Materials and Methods
This study was approved by the University of Iowa Human Subjects Committee. Asymptomatic premenopausal women age 30 and older were recruited through local advertisements to be studied by sonohysterography. They were compensated for participation. All women completed questionnaires detailing their menstrual, gynecologic, and obstetric histories. Women with abnormal menstrual bleeding (menses longer than 7 days, bleeding at intervals of 21 days or less, or menstrual bleeding leading to anemia) were excluded from the control arm. For comparison, we reviewed the charts of all premenopausal women evaluated by sonohysterography for abnormal uterine bleeding between January 1997 and January 1998 in our gynecologic ultrasound unit.
Standardized sonohysterography was done after a speculum was placed and the cervix was cleaned with an iodine-based solution. A 5.3-Fr Soules intrauterine insemination catheter (Cook Co., Spencer, IN) was placed in the uterine fundus. The speculum was removed and transvaginal ultrasonography was done. The uterine cavity was distended with 3–10 mL of sterile saline while the endometrium was viewed. All procedures were videotaped, and all lesions were recorded and confirmed by one of the authors (BJV). In asymptomatic women, sonohysterography was done at any time of the menstrual cycle other than during menses.
Intrauterine filling defects were noted during sonohysterography and were measured in the maximal two diameters. Defects were classified as polyps or myomas on the basis of sonographic findings. In general, polyps were considered present if a pedunculated or sessile and relatively hyperechoic mass was identified in the endometrial cavity. Submucous myomas were distinguished from polyps by their relatively hypoechoic echotexture, with sound attenuation in some cases. After sonohysterography, a routine transvaginal ultrasound examination of the pelvis was done and any myomas within the wall of the uterus were noted and measured.
To assess the accuracy of sonohysterography at our institution, we reviewed charts of all women who underwent sonohysterography, then had subsequent surgical procedures, to compare pathologic diagnoses with sonohysterographic diagnoses.
We calculated the necessary sample size for this study to be 65 for each group (α = .05, β = .80, calculation assumed a two-sided test, estimated prevalence of intrauterine lesions [polyps and myomas] in case patients 50%, estimated prevalence of lesions in control women 25%). To be certain that our control size was adequate to detect a significant difference, we recruited 100 asymptomatic premenopausal women.
To assess differences in prevalence rates of pathologic conditions between groups, we used the χ2 test with Yates' correction for continuity and the Fisher exact test (when n was less than 5). All other differences were analyzed using the Student t test. P < .05 was considered statistically significant.
Premenopausal women with abnormal uterine bleeding had a higher prevalence of endometrial polyps than did asymptomatic women the same age (Table 1). Uterine polyps are frequently asymptomatic and 10% of controls had them. (An example of one of these polyps is shown in Figure 1). Polyps in women with abnormal uterine bleeding tended to be larger than those in asymptomatic women (mean diameter 13.9 versus 8.5 mm, P = .064), although that difference failed to reach statistical significance. Polyps were associated significantly with uterine myomas, and 36% of women with myomas had polyps, compared with a 13% prevalence of polyps among women with no myomas (P ≤ .001). When women with myomas were excluded, polyps were still more common among women with abnormal uterine bleeding than among controls. Polyps also were more common in relatively older premenopausal women in case patients and controls. Only 3% of women under age 35 had polyps, compared with 23% of women 35 and older (P < .02).
Intramural and intracavitary myomas were significantly more prevalent in women with abnormal bleeding compared with asymptomatic women (Table 1). Among women with myomas, there was no significant difference in the mean number of myomas (2.0 versus 1.5) or mean diameter of myomas (21.5 versus 20.8 mm) between symptomatic and asymptomatic women, respectively.
We reviewed the charts of 48 women who had surgery after sonohysterography to determine its accuracy. In 35 women, intracavitary lesions were detected by sonohysterography and in 33 of these women (94.3%) the presence of intracavitary lesions was confirmed. Only two (5.7%) had sonohysterographically detected lesions without evidence of intracavitary lesions at surgery. On review, in both cases what had been called an intracavitary lesion was actually a blood clot. Both women had been bleeding heavily at the time of sonohysterography. Thirteen women had normal sonohysterographic findings and subsequently underwent surgery for bleeding abnormalities. In 12 (92.3%), no identifiable intracavitary lesions were noted at surgery or during pathologic evaluation. One woman (7.7%) had a small polyp in her hysterectomy specimen that was not seen during sonohysterography 1 month before surgery. Compared with surgical and pathologic findings as the criterion standard, sonohysterography had a sensitivity of 97%, a specificity of 86%, a positive predictive value of 94%, and a negative predictive value of 92%.
The ability of sonohysterography to differentiate intracavitary myomas from polyps was evaluated in the 35 women who had surgery after intracavitary lesions were found by sonohysterography. Polyps were considered present in 25 cases, and this presence was confirmed in 18 (72%). Two blood clots and five small myomas were called polyps mistakenly. The presence of intracavitary myomas detected by sonohysterography was confirmed 100% of the time (ten of ten cases).
