Gymnastics Did Not Make You Immortal. It May Have Bought You a Few Years.

Every so often a paper arrives that feels personally targeted. In 2024, researchers at the University Medical Center Groningen published an observational study of 95,210 former international athletes across 44 sports and 183 countries. They asked a simple, rude question: compared with people of the same sex, in the same country, dying in the same year, did the sport you lived inside change how long you lasted?

For male athletes, pole vaulting came first: about 8.4 extra years versus the reference population. Gymnastics was second, at 8.2 years. Racquet sports, mixed track and field, fencing, and a cluster of other mixed-skill disciplines followed. Sumo wrestling sat at the bottom, nearly a decade shorter. Volleyball was not far behind it.

If you spent a childhood in a gym, that ranking is allowed to produce a small, private smile. It is not allowed to produce a protocol. This is an association study built from public biographies. It is also one of the largest sport-by-sport lifespan maps we have. Both things can be true.

What they actually measured

Abdullah Altulea, Marco Demaria, and colleagues pulled athletes from Wikidata and English Wikipedia: humans with a recorded sport, birth year, death year, sex, and country. They dropped living people, accidental and unnatural deaths, doping mentions, Paralympic classifications, and sports they judged non-demanding. Categories with fewer than 100 observations per sex were excluded. What remained was 90,931 men and 4,279 women, born as early as 1862, dying between 1960 and 2021 — the window of the World Bank life-expectancy tables they used as the yardstick.

The outcome was age Δ: athlete age at death minus life expectancy for that sex, country, and year. Positive Δ means the athlete outlived the official average. They then ran a robust linear regression of age Δ on sport, separately by sex, with a zero intercept — the modeling choice that says, in the absence of a sport, you would expect no difference from the population tables.

Two numbers from that model should stay taped to the monitor. Sport type explained about 6.2% of the variance in male age Δ, and 3.5% in females. The rest of a life is still diet, infection, war, money, cigarettes, genes, and luck. Sport is not a small thing. It is also not most of the thing.

The leaderboard, without the victory lap

In men, the sports with the largest positive coefficients were mixed more than they were pure. Pole vault and gymnastics demand power, coordination, repeated high-force efforts, and a surprising amount of aerobic work if you have ever watched a full optional set. Targeting sports and fencing — which also track with affluence — landed around six extra years. Race walking, running, sprinting, hurdling, combined events, jumping, rowing, cricket, diving, and sailing occupied the four-to-six-year band. Cycling, which matters in this building, sat in the modest-positive group: roughly two extra years in the regression, with a median age Δ around 3.3. Soccer, gridiron football, basketball, rugby, and weightlifting were smaller still. Boxing and wrestling were essentially flat. Handball, martial arts, mountaineering, volleyball, and sumo pointed the other way.

That ranking does not crown gymnastics as a longevity drug. It says that, among famous dead men who were documented well enough to land in Wikipedia, the ones labeled gymnast or vaulter died later than their national tables predicted, after a crude country-year-sex adjustment. The authors’ own theory is that mixed aerobic-anaerobic sports capture two aging-relevant adaptations at once: the cardiorespiratory and metabolic benefits of endurance work, plus the muscle mass, bone density, and neuromuscular reserve that resist sarcopenia and falls. That theory is plausible. It is not proven by a Wikipedia regression.

This is not a new argument. It is a bigger map.

A 2010 review by Teramoto and Bungum had already concluded that elite endurance and mixed-sport athletes tend to outlive the general population, with lower cardiovascular mortality as the likely engine, while power-sport findings were inconsistent. Lemez and Baker’s later systematic review of hundreds of thousands of athletes pointed the same direction. In 2021, Runacres, Mackintosh, and McNarry meta-analyzed roughly 165,000 former athletes and reported standardized mortality ratios of 0.69 in men and 0.51 in women versus the general population — athletes dying less often than expected — and again found the survival benefit concentrated in endurance and team sports, not in power athletes, whose all-cause and cardiovascular mortality looked like everyone else’s.

