Is it actually true the commonest diabetes tablet started as a weed?
Order the plant file before the 850 mg chip
Farmers learned the plant first. Physicians arrived late, and they arrived because animals died.
Galega officinalis spread through southern and central Europe under two names that still sit on herb cards: goat's rue and French lilac. Livestock that overgrazed it fell ill. Death in a pasture is a hard ping. Something in the plant was chemically real. Medieval herbalists were already using that same plant against heavy urination, plague fever, and thin milk. They could not isolate a dose. They could only notice an effect and pass the rumor along.
The useful fraction was guanidine and a related alkaloid, galegine. By 1918 the glucose-lowering action of guanidine was measurable in animals. The catch arrived in the same notebook. Raw guanidine was too toxic to hand to a person. Galegine was milder and still too rough for a lifelong tablet. The plant marked a direction. It did not deliver an 850 mg chip. Where that direction later sat on a label is filed in the Glucophage 850 mg monograph.
A second early clue sat in wartime chemistry. Paludrine, an antimalarial, lowered glucose as a side effect, and flumamine, a guanidine analogue used against influenza fever, nudged the same labs back toward galegine's shape. None of those sidelines produced a durable diabetes tablet. They kept the pasture signal alive while insulin owned the wards. The 850 mg chip is what survived that sorting, not a pressed leaf.
Guanidine had to be cut before anyone could swallow it
Chemists did not bottle the weed. They bolted two guanidine units together and built a biguanide backbone that kept the glucose drop while dulling the poison. Galegine's structure was worked out in Edinburgh after French labs had already started the toxicology. Dimethylbiguanide - metformin - was synthesised in Dublin in 1922 by Emil Werner and James Bell. That date is not folklore. It is a bench date, and it is the same year insulin was isolated.
Insulin swallowed the decade. A hormone that pulled dying children with type 1 diabetes back from ketoacidosis absorbed the money, the wards, and the headlines. A mild oral compound that eased sugar down in adults looked dull beside that. Metformin sat on a shelf for the better part of four decades, chemically interesting and clinically ignored. The plant had been traced. The molecule had been made. Nobody had posted it as a drug.
Sterne posted the name. The molecule was already thirty-five.
A shelf compound needs an advocate. Metformin found one in 1950s Paris, not in a marketing deck.
Jean Sterne trained in diabetology under Francis Rathery, then held posts at Aron Laboratories and Hopital Laennec. With Denise Duval he ran several biguanides. He picked dimethylbiguanide. In 1957 he published the clinical work and coined Glucophage - glucose eater - in a short paper in Maroc Medical. The name described the effect he watched, not a mechanism. The tablet does not devour sugar. It mainly cuts the glucose the liver exports and sharpens the body's response to its own insulin.
France licensed it first. Britain followed in the late 1950s. Canada posted it in 1972. Sterne did not invent the molecule. He read the signal correctly: this particular biguanide struck the best balance of effect against gut cost, and it did not drive blood sugar into the floor the way insulin or a sulfonylurea can. The 850 mg tablet that later became this site's SERP lock is one labeled immediate-release strength on that same Glucophage line - 500, 850, and 1000 mg. The climb rules live in the dosing and safety journal.
Phenformin flagged the whole class. Metformin almost drowned with it.
1957 was a crowded year. Ungar's group posted phenformin, phenylethylbiguanide, in the same season. Buformin followed. Both cousins were more potent. Both were marketed harder. Both could set off lactic acidosis often enough to kill. Case reports stacked through the 1960s and 1970s. The United States pulled phenformin in 1977. Buformin dropped out wherever it had been adopted. Regulators then treated every biguanide as the same animal.
That was the wrong flag. Phenformin is more lipophilic, cleared in part by the liver, and its handling swings with CYP2D6. It can accumulate and interfere with cellular energy in ways that drive lactate up. Metformin leaves the body almost entirely unchanged through the kidneys. Once the drug is kept off failing kidneys, its lactic-acidosis rate sits far below phenformin's - later reviews put it near one-tenth. Guilt by family resemblance still stalled the American file for years. The safety rules that keep the rare event rare are the ones in the eGFR and gut guide.
Europe posted it in the 1950s. The US waited until 1995.
