Trial Design

This double-blind, multicenter, randomized, controlled trial evaluated the effect of adjuvant clodronate on the incidence of bone metastases, other metastases, and survival in patients having surgery, radiotherapy, and/or drug treatment for primary operable breast cancer. The trial centers were the Royal Marsden Hospital, London and Sutton, United Kingdom, the Tom Baker Cancer Centre, Calgary, Canada, 20 hospitals in Norway, the Hôpital St Luc, Montreal, Canada, and the Central Hospital, Jyväskylä, Finland. The University of Sheffield World Health Organization Metabolic Bone Unit, Sheffield, United Kingdom, analyzed and reviewed the trial radiologic data.

Inclusion criteria required that patients should have histologically or cytologically confirmed operable primary breast cancer with no evidence of metastatic disease or significant renal, hepatic, or nonmalignant bone disease. They should be psychologically and physically suitable for 2 years of oral clodronate or placebo, give written informed consent as approved by the research ethics committee for each center, and have no previous history of malignant disease or bisphosphonate use.

Eligible, consenting patients were randomized to receive either clodronate (Bonefos; Leiras Oy, Helsinki, Finland) 1,600 mg/d or an identical placebo regimen as four capsules once or two capsules bid, not taken with food, for 2 years. The study design required that randomization and medication should start within 6 months of primary treatment (surgery or preoperative chemotherapy) to avoid the problems of multiple medication schedules in patients receiving chemotherapy, endocrine therapy, radiation, and surgery. Medication for each patient was packed, coded, labeled, and supplied to each clinical center by Leiras Oy. Unused medication was recorded and returned.

Prerandomization assessment included clinical history and physical examination, hematologic, renal, and hepatic blood tests, urinary calcium, hydroxyproline and creatinine, and skeletal radiographs. Other investigations that included computed tomography scanning, magnetic resonance imaging, and bone scintigraphy were undertaken as clinically indicated. In a subset of patients, bone mineral density was measured by dual energy x-ray absorption.¹¹ Radiologic assessments for bone metastases were repeated at 24 and 60 months and also when clinically indicated. Hematologic and biochemical tests were repeated at 3-month intervals for the first year, 6-month intervals until 5 years, and then annually.

Side effects, adverse events, and compliance were evaluated at 6 weeks, at 3-month intervals for the first year, at 6-month intervals until 5 years, and then annually. All centers were regularly monitored on site and the case report forms verified with the appropriate source documents in accordance with standards for Good Clinical Practice. Patients were randomized using random numbers tables and random permutated blocks by Leiras Oy. Patients were registered and randomized in each participating center by numerically ordered and coded drug packages labeled with the patient’s initials, study number, and hospital number.

Patients

One thousand sixty-nine patients were randomized to receive clodronate (n = 530) or placebo (n = 539) between November 28, 1989, and July 3, 1995, in the United Kingdom, Canada, and Scandinavia. A further 10 patients were randomized (clodronate, n = 8 and placebo, n = 2) but withdrew their consent before the initial assessment and start of medication. All 1,069 patients were assessable for clinical characteristics and follow-up and have been included in this analysis on an intent-to-treat basis (Table 1). Primary tumor surgical treatment consisted of wide excision or segmental mastectomy, with or without axillary dissection, or mastectomy according to the protocols of the participating centers. Some patients received preoperative chemotherapy. Radiotherapy was given according to local protocols. Tamoxifen 20 mg/d was given to 850 (80%) patients usually for 5 years, and chemotherapy was received by 683 (64%) patients (31 of whom received two types of chemotherapy) according to local protocols (Table 2). The chemotherapy was usually the standard mitoxantrone/methotrexate, cyclophosphamide/methotrexate/fluorouracil, mitoxantrone/methotrexate/mitomycin, doxorubicin/cyclophosphamide, epirubicin/cisplatin/fluorouracil infusion, or fluorouracil/epirubicin/cyclophosphamide regimens.

At relapse, appropriate local or systemic therapies were used according to the treatment protocols of the center. Clodronate/placebo medication was discontinued if bone relapse occurred, and bisphosphonate therapy was used as indicated for the skeletal complications of bone metastases.

The diagnosis of bone metastases required radiologic confirmation found either by routine skeletal x-rays at 24 and 60 months or by interval investigations, including bone scintigram and x-rays, because of symptoms or relapse at other sites. Vertebral collapse required confirmation of bone metastases by magnetic resonance imaging. Nonosseous metastases were defined as soft tissue (skin, lymph node, or other breast) or visceral (lung, pleura, liver, or CNS).

Statistics

The end points for this trial were bone metastases, other metastases, and survival. Analysis was based on intent-to-treat according to a predetermined analysis plan. It had been estimated that 1,000 patients would be required with a follow-up of 3 years to detect a 50% reduction and 5 years to detect a 25% reduction in the incidence of bone metastases.

The final analysis for the primary events of metastatic relapse and survival with a data cutoff date of June 30, 2000, has been completed for this report. (The data cutoff for toxicity and confounding factors was June 30, 1997.) Analysis and figures were made using the SAS system (version 6.12 for Windows; SAS Institute, Cary, NC). P values were considered significant if they were ≤ .05 and marginally significant if between .05 and .10. All tests were two sided and calculated confidence intervals (CIs) were 95%.

All survival analyses were performed using the Cox proportional hazards model. The results were controlled for treatment center, stage, menopausal status, and hormone receptor status. The total follow-up period (median follow-up, 2007 days) and the medication period (the time to completion of medication; median, 730 days) were considered for analysis of events. Although this was not specified in the original protocol, this analysis plan was agreed on by the investigators before the analysis. Any interaction between the treatment effect and time was included in the Cox model, and the hazards ratio (HR) was estimated and tested separately for the medication period and postmedication period. Estimates for survival functions presented in figures were calculated by the Kaplan-Meier method. The event dates for different sites of metastases were the first recorded date for relapse at these sites independent of relapse at other sites. The censor date was the date of the last clinic visit or date of death either before the end of the medication period or to the data closure date (June 30, 2000).

Frequencies of toxicity, hematologic, and biochemical abnormalities between the treatment groups were tested using Fisher’s exact test. The data have been kept under annual review by an external data monitoring committee. No stopping rules were defined in the protocol, and the study is now closed.