Ultrasound and intrauterine adhesions: a novel structured approach to diagnosis and management.

Abstract

In anatomical terms, the uterus is often perceived as a simple organ that is easily accessible on clinical examination and can be evaluated in great detail using various diagnostic modalities. It therefore comes as a surprise that there are still many uncertainties regarding the effects of congenital and acquired uterine abnormalities on women's reproductive health. Intrauterine adhesions (IUAs), which involve both the myometrial and the endometrial layers, are relatively rare in comparison with other abnormalities such as uterine fibroids and adenomyosis, and our understanding of their causes and clinical significance is particularly sparse. IUAs were first described by Heinrich Fritsch in the late 19th century1, but it was Joseph Asherman who, in 1948, further defined this eponymous condition as a traumatic amenorrhea caused by obstruction of the internal cervical os following complicated labor or abortion requiring intrauterine instrumentation2. The terms ‘Asherman syndrome’ and ‘intrauterine adhesions’ are often used interchangeably, although the original definition of the syndrome includes amenorrhea, with or without cyclical abdominal pain. Over the past 70 years there have been a number of studies focusing on IUAs, but there remain many unanswered questions regarding their etiology, prevalence and effect on reproductive outcomes3. Historically, hysterosalpingography (HSG) has been the main tool for the diagnosis of IUAs. With the development and widespread use of endoscopic techniques, hysteroscopy is now considered the ‘gold standard’ for diagnosis of this condition and also enables us to treat it4, 5. High-resolution ultrasound has, in recent years, largely replaced diagnostic hysteroscopy as the primary tool for assessment of the uterine cavity. Although ultrasound is one of the most commonly used diagnostic tests in gynecology, there is a paucity of data regarding its value for the detection and management of IUAs. The recent advent of three-dimensional (3D) ultrasound has further enhanced the role of ultrasound in the diagnosis of uterine anomalies, as it enables imaging of the uterus in planes that cannot be obtained on standard two-dimensional (2D) imaging. In this Editorial, we review current thinking on the etiology and prevalence of IUAs, and propose a novel structured ultrasound-based approach to the diagnosis and management of this complex condition. The prevalence of IUAs is difficult to determine with certainty, but it is thought to vary between 0.3% and 21.5%, with the highest rates being reported in Israel, Greece and South America6. These geographical variations could be explained by the differences in the number of women receiving surgical treatment for miscarriage or surgical termination of pregnancy, prevalence of puerperal infection/genital tuberculosis and the diagnostic methods used to detect IUAs7. A recent meta-analysis found that the prevalence of IUAs following a miscarriage varied between 3% and 38%, with a pooled prevalence of 19%8. IUAs were significantly more frequent in women with a history of multiple miscarriages. There were no cases of IUAs in women who underwent conservative management of miscarriage, and the authors hypothesized that repeated uterine curettage is the most important risk factor for formation of IUAs. One of the main limitations in determining the prevalence of IUAs is the reliance on invasive diagnostic methods such as HSG and hysteroscopy to detect them. These tests tend to be offered selectively to women with a history of menstrual symptoms or infertility, while asymptomatic adhesions remain undetected. In view of this, it is likely that the prevalence of IUAs is underestimated and their clinical significance overstated. The endometrium consists of a basal layer, adjacent to the myometrium, and a superficial functional layer. The functional layer responds to the cyclical changes in the secretion of ovarian sex hormones. The increased hormone production during the proliferative phase of the cycle stimulates endometrial growth, while its decrease in the late luteal phase causes endometrial shedding and menstruation. The basal layer is not shed at any time during the menstrual cycle and provides the basis for regeneration of the functional layer, which is completed without formation of adhesions. Endometrial injury that results in the loss of stroma is believed to be the key event leading to the development of IUAs. Lost stroma is replaced by fibrous tissue, which causes the opposing endometrial surfaces to bond to one another, resulting in the formation of IUAs. The distinction between the functional and basal layers of the endometrium is lost and they are both replaced by an avascular and atrophic epithelial monolayer that is unresponsive to hormonal stimulation7. Injury to the gravid uterus is the main risk factor for developing IUAs. A literature review including 1856 women with IUAs found that 67% of them had undergone uterine curettage for termination of pregnancy or for the treatment of spontaneous miscarriage, 22% of the IUAs were secondary to postpartum curettage, 2% occurred following Cesarean section and 0.6% following uterine evacuation for a hydatidiform mole6. Instrumentation of the uterus between the second and fourth week postpartum is particularly hazardous, resulting in more frequent and more severe IUAs6. It has been hypothesized that this is due to the low estrogen level in