Is it actually true the world's biggest diabetes tablet started as a weed?
The weed that cut both ways
Readers bring a lab-origin story. The first source is a field, and the field was not kind. A clean-bench myth does not survive the livestock note.
Galega officinalis grows across southern Europe and the Near East. English speakers called it goat's rue. Older herbals also say French lilac. Herbalists listed it for the thirst and ceaseless urination we now read as uncontrolled diabetes. Shepherds listed something else: animals that ate too much of it died. Medieval and later European texts keep both observations. Nobody had a mechanism. They had a plant that moved water and sugar, and a dose that could wreck a flock.
That is the claim worth filing, not a romance about a miracle herb. The plant is rich in guanidine and close cousins. Guanidine lowers glucose. Guanidine is also poisonous. The same chemistry that hinted at a medicine explained the dead sheep. A century of work after that was not invention from nothing. It was an attempt to keep the sugar drop and lose the kill. The 500 mg tablet on a Manchester blotter is what survived that filter. If a history skips the livestock deaths, it is selling a garden story, not a taming.
Guanidine was the clue, not the pill
Early-1900s chemists isolated guanidine from goat's rue and confirmed the glucose drop in animals. Label the next strike: raw guanidine was never a candidate for a human tablet. The therapeutic window was too narrow. A dose that moved sugar sat too close to a dose that wrecked the host. Work shifted to relatives that kept the sugar effect without the acute toxicity. Linking two guanidine units produced the biguanides. Dimethylbiguanide - metformin - was among compounds first made in the 1920s, often credited to Werner and Bell in 1922. Synthesis is not a launch. It is a formula in a journal.
Then the room changed. Insulin arrived in the early 1920s and rescued people who were dying. Money, attention, and prestige followed the injection. Oral leftovers from a pasture weed looked minor next to a hormone that could pull someone out of ketoacidosis. Metformin sat in the chemical literature, described and unused, for a generation. Recurring pattern in this class of stories: a molecule is made, briefly noticed, then waits for someone to ask a clinical question the field is ready to hear. Sterne later asked it. The 1920s bench did not.
Sterne pulled it off the shelf and named the eater
The claim that synthesis equals launch does not hold. A French physician had to pick the compound back up.
Jean Sterne, a diabetologist, returned to the biguanides in the 1950s and studied metformin in people rather than leaving it a curiosity. He documented falling blood sugar at doses that did not carry guanidine's acute poison. Mid-1950s clinical work turned a known molecule into a medicine. That is the fileable step: not discovery of a new atom, but proof that a shelf chemical could be given and watched. Residents still skip this and credit 'the 1920s' as if a formula treated a patient. It did not.
Sterne also supplied the name that stuck. Glucophage - glucose eater. France marketed it under that brand in 1957. The launch was modest. Nobody billed it as a drug for hundreds of millions. It behaved like one more oral option in a cautious field for years. The molecule in today's Glucophage 500 mg reference is still dimethylbiguanide. The brand on the old French box is the same word. What changed after 1957 was not the atom. It was who was allowed to swallow it, and under which kidney rule.
The cousin that tarred the whole class
Metformin did not travel alone. Phenformin and buformin reached markets in the same era. Phenformin saw heavy US use. They lowered glucose. They also carried a much higher risk of lactic acidosis - rare, often fatal acid build-up in blood. Death reports stacked through the 1960s and 1970s. In 1977 the United States pulled phenformin. That withdrawal was correct for phenformin. It was then wrongly read as a verdict on every biguanide.
Strike the lazy class verdict. Guilt by structure kept metformin out of the largest market for another two decades, even though the clearance paths differ. Phenformin is partly handled by the liver and can accumulate when that handling fails. Metformin leaves largely unchanged through the kidneys. With decent kidney function it does not pile up the same way, so lactic acidosis risk is far lower. Regulators and clinicians still heard 'biguanide' and flinched. Reputation delayed a usable tablet more than chemistry did. The boxed warning on today's label is the memory of that cousin, written as an accumulation rule, not as a poison stamp.
