Effect of hysteroscopic septum resection on IVF/ICSI outcomes in women with septate uterus: a propensity-score retrospective matching study

Abstract

Background

Hysteroscopic septum resection has been a widespread technique throughout the world in women with a septate uterus. The effectiveness of hysteroscopic septum resection before IVF was equivocal. This study aimed to investigate whether hysteroscopic septum resection could improve IVF outcomes for women with a septate uterus.

Methods

This retrospective case-control study was carried out from January 2017 to December 2022. We included subfertile women with septate uterus (n = 147), including 113 women undergoing hysteroscopic septum resection and 34 women opting for expectant management before IVF/ICSI. In addition, we selected 147 women with normal uteri using propensity score matching (PSM). The clinical outcomes of IVF/ICSI were compared by logistic regression and Cox proportional regression in each group.

Results

During follow-up time, clinical pregnancy occurred in 66 women who underwent hysteroscopic septum resection (58.41%), compared to 23 women opting for expectant management (67.65%, OR 1.783, 95% CI 0.725–4.380) and 87 women with normal uteri (59.18%, OR 2.601, 95% CI 1.202–5.631). Miscarriage occurred in 12 women who underwent hysteroscopic septum resection (18.18%) versus 5 women who had expectant management (21.74%, OR 1.098, 95% CI 0.323–3.734) and 8 women with normal uteri (9.20%, OR 0.278, 95% CI 0.078–0.997), while preterm birth occurred in 7 women who underwent hysteroscopic septum resection (10.61%) versus 1 women who had expectant management (4.35%, OR 0.509, 95% CI 0.056–4.666) and 9 women with normal uteri (10.34%, OR 1.472, 95% CI 0.291–7.447). Live birth occurred in 54 women following hysteroscopic septum resection (47.79%) compared to 18 women following expectant management (52.94%, OR 1.368, 95% CI 0.589–3.175) and 78 women with normal uteri (53.06%, OR 1.641, 95% CI 0.941–2.864).

Conclusions

Women who underwent hysteroscopic septum resection had lower clinical pregnancy rate and higher risk of miscarriage compared to those with normal uteri, but there was no significant difference in live birth rate between the groups.

Trial registration

This is a retrospective case-control study.

Background

A septate uterus is an anatomical anomaly resulting from the maldevelopment of the embryologic Müllerian. Septate uteri have a prevalence of 2.3% in the unselected population [1] and is estimated to be up to 1.5% in reproductive-age females [2]. A septate uterus is accounting for approximately 55% of Müllerian anomaly [3].

The specific cause and genetic relationship of the anomaly is unknown. The etiology of Müllerian anomalies is thought to be multifactorial [4]. Many patients with a septate uterus are asymptomatic. During the seventh week of female embryonic development, the absence of the Müllerian inhibitory factor triggered the regression of the mesonephric ducts and stimulated the development of two paramesonephric ducts [5]. These paramesonephric ducts eventually formed the upper two-thirds of the vagina, cervix, uterus, and fallopian tubes [6]. In the eighth week of development, the paramesonephric ducts underwent a vertical fusion that created a midline septum. In the twentieth week of embryonic development, the septum usually underwent complete bidirectional resorption. If resorption was halted, the septum remained, and the uterus would be septate uterus [7].

There are two main classification systems: the American Society of Reproductive Medicine (ASRM) classification, published by the American Fertility Society in 1988 [8] and modified in 2016 [9], 2021 [10] and 2024 [11], and the European Society of Human Reproduction and Embryology (ESHRE) and the European Society for Gynecological Endoscopy (ESGE) classification, published in 2013 by the European Societies of Endoscopic and Reproductive Gynecology [12]. ASRM defined a septate uterus with an indentation depth of > 15 mm and an indentation angle of < 90°. ESHRE defined a septate uterus using indentation-to-wall-thickness (I: WT) > 50%. In addition, the Congenital Uterine Malformation by Experts (CUME) suggested using an indentation depth of ≥ 1 cm [13].

