Prediabetes Drugs: Metformin, GLP-1s, and Guidelines

Metformin is the most widely studied and commonly prescribed drug for prediabetes, but it is far from the only option, and the landscape has shifted dramatically in recent years with the arrival of newer medications. The American Diabetes Association recommends that metformin “should be considered” for people with prediabetes who have a BMI above 35, though lifestyle changes remain the default first step for most people. What makes this topic interesting is how rapidly the evidence is evolving: drugs originally designed for established type 2 diabetes or obesity are now showing striking results in preventing diabetes from developing in the first place.

Metformin and the Evidence Behind It

Metformin has the longest track record of any drug studied for prediabetes prevention. The landmark Diabetes Prevention Program (DPP), a large U.S. trial, found that metformin reduced the rate of progression to type 2 diabetes by about 31% compared to placebo, translating to roughly 14 people needing treatment for three years to prevent one case of diabetes.¹ A more recent meta-analysis pooling multiple trials confirmed a roughly 35% lower odds of developing diabetes among people with prediabetes who took metformin.²

The DPP also tested an intensive lifestyle intervention, which cut diabetes risk by 58%, roughly twice the effect of metformin alone.³ That result has shaped clinical guidelines ever since: lifestyle changes (weight loss, increased physical activity, dietary improvements) are recommended first, and metformin is added when lifestyle alone is insufficient or when someone has additional risk factors like a very high BMI. More recent pooled analyses have found that the head-to-head difference between metformin and lifestyle intervention did not always reach statistical significance across trials, suggesting the gap between them may not be as wide as the original DPP implied.⁴ In practice, many clinicians combine both, and there is no reason to think of them as competing strategies.

How Long the Benefits Last

One of the most important questions about any prediabetes drug is whether the benefits stick around over the long haul. Follow-up data from the DPP, now stretching past 21 years, shows that both the lifestyle and metformin arms maintained meaningful reductions in cumulative diabetes incidence decades after the original trial began.⁵ Most of the prevention or delay happened during the first three years of the trial, but the gap between the treatment groups and placebo persisted over an additional 12 years of follow-up.⁶

This is worth pausing on, because many interventions in medicine show a benefit during the treatment period that fades afterward. With metformin and lifestyle changes for prediabetes, the initial effort seems to buy lasting protection, even though some people in the treatment groups eventually developed diabetes too. The takeaway is not that these interventions eliminate risk forever, but that they meaningfully delay progression, and delaying diabetes by years matters because it means fewer years of the complications diabetes brings.

Metformin’s Side Effects and Practical Trade-Offs

Metformin is generally well tolerated, which is part of why it occupies the first-line position. The most common complaints are gastrointestinal: nausea, diarrhea, and stomach discomfort, especially in the first weeks. Extended-release formulations reduce these symptoms for many people. The drug is inexpensive, widely available as a generic, and does not cause low blood sugar on its own.

The less obvious trade-off is vitamin B12. Data from the DPP follow-up showed that long-term metformin use was associated with an increased risk of low or borderline-low B12 levels. At five years, combined low and borderline-low B12 was about twice as common in the metformin group compared to placebo, and the risk climbed with each additional year of use.⁷ B12 deficiency can cause fatigue, tingling in the hands and feet, and cognitive changes that mimic other conditions, so anyone on long-term metformin should have their B12 checked periodically.

GLP-1 Receptor Agonists and the New Wave

The biggest shift in the prediabetes drug conversation has come from GLP-1 receptor agonists, the same class of medications that includes semaglutide (Ozempic, Wegovy) and liraglutide (Saxenda, Victoza). These drugs were developed for type 2 diabetes and obesity, but recent evidence shows they can powerfully reverse prediabetes too.

A meta-analysis of trials found that people with prediabetes who received GLP-1 receptor agonists were over four times more likely to return to normal blood sugar levels compared to those on placebo. Both semaglutide and liraglutide individually showed strong effects, though the evidence for exenatide (an older drug in this class) was not significant.⁸ Another pooled analysis found that GLP-1 receptor agonists nearly tripled the rate of prediabetes reverting to normal glycemia and reduced new-onset diabetes by about 72%.⁹

In the SELECT trial, which enrolled people with overweight or obesity but without established diabetes, semaglutide pushed roughly 70% of participants back to normal blood sugar at three years, compared to about 36% on placebo. The proportion who progressed to full diabetes was also dramatically lower: about 1.5% with semaglutide versus nearly 7% with placebo.¹⁰ The researchers noted that weight loss accounted for roughly a quarter of the glycemic improvement, meaning the drug’s effects on blood sugar go beyond what weight loss alone explains.