America said yes late, after Europe had already used it
Open the US date and it looks like a 1990s drug. Open the French date and the gap is almost forty years.
Medieval herbals
Goat's rue listed for thirst and heavy urination; shepherds already knew the plant could kill livestock.
Early 1900s
Guanidine isolated from Galega officinalis; glucose drop confirmed; raw alkaloid too toxic to swallow as a drug.
1922
Dimethylbiguanide synthesised; insulin then swallows the field and the oral candidate sits unused.
1957
Jean Sterne's work reaches the French market as Glucophage, the glucose eater.
1977
US withdraws phenformin for lactic acidosis; the biguanide class is tarred by association.
1995
US practice treats metformin as newly available after the late-1994 FDA clearance.
Late 1990s
UKPDS 34 publishes event reductions that lock the tablet into first-line teaching.
The FDA cleared metformin at the end of 1994. US shelves and teaching shorthand settled on 1995. France had been using Glucophage since 1957. That lag is one of the odder modern delays: a tablet used routinely across Europe while the largest market stayed spooked by phenformin. Once the American door opened, uptake was fast. Clinics that had been told 'biguanide means danger' had to relearn a cousin with a different clearance path. The late-1990s outcomes paper then gave prescribers the event data they wanted - filed on this desk in the UKPDS event review, not here.
From that point the tamed plant became first-line in essentially every major type 2 guideline, cheap because patents are long gone, taken by hundreds of millions. How to start the 500 mg chip without wrecking the gut, and which eGFR number stops the file, sits in the titration and eGFR walkthrough. Chemistry was never the whole story. Evidence and reputation decided when the weed was allowed to be a pill. A patient who asks 'why did nobody give this to my parent in the 1980s in the States?' is asking about phenformin, not about a missing invention.
What 'tamed' means on a 500 mg chip
Tamed does not mean harmless folklore made official. It means the glucose-lowering idea survived a toxicity filter that the plant itself failed. You do not brew goat's rue. You do not forage a 'natural' metformin. The raw plant could drop a flock. The labeled tablet is among the most used chronic drugs we have, with kidney rules that exist because the molecule still leaves through the same organ that would let it accumulate. A 500 mg chip is a starting strength on the US Glucophage label, not a folk measure and not a weight-loss souvenir.
File this for patients who distrust 'chemicals' and for patients who think a weed origin makes the pill gentle. Both claims need a strike. Lineage is a real plant remedy that worked for a real reason. Refinement is the point. The 500 mg Glucophage start is a dose chosen so the gut and the kidneys can be watched, not a souvenir of a medieval garden. If someone offers goat's rue capsules as 'the original Glucophage,' that is a different product with none of the labeled holds. Mayo Clinic keeps a plain overview at Mayo Clinic. For the live US wording, use the FDA label.
Reader mail
Reader questions on this article
Answered by Dr. Sofia Kessler, MD · Internal medicine & clinical pharmacology
Mail after this history file. Claim on the left, strike on the right, Sofia signs the keep.
Yes - and it is useful to say out loud to people who flinch at 'chemicals.' Goat's rue, Galega officinalis, sat in European folk lists for thirst and heavy urination long before blood sugar had a name. The plant is rich in guanidine, which does lower glucose. Metformin is a refined, far safer descendant of that chemistry, not a ground-up sprig in a capsule. Lineage is real. Foraging is not. Raw plant could kill grazing sheep; the 500 mg chip is among the most watched chronic tablets we have. The pharmacology of the finished molecule is in the Glucophage note. Mayo Clinic's public pages at Mayo Clinic stay readable if you want a second plain pass.
Why did the US wait until the mid-1990s when France had Glucophage in 1957?
Fear by association, mostly. Metformin is a biguanide. Cousin phenformin caused dangerous lactic acidosis and was pulled in the United States in 1977. That tarred the class in American eyes even though metformin's risk is much lower, because the clearance paths are not the same. France and much of Europe used metformin routinely for decades. The US held back. When the FDA reviewed the file at the end of 1994, the overseas safety record was the reassurance that chemistry alone had not provided. Approval came; shelves and teaching shorthand settled on 1995. Reputation delayed good care more than a missing molecule did.