According to ESHRE/ESGE classification, septate uteri is further divided into two sub-classes according to the degree of the uterine corpus deformity. A partial septate uterus characterizes by the existence of a septum dividing partly the uterine cavity above the level of the internal cervical os. A complete septate uterus characterizes by the existence of a septum fully dividing the uterine cavity up to the level of the internal cervical os [12]. Three-dimensional transvaginal ultrasound (3D transvaginal ultrasound) has the best performance capacity for the diagnosis of the septate uterus compared with two-dimensional transvaginal ultrasonography, two-dimensional transvaginal sonohysterography and three-dimensional transvaginal sonohysterography [14]. Septate uterus has been associated with an increase in the risk of miscarriage, premature delivery and malpresentation. However, there is insufficient evidence that a uterine septum is associated with infertility.

Current guidelines have different recommendations: the ASRM guideline for septate uterus recommends to remove the intrauterine septum [9]; the latest ASRM guideline for septate uterus recommendes to use a shared decision-making model after appropriate counseling to determine whether or not to proceed with septum resection [11].

In view of the persisting uncertainty around the effect of a septate uterus and hysteroscopic septum resection on clinical outcomes, we performed the study in subfertile women with a septate uterus. Women with a septate uterus chose for hysteroscopic septum resection or expectant management before in Vitro Fertilization (IVF) / Intracytoplasmic Sperm Injection (ICSI) - Embryo Transfer (ET). Our aim was to investigate whether hysteroscopic septum resection could improve reproductive outcomes in these women.

Methods

Trial design and participants’ characteristics

This study was a retrospective case-control study and was carried out in the period of January 2017 to December 2022 in Yantai Yuhuangding Hospital Affiliated to Qingdao University. Subjects were included if they: (1) were diagnosed with a septate uterus according to the ASRM classification [9]; (2) had a history of infertility; (3) underwent their first IVF/ICSI cycle. Infertility was a disease of the male or female reproductive system efined by the failure to conceive after at least one year of unprotected coitus [15]. Patients with infertility underwent IVF/ICSI due to fallopian tubes factor, male factor, chromosome abnormality, monogenic disorders and recurrent spontaneous abortion. Additionally, subjects were excluded if they suffered from intrauterine adhesion, chromosome abnormality and preimplantation genetic testing (PGT), immune and metabolic disease and endometrial lesions. The study protocol was approved by the Ethical Committee of Yantai Yuhuangding Hospital. Patients’ characteristics and cycle parameters were obtained from patient medical records.

Allocation to hysteroscopic septum resection or expectant management was dependent on the reproductive history and patient preference. The septate uterus was mainly ascertained by three-dimensional ultrasound screening (3D-US) and hysteroscopy. Hysteroscopic septum resection was performed before IVF/ICSI-ET.

Treatment protocols

Hysteroscopic uterine septum resection was performed in the operating room under general anesthesia, using an operative hysteroscope fitted with a bipolar needle and a loop with a saline infusion, after cervical dilatation. The surgical procedure commenced with identifying both tubal ostia. In partial septate uterus, the next step is the horizontal incision of the septum with the bipolar needle equidistantly between the anterior and posterior uterine walls, starting from the lower part of the septum and continuing upwards to reach the uterine fundus, preserving 10–15 mm of the fundic myometrial wall. A loop removed residual septum in anterior and posterior uterine walls until the opening of the fallopian tubes on both sides was visible to restore normal uterine cavity. In complete septate uterus, transverse incision of septum was carried out from 5 mm above the internal opening of the cervix. The following operation was performed in the same manner as partial septate uterus and the cervical septate was retained to avoid the incidence of cervical incompetence.

