The Diabetes Prevention Program randomised adults with prediabetes to intensive lifestyle, metformin or placebo in the late 1990s. This report links 1,173 of those participants to United States Medicare claims through 2021, by which point the median age was 74, and asks who accumulated multiple chronic conditions — defined as two or more from a list of 15.
By the end of follow-up 85% had at least two conditions, with a median of five. Allocation to lifestyle was associated with a 21% lower risk of multimorbidity than placebo (HR 0.79, 95% CI 0.68 to 0.93). Metformin was no different from placebo (HR 0.91, 95% CI 0.78 to 1.07). The gap widened when the analysis was restricted to pairs of the costliest conditions (HR 0.57, 95% CI 0.38 to 0.85). It held when diabetes itself was removed from the definition, so this is not simply delayed diabetes being counted twice.
This is observational follow-up in consenting survivors, not a randomised comparison of this outcome, and only about a third of the original cohort contributed — selection is a real concern. But it is the longest evidence we have that a structured lifestyle programme still shows up decades later, and it argues against treating metformin as an equivalent substitute for the programme in prediabetes. In practice it supports spending the clinic time and the referral effort, rather than writing metformin and moving on.
- In prediabetes, offer a structured lifestyle programme first; metformin is an addition to it, not a replacement for it.
- Record which patients actually enrolled and attended, not just that a programme was advised.
- Follow up on weight and activity, not HbA1c alone — the long-term signal here tracked the intervention, not the glucose.
- Ask about the rest of the multimorbidity list — hypertension, lipids, heart failure, depression, chronic kidney disease — at prediabetes visits, not only once diabetes is diagnosed.
- Where no formal programme exists, agree one specific measurable change per visit rather than giving general advice.
The statistics, in plain English
A hazard ratio of 0.79 means the lifestyle group reached multimorbidity at about four-fifths the rate of the placebo group over follow-up. Its interval, 0.68 to 0.93, lies below 1.0 throughout, so a real reduction is likely. Metformin's interval, 0.78 to 1.07, crosses 1.0 — the data are compatible with a modest benefit, no effect at all, or slight harm, so it cannot be called effective here. The costliest-pairs result of 0.57 looks larger but rests on fewer events, which is why its interval is so much wider. And because this is observational follow-up of consenting survivors rather than the original randomised comparison, the groups may differ in ways no statistical adjustment can fix.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for diabetes & endocrinology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free