Why Thalidomide Looked Safe at First
In the late 1950s, thalidomide was widely prescribed in several countries as a sedative and as a treatment for nausea during pregnancy. Its apparent safety was misleading: tests available at the time did not adequately reveal how a medicine could affect a developing embryo. A drug can leave an adult unharmed yet interfere with the formation of a fetus’s limbs, organs, or other structures, especially during a narrow window of development.
Thalidomide disrupted the growth of limbs when exposure occurred at a critical stage. The resulting injuries could be severe, including arms or legs that were extremely shortened or absent. Because the damage depended on when during pregnancy exposure occurred, the link was not immediately obvious from ordinary reports of side effects.
The Reviewer Who Asked for Better Evidence
Frances Kelsey, a physician and pharmacologist at the U.S. Food and Drug Administration, reviewed the application to market thalidomide in the United States. She did not consider the submitted evidence sufficient and held up approval, including because reports raised concerns about nerve damage and the available safety information was inadequate. Her decision kept the drug from widespread U.S. use before the birth-defect crisis became clear elsewhere.
The episode exposed a weakness in treating a medicine’s lack of obvious harm in routine use as proof that it was safe. Evidence must address who will take the drug, how it behaves in the body, and whether it can affect pregnancy and fetal development—not just whether adults tolerate it.
From Safety Checks to Proof of Benefit
The scandal helped drive major changes in U.S. drug regulation. The 1962 Kefauver–Harris amendments required manufacturers to provide evidence that drugs were effective as well as safe, and strengthened the FDA’s role in reviewing clinical trials. That distinction matters: a medicine may have risks, but regulators need evidence that its benefits justify them and that the risks are understood.
The lasting lesson is not that testing can eliminate every danger. It is that testing must be designed to find harms that ordinary use may conceal—and that approval should depend on demonstrated benefit, not reassurance alone.