What made phenformin so much more dangerous if they are chemical cousins?
How the body clears each drug. Phenformin is partly broken down by the liver and can build up when that handling is poor - and accumulation is what drives lactic acidosis. Metformin: kidneys remove it largely unchanged. With reasonable kidney function it does not pile up the same way. Metformin's lactic acidosis risk is genuinely low; phenformin's was high enough to justify withdrawal. That is also why kidney function is the single most important check before and during metformin. The eGFR holds live in the titration and eGFR walkthrough, not in this history file. Do not treat 'same class' as 'same risk.'
Who named it Glucophage, and does the name still mean anything?
Jean Sterne, a French diabetologist, in the 1950s. He took metformin - sitting in chemical literature since the 1920s - and studied it properly in patients. He coined Glucophage, roughly glucose eater. The drug reached France under that name in 1957. The molecule was synthesised decades earlier; Sterne is fairly credited with turning it into a medicine. The name outlasted almost everyone who remembers the weed. It is still a fair teaching hook: the tablet does not 'eat' sugar like a cartoon, but it cuts the liver's glucose output, which is the clinical meaning behind the brand.
If chemists already had metformin in the 1920s, why did it sit unused?
Insulin, mostly. The early-1920s injection was such a dramatic rescue for people dying of diabetes that it absorbed nearly all attention and money. Oral candidates from goat's rue chemistry, metformin included, looked minor and got shelved. Common delay: made, briefly noted, then waits years for the right clinical question. Metformin waited about thirty years for Sterne to ask it in people rather than in a flask. That wait is not a conspiracy and not proof the drug was weak. It is what happens when one breakthrough swallows a field.
Is the metformin I swallow today the same drug Sterne used in 1957?
Chemically yes - same molecule, dimethylbiguanide. What changed is the surrounding file: kidney-function rules, food-with-dose habits, a later extended-release form that is gentler on the stomach, and decades of event data that Sterne did not have. Dose ranges and eGFR holds are refinement layered on his drug, not a new compound. You are taking essentially what he introduced, with sixty-odd years of experience about how to give it without repeating phenformin's mistakes. Regulatory wording lives at the FDA. A readable patient start is MedlinePlus.
Does a plant origin mean I can treat this as a gentle herbal?
No. That is the claim I strike hardest on this page. Goat's rue was never gentle. It lowered sugar and it killed livestock. The tablet exists because chemists stripped the useful idea out of a toxic plant and then clinicians spent decades learning who can take it. 'Natural' is not a safety grade. If someone is selling goat's rue tea as a metformin substitute, that is a different product with none of the labeled controls. Stay with the prescribed 500 mg chip and the kidney checks. Do not forage a first-line diabetes drug.
What did UKPDS actually add if Europe already trusted the tablet?
Events, not a new molecule. Europe already had years of use. What it did not have, in a form that moved guidelines worldwide, was a long randomised comparison showing fewer diabetes-related deaths and fewer heart attacks in overweight people on metformin versus conventional policy. That paper - UKPDS 34, late 1990s - is why the old, slightly suspect biguanide became first-line teaching rather than a cheap European habit. I file the numbers on the UKPDS event review. This history page only needs the date: after the trial, reputation finally matched the chemistry.
Why do you keep saying 500 mg if the history is about a plant?
Because this desk locks the labeled chip for the domain, and 500 mg is the Glucophage start that survived the taming. History without a strength becomes a folklore reel. The plant had no milligram. Sterne's medicine did. The US label still opens adult immediate-release at 500 mg with meals, or 850 mg once daily, then climbs. I keep 500 mg in the title so the file points at a real tablet, not a weed sketch. How you climb from that chip is a different article. The origin story ends when the dose becomes something a kidney number can govern.
General education from a clinician, not personal medical advice. Bring your own history to your own prescriber.
Last Updated