Abstract
Objective
To develop a risk prediction model for preterm premature rupture of membranes (PPROM) after fetal endoscopy laser coagulation (FLC) for Twin-to-Twin Transfusion Syndrome (TTTS), identifying related influencing factors.
Methods
A retrospective analysis was conducted on 414 parturients from six Chinese hospitals treated with FLC for TTTS between January 2016 and January 2025. Patients were divided into non-occurrence (n = 263) and occurrence groups (n = 151) based on PPROM during pregnancy. Logistic regression identified predictors, establishing the risk prediction model.
Results
Univariate and multivariate logistic analyses revealed postoperative amniotic fluid leakage, operation time, and intraoperative amniotic fluid reduction rate as risk factors, while preoperative cervical canal length and Trocar insertion distance from the internal cervical OS were associated with reduced risk (all P < 0.05). ROC curve analysis showed an area under the curve of 0.802 (95% CI: 0.700–0.904), indicating good predictive efficacy. The goodness-of-fit test (P = 0.165 > 0.05) confirmed model fit, and calibration curves demonstrated acceptable accuracy. Clinical decision curves indicated net benefit.
Conclusion
This model, incorporating postoperative amniotic fluid leakage, operation time, intraoperative amniotic fluid reduction rate, preoperative cervical canal length, and Trocar insertion distance, effectively predicts PPROM risk after FLC for TTTS. It can assess risks clinically and guide interventions to prolong pregnancy and improve outcomes.
Keywords: Fetal endoscopy laser coagulation, Twin-to-twin transfusion syndrome, Preterm premature rupture of membranes, Regression analysis, Prediction model
What does this study add to the clinical work
| Our study suggests preterm premature rupture of membranes (PPROM) after fetal endoscopy laser coagulation (FLC) for Twin-to-Twin Transfusion Syndrome (TTTS) could be predicted using a model. This model incorporates postoperative amniotic fluid leakage, operation time, intraoperative amniotic fluid reduction rate, preoperative cervical canal length, and Trocar insertion distance to assess the PPROM risk after FLC for TTTS. Given the sound performance of this mode, it can help guide interventions to prolong pregnancy and improve outcomes. |
This model effectively predicts PPROM risk after FLC for TTTS using postoperative amniotic fluid leakage, operation time, intraoperative amniotic fluid reduction rate, preoperative cervical canal length, and Trocar insertion distance. This proposed model can assess risks clinically and guide interventions to prolong pregnancy and improve outcomes.
Introduction
Twin-twin transfusion syndrome (TTTS) represents one of the most severe complications in monochorionic twin pregnancies and remains a focal point of contemporary research on complex twin gestations. Occurring in approximately 10–15% of monochorionic multiple pregnancies [1–3], TTTS arises primarily from an imbalance in blood perfusion due to abnormal numbers or distributions of placental vascular anastomoses between the two fetuses [4]. This pathological cascade manifests as polyuria, polyhydramnios, edema, myocardial hypertrophy, and heart failure in recipient fetuses, while donor fetuses suffer oliguria, oligohydramnios, intrauterine growth restriction, and even intrauterine fetal demise. Left untreated, TTTS not only jeopardizes fetal survival but also predisposes surviving fetuses to a spectrum of neurological insults, including periventricular leukomalacia, intracranial hemorrhage, and hydrocephalus, profoundly impairing neurodevelopmental outcomes [5]. Fetoscopic laser coagulation (FLC), widely regarded as the gold-standard intervention for TTTS [6], addresses the underlying pathophysiology by selectively ablating aberrant vascular connections [7]. This therapeutic approach significantly mitigates the risk of central nervous system injury and enhances both short-term neonatal survival and long-term neurodevelopmental prospects [8]. Following FLC, the perinatal survival rate of twins exceeds 60%, with over 90% of cases preserving at least one viable fetus [9, 10]. Advances in FLC technology have shifted postoperative complications away from life-threatening events, such as intrauterine fetal death and miscarriage toward obstetric challenges like preterm premature rupture of membranes (PPROM) and preterm birth, which nonetheless remain critical determinants of pregnancy outcomes [11]. PPROM, currently the most prevalent complication following FLC, occurs in 11– 50% of cases and correlates strongly with procedural factors, such as membrane perforation during surgery and gestational age at intervention [12]. Notably, earlier interventions—particularly those performed before 17 week gestation [13]—are associated with heightened risks of PPROM. The resultant preterm membrane rupture complicates subsequent fetal endoscopic procedures, thereby constraining the broader clinical applicability of this otherwise transformative technology.
With respect to the etiology of preterm premature rupture of membranes (PPROM) following fetoscopic laser coagulation (FLC) surgery, the prevailing hypothesis attributes it to trauma induced by fetal endoscopy [14]. Prior investigations have posited that a larger diameter of the insertion port for the endoscope, along with chorionic separation at this site, constitute high-risk factors predisposing to PPROM post-surgery [15, 16]. Although both in vitro and in vivo animal studies have corroborated the absence of healing capacity in the amniotic membranes [17, 18], most cases of amniotic membrane rupture during routine amniocentesis are self-limiting and tend to resolve spontaneously, leading to favorable pregnancy outcomes [19]. Despite this, the incidence of PPROM after fetal endoscopy remains relatively elevated, and its underlying causes, as well as associated high-risk factors, remain underreported and lack consensus within the scientific community. In a previous study, it has been documented that there were no statistically significant differences between the PPROM group and the non-PPROM group in maternal demographics or preoperative parameters, including the amount of amniotic fluid in the gestational sac of the twin pregnancy, the amount of reduced amniotic fluid, and the position of the placenta [20]. A later study showed that early surgery (< 17 weeks) carries a higher risk of postoperative PPROM [13]. However, no study has been conducted to fully appreciate all risk factors for PPROM following FLC. In light of these gaps, the present study undertakes a comprehensive review of the pathogenic factors contributing to PPROM following FLC. By treating these pathogenic factors as influencing variables, a predictive model was developed to swiftly identify the high-risk population susceptible to PPROM post-FLC, thereby enabling precise management strategies to mitigate hazard factors and avert adverse sequelae. The findings of this research are summarized below.
