E-submission
KOTA
TOTA
TOTS
Articles
- Page Path
- HOME > J Musculoskelet Trauma > Volume 38(4); 2025 > Article
-
Original Article
- Risk factors for ankle fractures in older adults based on clinical components of the Fracture Risk Assessment (FRAX) tool and comorbidities in Korea: a retrospective case-control study
-
Myeong Jun Song1
, Se Woong Jang1, Jun Young Lee1, Seojin Park2 -
Journal of Musculoskeletal Trauma 2025;38(4):193-202.
DOI: https://doi.org/10.12671/jmt.2025.00143
Published online: October 24, 2025
1Department of Orthopedic Surgery, Chosun University College of Medicine, Gwangju, Korea
2Department of Nursing Science, Gwangju University, Gwangju, Korea
- Correspondence to: Jun Young Lee Department of Orthopedic Surgery, Chosun University Hospital, 365 Pilmun-daero, Dong-gu, Gwangju 61453, Korea Tel: +82-62-220-3140 Email: leejy88@chosun.ac.kr
• Received: March 10, 2025 • Revised: July 14, 2025 • Accepted: August 4, 2025
© 2025 The Korean Orthopaedic Trauma Association
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
- 2,704 Views
- 47 Download
Abstract
-
Background Ankle fractures are common in older adults; however, their relationship with osteoporotic fractures remains unclear. This study aimed to evaluate potential risk factors for ankle fractures in older adults by analyzing individual clinical components of the Fracture Risk Assessment (FRAX) tool and comorbidities.
-
Methods We conducted a retrospective case-control study including 84 patients aged ≥65 years with ankle fractures and 150 controls who underwent bone mineral density (BMD) testing without prior ankle fractures. The variables analyzed included age, sex, body mass index, smoking, alcohol consumption, prior fracture history, and comorbidities such as hypertension, diabetes mellitus, and dementia. BMD was measured at the spine, total hip, and femoral neck.
-
Results Univariate analysis showed that alcohol consumption, diabetes mellitus, and total hip T-score categories were significantly associated with ankle fractures. In binary logistic regression, alcohol consumption remained significantly associated with lower ankle fracture risk (odds ratio [OR], 0.196; 95% confidence interval [CI], 0.066–0.589; P=0.004), and osteoporosis at the total hip was associated with decreased risk (OR, 0.253; 95% CI, 0.069–0.934; P=0.039), whereas osteopenia was not (OR, 1.258; 95% CI, 0.700–2.263; P=0.443). Diabetes mellitus was marginally significant in the adjusted model (OR, 1.838; 95% CI, 0.973–3.472; P=0.061). The Hosmer-Lemeshow test showed no evidence of poor fit (P=0.214).
-
Conclusions These findings suggest that alcohol consumption and total hip osteoporosis were inversely associated with ankle fractures in older adults; however, these associations should be interpreted with caution given the hospital-based selection of controls and the small number of exposed cases, and further studies are needed to clarify them. The FRAX score itself was not calculated; instead, this study focused on analyzing selected clinical components. Limitations include the retrospective design, lack of fall and medication data, and cross-sectional BMD assessment.
-
Level of evidence III.
Introduction
Ankle fractures are increasing in incidence due to aging, increased sports activities, industrial accidents, and traffic accidents, making them one of the most common fractures encountered in orthopedic practice [1]. In particular, among elderly individuals aged ≥65 years, ankle fractures occur not only due to high-energy trauma (e.g., traffic accidents, falls) but also from low-energy injuries, such as tripping during ambulation [2]. In Finland, which has experienced aging earlier than Korea, the number of ankle fractures in individuals aged 60 years and older increased from 369 cases in 1970 to 1,545 cases in 2000, representing an approximately 319% increase. If this trend continues, it is projected that the number of low-trauma ankle fractures in this population will triple by 2030 [3]. Consequently, evaluating the risk of ankle fractures and establishing prevention strategies have become critical issues. To assess fracture risk, the Fracture Risk Assessment (FRAX) tool, developed by a team at the University of Sheffield (formerly a World Health Organization [WHO] Collaborating Centre), estimates the 10-year probability of major osteoporotic fractures in individuals aged 40–90 years. The FRAX score considers various clinical factors, including age, sex, body mass index (BMI), previous fractures, bone mineral density (BMD), smoking, alcohol consumption, and underlying diseases. It incorporates epidemiological data from different countries to predict fracture risk, playing a crucial role in osteoporosis prevention and treatment decisions [4].
However, it remains unclear whether ankle fractures occur via the same mechanism as osteoporotic fractures [5]. Osteoporotic fractures typically result from diminished bone mass and structural integrity, often occurring following minimal trauma. These fragility fractures are most commonly observed in the vertebrae, hip, distal radius, and proximal humerus, where the link to low BMD is well established. In contrast, the association between ankle fractures and decreased BMD remains controversial. For example, Stein et al. [6] found no significant difference in BMD between postmenopausal women with and without ankle fractures. Similarly, Therdyothin et al. [7] and Seeley et al. [8] found no meaningful association between BMD and ankle fractures in older adults. However, Biver et al. [9] reported lower BMD and altered trabecular structure in women with ankle fractures. Lee et al. [10] also found that elderly patients with ankle fractures had reduced bone attenuation and more complex fracture patterns. These findings suggest that ankle fractures in the elderly may share features with osteoporotic fractures.
