almost nothing to 40%+.
decline per decade after ~30 in typical adults (cross-sectional). Longitudinal data from the Baltimore Longitudinal Study of Aging (Fleg 2005) show it's not linear: ~3–6% per decade in the 20s–30s, accelerating to more than 20% per decade after 70.
per decade in masters athletes who keep training hard. You can't stop the clock — max heart rate falls no matter what — but you can roughly halve the slope, and the gap compounds.
VO2max climbs through puberty as the heart, blood volume, and muscle mass grow, peaking around 17–20 in boys and 14–16 in girls, then plateauing until the decline starts in the late 20s. Teens in structured endurance sport routinely sit above 60 mL/kg/min; the international norms (Tomkinson 2017) put an average 15-year-old boy near 48 and girl near 38. Interval training works in adolescents — meta-analyses show gains of roughly 5–10% in 6–12 weeks — but the bigger lever is simply total activity: youth fitness has declined about 1% per year across 50 countries since the 1980s, almost entirely from inactivity, not genetics. The number you carry into your 20s sets the starting height for the whole decline curve.
Three findings show up in nearly every trial and meta-analysis, and they organize everything below.
Helgerud et al. (2007) put moderately trained men on four matched-work protocols for 8 weeks, 3×/week. 4×4-minute intervals at 90–95% HRmax raised VO2max 7.2%; 47 × 15-second intervals at the same intensity, 5.5%; long slow distance and threshold running, roughly nothing. The gain tracked the increase in stroke volume. Bacon et al.'s 2013 meta-analysis of 37 studies in sedentary-to-recreational adults found the average interval program adds ~3.5 mL/kg/min, and the best-performing protocols — 3–5 minute intervals, ≥12 weeks — approached +7. Milanović (2015) and Wen (2019) put the HIIT-over-moderate-continuous edge at about 1–1.5 mL/kg/min. The operative variable appears to be time accumulated at ≥90% VO2max (Buchheit & Laursen 2013), which is why long intervals and short-interval formats with brief recoveries both work: they keep you up there.
Elite endurance athletes across sports spend ~80% of sessions at low intensity and ~20% at high — "polarized" training (Seiler). Stöggl & Sperlich (2014) randomized trained athletes to four 9-week distributions: polarized produced the biggest VO2max gain (+11.7%), beating high-volume, threshold, and HIIT-only. Easy volume expands capillaries, mitochondria, and plasma volume, raises the ceiling on how much hard work you can absorb, and is what lets the intervals be truly hard. If you only have 3 hours a week, intervals win; if you have 6+, the split matters.
Montero & Lundby (2017) took people who had failed to improve on 1–3 sessions a week and added two more sessions. Every "non-responder" responded. Individual variability is real, but most of it is dose, sleep, and iron — not genetics.
4 × 4 minutes at 90–95% HRmax, 3 minutes active recovery at ~70% HRmax between. Warm up 10 minutes. Total ~35 minutes.
3 sets of (13 × 30 s at ~100% of VO2max power/pace, 15 s easy), 3 min between sets. Or 2–3 sets of 10 × 30 s on / 30 s off. Rønnestad (2015) found 30/15s produced larger VO2max gains than 4×4 in trained cyclists over 10 weeks — the short recoveries keep oxygen uptake pinned near max while limiting lactate. Less mentally brutal than long intervals; needs a bike/erg with power or a good sense of pace.
8 × 20 s at ~170% of VO2max power, 10 s rest — 4 minutes of genuine misery, 5×/week, plus one steady session. Tabata (1996) reported +7 mL/kg/min in 6 weeks in already-trained athletes. Most "Tabata" classes are nowhere near the intensity. Excellent when time-poor; hard to recover from more than 2×/week for most people over 40.
20–40 minutes at or just below lactate threshold (~85–90% HRmax, "comfortably hard"). Raises the fraction of VO2max you can sustain more than VO2max itself. Valuable for racing; secondary for lifespan. Once a week is plenty.
Conversational pace, ~60–70% HRmax, 45–90+ minutes. Nose-breathing possible; you could recite a paragraph. This is the base. It's also where most people go wrong in the other direction: too hard on easy days, then too tired to go hard on hard days. If your "easy" run leaves you unable to talk, it's not zone 2.
Strength work does not blunt VO2max gains in adults training a few hours a week (the "interference effect" only appears at high endurance volumes). It preserves the lean mass that keeps mL/kg/min up, improves running economy, and is itself a mortality predictor. Two full-body sessions a week; put them on easy days or after intervals, not before.
| Starting point | Realistic gain, 12 weeks of 2 quality sessions/week | Ceiling with a year of structured work |
|---|---|---|
| Sedentary | +15–25% (often +5–8 mL/kg/min) | +30–40% |
| Recreationally active | +8–15% | +15–25% |
| Trained (5+ h/week) | +3–8% | +5–12%; gains now come from volume and years |
| Masters athlete, 60+ | +5–12% | Similar in %; recovery is the limiter |
Ranges synthesized from Bacon 2013, Milanović 2015, Wen 2019, Helgerud 2007, Marriott 2019, and the HERITAGE study. Plasma volume gains show up in 2–3 weeks; stroke volume and mitochondrial changes over 6–12; hemoglobin mass over months.
Heart-rate zones are a good-enough proxy for most people. Use a measured max if you have one (the last minute of a 4×4 is a decent field test); the formula is an estimate with ±10 bpm of scatter.
Zone 5 is the 4×4 target. On a bike with power, VO2max intervals sit at roughly 105–120% of FTP; 30/15s at ~110–125%.
Polarized structure: most sessions easy, one or two genuinely hard, strength on the side. The generator progresses interval dose gradually and builds in a lighter week every fourth.
Graded treadmill or bike test to exhaustion with a mask analyzing expired gas. ±2–3% test-retest. Also gives you ventilatory thresholds, which set your zones precisely. Worth doing once as a baseline if you're serious; many university exercise physiology labs run them for $150–300.
Derived from pace-to-heart-rate relationships on outdoor runs and rides. Validation studies put Garmin within about ±5% on average for runners but with individual errors up to 15%; Apple Watch tends to underestimate. They are excellent for trend, unreliable for the absolute number, and blind on the bike unless you have a power meter. Log outdoor runs with GPS and HR to make the estimate better.
Run as far as you can in 12 minutes on a track. VO2max ≈ (distance in meters − 504.9) ÷ 44.73. Correlates ~0.9 with lab values in fit people; underestimates in poor pacers. Repeat every 8–12 weeks under the same conditions.
VO2max ≈ 15.3 × (HRmax ÷ HRrest). Needs a real max heart rate to be useful. Good for a ballpark, useless for tracking small changes.
All validated, all ±10%. The Rockport 1-mile walk test is the safe option for sedentary or older starters. Pick one test and stick with it; the absolute number matters less than the direction.