Bone Abstracts

The origin of the recent levelling-off in hip-fracture incidence in several settings is unknown.

Methods: Using Danish national inpatient data for individuals aged ≥50 years during 1980–2010, we examined annual number and incidence of hip fractures and age, period, and cohort effects by log-likelihood estimates in Poisson regression models. Age adjustment was done by direct standardization, time-trend analysis by linear regression, and identification of breakpoints in linear trends by join-point analyses.

Results: There were 240 121 hip fractures, 74% in women. Before 1993, the annual age-standardized hip fracture incidence increased (2.8% per year (95% CI 2.3 to 3.3%)), and the annual number of hip fractures increased (4.4% per year (3.8 to 5.0%)). After 1993, the age-standardized hip fracture incidence decreased (−1.2% per year (−1.5 to −0.9%)) and the number of hip fractures was stable (−0.3% per year (−0.7 to −0.0%)).

The combined period+cohort effects were more marked in men, with an incidence rate ratio (IRR) ranging from 0.4 to 1.2 depending on 6-year birth cohort and 0.7 to 1.1 depending on 3-year period. In women the corresponding results were IRR 0.8 to 1.4 and 0.9 to 1.2.

Analyses of specific cohort effects (estimated by deviations from underlying linear trends in cohort) in the full APC-model showed higher risk in men born 1900–1926 and in women born 1897–1938. The corresponding specific period analyses revealed increasing risk for men with later period while a higher risk was evident in the middle of the examination period for women.

Conclusion: The annual age-standardized hip fracture incidence has decreased in Denmark since 1993, resulting in a stable annual number of hip fractures. The magnitude of the period+cohort effects suggests a risk modulation in parity with other established risk factors for hip fracture. Gender-specific differences may partly result from changes in hormone-replacement or anti-osteoporosis therapy.