Home: Educational Supplement: Appendix

Duration of Adjuvant Hormonal Treatment

Christina Davies, MBChB, M.Sc.

Summary. Large trials of 10 years versus 5 years of adjuvant tamoxifen therapy are still in progress. Other large trials have shown that 7 to 8 years of adjuvant tamoxifen treatment are significantly better than 2 to 3 years, with much of the additional benefit emerging after year 5. Large trials have also shown that 5 years of tamoxifen is significantly better than 2 to 3 years. Whatever the hormonal treatment to be used, large-scale randomized evidence is needed as to whether the duration of hormonal therapy should in general be at least 10 years, or whether 7.5 years, or even 5 years, can suffice.

Five years of tamoxifen versus a shorter period. In trials of tamoxifen versus no tamoxifen and of one tamoxifen duration versus another duration (shorter versus longer), the overview has demonstrated that for women with potentially hormone-sensitive disease, tamoxifen is of substantial benefit (EBCTCG, unpublished data, 2000). Five years of tamoxifen appears to be better than shorter regimens at least in terms of recurrence, although this has not been shown so far to be translated into a survival benefit. A French trial of 7 to 8 years of tamoxifen versus 2 to 3 years also shows that, with regard to recurrence, longer is better (Delozier, Spielmann, Mace-Lesec’h, et al., 1997). However, there is inadequate randomized evidence about the effects of prolonging tamoxifen beyond 7 to 8 years. Five years of tamoxifen increases endometrial cancer mortality, and this adverse effect increases with a longer duration of tamoxifen (EBCTCG, 2000; EBCTCG, 1998). In addition, tamoxifen causes a slight increase in the risk of pulmonary embolus (EBCTCG, 2000; EBCTCG, 1998). But tamoxifen reduces the risk of new cancers in the opposite breast, and this effect is also increased with longer duration (EBCTCG, 2000; EBCTCG, 1998). There was no evidence of an effect on mortality from causes other than breast or endometrial cancer. In terms of the 10-year incidence of new cancers, the extra number of endometrial cancers caused by tamoxifen is smaller than the number of new cancers prevented in the opposite breast; the overall net survival benefit is 30 times greater than the hazard.

Five years of tamoxifen versus a longer period. The question of the value of 5 years of tamoxifen versus a longer period is still unanswered in terms of recurrence and survival. So far, that question has not been properly studied, either through indirect comparisons of duration between trials of tamoxifen versus no tamoxifen, or through direct comparisons in trials which compare, within the same study, 5 years of tamoxifen versus longer treatment (EBCTCG, 2000; Peto, 1996; Swain, 1996). Concerns have been expressed about resistance to tamoxifen with more prolonged treatment, but the mechanisms of resistance are poorly understood and, more importantly, this concern has not been supported by randomized evidence. The current trials are of insufficient size—even in combination—to detect the moderately sized difference that might exist. The three directly randomized comparisons of 5 years versus 10 years of adjuvant tamoxifen that started long enough ago to have produced some results have now ended (Tormey, Gray, Falkson, 1996; Stewart, Forrest, Everington, et al., 1996; Fisher, Dignam, Bryant, et al., 1996). All three (known colloquially as ECOG, Scottish Cancer Trial, and NSABP B-14) involved only small numbers of breast cancer recurrence or death after year 5. It remains quite possible, based on current evidence, to hope for additional benefit from longer treatment, and the results from the French trial support this (Delozier, Spielmann, Mace-Lesec’h, et al., 1997).

The need for large-scale randomized evidence. Studies assessing the optimal duration of adjuvant tamoxifen need to be much larger than has generally been recognized. Small-scale studies carry the substantial risk of getting the wrong answer because their results are unduly influenced by favorable or unfavorable random fluctuations. That is particularly the case if frequent interim analyses are carried out before sufficient numbers of events have been allowed to occur. As a consequence, chance “blips,” suggesting either benefit or hazard, are then inappropriately emphasized.