Based on this perspective, we aimed to investigate the potential risk factors for ankle fractures in older adults by analyzing the individual components of the FRAX tool. However, due to the retrospective nature of our study and the unavailability of some required FRAX input variables, we focused on selected FRAX clinical components and comorbidities, rather than the total FRAX score, to assess their potential relevance to ankle fracture risk.
Methods
This study was approved by the Institutional Review Board (IRB) of Chosun University Hospital (IRB No. 2024-10-024). The requirement for informed consent was waived due to the retrospective use of de-identified medical records.
This was a retrospective case-control study. It was described according to the STROBE statement (https://www.strobe-statement.org/).
The study was conducted at Chosun University Hospital and Chonnam National University Hospital in Gwangju, Korea, from 2020 to 2022.
The case group included 84 patients aged 65 years and older who were diagnosed with an ankle fracture using plain radiographs and computed tomography. The control group comprised 150 patients aged ≥65 years with no history of ankle fractures who underwent BMD testing during their hospital visit. Patients were excluded if they had high-energy trauma, multiple trauma, or prior surgeries affecting dual-energy X-ray absorptiometry (DXA) reliability (Fig. 1). Among the initial participants, eight patients with multiple trauma and five patients with high-energy injuries were excluded from the case group. In the control group, nine patients were excluded due to inaccurate DXA measurements related to prior orthopedic surgeries.
The primary outcome variable was the presence of ankle fracture. The primary predictive variables included age, sex, BMI, previous fracture history, BMD, smoking, alcohol consumption, and underlying diseases (hypertension, diabetes mellitus [type 2 diabetes mellitus in all cases], and dementia). Explanatory variables were BMD at the spine, femoral neck, and total hip.
Data were extracted from the electronic medical records, including DXA findings. Ankle fractures can be classified based on anatomical location and injury mechanism, which are important for determining treatment strategies. However, our primary goal was to identify relevant risk factors in elderly patients; therefore, detailed classification was not included in the present analysis.
BMD was measured at the spine, total hip, and femoral neck, with the lowest T-score used for osteoporosis status. Each site was analyzed independently to determine site-specific associations. Hypertension, diabetes mellitus, and dementia were classified based on documented prior diagnoses in the medical records.
Prior osteoporosis treatment history and BMD trends were unavailable due to the retrospective design, representing a limitation. We did not compute the FRAX score but analyzed selected FRAX-related clinical variables along with comorbidities such as diabetes mellitus, hypertension, and dementia as exploratory variables.
Sex was classified as male or female; age was grouped as 65–69, 70–79, or ≥80 years. Smoking and alcohol consumption were defined according to FRAX definitions. Alcohol consumption was classified as ≥3 units/day or <3 units/day based on the current intake recorded in the medical records. Due to limitations in medical records, alcohol intake was categorized as a binary variable, which may not reflect true exposure levels or severity.
BMI was categorized using the WHO Asia-Pacific criteria. BMD was classified per WHO criteria: normal (T≥–1.0), osteopenia (–2.5<T<–1.0), and osteoporosis (T≤–2.5). Previous fracture history was included.
Fall history and functional status were not available from records and could not be included, which is a key limitation.
Selection bias was minimized by applying strict inclusion and exclusion criteria and enrolling consecutive patients from two tertiary hospitals. However, because the controls were patients who underwent DXA during hospital visits, they may not be representative of the general older population. Information bias was reduced through standardized DXA protocols and verification of comorbidity diagnoses using electronic medical records. However, due to the retrospective design, unmeasured confounding factors such as fall history, medication use, and functional status may have influenced the results.
No a priori sample size calculation was performed because of the retrospective design. All eligible patients during the study period were included, and the precision of the estimates is presented as 95% confidence intervals.
This study did not employ case-control matching based on age, sex, or BMI due to a limited sample size (n=234). Because these variables did not differ significantly between the groups, they were not included in the logistic regression model. Analyses were performed using IBM SPSS ver. 27.0 (IBM Corp.). Categorical variables were compared using the chi-square test, and the two-sided Fisher exact test was applied when 20% or more of cells had expected counts less than five. Continuous variables were compared using the independent t-test. Variables that were significant at P<0.05 in the univariate analysis (diabetes mellitus, alcohol consumption, and total hip T-score category) were entered simultaneously (enter method) into a binary logistic regression model, with the absence of diabetes mellitus, alcohol intake <3 units/day, and normal total hip T-score as the reference categories. Model fit was assessed using the omnibus test of model coefficients, Nagelkerke R², and the Hosmer-Lemeshow goodness-of-fit test. P<0.05 was considered statistically significant, and P<0.10 was considered marginally significant.
Results
A total of 234 participants were included in the final analysis, consisting of 84 patients in the ankle fracture group and 150 in the control group. Tables 1 and 2 summarize the baseline and BMD characteristics, respectively. There were no significant differences between the groups in age, sex, or BMI (Table 3).
