Twenty one studies (n=1,323, range 10 to 155) and 11 case studies met the inclusion criteria. Five studies had no comparator group, four had non-comparable controls and six had comparable controls. Six studies were retrospective. Overall study quality was rated: C for retrospective studies; 3As, 3Bs and 3Cs for before-after studies; and 2As and 2Cs for controlled trials. The risk of bias was usually C regardless of study design.
Quality of life was significantly increased with exercise in two poor quality studies, but not in a good-quality RCT. Upper and lower body strength was significantly increased with exercise in 10 controlled studies; effect sizes were smaller where comparable controls were used (three studies). Habitual activity and/or exercise significantly increased maximum oxygen intake (10 studies), aerobic capacity (three studies) and maximum metabolic equivalents (seven studies).
Patients in the exercise groups showed: lower incidence of metabolic syndrome (one study); greater incidence of glucose tolerance (one study), but higher fasting glucose (one RCT) and lower postprandial glucose (one study); no difference in blood pressure (three studies) or hypertensive medication use (two RCTs); and increased muscle parameters (three studies), bone gain (one study) and serum alkaline phosphatase (one study).
Results for blood lipid measures (six studies), body fat (five studies), immune function (two studies) and adverse events (two studies) were inconsistent. Four out of five studies reported no positive effect of exercise on kidney function.
Results of the case studies were summarised in an online supplementary file.