All studies that investigated exercise as a rehabilitation intervention in adult patients with knee or hip osteoarthritis were eligible for inclusion in the review.
Included studies were heterogeneous with respect to the study population, interventions and outcome measures used. The different modalities of exercise studied included static or dynamic, resistive, isokinetic, with immersion, global or analytical lower-limb muscle strengthening, endurance training, balance training, passive mobilisation, muscle and tendon lengthening.
Outcome measures used included analytical evaluations (range of motion, isokinetic or isometric maximal voluntary force, balance tests and posturography, VO2 (oxygen uptake) max, level of spontaneous pain or pain while walking); functional evaluations (six-minute walk test, timed get-up-and-go test, ascending and descending stairs), self-reported measures (Western Ontario and McMaster Universities Osteoarthritis Index, Arthritis Impact Measurement Scale, Activities of Daily Living) or non-steroidal anti-inflammatory drug therapy. Some studies evaluated individual and group exercises separately; others used a combination of group and individual sessions. Some of the exercise programmes were delivered in the community or in the home. The frequency of exercise sessions was one to three per week with a duration of 20 to 90 minutes over a period of five weeks to 24 months.
Two reviewers selected the included articles.