Patients undergoing IVF cycles received luteal phase gonadotrophin-releasing hormone (GnRH) agonist protocol, GnRH-antagonist protocol agonist or mild ovarian stimulation protocol. During gonadotrophin (Gn) stimulation, serum estradiol (E₂), follicle-stimulating hormone (FSH), luteinizing hormone (LH), progesterone (P), follicle size and endometrial thickness were monitored until the day of human chorionic gonadotropin (hCG) trigger. When two leading follicles reached a mean diameter of 18 mm, hCG was given to trigger ovulation. Oocytes were retrieved transvaginally 35–37 h after hCG administration. IVF or ICSI was performed. Ultrasound-guided ET was performed on the third or fifth day after oocyte retrieval, and excess viable embryos were cultured to blastocyst stage and cryopreserved for subsequent frozen embryo transfer (FET). ET was cancelled if the embryo and endometrium were not synchronous or women had a high risk of ovarian hyperstimulation syndrome (OHSS) or factors seriously affecting embryo implantation. In FET cycles, natural cycles, hormone replacement treatment (HRT) or ovulation stimulation were used in the endometrial preparation process. The luteal phase was supported with 200 mg progesterone (Utrogest™ 200, Besins-Iscovesco, France) vaginal medication three times daily from the day of oocyte retrieval in ET cycles or the day of progesterone administration in FET cycles. A quantitative early pregnancy test was performed on the 14th day after ET or FET. Clinical pregnancy was confirmed if gestation sac was observed by transvaginal ultrasound after 28 days from ET or FET. Miscarriage was defined as the spontaneous demise of a pregnancy before 24 weeks of gestation. Preterm birth was defined as birth before a gestational age of 37 complete weeks. Live birth was defined as the birth of a live child after at least 28 weeks of gestation.

The moment of initiation of Gn was time zero. All conceptions during follow-up were registered and only the first live birth was used. The primary outcome measure was cumulative live birth rate. Secondary outcomes were cumulative clinical pregnancy rate, miscarriage rate and preterm birth rate.

Statistical analyses

Statistical analysis was performed in SPSS 26.0 (IBM Corp., Armonk, NY, USA). Continuous variables with normal distribution were expressed as mean ± standard deviation, non-normal continuous variables as median (interquartile range). Kruskal-Wallis test or one-way analysis of variance was performed to compare continuous variables. Categorical variables were presented in terms of frequency and percentages. The chi-squared or Fishers’ exact test was employed to compare the difference of categorical variables. Propensity Score Matching (PSM) was used to adjust for confounding factors related to achieving pregnancy outcomes, the variables in the PSM included age, infertility duration, BMI, AMH and basal AFC, which allowed a part of women in two groups can be matched reciprocally with similar characteristics. To optimize the precision of the study, the match was conducted in a 1:1 matching ratio without replacement, and with a caliper width equal to 0.02 of the standard deviation of the logit of the propensity score. Standardized diferences were estimated. Univariate and multivariate logistic regression and Cox proportional regression were performed to explore the effects of different treatments on clinical outcomes. We plotted adjusted survival curves. Age, infertility duration, AMH and basal AFC were adjusted. The difference was considered to be significant if P < 0.05.

Results

Baseline characteristics

10,667 patients were recruited in the study. 236 patients suffered from uterine anomaly. Patients with unicornuate (n = 78) and double uterus (n = 11) were excluded. Finally, 147 patients with a septate uterus were included in the study. Of the 147 women with septate uterus, 113 women underwent hysteroscopic septum resection before IVF-ET (Group 1) and 34 women had expectant management (Group 2). At the same time, 10,431 patients were with normal uterus. Subjects were excluded if they suffered from intrauterine adhesion (n = 357), chromosome abnormality and preimplantation genetic testing (PGT) cycles (n = 468), immune and metabolic disease (n = 320) and endometrial lesions (n = 3413). 5873 patients without uterine anomaly were recruited in original cohort (Group 3). After propensity score matching (PSM), 147 women with septate uterus and 147 women with normal uterus (Group 4) were in PSM cohort. In total, 294 women were confirmed eligible and were included in the study for analysis (Fig. 1).

Fig. 1

Flowchart of the patient selection process

Baseline characteristics before and after PSM were summarized in Tables 1 and 2. In original cohort, serum AMH level was significantly lower in Group 2 than that in Groups 3 (P = 0.017 ). Women in Group 1 had more basal AFC compared to women in Group 3 (P = 0.001 ). No significant differences were found in age, infertility duration, BMI, previous pregnancy, previous live birth and infertility diagnosis among three groups. There was no statistically significant differences in classification of septate uterus between Group 1 and Group 2. After PSM, the baseline characteristics between the three groups reached a well balance in age, infertility duration, BMI, AMH, basal AFC, previous pregnancy, previous live birth and infertility diagnosis.

Table 1 Baseline characteristics before propensity score matching (PSM) among the three groups^a

Table 2 Baseline characteristics after propensity score matching (PSM) among the three groups^a

Treatment characteristics in IVF/ICSI

Treatment information were exhibited in Table 3. There were no statistically significant differences in Gn dosage, duration of stimulation, number of oocytes retrieved, fertilization type and number of available embryos among the three groups. There was no statistically significant differences in the number of treatment cycles to achieve a pregnancy among the three groups (P = 0.288) (Table 3).

