The Drug That Survived Its Own Family
The FDA pulled metformin from the American market in 1977 alongside phenformin — a drug with a fatal safety record. Metformin had no such record. American diabetics would wait seventeen years for a drug Europeans had been prescribing since 1958.
Written by Seth Collins, Pharm.D.
Updated on May 28, 2026
Drug History
0 viewsIn 1978, the FDA pulled phenformin from the American market after a decade of watching it kill diabetic patients through lactic acidosis, a condition in which acid accumulates in the blood faster than the body can clear it. Mortality once the cascade was established ran between thirty and fifty percent. The withdrawal was correct. The decision the FDA had already made the year before, in 1977, was not.
Metformin, a chemical cousin in the same drug family, had been pulled from the American market alongside phenformin under regulatory pressure on the manufacturer.
Metformin had no comparable safety problem. It does not accumulate in the body the way phenformin does and had never produced a meaningful lactic acidosis signal in the substantial European clinical data already available by 1977. French and British physicians had been prescribing it as first-line oral therapy for nearly twenty years without producing the case reports that had killed phenformin. The FDA knew this. The decision to pull metformin anyway was a combination of regulatory caution, manufacturer liability concern, and the practical difficulty of defending any biguanide to a public that had just watched another biguanide kill diabetic patients. American type 2 diabetics would go without the best oral diabetes medication in the world for the next seventeen years because of a drug that was not the one being denied to them.
The plant came first.
Galega officinalis grows across Europe and western Asia, a leguminous perennial with pale purple flowers and a long history in folk medicine. In medieval Europe it was used under several names, goat's rue among them, and it carried a particular reputation in communities where diabetes-like symptoms were recognized: the sweetness of urine, the unquenchable thirst, the wasting that followed. Practitioners brewed preparations from the plant and gave them to patients for five hundred years without knowing what the plant actually contained. In 1918, chemists finally identified the active compound as guanidine, a molecule with measurable blood-sugar-lowering properties. The folk medicine had been pharmacologically real the entire time.
Guanidine itself was too toxic for clinical use, so chemists in the 1920s began synthesizing safer derivatives. The family that resulted was called the biguanides. Two chemists named Werner and Bell made metformin in 1922. Phenformin and buformin came out of the same period. They were published in the literature, studied intermittently for the next thirty years, and largely set aside while the pharmaceutical industry chased insulin, which had arrived in 1921 and reorganized every diabetes research priority for a generation.
Jean Sterne was a French physician in Paris in the 1950s who had been reading the biguanide literature and believed metformin deserved a serious clinical trial. He ran one. In 1957 he published the first human data on metformin in type 2 diabetes and named the drug Glucophage. Glucose eater.
The results held up. Metformin lowered blood sugar without causing the hypoglycemia that insulin produced. It was oral, inexpensive, and well-tolerated. The United Kingdom approved it in 1958. Most of Europe followed within a few years. By the early 1960s metformin was routine first-line care for type 2 diabetes across the continent.
The American story moved differently. The FDA approved phenformin instead of metformin in 1959, partly because phenformin was more potent on paper and partly because the manufacturer pursuing American approval had bet on the wrong biguanide. American physicians began prescribing phenformin. Case reports of lactic acidosis accumulated through the 1960s and into the 1970s. The pattern was clear by the mid-decade, but the FDA did not act until 1977. When it finally did, it pulled the entire category.
The seventeen years that followed are the most consequential gap in the history of American diabetes care. European physicians had a working drug. American physicians did not. American type 2 diabetic patients were managed on sulfonylureas (which work but cause hypoglycemia and weight gain), insulin (effective but injection-based and harder to dose safely), or dietary intervention alone. None of these approaches matched what metformin was doing in Europe. The drug was sitting on pharmacy shelves in London and Paris while American patients were progressing to complications it could have prevented.
The UK Prospective Diabetes Study changed the calculation when it published in 1998. The trial was large, long-running, and designed to examine how different diabetes treatments affected outcomes over time rather than just blood sugar numbers. The data showed that metformin reduced cardiovascular mortality and all-cause mortality in overweight type 2 diabetic patients in a way that went beyond glucose control. The drug was keeping people alive through a mechanism that sulfonylureas and insulin did not match.
