Hip Fracture Incidence and Prevalence of Osteoporosis in Sweden in Recent Decades

Abstract: An increase in hip fracture incidence during the recent half-century has been reported worldwide and has been predicted to continue. Recent publications, however, have reported stable or even decreasing hip fracture incidence. No reports evaluate whether these changes in hip fracture incidence are associated with changes in bone mineral density (BMD), if they apply in both urban and rural settings, and few have tried to analyze whether they are attributable to cohort and/or period effects. In population-based samples of urban and rural women (n=437, age ≥50 years) and men (n=277, age ≥60 years) risk factors for hip fracture were evaluated in 1988/89 and 1998/99 as was BMD in Malmö women from 1970 to 1999 (n=459, age ≥50 years). The annual number of hip fractures in the respective region, gender and age group as well as nationwide in all individuals aged ≥50 years were obtained from central national inpatient data from 1987 to 2002 and in Malmö women from 1967 to 2001 by local records. Age-Period-Cohort (APC) effects were evaluated by log likelihood estimates in Poisson regression models. In all local cohorts the age-standardized hip fracture incidence was stable from 1987 to 2002 and from 1969 to 2001, and BMD and prevalence of osteoporosis were stable between 1974 and 1999 and between 1988/89 and 1998/99. In nationwide Swedish data hip fracture incidence was stable from 1987 to 1996, after which a decrease was evident. There were evident period+cohort effects in women, with a major reduction in hip fracture incidence by subsequent birth cohorts (estimated incidence rate ratio 2.2 comparing women born 1889–96 to women born 1945–52). This thesis infers that the hip fracture incidence has levelled off also in Sweden and has since the mid 1990s been replaced by a decreasing annual incidence. In women this appear to be the result of cohort+period effects, in that that those born more recently had a lower incidence than those born earlier. As no differences in BMD or prevalence of osteoporosis were apparent, the changes in hip fracture incidence seem attributable to other factors, probably linked to the major changes in society during the century-long lifespan of the examined individuals. The data presented in this thesis are important both for projecting future hip fracture burden and when planning for hip fracture resources in the future.