the postpartum period, which fails to stimulate endometrial growth and regeneration. The risk of developing IUAs following trauma to the non-gravid uterus is lower than that for the gravid uterus. IUAs occurred in 1.6% of women following diagnostic curettage, 1.3% after abdominal myomectomy, 0.5% after cervical biopsy or polypectomy and 0.2% after insertion of an intrauterine device6. A prospective study found IUAs in 50% of women at follow-up outpatient hysteroscopy, 3 months after an open myomectomy. The number of fibroids removed was significantly higher in women who developed adhesions, while preoperative ovarian suppression, uterine size and opening of the uterine cavity during surgery were not associated with an increased risk of IUAs9. In recent years, hysteroscopic resection has been widely used to treat submucous fibroids. Taskin et al.10 found frequencies of postsurgical IUAs of 31.3% and 45.4% after resection of single and multiple fibroids, respectively. The relationship between multiple fibroids may be an important factor, as IUAs were more common in women with apposing fibroids11. Hysteroscopic resection of the uterine septum is also a known risk factor, with the reported prevalence of postoperative IUAs varying between 6 and 88%10-12. The risk appears to be higher in the presence of a wide septum with a large surface area13. It has been postulated that rough surfaces in the uterine cavity, following division of the septum, can fuse together, resulting in IUA formation resembling a residual septum. Formation of IUAs following endometrial ablation is further proof that surgical trauma can result in their development. Adhesions can also form following non-surgical trauma to the endometrium. Mara et al.14 showed that 14% of patients had IUAs following uterine artery embolization, suggesting that compromised uterine blood supply resulting in hypoxia could also lead to endometrial injury. There is no agreement regarding the possible role of infection in the formation of IUAs. Some authors maintain that the presence of peritubal adhesions, chronic or subacute endometritis and bacterial isolation, in cases of Asherman syndrome, confirms that infection could be a predisposing factor15, but others disagree16. Infection with Mycobacterium tuberculosis, however, is a recognized cause of severe IUAs, accounting for 4% of cases6. It often results in complete obliteration of the uterine cavity and destruction of the functional endometrium. In view of this, in such cases the prognosis is poor and surgical correction is difficult. Obliteration of the uterine cavity can be focal, partial or complete. Focal adhesions causing minimal obliteration of the cavity are most likely to develop in the areas of maximal endometrial trauma. The surrounding endometrium is usually normal and responsive to the circulating ovarian sex hormones. Focal adhesions are typically asymptomatic and are diagnosed incidentally on ultrasound or on diagnostic hysteroscopy. More severe adhesions result in partial or complete obliteration of the uterine cavity. In women with partial obliteration, the location of IUAs is important, as those located close to the tubal ostia may cause infertility or ectopic pregnancy6. The unaffected endometrium may be either normal or atrophic and therefore some women may be asymptomatic while others may experience reduced menstrual loss, menstrual irregularities or infertility. In these women, surgery may result in successful division of the IUAs and complete resumption of normal endometrial function. Women with complete obliteration of the uterine cavity present typically with amenorrhea and infertility. In most cases, however, there are small pockets of preserved functional endometrium, which can cause dysmenorrhea and chronic pelvic pain. Although surgical correction is achievable, the risk of recurrence is high and the long-term prognosis is poor. Obstruction to menstrual flow can be complete or partial. Complete obstruction is usually caused by adhesions obliterating the lower part of the uterine cavity and/or cervical canal. The typical presentation is amenorrhea and cyclical lower abdominal discomfort. Hematometra is often, but not always, present. The endometrium above the level of the obstruction may appear normal or abnormally thin. The reasons for hematometra developing in some, but not all, women with IUAs are unclear, as there is usually some functional endometrium preserved. It is possible that the presence of obstruction somehow suppresses the endometrial function above it. This type of IUA may occur after dilatation and curettage or at the level of a previous Cesarean section scar17. The prognosis is usually good, with restoration of normal menstruation after treatment. Reduced menstrual flow is a common presenting symptom in cases of partial blockage of the cavity. These types of IUAs can be located anywhere within the uterine cavity, with menstrual blood accumulating above the level of the blockage. The endometrium can vary in appearance from normal to atrophic. In women with IUAs, the surrounding endometrium may appear normal or atrophic (dysfunctional). Studies using pelvic angiography have shown that endometrial and myometrial vascularity are compromised in women with IUAs18, but there have been no comparative studies of uterine circulation using Doppler ultrasound. Women with IUAs in whom endometrial function is partially preserved continue to menstruate without the formation of new adhesions. Women with atrophic (dysfunctional) endometrium