Osteoporosis and Fracture Risk among Older US Asian Adults

Abstract

Purpose of Review

This review summarizes the current knowledge regarding osteoporosis and fracture among older US Asian adults.

Recent Findings

Asian adults have lower (areal) bone density than non-Hispanic White adults and thus are more likely to be diagnosed and treated for osteoporosis, despite their lower risk of hip fracture. The latter may relate to favorable characteristics in hip geometry, volumetric bone density, and bone microarchitecture; lower risk of falls; and other clinical factors. The fracture risk calculator FRAX accounts for the lower risk of hip fracture among US Asian adults. However, data on major osteoporotic fracture risk remain limited. Fracture rates also vary by Asian subgroup, which may have implications for fracture risk assessment. Furthermore, among women receiving bisphosphonate drugs, Asian race is a risk factor for atypical femur fracture, an uncommon complication associated with treatment duration. Recent clinical trial efficacy data pertaining to lower bisphosphonate doses and longer dosing intervals may be relevant for Asian adults.

Summary

More research is needed to inform osteoporosis care of US Asian adults, including risk-benefit considerations and the optimal duration of bisphosphonate treatment. Greater evidence-based guidance for primary fracture prevention among US Asian adults will ensure health equity in the prevention of osteoporotic fractures.

INTRODUCTION

The past fifty years have been marked by tremendous increase in the racial and ethnic diversity of the US population, with the Asian population now the fastest growing race group since the turn of the century [1, 2]. Following the repeal of the Chinese Exclusion Act in 1943, the 1965 Immigration Act, major wars in East and Southeast Asia, changing international relations, labor migration, and economic opportunities led to an early growing Asian population in the US [3] that has now nearly doubled from 2000 to 2020 [2]. In 2021, one-fifth of all US immigrants were from India (6.0%), China (5.3%), the Philippines (4.4%), Vietnam (3.0%), and Korea (2.2%) [4, 5]. Currently, 6.0% of the US population is of Asian race [6]. Chinese, Asian Indian, and Filipino ethnicities are the three largest subgroups and together with Vietnamese, Korean, and Japanese ethnicities account for 85% of the US Asian population [1, 7]. Native Hawaiian and Pacific Islander (PI) groups comprise an additional 0.3–0.5% of the overall US population [7].

In another demographic shift, one in five US residents will be over age 65 years by 2030 [8], with even greater ethnic diversity. In California, where nearly one third of all US Asians reside [1, 9], Asian/PI adults represent 16.9% of the state’s older population [10]. Osteoporosis is a serious public health issue for older adults and the burden of osteoporosis among Asian adults will increase substantially as the population ages. Yet the relative dearth of data pertaining to the epidemiology of osteoporosis and fracture among Asian ethnic groups presents unique healthcare challenges for this diverse population [11, 12]. The US Census has tracked Chinese, Japanese, Filipino, Korean, Asian Indian, and Vietnamese ethnicities since 1980 and earlier [13], but for the vast majority of epidemiologic studies focused on osteoporosis and fracture outcomes, this heterogeneous population has been grouped more broadly as “Asian”.

A challenge in clinical management is that Asian adults tend to have lower bone mineral density (BMD) than non-Hispanic White (NHW) adults, due in part to measurement of areal BMD by dual energy x-ray absorptiometry (DXA), which is influenced by bone size [14]. As such, Asian women are over-represented among post-menopausal women receiving osteoporosis therapy for primary fracture prevention [15, 16]. Yet hip fracture rates among US Asian adults are much lower than among NHW adults [17–21]. Bone structural properties and other factors that may explain these observed differences are briefly summarized in the current report and have been extensively reviewed elsewhere [22–24]. Another concern is the more recent recognition that Asian race is a risk factor for atypical femur fracture (AFF) among women who received bisphosphonate therapy [15, 25–28]. This uncommon treatment complication, first reported in the US [29–31] and Singapore [32, 33], is an additional consideration when determining the optimal length of bisphosphonate therapy for Asian women. In this review, we discuss the skeletal health of older US Asian adults, Asian subgroups, and clinical management considerations in primary care and suggest future research to guide evidence-based practice for this understudied population.

BONE STRUCTURE AND STRENGTH

Bone Mineral Density

A major determinant of bone strength is BMD which is typically measured by DXA, reflecting areal BMD (g/cm²) [34]. But in persons with small bone size, this two-dimensional measurement may underestimate true volumetric BMD, a three-dimensional property that can be assessed by quantitative computerized tomography (QCT) [14]. In the Osteoporotic Fractures in Men (MrOS) study, femoral neck areal BMD was lower in Asian compared to White men but femoral neck QCT data showed higher trabecular volumetric BMD and thicker cortices [35]. An estimate of volumetric BMD (bone mineral apparent density, BMAD) has also been calculated from DXA measurement of vertebral size and mineral content. The Study of Women’s Health Across the Nation (SWAN) found that areal BMD was lower in Japanese and Chinese women than in White women, but BMAD was higher in Asian women [36]. Others have also observed that calculation of BMAD reduces or eliminates BMD differences among Asian subgroups and White adults [37].

