The objective of this study was to determine the relationship between serum estradiol levels on day 2 of an in vitro fertilization and embryo transfer (IVF-ET) cycle and the chance of pregnancy and implantation rates according to age. Two hundred and forty-eight cycles of IVF-ET in patients treated with gonadotropin-releasing hormone analog (GnRHa, follicle-stimulating hormone (FSH) and human menopausal gonadotropin (hMG) were divided into several groups according to the basal serum concentrations of estradiol (E2):A < 25 pg/ml, B 26-50 pg/ml, C 51-75 pg/ml, D 76-100 pg/ml, E > 100 pg/ml. Furthermore, the patients were subdivided into two subgroups according to age: I younger; and II, older than 35 years. Pregnancy and implantation rates were analyzed for each subgroup according to basal E2 level and age. Statistical analysis was performed using analysis of variance, chi2 and Fisher's test. The number(s) of cycles for each subgroup were 164 and 84 for I and II, respectively. Even though there were no significant differences among groups for the pregnancy rates for individual groups, there was a tendency for a decreased rate with increasing levels of E2 only in women older than 35 years of age. When pregnancy rates of all women with E2 levels above 25 pg/ml were calculated, the differences between those observed in subgroup I (< or = 35 years) and II (> 35 years) was significant: p = 0.02 (38.7% vs. 18.7%, respectively). When implantation rates in the same groups and subgroups were analyzed, we found that again the differences were statistically significant: p = 0.001 (13.1%vs, 4.3% for I and II). The results of the present studies reveal that basal levels of E2 are a marker of poor prognosis for implantation and pregnancy in assisted reproduction (ART) cycles only in patients over the age of 35 years. These data cast doubt on the use of basal levels of E2 to screen women below the age of 35 prior to initiating an ART cycle. In addition, high levels of basal E2 per se may not be deleterious for reproductive efficiency in younger women, since pregnancy and implantation rates did not differ among patients younger than 35 years old (IA to E).