Significant differences in serum selenium concentration according to contraceptive treatment and age have been evidenced in women of the SU.VI.M.AX cohort. This study aimed at verifying the physiopathological hypothesis that the observed increase in serum selenium concentration could be related to serum lipid increase and/or bleeding decrease. Women were divided into six groups: menopausal with or without hormonal replacement therapy; non-menopausal using contraceptive pills; intrauterine device; other contraceptive treatment or no contraceptive treatment. Adjusted linear regression indicated positive associations between selenium and apolipoprotein A1 (r(2) from 0.038 to 0.074, p<0.07 depending on groups) or ferritin in serum (r(2) from 0.032 to 0.075, p<0.07 depending on groups). These relationships could explain the differences observed according to hormonal treatment and age in the SU.VI.MAX study.