Determinants of low bone mineral density in people with multiple sclerosis: Role of physical activity

•Femoral neck bone mineral density is decreased in people with multiple sclerosis.•Physical activity, depression, and fatigue were the greatest contributors.•Other independent contributors were disability and inflammation. People with multiple sclerosis (PwMS) have reduced bone mineral density (BMD), but the causes are unclear. Some factors that may cause reduced BMD in PwMS have been understudied, including physical activity, inflammation, cortisol, symptomatic fatigue, and depression. The aim of this study was to investigate factors that may uniquely contribute to reduced BMD in PwMS as compared to people without MS. We hypothesized that physical activity would be the primary determinant of low BMD in PwMS, with additional contributions from inflammation and sympathetic nervous system activation. We tested 23 PwMS (16 women; median EDSS: 2) and 22 control participants (16 women). BMD was measured from the femoral neck and lumbar spine with dual x-ray absorptiometry. Disability was measured with the Expanded Disability Status Scale, and functional capacity was measured with the Multiple Sclerosis Functional Composite. Questionnaires measured symptomatic fatigue and depression. A blood draw was used to measure calcium, phosphate, vitamin D, N-terminal telopeptide, osteopontin, and cytokine markers of inflammation. Physical activity was measured with accelerometry. Salivary cortisol and cardiac heart rate variability also were obtained. All outcome variables were compared between groups with independent samples t-tests. Variables that were different between groups and significantly correlated (Pearson product-moment) with femoral neck BMD, were included in a theoretical model to explain femoral neck BMD. The expected direction of relations in the theoretical model were developed based upon the results of previous research. A Bayesian path analysis was used to test the relations of predictive variables with femoral neck BMD and interrelations among predictive variables, as detailed in the theoretical model. PwMS had lower BMD at the femoral neck than controls (p = =0.04; mean difference: -0.09; 95% CI: -0.2, -0.004; Cohen's d = =0.65), and there was a smaller, statistically non-significant difference in BMD at the lumbar spine (p = =0.07; mean difference: -0.08; 95% CI: -0.17, 0.007; Cohen's d = =0.59). PwMS also had lower functional capacity (p ≤ 0.001; Cohen's d = =1.50), greater fatigue (p