Few studies have compared (areal) BMD among other US Asian subgroups. Table 1 summarizes population studies that report areal BMD data for US Asian ethnic subgroups. In an early study of 449 South Asian women and 2245 age-matched Chinese women, age-specific femoral neck BMD was slightly lower for Chinese compared to South Asian women age 50–69 years [38]. In a more recent report, mean femoral neck BMD among 11,147 Filipina, 10,648 Chinese, and 2,519 Japanese women generally differed by ≤3%, but were collectively lower than NHW counterparts – at least 6–8% lower among older women [39]. While adjustment for stature (which was 6–8 cm lower among Asian women) reduced the Asian-NHW BMD differences by 30–40%, persisting BMD differences observed among older women suggested a potential age cohort effect [39].

Table 1.

Population studies examining bone mineral density (BMD) or fracture incidence among adult and older US Asian ethnic subgroups

Author (Year)	Cohort Setting & Years of Study	Sex and Age	Racial and Ethnic Subgroups (N)	Outcome: BMD or Fracture	Primary findings pertaining to the US Asian subgroups studied
Lauderdale, DS, et al. (1997) Reference [17]	US Medicare 1992–1993	women men ≥65 y	Chinese (24,366) Japanese (28,762) Korean (5470)	hip fracture	Age-adjusted incidence of hip fracture was generally lower for Chinese, Japanese, and Korean vs White adults, with incidence lowest for Chinese adults.
Ross PD, et al. (1991) Reference [99]	Oahu, Hawaii 1979–1981 Okinawa, Japan 1984–1985 Rochester, MN 1978–1982	women men 50–84 y	Japanese (Hawaii) Japanese (Okinawa) Caucasian (Hawaii) Caucasian (Rochester, MN)	hip fracture	Hip fracture incidence did not differ between Japanese adults in Hawaii vs Japanese adults in Okinawa Japan but were about half the hip fracture incidence of Caucasian adults in Hawaii and Rochester, MN.
Kin K, et al. (1993) Reference [137]	San Diego county, CA 1991–1992	women 18–89 y	US-born Japanese (151) Japan-born immigrant (137)	femoral neck BMD lumbar spine BMD	US-born Japanese women had higher BMD and higher percentage of body fat than immigrant Japanese women
Marquez, MA, et al. (2001) Reference [37]	Rochester, MN 1997 (Asian) 1989–1992 (White)	women men ≥20 y	Vietnamese (172) Cambodian (171) Laotian (53) White (684)	femoral neck BMD lumbar spine BMD	BMD was lower in Southeast Asian vs White adults. Some differences were reduced or eliminated when bone mineral apparent density (BMAD) was calculated, depending on sex and age/menopause status.
Morton DJ, et al. (2003) Reference [157]	San Diego county, CA 1992–1997 1995–1998	women 50–69 y	Filipina (285) Hispanic (164) Caucasian (354) Northern European	femoral neck BMD total hip BMD lumbar spine BMD	Adjusting for body size and lifestyle factors minimized BMD differences between Filipina, Hispanic, and Caucasian women.
Walker MD, et al. (2006) Reference [48]	New York City 2002–2004	women 20–90 y	Chinese (359)	femoral neck BMD total hip BMD lumbar spine BMD	US Chinese referent BMD values (for T-score calculation) are lower than Caucasian referent BMD values and would result in half reclassified from osteoporosis to osteopenia.
Melamed A, et al. (2010) Reference [47]	MASALA study; San Francisco Bay Area, CA 2006–2007	women men 45–79 y	South Asian Indian (150)	total hip BMD lumbar spine BMD	13–40% of South Asian adults were reclassified to normal BMD with a North or South Indian-based referent population instead of the NHW referent population
Finkelstein JS, et al. (2002) Reference [36]	Study of Women’s Health Across the Nation (SWAN) 1996–1997	women 42–52 y	Chinese (232–235) Japanese (257) African American (591–608) Caucasian (1051–1076)	femoral neck BMD lumbar spine BMD	BMD was highest in African American followed by Caucasian women and lowest in Chinese and Japanese women. Adjusting for body weight and covariates resulted in similar BMD for Chinese, Japanese, and Caucasian women.
Finkelstein JS, et al. (2008) Reference [91]	Study of Women’s Health Across the Nation (SWAN) 1996–2002 (includes follow-up)	women 42–52 y (baseline)	Chinese (221) Japanese (243) African American (494) Caucasian (944)	total hip BMD lumbar spine BMD	Japanese and Chinese women had greater spine BMD loss than Caucasian women; African American women had the least. Ethnic variations in bone loss were mainly attributed to body weight differences and were no longer evident after restricting analyses to women weighing 50–78 kg.
Shieh A, et al. (2022) Reference [105]	Study of Women’s Health Across the Nation (SWAN) 1996–2016 (includes follow-up)	women 42–52 y (baseline)	Chinese (147) Japanese (178) Black (443) Hispanic (105) White (681)	clinical fracture	Fracture risk for Black (HR 0.55 [CI 0.46–0.75]) and Japanese (HR 0.47 [CI 0.27–0.84]) women was lower than White women (reference), but non-significant for Chinese (HR 0.73 [CI 0.41–1.31]) and Hispanic women (0.50 [CI 0.22–1.14]).
Khandelwal S, et al. (2012) Reference [38]	Kaiser Permanente Northern California 1997–2010 (includes follow-up)	women 50–85 y (baseline)	Chinese (2,245) South Asian (449) NHW (4,490)	femoral neck BMD wrist fracture humerus fracture (proximal, shaft) hip fracture	BMD-osteoporosis prevalence tended to be higher for Chinese vs. South Asian women, both higher than NHW women. Wrist fracture incidence was similar for South Asian vs NHW women, both higher than Chinese women.
Lo JC, et al. (2016) Reference [49]	Kaiser Permanente Northern California 1997–2003	women 50–79 y	Chinese (4,039) NHW (20,195)	femoral neck BMD total hip BMD lumbar spine BMD FRAX score	A US Chinese T-score reduced osteoporosis prevalence by about half for US Chinese women. For those age <65y reclassified to femoral neck osteopenia, most had FRAX hip fracture risk <3%.
Lo JC, et al. (2020) Reference [39]	Kaiser Permanente Northern California 1998–2017	women 50–79 y	Chinese (10,648) Filipina (11,147) Japanese (2,519) NHW (115,318)	femoral neck BMD	Mean femoral neck BMD differed by ≤3% for older Chinese, Filipino, and Japanese women but was 6–8% lower than NHW women. Asian-White differences were partly attenuated with height adjustment but remained greater for older women.
Lo JC, et al. (2023) Reference [20]	Kaiser Permanente Northern California 2000–2021 (includes follow-up)	women men ≥50 y	Chinese (91,556) Filipino (106,643) South Asian (24,359) Japanese (19,782) Other/unspecified Asian/PI (161,347) NHW (1,473,885)	hip fracture	Hip fracture incidence rate ratios for US Asian/PI adults in aggregate vs NHW were 0.45 (women) and 0.34 (men) but varied among the four major Asian subgroups: ranging 0.45–0.64 for women (lower for Filipina, higher for Japanese and South Asian vs Chinese) and 0.32–0.55 for men (lower for Filipino, Japanese, Chinese vs South Asian)
Lo JC, et al. (2023) Reference [104]	Kaiser Permanente Northern California 2000–2021 (includes follow-up)	women men ≥50 y	Chinese (91,556) Filipino (106,643) South Asian (24,359) Japanese (19,782) Other/unspecified Asian/PI (161,347) NHW (1,473,885)	wrist fracture humerus fracture (proximal)	Humerus fracture incidence was lower for Filipino, higher for Japanese adults, and even higher for South Asian men vs. Chinese adults. Wrist fracture incidence was lower for Filipino and higher for South Asian adults and Japanese women vs. Chinese adults. Wrist fracture incidence for South Asian men approached that of NHW men.