The FDA had already approved metformin by then. Bristol-Myers Squibb launched it in the United States as Glucophage in 1994, thirty-seven years after Sterne published in Paris. The UKPDS data published four years later confirmed what European clinicians had been observing for a generation. American guidelines shifted, metformin moved to first-line therapy, and it has held that position ever since.
What the UKPDS data also implied, although the trial was not designed to measure it, is that the seventeen-year American gap was not just a matter of suboptimal glycemic control. It was a matter of cardiovascular mortality. American type 2 diabetic patients during those seventeen years died at rates that European patients on metformin did not. The exact body count cannot be computed. It was not zero.
After more than sixty years of continuous clinical use, why metformin works remains incompletely understood.
The cellular pathway involves an enzyme called AMPK, which acts as a sensor of cellular energy status. Metformin activates AMPK indirectly through effects on mitochondrial function, which then signals the liver to reduce glucose output. The molecular details of exactly how metformin reaches AMPK are still being worked out in research published in the last five years. The pattern is familiar in pharmacology. Aspirin was used for decades before the prostaglandin mechanism was characterized. Colchicine's tubulin binding was not described until the 1970s after three thousand years of clinical use. Metformin is unusual mainly in how recent the gap is. The drug entered European practice in 1958 and the underlying molecular biology is still producing papers in 2024.
A parallel research program now examines whether metformin does something beyond diabetes care. Diabetic patients on metformin appear to have lower rates of certain cancers and cardiovascular conditions even after correcting for glucose control, which suggests the drug is hitting something more upstream than blood sugar. The TAME trial, Targeting Aging with Metformin, is an ongoing clinical trial designed to test whether the drug slows biological aging in non-diabetic adults. If it succeeds, metformin will be the first drug formally trialed and approved as an anti-aging intervention. Sterne named it glucose eater in 1957, and current research is examining whether the name was incomplete.
Metformin costs between four and twelve dollars for a month's supply at most pharmacies today. It has been generic for decades. The molecule is the same one Sterne gave his patients in Paris in 1957.
If you are paying significantly more than that, the difference is not the drug. The difference is the system stacked around the drug: the office visit required to renew the prescription, the insurance gating, the pharmacy benefit manager negotiations on a generic that costs almost nothing to manufacture, the brand-name markups for products pharmacologically identical to the generic version. None of this has anything to do with metformin itself, or with the European clinical practice that proved it worked, or with the seventeen years American patients spent without it.
Pillar Drug Club was built on the premise that the people who take these medications every year should understand what they are actually taking, and should not pay more than what the molecule costs to manufacture and dispense. If you take metformin, you are taking a drug that American regulators withheld for seventeen years for reasons that were not pharmacological. The price you pay for it now should reflect what the molecule actually is, which is to say, it should be cheap.
Sources
Sterne J. (1957). Du nouveau dans les antidiabetiques. La NN dimethylamine guanyl guanide (N.N.D.G.). Maroc Medical, 36, 1295-1296.
Bailey CJ, Turner RC. (1996). Metformin. New England Journal of Medicine, 334(9), 574-579. https://doi.org/10.1056/NEJM199602293340906
UK Prospective Diabetes Study (UKPDS) Group. (1998). Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes. Lancet, 352(9131), 854-865. https://doi.org/10.1016/S0140-6736(98)07037-8
Witters LA. (2001). The blooming of the French lilac. Journal of Clinical Investigation, 108(8), 1105-1107. https://doi.org/10.1172/JCI14178
Foretz M, Guigas B, Viollet B. (2019). Understanding the glucoregulatory mechanisms of metformin in type 2 diabetes mellitus. Nature Reviews Endocrinology, 15(10), 569-589. https://doi.org/10.1038/s41574-019-0242-2
Barzilai N, Crandall JP, Kritchevsky SB, Espeland MA. (2016). Metformin as a tool to target aging. Cell Metabolism, 23(6), 1060-1065. https://doi.org/10.1016/j.cmet.2016.05.011
Misbin RI. (2004). The phantom of lactic acidosis due to metformin in patients with diabetes. Diabetes Care, 27(7), 1791-1793. https://doi.org/10.2337/diacare.27.7.1791
The Drug Files is a series on the molecules behind modern medicine. Written by Seth Collins, PharmD.
