Elsevier

The Lancet

Volume 365, Issue 9472, 14–20 May 2005, Pages 1687-1717
The Lancet

Fast track — Articles
Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials

https://doi.org/10.1016/S0140-6736(05)66544-0Get rights and content

Summary

Background

Quinquennial overviews (1985–2000) of the randomised trials in early breast cancer have assessed the 5 year and 10-year effects of various systemic adjuvant therapies on breast cancer recurrence and survival. Here, we report the 10-year and 15-year effects.

Methods

Collaborative meta-analyses were undertaken of 194 unconfounded randomised trials of adjuvant chemotherapy or hormonal therapy that began by 1995. Many trials involved CMF (cyclophosphamide, methotrexate, fluorouracil), anthracycline-based combinations such as FAC (fluorouracil, doxorubicin, cyclophosphamide) or FEC (fluorouracil, epirubicin, cyclophosphamide), tamoxifen, or ovarian suppression: none involved taxanes, trastuzumab, raloxifene, or modern aromatase inhibitors.

Findings

Allocation to about 6 months of anthracycline-based polychemotherapy (eg, with FAC or FEC) reduces the annual breast cancer death rate by about 38% (SE 5) for women younger than 50 years of age when diagnosed and by about 20% (SE 4) for those of age 50–69 years when diagnosed, largely irrespective of the use of tamoxifen and of oestrogen receptor (ER) status, nodal status, or other tumour characteristics. Such regimens are significantly (2p=0·0001 for recurrence, 2p<0·00001 for breast cancer mortality) more effective than CMF chemotherapy. Few women of age 70 years or older entered these chemotherapy trials.

For ER-positive disease only, allocation to about 5 years of adjuvant tamoxifen reduces the annual breast cancer death rate by 31% (SE 3), largely irrespective of the use of chemotherapy and of age (<50, 50–69, ≥70 years), progesterone receptor status, or other tumour characteristics. 5 years is significantly (2p<0·00001 for recurrence, 2p=0·01 for breast cancer mortality) more effective than just 1–2 years of tamoxifen. For ER-positive tumours, the annual breast cancer mortality rates are similar during years 0–4 and 5–14, as are the proportional reductions in them by 5 years of tamoxifen, so the cumulative reduction in mortality is more than twice as big at 15 years as at 5 years after diagnosis.

These results combine six meta-analyses: anthracycline-based versus no chemotherapy (8000 women); CMF-based versus no chemotherapy (14 000); anthracycline-based versus CMF-based chemotherapy (14 000); about 5 years of tamoxifen versus none (15 000); about 1–2 years of tamoxifen versus none (33 000); and about 5 years versus 1–2 years of tamoxifen (18 000). Finally, allocation to ovarian ablation or suppression (8000 women) also significantly reduces breast cancer mortality, but appears to do so only in the absence of other systemic treatments.

For middle-aged women with ER-positive disease (the commonest type of breast cancer), the breast cancer mortality rate throughout the next 15 years would be approximately halved by 6 months of anthracycline-based chemotherapy (with a combination such as FAC or FEC) followed by 5 years of adjuvant tamoxifen. For, if mortality reductions of 38% (age <50 years) and 20% (age 50–69 years) from such chemotherapy were followed by a further reduction of 31% from tamoxifen in the risks that remain, the final mortality reductions would be 57% and 45%, respectively (and, the trial results could well have been somewhat stronger if there had been full compliance with the allocated treatments). Overall survival would be comparably improved, since these treatments have relatively small effects on mortality from the aggregate of all other causes.

Interpretation

Some of the widely practicable adjuvant drug treatments that were being tested in the 1980s, which substantially reduced 5-year recurrence rates (but had somewhat less effect on 5-year mortality rates), also substantially reduce 15-year mortality rates. Further improvements in long-term survival could well be available from newer drugs, or better use of older drugs.

Introduction

In early breast cancer, disease is detected only in the breast or, in the case of women with node-positive disease, the breast and locoregional lymph nodes, and all detected disease can be removed surgically. However, undetected deposits of disease may remain either locally or at distant sites that, if untreated, could over the next 5, 10, 15, or more years develop into a life-threatening clinical recurrence. Breast cancer is unusual in that although the risk of distant recurrence is greatest during the first decade, it may still be substantial during the second decade after diagnosis. The main aim of systemic adjuvant treatment is to control any remaining deposits of disease, reduce the recurrence rate, and improve long-term survival.

Over the past few decades, many randomised trials have been undertaken of various treatments for early breast cancer, but the duration of follow-up differs greatly between different trials and between different patients in the same trial. Hence, meta-analyses of the effects of such treatments on long-term outcome (during and, where possible, after the first decade) in various types of patient should ideally involve central review of data on time to recurrence, death, or end of follow-up from every individual patient in every trial. Moreover, as the numbers randomised continue to increase, and follow-up on those already randomised continues to accumulate in many trials, such meta-analyses should ideally be updated every few years.

