Context: A low response to controlled ovarian hyperstimulation implies a reduced number of embryos and impaired pregnancy rate. Follicular priming with steroids before controlled ovarian hyperstimulation has been suggested to improve the subsequent ovarian response.Objective: The purpose of this study was to determine the best follicular priming protocol in low responders and to investigate the intrafollicular mechanisms triggered by steroid hormone priming. Design: This was a single-center, randomized, parallel, open-label, controlled trial, in two phases.Setting: The setting was a university-based in vitro fertilization unit.Patients: Potential low responders (n = 99) underwent a first intracytoplasmic sperm injection cycle. Confirmed low responders (n = 66) were randomized to different priming protocols before a new intracytoplasmic sperm injection.Interventions: Randomized patients underwent one of the following priming strategies: transdermal testosterone (20 μg/kg/d), transdermal estradiol (200 μg/d), or combined estrogens and oral contraceptive pills (30 μg of ethinyl estradiol plus 150 μg of desogestrel administered during the luteal phase of two consecutive cycles) and 4 mg/d of estradiol valerate during the follicular phase between them.Main Outcomes Measures: Metaphase II (MII) oocytes were retrieved. Gene expression levels in the granulosa cells of steroidogenesis enzymes and FSH, LH, and androgen receptors were measured.Results: The number of retrieved MII oocytes did not differ between the interventional groups (testosterone, 2.2 ± 2.0; estrogen, 2.7 ± 1.7; and combined estrogens and oral contraceptive pills, 2.0 ± 1.3; not significant). Compared with those in nonprimed cycles, estradiol pretreatment yielded more MII oocytes (primed, 2.7 ± 1.7; nonprimed, 1.6 ± 1.2; P = .029) although the clinical pregnancy rate was higher in patients treated with testosterone (P = .003). Testosterone pretreatment increased androgen receptor expression (P = .028) compared with that for the previous cycle without priming.Conclusions: The results of the present trial do not support the superiority of one priming strategy over the others.