REF-1465 · Sandvik et al., 1995 · BMJ · Paper · PubMed · Free full text

Does smoking speed fitness loss with age?

Over 7.2 years, 347 men who kept smoking lost 217 J/kg vs 86 in 791 non-smokers; 199 who quit tracked non-smokers. Healthy men 40-59, not athletes; observational. Abstract.

Main caveat: Abstract only: the PMC article is a page scan; its PDF route stopped at a download check and the scan was not read.
moderate evidenceMeasured on non-athletes and recreational people

Sandvik and colleagues followed 1,393 men aged 40-59 at National University Hospital of Oslo, all healthy at baseline and at follow-up about 7 years later. Fitness was total work done in a symptom-limited bicycle ergometer test divided by body weight, alongside FEV1.

At baseline, 347 persistent smokers were less fit than 791 persistent non-smokers (1349 vs 1618 J/kg) and had lower FEV1 (3341 vs 3638 mL). Fitness fell by 217 (95% CI 185 to 249) J/kg in smokers and 86 (59 to 113) J/kg in non-smokers (P < 0.001); FEV1 by 271 (226 to 316) mL and 116 (85 to 147) mL (P < 0.001).

The differences were practically unchanged after adjustment for age and physical activity. Changes in 199 men who stopped mimicked those of non-smokers, and 56 men who started were close to persistent smokers. Our reading: the longest-horizon number in the base for what smoking costs fitness, with the group that stopped pointing the right way.

It is observational, the men who stopped chose to, and the abstract does not say when they stopped.

Protocol

Participants: n = 1393 men, 40-59 years at baseline, healthy at baseline and follow-up. Groups by smoking over the period: persistent smokers 347, persistent non-smokers 791, stopped 199, started 56.
Timing: baseline 1972-5, follow-up 1980-2; mean 7.2 years.
Outcomes: physical fitness as total work in a symptom-limited bicycle ergometer test divided by body weight; FEV1.
Adjustment: age and level of physical activity.
Funding and interests: none listed in the PubMed record; no statement in the abstract.

Key Results

OutcomePersistent smokers (n = 347)Persistent non-smokers (n = 791)P
Initial fitness (J/kg)13491618
Fall in fitness over 7.2 years (J/kg, 95% CI)217 (185 to 249)86 (59 to 113)<0.001
Initial FEV1 (mL)33413638
Fall in FEV1 (mL, 95% CI)271 (226 to 316)116 (85 to 147)<0.001
Stopped (n = 199) and started (n = 56): changes "mimicked" non-smokers and were "close to" persistent smokers respectively; the abstract gives no numbers for either.
In proportion (our arithmetic): the fall is about 16% of initial fitness in smokers and about 5% in non-smokers.

Mechanisms

Measured: faster falls in work capacity and FEV1 in men who kept smoking.
Authors' interpretation: the decline was considerably greater among smokers and could not be explained by age and physical activity.
Inference: a faster FEV1 fall and a faster fitness fall fit together, but lungs rarely limit healthy people at sea level (PHYS-15, well supported), so the FEV1 fall may be a companion rather than the cause; the abstract does not separate them.

Practical Application

Does smoking speed fitness loss with age? Over 7.2 years, 347 men who kept smoking lost 217 J/kg vs 86 in 791 non-smokers; 199 who quit tracked non-smokers. Healthy men 40-59, not athletes; observational. Abstract.
What stopping may give: in this cohort, men who stopped followed the non-smokers' slope. That is an association over years and does not say how fast; for a one-week measurement see REF-1467.
Reading a masters athlete: a smoker's slower fitness trend may have a cause other than training (SYNTH-35 notes the same kind of caveat for smoking as an uncontrolled factor among sub-4 milers). This is an inference; no athletes were measured here.
What it does not show: that smoking caused the gap beyond doubt; whether training offsets it; the size in women; race performance.

Limitations

Abstract only: the PMC article is a page scan; its PDF route stopped at a download check and the scan was not read. Tables, how activity was measured and the timing of stopping were not seen.
Observational: smoking and stopping were chosen; residual confounding remains after two adjustments.
Selection: only men healthy at both time points, which may understate harm.
Who was measured: middle-aged men, not athletes, in the 1970s and 1980s.
Smaller groups: 199 and 56 for stopped and started, with no numbers in the abstract.
Funding and interests: none in the record.
How this entry was checked and changed
Why this level

(from the abstract) — a large observational cohort of at least 200 (rubric B+); it cannot show cause, the men were not randomised to smoking or stopping, and the outcome is a symptom-limited bicycle test, not a race