Conversion of measured bone mineral density T-scores of Chinese women to equivalent Caucasian women’s T-score values

The prevalence of densitometric osteoporosis (DOP) in a specific population should be in proportion to its relative risk of fragility fracture (FF), using Caucasian data as a reference (1,2). While the cutpoint T-score value of −2.5 for defining DOP based on dual-energy X-ray absorptiometry (DXA) measurements of hip and spine bone mineral density (BMD) was developed using Caucasian female data, the application of the same cutpoint value of −2.5 to Caucasian men and Black Americans appears reasonable (3). However, the application of the same cutpoint value of −2.5 to East Asian populations leads to the current dilemma that “the prevalence of DOP for East Asians is high (even when an East Asian BMD reference range is applied), but the osteoporotic fracture incidence is low” (4-7). We have proposed that directly measured BMD T-scores of Chinese women should be converted (or ‘normalised’) to equivalent Caucasian women’s T-score values (8). In this letter we propose provisional formulas for such a conversion.

It is well noted that the use of a Caucasian BMD reference database in East Asians will lead to systematically lower T-scores and hence increase the prevalence of DOP (6,9). The use of Asian-specific BMD reference data was recommended by the International Society for Clinical Densitometry Asia-Pacific Region 2010 Consensus (10) and recently further argued for by us (6). Thus, all the discussions in this letter deal with results derived using T-scores calculated with Chinese (or East Asian) local BMD reference data.

We analysed multiple published BMD databases for Caucasian and Chinese populations and conducted statistical modelling to find more suitable threshold T-scores for older Chinese (5). We used a T-score threshold of ≤−2.5 and its equivalent BMD cutpoints to estimate DOP prevalence for Caucasians assuming a Gaussian distribution; then, assuming that, for consistency with data on fracture incidence, the prevalence of DOP amongst Chinese is half that of Caucasians (11-25), data from BMD databases for Chinese were analysed to estimate revised BMD thresholds and their corresponding T-scores consistent with the reduced prevalence (5). For most Chinese female BMD reference databases, our proposed osteoporosis cutpoint T-score_neck (femoral neck T-score) differs by −0.2 to −0.25 from the conventional value of −2.5 (5). According to our estimation for Chinese women, for the local BMD reference database of Lynn et al. (26), we found osteoporosis cutpoint values for T-score_neck, T-score_hip (total hip T-score), and T-score_spine (lumbar spine T-score) of −2.7, −2.6, and −3.7, respectively (5).

In addition to the analysis of multiple Caucasian BMD reference databases from Europe, North America, and Australia, we also evaluated multiple databases using different DXA manufacturers’ scanners from East Asia and found similar trends requiring the threshold T-score_neck, T-score_hip, and T-score_spine for East Asians to be adjusted to a lower value than the conventional cutpoint of ≤−2.5 (5). For the Japanese female BMD reference database of Iki et al. (27), we found the osteoporosis cutpoint values for T-score_neck, T-score_hip, and T-score_spine to be −2.75, −3.0, and −3.9 respectively (5). Note that, we recommend a larger downward adjustment for T-score_spine than for T-score_neck and T-score_hip (T-score_neck and T-score_hip are highly correlated). This is consistent with the observation that, following natural aging, T-score_spine and T-score_neck (and T-score_hip) decline with a similar rate among Caucasian women (2,24,28-30), while T-score_spine declines at a much faster rate than T-score_neck (and T-score_hip) among East Asian women (8,24,27,30). The T-score_neck is adjusted downward by only −0.2 to −0.25. For individual patients, the clinical implication of such a T-score_neck adjustment is not substantial; however, the impact on epidemiological studies will not be trivial.

