Background
For patients with hormone-positive early breast cancer, adjuvant endocrine therapy undoubtedly prolongs the time to recurrence [
1‐
3]. Moreover, 5 years of adjuvant endocrine treatment has been verified to be more effective than 1–2 years of treatment [
3]. However, the recurrence rate of patients receiving tamoxifen (TAM) increases from 15% at 5 years to 33% at 15 years, and cancer mortality increases from 8.3% at 5 years to 26% at 15 years [
3]. To control the increased recurrence rate and mortality rate even after receiving 5-year adjuvant endocrine treatment, extended endocrine therapy to 10 years or longer is expected to bring more clinical advantage. However, the research conclusions are controversial.
Both TAM and aromatase inhibitors (AIs) are used as extended adjuvant endocrine regimens. In MA.17, NSABP-B42, MA-17R trials, the prolonged application of AI to 10 years significantly reduced the recurrence risk after 5 years of adjuvant TAM and/or AI treatment [
4‐
6]. However, in the NSABP-B33 trial, prolonged exemestane (EMT) for 5 years did not significantly decrease the recurrence compared with placebo [
7]. In the IDEAL (S1–04) trial, prolonged letrozole (LET) for 5 years also did not prolong the disease-free survival (DFS) compared with prolonged LET for 2.5 years [
8]. An extended 10 years of adjuvant TAM compared with 5 years TAM showed mixed results. The ATLAS, aTTom, E4181/E5181 trials showed significant recurrence reduction by 10 years of TAM treatment compared with 5 years of TAM treatment; however, the Scottish trial demonstrated no benefit of extended adjuvant TAM [
9‐
12]. Additionally, the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-14 reported extended TAM by more than 5 years led to a shorter DFS. Furthermore, all extended endocrine treatment did not bring an overall survival (OS) benefit [
13].
The real benefit of extended adjuvant endocrine is unclear. The objectives of the present study were to compare the clinical outcomes of extended 10 y versus 5 y of adjuvant endocrine therapy in patients with early breast cancer.
Methods
Literature search strategy
This comprehensive analysis was performed according to the Preferred Reporting Items for Systematic Reviews [
14]. PubMed (1966–2017), Embase (1974–2017), the annual meeting abstracts of the European Society of Medical Oncology and American Society of Clinical Oncology were searched. Only prospective studies were permitted to be included in the assessment. The initial search used the following MeSH terms: “Breast cancer OR Breast cancers OR Breast carcinoma” AND “Adjuvant endocrine OR Extended endocrine treatment OR Extended adjuvant endocrine treatment OR Prolonged endocrine treatment OR Prolonged adjuvant endocrine treatment”. We also used the following MeSH terms: “Breast cancer OR Breast cancers OR Breast carcinoma” AND “Tamoxifen OR Aromatase Inhibitor OR Exemestane OR letrozole OR anastrozole” AND “Clinical Trial”. The PRISMA Checklist is described in Additional file
1: Table S1.
Inclusion and exclusion criteria of the trials
All the included studies met the following criteria: 1) All trials were prospective, properly randomized controlled and well matched for factors such as age, gender, tumor stage, performance status, clinical stage, treatment regimen, menopausal status, lymph node status, and hormone status. 2) When the same trial was summed up using different time points, only the trial with complete results and the longest follow-up time was included. 3) The primary endpoint was DFS, and the secondary endpoint was OS. 4) After the patients with early breast cancer had completed 5 years of adjuvant endocrine treatment, they were randomly allocated to continue adjuvant endocrine treatment for 5 years or receive placebo for 5 years (or only undergo observation for 5 years).
The exclusion criteria were as follows. 1) All trials concerning neo-adjuvant endocrine treatment were excluded. 2) Ongoing clinical trials without the results of DFS and OS were excluded. 3) Trials involving concomitant interventions, such as adjuvant chemotherapy or radiotherapy, were excluded.