y = year; BMD = bone mineral density; NHW = non-Hispanic White; PI = Pacific Islander; HR = hazard ratio; CI = confidence interval;

MASALA = Metabolic Syndrome and Atherosclerosis in South Asians Living in Americas (recruited from San Francisco Bay Area, CA)

This Table focuses on studies of central areal BMD (dual energy x-ray absorptiometry) and fracture incidence among US Asian ethnic subgroups.

Studies examining US Asian adults as an aggregate group are described in a recent comprehensive review by Noel, et. al. [11] and not included in this table.

The Study of Women’s Health Across the Nation (SWAN) examined BMD and fracture in midlife Chinese (recruited in Oakland, CA) and Japanese (recruited in Los Angeles, CA) women. https://www.swanstudy.org/publications/

Areal BMD has been used to define osteoporosis, but because BMD differs depending on the densitometer manufacturer, it is reported in relative terms as a T-score, representing standard deviations from peak bone mass [40]. In 1994, an expert committee sponsored by the World Health Organization recommended that osteoporosis be defined as a BMD T-score ≤ −2.5, osteopenia between −1.0 and −2.5, and normal ≥ −1.0 [41]. These definitions relied on reference populations of White women. When men were studied, new reference data allowed T-scores by sex, and additional race and ethnicity data added early sets of sex-specific ethnic T-scores, depending on the densitometer [42]. Once the uniform NHW femoral neck BMD reference from National Health and Nutrition Examination Survey (NHANES) III was used for all densitometers [43, 44], the NHW T-score was reported for Asian adults. NHANES III (1988–1994) studied NHW, Black, and Mexican men and women but not Asian adults [45, 46].

