Design: Cochrane systematic review and meta-analysis of randomised trials, GRADE assessed
Population: 26,886 adults with coronary heart disease across 107 trials, mostly post-infarct or post-revascularisation
Primary outcome: mortality, myocardial infarction, revascularisation and hospitalisation at 6–12 months
Effect: MI RR 0.72 (95% CI 0.54–0.95, NNT 71, high certainty); all-cause admission RR 0.65 (0.51–0.82, NNT 17); all-cause mortality RR 0.86 (0.74–1.00)

The Cochrane review of exercise-based cardiac rehabilitation has been updated to 107 randomised trials and 26,886 adults with coronary heart disease, adding 22 trials and 3,456 participants since the last version. Most participants were after a myocardial infarction or revascularisation. Comparison was against a no-exercise control, with follow-up of at least six months.

At 6 to 12 months, exercise-based rehabilitation produced a large reduction in myocardial infarction (RR 0.72, 95% CI 0.54 to 0.95; 25 trials, 8,584 participants; number needed to treat 71), graded high-certainty. It likely reduced all-cause hospital admission (RR 0.65, 0.51 to 0.82; number needed to treat 17, moderate certainty) and likely reduced all-cause mortality (RR 0.86, 0.74 to 1.00, moderate certainty), with cardiovascular mortality in the same direction but less certain (RR 0.87, 0.68 to 1.11). It made little or no difference to CABG or PCI. Quality of life improved across SF-36 and EQ-5D domains, and eight economic evaluations found it cost-effective.

Two features matter for reading it here. Twenty-six of the trials, eleven of them new, came from low- and middle-income countries, and recent trials have used home-based and digitally supported delivery — which removes the usual objection that the evidence came from centre-based programmes in settings unlike ours. Women were 17% of participants overall despite appearing in 78% of trials, so the evidence in women remains thinner than the headline suggests.

Refer every patient after myocardial infarction or revascularisation, not only those who ask
Offer home-based or digitally supported rehabilitation where a centre programme is impractical
Record the referral and whether it was taken up — uptake, not availability, is usually the gap
Be explicit that the evidence base in women is thinner, and do not let that become a reason not to refer

Why it matters

The usual objection — that the evidence comes from unlike settings and centre-based programmes — no longer holds.

The statistics, in plain English

High-certainty for myocardial infarction means further research is unlikely to change the estimate — the strongest grade Cochrane awards. The mortality interval of 0.74 to 1.00 touches 1.0, so the survival benefit is likely but not established, which is why it is described as a small reduction rather than a demonstrated one. A number needed to treat of 17 for hospital admission is a very favourable figure for a non-drug intervention. About a quarter of trials predate modern optimal medical therapy, which may inflate the estimates.