Data and methods
Clinical information
A retrospective analysis was performed on the clinical data of 414 parturients who underwent fetoscopic laser coagulation (FLC) treatment for twin–twin transfusion syndrome (TTTS) across six medical institutions in China between January 2016 and January 2025. Based on the occurrence or non-occurrence of preterm premature rupture of membranes (PPROM) during pregnancy, the cohort was stratified into two groups: 263 cases in the non-occurrence group and 151 cases in the occurrence group.
The diagnostic criteria for TTTS Presently, the globally recognized diagnostic criteria are as follows: Ultrasound examination reveals that in one of the monochorionic twins (the recipient fetus), the maximum vertical pocket (MVP) of amniotic fluid measures ≥ 8 cm (≥ 10 cm after 20 weeks of gestation), while in the other fetus (the donor), the MVP is ≤ 2 cm. For staging, the Quintero clinical staging standard was adopted: Stage I: The largest amniotic fluid pool in the recipient fetus exceeds 8 cm (≥ 10 cm after 20 weeks of gestation), and the largest amniotic fluid pool in the donor fetus is ≤ 2 cm. Stage II: The bladder of the donor fetus is persistently not visualized under ultrasound imaging. Stage III: Doppler ultrasound detects blood flow abnormalities in one of the twins, including absent or reversed end-diastolic flow in the umbilical artery, reverse A-wave in the ductus venosus, or pulsatile flow in the umbilical vein. Stage IV: Edema develops in one or both fetuses. Stage V: Intrauterine fetal demise occurs in at least one fetus.
Inclusion criteria (1) Fetoscopic laser coagulation (FLC) treatment was administered between 16 and 26 weeks of gestation (in accordance with the recommended gestational age guidelines issued by the Chinese Medical Association in 2015); (2) cases included those classified as Quintero stages II–IV, or stage I accompanied by abdominal distension and progressively worsening disparities in amniotic fluid volume; (3) following a definitive diagnosis, patients underwent treatment and were monitored through regular follow-ups until the conclusion of pregnancy; (4) clinical data records were comprehensive, with all relevant symptoms, physical examinations, and ancillary diagnostic tests meticulously documented and verified for accuracy; (5) participants voluntarily consented to the study, signing an informed consent form authorizing clinical data analysis and longitudinal follow-up.
Exclusion criteria (1) Individuals who underwent sequential amniotic fluid reduction or amniocentesis either prior to or following fetoscopic laser coagulation (FLC) surgery; (2) cases involving selective fetal reduction surgery performed subsequent to FLC treatment; (3) patients experiencing severe post-FLC surgical complications, such as life-threatening infections or significant hemorrhage; (4) instances where the pregnancy was terminated prematurely due to other maternal medical complications of grave concern; (5) pregnancies voluntarily discontinued by the mother for non-medical reasons; (6) perinatal clinical data that are either incomplete or contain documented errors.
The research followed the 2013 Declaration of Helsinki ethical guidelines, was approved by the Ethics Committee, and obtained voluntary informed consent from patients and their families after fully explaining the study details.
Surgical methods
Following a comprehensive ultrasound evaluation to assess the severity of the condition, the intrauterine status of the fetus, and the spatial positioning of the placenta and umbilical cord, patients and their families were thoroughly informed about the surgical plan, perioperative risks, potential postoperative complications, and the long-term prognosis for the newborn. Upon obtaining informed consent via the surgical consent form, fetoscopic laser coagulation (FLC) surgery was performed under local anesthesia. In cases where preoperative ultrasound revealed a maternal cervical length ≤ 2 cm, progressive shortening of the cervix, or an open internal os necessitating concomitant cervical cerclage during the procedure, epidural anesthesia was administered. Each operation was conducted by an experienced physician using selective vascular communication branch coagulation. For challenging cases requiring a straight fetal endoscope (Trocar diameter 8Fr), such as anterior or lateral placentation, a curved fetoscope (Trocar diameter 10Fr) was employed. During the procedure, intravenous sedatives were administered as an adjunct for frequent fetal movements that could interfere with surgical precision. If turbid amniotic fluid was observed within the amniotic cavity, concurrent amniotic fluid replacement was performed. Following laser coagulation of the vessels, amniotic fluid reduction was achieved through the Trocar puncture channel, ultimately reducing the maximum depth of amniotic fluid in the recipient fetus to 7 cm. Postoperatively, both fetal hearts were confirmed to be functioning normally. Contraindications for FLC surgery include the presence or suspicion of intrauterine infection, abnormal coagulation function, severe structural or chromosomal abnormalities in the fetus, an excessively large anterior wall placental area rendering surgery unfeasible, and threatened miscarriage.
