- © 2010 by American Society of Clinical Oncology
Bisphosphonate Risks and Benefits: Finding a Balance
Bisphosphonates have several important therapeutic roles in patients with cancer. The primary pharmacologic action of these drugs involves inhibition of osteoclast-mediated bone resorption, and their use in oncology revolves around this mechanism. The two intravenous bisphosphonates approved by the US Food and Drug Administration for cancer indications, pamidronate and zoledronic acid, were first proven to be effective in the management of hypercalcemia of malignancy. They are most commonly used to reduce skeletal-related events in patients with bone involvement from multiple myeloma and solid tumors. They have been shown to reduce pain from skeletal metastases and to improve quality of life. In addition to effects on bone and the osteoclast, there is emerging evidence for direct and indirect effects of bisphosphonates on cancer cells, and possible synergy with anticancer agents.1–3 Randomized trials suggest that adjuvant bisphosphonates in early-stage breast cancer may prevent recurrence, and intriguing recent population-based studies hint that this class of drugs may reduce the incidence of breast cancer in postmenopausal women.4–9 Both oral and intravenous bisphosphonates are highly effective and commonly used in the prevention and treatment of osteoporosis, including the treatment of bone loss associated with cancer treatment. The potential clinical indications for bisphosphonates continue to expand, making it increasingly important to understand the risks of these drugs as well as the benefits.
Bisphosphonates are generally well tolerated; a relatively low risk of serious adverse effects has been reported in clinical trials. The adverse effect profile of bisphosphonates is still not fully clear, however, and several safety concerns have emerged in postmarketing follow-up. The intravenous bisphosphonates are associated with renal toxicity, electrolyte imbalances, and first-infusion acute phase reactions typified by fever and flu-like symptoms. GI adverse effects, common with oral formulations, are generally avoided when these agents are given intravenously. Examples of rare but serious complications of bisphosphonates that have come to light after drug approval include ocular inflammation and osteonecrosis of the jaw.10–12 There were no reported cases of osteonecrosis of the jaw in any trial of intravenous bisphosphonates before US Food and Drug Administration approval, which highlights the difficulty of identifying unexpected, low frequency but serious adverse events in clinical trials. In 2008, the US Food and Drug Administration issued a communication highlighting the possibility of severe and sometimes incapacitating bone, joint, and or muscle pain in patients receiving bisphosphonates; the communication described the pain as distinct from the more common and well-recognized acute phase reaction.13 Additionally, a small but disquieting number of cases of atypical subtrochanteric hip fractures have been reported with long-term oral bisphosphonate therapy, although a recent review of two large, observational studies of patients with osteoporosis did not support an increased risk for patents receiving bisphosphonates.14–16
Whether a causal relationship exists between bisphosphonates and atrial fibrillation has been a question since 2007, when a large, placebo-controlled trial of one time per year zoledronic acid for the treatment of postmenopausal osteoporosis reported a small but significant increase in serious atrial fibrillation (defined as life-threatening or as resulting in hospitalization or as disability).17 An electrocardiography (ECG) substudy performed in a subset of these patients on days 9 through 11 after infusion showed no difference in atrial fibrillation or other ECG abnormalities between the zoledronic acid and placebo groups, and serum calcium measured at that time point also showed no difference. There was no apparent correlation between the timing of the zoledronic acid infusion and the occurrence of atrial fibrillation; the events were evenly distributed over time, with the vast majority occurring more than 30 days after infusion. Although unexpected and unexplained, this was the first report of a possible association between atrial fibrillation and bisphosphonates. A letter to the editor published at the same time reported a similar although insignificant trend for an increase in serious atrial fibrillation in a placebo-controlled osteoporosis trial of alendronate.18 It is important to note that, in both of these studies, the rates of all atrial fibrillation events (serious and nonserious) and stroke were similar to those in patients treated with bisphosphonates versus those in patients treated with placebo; this was also the case with rates of death from cardiovascular causes. Results were published in 2008 from a trial of patients undergoing surgical repair of a hip fracture in which annual zoledronic acid infusions compared with placebo showed no difference in rates of cardiovascular events, including atrial fibrillation and stroke.19 The US Food and Drug Administration responded to these reports by requesting placebo-controlled clinical trial information from the sponsors of all bisphosphonates approved for osteoporosis and by reviewing spontaneous postmarketing reports of atrial fibrillation reported in association with oral and intravenous bisphosphonates. Data submitted by the sponsors included almost 20,000 patients treated with bisphosphonates and an additional 20,000 patients treated with a placebo. In 2008, the US Food and Drug Administration concluded that, across all studies, no clear association between bisphosphonate exposure and the rate of serious or nonserious atrial fibrillation was observed.20 In addition, increasing dose and duration of bisphosphonate therapy was not associated with an increased rate of atrial fibrillation. Subsequently, several epidemiologic studies and at least two meta-analyses have further examined patient databases to explore a possible relationship between bisphosphonates and atrial fibrillation in patients with osteoporosis, reaching conflicting conclusions.21–28 The association between atrial fibrillation and bisphosphonates in patients with osteoporosis remains unclear.
