Potential association of bone mineral density loss with cognitive impairment and central and peripheral amyloid-β changes: a cross-sectional study

BACKGROUND: There is some evidence in the literature that older adults with cognitive impairments have a higher risk for falls and osteoporotic hip fractures. Currently, the associations between bone health and cognitive health have not been extensively studied. Thus, the present cross-sectional study aims to investigate the relationship between markers of bone loss and cognitive performance in older adults with and without osteopenia as well as older adults with cognitive impairments (i.e., Alzheimer’s disease [AD]). METHODS: Sixty-two non-osteopenia participants and one hundred three osteopenia participants as the cohort 1 and 33 cognitively normal non-AD participants and 39 AD participants as the cohort 2 were recruited. To assess cognitive and bone health, hip bone mineral density (BMD) and cognitive performance (via Minimal Mental State Examination [MMSE] and/or Auditory Verbal Learning Test-delayed recall [AVLT-DR]) were assessed. Furthermore, in cohort 1, plasma amyloid-β (Aβ) levels, and in cohort 2, cerebrospinal fluid (CSF) Aβ levels were determined. RESULTS: We observed that (1) compared with non-osteopenia participants, BMD values (t = − 22.806; 95%CI: − 1.801, − 1.484; p < 0.001), MMSE scores (t = − 5.392; 95%CI: − 3.260, − 1.698; p < 0.001), and AVLT-DR scores (t = − 4.142; 95%CI: − 2.181, − 0.804; p < 0.001), plasma Aβ42 levels (t = − 2.821; 95%CI: − 1.737, − 0.305; p = 0.01), and Aβ42/40 ratio (t = − 2.020; 95%CI: − 0.009, − 0.001; p = 0.04) were significantly lower in osteopenia participants; (2) plasma Aβ42/40 ratio showed a mediate effect for the association between BMD values and the performance of cognitive function in osteopenia participants by mediation analysis, adjusting age, sex, years of education, and body mass index (BMI); (3) BMD values (95%CI: − 1.085, 0.478; p < 0.001) were significantly reduced in AD participants as compared with cognitively normal non-AD participants; (4) in AD participants, the interactive effects of BMD and CSF Aβ42/40 ratio on MMSE scores was found by regression analysis, controlling age, sex, years of education, and BMI; (5) BMD can distinguish AD participants from cognitively normal non-AD participants with AUC of 0.816 and distinguish participants with the cognitive impairment from cognitively normal participants with AUC of 0.794. CONCLUSION: Our findings suggest a relationship between bone health and cognitive health. Given the correlations between BMD and important markers of cognitive health (e.g., central and peripheral pathological change of Aβ), BMD might serve as a promising and easy-accessible biomarker. However, more research is needed to further substantiate our findings.