Association of white matter hyperintensities with BMD, incident fractures, and falls in the UK Biobank cohort

Abstract

Osteoporosis is the most common metabolic bone disease globally, which increases the healthcare service burden. Recent studies have linked higher white matter hyperintensities (WMH) to reduced BMD, increasing the risk of fractures and falls in older adults. However, limited evidence exists regarding the dose-response relationship between WMH and bone health in a larger and younger population. Our study aimed to examine the association of WMH volume with BMD, incident fractures and falls, focusing on dose-response relationship with varying levels of WMH volume. We included 26 410 participants from the UK Biobank. The association between WMH volume and BMD was analyzed using multiple linear regression. Cox regression models were used to estimate the hazard ratios of incident fractures and falls. Restricted cubic spline (RCS) fitted for linear and Cox regression models were employed to explore potential non-linearity. Over a mean follow-up time of 3.8 yr, we documented 59 hip fractures, 392 all fractures, and 375 fall incidents. When applying RCS, L-shaped relationships were identified between WMH volume and BMD across all 4 sites. Compared with those in the lowest fifth of WMH volume, individuals in the second to the highest fifths were associated with a reduction of 0.0102-0.0305 g/cm2 in femur neck BMD, 0.0075-0.0273 g/cm2 in femur troch BMD, 0.0173-0.0345 g/cm2 in LS BMD, and 0.0141-0.0339 g/cm2 in total body BMD. The association was more pronounced among women and younger participants under age 65 (Pinteraction < .05). Per 1 SD increment of WMH volume was associated with 36.9%, 20.1%, and 14.3% higher risks of incident hip fractures, all fractures, and falls. Genetically determined WMH or apolipoprotein E genotypes did not modify these associations. We demonstrated that a greater WMH was associated with BMD in an L-shaped dose-response manner, especially in women and those under 65 yr.

Keywords: BMD; aging; fall; fracture; osteoporosis; white matter hyperintensities.

Figure 1

Nonlinear associations between the volume of WMH and BMD. RCS models fitted for linear regression for BMD at total body, LS, femur neck, and femur troch. The dashed vertical lines represent the mean values of the WMH volume (4.8 cm³). Model I: adjustment for age, sex, race/ethnicity, assessment center, BMI, Townsend deprivation index, and the volume of estimated total intracranial. Model II: additional adjustment for physical activity (MET-min/wk), smoking status, alcohol intake, serum vitamin D, serum calcium, and serum phosphate.

Figure 2

Nonlinear associations between the volume of WMH and incident fractures and falls. RCS models fitted for Cox proportional hazards models for fractures and falls. The vertical dashed lines represent the mean values of the WMH volume (4.8 cm³). The horizontal dashed line indicates HR = 1. Model I: adjustment for age, sex, race/ethnicity, assessment center, BMI, Townsend deprivation index, and the volume of estimated total intracranial. Model II: additional adjustment for physical activity (MET-min/wk), smoking status, alcohol intake, serum vitamin D, serum calcium, and serum phosphate.

Figure 3

Association of total volume of WMH with bone health and diseases stratified by sex. The HR is presented as HR per 1 SD in WMH volume. The model was adjusted for age, sex, race/ethnicity, assessment center, BMI, Townsend deprivation index, and the volume of estimated total intracranial. All adjusted P values were corrected using the Benjamini–Hochberg (B–H) method for multiple hypothesis correction.