2004 OPEN FORUM Abstracts
Smoking and Bone Density in Men: Results from a National Survey
E. Anders, M.P.H., RRT,* Lori W. Turner, Ph.D, RD,†
and Horace Spencer, M.S.* *University of Arkansas for Medical
Sciences, Little Rock, Arkansas University
of Arkansas, Fayetteville, Arkansas
Background. Osteoporosis is an insidious, chronic disease that precipitates debilitating, costly, and sometimes lethal fractures. One-eighth of men will suffer osteoporotic fractures in their lifetime. Though men have a lower incidence of hip fractures than women, they are twice as likely to die in the year afterward. National selection criteria for osteoporosis screening with bone density measurement for women rely on identification of major risk factors, including smoking, but selection criteria for men are lacking. The purpose of this study was to determine correlates of bone density in men.
Methods. Sample: The Centers for Disease Control collected data from a nationally representative sample via a stratified, multistage probability design. The study extracted data using inclusion criteria of males, ages 20 through 90 years, for who bone density data were available. The study excluded those who were institutionalized and minorities. Testing instrument: The testing instrument was the National Health and Nutrition Examination and Survey III. Dual energy x-ray absorptiometry measured bone density. The Mayo Clinic reviewed each test. Random test-retest demonstrated reliability. Statistical analysis: The study employed statistical software, SPSS™, to compute all statistical tests. Univariate linear regression tested each of the following as correlates of bone density: age, weight, smoking pack-years, physical activity, chronic disease, weight loss, calcium intake, alcohol intake, serum retinyl esters, and serum vitamins C and E. Each independent variable with a p value < 0.25 in the univariate analysis was then retested for significance in a full-model multivariable linear regression and subsequently in a reduced-model multivariable linear regression at an alpha = .01. Covariates in the reduced-model were tested for both collinearity and two-way interactions.
The study sample included 2,930 subjects, with a mean age = 54.87 (+ 19.61) years and a mean weight = 81.49 (+ 15.18) kg.
Bone Density: Reduced Multivariable Model
|Standardized β||p value|
|Age||(-) .630||< .001|
|Age x weight||.352||< .001|
|Physical activity||.074||< .001|
|Smoking pack-years||(-) .060||< .001|
|Vitamin C||.057||< .001|
Our study results provided sufficient evidence to suggest that
in men: (a) weight, physical activity, and Vitamin C were associated
with increased bone density; and (b) age and smoking pack-years were
associated with decreased bone density. Interaction existed between
age and weight; however the slopes resulting from regression models
of bone density as functions of quartiles of age and weight were