REF-1457 · Haasova et al., 2013 · Addiction · Paper · PubMed

I crave a cigarette mid-session: will it pass if I keep going?

Short walks or rides cut the urge by about 2 SD against sitting in abstaining smokers (19 trials). Rated right after, not mid-session; not athletes, not quit rates.

Main caveat: Who was measured: abstaining smokers in laboratories, not athletes.
fairly strong evidenceMeasured on non-athletes and recreational people

Haasova and colleagues (Exeter and partners) ran the first meta-analysis of the acute effects of short bouts of physical activity on two ratings of cigarette craving, strength of desire and desire to smoke, using raw data from study authors. Eligible trials enrolled temporarily abstaining smokers using no pharmaceutical aid.

Thirty-six studies were identified, 20 were eligible, and raw data came from 19. All control conditions were passive, mostly sitting; some added a recording, a cognitive task, a video or body scanning. Bouts lasted 5-40 min, and craving was rated immediately before and immediately after, or 5 min after, the bout.

Two-stage random-effects models gave −1.91 (95% CI −2.59 to −1.22) for strength of desire and −2.03 (−2.60 to −1.46) for desire to smoke; one-stage models, parallel-arm trials alone, published and unpublished trials separately, and moderate exercise alone (−2.20 and −2.14) agreed. Heterogeneity was high throughout, but every study pointed the same way.

Isometric exercise had the smallest effects, with intervals crossing zero in some cases, and light cycling was not different from passive control. The authors call the effect large in their conclusions and suggest a role in cessation.

Our reading: strong evidence that moving lowers the urge in the short term in people who are not athletes; nothing here says it helps anyone stop for good.

Protocol

Search and selection: 544 database items and other sources, 411 titles, 58 abstracts; 36 studies identified, 20 eligible, raw data unavailable for one (eight participants).
Participants: temporarily abstaining smokers not using cessation drugs; 10-84 per study.
Interventions: moderate walking, running, moderate and vigorous cycling, light cycling, isometric exercise; 5-40 min.
Controls: all passive; mostly sitting.
Outcomes: strength of desire and desire to smoke (baseline-adjusted post-intervention ratings), immediately after or 5 min after the bout.
Funding and interests: internal institutional funds; the authors state no other support and no connection with the tobacco or pharmaceutical industries.

Key Results

AnalysisStrength of desire: SMD (95% CI); I²Desire to smoke: SMD (95% CI); I²
All studies, two-stage−1.91 (−2.59 to −1.22); 94.2%−2.03 (−2.60 to −1.46); 92.0%
All studies, one-stage−1.89 (−2.52 to −1.26)−2.03 (−2.54 to −1.51)
Parallel-arm only−1.78 (−3.17 to −0.40); 96.5%−2.27 (−3.82 to −0.72); 96.8%
Moderate exercise vs control−2.20 (−2.89 to −1.51); 92.1%−2.14 (−2.71 to −1.57); 89.7%
Largest and smallest: one study reached −4.54 and −4.27 and had the highest baseline craving; isometric exercise gave the smallest effects.
Numbers that do not add up: the text gives 15 studies for strength of desire while Table 2 prints N = 16 for the same analysis; the Discussion calls the decrease "moderate" and the Conclusions call the effects "large".

Mechanisms

Measured: lower craving ratings after a bout than after a passive control.
Authors' interpretation: effects may depend on the type, intensity and duration of activity, and on age and dependence (judged by baseline craving); they call for moderator analyses. No biological mechanism is tested.
Inference: craving was rated right after the bout, so "the urge drops as you move" is the nearest reading; how long the drop lasts cannot be taken from these pooled results.

Practical Application

I crave a cigarette mid-session: will it pass if I keep going? Short walks or rides cut the urge by about 2 SD against sitting in abstaining smokers (19 trials). Rated right after, not mid-session; not athletes, not quit rates.
For the athlete with an urge: moderate walking or cycling did at least as well as all activity pooled, so a moderate walk or spin counts; it need not be a hard effort. Isometric exercise did less, and light cycling did not differ from passive control.
What to say about quitting: this is about the urge. Not from this paper; the base's own practice: an athlete who wants to stop is pointed to a doctor or stop-smoking service, as SYNTH-77 does for alcohol dependence, with no product or dose advice.
What it does not show: that exercise raises long-term quit rates; any effect during hard training or in trained athletes; how long the drop lasts; effects in people using nicotine products or cessation drugs.

Limitations

Who was measured: abstaining smokers in laboratories, not athletes.
Very high heterogeneity (I² about 90-97%); the pooled size is an average of very different trials, and the same-direction result is the firmer part.
Self-rated craving in unblinded trials against sitting, so expectation can contribute; one trial used recruitment order for randomisation.
Short horizon: only the minutes after a bout.
Source read is the accepted postprint, not the copy-edited article; five of the 19 trials are theses; supplementary tables were not read.
Funding and interests: internal institutional funds; no tobacco or pharmaceutical ties stated.
How this entry was checked and changed
Why this level

a meta-analysis of randomised trials (rubric A) held one step down: heterogeneity is very high (I² about 90-97%), trials are small and unblinded with self-reported craving, and the outcome is an urge in the minutes after a bout, not smoking behaviour

Full text (checked 2026-10-30)
Among the 19 RCTs included in the meta-analysis, seven studies used a parallel arm design and 12 studies used a cross-over design (full text)
The duration of the PA/control interventions ranged from 5–40 minutes; the number of participants in each study varied from 10–84 (full text)
All craving measures were taken immediately before the intervention and immediately after or 5 minutes after the intervention (full text)
All control conditions were passive (full text)
A summary result of -1.91 (95% CI:-2.59 to -1.22) for SoD and -2.03 (95% CI: -2.60 to -1.46) for DtS; I2 = 94.2% and I2 = 92.0% (full text)
Moderate PA compared with a control condition: -2.20 (95% CI:-2.89 to -1.51) for SoD and -2.14 (95% CI: -2.71 to -1.57) for DtS (full text)
The study investigating light cycling reported no significant differences in change scores between the light cycling and passive conditions (full text)
The research was conducted with the support of internal institutional funds; none of the researchers have any connection with the tobacco or pharmaceutical industries (full text)