Two different regimens of luteal support in gonadotrophin hormone-releasing hormone (GnRH) analogue/human menopausal gonadotrophin (GnRHa/HMG)-induced in-vitro fertilization cycle (IVF) were compared in a randomized clinical trial. After embryo transfer, either vaginal progesterone alone was administered (n = 89, P group), or a combination of vaginal progesterone and human chorionic gonadotrophin (n = 87, P/HCG group). The primary aim of this study was to assess the effect of the different regimens of luteal support on the pregnancy rate. The secondary aim was to compare oestradiol and progesterone concentrations in the luteal phase between the two groups, and assess their effect on the pregnancy rate. A clinical pregnancy rate of 15% was found in the P/HCG group in comparison with 26% in the P group (odds ratio 0.49; 99% confidence interval: 0.18-1.3). The luteal serum oestradiol and progesterone values in the P/HCG group were significantly higher when compared with the P group on the 6th, 9th and 12th day after oocyte retrieval (Wilcoxon P < 0.001). In accordance with the high oestradiol concentrations, more cases of ovarian hyperstimulation syndrome (OHSS) were found in the P/HCG group. Oestradiol values on the 9th day after oocyte retrieval, presumably the day of implantation, appeared to be higher in women who did not become clinically pregnant. We conclude that vaginal progesterone alone provides sufficient luteal support in GnRHa/HMG induced IVF cycles. The combination of vaginal progesterone and HCG as luteal support leads to significant high luteal oestradiol and progesterone concentrations. But a high concentration of oestradiol seems to have a deleterious effect on the implantation process, resulting in a low pregnancy rate.