key: cord-0860396-mw2mn6bv authors: Clemons, Mark; Ong, Michael; Stober, Carol; Ernst, Scott; Booth, Christopher; Canil, Christina; Mates, Mihaela; Robinson, Andrew; Blanchette, Phillip; Joy, Anil Abraham; Hilton, John; Aseyev, Olexiy; Pond, Gregory; Jeong, Ahwon; Hutton, Brian; Mazzarello, Sasha; Vandermeer, Lisa; Kushnir, Igal; Fergusson, Dean title: A randomised trial of 4- versus 12-weekly administration of bone-targeted agents in patients with bone metastases from breast or castration-resistant prostate cancer date: 2020-10-03 journal: Eur J Cancer DOI: 10.1016/j.ejca.2020.08.019 sha: 3e5b61e598e412f07853c90451e756b10bf3cea9 doc_id: 860396 cord_uid: mw2mn6bv BACKGROUND: Optimal dosing of bone-targeted agents (BTAs), in patients with bone metastases remains an important clinical question. This trial compared 4-weekly versus 12-weekly therapy. PATIENTS AND METHODS: Patients with bone metastases from breast or castration-resistant prostate cancer (CRPC), who were going to start or already on BTAs, were randomised 1:1 to 4-weekly or 12-weekly BTA treatment for one year. Primary end point was change in health-related quality of life (HRQoL)-physical function European Organisation for Research and Treatment of Cancer (EORTC)-QLQ-C30). Secondary end points included pain (EORTC-QLQ-BM22), global health status (EORTC-QLQ-C30), symptomatic skeletal events (SSEs) rates and time to SSEs. Primary analysis was per protocol and a non-inferiority margin of 5 points was used. RESULTS: Of 263 patients (160 breast cancer, 103 CRPC), 133 (50.6%) and 130 (49.4%) were randomised to the 4- and 12-weekly groups, respectively. BTAs included denosumab (56.3%), zoledronate (24.0%) and pamidronate (19.8%). Using repeated-measures analysis, across all time points, patients in the 4-weekly arm had a mean HRQL-physical subdomain score which was 1.2 (95% confidence interval: -1.6 to 4.0) higher than the 12-weekly arm. The study met the definition of non-inferiority for our primary outcome. Secondary outcomes showed no significant difference in scores for pain, global health status, SSE rates and SSE-free survival between arms. Subgroup analyses for cancer type, prior BTA use or BTA type showed no significant difference between arms. CONCLUSION: These results in addition to those previously reported for de-escalating zoledronate and systematic reviews in both breast and prostate cancers, would support that de-escalation of commonly used BTAs is a reasonable treatment option. The use of bone-targeted agents (BTAs), such as bisphosphonates and Receptor activator of nuclear factor kappa-B (RANK) ligand inhibitors, in patients with bone metastases from breast and castration-resistant prostate cancer (CRPC) is associated with improvements in morbidity, pain, quality of life (QoL) and skeletal-related events (SREs). They have shown no improvement in either progression-free or overall survival. As supportive care drugs, questions remain around their optimal dosing [1] . De-escalation studies with pamidronate [2] , zoledronate [3e5] and denosumab [6] have shown similar efficacy outcomes across a broad range of end points in the 4-weekly and 12-weekly groups. Systematic reviews in both breast [7] and prostate cancer [8] confirmed equivalent outcomes with 4-weekly versus 12-weekly dosing. Despite these findings and evidence-based guidelines stating the efficacy of 12-weekly zoledronate [9] , surveys of both patients and oncologists confirmed that significant clinical equipoise still exists with patients receiving different BTAs at different dosing intervals [10, 11] . Although a comparison of SREs rates would be the preferred primary end point for studies comparing dosing intervals of BTAs, the use of this end point would require a study sample size in the thousands. However, as BTAs are supportive care interventions, health-related quality of life (HRQoL) is also important and validated patent reported outcome. In this pragmatic non-inferiority trial, we evaluated the effects of 4versus 12-weekly dosing of three commonly used BTAs in terms of HRQL, pain, symptomatic skeletal events (SSEs) and toxicity. We conducted this pragmatic, randomised, open-label, non-inferiority trial at five Canadian centres. Patients with bone metastases from either metastatic breast or CRPC, who were either going to start or were already on BTAs (either denosumab, pamidronate, or