Majority of hip fragility fractures among older people can be predicted by a DXA examination: an updated analysis of literature results and empirical Chinese data with a focus on the validation of the newly proposed osteofrailia criterion for men

Introduction

Osteoporosis is a systemic skeletal disease characterised by a reduction in bone mass and qualitative skeletal changes that cause an increase in bone fragility and a higher fracture risk. The clinical significance of osteoporosis lies in the occurrence of fragility fractures (FFx), and the most relevant fracture site is the hip. Because hip FFx typically necessitates hospitalization, data on their incidence is more reliable than data on other types of fractures. According to the criteria set by the 1994 World Health Organization (WHO) Study Group, the T-score is defined as: (BMD_patient − BMD_{young adult mean})/SD_{young adult population,} where BMD is bone mineral density and SD is the standard deviation (1). For a variety of reasons, including differences in X-ray energy generation, bone edge detection algorithms, region of interest placement, and methods of calibration, BMD by dual-energy X-ray absorptiometry (DXA) in g/cm² differs among DXA manufacturers. To avoid the confusion that would result from instrument-specific numerical BMD cutpoint values, the T-score concept was proposed, whereby each patient’s value is compared with a young normative database generated on the same device. Although BMD is recognized as a continuous risk factor for fracture (i.e., no natural fracture threshold exists), operational ranges for the T-score were proposed for epidemiologic purposes. When the femoral neck (FN) is measured in adult Caucasian women, a cutpoint value of patient BMD of 2.5 SD below the young adult mean BMD results in a prevalence of osteoporosis for those aged ≥50 years of about 16.2%, the same as the lifetime risk of hip FFx for Caucasian women (1). This definition is the cornerstone of densitometric osteoporosis (DOP) classification. However, it has been frequently cited that one of the limitations of BMD measurement is ‘the majority of fragility fractures occur at non-osteoporotic BMD (T-scores >−2.5), compromising its role as a screening tool’ (2-7).

We have recently conducted literature analyses and showed that, for older women, the majority of hip FFx occurs among the DOP population (8). For female hip FFx, the notion that the majority of fractures occur at non-osteoporotic BMD (T-scores >−2.5) was partially derived from a few prospective epidemiology studies of older populations, where BMD was measured at baseline, rather than at the timepoint when a hip FFx occurred (9,10). A closer look at these reports suggests it is likely when hip FFx actually occurred, >50% of the study participant patients were with FN DOP (8). A few prospective epidemiology studies of older population also showed that >50% study participants with hip FFx during the follow-up already had DOP at baseline (11-13).

Following the 1994 WHO definition, DOP prevalence among a specific population should be in proportion to its relative osteoporotic fracture risk with Caucasian female data as reference (1,14,15). It is well noted that the total hip (TH) T-score tends to measure higher than FN T-score, then if TH T-score is solely considered (rather than the lower of FN or TH T-scores) and T-score DOP cutpoint of ≤−2.5 is used, then DOP will be under-diagnosed (16). For men, we recently described that older men suffer from hip FFx at FN or TH T-score approximately 0.5–0.6 higher than older women, thus we proposed a new category of low BMD status, osteofrailia, for older Caucasian men with T-score ≤−2.0 (T-score ≤−2.1 for older East Asian men) who are likely to suffer from hip FFx (16). Moreover, we noted that, due to the lower prevalence of FFx, the T-score cutpoint values for East Asians should be adjusted, rather than using the Caucasian values (14,15,17).

Based on the considerations noted above, in this article we present an updated analysis of hip FFx prediction among older people by a DXA measure, using literature results and our own Chinese data. We argue that, if DXA is applied appropriately, the majority of hip FFx among older population can be predicted.

Methods

Estimations with literature data on the sensitivity of hip region DXA to predict short-term hip FFx

The recent literature research results described were re-used in this study (8,16,18). We systematically searched literature reports on DXA T-score results measured at the timepoint of a hip FFx. We aimed to only include studies concerning the contralateral FN/TH DXA BMD measured right after the time of fracture, and we only included reports with data for female patients and male patients described separately. Additional exclusion criteria were (I) articles concerned with patients group-wise under a specific anti-osteoporotic treatment regime; (II) articles concerned with specific types of patients such as those with diabetes mellitus type 2; (III) articles only concerned with hip re-fracture patients; (IV) articles concerned with atypical femur fracture; (V) articles concerned with femoral head subchondral insufficiency fracture; and (VI) study cohorts with fewer than ten cases. For articles from East Asia, we only included studies which used a local or an East Asian gender-specific BMD reference to calculate the T-score. We were able to identify 12 studies in English that reported both women’s and men’s FN and/or TH T-score data separately (19-30). The report of Wilson et al. (29) had been noted to be an outlier [see (16)] and thus excluded. All included Caucasian data were from Europe or Turkey.

The estimation was based on the reported T-score data (mean, number of cases, SD or 95% confidence interval) and assuming the normal distribution of T-score measures. The report of Yeo et al. (30) only described mean values and range and was thus further excluded. The first question was: if a hip region DXA measure were taken shortly before the FFx accident, how many percentages of the hip FFx could be predicted? For this, the reported T-score data at the timepoint of hip FFx were used, and the DOP thresholds are listed in Table 1. The second question was: if a hip region DXA measure were taken 1 year before the FFx accident, how many percentages of the hip FFx could be predicted? For this, an approximate assumption was made that, both FN and TH T-score of the patients would be 0.02 higher than the values at the FFx timepoint value (Appendix 1) (31,32).

