Race and Ethnicity and Fracture Prediction Among Younger Postmenopausal Women in the Women’s Health Initiative Study

Abstract

ImportanceThe best approach to identify younger postmenopausal women for osteoporosis screening is uncertain. The Fracture Risk Assessment Tool (FRAX), which includes self-identified racial and ethnic information, and the Osteoporosis Self-assessment Tool (OST), which does not, are risk assessment tools recommended by US Preventive Services Task Force guidelines to identify candidates for bone mineral density (BMD) testing in this age group.ObjectiveTo compare the ability of FRAX vs OST to discriminate between younger postmenopausal women who do and do not experience incident fracture during a 10-year follow-up in the 4 racial and ethnic groups specified by FRAX.Design, Setting, and ParticipantsThis cohort study of Women’s Health Initiative participants included 67 169 women (baseline age range, 50-64 years) with 10 years of follow-up for major osteoporotic fracture (MOF; including hip, clinical spine, forearm, and shoulder fracture) at 40 US clinical centers. Data were collected from October 1993 to December 2008 and analyzed between May 11, 2022, and February 23, 2023.Main Outcomes and MeasuresIncident MOF and BMD (in a subset of 4607 women) were assessed. The area under the receiver operating characteristic curve (AUC) for FRAX (without BMD information) and OST was calculated within each racial and ethnic category.ResultsAmong the 67 169 participants, mean (SD) age at baseline was 57.8 (4.1) years. A total of 1486 (2.2%) self-identified as Asian, 5927 (8.8%) as Black, 2545 (3.8%) as Hispanic, and 57 211 (85.2%) as White. During follow-up, 5594 women experienced MOF. For discrimination of MOF, AUC values for FRAX were 0.65 (95% CI, 0.58-0.71) for Asian, 0.55 (95% CI, 0.52-0.59) for Black, 0.61 (95% CI, 0.56-0.65) for Hispanic, and 0.59 (95% CI, 0.58-0.59) for White women. The AUC values for OST were 0.62 (95% CI, 0.56-0.69) for Asian, 0.53 (95% CI, 0.50-0.57) for Black, 0.58 (95% CI, 0.54-0.62) for Hispanic, and 0.55 (95% CI, 0.54-0.56) for White women. For discrimination of femoral neck osteoporosis, AUC values were excellent for OST (range, 0.79 [95% CI, 0.65-0.93] to 0.85 [95% CI, 0.74-0.96]), higher for OST than FRAX (range, 0.72 [95% CI, 0.68-0.75] to 0.74 [95% CI, 0.60-0.88]), and similar in each of the 4 racial and ethnic groups.Conclusions and RelevanceThese findings suggest that within each racial and ethnic category, the US FRAX and OST have suboptimal performance in discrimination of MOF in younger postmenopausal women. In contrast, for identifying osteoporosis, OST was excellent. The US version of FRAX should not be routinely used to make screening decisions in younger postmenopausal women. Future investigations should improve existing tools or create new approaches to osteoporosis risk assessment for this age group.

Potential Discrepancies in Sensitivity of Fracture Prediction Tools

“…To the Editor We read with great enthusiasm the recently published original investigation by Dr Crandall and colleagues that evaluated the Fracture Risk Assessment Tool (FRAX) and the Osteoporosis Self-Assessment Tool and found poor performance in identifying major osteoporotic fracture (MOF) in young postmenopausal women of certain racial and ethnic categories. This discovery is very important for guiding early screening of osteoporosis in women, given that one-third of women globally will experience at least 1 fracture in their lives after age 50 years .…”

“…This discovery is very important for guiding early screening of osteoporosis in women, given that one-third of women globally will experience at least 1 fracture in their lives after age 50 years . The findings of Dr Crandall and colleagues challenged the validity of the traditional FRAX scale and the Osteoporosis Self-Assessment Tool and offered valuable insights for developing new screening guidelines with higher sensitivity and specificity for osteoporosis in the future.…”

“…Second, fall frequency is a significant risk factor related to fractures but is currently excluded from the standard quantifiable variables used with the current analysis of FRAX model . Although the researchers counted and grouped participants’ fracture history in detail, the application of this scale is more inclined toward populations without a history of fractures.…”

Potential Discrepancies in Sensitivity of Fracture Prediction Tools

“…Although reliable data indicate that fracture risk differs among racial and ethnic groups—even when bone mineral density (BMD) is the same—we believe that the inclusion of BMD information is crucial. However, the study by Dr Crandall and colleagues mainly evaluated US FRAX without BMD information. We assert that a comprehensive approach should involve multiple clinical tools, such as BMD and FRAX, to identify high-risk patients and ensure appropriate treatment rather than labeling a single tool as suboptimal or recommending against its routine use.…”

“…Second, the study focused on incident MOFs as the primary end point. However, it is worth considering that outcomes may vary following MOFs, with racial and ethnic disparities being observed.…”