Concentrations of various proteins in uterine flushings have been described as a direct method for assessment of the secretory activity of the endometrium. We investigated levels of the endometrial protein known as placental protein 14 (PP14) in flushings obtained from 271 infertile women. Under transvaginal ultrasonographic control, 2 ml of 0.154 M sodium chloride solution were injected into the uterine cavity and re-aspirated, five times. In contrast to previous studies the recovered volume of each flushing was not consistent (range: 0.05-2.1 ml); the volume varied significantly between serial samples obtained from an individual (P = 0.02, one-way ANOVA), different cycle days (P < 0.0001, one-way ANOVA) and women with bilaterally blocked versus patent Fallopian tubes (P < 0.05, Student's t-test). Concentrations of PP14 showed a better correlation with protein content (r = 0.506, P < 0.0001) than with the recovered volume (r = 0.087, P = 0.095). We therefore corrected PP14 concentrations for total protein content as an indicator of the efficiency of the flushing process. Corrected PP14 concentrations varied significantly relative to time since the onset of menstruation (P = 0.001, Kruskal Wallis ANOVA) with higher levels on days 1-8, as previously observed in plasma samples. No significant difference in PP14 levels was found with different causes of infertility. This study shows that uterine flushing is not a consistent process in women with differing physical characteristics and at varying times throughout the menstrual cycle.