Duration of osteoporosis treatment to reduce the risk of subsequent osteoporotic fracture and all-cause mortality in elderly hip fracture patients in a Korean real-world study

Hip fracture rates have attained epidemic proportions worldwide, particularly in the elderly population, and are a major public health concern [1]. Globally, the incidence of hip fracture was 14.2 million in 2019, an increase of approximately 93.0% since 1990 [1]. Among 271,197 South Korean women at the age of 66 years between January 1, 2008 and December 31, 2015 who were screened for bone mineral density, the 10-year cumulative incidence of fragility fractures was higher in women with osteoporosis (hazard ratio [HR] [95% confidence interval; CI]: 1.68 [1.64–1.72]) compared with women who had a normal bone mineral density [2]. With an increase in the aging population, the economic burden of hip fracture is predicted to increase further in the future [3]. Since the risk of hip fracture increases immediately after a preceding fracture [4], a multistakeholder coalition was assembled by the American Society for Bone and Mineral Research to develop clinical recommendations for the optimal prevention of secondary fractures among people aged ≥ 65 years with a hip or vertebral fracture [5]. The coalition recommended that pharmacological therapy for osteoporosis must be offered to individuals aged ≥ 65 years to reduce the risk of additional fractures [5].

In real-world studies conducted in patients with hip fracture, osteoporosis medications have reduced the incidence of fracture and all-cause mortality over time [6, 7]. Despite good evidence showing an increased risk of hip fracture in patients with osteoporosis, the mean rate of global treatment initiation with osteoporosis medications in patients with hip fracture is low (6.5%–33.5%) [8‐12]. In Korea, approximately 42.0% of patients had initiated osteoporosis medication in the first 12 months after an osteoporotic fracture. Moreover, the rate of prescription of osteoporosis medications was lower in patients with a hip fracture than in those with a vertebral fracture (36.6% versus 53.2%) [13]. In a study conducted in the United States of America, 19.0%–26.0% of patients discontinued osteoporosis treatment at an early stage [14], potentially leading to a higher risk of subsequent fracture or death. Long-term osteoporosis treatment (> 1 year) might reduce the risk of fracture and all-cause mortality [6, 7]. In a real-world study from Spain, 28.3% of patients were prescribed an osteoporotic medication within 6 months since the hip fracture and 54.5% of naïve patients had discontinued the treatment 1 year after the fracture [15]. Since only a few studies have evaluated the impact of osteoporosis medication duration on the incidence of subsequent fracture or death after a hip fracture [16, 17], evidence on treatment duration focusing on early discontinuation and long-term continuation needs to be investigated.

Therefore, we proposed a study to evaluate the association between treatment duration of osteoporosis medications and clinical outcomes of patients with a hip fracture. We also questioned whether there was a difference in the risk of subsequent osteoporotic fracture (SOF) and mortality among patients who discontinued the osteoporosis treatment in the first year of treatment or who continued the treatment for more than 1 year according to the osteoporosis treatment duration. Consequently, this study aimed to estimate the risk of SOF and all-cause mortality following an initial hip fracture and to evaluate the risk reduction in SOF and all-cause mortality by osteoporosis treatment duration among the elderly patients with hip fracture in South Korea.

The findings from this study revealed that approximately 30% of patients (6,217/19,390) with a hip fracture who were treated with osteoporosis medications initiated the treatment within 3 months of the initial fracture. The risk of SOF showed a trend toward reduction in patients in the TC group, and the risk of all-cause mortality was significantly reduced in patients in the TC and LD groups compared with those in the ST group. A reduced risk of SOF and all-cause mortality was associated with longer treatment with osteoporosis medications from the treatment initiation. Consequently, there is a need to increase the implementation of long-term prescription of osteoporosis medications in South Korea for patients who are predisposed to hip fractures.

Patients with more than 12 months of treatment with osteoporosis medications (TC group) showed a decreased risk of SOF compared with those who received treatment for less than 3 months (ST group). Our results corroborate the findings of previous studies that reported a lower risk of recurrent fractures in patients with hip fracture who took osteoporosis medications long-term [16, 17, 22]. In the TC group, the risk of subsequent vertebral fracture was significantly reduced by 35% and that of hip fracture was reduced by 21%; the reduction in the risk of hip fracture was not statistically significant. Our findings were similar to those reported in trials of a few osteoporosis medications where a significant reduction was noted in the risk of subsequent vertebral fracture but not subsequent hip fracture [23, 24]. However, there are several types of osteoporosis medications and each of these medications has a different potency, which can affect the efficacy in the prevention of fractures [25]. For example, the use of certain osteoporosis medications decreased the risk of subsequent hip fracture in patients aged > 80 years [26]. In a meta-analysis that evaluated the efficacy of osteoporosis treatment in patients aged > 75 years, osteoporosis treatment was significantly associated with a decreased risk of hip fracture at 1 and 3 years [27]. Thus, to reduce the risk of subsequent hip fractures, it is essential to choose the appropriate type of osteoporosis medication.