zoledronate) were eligible. Other eligibility requirements included Eastern Cooperative Oncology Group performance status 0e2, estimated life expectancy of >12 months and able to provide verbal consent. Exclusion criteria included history of or current evidence of osteonecrosis of the jaw (ONJ), or planned radiotherapy or surgery to the bone within 4 weeks of randomisation. There were no study-mandated changes in the type of BTA the patient received, no prior maximum duration of BTA use and patients could have had prior SSEs. All patients provided verbal consent following the integrated consent model [12] . The study was approved by the Ontario Research Ethics Board and registered on clinicaltrials.gov (NCT02721433). This report was prepared following the CONSORT extension for noninferiority trials [13] . Eligible and consented patients were randomised in a 1:1 ratio using a permuted block design with variable block sizes of 2 and 4 to either 4-weekly or 12-weekly BTAs for 1 year via a web-based randomisation system. Stratification was based on tumour type (breast vs. CRPC) and centre. After enrolment, neither investigators nor participants were masked to treatment allocation. The choice of BTA (either denosumab, pamidronate and zoledronate) was made before randomisation and was left to the patient and treating physician. Patients were instructed to take calcium and vitamin D as per guidelines. As the trial was pragmatic in nature, if a patient was receiving another systemic therapy every 3 weeks and therefore randomisation to the 4-weekly BTA arm would be more inconvenient, then the BTA could be administered every 3 or 6 weeks. End point data were collected from self-completed patient questionnaires European Organisation for Research and Treatment of Cancer (EORTC)-QLQ-C30 and EORTC-QLQ-BM22) [14, 15] . Study data were also obtained and verified through the patient's electronic medical record and emails to treating physician. Patients were assessed at their usual clinic visits. No radiological assessments beyond conventional practice were mandated. The primary outcome was the change in patient HRQoL scores (EORTC-QLQ-C30 Physical Subdomain) during the first year post-randomisation. This end point was chosen as it is patient-centred, clinically meaningful and validated in patients with bone metastases [15e17]. Information was collected at baseline and weeks 12, 24, 36 and 48. A decrease in the HRQL score is associated with worsening quality of life. Secondary end points evaluated pain (BM22 pain domain), global health status (C30), SSEs and toxicity. SSEs were defined as radiotherapy to bone, new symptomatic pathological bone fractures, spinal cord compression, tumour-related orthopaedic surgery, and hypercalcaemia. Time to development of SSEs was calculated from the date of randomisation until the first date a patient experienced an SSE and presented as a cumulative incidence rate. Any patient not experiencing an SSE was censored on the last date that they were confirmed to be SSE-free. Total number of SSEs and time to subsequent on-study SSE were used to calculate the skeletal morbidity rates (SMR, mean number of SSEs per year). Adverse events, toxicity profiles and bone metastaseserelated hospitalisation were collected from electronic health records. It was assumed there would be no difference between the 12-weekly and 4-weekly treatment arms, and a noninferiority bound of 5 points on the C30 Physical Subdomain at 48 weeks used. This 5-point inferiority margin has previously been shown to be clinically meaningful in patients receiving BTAs [15e17]. Furthermore, a 5-point change in the C30 was expected to be equivalent to a change of roughly 0.5 standard deviations, which is often used as a guideline to measure a clinically important difference. To have 80% statistical power in ensuring the lower limit of a one-sided 95% confidence interval (CI) (or equivalently a 90% twosided CI) will be above the non-inferiority limit, a minimum of 224 patients was required. To allow for an expected 10% non-compliance rate and account for stratification, enrolment of 250 total patients was planned. Additional statistical power would be realised through the use of a repeated measures analysis. Baseline characteristics are summarised using means and standard deviations (continuous measures) or proportions (categorical