REF-1466 · Marti et al., 1988 · Preventive Medicine · Paper · PubMed

Does smoking slow a recreational runner?

Among 4,358 male 16-km racers, smokers were slower even after their lower training was controlled (coefficient 0.10). One-time survey, so no proof of cause. From the abstract.

Main caveat: Abstract only; no effect sizes in time or distance.
moderate evidenceMeasured on recreational to trained athletes

Marti and colleagues (University of Berne) studied smoking, alcohol and endurance capacity in two Swiss samples. Among conscripts, distance in a 12-min endurance run was inversely related to daily cigarette consumption and to years of smoking (P < 0.001), and the association was apparent even in light smokers who had smoked for less than 2 years.

Among race participants, smoking kept a negative, independent association with 16-km race time even when the smokers' lower training activity was controlled. In multiple regression the standardised coefficient for smoking was −0.14 for 12-min distance and 0.10 for 16-km time, the latter despite only 6.9% regular smokers among the joggers.

Alcohol correlated with endurance capacity among joggers in univariate analysis and lost significance in multivariate analysis. Seventy percent of race participants who smoked around the time they took up jogging quit as joggers. The authors conclude, within the limits of the design, that smoking exerts a direct, biologically mediated, deleterious effect.

Our reading: a small standardised association in large samples, which training does not fully explain, in the kind of runner the app serves. The authors' causal wording goes further than a one-time comparison allows.

Protocol

Conscripts: n = 6592, a 20% random sample of all 19-year-old Swiss army conscripts; 12-min endurance run; cigarettes a day and years of smoking.
Race participants: n = 4358, all male participants in a 16-km popular race; 16-km race time; training activity, smoking, alcohol.
Analysis: univariate and multivariate; standardised regression coefficients. Which other variables were in the models is not in the abstract.
Funding and interests: none listed in the PubMed record; no statement in the abstract.

Key Results

SampleOutcomeResult
Conscripts (n = 6592)12-min run distance vs cigarettes a day and years of smokinginverse (P < 0.001); apparent even in light smokers of under 2 years, against non-smokers
ConscriptsStandardised coefficient for smoking−0.14
Race participants (n = 4358)16-km time, with training controllednegative independent association; standardised coefficient 0.10 (higher = slower)
Race participantsRegular cigarette smokers6.9%
Race participantsSmoked when they took up jogging and later quit70%
Race participantsAlcoholsignificant in univariate analysis, not in multivariate
Not in the abstract: the difference in time or distance, confidence intervals, how training was measured, and cigarette counts among the runners.
Numbers that do not add up: the title says "4,100 joggers" and the abstract 4,358 male participants; the analysed subset is not explained.

Mechanisms

Measured: statistical associations in two cross-sections.
Authors' interpretation: a direct, biologically mediated, deleterious effect on endurance capacity, since lower exercise levels did not entirely explain the difference.
Inference: the CO mechanism is plausible (REF-1456, REF-1459), but a survey cannot separate it from what else differs between runners who smoke and those who do not. Quitting by 70% of runners who smoked may also leave a selected group of smokers.

Practical Application

Does smoking slow a recreational runner? Among 4,358 male 16-km racers, smokers were slower even after their lower training was controlled (coefficient 0.10). One-time survey, so no proof of cause. From the abstract.
Light smokers: the "even light, under 2 years" finding is in 19-year-old conscripts on a 12-min run, not in the race runners.
Runners tend to stop: 70% of those who smoked when they took up jogging had quit; this describes this group and is not evidence that running makes people quit.
Alcohol: did not hold up after adjustment here, which sits alongside SYNTH-77 but does not test next-day performance.
What it does not show: how many seconds a cigarette costs; that stopping would recover a given amount; anything in women, in younger trained athletes or in other countries today.

Limitations

Abstract only; no effect sizes in time or distance.
Cross-sectional: no cause, whatever the authors' wording; reverse selection is possible (people who run well may be less likely to keep smoking).
Few smokers among runners (6.9%), and residual confounding is possible: the abstract names only training among the covariates, so training quality, body mass and years of running may or may not be in the model.
Men only; 1980s Switzerland; race entrants self-selected.
Standardised coefficients do not translate to an individual's time.
Funding and interests: none in the record.
How this entry was checked and changed
Why this level

(from the abstract) — a cross-sectional measurement study of at least 100 (rubric B+), with real performance outcomes; it cannot show cause, and the covariates beyond training are not listed in the abstract