Exercise Looks Like a Powerful Adjunct for Craving. The Authors Rate the Evidence Very Low.

A systematic review and meta-analysis published October 30 in the International Journal of Mental Health and Addiction pooled 76 studies of structured exercise in adults with substance use disorders and came back with two findings that sit awkwardly beside each other. Exercise was associated with substantially lower depressive symptoms and lower craving in people whose primary substance was not nicotine or tobacco. And the authors graded the certainty of that evidence as low to very low.

The effect sizes are the reason the paper is circulating. Among 40 studies of non-nicotine substance use disorders, exercise was associated with a reduction in depressive symptoms of g = −1.29, with a 95 percent confidence interval running from −1.91 to −0.67. For craving, the pooled estimate was g = −0.78, confidence interval −1.26 to −0.30. In conventional terms, a standardised mean difference above 0.8 is usually called large. The depression figure is well past that.

The team, led by Guillermo Francisco Martínez-Montás, searched five databases through September 21, 2026, and split the literature in two: 36 studies of nicotine or tobacco use, and 40 of everything else. The authors reasoned that smoking trials constitute a large and methodologically distinct evidence base. The two halves did not agree.

The half that didn't work

In the tobacco group, the pooled estimates for depressive symptoms and craving were imprecise and the confidence intervals spanned no effect. On the outcome that actually matters for a smoking trial — abstinence — exercise showed no consistent benefit at short-, medium- or long-term follow-up. That null is the cleanest result in the paper, and it will be the one most quoted.

Read against the non-nicotine half, it also argues against the simplest mechanism. If exercise worked chiefly by relieving dysphoria or occupying time, the smoking trials should have shown something on craving. They did not. What differs between the two bodies of evidence is less the substance than the trial: nicotine studies are large, well funded and often run by researchers with decades of infrastructure behind them, while exercise trials in illicit drug use are mostly small pilots with unblinded designs.

That is also the most defensible explanation for the size of the non-nicotine effects. Pooled exercise effects on depression in general adult populations usually land around half a standard deviation. A figure of 1.29 in substance use disorder is more plausibly a signature of small trials, weak blinding and selective reporting than a treatment that outperforms psychotherapy.

What a program should do with it

The authors' own conclusion is narrow and should be quoted that way: exercise may be a useful adjunct for depressive symptoms and craving, and the evidence does not establish a durable abstinence benefit.

For anyone treating dual diagnosis care — depression tangled with substance use — that is still worth having. Exercise is cheap, carries no diversion risk, and targets the anhedonia that makes early recovery miserable. As an add-on to medication and counseling, the risk-benefit calculation is easy. As a replacement for either, nothing in this paper supports it.

There is a second-order observation in the reference list. This is at least the third meta-analysis of exercise for substance use disorders in two years, after 2025 reviews in Experimental and Clinical Psychopharmacology and Behavioral Sciences. The reviews keep accumulating while the field still lacks the one thing that would settle the question: a large, blinded, adequately powered trial with abstinence as the endpoint.

For people looking at treatment options that include physical activity, the practical read is unglamorous. Structured exercise belongs in the schedule. It is not the treatment.