Background: Discriminant analysis (DA) was performed on data of two combined hormonal contraceptives (CHC) differing in estrogen ratio to explore whether a combination of variables rather than a single variable distinguishes CHCs better.
Study design: Data were used of a parallel study in premenopausal women treated for three cycles (21 days on, 7 days off) with a contraceptive vaginal ring delivering Nestorone and ethinyl estradiol (EE) or an oral contraceptive containing levonorgestrel and EE. DA was performed on the change from baseline (CFB) and the end-of-treatment values at 3 months for lipids, sex-hormone binding globulin (SHBG), C-reactive protein, angiotensinogen, blood pressure and hemostasis variables, and on the hemostasis variables only.
Results: For the complete set, the CFB for factor VII (FVII), SHBG and plasminogen (PLG), or end-of-treatment SHBG- and FVII level discriminated the treatments best. Maximal discrimination for the hemostasis data was by CFB for FVII and PLG or end-of-treatment FVII level.
Conclusions: DA identifies differences between CHCs and may provide information on the factors associated with thrombotic risk.
Copyright © 2012 Elsevier Inc. All rights reserved.