BackgroundGuidance from the National Institute for Health and Care Excellence stipulates exercise and weight loss as core treatments for osteoarthritis. Typically, NHS physiotherapists and dietitians deliver these treatments within the secondary health-care setting. In view of increasing demands placed upon secondary health-care professionals, we undertook a novel pilot programme consisting of a community-based exercise and dietary intervention led by registered exercise professionals in obese patients with established knee osteoarthritis, to assess the efficacy and potentially widespread use of the intervention within the National Exercise Referral Scheme (NERS) in Wales. Methods80 white patients (body-mass index [BMI] ≥30 kg/m2) from Cwm Taf Health Board in Wales with radiological evidence of unilateral or bilateral knee osteoarthritis were invited (mean age 57·2 years [SD 7·4], BMI 38·5 kg/m2 [7·2]) from orthopaedic, rheumatology, and primary care settings in December, 2012. The exclusion criteria already in use by NERS were adhered to. Patients meeting the eligibility criteria were told by letter about the nature of the programme and were invited to attend an initial assessment. The 16-week pilot delivered dietary and exercise programmes in the community, with two 1-h sessions per week of aerobic and resistance exercise, 1 h per week of dietary education based on Public Health Wales' Foodwise for Life programme, and four practical cookery skills classes. Patients were separated into four groups, containing a maximum of eight patients per group. Primary outcome measures were Oxford knee score, quality of life (EQ-5D and EQ visual analogue scale [EQ-VAS]), and function (6-min walk test and 30 s sit–stand). Secondary measures were bodyweight, resting blood pressure, and resting heart rate. Outcome data were collected at week 16 of the pilot programme (June, 2013). Findings27 patients attended an initial assessment, four of whom declined to attend the programme and one of whom was excluded owing to a newly diagnosed medical disorder. 18 of the initial starting 22 patients completed the 16-week programme (nine men, mean age 59·7 years [SD 5·1], starting BMI 39·5 kg/m2 [8·0]). Seven of the 18 programme completers missed eight sessions or fewer. Statistically significant improvements between baseline and week 16 occurred in the Oxford knee score (mean 3·4, 95% CI 0·0–6·8; p=0·048), 6-min walk test (135·5 m, 92·0–179·0; p<0·0001), 30 s sit–stand (4·6 reps, 3·0–6·2; p<0·0001), EQ-VAS (15·5, 5·6–25·5; p=0·01), weight loss (−3·08 kg, −4·64 to −1·52; p=0·001), and systolic blood pressure (−22·4 mm Hg, −29·8 to −15·0) in those who completed the programme. Non-significant changes occurred in diastolic blood pressure (p=0·96), resting heart rate (p=0·11), and EQ-5D (p=0·11). No adverse events or side-effects were detected in the programme participants. InterpretationA community-based intervention led by exercise professionals for clinically obese patients with knee osteoarthritis could be a useful therapeutic resource for health professionals. Significant improvements in pain scores, function, and quality of life occurred in those who completed the programme. However, there are limitations in the study design because of the absence of a comparator group, the small sample size, and the retrospective nature of the study. The recruitment process used was suboptimal, and further investigation into the optimum referral pathway into such a programme is suggested. Furthermore, the long-term effect of this programme on participants needs detailed ongoing assessment. The programme costs were calculated at £1222 per patient. This substantial expenditure could be reduced to £687 per patient if the programme ran at maximum capacity (eight patients per session). FundingCwm Taf Health Board.