Chinese studies on the sensitivity of DXA examination to predict future hip FFx

Osteoporotic fractures in women (MsOS) and in men (MrOS) Hong Kong studies represent the first large-scale prospective cohort studies conducted on bone health in East Asians. At baseline, 2,000 Chinese men and 2,000 Chinese women ≥65 years were recruited from the local communities from August 2001 to March 2003, to determine the relationship between anthropometric, lifestyle, medical, and other factors with BMD measured at the hip and spine. The subjects had a baseline mean age of 72.3 years (range, 65–92 years) for men and 72.5 years (range, 65–98 years) for women. The recruitment criteria were structured so that the study results would represent similarly aged community-dwelling ethnic Chinese men and women in Hong Kong. All subjects were able to walk without assistance, without bilateral hip replacement, and have the potential to survive the duration of primary study for at least 4 years as judged by their pre-existing medical status. Men and women of similar age and from the same community-based population were investigated using the same methodology, thereby enabling a comparison of the results for men and for women. At baseline, BMD at the hip and lumbar spine (LS, L1–L4) was measured by Hologic QDR-4,500 W densitometers (Hologic, Inc., Waltham, MA, USA). Local reference data were used for the T-score calculation (33). Left lateral thoracic and LS radiographs were also obtained.

In addition to the primary studies of four years, in an observational manner, female participants were followed up for a total 8.8±1.5 years, and 69 FFx (fracture age: 82.0±5.9 years) were recorded. Male participants were followed up for a total of 9.9±2.8 years, and 63 hip FFx (mean fracture age: 82.5±5.7 years) were recorded. For these hip FFx patients, their baseline FN T-score, TH T-score, and LS T-score were extracted. As an alternative to DXA image-based vertebral fracture assessment (VFA), their spine radiographs were evaluated for osteoporotic-like vertebral fractural deformity (OLVF) (34,35). OLVF sum score (OLVFss) is a biomarker of spine bone strength. For each vertebra, according to the extended semi-quantitative VFA scheme, a score of 0, −0.5, −1, −1.5, −2, −2.5, and −3 was assigned for no OLVF or OLVF of <20%, 20% to <25%, 25% to <33%, 33% to <40%, 40% to <67%, and ≥67% vertebral height loss, respectively (34). Two adjacent minimal OLVFs were assigned as −0.5, and three adjacent minimal OLVFs were assigned to be −1. The OLVFss is calculated by summing up the scores of vertebrae T3 to L5. For Chinese women, OLVFss ≤−1.5 suggests this subject being osteoporotic (36); for Chinese men, OLVFss ≤−2.5 suggests this subject being osteoporotic (35,37). Various combinations of DXA T-score and OLVFss were used to test their sensitivity and positive prediction value for hip FFx prediction.

Relative relevance of FN osteoporosis threshold and osteofrailia threshold among older Chinese men

For imperfect tests, the selection of a cutpoint value is always a compromise of mutual constrain of sensitivity and specificity. To capture the majority of FFx in males, we proposed osteofrailia category with FN T-score ≤2.1 for East Asians (16). Conventionally, osteoporosis criterion (FN T-score ≤−2.7 for East Asians) has been applied (15). With MrOS Hong Kong follow-up results, we tested the relationship between various baseline FN T-score groupings and the related percentage of hip FFx incidents during follow-up. Moreover, with a random selection of 802 cases with FN T-score ≤0.5 from MrOS Hong Kong study, we tested the relationship between various baseline FN T-score groupings and severity of OLVFss (baseline measures).

Results

Ten published articles with FN and/or TH T-score at the timepoint of a hip FFx were analysed, and the results are shown in Table 2 and Figure 1. If a hip region DXA measure were taken shortly before the hip FFx accident (prediction-1), females’ FN, females’ TH, males’ FN, or males’ TH T-score osteoporosis/osteofrailia cutpoints on average predicted 66.9%, 70.4%, 66.5% and 67.8% of the hip FFx. If the DXA was taken 1 year before the hip FFx accident and assuming the T-scores were all 0.02 higher (prediction-2), the prediction sensitivities were only slightly inferior to prediction-1 results (Table 2).

Figure 1 Hip FFx prediction sensitivity with the data shown in prediction-1 of Table 2, assuming DXA was taken shortly before the hip FFx accident. Data are from Li et al. (19), Zhu et al. (20), Gani et al. (21), Lee et al. (22), Ho et al. (23), Di Monaco et al. (24,25), Schnabel et al. (26), Cesme et al. (27), Vlachos et al. (28). Each dot represents one study, and East Asian data marked with #. Bars: mean values, with each study giving an equal weight (i.e., sample size for each study was not weighted). F, females; M, males; FN, based on femoral neck T-score; TH, based on total hip T-score; FFx, fragility fracture.

For the MsOS and MrOS Hong Kong studies, Table 3 shows a baseline hip region DXA (a combination of FN and TH T-score) predicted >50% of the cases with a follow-up hip FFx. Table 4 shows a combination of baseline hip region DXA, LS DXA and OLVFss predicted 68.1% of the female cases with a follow-up hip FFx, and 63.4% of the male cases with a follow-up hip FFx. For females, LS <−3.7 predicted 23.2% of the follow-up hip FFx with a positive predictive value of 5.5%. For males, LS <−2.5 and −2.2 predicted 27.0% and 30.2% of the follow-up hip FFx with positive predictive values of 8.85%, and 6.86% respectively. Baseline LS T-score alone was inferior to hip region T-scores in prediction hip FFx. Baseline OLVFss was more useful for female patients’ FFx prediction than for male patients’ FFx prediction (Table 4, Figure 2).

