Recent observations suggest that the risk of coronary artery disease (CAD) is associated with both the level and composition of the two major populations of apolipoprotein (apo)-defined high-density lipoprotein (HDL) particles: those containing both apo A-I and apo A-II [Lp(AI,AII)] and those containing apo A-I without apo A-II [Lp(AI)]. While sex hormones are known to affect HDL, their influence on these apo-defined HDL particles is not known. We have determined the effects of two triphasic oral contraceptive (OC) formulations on these HDL particles in healthy normolipidemic women aged 21 to 35 years. The formulations contain comparable quantities of ethinyl estradiol (EE) and either desogestrel (DG), a minimally androgenic progestin, or levonorgestrel (LN), a more androgenic progestin. Lipid and lipoprotein levels were measured during the third week of the normal menstrual cycle and the sixth month of OC use. The DG/EE formulation significantly increased total cholesterol (C) 15%, triglyceride (TG) 99%, phospholipid (PL) 17%, apo A-I 28%, apo A-II 34%, apo B 21%, very-low-density lipoprotein cholesterol (VLDL-C) 238%, HDL-C 20%, and HDL3-C 28% (P < .02 to .005, n = 11), but not low-density lipoprotein cholesterol (LDL-C). The LN/EE formulation also increased total C 15%, TG 33%, apo A-I 15%, HDL3-C 21% (P < .05, n = 10), apo B 30% (P < .005), and, additionally, LDL-C 19% (P < .05). Both formulations increased Lp(AI,AII) (DG/EE, 34%, P < .005; LN/EE, 24%, P < .01). These changes reflected comparable increases of small (7.0 to 8.2 nm) and medium (8.2 to 9.2 nm) particles in the LN/EE group and a predominant increase of medium-sized particles in the DG/EE group. Also, in the LN/EE group but not the DG/EE group, there were fewer large (9.2 to 11.2 nm) particles. Lp(AI) increased only in the DG/EE group (25%, P = .075) and was due to the presence of more large particles. The level of Lp(AI) did not change in the LN/EE group, but the lipid/A-I ratio of these particles was lower (P = .012) and there were more small particles. Thus, triphasic OC formulations with progestins of different androgenicity had different effects on VLDL, LDL, and the level and composition of HDL particles with and without apo A-II, possibly reflecting estrogen/progestin/androgen balance. Estrogen dominance increases both Lp(AI,AII) and Lp(AI) and favors large Lp(AI) particles, while progestin/androgen dominance increases only Lp(AI,AII) and favors small particles. Because of the importance of HDL in the arterial wall physiology, OC formulations with different estrogen and progestin content may affect arterial wall health to a different extent.