Table 3 Ovarian stimulation parameters and embryology outcomes among the three groups after PSM^a

Treatment characteristics in hysteroscopic uterine septum resection

In Group 1, 100 patients underwent diagnostic hysteroscopy within one month to assess the uterine cavity. Another 13 patients who rejected diagnostic hysteroscopy were monitored with transvaginal ultrasound to assess the endometrium. All patients received evaluation. No operative complication such as intrauterine adhesion, perforation of uterus and fluid overload occurred.

Pregnancy outcomes

The pregnancy outcomes were presented in Table 4. Women who underwent hysteroscopic septum resection had lower cumulative clinical pregnancy rate and higher risk of miscarriage compared to those with normal uteri (58.41% vs. 59.18%) (OR 2.601, 95% CI 1.202–5.631, P = 0.015) (18.18% vs. 9.20%) (OR 0.278, 95% CI 0.078–0.997, P = 0.049). There were no significant differences between the groups on preterm delivery rate (10.61%, 4.35% and 10.34%, respectively) (OR 0.509, 95% CI 0.056–4.666; OR 1.472, 95% CI 0.291–7.447) and cumulative live birth rate (47.79%, 52.94% and 53.06%, respectively) (OR 1.368, 95% CI 0.589–3.175; OR 1.641, 95% CI 0.941–2.864).

Table 4 In vitro fertilization cycle clinical outcomes among the three groups after PSM: univariate and multivariate logistic analysis^a

Long-term reproductive outcomes

The flow of patients during the follow-up of the study is shown in Fig. 2. The duration of follow-up was 96 months. Overall, 54 of 131 patients who underwent hysteroscopic septum resection had at least one live birth, compared to 18 of 34 patients opting for expectant management (HR 0.680, 95% CI 0.392–1.179) and 78 of 147 patients with normal uteri (HR 0.880, 95% CI 0.611–1.268) (Table 5). There were no significant differences in cumulative live birth among the three groups, as can be seen in the survival plot adjusted for confounders (Fig. 3). Other obstetrics results were reported in Table 6. There were no significant differences in the obstetric complications among the three groups (Table 6).

Fig. 2

Flow diagram of live birth

Table 5 Adjusted cumulative live birth among the three groups after PSM: Cox proportional regression^a

Fig. 3

Adjusted cumulative live birth. Adjusted for age, AMH, basal AFC

Table 6 Obstetric complications among the three groups after PSM^b

Reproductive outcomes of women experiencing recurrent pregnancy loss (RPL)

In Group 1, two patients with a septate uterus experienced two consecutive miscarriages before hysteroscopic septum resection. In Group 2, only one patient presented the history of two consecutive miscarriages. Fortunately, all of them gave birth to term live singleton.

Discussion

A septate uterus might result in adverse reproductive outcomes such as infertility, recurrent pregnancy loss (RPL), or preterm delivery. When a uterine septum was diagnosed in patients with infertility or a history of poor obstetrical outcomes, septum resection could be considered with the appropriate patient counseling and shared decision-making [9], and one study supported hysteroscopic resection of the septum [16]. Nevertheless, the current evidence that such intervention improved reproductive outcomes was equivocal.

In observational studies, patients with a septate uterus were found to have higher rates of first-trimester miscarriage than those with a normal uterine cavity (42% vs. 12%); whereas, a septate uterus did not increase the risk of a second-trimester miscarriage or preterm delivery [17]. In our retrospective case-control study, after adjusted for age, infertility duration, AMH and basal AFC, women with normal uteri conducted higher cumulative clinical pregnancy rate and lower miscarriage rate when compared to women with a septate uterus and underwent hysteroscopic septum resection (P < 0.05). A recent 20-year retrospective analysis in Slovenia found 4 cases of uterine rupture in patients who achieved term pregnancy after undergoing hysteroscopic uterine septum resection [18]. Uterine rupture in pregnancy after a previous hysteroscopic septum resection was a very rare, but life-threatening event. Physiologic distention of the uterus during pregnancy might increase the risk of rupture in scarred uteruses. The course of uterine rupture in pregnancy after hysteroscopic surgery might depend on many variables, such as the type of surgery, presence or absence of complications such as uterine perforation, the interval between hysteroscopy and pregnancy and obstetric history [19]. Avoiding myometrial incision through judicious septal transection might prevent this catastrophic complication [20]. Prompt diagnosis could ensure successful management and avoided adverse maternal-fetal outcomes. In our study, up to now, no uterine rupture occurred in women who underwent hysteroscopic septum resection.

