VO2 Max and Longevity: The 122,000-Patient Mortality Study

The longevity conversation of the last two years has fixed on one number: VO2 max, the peak rate at which a person can take in and use oxygen during all-out exercise. Wearables now print an estimate on the wrist, and that exposure has pulled a piece of exercise physiology into mainstream health talk. The primary evidence behind the claim is older and quieter than the podcast version — and it is worth reading at the source rather than through three layers of summary.

This is the academic entry in our fitness-and-mortality series: it walks the 122,007-patient Mandsager 2018 cohort, the 2016 American Heart Association statement that argued for treating fitness as a vital sign, and Kodama's 2009 meta-analysis that fixed the per-MET dose-response. It then states plainly what the data does and does not license you to conclude.

The Mandsager 2018 cohort: 122,007 patients

Mandsager and colleagues at the Cleveland Clinic studied 122,007 consecutive adults who underwent a symptom-limited exercise treadmill test between January 1991 and December 2014. Mean age was 53.4 years (SD 12.6) and 59.2% were men.^[1] Fitness was recorded as peak estimated metabolic equivalents (METs) achieved on the treadmill, then converted to age- and sex-matched percentiles so a 60-year-old woman and a 35-year-old man were each scored against their own reference group.

Patients were sorted into five fitness bands by percentile, and all-cause mortality was tracked over a median follow-up of 8.4 years (interquartile range 4.3 to 13.4):

Fitness band     Percentile        vs Elite   Adjusted HR   95% CI
──────────────────────────────────────────────────────────────────
Low              under 25th        reference   5.04          4.10–6.20
Below average    25th–49th
Above average    50th–74th
High             75th–97.6th        Elite      0.77          0.63–0.95
Elite            97.7th and up      —          1.00 (ref)    —

Below average vs above average:    HR 1.41    (95% CI 1.34–1.49)

Two readings matter. First, the size of the gap: the least-fit quartile died at 5.04 times the rate of the elite group.^[1] Second, and the headline the authors emphasized, there was no upper limit. Elite fitness beat even the high-fitness band (HR 0.77, 95% CI 0.63–0.95), so the curve kept bending downward at the extreme rather than flattening into a plateau.

Why "no upper limit" was the surprise

A long-running worry held that extreme endurance training might carry cardiac costs — atrial fibrillation, coronary calcification — that could erase the survival edge at the top. The Mandsager data found no such reversal in this cohort: the most-fit 2.3% had the lowest mortality of anyone, and senior author Wael Jaber framed the result as showing no observed ceiling to the benefit.^[1]

The careful version: "no upper limit" describes this referral population over this follow-up, not a license to chase unlimited training volume. The cohort was patients sent for stress testing, the comparison is between percentile bands rather than absolute VO2 max values, and the result speaks to fitness level, not the dose of exercise that built it.

Fitness against the traditional risk factors

The reason this study traveled is the comparison it ran against the risk factors clinicians already screen for. In the same adjusted model, the mortality risk of being unfit was on par with or larger than the classic ones:^[1]

Risk factor               Adjusted HR for all-cause mortality
──────────────────────────────────────────────────────────────
Low vs elite fitness       5.04   (95% CI 4.10–6.20)
Coronary artery disease    1.29   (95% CI 1.24–1.35)
Diabetes mellitus          1.40   (95% CI 1.34–1.46)
Smoking (current)          1.41   (95% CI 1.36–1.46)

Read carefully, the comparison is between a five-band fitness contrast and single binary risk factors, so the numbers are not strictly like-for-like. The defensible takeaway is the one the AHA drew: poor cardiorespiratory fitness sits in the same risk tier as the conditions medicine treats aggressively, yet it goes unmeasured in most clinical encounters.^[2]

The 2016 AHA statement: fitness as a vital sign

Two years before Mandsager, the American Heart Association published a scientific statement, first author Robert Ross. It argued that cardiorespiratory fitness, quantified as VO2 max or peak, should be assessed routinely as a clinical vital sign alongside blood pressure and heart rate.^[2] The case rested on the consistency of the fitness-mortality association across decades of cohort data, and on fitness adding predictive value beyond the standard risk panel.

The statement also conceded the practical obstacle: a full cardiopulmonary exercise test with gas exchange is expensive and not available in primary care. Its compromise was to endorse submaximal and non-exercise estimating equations as a first pass — the same family of estimates our VO2 Max Estimator produces from a field test — accepting their error in exchange for getting a number into the chart at all.

