Differential physical activity responses to diabetes risk reduction interventions according to prediabetes phenotype in at-risk individuals: pooled analysis from three randomised controlled trials

Abstract Aims/hypothesis Emerging research suggests that the effects of lifestyle interventions to reduce the risk of type 2 diabetes progression differ by prediabetes phenotype. We investigated whether this divergent effectiveness is accompanied by differing physical activity responses to the intervention. Methods Pooled data from three randomised controlled trials were used ( n =2150), with up to 48 months annual follow-up. Participants at risk of type 2 diabetes were randomised to either: (1) usual care (control) or (2) similar interventions, in which they received structured group-based lifestyle interventions to promote increases in physical activity and other behaviours associated with diabetes risk reduction, supported by education, goal setting, pedometer use and maintenance support. Change in physical activity and glucose levels were stratified by the presence of impaired fasting glucose (IFG) and/or impaired glucose tolerance (IGT). Results Participants in the group with normoglycaemia were 63 years old (mean), with a mean BMI of 29.9 kg/m 2 ; 56.4% were men and 76.7% were White European. Participants in the IFG group were 63 years old, with a BMI of 29.8 kg/m 2 ; 70.0% were men and 79.1% were White European. Participants in the IGT group were 65 years old, with a BMI of 31.1 kg/m 2 ; 60.4% were men and 80.5% were White European. Participants in the group with both IFG and IGT were 65 years old, with a BMI of 31.7 kg/m 2 ; 62.7% were men and 82.6% were White European. The number of baseline steps/day was 6130 ± 3201 in the normoglycaemia group, 7244 ± 3489 in the IFG group, 5984 ± 2904 in the IGT group and 6310 ± 2872 in the IFG+IGT group (means ± SD). The intervention increased the mean number of steps/day relative to usual care by 247 (95% CI 4, 491) in the normoglycaemia group, 523 (95% CI 229, 818) in the IGT group, and 808 (95% CI 402, 1213) in the IFG+IGT group, but not in the IFG group (mean 71; 95% CI −429, 571). The intervention improved 2 h post-challenge glucose levels in the IFG+IGT group only (mean −0.5 mmol/l; 95% CI −0.9, 0.0). Conclusions/interpretation The effectiveness of structured, group-based interventions with goal setting to promote physical activity differs by glycaemic status; those with IFG were least responsive. This may contribute to previously observed differences in the clinical effectiveness of diabetes prevention programmes.