Cookies on this website

We use cookies to ensure that we give you the best experience on our website. If you click 'Accept all cookies' we'll assume that you are happy to receive all cookies and you won't see this message again. If you click 'Reject all non-essential cookies' only necessary cookies providing core functionality such as security, network management, and accessibility will be enabled. Click 'Find out more' for information on how to change your cookie settings.

Accept all cookies Reject all non-essential cookies Find out more

9.3 The second cycle (1990)

Breast cancer is unusual in that survival without apparent recurrence for the first five years after diagnosis is by no means a guarantee that cure has been achieved. Indeed, among women diagnosed with early breast cancer, mortality rates from the disease remain substantially elevated for at least the next ten to twenty years (Quinn et al. 2001). In the first cycle of the EBCTCG the 8000or so deaths in the randomized women were approximately evenly distributed over years 1, 2, 3, 4, and 5+ of follow-up, but there was little useful information on the effects of the treatments being compared after year 5. Therefore, the second cycle of the EBCTCG included trials that began before 1985, as in the first cycle, but the follow-up was now extended for a further five years, to 1990. This enabled the effects of treatment on mortality to be evaluated not just to five, but also to ten years after the diagnosis of breast cancer (EBCTCG 1992). Data were available for a total of 40 trials of the effect of tamoxifen versus the same standard treatments but without tamoxifen. These trials included nearly 30,000 women, comprising approximately 98% of all those randomized into eligible trials, and just over 8000 of them had died. A highly significant effect of tamoxifen was once again apparent (p < 0.00001), and women who had received the drug had 17% (standard error 2%—here and below we give standard errors in parentheses after the corresponding estimates) lower mortality rate over the entire period of follow-up than those who did not. It was also clear that tamoxifen substantially reduced—by 39% (9%), giving p < 0.00001—the risk of development of ‘contralateral’ breast cancer, that is, of a completely new cancer arising in the previously unaffected breast.

In the second cycle of the EBCTCG, data were available for 11,000 women randomized in a total of 31 trials of adjuvant polychemotherapy versus no chemotherapy—79% of all women randomized into eligible trials—of whom over 3500 had died. A highly significant effect of polychemotherapy was demonstrated (p < 0.00001) and women who had received polychemotherapy had mortality rate 16% lower than those who did not.

Treatment of pre-menopausal women with ovarian ablation, which destroys ovarian function, thus altering sex hormone levels and inducing an artificial menopause, may affect the progression of breast cancer and also survival. Data were available for just over 3000 women in ten randomized trials of the effect of ovarian ablation and these demonstrated a highly significant (p = 0.0004) reduction in mortality of 25% (7%) for the 1800 or so women treated under the age of 50. Curiously, this is the one EBCTCG result that did not prove durable. For the 1326 women randomized to ovarian ablation when aged over 50, ovarian ablation had no significant effect either on overall mortality or on recurrence-free survival. This is possibly due to the fact that the majority of such women would have been post-menopausal at the time they were randomized.

In the second cycle of the EBCTCG, data were also available on 6300 women randomized in 24 trials of immunotherapy given to increase the immune response of the woman’s body to the tumour; eight trials of bacillus Calmette–Gu´erin, nine of levamisole, and seven of other agents. Neither in total, nor in any of the three subgroups, nor in any of the 24 separate trials was there any significantly favourable difference between treatment or control in either recurrence-free or overall survival—indeed, women given immunotherapy had 3% (4%) higher rates for both endpoints. Perhaps by chance, for bacillus Calmette–Gu´erin the adverse effect of 20% (8%) just reached a conventional level of statistical significance (p = 0.02). These results show that an overview can yield a strongly null result and, in particular, the results for bacillus Calmette–Gu´erin in the overview contrasted with previous claims of a benefit from it in ‘historically controlled’ comparisons (EBCTCG 1992, Hortobagyi et al. 1978).

In the second cycle of the EBCTCG, the reduction in mortality following either tamoxifen or the polychemotherapy regimens tested in the trials available at that time was highly significant both during and after years 0–4, so the cumulative differences in survival are larger at ten than at five years after initial treatment. Both direct and indirect randomized comparisons indicate long-term polychemotherapy, over 12 months, for example, to be no better than shorter, for example, of six months. However, indirect comparisons suggested that longer-term tamoxifen, such as daily for two or for five years, is significantly more effective than shorter tamoxifen regimens, such as daily for only 1 year. This observation, together with promising preliminary findings from several trials, which randomly assign women to different durations of adjuvant tamoxifen therapy—such as two versus five years or five versus ten years (Peto 1996)—led in the 1990s to the establishment of the ATLAS and aTTom trials (www.ctsu.ox.ac.uk/atlas/) that are seeking to establish reliably the effects of prolonging adjuvant tamoxifen by an extra five years among women who have already received a few years—usually five or so—of tamoxifen prior to being enrolled in the trials. These two trials address the question of the appropriate duration of adjuvant hormonal therapy in general. Therefore, their results should be relevant not only to tamoxifen itself but also to other hormonal treatments as they become available.

In women aged over 70 at diagnosis, the second cycle of the EBCTCG showed the efficacy of tamoxifen, but few women in this age range had been included in the trials of polychemotherapy. Among women aged 50–69 not only were tamoxifen and polychemotherapy both demonstrated to have a beneficial effect when given individually, but a directly randomized comparison also demonstrated polychemotherapy plus tamoxifen to be better than polychemotherapy alone; a 20% (4%) reduction in mortality rate was seen (p < 0.00001). Directly randomized comparisons also suggested that polychemotherapy plus tamoxifen may be better than tamoxifen alone, but at that time the reduction in mortality of 10% (4%) was not highly significant. At ages below 50, although both polychemotherapy and ovarian ablation were demonstrated to have a beneficial effect when given individually, the numbers available for the examination of their effects when given in combination were small and the associated standard errors correspondingly large.

In addition to examining data on the effects of systemic treatments, the second cycle of the EBCTCG also considered trials of the effects of different local therapies (EBCTCG 1995). Data were available from 36 trials comparing radiotherapy plus surgery with the same type of surgery alone, ten comparing more extensive surgery with less extensive surgery and 18 comparing more-extensive surgery with less-extensive surgery plus radiotherapy. Information was available on mortality for 28,500 women, over 97% of the women randomized to eligible trials. Some of these local therapies had substantially different effects on the rates of local recurrence of the breast cancer: in particular, the addition of radiotherapy to surgery reduced the rate of local recurrence by a factor of three, and breast-conserving surgery involved some risk of recurrence in the remaining tissue. However, at least in the 1990 overview, these trials did not demonstrate any differences in overall survival at ten years.