乌鸦传媒

The relative risk of an atypical femoral fracture with bisphosphonate use related to duration of therapy, with the highest risks being seen after 5 years of exposure, a nationwide study in Denmark found.

Compared with no bisphosphonate use, the relative risk of an atypical femoral fracture after 5 to 7 years of use in individuals older than 50 was 35.57 (95% CI 21.65-58.45) and nearly 40 after 7 years (RR 38.29, 95% CI 24.17-60.66), said Douglas Bauer, MD, of the University of California San Francisco.

However, the absolute risk was low, at 4.19/10,000 person-years after 5 to 7 years and 4.63/10,000 after 7 years, while the rate of classic hip fractures in Danish adults older than 50 was 43/10,000 person-years, Bauer reported at the American Society for Bone and Mineral Research (ASBMR) virtual meeting.

"Since first reported more than 10 years ago, it has become clear that atypical femoral fractures are a rare but serious complication of bisphosphonate therapy, and fear of these events discourages the use of osteoporosis medications," he said.

Considerable uncertainty exists regarding the absolute risks, relationship with bisphosphonate exposure, and clinical risk factors associated with these fractures, which are defined by the ASBMR as being located in the subtrochanteric region and diaphysis, an absence of a history of trauma, and a transverse or short oblique configuration.

To address these uncertainties, the researchers analyzed data from the Danish National Healthcare database, which maintains longitudinal records and radiographs of medication use, fracture events, and comorbidities for the entire population.

They identified 4,973 subtrochanteric/femoral shaft fractures among Danish adults older than 50 from 2010 to 2015, performing blinded central review of radiographs and identifying 189 as being atypical femoral fractures.

A comparison group included 2,397 individuals with subtrochanteric/femoral shaft fractures not meeting the ASBMR criteria for atypical femoral fractures, and an additional comparison group consisted of a random sample of 35,000 older Danish adults. A case-cohort design was used to examine patterns of bisphosphonate use, accounting for demographic factors, comorbidities, and other medication use.

The atypical fracture group averaged 71.1 years at the time of the event, while mean ages in the non-atypical fracture group and the healthy controls were 76.5 and 65.5 years, respectively.

Bisphosphonate exposure was reported in 58% of the atypical fracture group, 19% of the non-atypical fracture group, and 9.9% of the healthy control group. Alendronate was the specific drug used in 85% of patients.

Other medication use also differed among the groups, with corticosteroids being used by 15.9% of the atypical fracture group, 9% of the non-atypical fracture group, and 5.5% of the healthy controls, and proton pump inhibitors by 35.4%, 22.1%, and 14.4%, respectively.

In a multivariate analysis adjusted for age, sex, comorbidities, and other medication use, the relative risk of atypical femoral fracture was increased with any duration of bisphosphonate exposure:

• 0-1 years, RR 5.32 (95% CI 2.28-12.42)
• 1-3 years, RR 10.01 (95% CI 5.39-18.59)
• 3-5 years, RR 18.53 (95% CI 10.67-32.19)

However, the adjusted relative risks fell with time since stopping the treatment:

• 0-1 year, RR 0.96 (95% CI 0.61-1.51)
• 1-3 years, RR 0.26 (95% CI 0.08-0.80)
• >3 years, RR 0.19 (95% CI 0.05-0.78)

"The finding that risks decreased rapidly after withdrawal of bisphosphonates suggests that drug holidays may be useful for some patients to reduce the risk of atypical femoral fractures," Bauer said.

A total of 34% of the atypical femoral fractures occurred in individuals who had never been exposed to bisphosphonates, which motivated the researchers to consider other factors that might influence risk. They found that risks were elevated among patients with rheumatoid arthritis (RR 2.02, 95% CI 1.25-3.26) and hypertension (RR 1.53, 95% CI 1.10-2.13), and with per-year cumulative protein pump inhibitor use (RR 1.05, 95% CI 1.02-1.08).

Limitations of the study were the use of alendronate by the majority of patients, so the data could not be analyzed according to specific agent, and the inclusion of individuals from a single Nordic country.

Comment