Using relative values for areal BMD (T-scores according to the distribution within a population), individuals with the same areal BMD but different ethnicity could have different ethnic T-scores that impact BMD classification. For example, a study of 150 South Asian women in California found that 13% were reclassified from low to normal BMD using a North India BMD reference (total hip), whereas 40% were reclassified using a South India BMD reference (lumbar spine) [47]. In 2006, Walker et al. published a Chinese American referent BMD database derived from 359 healthy Chinese American women aged 20–90 years in New York City [48]. Using this reference, the prevalence of T-score ≤ −2.5 in Chinese women fell from 43% to 21% based on the lowest of femoral neck, total hip, and lumbar spine T-scores [48]. This same Chinese BMD reference applied to Chinese women in northern California [49] showed similarly large shifts in T-score classification (Figure 1). Although Chinese women in New York may differ from those in California, the question remains how to best risk stratify Asian adults with BMD measured by DXA.

Figure 1.

Change in the percentage of Chinese women with T-score ≤−2.5 when bone mineral density T-score is calculated using US Chinese reference data compared to US Non-Hispanic White (NHW) reference data.

A much lower percentage of Chinese women are classified with bone mineral density (BMD) T-score ≤ −2.5 when a US Chinese BMD reference [48] is used for T-score calculation compared to the NHW BMD reference [43,48,49].

Figure redrawn from reference 49 (Lo JC, et al. Applying ethnic-specific bone mineral density T-scores to Chinese women in the USA. Osteoporos Int. 2016; 27:3477–3484) with permission.

Since 2002–2006 for women [42, 50] and 2013 for men [51], the International Society for Clinical Densitometry has recommended using the NHW female reference to calculate BMD T-scores for adults of all races and ethnicities [52]. Although local recommendations may vary with regard to sex- or race-specific T-scores [52, 53], the general consensus was to use an absolute definition for osteoporosis: areal BMD T-score relative to the peak bone mass of NHW women in the NHANES III, thereby providing a standard referent [54]. At the femoral neck, this osteoporosis threshold is an areal BMD of 0.577 g/cm² (Hologic densitometer; cross-calibration equations convert to other manufacturer densitometers) [43, 54]. The absolute value was defined in NHW women but is independent of age, sex, and ethnicity. However, this is an areal BMD and size differences must be considered. Even with volumetric bone density, BMD assignment does not fully convey the risk of fracture or whether any treatment will reduce the risk of fracture.

For NHW women, large population studies were conducted to enable estimation of fracture risk from areal BMD and other risk factors. The NIH-funded Study of Osteoporosis Fractures (SOF) enrolled nearly 10,000 women (99.7% White) in 1986–1988 [55] . Black women were excluded because of their low incidence of hip fractures; Asian and Hispanic women were not mentioned. The only “ethnic origin” information was Northern or Southern European ancestry [55]. Since then, more studies have included US Asian adults (as an aggregate group), but disparities remain in the amount of data examining the relationship of BMD and other risk factors to fracture. A major gap in the field is the lack of data pertaining to US Asian subgroups. The National Osteoporosis Risk Assessment (NORA) study characterized low peripheral BMD (heel, forearm, finger), risk factors, and fracture incidence in postmenopausal women, but only 1% of subjects were of Asian race [56]. The mean T-score for Asian women was lower compared to White women in each age decade, whereas their relative risk for fracture at one year was 0.32, adjusting for BMD, weight, and other covariates [57]. Within each racial and ethnic group in the NORA study, the associations of peripheral BMD (T-score) and fracture were similar [57].

Bone Microarchitecture

In addition to bone mass, bone microarchitecture plays an important role in bone strength [58]. Bone microarchitecture assessed by high-resolution peripheral QCT (HR-pQCT) of the distal radius and tibia demonstrates higher (volumetric) trabecular and cortical bone density and thickness in premenopausal Chinese American women compared to White women, despite smaller bone area [59], as well as better trabecular microstructure [60]. In postmenopausal women, higher cortical bone density and thickness and lower trabecular number but higher trabecular thickness have been observed in Chinese compared to White women, with similar estimates of whole bone stiffness [61]. Recent interest has focused on trabecular bone score (TBS), an index derived from lumbar spine DXA images [62] that correlates with bone microarchitectural properties at the spine and hip [63]. Leslie, et al. [64] examined data from 29,407 women in Manitoba, Canada and found that lumbar spine TBS predicted major osteoporotic fracture in women with diabetes and captured diabetes-associated fracture risk more optimally than BMD (BMD tends to be higher in patients with diabetes). Whether these same relationships exist for Asian adults, a demographic group with prevalent diabetes and lower body weight [65], has not been examined. Racial and ethnic differences in TBS have been reported [66–68], and more studies examining TBS in racial and ethnic minority populations have been advocated [69].

Hip Geometry

Faulkner and colleagues [70] were the first to report that longer hip axis length predicts increased hip fracture risk among postmenopausal White women. Cummings, et al. [71] further observed that US Asian women (89% were Japanese-American) had shorter hip axis length independent of height differences, which could be a potential explanation for their lower risk of hip fracture. Other subsequent studies have also identified shorter hip (or femoral neck) axis length in Asian compared to White adults, as summarized in prior reviews [22, 72]. In the SWAN study, differences in hip structure analysis measures and higher composite strength indices which may confer lower hip fracture risk were observed in midlife Chinese and Japanese women when compared to White women [73–75].

