Adherence to a Mediterranean diet and risk of fractures in French older persons

Prevention of fractures is a considerable public health challenge. In a population-based cohort of French elderly people, a diet closer to a Mediterranean type had a borderline significant deleterious effect on the risk of fractures, in part linked to a low consumption of dairy products and a high consumption of fruits. Introduction Higher adherence to the Mediterranean diet (MeDi) is linked to a lower risk of several chronic diseases, but its association with the risk of fractures is unclear. Our aim was to investigate the association between MeDi adherence and the risk of fractures in older persons. Methods The sample consisted of 1,482 individuals aged 67 years or older, from Bordeaux, France, included in the Three-City Study in 2001-2002. Occurrences of hip, vertebral and wrist fractures were self-reported every 2 years over 8 years, and 155 incident fractures were recorded. Adherence to the MeDi was evaluated at baseline by a MeDi score, on a 10-point scale based on a food frequency questionnaire and a 24-h recall. Multivariate Cox regression tests were performed to estimate the risk of fractures according to MeDi adherence. Results Higher MeDi adherence was associated with a nonsignificant increased risk of fractures at any site (hazard ratio [HR] per 1-point increase of MeDi score=1.10, P=0.08) in fully adjusted model. Among MeDi components, higher fruits consumption (>2 servings/day) was significantly associated with an increased risk of hip fractures (HR=1.95, P=0.04), while low intake of dairy products was associated with a doubled risk of wrist fractures (HR=2.03, P=0.007). An inverse U-shaped association between alcohol intake and risk of total fracture was observed (HR high vs. moderate=0.61, P for trend=0.03). Conclusions Greater MeDi adherence was not associated with a decreased risk of fractures in French older persons. The widely recognized beneficial effects of the MeDi do not seem to apply to bone health in these people.