Our proposal to convert raw (i.e., directly measured, with a Chinese women’s BMD reference database) T-score_spine to normalised T-score_spine is illustrated in Figure 1. Figure 1A shows our empirical data based on Hong Kong Chinese and Italian Caucasians. Spine radiograph osteoporotic-like vertebral fracture (OLVF) was considered as a relevant surrogate endpoint. For each vertebra, a score of 0, −0.5, −1, −1.5, −2, −2.5, and −3 is assigned for no OLVF or an OLVF of <1/5, ≥1/5–1/4, ≥1/4–1/3, ≥1/3–2/5, ≥2/5–2/3, and ≥2/3 vertebral height loss, respectively. The OLVF sum score (OLVFss) was calculated by summing up the scores for vertebrae T1 to L5. Counting each study case, the OLVFss, T-score_neck and T-score_spine were all independently ranked from the smallest value to the largest [for details, see (31,32)]. In this way, it was noted that, when OLVFss =−2.5, T-score_neck was −2.60 for Italians and −2.77 for Chinese, and T-score_spine was −2.44 for Italians and −3.75 for Chinese. These data support our statistical modeling results (5). Figure 1B illustrates that, following natural aging, the decline of the population mean raw T-score_spine of Chinese women is faster than that of US Caucasian women [data from Wu et al. (30)]. Plotting the data from Figure 1A and Figure 1B together shows good agreement (Figure 1C), supporting the reliability of the data. Data from Figure 1A and Figure 1B, together with some statistical modeling results around a raw T-score_spine =−2.5 reported by Wáng and Xiao (5) for Hong Kong Chinese women, were used to fit a quadratic relationship between the raw T-score_spine and the normalised T-score_spine. In addition, we enforced when raw T-score_spine =−3.7 then normalized T-score_spine =−2.5 (Figure 1D), and final conversion formula for T-score_spine is:

NormalizedspineT_score=−2.013+3.9438+0.7616*(rawspineT_score)0.3808[1]

Figure 1 Conversion of raw LS T-score to normalised T-score for Chinese women. (A) For each OLVFss value, the corresponding raw LS T-scores of Hong Kong Chinese women and Italian Caucasian women. Data from Wáng et al. (31,32). (B) Decline of the population mean raw LS T-score of Chinese women and US Caucasian women during natural aging. Data from Wu et al. (30). (C) Relationship between raw LS T-scores and normalised LS T-scores (i.e., values equivalent to Caucasian T-scores) for Chinese women. Data based on those of (A) (red dots) and (B) (yellow dots) and also the statistical modeling of Wáng and Xiao (5) for Hong Kong Chinese women (green dots). (D) All the data in (C) are used to fit a quadratic relationship between the raw LS T-scores and normalised LS T-scores. In addition, it is enforced that when raw LS T-score =−3.7 then normalized LS T-score =−2.5. OLVFss, osteoporotic-like vertebral fracture sum score; LS, lumbar spine; IL, Italian; HK, Hong Kong.

To further support our argument that the osteoporosis cutpoint T-score_neck value among East Asian populations should be lower than the conventional value of ≤−2.5, we conducted a literature analysis (33). For hip FF patients, there is a trend for East Asian women to have a lower T-score_neck and T-score_hip than Caucasians. If we assume that hip FF in East Asians and Caucasians is equally relevant as a clinical endpoint for osteoporosis, then the results of this literature analysis suggest that a ‘lower’ T-score in East Asian women corresponds to a ‘higher’ T-score in Caucasian women, supporting our argument that to achieve the equivalent of a Caucasian T-score of −2.5, the corresponding T-score for East Asian women should be lower. Our proposal to convert raw (directly measured, with a Chinese women’s BMD reference database) T-score_neck to normalised T-score_neck is shown in Figure 2. Figure 2A shows the OLVFss values and the corresponding raw T-score_neck of Hong Kong Chinese women and Italian Caucasian women (31,32). Figure 2D shows the relationship between of raw T-score_neck and normalised T-score (i.e., values equivalent to Caucasian T-score) for Chinese women based on the data in Figure 2A, as well as some statistical modeling results around a raw T-score_neck of −2.0 to −2.5 of Wang and Xiao (5) for Hong Kong Chinese women. These data were then used to fit a quadratic relationship between the raw T-score_neck and the normalised T-score_neck. In addition, we enforced when raw T-score =−2.7 then normalized T-score =−2.5 (Figure 1D), and the final conversion formula for T-score_neck is:

NormalizedneckT_score=−1.1040+1.2317−0.0736*(rawneckT_score)−0.0368[2]