The need for long-term followup. Tamoxifen has a substantial carry-over benefit, with fewer recurrences seen for a few years after the end of the treatment period. This means that, in trials of tamoxifen versus no tamoxifen, the benefits of tamoxifen are enhanced compared with the no tamoxifen arm, but in trials of longer versus shorter tamoxifen there may be little additional benefit apparent during the first few years of continued treatment due to the carry-over benefit in women in the shorter tamoxifen arm of the trial. Any worthwhile benefit will emerge later on, provided that the study is large enough. This means that in trials of longer versus shorter treatment, long-term followup is required, because the expected balance of risks and benefits may change during the followup period. If the question of 5 years versus 10 years is going to be reliably answered, tens of thousands of women may need to be randomized and followed up for at least 10 years. It will probably not be until 2005, or more likely 2010, that there will be sufficient randomized evidence on 5 versus 10 years of tamoxifen to merit a review by the EBCTCG.

Current large-scale trials of 5 versus 10 years of tamoxifen. ATLAS and aTTom are both large-scale randomized trials of longer versus shorter hormonal therapy to assess reliably the effects of an extra 5 years of tamoxifen in women who have had some years of treatment and for whom there is uncertainty as to whether they should stop or continue (Rea, Poole, Gray, 1998). About 20,000 women will be randomized, usually after about 5 years of tamoxifen, to either stop tamoxifen or to continue it for 5 more years. The main analyses in ATLAS and aTTom will be of all-cause mortality, but they will also provide information on cause-specific mortality and on nonfatal but important events. If, by 2010, patients show improved long-term survival with 10 years of tamoxifen, this result will save thousands of lives annually, and will be relevant to the use of hormonal therapies in general. By November 2000, more than 7,000 women will have entered ATLAS, making it the largest-ever study of tamoxifen duration. Some 4,000 have been randomized in aTTom. Still, many thousands more need to be studied in randomized trials to obtain reliable answers about treatment of an extra 5 years with tamoxifen.

Conclusion. There is now reliable evidence that 7 to 8 years of tamoxifen is better than shorter regimens in women with hormone-sensitive breast cancer. There is insufficient evidence to draw reliable conclusions about whether 10 years of tamoxifen compared with 5 years confers additional benefits. ATLAS and aTTom, in combination with the trials of 5 years versus longer adjuvant tamoxifen that have now ended, should help to answer this question, but it is unlikely that a reliable answer will emerge before 2010.

References

Delozier T, Spielmann J, Mace-Lesec’h M, Janvier M, et al. Short-term versus lifelong adjuvant tamoxifen in early breast cancer: a randomised trial (TAM-01). [abstract]. Proc Am Soc Clin Oncol 1997. Abstract.

Early Breast Cancer Trialists’ Collaborative Group (EBCTCG). Tamoxifen for early breast cancer: an overview of the randomised trials. Lancet 1998;351:1451-67. Abstract.

Early Breast Cancer Trialists’ Collaborative Group, 2000. Unpublished data. Fisher B, Dignam J, Bryant J, DeCillis A, Wickerham DL, Wolmark N, et al. Five versus more than five years of tamoxifen therapy for breast cancer patients with negative lymph nodes and estrogen receptor-positive tumors. J Natl Cancer Inst 1996;88:1529-42. Abstract.

Peto R. Five years of tamoxifen—or more? J Natl Cancer Inst 1996;88:1791-3. Rea D, Poole C, Gray R. Adjuvant tamoxifen: how long before we know how long? BMJ 1998;316:1518-9. No Abstract Available.

Rea D, Poole C, Gray R. Adjuvant tamoxifen: how long before we know how long? BMJ 1998;316:1518-9. Abstract.

Stewart HJ, Forrest AP, Everington D, McDonald CC, Dewar JA, Hawkins RA, et al. Randomised comparison of 5 years of adjuvant tamoxifen with continuous therapy for operable breast cancer. Br J Cancer 1996;74:297-9. Abstract.

Swain SM. Tamoxifen: the long and short of it. J Natl Cancer Inst 1996;88:1510-2. No Abstract Available.

Tormey DC, Gray R, Falkson HC. Postchemotherapy adjuvant tamoxifen therapy beyond five years in patients with lymph node-positive breast cancer. [Eastern Cooperative Oncology Group]. J Natl Cancer Inst 1996;88:1828-33. Abstract.