Mineral Metabolism

Important racial and ethnic differences are seen in vitamin D metabolism [76], along with variations in sun exposure and supplement use [77, 78]. In the Multi-Ethnic Study of Atherosclerosis (MESA), community-dwelling Chinese adults had 25OH-vitamin D levels that were lower than NHW adults, similar to Hispanic adults, and higher than Black adults [76]. Increased prevalence of vitamin D deficiency has been reported in some but not all US Asian subgroups compared to NHW adults [79–82], although Filipino adults are less likely than East Asian adults to have low 25OH-vitamin D levels [77, 81]. As a group, Asian adults with vitamin D deficiency are less likely than NHW adults to achieve successful repletion after initial pharmacologic therapy [83]. Serum vitamin D levels also depend on binding protein levels, which are diverse. Despite widespread interest in vitamin D, there is insufficient data about the optimal vitamin D level for adults of Asian race. Asian subjects (1.5%) were under-represented in the recent large Vitamin D and Omega-3 Trial (VITAL) of vitamin D [84]. Few studies have examined cultural variation in diet and relation to fracture risk among the Asian diasporas. Traditional diets in many Asian countries are low in calcium [85], but calcium intake among Asian Americans may be higher, with variation by subgroup and generational status [86].

Bone Remodeling

Multiple studies from China, Japan, and Singapore confirm a positive relationship between bone turnover markers and risk of fracture [87–89]. In the SWAN study, increases in the bone resorption marker urinary crosslinked N-telopeptide of type I collagen (NTX) across the menopause transition were greatest for Japanese followed by Chinese women compared to White women, but ethnic differences were attenuated after adjusting for BMI and other covariates [90]. These findings extend earlier observations that ethnic variation in perimenopausal bone loss appear to be largely attributable to differences in body weight [90, 91]. The SWAN study also found that pre- and early perimenopausal levels of serum osteocalcin, a bone formation marker, were lower in Chinese and Japanese women compared to White women [92, 93]. Serum sclerostin, a protein that inhibits osteoblast differentiation and bone formation, has been associated with hip fracture risk in White women [94], but levels do not appear to differ between White and Chinese women and are not associated with volumetric BMD or microarchitecture [95]. To our knowledge, no studies in the US have examined bone biopsies in a population of Asian adults to directly measure their bone turnover rates.

FRACTURE EPIDEMIOLOGY

Hip Fracture

Epidemiologic research over four decades consistently demonstrate much lower hip fracture incidence among US Asian compared to NHW adults [17–21, 96–99]. These reports are summarized in a recent comprehensive review by Noel and colleagues [11]. However, the Asian population has been largely studied in aggregate [11], with or without inclusion of Native Hawaiian/PI adults, comprising ≤7% of the Asian/PI population [7]. Overall, among older US Asian adults, hip fracture rates have generally ranged about 35–65% lower than in NHW adults, varying by sex, ethnicity, and era.

Few studies have compared hip fracture incidence among US Asian subgroups. Table 1 summarizes the existing US studies that have reported on fracture outcomes among one or more Asian subgroups. In an early study of East Asian adults using 1992 US Medicare data, sex-specific hip fracture incidence was higher for Japanese and Korean adults compared to Chinese adults, with all three groups generally lower than White adults [17]. The standardized fracture ratios relative to White women/men were 0.30/0.42 for Chinese, 0.73/0.58 for Japanese, and 0.53/0.91 for Korean adults, respectively [17]. In a more contemporary population of northern California adults aged ≥50 years in an integrated healthcare system, age-adjusted hip fracture incidence ratios for Asian/PI compared to NHW adults were 0.45 for women and 0.34 for men, with heterogeneity by Asian subgroup [20]. Compared to Chinese women, hip fracture incidence was lower for Filipina and higher for Japanese and South Asian women, and compared to Chinese men, fracture incidence was similar for Filipino and Japanese but higher for South Asian men [20]. The corresponding hip fracture incidence rate ratios compared to NHW women/men were: 0.45/0.35 for Chinese, 0.64/0.39 for Japanese, 0.37/0.32 for Filipino, and 0.56/0.55 for South Asian adults among the four major Asian subgroups examined [20].

Several factors may explain the lower risk of hip fracture in Asian populations, including differences in hip geometry and other bone structure or strength indices that confer additional skeletal advantages (as previously discussed). Lower stature and weight may result in less impact during ground level falls, reducing the likelihood of fracture [61]. Other health or lifestyle factors such as balance, physical mobility, nutrition, diet, smoking, alcohol intake, and clinical factors [53] that differ by race and ethnicity may also contribute, despite the lower weight and lower BMI typically observed in Asian adults [39, 65, 100, 101]. The risk of falls also differs by race and ethnicity; several studies report that US Asian women are one-third less likely to have ≥1 or ≥2 falls within the past year compared to NHW women [102, 103], although subgroup differences have not been studied. Anthropometric differences include higher BMI among Filipino adults and recent trends in younger populations also suggest that BMI is higher among Japanese and South Asian adults compared to Chinese, Vietnamese, and other Southeast Asian adults [65, 100, 101].

