Introduction

Although osteoporosis is a silent disease, it is an important public health concern around the world. Osteoporosis-related fracture is one of the leading causes of significant morbidity and disability in old patients, and it increases economic burden on the health care system [1–4].

It is important to determine the incidence of osteoporosis-related fracture and trace its trends, because they are the first step in providing baseline data for establishing policy in public health care.

Several studies have suggested that incidences of osteoporosis-related fractures (hip, spine, distal radius, and humerus) vary across the world. Although Korea might be expected to be at low risk for osteoporosis because it is an Asian country, there are few data available to confirm this in terms of the incidence of osteoporosis-related fractures in Korea. In addition, comparisons with other countries have not been established.

For comparisons across countries and regions, the determination of residual lifetime risk is also an epidemiologically useful tool. The WHO definition of osteoporosis, which is based on bone mineral density (BMD), has shortcomings with regard to risk assessment and as a basis for treatment decisions; analysis of osteoporosis-related fractures has been proposed as an alternative that overcomes these drawbacks [5]. Gender- and age-specific lifetime absolute risk of osteoporosis-related fractures are reported in studies from other region of the world [6–12], but no such estimations are published regarding the Korean population [13].

The purpose of this study was to determine the incidence and lifetime risk of major osteoporotic (hip, spine, distal radius, and humerus) fractures in a nationwide cohort study of the Korean population, using data from the Health Insurance Review and Assessment Service (HIRA).

Materials and methods

Subjects

We analyzed data from the nationwide claims database of the HIRA of Korea, which is comprised of adjusted medical and pharmacy claims for all Korean citizens, between the years of 2005 and 2008. In Korea, 97.0% of the population is legally obliged to enroll in the Korea National Health Insurance Program. Patients pay an average of 30% of the total medical costs to clinics or hospitals that manage almost all diseases, except for those not covered by insurance, such as cosmetic surgery or some new unproven therapies. All clinics and hospitals then submit claims data for inpatient and outpatient care, including diagnoses [in International Classification of Diseases, 10th revision (ICD-10)], procedures, prescription records, demographic information, and direct medical costs, to the Korean HIRA to obtain 70% reimbursement of the total medical cost from the government. The remaining 3% of the population not insured by the Korean National Health Insurance Program are either covered by a Medical Aid Program or are temporary or illegal residents. The claims data covered by the Medical Aid program were also reviewed by the HIRA. Therefore, virtually all information about patients and their medical records is available from the Korean HIRA database, which has been used on several occasions for epidemiological studies [14–18]. All new visits or admissions to Korean hospitals for osteoporosis-related fracture were recorded prospectively in a nationwide cohort by the above-described system using ICD-10 codes and procedures. Each patient has a unique identifier, making it feasible to track a specific patient over time. The total number of men and women over the age of 50 in the Korean population was obtained from the web site of the Statistics Korea (http://www.kosis.kr/), which is the central government organization for statistics [19].

To identify osteoporosis-related fractures, selected ICD-10 codes and a cut-off value of 50 years of age were used [14, 16, 20].

Fractures were identified on the basis of selected ICD-10 codes; hip [ICD-10 codes S72.0 (fracture of the femoral neck), S72.1 (pertrochanteric fracture) and 7 procedures (open reduction of fractured extremity-femur, closed pinning-femur, external fixation-pelvis/femur, closed reduction of fractured extremity-pelvis/femur, bone traction, skin traction, hemiarthroplasty-hip)]; spine [S22.0 (fracture of the thoracic spine), S22.1 (multiple fractures of the thoracic spine), S32.0 (fracture of the lumbar spine), M48.4 (fatigue fracture of vertebra) and M48.5 (collapsed vertebra, NEC)]; distal radius [S52.5 (fracture of the distal radius) and S52.6 (combined fracture of the distal radius/ulna)]; humerus [S42.2 (fracture of the proximal humerus) and S42.3 (fracture of shaft of humerus)]; and overall fractures.

When an individual had more than one fracture with the same encoding, the patient was followed from the first encoding and recounted if a further encoding occurred six months or later after the first event [14, 21]. In overall fractures, we counted plural ICD-10 codes occuring on the same date as a single event.

These data were retrospectively evaluated to determine the trend of annual incidence and lifetime risk of osteoporosis-related fractures (hip, spine, distal radius, and humerus fracture), in men and women aged 50 years or more between the years of 2005 and 2008.

The study was subject to control by the National Board of the HIRA, and approved by the institutional review board.

