9.2 The first cycle of the EBCTCG (1984 and 1985)

The first meeting of the nascent EBCTCG took place in London in October 1984 and preliminary analyses of data from two categories of trial were presented. Both were concerned with the evaluation of ‘systemic’ treatments—those involving drugs that would reach not just the breast and local tissues, but all parts of the body to which microscopic deposits of the cancer might already have spread. The first category included trials in which women in one treatment arm received only the standard treatment schedule at the centre involved in terms of the extent of surgery and whether or not radiotherapy was given, while women in the other treatment arm received the same standard treatment plus treatment with the anti-oestrogen agent tamoxifen, which can inhibit the growth of tumours with appreciable oestrogen receptor expression—so-called ‘ER-positive’ tumours. The second category of trials included patients in which the two trial arms differed only by the addition of some form of long-term cytotoxic chemotherapy, consisting of the administration of one or more drugs intended to kill cancer cells.

The data were analysed using standard statistical methods comprising significance tests and appropriately weighted estimates of treatment effects, based on log-rank analyses, together with life-table estimates of survival, which have been described elsewhere (EBCTCG 1990). Before these analyses, no generally agreed conclusions had emerged about the effects on mortality of either of these types of systemic therapy in early breast cancer, but it was apparent even from the preliminary results presented at the 1984 meeting that there were clearly significant reductions in short-term mortality among those receiving either tamoxifen or chemotherapy (Anonymous 1984). The final published data confirmed these preliminary findings (EBCTCG 1990, 1988). Individual data were available in 1985, when the EBCTCG was formally established, from a total of 28 trials of the effects of tamoxifen in which a total of 16,500 women had been randomized, of whom nearly 4000 had died by the end of the available follow-up. When all ages at diagnosis were combined, there was a highly significant reduction (p < 0.0001) in mortality for the women who had been allocated to tamoxifen compared with the women who had not. More detailed analysis indicated—misleadingly, as became apparent from later analyses of larger datasets—that the beneficial effect of tamoxifen was concentrated in women who were aged 50 or more at diagnosis, in whom the annual death rate from all causes combined was reduced by about one fifth. Furthermore, since not all patients complied with the treatment assigned, the true beneficial effect of tamoxifen is likely to be somewhat greater than this.

As regards the trials of cytotoxic chemotherapy, data were available from a total of 40 trials involving over 13,000 randomized women, of whom just over 4000 had died. There was a highly significant reduction in the overall death rate in trials of any chemotherapy versus no chemotherapy (p = 0.003) and also in trials of polychemotherapy—that is, chemotherapy with more than one drug for more than one month—versus single-agent chemotherapy (p = 0.001). In contrast to the tamoxifen trials, the beneficial effect of the chemotherapy appeared to be concentrated in women aged under 50 at diagnosis, in whom the annual death rate was reduced by about one quarter. The chemotherapy trials also suggested that administration of chemotherapy for 12–24 months might offer little survival advantage over administration of the same chemotherapy for about six months.