data). As a non-inferiority study, the primary analysis was based on a per protocol data set, defined as those patients who completed all 48 weeks of allocated treatment. Supportive analyses were conducted after the intention-to-treat principle, using data from all patients in accordance with the allocated treatment, irrespective of missing or incomplete data. The mean difference in the C30 Physical Subdomain calculated between treatment groups using repeatedmeasure analysis of variance was used as the primary analysis. Secondary analyses were performed looking at the difference between interventions for different C30 domains at each time point separately. Secondary outcomes were compared using Fisher's exact tests, Cochran-Armitage test for trends and Wilcoxon rank sum tests as appropriate. Relative risk (dichotomous) and mean differences (continuous) along with their 95% CIs were calculated and presented. Subgroup analyses were performed to explore differences based on disease, type of BTA and prior exposure; however, there was insufficient power available to perform definitive analyses within subgroups, and these are considered exploratory analyses only. All analyses were performed using SAS, version 9.2 (SAS Institute, North Carolina, USA). Except where indicated, all tests were two-sided at the a Z 0.05 level of significance. Between August 3, 2016 and June 5, 2018, 263 patients were enrolled and 133 (50.6%) randomised to 4-weekly and 130 (49.4%) to 12-weekly therapy ( Fig. 1 , CON-SORT diagram). Baseline characteristics are in shown in Table 1 , where 60.8% (160) of patients had breast cancer and 39.2% (103) of patients had CRPC. Median patient age was 68 (interquartile range: 57 to 75). Patients received denosumab (n Z 148, 56.3%), zoledronate (n Z 63, 24.0%) and pamidronate (n Z 52, 19.8%). There were more patients with CRPC on denosumab (85.4%) compared with breast cancer (37.5%) ( Table 1) . Amongst all the randomised patients, 136 (51.7%) were bone-agent naïve. Amongst patients with prostate cancer, 78% were BTA-naïve in the 12-weekly cohort, whereas 62% were BTA-naïve in the 4-weekly cohort. This difference was not statistically significant (pvalue Z 0.086). As baseline anti-cancer treatments could affect HRQoL, data are presented for the highest accruing site (Table 1) , and these were balanced. For patients at this site who were receiving trastuzumab (and the protocol allowed patients randomised to 4-weekly BTA to receive their BTA every 3 or 6 weeks for convenience), BTAs were administered 3-weekly (n Z 1), 4weekly (n Z 4), 6-weekly (n Z 3) and one patient was treated every 12-weeks due to physician error. Just more than three-fourth of patients (201) completed the study at 1 year, reasons for attrition were death (54), withdrawal or refusal (5) and lost to follow-up (3). Across all time points, the mean (sd) physical functioning score was 75.0 (1.3) for patients in the 4-weekly, and 73.8 (1.3) for the 12-weekly arm, resulting in a mean (sd) difference of À1.2 (1.4). The lower bound for the one-sided, 95% CI is À3.6, which was greater than the a priori defined boundary for declaring non-inferiority. Therefore, the 12-weekly was concluded to be noninferior to the 4-weekly arm in terms of physical functioning for the first 48 weeks after treatment initiation (Online Supplemental Table 1 ). The mean (sd) change in physical functioning from baseline was À6.4 (1.3) for 4-weekly and À4.5 (1.3) for 12-weekly patients, resulting in an estimated (95% CI) difference of À1.9 (À4.8 to 1.0) change in physical functioning ( Table 2) . No significant differences in the number of painful sites, nor painful characteristics was observed between the two treatment groups. Patients in the 4-weekly group had a higher baseline global QoL domain score (Online Supplemental Table 2 ), resulting in a higher mean score (difference of 5.3, 95% CI Z 2.4 to 8.1) across all time points. At the same time, the change in global QoL domain scores from baseline to post-baseline time points amongst 4-weekly patients was greater (À4.7, 95% CI Z À8.0 to À1.3). Subgroup analyses comparing BTA-naïve verses pretreated groups, as well as patients receiving denosumab versus zoledronate versus pamidronate showed no significant differences between the subgroups (Table 3) . Similarly, no significant difference was observed between groups in