Early and locally advanced breast cancer: diagnosis and management

Information considered when developing the guideline

The evidence review on the use of adjuvant bisphosphonates in people with early and locally advanced breast cancer was concerned with the effect of bisphosphonates on breast cancer specific outcomes (see evidence review G: adjuvant bisphosphonates). The critical outcomes were overall survival (OS), disease-free survival (DFS), and treatment-related morbidity (particularly osteonecrosis of the jaw, because of its severity).

Bone health, treatment-related mortality and health-related quality of life (HRQoL) were identified as important outcomes; however bone health was only included to check whether the new evidence was consistent with existing recommendations for the use of bisphosphonate treatment for bone loss (see recommendation 1.9.6 which references Guidance for the management of breast cancer treatment-induced bone loss: a consensus position statement from a UK expert group).

A literature search was undertaken for systematic reviews or meta-analysis of randomised control trials (RCTs) and RCTs published up to September 2017 that could answer the review question 'What are the indications for using adjuvant bisphosphonates in people with early and locally advanced breast cancer?' Included bisphosphonates were: alendronic acid/alendronate, sodium clodronate, pamidronate disodium, ibandronic acid/ibandronate, zoledronic acid/zoledronate, and risedronate sodium/risedronate.

Twenty articles (number of participants, n=33,051) were included in the review of evidence reported in RCTs and a systematic review). The quality of the evidence was assessed using GRADE.

Bone health evidence was of mixed quality (high to very low) and was found to be consistent with existing recommendations to use bisphosphonate treatment for bone loss. There was limited, very low-quality evidence on HRQoL which reported no effect of bisphosphonate treatment on HRQoL. No evidence was found for treatment-related mortality.

For the critical outcomes of OS and DFS the evidence was assessed as moderate to high quality. However, it was reported in the evidence review that 'it was not possible to judge the quality of evidence for a number of the subgroups as the number of people and/or number of events of interest were not reported in some papers, and so it was not possible to determine the imprecision around the estimate, and therefore the overall quality'.

The only clinical meaningful effects found for OS and DFS were for the bisphosphonates zoledronic acid and sodium clodronate:

An economic analysis was developed to estimate the cost-effectiveness of various bisphosphonates in the treatment of early and locally advanced breast cancer (all cases), and in node-positive and postmenopausal populations. The analysis was based on OS and DFS estimates for each of the treatments included in the analysis. The following were found to be cost-effective at the NICE threshold of £20,000 per QALY:

The guideline development group (GDG) noted that while the health economic results showed that bisphosphonates may be cost-effective, especially in higher risk populations, there was a high degree of uncertainty around the clinical inputs upon which the analysis was based; but that 'the analysis gives an indication that the cost-effectiveness results largely mirror the clinical effectiveness inputs. Therefore, if bisphosphonates were shown to improve overall and disease-free survival then it is likely that their use would be cost-effective.'

Based on the effectiveness and cost-effectiveness evidence the GDG made recommendation 1.9.1 to 'offer bisphosphonates (zoledronic acid or sodium clodronate) as adjuvant therapy to postmenopausal women with node‑positive invasive breast cancer'.

Recommendation 1.9.2 to 'consider bisphosphonates (zoledronic acid or sodium clodronate) as adjuvant therapy for postmenopausal women with node‑negative invasive breast cancer and a high risk of recurrence' was supported by the high quality evidence that sodium clodronate produced benefits in OS in mixed populations, but the GDG decided that 'a strong 'offer' recommendation could not be made due to the fact that for a number of other bisphosphonate comparisons a clinical benefit was not shown'.

Treatment-related morbidity evidence was of mixed quality (high to very low), but there was moderate quality evidence that IV zoledronic acid was associated with a clinically meaningful 1% increase in osteonecrosis of the jaw at 5 years follow-up compared with no treatment control for people with invasive breast cancer (n=3,359). There was no evidence available for osteonecrosis rates after treatment with other bisphosphonates, but the evidence review reported that 'it is known that the risk is greatest following IV bisphosphonates' (zoledronic acid is only given by intravenous infusion, whereas sodium clodronate is given orally); and highlighted that 'jaw osteonecrosis is a very serious adverse event, can be life changing, and there is no effective treatment, with only conservative management available'. The GDG therefore decided that it was important that the risk of jaw osteonecrosis is discussed with people considering bisphosphonate treatments, and therefore made recommendation 1.9.3 to 'discuss the benefits and risks of bisphosphonate treatment with women, particularly the risk of osteonecrosis of the jaw, atypical femoral fractures and osteonecrosis of the external auditory canal' and added a cross reference to the warning from the Medicines and Healthcare products Regulatory Agency/Commission on Human Medicines (MHRA/CHM) advice on bisphosphonates which highlights that 'risk factors for developing osteonecrosis of the jaw that should be considered are: potency of bisphosphonate (highest for zoledronate), route of administration, cumulative dose, duration and type of malignant disease, concomitant treatment, smoking, comorbid conditions, and history of dental disease'.

