A 21-year follow-up of the Diabetes Prevention Program and its outcomes study has added a new question to familiar prediabetes evidence: not only whether diabetes is delayed, but whether several chronic conditions accumulate over time. The analysis covered 1,173 high-diabetes-risk adults enrolled in Medicare who had originally been randomized at 27 US sites to intensive lifestyle intervention, metformin or placebo. [1][2]
What “lifestyle intervention” meant
This was not simply a prompt to exercise more. The lifestyle arm began with 16 individual sessions, then monthly support, with targets to reduce calories and fat, complete at least 150 minutes of activity each week and lose at least 7% of starting weight. [1]
By the end of follow-up, 997 participants — 85% of the cohort — had developed at least two chronic conditions. The proportion was 82% in the lifestyle group, compared with 85% in the metformin group and 87% in the placebo group. In the adjusted analysis, the lifestyle group had a hazard ratio of 0.79 versus placebo for developing at least two conditions; NIH describes that as a 21% lower risk. It also reported a 25% lower risk of developing at least three conditions. [1][2]
Metformin did not show a statistically significant difference from placebo for this multimorbidity outcome. That distinction matters because the trial is not a contest between a “natural” option and a drug: it is a long-term comparison of three original randomized strategies, measured here through later Medicare claims. [1][2][3]
What the result cannot settle
The study does not show that every person with prediabetes should receive the same programme, or that a single behaviour prevents all chronic disease. It assessed 15 claims-defined conditions in an older, Medicare-linked cohort with a median age of 74, so its clearest inference is about that population and outcome. [1][2]
For learners, the useful lesson is to separate a structured, supported intervention from a simplistic calorie-burn story — and to ask which outcome, population and follow-up period a result actually covers.