Tirzepatide and Dual-Action Drugs

Tirzepatide (Mounjaro, Zepbound) works on two gut-hormone pathways simultaneously: GLP-1 and GIP. In people with prediabetes and obesity, the diabetes prevention results have been extraordinary. In one trial, only about 1.3% of participants on tirzepatide progressed to type 2 diabetes, compared to 13.3% on placebo, a roughly 93% reduction in risk.¹¹ Even 17 weeks after stopping the drug, the difference persisted: about 2.4% of the tirzepatide group versus nearly 14% of placebo had diabetes.¹²

These numbers are striking enough that some researchers describe tirzepatide as the most potent diabetes-prevention drug studied to date.¹³ But the practical question is more complicated. These drugs are expensive, typically require injections (though oral semaglutide exists), carry their own side effects including nausea and gastrointestinal discomfort, and are not yet approved specifically for prediabetes treatment. Weight regain after stopping GLP-1 drugs is well documented, and the long-term effects of using them purely for prediabetes, especially in people who are not obese, remain unclear.

SGLT-2 Inhibitors

Sodium-glucose cotransporter 2 (SGLT-2) inhibitors, drugs like empagliflozin, dapagliflozin, and canagliflozin, work by causing the kidneys to excrete excess glucose in the urine. Originally designed for type 2 diabetes, they have accumulated growing evidence in prediabetes as well. Reviews of the data suggest they improve blood sugar markers, delay progression to diabetes, and may reduce cardiovascular and kidney complications, particularly in people who already have heart or kidney disease.¹⁴¹⁵

The appeal of SGLT-2 inhibitors lies partly in their side benefits. In people with type 2 diabetes, they have shown clear reductions in heart failure hospitalization and kidney disease progression. Whether those same benefits apply to someone who only has prediabetes and no established cardiovascular disease is less certain. The evidence base is thinner than for metformin or GLP-1 drugs in this specific population, and no major clinical guideline currently recommends SGLT-2 inhibitors as a standard prediabetes intervention. Their main niche right now may be in people with prediabetes who also have heart failure or chronic kidney disease, where the organ-protective effects overlap with the glycemic benefit.

Pioglitazone and Acarbose

Not all prediabetes drug research centers on the newest medications. Pioglitazone, an older insulin-sensitizing drug in the thiazolidinedione class, showed impressive results in a dedicated prediabetes trial: the annual rate of developing diabetes was about 2% with pioglitazone versus nearly 8% with placebo, a 72% reduction. The catch was meaningful side effects, including an average weight gain of about 4 kilograms and a higher rate of fluid retention.¹⁶ Those side effects, along with historical concerns about bone fractures and a possible link to bladder cancer (which remains debated), have kept pioglitazone out of routine prediabetes use despite its effectiveness at lowering glucose.

Acarbose works differently from most prediabetes drugs. Instead of improving insulin sensitivity or boosting insulin secretion, it slows the digestion of carbohydrates in the gut, blunting the blood sugar spike that follows a meal. Studies in people with impaired glucose tolerance have found that acarbose can reduce progression to diabetes, and research in healthy volunteers has confirmed it can dampen post-meal glucose spikes by over 17%.¹⁷ Acarbose tends to cause gas and bloating, which limits adherence, but it remains a viable option in some settings, particularly in parts of Asia where it is used more widely than in Western countries.

When Do Guidelines Actually Recommend a Drug?

Despite the growing list of medications that can delay or prevent type 2 diabetes, clinical guidelines remain relatively conservative about prescribing them for prediabetes. The American Diabetes Association’s position, reflected in its standards of care, is that metformin “should be considered” specifically for people with prediabetes whose BMI exceeds 35, along with those who are younger, have a history of gestational diabetes, or whose blood sugar continues to rise despite lifestyle changes.¹⁸ For most people with prediabetes, the initial recommendation is still a structured lifestyle program targeting about 7% body weight loss and at least 150 minutes of moderate physical activity per week.

GLP-1 receptor agonists and tirzepatide are not yet guideline-recommended specifically for prediabetes prevention, even though the trial data are compelling. This gap between the evidence and the guidelines exists partly because regulatory approval for prediabetes indications is still catching up, and partly because cost and access create real-world barriers. Prescribers sometimes use these drugs off-label for patients with prediabetes and obesity, but insurance coverage is inconsistent.