The classic Finnish cohort of Sarna, Sahi, Koskenvuo, and Kaprio still earns its keep here. World-class Finnish men competing from 1920 to 1965 lived longest if they were endurance athletes (about 75.6 years), then mixed and team-sport athletes (73.9), then power athletes (71.5), then referents (69.9), after adjustment for occupation and marital status. Cardiovascular deaths drove the gap. Maximum lifespan did not. You can live longer, on average, without moving the ceiling.

Altulea’s paper does not overthrow that literature. It slices it thinner. Where Sarna had a handful of sport families, this study has 44 labeled disciplines and a sample large enough to put pole vault and sumo on the same forest plot. The price of that resolution is the data source.

The part a gymnast should not skip

Wikipedia is not a vital-statistics registry. Athletes enter that corpus because they were visible. Visibility tracks medals, wealth, language, and whether your country bothered to write you down. Sixty-three percent of the observations came from nine countries, mostly in Europe and North America. Japan’s median age Δ was negative for both sexes. Female athletes were 4.5% of the file, missing from large regions of the world, and most female sport categories could not even be modeled. The authors are explicit: interpret the women’s results with caution. Earlier reviews that did have female cohorts often found a survival advantage. This paper’s female signal is weaker and, in several sports, negative. That is more likely a data problem than a biological verdict.

They also could not adjust for education, income, or ethnicity — the variables that decide who gets coached in fencing and who never sees a vault runway. Fencing and archery looking long-lived is consistent with mixed training. It is also consistent with money. Sumo looking short-lived is consistent with extreme mass and a constrained diet. It is also a small, culturally specific professional world. Volleyball’s negative coefficient is the odd one; the authors float chronic musculoskeletal trauma and then admit they cannot close the case.

Selection cuts the other way, too. Only the dead are in the age-Δ calculation. Only the documented dead. Accidental deaths were removed, which is reasonable and also removes a piece of what some sports actually do to people. Doping mentions were removed, which is tidy and incomplete. An eight-year coefficient is not a promise that your optional routine purchased eight years. It is a residual after a blunt adjustment, in a sample that already had to be famous.

What a midlife athlete is supposed to do with this

If the question is “does a serious athletic life look compatible with a longer one,” the weight of evidence — Sarna, Teramoto, Runacres, now Altulea — says yes for endurance and mixed work, and less clearly for pure power sports built around mass and collision. If the question is “should I switch from the bike to the rings because gymnastics won the forest plot,” no. Cycling still sits on the beneficial side of zero. The people in that cycling cell were professionals and internationals, not the Saturday group ride, and they still cleared the population tables.

The transferable pieces are less romantic than a medal count. Keep a large aerobic engine. Keep muscle and bone on purpose, not as an afterthought to zone-2. Do not treat contact, extreme mass, or an unexamined diet as neutral. The same paper that flatters gymnastics also notes that ultra-endurance volume can cut telomeres in the acute setting even as master endurance athletes, in other studies, show longer ones. Dose is a real variable. So is what you do after you stop competing. Former athletes who stay active and smoke less explain part of every longevity gap in this literature. The career is a chapter. The decades after it are the book.

Project Blue’s reading is the unglamorous one. Measure the physiology you can still change: VO2, lactate, strength, lean mass, blood pressure, sleep. Use sport as the vehicle, not the identity. Gymnastics may have been an unusually complete mixed stimulus — power, skill, repeated efforts, some real aerobic demand, and a body that had to stay light. That is a training principle. It is not a prophecy, and it does not survive a knee replacement on its own.

The grown-up summary

A very large, very messy international sample found that male gymnasts and pole vaulters died later than their national averages by a margin that would make any coach sit up. Mixed sports clustered high. Collision, extreme mass, and a few team sports clustered low. Women are still undercounted. Sport explained a sliver of the variance. Famous people on the internet are not a randomized trial.

If you came from the gym, take the compliment. Then go train like someone who intends to use the years.