Epidemiology of risk factors for osteoporosis and osteoporotic fractures

1,511 Men and women aged >65 years in the Norfolk arm of the European Prospective Investigation of Cancer (EPIC - Norfolk, UK) cohort study were prospectively followed up for 8 years since April 1995.  Data for prevalent and incident limb and vertebral fractures were obtained from 7223 men and women aged >50 years in the multicentre multinational European Prospective Osteoporosis Study (EPOS) and were used to investigate determinants of fracture risk. In the EPIC-Norfolk study, weight maintenance or gain and commonly practised forms of physical activities protected against BMD loss (P<0.013) but nutritional variation had less impact.  Men had higher values than women for bone geometry variables namely;  cross-sectional area, subperiosteal diameter, endosteal diameter, section modulus, average cortical thickness and lower buckling ratio after adjusting for age, weight, and height, (all P<0.0001) and could explain the lower prevalence of hip fractures in men.  Longitudinally the subperiosteal diameter, endosteal diameter and buckling ratio increased at a faster rate in women than men (P<0.05), while BMD and cortical thickness decreased faster in women (P<0.037).  There was evidence of infero-medial shifting of the centre of mass with ageing in hip regions for both genders.  The section modulus, a measure of bending resistance, was more strongly related to reported physical activity than BMD, and a higher level of reported lifetime physical activity was a key positive determinant of subperiosteal diameter.  The rate of subperiosteal diameter expansion was significantly faster in women with higher levels of sex hormone binding globulin (SHBG) adjusted for other key determinants (P=0.001).  In the EPOS study, between centre variation in BMD explained les of the variation in incidence of limb fractures in women across Europe compared to variation in fall rates.  In both men and women, the risk of prevalent vertebral fracture significantly increased with age, statural height loss, self-reported history of spine fracture and history of other major fracture (all P<0.05).  Higher body weight reduced risk.