My clinician started metformin the day I was diagnosed. Why not try diet for a few months first?
What UKPDS 34 counted
Start with the paper, not with a guideline slogan. The slogan came after the counts.
| UKPDS 34 endpoint | Metformin vs conventional | Why it still gets cited |
|---|---|---|
| Any diabetes-related endpoint | 32% fewer (p=0.002) | Composite of the events the study was built to catch |
| Diabetes-related death | 42% fewer (p=0.017) | The line that changed reputation |
| All-cause mortality | 36% fewer (p=0.011) | Rare for an old oral agent |
| Myocardial infarction | 39% fewer (p=0.01) | Heart attack, not a lab surrogate |
| Median HbA1c | 7.4% vs 8.0% | Real, but not the keep |
UKPDS 34, Lancet 1998, sat inside a larger British program that followed newly diagnosed type 2 diabetes. The metformin question used overweight patients - more than 120 percent of ideal body weight. Of 1,704 such people, 753 entered the randomised comparison: 342 to intensive policy with metformin, 411 to conventional policy built mainly on diet. Median follow-up 10.7 years. Median HbA1c landed at 7.4 percent on metformin versus 8.0 percent on conventional care. That is a real glucose gap. It is not the line that moved the field. A 0.6-point A1c difference does not, by itself, explain a first-line habit that outlasted a dozen newer orals.
Versus conventional policy, metformin cut any diabetes-related endpoint by 32 percent, diabetes-related death by 42 percent, all-cause mortality by 36 percent, and myocardial infarction by 39 percent. Those are the numbers I file when someone says 'it only lowers sugar.' Plenty of tablets move HbA1c. Far fewer, in a randomised setting of that era, moved deaths and heart attacks. A secondary look against overweight people on intensive sulphonylurea or insulin also favoured metformin on several aggregates, including all-cause mortality and stroke. That is why a 500 mg start still opens so many first scripts. Reviews that never name those endpoints are not reviews. They are price lists.
What the design will not carry
Strike the overclaim. Benefit came from an overweight subgroup of a larger program, not from a modern placebo-controlled cardiovascular outcomes trial built only to re-prove metformin. You could not ethically withhold an established first-line tablet today to rerun that experiment. Sample sizes look small next to contemporary CVOTs. The conventional arm was diet-first, not a slick active comparator. Those limits are real. They do not erase the signal. They stop you from talking as if every person on a 500 mg chip has a personalised 42 percent death reduction stamped on the bottle. A review that quotes the 42 percent and hides the overweight frame is doing marketing.
There is a second wrinkle I will not hide. A supplementary randomisation added metformin early onto maximum sulphonylurea and saw more diabetes-related death in that small add-on group. Combined analyses and later epidemiologic looks did not confirm a class-wide poison story, and practice did not abandon the combination. Still: the paper is not a cartoon of unbroken benefit. File the main overweight comparison as the reason metformin outranked old orals. File the add-on caution as a reminder that trial appendices exist. Honesty about a 1998 appendix is part of why this desk still trusts the main result.
Not a weight-loss cart
The other claim on this slug is a cart: take 500 mg, watch the scale, skip the harder drugs.
Metformin is weight-neutral or mildly down. Some people lose a few kilograms, partly because early gut effects blunt appetite. That is not an obesity indication. The US label is type 2 diabetes as an adjunct to diet and exercise. Reviews that treat a 500 mg tablet like a GLP-1 injector are selling a different product. Dedicated weight-loss medicines move the scale in a range metformin never claimed and never showed. A friend who 'dropped a stone on Glucophage' may have dropped water, meals, and a few kilos of gut-driven appetite loss. That anecdote is not a trial endpoint.
Why the confusion lasts: first-line diabetes care and weight sit in the same clinic visit, and metformin does not cause the gain that sulphonylureas and insulin often do. 'Does not make you heavier' is a genuine advantage. 'Is a weight-loss cart' is a strike. If weight is the primary goal without diabetes or a related metabolic brief, this is the wrong file. If diabetes is the brief, the 500 mg start is still a foundation, not a substitute for the agents that now carry dedicated heart-failure, kidney, and weight trial programs. Keep the chip. Do not ask it to be a weekly pen.
