Associations between dietary patterns and osteoporosis-related outcomes in older adults: a longitudinal study

Background/Objectives To describe the associations of baseline dietary pattern scores with falls risk, bone mineral density (BMD), and incident fractures measured over 10 years in older adults. Subjects/Methods Dietary patterns were identified using exploratory factor analysis. Femoral neck (FN), hip, and lumbar spine (LS) BMD were measured using dual-energy X-ray absorptiometry, falls risk z-score using the Physiological Profile Assessment, and incident fractures by self-report. Linear mixed-effects models and log-binomial regression were used to estimate associations between baseline dietary pattern z-scores and outcomes. Results Of 1098 participants at baseline, 567 were retained over 10 years. Four dietary patterns were derived: fruit and vegetable (FV), animal protein (AP), snack, and Western. FV pattern reduced falls risk at baseline by beta = 0.05-0.08/SD and the annual decreases of FN and hip BMD were less for higher Western or AP pattern scores in all populations and women. The annual increase in LS of the entire population was greater with higher scores of FV, AP, and Western patterns (all beta = 0.001 g/cm(2)/year/SD, p < 0.05). Higher scores of FV and snack were associated with a higher risk of LS BMD increasing over 10 years (p < 0.05 for all, except snack pattern in men) and incident fracture was not associated with any dietary pattern in the overall cohort and both men and women separately. Conclusions An FV dietary pattern may be beneficial for reducing falls risk. The associations of dietary patterns and BMD are modest in magnitude and did not translate into an improved fracture risk. Associations between diet and LS BMD may reflect osteoarthritis rather than osteoporosis.