REPRODUCTIVE SURGERY Comparison of transvaginal sonography, saline infusion sonography, and ofce hysteroscopy in reproductive-aged women with or without

abnormal uterine bleeding

Sefa Kelekci, M.D., Erdal Kaya, M.D., Murat Alan, M.D., Yasemin Alan, M.D., Umit Bilge, M.D., and Leyla Mollamahmutoglu, M.D.
Zekai Tahir Burak Womens Health Education and Research Hospital, Ankara, Turkey

Objective: To compare the diagnostic accuracy and acceptability of transvaginal sonography (TVS), saline infusion sonography (SIS), and ofce hysteroscopy (OHS) for detecting intracavitary abnormalities in women with or without abnormal uterine bleeding (AUB). Design: Prospective double-blind study. Setting: Zekai Tahir Burak Womens Health Education and Research Hospital, Gynecology Clinic (Ankara, Turkey). Patient(s): A total of 26 women with AUB and 24 women without AUB were enrolled in this study. Intervention(s): Transvaginal sonography, SIS, and OHS were performed on women scheduled for hysterectomy. Main Outcome Measure(s): Sensitivity, specicity, and positive and negative predictive values of TVS, SIS, and OHS to detect intracavitary abnormalities (with histopathologic ndings used as the gold standard), duration of procedure, and pain scores. Result(s): The sensitivity and specicity of TVS, SIS, and OHS in detecting intracavitary abnormalities were 56.3% and 72%, 81.3% and 100%, and 87.5% and 100%, respectively. The prevalence of endometrial polyps was not different in women with and without AUB. Saline infusion sonography was less painful than OHS (pain scores of 4.3 and 7.2, respectively). Conclusion(s): The diagnostic accuracy of SIS was equal to that of OHS in diagnosing intracavitary abnormalities. Moreover, SIS was less painful than OHS for patients. (Fertil Steril 2005;84:682 6. 2005 by American Society for Reproductive Medicine.) Key Words: Endometrial polyp, ofce hysteroscopy, saline infusion sonography, submucous myoma, transvaginal sonography

Abnormal uterine bleeding (AUB) is one of the most common gynecological complaints. Up to 33% of women referred to gynecological outpatient clinics have AUB, and this proportion rises to 69% in a perimenopausal or postmenopausal group (1). Local causes include broids, endometrial polyps, cervical polyps, endometrial hyperplasia, and endometrial carcinoma. In premenopausal women with complaints of AUB, the prevalence of benign intracavitary abnormalities, such as submucous myomas and endometrial polyps, is approximately 35% (2). The prevalence of endometrial polyps and submucous myomas in women without AUB is not known completely.
Received July 31, 2004; revised and accepted March 3, 2005. Presented at the Fifth World Congress on Controversies in Obstetrics, Gynecology & Infertility, Las Vegas, Nevada, June 3 6, 2004. Reprint requests: Sefa Kelekci, M.D., Sakarya Mahallesi, Bassehir Sokak 16/11, Cebeci, Ankara 06340, Turkey (FAX: 90-312-319-18-49; E-mail: sefakelekci@ttnet.net.tr).

Transvaginal sonography (TVS) is increasingly being used as a rst line of investigation of patients with abnormal bleeding (3). The uterus and ovaries can be visualized clearly, and their pathologic lesions can be identied. However, reports on the diagnostic accuracy of TVS are conicting (3, 4). Saline infusion sonography (SIS) is a relatively new diagnostic technique, in which the uterine cavity is distended, thereby enabling the visualization of the endometrial surface (4). In addition, SIS seems to be less painful for patients than is hysteroscopy, and it is performed more quickly (5). Ofce hysteroscopy (OHS) has the advantage of directly visualizing the uterine cavity and endometrium, but it cannot comment on myometrial pathology. The choice of diagnostic procedure seems to be determined largely by clinicians preference. However, acceptability of the procedure by subjects is very important, and
0015-0282/05/$30.00 doi:10.1016/j.fertnstert.2005.03.036