either the breast or CRPC populations (Table 3) . SSE data are presented in Table 2 with the 1-year cumulative incidence rate presented in Fig. 2 , whereas SSE-free survival was 79.1% (95% CI: 71.0 to 85.2) and 72.4% (95% CI: 63.5 to 79.5) at 1-year in the 4-and 12weekly groups, respectively (p Z 0.31; Fig. 3 ). No differences were noted between groups by the type of BTA (p-value Z 0.65 for all patients combined, or pvalue Z 0.92 and 0.32 within patients with breast and CRPC cohorts, respectively) or whether the patients were BTA-naïve or pre-treated or amongst patients with CRPC. Among patients with breast cancer, the 4-weekly group had a greater SSE-free survival of 82.2% (95% CI: 71.7 to 89.1) than the 12-weekly group with 68.0% (95% CI: 56.1 to 77.3), which approached but did not attain statistical significance (p Z 0.051). A total of 44 SSEs were experienced by patients in the 12-weekly arm during the 1-year follow-up, with a mean 0.34 SSEs per patient and maximum 5 per patient. This was similar in the 4-weekly arm, where a total of 44 SSEs were experienced, with a mean of 0.33 SSEs per patient and maximum 5 per patient. The physician reported incidence of ONJ, renal impairment, and bone metastasiserelated hospitalisations were similar between the groups (Online Supplemental Table 3 ). Biochemical hypocalcemia was evaluated in patients with CRPC treated with denosumab (n Z 51) at the Ottawa Hospital. There was no significant difference in hypocalcemia rate between the 2 groups. Eight patients (31%) treated with 4-weekly denosumab and 10 patients (40%) treated with 12weekly had hypocalcemia. There was a change in BTA (including change in frequency or agent or discontinuation of BTA altogether) in 27 (20.8%) patients in the 4-weekly compared with 49 (36.9%) patients in the 12-weekly group. Frequency of drug administration was changed in 11 patients (8.5%) in the 12-weekly group changing to 4weekly therapy, whereas 18 patients (13.5%) in the 4weekly group changed to 12-weekly dosing (Fig. 1 , CONSORT diagram). This is the first prospective randomised, open label, clinical trial involving patients with bone metastases from either breast or CRPC, comparing 4-weekly versus 12-weekly dosing of three commonly used BTAs. The change in physical function met our definition of noninferiority looking at the primary outcome of the physical function domain. No statistically significant differences in secondary outcomes were observed between patients on the two different intervention arms. These results are consistent with those previously reported for de-escalating zoledronate and add to the literature about the acceptability of de-escalating denosumab and pamidronate. This data, in addition to systematic reviews in both breast [7] and prostate cancer [8] , indicate no statistically significant evidence against the suggestion that de-escalation of BTA is a reasonable treatment option. These finding are especially important at the moment due to the COVID-19 pandemic, when both patients and physicians are trying to safely reduce the number of visits patients make to healthcare providers. The observed difference in SSE rates between the two study arms is notable. Previous studies have shown similar rates of SREs and SSEs between 4-weekly versus 12-weekly dosing of pamidronate [2, 4] , zoledronate [3, 5, 18] and denosumab [6] . Prior evidence from systematic reviews of de-escalation in both breast [7] and prostate [8] populations showed no significant difference in SSE rates. To date the largest data set is from three trials evaluating the de-escalation of zoledronate in patients with breast cancer [3, 5, 18] and showed no significant difference in SSE rates. This is reflected in recommendations from evidence-based guideline groups [19] . In the current trial, the 1-year cumulative incidence rate of SSE was 7.6% (95% CI: 4.3 to 10.9) and 16.6% (95% CI: 12.0 to 21.2), whereas the 1-year SSE-free survival was 79.1% (95% CI: 71.0 to 85.2) and 72.4% (95% CI: 63.5 to 79.5) in the 4-weekly and 12-weekly arms, respectively. Despite these observed differences, none of these were statistically significant. Although the study was not powered to show a difference in SSEs and these results are not incompatible with chance results. Given the additional evidence from other trials indicating no difference in SSE rates, it is believed that