Table II.

Results of explanatory multiple linear regression analyses: SI (log) models

Main exposure variables	Model 1a Nc=1,041 (Nd=180)			Model 2b Nc=1,135 (Nd=194)
Contaminant	R2	Regression coefficient (SE)e	p-Valuef	R2	Regression coefficient (SE)e	p-Valuef
DLCs (pg TCDD-EQ/L)	0.4184	1.6468 (1.3199)	0.2128	0.3303	−0.1245 (1.5045)	0.9341
DL-PCB 105 (ng/L)	0.4170	1.8651 (1.7705)	0.2926	0.3307	−0.5484 (1.7947)	0.7601
DL-PCB 118 (ng/L)	0.4163	0.3552 (0.4085)	0.3850	0.3339	−0.3153 (0.3618)	0.3840
DL-PCB 156 (ng/L)	0.4141	−0.1520 (0.7529)	0.8401	0.3462	−1.5856 (0.6564)	0.0161
DL-PCB 157 (ng/L)	0.4242	−4.5409 (2.3586)	0.0548	0.3719	−7.1843 (1.6857)	<0.0001
DL-PCB 167 (ng/L)	0.4143	−0.9897 (3.2387)	0.7601	0.3499	−6.5369 (2.6521)	0.0140
DL-PCB 189 (ng/L)	0.4217	−7.7779 (5.0742)	0.1260	0.3673	−14.2291 (4.2063)	0.0008
ΣDL-PCBsg (ng/L)	0.4142	0.0712 (0.2391)	0.7660	0.3422	−0.3521 (0.2110)	0.0958

^aModel adjusted for age, total plasma lipids, menopausal status, education level, supplement use (calcium and vitamin D), blood mercury levels, blood lead levels, omega-3 PUFA, omega-6 PUFA, vitamin A (retinol), vitamin D, tobacco use, parity, height, weight

^bmodel adjusted for age and total plasma lipids

^cweighted and rounded size

^dsample size

^ethe values of the regression coefficients and standard errors were multiplied by 10⁴

^fWald Chi square test with Satterthwaite correction for the degrees of freedom

^gsum of DL-PCBs nos. 105, 118, 156, 157, 167 and 189