Metformin, a chemical cousin in the same drug family, had been pulled from the American market alongside phenformin under regulatory pressure on the manufacturer.
Metformin had no comparable safety problem. It does not accumulate in the body the way phenformin does and had never produced a meaningful lactic acidosis signal in the substantial European clinical data already available by 1977. French and British physicians had been prescribing it as first-line oral therapy for nearly twenty years without producing the case reports that had killed phenformin. The FDA knew this. The decision to pull metformin anyway was a combination of regulatory caution, manufacturer liability concern, and the practical difficulty of defending any biguanide to a public that had just watched another biguanide kill diabetic patients. American type 2 diabetics would go without the best oral diabetes medication in the world for the next seventeen years because of a drug that was not the one being denied to them.
The plant came first.
Galega officinalis grows across Europe and western Asia, a leguminous perennial with pale purple flowers and a long history in folk medicine. In medieval Europe it was used under several names, goat's rue among them, and it carried a particular reputation in communities where diabetes-like symptoms were recognized: the sweetness of urine, the unquenchable thirst, the wasting that followed. Practitioners brewed preparations from the plant and gave them to patients for five hundred years without knowing what the plant actually contained. In 1918, chemists finally identified the active compound as guanidine, a molecule with measurable blood-sugar-lowering properties. The folk medicine had been pharmacologically real the entire time.
Guanidine itself was too toxic for clinical use, so chemists in the 1920s began synthesizing safer derivatives. The family that resulted was called the biguanides. Two chemists named Werner and Bell made metformin in 1922. Phenformin and buformin came out of the same period. They were published in the literature, studied intermittently for the next thirty years, and largely set aside while the pharmaceutical industry chased insulin, which had arrived in 1921 and reorganized every diabetes research priority for a generation.
Jean Sterne was a French physician in Paris in the 1950s who had been reading the biguanide literature and believed metformin deserved a serious clinical trial. He ran one. In 1957 he published the first human data on metformin in type 2 diabetes and named the drug Glucophage. Glucose eater.
The results held up. Metformin lowered blood sugar without causing the hypoglycemia that insulin produced. It was oral, inexpensive, and well-tolerated. The United Kingdom approved it in 1958. Most of Europe followed within a few years. By the early 1960s metformin was routine first-line care for type 2 diabetes across the continent.
The American story moved differently. The FDA approved phenformin instead of metformin in 1959, partly because phenformin was more potent on paper and partly because the manufacturer pursuing American approval had bet on the wrong biguanide. American physicians began prescribing phenformin. Case reports of lactic acidosis accumulated through the 1960s and into the 1970s. The pattern was clear by the mid-decade, but the FDA did not act until 1977. When it finally did, it pulled the entire category.
The seventeen years that followed are the most consequential gap in the history of American diabetes care. European physicians had a working drug. American physicians did not. American type 2 diabetic patients were managed on sulfonylureas (which work but cause hypoglycemia and weight gain), insulin (effective but injection-based and harder to dose safely), or dietary intervention alone. None of these approaches matched what metformin was doing in Europe. The drug was sitting on pharmacy shelves in London and Paris while American patients were progressing to complications it could have prevented.
The UK Prospective Diabetes Study changed the calculation when it published in 1998. The trial was large, long-running, and designed to examine how different diabetes treatments affected outcomes over time rather than just blood sugar numbers. The data showed that metformin reduced cardiovascular mortality and all-cause mortality in overweight type 2 diabetic patients in a way that went beyond glucose control. The drug was keeping people alive through a mechanism that sulfonylureas and insulin did not match.
The FDA had already approved metformin by then. Bristol-Myers Squibb launched it in the United States as Glucophage in 1994, thirty-seven years after Sterne published in Paris. The UKPDS data published four years later confirmed what European clinicians had been observing for a generation. American guidelines shifted, metformin moved to first-line therapy, and it has held that position ever since.