the observed numerical differences in SSE rates between 4-week and 12-week BTA is due to chance; however, further evidence is required to confirm this hypothesis. Our rate of asymptomatic hypocalcemia was similar to that recently reported [20] , but we did not detect any difference between treatment arms in this small sample size. As this was a pragmatic trial, patients did not undergo extensive toxicity assessments, and therefore the incidence of ONJ of 2 of 263 (0.76%) patients and renal dysfunction of 8 of 263 (3.0%) patients is lower than previous trials. Although toxicity rates were similar, patients in the 4-weekly group were less likely to stay on the randomised arm and often changed to a de-escalated schedule. This study has both strengths and limitations. Although there is considerable heterogeneity in the study subgroups (2 different tumour types, 3 different drugs, 2 different settings with patients pre-treated with BTA and BTA-naive), the study was designed to be broadly applicable to real-world practice by including patients with breast and CRPC. These differences in baseline characteristics could have led to differences in the results. For example, patients who have been on BTAs might be assumed to have a different risk profile for SSEs compared with BTA-naive subjects. The pragmatic nature of the trial with very few restrictive inclusion criteria also means that the study population more readily reflects that in clinical practice. The pragmatic nature did however mean that factors like burden of bone metastases (e.g. <3 bone metastases vs 3e10 metastases vs > 10 or similar) was not available for the analysis as this level of data is not routinely collected at the study sites. This reality of real-world practice is noticeable in the number of patients who died during the study period (n Z 54), as well as the number of patients who changed from 12-weekly to 4weekly dosing (8.5%) and from 4-weekly to 12-weekly (13.5%) dosing (refer Consort Fig. 1 ). Although the sample size calculation included a 10% adjustment for patient attrition (for a total sample size of 250 patients) and 263 patients were accrued, it is clear that in reality in the setting of a real-world trial in patients with metastatic cancer that this value should have been higher. Use of HRQoL was also important as ultimately with any palliative intervention, we need to ensure quality of life of our patients. The study was designed as open-label because the use of sham to enable double-blinding would have added considerable expense and likely have hindered accrual. The sample size was based on HRQoL analysis as the primary end point. Use of SSE as the primary end point would have significantly increased the sample size, the duration and the cost of the trial to such an extent to make it unlikely to ever occur in our current research funding environment. As BTAs are supportive care interventions, HRQoL is also important and validated patent reported outcome. A challenge with all studies is the choice of HRQoL score to use, especially as some are long for patients to complete and missing data due to patient 'questionnaire fatigue' becomes very challenging. Although several bone metastasisespecific tools exist, we felt as the QLC-30 has been validated in patients receiving BTAs [15, 21] (as well as radiotherapy) [22] and its relative ease of completion, that this tool would enable us to achieve the goals of the study. Another limitation could be viewed as the inclusion of hypercalcaemia of malignancy as an SSE in the present study. However, as hypercalcaemia can lead to hospitalisations and incurs costs to both the patient and the healthcare system, we felt it was important to include it as an end point. Given the lack of any clinically meaningful difference in the primary analysis, even with greater sample size, it is highly unlikely to affect the study results substantially. Finally, the study follow-up was only for 1 year; however the protocol has been updated to allow patients to remain on study for 2 years, which will be analysed at a later date. In this pragmatically designed randomised clinical trial for metastatic breast or CRPC, 12-weekly BTAs were non-inferior to 4-weekly BTAs for our HRQoL primary outcome. This trial's results are consistent with those previously reported for de-escalating zoledronate and add to the literature