Candidate predictor
According to the literature, readily accessible clinical, ultrasound, and laboratory indicators were selected as candidate variables for analysis. These included maternal age, height, body mass index (BMI), gravidity, parity, history of adverse pregnancy or childbirth outcomes, family medical history, history of uterine surgery, uterine malformations, mode of conception, history of threatened miscarriage, time interval from symptom onset to surgery (in days), percentage difference in twin weights, Quintero clinical staging, presence of trace-element deficiencies, maximum amniotic fluid depth (DVP) for recipient and donor fetuses prior to surgery, and preoperative maternal blood test results [including red blood cell count (RBC), white blood cell count (WBC), platelet count (PLT), hemoglobin level (HGB), neutrophil count (ANC) and percentage (NE%)]. Additionally, four coagulation parameters were assessed: prothrombin time (PT), activated partial thromboplastin time (APTT), thrombin time (TT), and fibrinogen (FIB). Preoperative cervical status was also documented, including whether the cervix was closed or open. For closed cervix cases, the length of the cervical canal was recorded without specifying its shape; for open cervix cases, both the length and shape (V or U) were noted. Other variables included placental position, use of prophylactic uterine contraction inhibitors, gestational age at the time of surgery, distance from the Trocar insertion point to the placenta during surgery, distance from the Trocar insertion point to the internal cervical os during surgery, use of an curved fetoscope during surgery, number of electrocoagulated anastomotic vessels during surgery, use of sedatives during surgery, and total operation time (in minutes). Intraoperative variables encompassed the volume of amniotic fluid reduction (mL), rate of amniotic fluid reduction (mL/min), middle cerebral artery MOM values for both recipient and donor fetuses, and whether combined cervical cerclage was performed. Postoperative assessments included blood test results on the first day after surgery [RBC, WBC, PLT, HGB, ANC, NE%, PT, APTT, TT, and FIB], cervical canal length on the first day after surgery, occurrence of postoperative uterine contractions, use of uterine contraction inhibitors postoperatively, survival of at least one fetus after surgery, incidence of postoperative amniotic fluid leakage, and duration of continued pregnancy following surgery.
Observation index
The follow-up information of treated patients throughout their pregnancies was comprehensively obtained through methods, such as telephone consultations, regular outpatient visits, and inpatient deliveries or abortions. Based on the occurrence or non-occurrence of preterm premature rupture of membranes (PPROM), the patients were categorized into two distinct groups: the non-PPROM group and the PPROM group.
Statistical analysis
The data were subjected to comprehensive statistical analysis using SPSS 25.0 and R 4.5.0 software. For continuous variables, those conforming to a normal distribution were described using the mean ± standard deviation, with intergroup comparisons performed via independent sample t tests. In cases where normal distribution was not observed, the median [P25, P75] was employed for descriptive purposes, and intergroup comparisons were conducted using rank sum tests. For categorical data, frequencies (%) were utilized for description, with Chi-square tests applied for intergroup comparisons. When the assumptions of the Chi-square test were not met, Fisher's exact probability method was employed as an alternative.
Construction of predictive model
First, potential predictors of outcome events (P ≤ 0.05) were identified through univariate logistic regression analysis. Subsequently, a multivariate logistic regression analysis employing the stepwise method (bidirectional) was conducted on the screened variables to further elucidate their roles. Based on the variables with P < 0.05 derived from the stepwise analysis, a nomogram for the predictive model was constructed. The robustness of the constructed nomogram was validated through 1000 bootstrap resampling operations. A calibration curve was generated to assess the accuracy and reliability of the model’s predictions. Additionally, the Hosmer–Lemeshow (HL) test was performed to evaluate the goodness of fit of the model. Furthermore, the receiver-operating characteristic (ROC) curve was analyzed to calculate key performance indicators, such as the area under the curve (AUC), sensitivity, and specificity. To evaluate its discriminatory power, a clinical Decision Curve Analysis (DCA) was constructed, enabling an assessment of the model’s clinical application value and quantifying its net benefit across the threshold probability range.
