Skip to content
Prof. Dr. Mustafa Özdoğan · Medical Oncologist

From a Failed Contraceptive to Tamoxifen: The Path of ICI 46,474, 1962-1977

From a Failed Contraceptive to Tamoxifen: The Path of ICI 46,474, 1962-1977
In this article

 

From a Failed Contraceptive to Tamoxifen: The Path of ICI 46,474, 1962-1977

Period: 1962-1998 (from the synthesis of the molecule to the 1977 US approval and the 1998 pooled analysis)
Place: Cheshire, England (ICI Pharmaceuticals); the Christie Hospital in Manchester; the Worcester Foundation in Massachusetts; Leeds and Cambridge

On Day 161 we looked for the cause of cervical cancer in a virus. Today we return to the hormone axis: to Beatson's ovary removal of 1896 on Day 91, to Lacassagne's estrogen-induced tumour in male mice on Day 105, to Huggins's hormonal suppression of prostate cancer on Day 112. In all of them the hormone was either removed from its source or balanced against another hormone. In today's story, for the first time, a drug tries to block the hormone's attachment to the cell. As promised in yesterday's tomorrow box, I verified separately what exactly 1977 is the date of, the dates of synthesis and first clinical use, and when the mechanism of action was understood; I did not mix chemistry with clinical evidence.

— A FAILED CONTRACEPTIVE PILL —

A number in a laboratory notebook

The story begins not with a search for a cancer drug but with a search for a contraceptive pill. According to a 2017 case study of drug development in Frontiers in Pharmacology (which I read with a summarising tool), the fertility regulation programme of ICI Pharmaceuticals was run by the biologist Arthur Walpole, and in 1962 the chemist Dora Richardson synthesised the compound numbered ICI 46,474. Scientific American's article on Richardson gives the same year. Only these two secondary sources give the year of synthesis; the 2023 review by Howell and Howell gives no exact year of synthesis and mentions only the isomer papers of 1966. So 1962 is a date "according to two secondary sources".

The compound was a mixture. According to Howell and Howell, the cis isomer (ICI 47,699) was a pure estrogen, while the trans isomer, that is tamoxifen, had a mixed estrogen and anti-estrogen action. According to the Frontiers article, Richardson separated the isomers, and an X-ray analysis in 1968 confirmed that the anti-estrogenic activity lay in the trans isomer. According to Scientific American, Richardson's name is on a patent of 1965, but her contribution was largely overlooked in later accounts; the same article names an indexing error as one reason. I relay this from a single secondary source.

The aim was to prevent pregnancy with an anti-estrogen. It did not work: according to the Frontiers article, trials in Sweden showed that tamoxifen stimulated ovulation instead of suppressing it. As a contraceptive the molecule had failed, and for the company it looked like an obstacle.

1969-1971

The first clinical study at the Christie Hospital

According to the Frontiers article, in 1969 the British drug safety committee approved therapeutic studies, among them a breast cancer study at the Christie Hospital in Manchester. The results were published in the British Journal of Cancer: Cole, Jones and Todd, "A New Anti-oestrogenic Agent in Late Breast Cancer: An Early Clinical Appraisal of ICI46474", 1971, volume 25, pages 270-275. According to the journal's abstract page, 46 patients were treated; 10 showed a good response, which the authors found similar to that seen with estrogens and androgens, and they noted that the drug's main advantage was a "low incidence of troublesome side effects". Howell and Howell also describe the same work as a single-arm phase 2 trial that began in 1969 with 46 patients and report a 22% response; for comparison they give 25% for diethylstilbestrol (DES, Day 110) and 18% for high-dose androgens.

This was not a striking result. According to Craig Jordan's own account (Cancerworld, 2020), the response rate in advanced disease was about 30% and lasted one to two years, which was no better than existing hormonal treatments, and ICI doubted it would be profitable. According to the same account Walpole argued against dropping the drug. This is a story told from Jordan's point of view; the Frontiers article gives a different figure for the response (measurable shrinkage in all forty women), and I recorded this conflict in the fact-check box.

SERM and the receptor

Estrogen acts by binding to a protein in the cell called the estrogen receptor. Tamoxifen blocks this binding, but not in the same way in every tissue: in some tissues it partly behaves like an estrogen. This is why it is called a "selective estrogen receptor modulator" (SERM). It is also a prodrug: in the liver the enzymes CYP2D6 and CYP3A4 convert it into the more potent 4-hydroxytamoxifen and endoxifen (Howell and Howell 2023; a drug data page from Washington University). This is a simplified definition that gives the main line.

