Off-trial evaluation of bisphosphonates in patients with metastatic breast cancer

Randomised controlled studies have consistently demonstrated that the skeletal complications of metastatic breast cancer can be reduced by the regular administration of intravenous bisphosphonates [1–7]. Given the potential impact of a pathological fracture, it is not surprising that bisphosphonate therapy has been readily accepted as standard of care for metastatic bone disease [8–10]. However, this therapy is expensive, and since bisphosphonates have no impact on survival, their cost-effectiveness is primarily justified by the avoidance of radiotherapy or surgery [11, 12].

To date, clinical trials of bisphosphonate therapy have failed to determine the optimal frequency of administration, timing of initiation or duration of use [9]. In practice, patients are treated on a three to four weekly basis for an indefinite period. Until their death, individuals with advanced cancer are therefore exposed to the risk of infusion-related adverse events, the possibility of nephrotoxicity, and the inconvenience of intravenous treatment. In an effort to promote rational prescribing of bisphosphonates, a number of medical and funding agencies have developed treatment guidelines which include rules for initiation and cessation of treatment [9, 10]. Unfortunately, recent studies have demonstrated that adherence to these guidelines is universally poor [10, 13, 14]. We propose that prescribing habits are unlikely to change without evidence of the efficacy, impact on quality of life, and cost-effectiveness of bisphosphonates in routine clinical practice.

To the knowledge of the investigators, there has been no report of the outcomes of patients receiving bisphosphonates outside the conduct of a clinical trial. Given the impact of inappropriate prescribing, it seems unreasonable to assume that the benefits observed in randomized controlled studies will be perfectly replicated in a clinical practice population. Furthermore, individuals in the key randomized studies underwent regular skeletal surveys, and the skeletal-related events (SRE) identified in these trials represented the composite of clinically relevant as well as asymptomatic radiological changes [1–6]. Skeletal surveys are not used routinely in clinical practice, and thus only clinically relevant skeletal events are likely to be identified in this setting. In this study, we report the results of an audit of intravenous bisphosphonate use in patients with metastatic breast cancer in the setting of routine clinical practice.

Over 80% of women with metastatic breast cancer have bone metastases [15, 16], yet a much smaller proportion of these will develop clinically apparent complications related to bone destruction. One large retrospective study performed before the introduction of bisphosphonates demonstrated that 29% of individuals develop a clinically significant SRE [15]. This figure is comparable to the findings of the current study (30%) yet different from the results of the randomized controlled trials of intravenous bisphosphonates [1–6]. In these latter studies, the proportion of individuals who experienced at least one skeletal complication at 12 months from bisphosphonate commencement was 43% in the treatment arm compared with 56% in the placebo arm. We propose that the discrepancy between the results of the randomized controlled studies and our current audit relates primarily to the use of regular radiographic skeletal surveys in the trial setting [4]. In practice, skeletal events are only identified on the basis of clinical suspicion and thus many bone lesions may be appropriately undetected and untreated. The corollary is of course that intensive exposure to skeletal surveys may result in the treatment of lesions which may never become clinically significant.

The baseline population risk of a skeletal event is a further factor which may explain the discrepancies between the current audit and the results of previous randomized controlled studies. Patients with metastatic disease limited to bone are four times more likely to fracture a long bone than those with concurrent liver and bony metastases, while individuals with extensive metastasis involving long bones are more likely to develop a fracture than those with solitary metastases [17]. In terms of these risk factors, our patient population was certainly at no lesser risk of fracture than the individuals enrolled in the studies of Hortobagyi et al [1, 2] and Theriault [3]. The percentage of individuals with metastatic disease limited to bone was comparable at 58.2% (current audit), 60% (Hortobagyi studies) and 70% (Theriault et al,). Furthermore, only 9.1% of the subjects in the current study had a solitary bone metastasis, whereas 43% of patients in the Hortobagyi studies had an isolated lesion [1, 2].

The endpoint used in this study, specifically the proportion of patients with more than one SRE is well accepted and provides readily assessable and comparable estimations of treatment effect. It however only captures information about the first event and does not measure the multiple skeletal events which sometimes occur in a single individual. While acknowledging that the impact of bisphosphonate therapy on recurrent events is important the best way to measure and analyse this data is controversial [18–21]. For this reason, we chose not to evaluate secondary endpoints such as total number of SRE, and event rate in the current audit.

Given the relative risk reduction in skeletal related events bisphosphonates clearly have a place in the management of women with metastatic bone disease. It is uncertain however what the absolute benefits of these reductions are, whether in terms of SRE, or economic endpoints. The rationale use of these agents is hindered by a lack of data concerning drug scheduling, duration of use, indications for initiation and cessation [9]. This audit highlights the fact that while these questions remain unanswered, clinicians will be reluctant to alter current prescribing habits. Most patients remained on bisphosphonates despite a declining performance status; the drugs were indefinitely administered on a three-four weekly basis and 67% of patients who experienced a second SRE continued bisphosphonate therapy.

The findings of this study provide an additional impetus to proceed with post-marketing evaluation of the use of bisphosphonates in clinical practice.