Figure 2 Conversion of raw femoral neck T-score to normalised T-score for Chinese women. (A) For each OLVFss value, the figure shows the corresponding raw femoral neck T-scores for Hong Kong Chinese women and Italian Caucasian women. Data from Wáng et al. (31,32). (B) A quadratic fit to the data in (A). Whether the data point marked with an arrow is reliable is uncertain. (C) If the latter data point is removed for curve fitting, the fitted curve is almost unchanged. Thus, this data point is considered reasonable. (D) Relationship between raw femoral neck T-scores and normalised T-scores (i.e., values equivalent to a Caucasian T-score) for Chinese women. Data based on those in (A) (red dots) and also the statistical modeling of Wáng and Xiao (5) for Hong Kong Chinese women (green dots). The black line indicates a theoretical one-to-one relationship between the raw T-scores and the normalised T-scores (i.e., they are the same). (E) All the data in (D) are used to fit a quadratic relationship between the raw T-scores and the normalised T-scores. In addition, it is enforced that when raw T-score =−2.7 then normalized T-score =−2.5. Note the focus of the fitting is to ensure its reliability for values of −2.0 to −3.0. OLVFss, osteoporotic-like vertebral fracture sum score; IL, Italian; HK, Hong Kong.

It is well-known that T-score_neck and T-score_hip are highly correlated (27,31,34,35). Using OLVFss as a surrogate endpoint, for our Hong Kong older women’s data, for each OLVFss the corresponding T-score_neck and T-score_hip are almost the same (Figure 3A). The good correlation of T-score_neck and T-score_hip in the Japanese data of Iki et al. (27) is also shown in Figure 3B. As a result, we believe that the relationship between the Chinese women’s T-score_hip and the Caucasian women’s T-score_hip follows the same pattern as the relationship between the Chinese women’s T-score_neck and the Caucasian T-score_neck. Since we recommended that a Chinese women’s T-score_hip of −2.6 is equivalent to a Caucasian women’s T-score_hip of −2.5 (5), our proposed conversion formula for Chinese women’s T-score_hip is:

NormalizedhipT_score=−1.1040+1.2390−0.0736*(rawhipT_score)−0.0368[3]

Figure 3 Femoral neck T-score and the total hip neck T-score are highly correlated. (A) For each OLVFss value, the corresponding femoral neck T-score and total hip neck T-score of Hong Kong Chinese women are almost the same. Data from Wáng et al. (31). (B) During natural aging, the population mean femoral neck and total hip neck T-scores decrease with a very similar trend. Data from Iki et al. (27). OLVFss, osteoporotic-like vertebral fracture sum score; HK, Hong Kong.

According to the 1994 World Health Organization (WHO) document (1), the definition of osteopenia (low bone mass) is a T-score between −1 and −2.5 below SD_{youngadultpopulation} (standard deviation of BMD of young adult population). However, there is no biological or epidemiological rationale for the threshold of −1 (36). The original intention of the WHO was to choose a threshold that would make osteopenia uncommon at the time of menopause. Actually, the prevalence of osteopenia is higher and can be more than 50% in postmenopausal women over the age of 50 years (5). This makes osteopenia a less relevant diagnosis for patients. In our modelling (5), we assumed osteopenia prevalence among Chinese women is also half that of Caucasian women. Based on the available BMD reference data, we recommend that, for Chinese women, osteopenia has a raw T-score range of −2.7 to −3.7, −2.0 to −2.7, and −1.8 to −2.6 for T-score_spine, T-score_neck, and T-score_hip, respectively.

Strictly speaking, different East Asian BMD reference databases could be associated with slightly different DOP cutpoint T-scores (5). However, we anticipate our conversion formulas will be broadly applicable to all East Asian female populations, and the application of these conversion formulas will not result in any important statistical errors.

Our suggested normalized DOP cutpoint T-scores are in line with the original WHO T-score definition of osteoporosis and will allow a more meaningful international comparison of disease burden and epidemiological studies.

