Chorioamniotic Membrane Separation after Fetal Spina Bifida Repair: Impact of CMS Size and Patient Management

In-utero fetal spina bifida (fSB) repair became a treatment option in selected cases of fSB since the Management of Myelomeningocele Study (MOMS) and later some of our own studies showed significant improvements in mental development and motor function, as well as reduced need for shunt placement in children with prenatal compared to postnatal spina bifida repair [1‒3]. Nevertheless, fSB repair is associated with an increased risk for several complications like postoperative chorioamniotic membrane separation (CMS), preterm premature rupture of membranes (PPROMs), or preterm birth (PTB) [1, 4, 5].

CMS is defined as the separation of the amniotic membrane from the chorion [6]. On ultrasound examination, it presents as a fine mobile line that is separated by a lucent space between the chorion and the uterine wall [6]. After fSB repair, incidence rates for CMS of up to 40% have been described [7‒10]. Soni et al. [8], as well as Wilson et al. [7], reported an increased risk for CMS after fSB repair if performed <23 gestational weeks (GW). CMS itself was found to increase the risk for PPROM and PTB [8, 11]. However, a study by Corroenne et al. [9] reported a higher risk for PPROM and PTB due to CMS only after fetoscopic but not after open fSB repair. CMS can cause umbilical cord injuries, due to wrapping of the amniotic bands around the cord, with the risk for fetal demise [12, 13]. Close follow-up of pregnancies complicated by CMS after fSB is therefore essential to assure quick and adequate action to ensure the most important goal, the fetal survival [7, 8, 13]. However, there is still no sound consent on how to manage women with postoperative CMS after fSB repair. Therefore, the aim of this study was to first analyze the frequency of CMS in a rather large single-center cohort, as well as its influence on the occurrence of PPROM and PTB, and second to assess if prophylactic hospitalization influences outcomes in women with CMS.

For this retrospective single-center cohort study, we analyzed data of 200 women, who consecutively underwent fSB repair at our center between December 2010 and January 2023. Two women were excluded due to missing informed consent and another four due to incomplete data regarding CMS occurrence, leaving a cohort of 194 women for further analysis. The study was approved by the local Ethic Committee Zurich (KEK-Nr. 2021-01101), and all women included gave their written informed consent.

Eligibility criteria for fSB repair, details on operative technique, as well as peri- and postoperative procedures have been published previously [14‒16]. Of importance for this study is that for the fSB repair we used two monofilament traction sutures that were placed through the entire uterine wall. In between all of these, the uterus was then first opened with electrocautery followed by extension of the hysterotomy that was performed using a stapling device with absorbable polyglycolic acid staples (Covidien Auto Suture, Norwalk, CT, USA) in the first 193 women. In all subsequent women, no stapling device was used. In these cases, the hysterotomy extension was performed with electrocautery followed by a suture incorporating the amniotic membrane and full thickness uterine wall to avoid CMS. After performing the fSB repair, closure of the uterus was done in a two-layered way using long-lasting polydioxanone sutures. These incorporated the uterine membranes, the absorbable stables (if used), and all of the myometrium.

At the very beginning of our program, a prospective data registry was created to record all pertinent data in a prospective and comprehensive way. All data for this study were retrieved from our REDCap^®-based repository. Baseline characteristics included maternal age and body mass index at surgery, parity, smoking status, and placental location. Perioperative findings comprised gestational age (GA) at fSB repair, preoperative fetal version, cervical length, uterine incision site, length of hysterotomy, durations of total surgery, hysterotomy, and fetal surgery. Regarding pregnancy outcomes, the following parameters were assessed: CMS, its severity, defined by the largest distance measured between chorion and amnion after fSB repair throughout the remaining course of the pregnancy (small = distance from chorion to amnion ≤20 mm; large = distance from chorion to amnion >20 mm) and its progression (increase of CMS >5 mm since diagnosis), and GA at occurrence, PPROM, GA at PPROM, duration from fSB repair to CMS and PPROM, histologically diagnosed chorioamnionitis (hCA), amniotic infection syndrome (AIS), amniotic fluid leakage, vaginal bleeding, abdominal pain, placental abruption, uterine rupture, amniotic fluid index 48 hours (h) postoperatively and before cesarean delivery, length of hospital stay (LOS) after fSB repair, overall rate of unplanned and planned prenatal readmissions, and hysterotomy site at delivery. Finally, fetal parameters included GA and weight at delivery, as well as live born and neonatal death rate.

