A modestly increased risk for atrial fibrillation was seen during the first year of treatment with zoledronic acid (Reclast) for osteoporosis, and a nonsignificant trend toward an increased risk when the drug was used for malignancy-related bone disease, a researcher reported.
Compared with patients given denosumab (Prolia), the incidence rate ratio (IRR) for atrial fibrillation among those receiving zoledronic acid for osteoporosis was 1.25 (95% CI 1.04-1.50), which was statistically significant, according to Sara J. Cromer, MD, of Massachusetts General Hospital in Boston.
And for those given zoledronic acid for malignant metastases to the bone, the IRR was a nonsignificant 1.20 (95% CI 1-1.44, P=0.06), she said in a presentation at the virtual American Society for Bone and Mineral Research meeting.
"Zoledronic acid is generally considered quite safe. However, in one of the earliest trials of the drug, the pivotal fracture trial, a signal emerged for increased rates of arrhythmias, driven primarily by rates of severe atrial fibrillation, defined as requiring hospitalization or resulting in permanent morbidity," she said.
In a continuation of the HORIZON trial, patients initially randomized to zoledronic acid for 3 years were then randomized to continue on the drug for another 3 years or to transition to placebo. While the results of this extension did not reach statistical significance, there were numerically higher rates of atrial fibrillation and severe atrial fibrillation in those treated longer with zoledronic acid.
"When the data initially surfaced, there were concerns that this could be a class effect," Cromer said, adding that "secondary analyses were done in the examining rates of atrial fibrillation in those treated with alendronate versus placebo. Although the study was not powered for this rare event, there was a nonsignificant trend seen towards increased atrial fibrillation and especially severe atrial fibrillation in the alendronate group."
The result was of concern to the medical community, and multiple subsequent observational studies were done, which had equivocal results.
"We sought to clarify the conflicting results by performing a propensity score-matched study with a large sample and real-world data using an active comparator design to decrease the risk of confounding by indication," Cromer explained.
The data were obtained from the Clinformatics Data Mart (Optum) and Medicare claims, with propensity score matching for demographics, concurrent diagnoses, medications, and utilization measures.
For the osteoporosis cohort, there were 16,235 matched pairs. During 13,496 person-years of follow-up, patients receiving zoledronic acid experienced 251 atrial fibrillation events, while for those given denosumab there were 201 events during 13,483 person-years of follow-up.
While the IRR was statistically significant, the absolute risk remained low, at approximately four events per 1,000 person-years and a number needed to harm of 325.
A sensitivity analysis that used two alternate definitions of atrial fibrillation had results similar to what was seen in the primary analysis. For inpatient atrial fibrillation, the IRR was 1.38 (95% CI 1.03-1.85), while for atrial fibrillation plus medication prescribed, it was 1.30 (95% CI 1.02-1.67).
On secondary endpoints, there was no difference in the occurrence of stroke or transient ischemic attack, or in nonvertebral osteoporotic fracture risk in the first year of treatment with zoledronic acid.
The malignancy cohort included 7,732 matched pairs. During 5,419 person-years of treatment with zoledronic acid, there were 254 atrial fibrillation events, while during 5,688 person-years of treatment with denosumab there were 222 events, which did not differ significantly.
As with the osteoporosis cohort, results on the sensitivity analysis for the two different definitions of atrial fibrillation were similar to the primary analysis, with IRRs of 1.08 (95% CI 0.79-1.46) for inpatient atrial fibrillation and 1.14 (95% CI 0.88-1.47) for atrial fibrillation plus medication prescription.
On the secondary endpoints, there were no differences in rates of stroke or transient ischemic attack among the malignancy cohort, but there was a difference in the rate of nonvertebral osteoporotic fractures, with greater risks seen for zoledronic acid (IRR 1.32, 95% CI 1.01-1.74).
Strengths of the study included its active comparator design, large sample size, and use of real-world data. A limitation was its observational design, allowing the possibility for residual confounding.
"Clinicians may consider atrial fibrillation risk when choosing whether to initiate zoledronic acid or denosumab. However, overall risk remains low and we do not think clinicians should avoid offering zoledronic acid to patients when indicated," she concluded.
Disclosures
Cromer disclosed a relevant relationship with Johnson & Johnson.
Primary Source
American Society for Bone and Mineral Research
Cromer S, et al "Risk of incident atrial fibrillation with zoledronic acid versus denosumab: a propensity-score matched cohort study" ASBMR 2020; Abstract 1066.