Patients at increased fracture risk: identification and pharmacological treatment

Abstract

Fragility fractures are common and are associated with a substantial burden for patients and the healthcare system. Hip fractures in particular are associated with increased morbidity, institutionalisation, and even mortality with a mortality rate between 20-30% in the first year. This burden is increasing due to the ageing of the population. It is therefore important to identify high-risk patients in order to implement preventive measures. One such measure is treatment with anti-osteoporosis drugs which may reduce (subsequent) fracture risk by 30–70%, depending on the drug and fracture site. Over the last years, identification of individuals at increased fracture risk is increasingly based on absolute fracture risk as predicted from clinical risk factors such as age, gender, use of certain drugs (e.g. glucocorticoids) and diseases (e.g. rheumatoid arthritis) optionally combined with bone mineral density. The overall thesis aim was to evaluate and to help improve prediction of absolute fracture risk and implementation of pharmacological treatment. The results from this thesis confirm the importance of fracture prevention where 1-year mortality after hip fracture was respectively 2.4 and 3.5-fold higher among females and males when compared to the general population which has not declined over the last decade. Furthermore, up to 33% of hip fracture patients sustained another fracture within 5 years. This thesis shows the importance of external validation of the most utilised fracture risk prediction model, FRAX, with assessment of calibration (agreement between predicted and actual risks) in order to avoid over –or under treatment. This is true for the general population but also for subpopulations where fracture risk prediction is relevant. Indeed, we show a significant overestimation of predicted risks from the FRAX model among patients with rheumatoid arthritis. This requires a complete and valid registration of fractures, but this is currently lacking for non-hip fractures in the Netherlands. Furthermore, anti-osteoporosis drug prescribing remains inadequate with 54% of hip fracture patients not receiving an anti-osteoporosis drug in the year 2010. In addition, 25% of fracture patients stopped drug treatment within 1 year which increased to 55% after 5 years while a treatment duration of 3-5 years has been advocated. Males in particular were significantly less likely (1.7-fold) to initiate drug treatment and risk of early discontinuation was 1.7-fold higher among elderly (≥ 80 years) fracture patients. Finally, in a randomised clinical trial it appeared that feedback by pharmacists to physicians of patients eligible for pharmacological prevention of glucocorticoid-induced fracture risk did not significantly increase the prescribing of anti-osteoporosis drugs over a period of 6 months (11.4% intervention group vs. 8.0% control group; HR 1.47, 95% CI: 0.91-2.39). In conclusion, fracture risk prediction models may be valuable tools for fracture prevention but we stress the importance of external validation with assessment of calibration before use in clinical practice which demands good registration of fractures. Furthermore, the inadequate uptake of preventive treatment with anti-osteoporosis drugs indicates a need for increased awareness about the consequences of fragility fractures among both physicians and patients and calls for additional measures to improve this.