Univariate analysis (Table 3) revealed no significant associations between ankle fracture occurrence and sex (χ²=0.579, P=0.447), age (χ²=0.783, P=0.676), BMI (χ²=3.114, P=0.638), hypertension (χ²=0.108, P=0.742), dementia (χ²=0.037, P=1.000), smoking (χ²=0.168, P=0.682), or prior fracture history (χ²=1.941, P=0.164). Fisher exact test was used for dementia and BMI category.
However, diabetes mellitus was significantly associated with ankle fracture occurrence (χ²=4.309, P=0.038). Alcohol consumption also showed a significant association (χ²=10.068, P=0.002) (Table 3). In terms of BMD, there were no significant differences in mean lowest T-score, spine T-score, or femoral neck T-score. However, total hip T-score category was significantly associated with ankle fracture (χ²=7.345, P=0.025), with a lower proportion of osteoporosis in the fracture group (3.6% vs. 12.7%), whereas the mean total hip T-score did not differ significantly (P=0.471) (Table 4).
Binary logistic regression analysis was performed including variables with P<0.05 from the univariate analysis (Table 5). Alcohol consumption remained significantly associated with a lower risk of ankle fracture, with alcohol intake <3 units/day as the reference group (odds ratio [OR], 0.196; 95% confidence interval [CI], 0.066–0.589; P=0.004).
Total hip osteoporosis was associated with decreased fracture risk compared with individuals with a normal total hip T-score (≥–1.0), the reference group, whereas osteopenia (–2.5<T-score<–1.0) was not. The odds ratio for osteopenia was 1.258 (95% CI, 0.700–2.263; P=0.443) and for osteoporosis was 0.253 (95% CI, 0.069–0.934; P=0.039); the overall effect of total hip T-score category was also statistically significant (P=0.047).
The association with diabetes mellitus was marginally significant (OR, 1.838; 95% CI, 0.973–3.472; P=0.061), with individuals without diabetes mellitus as the reference group, suggesting a possibly higher risk of ankle fracture in the diabetes group; this finding should therefore be interpreted with caution.
The logistic regression model was statistically significant (χ²=23.473, df=4, P<0.001), with a Nagelkerke R² of 0.131 and an overall classification accuracy of 65.0%. The Hosmer-Lemeshow goodness-of-fit test showed no evidence of poor fit (χ²=7.097, df=5, P=0.214).
Discussion
Among 234 older adults (84 cases, 150 controls), logistic regression identified alcohol consumption and total hip osteoporosis as factors associated with ankle fractures. Alcohol use was associated with a lower fracture risk (OR, 0.196; 95% CI, 0.066–0.589; P=0.004). Total hip osteoporosis was associated with a lower risk of ankle fractures (OR, 0.253; P=0.039), whereas osteopenia was not (OR, 1.258; P=0.443), and diabetes mellitus was marginally significant (P=0.061). These associations should be interpreted with caution, as discussed below.
Moayeri et al. [11] conducted a meta-analysis of 30 studies and found increased fracture risk in patients with type 2 diabetes mellitus, though not at the ankle. Wang et al. [12] later found a significant association between diabetes and ankle fractures. However, few studies have specifically examined this relationship in elderly individuals, supporting its inclusion here.
Du et al. [13] demonstrated that hypertension was associated with increased future fracture risk, and Li et al. [14] confirmed a higher risk of osteoporotic fractures in hypertensive individuals. Still, limited research exists on ankle fractures specifically, particularly in the elderly.
Dementia was included due to its known link to fall risk. Dumurgier and Tzourio [15] found dementia to be the most prevalent neurological condition in older adults. Phelan et al. [16] and Simpkins et al. [17] both showed that fall incidence increases in individuals with Alzheimer’s disease, the most prevalent type of dementia. Liu et al. [18] reported that over three-quarters of ankle fractures resulted from low-energy trauma, highlighting the relevance of fall-related conditions. Based on these considerations, hypertension, diabetes mellitus, and dementia were included as variables in this study.
Interestingly, while several components such as age, sex, BMI, prior fractures, and BMD variables other than total hip T-score were not consistently associated with ankle fracture occurrence, certain findings, including alcohol consumption, total hip T-score, and diabetes mellitus status, were found to be associated with ankle fractures in the elderly, but should be interpreted with caution.
In our study, alcohol consumption was significantly associated with a decreased risk of ankle fractures. Specifically, individuals with alcohol intake ≥3 units/day had a lower risk of ankle fractures than those with <3 units/day (OR, 0.196; 95% CI, 0.066–0.589; P=0.004). This finding is inconsistent with previous research investigating alcohol-related injuries to the foot and ankle. Pirruccio and Farber [19] reported that alcohol use is mechanistically linked to traumatic foot and ankle fractures through increased incidence of falls, particularly in emergency department visits across the United States. Alcohol consumption has been repeatedly associated with increased fall risk in older adults, primarily due to its effects on sensorium, balance, and concentration, which are further exacerbated by age-related physiological decline [19]. In a nationwide longitudinal study by Wang et al. [20], the authors reported that the pattern of association between alcohol consumption and fracture risk varies depending on the anatomical site. Specifically, vertebral fractures demonstrated a J-shaped association with alcohol intake. Mild to moderate drinking was associated with reduced risk, while excessive intake increased the risk. In contrast, limb fractures, including ankle fractures, showed a dose-dependent linear increase in fracture risk with increasing alcohol consumption. Given the observational nature of the study and the presence of multiple confounding factors—including physical frailty, comorbidities, and lifestyle behaviors—the inverse association for alcohol consumption should be interpreted with caution and should not be considered a protective factor. Further prospective studies with more precise measurement of alcohol exposure are needed to clarify this relationship.
