Arguments in editorial were not “biologically implausible”

BMJ 1999; 319 doi: https://doi.org/10.1136/bmj.319.7201.57 (Published 03 July 1999) Cite this as: BMJ 1999;319:57
  1. Klim McPherson, Professor of public health epidemiology (k.mcpherson{at}lshtm.ac.uk)
  1. Cancer and Public Health Unit, London School of Hygiene and Tropical Medicine, London WC1E 7HT

    EDITOR—In a news article Woodman reports that “controversy about the contraceptive pill and its effect on the risk of breast cancer was rekindled” in an editorial that I wrote in the Journal of Epidemiology and Community Health.1 The editorial was intended in part to reduce undue complacency about the safety of the pill. Ex cathedra statements by Professor John Guillebaud in Woodman's article—that my argument “is biologically implausible”—need to be well justified, since many people rely on him for advice on matters of contraception. Biological plausibility—an elusive concept, which keeps epidemiologists in business, of course—is never well served by unjustified assertions.

    In this case, use of oral contraception could be an important part of a breast cancer initiation process without being an initiator itself, simply by speeding up cell proliferation rates at vulnerable times, whatever they may be.2 This could happen with a wholly different potency before first term pregnancy than afterwards, and this is well supported by biological and epidemiological data. Firstly, the epidemiology of the effect of menarche and first pregnancy on the risk of breast cancer is already well understood.3 Secondly, the only large study to investigate the complex determinants of mitotic activity in resting breast tissue samples found a dramatic difference in the adjusted proliferation rate attributable to the oestrogen dose of current oral contraception, but only among nulliparous women.4 Among parous women no effect whatsoever was observed, and this difference was highly significant.

    Cell proliferation rates are widely thought to be a strong determinant of the risk of breast cancer.5 If they are then the latency problem is indeed only going to exist among users before first pregnancy and could be as long as 30 years. Such a postulate is not biologically implausible, and to say that it is is misleading, particularly without evidence, and is likely to induce more undue complacency.

    An important issue, of course, is that such “effect modifications” are usually difficult to detect because most studies lack sufficient power. When they are detected, as in this case—rather starkly—they should be heeded, not dismissed. But as Gillian Reeves says in Woodman's article, we certainly do need more data; meanwhile I repeat that we also need less illegitimate certainty.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    View Abstract

    Sign in

    Log in through your institution

    Free trial

    Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
    Sign up for a free trial