By the early 1990s French and Scottish clinics had decades of everyday metformin use. American patients still could not get the tablet.
Medieval Europe
Goat's rue used for heavy urination and other folk complaints. Livestock poison is the harder signal.
1918
Guanidine lowers glucose in animals and is too toxic to keep.
1922
Werner and Bell synthesise dimethylbiguanide in Dublin. Insulin is isolated the same year.
1957
Sterne names Glucophage. Phenformin is posted in the same season.
1972
Canada licenses metformin, more than twenty years ahead of the US.
1977
The US withdraws phenformin. The class inherits the stain.
1994-1995
FDA approval on 29 December 1994. Glucophage launch in 1995. The 850 mg IR tablet is one labeled chip.
Two forces stretched the lag. The phenformin scare left the FDA cautious toward the whole class. And no patent-holding sponsor wanted to fund a long American application for a molecule whose original protection had lapsed. Lipha, which had absorbed Aron, finally pushed the file. Gerard Daniel and Anita Goodman answered an avalanche of agency questions. Independent clinicians - Reaven, DeFronzo, Bailey, later Turner and Garber - helped design the US trials. The FDA approved metformin on 29 December 1994. Bristol-Myers Squibb launched Glucophage in 1995.
The launch came with a black-box reminder and a safety-first education push: this is not a sulfonylurea, hold it when kidneys fail or oxygen falls, and do not treat the class as interchangeable with phenformin. An extended-release form followed around 2000, mainly to blunt the gut. Within a few years the cheap tablet had become the default first prescription for type 2 diabetes in the country with the largest burden of the disease. The event data that later locked that default are in the UKPDS evidence journal.
The 850 mg chip is factory work. Calling it herbal is a miss.
Readers still ask whether a plant origin makes metformin 'natural'. It does not. The 850 mg tablet is a fully synthetic dimethylbiguanide made in a plant of a different kind - a factory. Goat's rue supplied the chemical clue. It does not supply the dose. You cannot titrate a pasture. You can titrate 500, 850, or 1000 mg and stop below an eGFR of 30. That is the whole point of the rewrite.
The modest traits that made the molecule forgettable next to insulin in 1922 are the ones that let it last. It lowers glucose without the weight gain and the hypoglycaemia that trail insulin and the sulfonylureas. It is cheap to manufacture. It does not flog the pancreas to secrete more insulin. For a lifelong disease carried by millions, a tablet that is safe, dull, and affordable outlasts a flashier cousin that is none of those. The modern evidence file - UKPDS events, not a slimming cart - is the reviews journal.
What the trace leaves on the blotter
The honest history is shorter than the myth and stranger than a brand story. A poisonous pasture weed pointed at guanidine. Chemists cut the poison. Insulin buried the result. A French physician posted the name in 1957. A more potent cousin nearly sank the class. The United States arrived in 1995 with an 850 mg chip already on the European shelf. None of that makes the tablet herbal. All of it explains why the label still talks about kidneys and lactate.
If you came here because a cart called the 850 mg tablet a 'plant medicine', flag that line and move on. The plant was the ping. The chip is the post. For labeled strengths, food, and the eGFR gate, stay on this site's Glucophage 850 mg page. For the trial numbers that later justified first-line use, open the event review.
Reader mail
Reader questions on this article
Answered by Dr. Nathan Whitfield, MD · Internal medicine & clinical pharmacology
Readers pinged the plant story. These are the flags worth answering in full.
Yes, and the weed was better known for killing animals than helping anyone. Goat's rue, Galega officinalis, packed enough guanidine to sicken livestock that overgrazed it. Herbalists used the same plant for heavy urination we now tie to diabetes. That is a folk ping, not a measured dose. Metformin is a later, synthetic rewrite of that guanidine clue. The 850 mg Glucophage chip does not grow on a stem. The plant had to be traced so chemists could cut the poison and keep the glucose drop. I file the finished molecule on the Glucophage 850 mg monograph.
If the molecule was made in 1922, why did nobody use it until 1957?