Cost and Value

A systematic review of cost-effectiveness studies found that both lifestyle programs and metformin were generally cost-effective for diabetes prevention in high-risk individuals. The median cost per quality-adjusted life-year was roughly £7,500 for lifestyle programs and about £8,400 for metformin, though estimates varied widely between studies.¹⁹ Metformin as a generic drug costs a few dollars per month in most markets, making it accessible even without insurance.

The newer drugs present a stark contrast. GLP-1 receptor agonists and tirzepatide can cost over a thousand dollars per month at list price, and their cost-effectiveness for prediabetes specifically has not been as thoroughly studied. If these drugs prevent or substantially delay diabetes in someone who would otherwise develop expensive complications, the long-term math could work out, but the upfront expense is a genuine barrier. This is one of the clearest cases in medicine where what works best biologically and what is practical for most patients are two different conversations.

Prediabetes in Younger Populations

Most of the major prediabetes drug trials have enrolled middle-aged and older adults, so it is worth asking what happens in younger people and adolescents. Childhood and adolescent prediabetes is increasing alongside rising obesity rates, and clinicians face the question of whether to start medications early.

The evidence here is much less encouraging. Clinical trials in youth are smaller and shorter than their adult counterparts. While metformin and other insulin sensitizers have shown modest improvements in insulin sensitivity and some short-term weight stabilization in adolescents, the beneficial effects on beta-cell function seen in adult studies have generally not been replicated in younger patients. Perhaps more concerning, the effects of metformin in youth did not appear to last after the medication was stopped.²⁰ Intensive lifestyle modification remains the primary approach for children and teens with prediabetes, with medications reserved for cases where lifestyle efforts fail and risk is high.

How Prediabetes Is Diagnosed and Why It Matters for Treatment

The drugs discussed above were tested in people identified through specific blood tests, and the diagnostic criteria matter more than you might expect. Prediabetes is generally diagnosed when fasting blood glucose falls between 100 and 125 mg/dL (5.6 to 6.9 mmol/L), or when a two-hour glucose tolerance test result lands between 140 and 199 mg/dL (7.8 to 11.0 mmol/L), or when HbA1c (a marker of average blood sugar over three months) falls between 5.7% and 6.4%.²¹

The wrinkle is that these different tests do not always agree. Research comparing HbA1c against glucose-based tests found that the standard HbA1c cutoff of 5.7% for prediabetes had low sensitivity, meaning it missed a large proportion of people who would be diagnosed by glucose testing. The false-negative rate was around 65%.²² In other words, a normal HbA1c does not guarantee normal glucose metabolism. This matters for drug decisions because if your doctor uses only HbA1c screening, your prediabetes might be missed entirely, meaning you never enter the conversation about whether medication is appropriate.

What Happens at the Cellular Level

Prediabetes is fundamentally a story of two breakdowns happening at once. First, cells in the muscles, liver, and fat tissue become less responsive to insulin, so the body needs more of it to keep blood sugar in check. Second, the insulin-producing beta cells in the pancreas eventually struggle to keep up with the increased demand. Different drugs attack different parts of this equation. Metformin primarily reduces the liver’s glucose output and improves insulin sensitivity. GLP-1 receptor agonists stimulate insulin release in a glucose-dependent way (so they rarely cause low blood sugar), slow stomach emptying, and reduce appetite. A small trial in people with prediabetes found that both moderate exercise and metformin improved beta-cell function compared to placebo after 12 weeks.²³ Preserving beta-cell function early is thought to be one of the reasons that intervening during prediabetes, rather than waiting for full diabetes to develop, may produce better long-term outcomes.

The Rebound Problem

One question that does not get enough attention is what happens when you stop the drug. The DPP follow-up data showed that the benefits of metformin and lifestyle intervention persisted for over two decades, even as the original structured interventions ended and participants were no longer in tightly controlled treatment programs.²⁴ But the newer, more potent drugs raise a different concern. The SELECT trial with semaglutide and tirzepatide trials both showed that some of the glycemic benefit waned after stopping treatment. In the tirzepatide trial, the proportion with diabetes after 17 weeks off treatment roughly doubled compared to the on-treatment numbers, though it was still far lower than placebo.²⁵ Similarly, the weight regain associated with stopping GLP-1 drugs is well established.

This creates a real clinical dilemma. If someone starts semaglutide or tirzepatide for prediabetes, are they committing to taking it indefinitely? Is that different from being on metformin indefinitely, which is cheap and well tolerated? These are questions the current evidence cannot fully answer, and they matter at least as much as the headline efficacy numbers.