850 mg reviews that count UKPDS events, not pounds
Open a cart review and you will hear 'lost ten pounds'. Open UKPDS 34 and you will hear endpoints.
Metformin drops HbA1c by roughly one to one-and-a-half percentage points. Used alone it almost never causes hypoglycaemia. It is weight-neutral to mildly weight-reducing. Those traits make it a sensible opening move. They do not make it a slimming SKU. The GLP-1 agonists produce large, dedicated weight-loss trials. Metformin does not. Anyone swallowing an 850 mg chip mainly to thin down is reading the wrong poster.
What moved guidelines from 'tolerable cheap tablet' to 'start here' was not a forum before-and-after. It was UKPDS 34 in 1998, then the prevention data below. Major diabetes societies still name metformin the preferred initial drug for most people with type 2 diabetes, and it usually stays in the regimen when a newer agent is layered on. The labeled pharmacology behind that glucose drop sits in the Glucophage 850 mg monograph.
UKPDS 34: 342 overweight adults, events on the primary line
The United Kingdom Prospective Diabetes Study ran newly diagnosed type 2 diabetes across 23 UK centres from 1977. In 15 of those centres, overweight patients - more than 120 percent of ideal body weight - could be randomised to metformin. The primary comparison in UKPDS 34 was 342 people on intensive metformin versus 411 on conventional diet-first care. Median follow-up was 10.7 years. Mean HbA1c landed at 7.4 percent on metformin and 8.0 percent on conventional care.
Against conventional care, metformin cut any diabetes-related clinical endpoint by 32 percent, diabetes-related death by 42 percent, all-cause mortality by 36 percent, and myocardial infarction by 39 percent. A secondary look against intensive sulfonylurea or insulin in other overweight patients also favoured metformin on the aggregate event line, mortality, and stroke. That is why the tablet became first-line. It hinted the drug was doing more than moving a lab value.
The caveat has to sit on the same blotter. Three hundred and forty-two is not a small n for 1970s diabetes work, but it is not a modern cardiovascular-outcomes trial either. Metformin was in the original protocol for overweight patients only, in 15 of 23 centres. No later study has reproduced the mortality benefit as cleanly. A supplementary add-on of metformin to maximum sulfonylurea in 537 people even associated with more diabetes-related death - a messy, under-powered signal that still keeps honest readers from treating UKPDS as a slogan. Encouraging event data. Not a closed case.
Ten more years of watching did not erase the event cut
When the randomised phase ended, glucose differences faded. The event gap in the metformin arm did not.
Holman and colleagues posted a ten-year post-trial watch in 2008. Between-group HbA1c differences were lost after the first year off protocol. In the overweight metformin group, relative cuts persisted for any diabetes-related endpoint (21 percent), myocardial infarction (33 percent), and death from any cause (27 percent). That is the 'legacy' line people quote. It is real. It is also still the same original 342-versus-411 randomisation, now followed longer.
Microvascular disease did not move the same way in the metformin arm, during the trial or after. The sulfonylurea-insulin intensive group later showed a microvascular cut and a delayed infarct and mortality signal of its own. So the fair post is not 'metformin uniquely saves every organ'. The fair post is: in overweight, newly diagnosed adults, an early metformin policy left an event scar that outlived the HbA1c gap. That is enough to keep the 850 mg chip at the front of most starts. It is not enough to sell it as a heart drug the way empagliflozin or semaglutide can now be sold from dedicated trials.
DPP used 850 mg twice daily. Lifestyle still won.
The Diabetes Prevention Program, posted in 2002, asked a different question: can you keep type 2 diabetes from starting? Knowler's group randomised 3,234 adults with elevated fasting and post-load glucose to placebo, metformin 850 mg twice daily, or an intensive lifestyle program aiming for at least 7 percent weight loss and 150 minutes of activity a week. Mean age 51. Mean BMI 34. Average follow-up 2.8 years.
Incidence was 11.0, 7.8, and 4.8 cases per 100 person-years in placebo, metformin, and lifestyle. Metformin cut progression by 31 percent versus placebo. Lifestyle cut it by 58 percent. To prevent one case over three years you needed about 13.9 people on metformin and 6.9 in the lifestyle program. Lifestyle beat the drug. The 850 mg twice-daily dose in that trial is the same labeled chip this site locks in SERP - used as a prevention tool, not as a slimming cart.