Prediabetes: the DPP ranks it second
The Diabetes Prevention Program is the second-strongest named trial on this board. People with prediabetes - high sugar, not yet diabetes - were randomised to intensive lifestyle, metformin, or placebo. Both active arms cut progression. Lifestyle cut it by about 58 percent. Metformin cut it by about 31 percent. Diet and exercise won. Long follow-up kept a benefit in view for years. That ranking is why guidelines put lifestyle first and reserve metformin for groups where the drug's effect was strongest - often younger people, higher BMI, prior gestational diabetes. A 500 mg prevention script that never mentions walking is a cart with a different label glued on.
File the honest sentence: metformin can delay diabetes. It does not buy anyone out of the work the lifestyle arm won on. Starting doses for that use still look like the labeled 500 mg climb, which belongs in the titration and eGFR walkthrough. Do not let a prevention script become a permission slip to ignore meals and walking. If the numbers are already in the diabetes range, this is no longer a DPP conversation. It is a type 2 start, and UKPDS 34 is the named file again.
PCOS is a side file, not the headline
Polycystic ovary syndrome is off-label territory with mixed trials. Insulin resistance is part of many PCOS pictures, so improving insulin sensitivity has a rationale. Metformin can help menstrual regularity, ovulation, and metabolic markers, especially when glucose handling is already impaired. That overlap is real. It is also narrower than early enthusiasm suggested.
For fertility alone, dedicated ovulation-induction drugs generally work better; metformin is no longer first-line fertility treatment by itself. For acne and excess hair, the effect is modest at best. Honest framing: a limited metabolic tool inside a syndrome, strongest where insulin resistance and glucose problems dominate. Not a cure for every feature. A clinician who follows current PCOS guidance should weigh it, not a cart review.
Newer agents layer on; they do not retire the 500 mg start
The last claim: SGLT2 inhibitors and GLP-1 agonists made metformin obsolete. Open the clinic, not the ad.
| Use | Best named evidence | How this desk files it |
|---|---|---|
| Type 2 diabetes | UKPDS 34 plus decades of use | First-line; event signal in overweight |
| Prediabetes | Diabetes Prevention Program | Works; second to lifestyle |
| PCOS, metabolic slice | Mixed RCTs | Adjunct where insulin resistance leads |
| PCOS, fertility alone | RCTs vs ovulation drugs | Second-line; not the headline |
| Weight as the only goal | Limited, not labeled | Not a cart; effect modest |
Two things still separate classes: hypoglycaemia when used alone, and weight. Metformin scores well on both, which is why it remains a foundation even as newer drugs stack on top. Sulphonylureas are cheap and effective but push insulin release, so lows and gain follow. SGLT2 inhibitors and GLP-1 receptor agonists avoid hypoglycaemia when used alone, tend to reduce weight, and have cardiovascular and kidney benefits in dedicated modern trials. Those benefits are stronger and cleaner than UKPDS 34 on those organs. Cost and route differ: metformin is a cheap tablet; many GLP-1 agonists are expensive and injected.
Practical file: start metformin; if control is short, or if heart failure, atherosclerotic disease, or kidney disease is already on the chart, add the newer class for the organ protection UKPDS was never built to isolate. Layering, not a replacement contest. Side-effect and eGFR rules that shape the metformin half of that stack are in the titration and eGFR walkthrough. How the molecule got from a weed to a chip is in the plant-to-pill file.
Reader mail
Reader questions on this article
Answered by Dr. Sofia Kessler, MD · Internal medicine & clinical pharmacology
Reviews after the UKPDS file. Events first. A scale does not get the last line.
For most newly diagnosed type 2 diabetes, current guidance starts metformin with lifestyle change rather than waiting for willpower to finish the job. Reasoning: the tablet is cheap, does not cause low blood sugar alone, and carries an event signal from UKPDS 34 - little downside to starting early, real cost to leaving high glucose untreated for months. That is not a verdict that you failed at diet. Lifestyle stays central. The 500 mg chip supports it. If numbers are only mildly raised, some clinicians still trial lifestyle alone briefly. Ask what yours had in mind. A readable public pass sits at Mayo Clinic.
Does metformin actually protect the heart, or does it just lower sugar?