Results
Examine the clinical characteristics of the sample
A total of 414 cases were enrolled in the study, of which 151 cases were identified in the preterm premature rupture of membranes (PPROM) group, corresponding to an incidence rate of 36.47% (151/414). The detailed research materials are presented in Table 1 (baseline table).
Table 1.
Baseline features of recruited subjects
| Variables | Total (n = 414) | PPROM 0 (n = 263) | PPROM 1 (n = 151) | P |
|---|---|---|---|---|
| Age | 31.89 ± 6.49 | 29.48 ± 5.54 | 29.48 ± 5.54 | 0.284 |
| Height | 29.42 ± 5.44 | 161.71 ± 5.06 | 161.47 ± 5.51 | 0.763 |
| BMI | 24.34 ± 3.22 | 24.35 ± 3.14 | 24.33 ± 3.38 | 0.966 |
| Gravida, median | 2.10 ± 1.25 | 2.04 ± 1.18 | 2.21 ± 1.37 | 0.184 |
| Parity, median | 0.63 ± 0.75 | 0.60 ± 0.72 | 0.69 ± 0.80 | 0.251 |
| Adverse pregnancy history, n (%) | 0.817 | |||
| No | 385 (93.0) | 232 (88.2) | 139 (92.1) | |
| Yes | 29 (7.0) | 31 (11.8) | 12 (7.9) | |
| Family history, n (%) | 0.956 | |||
| No | 371 (89.6) | 101 (70.1) | 16 (69.6) | |
| Yes | 43 (10.4) | 43 (29.9) | 7 (30.4) | |
| Uterine operation history, n (%) | 0.414 | |||
| No | 360 (87.0) | 226 (85.9) | 134 (88.7) | |
| Yes | 54 (13.0) | 37 (14.1) | 17 (11.3) | |
| Uterine malformation, n (%) | 0.726 | |||
| No | 387 (93.5) | 245 (93.2) | 142 (94.0) | |
| Yes | 27 (6.5) | 18 (6.8) | 9 (6.0) | |
| Conception mode, n (%) | 0.873 | |||
| No | 328 (79.2) | 209 (79.5) | 119 (78.8) | |
| Yes | 86 (20.8) | 54 (20.5) | 32 (21.2) | |
| Threatened abortion history, n (%) | 0.341 | |||
| No | 304 (73.4) | 189 (71.9) | 115 (76.2) | |
| Yes | 110 (26.6) | 74 (28.1) | 36 (23.8) | |
| Time interval from onset to surgery | 15.21 ± 5.83 | 15.37 ± 6.15 | 14.93 ± 5.23 | 0.465 |
| Twin weight differences | 161.62 ± 5.22 | 32.15 ± 6.70 | 31.44 ± 6.12 | 0.657 |
| Quintero clinical staging, n (%) | 0.945 | |||
| I | 25 (6.0) | 17 (6.5) | 8(5.3) | |
| II | 54 (13.0) | 33 (12.5) | 21 (13.9) | |
| III | 287 (69.3) | 182 (69.2) | 105 (69.5) | |
| IV | 48 (11.6) | 31 (11.8) | 17 (11.3) | |
| Trace-element deficiency, n (%) | 0.876 | |||
| No | 347 (83.8) | 221 (84.0) | 126 (83.4) | |
| Yes | 67 (16.2) | 42 (16.0) | 25(16.6) | |
| Blood recipients DVP | 8.56 ± 0.43 | 8.55 ± 0.45 | 8.58 ± 0.40 | 0.450 |
| Blood donor infant DVP | 1.03 ± 0.48 | 1.06 ± 0.47 | 0.99 ± 0.48 | 0.168 |
| Preoperative RBC | 4.15 ± 0.55 | 4.14 ± 0.55 | 4.15 ± 0.54 | 0.860 |
| Preoperative WBC | 5.67 ± 1.04 | 5.69 ± 1.06 | 5.63 ± 1.01 | 0.570 |
| Preoperative PLT | 209.35 ± 55.56 | 209.98 ± 56.19 | 208.26 ± 54.62 | 0.762 |
| Preoperative HGB | 120.86 ± 18.74 | 121.01 ± 18.22 | 120.60 ± 19.68 | 0.833 |
| Preoperative ANC | 3.79 ± 1.24 | 3.73 ± 1.26 | 3.89 ± 1.21 | 0.206 |
| Preoperative NE | 61.50 ± 6.39 | 61.76 ± 6.60 | 61.05 ± 6.02 | 0.279 |
| Preoperative PT | 11.95 ± 0.62 | 11.94 ± 0.64 | 11.96 ± 0.60 | 0.779 |
| Preoperative APTT | 24.90 ± 1.16 | 24.95 ± 1.18 | 24.81 ± 1.13 | 0.224 |
| Preoperative TT | 15.53 ± 2.24 | 15.57 ± 2.22 | 15.46 ± 2.28 | 0.629 |
| Preoperative FIB | 2.99 ± 0.65 | 3.00 ± 0.62 | 2.98 ± 0.70 | 0.860 |
| Preoperative cervical status, n (%) | 0.304 | |||
| Closed | 364 (87.9) | 236 (89.7) | 128 (84.8) | |
| Type V | 29 (7.0) | 15 (5.7) | 14 (9.3) | |
| Type U | 21 (5.1) | 12(4.6) | 9 (6.0) | |
| Preoperative cervical length | 1.90 ± 0.98 | 2.01 ± 0.91 | 1.70 ± 1.07 | 0.002 |
| Placental position, n (%) | 0.696 | |||
| Anterior | 76 (18.4) | 47 (17.9) | 29 (19.2) | |
| Lateral | 54 (13.0) | 38 (14.4) | 16 (10.6) | |
| Fundus | 32 (7.7) | 19 (7.2) | 13(8.6) | |
| Posterior | 252 (60.9) | 159 (60.5) | 93 (61.6) | |
| Preoperative uterine contraction inhibitor, n (%) | 0.861 | |||
| No | 305 (73.7) | 193 (73.4) | 112 (74.2) | |
| Yes | 109 (26.3) | 70 (26.6) | 39 (25.8) | |
| Gestational weeks of surgery | 20.89 ± 2.15 | 20.92 ± 2.19 | 20.84 ± 2.09 | 0.732 |
| Trocar insertion point to the placenta | 6.09 ± 1.14 | 6.11 ± 1.11 | 6.07 ± 1.20 | 0.722 |
| Distance from trocar insertion site to internal cervical os | 14.52 ± 3.69 | 14.91 ± 3.13 | 13.82 ± 4.44 | 0.004 |
| Curved fetoscope, n (%) | 0.716 | |||
| No | 325 (78.5) | 238 (90.5) | 143 (94.7) | |
| Yes | 89 (21.5) | 25 (9.5) | 8 (5.3) | |
| Electrocoagulation anastomosis branches | 8.10 ± 2.63 | 8.08 ± 2.62 | 8.13 ± 2.64 | 0.834 |
| Sedative, n (%) | 0.128 | |||
| No | 381 (92.0) | 136 (94.4) | 21 (91.3) | |
| Yes | 23 (8.0) | 8 (5.6) | 2 (8.7) | |
| Operation time | 31.49 ± 5.76 | 30.98 ± 5.77 | 32.36 ± 5.67 | 0.019 |
| Amniotic fluid reduction volume | 1181.11 ± 581.20 | 1172.17 ± 453.10 | 1196.69 ± 755.71 | 0.680 |
| Amniotic fluid reduction rate | 58.83 ± 9.80 | 57.77 ± 8.14 | 60.69 ± 11.96 | 0.003 |
| Recipient MOM value | 1.02 ± 0.30 | 1.00 ± 0.30 | 1.04 ± 0.30 | 0.180 |
| Donor MOM value | 0.99 ± 0.31 | 0.98 ± 0.30 | 1.00 ± 0.32 | 0.472 |
| Cervical cerclage, n (%) | 0.894 | |||
