Table 4.

Linear regression coefficients (brute and adjusted) for bone mineral density according to sociodemographic, lifestyle, anthropometric, and body composition variables, in elderly men. Goiânia, GO, 2009. n = 132.

Variables	β brute	p value#	Hierarchic multivariate analysis
			β Aj ‡	p-value#	R 2
1st level					0.29
Age
60–64	1.00		1.00
65–69	−0.046	0.355	−0.047	0.323
70–74	−0.047	0.323	−0.030	0.520
75–79	−0.064	0.274	−0.055	0.320
Over 80	−0.166	0.013	−0.158	0.012
Years of school
0–4	−0.102	0.009	−0.090	0.018
5–7	−0.115	0.008	−0.120	0.005
Over 8	1.000	—	1.00
2nd level					0.44
Use of tobacco∗
Nonsmoker	1.000		1.000
Ex-smoker	−0.025	0.470	0.004	0.896
Smoker	−0.102	0.063	−0.068	0.221
BMI (WHO)
Low weight	−0.115	0.299	−0.089	0.400
Eutrophic	−0.154	0.001	−0.141	0.003
Overweight	−0.032	0.461	−0.039	0.376
Obese	1.00	—	1.00	—
BMI (Lipschitz)
Low weight	−0.161	0.000	−0.149	0.002
Eutrophic	−0.096	0.004	−0.059	0.110
Overweight	1.00	—	1.00	—
BFP∗
Adequate	−0.044	0.185	0.488	0.246
Augmented	1.000	—	1.000
WC∗
Adequate	−0.108	0.007	−0.020	0.720
Enlarged	−0.062	0.116	−0.017	0.718
Very enlarged	1.000	—	1.000
WHR∗
No risk for CNTD	0.064	0.158	0.059	0.316
At risk for CNTD	1.000	—	1.000

^∗Due to loss of statistical significance, was not maintained in the final model of multivariate linear regression (p > 0.05).

^#Linear regression test.

^‡Adjusted by age and years of school (education level).