Figure 2 MsOS (A) and MrOS (B) Hong Kong studies, prediction of hip FFx during the follow-up period based on a combination of baseline measures of FN T-score, TH T-score, LS T-score, and OLVFss. Only the cases with a follow-up hip FFx are presented. (A) For the X-axis, cases with FN T-score ≤2.7 were firstly selected and presented; after that, if FN T-score was >2.7, the cases with TH T-score ≤2.7 were selected and presented, then LS T-score ≤3.7 were selected and presented if both FN and TH was >2.7. For the remaining cases did not meet these criteria, the FN T-scores were presented. The sample principle as in (A) was applied in (B). Osteoporotic/osteofrailia cases by either FN T-score or TH T-score ≤2.7 (≤−2.1 for males), or LS T-score ≤3.7 (≤−2.5 for males) or OLVFss (≤1.5 for females and −2.5 for males) are labeled as green (based on FN T-score or TH T-score) or blue (based on LS T-score). Non-osteoporotic/osteofrailia cases are labeled as red. Blue dotted lines: demarcation for osteoporosis threshold for women or osteofrailia threshold for men. OLVFss, osteoporotic-like vertebral fractural deformity sum score; FN, femoral neck; LS, lumbar spine; TH, total hip; FFx, fragility fracture; MrOS, osteoporotic fractures in men; MsOS, osteoporotic fractures in women.

For the Chinese men in MrOS Hong Kong study, Figure 3 shows that when baseline FN T-score was higher than −2.1, then hip FFx prevalence was low during the follow-up period. After the baseline FN T-score was lower than −2.1, hip FFx prevalence increased substantially. After the FN T-score was lower than −2.7, hip FFx prevalence increased further but only relatively slightly. Note that FN T-score ≤−2.1 predicted 46.0% (29/63) of male hip FFx cases, while FN T-score ≤−2.7 predicted only 15.9% (10/63) of male hip FFx cases (16). Therefore, the hip FFx prediction sensitivity analysis did not suggest a −2.7 (or −2.5) to be a favorable threshold. Figure 4 shows that, with FN T-score decreased from −2.0 to −3.0, there was not an accelerated increase in OLVF severity, though a lower FN T-score is associated with an overall higher OLVF severity (i.e., lower OLVFss values). However, Figure 4 suggests the possibility of an accelerated increase in OLVF severity after the FN T-score was lower than −3.0, and this was consistent with the data in Figure 3 that FN T-score lower than −3.0 (together with a more advanced age) was associated with a further elevated hip FFx risk.

Figure 3 Hip FFx prevalence (Y-axis) among Chinese men associated with various groupings (A,B) of baseline FN T-score (blue dots). Data are from MrOS Hong Kong study follow-up results for males (n=1,260 cases with FN T-score <−1). X-axis is the mean age for each FN T-score grouping. FFx, fragility fracture; FN, femoral neck; MrOS, osteoporotic fractures in men.

Figure 4 Relationships between various groupings of FN T-score (X-axis) and the severity of OLVFss (Y-axis) among Chinese men. Data are from MrOS Hong Kong study baseline data for males with FN T-score ≤−0.5 (total n=802 cases). (A,B) Mean and 95% confidence interval with scatter plot; (C) a linear fitting is taken. FN, femoral neck; OLVFss, osteoporotic-like vertebral fractural deformity sum score; MrOS, osteoporotic fractures in men.

Discussion

Among all FFx, hip fracture incurs the greatest morbidity, mortality, and costs. Though the prevalence of hip FFx is lower among men than among women, multiple reports described that, once hip FFx occurs, the post-fracture mortality is substantially higher among men than among women (38-41). In addition, some studies suggest that males tend to suffer from hip FFx at a younger age than females [reviewed in (16)]. Thus, it is important to identify hip FFx high risk populations both for females and for males. Other ‘major osteoporotic fractures’ include those of distal forearm and proximal humerus and vertebral. However, traumatological literature shows that distal forearm, proximal humerus and vertebral fractures can also occur in young subjects with normal bone strength and with low energy trauma (35,42-46). For example, Lindau et al. (42) and Brogren et al. (43) described that, among young subjects, a notable portion (approximately 30%) of the distal forearm fractures was due to low energy trauma. Wong (44) and Rose et al. (45) described that, among young subjects, a notable portion (also approximately 30%) of the humeral fractures was due to low energy trauma. It has been well documented that low-energy trauma induced vertebral deformities are common among young subjects with normal bone strength (35,46). Compared with the cases of hip fracture, distal forearm fracture occurs at ‘younger’ age and ‘higher’ BMD (8,47). Compared with the cases with hip fracture, distal forearm fracture and proximal humerus are more likely to be associated with a ‘higher’ energy level (Figure 5). A fall involving a hip fracture is also associated with a larger ‘contact surface’ than a fall involving the distal forearm, and also the hip region has more soft tissues and muscles functioning as a cushion. While it is not necessary that all low-energy trauma among older population are osteoporotic fractures, hip fractures are most likely to be associated with a lower energy level, and low-energy induced hip fracture suggests that the bone strength is indeed much compromised. Thus, our working hypothesis is that in vivo imaging would likely be able to detect such a weakness of the femur.

