Both female sex and intake of vitamin supplements are known to be associated with a low mean plasma homocysteine level. In the present study, we used multiple logistic regression analyses to investigate the relation between these two variables and plasma homocysteine at the extreme ends of the plasma homocysteine distribution curve. We propose that the obtained set of odds ratios for hyper- and hypohomocysteinemia, referred to as the opposite tails odds ratios, may be an alternative approach to study determinants of plasma homocysteine.