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-23-1090/coif). YXJW serves as the Editor-in-Chief 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. World Health Organization. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: report of a WHO study group [meeting held in Rome from 22 to 25 June 1992]. 1994. Available online: https://apps.who.int/iris/handle/10665/39142
2. Wáng YXJ. Estimation of osteoporosis prevalence among a population is reasonable only after the concerned reference bone mineral density database and cutpoint T-score have been validated. Osteoporos Int 2023;34:417-8. [Crossref] [PubMed]
3. Wright NC, Looker AC, Saag KG, Curtis JR, Delzell ES, Randall S, Dawson-Hughes B. The recent prevalence of osteoporosis and low bone mass in the United States based on bone mineral density at the femoral neck or lumbar spine. J Bone Miner Res 2014;29:2520-6. [Crossref] [PubMed]
4. Noel SE, Santos MP, Wright NC. Racial and Ethnic Disparities in Bone Health and Outcomes in the United States. J Bone Miner Res 2021;36:1881-905. [Crossref] [PubMed]
5. 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]
6. 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 2023; Epub ahead of print. [Crossref]
7. Wáng YXJ, Xiao BH, Su Y, Leung JCS, Lam PMS, Kwok TCY. Fine-tuning the cutpoint T-score as an epidemiological index with high specificity for osteoporosis: methodological considerations for the Chinese population. Quant Imaging Med Surg 2022;12:882-5. [Crossref] [PubMed]
8. Wáng YXJ, Guglielmi G, Guermazi A, Kwok TCY, Griffith JF. Much lower prevalence and severity of spine degenerative changes among older Chinese women than among older Caucasian women and its implication for the interpretation of lumbar spine BMD T-score for Chinese women. Skeletal Radiol 2023; Epub ahead of print. [Crossref]
9. Lo JC, Kim S, Chandra M, Ettinger B. Applying ethnic-specific bone mineral density T-scores to Chinese women in the USA. Osteoporos Int 2016;27:3477-84. [Crossref] [PubMed]
10. Kung AW, Wu CH, Itabashi A, Lee JK, Park HM, Zhao Y, Chan WP, Kendler DL, Leib ES, Lewiecki EM, Bilezikian JP, Baim SAsia Pacific Panel of ISCD. International Society for Clinical Densitometry official positions: Asia-Pacific Region consensus. J Clin Densitom 2010;13:346-51. [Crossref] [PubMed]
11. Wáng YXJ. Fragility fracture prevalence among elderly Chinese is no more than half of that of elderly Caucasians. Quant Imaging Med Surg 2022;12:874-81. [Crossref] [PubMed]
12. Ho SC, Bacon WE, Harris T, Looker A, Maggi S. Hip fracture rates in Hong Kong and the United States, 1988 through 1989. Am J Public Health 1993;83:694-7. [Crossref] [PubMed]
13. Xu L, Lu A, Zhao X, Chen X, Cummings SR. Very low rates of hip fracture in Beijing, People's Republic of China the Beijing Osteoporosis Project. Am J Epidemiol 1996;144:901-7. [Crossref] [PubMed]
14. Fang J, Freeman R, Jeganathan R, Alderman MH. Variations in hip fracture hospitalization rates among different race/ethnicity groups in New York City. Ethn Dis 2004;14:280-4.
15. Zingmond DS, Melton LJ 3rd, Silverman SL. Increasing hip fracture incidence in California Hispanics, 1983 to 2000. Osteoporos Int 2004;15:603-10. [Crossref] [PubMed]
16. Barrett-Connor E, Siris ES, Wehren LE, Miller PD, Abbott TA, Berger ML, Santora AC, Sherwood LM. Osteoporosis and fracture risk in women of different ethnic groups. J Bone Miner Res 2005;20:185-94. [Crossref] [PubMed]
17. Lofthus CM, Frihagen F, Meyer HE, Nordsletten L, Melhuus K, Falch JA. Epidemiology of distal forearm fractures in Oslo, Norway. Osteoporos Int 2008;19:781-6. [Crossref] [PubMed]
18. Khandelwal S, Chandra M, Lo JC. Clinical characteristics, bone mineral density and non-vertebral osteoporotic fracture outcomes among post-menopausal U.S. South Asian Women. Bone 2012;51:1025-8. [Crossref] [PubMed]
