Task force recommends population-based tool to better assess fracture risk in US
Click Here to Manage Email Alerts
Key takeaways:
- A systematic review found limited evidence to justify the use of racial-ethnic adjustments in FRAX.
- New population-based prediction tools are needed to better assess fracture risk in the U.S.
The use of racial-ethnic adjustments in the U.S. version of the Fracture Risk Assessment Tool may be underestimating fractures among underrepresented groups, and new, more accurate tools are needed, according to a task force report.
The American Society for Bone and Mineral Research (ASBMR) convened a task force to assess how well the U.S. version of the Fracture Risk Assessment Tool (FRAX) performs with predicting the risk for incident fractures over 10 years for Asian, Black, Hispanic and white adults. The task force found that few studies have been conducted among diverse populations, with most participants consisting of non-Hispanic white women. Additionally, the evidence that has been published found FRAX is limited in its ability to determine fracture risk among Black and Hispanic women.
“The task force concluded that research is needed to create fracture risk prediction tools that do not include race and ethnicity, and which are representative from a race and ethnicity, gender and age perspective,” Sherri-Ann Burnett-Bowie, MD, MPH, associate professor of medicine at Harvard Medical School and Massachusetts General Hospital, told Healio.
Data on racial-ethnic groups lacking
The University of Minnesota Evidence-based Practice Core conducted a systematic review for the ASBMR task force. The review included studies investigating the performance of prediction tools for incident fractures across racial-ethnic groups in the U.S. Six studies were included in the review: Three analyzed data from the Women’s Health Initiative and three assessed data from the Study of Osteoporotic Fractures (SOF) cohort.
The first WHI report assessed fractures among younger postmenopausal women aged 50 to 64 years. Discrimination of US-FRAX without bone mineral density included was deemed poor to fair by the researchers. In assessing risk for major osteoporotic fractures, the area under the curve was 0.53 for Black women, 0.57 for Hispanic women and 0.57 for white women. AUC improved to 0.54 for Black women and 0.66 for white women while declining to 0.53 for Hispanic women when assessing risk for hip fractures.
The second WHI study assessed fracture risk among postmenopausal women aged 50 to 79 years. For predicting major osteoporotic fractures without BMD, the AUC was 0.61 for Black women and 0.64 for white women. AUC increased for hip fractures to 0.81 for Black women and 0.76 for white women. When BMD was included in FRAX, AUC was 0.66 for Black women and 0.68 for white women for predicting major osteoporotic fractures and 0.85 for Black women and 0.78 for white women for predicting hip fractures.
The third WHI study included women aged 50 to 79 years in four WHI substudies with genomic data. US-FRAX overpredicted major osteoporotic fracture risks for all racial-ethnic groups, particularly for American Indian, Asian, Black and Hispanic women.
All three SOF papers examined the performance of FRAX for predicting fracture risk among women aged 65 years and older. AUC for predicting fractures with and without BMD included were similar to what was observed in the WHI studies. The second SOF report found FRAX with BMD included closely predicted the number of women with an incident major osteoporotic or hip fracture, though no data were reported on whether the accuracy varied across predicted fracture risk. Additionally, researchers could not assess US-FRAX’s performance among different racial-ethnic groups in the SOF papers because more than 99% of participants were non-Hispanic women.
Burnett-Bowie said the findings from the systematic review revealed there was little justification for including racial-ethnic adjustments in the U.S. version of FRAX.
“In that systematic review there was only one study, the WHI, where FRAX was studied in women from different racial and ethnic groups,” Burnett-Bowie said. “In the relevant WHI publications by Carolyn Crandall, MD, and team, there were few fractures in American Indian, Asian, Black, Hispanic, Native Hawaiian or Pacific Islander women and/or FRAX did not predict future fractures well.”
Population-based prediction tool needed
The task force said the dynamic social construct nature of race and ethnicity, the fact that race and ethnicity do not reflect biology or genetics and the increasing number of people in the U.S. who identify as mixed-race heritage are among the reasons a population-based fracture prediction algorithm should be created. The task force wrote that research is “urgently needed” to create such a tool.
“We need to conduct research studies in cohorts that more align with U.S. population demographics,” Burnett-Bowie said. “Having such improved racial and ethnic representation means that the findings of those studies will be more applicable. Such cohorts might be generated from de-identified medical records, with the appropriate prior consent, or through enrolling new cohorts. We need to ask research participants a wider range of questions, so that we can identify clinical, behavioral and social determinants of fracture risk. As precision medicine advances, we will hopefully be able to identify genetic and hormonal factors that predict a patient’s risk of fracture, in combination with clinical, behavioral and social determinants.”
To close gaps in osteoporosis care research, the task force said, race and ethnicity data need to be standardized across studies, social and behavioral determinants of health should be assessed to determine how they impact fracture risk, and postfracture outcomes should be collected among diverse groups to better understand differences and identify potential interventions. To advance fracture prediction, the task force encouraged the creation of a population-based risk calculator that does not adjust for race and ethnicity as well as a better description of fracture rates and BMD among underrepresented groups.
As more research is conducted, Burnett-Bowie said, providers need to inform people from underrepresented groups about the shortcomings of FRAX, so they have a better understanding of their fracture risk.
“In patients with osteopenia with a DXA-measured T-score of –1 to –2.5, providers should share with Asian, Black or Hispanic patients that FRAX may underestimate their risk for fracture and that their fracture risk may be higher than what is calculated using the Asian, Black or Hispanic U.S. FRAX calculators,” Burnett-Bowie said. “For Black patients, providers should advise that Black patients may be at increased risk for complications after a fracture and thus it is that much more important to start anti-osteoporosis prescription medication if they have experienced prior fracture, have osteoporosis based on DXA T-score or have increased fracture risk on FRAX.”
For more information:
Sherri-Ann Burnett-Bowie, MD, MPH, can be reached at sburnett-bowie@mgh.harvard.edu.
Collapse
Click Here to Manage Email Alerts