The objectives of this study were to define the "normal" position of the T-shaped intrauterine device (IUD) in the endometrial cavity, using transvaginal ultrasound, and to compare the IUD's position in women with and without complaints of bleeding and/or pain. Vaginal ultrasound was carried out in women who had used a T-Cu 200 or T-Cu 380 for at least 6 months. Of these women, 236 registered bleeding complaints and/or pain and 245 had no complaints. The distances, IUD-endometrium, IUD-myometrium, and IUD-fundus, were measured in each woman. Our results showed that women with and without complaints presented no significant differences in age, parity, type of IUD, duration of use, previous contraceptive method used, hysterometry, uterine position, or period of the cycle in which the ultrasound was performed. The values of the three distances in users without complaints were widely dispersed. The 90th percentile of the IUD-endometrium, IUD-myometrium, and IUD-fundus distances were 7, 11, and 27 mm, respectively. The curve of the values for the three distances among women with complaints was similar to that of the subjects without complaints. Using the generally accepted measurement of 20 or 25 mm as the limits of normality of the distance IUD-fundus, and the 90th percentile of the distance IUD-endometrium as the gold standard, 77% and 43% of the women were false positives for "incorrect IUD position." Similar results were obtained when the 90th percentile of the distance IUD-myometrium was used as a gold standard. We concluded that the 90th percentiles of the distance IUD-endometrium, IUD-myometrium, and IUD-fundus were 7, 11 and 27 mm, respectively, among users without complaints. The ultrasonographic diagnosis of the IUD position was unable to discriminate between women who did or did not have complaints of bleeding and/or pain. If ultrasound were to be performed in T-shaped IUD users, the IUD-myometrium distance is likely to be the most reliable measurement.