682

Fertility and Sterility Vol. 84, No. 3, September 2005 Copyright 2005 American Society for Reproductive Medicine, Published by Elsevier Inc.

little comparative research dealing with patients choice is available. The purpose of this study was to compare the diagnostic accuracy, pain scores, and acceptability of TVS, SIS, and OHS for detecting intracavitary abnormalities in women with or without AUB scheduled for hysterectomy, according to the nal diagnosis given at hysterectomy. MATERIALS AND METHODS This prospective double blinded study was conducted in our gynecology clinic between May 2003 and April 2004. The institutional review board approved the study. All candidates were patients scheduled for hysterectomy as nal diagnosis. The study and procedures were explained to the patients before entering the study, and written consent was obtained from women willing to participate in the study. For inclusion in the study, the following criteria were required: normal cervical pathology, no pregnancy, age 35 years, and being scheduled for hysterectomy for variable indications. Those with a previous history of cervical surgery, previous difculties with hysteroscopy, endometrial biopsy within the previous year, and with postmenopausal status were excluded from the study. No patient had received hormonal therapy 1 month before surgery. A total of 50 patients met the study criteria. Twenty-six women with abnormal vaginal bleeding and 24 women without abnormal vaginal bleeding, willing to participate, were enrolled in the study. The examinations were performed in early proliferation phase of the menstrual cycle, if possible. The assessment of the uterine cavity consisted of four steps: [1] transvaginal sonography, [2] saline infusion sonography, [3] ofce hysteroscopy, and [4] hysterectomy because of AUB and benign pathologies. Both TVS and SIS were performed with a 5.0-MHz vaginal probe (GE Logic 200, Milwaukee, WI). For TVS, the endometrium and cavity were considered normal if sonograms showed a hyperechoic line in the middle of the uterus with a homogenous endometrial lining and distinct margins to the myometrium. All other ndings, such as deformations of the endometrial lining, absence of central hyperechoic line, and the appearance of any structure with or without well-dened margins or variable echogenicity, were considered abnormal. An endometrial polyp was dened as a smooth-margined, echogenic mass of variable size and shape with a fairly homogenous texture; it emerged from the endometrium and did not disrupt the myometrial endometrial interface. A submucosal myoma was dened as a solid, round structure of mixed echogenicity emanating from the myometrium, disrupting the inner circular muscle layer, and protruding into the uterine cavity (4). The measurements of the endometrium were performed at the thickest part from cornu to
Fertility and Sterility

cornu in the longitudinal plane and included both endometrial layers. The presence of the intracavitary abnormalities myomapolyp was recorded. To perform SIS, a sterile 8-F Foley catheter (length, 30 cm; diameter, 2.7 mm; Foleycath, Asia Pacic, Malaysia) was introduced through the cervical orice until it reached the fundus. The speculum was withdrawn, and the ultrasound probe was reintroduced into the vaginal canal. A 50-mL syringe containing sterile normal saline was then attached to the catheter. Saline instillation and distention of the uterine cavity with the saline were sonographically observed. Generally, approximately 20 mL of saline was used. The measurements of the endometrium were performed at the thickest part from cornu to cornu in the longitudinal plane in the single endometrial layer. The uterine cavity contours were inspected for irregularities, and suspicious intracavitary lesions were recorded. The same criteria for abnormal ndings were used for SIS as for TVS. Ofce hysteroscopy was performed with a rigid microhysteroscope with a 3.5-mm diagnostic sheath (Karl Storz, Tuttlingen, Germany) in an outpatient setting without anesthesia. We used sterile saline solution as the distention medium. A maximum intrauterine pressure of 100 mm Hg was allowed. This procedure was performed immediately after TVS and SIS. The cavity was rst evaluated visually, with both the tubal ostia being noted and the endometrial appearances documented. Visual analogue scores for pain were recorded separately after the three diagnostic procedures. Subjects rated their overall pain by means of the numerical rating scale from 0 (none) to 10 (most) (6, 7). The nal diagnosis was based on the combined visual and histopathological examination of the specimens. Sonographic, hysteroscopic, and histologic examinations were completed, and the ndings were recorded by different investigators without the knowledge of each others ndings. Sensitivity, specicity, and positive and negative predictive values (PPV, NPV) were calculated from 2 2 tables for each method, with histopathological ndings used as the gold standard. All data were entered into SPSS 11.5 for Windows (SPSS, Chicago, IL), and P .05 was considered signicant. We compared diagnostic accuracy, pain score, and the duration of procedure among the three diagnostic approaches. In subgroups of patients with AUB and without AUB, we compared demographic variables and prevalence of intracavitary abnormalities. RESULTS A total of 50 patients were enrolled in the study. In 9 patients (18%), we were unable to perform either procedure or both procedures (SIS and OHS). In 5 of the women with AUB (3 in whom we were unable to introduce the hysteroscope from the cervical canal without cervical dilatation and 2 in whom we were unable to have a satisfactory view because of
683

