A prediabetes result raises a reasonable question: would taking metformin now prevent diabetes later? There is trial evidence for delaying the diagnosis in selected adults at high risk. There is also a large gap between that finding and the claim that metformin is a longevity medicine. The useful discussion begins with the outcome being sought, rather than the reputation of the drug.
First, establish what is being prevented
Prediabetes describes glucose results above the usual range but below the threshold for diabetes. It is a risk category, not a prediction that an individual will inevitably develop the condition. The NIDDK explanation of insulin resistance and prediabetes describes how clinicians use HbA1c, fasting plasma glucose and sometimes an oral glucose tolerance test to assess it. These tests measure different aspects of glucose regulation.
That distinction matters before discussing a prescription. A laboratory result, its trajectory and the person’s wider history belong in the same conversation. The question is whether someone has a sufficiently high likelihood of progression for the expected benefit of medication to justify its burdens. A label on a report cannot answer that on its own.
Nor should a prevention discussion replace assessment of symptoms. NHS guidance on diabetes symptoms advises seeing a GP for possible diabetes, including persistent thirst, passing urine more often and unexplained weight loss. Some people have no symptoms. Waiting for a noticeable change is therefore not a substitute for agreed follow-up testing.
What the prevention trial actually tested
The US Diabetes Prevention Program compared an intensive lifestyle programme, metformin and placebo in adults already at high risk. In its initial approximately three-year phase, lifestyle intervention reduced diabetes incidence by 58% relative to placebo, and metformin by 31%. These are relative reductions, not the proportion of participants permanently protected. The NIDDK summary of the DPP explains the original findings.
The lifestyle arm was a supported programme, rather than a leaflet telling people to try harder. The CDC description of the trial and prevention programme reports a focus on dietary change, at least 150 minutes of weekly activity and weight loss of around 5–7%. Those were study goals within an organised intervention, not a universal exercise or weight-loss prescription.
This is an important comparison to preserve. A trial of structured support does not tell us that a person who struggles with unsupported changes has failed. It tells us what was tested. Access, disability, caring responsibilities and the feasibility of a programme belong in a clinical discussion about what support can realistically look like.
Delay can matter without being a cure
A 2025 analysis in The Lancet Diabetes & Endocrinology examined the DPP and its follow-up over approximately 21 years. Compared with original placebo assignment, diabetes incidence rates were 24% lower in the original lifestyle group and 17% lower in the metformin group. Median diabetes-free survival was extended by 3.5 and 2.5 years, respectively.
The follow-up was not an unchanged, blinded trial lasting two decades. Placebo was discontinued, metformin continued openly, and lifestyle support was offered across groups. The authors reported that much of the overall separation came from the early trial period. These details make the long-term results informative, but harder to read as a simple comparison of two treatments continuously delivered under identical conditions.
What we have is evidence that the initial interventions shifted the timing and cumulative incidence of diabetes. That is useful. It does not mean everyone avoided diabetes, or that a particular reader can expect exactly another two and a half years without it. Group averages describe the study; they do not specify an individual’s future.
Where a UK prescribing discussion fits
NICE’s metformin recommendation, updated in 2017, places clinical judgement at the centre. It concerns people whose HbA1c or fasting plasma glucose results have deteriorated and who either experienced that deterioration despite participating in an intensive lifestyle-change programme or are unable to participate in one. NICE highlights particular consideration for people with a BMI greater than 35. Metformin supports lifestyle change within that decision.
This is narrower than offering a tablet to every person with a borderline result. It also avoids treating access to a programme as a moral test. The discussion needs to cover potential benefit, limitations and whether the proposed approach is feasible. The recommendation remains guidance for a prescriber to apply to an individual, rather than an eligibility rule readers can use to prescribe for themselves.
The DPP follow-up used standard-release metformin at 850 mg twice daily as tolerated. That describes the research exposure, not a starting dose to copy. Product, dose and review arrangements require an individual prescription. Anyone already taking metformin for another indication needs their own treatment plan considered before drawing conclusions from a prevention article.
The safety discussion deserves equal space
NHS medicine information lists nausea, diarrhoea, abdominal pain and appetite changes among metformin’s common adverse effects. Kidney or liver problems, severe infection and heavy alcohol use can affect suitability. The medication review therefore needs to include other medicines and supplements, as well as diagnoses and current illness.
A prevention decision still involves a real medicine. It is reasonable to ask how tolerability will be reviewed and whom to contact if problems develop. A familiar drug is not automatically appropriate for every person, and a favourable trial result does not remove the need for those checks. Do not start, stop or change prescribed treatment on the basis of this article.
Vitamin B12 is another consideration. NICE’s 2024 vitamin B12 guideline identifies metformin as a risk factor for deficiency. Symptoms and risk factors inform assessment; unexplained neurological symptoms need attention rather than being dismissed as ageing. The practical question is whether the prescriber has considered B12 risk and what symptoms should prompt contact, not whether every reader should add a supplement.
Preventing diabetes is a different endpoint from longevity
Metformin affects glucose regulation, including reducing the liver’s glucose production and improving how the body uses insulin, as NIDDK’s medicine overview explains. A biological effect is a reason to investigate an outcome. It is not evidence that every outcome linked to metabolism will improve.
A 2022 DPP follow-up analysis in Circulation found no significant reduction in major cardiovascular events with original metformin or lifestyle assignment compared with placebo over a median 21 years. Other cardiovascular treatments and changes during follow-up complicate interpretation. The finding does not erase the diabetes result, but it does prevent us from translating it directly into a proven heart-attack benefit.
The editorial distinction is straightforward: name the outcome measured. A lower rate of diabetes diagnosis is evidence about diabetes diagnosis. A claim about living longer needs evidence about survival. Combining the two into a general promise of slower ageing asks the data to carry more than this trial established.
What this means in practice
- Bring the dates and values of previous glucose tests to a review, so the conversation can address the pattern rather than a label alone.
- Ask what structured prevention support is available and explain any barriers to participating.
- If metformin is proposed, ask what benefit is expected in your circumstances and when the decision will be reviewed.
- Bring a current medicine and supplement list, and ask about suitability, adverse effects and B12 assessment.
- Arrange the recommended follow-up; seek assessment for possible diabetes symptoms instead of waiting for the next routine test.
What we don’t know
The DPP recruited adults at high risk, so its results cannot establish a benefit for people with normal glucose regulation taking metformin for ageing. Long-term treatment changes also limit how precisely the follow-up can isolate continuous drug exposure. Nor can subgroup findings guarantee benefit for an individual. A credible prevention discussion can acknowledge these limits without dismissing the evidence.
Metformin has a place in selected diabetes-prevention decisions. The useful promise is a possible delay in disease, weighed against treatment burdens with a clinician. That is a more precise claim than a longer life, and a better basis for care.
Photo: SHVETS production on Pexels. Illustrative consultation; the people pictured are not participants in the research.
