Lifestyle intervention, not metformin, linked to less multimorbidity decades later

Design

Observational long-term follow-up of a randomised trial (DPP/DPPOS)

Population

1173 US adults with prediabetes at baseline, median age 74 at analysis

Primary outcome

Multimorbidity (two or more of 15 chronic conditions) through 2021

Effect

Lifestyle vs placebo HR 0.79 (95% CI 0.68–0.93); metformin vs placebo HR 0.91 (0.78–1.07)

The Diabetes Prevention Program randomised 3234 US adults with prediabetes to intensive lifestyle intervention, metformin or placebo in 1996–1999. This analysis followed 1173 participants with Medicare morbidity data through 2021, counting 15 chronic conditions.

By the end, 85% had two or more conditions. The lifestyle group had a lower risk of multimorbidity than placebo (HR 0.79), and the association held when diabetes itself was excluded from the count. Metformin showed no difference from placebo. For dyads of the costliest conditions, the lifestyle hazard ratio was 0.57.

This supports structured lifestyle programmes as the first offer in prediabetes, with benefits that may reach well beyond glucose. Metformin remains reasonable for delaying diabetes in selected patients, but this analysis gives no reason to expect wider protection from it.

• Offer a structured lifestyle programme first in prediabetes, not a prescription alone
• Frame the goal to patients as fewer chronic conditions, not only avoiding diabetes
• Do not present metformin as protection against broader chronic disease
• Note that only about a third of the original cohort contributed data here

Why it matters

Diabetes prevention programmes may be justified by their effect on overall chronic disease, not just diabetes incidence.

Don't overread it

Only 1173 of 3234 participants had outcome data, so this follow-up is observational and selection could explain part of the result.

The statistics, in plain English

HR 0.79 (95% CI 0.68–0.93) means about a 21% lower rate of developing multimorbidity with lifestyle intervention; the interval stays below 1, so chance is an unlikely explanation. Metformin's HR 0.91 (0.78–1.07) crosses 1, so no difference was shown. The randomised groups were compared, but after unmasking and loss of most participants the analysis is observational in character.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.