TABLE 1
Comparison of results from TVS, SIS, and OHS with nal diagnosis obtained from hysterectomy specimens. Sensitivity TVS SIS OHS 56.3 (0.410.719) 81.3 (0.690.93) 87.5 (0.770.97) Specicity 72.0 (0.580.85) 100 100 PPV 56.3 (0.410.719) 100 100 NPV 72.0 (0.580.85) 88.9 (0.790.97) 92.6 (0.841.0) Accuracy 65.8 (0.510.80) 92.5 (0.841.0) 95.0 (0.881.0)

Note: Data are expressed as % (95% CI).

Kelekci. Diagnosis of intracavitary lesions. Fertil Steril 2005.

bleeding) and 4 women without AUB (in all cases we were unable to introduce the hysteroscope from the cervical canal without cervical dilatation), we could not perform ofce hysteroscopy. Neither procedure could be performed in 2 of 9 patients because of cervical stenosis. These women were not included in the statistical analyses. Indications for hysterectomy (n 41) were menorrhagia (n 9), metrorrhagia (n 12), descensus uteri (n 2), descensus uteri with cystorectocele (n 7), descensus uteri with enterocele (n 2), chronic pelvic pain (n 2), and endometrioma (n 3). Four subjects had benign adnexal pathologies: 1 with serous cyst and 3 with recurrent dermoid cyst. Demographic characteristics of women with or without AUB were not signicantly different. The mean ( SD) age, parity, and body mass index for the women with AUB (n 21) were 48.5 4.1 years, 2, and 26.8 5.36 kg/m2, respectively; for the patients without AUB (n 20), these were 46.4 5.4 years, 2, and 26.6 4.36 kg/m2, respectively. Histologic examination of the hysterectomy specimens revealed normal endometrial histology (endometrial proliferation, secretion, atrophia) in 17 patients, submucous myomas in 3 patients, intracavitary polyps in 10 patients, en-

dometrial hyperplasia in 3 patients, endocervical polyp in 2 patients, intramural myoma in 2 patients, synechiae in 3 patients, and endometrial cancer in one patient. Table 1 shows the results of TVS, SIS, and OHS compared with the histologic diagnosis obtained after hysterectomy. Transvaginal sonography demonstrated sensitivity in diagnosing intracavitary abnormalities of 56.3% for a specicity of 72%. The likelihood ratio of a positive test result was 3 (95% condence interval [CI] 1.8 4.4), and the likelihood ratio of a negative test result was 0.93 (95% CI 0.90 0.96). In detecting endometrial polyps, sensitivity, specicity, and PPV and NPV were 100%, 96.8%, and 50% and 76.9%, respectively. In detecting endometrial hyperplasia (cut-off, 8 mm), sensitivity, specicity, and PPV and NPV were 100%, 78.9%, and 27.3% and 100%, respectively. Saline infusion sonography demonstrated sensitivity in diagnosing intracavitary abnormalities of 81.3% for a specicity of 100%. In detecting endometrial polyps, sensitivity, specicity, and PPV and NPV were 70%, 100%, and 100% and 90.9%, respectively. In detecting submucous myoma, sensitivity, specicity, and PPV and NPV were 100%, 100%, and 100% and 100%, respectively.

