When Chemie Grünenthal launched thalidomide in West Germany as Contergan on 1 October 1957, it sold the compound as a sedative so safe that a child could swallow the whole bottle without harm, marketed it without prescription, and promoted it specifically to pregnant women for morning sickness; the documented record shows that the drug was instead one of the most powerful human teratogens ever distributed, and that it reached pregnant patients across roughly 46 countries before any company or regulator had required a single study of its effects on the unborn. Thalidomide, sold in the United Kingdom and Commonwealth by the Distillers Company as Distaval and in dozens of other markets under more than forty trade names, was withdrawn in West Germany on 26 November 1961 after the Hamburg pediatrician Widukind Lenz formally linked it to an epidemic of phocomelia — infants born with foreshortened or flipper-like limbs.
The gap between the promise and the harm was total and structural. Grünenthal advertised the drug as “completely safe” and “non-toxic,” yet it had performed no studies of the drug crossing the placenta and no testing in pregnant animals before approval; the standard toxicity tests of the era, measured by acute lethal dose, registered thalidomide as remarkably benign precisely because its danger lay not in poisoning the patient but in disrupting the developing embryo during a narrow window early in pregnancy. The verdict here is therefore plain at the outset: a non-prescription product advertised as harmless to mother and child was, in fact, causing catastrophic limb, ear, eye, and internal-organ malformations, and the absence of reproductive-safety testing meant the signal could only surface after thousands of affected infants had already been born.
The reckoning that followed reshaped how the world licenses medicines. William McBride’s letter in The Lancet on 16 December 1961 and Lenz’s epidemiological work converted a cluster of “mysterious” birth defects into a recognized drug catastrophe. In the United States, the drug was never approved: FDA medical reviewer Frances Oldham Kelsey had refused to clear Richardson-Merrell’s application (trade name Kevadon) for over a year, repeatedly demanding the safety data the company could not supply. The episode produced the Kefauver-Harris Amendments, signed on 10 October 1962, which for the first time required proof of efficacy as well as safety and established the modern controlled-trial regime.
The Contergan name became the permanent byword for what happens when marketing velocity, over-the-counter access, and the complete absence of reproductive testing converge on a vulnerable population — and for how a single regulator’s refusal can avert a national disaster.
When the U.S. Food and Drug Administration asked manufacturers to pull propoxyphene from the American market on 19 November 2010, it was retiring a painkiller that Eli Lilly had introduced in 1957 and that had been prescribed to tens of millions of people across two regulatory generations; the documented record shows the agency acted only after a study it had itself ordered proved the harm, and decades after the first petition to ban the drug was filed. Propoxyphene — marketed alone as Darvon and combined with acetaminophen as Darvocet — was promoted for half a century as a mild, well-tolerated opioid for moderate pain. The gap between that gentle reputation and the molecule’s actual pharmacology was the whole story: at or near ordinary therapeutic doses the drug prolonged the PR interval, widened the QRS complex, and lengthened the QT interval, the electrocardiographic signature of a compound that can stop a heart.
The harm was not subtle and not new. Propoxyphene blocks cardiac sodium channels more potently than the antiarrhythmic agents lidocaine, quinidine, and procainamide, and its long-lived metabolite norpropoxyphene accumulates — especially in the elderly and in patients with impaired kidneys — pushing toxic effects past the point of reversal. The consumer group Public Citizen petitioned the FDA to ban propoxyphene in 1978 and again in February 2006. The drug had been associated with more than 2,000 accidental U.S. deaths since 1981 and ranked among the most common drugs found in fatal overdoses, a mortality profile that prompted the United Kingdom to begin withdrawing the equivalent product, co-proxamol, in January 2005.
The verdict here is therefore plain at the outset: an approved, familiar, “weak” medicine reached enormous populations while the evidence that would eventually condemn it — sodium-channel data, autopsy series, decades of overdose statistics, and a foreign regulator’s reversal — sat fully legible in the literature. The thing that finally moved the FDA was not new theory but a single dedicated experiment: a controlled electrocardiographic study in healthy volunteers, completed in 2010, showing that even at labeled doses propoxyphene measurably deranged cardiac conduction.
What followed the withdrawal was less a courtroom reckoning than a public-health correction visible in the morgue. In the state of Florida, where overdose deaths are tracked closely, fatalities involving propoxyphene fell on the order of 84 percent after the drug left the market — a natural experiment that quantified, in lives, the cost of every year the agency had waited. Propoxyphene became the standard byword for regulatory delay: a case in which the question was not whether a drug was dangerous but how long an agency could decline to answer a petition it had already received twice.
When Joseph Califano, Secretary of Health, Education and Welfare, signed the order suspending phenformin on 25 July 1977, he was invoking a power Congress had granted the Food and Drug Administration fifteen years earlier and which no official had ever used: the declaration that an approved medicine was an “imminent hazard to the public health.” Phenformin (phenethylbiguanide), synthesized in 1957 by a team including Seymour Shapiro at the U.S. Vitamin Corporation and marketed in the United States from 1959 by Ciba-Geigy as DBI, had been prescribed to control adult-onset diabetes for nearly two decades. It was withdrawn not because it failed to lower blood sugar — it did that adequately — but because it could quietly poison the body with lactic acid, a complication that killed roughly one of every two patients it reached.
The gap between the promise and the harm was the gap between a surrogate benefit and a survival outcome. Phenformin reliably reduced glycemia, the number physicians could see on a chart; what it did not reliably do was keep patients alive. The biguanide’s mechanism — pushing cellular metabolism toward anaerobic glycolysis — produced excess lactate, and in patients with even modest kidney or heart impairment that lactate could accumulate into a metabolic catastrophe. Phenformin-associated lactic acidosis ran at roughly 40 to 64 cases per 100,000 patient-years, and once it occurred it was fatal in approximately half of cases. The drug treated a chronic disease by trading a visible, manageable problem for an occasional, lethal one.
The verdict was therefore plain long before the suspension. As early as 1959 — the same year U.S. marketing began — the literature linked biguanides to lactic acidosis. In 1971 the federally funded University Group Diabetes Program (UGDP) terminated its phenformin arm after finding excess total and cardiovascular mortality, roughly double that of the placebo and insulin groups combined. By October 1976 the FDA’s own Endocrinology and Metabolism Advisory Committee had recommended removal. The harm sat in the published record for years while ~385,000 Americans were still taking the drug in 1977.
What forced the cord was not the agency but a consumer group and a lawyer’s petition. Ralph Nader’s Health Research Group petitioned for suspension on 22 April 1977; Califano acted in July; the drug was off the U.S. market by 15 November 1978. The episode became the founding precedent for emergency drug suspension, the cautionary prologue to its safer sister biguanide metformin, and the textbook case of how a medicine can be effective on the metric it was sold on and still be a net killer.