Next, osteoporosis, as defined by total hip T-score, was significantly associated with a decreased risk of ankle fractures compared with the normal total hip T-score group. Specifically, the risk of ankle fracture was lower in the osteoporosis group (OR, 0.253; 95% CI, 0.069–0.934) but not in the osteopenia group (OR, 1.258; 95% CI, 0.700–2.263); the overall effect of total hip T-score category was statistically significant (P=0.047). Cummings and Melton [21] reported that, in addition to the presence of low BMD itself, the anatomical site of BMD measurement may influence its predictive value for fracture risk. In particular, BMD at the proximal femur provides better predictive accuracy for overall fracture risk than spinal measurements, likely due to fewer degenerative artifacts and greater measurement reliability at the hip.
A cohort study by Leslie et al. [22] further explored this issue by examining the clinical implications of discordance between total hip and femoral neck T-scores. They suggested that the total hip T-score may better reflect susceptibility to osteoporotic fractures compared to the femoral neck. For example, individuals whose total hip T-score was at least one unit lower than their femoral neck T-score had significantly higher risks of major osteoporotic fractures. They proposed that although hip sites are highly correlated, there are different proportions of cortical and trabecular bone, which may contribute differently to bone strength across the skeleton. Furthermore, the femoral neck is more prone to measurement artifacts due to osteoarthritic changes such as medial buttressing, especially in older individuals. In our study, total hip osteoporosis was associated with a decreased ankle fracture risk, suggesting that the predictive value of total hip BMD for osteoporotic fractures may not directly extend to ankle fractures. Nevertheless, this interpretation must be approached with caution. In a recent systematic review and meta-analysis, So et al. [23] reported that elderly patients with ankle fractures had significantly lower femoral neck BMD compared to those without fractures. However, this relationship remains controversial and is not yet fully established [7]. Also, due to the retrospective design of our study and lack of information, we were unable to account for potential confounding factors such as prior use of osteoporosis medications, adherence to treatment, or longitudinal changes in BMD over time. Additionally, Shevroja et al. [24] argued that DXA-derived T-scores do not account for other important aspects of bone health, such as microarchitecture, bone turnover, or mineralization. They further noted that unmeasured factors, including fall history, balance, muscle strength, and physical function, may also play a significant role in fracture occurrence [24]. Therefore, the role of total hip T-score as an indicator of ankle fracture risk in older adults remains uncertain, which is in line with earlier studies reporting no clear association between BMD and ankle fractures [6-8]. However, further prospective studies incorporating treatment history, bone quality parameters, and functional risk factors are warranted to clarify this association.
Next, diabetes mellitus was associated with a higher risk of ankle fracture, which was marginally significant (P=0.061) but did not reach conventional statistical significance. This finding is in line with prior studies that have consistently reported diabetes mellitus as a risk factor for fractures, attributed to factors such as impaired bone quality, microvascular complications, neuropathy, and increased fall risk [11,12]. Given that the observed association may be influenced by unmeasured or residual confounders, this finding should not be disregarded and warrants further investigation. On the other hand, the inverse associations for alcohol consumption and total hip osteoporosis observed in our study may not reflect a true protective effect, but rather a result of selection bias, confounding, or sample-specific variation such as treatment confounding and healthy user bias.
Several factors may have contributed to these findings. First, the controls were not drawn from the general older population but were patients who underwent DXA during hospital visits; therefore, individuals with osteoporosis or at increased risk of osteoporosis may have been overrepresented among the controls. Second, the number of exposed cases was small (four patients with alcohol intake ≥3 units/day and three with total hip osteoporosis in the fracture group), which may have produced biased estimates [25]. Third, alcohol consumption was classified according to current intake, so former heavy drinkers could not be distinguished from those who had never been heavy drinkers, and the <3 units/day group may have included individuals who had stopped or reduced drinking because of poor health. Fourth, information on osteoporosis treatment, which may affect subsequent fracture risk, was not available. Finally, central DXA measurements may not fully reflect bone quality at the ankle [6,24]. These possibilities could not be verified in the present study and require confirmation in future studies.
We were unable to assess fall history, functional status, or osteoporosis treatment adherence, which are important confounders in FRAX. In addition, the hospital-based selection of controls and the small number of exposed cases in some categories may limit the generalizability of our findings. Alcohol consumption was classified according to current intake (≥3 vs. <3 units/day), so former heavy drinkers could not be distinguished from those who had never been heavy drinkers. Age, sex, and BMI were not included in the multivariable model, and residual confounding cannot be excluded.
In conclusion, our findings suggest that alcohol consumption, a FRAX clinical component, and total hip osteoporosis were inversely associated with ankle fractures in this sample; however, these associations should be interpreted with caution given the hospital-based selection of controls and the small number of exposed cases. Notably, the current FRAX tool estimates the 10-year probability of major osteoporotic fractures based on a composite risk score incorporating clinical variables and femoral neck BMD. Several FRAX components—including prior glucocorticoid use and secondary osteoporosis—could not be assessed in our study. Future prospective studies incorporating longitudinal BMD data, treatment history, and functional risk factors are warranted to clarify these relationships and facilitate the development of an ankle fracture–specific risk prediction model.