— THE LABORATORY QUESTION: IN WHOM, AND FOR HOW LONG? —

The rat model and the receptor

Laboratory questions shaped how the drug settled into the clinic. According to the obituary that the University of Leeds published for Craig Jordan (written by a former student, with no independent reference list), between 1972 and 1974 Jordan set up a DMBA rat model of mammary tumours at the Worcester Foundation in Massachusetts, tested the effect of ICI 46,474 on cancer, and found that the drug worked only in estrogen receptor positive tumours. According to Jordan's own account, in 1973 he visited Elwood Jensen's laboratory in Chicago, and by 1975 his own laboratory had shown that tamoxifen blocks the binding of estradiol to the estrogen receptor in human tumours. The same account says that making receptor testing standard practice then took about another ten years.

Two things have to be kept apart. First, these are animal and laboratory findings; what showed the same effect in humans was the clinical studies done afterwards. Second, much of this information comes from Jordan's own accounts and from texts that honour him; that is, the person telling the story is a participant in it. I kept this separate with attributions. In a 1997 review Jordan summarises his 1972-74 work as showing that tamoxifen inhibited chemically induced rat mammary tumours, favoured ER-positive tumours, and blocked [3H]estradiol binding in human tumours; the 1975 date appears in the Cancerworld account, and the dates of the two sources do not fully coincide.

Jordan's main proposal concerned duration. According to his account in Cancerworld, in the rat experiments the earlier the treatment began the better the result, and continuous dosing gave lasting protection; in 1977 at a symposium in Cambridge he argued for this as long-term adjuvant use, but most clinicians rejected it out of concern about resistance. Jordan's 1997 review places the symposium in September 1977; the Leeds obituary says that data presented in the same year encouraged the Royal Marsden, NSABP, NATO and Scottish trials. The identification of 4-hydroxytamoxifen in 1977 also appears in the same sources. All of this is the account of one person and of his student; I did not support it with independent confirmation.

1973-1977

First Britain, then the US: what is 1977 the date of?

According to the Frontiers article, ICI applied for a product licence in February 1973, the licence was granted a few months later, and Nolvadex was launched in Britain in October 1973 for anovulatory infertility and for palliative treatment of breast cancer. Jordan's 1997 review also gives 1973 for the first approval (advanced breast cancer in postmenopausal women). So 1977 is not the first approval of tamoxifen. According to the same article, ICI made a New Drug Application to the US in 1976, and adjuvant trials began in the same year through the Nolvadex Adjuvant Trial Organisation in Britain.

1977 is the year of three separate things: the first US approval, Jordan's proposal of long-term adjuvant use in Cambridge, and, according to the Frontiers article, Walpole's sudden loss of life. For the day of the US approval, a drug database listing (drugfuture.com, NDA 017970) gives "original approval date: 1977/12/30"; in the table of Jordan's 1997 review the US approval appears as 1978. I could not reconcile these two sources; I give both values in the fact-check box. According to Jordan, the scope of the approval was advanced breast cancer in postmenopausal women: a narrow beginning compared with the drug's broad use today.

The arrival of the evidence: 1998

The drug's real identity was settled by its use as an adjuvant after surgery in early-stage patients (for the concept of adjuvant therapy see Day 159). The evidence for this came not from a single trial but from pooling many. In 1998 the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) combined, in The Lancet (volume 351, pages 1451-1467), the individual data of about 37,000 women from 55 randomised trials comparing women who took tamoxifen with women who did not; this was about 87% of the worldwide evidence. At about ten years of follow-up, the reduction in recurrence was reported as 21% with one year, 29% with two years and 47% with about five years; the reduction in loss of life was 12%, 17% and 26% respectively. These are relative reductions. In absolute terms, about five years of tamoxifen raised ten-year survival from 50.5% to 61.4% (10.9 points) in node-positive patients and from 73.3% to 78.9% (5.6 points) in node-negative patients. In the roughly 8,000 women whose estrogen receptor level was low or zero the effect appeared small; for this reason the recurrence and loss-of-life analyses were limited to about 30,000 women.