19. Wright NC, Saag KG, Curtis JR, Smith WK, Kilgore ML, Morrisey MA, Yun H, Zhang J, Delzell ES. Recent trends in hip fracture rates by race/ethnicity among older US adults. J Bone Miner Res 2012;27:2325-32. [Crossref] [PubMed]
20. Lo JC, Zheng P, Grimsrud CD, Chandra M, Ettinger B, Budayr A, Lau G, Baur MM, Hui RL, Neugebauer R. Racial/ethnic differences in hip and diaphyseal femur fractures. Osteoporos Int 2014;25:2313-8. [Crossref] [PubMed]
21. Leslie WD, Morin SN, Lix LM, McCloskey EV, Johansson H, Harvey NC, Kanis JA. Fracture prediction from FRAX for Canadian ethnic groups: a registry-based cohort study. Osteoporos Int 2021;32:113-22. [Crossref] [PubMed]
22. Wáng YXJ, Diacinti D, Leung JCS, Iannacone A, Kripa E, Kwok TCY, Diacinti D. Much lower prevalence and severity of radiographic osteoporotic vertebral fracture in elderly Hong Kong Chinese women than in age-matched Rome Caucasian women: a cross-sectional study. Arch Osteoporos 2021;16:174. [Crossref] [PubMed]
23. Wáng YXJ, Deng M, Griffith JF, Kwok AWL, Leung JCS, Lam PMS, Yu BWM, Leung PC, Kwok TCY. 'Healthier Chinese spine': an update of osteoporotic fractures in men (MrOS) and in women (MsOS) Hong Kong spine radiograph studies. Quant Imaging Med Surg 2022;12:2090-105. [Crossref] [PubMed]
24. Wáng YXJ. The definition of spine bone mineral density (BMD)-classified osteoporosis and the much inflated prevalence of spine osteoporosis in older Chinese women when using the conventional cutpoint T-score of -2.5. Ann Transl Med 2022;10:1421. [Crossref] [PubMed]
25. Morin SN, Berger C, Liu W, Prior JC, Cheung AM, Hanley DA, Boyd SK, Wong AKO, Papaioannou A, Rahme E, Goltzman D. CaMos Research Group. 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]
26. 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]
27. Iki M, Kagamimori S, Kagawa Y, Matsuzaki T, Yoneshima H, Marumo F. Bone mineral density of the spine, hip and distal forearm in representative samples of the Japanese female population: Japanese Population-Based Osteoporosis (JPOS) Study. Osteoporos Int 2001;12:529-37. [Crossref] [PubMed]
28. Mulder JE, Michaeli D, Flaster ER, Siris E. Comparison of bone mineral density of the phalanges, lumbar spine, hip, and forearm for the assessment of osteoporosis in postmenopausal women. J Clin Densitom 2000;3:373-81. [Crossref] [PubMed]
29. Looker AC, Borrud LG, Hughes JP, Fan B, Shepherd JA, Melton LJ 3rd. Lumbar spine and proximal femur bone mineral density, bone mineral content, and bone area: United States, 2005-2008. Vital Health Stat 11 2012;1-132.
30. Wu XP, Liao EY, Huang G, Dai RC, Zhang H. A comparison study of the reference curves of bone mineral density at different skeletal sites in native Chinese, Japanese, and American Caucasian women. Calcif Tissue Int 2003;73:122-32. [Crossref] [PubMed]
31. 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]
32. Wáng YXJ. A summary of our recent evidence-based works on radiographic diagnostics of prevalent osteoporotic vertebral fracture in older men and women. Quant Imaging Med Surg 2023;13:1264-85. [Crossref] [PubMed]
33. 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]
34. Namwongprom S, Ekmahachai M, Vilasdechanon N, Klaipetch A, Wongboontan C, Boonyaprapa S. Bone mineral density: correlation between the lumbar spine, proximal femur and Radius in northern Thai women. J Med Assoc Thai 2011;94:725-31.
35. Leslie WD, Lix LM, Tsang JF, Caetano PAManitoba Bone Density Program. Single-site vs multisite bone density measurement for fracture prediction. Arch Intern Med 2007;167:1641-7. [Crossref] [PubMed]
36. Roux C. Osteopenia: is it a problem? Int J Clin Rheumatol 2009;4:651-5.