The objective of this study was to compare the implantation rate, pregnancy rate and endometrial thickness of frozen-thawed embryo transfers using endometrial preparation with either an artificial cycle or stimulated cycle. This was a prospective randomized trial at a single academic IVF centre. Seventy-seven patients undergoing artificial cycles received oral oestradiol; patients with endometrium < 7 mm on day 9-10 were switched to vaginal oestradiol. Eighty-six patients undergoing stimulated cycles received recombinant FSH followed by human gonadotrophin hormone injection. Vaginal progesterone was begun 2 or 3 days prior to embryo transfer. There was no difference in implantation rate (8.5% versus 7.3%), pregnancy rate (16% versus 13%), cancellation rate (both 23%) or endometrium thickness (8.7 +/- 1.1 mm versus 8.7 +/- 1.0 mm) between artificial and stimulated cycles. Stimulated cycles had a higher incidence of thin endometrium (27% versus 5%, P < 0.01). In artificial cycles, patients switched to vaginal oestradiol had improved pregnancy rate (31%) versus patients who received oral oestradiol alone (13%) (P = 0.05). It is concluded that artificial and stimulated cycles produce comparable pregnancy rates, implantation rates, cancellation rates and endometrial thickness, although stimulated cycles have a higher incidence of thin endometrium. Vaginal oestradiol supplementation improved implantation rates.