Zoledronate and osteonecrosis of the jaw in osteoporosis: incidence and risk factors. Analysis of the French Pharmacovigilance Database

INTRODUCTION

Bisphosphonate-related osteonecrosis of the jaw (BRONJ) have been characterized with the use of oral bisphosphonates in osteoporosis and zoledronate in oncology. Uncertainties remain, though, with the occurrence of BRONJ related to the use of zoledronate in osteoporosis.

OBJECTIVES

We aimed to estimate the incidence and characterize the risk factors of zoledronate-associated BRONJ in osteoporosis as compared with oral bisphosphonates in real life setting.

METHODS

Cases of BRONJ associated with zoledronate, alendronate or risedronate were extracted from the French pharmacovigilance database up to 2020. The incidence of BRONJ was estimated as their respective numbers related to cases of BRONJ in patients treated with bisphosphonates for osteoporosis, over the same period, according to the Medic'AM database.

RESULTS

Between 2011 and 2020, BRONJ incidence with zoledronate was 9.6/100,000 patient-year (PY), significantly higher than with alendronate (5.1/100,000 PY, P<0.001), and risedronate (2.0/100,000 PY, P<0.001). The number of patients treated with bisphosphonates has steadily decreased by 44.5% over 10 years. Meanwhile, the incidence of BRONJ decreased (5.8/100,000 PY in 2011; 1.5/100,000 in 2020), although a rebound was observed in 2018, including 47.6% of BRONJ following denosumab. Apart from classical risk factors, recent dental cares stood out in more than 40% of BRONJ, and zoledronate had a shorter exposure time than oral bisphosphonates.

CONCLUSIONS

In a real-life setting, our data confirm that zoledronate-associated BRONJ in osteoporosis is scarce, seeming slightly more common compared with oral bisphosphonates. We also raise awareness of dental care guidelines and greater vigilance when using bisphosphonates in patients with previous exposure to denosumab.

OP0240 ZOLEDRONATE-RELATED OSTEONECROSIS OF THE JAW IN OSTEOPOROSIS: INCIDENCE, RISK FACTORS AND COMPARISON TO ORAL BISPHOSPHONATES

Background

Mediatization of bisphosphonate(BP)-related osteonecrosis of the jaw (ONJ) has hampered the treatment of osteoporosis (OP). The risk of ONJ associated with oral BP in OP and zoledronate (ZOL) in oncology has previously been reported. However, the risk of ZOL-related ONJ in OP has only been reported in Randomized clinical trials.

Objectives

Therefore, we aimed to characterize ZOL-related ONJ in OP and compare it’s incidence with oral BP in real life setting.

Methods

All reports of adverse events (AEs) of BP (ZOL, alendronate (ALN), risedronate (RIS)) were extracted from the French pharmacovigilance database (FPD) since its origin till 2020. For ZOL, cases were separated by indication, rheumatological (ZOL-R) or oncological (ZOL-K). The risk factors analyzed for ONJ were age, sex, active smoking, alcoholism, diabetes, corticotherapy, neoplasia, previous treatment by chemo or immunotherapy, recent dental care and time of exposure to BP. The association between the occurrence of ONJ and the associated BP was assessed by calculating the reporting odds ratio (ROR) in a case/non-case study. Stratification on assumed risk factors was performed to assess their impact on the risk of ONJ.

The incidence of ONJ under BP between 2011 and 2020 was estimated by relating the number of ONJ cases under BP reported to the FPD to the estimated number of patients treated with ZOL, ALN and RIS over the same period according to health insurance reimbursement data of Medic AM database. Incidence rate calculation, confidence interval calculation, and comparison of incidence rates (Fischer exact test) were performed.

Results

For ZOL 2254 AEs were reported: 568 ONJ/1103 AES with ZOL-K and 70 ONJ/1151 AEs with ZOL-R. For ZOL-R-related ONJ, 30/70 cases had recent dental care and 48.7 months of mean time of exposure to BP. Risk factors for ONJ were smoking, history of neoplasia and chemotherapy. For ALN, 1,010 AEs were reported, including 188 ONJ. 91/188 ONJ had recent dental care and 70.9 months of mean time of exposure to BP. Risk factors for ONJ were age ≥65 years, diabetes, corticotherapy, and history of neoplasia. For RIS, 771 AEs were reported including 68 cases of ONJ. 28/68 ONJ had recent dental care and 53.6 months the average time of exposure to BP. The risk factors were age ≥ 65 years, smoking, corticotherapy, history of neoplasia and chemotherapy. ROR calculation shows that corticotherapy was associated more frequently with RIS (2.10 [1.64-2.69]) and ALN (1.33 [1.04-1.70]) as compared with ZOL-R, with no other significant difference.

The incidence of ONJ was significantly higher with ZOL-R than with RIS (p<0.001) and ALN (p<0.001). Indeed, between 2011 and 2020, 614 932 patients were treated with ZOL-R and 59 cases of ONJ reported yielding an incidence of 9.6/100,000 person-years; 2 233 536 patients were treated with RIS and 44 cases of ONJ reported yielding an incidence of 2.0/100 000 person-years; 2 432 373 patients were treated with ALN and 125 cases of ONJ reported yielding an incidence of 5.1/100 000 person-years.

Conclusion

Our data confirm in real-life settings and a large population, that BP-related ONJ is a rare event associated with OP. The risk of BP-related ONJ appears related to the potency of bone resorption inhibition, as RIS has the lowest and ZOL-R the highest risk. Smoking appears a consistent risk factor of ONJ, as is recent dental care. However, we must stress out the usual limitations of pharmacovigilance studies, their retrospective nature, common under-reporting of adverse-events, and the fact that the Medic AM database only refers to city reimbursements.

To our knowledge, this is the first study reporting the risk of ONJ under ZOL from the national pharmacovigilance database. Our study confirms the rarity, in OP, of BP-related ONJ. The risk of incidence of ONJ on ZOL-R is higher than on RIS and ALN, suggesting a risk associated with the inhibitory power of bone resorption.

Acknowledgements

Acknowledgements to the French Network of Pharmacovigilance Centers

Disclosure of Interests

Claire Amigues: None declared, Audrey Fresse: None declared, Christian Roux: None declared, Sophie Gauthier: None declared, Marie-Hélène Vieillard: None declared, Milou-Daniel Drici: None declared, Véronique Breuil Speakers bureau: AMGEN MSD UCB THERAMEX LILLY, Consultant of: AMGEN MSD UCB THERAMEX LILLY, Grant/research support from: AMGEN LILLY