The purpose of the present study was to evaluate the use of colour Doppler sonography of ascendent uterine artery perfusion in 91 patients undergoing in-vitro fertilization (IVF)-embryo transfer treatment after ovarian stimulation with a depot formulation of goserelin and recombinant human follicle stimulating hormone according to the long protocol. Resistance index (RI), pulsatility index (PI), maximum peak velocity (Vmax) and minimum diastolic velocity (Vmin) were assessed for the left and right ascending uterine artery on day 1 of ovarian stimulation, on day -2 [the day of human chorionic gonadotrophin (HCG) application] and on day +14 (12 days after embryo transfer). The data of 75 patients who had at least two cleaved preimplantation embryos available for transfer were analysed: 21 patients became pregnant resulting in a pregnancy rate of 28% (21/75). After exclusion of biochemical, ectopic and abortive pregnancies (n = 5), the data from 54 non-pregnant patients and 16 pregnant patients were analysed. No differences with respect to patient age, current cycle number, indication for IVF treatment, endometrial thickness at day -2 and serum oestradiol and serum progesterone concentrations at day -2 were found between the pregnant and non-pregnant groups. Compared to the non-pregnant patients the ascendent uterine artery flow of the pregnant patients showed significantly lower RI (P < 0.009) and PI (P < 0.03) values at the beginning of ovarian stimulation. Vmax and Vmin did not differ between the two groups. On day -2 no differences in RI, PI, Vmax and Vmin were found between pregnant and non-pregnant patients. On day +14 the flow in the ascendent uterine arteries of the pregnant patients showed significantly lower RI (P < 0.008) and PI (P < 0.03) values and significantly higher Vmax (P < 0.003) and Vmin (P < 0.0001) values. RI (P < 0.009) and PI (P < 0.003) values had decreased significantly and Vmax (P < 0.0002) and Vmin (P < 0.0001) had increased significantly on day +14 compared to the previous observation times in both the pregnant and non-pregnant groups. A significant correlation between the increase of serum progesterone concentrations and the decrease of RI (r = 0.68, P < 0.009), and the increase of Vmin (r = 0.67, P < 0.01) was only detected in the pregnant group. In conclusion, the differences found in RI and PI values at the beginning of ovarian stimulation were not clinically helpful as there was a wide overlap between non-pregnant and pregnant patients. The parameters currently used in colour Doppler assessment of uterine artery perfusion are not clinically helpful in discriminating prospectively which patients will and will not become pregnant in an IVF programme. In pregnant patients, increasing progesterone concentration is correlated with a significant decrease in impedance to uterine perfusion in the late luteal phase.