At our center, women with large CMS were generally hospitalized, whereas women with small CMS were usually monitored for some time and then discharged, unless further complications did occur. Delivery of women who previously had fSB repair was regularly scheduled at 37 0/7 GW, unless additional complications did occur, that require earlier delivery. Reasons for earlier delivery were the following: pathologic cardiotocography (CTG), failure of inhibiting preterm uterine contractions, strong vaginal bleeding, placental abruption, uterine rupture, and AIS. PPROM itself did not lead to changes regarding the timing of delivery.

Outcomes of women with CMS (group A) were compared to the ones without CMS (group B). Additionally, a regression analysis was conducted to assess independent risk factors for CMS. Furthermore, a first subgroup analysis was performed comparing outcomes of women with small CMS (group A-small) to the ones with large CMS (group A-large). Additionally, a second subgroup analysis was performed in women with CMS to analyze the value of a prophylactic hospitalization (group A1) compared to no hospitalization (group A2) (shown in Fig. 1). Women in group A1 remained hospitalized after fSB repair or were re-hospitalized due to CMS diagnosis. Women in group A2 were discharged although CMS was diagnosed or, respectively, not re-hospitalized due to CMS diagnosis but received regular check-ups once a week with ultrasound and CTG.

Statistical analysis was performed using IBM SPSS Statistics (version 29, IBM, USA). Shapiro-Wilk test was used to test for normal distribution. Descriptive statistics were used to analyze outcome parameters of the cohort. To compare the groups, the Mann-Whitney-U test was used for quantitative data, whereas the Pearson-Chi-square test or Fisher’s exact test were used to compare categorical variables. Furthermore, a logistic regression analysis was performed to analyze risk factors for CMS. Quantitative data are presented as mean ± standard deviation in case of normal distribution or median with interquartile range (25 and 75 percentiles) if normal distribution was not given. Categorical variables are given as numbers (N) with percentages (%). A p value of <0.05 was defined as statistically significant.

Of 194 women (100%), 23 (11.9%) were diagnosed with CMS (group A), whereas in 171 (88.1%) women, no CMS was detected (group B) (shown in Fig. 1). Regarding baseline characteristics, presented in Table 1, as well as perioperative findings shown in Table 2, no significant differences were seen between group A and B.

CMS in group A was small in 13 (56.5%) of cases, and large CMS occurred in 10 (43.5%) cases. GA at CMS diagnosis was 29.2 ± 3.2 GW. Regarding CMS progression, 5/23 (21.7%) women with CMS showed progression after CMS diagnosis. Thereof, three women had small CMS at initial diagnosis but progressed to large CMS over the course of the pregnancy, and two women with small CMS showed progression but not to the extent to reach large CMS. When comparing group A and B regarding further pregnancy outcomes (shown in Table 3), the following significant differences were seen: group A had a significantly higher rate of PPROM (69.6% vs. 24.1%, p = <0.001), shorter duration from fSB repair to PPROM (46.8 ± 17.7 vs. 59 [50.5–69.5] days, p = 0.03), higher rate of AIS (22.7% vs. 7.1%, p = 0.03), and hCA (40.0% vs. 14.7%, p = 0.03). The occurrence of vaginal bleeding did not significantly differ. There was a higher placental abruption rate in group A compared to B (17.4% vs. 4.7%, p = 0.05) that almost reached statistical significance. Furthermore, group A reported a significantly longer LOS after fSB repair (35 [19–65] vs. 17 [14–27] days, p < 0.001) and overall LOS (43 [33–71] vs. 35 [27–46] days, p = 0.004) compared to group B. On the other hand, the rate of unplanned re-hospitalization due to spontaneous labor as well as planned re-admission was significantly lower in group A (4.3% vs. 24.3%, p = 0.02 and 21.7% vs. 56.5%, p = 0.003). No significant differences existed regarding further pregnancy complications like abdominal pain, uterine rupture, amniotic fluid index, or hysterotomy status at delivery.