Because of a lack of conclusive evidence on OS and DFS for bisphosphonates other than zoledronic acid and sodium clodronate, the committee decided to make a recommendation for research on which groups of people with early and locally advanced breast cancer would benefit from the use of adjuvant bisphosphonates. This was to encourage research to determine the long-term survival benefits for a wider number of bisphosphonates; and to focus on which subgroups of people with breast cancer (such as premenopausal women, premenopausal women on ovarian suppression, those with node‑positive or node‑negative disease, and those with positive or negative oestrogen or progestogen statuses) may benefit from adjuvant bisphosphonates.

Methods

This is a phase 3 North American open-label trial comparing the efficacy of 3 years of treatment of either intravenous (IV) zoledronic acid (given monthly for 6 months, then every 3 months; standard dosage was 4 mg, with graduated reduction to 3 mg for renal impairment), oral clodronate (1,600 mg daily), or oral ibandronate (50 mg daily) in 6,097 women aged 18 years or older with stage I-III breast cancer. Patients were recruited between January 2006 and February 2010 (with random assignments to ibandronate stopping in August 2009 because of plans to market the drug at the trial dose being abandoned in North America). Participants were then followed-up for 5 years before final analysis (with all patients followed for 10 years from treatment assignment).

To be included, patients must have received, or planned to receive, systemic adjuvant therapy. Patients were excluded if they had renal failure or history of prior malignancy (except for specified in situ cancers or other cancers from which they were disease free for 5 years or over).

The study changed the target for sample size recruitment (after changes in plans to market ibandronate at 50 mg daily) to 2,000 patients receiving zoledronic acid, 2,000 receiving clodronate, and 1,400 receiving ibandronate over 4 years. The authors reported that the 'study was powered to find a statistically significant difference among the 3 arms at two-sided alpha=0.05. This assumed that the worst treatment would have a 5-year DFS of 80% and that the best treatment compared to the worst treatment would have a hazard ratio (HR) of 0.80 (justification for selecting this value was not provided).' The authors reported that they had initially planned to a have a no treatment control arm, but decided that clodronate (a nonaminobisphosphonate) would serve as the baseline to compare with the newer aminobisphosphonates (nitrogen-containing bisphosphonates) zoledronic acid and ibandronate and hypothesised that these may be more effective at preventing metastases compared with clodronate.

Results

Of the 6,097 recruited patients, 73 were found to be ineligible and 6 withdrew consent, resulting in data from 6,018 patients being included in the analysis: n=2,231 for zoledronic acid, n=2,235 for clodronate and n=1,552 for ibandronate.

Only 60.3% completed all 3 years of bisphosphonate therapy. The study authors reported that the difference between treatment compliance was 'small': for zoledronic acid n=1,410 (63.2%), for clodronate n=1,276 (57.1%) and for ibandronate n=943 (60.8%).

The baseline characteristics of patients (demographics, tumour characteristics, breast cancer stage) did not significantly differ between the treatment arms. Menopausal status was not reported, only whether women were aged less than 55 years of age (57.6% of participants) or 55 years and older (42.4%). The majority of tumours (78.5%) were hormone receptor positive (oestrogen or progesterone receptor positive) and/or human epidermal growth receptor 2 (HER2) status was negative (80.1%). Breast cancer stage was unknown for 142 participants (2.3%), 33.2% were stage 1, 43.9% stage 2 and 20.5% stage 3. Chemotherapy and/or endocrine therapy was used or planned in 79.6% and 75.2% of women respectively. DFS (starting at 3 years) was compared across treatment arms and according to whether women had completed therapy ('completers') or not ('non-completers'). A log-rank test at 3 years found no significant differences in DFS between the treatment arms for completers compared with non-completers; but did find that, when adjusted for treatment, completers were significantly less likely than non-completers to have a DFS event after 3 years. This comparison between completers and non-completers was not reported for the 5-year DFS. The study authors reported that there were no significant differences between treatment arms in 5-year DFS: in the zoledronic acid treatment group DFS was 88.3% (95% confidence interval [CI]=86.9% to 89.6%), for clodronate 87.6% (95% CI=86.1% to 88.9%), and ibandronate 87.4% (95% CI=85.6% to 88.9%). A univariate Cox model comparing treatments, gave the following (non-significant) pairwise HRs: clodronate versus zoledronic acid HR=1.09, 95% CI=0.94 to 1.26; ibandronate versus zoledronic acid HR=1.06, 95% CI=0.90 to 1.24 (comparison of clodronate versus ibandronate not reported).