Statistical analysis

To determine the recent trends of fracture incidences, the patients were divided into groups according to their age (subdivided into 5-year increments), gender, and fracture site (hip, spine, distal radius, and humerus). The age-adjusted and gender-specific incidence rates (per 100,000 person years) were calculated based on the general population of Korea during each year. Cochran–Armitage test for trends was performed to evaluate whether the incidence rate changed from 2005 to 2008. The age-adjusted and gender-specific incidences from 2005 to 2008 were used for this analysis. P values of <0.05 were considered significant.

To compare the incidence of the osteoporosis-related fractures between Korea and other geographic areas [20, 22–28], we used age-standardized incidence rates, which are weighted averages of the age-specific incidence rates of people in the corresponding age groups in a standard population, which in this case was estimated for the population in the United States on 2008 [29, 30].

The residual lifetime risks of osteoporosis-related fractures were estimated by simple approximation using the age- and gender-specific incidence rate in 2008 and life tables for the Korean population in 2008 released by Statistics Korea.

To compute the residual lifetime risk for each ([k]) age, Pr[k] was defined as the probability of having no osteoporosis-related fracture for entire lifetime starting at age of k. This probability may be discretely approximated using the following formula [27]:

$$ Pr\left[ k \right] = d\left[ k \right]\left( {1-I\left[ k \right]} \right) + \left( {1 - d\left[ k \right]} \right)\left( {1 - I\left[ k \right]} \right)d\left[ {k + 1} \right]\left( {1 - I\left[ {k + 1} \right]} \right) + \left( {1 - d\left[ k \right]} \right)\left( {1 - I\left[ k \right]} \right)\left( {1 - d\left[ {k + 1} \right]} \right)\left( {1 - I\left[ {k + 1} \right]} \right)d\left[ {k + 2} \right]\left( {1 - I\left[ {k + 2} \right]} \right) + \cdots $$

where d[k] is the probability of dying at k-year-old, I[k] is the probability of having a fracture at k-year-old, and so on.

The probability of not having an osteoporosis-related fracture during the remaining lifetime also can be written as follows:

$$ Pr\left[ k \right] = \sum\limits_{s = k}^{100} {\left[ {\left\{ {\prod\limits_{t = k}^{s - 1} {\left( {1 - d\left[ t \right]} \right)\left( {1 - I\left[ t \right]} \right)} } \right\}d\left[ s \right]\left( {1 - I[s]} \right)} \right]} $$

These values were replaced with the corresponding incidence or mortality rates in this study. The residual lifetime risk of an individual aged k years experiencing an osteoporosis-related fracture is then estimated by 1 − Pr[k] [27]. We assumed that mortality is 1 in 100 years. Since the numbers of incidence and population data are not available, we calculated the incidence rate only for those in the ages of 95–100. All database management and analysis were performed using SAS statistical package version 9.1.3 (SAS Institute Inc., Cary, NC, USA).

Results

Among those aged 50 years or older, the crude overall incidence of osteoporosis-related fractures was 1,635 per 100,000 person years from 2005 to 2008, and the gender-specific incidence was 725 per 100,000 person years for men and 2,408 per 100,000 person years for women (Table 1).

Table 1 Age-adjusted and gender-specific incidence (per 100,000 persons) of osteoporosis-related fractures in Korean over 50 years old
Full size table

Although the absolute number of osteoporosis-related fractures increased, age-adjusted incidence rates were stable for most fractures during the period of 2005–2008 (Table 1). However, the age-adjusted incidence of distal radius fracture decreased in women (p = 0.0183) (Table 2). Among those aged 50 years or older, the incidence of spine fractures was highest (969 per 100,000 persons), followed by distal radius (422), hip (157), and humerus (81) in 2008, sequentially.

Table 2 Trend test of age-adjusted and gender-specific incidence of osteoporosis-related fractures from 2005 to 2008
Full size table

Figure 1 shows the age- and gender-specific incidences for osteoporosis-related fracture from 2005 to 2008. For both genders, the incidence rate of most osteoporosis-related fractures increased with age. The gender ratio of the female to male incidence of osteoporosis-related fractures was >1.0 for all age groups, especially in the distal radius and the spine.

Fig. 1
figure 1

Age-specific incidences of osteoporosis-related fractures in men and women between 2005 and 2008

Full size image

The incidence of hip fracture which is standardized to population in 2008 in the United States was lower than that of Japan in women, and the incidence of distal radius fracture was comparable with Switzerland in men and women (Table 3).

Table 3 Age-adjusted incidence (per 100,000) of osteoporosis-related fractures over 50 years old in different populations
Full size table

Figure 2 shows the residual lifetime risk of osteoporosis-related fracture for individuals aged 50 years or more in men and women. The residual lifetime risk of osteoporosis-related fracture for individuals with age 50 was estimated to be 23.8% for men and 59.5% for women.