We considered that confounding factors could explain the differences we found in the prevalence of uterine lesions between case patients and controls. Risk factors for myomas include older age before menopause and black race.15 There was no difference in mean age between symptomatic women and asymptomatic women (41.1 versus 39.5 years, respectively; P = .83). There was also no difference in race (more than 90% of women in both groups were white). In the control group, only 3% of women had ever been treated or evaluated for abnormal bleeding at a prior clinic visit. They had been treated only with oral contraceptives (OCs) and currently were not having bleeding problems. Only one of the three women was still taking OCs. The symptomatic women were a highly selected population, 81% of which had been evaluated for abnormal bleeding. Half of the symptomatic women had had endometrial biopsies or dilation and curettage. There was a significant difference in the percentage of women prescribed hormonal medications between symptomatic and asymptomatic women (54 versus 21%, respectively, P < .001). Asymptomatic women were using hormonal medications for contraception only, whereas most of the symptomatic women were prescribed hormonal medications to control abnormal bleeding.
Sonohysterography is effective for diagnosing intracavitary uterine lesions. Using pathologic findings as the standard, we found sonohysterography to have very good sensitivity, specificity, positive predictive value, and negative predictive value for evaluating premenopausal women with abnormal uterine bleeding. Our findings are similar to previous sonohysterographic findings in premenopausal women. Gaucherand et al9 compared sonohysteroscopy with hysterosalpingography and compared results with pathologic findings at hysteroscopy. They found sensitivity and specificity for sonohysterography to be greater than 90% and found this technique to be superior to hysterosalpingogram for detecting intrauterine filling defects. Saidi et al14 compared diagnoses by sonohysterography with findings during office hysteroscopy and found higher sensitivity and specificity for sonohysterography. Widrich et al11 compared sonohysterography and office hysteroscopy in premenopausal women and found nearly identical sensitivity and specificity, but sonohysterography was less painful for the women.
To avoid false-positive findings, we only did sonohysterography when control women were not menstruating. Some investigators1,7 restricted use of sonohysterography to the early proliferative phase, when the endometrium should be thinnest. Others5 reported that sonohysterographic findings were independent of cyclic endometrial changes. When sonohysterography is done in the secretory phase of the cycle, the endometrium is thickened and one must carefully differentiate focal regions of endometrial thickening from polyps. We found it best to view the area in question in several planes. Polyps will be surrounded by saline in multiple planes, whereas areas of endometrial thickening will not. Among women who subsequently had surgery, we noted no false-positive results for sonohysterography, except in women studied during heavy menses.
We found a very high prevalence (54%) of intracavitary lesions with sonohysterography in women with abnormal uterine bleeding. In studies of sonohysterography in selected premenopausal women,5,7–9,11,13,14 similarly high prevalences of polyps and myomas have been reported in women with abnormal uterine bleeding. Women referred for sonohysterography in our study and previous studies were selected women who often had bleeding refractory to initial management. Half of our subjects had had previous office endometrial biopsies or dilation and curettage to rule out endometrial hyperplasia and cancer. Intracavitary myomas and polyps frequently are missed in nondirected biopsies, and women with myomas are known to respond poorly to medical management for abnormal bleeding, so we might have selected women with a high prevalence of those conditions. It is likely that the prevalence of uterine lesions would be significantly less in an unselected population with abnormal uterine bleeding. Goldstein et al16 evaluated transvaginal ultrasonography and sonohysterography in initial management of perimenopausal uterine bleeding and found prevalences of polyps and submucous myomas in their unselected population of 431 women to be 13 and 5%, respectively.
A unique aspect of our study was the asymptomatic controls. We found that 10% of asymptomatic premenopausal women age 30 and older had polyps detected by sonohysterography, a higher prevalence than that reported in a study of asymptomatic women who underwent hysteroscopy. Cooper et al17 found a 1% prevalence of polyps in women between ages 19 and 44 who presented for hysteroscopic sterilization. Differences in subject age and sensitivity of techniques might explain differences in the prevalence of asymptomatic polyps found in that study and ours. Our study confirmed that polyps frequently can be asymptomatic,17,18 which is important because transvaginal ultrasonography and sonohysterography are being used increasingly for evaluating abnormal uterine bleeding, recurrent pregnancy loss, and infertility and for evaluating normal women presenting for annual examinations. Our study raises questions about the significance of polyps in women with infertility, because the prevalence of polyps found by sonohysterography does not appear to be different in infertile women4 compared with fertile controls. That raises questions about whether to remove polyps in infertile women. The answer likely will come only with a randomized clinical trial. One concern of women after detection of polyps is whether the polyps could be associated with malignancy. In premenopausal women, endometrial cancer appears rarely (less than 1%) to be associated with polyps, although in postmenopausal women the association is more likely (9.6%).19
We found that polyps were more prevalent in women age 35 and older and were associated strongly with myomas, confirming results of a previous study of hysterectomy specimens, performed at Johns Hopkins.19 Determining whether polyps cause abnormal bleeding becomes difficult because of the frequent association of polyps and fibroids. After excluding women with myomas, we found a higher prevalence of polyps among women with abnormal bleeding than among controls. Polyps in bleeding women tended to be larger than those in asymptomatic women, which suggests that the lesions are causative.