Afterwards the question of duration was reopened. As Howell and Howell relay, in the ATLAS trial (12,894 patients) 10 years of treatment gave less recurrence and less loss of life than 5 years (recurrence 21.4% versus 25.0%; loss of life 12.2% versus 15.0%; the follow-up time is not stated in the review, so I relay these numbers with caution). This means that the idea of long duration that Jordan argued for in 1977 found clinical evidence about thirty years later; but the arrival of the evidence does not erase the fact that on that day the idea was not accepted.

Limits and cost

The drug does not work in every patient. In the EBCTCG the effect appeared small in women whose receptor level was low or zero; in the first study too the response lay between about one fifth and one third (22% or about 30%, depending on the source). The effect depended on the cell carrying the receptor.

There is also a cost. According to the EBCTCG, the incidence of cancer of the uterine lining (endometrium) roughly doubled with one to two years of tamoxifen and roughly quadrupled with about five years (case numbers were small, and the ratios did not differ significantly from each other); the absolute reduction in cancer of the opposite breast was about twice the absolute increase in endometrial cancer. Jordan's 1997 review gives the absolute figure as about 2 per 1,000 women per year in tamoxifen users and about 1 in breast cancer patients not taking it (an estimate from 1997). In the preventive NSABP P-1 trial (13,388 women; Fisher et al., JNCI 1998;90:1371-1388), the risk of invasive breast cancer fell by 49% (43.4 versus 22.0 per 1,000 over 69 months), but the risk ratio for endometrial cancer was 2.53, and stroke, pulmonary embolism and deep vein thrombosis were more common in the tamoxifen arm. In ATLAS, with extended treatment, the hazard ratio was 1.87 for pulmonary embolism and 1.30 for endometrial cancer, and 0.75 for ischaemic heart disease (Howell and Howell). According to Jordan's 1997 review, in 1984 tamoxifen was shown to partly stimulate a human endometrial tumour in mice; a mouse finding cannot be transferred directly to humans, but a similar signal was later seen in the clinic.

There is also the matter of the drug's metabolism: a data page reports that the drug may give less benefit in patients with CYP2D6 variants; this article does not assess the clinical importance of that. Whether the drug is suitable, for whom, for how long and in which receptor status is an individual decision; it should be made by a physician.

What remained
  • The estrogen receptor as a treatment target: the drug works in tumours that carry the receptor
  • The importance of duration: about five years instead of one or two, and longer in some patients
  • Re-evaluating a drug, independently of its original aim, in a different disease
What fell or was limited
  • The compound's use as a contraceptive (it stimulated ovulation)
  • The assumption that short courses are enough
  • The impression that a hormonal drug is "cost-free": the risks of endometrial cancer and of blood clots
Synthesis: ICI 46,474 (1962)First clinical study at the Christie Hospital (1969-1971, range)Jordan: rat model, Worcester Foundation (1972-1974, range)Launch in the UK (Oct 1973)Adjuvant trials begin (1976)Jordan: long-term adjuvant idea, Cambridge (Sep 1977)First US approval (30 Dec 1977; 1978 in one source)EBCTCG: overview of 55 trials, Lancet (1998)19601970198019902000year-only dates are drawn mid-year; bars are ranges; axis: year

The path of tamoxifen: 1962-1998. The axis runs from 1960 to 2005. Green marks are dates confirmed in several records or in the primary record; blue marks and bars rest on a single source or are ranges. Dates known only by year are drawn at mid-year.

— TÜRKİYE PERSPECTIVE —

An old and widespread drug, with a new question

I could not find a verified source on the history of tamoxifen in Türkiye or on usage numbers, and I do not construct such a link. I can write a general inference: Howell and Howell report that the drug remains on the World Health Organization's list of essential medicines; so this molecule, about half a century old, is a learned tool that can be used anywhere in the world. I do not give here the data showing how this inference plays out in Türkiye (how often hormone receptors are tested, access to the drug); for that, national registries have to be consulted.

