REF-1579 · Marijon et al., 2013 · European Heart Journal · Paper · PubMed

Does pro-level cycling shorten life?

Not shown: French Tour men's mortality was 41% below French men's (SMR 0.59, 0.51-0.68), but selection could hide harm and doping is not separated. Heart symptoms: a doctor.

Main caveat: Abstract only. No full text is held; a published editorial was not read.
limited evidenceMeasured on elite to world-class athletes

Vital status and characteristics of all French participants were collected for 1947-2012, with causes of death available from 1968. By 1 September 2012, 208 (26%) had died; neoplasms and cardiovascular disease accounted for 61% of deaths. The advantage did not change over time (P = 0.70). For external causes no difference was shown (SMR 1.06, 0.71-1.53).

Our reading: the use here is as an answer to a fear, not as a recommendation. Reaching the Tour, across eras, was not followed by shorter lives on average.

Protocol

Who: all French men who started at least once; one start is enough to count.
Comparator: French male population, overall and cause-specific rates.
Not in the abstract: how vital status was ascertained, when follow-up began for each rider, career length, smoking, wealth, and any measure of doping.
Funding and interests: not in what was read.

Key Results

Cause of deathSMR (95% CI)P
All causes0.59 (0.51-0.68)<0.0001
Neoplasms0.56 (0.42-0.72)<0.0001
Cardiovascular0.67 (0.50-0.88)0.004
External causes1.06 (0.71-1.53)0.80
Time trend: no change over time was detected (P = 0.70), which does not show the eras were alike.
Cause data start in 1968, so the cause-specific rows do not cover the deaths of earlier years.
External causes is a non-significant result, so a difference from the general population is not shown; it is not shown to be equal.

Mechanisms

Measured: none; only survival and registered cause of death.
Authors' interpretation: the balance between the positive effects of high-level activity, selection of healthy athletes and the possible harm from excess exercise or alleged doping cannot be assessed in detail.
Inference: a healthy-entrant effect is likely, since anyone able to start the Tour is already free of many diseases, and the population rates include people ill from youth. The lower cancer mortality may also reflect low smoking in this group, which the abstract does not measure.

Practical Application

Does pro-level cycling shorten life? Not shown: French Tour men's mortality was 41% below French men's (SMR 0.59, 0.51-0.68), but selection could hide harm and doping is not separated. Heart symptoms: a doctor.
Good for: an athlete or family member afraid that very hard endurance training must shorten life. This group gives no average signal of that.
Not good for: any claim that racing or high volume lengthens life. The design cannot show it, and the authors say so.
Not good for: individual risk: an average advantage can coexist with an individual heart problem (REF-1185, REF-1577, REF-1578). Palpitations, chest pain, fainting: a doctor.
Not good for: women or amateurs: no data.

Limitations

Abstract only. No full text is held; a published editorial was not read.
General population as comparator: healthy-entrant selection, and likely differences in smoking, income and access to medicine, with no individual adjustment read.
Doping is unmeasured and named by the authors as unassessable.
Cause-of-death data only from 1968.
Men only, one country, one race, with one start enough to enter.
No training exposure, so no dose can be read.
How this entry was checked and changed
Why this level

(from the abstract) — n ≥ 200 would earn the observational B+; one step down because only the abstract was read and the comparator is an external general population with no matched controls and no individual covariates reported, as in the sub-4-minute-mile entry REF-1186, also B