typically complain of absent or very light periods. In some women, the uterine cavity contains areas of both normal functional and non-functional endometrium, known as ‘skip lesions’. In general, the presence of functional endometrium is associated with a better long-term prognosis, while for those presenting with complete endometrial dysfunction the prognosis is worse. Although menstrual disturbances are the most common clinical manifestation of IUAs, subfertility and recurrent miscarriage are also common reasons for women to seek medical help. Women with complete obstruction and/or obliteration of the uterine cavity remain amenorrheic and cannot achieve a pregnancy. Those women who have either focal or partial obliteration and/or obstruction continue to menstruate and are therefore able to conceive. Conception rates vary between 40.4% and 66.1%19, 20, with live-birth rates varying between 64% and 86%19, 21. This variation in success rates could be correlated to the severity of adhesions, those with more extensive IUAs having a lower pregnancy rate5. In women with untreated IUAs, the rate of spontaneous miscarriage is 40%6. This could be explained by compromised endometrial function associated with IUAs19. After successful treatment, miscarriage rates decrease to 11.1–20.5%, similar to that of the general population19, 21. In pregnancies with IUAs that progress into the third trimester, abnormal placentation, including placenta accreta, occurs in 13% to 20% of cases, and is thought to be the consequence of poor decidualization6, 19, 21. Preterm labor is more common in women with IUAs, even after successful adhesiolysis. This could be owing to an increased risk of antepartum hemorrhage and intrauterine growth restriction that is associated with abnormal placentation8. Women with IUAs obstructing tubal ostia are at risk of an ectopic pregnancy, as the communication between the uterine cavity and Fallopian tube is blocked6. Although rare, intramural ectopic pregnancy may be more common in cases of IUAs owing to the absence of decidua and gaps in the endometrial–myometrial junction facilitating propagation of the trophoblastic tissue deep into the myometrium22. Diagnosis of IUAs has become more common because of better awareness of the condition and improvements in diagnostic techniques. Hysteroscopy has been traditionally considered the gold standard for the diagnosis of IUAs, as it allows direct visualization of the uterine cavity4, 5. As a result, hysteroscopic findings are used as a basis for most classification systems for IUAs5, 23-27. At hysteroscopy, adhesions may appear as thin or filmy avascular strands or thick bands between the anterior and posterior walls, or adherence (agglutination) of the walls, resulting in partial or complete obliteration of the cavity. Hysteroscopy, however, is a relatively invasive test that can only be used selectively in women with a high index of clinical suspicion for IUAs. Its reproducibility has been questioned, as it relies solely on the subjective impression of the operator performing the test28. Differential diagnosis between IUAs and uterine anomalies such as unicornuate uterus may be difficult to achieve at hysteroscopy in a small number of cases in which only one side of the cavity can be seen. Hysteroscopic examination of the uterine cavity may not be possible in women with severe IUAs resulting in obliteration of the lower part of the uterine cavity because of the inability to correctly identify surgical planes. Prior to the advent of hysteroscopy, HSG was the key method for the diagnosis of IUAs. HSG reveals filling defects, and outlines of the uterine cavity may appear irregular in the presence of IUAs. Similar to hysteroscopy, HSG cannot be performed in women with complete cervical obstruction or extensive obliteration of the lower part of the uterine cavity7. In the absence of complete obstruction of the lower uterine cavity and/or cervical canal, HSG can be useful for outlining the uterine cavity, as the contrast medium can find its way around the adhesions. This can be useful in determining the extent of adhesions and their location before hysteroscopy. Using hysteroscopy as the reference standard for the diagnosis of IUAs, HSG had a sensitivity of 75–81%, specificity of 80% and positive predictive value of 50%4, 29. The high rate of false-positive diagnoses, coupled with radiation exposure and invasiveness, are the main disadvantages of HSG. Magnetic resonance imaging has been used to investigate IUAs, especially when cervical obstruction is suspected, but experience with this technique is limited7. 2D ultrasound provides clear images of the uterus and is used routinely as a first-line diagnostic tool for the assessment of uterine and endometrial morphology. It also enables accurate measurement of endometrial thickness. The integrity of the endometrial layer can be assessed including disruptions to the endometrial–myometrial junction. Adhesions on ultrasound are seen as bands of myometrial tissue traversing the endometrial cavity and adjoining the opposing uterine walls (Figure 1a). The bands may vary in thickness, but their echogenicity is usually the same as that of the adjacent myometrium30. In cases of severe IUAs, the uterine cavity may appear irregular with a loss of endometrial echo31 (Figure 1b). Partial or complete obstruction of the uterine cavity results in the accumulation of menstrual blood, which can also be seen on 2D ultrasound (Figures 1c and 2). The typical ultrasound finding is the