| No | 325 (78.5) | 207 (78.7) | 118 (78.1) | |
| Yes | 89 (21.5) | 56 (21.3) | 33 (21.9) | |
| Postoperative RBC | 3.75 ± 0.65 | 3.76 ± 0.66 | 3.74 ± 0.62 | 0.729 |
| Postoperative WBC | 9.70 ± 2.54 | 9.66 ± 2.58 | 9.76 ± 2.48 | 0.703 |
| Postoperative PLT | 201.03 ± 58.37 | 203.85 ± 58.48 | 196.13 ± 58.06 | 0.196 |
| Postoperative HGB | 104.88 ± 15.02 | 105.73 ± 14.93 | 103.40 ± 15.09 | 0.128 |
| Postoperative ANC | 3.55 ± 1.38 | 3.53 ± 1.36 | 3.59 ± 1.40 | 0.683 |
| Postoperative NE | 74.15 ± 5.16 | 74.37 ± 5.05 | 73.76 ± 5.34 | 0.242 |
| Postoperative PT | 13.14 ± 1.09 | 13.11 ± 1.12 | 13.18 ± 1.06 | 0.525 |
| Postoperative APTT | 26.99 ± 2.92 | 26.93 ± 3.03 | 27.09 ± 2.72 | 0.592 |
| Postoperative TT | 13.42 ± 1.59 | 13.44 ± 1.58 | 13.37 ± 1.62 | 0.661 |
| Postoperative FIB | 2.82 ± 0.55 | 2.83 ± 0.56 | 2.80 ± 0.54 | 0.642 |
| Postoperative cervical length | 2.62 ± 0.91 | 2.65 ± 0.88 | 2.58 ± 0.96 | 0.412 |
| Uterine contractions, n (%) | 0.533 | |||
| No | 388 (93.7) | 245 (93.2) | 143 (94.7) | |
| Yes | 26 (6.3) | 18 (6.8) | 8 (5.3) | |
| Postoperative uterine contraction inhibitor, n (%) | 0.277 | |||
| No | 291 (70.3) | 180 (69.4) | 111 (73.5) | |
| Yes | 123 (29.7) | 83 (31.6) | 40 (26.5) | |
| At least one fetus survived, n (%) | 0.086 | |||
| No | 27 (6.5) | 13 (4.9) | 14 (9.3) | |
| Yes | 387 (93.5) | 250 (95.1) | 137 (90.7) | |
| Amniotic fluid leakage, n (%) | < 0.001 | |||
| No | 344 (83.1) | 233 (88.6) | 111 (73.5) | |
| Yes | 70 (16.9) | 30 (11.4) | 40 (26.5) | |
| Duration of continued pregnancy | 64.91 ± 22.42 | 66.32 ± 22.38 | 62.46 ± 22.36 | 0.092 |
Logistic regression analysis
The regression coefficient (β), standard error (S.E.), Z value, P value, odds ratio (OR), and 95% confidence interval (CI) were calculated for variables associated with preterm premature rupture of membranes (PPROM) following fetoscopic treatment for twin–twin transfusion syndrome (TTTS). The results demonstrated that in the univariate logistic analysis, postoperative amniotic fluid leakage, operation duration, and the rate of intraoperative amniotic fluid reduction emerged as significant risk factors for postoperative PPROM, whereas preoperative cervical canal length and the distance between the Trocar insertion point and the internal cervical os during surgery were associated with reduced risk against postoperative PPROM (all P < 0.05). Detailed findings are presented in Table 2. Significant variables identified from the univariate analysis were subsequently incorporated into a multivariate logistic regression model. The results consistently revealed that postoperative amniotic fluid leakage, operation duration, and the rate of intraoperative amniotic fluid reduction remained critical risk factors for postoperative PPROM. Conversely, preoperative cervical canal length and the distance between the Trocar insertion point and the internal cervical os during surgery continued to be associated with reduced risk against PPROM occurrence postoperatively (all P < 0.05). Specific results are summarized in Table 3 (multivariate logistic regression results).
Table 2.
Single risk factor analyzed with univariate logistic regression
| Variable | Β | SE | Z | OR | P |
|---|---|---|---|---|---|
| Age | −0.007 | 0.019 | 0.137 | 0.993 | 0.711 |
| Height | −0.010 | 0.020 | 0.238 | 0.990 | 0.626 |
| BMI | −0.003 | 0.032 | 0.008 | 0.997 | 0.930 |
| Gravida, Median | 0.701 | 0.384 | 3.325 | 2.015 | 0.068 |
| Parity, Median | 0.250 | 0.307 | 0.664 | 1.284 | 0.415 |
| Adverse pregnancy history, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −0.113 | 0.406 | 0.078 | 0.893 | 0.780 |
| Family history, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −0.429 | 0.358 | 1.439 | 0.651 | 0.230 |
| Uterine operation history, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −0.232 | 0.314 | 0.548 | 0.793 | 0.459 |
| Uterine malformation, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −0.146 | 0.423 | 0.119 | 0.864 | 0.730 |
| Conception mode, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | 0.055 | 0.252 | 0.048 | 1.057 | 0.827 |
| Threatened abortion history, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −0.224 | 0.235 | 0.909 | 0.799 | 0.340 |
| Time interval from onset to surgery | −0.012 | 0.018 | 0.461 | 0.988 | 0.497 |
| Twin weight differences | −0.017 | 0.016 | 1.100 | 0.983 | 0.294 |
| Quintero clinical staging, n (%) | |||||
| I | 0.000 | Reference | |||
| II | 0.302 | 0.512 | 0.348 | 1.352 | 0.555 |
| III | 0.204 | 0.446 | 0.209 | 1.226 | 0.648 |
| IV | 0.153 | 0.524 | 0.085 | 1.165 | 0.770 |
| Trace-element deficiency, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | 0.043 | 0.276 | 0.024 | 1.044 | 0.876 |
| Blood recipients DVP | 0.185 | 0.235 | 0.616 | 1.203 | 0.432 |
| Blood donor infant DVP | −0.337 | 0.222 | 2.315 | 0.714 | 0.128 |