TABLE 2
Final diagnoses of patients in both groups. AUB present (n 21) Normal cavity Submucous myomaa Endometrial polyp Endometrial hyperplasiaa Endometrial synechiae Atrophia Endocervical polyp Intramural myoma
Note: Data are expressed as n (%). a 2 , P .125.
Kelekci. Diagnosis of intracavitary lesions. Fertil Steril 2005.

AUB absent (n 20) 10 (50) 0 (0) 5 (25) 0 (0) 2 (10) 1 (5) 1 (5) 1 (5)

Total (n 41) 17 (41.4) 3 (7.3) 10 (24.4) 3 (7.3) 3 (7.3) 1 (2.4) 2 (4.8) 2 (4.8)

7 (33.3) 3 (14.3) 5 (23.8) 3 (14.3) 1 (4.7) 0 (0) 1 (4.7) 1 (4.7)

684

Kelekci et al.

Diagnosis of intracavitary lesions

Vol. 84, No. 3, September 2005

TABLE 3
Comparison of pain scores and operation time between TVS, SIS, and OHS. TVS Pain score Operation time (min) 1.1 (02) 5.1 (39) SIS 4.3 (110) 6.0 (320) OHS 7.2 (110) 6.0 (320) P value .042a .702a

Note: Data are expressed as median (range) for pain score and mean (range) for operation time. a Wilcoxon signed rank test (SIS vs. OHS).
Kelekci. Diagnosis of intracavitary lesions. Fertil Steril 2005.

Ofce hysteroscopy revealed a sensitivity of 87.5% and a specicity of 100%. For detecting the endometrial polyps, it has a sensitivity of 80%, a specicity of 80%, PPV of 100%, and NPV of 93.9%. Ofce hysteroscopy demonstrated sensitivity in diagnosing submucous myomas of 100% for a specicity of 100%. In detecting endometrial polyps, the diagnostic accuracies of TVS, SIS, and OHS were 76.5%, 92.5%, and 95%, respectively. All three diagnostic approaches have a similar diagnostic accuracy (100%) in detecting submucous myomas. The distributions of all pathologic conditions in women with or without AUB are summarized in Table 2. Among women with AUB, the endometrial cavity was normal in seven patients (33.3%). There was no difference between the two groups in terms of endometrial polyp prevalence. Submucous myoma and endometrial hyperplasia were more common in women with AUB (both P .125). Endometrial synechiae was diagnosed in one woman with AUB and two women without AUB. Neither TVS nor SIS could diagnose these cases. The pain score for SIS was 4.3 and for OHS was 7.2. According to a Wilcoxon rank test, there was a signicant difference in pain scores between the two techniques (P .042). The examination took a mean of 5.1 minutes, 6.0 minutes, and 6.0 minutes in TVS, SIS, and OHS, respectively (Table 3). DISCUSSION In this study, we compared the diagnostic accuracy of TVS, SIS, and OHS for the detection of intracavitary abnormalities. The diagnostic accuracies of SIS and OHS in diagnosing intracavitary abnormalities were higher than that of TVS. The sensitivity of TVS in diagnosing intracavitary abnormalities by direct observation was 56% for a specicity of 72% in our study. Some studies reported a similar diagnostic accuracy (8, 9). Emanuel et al. (10) reported a higher sensitivity (96%) for a specicity of 89%; however, this unblinded study used hysteroscopy as the reference test and included premenopausal as well as postmenopausal patients. The latter might be an explanation for the better perforFertility and Sterility

