Table 3.

Results From Marginal Structural Models for Estimating the Controlled Direct Effect of Fiber Intake (≥22 g/day vs. <22 g/day) on Log Lipoprotein Levels (mg/dL) Among Women Participating in the BioCycle Study, Buffalo, New York, 2005–2007

Lipoprotein and Estradiol Level, pg/mL	Estimate of Controlled Direct Effecta	95% Confidence Interval	P for Interactionb
Total cholesterol			0.3
30	−0.048	−0.081, −0.014
45	−0.045	−0.076, −0.013
110	−0.039	−0.070, −0.006
High density lipoprotein cholesterol			0.9
30	−0.013	−0.059, 0.032
45	−0.012	−0.056, 0.033
110	−0.009	−0.053, 0.035
Low density lipoprotein cholesterol			0.7
30	−0.062	−0.111, −0.015
45	−0.062	−0.111, −0.018
110	−0.064	−0.110, −0.020
Triglycerides			0.03
30	−0.078	−0.184, 0.027
45	−0.053	−0.151, 0.044
110	0.004	−0.092, 0.099

^aEstimates were adjusted for age, body mass index, physical activity, total energy intake, vitamin E intake, luteinizing hormone, follicle-stimulating hormone, and progesterone levels through the use of inverse probability weights. Direct effects are the effects of fiber on lipoprotein cholesterol levels that are not mediated through estradiol. The total effects shown in Table 2 include the mediating effects. The direct effects shown here are larger than the total effects shown in Table 2, suggesting that estradiol mediates at least part of the association.

^b P value for the interaction between fiber and estradiol.