Study on hip fracture and the treatment effect of alendronate among the elderly in Taiwan : a nationwide population-based retrospective cohort study.

• Main Authors: Chang-Bi Wang, 王昶弼
• Other Authors: Wen-Miin Liang
• Format: Others
• Language: en_US
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/cq2968

博士 === 中國醫藥大學 === 公共衛生學系博士班 === 102 === Background Osteoporotic hip fractures cause high mortality in the elderly population. However, few population studies reported the long-term mortality, hip refracture of hip fracture among the elderly in Asian population. In addition, the dose response of the use of bisphosphonates, one kind of drug to prevent the loss of bone density, on the prevention of refracture remained unclear. There are three aims in this study. First, we assessed the incidence, excess mortality, and risk factors after osteoporotic hip fractures through inpatients aged 60 years or older. Second, we considered death as a competing event alongside age, gender, surgery type and comorbidities in the analysis of retrospective data of the hip fracture population and to calculate the risk of hip refracture. Third, we tried to provide evidence to support the view that bisphosphonates may be an effective treatment for the prevention of refracture in Asian osteoporosis patients. Materials and Methods For the first aim, a total of 143,595 patients with hip fracture were selected from Taiwan National Health Insurance database in the years 1999 to 2009 and followed up until the end of 2010. Annual incidence, mortality and SMR, and mortality and SMR at different periods after fracture were measured. For the second aim, we selected subjects with first discharge diagnosis code of hip fracture, aged 60 years or older, who were admitted to hospitals between 2000 and 2005. In total, 77,795 subjects with hip fracture were enrolled in the study and followed up until exiting the NHI program, death, or the end of 2010. We considered death using the Fine and Gray Regression Model to calculate subdistribution hazards for hip refracture. For the third aim, we conducted a population-based, retrospective case-control cohort study of 1,392 patients in Taiwan. Case cohorts (with alendronate drugs) and control cohorts (without alendronate drugs) were selected from among patients with a primary diagnosis of fracture. Controls were matched by age, gender, index day and fracture history at a ratio of 1∶4 (case : controls) during the same time period. Odds ratios (ORs) were calculated to determine the risk of refracture by conditional logistic regression model. Results First, from 1999 to 2005, hip fracture incidence gradually increased and then fluctuated after 2006. From 1999 to 2009, the male-to-female ratio of annual incidence increased from 0.60 to 0.66, annual mortality for hip fracture decreased from 18.10% to 13.98%, male-to-female ratio of annual mortality increased from 1.38 to 1.64, and annual SMR decreased from 6.15 to 2.94. Follow-up SMR at one, two, five, and ten years post-fracture was 5.24, 3.63, 3.56, and 2.88, respectively. Females had higher follow-up SMR in the younger age groups (60–69 yr of age) but lower follow-up SMR in the older age groups (over 80 yr of age) compared with males. Second, the results of the Fine and Gray regression model determined the subdistribution hazards for gender(female vs male, HR=0.79, 95% CI=0.67-0.93), hip fracture type (femoral neck vs trochontric, HR=1.48, 95% CI=1.17-1.89), surgery type (internal fixation vs. replacement, HR=1.46, 95% CI=1.15-1.91), and number of Charlson comorbidities( 1 vs 0, HR=0.97, 95% CI=0.78-1.19; >=2 vs 0, HR=0.80, 95% CI=0.65-0.99). In the traditional Cox model showed that number of comorbidities had high risk of hip refracture, however, the Fine and Gray regression model had the reversed results. Third, after adjusting for age, gender, and comorbidities, the case cohort patients who took a “≧183 d” course of alendronate showed ORs of refracture of 0.14, 0.36, and 0.43 at the end of the one-, two-, and three-year intervals when compared with control cohort, respectively. However, the case cohort patients who took a “under 182 d” course showed ORs of refracture of 0.76, 0.86, and 0.86 at the end of the one-, two-, and three-year intervals when compared with control cohort, respectively. Overall, patients in the case cohort with low-frequency alendronate use carried the highest risk for refracture, especially those with hypertension and Diabetes Mellitus. Conclusions Among the studied patients, incidence is gradually decreasing along with annual mortality and SMR. Hip fracture affects short-term especially the first year but not long-term mortality. The Fine and Gray regression model is a useful tool that allows better and more accurate estimation of hazard ratio. Our results provided evidence to support the viewpoint that alendronate may be a feasible method for the prevention of refracture.