Association of overall diet quality with falls and physical function among community dwelling older adults - results from MrOS study

摘要

Falls among people aged 65 and older are recognized as an issue of public health significance in the United States which faces an increasingly aging society. The current study tests the hypothesis that lower diet quality is prospectively associated with incident falls and whether physical function as assessed by gait speed is a mediator of this association. For our prospective analysis, we studied 5849 men enrolled in the Osteoporotic Fractures in men study (MrOS). All men in MrOS were ≥ age 65 at baseline. Overall diet quality as assessed by Healthy Eating Index (HEI-2010) scores were calculated from a reduced version Block 98 food-frequency questionnaire and categorized into quartiles. An indicator variable for incident falls was computed using falls categorized by a yes/no response and number of falls reported at follow-up of 12 months. Recurrent falls were classified as (≥ 2) vs single (1) or no fall (0). Gait speed was measured in meters/second on a standard six meter walking course. Multiple logistic regression methods were used to test the association between HEI-2010 scores and incident falls after adjusting for age, race, clinic site, smoking status, physical activity, Body Mass Index, diabetes, stroke, parkinson’s disease, glaucoma, selective serotonin reuptake inhibitor antidepressants, tricyclic antidepressants, antidepressants, benzodiazepines and gait speed.udAt the end of the follow-up period of 1 year, 1473 (25.7%) experienced a fall (≥ 1), 685 (11.7%) experienced recurrent falls (≥ 2). Compared to older men with the highest HEI-2010 scores, those with the lower HEI-2010 scores had a similar risk of falls. There was no significant association with falls in the unadjusted or adjusted models. Additional adjustment for gait speed showed similar associations. For recurrent falls, compared to men with the highest HEI-2010 scores, older men with the lower HEI-2010 scores had a higher risk of recurrent falls in the unadjusted model which remained significant after adjustment for age, race, clinic site (p trend=0.02), lifestyle factors including smoking status and physical activity (p trend=0.03) and BMI (p trend=0.04). Further adjustments for comorbidities reduced the overall significance of the trend (p trend=0.07). Finally adjusting for gait speed, there was no significant association although the risk for falls marginally increased (p trend=0.12). We concluded that there was no significant association between diet quality and incident falls.ud