-
Author contribution
Conceptualization: MJS, JYL. Data curation: MJS, SWJ, SP. Formal analysis: SWJ, JYL. Methodology: SWJ, SP. Project administration: JYL. Funding acquisition: JYL. Investigation: MJS. Supervision: JYL. Writing-original draft: MJS, SP. Writing-review & editing: MJS, SWJ, JYL. All authors read and approved the final manuscript.
-
Conflict of interests
No potential conflict of interest relevant to this article was reported.
-
Funding
This study was supported by a research fund from Chosun University Hospital (2023).
-
Data availability
Contact the corresponding author for data availability.
-
Acknowledgments
None.
-
Supplementary materials
None.
Article Information
Fig. 1.
Flowchart depicting the participant selection process for inclusion in the study. Participants were selected from two hospitals between January 2020 and December 2022, with exclusion of patients with multiple trauma or high-energy trauma from the case group and those with inaccurate DXA measurements due to prior orthopedic surgery from the control group. BMD, bone mineral density; DXA, dual-energy X-ray absorptiometry.
Table 1.
Demographics of participants (n=234)
Table 2.
Bone mineral density characteristics of participants
Values are presented as mean±standard deviation or number (%).
T-scores were classified according to the World Health Organization criteria: normal (T≥−1.0), osteopenia (−2.5<T<−1.0), and osteoporosis (T≤−2.5). Lowest T-score indicates the lowest value among the spine, femoral neck, and total hip T-scores.
Table 3.
Presence of ankle fractures according to participants' characteristics
| Variable | With ankle fracture | Without ankle fracture | χ² or t | P-value |
|---|---|---|---|---|
| Sex | 0.579 | 0.447 | ||
| Male | 31 (36.9) | 48 (32.0) | ||
| Female | 53 (63.1) | 102 (68.0) | ||
| Age (yr) | 71.64±5.67 | 72.71±5.84 | 1.350 | 0.178 |
| 65–69 | 35 (41.7) | 54 (36.0) | 0.783 | 0.676 |
| 70–79 | 38 (45.2) | 76 (50.7) | ||
| ≥80 | 11 (13.1) | 20 (13.3) | ||
| Underlying diseases | ||||
| Diabetes mellitus | 4.309 | 0.038 | ||
| Negative | 57 (67.9) | 120 (80.0) | ||
| Positive | 27 (32.1) | 30 (20.0) | ||
| Hypertension | 0.108 | 0.742 | ||
| Negative | 39 (46.4) | 73 (48.7) | ||
| Positive | 45 (53.6) | 77 (51.3) | ||
| Dementia | 0.037 | 1.000a) | ||
| Negative | 82 (97.6) | 147 (98.0) | ||
| Positive | 2 (2.4) | 3 (2.0) | ||
| Health behavior | ||||
| Smoking | 0.168 | 0.682 | ||
| Absent | 78 (92.9) | 137 (91.3) | ||
| Present | 6 (7.1) | 13 (8.7) | ||
| Alcohol consumption | 10.068 | 0.002 | ||
| <3 units/day | 80 (95.2) | 120 (80.0) | ||
| ≥3 units/day | 4 (4.8) | 30 (20.0) | ||
| Body mass index (kg/m2) | 24.86±3.13 | 24.97±3.31 | 0.247 | 0.805 |
| Underweight (<18.5) | 0 (0.0) | 4 (2.7) | 3.114 | 0.638a) |
| Normal (18.5–22.9) | 21 (25.0) | 36 (24.0) | ||
| Overweight (23.0–24.9) | 22 (26.2) | 40 (26.7) | ||
| Obesity (25.0–29.9) | 37 (44.0) | 59 (39.3) | ||
| Severe obesity (≥30.0) | 4 (4.8) | 11 (7.3) | ||
| Previous fracture history | 1.941 | 0.164 | ||
| Absent | 69 (82.1) | 133 (88.7) | ||
| Present | 15 (17.9) | 17 (11.3) |
Values are presented as number (%) or mean±standard deviation. P-values were obtained using the chi-square test for categorical variables and the independent t-test for continuous variables.
a)Fisher exact test was used because 20% or more of cells had expected counts less than 5; the χ² value shown is the Pearson chi-square.
Table 4.
Presence of ankle fractures according to participants' bone mineral density
Values are presented as number (%) or mean±standard deviation. P-values were obtained using the chi-square test for categorical variables and the independent t-test for continuous variables. Lowest T-score indicates the lowest value among the spine, femoral neck, and total hip T-scores. T-scores were classified according to the World Health Organization criteria: normal (T≥−1.0), osteopenia (−2.5<T<−1.0), and osteoporosis (T≤−2.5).
Table 5.
