. Author manuscript; available in PMC: 2012 Oct 1.

Published in final edited form as: Bone. 2011 Jul 19;49(4):886–894. doi: 10.1016/j.bone.2011.07.006

Table 2.

a. Multiple regression models for E_FEM, VMCV and VMExp/σ_app. For comparison with previous work [59], models were constructed using a stepwise regression that included BV/TV.Av, BV/TV.SD, Tb.N.Av, Tb.N.SD, -#Euler/Vol and DA as explanatory variables. BV/TV.Av was forced to stay in the model in order to determine contributions from other variables that are independent of BV/TV.

b. Multiple regression models for E_FEM^(λ) VMCV^(λ) and (VMExp/σ_app) ^(λ) after elimination of Tb.N.Av. Models were constructed using a stepwise regression that included BV/TV.Av^(λ), BV/TV.SD^(λ), Tb.N.SD^(λ), -#Euler/Vol^(λ) and DA^(λ) as explanatory variables. BV/TV.Av^(λ) was forced to stay in the model in order to determine contributions from other variables that are independent of BV/TV.

Response	Term	Estimate	t ratio	p<

E_FEM (r²_adj=0.93; p<0.0001)	BV/TV.Av	2650.5	28.07	0.0001
	BV/TV.SD	−1962.5	−2.79	0.007
	DA	52.73	2.47	0.02
	Intercept	−190.3	−4.71	0.0001

VMCV (r²_adj=0.68; p<0.0001)	BV/TV.Av	−1.865	−10.95	0.0001
	BV/TV.SD	3.611	2.84	0.006
	DA	−0.2347	−6.10	0.0001
	Intercept	1.312	18.01	0.0001

VMExp/σ_app (r²_adj=0.78; p<0.0001)	BV/TV.Av	−40.99	−2.94	0.005
	Tb.N.Av	−5.764	−2.89	0.005
	DA	−2.992	−2.38	0.02
	Intercept	28.63	12.75	0.0001

Response	Term	Estimate	t ratio	p<

E_FEM (r²_adj=0.93; p<0.0001)	BV/TV.Av^(λ)	2852.5	24.03	0.0001
	BV/TV.SD^(λ)	−1915.0	−2.39	0.02
	DA^(λ)	59.27	2.76	0.008
	-#Euler/Vol^(λ)	−99.56	−3.15	0.003
	Intercept	−1915.0	−2.39	0.0001

VMCV^(λ)(r²_adj=0.74; p<0.0001)	BV/TV.Av^(λ)	−1.837	−12.95	0.0001
	BV/TV.SD^(λ)	4.209	3.55	0.0007
	DA^(λ)	−0.1781	−5.59	0.0001
	Intercept	−0.6122	11.89	0.0001

(VMExp/σ_app)^(λ) (r²_adj=0.98; p<0.0001)	BV/TV.Av^(λ)	−75.75	−45.62	0.0001
	DA^(λ)	−1.138	−3.83	0.0003
	-#Euler/Vol^(λ)	1.094	2.26	0.03
	Intercept	28.63	12.75	0.0001