Two-Thirds of All Fractures Are Not Attributable to Osteoporosis and Advancing Age: Implications for Fracture Prevention

Context Although bone mineral density (BMD) is strongly associated with fracture and postfracture mortality, the burden of fractures attributable to low BMD has not been investigated. Objectives We sought to estimate the population attributable fraction of fractures and fracture-related mortality that can be attributed to low BMD. Design and Setting This study is a part of an ongoing population-based prospective cohort study, the Dubbo Osteoporosis Epidemiology study. In total, 3700 participants aged ≥50 years participated in the study. Low-trauma fracture was ascertained by X-ray reports, and mortality was ascertained from the Birth, Death and Marriage Registry. Results Overall, 21% of women and 11% of men had osteoporotic BMD. In univariable analysis, 21% and 16% of total fractures in women and men, respectively, were attributable to osteoporosis. Osteoporosis combined with advancing age (>70 years) accounted for 34% and 35% of fractures in women and men, respectively. However, these two factors accounted for ∼60% of hip fractures. About 99% and 66% of postfracture mortality in women and men, respectively, were attributable to advancing age, osteoporosis, and fracture; however, most of the attributable proportion was accounted for by advancing age. Conclusions A substantial health care burden of fracture is on people aged <70 years or nonosteoporosis, suggesting that treatment of people with osteoporosis is unlikely to reduce a large number of fractures in the general population.

Secular Changes in Postfracture Outcomes Over 2 Decades in Australia: A Time-Trend Comparison of Excess Postfracture Mortality in Two Birth Controls Over Two Decades

Context: Hip fracture incidence has been declining and life expectancy improving. However, trends of postfracture outcomes are unknown. Objectives: The objective of the study was to compare the refracture risk and excess mortality after osteoporotic fracture between two birth cohorts, over 2 decades. Design: Prospective birth cohorts were followed up over 2 decades (1989–2004 and 2000–2014). Setting: The study was conducted in community-dwelling participants in Dubbo, Australia. Participants: Women and men aged 60–80 years, participating in Dubbo Osteoporosis Epidemiology Study 1 (DOES 1; born before 1930) and Dubbo Osteoporosis Epidemiology Study 2 (DOES 2; born after 1930) participated in the study. Main Outcome Measure: Age-standardized fracture and mortality over two time intervals: (1989–2004 [DOES 1] and 2000–2014 [DOES 2]) were measured. Results: The DOES 2 cohort had higher body mass index and bone mineral density and lower initial fracture rate than DOES 1, but similar refracture rates [age-standardized refracture rates per 1000 person-years: women: 53 (95% confidence interval [CI] 42–63) and 51 (95% CI 41–60) and men: 53 (95% CI 38–69) and 55 (95% CI 40–71) for DOES 2 and DOES 1, respectively). Absolute postfracture mortality rates declined in DOES 2 compared with DOES 1, mirroring the improvement in general-population life expectancy. However, when compared with period-specific general-population mortality, there was a similar 2.1- to 2.6-fold increased mortality risk after a fracture in both cohorts (age-adjusted standardized mortality ratio, women: 2.05 [95% CI 1.43–2.83] and 2.43 [95% CI 1.95–2.99] and men: 2.56 [95% CI 1.78–3.58] and 2.48 [95% CI 1.87–3.22] for DOES 2 and DOES 1, respectively). Conclusion: Over the 2 decades, despite the decline in the prevalence of fracture risk factors, general-population mortality, and initial fracture incidence, there was no improvement in postfracture outcomes. Refracture rates were similar and fracture-associated mortality was 2-fold higher than expected. These data indicate that the low postfracture treatment rates are still a major problem.