mance. Another explanation is that small intracavitary abnormalities that were detected in the present study were missed by Emanuel et al. (10) at hysteroscopy because we used hysterectomy specimens as the reference. Where TVS is used to triage premenopausal women with AUB, its clinical value would be the prevention of a diagnostic hysteroscopy. The high sensitivity that is required for such a purpose could not be observed in our study. It might have been better to comment on whether there were differences in sensitivity and specicity of the vaginal probe ultrasound performed in early proliferative, mid-cycle, and late secretory endometrium. It is not easy to comment on this, not only because we had to perform SIS and OHS sequentially in the same cycle but also because we had difculty separating the cycle to different phases, particularly in patients with AUB. Moreover, the use of threedimensional ultrasound might have changed the results, but it was not available in our institution. In our study, the endometrial thickness measurement alone indicated poor diagnostic results. Using a similar cutoff level of endometrial thickness, Bronz et al. (11) reported a sensitivity of 92% for a specicity of 22%. A cut-off level of 5 mm (double layer) would increase sensitivity to 86%, but specicity would decrease to 49% in our study. This level is not useful for an ideal screening test. Saline infusion sonography demonstrated a sensitivity of 81% and a specicity of 100%. Although the number involved is small, our data correspond to those of the literature, in which the sensitivity of SIS for diagnosing intracavitary abnormalities varies from 88% to 100% for a specicity of 76% to 96% (8, 11, 12). Saline infusion sonography was very effective for evaluating uterine myomas. It equalled hysteroscopy in detecting submucous myomas and in addition was able to evaluate the uterine wall, determine the intramural component of submucous myomas, and locate other intramural myomas. Important information is also obtained by adnexal evaluation. Hysteroscopy allows direct visualization of the uterine cavity. To be expected, endometrial synechiae was diagnosed by OHS with 100% sensitivity and 100% specicity, whereas it has similar diagnostic results to SIS for
685

detecting endometrial polyps and submucous myoma. The main result of our study is that OHS yielded no additional information in women with AUB compared with that obtained by SIS. This nding was comparable to those from other studies (7, 13, 14). Saline infusion sonography was less painful for patients than OHS in our study. Rogerson et al. (15) reported a mean pain score of 3.6 for hysteroscopy and a lower score of 2.5 for SIS. This nding is comparable to that of Widrich et al. (5). Because SIS seems to be less painful for patients than hysteroscopy and is performed quickly, SIS might replace OHS for many patients. A systematic review and meta-analysis (16) provided the strongest evidence to date that saline hysterosonography is both feasible and accurate. Therefore, SIS might be useful for gynecologists who do not perform diagnostic hysteroscopy because SIS can be performed in an outpatient setting. However, hysteroscopy was superior to SIS because it enables simultaneous diagnosis and treatment of focal endometrial pathologies, such as broids and polyps. The most surprising nding in our study was the prevalence of endometrial polyps in women with or without AUB. The prevalence was similar in the two subsets of patients. To the best of our knowledge, this is the rst report evaluating the endometrial cavity in women without AUB. Endometrial polyps might not be symptomatic in all women, though our sample volume was small, and this group of women might not reect the whole population. Submucous myoma and endometrial hyperplasia were more common in women with AUB. The prevalence of submucous myomas has been estimated as 5%18% of all myomas (17, 18). In our study, submucous myomas were seen in 7.3% of all patients and in 14.3% of women with AUB. The latter prevalence of submucous myomas could be higher because all the patients had menorrhagia and metrorrhagia. Finally, SIS has a diagnostic performance similar to that of OHS in the evaluation of women aged 35 years with or without AUB. The limiting factor of TVS is incomplete visualization of the uterine cavity. Because SIS seems to be less painful for patients than hysteroscopy and is performed quickly, it might be a useful and noninvasive tool in the evaluation of the intracavitary lesions of the uterus.
Acknowledgement: The authors thank Miss Begum Kaya for the nal linguistic revision.