| Preoperative RBC | 0.036 | 0.187 | 0.038 | 1.037 | 0.846 |
| Preoperative WBC | −0.087 | 0.101 | 0.741 | 0.916 | 0.389 |
| Preoperative PLT | 0.000 | 0.002 | 0.061 | 1.000 | 0.805 |
| Preoperative HGB | −0.001 | 0.005 | 0.052 | 0.999 | 0.820 |
| Preoperative ANC | 0.103 | 0.082 | 1.568 | 1.109 | 0.211 |
| Preoperative NE | −0.018 | 0.016 | 1.233 | 0.982 | 0.267 |
| Preoperative PT | 0.040 | 0.164 | 0.058 | 1.040 | 0.810 |
| Preoperative APTT | −0.109 | 0.089 | 1.505 | 0.897 | 0.220 |
| Preoperative TT | −0.022 | 0.046 | 0.217 | 0.979 | 0.642 |
| Preoperative FIB | −0.019 | 0.159 | 0.014 | 0.981 | 0.906 |
| Preoperative cervical status, n (%) | |||||
| Closed | 0.000 | Reference | |||
| Type V | 0.543 | 0.387 | 1.962 | 1.721 | 0.161 |
| Type U | 0.324 | 0.454 | 0.509 | 1.383 | 0.476 |
| Preoperative cervical length | −0.324 | 0.107 | 9.159 | 0.723 | 0.002 |
| Placental position, n (%) | |||||
| Anterior | 0.000 | Reference | |||
| Lateral | −0.399 | 0.382 | 1.089 | 0.671 | 0.297 |
| Fundus | 0.075 | 0.436 | 0.029 | 1.077 | 0.864 |
| Posterior | −0.061 | 0.271 | 0.051 | 0.941 | 0.821 |
| Preoperative uterine contraction inhibitor, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −0.053 | 0.234 | 0.052 | 0.948 | 0.820 |
| Gestational weeks of surgery | −0.032 | 0.049 | 0.436 | 0.968 | 0.509 |
| Trocar insertion point to the placenta | −0.045 | 0.092 | 0.237 | 0.956 | 0.626 |
| Distance from trocar insertion site to internal cervical os | −0.081 | 0.029 | 7.884 | 0.923 | 0.005 |
| Curved fetoscope, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −0.094 | 0.254 | 0.139 | 0.910 | 0.710 |
| Electrocoagulation anastomosis branches | 0.008 | 0.039 | 0.044 | 1.008 | 0.833 |
| Sedative, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −1.366 | 0.827 | 2.729 | 0.255 | 0.099 |
| Operation time | 0.044 | 0.018 | 5.747 | 1.045 | 0.017 |
| Amniotic fluid reduction volume | 0.000 | 0.000 | 0.029 | 1.000 | 0.866 |
| Amniotic fluid reduction rate | 0.031 | 0.011 | 8.608 | 1.031 | 0.003 |
| Recipient MOM value | 0.497 | 0.349 | 2.027 | 1.644 | 0.155 |
| Donor MOM value | 0.298 | 0.334 | 0.798 | 1.347 | 0.372 |
| Cervical cerclage, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | 0.070 | 0.250 | 0.079 | 1.073 | 0.779 |
| Postoperative RBC | −0.053 | 0.159 | 0.113 | 0.948 | 0.737 |
| Postoperative WBC | 0.018 | 0.040 | 0.201 | 1.018 | 0.654 |
| Postoperative PLT | −0.002 | 0.002 | 1.573 | 0.998 | 0.210 |
| Postoperative HGB | −0.010 | 0.007 | 2.104 | 0.990 | 0.147 |
| Postoperative ANC | 0.032 | 0.074 | 0.185 | 1.033 | 0.667 |
| Postoperative NE | −0.024 | 0.020 | 1.408 | 0.977 | 0.235 |
| Postoperative PT | 0.058 | 0.094 | 0.382 | 1.060 | 0.537 |
| Postoperative APTT | 0.017 | 0.035 | 0.224 | 1.017 | 0.636 |
| Postoperative TT | −0.028 | 0.064 | 0.192 | 0.972 | 0.661 |
| Postoperative FIB | −0.082 | 0.188 | 0.191 | 0.921 | 0.662 |
| Postoperative cervical length | −0.099 | 0.112 | 0.771 | 0.906 | 0.380 |
| Uterine contractions, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | 0.951 | 0.867 | 1.205 | 2.590 | 0.272 |
| Postoperative uterine contraction inhibitor, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −0.247 | 0.233 | 1.132 | 0.781 | 0.287 |
| At least one fetus survived, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | −0.725 | 0.406 | 3.182 | 0.485 | 0.074 |
| Amniotic fluid leakage, n (%) | |||||
| No | 0.000 | Reference | |||
| Yes | 1.043 | 0.269 | 15.031 | 2.838 | < 0.001 |
| Duration of continued pregnancy | −0.008 | 0.005 | 2.918 | 0.992 | 0.088 |
Table 3.
Single risk factor analyzed with multivariate logistic regression
| Variable | β | SE | Z | OR (95% CI) | P |
|---|---|---|---|---|---|
| Amniotic fluid leakage | |||||
| No | 0.000 | Reference | |||
| Yes | 0.936 | 0.281 | 11.058 | 2.549 (1.468,4.424) | 0.001 |
| Preoperative cervical length | −0.237 | 0.111 | 4.542 | 0.789 (0.635,0.981) | 0.033 |
| Distance from trocar insertion site to internal cervical os | −0.060 | 0.030 | 4.138 | 0.942 (0.889,0.998) | 0.042 |
| Operation time | 0.042 | 0.019 | 4.943 | 1.043 (1.005,1.082) | 0.026 |
| Amniotic fluid reduction rate | 0.028 | 0.011 | 6.633 | 1.028 (1.007,1.051) | 0.010 |
Nomogram
The five predictors identified in the multi-factor prediction model were incorporated into a nomogram constructed using R software, as illustrated in Fig. 1. The first row of the nomogram displays the scores associated with each individual variable. For each variable, draw a vertical line upward to determine its corresponding score. Sum the scores attributed to all variables to derive the total score. Next, locate the total score in the final row of the nomogram and then draw a vertical line downward to identify the corresponding point in the last row, thereby obtaining the predicted risk probability for the outcome of interest.