Intramural myomas and submucous myomas were significantly more prevalent in women with abnormal bleeding than in asymptomatic women of the same age. Myomas also can be asymptomatic, as was the case in 13% of our controls. Why some women with myomas have bleeding while others do not is unclear. There was no difference in numbers or size of myomas between bleeding women and controls.
1. Deichert U, Schleif R, van de Sandt M, Juhnke I. Transvaginal hysterosalpingo-contrast-sonography (Hy-Co-Sy) compared with conventional tubal diagnostics. Hum Reprod 1989;4:418–24.
2. Mitri FF, Andronikou AD, Perpinyal S, Hofmeyr GJ, Sonnendecker EW. A clinical comparison of sonographic hydrotubation and hysterosalpingography. Br J Obstet Gynaecol 1991;98:1031–6.
3. Bonilla-Musoles F, Simón C, Serra V, Sampaio M, Pellicer A. An assessment of hysterosalpingosonography (HSSG) as a diagnostic tool for uterine cavity defects and tubal patency. J Clin Ultrasound 1992;20:175–81.
4. Syrop CH, Sahakian V. Transvaginal sonographic detection of endometrial polyps with fluid contrast augmentation. Obstet Gynecol 1992;79:1041–3.
5. Parsons AK, Lense JJ. Sonohysterography for endometrial abnormalities: Preliminary results. J Clin Ultrasound 1993;21:87–95.
6. Balen FG, Allen CM, Siddle NC, Lees WR. Ultrasound contrast hysterosalpingography—Evaluation as an outpatient procedure. Br J Radiol 1993;66:592–9.
7. Goldstein SR. Use of ultrasonohysterography for triage of perimenopausal patients with unexplained uterine bleeding. Am J Obstet Gynecol 1994;170:565–70.
8. Di Naro E, Bratta FG, Romano F, Caradonna F, Loizzi P. The diagnosis of benign uterine pathology using transvaginal endohysterosonography. Clin Exp Obstet Gynecol 1996;23:103–7.
9. Gaucherand P, Piacenza JM, Salle B, Rudigoz RC. Sonohysterography of the uterine cavity: Preliminary investigations. J Clin Ultrasound 1995;23:339–48.
10. Wolman I, Jaffa AJ, Hartoov J, Bar-Am A, David MP. Sensitivity and specificity of sonohysterography for the evaluation of the uterine cavity in perimenopausal patients. J Ultrasound Med 1996;15:285–8.
11. Widrich T, Bradley LD, Mitchinson AR, Collins RL. Comparison of saline infusion sonography with office hysteroscopy for the evaluation of the endometrium. Am J Obstet Gynecol 1996;174:1327–34.
12. Lev-Toaff AS, Toaff ME, Liu J-B, Merton DA, Goldberg BB. Value of sonohysterography in the diagnosis and management of abnormal uterine bleeding. Radiology 1996;201:179–84.
13. Bronz L, Suter T, Rusca T. The value of transvaginal sonography with and without saline instillation in the diagnosis of uterine pathology in pre- and postmenopausal women with abnormal bleeding or suspect sonographic findings. Ultrasound Obstet Gynecol 1997;9:53–8.
14. Saidi MH, Sadler RK, Theis VD, Akright BD, Farhart SA, Villanueva GR. Comparison of sonography, sonohysterography and hysteroscopy for evaluation of abnormal uterine bleeding. J Ultrasound Med 1997;16:587–91.
15. Marshall LM, Spiegelman D, Barbieri LR, Goldman MB, Manson JE, Colditz GA, et al. Variation in the incidence of uterine leiomyoma among premenopausal women by age and race. Obstet Gynecol 1997;90:967–73.
16. Goldstein SR, Zeltser I, Horan CK, Snyder JR, Schwartz LB. Ultrasonography-based triage for perimenopausal patients with abnormal uterine bleeding. Am J Obstet Gynecol 1997;177:102–8.
17. Cooper JM, Houck RM, Rigberg HS. The incidence of intrauterine abnormalities found at hysteroscopy in patients undergoing elective hysteroscopic sterilization. J Reprod Med 1983;28:659–61.
18. Scott RB. The elusive endometrial polyp. Obstet Gynecol 1953;1:212–8.
19. Peterson WF, Novak ER. Endometrial polyps. Obstet Gynecol 1956;8:40–9.