Factors Associated With Inadequate Response to Bisphosphonate Therapy in Patients With Osteoporosis in Real-Life Clinical Practice: a Single-Center Retrospective Analysis of 300 Patients

Introduction: Bone mineral density (BMD) measurement by dual X-ray absorptiometry (DXA) is a useful tool to monitor response to osteoporosis treatment in clinical practice. Despite bisphosphonates therapy, some patients may exhibit bone loss during treatment for different reasons. These patients may have greater fracture risk than responders and may have unrecognized secondary causes that require further attention and treatment. Objectives: To identify factors associated with inadequate response (IR) to bisphosphonates therapy in patients with osteoporosis in real-life clinical practice. Methods: This is a single-center case-control study of patients with osteoporosis treated with bisphosphonates as recommended. Baseline and follow-up (12–24 months/apart) DXA scans were performed on same device (GE-Lunar Prodigy). IR was defined as loss of BMD greater than the least significant change (LSC) on the follow-up DXA. Clinical, biochemical and densitometric parameters of patients with IR were compared to responders using t-test or Mann-Whitney test (continuous), or chi-square test (categorical variables), as appropriated. We used logistic regression to assess the association magnitude between exposures and IR. Results: From 300 patients included from 2014 to 2018 (13% males, mean age 68 ±10 years), 198(66%) were treated with oral bisphosphonates and 102(34%) with zoledronic acid (ZA). IR was observed in 44(15%) patients. All parameters were similar at baseline, except for greater prevalence of oral bisphosphonates (82% vs 63%, p=0.016) and anticonvulsants use (18% vs 7%, p=0.015) in patients with IR compared to responders. Additionally, patients with IR exhibited a lower % change in CTX following therapy in comparison to responders (median -37% [IQR -68; -16%] vs -57% [-74; -32], p=0.029, respectively), and higher serum CTX levels after treatment (median 236pg/mL [IQR 162; 344] vs 165pg/mL [119; 254], p=0.004). The likelihood of IR was greater with oral bisphosphonates then with ZA (OR 2.61, IC95% 1.16–5.81, p=0.002), and with anticonvulsants use (OR 2.94. IC95% 1.19–7.25, p=0.019). The association with IR persisted for both variables (p≤0.01), when accounted simultaneously in the same model, along with age and gender. Conclusion: Inadequate bisphosphonate response was present in 15% of individuals, which was independently associated with anticonvulsant use and particularly among those on oral bisphosphonate therapy rather than ZA. This knowledge may help to clinically identify potential modifiable factors related to unresponsiveness and to optimize treatment accordingly.