Twenty studies (24 articles) were included in the review: 10 randomised controlled trials (n between 614 and 721, duplicate numbers unclear); seven non-randomised controlled trials; and three uncontrolled trials. Sample size among the randomised controlled trials varied from 15 to 251. Only one trial reported adequate allocation concealment. Three trials conducted blinded assessment. Most randomised controlled trials adequately described eligibility criteria, intervention and outcomes measures. Most randomised controlled trials used suitable statistical methods. Only about half reported measures of variability. Only four studies in the review reported sufficient data to permit calculation of effect sizes. Few trials reported how missing data were handled.
Glycaemic control: Two out of nine randomised controlled trials reported clinically relevant reductions in glycosylated haemoglobin of 1 to 1.2 per cent from over 8.0 per cent. Two randomised controlled trials reported that resistance training was as effective or more effective than aerobic training in reducing glycosylated haemoglobin. Improvements diminished during long-term maintenance programmes (two randomised controlled trials). One randomised controlled trial (out of seven) reported a statistically significant improvement in fasting blood glucose level of 3.2 millimoles per litre (which was not evident in controls having aerobic training), but the authors noted clinical differences and possible bias in this study.
Insulin sensitivity: Resistance training significantly improved oral glucose tolerance when compared with sedentary controls (p<0.05, one randomised controlled trial), but not when compared to active controls (one randomised controlled trial). Four out of five randomised controlled trials using homeostasis model assessment to determine insulin resistance reported statistically significant improvements in the resistance training group. Outcomes in these four trials were reported: as changes from baseline (two randomised controlled trials); in comparison with inactive controls (one randomised controlled trial); and in comparison with aerobic training (one randomised controlled trial). One randomised controlled trial using the insulin sensitivity index reported no statistically significant findings.
Insulin signalling (one randomised controlled trial): Muscle glycogen levels improved significantly in the resistance training group compared to controls (p=0.04), but there was no evidence of change in GLUT4 gene or protein expression.
Adverse events (six randomised controlled trials): Hypoglycaemic events were reported in all study groups, but only one case required medical attention. Resistance training appeared to be well tolerated.
Results of other outcomes and of non-randomised trials were also reported in the review.