Other Major Osteoporotic Fracture

Fewer studies have examined humerus and wrist fractures in US Asian populations, and findings for Asian subgroups remain sparse (Table 1). In a study using 2000–2005 Medicare claims data, the fracture incidence ratio comparing Asian to White adults was 0.63 for distal radius or ulna fracture and 0.52 for humerus fracture, adjusting for age, sex, and other factors [19]. Differences between Asian men and women were not characterized, but in the same population, the fracture incidence ratio for hip fracture was 0.61 [19]. Among older South Asian women and age-matched Chinese and NHW women in an integrated Northern California healthcare system, Khandelwal et al. observed that South Asian women had a higher incidence and relative rate of wrist fracture compared to Chinese women, with rates approaching that of NHW women [38]. In a much larger and more contemporary population of Asian adults age ≥50 years from the same Northern California healthcare system (Table 1), Lo and colleagues observed that wrist fracture rates were lower for Filipino compared to Chinese adults, but higher for Japanese women and South Asian adults, with South Asian men not significantly different from NHW men [104]. The age-adjusted wrist fracture incidence rate ratios compared to NHW women/men were: 0.62/0.68 for Chinese, 0.82/0.75 for Japanese, 0.27/0.42 for Filipino, and 0.76/0.90 for South Asian adults, respectively [104]. In this same study, proximal humerus fracture rates also varied but were lower than NHW adults, except for South Asian men where differences did not reach statistical significance; the incidence rate ratios compared to NHW women/men were: 0.40/0.35 for Chinese, 0.56/0.50 for Japanese, 0.30/0.25 for Filipino, and 0.41/0.79 for South Asian adults [104]. These results highlight differences in upper extremity fracture incidence by Asian ethnicity, lower for Filipino and higher for selected Japanese and South Asian subsets, depending on the skeletal site and sex [104]. In addition, findings from the SWAN study suggest that ethnic patterns in fracture risk further differ for the outcome of any clinical fracture [105].

Studies from countries in Asia suggest that vertebral fracture prevalence and incidence are high in older Asian adults, but methodologic approaches vary, including methods for morphometric ascertainment [106–108]. To our knowledge, there are no data comparing vertebral fracture incidence and risk among US Asian subgroups, and the vertebral fracture incidence ratios comparing US Asian and Asian subgroups to NHW adults remain somewhat uncertain.

Atypical Femur Fracture

Atypical femur fractures (AFF) are an uncommon complication of potent antiresorptive therapy, first reported with bisphosphonate drugs [29–33, 109]. Prolonged suppression of bone turnover coupled with impaired micro-crack repair in areas of higher mechanical stress are hypothesized to predispose susceptible patients to AFF [30, 110, 111]. The risk of AFF is evident after three years of bisphosphonate treatment [112], increases substantially with treatment duration [27, 111, 113], and varies by race and ethnicity. Notably, 5- to 6-fold greater risk is observed in US Asian women [15, 26, 27]. Two independent California populations of primarily women who received bisphosphonate therapy (13–17% Asian) demonstrate an age-adjusted incidence of AFF ranging from 2–3 per 100,000 person-years for <2 years treatment to 112–113 per 100,000 person-years for ≥8 years [111, 113]. Among women with any bisphosphonate use, Black et al. reported 9-fold and 43-fold higher risk of AFF after 3–5 years and ≥8 years of treatment, respectively, compared to ≤3 months [27]. The risk of AFF appears to decline rapidly following bisphosphonate cessation [27, 114]. Atypical femur fracture can occur with denosumab, but the additional risk associated with Asian race has not been well characterized.

The mechanisms underlying the increased risk of AFF in Asian women with bisphosphonate exposure have not been clearly elucidated. Although BMD may be associated with AFF location in the diaphyseal femur [115], the risk of AFF appears to be independent of hip BMD [26, 27]. One hypothesis is that differences in femur geometry and greater femur bowing in Asian populations result in biomechanical forces that predispose to peak tensile stress and formation of (atypical) stress fractures [15, 26, 115–120]. This could explain the frequent symmetric bilaterality of AFF findings (incomplete AFF or stress reaction in the contralateral femur in a patient with complete AFF). A recent Australian study examined women with typical femur fracture and those with AFF and found that Asian ethnicity remained a strong independent predictor of AFF after accounting for differences in femur geometry [121]. However, AFF risk has not been examined in prospective cohorts with respect to measures of femur geometry [120], bone turnover [122], and genetic factors [123].