When bisphosphonates are used for bone metastasis indications, the relative dose density and total exposure to drug are substantially higher than when bisphosphonates are used for the treatment of osteoporosis. If there is a causal relationship between treatment with bisphosphonates and atrial fibrillation, it might be assumed that it would be easier to identify in patients with cancer receiving higher doses of the drugs. The randomized, phase III registration trials that used one dose of intravenous pamidronate or zoledronic acid every 3 to 4 weeks to reduce skeletal-related events in multiple myeloma and solid tumors reported no cardiac safety issues.29–35 For a proposed rare event, however, these trials might not have been large enough to pick up small differences in rates between arms. In one small trial specifically evaluating cardiac effects of bisphosphonates in patients with cancer with bone metastases receiving zoledronic acid once every 4 weeks, ECG evaluation showed no evidence of atrial fibrillation.36
In this issue of Journal of Clinical Oncology, Wilkinson et al37 use the Surveillance, Epidemiology, and End Results (SEER) -Medicare database to evaluate the incidence of atrial fibrillation, supraventricular tachycardia (SVT), and stroke in older patients with cancer receiving intravenous bisphosphonates.37 The authors report a modestly increased risk for atrial fibrillation, all SVT, and stroke in patients with cancer receiving intravenous bisphosphonates. There was no dose-response effect seen for atrial fibrillation or stroke, although a 7% risk for all SVT is reported for each increase of five bisphosphonate dose equivalents. To my knowledge, the Wilkinson et al study is the first reported large-scale, population-based study to evaluate atrial fibrillation and stroke in patients with cancer receiving intravenous bisphosphonates. How does this study help in defining a possible relationship between pamidronate and zoledronic acid and cardiac effects? An important advantage of using the SEER-Medicare database is the large number of patients with a small loss to follow-up, although there are clear limitations in the data available. Patients were identified on the basis of drug administration codes for pamidronate and zoledronic acid. Eligibility requirements included that the patients be age 65 years or older with a diagnosis of cancer who had received a bisphosphonate infusion between 1995 and 2003 and then matched to patients with cancer who had not received intravenous bisphosphonates. Patients with lung cancer were excluded because of short life span, and patients with multiple myeloma, a group that frequently receives intravenous bisphosphonates, were excluded because of a difficulty in matching. Matching included cancer type but did not include cancer stage, presumably because stage at diagnosis is available in the SEER-Medicare database, but subsequent relapse is generally unavailable. An attempt was made to match for presence or absence of bone metastases on the basis of a diagnosis code from the International Classification of Diseases (ninth revison). Only 70% of patients receiving bisphosphonates were classified as having bone metastases, and yet there would have been essentially no other indication for intravenous bisphosphonate use, except the relatively rare case of hypercalcemia of malignancy (which would be commonly associated with bone metastases). Of the patients not receiving bisphosphonates, 53% were classified as having bone metastases—an obvious imbalance that suggests that this group had significantly less metastatic disease in general, which could substantially impact differences in the overall health status and concurrent medication use between the groups. Why so many patients coded as having bone metastases were not receiving bisphosphonates is puzzling, especially during the later years of the study inclusion period when treatment with bisphosphonates had become standard of care for most cancer types. Higher rates of metastatic disease and concurrent anticancer therapy in patients receiving bisphosphonates could clearly impact rates of cardiac and cerebrovascular events. Indeed, large imbalances in patient characteristics included treatment with anthracycline/taxane chemotherapy (28% v 12.6%) and receipt of intravenous corticosteroids (29.3% v 11.2%). Several risk factors for cardiac events and stroke were found to be balanced between the two cohorts (including history of hypertension, myocardial infarction or arrhythmia, heart failure, smoking, alcohol use); however, small but significant imbalances were reported for the presence of valve disease, enlarged heart, and pulmonary and thyroid disease. Concurrent medication use in the SEER-Medicare database, which can also impact atrial fibrillation rates, is not available for oral drugs.
