Description
This study used data from the Third National Health and Exam Survey in order to examine the association between weight cycling and low bone density. A household interview and a physical examination were conducted for each survey participant. Physical examinations were conducted in specially equipped mobile examination centers (MEC); examinations were conducted in specially equipped mobile examination centers (MEC); were taken using the DXA by trained technicians. After combining the data from the exam, lab, and adult household interview surveys there were 31,808 observations available for this study. Acceptable bone mineral density (BMD) scans were performed on 14,646 men and women aged 20 and older. Only non-Hispanic white females aged 40-69 years of age were included in the current study. This left a sample size of 1348 women. From the home interview information on health practices, alcohol, tobacco, drug use, and physical activity was obtained from study participants. A family questionnaire asked questions about race and ethnicity, family income, and other demographic characteristics of the household. As part of the exam component women were asked several questions which were used to determine menopausal status. Univariate analyses and multivariate analyses were used to examine the independent and adjusted associations between risk factors and the outcome. Significant independent associations with bone mineral density (BMD) were found with weight cycling, age category, menopause, body mass index (BMI), maternal family history of hip fracture or osteoporosis, and smoking. Prior fracture, exercise, and alcohol use were not statistically significant. Weight cycling, BMI, and exercise had unadjusted T-score means positively associated with BMD; age, menopause status, maternal family history of hip fracture or osteoporosis, prior fracture, smoking, and alcohol use had unadjusted T-score means negatively associated with BMD. Because all covariates were well known risk factors for low bone density they were left in the final model. In the multivariate analysis of covariance model, weight cycling was statistically significant (p-value .0057) and had adjusted T-score means which were positively associated with BMD. Age category (p-value <.0001), post-menopause status (p-value <.0001), family history (p-value .0009), prior fracture (p-value .0239), and smoking (p-value <.0001) had significant inverse associations with BMD. Alcohol was not statistically significant (p-value 0.8428). Future research on the association of weight cycling and bone mineral density should include a spinal BMD measurement since the literature suggests that weight loss related bone loss is not as easily restored upon weight regain in the spine. Trabecular bone, which is found mostly in the spine, is also the type of bone which is lost more rapidly upon the transition towards menopause.