UKPDS 34 is the named reason I will not say 'just sugar.' In the overweight comparison, metformin cut myocardial infarction by 39 percent and diabetes-related death by 42 percent versus conventional policy. That is a cardiovascular signal, not a marketing slogan. Limits: overweight subgroup, older design, no modern placebo CVOT that isolates metformin. I will not pretend every 500 mg start reprints those percentages on your chart. I will also not pretend any other old oral agent has a comparable positive event file. For heart-failure and kidney protection beyond that, SGLT2 and GLP-1 drugs have stronger dedicated trials - which is why we often add one rather than retire metformin.
I have prediabetes. Should I be asking for metformin?
Maybe - after lifestyle, not instead of it. The DPP showed intensive diet and exercise cut progression about 58 percent; metformin about 31 percent. Lifestyle won. Guidelines put diet, weight loss, and activity first, then metformin for groups where the drug helped most - younger people, higher BMI, gestational diabetes history. If you sit in one of those groups, or if lifestyle alone is not holding the numbers, a 500 mg start is reasonable. It is a supplement to the work the winning arm did, not a substitute. Do not let a prevention script become a cart.
I was given metformin for PCOS, not diabetes. Is that a real use?
Recognised off-label use. Whether it helps depends on which PCOS features actually trouble you. The syndrome often includes insulin resistance, so a drug that improves insulin sensitivity has a rationale. It can help cycles, ovulation, and metabolic markers, especially when glucose handling is already off. It disappoints as a solo fertility drug - dedicated ovulation medicines work better - and it is modest at best for acne and excess hair. Legitimate tool for the metabolic slice. Not a fix for every symptom. A specialist who follows current PCOS guidance should weigh it for your case, not a forum review.
A friend takes 500 mg to lose weight. Will it do that for me?
This is the claim I strike on this slug. Metformin is weight-neutral or mildly down - not approved as a weight-loss drug. Some people drop a few kilograms. That is nowhere near the range of GLP-1 medicines built for weight. Taking a 500 mg tablet purely to lose weight, without diabetes or a related metabolic brief, buys gut side effects without a labeled indication. If the scale is the only goal, that is a different conversation about drugs designed for it. If you already have type 2 diabetes, mild loss is a bonus, not the reason the tablet leads.
How does metformin stack up against the newer diabetes drugs I keep hearing about?
Foundation, with newer drugs added on top - not a retirement. Metformin strengths: no low blood sugar alone, no weight gain, very cheap, UKPDS event signal. Sulphonylureas, the older add-on, cause both lows and gain and have fallen from favour. SGLT2 inhibitors and GLP-1 agonists avoid hypoglycaemia, help weight, and have real heart and kidney trial benefits that are cleaner than 1998 data. Usual file: keep metformin as the base, add one when more control is needed or when heart or kidney disease is already present. Layering. Not a contest with a single winner.
If the newer drugs are so good, why is a 500 mg metformin tablet still first?
Because it does the fundamentals well at almost no cost with the longest safety track record we have in this class. It lowers glucose reliably, does not cause hypoglycaemia alone, does not drive weight gain, and still carries the UKPDS event file. Newer drugs add things metformin cannot - especially dedicated heart-failure, kidney, and weight programs - but they are expensive and many GLP-1 agonists are injected. For systems treating millions, a cheap safe tablet first, then escalate, is still sound. The live label sits at the FDA. The molecule itself is unpacked in the Glucophage note.
Was UKPDS 34 only in overweight people? Does that weaken the first-line habit?
The randomised metformin-versus-conventional comparison was in overweight patients - more than 120 percent of ideal body weight, 753 people, median 10.7 years. That is a real limit. It is also the group that produced the death and heart-attack reductions. Guidelines still put metformin first for most new type 2 diabetes because the glucose effect, the hypoglycaemia profile, the weight profile, and decades of use extend beyond that subgroup, while no other old oral has a matching event signal. I will not hide the overweight frame. I will not use it to pretend the tablet is unproven. I will use it to stop cartoon percentages on every bottle.
What should I ignore in online 'metformin 500 mg reviews'?
Ignore the cart. Ignore before-and-after scale photos that treat a diabetes tablet like an injector. Ignore anyone who says UKPDS 'proved' a personal 42 percent death cut for every starter. Keep the named endpoints, the overweight frame, the DPP ranking, and the PCOS limits. Keep the gut and eGFR rules on the sister page. If a review never mentions lactic acidosis rarity, kidney clearance, or food-with-dose, it is not a review - it is a pitch. Named-trial language is the filter on this desk.
General education from a clinician, not personal medical advice. Bring your own history to your own prescriber.
Last Updated