Figure 5 Compared with cases with hip fracture, distal forearm fracture is more likely associated with a ‘higher’ trauma energy level. (A) A standing person; (B,C) a fall will lead to a distal forearm (wrist) hitting the ground; (D) a fall will lead to a buttock hitting the ground. The net external force equals the change in momentum of a system divided by the time over which it changes. If we assume that the angular velocity is the same for a forward fall (B,C) as for a backward fall (D), since the radius r1 > r2, velocity at the impact with the ground for a forward fall (v1) will be larger than that for a backward fall (v2, v1 > v2) (E). In addition, due to the buffering of muscle and fat at the buttock, the impact time for backward fall (t2) will be larger than that of forward fall (t1, t1 < t2). Thence, the impact force for a forward fall (B', F1 = v1/t1) will be larger than that of backward fall (D', F2 = v2/t2). Reused with permission from (8).

In this study, Figure 1 shows, if a hip region DXA was performed shortly before a hip FFx, then by either FN T-score or TH T-score, approximately 68% of the hip FFx could be predicted. Then, if a combination of FN T-score and TH T-score was applied, we can postulate that approximately >70% of the hip FFx might be predicted (see the related results in Table 3). Table 2 further shows that, if a hip region DXA was performed 1 year prior to the hip FFx incident, a combination of FN T-score and TH T-score might still be able to predict 70% of the hip FFx incidents. Figure 1 also suggests that, with cutpoints of T-scores used in Table 1, the percentages of prediction were broadly comparable for females’ data and males’ data, and for the FN data and TH data, tentatively suggesting the agreement among various T-score cutpoints applied in the current study. There is evidence that hip region quantitative computed tomography (QCT) is able to satisfactorily predict hip FFx only when gender-specific volumetric BMD thresholds are applied (48,49).

In this study, Table 3 shows, for Chinese females, FN T-score alone, a combination of FN and TH T-scores, and a combination of FN, TH, and LS T-scores, can predict 47.8%, 50.7%, and 52.2% hip FFx during the 8-year observation; for Chinese males, FN T-score alone, a combination of FN and TH T-scores, and a combination of FN, TH, and LS T-scores, can predict 46.0%, 54.0%, and 58.5% during the 8-year observation. These results are likely consistent with the results in Table 2 and Figure 1, as results in Table 3 were from a long observation period. As expected, the prediction of hip FFx by only LS T-score was inferior to those of hip region DXA. For male’s results, the results in Table 2 suggest the cutpoint value of LS ≤−2.5 may be favored over LS ≤−2.2, considering the compromise between sensitivity and the positive prediction value. LS T-score of −2.5 in Chinese men would be equivalent to QCT BMD of 68 mg/mL according to our earlier analysis (14,50). Table 4 further shows, a combination of hip region DAX, LS DAX, and OLVF assessment can predict 68.1% of the hip FFx for females and 63.4% of hip FFx for males during the 8-year observation period. Table 3 shows OLVF analysis contributed better to females’ hip FFx prediction than to males’ hip FFx prediction. We have also reported that, in Chinese men, OLVF at baseline did not predict vertebral radiographical FFx during 4-year follow-up (51).

Compared with FN/TH DXA, LS DXA is inferior in predicting hip FFx which is the most important FFx. Thus, LS BMD defined DOP and hip BMD defined DOP have different clinical relevance, with hip BMD defined DOP being more clinically relevant. Currently spine QCT is more commonly conducted due to its conveniency than hip QCT, the prediction power (sensitivity and positive prediction value) of spine QCT for hip FFx shall require further studies.

In this study, the positive predictive value for hip DXA is slightly less than 10% for both Chinese women and men (Table 3). The current guidelines recommend that postmenopausal women and older men should be diagnosed as with osteoporosis if there is 3% or more 10-year risk for hip fracture. Positive predictive value of around 10% for FN T-score ≤2.1 for Chinese men reaches the treatment threshold (52). Note that, the relevance of a positive predictive value is related to intervention regimens. For example, if a drug is effective, safe, and low cost, then a lower positive predictive value will be more acceptable. In the meantime, further studies to increase the sensitivity and specificity of a diagnostic imaging method for predicting hip FFx should be pursued. One simple and practically feasible approach is to improve the BMD reference database and to use a higher quality local BMD reference database (14,15,53). T-score estimation is very sensitive to the value of SD_{young adult population}. It has been noted that many East Asian BMD databases included relatively few participants, particularly in the young adult group, a factor that is critical in determining the statistical accuracy of the SD_{young adult population} (14,15). With T-score definition being: (BMD_patient − BMD_{young adult mean})/SD_{young adult population}, to construct reliable BMD databases it is more important to sample a large number of representative young subjects. Northern Europeans and Mediterranean Europeans may better use region-specific BMD references. The hip fracture incidence rates are the highest in the Scandinavian countries particularly those of Norway, Denmark, Sweden, and Iceland, while lower among Southern Europeans (54,55). Lucas et al. (55) predicted that, the maximum hip fracture incidence rate (per 100,000 subjects) is 1,389.8 for Swedish women and 1,089.7 for Danish women (742.4 for Swedish men, 551.1 for Danish men), 376.0 for Portuguese women and 420.0 for Spanish women (156.9 for Portuguese men, and 195.0 for Spanish men). Among Scandinavian countries, the Finns have lower hip FFx risk. Lucas et al. (55) predicted that, the maximum hip fracture incidence rate was 649.5 for Finn women and 429.8 for Finn men.

