OBJECTIVE: To examine the effects of a GP exercise referral programme on modifiable coronary heart disease risk factors. DESIGN: Randomised controlled trial. A battery of validated measures were carried out at 0, 8, 16, 26, and 37 weeks. SETTING: Two community health centres and a leisure centre in Hailsham, East Sussex. SUBJECTS: 389 patients (smokers, hypertensive or overweight) were selected from medical records, screened for contraindications to exercise and 345 were invited into the study. Of 142 patients randomly allocated, 40 (41%) completed the study in the exercise group and 31 (69%) in the control group. Sixty (35%) invited smokers (48% of non-smokers), 71 (38%) invited hypertensive patients (45% of non-hypertensive patients), and 107 (45%) overweight patients (33% of non-overweight patients) were randomised. Of those randomised, 27 (45%) smokers, 52 (48%) overweight, and 43 (61%) hypertensive patients completed the study. INTERVENTION: The exercise group was offered 20, half price sessions over 10 weeks at a leisure centre. Patients engaged in moderate and vigorous aerobic type activity on various exercise machines, in a semi-supervised, informal environment. RESULTS: 87% of those referred used the prescription and 28% (high adherers)(45% of obese patients) did at least 15 sessions. The exercise group reduced sum of skinfolds by 8.1% (2.9 to 13.3, 95% confidence intervals) more than the control group, up to 16 weeks after baseline. High adherers reduced sum of skinfolds by 9.2% (0.9 to 17.5) more than the control group, up to 26 weeks. High adherers reduced systolic blood pressure by 7.2% (-0.7 to 14.9) (that is, 9 mm Hg) more than low adherers, up to 37 weeks. Non-smokers and obese patients attended more prescribed sessions than smokers and non- overweight patients. CONCLUSIONS: Reduction in sum of skinfolds was maintained up to 26 weeks, among high adherers compared with controls. Reduction in systolic blood pressure was evident up to 37 weeks among high adherers, but only in comparison with low adherers. Selection of appropriate referees and use of other strategies to improve exercise adherence will help to maximise the benefits from GP exercise prescription schemes.