presence of hypoechoic areas with interruptions of the endometrial layer (skip lesions representing entrapped menstrual blood or secretions from preserved functioning endometrium)7 (Figure 3). In cases with adhesions that obliterate the internal os, the whole of the uterine cavity may be distended with uterine secretions or blood, but endometrial bands are rarely seen. Thin atrophic endometrium is a common feature when significant IUAs are present and indicates its unresponsiveness to estrogens. Following successful division of adhesions, normal endometrial function tends to be restored. IUAs should not be confused with subendometrial fibrosis. Fibrosis occurs after invasive intrauterine procedures such as dilatation and curettage, and is particularly common and extensive after manual removal of the placenta31. Fibrosis appears typically on 2D ultrasound as echogenic foci in the basal endometrial layer with no loss of functional endometrium (Figure 4). Fibrosis is typically asymptomatic and there is no evidence that it has any adverse effect on women's menstrual periods and reproductive health. However, as they are both caused by surgical trauma, it is not unusual to find subendometrial fibrosis concomitant with IUAs. The reported sensitivity of 2D ultrasound in detecting IUAs varies. Fedele et al.32 reported a sensitivity of 97%, which differs greatly from the 0% reported by Soares et al.4; however, their study included only four patients. These differences suggest that the skill of the operator and quality of the ultrasound equipment could be important factors that affect the accuracy of ultrasound diagnosis. In recent years, 3D ultrasound has become the method of choice for the diagnosis and assessment of congenital uterine anomalies33. 3D reconstruction of the uterine anatomy provides panoramic views of the uterine cavity in the coronal plane, with much clearer views of the endometrial–myometrial junction than those seen on 2D ultrasound. This is particularly helpful for the detection of adenomyosis31. 3D ultrasound also facilitates mapping of uterine fibroids34. Another advantage of 3D ultrasound is the ability to measure both the endometrial surface area and volume in standardized planes, which provides the option of describing the observed uterine abnormalities in quantitative terms35. The experience of 3D ultrasound for the detection and evaluation of IUAs is limited, but some studies have suggested that it may be superior to HSG for the detection of uterine intracavitary pathology36. 2D saline or gel infusion sonohysterography (2D-SIS, 2D-GIS or 2D-SHG) can also be used to detect IUAs, with comparable sensitivity and specificity to those of HSG4, 37. Some studies have shown that 3D-SIS is more accurate than is 2D-SIS for the diagnosis of IUAs38. Combining 3D-SIS with power Doppler yields a sensitivity and specificity of 91.1% and 98.8%, respectively, for all intrauterine lesions, including adhesions39. Ludwin et al.40 demonstrated that 3D-SIS had a sensitivity of 90% and specificity of 95% for the diagnosis of IUAs following hysteroscopic metroplasty. Mild adhesions appear typically as mobile, thin echogenic bands bridging a normally distensible endometrial cavity with pockets of fluid trapped between them41 (Figure 5). As the severity of adhesions increases they appear as thick, broad-based bridging bands. In women with complete obliteration of the cavity, SIS encounters the same limitation as do hysteroscopy and HSG42. Hysteroscopic adhesiolysis is the mainstay of treatment of IUAs (Figure 6). Various surgical techniques have been described using scissors, electrosurgery or laser vaporization, however no method is favored universally. Restoration of normal menses, conception and live-birth rates have been the main primary end-points and markers of successful treatment, with many adopting a ‘second-look’ hysteroscopy to assess the success of the treatment11, 21, 43, 44. Despite the use of ancillary treatments, such as postoperative insertion of an intrauterine contraceptive device, anti-adhesion barrier agents and high-dose estrogens, recurrence of IUAs is common and represents the major limitation of the procedure. The reported recurrence rate varies from 3.1% to 23.5% for all types of IUAs5, 45, 46, with severe forms having a higher recurrence rate of 20–62.5%5, 45-47. In these cases, multiple surgical procedures are often required. It is therefore important to identify preoperatively women with severe IUAs to provide appropriate counseling and consider referral to expert hysteroscopic surgeons with experience in complex intrauterine surgery. Ultrasound guidance has been used in recent years to increase the safety and efficacy of various intrauterine procedures including surgical treatment of miscarriage, evacuation of Cesarean scar ectopic pregnancy, hysteroscopic metroplasty and resection of submucous fibroids48, 49. There are also reports of using ultrasound guidance to assist with difficult hysteroscopy due to cervical stenosis or IUAs. Kresowik et al.50 compared surgical complications from hysteroscopic adhesiolysis using laparoscopic, ultrasound or no guidance. They found that ultrasound provided a reliable, safe method of assistance, while being cost-effective50. In addition, ultrasound is the ideal method of assessing the uterine cavity postoperatively and checking for recurrent IUAs. The demonstration of a normal cavity on ultrasound could avoid the need for routine post-treatment hysteroscopy, thus reducing the overall cost of treatment