REFERENCES
1. Mencaglia L, Perino A, Hamou J. Hysteroscopy in perimenopausal and postmenopausal women with abnormal uterine bleeding. J Reprod Med 1987;32:577 82. 2. Emanuel MH, Verdel MJC, Stas H, Wamsteker K, Lammes FB. An audit of true prevalence of intrauterine pathology: the hysteroscopical ndings controlled for patient-selection in 1202 patients with abnormal uterine bleeding. Gynecol Endosc 1995;4:237 41. 3. Dijkhuizen FPHLJ, Brlman HAM, Potters AE, Bongers MY, Heintz APM. The accuracy of transvaginal ultrasonography in the diagnosis of endometrial abnormalities. Obstet Gynecol 1996;87:3459. 4. Parsons AK, Lense JJ. Sonohysterography for endometrial abnormalities: preliminary results. J Clin Ultrasound 1993;21:8795. 5. Widrich T, Linda D, Bradley LD, Mitchinson AR, Collins RL. Comparison of saline infusion sonography with ofce hysteroscopy for the evaluation of the endometrium. Am J Obstet Gynecol 1996;174:132734. 6. Yeager KA, Miaskowski C, Dibble S, Wallhagan M. Differences in pain knowledge in cancer patients with and without pain. Cancer Pract 1997;5:39 45. 7. Salomaki TE, Kokki H, Turunen M, Havukainen U, Nuutinen LS. Introducing epidural fentanyl for on-ward pain relief after major surgery. Acta Anaesthesiol Scand 1996;40:704 9. 8. Djkhuizen FPHLJ, De Vries LD, Mol BWJ, Brlmann HAM, Peters HM, Moret E, et al. Comparison of transvaginal ultrasonography and saline infusion sonography for the detection of intracavitary abnormalities in premenopausal women. Ultrasound Obstet Gynecol 2000;15: 372 6. 9. Williams CD, Marshburn PB. A prospective study of transvaginal hydrosonography in the evaluation of abnormal uterine bleeding. Am Obstet Gynecol 1998;179:292 8. 10. Emanuel MH, Verdel MJ, Wamsteker K, Lammens FB. A prospective comparison of transvaginal ultrasonography and diagnostic hysteroscopy in the evaluation of patients with abnormal uterine bleeding: clinical implications. Am J Obstet Gynecol 1995;172:54752. 11. Bronz L, Suter T, Rusca T. The value of transvaginal sonography with and without saline installation in the diagnosis of uterine pathology in pre- and postmenopausal women with abnormal bleeding or suspect sonographic ndings. Ultrasound Obstet Gynecol 1997;9:53 8. 12. Goldstein SR, Zeltser I, Horan CK, Synder JR, Schwartz LB. Ultrasonography-based triage for perimenopausal patients with abnormal uterine bleeding. Am J Obstet Gynecol 1997;177:102 8. 13. Krampl E, Bourne T, Solbakken HH, Istre O. Transvaginal ultrasonography, sonohysterography and operative hysteroscopy for the evaluation of abnormal uterine bleeding. Acta Obstet Gynecol Scand 2001; 80:616 22. 14. Dueholm M, Lundorf E, Hansen ES, Ledertough S, Olesen F. Evaluation of the uterine cavity with magnetic resonance imaging, transvaginal sonography, hysterosonographic examination and diagnostic hysteroscopy. Fertil Steril 2001;76:350 7. 15. Rogerson L, Bates J, Weston M, Duffy S. A comparison of outpatient hysteroscopy with saline infusion hysterosonography. BJOG 2002;109: 800 4. 16. de Kroon CD, de Bock GH, Dieben SWM, Jansen FWJ. Saline contrast hysterosonography in abnormal uterine bleeding: a systematic review and meta-analysis. BJOG 2003;110:938 47. 17. Baggish MS. Mesenchymal tumors of the uterus. Clin Obstet Gynecol 1974;15:51 64. 18. Fedele L, Bianchi S, Dorta M, Brioschi D, Zanotti F, Vercellini P. Transvaginal ultrasonography versus hysteroscopy in the diagnosis of the uterine submucous myomas. Obstet Gynecol 1991;77:7459.

686

Kelekci et al.

Diagnosis of intracavitary lesions

Vol. 84, No. 3, September 2005