In Asian women receiving bisphosphonate therapy, the risk benefit considerations currently favor inclusion of drug holidays to limit the duration uninterrupted therapy beyond five years. After three years of treatment for Asian women, Black et al. estimated that 91 hip fractures and 330 clinical fractures would be prevented compared to 8 AFF events, whereas this difference narrowed to 360 hip fractures and 831 clinical fractures prevented compared to 236 AFF events after ten years of treatment [27]. These projections optimistically assumed that the benefit of bisphosphonate therapy beyond five years is the same as during the first three years, and that the fracture reduction in Asian women is the same as in NHW women [124]. However, the efficacy of neither short nor long-term bisphosphonate therapy has been examined in US Asian women, a population with lower hip fracture risk than NHW women and one scarcely represented in the major US osteoporosis clinical trials [125–128].

FRACTURE RISK ASSESSMENT

Although BMD-based treatment recommendations (T-score ≤ −2.5, NHW reference) do not account for racial and ethnic differences in fracture risk, the fracture risk assessment tool FRAX (https://frax.shef.ac.uk/FRAX/) currently considers the lower population fracture risk in Asian adults by including a calibration factor [129] for the US Asian FRAX (0.50 for women, 0.64 for men). These estimates are based on epidemiologic studies demonstrating lower hip fracture incidence among US Asian compared to White adults [17, 96, 97], thereby reducing potential over-estimation of fracture risk which is relevant for populations with lower BMD and lower fracture incidence. The Canadian FRAX does not account for Asian race and can overestimate their fracture risk [130]. While ongoing refinement of FRAX is expected [131, 132], including new risk factors and efforts to examine FRAX and the efficacy of intervention [133], studies examining how fracture risk prediction can be optimized for US Asian subgroups and accounting for factors such as length of residence in the US are likely to be relevant [134, 135]. Data also suggest there are first immigrant and generational effects on skeletal health [135–137]. Prevalent comorbidities should also be studied; for instance, diabetes has been identified as a rheumatoid arthritis risk equivalent in FRAX based on studies conducted in primarily White populations [138, 139]. On a much larger scale, an updated version of FRAX is planned, that will be informed by pooling multiple large and diverse prospective cohorts with baseline risk factor assessment and fracture outcomes, where ethnicity-specific differences will also be examined [140].

In the meantime, clinicians should be aware that fracture risk may differ amongst Asian ethnicities due to unmeasured risk factors and this should be considered when interpreting FRAX scores calculated under the umbrella of “US Asian”. While data informing major osteoporotic fracture risk for US Asian subgroups remain sparse, some evidence points to greater differences among Asian subgroups at other skeletal sites beyond hip fracture, including differences among East Asian adults [104]. South Asian adults, especially men, may have upper extremity fracture incidence rates closer to the NHW population [38, 104] and some propose that the US Caucasian FRAX is more appropriate for assessing major osteoporotic fracture risk in South Asian men [141]. As we strive for greater health equity in research and clinical practice, the question remains how the known heterogeneity among US Asians can be addressed through FRAX and similar fracture risk calculators, in order to unmask meaningful differences in clinical outcomes.

For first generation adult immigrants, the native country FRAX calculator has been considered [142, 143], but this has not been studied for Asian adults in the US and may yield conflicting guidance for Chinese immigrants from highly industrialized regions in Asia (Taiwan, Hong Kong, and Singapore); their country-region FRAX scores for hip fracture risk are much higher than the US Asian, China, and US Caucasian FRAX scores for the same clinical profile [134, 145] and country-specific intervention thresholds may differ [144]. As the US Asian population ages, there may be further bone health differences among US native-born, foreign-born acculturated, and foreign-born adult immigrant populations that warrant consideration, including the role of biology and social, behavioral, and cultural factors [146–148].

TREATMENT CONSIDERATIONS

What Does Guideline Concordant Care Look Like for the Asian Woman?

Currently, guideline [53] concordant care in the US results in the treatment of Asian women aged 65–70 years with areal BMD T-score ≤ −2.5 and no prior fracture or other risk factors. The vast majority of these women have discordant BMD- and FRAX-based treatment indication that is generally not seen for NHW female counterparts (Table 2). Key questions are whether FRAX, which is clinically used for BMD T-score between −1.0 and −2.5 [53], can help risk stratify postmenopausal Asian women under age 70 years with no risk factors except T-score ≤ −2.5, and whether alternative T-score thresholds might address the lower areal BMD in Asian women with short stature. These questions target the concern of potential overtreatment of Asian women who are otherwise at lower fracture risk except for their BMD categorization [12].

Table 2.

