In the present study, the reproductive outcomes of 128 cases of IUA with different severities were reported. The predictive value of live birth rate using five evaluation systems including AFS, ESGE, March, Nasr and Chinese classification systems were compared to figure out an optimal evaluation system for IUA. As expected, the live birth rate following spontaneous conception closely related with the severities of IUA. Furthermore, although all five classification systems were capable to predict live birth rate, Nasr’s classification, with the highest AUC, might be the optimal prediction system for live birth after natural conception.
Previous reports implied that the severity of adhesions was negatively correlated with the reproductive outcome after hysteroscopic adhesiolysis[16,17]. There were 83.3% of women with mild adhesions achieved live births, whereas only 25% women with severe adhesion had live births[16]. American Association of Gynecological Laparoscopists reported a practice guideline on intrauterine adhesions which was developed in collaboration with ESGE in 2017. This guidelines also proposed that IUA should be accurately classified because prognosis of live birth is highly related to severities of adhesion, and it is impossible to endorse any specific system because it is difficult to compare various evaluation systems between studies[10]. Therefore, an accurate assessment of the IUA severities is one key factor for management.
The five evaluation systems in the present study evaluate IUA severities mainly based on hysteroscopic findings, but the specific evaluation items involved in the classifications are different. The earliest evaluation system of IUA, March classification[13] classifies adhesions as mild, moderate or severe types only based on the extent and type of adhesions observed during hysteroscopy, which is simple and practical. AFS classification[11] defined the severities of IUA into three stages (I - III) based on a combined evaluation of hysteroscopic or HSG findings and the menstrual pattern. It was the first evaluation system which takes the menstrual pattern into consideration. ESGE classification[12], classifies IUA as stages I to V with subtypes (seven stages in all) based on hysteroscopy and HSG findings. This classification is more complex but more accurate in the aspect of the extent, location and type of adhesions, hence experienced operational skill of hysteroscopy operator is required. Nasr’s classification[14] is the first classification that incorporate hysteroscopic findings with menstrual pattern as well as obstetric history, and correlates them with the prognosis of IUA. Nasr et al. considered the menstrual pattern as an important prognostic point. Besides, they believed that the type of adhesions, especially the tubular cavity, was more decisive to the severity and prognosis of IUA. The latest Chinese classification[15] was proposed based on AFS and ESGE classifications, including menstrual pattern and reproductive history, besides of observations under hysteroscopy.
It is reasonable to include risk factors affecting the reproductive outcomes as the basis of IUA classcification systems. The extent of uterine cavity involved and type of adhesions are included in all five classifications. This may be due to the fact that the extent and location of IUA are the dominant factors affecting reproductive outcomes[18], as also suggested in our results (Table 2). Taylor et al.[19] suggested that filmy IUA may lead to infertility, in addition to dense IUA. Nasr et al. believed that the prognosis was associated with the type of adhesions and whether or not they cover the tubal ostia[14]. An early study reported that a classification of IUA based on both adhesion types and of uterine cavity occlusion extent, is valuable in the prediction of reproductive outcome[20]. It is not clear the live birth outcomes of each IUA adhesion type. Here, in the present study, we did found that women with fibrous adhesion end up with high live birth rate.
Yu et al.[21] and Roy et al.[22] suggested that the menstrual pattern after hysteroscopic surgery had a significant impact on reproductive prognosis but no such correlation were observed between the menstrual pattern before hysteroscopic surgery and the prognosis of IUA. This could be explained by the fact that the menstrual pattern after surgery reflects the remaining amount of endometrium available for regeneration, which is critical for the implantation ability of endometrium[14,21]. However, conflicting with an earlier report[23], March et al. concluded that there was no perfect correlation between the extent of IUA and menstrual pattern[24], hence menstrual pattern was not contained in its classification system. Still, AFS, Nasr and Chinese evaluation systems included the menstrual pattern before hysteroscopic surgery as one factor to determine the IUA classification. Here in this study, among the 128 women who desired to conceive spontaneously in the following 24 months after TCRA, 93.8% of them achieved improvement in menstruation pattern, which is similar to He’s report that 91.30% of IUA patients showed with improvement in menstrual amount after TCRA[25]. However, similar proportions of women with and without live birth showed increased amount of menstrual blood following TCRA, which might indicate that improvement in menstrual amount is not closely related with reproductive outcomes. The association between menstrual pattern and reproductive outcome of IUA is still controversial, which might be because the report of menstrual pattern is quite subjective based on patients’ own evaluation.
