LL1-08 digest — Ch8 The DES story: long-term consequences of prenatal exposure#

Late lessons from early warnings (EEA, 2001), pp. 84–92.

• Authors: Dolores Ibarreta (EC JRC-IPTS; endocrine-disrupter risk assessment) and Shanna H. Swan (University of Missouri; DES reproductive epidemiologist, described in her bio as “always” precautionary, p. 199).
• No panels or dissent. The editors warn generally that case-study authors “are not without strong views” (p. 12).
• Full notes: notes/LL1-08.md.

Core story#

DES was a cheap, unpatented, potent synthetic oestrogen, first made in 1938 and sold under more than 200 brand names (pp. 84, 88). From the 1940s it was prescribed to prevent miscarriage (US approval 1947). The rationale was a theory that got cause and effect backwards, and the “proof” came from proponents’ uncontrolled trials (p. 84).

• Warning signs before and during use. Oestrogen carcinogenicity was known from 1933. DES caused cancer in rodents from 1938 onward. Developmental effects in animals (1939), oestrogenic signs in exposed newborns (1945), a proponent’s caution about latent effects (1948) and further animal cancers at the relevant sites (1959, 1963) were reported. The chapter says these were “largely disregarded” or “largely ignored” (pp. 85, 88).
• Inefficacy ignored. Randomised trials in the early 1950s (Dieckmann et al., 1953) showed DES did not work. Promotion continued anyway, including “routine prophylaxis in all pregnancies” (p. 86; Fig. 8.2). The chapter calls two decades of prescribing after inefficacy was shown “a massive failure of the system” (p. 86). A 1978 reanalysis of the 1953 data, using methods available at the time, found DES had increased the very risks it was sold to prevent (p. 86). There was no regulatory action, though the chapter’s own Figure 8.1 shows one manufacturer’s sales falling by about three-quarters in the mid-1950s, a decline the chapter does not discuss (p. 85; auditor’s reading).
• Accidental detection. What ended use in pregnancy was a “fortuitous accident” (p. 86). A cluster of a very rare vaginal cancer in young women at one hospital was traced to prenatal exposure (7 of 8 cases vs 0 of 32 controls; p. 84).
• Regulatory lag. The FDA acted within about seven months (pp. 84, 86). Europe acted 3–6 years later (Netherlands 1974, France 1977), with use reported in Spain until 1983, a delay “as long as 12 years” (p. 89). Use continued in developing countries to 1985, and in Mexico, Uganda and Poland into the early 1990s (p. 89, one secondary source).
• Scale of harm. The cancer affects fewer than 1 in 1,000 exposed daughters. The common reproductive-tract and pregnancy harms were found because the cancer triggered follow-up; the authors judge it “unlikely” they would ever have been identified otherwise (pp. 86–87). An estimated 2–10 million people were exposed in the womb, and mouse studies suggest effects may extend to a third generation (pp. 87–88).

The authors’ lessons (pp. 88–90)#

1. Hidden, long-delayed effects of hormonal exposure during development can be devastating, so “extreme caution” is needed.
2. The absence of visible, immediate teratogenic effects is not proof that there is no reproductive toxicity.
3. Signals of harm before and during marketing were ignored; in hindsight, scientists’ concerns “should have been heeded” (p. 90).
4. A proper efficacy trial might have prevented the whole episode.
5. Strongest: even if 1947 marketing “could have been justified” (a hypothetical concession), once DES was shown ineffective in 1953 there was no justification for any cancer risk, so use in pregnancy should have been contraindicated then.
6. Surveillance must look beyond gross malformations and cancers. Legacy drugs and the rare use of formal risk/benefit analysis remain gaps, despite real progress since 1970 (pp. 89–90).
7. After 1971, “economic interests predominated”, producing a “wait and see” approach by manufacturers and regulators.
8. DES as a warning about hormone-disrupting chemicals generally. This is an advocacy-leaning extrapolation beyond the case; the authors acknowledge the dose gap but do not bridge it.

Main mechanisms#

• Theory over evidence. A plausible but reversed causal theory, plus enthusiast-produced evidence, stood in for proof of outcomes.
• Uptake without evidence. Uptake ran ahead of evidence: efficacy was not required, promotion was heavy, lay publicity drove patient demand, and peer pressure meant refusing took a “courageous physician” (p. 88). Clinicians also wanted to act.
• Doubts absorbed. Several warnings came from proponents themselves (Dodds’s co-authored 1933 paper, Karnaky, Laplan) and were “largely ignored” (p. 88); the chapter describes no suppression.
• Asymmetric response. Proof that DES did not work produced no regulatory action for about 18 years; a distinctive harm got US action in months. In Europe, even the harm signal took years to act on.
• Detection by chance. Detection hinged on a rare signature outcome clustering at one hospital, where, per the chapter, local use had been high because the Smiths had experimented there (p. 86). The proponents’ concentrated use created the detectable signal.
• Uneven, slow exit. Jurisdictions diverged, and use continued for years in countries that had not acted (the chapter documents continued use, not redirected sales). Each use was restricted separately, with US production ending only in 1997 (p. 89).

Transferable insights (technology-neutral)#

Numbered as in the notes (1–16).

Main caveats#

• Motives are asserted, not documented. There are no internal documents, and “undoubtedly” does much of the causal work (p. 88).
• Hindsight. The claim that trial methods were available before 1947 is shaky on background knowledge (to verify). The post-1953 argument is unaffected. The authors frame their judgement explicitly as hindsight (p. 90), which sits uneasily with the editors’ brief to judge by “the spirit of the times”, not “the luxury of hindsight” (p. 11).
• Unquantified counterfactual. The claim that withdrawal in 1953 would have spared “millions” (p. 90) is not quantified. Figure 8.1 suggests much US exposure preceded 1953–55, while most European use came later (Box 8.1).
• Early animal evidence partly generic. Two of the four 1938–41 papers cited as showing DES carcinogenicity concern oestrogens or sex hormones generally, by reference title (pp. 85, 91–92).
• Internal inconsistencies:
• exposure counts (pp. 86–87), with a worldwide figure referred to a Box that does not contain it;
• start dates of use;
• “last use 1985” in Table 8.1 vs early-1990s use in the text (pp. 89–90);
• the UK described as a top user but credited with only 7,000–8,000 pregnancies (p. 87);
• “seven” clear-cell cases in the text vs “6 clear-cell” in the 1970 reference title (pp. 84, 91);
• the “12-year” European delay measured to last reported use, not withdrawal (p. 89);
• a mislabelled axis on Figure 8.1.
• Tension within the report. The editors call DES’s next-generation effects “a complete surprise” (p. 170); the chapter argues the warnings were there and ignored.
• Omissions: costs (the editors had asked for brevity, p. 12), litigation, the role of affected people’s groups, long-term harms to mothers, the legal basis for approving DES without proof of efficacy, and why Europe lagged.
• Source provenance. Several European and post-1971 claims (the 18-country survey, Spanish use to 1983, continued use abroad, the end of US production) rest on DES Action or consumer-organisation reports, which the chapter does not flag. Box 8.1’s national figures draw partly on peer-reviewed studies.
• An easy case. With zero benefit, DES does not test the harder trade-off where benefits are real.