Figure 3 shows that, while Chinese men with FN T-score of less than −3.0 were associated with an elevated hip FFx prevalence, there was no advantage in selecting −2.7 cutpoint value (or −2.5) over selecting −2.1 cutpoint, i.e., no advantage in substantially improved specificity at the cost of lower sensitivity. Figure 4 also shows, while a lower FN T-sore was overall associated with lower OLVFss score, subjects with FN T-score of less than −3.0 were associated with accelerated lower OLVFss. Therefore, FN T-score ≤−3.0 might suggest ‘severe’ osteoporosis among Chinese men.

There are many limitations to this study. With the literature analysis, annual loss of 0.02 T-score for hip region DXA measure was an estimation for average older populations (Appendix 1), there might be some BMD ‘fast losers’ among hip FFx patients. However, the results in Tables 3,4 did not disagree with the results in Table 2 at a group level. In the Hong Kong follow-up studies, the proportion of male hip FFx patients relative to female patients was high. It is possible that after the female participant was informed of being osteoporotic with their baseline T-scores, they might have taken some measures to lower future FFx risk, while male participants would not be informed of being osteoporotic when their baseline T-score was between −2.5 (i.e., the conventional osteoporosis cutpoint) and −2.0. If a study participant was informed that he or she was osteoporotic and then this participant took measures to lower future FFx risk, this might lower the incidences of hip FFx during the observation period, thus leading to a lower sensitivity and a lower positive predictive value for the results in Tables 3,4, than it would be if it was a natural experiment without any intervention. However, it was known to us that regular formal medication was not common among our study participants during the 8-year follow-up observation. Another issue is that some participants had died during the observation period. If the participants who would develop FFx had died in a higher proportion than the participants who would not develop FFx during the observation period, then the sensitivity and the positive predictive value of baseline DXA would have been under-estimated. When the Hong Kong study participants were recruited, one of the recruit criteria was that they would survive the duration of primary study for at least 4 years as judged by their pre-existing medical status. In addition, none of the participants recruited had hip fracture history before the enrolment, though the recruit criterion was that they should be ‘without bilateral hip replacement’. There was thus the possibility that our participants were somewhat healthier than the general community population. If so, then our positive predictive value might have been again underestimated. However, as a part of the Asian Osteoporosis Study (AOV), spine radiographs were compared with age-matched Thais and Indonesians in community settings, and it was noted that our Chinese participants had higher prevalence and severity of spine degenerative changes and spine fractural deformities, suggesting our participants might not be overrepresented by participants with much better bone heath (56,57). Positive predictive value is related to the observation period and FFx prevalence in a population. It is well known that, for both females and males, the hip FFx prevalence among Chinese is only half of that those of their Caucasian counterparts (15). If we extrapolate the value in the Hong Kong studies to predict older Caucasians, the positive predictive value would be higher for older Caucasians. Yet another limitation is while we were able to register the hip FFx incidents during the observation period, the trauma energy levels were not always well defined. Johansen et al. (58) estimated that, for subjects aged over 80 years, 4% of hip fractures could be due to high energy trauma. It is possible that if a small portion of our cases had high energy trauma and we were able to exclude these cases, then the DXA diagnostic performance might be further slightly improved.

In conclusion, if DXA scan is regularly performed for at-risk populations, with proper selection of cutpoint T-score values, approximately 70% of the hip FFx incidents can be predicted for older women and men. Our results support regular DXA scan screening for at-risk subjects. OLVF analysis may benefit hip FFx prediction more for females than for males. While osteoporosis remains relevant as an epidemiological concept, there might be no particular benefit in maintaining the osteoporosis threshold for diagnostic DXA-screening for the at-risk older male population.

Acknowledgments

Funding: None.