Binary logistic regression analysis of factors associated with ankle fractures (n=234)
Reference groups: no diabetes mellitus, alcohol intake <3 units/day, and normal total hip T-score. Model χ²=23.473 (df=4, P<0.001); Cox & Snell R²=0.095; Nagelkerke R²=0.131; Hosmer-Lemeshow test, χ²=7.097 (df=5, P=0.214); classification accuracy, 65.0%. The P-value for total hip T-score (P=0.047) indicates the overall effect of the variable.
B, regression coefficient; SE, standard error; CI, confidence interval.
- 1. Kang HJ, Lee JW, Kwon YM, Kim SJ. Epidemiology of ankle fractures in Korea: a nationwide population-based study. J Korean Med Sci 2022;37:e288.ArticlePubMedPMCPDF
- 2. Scheer RC, Newman JM, Zhou JJ, et al. Ankle fracture epidemiology in the United States: patient-related trends and mechanisms of injury. J Foot Ankle Surg 2020;59:479-83.ArticlePubMed
- 3. Kannus P, Palvanen M, Niemi S, Parkkari J, Jarvinen M. Increasing number and incidence of low-trauma ankle fractures in elderly people: Finnish statistics during 1970-2000 and projections for the future. Bone 2002;31:430-3.ArticlePubMed
- 4. Kanis JA, Johnell O, Oden A, Johansson H, McCloskey E. FRAX and the assessment of fracture probability in men and women from the UK. Osteoporos Int 2008;19:385-97.ArticlePubMedPMCPDF
- 5. Seeley DG, Kelsey J, Jergas M, Nevitt MC. Predictors of ankle and foot fractures in older women. J Bone Miner Res 1996;11:1347-55.ArticlePubMedPDF
- 6. Stein EM, Liu XS, Nickolas TL, et al. Abnormal microarchitecture and stiffness in postmenopausal women with ankle fractures. J Clin Endocrinol Metab 2011;96:2041-8.ArticlePubMedPMC
- 7. Therdyothin A, Phiphopthatsanee N, Wajanavisit W, Woratanarat P, Laohajaroensombat S, Tawonsawatruk T. Is ankle fracture related to low bone mineral density and subsequent fracture? A systematic review. Osteoporos Sarcopenia 2020;6:151-9.ArticlePubMedPMC
- 8. Seeley DG, Browner WS, Nevitt MC, Genant HK, Scott JC, Cummings SR. Which fractures are associated with low appendicular bone mass in elderly women? The Study of Osteoporotic Fractures Research Group. Ann Intern Med 1991;115:837-42.ArticlePubMed
- 9. Biver E, Durosier C, Chevalley T, Herrmann FR, Ferrari S, Rizzoli R. Prior ankle fractures in postmenopausal women are associated with low areal bone mineral density and bone microstructure alterations. Osteoporos Int 2015;26:2147-55.ArticlePubMedPDF
- 10. Lee KM, Chung CY, Kwon SS, et al. Ankle fractures have features of an osteoporotic fracture. Osteoporos Int 2013;24:2819-25.ArticlePubMedPDF
- 11. Moayeri A, Mohamadpour M, Mousavi SF, Shirzadpour E, Mohamadpour S, Amraei M. Fracture risk in patients with type 2 diabetes mellitus and possible risk factors: a systematic review and meta-analysis. Ther Clin Risk Manag 2017;13:455-68.ArticlePubMedPMCPDF
- 12. Wang H, Ba Y, Xing Q, Du JL. Diabetes mellitus and the risk of fractures at specific sites: a meta-analysis. BMJ Open 2019;9:e024067.ArticlePubMedPMC
- 13. Du XP, Zheng ML, Yang XC, Zheng ML. High blood pressure is associated with increased risk of future fracture, but not vice versa. Sci Rep 2024;14:8005.ArticlePubMedPMCPDF
- 14. Li C, Zeng Y, Tao L, et al. Meta-analysis of hypertension and osteoporotic fracture risk in women and men. Osteoporos Int 2017;28:2309-18.ArticlePubMedPDF
- 15. Dumurgier J, Tzourio C. Epidemiology of neurological diseases in older adults. Rev Neurol (Paris) 2020;176:642-8.ArticlePubMed
- 16. Phelan EA, Mahoney JE, Voit JC, Stevens JA. Assessment and management of fall risk in primary care settings. Med Clin North Am 2015;99:281-93.ArticlePubMedPMC
- 17. Simpkins C, Khalili SM, Yang F. Meta-analysis-based comparison of annual fall risk between older adults with Alzheimer's disease and mild cognitive impairment. Adv Geriatr Med Res 2024;6:e240002.PubMedPMC
- 18. Liu S, Zhu Y, Chen W, Wang L, Zhang X, Zhang Y. Demographic and socioeconomic factors influencing the incidence of ankle fractures, a national population-based survey of 512,187 individuals. Sci Rep 2018;8:10443.ArticlePubMedPMCPDF
- 19. Pirruccio K, Farber DC. Rising burden of alcohol-associated foot and ankle fractures presenting to U.S. emergency departments: 2000-2017. Foot Ankle Orthop 2019;4:2473011419S00341.ArticlePMCPDF