In this study, septum resection did not improve cumulative live birth rate, nor did it decrease preterm birth rate. This results were similar to the study of Rikken [21]. Another study reported that the proportion of live birth rate (LBR) was higher after the removal of septum and the proportion of spontaneous abortion (SA) and preterm labor (PL) was lower after the removal of the septum [22].

The risk of recurrent pregnancy loss was increased in women with congenital uterine malformations [23]. The septate uterus was the congenital malformation most commonly associated with RPL, being found in 6-16% of cases [24]. Miscarriages in case of septate uteri likely resulted from inadequate implantation of the embryo on a poorly vascularized septum. Furthermore, the septum could alter the preovulatory and postovulatory changes of the endometrium under the influence of estradiol and progesterone, uterine contractibility, and/or disruption of the physiology of endometrial factors such as, notably, vascular endothelial growth factor [2]. Complete uterine assessment should be performed in all women who present with a history of RPL. Much work remained to be done to better understand the clinical process of RPL. Resection of uterine septa, IUAs, submucosal myomas and endometrial polyps were indicated in women with RPL [25]. Hysteroscopic septoplasty resulted in an overall pregnancy rate of 71% for women with either primary infertility or secondary infertility with only one miscarriage and women who had experienced ≥ 2 consecutive miscarriages before 24 weeks of gestation. The spontaneous pregnancy rate was 45% and 8 pregnancies resulted from assisted reproductive technologies (ART) (26%). The overall first live birth rate was 51.6%. A decrease was seen in the miscarriage rate from 95.24 to 24% (p < 0.001) in the overall population. The results pointing out the benefits of surgery on the reproductive outcomes as well as the relatively simple and safe surgery made the intervention attractive [26]. Both abdominal and hysteroscopic metroplasty for septate uterus increased the take-home baby (THB) rate by preventing pregnancy loss without affecting the chance of pregnancy in women with the history of ≥ 3 miscarriages, or ≤ 2 miscarriages and unwilling to try the next conception without surgery [27]. In our study, only three woman presented the history of RPL. Two women accepted septum resection and all of them gave birth to term live singleton. Nevertheless, our study was not powered to detect small differences due to the small number women in the expectant group. The sample size of woman presented the history of RPL should be necessary to expand to assess whether this subgroup of women could derive more benefit from resection. This study revealed that live birth occured within 2 year in women who underwent septum resection and women with normal uteri through IVF/ICSI-ET treatment.

Strengths and limitations

Our study had a number of strengths. First, we collected data in recent six years, with a specifically large proportion of women. The relatively large sample size in combination with a long follow-up resulted in a unique study population. Second, since a retrospective study was by design at risk of selection bias, we adjusted for possible differences in prognostic factors using logistic regression and Cox proportional regression and matched patients cycles using PSM. Our study had also limitations. First, an adequately powered randomized controlled trial (RCT) obviously offered the best way to evaluate the effectiveness of septum resection. This was a propensity-score retrospective matching study. We tried to correct possible differences for patient characteristics, but our results could still be influenced by residual confounding. The retrospective nature of the study, possibility of bias which could not be corrected for and the use of PSM were important limitations. Second, the size of the septum could have influenced the choice of treatment of the patient or physician and possibly led to confounding by indication.

Conclusions

In summary, hysteroscopic septum resection did not lead to improved cumulative live birth rate nor did it decrease preterm birth rate in women with a septate uterus who underwent hysteroscopic septum resection compared with women with a septate uterus opting for expectant management and women with normal uteri. However, data regarding the efect of hysteroscopic septum resection on subfertile women with a septate uterus was not suffcient to draw a consensual conclusion. Future studies should continue to assess the efect of hysteroscopic septum resection on subfertile women with a septate uterus.