Two limits belong next to the 31 percent. Prevention here mostly means delay: when the tablet stopped, part of the protection faded. And metformin is not FDA-approved for diabetes prevention, so this use, evidence-based as it is, stays off-label. Later DPP follow-up still favoured lifestyle over the long haul. The drug is the runner-up you offer when coaching cannot be staffed the way a trial staffs it. The labeled start and the gut climb for that 850 mg chip are in the dosing journal.
PCOS, pregnancy sugar, and other off-label edges
| Claim you will see | What actually moved | Honest post |
|---|---|---|
| Type 2 glucose control | Large, long, first-line use | Strong. The approved core. |
| Heart and death benefit | UKPDS 34 overweight arm, 342 on drug | Event signal. Not a modern CVOT. |
| Slimming cart | Modest, inconsistent pounds | Not a weight-loss SKU. |
| Diabetes prevention | DPP: 31% vs placebo, 850 mg BID | Real, off-label. Lifestyle won. |
| PCOS | Cycles, ovulation, insulin numbers | Helpful adjunct. Off-label. |
Polycystic ovary syndrome shares a thread with type 2 diabetes: insulin resistance in a large fraction of women. Metformin can regularise cycles, lift ovulation odds, and tidy metabolic numbers. For fertility, letrozole generally outperforms it. For androgen-driven skin and hair, other treatments do more. The role is supporting cast, and it is off-label in the United States. That is still a reasonable clinic use. It is not a second approved indication hiding in the carton.
Gestational diabetes and antipsychotic weight-gain blunting sit in the same drawer: trial-supported in places, used widely, still outside the US type 2 label. None of those edges should be sold as if they carried UKPDS event weight. The plant-to-pill file that explains why the class even survived is the goat's rue trace.
SGLT2 and GLP-1 outrun it on dedicated heart and kidney files
First-line is not 'strongest on every axis'. The 850 mg chip lost that contest years ago.
| Axis | Metformin 850 mg class | Sulfonylurea | SGLT2 | GLP-1 |
|---|---|---|---|---|
| HbA1c | ~1 to 1.5% | ~1 to 1.5% | ~0.5 to 0.8% | ~1 to 1.5% |
| Weight | Neutral / slight loss | Gain | Loss | Loss |
| Hypoglycaemia alone | No | Yes | No | No |
| Dedicated CV trial | No (UKPDS events) | No | Yes | Yes |
| Cost | Very low | Low | Higher | Higher |
Line metformin up against a sulfonylurea and it wins on what patients feel: no weight gain, no lone-drug hypoglycaemia. Line it up against an SGLT2 inhibitor or a GLP-1 receptor agonist and the newer classes win on proven cardiovascular and kidney endpoints plus weight. Those benefits come from dedicated outcome trials. UKPDS was not built as a modern MACE study. Pretending otherwise is how a review turns into a cart.
So why does the cheap tablet keep the front slot? Cost, gut-manageable safety, and decades of mapped behaviour. The modern move is usually keep metformin and add the newer agent when heart failure, atherosclerotic disease, or chronic kidney disease is the reason you are in the room. Swap-out happens. Foundation-plus is more common. The practical climb that keeps people on the foundation is in the eGFR and gut journal.
What an 850 mg review owes the reader
A fair review says: solid glucose drop, excellent price, rare lone-drug lows, a real UKPDS event signal in overweight new diagnoses, a real DPP delay at 850 mg twice daily, and a modest weight nudge that should not be the reason you start. An unfair review says: 'this is the weight-loss pill doctors hide' or 'proven to save every heart'. The first set is enough. The second set is a cart.
The overstatement slips in when 342 people become a slogan, when lifestyle's 58 percent is memory-holed, or when PCOS is treated as if it sat on the FDA label. Read the event file. Ignore the slimming pitch. Then decide whether the 850 mg chip is your foundation or someone else's leftover story. Primary trial pages worth a nofollow look: the Lancet UKPDS 34 abstract and the NEJM DPP paper.