Diet, exercise, drugs may prevent diabetes in high-risk patients

Clinical Question

Can the onset of diabetes be delayed or prevented in people with impaired glucose tolerance?

Bottom Line

Diet, exercise, or diet and exercise changes, at least those in study situations, will slow the progression of diabetes by approximately 50% in patients with impaired glucose tolerance. Drug therapy with either oral diabetes drugs or the weight loss drug orlistat (Xenical) will also slow progression. The preventive effect of the drugs is not maintained when they are stopped, and research has not been conducted for long enough to determine whether diabetes onset is prevented or just delayed. (LOE = 1a)


Gillies CL, Abrams KR, Lambert PC, et al. Pharmacological and lifestyle interventions to prevent or delay type 2 diabetes in people with impaired glucose tolerance: systematic review and meta-analysis. BMJ 2007;334:299.  [PMID:17237299]

Study Design

Systematic review




Various (meta-analysis)


To answer this question, researchers completed a thorough search of 4 databases, contacted experts, and checked references of identified studies. They included only randomized controlled trials -- most were not blinded -- that evaluated medicine or lifestyle changes to prevent the onset of type 2 diabetes in patients with impaired glucose tolerance. They included research in all languages. Two authors independently assessed the validity of and abstracted data from the studies, excluding 6 studies not considered appropriately randomized and 3 studies for which they could not obtain additional data from the investigators. Since not all data were reported in the same way in all of the studies, the authors had to estimate some data . However, they used the more conservative random effect model when combining the data. The authors included 17 studies of 8084 patients in their analysis, including 2 studies conducted in Japan and 3 done in China. Most of the studies were several years in length. The baseline risk of diabetes in the studies was 37% over 5 years. Overall, the interventions decreased the onset of frank diabetes by approximately half (hazard ratio [HR] = .51; 95% CI, .44 - .60). Diet changes, exercise, or the combination produced a similar reduction in risk (HR = ~.50). Two types of drug therapy were used. The oral diabetes drugs acarbose (Precose), glipizide (Glucotrol), metformin (Glucophage), or the biguanide flumamine decreased the onset of diabetes by 30% (HR = .7; .62 - .79). The anti-obesity drug orlistat also decreased the likelihood by a similar amount (HR = .44; .28 - .69). Using rough estimates, the number needed to treat to prevent one patient from developing diabetes would be in the range of 5 to 10 for lifestyle changes or drug treatment. Although less well studied, it seems that the rate of diabetes returns to baseline once drug therapy is stopped, and no studies have been conducted for long enough to determine whether diabetes is truly prevented or simply delayed.