Regarding fetal parameters (shown in Table 4), GA at delivery was significantly lower in group A compared to group B (35.3 [32.3–36.3] GW vs. 36.7 [34.9–37.0] GW, p = 0.006), whereas no significant differences existed regarding fetal weight at delivery as well as live born rate or neonatal death rate. There was no fetal or neonatal death in group A, whereas group B reported one each. Regression analysis did not show any independent risk factor for CMS occurrence (shown in Table 5).

Subgroup analysis, comparing women with small CMS (group A-small) to the ones with large CMS (group A-large), shown in Table 6, showed a significantly higher rate of AIS in group A-small compared to group A-large (41.7% vs. 0.0%, p = 0.04). When analyzing their management after CMS diagnosis in more detail, results showed that 2/5 women with small CMS and AIS were not hospitalized due to CMS, whereas 3/5 were hospitalized due to CMS, however two of them only temporary with discharge before delivery. The only other significant differences were a significantly shorter LOS after fSB repair in group A-small (22 [16–43] days vs. 60 ± 24 days, p = 0.003) and significantly shorter overall LOS in group A-small compared to group A-large (35 [32–45] days vs. 66 ± 18 days, p = 0.002). No differences were seen regarding baseline and perioperative characteristics.

A further subgroup analysis, comparing women with CMS (group A) according to their management (group A1 = prophylactic hospitalization vs. group A2 = no hospitalization) showed the following: of 23 women with CMS, 18 (78.3%) were prophylactically hospitalized (group A1), whereas five (21.7%) were not but underwent regular check-ups. A total of 10/18 (55.6%) women that were hospitalized had large CMS, whereas 8/18 (44.4%) had small CMS. The comparison of group A1 with A2 did not reveal any significant differences in regard of baseline characteristic as well as perioperative findings. Table 7 shows pregnancy outcomes of group A1 and A2. Apart from significantly longer LOS after fSB repair (49.3 ± 22.8 days in group A1 vs. 15 (15–17.5) days in group A2, p < 0.001), as well as overall LOS (55.4 ± 20.1 days in group A1 vs. 34.0 ± 2.7 days in group A2, p = 0.04), and lower rate of planned re-admission (5.6% in group A1 vs. 80.0% in group A2, p = 0.003), no significant differences existed between groups A1 and A2.

Results of this study confirmed that CMS represents a relevant complication after open fSB repair and its occurrence increases the risk for additional complications like PPROM and PTB that will be discussed in more detail. The CMS rate in our cohort of 194 women was 12%, whereof 57% were diagnosed with small CMS and 43% with large CMS. These numbers are considerably lower compared to other studies. A review by Licci et al. [10] reported CMS rates of up to 66% in women after fSB repair; however, the review included both studies with open and fetoscopic fSB repairs. When only including studies with open fSB repair, the CMS rate was 20% [10], still remarkably higher than ours. Soni et al. [8], who also analyzed CMS rates after open fSB repair in 88 patients, reported an even higher rate of 24%, with 48% global and 52% local CMS. Various studies did not find any relationship between CMS extension and the occurrence of PPROM or preterm labor/birth [8, 9]. However, multiple studies reported that GA at fetal surgery is a significant risk factor for CMS occurrence, especially if performed <23 GW [7, 8, 17]. In contrast to these studies, our study did not show any significant relationship between the GA at fSB repair and CMS occurrence. This is most probably because in our cohort fSB repair was done relatively late, i.e., around 25 GW in both women with and without CMS. Only one women was operated <23 GW at the beginning of our fetal surgery program. However, she was not diagnosed with CMS.