The Wilkinson et al study does not distinguish between pamidronate and zoledronic acid in the analysis or reporting. On the basis of a selected date range of 1995 to 2003 for bisphosphonate infusion and a cutoff of 2003 for patient follow-up, the majority of bisphosphonate use in this study was likely pamidronate. Pamidronate was approved in 1991 for hypercalcemia of malignancy, in 1995 for multiple myeloma, and in 1996 for bone metastases in breast cancer. Zoledronic acid was approved in 2001 for hypercalcemia of malignancy. In 2002, the label was expanded to include prevention of skeletal-related events in patients with cancer with bone metastases, although that use did not receive a separate billing code until January 2003. Of the patients included in this trial, 60% initiated treatment with a bisphosphonate before 2003. On the basis of US Food and Drug Administration approval dates and billing code availability, the 1995 to 2002 time period would have been restricted to pamidronate infusion, and therefore the majority of patients, and those with longer-term follow-up, would have received pamidronate. A median of 15 dose equivalents of bisphosphonates per patient per year were reported for the first year of follow-up with a range of from 5 to 32 doses. This seems too high and calls into question the ability to accurately evaluate a dose-response effect. Although both drugs are approved at one dose in each 3- to 4-week infusion interval in patients with bone metastases, intervals of one dose per 4 weeks or one dose per month is the most common clinical practice. Giordano et al,38 using the same SEER-Medicare database over a similar time range to evaluate patients with breast cancer with bone metastases, report that the mean number of doses of intravenous bisphosphonates received at 12 months was 8, and at 2 years, it was 15. Median survival in this study population was only 21 months from the first dose of bisphosphonates, which supports a late-stage cancer diagnosis in these patients.
Do bisphosphonates increase rates of cardiac arrhythmia and stroke? The jury is still out. The Wilkinson et al study and the evidence available to date support further evaluation in rigorous, prospective trials focused on the issue. As a result of the limitations of health care utilization data without other supporting clinical information and clear differences between patients treated with bisphosphonates and those not treated with bisphosphonates in this study—which could affect cardiac and cerebrovascular events—the question remains unanswered. The lack of any robust dose-response or duration of use relationship between bisphosphonates and atrial fibrillation in the majority of studies and conflicting findings in the evidence available to date leaves the question in doubt, but it begs for further evaluation.
Is there a biologic plausibility? Intravenous bisphosphonates can stimulate the release of inflammatory cytokines and cause shifts of calcium within atrial cells, which could certainly predispose patients to atrial fibrillation. In the osteoporosis study that first reported an association,17 however, atrial fibrillation occurred with no correlation in time to exposure to zoledronic acid, which seems incompatible with a simple acute phase effect or with short-term changes in electrolytes.
We continue to learn about the long-term safety of bisphosphonates. Evidence of both common and uncommon adverse effects of these drugs continues to accumulate. Bisphosphonates can offer substantial clinical benefit to patients in both the osteoporosis and cancer settings. The potential risks of bisphosphonates must be weighed against the significant reductions in risk of clinically important events, including fractures and skeletal-related events. Osteoporosis treatments will be used with increasing frequency as the population ages and the clinical indications for bisphosphonates in oncology continue to expand. The hypothetical benefits of these agents as anticancer therapies may soon extend well into the early stages of the continuum of cancer development and progression. New osteoclast-targeted therapies are on the horizon, with the receptor activator of NF-κB (RANK) ligand inhibitor denosumab now approved for treatment of osteoporosis and a bone metastases indication likely in the near future. How will the adverse effect profiles of these newer agents compare with bisphosphonates?
A causal relationship between atrial fibrillation, stroke, and bisphosphonates is unproven by existing evidence. Until further data are available, providers should take all potential adverse effects into account. All efforts should be made to prevent, recognize, and effectively manage complications of these drugs, including repleting calcium and vitamin D, treating acute phase reaction symptoms and bone pain, and monitoring renal function with the intravenous agents. Before prescribing any bisphosphonates, risks and benefits must be carefully weighed by physicians and well-informed patients.
AUTHOR'S DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: None Research Funding: Julie R. Gralow, Roche, Amgen, Novartis, Genentech, Bristol-Myers Squibb Expert Testimony: None Other Remuneration: None