Fig. 2
figure 2

The residual lifetime risk of osteoporosis-related fractures in men and women by age

Full size image

Discussion

Although there were several epidemiologic studies on hip fractures in Korea, to the best of our knowledge this is the first nationwide, population-based, epidemiological study of osteoporosis-related fractures (hip, spine, distal radius, and humerus). Moreover, this is one of the few studies worldwide presenting the lifetime risks of osteoporosis-related fractures based on national health insurance claims data.

With regard to hip fracture, the incidence rate was stable during the period from 2005 to 2008 in Korea, and the reported incidence of hip fracture in this study is in the middle of the range observed in the previous cohort studies in Korea [31, 32]. Future studies will be necessary to determine the trends of hip fracture incidences over longer periods by using consistent protocols.

Several studies have presented a wide geographic variation in the incidence of osteoporotic fractures between different populations [13, 27, 33–35]. In general, populations in higher latitudes seem to show a higher incidence of osteoporotic fracture [13, 27, 33–35]. Asian populations, including the Chinese and Japanese, were considered to have low- to moderate-risk for osteoporotic fracture [13]. In this study, we could not directly compare the incidence of osteoporosis-related fractures in Korea with that of other countries, because the definition of osteoporosis-related fractures varied in each study. However, the incidence of hip fracture in Korean women was between those of Singapore and Thailand, which were considered countries with moderate risk for osteoporotic fracture (Table 3) [13, 27, 28, 31]. Although the residual lifetime risk of a hip fracture at age 50 was 5.23% in men and 12.31% in women, lower than that of Japan (Table 4) [27], the residual lifetime risk of osteoporosis-related fractures for individuals aged 50 is estimated to be 59.5% for women and 23.8% for men in Korea.

Table 4 Comparison of gender-specific residual lifetime risk (%) of hip and distal radius fractures across countries at age 50 years
Full size table

On the other hand, it was interesting that the age-adjusted incidence of the distal radius fracture was as great as that of Switzerland, a population with high risk for osteoporotic fractures [20] (Table 3). In addition, the residual lifetime risk of a distal radius fracture was also similar to that of Sweden, a population with high risk for osteoporotic fractures [8] (Table 4). We do not know the exact reason for the high incidence and lifetime risk of distal radius fracture in Korea, but there could be several explanations for these findings. First, in this study, ICD-10 codes for distal radius contained not only S52.5 (fracture of the distal radius) but also S52.6 (combined fracture of the distal radius/ulna), and this made it possible to count cases of distal radius fracture that were missed in other studies. Second, not all distal radius fracture is due to osteoporosis.

Understanding osteoporosis as a problem predominantly occuring in elderly women, it can be also well understood from the study that women show consistently higher incidence of osteoporosis-related fractures than men; this is comparable to previous studies performed in other countries [6, 8, 36].

Based on these results, the incidence rate of the osteoporosis-related fractures increased with age after age 50, and the residual lifetime risk decreased after the age of 70. Considering the lifetime risk as a more intuitive parameter, a concentrated management for osteoporosis is necessary to reduce the incidence and socio-economic burden of the osteoporosis-related fractures in Korea, especially in these populations of elderly women.

Our study has some limitations. First, patients that had multiple osteoporosis-related fractures during the observational period were not specifically identified in the present survey, which may have affected the result of the overall fractures. This was why total numbers of overall fractures are different from summations of each fracture. Second, all patients with osteoporosis-related fractures may not be coded in this nationwide database. For example, many patients with asymptomatic spine fracture may not visit medical institutions in Korea. This can cause incidence figures based on the HIRA database to be underestimated. Third, when identifying osteoporotic fracture by means of ICD-10 codes, it is difficult to distinguish a new fracture from deterioration of an old one. Prevalent spine fractures, especially, might be counted as new fractures, and this might give misleading results. Forth, BMD of patients was not available in this study due to our study design based on the National Claim Registry. It was possible that fractures due to high-energy trauma were included in this study, because the distinction between high and low-energy fractures could not be made by using the ICD-10 coding system. However, we used the additional criteria of “aged 50 years or more” to exclude non-osteoporotic fractures. In addition, these inclusion criteria including ICD code and age without BMD measurement could be found in several studies on osteoporosis-related fractures [8, 11, 12]. Therefore, our study without BMD measurement could be justified.

Despite these limitations, the results of this study show that age-adjusted incidence rates were stable for most fractures during the study period. In terms of residual lifetime risk, Korea was comparable with countries with moderate risk for osteoporosis-related fracture such as Japan.