- 20. Wang SM, Han KD, Kim NY, et al. Association of alcohol intake and fracture risk in elderly varied by affected bones: a nationwide longitudinal study. Psychiatry Investig 2020;17:1013-20.ArticlePubMedPMCPDF
- 21. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet 2002;359:1761-7.ArticlePubMed
- 22. Leslie WD, Morin SN, Lix LM, et al. Effect of discordant hip bone density on incident fracture risk: a registry-based cohort study. J Bone Miner Res 2022;37:2018-24.PubMed
- 23. So E, Rushing CJ, Simon JE, Goss DA Jr, Prissel MA, Berlet GC. Association between bone mineral density and elderly ankle fractures: a systematic review and meta-analysis. J Foot Ankle Surg 2020;59:1049-57.ArticlePubMed
- 24. Shevroja E, Cafarelli FP, Guglielmi G, Hans D. DXA parameters, trabecular bone score (TBS) and bone mineral density (BMD), in fracture risk prediction in endocrine-mediated secondary osteoporosis. Endocrine 2021;74:20-8.ArticlePubMedPMCPDF
- 25. Greenland S, Mansournia MA, Altman DG. Sparse data bias: a problem hiding in plain sight. BMJ 2016;352:i1981.ArticlePubMed
References
Figure & Data
REFERENCES
Citations
Citations to this article as recorded by 
ePub Link
Cite
XML DownloadRisk factors for ankle fractures in older adults based on clinical components of the Fracture Risk Assessment (FRAX) tool and comorbidities in Korea: a retrospective case-control study
Fig. 1. Flowchart depicting the participant selection process for inclusion in the study. Participants were selected from two hospitals between January 2020 and December 2022, with exclusion of patients with multiple trauma or high-energy trauma from the case group and those with inaccurate DXA measurements due to prior orthopedic surgery from the control group. BMD, bone mineral density; DXA, dual-energy X-ray absorptiometry.
Fig. 1.
Risk factors for ankle fractures in older adults based on clinical components of the Fracture Risk Assessment (FRAX) tool and comorbidities in Korea: a retrospective case-control study
| Variable | Value |
|---|---|
| Sex | |
| Male | 79 (33.8) |
| Female | 155 (66.2) |
| Age (yr) | 72.32±5.79 |
| 65–69 | 89 (38.0) |
| 70–79 | 114 (48.7) |
| ≥80 | 31 (13.2) |
| Underlying diseases | |
| Diabetes mellitus | |
| Negative | 177 (75.6) |
| Positive | 57 (24.4) |
| Hypertension | |
| Negative | 112 (47.9) |
| Positive | 122 (52.1) |
| Dementia | |
| Negative | 229 (97.9) |
| Positive | 5 (2.1) |
| Health behavior | |
| Smoking | |
| Absent | 215 (91.9) |
| Present | 19 (8.1) |
| Alcohol consumption (units/day) | |
| <3 | 200 (85.5) |
| ≥3 | 34 (14.5) |
| Body mass index (kg/m2) | 24.93±3.24 |
| Underweight (<18.5) | 4 (1.7) |
| Normal weight (18.5–22.9) | 57 (24.4) |
| Overweight (23.0–24.9) | 62 (26.5) |
| Obesity (25.0–29.9) | 96 (41.0) |
| Severe obesity (≥30.0) | 15 (6.4) |
| Previous fracture history | |
| Absent | 202 (86.3) |
| Present | 32 (13.7) |
| Variable | Value |
|---|---|
| Lowest T-score | −2.29±1.02 |
| Normal | 27 (11.5) |
| Osteopenia | 101 (43.2) |
| Osteoporosis | 106 (45.3) |
| Spine T-score | −1.75±1.46 |
| Normal | 68 (29.1) |
| Osteopenia | 87 (37.2) |
| Osteoporosis | 79 (33.8) |
| Femoral neck T-score | −1.84±0.95 |
| Normal | 43 (18.4) |
| Osteopenia | 127 (54.3) |
| Osteoporosis | 64 (27.4) |
| Total hip T-score | −1.22±0.97 |
| Normal | 97 (41.5) |
| Osteopenia | 115 (49.1) |
| Osteoporosis | 22 (9.4) |
| Variable | With ankle fracture | Without ankle fracture | χ² or t | P-value |
|---|---|---|---|---|
| Sex | 0.579 | 0.447 | ||
| Male | 31 (36.9) | 48 (32.0) | ||
| Female | 53 (63.1) | 102 (68.0) | ||
| Age (yr) | 71.64±5.67 | 72.71±5.84 | 1.350 | 0.178 |
| 65–69 | 35 (41.7) | 54 (36.0) | 0.783 | 0.676 |
| 70–79 | 38 (45.2) | 76 (50.7) | ||
| ≥80 | 11 (13.1) | 20 (13.3) | ||
| Underlying diseases | ||||
| Diabetes mellitus | 4.309 | 0.038 | ||
| Negative | 57 (67.9) | 120 (80.0) | ||
| Positive | 27 (32.1) | 30 (20.0) | ||
| Hypertension | 0.108 | 0.742 | ||
| Negative | 39 (46.4) | 73 (48.7) | ||
| Positive | 45 (53.6) | 77 (51.3) | ||
| Dementia | 0.037 | 1.000 |
||
| Negative | 82 (97.6) | 147 (98.0) | ||
