Associations of ethnicity, skin tone, and genome-wide sequencing with bone mineral density in adolescents.

Background: Dual-energy x-ray absorptiometry reference data designate Black and non-Black categories, as higher BMD has been documented among Black youth. We examined associations of race, skin tone, and genetic factors with bone mineral density (BMD).

Methods: 557 adolescents were followed longitudinally. Exposures included race, skin tone, and principal components (PC) from genome-wide arrays. Total body BMD Z-score (BMD-Z) was the primary outcome using linear regression.

Results: 359 adolescents identified as non-Hispanic White (NHW) and 75, non-Hispanic Black (NHB). BMD-Z was higher in NHB vs. NHW (β: 0.92 units, 95% CI: 0.64, 1.19) or those with darker skin (0.79, 95% CI: 0.49, 1.08 for brown vs. medium). The first genetic PC (PC1) correlated with identification as NHB. PC1 was associated with higher BMD-Z (0.09, 95% CI: 0.06, 0.12), even after including race (0.07, 95% CI: 0.00, 0.14) or skin tone (0.10, 95% CI: 0.05, 0.15); both race (0.26, 95% CI: -0.49, 1.01 for NHB vs. NHW) and skin tone (-0.08, 95% CI: -0.59, 0.44 for brown vs. medium) no longer predicted BMD-Z after adjustment for PC1.

Conclusion: Genetic similarity was robustly associated with BMD, prompting a reevaluation of adolescent BMD reference data to exclude the consideration of race.

Impact: Current bone density reference databases include a binary assignment of patients into "Black" and "non-Black" categories, as a higher BMD has been documented among those identifying as Black compared with individuals of other racial and ethnic backgrounds. This study found genetic similarity to be more strongly associated with bone density by dual-energy x-ray absorptiometry than race or skin tone. These data emphasize a need to reevaluate how bone density measurements are interpreted, including exploring reference data that exclude the consideration of race.