Immune checkpoint inhibitors (ICIs) are first-line treatments for melanoma in both the metastatic and adjuvant settings.1 The success of ICIs in the treatment of melanoma is unparallelled yet tempered by the potential development of off-target effects, termed immune-related adverse events (irAEs).2 IrAEs affecting bone and its metabolism, including osteoporosis and fracture, remain infrequently described and poorly understood. Cancer treatment-induced bone loss due to androgen deprivation therapy, aromatase inhibitors and ovarian suppression therapy is well recognised, with established guidelines to assist clinicians in making informed decisions regarding prevention, monitoring and treatment.3 Given the increasing survival associated with ICIs and the potential morbidity and mortality associated with major osteoporotic fractures (MOFs), there is a growing urgency to quantify the potential risk ICIs may pose to bone health and mitigate it if possible.
The last decade has heralded a renewed interest in the role of the immune system in bone metabolism and the emergence of a new field ‘immunoporosis’, highlighting the intricate role of innate and adaptive immune cells in osteoporotic conditions.4 T-cell activation, inherently increased with ICI use, can lead to the production of various pro-osteoclastogenic mediators including tumour necrosis factor-alpha and receptor activator of nuclear factor kappa-B ligand, with the potential to tip the balance of bone remodelling towards resorption and ultimately fracture.5 Thus, while the pathophysiological mechanism by which ICI could induce fracture is well seeded, studies evaluating ICIs effect on bone health remain scarce.
Until recently, reports of fracture in the setting of ICI use were limited to several small case series and one pharmacovigilance study.6 7 A recent population-level study in Canada using administrative data was the first attempt to quantify fracture rates in the setting of ICIs and showed more than a twofold increase in fracture rate in the year following ICI initiation as compared with the year prior (incidence rate ratio (RR) 2.43, 95% CI 1.34 to 4.27).8 Limitations of this study included the relatively small sample size, a heterogeneous group of tumour types and an inability to distinguish osteoporotic from pathological fractures.
The objective of the current study by Ye et al was to compare incidence rates of MOF in melanoma patients treated with ICI, both before and after ICI initiation. To achieve this, the authors used administrative data obtained from a large commercial healthcare claims dataset in the USA and a before-and-after retrospective cohort design, using both time-to-event and incidence rate analysis. As compared with the previous population-based study, it had twice the sample size (3137) and included only melanoma patients, to create a more homogeneous population of patients with a lower propensity to metastasise to bone, and thus aid differentiation of pathological from MOF. Rates of bone metastasis in the cohort were low at less than 5% with no difference between those who fractured and those who did not. The HR of suffering an incident MOF in the first year post-ICI as compared with the pre-ICI year was 1.82 (95% CI 1.24 to 2.66). Like the previous study, there was a higher fracture rate in the year following ICI initiation (24.2 per 1000 patient-years) as compared with the year prior (12.8 per 1000 patient-years) yielding just shy of a twofold increased incidence of MOF in the year following ICI initiation (RR 1.90, 95% CI 1.25 to 2.91). In multivariable regression analysis, risk factors for MOF included history of prior fracture, older age and female sex, which are consistent with known risk factors for MOF outside the ICI setting. A new finding was the potential increased association of MOF with combination ICI therapy which has not been reported to date. Interestingly, there was no association between glucocorticoid use and MOF, however, glucocorticoid use in this study was surprisingly low and potentially under-reported given only outpatient prescriptions were accounted for.
A limitation of this study, given its use of administrative data, is the risk of a coding error, particularly in differentiating pathological fractures from MOF. Similarly, accounting for concomitant irAE, which lack unique International Classification of Diseases (ICD) codes, and could be an important predictor of risk, particularly given the association with glucocorticoid use, is difficult to ascertain in a population-based study such as this. As such, fully elucidating the association between fracture risk, glucocorticoid use and metastatic disease in this unique patient population will require a different study design.
This study is an important addition to the growing body of literature on fracture risk in patients treated with ICI with important implications for practice. Increased vigilance, at a minimum, in screening and preventative treatments based on current guidelines and traditional risk factors is needed, until further research can support guideline formation for this patient population. Future preclinical and prospective studies which include bone mineral density measurement and bone turnover markers will pave the way for a deeper understanding of the effect of ICI on bone and its metabolism.