Similarly to the results of Soni et al. [8], who reported an increased rate of PPROM (59.1%) in women with CMS after open fSB repair compared to the ones without (21.2%) (p = 0.008), our results showed higher rates of PPROM in women with CMS (69.6%) that differed significantly from women without CMS (24.1%) (p < 0.001). Additionally, GA at PPROM was significantly lower in women with CMS and occurred around 2 weeks earlier compared to women without CMS. Apart from CMS, nulliparity and GA at fSB repair have been described as risk factors for PPROM after open fSB repair, while another study only reported an anterior placenta to be a risk factor for CMS in women after fetoscopic fSB repair [8, 9]. In contrast, none of these factors were associated with CMS occurrence in our cohort. In fact, we did not detect any risk factor for CMS occurrence.

According to a review by Goldenberg et al. [18], 25% of all PTBs are caused by PPROM. It is therefore not surprising that in our cohort, women with CMS, who were significantly more often affected by PPROM, also had a significantly lower GA at delivery (around 35 GW) in comparison to women without CMS (around 37 GW) (p = 0.006). Nevertheless, GA at delivery in women with CMS in our cohort was higher when compared to results of Soni et al. [8], who reported a GA at delivery of around 32 GW in women with CMS after open fSB repair.

It is generally known that infections and inflammatory processes promote PPROM [18]. Accordingly, since our results showed higher PPROM rates in women with CMS, the rate of AIS as well as hCA was significantly higher in women with CMS. This matches the findings of Sydorak et al. [19] who likewise reported a higher rate of chorioamnionitis in women with CMS after fetal surgery (including open and fetoscopic fetal surgeries). Furthermore, the rate of placental abruption in women without CMS in our cohort was 4.7% and therefore similar to previously reported numbers [10]. However, the placental abruption rate in women with CMS was more than 3 times higher (17.4%) compared to the ones without CMS and almost reached statistical significance (p = 0.05).

When reflecting on ways to reduce further complications after CMS, especially PPROM, PTB, and fetal cord injuries, we assumed that a prophylactic hospitalization, namely, putting these women on partial bed rest and monitoring the fetus regularly with CTG, might work. Ultimately, this management enabled quick and adequate action in case further complications would occur. Since umbilical cord injuries are especially feared in case of global CMS with consecutive amniotic band formation [12], these patients are most commonly hospitalized for closer surveillance, so they were at our center. Luckily, results of this study did not show any umbilical cord injury in women with CMS, irrespectively of CMS size. However, the CMS rate in our cohort was overall low. Results furthermore showed that five women with small CMS were hospitalized. When looking into more detail, two of these women were only diagnosed with small CMS when presenting for planned readmission around 34 0/7 GW. The other three women stayed hospitalized for some time after CMS diagnosis, when further complications (PPROM in 2 cases and pulmonary lung embolism in one case) occurred, which caused continuous hospitalization. When comparing pregnancy outcomes among women with CMS that were hospitalized, to the ones that were not, the only significant differences between these two groups were a longer LOS after fSB repair, as well as overall LOS in women that were hospitalized, as well as a higher unplanned and planned readmission rate in women who were not hospitalized. The subgroup analysis that compared outcomes according to CMS size showed similar results. LOS after fSB repair and overall LOS were longer in women with large CMS since all women with large CMS were hospitalized. Apart from that, the only other significant difference was a higher rate of AIS in women with small CMS compared to the ones with large CMS (42% vs. 0%, p = 0.04). Since more of these women were treated as outpatients, future studies with larger sample sizes should analyze if hospitalization might help to prevent AIS. Knowing that pregnancy outcomes were otherwise not altered due to prophylactic hospitalization, further larger studies are needed, to confirm if women with CMS, irrespectively of CMS size, might be safely managed as outpatients with regular check-ups as long as no further complications occur that require a different management.

