Twelve studies were included in the review (n=87,577), which appeared to comprise: 10 controlled clinical trials (n=5,648), at least two of which were randomised; one prospective cohort study (n=70,102); and one post-hoc cohort analysis (n=11,827). Duration of follow up varied from one to eight years.
Short term outcomes, diet or exercise alone versus active comparators or controls: Statistically significant decreases in plasma glucose (p<0.05) were noted initially in the intervention groups in four out of five studies of diet. A statistically significant decrease in weight (but not plasma glucose) was noted in the intervention group in one of three studies of exercise.
Short term outcomes, diet plus exercise versus active comparators or controls: Five out of seven studies initially reported a statistically significant reduction in plasma glucose and five reported a statistically significant decrease in weight or body mass index in the intervention groups.
Counselling: Four studies included counselling components. One randomised controlled trial reported increased weight reduction when behavioural email counselling was added to an internet weight loss programme.
Longer term outcomes (12 studies): Three studies of diet and/or exercise versus controls included one (n=577) that reported the interventions significantly decreased diabetes incidence at six years. A second study (n=3,234) reported lower diabetes incidence in the intervention group over 2.8 year follow-up (cumulative incidence 14.4% versus 28.9%), but it was unclear whether this was statistically significant. Post hoc analysis in the third study (n=1,1827) reported higher rates of diabetes in the intervention groups. Five of eight studies of lifestyle interventions reported initial positive results, but noted reduced benefits over time. Two studies of diet and exercise and one of diet alone reported statistically significant weight loss at two years or more. There was no statistically significant difference between the groups at five years (one study).