Large discordance in osteoporosis treatment indication is evident for US Asian women aged 65–70 years old with femoral neck T-score ≤ −2.5 and no other clinical risk factors compared to non-Hispanic White women with the same clinical profile

10-year risk of hip fracture estimated by FRAX* for women with femoral neck BMD T-score ≤ −2.5 and no clinical risk factors
120 pounds, 62 inches	Non-Hispanic White Women (US Caucasian FRAX)						US Asian Women (US Asian FRAX)
Age (years)	T −2.5	T −2.6	T −2.7	T −2.8	T −2.9	T −3.0	T −2.5	T −2.6	T −2.7	T −2.8	T −2.9	T −3.0
65	2.7	3.1	3.4	3.8	4.2	4.7	1.5	1.7	1.9	2.1	2.4	2.6
66	2.9	3.2	3.6	4.0	4.4	4.9	1.6	1.8	2.0	2.2	2.5	2.8
67	3.1	3.4	3.8	4.2	4.7	5.2	1.7	1.9	2.1	2.4	2.6	2.9
68	3.3	3.6	4.0	4.5	4.9	5.5	1.9	2.1	2.3	2.5	2.8	3.1
69	3.5	3.9	4.3	4.7	5.2	5.8	2.0	2.2	2.4	2.7	3.0	3.3
70	3.7	4.1	4.5	5.0	5.5	6.1	2.1	2.3	2.6	2.8	3.1	3.5
71	4.0	4.4	4.8	5.3	5.8	6.4	2.3	2.5	2.8	3.0	3.3	3.7
72	4.2	4.7	5.1	5.6	6.2	6.8	2.4	2.7	2.9	3.2	3.6	3.9
73	4.5	5.0	5.5	6.0	6.6	7.2	2.6	2.9	3.2	3.5	3.8	4.2
74	4.8	5.3	5.8	6.3	6.9	7.6	2.8	3.1	3.4	3.7	4.1	4.5
75	5.2	5.6	6.2	6.7	7.3	8.0	3.0	3.3	3.6	4.0	4.3	4.7

Shaded areas indicate when the Bone Health and Osteoporosis Foundation (BHOF) treatment threshold of ≥3% for FRAX-calculated 10-year risk of hip fracture is reached or exceeded, evident for most non-Hispanic White women aged >65 years with BMD T-score ≤ −2.5 (non-Hispanic White reference) and with no other clinical risk factors (body mass index 21.9 kg/m2).

^*The FRAX scores for 10-year risk of hip fracture are calculated using the FRAX Tool: https://frax.shef.ac.uk/FRAX/ (web version 4.3)

Bisphosphonate Dosing and Drug Holidays

When bisphosphonates were first approved for osteoporosis, guidelines recommended treatment at an areal BMD T-score ≤ −2.0 or when the T-score was ≤ −1.5 with risk factors which could include body weight <127 pounds [149]; in an early era, low body weight was also an osteoporosis screening indication [43]. This may have resulted in disproportionate unnecessary treatment of postmenopausal Asian women undergoing BMD testing. As BMD testing now targets all women aged 65–75 years [150], more evidence-based guidance is needed for primary fracture prevention among Asian women.

The majority of postmenopausal women receiving osteoporosis therapy are treated with bisphosphonate drugs. In NHW women, there is strong clinical trial efficacy data for the first three years. However, over time, prolonged inhibition of bone remodeling may compromise bone micromechanical properties [151, 152] and over-suppression of bone turnover may predispose to AFF. An important area lacking evidence is the bone turnover rate in Asian women and the degree to which bisphosphonate treatment suppresses their bone formation. While drug holidays are now considered after five years of treatment [153], future trials should examine whether earlier drug holidays (e.g., temporary cessation after three years therapy) might result in similar efficacy while limiting the initial length of uninterrupted treatment. Lower dosing regimens have been proposed, based on early comparability of standard and lower oral bisphosphonate dose for fracture prevention [152]. In Japan, the approved alendronate dose for osteoporosis is lower at 5 mg/day [154]. Randomized clinical trial data demonstrate that 5 mg of zoledronate given at wider 18-month intervals over six years reduces the risk of both nonvertebral and vertebral fractures in women with osteopenia [155]. Furthermore, a single dose of 5 mg zoledronate resulted in an increase in spine and hip BMD that persisted for five years and then returned to baseline after nine years [156]. Zoledronate doses as low as 1 mg at baseline result in higher BMD compared to placebo four years later [156]. These findings support efforts to limit the bisphosphonate dose and dosing frequency, which may be especially relevant for Asian women with smaller body size. Finally, studies examining baseline bone turnover and fracture efficacy of osteoporosis drugs in Asian adults may inform individualized treatment considerations.

CONCLUSION

Despite substantial advances in our understanding of bone fragility and fracture risk, there are serious knowledge gaps concerning skeletal health in US Asian adults. More research is needed to inform fracture prevention care of this diverse and heterogeneous population that generally has lower areal BMD, yet lower risk of hip fracture, and important differences by ethnicity, sex, and fracture site. The most critical knowledge gaps concern primary fracture prevention, risk benefit treatment considerations, and counseling of fracture risk, where data pertaining to the association of areal BMD, clinical risk factors, and fracture outcome are lacking for the Asian population and especially for Asian ethnic subgroups. Future studies should target high risk fracture sites, refinement of fracture risk assessment tools, safety and efficacy of osteoporosis medications, evidence-based guidance regarding treatment, and approaches to optimizing the benefit to risk ratio of pharmacologic therapy in this population.

Code Availability:

Not applicable