Background
Hip fractures are a major public health problem worldwide, contributing to decreased quality of life and premature death. In the United States, over 280,000 people over the age of 65 experienced a hip fracture in 2007. In 2005, the estimated total cost of hip fractures in the US was $12 billion and was estimated to increase 50% by 2025. More than two-thirds of all hip fractures occur in women and older age significantly increases the risk of fracture with those 85 and older having a ten-fold risk compared to those who are 60–65. Other factors known to be associated with hip fracture incidence include low body mass index, European or Asian race/ethnicity, smoking and less physical activity. A previous analysis of Women's Health Initiative (WHI) Observational Study (OS) data identified each of these factors as important predictors of hip fracture.
Light-to-moderate alcohol consumption has been shown to be associated with reduced risk of hip fracture and increased bone density. More precisely, a U-shaped relationship has been observed in which non-drinkers and heavy drinkers have an elevated risk of hip fracture compared to light-moderate drinkers. The 2010 Dietary Guidelines for Americans defines moderate alcohol consumption as up to one drink per day for women.
While various studies suggest that alcohol consumption may be related to hip fracture, the risk of hip fracture may be different for those who consume beer, wine, and hard liquor as was observed for cardiovascular disease and overall mortality. In a study of 31,785 men and women in Denmark, Høidrup et al. assessed preference of alcohol type among current alcohol consumers and found that those who preferred wine had a reduced risk of hip fracture compared to those who preferred beer or liquor. In contrast, in their analysis of the Cardiovascular Health Study, Mukamal et al. found that the reported consumption of beer, wine and hard liquor did not have a significant association with hazard of hip fracture. The discrepancy may be due to differences in the populations studied; Høidrup et al. studied Danish adults and Mukamal studied US adults over 65 from four communities. Another possible reason for the discrepancy may lie in how the exposure variable was defined. Whereas Høidrup et al. modeled alcohol preference, Mukamal et al. modeled consumption of type using indicator variables. Our study seeks to reconcile inconsistencies in past studies of alcohol preference and to take past drinkers and non-drinkers into account. We will investigate the relationship between type of alcohol consumed and the risk of hip fracture in a large, ethnically and geographically diverse cohort of postmenopausal women in the United States with available data on potentially important confounders of the relationship including physical activity and falls.