Fig. 1.
Nomogram constructed using R software with five predictors identified in the multi-factor prediction model
The receiver-operating characteristic (ROC) curve
This predictive model demonstrates a relatively robust discriminatory capacity. The results of the receiver-operating characteristic (ROC) curve analysis revealed that the area under the ROC curve for the model was 0.802 (95% CI: 0.700, 0.904), indicating that the model exhibits commendable predictive efficacy, as shown in Fig. 2. The goodness-of-fit test yielded a P value of 0.165 > 0.05, suggesting an excellent fit of the model. The sensitivity and specificity were determined to be 0.856 (95% CI: 0.721, 0.938) and 0.755 (95% CI: 0.626, 0.865), respectively. Additionally, the median predicted success probability was 0.876, with a cut-off value of 0.655, a maximum approximate index of 0.652, a positive predictive value of 0.583, and a negative predictive value of 0.893.
Fig. 2.
ROC curve of the predictive model
Calibration curve
The constructed nomogram model underwent rigorous bootstrap sampling validation (with 1000 internal bootstrap resampling verifications), and a calibration curve was subsequently plotted. The results demonstrated that the predicted probability of the model closely aligned with the actual incidence rate, exhibiting an average absolute difference of merely 0.043, thereby indicating the model’s exceptional accuracy. In Fig. 3, the horizontal axis denotes the predicted probability, while the vertical axis represents the actual probability. Specifically, “Apparent” corresponds to the actual sampling curve of the model, “bias-corrected” signifies the curve after deviation correction of the actual curve, and “Ideal” represents the ideal line where the predicted probability perfectly matches the actual probability—an indication of the optimal scenario.
Fig. 3.
Calibration curve for the constructed nomogram model
Clinical decision curve
The vertical axis represents the net benefit, which quantifies the overall advantage derived from the model in terms of patient outcomes. Specifically, the net benefit reflects the true positive rate of the model, as these patients are accurately predicted and thus can derive clinical benefits. Conversely, the cost is represented by the false-positive rate, where patients receive unnecessary clinical interventions due to incorrect predictions by the model. Therefore, a higher net benefit indicates a greater practical application value of the model. The horizontal axis denotes the threshold probability, which serves as the critical cutoff used within the model to determine whether an event has occurred. In the Fig. 4, the red curve illustrates the net benefit of intervention for patients across various risk thresholds within the constructed prediction model. Within the risk threshold probability range of 0.03–0.96, the model demonstrates a significant net benefit. Notably, the net benefit of intervention based on the constructed model surpasses both the universal intervention approach and the no-intervention approach.
Fig. 4.
Clinical decision curve for the constructed prediction model
Discussion
Compared with spontaneous premature rupture of membranes, invasive prenatal interventions often lead to a condition referred to as iatrogenic premature rupture of membranes. Over the past few decades, advancements in fetal endoscopy and open fetal surgery have introduced groundbreaking treatment options for conditions, such as lower urinary tract obstruction, complex monochorionic pregnancies, and congenital diaphragmatic hernia [21]. Currently, the efficacy of fetal endoscopic laser therapy in many medical institutions has achieved remarkable results, with survival rates of at least one fetus ranging from 75 to 90%, and both twins surviving in 60–70% of cases. The incidence of neurological complications has been reduced to approximately 5–20% [22]. However, studies have also revealed [23] that the complication rate associated with fetal endoscopic laser surgery remains notably high, significantly impacting pregnancy outcomes. Among these complications, the incidence of preterm premature rupture of membranes (PPROM) is particularly alarming, accounting for 17–30% of all complications. Furthermore, PPROM carries potential long-term sequelae, including pulmonary hypoplasia secondary to oligohydramnios, chorioamnionitis, preterm birth, and developmental delays. These complications are among the leading causes of adverse perinatal outcomes, contributing to neonatal morbidity and even mortality [24].
