Exercise matched cognitive therapy for depression, but the trials were shaky

Exercise is one of the most commonly suggested non-drug approaches for low mood, and it is cheap, accessible and carries side benefits for physical health. But a recommendation is only as good as the trials behind it, and for years the evidence on exercise for depression came from scattered small studies. This review set out to answer two questions at once: does exercise actually reduce depressive symptoms compared with doing nothing, and how does it stack up against an established treatment, cognitive therapy? Just as importantly, it asked whether the studies making those claims were any good.

The authors ran a systematic review and meta-regression of randomised controlled trials gathered from five electronic databases, Medline, Embase, Sports Discus, PsycLIT and the Cochrane Library, plus expert contacts, bibliographic searches and hand searches of journals. Fourteen trials were analysed. The main outcomes were the standardised mean difference in effect size and the weighted mean difference in Beck depression inventory score, comparing exercise against no treatment and against cognitive therapy. The reviewers also formally assessed each trial's methods, checking randomisation concealment, intention to treat analysis and blinding of outcome assessment.

Compared with no treatment, exercise reduced depression symptoms with a standardised mean difference of 1.1 in favour of exercise, and Beck depression inventory scores fell by 7.3 points more than in untreated groups. Against cognitive therapy, exercise looked similar, with a standardised mean difference of 0.3 whose confidence interval crossed zero. But quality problems ran through everything: randomisation was adequately concealed in only three studies, intention to treat analysis appeared in only two, and outcome assessment was blinded in just one. Effects were also significantly larger in trials with shorter follow-up and in two trials reported only as conference abstracts, a pattern that often signals inflated results.

This review complicated a popular belief rather than confirming it. Exercise clearly performed well in the available trials, both against no treatment and alongside cognitive therapy, yet the reviewers judged the evidence base too weak to declare effectiveness, particularly for clinically diagnosed populations with adequate follow-up. Most participants were community volunteers whose depression was defined by a questionnaire score rather than a clinical diagnosis. The open questions it left were pointed: whether the benefit survives in rigorous trials, whether it lasts beyond short follow-up, and whether it holds in people with formally diagnosed depression.

There is a genuine signal here that moving your body is associated with better mood in trial settings, and that in the studies available it kept pace with a recognised psychological therapy. At the same time, this review is a lesson in reading evidence: large effects from flawed, short trials tend to shrink when the science tightens up. It is a reason to take exercise seriously as one plausible support for mental well-being, not proof of a treatment effect of any particular size. Anyone dealing with significant depression is in territory this review explicitly says the trials did not properly cover.