Footnote

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-24-2568/coif). Y.X.J.W. serves as the Editor-in-Chief of Quantitative Imaging in Medicine and Surgery. J.F.G. serves as an unpaid editorial board member of Quantitative Imaging in Medicine and Surgery. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Kanis JA. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: synopsis of a WHO report. WHO Study Group. Osteoporos Int 1994;4:368-81. [Crossref] [PubMed]
2. Brown SE. What a diagnosis of osteopenia means for you. Available online: https://www.betterbones.com/osteopenia/about-osteopenia/ (accessed 20 Sep 2024).
3. Lespessailles E, Cortet B, Legrand E, Guggenbuhl P, Roux C. Low-trauma fractures without osteoporosis. Osteoporos Int 2017;28:1771-8. [Crossref] [PubMed]
4. Dimai HP. New Horizons: Artificial Intelligence Tools for Managing Osteoporosis. J Clin Endocrinol Metab 2023;108:775-83. [Crossref] [PubMed]
5. Mai HT, Tran TS, Ho-Le TP, Center JR, Eisman JA, Nguyen TV. Two-Thirds of All Fractures Are Not Attributable to Osteoporosis and Advancing Age: Implications for Fracture Prevention. J Clin Endocrinol Metab 2019;104:3514-20. [Crossref] [PubMed]
6. Siris ES, Miller PD, Barrett-Connor E, Faulkner KG, Wehren LE, Abbott TA, Berger ML, Santora AC, Sherwood LM. Identification and fracture outcomes of undiagnosed low bone mineral density in postmenopausal women: results from the National Osteoporosis Risk Assessment. JAMA 2001;286:2815-22. [Crossref] [PubMed]
7. Sornay-Rendu E, Duboeuf F, Chapurlat RD. Postmenopausal women with normal BMD who have fractures have deteriorated bone microarchitecture: A prospective analysis from The OFELY study. Bone 2024;182:117072. [Crossref] [PubMed]
8. Wáng YXJ. For older women, the majority of hip fragility fractures and radiographic vertebral fragility fractures occur among the densitometrically osteoporotic population: a literature analysis. Quant Imaging Med Surg 2024;14:4202-14. [Crossref] [PubMed]
9. Wainwright SA, Marshall LM, Ensrud KE, Cauley JA, Black DM, Hillier TA, Hochberg MC, Vogt MT, Orwoll ES. Hip fracture in women without osteoporosis. J Clin Endocrinol Metab 2005;90:2787-93. [Crossref] [PubMed]
10. Hillier TA, Cauley JA, Rizzo JH, Pedula KL, Ensrud KE, Bauer DC, Lui LY, Vesco KK, Black DM, Donaldson MG, Leblanc ES, Cummings SR. WHO absolute fracture risk models (FRAX): do clinical risk factors improve fracture prediction in older women without osteoporosis? J Bone Miner Res 2011;26:1774-82. [Crossref] [PubMed]
11. Sanders KM, Nicholson GC, Watts JJ, Pasco JA, Henry MJ, Kotowicz MA, Seeman E. Half the burden of fragility fractures in the community occur in women without osteoporosis. When is fracture prevention cost-effective? Bone 2006;38:694-700. [Crossref] [PubMed]
12. Schuit SC, van der Klift M, Weel AE, de Laet CE, Burger H, Seeman E, Hofman A, Uitterlinden AG, van Leeuwen JP, Pols HA. Fracture incidence and association with bone mineral density in elderly men and women: the Rotterdam Study. Bone 2004;34:195-202. [Crossref] [PubMed]
13. Schott AM, Cormier C, Hans D, Favier F, Hausherr E, Dargent-Molina P, Delmas PD, Ribot C, Sebert JL, Breart G, Meunier PJ. How hip and whole-body bone mineral density predict hip fracture in elderly women: the EPIDOS Prospective Study. Osteoporos Int 1998;8:247-54. [Crossref] [PubMed]
14. Wáng YXJ, Xiao BH. Estimations of bone mineral density defined osteoporosis prevalence and cutpoint T-score for defining osteoporosis among older Chinese population: a framework based on relative fragility fracture risks. Quant Imaging Med Surg 2022;12:4346-60. [Crossref] [PubMed]
15. Wáng YXJ, Griffith JF, Blake GM, Diacinti D, Xiao BH, Yu W, Su Y, Jiang Y, Guglielmi G, Guermazi A, Kwok TCY. Revision of the 1994 World Health Organization T-score definition of osteoporosis for use in older East Asian women and men to reconcile it with their lifetime risk of fragility fracture. Skeletal Radiol 2024;53:609-25. [Crossref] [PubMed]
16. Wáng YXJ, Xiao BH, Leung JCS, Griffith JF, Aparisi Gómez MP, Bazzocchi A, Diacinti D, Chan WP, Guermazi A, Kwok TCY. The observation that older men suffer from hip fracture at DXA T-scores higher than older women and a proposal of a new low BMD category, osteofrailia, for predicting fracture risk in older men. Skeletal Radiol 2024; Epub ahead of print. [Crossref]
17. Wáng YXJ, Blake GM, Tang SN, Guermazi A, Griffith JF. Quantitative CT lumbar spine BMD cutpoint value for classifying osteoporosis among older East Asian women should be lower than the value for Caucasians. Skeletal Radiol 2024;53:1473-80. [Crossref] [PubMed]
18. Wáng YXJ. Around the time of a hip fracture, older East Asian female patients tend to measure lower densitometric femoral neck and total hip T-scores than older Caucasian female patients: a literature analysis. Quant Imaging Med Surg 2023;13:2772-9. [Crossref] [PubMed]