Three reviewers (Li Li or Bingmei Chang, Xiaoyue Jiang) independently searched the articles. They screened the articles by reading the titles, abstracts or full texts. Any discrepancy was determined by a third reviewer (Shanshan Wu). The hazard ratio (HR) or risk ratio (RR) and 95% confidence interval (CI) of DFS and OS in each trial were extracted. If a trial only provided a Kaplan-Meier curve, the HR and 95% CI were estimated using the Engauge Digitizer V4.1 screenshot tool and a formula proposed by Parmar [
15,
16]. Related statistical data were extracted by an expert at the Statistics Center (Shanshan Wu). The following information were also extracted and summarized: journal name, publication year, author’s name, type of clinical trial, follow-up time, previous endocrine treatment regimens, extended endocrine treatment regimens, primary endpoints, second endpoints, lymph node status, estrogen (ER) + and/ or progesterone (PR) status, and menopausal status. The qualities of the enrolled trials were assessed according to the Cochrane Handbook 4.2.6 for Systematic Reviews of Interventions [
17].
Statistical analysis
Systematic assessment was performed using Stata version 12.0 software (Stata Corporation, College Station, Texas, USA). HR/RR and 95% CI were collected to estimate the clinical efficacy of DFS and OS. An HR > 1.0 indicated more recurrence risk or death risk in the extended endocrine treatment group. In each systematic review, Cochrane’s χ
2 test was used to evaluate the heterogeneity of the included clinical trials. When the
P-value of heterogeneity was < 0.05 or I
2 was > 50%, the random-effects model (REM) was used; otherwise, the fixed-effects model (FEM) was used. Begg’s and Egger’s tests were used to evaluate the publication bias of these trials [
18,
19].
Discussion
Five years of adjuvant endocrine therapy has been verified to significantly reduce the recurrence risk and cancer mortality in ER+, early-stage breast cancer [
20]. Whether extended endocrine therapies could further increase the clinical benefit has always been a controversial topic. Some trials have shown that extended endocrine therapies could further lower the recurrence risk. However, IDEAL trials demonstrated that (AI/TAM/TAM-AI) 5y - LET 5y did not bring significant prolongation for DFS and OS compared with (AI/TAM/TAM-AI) 5y - LET 2.5y [
8]. Additionally, in the Scottish and NSABP-B14 trials, extended TAM to 10 years was confirmed to decrease DFS compared with 5 years of treatment [
12,
13]. The optimal time of extended endocrine treatment is a controversial hot topic for oncologists. Our study comprehensively analyzed the related clinical trials, and the results showed that extended 10 years of adjuvant endocrine treatment was more efficacious than “standard” 5 years of endocrine treatment in preventing recurrence. Unfortunately, extended endocrine therapy to 10 years are not verified to improve OS, and this might mainly influenced by the efficacy of multiline therapy after relapse. It should not be overlooked that the overall follow-up time is not long, so a very long term follow-up is needed in these populations to prove OS differences. Considering the difference in the research populations and backgrounds in different trials, clinical application of extended endocrine treatment should be carefully weighed.
The debates concerning extended TAM treatment is obvious. The time of extended endocrine therapy is an important factor affecting the conclusion. The ATLAS and aTTom trials reported that extended adjuvant TAM for more than 10 years provides further protection against recurrence; however, 5–9 years of application did not exert a positive effect [
9,
10]. Saphner reported that the mean recurrence rate of 5–10 years was 4.3% per year; 5–9 years of extended endocrine treatment might not be the most advantageous with a lower recurrence risk [
21]. With the increase in the recurrence risk after 10 years, extended adjuvant TAM for more than 10 years probably shows significant clinical efficacy. In the process, identifying patients with high-risk recurrence is the key factor to guide treatment.
The difference in drugs in extended endocrine treatment might affect the conclusion. Coombes RC reported that the DFS and OS in the (TAM-EXE) 5y group were significant improved compared with those in the TAM 5y group [
22]. The 10-year breast cancer mortality was 12.1% after 5 years of AI treatment, a value that was lower than 14.2% after 5 years of TAM treatment [
23], and 5 years of AI showed an advantage over 5 years of TAM in decreasing the 10-year mortality risk. The extended 3 years of ANA after 5 years of TAM showed a significant DFS reduction of 36% (ABCSG-6a) [
24], and the extended 5 years of LET after 5 years of TAM showed a significant DFS reduction of 42% [
3]. Our subgroup analysis showed that TAM 5y - AI 5y and (AI and/or TAM) 5y - LET 5y significantly prolonged DFS compared with TAM 5y. These data indicated that extended AI treatment is a valid therapeutic option for early breast cancer. However, the conclusions of continuous 10 years of TAM were not consistent. The NSABP-B14, Scottish and E4181/E5181 trials reported that an additional benefit was not observed in the TAM 5y - TAM 5y arms [
11‐
13]. However, the ATLAS and aTTom trials reported that a prolonged 10 years of TAM significantly reduced the recurrence risk [
9,
10]. Our subgroup analysis verified that continuous 10 years of TAM did not improve DFS compared with 5 years of TAM.