about the acceptability of deescalating denosumab and pamidronate. The standard incorporation of 12-weekly dosing of BTAs into routine clinical practice could substantially benefit both patients and the healthcare system. While awaiting the results of the REDUSE trial [23] which will definitively answer the question of de-escalating 4-weekly to 12-weekly denosumab in our opinion, de-escalation of all commonly used bone-targeted agents is a reasonable clinical decision. Mates Dosing strategies of bonetargeting agents A phase II, multicentre trial evaluating the efficacy of de-escalated bisphosphonate therapy in metastatic breast cancer patients at low-risk of skeletal-related events Efficacy and safety of 12-weekly versus 4-weekly zoledronic acid for prolonged treatment of patients with bone metastases from breast cancer (ZOOM): a phase 3, open-label, randomised, non-inferiority trial Randomized feasibility study of de-escalated (every 12 wk) versus standard (every 3 to 4 wk) intravenous pamidronate in women with low-risk bone metastases from breast cancer Effect of longer-interval vs standard dosing of zoledronic acid on skeletal events in patients with bone metastases: a randomized clinical trial Randomized active-controlled phase II study of denosumab efficacy and safety in patients with breast cancerrelated bone metastases De-escalation of bonemodifying agents in patients with bone metastases: the best of times and the worst of times? De-escalation of bone-targeted agents for metastatic prostate cancer Use of adjuvant bisphosphonates and other bone-modifying agents in breast cancer: a cancer care Ontario and American society of clinical Oncology clinical practice guideline De-escalated administration of bone-targeted agents in patients with breast and prostate cancer-A survey of Canadian oncologists Bone-targeted agent use for bone metastases from breast cancer and prostate cancer: a patient survey Creating a pragmatic trials program for breast cancer patients: Rethinking Clinical Trials (REaCT) Reporting of noninferiority and equivalence randomized trials: extension of the CONSORT 2010 statement Validation of the EORTC QLQ-C30 quality of life questionnaire through combined qualitative and quantitative assessment of patient-observer agreement The European organisation for research and treatment of cancer quality of life questionnaire for patients with bone metastases: the EORTC QLQ-BM22 Minimal clinically important differences in the EORTC QLQ-C30 and brief pain inventory in patients undergoing re-irradiation for painful bone metastases Predictive factors of overall quality of life in advanced cancer patients using EORTC QLQ-C30 Continued treatment effect of zoledronic acid dosing every 12 vs 4 Weeks in women with breast cancer metastatic to bone: the OPTIMIZE-2 randomized clinical trial Role of bone-modifying agents in metastatic breast cancer: an American society of clinical oncology-cancer care Ontario focused guideline update Incidence of hypocalcemia in patients with castration-resistant prostate cancer treated with denosumab: data from a non-inferiority phase III trial assessing prevention of symptomatic skeletal events (SSE) with denosumab administered every four weeks (q4w) versus every 12 weeks (q12w)dSAKK 96/12 (REDUSE) Zoledronic acid significantly improves pain scores and quality of life in breast cancer patients with bone metastases: a randomised, crossover study of community vs hospital bisphosphonate administration Determinants of quality of life in advanced cancer patients with bone metastases undergoing palliative radiation treatment. Support Care Canc Prevention of symptomatic skeletal events with denosumab administered every 4 weeks versus every 12 weeks: a noninferiority phase III trial (SAKK 96/12, REDUSE) The authors are grateful for patients and their families for their assistance with this study, as well as physicians for approaching patients. Main accrual by physician was as follows: Clemons (83), Ong (21) , Ernst All procedures performed in this study involving human participants were in accordance with the ethical standards of the institution, the Ontario Cancer Research Ethics Board and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all trial participants included in the study. Supplementary data to this article can be found online at https://doi.org/10.1016/j.ejca.2020.08.019.