What the UKPDS data also implied, although the trial was not designed to measure it, is that the seventeen-year American gap was not just a matter of suboptimal glycemic control. It was a matter of cardiovascular mortality. American type 2 diabetic patients during those seventeen years died at rates that European patients on metformin did not. The exact body count cannot be computed. It was not zero.
After more than sixty years of continuous clinical use, why metformin works remains incompletely understood.
The cellular pathway involves an enzyme called AMPK, which acts as a sensor of cellular energy status. Metformin activates AMPK indirectly through effects on mitochondrial function, which then signals the liver to reduce glucose output. The molecular details of exactly how metformin reaches AMPK are still being worked out in research published in the last five years. The pattern is familiar in pharmacology. Aspirin was used for decades before the prostaglandin mechanism was characterized. Colchicine's tubulin binding was not described until the 1970s after three thousand years of clinical use. Metformin is unusual mainly in how recent the gap is. The drug entered European practice in 1958 and the underlying molecular biology is still producing papers in 2024.
A parallel research program now examines whether metformin does something beyond diabetes care. Diabetic patients on metformin appear to have lower rates of certain cancers and cardiovascular conditions even after correcting for glucose control, which suggests the drug is hitting something more upstream than blood sugar. The TAME trial, Targeting Aging with Metformin, is an ongoing clinical trial designed to test whether the drug slows biological aging in non-diabetic adults. If it succeeds, metformin will be the first drug formally trialed and approved as an anti-aging intervention. Sterne named it glucose eater in 1957, and current research is examining whether the name was incomplete.
Metformin costs between four and twelve dollars for a month's supply at most pharmacies today. It has been generic for decades. The molecule is the same one Sterne gave his patients in Paris in 1957.
If you are paying significantly more than that, the difference is not the drug. The difference is the system stacked around the drug: the office visit required to renew the prescription, the insurance gating, the pharmacy benefit manager negotiations on a generic that costs almost nothing to manufacture, the brand-name markups for products pharmacologically identical to the generic version. None of this has anything to do with metformin itself, or with the European clinical practice that proved it worked, or with the seventeen years American patients spent without it.
Pillar Drug Club was built on the premise that the people who take these medications every year should understand what they are actually taking, and should not pay more than what the molecule costs to manufacture and dispense. If you take metformin, you are taking a drug that American regulators withheld for seventeen years for reasons that were not pharmacological. The price you pay for it now should reflect what the molecule actually is, which is to say, it should be cheap.
Sources
Sterne J. (1957). Du nouveau dans les antidiabetiques. La NN dimethylamine guanyl guanide (N.N.D.G.). Maroc Medical, 36, 1295-1296.
Bailey CJ, Turner RC. (1996). Metformin. New England Journal of Medicine, 334(9), 574-579. https://doi.org/10.1056/NEJM199602293340906
UK Prospective Diabetes Study (UKPDS) Group. (1998). Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes. Lancet, 352(9131), 854-865. https://doi.org/10.1016/S0140-6736(98)07037-8
Witters LA. (2001). The blooming of the French lilac. Journal of Clinical Investigation, 108(8), 1105-1107. https://doi.org/10.1172/JCI14178
Foretz M, Guigas B, Viollet B. (2019). Understanding the glucoregulatory mechanisms of metformin in type 2 diabetes mellitus. Nature Reviews Endocrinology, 15(10), 569-589. https://doi.org/10.1038/s41574-019-0242-2
Barzilai N, Crandall JP, Kritchevsky SB, Espeland MA. (2016). Metformin as a tool to target aging. Cell Metabolism, 23(6), 1060-1065. https://doi.org/10.1016/j.cmet.2016.05.011
Misbin RI. (2004). The phantom of lactic acidosis due to metformin in patients with diabetes. Diabetes Care, 27(7), 1791-1793. https://doi.org/10.2337/diacare.27.7.1791
The Drug Files is a series on the molecules behind modern medicine. Written by Seth Collins, PharmD.
Tags
Issue 005
Metformin
FDA
Diabetes
Drug History