Introduction Oral contraceptives (OC) are currently being used by 65 million women worldwide, approximately 6% of all women of reproductive age.1 One-fourth of all women of reproductive age in the United States use OCs, which are now more widely used in the United States than any other method, including sterilization.2 Their widespread use throughout the world for several decades suggests that both medical professionals and the lay public believe that the benefits of OCs outweigh potential side effects.1 Greater awareness of the metabolic effects of OCs and the role of smoking and hypertension in the development of cardiovascular disease has led to more selective prescribing and a substantial reformulation of the steroidal components.3 A variety of formulations are available, but all combination OCs contain two components: an estrogen, which maintains the endometrium and prevents breakthrough bleeding, and a progestin with sufficient activity to suppress pituitary gonadotropin secretion. Each of the steroids also has additional functions. The progestin component causes changes in the cervical mucus and the endometrium that enhance the antifertility effects of OCs should ovulation occur4 and the estrogen acts synergistically with the progestin to inhibit the pituitary.4 Since OCs were first introduced, both hormonal components have been substantially reduced in amount and reformulated. Prior to 1992, OCs in the United States contained one of two estrogens, either ethinyl estradiol (EE) or mestranol. Since then, all low-dose OCs (ie, preparations containing ,50 mg) have contained EE exclusively. Low-dose formulations are now the standard, accounting for almost all OC prescriptions. Prior to 1992, OCs in the United States contained one of four progestins: norethindrone; norgestrel and its active isomer, levonorgestrel; norethindrone acetate; or ethynodiol diacetate. All of these mediumand high-androgenic progestins were less selective than later formulations, and they were associated with a variety of adverse androgenic effects. They had a negative effect on carbohydrate and lipid metabolism3,5 and often produced acne,6 hirsutism,7 and weight gain.8,9 In 1992, low-androgenic progestins, norgestimate and desogestrel, were introduced in the United States. Because the affinity of these progestins for progesterone receptors is higher than for androgen receptors, their activity is more selective. Consequently, androgenic side effects associated with OCs containing these progestins are less common compared with the side effects of OCs containing high doses of the more androgenic progestins.10–14 All currently available OCs are highly effective for preventing pregnancy, but variations in their hormonal components are clinically important. Variations in progestin formulations, for example, modulate the overall metabolic and clinical effect of a given combination OC.3 Variations in estrogen dose among formulations containing ,50 mg EE may also affect cycle control. However, all low-estrogen dose formulations are safe for use by nonsmoking women without major cardiovascular risk factors such as uncontrolled hypertension. To assist clinicians in selecting the most appropriate OC for their patients, a panel of experts in gynecology and contraception has developed a set of practice guidelines for OC selection. These physicians, who are members of the editorial board of Dialogues in Contraception, reviewed the guidelines with a panel of 70 visiting faculty members.15,16 Together, they reviewed the clinical effects of the variations in OC formulations on safety, cycle control, side effects, and noncontraceptive benefits—the factors which, collectively, provide the basis for selecting one OC over another for a particular patient. (It must be recognized that these guidelines were developed within the context of countries with advanced medical systems and low maternal mortalities. In less developed countries, with high rates of maternal morality and limited health care resources, the use of any form of oral contraceptive with limited screening and follow-up may be quite appropriate.) Department of Obstetrics and Gynecology, Baystate Medical Center, Springfield, Massachusetts; and Department of Obstetrics and Gynecology, University of Tennessee at Memphis, Memphis, Tennessee Name and address for correspondence: Ronald T. Burkman, Chairman, Department of Obstetrics and Gynecology, Baystate Medical Center, 765 Chestnut St., Springfield, MA 01109