The maximum endometrial thickness is another critical factor predicting the live birth outcomes of IUA patients. IUA occurs due to the disruption of the basalis layer of the endometrium[2]. Malhotra et al. believed that thin and damaged endometrium lost its endometrial receptivity and hampers implantation in women with Asherman’s syndrome[26]. In their series, by analyzing the endometrial thickness and Doppler flows in infertile women with IUA both pre- and post-hysteroscopic, they observed patients with flimsy adhesion achieved an improvement in the endometrial thickness, while patients with dense adhesions had no change in the endometrial thickness and blood flow and the pregnancy outcome remained poor. Bhandari et al.[27] checked changes in endometrial thickness with ultrasonography and found that the pregnancy outcomes of IUA patients were associated with postoperative endometrial echo pattern, which reflected that endometrial function after hysteroscopic adhesiolysis plays an important role in determining the reproductive outcome. Baradwan et al. reported that pregnancy rates were higher in endometrial thickness > 5 mm group (80%) than group with < 5 mm endometrium (38.46%)[28]. Researchers[29,30] also found that endometrial thickness was a significant and independent predictor of live birth after the in vitro fertilization/intracytoplasmic sperm injection (IVF/ICSI) treatment, and the live birth rate rise with endometrial thickness increasing. Live birth rate was significantly reduced when endometrial thickness less than 8 mm, and no adverse pregnancy outcome was observed when endometrial thickness more than 14mm[31]. However, lower implantation and pregnancy rates when endometrial thickness more than 14mm on the day of HCG administration[32]. Controversially, some researchers concluded that endometrial thickness has limited capacity to identify women who have a low pregnancy rate after IVF[33], and pregnancy outcome could not be predicted by endometrial thickness alone[34]. Overall, adequate endometrial thickness is required for successful pregnancy. According to the analysis above, we believe that endometrial thickness before and after adhesiolysis can be taken into consideration as prognosis points for live birth. Nevertheless, among the five evaluation systems in the present study, only Chinese classification includes the endometrial thickness as one of the parameters. Further studies are required to determine the optimal endometrial thickness for satisfactory reproductive outcomes of IUA women.
An ideal and clinically practical evaluation system of IUA should be simple, reproducible, quantifiable, and capable of predicting the reproductive prognosis. To date, no evaluation system obtained universal recognition, which might reflects inherent deficiencies in all of these proposed systems. In the present study, all the five evaluation systems were demonstrated to be valuable of predicting live birth. Among them, Nasr classification system demonstrated the highest AUC value, which might be due to its inclusion of most associated factors with live birth. On the contrary, March’s classification, which was established four decades ago, showed a low predictive value of live birth, which could be attributed to the simplified evaluation approach based only on hysteroscopic features.
It is difficult to compare live birth rates with other studies due to the different durations of follow-up, variable proportions of different severities, different post-surgery treatments and retrospective bias. One study reported that the live birth rate after hysteroscopic treatment was 63.7% (79/124) with 14-year follow-up[35], whereas the live birth rate in our series was 53.1% (68 /128), which may be in part due to the shorter duration of follow-up. Another study from one hospital in China, with the similar inclusion criteria and management of IUA to our study, reported a similar live birth rate of 42.2% (140/332) as in our study after following up for 27±9 months[36]. A pooled live birth rate, 46.3% (696/1504), was reported from a systematic review containing 36 studies, which was also similar to our study[37].
To our knowledge, this is the first study to compare the predictive value of live birth rate based on the five evaluation systems including AFS, ESGE, March, Nasr and Chinese classifications. However, this study is limited by its retrospective design and relatively small number of patients included. Besides, certain clinical characteristics, including the gravidity, parity, and body mass index was not obtained due to its retrospective nature.