19. Li HL, Shen Y, Tan LH, Fu SB, Dai RC, Yuan LQ, Sheng ZF, Xie ZJ, Wu XP, Liao EY, Tang XL, Wu XY. Relationship between bone mineral density and fragility fracture risk: a case-control study in Changsha, China. BMC Musculoskelet Disord 2021;22:728. [Crossref] [PubMed]
20. Zhu TY, Hung VW, Cheung WH, Cheng JC, Qin L, Leung KS. Value of Measuring Bone Microarchitecture in Fracture Discrimination in Older Women with Recent Hip Fracture: A Case-control Study with HR-pQCT. Sci Rep 2016;6:34185. [Crossref] [PubMed]
21. Gani LU, Saripalli KR, Fernandes K, Leong SF, Tsai KT, Tan PT, Chong LR, King TFJ. Bone mineral density and trabecular bone score in elderly type 2 diabetes Southeast Asian patients with severe osteoporotic hip fractures. PLoS One 2020;15:e0241616. [Crossref] [PubMed]
22. Lee KH, Park JW, Kim S, Lee GY, Park SB, Yang DB, Ha YC. Prevalence, Clinical Implication, and Cause of Spine Hip Discordance in Elderly Patients with Fragility Hip Fracture. J Bone Metab 2022;29:51-7. [Crossref] [PubMed]
23. Ho AW, Lee MM, Chan EW, Ng HM, Lee CW, Ng WS, Wong SH. Prevalence of pre-sarcopenia and sarcopenia in Hong Kong Chinese geriatric patients with hip fracture and its correlation with different factors. Hong Kong Med J 2016;22:23-9. [Crossref] [PubMed]
24. Di Monaco M, Castiglioni C, Bardesono F, Milano E, Massazza G. Sarcopenia, osteoporosis and the burden of prevalent vertebral fractures: a cross-sectional study of 350 women with hip fracture. Eur J Phys Rehabil Med 2020;56:184-90. [Crossref] [PubMed]
25. Di Monaco M, Castiglioni C, Milano E, Massazza G. Is there a definition of low lean mass that captures the associated low bone mineral density? A cross-sectional study of 80 men with hip fracture. Aging Clin Exp Res 2018;30:1429-35. [Crossref] [PubMed]
26. Schnabel M, Eser G, Ziller V, Mann D, Mann E, Hadji P. Bone mineral density in postmenopausal women with proximal femoral fractures--comparative study between quantitative ultrasonometry and gold standard DXA. Zentralbl Chir 2005;130:469-75. [Crossref] [PubMed]
27. Cesme F, Esmaeilzadeh S, Oral A. Discriminative ability of calcaneal quantitative ultrasound compared with dual-energy X-ray absorptiometry in men with hip or distal forearm fractures. Acta Orthop Traumatol Turc 2016;50:548-53. [Crossref] [PubMed]
28. Vlachos C, Ampadiotaki MM, Papagrigorakis E, Galanis A, Patilas C, Sakellariou E, Rodis G, Vasiliadis E, Kontogeorgakos VA, Pneumaticos S, Vlamis J. Is Regional Bone Mineral Density the Differentiating Factor Between Femoral Neck and Femoral Trochanteric Fractures? Cureus 2024;16:e53003. [Crossref] [PubMed]
29. Wilson J, Bonner TJ, Head M, Fordham J, Brealey S, Rangan A. Variation in bone mineral density by anatomical site in patients with proximal humeral fractures. J Bone Joint Surg Br 2009;91:772-5. [Crossref] [PubMed]
30. Yeo AK, Ahrberg AB, Theopold JD, Ewens S, Borte G, Josten C, Fakler JK. Are radiographic indices reliable indicators for quantitative bone mineral density and vitamin D status after femoral neck fractures? A retrospective study in 112 elderly patients. Patient Saf Surg 2015;9:39. [Crossref] [PubMed]
31. Morin SN, Berger C, Liu W, Prior JC, Cheung AM, Hanley DA, Boyd SK, Wong AKO, Papaioannou A, Rahme E, Goltzman D. Differences in fracture prevalence and in bone mineral density between Chinese and White Canadians: the Canadian Multicentre Osteoporosis Study (CaMos). Arch Osteoporos 2020;15:147. [Crossref] [PubMed]
32. Lloyd JT, Alley DE, Hochberg MC, Waldstein SR, Harris TB, Kritchevsky SB, Schwartz AV, Strotmeyer ES, Womack C, Orwig DL. Changes in bone mineral density over time by body mass index in the health ABC study. Osteoporos Int 2016;27:2109-16. [Crossref] [PubMed]
33. Lynn HS, Lau EM, Au B, Leung PC. Bone mineral density reference norms for Hong Kong Chinese. Osteoporos Int 2005;16:1663-8. [Crossref] [PubMed]
34. Wáng YXJ, Diacinti D, Aparisi Gómez MP, Santiago FR, Becce F, Tagliafico AS, Prakash M, Isaac A, Dalili D, Griffith JF, Guglielmi G, Bazzocchi A. Radiological diagnosis of prevalent osteoporotic vertebral fracture on radiographs: an interim consensus from a group of experts of the ESSR osteoporosis and metabolism subcommittee. Skeletal Radiol 2024;53:2563-74. [Crossref] [PubMed]
35. Wáng YXJ. Radiographic Diagnosis of Osteoporotic Vertebral Fracture in Older Women and Men Based on Statistical Probability. Semin Musculoskelet Radiol 2024;28:628-40. [Crossref] [PubMed]
36. Wáng YXJ, Diacinti D, Leung JCS, Iannacone A, Kripa E, Kwok TCY, Diacinti D. Conversion of osteoporotic vertebral fracture severity score to osteoporosis T-score equivalent status: a framework and a comparative study of Hong Kong Chinese and Rome Caucasian older women. Arch Osteoporos 2022;18:1. [Crossref] [PubMed]