Timing, not chemistry. Werner and Bell synthesised dimethylbiguanide in Dublin in 1922, the same year insulin was isolated. Insulin looked like a miracle beside it. A mild oral compound that eased sugar down in adults drew almost no funding. The molecule sat on a shelf for thirty-five years. Jean Sterne and Denise Duval came back to several biguanides in 1950s Paris, picked this one, and posted the clinical work in 1957 under the name Glucophage. The right advocate and the right question arrived late. The bench date was never in doubt.
What does Glucophage actually mean, and is that how it works?
It is as literal as drug names get. Glucophage reads as 'glucose eater', from the Greek roots for glucose and for eating. Sterne picked it to name the effect he watched in clinic, not a mechanism. The tablet does not devour sugar. It mainly cuts the glucose the liver releases and helps muscle respond to the insulin a person already makes. That is why it rarely drives blood sugar into the floor when used alone. The name stuck because it was honest about the effect and short enough to print on a European carton. The 850 mg chip still carries that 1957 name in a lot of pharmacies.
How is phenformin different, and why did it matter so much?
Same family, different risk. Phenformin is a more potent, more fat-soluble biguanide that the liver handles in part, so levels swing from person to person. It caused lactic acidosis often enough that the United States pulled it in 1977. Metformin is cleared unchanged by the kidneys and triggers that crisis far more rarely when eGFR is respected. The damage was guilt by association. After phenformin left, regulators treated every biguanide as the same animal. That stain is a large reason the American 850 mg tablet waited until 1995 even though France had been writing Glucophage since 1957.
Why did America wait until 1995 if Europe had it in the 1950s?
Two stacked problems. The phenformin withdrawal left the FDA wary of the whole class, metformin included. And no commercial sponsor wanted to fund a long, expensive US file for a drug whose original patent had lapsed. Lipha finally pushed it. The agency approved metformin on 29 December 1994. Bristol-Myers Squibb launched Glucophage in 1995 with a black-box reminder and a safety-first education push. European patients had decades of everyday use by then. The 850 mg IR tablet that this site locks in SERP was already a labeled European chip. The lag was regulatory and commercial, not a late discovery.
Was goat's rue ever a real medicine, or is that just a story?
It was used, in the loose medieval sense. Herbalists gave it for heavy urination, fever, and to boost milk in nursing mothers and animals. People noticed an effect. They could not separate the useful fraction from the toxic one or control a dose. That is exactly the problem modern chemistry solved by building biguanides out of the plant's guanidine. Calling the old use 'medicine' in the modern sense oversells it. Calling the 850 mg tablet a herbal product undersells the rewrite. The pasture was a clue. The factory is the source.
Does a plant origin mean I can treat this as a natural supplement?
No. That is the miss I see most often in comment threads. Metformin is a synthetic dimethylbiguanide made to a labeled strength - 500, 850, or 1000 mg immediate-release, plus extended-release chips. It is not an extract of goat's rue. You cannot titrate a pasture against eGFR. You can hold a tablet when kidneys fail, around iodinated contrast, or in a severe acute illness. Treat it as a prescription glucose drug with a plant-shaped backstory, not as a wellness scoop. If someone is selling 'natural Glucophage', flag the cart and walk away.
Who actually brought the US tablet to market, and does the brand still matter?
Aron Laboratories held Sterne's early work. Lipha absorbed Aron and ran the FDA file. Bristol-Myers Squibb launched Glucophage in the United States in 1995. The original brand still exists in some markets, but most 850 mg chips dispensed today are generic metformin hydrochloride from many manufacturers. Clinically the molecule is the same. The brand name matters for history and for spotting an old carton. The labeled rules - food, climb, eGFR below 30 - do not change with the logo. I keep those rules on the climb and eGFR journal.
If the plant was the clue, why do you keep saying the tablet had to be traced?
Because the useful story is the chain, not the romance. Goat's rue pointed at guanidine. Guanidine was too toxic. The 1922 synthesis sat unused. Sterne posted the name. Phenformin almost erased the class. The US arrived decades late. Each step is a place people invent a cleaner myth - herbal, accidental, American, or 'always first-line'. Tracing the plant keeps the 850 mg chip honest. It is a cheap, synthetic, kidney-cleared tablet with a long European file and a late American stamp. That is enough. It does not need a fairy tale.
General education from a clinician, not personal medical advice. Bring your own history to your own prescriber.
Last Updated