| Positive | 2 (2.4) | 3 (2.0) | ||
| Health behavior | ||||
| Smoking | 0.168 | 0.682 | ||
| Absent | 78 (92.9) | 137 (91.3) | ||
| Present | 6 (7.1) | 13 (8.7) | ||
| Alcohol consumption | 10.068 | 0.002 | ||
| <3 units/day | 80 (95.2) | 120 (80.0) | ||
| ≥3 units/day | 4 (4.8) | 30 (20.0) | ||
| Body mass index (kg/m2) | 24.86±3.13 | 24.97±3.31 | 0.247 | 0.805 |
| Underweight (<18.5) | 0 (0.0) | 4 (2.7) | 3.114 | 0.638 |
| Normal (18.5–22.9) | 21 (25.0) | 36 (24.0) | ||
| Overweight (23.0–24.9) | 22 (26.2) | 40 (26.7) | ||
| Obesity (25.0–29.9) | 37 (44.0) | 59 (39.3) | ||
| Severe obesity (≥30.0) | 4 (4.8) | 11 (7.3) | ||
| Previous fracture history | 1.941 | 0.164 | ||
| Absent | 69 (82.1) | 133 (88.7) | ||
| Present | 15 (17.9) | 17 (11.3) |
| Variable | With ankle fracture | Without ankle fracture | χ² or t | P-value |
|---|---|---|---|---|
| Lowest T-score | −2.37±0.96 | −2.24±1.05 | 0.943 | 0.347 |
| Normal | 10 (11.9) | 17 (11.3) | 3.150 | 0.207 |
| Osteopenia | 30 (35.7) | 71 (47.3) | ||
| Osteoporosis | 44 (52.4) | 62 (41.3) | ||
| Spine T-score | −1.90±1.41 | −1.66±1.49 | 1.205 | 0.229 |
| Normal | 20 (23.8) | 48 (32.0) | 3.918 | 0.141 |
| Osteopenia | 29 (34.5) | 58 (38.7) | ||
| Osteoporosis | 35 (41.7) | 44 (29.3) | ||
| Femoral neck T-score | −1.79±0.95 | −1.87±0.95 | 0.619 | 0.537 |
| Normal | 20 (23.8) | 23 (15.3) | 3.246 | 0.197 |
| Osteopenia | 40 (47.6) | 87 (58.0) | ||
| Osteoporosis | 24 (28.6) | 40 (26.7) | ||
| Total hip T-score | −1.16±0.94 | −1.26±0.99 | 0.721 | 0.471 |
| Normal | 32 (38.1) | 65 (43.3) | 7.345 | 0.025 |
| Osteopenia | 49 (58.3) | 66 (44.0) | ||
| Osteoporosis | 3 (3.6) | 19 (12.7) |
| Variable | B | SE | P-value | Odds ratio | 95% CI |
|---|---|---|---|---|---|
| Underlying diseases | |||||
| Diabetes mellitus (ref. negative) | 0.609 | 0.325 | 0.061 | 1.838 | 0.973–3.472 |
| Health behavior | |||||
| Alcohol consumption ≥3 units/day (ref. <3 units/day) | −1.627 | 0.560 | 0.004 | 0.196 | 0.066–0.589 |
| Total hip T-score | 0.047 | ||||
| Osteopenia (ref. normal) | 0.230 | 0.300 | 0.443 | 1.258 | 0.700–2.263 |
| Osteoporosis (ref. normal) | −1.374 | 0.666 | 0.039 | 0.253 | 0.069–0.934 |
Table 1. Demographics of participants (n=234)
Values are presented as number (%) or mean±standard deviation. Body mass index was classified according to the World Health Organization Asia-Pacific criteria.
Table 2. Bone mineral density characteristics of participants
Values are presented as mean±standard deviation or number (%). T-scores were classified according to the World Health Organization criteria: normal (T≥−1.0), osteopenia (−2.5<T<−1.0), and osteoporosis (T≤−2.5). Lowest T-score indicates the lowest value among the spine, femoral neck, and total hip T-scores.
Table 3. Presence of ankle fractures according to participants' characteristics
Values are presented as number (%) or mean±standard deviation. P-values were obtained using the chi-square test for categorical variables and the independent t-test for continuous variables. Fisher exact test was used because 20% or more of cells had expected counts less than 5; the χ² value shown is the Pearson chi-square.
Table 4. Presence of ankle fractures according to participants' bone mineral density
Values are presented as number (%) or mean±standard deviation. P-values were obtained using the chi-square test for categorical variables and the independent t-test for continuous variables. Lowest T-score indicates the lowest value among the spine, femoral neck, and total hip T-scores. T-scores were classified according to the World Health Organization criteria: normal (T≥−1.0), osteopenia (−2.5<T<−1.0), and osteoporosis (T≤−2.5).
Table 5. Binary logistic regression analysis of factors associated with ankle fractures (n=234)
Reference groups: no diabetes mellitus, alcohol intake <3 units/day, and normal total hip T-score. Model χ²=23.473 (df=4, P<0.001); Cox & Snell R²=0.095; Nagelkerke R²=0.131; Hosmer-Lemeshow test, χ²=7.097 (df=5, P=0.214); classification accuracy, 65.0%. The P-value for total hip T-score (P=0.047) indicates the overall effect of the variable. B, regression coefficient; SE, standard error; CI, confidence interval.