and social inconvenience to women. On the other hand, residual IUAs can be described in great detail and aid in the planning of further treatment. A clinically useful classification of IUAs has to be simple, quantifiable, reproducible and understandable to women and health professionals. The classification should also provide information on the probable impact of IUAs on women's reproductive health and the probability of successful surgical treatment. Current classification systems are mainly based on hysteroscopic findings, taking into account the location, extent and appearance of adhesions. In 1978, Toaff and Ballas51 and March et al.23 independently published their descriptions of IUAs based on findings at HSG and hysteroscopy, respectively. A decade later, the American Fertility Society (AFS) devised a classification system that, for the first time, took into consideration the menstrual pattern as a prognostic marker25. The scoring system encompasses the extent of the cavity involved and the type of adhesions based on HSG and hysteroscopic findings, as well as menstrual symptoms25. Although other classification systems of IUAs have been proposed5, 24, 26, including one by the European Society of Hysteroscopy and European Society of Gynaecological Endoscopy27, the AFS system appears to be the most widely adopted. To date, there have been no comparative analyses of different classification systems taking into account surgical and reproductive outcomes. Although hysteroscopy is undoubtedly a reliable method for describing and treating IUAs, its relatively invasive nature is a major limitation for its use as a diagnostic test. An optimal diagnostic test should be non-invasive, safe, painless, inexpensive and applicable to all women regardless of their pretest probability of having a particular condition of interest. Diagnostic ultrasound is widely available and is the only test that satisfies these criteria. Despite this, sonographic features of adhesions do not feature in any of the current classification systems despite suggestions that the preoperative endometrial thickness on transvaginal scan may be of some prognostic value52. In addition, ultrasound is the only method that provides comprehensive information about all aspects of IUAs including cavity obliteration, obstruction and the functional state of the endometrium. 2D ultrasound is a sensitive method for the detection of IUAs and should be the first method employed. In women with concomitant pathology, such as fibroids or adenomyosis, visualization of the uterine cavity is sometimes poor. In these cases, SIS or GIS may be helpful in increasing diagnostic confidence. The endometrial thickness should be measured first on 2D ultrasound to obtain information about the endometrial responsiveness. It is important to state whether all endometrium appears fibrous or whether there are parts of the cavity in which the endometrium is normal. The intracavitary fluid, if present, should be excluded from the measurement. Similar to the ultrasound approach for the diagnosis of congenital uterine anomalies, all women with evidence of intrauterine adhesions on 2D ultrasound should be offered a 3D examination, with or without contrast. The aim of 3D scanning is to provide women and clinicians with detailed, self-explanatory images that contain information regarding the location and extent of cavity obliteration and obstruction. Obliteration should be described subjectively as focal, partial or complete (Figure 7). In addition, the extent of uterine cavity damage should be quantified by measuring the size of the obliterated area in the coronal plane and comparing it with the total surface area of the uterine cavity. This could be expressed as an obliteration ratio: O/C × 100 (%), where O is the obliterated area and C is the total cavity area (Figure 8). The areas of obstruction could be described in a similar fashion, stating the level of obstruction and percentage of the uterine cavity involved. The same approach should be used for both preoperative assessment and following surgical division of IUAs (Figure 9). The quality of ultrasound imaging has improved significantly in recent years, and has reached a level that enables its use as the primary diagnostic tool for IUAs. Ultrasound scanning is non-invasive, inexpensive and widely available, which makes it suitable for use in both low- and high-risk women. We believe that wider acceptance of the proposed ultrasound-based approach would facilitate future large-scale prevalence studies, which should help provide further information on the risk factors for developing IUAs, their early detection and the efficacy of different management strategies for preventing their formation. The ability of 3D ultrasound to provide clear panoramic images of the uterine cavity and express the extent of adhesions in quantitative terms is a major advantage, which could form the basis for more robust studies on the efficacy of both surgical and medical management of IUAs. Hysteroscopic adhesiolysis remains the best method for treating IUAs, but ultrasound has the potential to improve both the safety and efficacy of this procedure by providing intraoperative guidance in difficult cases and fast and objective feedback on the success of surgery, without the need for additional invasive procedures. This has the potential to decrease further the risks associated with repeated surgery, contribute to the cost-effectiveness of treatment and reduce discomfort and social inconvenience to women.