37. Wáng YXJ, Leung JCS, Lam PMS, Kwok TCY. Conversion of osteoporotic-like vertebral fracture severity score to osteoporosis T-score equivalent status: A framework study for older Chinese men. Osteoporos Sarcopenia 2023;9:14-21. [Crossref] [PubMed]
38. Sullivan KJ, Husak LE, Altebarmakian M, Brox WT. Demographic factors in hip fracture incidence and mortality rates in California, 2000-2011. J Orthop Surg Res 2016;11:4. [Crossref] [PubMed]
39. Kannegaard PN, van der Mark S, Eiken P, Abrahamsen B. Excess mortality in men compared with women following a hip fracture. National analysis of comedications, comorbidity and survival. Age Ageing 2010;39:203-9. [Crossref] [PubMed]
40. Man LP, Ho AW, Wong SH. Excess mortality for operated geriatric hip fracture in Hong Kong. Hong Kong Med J 2016;22:6-10. [Crossref] [PubMed]
41. Wang PW, Yao XD, Zhuang HF, Li YZ, Xu H, Lin JK, Liu WG. Mortality and Related Risk Factors of Fragile Hip Fracture. Orthop Surg 2022;14:2462-9. [Crossref] [PubMed]
42. Lindau TR, Aspenberg P, Arner M, Redlundh-Johnell I, Hagberg L. Fractures of the distal forearm in young adults. An epidemiologic description of 341 patients. Acta Orthop Scand 1999;70:124-8. [Crossref] [PubMed]
43. Brogren E, Petranek M, Atroshi I. Incidence and characteristics of distal radius fractures in a southern Swedish region. BMC Musculoskelet Disord 2007;8:48. [Crossref] [PubMed]
44. Wong PC. An epidemiological study of humeral fractures in Singapore. Med J Malaya 1966;20:300-5.
45. Rose SH, Melton LJ 3rd, Morrey BF, Ilstrup DM, Riggs BL. Epidemiologic features of humeral fractures. Clin Orthop Relat Res 1982;24-30.
46. Ma JB, Wáng YXJ. Chest radiograph prevalence of vertebral deformity among young and middle-aged population of mixed city dwellers and rural residents. J Thorac Dis 2022;14:4685-98. [Crossref] [PubMed]
47. Boschitsch EP, Durchschlag E, Dimai HP. Age-related prevalence of osteoporosis and fragility fractures: real-world data from an Austrian Menopause and Osteoporosis Clinic. Climacteric 2017;20:157-63. [Crossref] [PubMed]
48. Xu Y, Zhu Y, Lin W, Lee DH, Yang F, Fan Y. Quantitative computed tomography analysis of proximal femur bone mineral density and its relation to hip fracture risk. Quant Imaging Med Surg 2024; Epub ahead of print. [Crossref]
49. Wáng YXJ. Osteopenic fractures, the importance of applying gender-specific bone mineral density thresholds in identifying hip fracture at-risk populations, and comments on the article of Xu et al. Quant Imaging Med Surg 2024; [Crossref]
50. Wáng YXJ, Chan WP, Yu W, Guermazi A, Griffith JF. Quantitative CT lumbar spine BMD cutpoint value for classifying osteoporosis among older Chinese men can be the same as that of older Chinese women, both much lower than the value for Caucasians. Skeletal Radiol 2024; Epub ahead of print. [Crossref]
51. Wáng YXJ, Che-Nordin N, Deng M, Griffith JF, Leung JCS, Kwok AWL, Leung PC, Kwok TCY. Elderly males with or without existing osteoporotic vertebral fracture have much lower future vertebral fracture risk than elderly females: the MrOS (Hong Kong) year-4 follow-up spine radiograph study. Osteoporos Int 2019;30:2505-14.
52. Tosteson AN, Melton LJ 3rd, Dawson-Hughes B, Baim S, Favus MJ, Khosla S, Lindsay RLNational Osteoporosis Foundation Guide Committee. Cost-effective osteoporosis treatment thresholds: the United States perspective. Osteoporos Int 2008;19:437-47. [Crossref] [PubMed]
53. Wáng YXJ, Blake GM, Xiao BH, Guglielmi G, Su Y, Jiang Y, Guermazi A, Kwok TCY, Griffith JF. East Asians' T-scores for the diagnosis of osteoporosis should be calculated using ethnicity- and gender-specific BMD reference ranges: justifications. Skeletal Radiol 2024;53:409-17. [Crossref] [PubMed]
54. Cheng SY, Levy AR, Lefaivre KA, Guy P, Kuramoto L, Sobolev B. Geographic trends in incidence of hip fractures: a comprehensive literature review. Osteoporos Int 2011;22:2575-86. [Crossref] [PubMed]
55. Lucas R, Martins A, Severo M, Silva P, Monjardino T, Gaio AR, Cooper C, Barros H. Quantitative modelling of hip fracture trends in 14 European countries: testing variations of a shared reversal over time. Sci Rep 2017;7:3754. [Crossref] [PubMed]
56. Tang SN, Li CY, Leung JCS, Kwok AWL, Kwok TCY, Wáng YXJ. Fewer spine degenerations among Southeast Asians than among Southern Chinese suggests populations from a warmer climate have inherently better spine health. Quant Imaging Med Surg 2024;14:6922-33. [Crossref] [PubMed]
57. Wáng YXJ, Tang SN, Leung JCS, Li CY, Kwok AWL, Kwok TCY. Lower osteoporotic-like vertebral fracture (OLVF) prevalence and severity among older Thais and Indonesians than among older Chinese. Quant Imaging Med Surg 2024; [Crossref]
58. Johansen A, Harding K, Evans R, Stone M. Trauma in elderly people: what proportion of fractures are a consequence of bone fragility? Arch Gerontol Geriatr 1999;29:215-21. [Crossref] [PubMed]