Introduction

Long-term androgen deprivation therapy (ADT) is the mainstay medical treatment for men with advanced and metastatic prostate cancer. Loss of bone mineral density is a well-recognized side effect occurring most rapidly within the first 12 months (between 5% and 10%),^1,2 and is ongoing and continuous throughout the duration of ADT.³ Net bone loss compounded by higher rates of osteoporosis at diagnosis⁴ leads to increased fracture risk.⁵ Results from the STAMPEDE trial demonstrated no overall survival benefit with the addition of zoledronic acid (ZA) to ADT.⁶ However, the long-term effects of ZA on bone health and fracture risk were not formally collected as part of the trial. Routinely collected health care records accessed through Hospital Episode Statistics (HES) for patients in England provide reliable, high-quality data⁷ permitting analysis of outcome data beyond standard trial follow-up to evaluate fracture incidence and the effect of ZA on fracture risk.

Method

HES data up to January 2018 were obtained for STAMPEDE trial participants from England randomized 2:1:1:1 between ADT (Arm A), ADT+ZA (Arm B), ADT+docetaxel (Arm C), and ADT+docetaxel+ZA (Arm E). ZA (4 mg IV infusion) was administered every 3 weeks for 6 doses, then every 4 weeks for 2 years (Figure 1). Clinical fracture events were identified from the Admitted Patient Care data set using a prespecified coding framework of International Classification of Diseases diagnosis and Classifications of Interventions and Procedures procedure codes. Multivariate Cox regression models including age (≥70 years, <70 years), nodal (N) stage, WHO performance status, Gleason score (≤7, 8-10), and NSAID or aspirin use were used to determine the effect of ZA on fracture risk.

Results

Figure 1. Trial design schema. ADT indicates androgen deprivation therapy; Doc, docetaxel; HSPC, hormone-sensitive prostate cancer; PSA, prostate-specific antigen; ZA, zoledronic acid.

Figure 2. CONSORT diagram. HES indicates Hospital Episode Statistics.

Figure 3. Cumulative fracture incidence of fracture by metastatic status.

Figure 4. Cumulative incidence of fracture by allocation to zoledronic acid (ZA; M1 only). ADT indicates androgen deprivation therapy; Doc, docetaxel.

Connected data sets were available for 2,145 (74%) of 2,962 trial participants recruited between October 2005 and March 2013 (Figure 2). Patients with nonmetastatic (M0, n=796) and metastatic (M1, n=1,349) disease were included in the analysis. Baseline characteristics were comparable to the overall trial cohort and balanced across trial arms. The 5-year fracture incidence was significantly lower in M0 participants compared to M1 participants (2.1% in M0 vs 9.6% in M1; P < .0001; Figure 3). In M1 participants the 5-year fracture incidence was significantly lower among those allocated to ADT±docetaxel+ZA (4.6%) compared to those allocated to ADT±docetaxel (12.9%; P < .0005; Figure 4). With a median follow-up of 5.3 years, ZA significantly reduced fracture risk in M1 participants (HR 0.36, 95% CI 0.22-0.57; P < .0005) but not in M0 participants (HR 0.67, 95% CI 0.32-1.39; P = .28). The low fracture event rate in M0 participants suggests longer follow-up is required.

Discussion

These data from a large, phase 3, randomized trial highlight a high 5-year fracture incidence in men with metastatic hormone-sensitive prostate cancer (mHSPC) treated with ADT±docetaxel (12.9%), with a 64% relative risk reduction with the addition of ZA. This is consistent with findings of a meta-analysis of 15 randomized trials on bisphosphonate therapies in men with prostate cancer which reported a 20% and 61% relative risk reduction in fracture incidence and osteoporosis, respectively.⁸ However, the CALBG90202 study, which tested the impact of ZA in mHSPC patients, did not demonstrate reduction in skeletal related events, which included pathological fracture.⁹ Current guidance therefore does not routinely recommend bone-protective agents in men with mHSPC, except for those at high risk of fracture.¹⁰ The role of fracture risk prediction tools including FRAX and bone mineral density assessment in identifying those at highest risk may be limited among men with prostate cancer starting lifelong ADT,¹¹ particularly for those with bone metastases. Exercise, calcium, and vitamin D supplementation are also unlikely to offer sufficient protection for skeletal integrity when used in isolation.¹²

It is important to note that the ZA dosing schedule in STAMPEDE was higher than is recommended for patients with osteoporosis: high doses over longer periods are associated with an increase in the risk of osteonecrosis of the jaw. ZA dosing of 5 mg every 6-12 months substantially reduces the risk of osteonecrosis of the jaw to between 0.001% and 0.01%.¹³ Alternative bone protective agents including oral bisphosphonates and the RANKL inhibitor denosumab at a bone protective dose have demonstrated similar efficacy: both of these agents are easily administered.¹⁴

Access to HES records provides reliable clinical outcome data which complement standard trial follow-up and are being used in ongoing audits¹⁵ and clinical trials.¹⁶ The true fracture incidence is likely to be even higher than reported in this study due to improvements in coding accuracy.¹⁷ In addition, STAMPEDE trial participants are younger (median age 65 years) and fitter than the average patient presenting with mHSPC starting lifelong ADT.¹⁸

Conclusions

We evaluated the long-term fracture risk in men with prostate cancer recruited to STAMPEDE through HES data linkage. Fracture rates overall were high in M1 patients and were significantly reduced in participants randomized to receive ZA. These data support the use of bone protective agents in men with mHSPC to reduce clinically significant fractures.