BACKGROUND Vascular endothelial growth factor (VEGF) has been investigated as a marker of ovarian response to controlled ovarian hyperstimulation and as a predictor of ovarian hyperstimulation syndrome (OHSS) in IVF cycles. In most studies, serum has been used for circulating VEGF concentration measurement, but it has been suggested that plasma is the preferred medium to measure VEGF levels because of the potential contribution of VEGF released from platelets during blood clotting. This study investigated VEGF concentrations in paired serum and plasma samples from patients undergoing controlled ovarian hyperstimulation for IVF. METHODS Serum and plasma VEGF levels, as well as the number of platelets, were measured in 30 IVF patients who comprised three study groups delineated according to the estradiol (E(2)) serum concentration reached on the day of HCG administration: 10 patients having low E(2) serum levels (<1500 pg/ml, group L), 10 patients having intermediate E(2) serum levels (1500-3000 pg/ml, group I) and 10 patients having high E(2) serum levels (>3000 pg/ml, group H). RESULTS There was a statistically significant correlation between plasma and serum VEGF levels (rho = 0.61; P < 0.005) for the entire population studied, although serum values were higher by a factor of approximately 6-fold. No significant correlation was found between peripheral blood VEGF concentrations and serum E(2) or follicle number on HCG day or the number of oocytes collected. Similarly, paired serum and plasma VEGF measurements did not correlate with platelet count. CONCLUSIONS Serum and plasma VEGF concentrations are strongly correlated in paired samples from infertile patients undergoing controlled ovarian hyperstimulation. However, neither serum nor plasma VEGF levels were correlated with parameters associated with ovarian follicular activity. Peripheral blood VEGF levels were not correlated with platelet count.