The following factors are closely related to the efficacy of extended endocrine therapy: (1) Hormone receptor status. In hormone receptor-positive patients, 5 years of TAM treatment decreased the 5-y and 10-y breast cancer recurrence rates from 26.1 and 37.7% to 15.4% and 24.8%, respectively; however, the treatment could not decrease the recurrence rates in hormone receptor-negative patients [
25]. In 6 ER+ and/or PR+ trials, five trials (ATLAS, MA.17, NSABP-B33, NSABP-B42, MA-17R) showed the superiority of 10 years of endocrine treatment [
4‐
7,
9], whereas only the NSABP-B14 trial reported an opposite conclusion [
13]. Our comprehensive analysis showed that 10 years of endocrine treatment further reduced the breast cancer recurrence rate and significantly improved DFS compared with 5 years of endocrine treatment (HR = 0.80, 95% CI: 0.67–0.96) in ER+ and/or PR+ patients. (2) Menstrual status. The ATLAS stratified study showed that extended TAM treatment seemed to have a beneficial effect in preventing recurrence in postmenopausal women (RR = 0.85,
P = 0.05) but has no effect in premenopausal women (RR = 0.81,
P = 0.15) [
9]. In postmenopausal women, 5y TAM - 5y AI led to significant improvement in DFS; in the MA.17 trial, it led to OS improvement in the high-risk node-positive subset [
4]. Our meta-analysis showed that extended endocrine therapy to 10 years significantly improved DFS in postmenopausal women (HR = 0.70, 95% CI: 0.58–0.85). (3) Lymph node status. The late recurrence risk in breast cancer patients with more than three positive nodes was 2.18-fold that in patients without lymph node metastasis after the completion of 5 years of adjuvant endocrine therapy [
26]. In theory, the extended endocrine therapy should exert a clinical benefit in these patients with a lymph node-positive status. Unfortunately, the ATLAS trial showed that 10-year TAM treatment could not improve DFS compared 5 years of TAM in lymph node-positive or -negative patients [
9]. 5y TAM-5y AI treatment only improved DFS compared with 5y TAM treatment in lymph node-positive patients, and the corresponding results in lymph node-negative patients were uniform [
4,
7]. The mode of TAM 5y -AI 5y is worth recommending in lymph node-positive patients.
The clinical efficacy and side effects are two determining factors for extended endocrine treatment. Serious adverse effects will interrupt the persistence of endocrine treatment [
27,
28]. The Ideal study reported that 15.7% patients discontinued extended 2.5 years of LET treatment because of toxicities [
8]. However, the MA17R trial showed that it was safe and beneficial for women with HR-positive breast cancer to receive AI for another 5 years after initial treatment; the incidence of toxic effects was lower except that bone-related events occurred more frequently than those in the placebo group (14.0% versus 9.0%, respectively) [
6]. Moreover, the discontinuation rate of extended AI for 5 years was 6.0% due to bone fracture, a value similar to that for the placebo. Extended AI to 10 years may be recommended when improved DFS is accompanied by a lower incidence rate of toxic effects. The reported ratios of endometrial cancer in the TAM 5y - TAM 5y group were 2.1, 1.7 and 2.2 folds compared with those in the TAM 5y group in the NSABP-B-14, ATLAS and aTTOM trials, respectively [
9,
10,
13]. The TMA 5y - TAM 5y strategy may not be recommended with no improvement in DFS and higher incidence rates of endometrial cancer. Sequential treatment with different types of endocrine drugs is also used to maintain the efficacy to a maximum and decrease adverse effects to a minimum.