— TIMELINE —

From 1962 to today

1962
Richardson, Walpole. Synthesis of ICI 46,474 in the fertility regulation programme (according to two secondary sources).
1966-1968
Isomers. Separation of the cis and trans isomers and the demonstration that the anti-estrogenic activity lies in the trans isomer.
1969-1971
Christie Hospital. First clinical study; Cole, Jones, Todd, Br J Cancer 25:270-275 (46 patients, 10 good responses).
1972-1974
Jordan, Worcester Foundation. DMBA rat model; effect in ER-positive tumours (account of Jordan and of the Leeds obituary).
Oct 1973
Britain. Nolvadex on the market; anovulatory infertility and palliative breast cancer treatment.
1976
US application and adjuvant trials. ICI's New Drug Application; start of the adjuvant trials in Britain.
Sep 1977
Cambridge. Jordan's proposal of long-term adjuvant use (Jordan's account).
30 Dec 1977
First US approval. According to a drug database listing; 1978 in Jordan's 1997 review.
1998
EBCTCG. 55 trials, about 37,000 women; Lancet 351:1451-1467. The same year NSABP P-1 (JNCI 90:1371-1388).
TODAY — DAY 162
A lock on the receptor. The hormone axis stretched from removing the source to blocking the receptor and proving the duration; the cost travelled along the same line.
— FACT-CHECK NOTES —

What was confirmed, what stayed uncertain

  • Day 158: Köhler and Milstein are still unwritten; the promise in the tomorrow box of Day 157 has not been kept.
  • What 1977 is the date of: Of three separate events: the first US approval, Jordan's September 1977 Cambridge proposal, and Walpole's sudden loss of life (Frontiers article). It is not the first approval of tamoxifen (Britain, 1973), not its synthesis (1962), and not its first clinical study (1969-71).
  • US approval date: A drug database listing (drugfuture.com, NDA 017970) gives "original approval date 1977/12/30"; in Jordan's 1997 review the US approval is 1978; another drug page gives the first product as "approved prior to 1 January 1982", which is a format used for old drugs and does not count as a conflict. I did not see the approval letter itself; this is why I wrote 30 December 1977 as "according to a listing". The scope of the approval (advanced breast cancer in postmenopausal women) is from Jordan 1997.
  • Year of synthesis: 1962, from the Frontiers case study and the Scientific American article (two secondary sources). Howell and Howell give no exact year; Jordan 1997 gives no synthesis date (the phrase "about 25 years ago" suggests a year like 1972, but that is an estimate and conflicts with the other sources). Richardson's patent (1965) and the overlooking of her contribution come from a single article.
  • First clinical results: According to the abstract page of Cole et al. 1971, 10 good responses in 46 patients (about 21.7%); Howell and Howell say 46 patients and 22%, consistent. The Frontiers article, however, says that all of the 40 women who remained in the study for more than 10 weeks showed measurable shrinkage; the same article states that the study approved in 1969 was for 30 patients and that 60 were admitted by the end of 1970, and does not explain these numbers. Jordan's 2020 account gives a response rate of about 30%. The four figures do not match; the sources may have described them with different definitions and in different subgroups, but I could not verify this. I could not reach the full text of the paper (PMC verification barrier); I am limited to the abstract.
  • Jordan's dates: 1972-74 at the Worcester Foundation (Leeds obituary, Jordan 1997); the 1973 visit to Jensen's laboratory and the demonstration in 1975 that estradiol binding is blocked (Cancerworld 2020); Jordan 1997, on the other hand, counts the binding blockade within the 1972-74 work. The dates of the two sources do not coincide. The identification of 4-hydroxytamoxifen in 1977 is from the obituary and from the 1977 J Endocrinol reference in Cancerworld. Jordan's accounts concern his own contribution; there is no independent confirmation.
  • Understanding of the mechanism: The finding that receptor binding is blocked dates to the mid-1970s (Jordan's account); choosing patients by receptor status becoming standard took "about ten more years" (same account). The SERM definition and CYP2D6 metabolism are from Howell and Howell and from a data page; both are secondary.
  • EBCTCG 1998: The figures are from the journal's abstract (WRAP and PubMed abstract pages, read with a summarising tool); the full text was not read. Jordan's 2020 account has the statement that "one year of treatment gave no benefit"; the EBCTCG abstract gives a 21% reduction in recurrence and a 12% reduction in loss of life at one year. The figures of the EBCTCG abstract were used in this article.
  • ATLAS: The numbers are from Howell and Howell (summarising tool); the follow-up time is not stated; the review also mentions aTTom data, but as unpublished.
  • NSABP P-1: From the abstract of Fisher et al. 1998 (13,388 women; 49% reduction in invasive cancer, 43.4 versus 22.0 per 1,000; endometrial RR 2.53); an absolute number for thromboembolic events was not in the text I read. The information that the FDA changed its approval wording from "prevention" to "risk reduction" is from the Frontiers article.
  • Türkiye: No verified data were found; there is no national number in the article.
  • Unreadable sources: PMC (the full text of Cole 1971, copies outside Frontiers) and Jordan's 2003 article in Nature Reviews Drug Discovery (the site refused access) could not be read; they were not used.
Tomorrow — Day 163