For secondary outcomes, there were no significant differences between the treatment groups for 5-year OS (for zoledronic acid this was 92.6%, 95% CI=91.4% to 93.6%; for clodronate 92.4%, 95% CI=91.2% to 93.5%, and for ibandronate 92.9%, 95% CI=91.5% to 94.1%). Nor were there any significant differences between the treatment groups for bone as first site of recurrence; and there were no treatment differences related to age, tumour subtypes, or any other baseline characteristics. While the study reported that 5-year DFS was 89.9% for patients with hormone receptor positive subtypes (irrespective of HER2 status), 85.3% for patients HER2 positive and hormone receptor negative and 78.4% for patients with triple-negative subtypes, no statistical analysis was reported on comparisons of efficacy between different patients with different baseline characteristics to determine whether bisphosphonates may be more beneficial in specific groups of people with breast cancer.

For treatment-related morbidity, osteonecrosis of the jaw was significantly higher in patients receiving zoledronic acid (1.26%) compared with clodronate (0.36%) or ibandronate (0.77%). Analysis of this data is described in a separate publication, see the evidence summary section and appendix A. In relation to non-completers, only 10.0% of women receiving zoledronic acid stopped treatment because of toxicity or serious adverse events, whereas this was 17.0% for clodronate and 17.2% for ibandronate (statistical analysis of difference not reported).

Discussion

While this RCT provides evidence which indicates that there is no difference in DFS or OS in people with stage I-III breast cancer treated with adjuvant zoledronic acid, clodronate or ibandronate, and that there are low levels of treatment-related morbidity, there are several limitations to the study. This phase 3 trial was an open-label trial, so both researchers and participants were aware of which treatment was being administered, it assessed the benefits of treatment under ideal conditions (efficacy rather than effectiveness), with neither a placebo nor 'no treatment' arm to compare results with, which meant, as noted by the study authors, that 'this trial does not allow assessment of the degree of benefit bisphosphonates offer, if any, in early-stage breast cancer'; and the assessment of DFS at 3 years between completers and non-completers of the adjuvant bisphosphonate treatment should not be considered as a proxy for a control group. In addition, while the high level of 5-year DFS across all groups is an encouraging result, the sample size and study power analysis assumed that 'the worst treatment would have a 5-year DFS of 80% and that the best treatment compared to the worst treatment would have a HR of 0.80'; as DFS values were considerably higher and did not differ between treatments, the study may be underpowered.

Although the study findings suggest that ibandronate might be an alternative treatment option to zoledronic acid or clodronate, analysis by breast cancer subgroups is not provided, thereby not addressing the recommendation for research in the NICE guideline; and the study limitations are such that, based on this study alone, we would not propose an update of recommendations on adjuvant bisphosphonate treatment in the NICE guideline to consider the inclusion of ibandronate as a treatment option for people with early or locally advanced cancer. However, given that the evidence on which recommendations 1.9.1 to 1.9.3 were based on has not been reviewed since 2017, we decided that we should assess whether there is additional RCT evidence that could inform our decision, and so we decided to undertake a search for new evidence.

Search and selection strategy

We searched for new evidence related to the evidence review question 'What are the indications for using adjuvant bisphosphonates in people with early and locally advanced breast cancer?' The search strategy was the same as that used in the development of recommendations 1.9.1 to 1.9.3 on adjuvant bisphosphonate therapy in 2018 for the NICE guideline (see appendix B of evidence review G: adjuvant bisphosphonates).

We found 735 studies in a focused search for systematic reviews, meta-analyses of RCTs, and RCTs published between 26 September 2017 (the end date for the search period for the NICE guideline) and 29 April 2022 (3 May 2022 for CENTRAL database due to the download function not working on 29 April 2022).

The selection criteria were the same as those in evidence review G: adjuvant bisphosphonates (see table 1 in the evidence review for the population, intervention, comparison and outcome inclusion criteria). Phase 2 trials were excluded.

We considered 15 studies to be relevant to the review question, including the S0307 RCT that triggered this evidence review (Gralow et al. 2020; see the section on evidence that triggered the exceptional review) and another study reporting in more detail on osteonecrosis of the jaw in patients in the S0307 RCT (Kizub et al. 2021). We also included 1 Cochrane review, 6 systematic reviews/meta-analyses/network meta-analyses and 6 studies reporting on individual RCTs.

See appendix A for summary details of included studies.

Evidence summary