In our study, the risk of all-cause mortality was significantly lower in the LD and TC groups compared with the ST group. Treatment with osteoporosis medications can reduce mortality in patients with hip fractures [28, 29]. The effect of treatment duration on cause-specific mortality stratified by the leading cause of death was statistically significant only for diseases of the circulatory system in the LD and TC groups. Osteoporosis and diseases of the circulatory system coexist in elderly patients, and there is evidence of a potential causal link between cardiovascular disease and bone loss [30]. However, additional analysis of the effect of bisphosphonates on mortality due to atherosclerotic cardiovascular disease revealed that the effect was not statistically significant. These findings concord with those of a meta-analysis that concluded that bisphosphonates do not have beneficial or harmful effects on atherosclerotic cardiovascular events [31]. As the cause of death in the elderly population is difficult to determine, it may be difficult to estimate the treatment effect of osteoporosis medications on cause-specific death. The reduced risk of mortality in observational studies such as this needs to be interpreted with caution as the results may be confounded by the high adherence rates to osteoporosis medications [32]. Nevertheless, this study supports the preventive effect of continuation of osteoporosis medications in reducing all-cause mortality.

The patients in the LD and TC groups who continued using osteoporosis medications for ≥ 6 months had a reduced risk of all-cause mortality than those in the ST group who were treated with osteoporosis medications for < 3 months. To the best of our knowledge, this is the first study to show a reduced risk of all-cause mortality but not SOF with a treatment duration of ≥ 6 months compared with a treatment duration of < 3 months. Patients’ adherence or compliance to osteoporosis medications is one of the most important factors to reduce the risk of fracture or mortality [16, 17, 33]. Previous studies showed that treatment duration of > 12 months with good adherence reduced the risk of SOF in patients with osteoporosis medications after hip fractures [16, 17]. Compared with the risk of SOF, a decrease in the risk of all-cause mortality was observed after the treatment period of ≥ 6 months, suggesting that a potential threshold for the treatment period that reduced the risk of all-cause mortality might be shorter than that for reducing the risk of SOF. The differences in the effects of osteoporosis medications may be attributed to the differences in insurance reimbursement policies, eligibility by country, time and insurance type. Further well-designed studies are warranted to evaluate the exact threshold for the duration of osteoporosis medications to reduce the risk of all-cause mortality. Nonetheless, it is important to continue treatment with osteoporosis medications to reduce the risk of SOF and all-cause mortality. In our study, treatment duration of > 6 months for osteoporosis medications did not significantly reduce the risk of SOF; however, it reduced the risk of all-cause mortality.

One of the strengths of this study is that we use of a nationwide representative population-based cohort. The incidence of both SOF and all-cause mortality was estimated in our study. A detailed cause of death was not provided due to the possibility of personal identification and because the number of patients associated with each cause of death was limited. Nevertheless, our study is meaningful in that the cause of death was analyzed. The study also applied a validated algorithm to identify outcome events to ensure the accuracy and completeness of claims-based information. Lastly, rigorous epidemiological methods have been used to reduce bias and confounding. By excluding patients who initiated the treatment after 3 months, we could minimize the confounding by different initiation times; by excluding the patients who experienced SOF or mortality within 1 year after the index date, we could minimize the immortal person-time. Moreover, we employed an additional propensity score-weighted model in a sensitivity analysis to minimize residual confounding in the regression analysis. Our study has a few limitations including the potential for unmeasured confounding due to a lack of clinical information, such as bone mineral density, body mass index, and incidence of fragility or traumatic fracture. Since the many clinical guidelines recommend initiating treatment regardless of BMD score in patients with an osteoporotic fragility fracture, we specifically included patients who began treatment within 3 months of a hip fracture [34]. Therefore, we believed that narrowed patient inclusion criteria can minimize the effect of potential for unmeasured confounding due to BMD. The sample size was insufficient to draw meaningful conclusions in some subgroups, especially in males. Factors related to treatment interruption, such as deterioration of existing health conditions and renal dysfunction, were not analyzed. However, as no association was found in the cause-specific mortality analysis, there is little possibility of discontinuation due to poor health.

The initial treatment rate with osteoporosis medications was low among elderly patients with hip fracture in South Korea. Compared with patients whose treatment duration was < 3 months (ST group), patients with treatment duration ≥ 6 months showed reduced all-cause mortality and patients with treatment duration ≥ 12 months showed reduced all-cause mortality and SOF. Furthermore, these results were evident in females compared with males and for a second vertebral fracture compared with other types of fractures. A longer duration of treatment with osteoporosis medications was beneficial in reducing the risk of SOF and all-cause mortality in elderly patients with hip fracture in Korean routine clinical practice. These findings underscore the importance of early and continuous osteoporosis treatment for elderly patients with hip fracture to improve health outcomes. The benefits of long-term osteoporosis treatment should be discussed in clinical practice to reduce the burden of disease and improve overall health outcomes.