Kodama 2009: the per-MET dose-response

Where Mandsager gives the categorical contrast, Kodama's 2009 JAMA meta-analysis gives the continuous slope. It pooled 33 cohorts: 102,980 participants and 6,910 deaths for the all-cause analysis, plus 84,323 participants and 4,485 events for cardiovascular endpoints.^[3]

Outcome              Per 1-MET higher fitness   Low vs high band
─────────────────────────────────────────────────────────────────
All-cause mortality  RR 0.87 (95% CI 0.84–0.90)  RR 1.70 (1.51–1.92)
CHD / CVD events     RR 0.85 (95% CI 0.82–0.88)  RR 1.56 (1.39–1.75)

MET bands used:  low under 7.9  ·  intermediate 7.9–10.8  ·  high 10.9+
1 MET = 3.5 mL O2 per kg per minute

Each 1-MET of additional fitness lowered all-cause mortality risk by about 13% (RR 0.87) and cardiovascular events by about 15% (RR 0.85).^[3] Because one MET equals 3.5 mL O2/kg/min,^[4] a single MET is a modest, trainable increment — the difference between a VO2 max of 35 and 38.5 ml/kg/min. Adding three or four METs over a training year is realistic for an untrained adult, and it moves the risk needle meaningfully.

Reading the studies in METs vs ml/kg/min

The literature toggles between two units, which is a frequent source of confusion. Treadmill cohorts like Mandsager report peak METs; the longevity-podcast framing usually quotes VO2 max in ml/kg/min. They convert directly:

VO2 max (ml/kg/min) = METs × 3.5
METs                = VO2 max (ml/kg/min) / 3.5

Examples
  7.9 METs  (Kodama low cutoff)        = 27.7 ml/kg/min
  10.9 METs (Kodama high cutoff)       = 38.2 ml/kg/min
  1 MET gained                         = +3.5 ml/kg/min

So Kodama's "low" fitness band tops out near a VO2 max of 28, and "high" starts around 38.^[3][4] The 13%-per-MET slope is the same fact stated per 3.5 ml/kg/min. Hold the units straight before comparing your wrist number to any published threshold.

The June 2026 angle: your wrist number vs the studies

The reason this physiology went mainstream is the wearable estimate. Apple, Garmin, and others now infer VO2 max from heart-rate and pace during ordinary walks and runs, which is what put a longevity metric on millions of wrists. The catch is accuracy. A 2025 PLOS ONE validation study tested the Apple Watch against laboratory cardiopulmonary exercise testing in 28 adults at University College Dublin.^[5]

Apple Watch vs lab CPET (Lambe 2025, n = 28)
  Mean bias              −6.07 ml/kg/min (watch underestimates)
  95% CI of bias         3.77 to 8.38 ml/kg/min
  Mean abs. % error      13.31%  (95% CI 10.01–16.61)
  Bland-Altman limits    −6.11 to +18.26 ml/kg/min

The watch underestimated true VO2 max by about 6 ml/kg/min on average — roughly 1.7 METs.^[5] That single-device error is wider than the gaps between several of the mortality categories above, and the Bland-Altman limits span more than 24 ml/kg/min. The authors concluded the estimate needs refinement before clinical use. The practical move: read the wearable as a relative tracker — is my number rising over months — and reserve the absolute value for a measured test if a clinical decision rides on it.

What the data licenses, and what it does not

Starting point

1. Get a baseline VO2 max from a field test — Cooper 12-minute, 1-mile walk, or step test — through the VO2 Max Estimator, and convert to METs with the ×3.5 relationship.
2. Locate yourself against the Kodama bands (low under 7.9 METs, high above 10.9) to see which percentile tier you are starting from.
3. Build the aerobic base that drives the number: most of the gain comes from sub-threshold volume, framed in Zone 2 methods and limits, set against your Zone 2 heart-rate ceiling.
4. Add the high-intensity work that lifts the ceiling, covered in How to Improve VO2 Max.
5. Re-test every 8 to 12 weeks. A 3-to-4 MET gain over a year is realistic for an untrained adult and, on Kodama's per-MET slope (RR ≈ 0.87 per MET), maps to roughly a 34–43% lower all-cause mortality risk — an observational association across populations, not a guaranteed individual effect.^[3]

Related in this series

• How to Improve VO2 Max — the interventions with the largest meta-analytic effect on aerobic capacity.
• Zone 2 Training Methods and Limits — how the aerobic-base zone is defined and measured.
• Concurrent Training Interference — what happens when endurance and strength share a training week.

Tools: VO2 Max Estimator, Zone 2 Heart Rate Calculator, Heart Rate Zone Calculator, Race Time Predictor.