Seeing inside the body without cutting: magnetic resonance imaging, 1977

Today we watched a drug put a lock on the hormone receptor; tomorrow we move to another turning point of the same year: magnetic resonance reaching the human body. The plan lists MRI as 1977. As on Days 161 and 162, I will first separate what 1977 is the date of: an idea, a paper, a first human image or a device. I will verify separately the contributions of Damadian, Lauterbur and Mansfield with their own dates and sources; and the dispute over who received the 2003 Nobel Prize I will also tell, opening the conflict if the sources disagree.

Author's note: What I watched most closely in this article was to separate clearly what "1977" is the date of, and not to mix chemistry (the 1962 synthesis) with clinical evidence (1969-71, 1998). My second care was not to present the own accounts of Jordan, a witness of the events, as independent confirmation, and to give both values of the US approval date together.

Sources and Further Reading

  1. Cole M, Jones C, Todd I. A new anti-oestrogenic agent in late breast cancer: an early clinical appraisal of ICI46474. Br J Cancer 1971;25(2):270-275, doi:10.1038/bjc.1971.33 (the journal's abstract page; the full text could not be read).
  2. Early Breast Cancer Trialists' Collaborative Group. Tamoxifen for early breast cancer: an overview of the randomised trials. Lancet 1998;351(9114):1451-1467 (abstract pages).
  3. Fisher B, Costantino JP, Wickerham DL et al. Tamoxifen for prevention of breast cancer: report of the National Surgical Adjuvant Breast and Bowel Project P-1 study. J Natl Cancer Inst 1998;90(18):1371-1388 (abstract).
  4. Tamoxifen from failed contraceptive pill to best-selling breast cancer medicine: a case-study in pharmaceutical innovation. Front Pharmacol 2017;8:620, doi:10.3389/fphar.2017.00620 (read with a summarising tool; secondary source).
  5. Howell A, Howell SJ. Tamoxifen evolution. Br J Cancer 2023 (read with a summarising tool; secondary source; volume and page citation not read).
  6. Jordan VC. Tamoxifen treatment for breast cancer: concept to gold standard. Oncology 1997, cancernetwork.com (read with a summarising tool; page citation not read; a review in which the author tells of his own contribution).
  7. Jordan VC. How to make precision drugs that work better. Cancerworld, 2020, cancerworld.net (read with a summarising tool; the author's own account).
  8. The forgotten developer of tamoxifen, a lifesaving breast cancer therapy. Scientific American, scientificamerican.com (read with a summarising tool; secondary source) and the V. Craig Jordan obituary, University of Leeds, secretariat.leeds.ac.uk (an obituary with no independent reference list).
  9. Tamoxifen (Nolvadex), NDA 017970. drugfuture.com (a drug database listing; the approval letter was not seen) and the active-ingredient page at cdek.wustl.edu (read with a summarising tool).
  10. Mukherjee S. The Emperor of All Maladies: A Biography of Cancer. New York: Scribner; 2010 (general background; the relevant pages were not re-read for this article).

This article is Day 162 of the "365 Days of Cancer History" series and presents the development of tamoxifen as a historical reading. It is historical-popular content; it is not diagnostic or treatment advice. Consult your physician about drug use and treatment decisions. Sources have been verified; details that rest on a single secondary source, on only an abstract, on a source's own account, or on a text that could not be read are marked in the fact-check notes.

Appointment request

DROZDOGAN AI

Explore this article together

Ask about the article and read answers alongside published sources from our site.

DROZDOGAN AI

Drozdogan Assistant

From a Failed Contraceptive to Tamoxifen: The Path of ICI 46,474, 1962-1977

Ask about the article and read answers alongside published sources from our site.

Conversation and privacy information

Drozdogan Assistant answers this conversation. Authorized site administrators may review it for quality. Do not share personal information.

Replies use OpenAI infrastructure. Questions and replies are processed by that service; conversation records are stored on our site. Do not share identifying, contact or personal health information.

0 / 1000

For information only; no personal diagnosis or treatment advice.

Related articles

Explore other articles on this topic.

Stay informed

Receive new articles and scientific updates by email.

Google Analytics runs on public pages independently of this window and may use analytics cookies. Your choice here controls external services only.

Cookie policy