In contrast to possible approaches trying to reduce complications after CMS diagnosis, the study group of Olejek et al. [20] analyzed a different hysterotomy technique with the goal to reduce CMS occurrence upfront. Their technique for open fSB repair consists of a hysterotomy of 6–7 cm, performed without stapler, as well as a three-layered closure of the uterus, including one first layer that fixes the membranes to the inner half of the uterine wall and therefore closes the amniotic cavity while preventing contact of amniotic fluid with the incised uterine muscle [20]. With this approach, they followed the hypothesis of Pomini et al. [21] that assumed that amniotic fluid infiltration, containing high amounts of prostaglandins, into the myometrium stimulates contractions and therefore promotes CMS, PPROM, and PTB [20]. Interestingly, using this technique in 49 women, they reported a considerably lower CMS rate of 6% and PPROM rate of 15% [20] compared to our results. Nevertheless, when comparing GA at delivery as one of the most important outcome parameters, results showed a lower GA at delivery within their total cohort (34.4 ± 3.4 GW), despite lower CMS and PPROM rates, when compared to women with CMS in our cohort (35.3 [32.3–36.3] GW). This finding points out the complex interaction of multiple factors influencing outcome variables including GA at delivery.

To our knowledge, this is the largest study analyzing CMS after fSB repair and the first one to assess a potential benefit of prophylactic hospitalization. The study confirmed the increased risk of PPROM as well as PTB in women with CMS. Additionally, we showed that outcomes of women with and without prophylactic hospitalization after CMS diagnosis did not differ, except for longer LOS. This knowledge is important to further improve patient management. Pregnancies with CMS need to be monitored closely; however, women may not need to stay in hospital for weeks but rather be managed as outpatients. Nevertheless, the study has limitations. One is the small sample size of the CMS cohort, which resulted from generally low CMS diagnosis within our whole cohort. This limits the generalizability of the results of the subgroup analyses. Additionally, GA at diagnosis of CMS might be overestimated since CMS was diagnosed during the scheduled regular check-ups after fSB repair but may have occurred at any time point since the last check-up. Furthermore, since this was a retrospective study, definition of CMS size was retrospectively made based on 2D ultrasound pictures. In addition, there might be a bias since in some cases, not only the CMS diagnosis but also additional concurrent complications led to prolonged hospitalization or re-hospitalization.

This study confirms that CMS after open fSB repair significantly increases the risk for PPROM and lowers GA at delivery. Furthermore, CMS increases the risk for AIS, as well as hCA, and significantly prolongs LOS. Apart from higher AIS rates and lower LOS in women with small CMS, outcomes did not differ to the ones with large CMS. A prophylactic hospitalization because of CMS may not be mandatory.

The authors want to first thank all families who participated in this research. In addition, we thank our contributing study group without which this research would not be possible. From the University Children’s Hospital, this includes Barbara Casanova, Thomas Dreher, Ruth Etter, Patrice Grehten, Domenic Grisch, Annette Hackenberg, Cornelia Hagmann, Maya Horst Luethy, Raimund Kottke, Niklaus Krayenbuehl, Claudia M. Kuzan-Fischer, Markus A. Landolt, Bea Latal Hajnal, Andreas Meyer-Heim, Theres Moehrlen, Svea Muehlberg, Beth Padden, Silke Quanz, Brigitte Seliner, Mithula Shellvarajah, Sandra P. Toelle, Julia Velz, Alexandra Wattinger, and Noemi Zweifel. From the University Hospital Zurich, our study group consists of Dirk Bassler, Lukas Kandler, Salome Meyer, and Christian Schaer. Additionally, we thank the Clinical Trial Center, University Hospital Zurich, for hosting the REDCap^® Database.

This study protocol was reviewed and approved by the local Ethic Commission Zurich (KEK-ZH, Nr. 2021-01101). The study was carried out in compliance with the Declaration of Helsinki and all women included in this study gave their written informed consent.

None of the authors has any conflict of interest to declare.

No funding sources exist for this study.

J.W.-Z. and N.O.-K. designed the study outline. N.S. and J.Z. did data collection and data quality control. J.Z. performed data analysis and wrote the manuscript. J.Z., L.V., L.R., F.K., N.S., L.M., M.M., U.M., and N.O.-K. participated in revising of the manuscript and contributed to its intellectual content. Prior to publication, all mentioned authors approved the final version of the manuscript.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from J.W.-Z. upon reasonable request. Further inquiries can be directed to the corresponding author.