The Diabetes Prevention Program (DPP) randomized trial assigned 3234 nondiabetic persons with elevated fasting and post-load plasma glucose concentrations to placebo, metformin (850 mg twice daily), or a lifestyle-modification program. It measured how often diabetes developed. The authors reported that the lifestyle intervention was significantly more effective than metformin.
What the original trial found
The trial reported that the incidence of diabetes was 11.0, 7.8, and 4.8 cases per 100 person-years in the placebo, metformin, and lifestyle groups, respectively. Compared with placebo, the lifestyle intervention reduced the incidence by 58 percent (95 percent confidence interval, 48 to 66 percent) and metformin by 31 percent (95 percent confidence interval, 17 to 43 percent).
A patient information page from NIDDK, part of the National Institutes of Health, reports that after about 3 years participants in the DPP Lifestyle Change Program lowered their chances of developing type 2 diabetes by 58 percent compared with placebo. It reports 31 percent for participants who took metformin, and says metformin's effect was "to a lesser degree" than lifestyle changes.
What the follow-up reports found
Researchers kept following the same participants. These reports cover longer periods.
| Follow-up | What the source says |
|---|---|
| 10 years | Diabetes incidence in the 10 years since randomisation was reduced by 34% in the lifestyle group and 18% in the metformin group compared with placebo. |
| 15 years, diabetes | Cumulative incidences of diabetes at year 15 were 55% in the lifestyle group, 56% in the metformin group, and 62% in the placebo group. NIDDK reports the same percentages. |
| 15 years, microvascular outcome | The aggregate microvascular outcome was not significantly different between groups (placebo 12·4%, metformin 13·0%, lifestyle intervention 11·3%). |
A 2025 long-term follow-up report of the randomized trial and its Outcomes Study (metformin 850 mg twice a day as tolerated) states that the metformin and intensive lifestyle (ILS) curves "progressively converged" with longer follow-up. It adds that the overall treatment effects "appeared to result from large early effects during the DPP." Its declaration of interests states that one author has stock in United Health Care Group and Merk & Co in an account administered by a financial advisor.
Subgroup and observational analyses
An analysis of the DPP and its outcomes study (metformin versus placebo) concluded, in the authors' words, that "metformin reduces the development of diabetes over 15 years." They added that the subsets who "benefitted the most" included people with higher baseline fasting glucose or HbA1c and women with a history of GDM (gestational diabetes). These are the authors' conclusions from an abstract.
A separate observational follow-up cohort study of the DPP randomized trial concluded that, among adults with prediabetes at baseline, lifestyle intervention, but not metformin, was associated with a lower burden of multimorbidity. An observational study cannot show cause and effect. The abstract does not state a metformin dose.
Weight loss and diabetes risk
A separate NIDDK page, Preventing Type 2 Diabetes, states that you may be able to prevent or delay diabetes by losing 5 to 7 percent of your starting weight. That statement is about weight loss, not metformin.
What this does not tell you
The DPP enrolled nondiabetic persons with elevated glucose and measured diabetes incidence. It is not a longevity or weight-loss trial, and it does not describe people who already have type 2 diabetes. The prescribing information for metformin ER tablets states use as an adjunct to diet and exercise to improve glycemic control in adults with type 2 diabetes mellitus. Use to lower diabetes incidence in people without diabetes is off-label.
The results moved over time: the lifestyle and metformin groups ended up close together at year 15. The excerpts used here do not say which tablet form people took, and they report diabetes incidence and the other outcomes named above, not side effects or risks. The observational result cannot show cause and effect. A licensed provider decides whether any medicine fits a person's situation.