The Early Breast Cancer Trialists' Collaborative Group (EBCTCG) was, therefore, set up in 1984–851 to coordinate quinquennial worldwide meta-analyses2, 3, 4, 5, 6, 7, 8, 9, 10 of centrally collected data from every woman in all randomised trials of the treatment of early breast cancer that had, at the time of the analysis, already been running for at least 5 years. The present report is of the final results from the year 2000 EBCTCG meta-analyses of the trials of systemic adjuvant treatments (chemotherapy, endocrine therapy, or chemoendocrine therapy) that had begun in or before 1995. The corresponding meta-analyses of the trials of local treatments (surgery or radiotherapy) will be reported separately.

This is the fourth quinquennial cycle of this worldwide collaboration. It addresses many of the same questions as the previous cycles, but with more trials, more patients, better ascertainment of causes of death, and, particularly, longer follow-up. Hence, there is now substantially more evidence than before8, 9, 10 comparing the effects on 10-year survival of different adjuvant regimens (eg, anthracycline-based vs other chemotherapy regimens; longer vs shorter tamoxifen regimens; ovarian ablation or suppression in addition to chemotherapy vs chemotherapy alone).

From the older trials of adjuvant treatment versus not, where the 10-year survival differences were already definite,7, 8 the 15-year differences between treatment and control are now stable enough to be compared usefully with the 10-year differences. Thus, from the first and second decades of follow-up in various types of trial, a clearer picture is now emerging of what the lifelong risks and benefits could eventually be.

Section snippets

Methods

Trial identification and data handling procedures have been described previously.3, 4 Information was sought from all randomised trials that had started by 1995. This report describes all the trials of more than 1 month9 of systemic adjuvant therapy in which two treatment arms provided an unconfounded comparison of: (a) single-agent chemotherapy versus no adjuvant chemotherapy; (b) polychemotherapy versus no adjuvant chemotherapy; (c) anthracycline-based polychemotherapy versus standard

Results

The panel describes the format of figures and tables in this report. The study website provides supplementary information to every figure and table in the form of annex-figures and table appendices, which are also available in webappendix 1 (with explanations of what they are in webappendix 2). A list describing every trial separately can be found in webappendix 3.

Results are given first for chemotherapy, then for tamoxifen, and then for ovarian ablation or suppression. Table 1 shows the

Tamoxifen versus no tamoxifen

Figure 7 summarises the effects of 1–2 years of tamoxifen and of about 5 years of tamoxifen in the trials that compared tamoxifen versus no adjuvant tamoxifen. Because of the established relevance of the hormone receptor status of the primary tumour, the analyses are subdivided by ER status, classified as ER-poor, ER-positive, and ER-unknown. Procedures for measuring receptor status continue to evolve, so current and future measurements could well be more predictive of response. But, even

Ovarian ablation or suppression

Almost 8000 women younger than 50 years of age with ER-positive or ER-unknown disease have been randomised into trials of ovarian ablation by surgery or irradiation (4317 women, 63% ER-untested, mean follow-up 8 woman-years) or of ovarian suppression by some years of treatment with a luteinising-hormone-releasing-hormone inhibitor (LHRHI; 3408 women, 26% ER-untested, mean follow-up 5 years; figure 12).

Overall, there is a definite effect of ovarian ablation or suppression both on recurrence

15-year survival

The present analyses of systemic adjuvant treatment for early breast cancer involve a total of almost 150 000 women in 200 randomised trials, many with long-term follow-up. This collaboration, which could at first assess only short-term survival differences, has now continued for 20 years, providing increasingly reliable evidence about the 15-year risks and benefits of various treatments that were being tested in the 1980s (eg, about 6 months of treatment with anthracycline-based combinations

References (22)

  • R Peto et al.

    UK and USA breast cancer deaths down 25% in 2000 at ages 20–69 years

    Lancet

    (2000)
  • Review of mortality results in randomised trials in early breast cancer

    Lancet

    (1984)
  • Effects of adjuvant tamoxifen and of cytotoxic therapy on mortality in early breast cancer: an overview of 61 randomised trials among 28 896 women

    N Engl J Med

    (1988)
  • Treatment of early breast cancer, vol 1: worldwide evidence, 1985–1990

    (1990)
  • Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy: 133 randomised trials involving 31 000 recurrences and 24 000 deaths among 75 000 women

    Lancet

    (1992)
  • Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy: 133 randomised trials involving 31 000 recurrences and 24 000 deaths among 75 000 women

    Lancet

    (1992)
  • Effects of radiotherapy and surgery in early breast cancer: an overview of the randomised trials

    N Engl J Med

    (1995)
  • Ovarian ablation in early breast cancer: overview of the randomised trials

    Lancet

    (1996)
  • Tamoxifen for early breast cancer: an overview of the randomised trials

    Lancet

    (1998)
  • Polychemotherapy for early breast cancer: an overview of the randomised trials

    Lancet

    (1998)
  • Favourable and unfavourable effects on long-term survival of radiotherapy for early breast cancer: an overview of the randomised trials

    Lancet

    (2000)
  • Cited by (6779)

    View all citing articles on Scopus

    Collaborators listed at end of report

    View full text