Both univariate and multivariate logistic analyses in this study consistently demonstrated that postoperative amniotic fluid leakage, operation duration, and the rate of intraoperative amniotic fluid reduction were significant risk factors for postoperative preterm premature rupture of membranes (PPROM). Conversely, the preoperative cervical canal length and the distance between the distance from trocar insertion site to internal cervical os during surgery was associated with reduced risk against PPROM. The preoperative cervical length and the presence or absence of uterine contractions exhibited a clear correlation with fetal outcomes following fetoscopic laser coagulation (FLC) surgery. Prior studies have established that the reliability of ultrasound-measured cervical length in predicting preterm birth is comparable between twin pregnancies and singleton pregnancies. However, the threshold for cervical length appears higher in twin pregnancies, likely due to their inherently elevated risk of preterm birth compared to singleton pregnancies [25]. This study confirmed an inverse correlation between cervical length and the occurrence of PPROM after surgery, yet no association was identified between postoperatively induced uterine contractions and PPROM incidence. Given the excessive emphasis on inhibiting uterine contractions in Chinese medical institutions included in this study and the relatively small sample size of uterine contraction cases, further evaluation of this factor remains challenging. A larger-scale study is warranted for definitive conclusions. The distance between the distance from trocar insertion site to internal cervical os was tightly associated with reduced risk against PPROM following FLC surgery. Specifically, a shorter distance correlates with a heightened risk of postoperative PPROM, findings consistent with previous studies [26]. This phenomenon may be attributed to the restricted range of motion of the Trocar, which applies greater external force to the device, thereby exacerbating fetal membrane damage—a mechanism corroborated by clinical observations. Thus, selecting an appropriate Trocar insertion position constitutes a key consideration in FLC surgery. In vitro studies on fetal membrane healing using human fetal membrane explants cultured under complete in vitro conditions revealed that although epithelial cells proliferated and migrated, and the explants survived up to 12 days, no healing of the central defect was observed [27]. Similarly, animal experiments provided no evidence of spontaneous healing, with proliferation limited to the chorion layer [28]. Consequently, amniotic fluid leakage during FLC surgery remains relatively common. Moreover, such leakage elevates the risks of infection, oligohydramnios, and increased stress resistance of the amniotic membrane, ultimately heightening the likelihood of PPROM. It is widely acknowledged that the localized nature of microscopic video technology makes it challenging to fully expose the entire placenta during fetal endoscopic surgery. The anterior wall of placental vessels, due to their anatomical positioning, poses an even greater challenge for exposure. This significantly elevates the complexity of fetal microsurgery, potentially leading to incomplete laser-induced vascular coagulation. Simultaneously, the heightened surgical difficulty increases the duration of the procedure, thereby amplifying the risks of intraoperative bleeding and infection. These factors collectively impact the prognosis of fetoscopic surgeries. However, in this study, no significant association was identified between placental position and the risk of postoperative preterm premature rupture of membranes (PPROM). Notably, arc-shaped fetoscopic treatment of anterior placentas shortened the operation time without increasing PPROM risk, although a positive correlation between operation time and surgical outcomes was observed. Therefore, the author posits that neither surgical instruments nor placental position are directly linked to PPROM risk. Nevertheless, attention must be directed toward operation time, as prolonged procedures extend the duration of changes in uterine cavity pressure and morphology, consequently heightening the likelihood of fetal membrane rupture. Furthermore, this study revealed a significant correlation between the rate of amniotic fluid reduction during fetoscopic laser coagulation (FLC) and the occurrence of postoperative PPROM. The author speculates that excessive amniotic fluid reduction during surgery may induce rapid alterations in uterine cavity pressure and morphology, increase the strain rate on the fetal membranes, and reduce their elasticity, thereby predisposing patients to PPROM. Previous studies have generated controversy regarding factors, such as maternal age, prior history of preterm birth, gestational age at surgery, and recipient infant deep venous pressure (DVP) ≥ 14 cm. In this study, clinical baseline characteristics, including age, height, BMI, gravidity, parity, history of adverse pregnancy and childbirth, family history, history of uterine surgery, uterine malformations, conception mode, history of threatened miscarriage, gestational age at surgery, and time from symptom onset to surgery were comprehensively analyzed. No association was found between these factors and the risk of postoperative PPROM. Prior research has confirmed that the Quintero clinical stage is closely correlated with fetal outcomes following fetoscopic intervention [29]. However, in this study, indicators, such as Quintero clinical stage, weight difference between twins, and DVP in preoperative recipient and donor infants for assessing the severity of twin–twin transfusion syndrome (TTTS), were not associated with postoperative PPROM occurrence. The author concludes that postoperative PPROM is not directly linked to fetal outcomes. For decades, deficiencies in trace elements, such as copper and zinc, have been implicated as potential causes of PPROM, given the critical role of ascorbic acid in collagen synthesis and immune response to infection. However, this study failed to identify any correlation between trace-element deficiency and postoperative PPROM risk. The author further suggests that Chinese obstetricians place significant emphasis on laboratory testing during pregnancy and provide timely pharmaceutical and dietary supplementation. Consequently, cases of long-term trace-element deficiency among parturients are relatively rare, necessitating a larger sample size to validate this factor. Additionally, since Chinese medical institutions rigorously monitor and promptly correct blood routine and coagulation parameters before and after FLC, cases of PPROM clearly attributable to infection or other factors were excluded from this study. As a result, no outcome-related indicators were identified within normal laboratory results. The preoperative open or closed status of the cervix, as well as the use of uterine contraction inhibitors before and after surgery, were not found to be significantly associated with the risk of postoperative preterm premature rupture of membranes (PPROM). This may stem from the fact that cases included in the study utilized a higher frequency of uterine contraction inhibitors, thereby reducing the incidence of uterine contractions. A simple open cervix was also not directly linked to PPROM risk. Furthermore, factors, such as the distance between the Trocar insertion point and the placenta during surgery, the number of electrocoagulation anastomoses performed, the use of sedatives during the procedure, the volume of amniotic fluid reduction (mL), and the concurrent application of cervical cerclage, were not significantly correlated with the occurrence of postoperative PPROM. These findings suggest that, compared to absolute volume changes, the speed and frequency of these changes may exert a more pronounced influence on PPROM risk. This hypothesis warrants further mechanistic investigation for validation. An increase in the peak cerebral artery velocity (MCA-PSV) of the fetus during the perioperative period has been shown to elevate the risk of intrauterine fetal demise following surgery. MCA-PSV is typically measured using median multiples of the median (MOM), and twin transfusion syndrome (TTTS)-related blood transfusions may result in fluctuations in hemoglobin levels among donor and recipient infants. Prior studies have confirmed that MOM values of the middle cerebral artery significantly impact fetal outcomes [30]. However, no correlation was identified between MCA-PSV and the occurrence of postoperative PPROM. The author concludes that the occurrence of postoperative PPROM is not directly tied to fetal outcomes. Additionally, three indicators closely related to fetal outcomes—cervical canal length on the first postoperative day, survival of at least one fetus after surgery, and duration of continued pregnancy post-surgery—were also not found to correlate with the occurrence of postoperative PPROM. These results further substantiate the notion that postoperative PPROM occurrence is not directly linked to fetal outcomes.
This study demonstrates that postoperative amniotic fluid leakage, operation duration, intraoperative amniotic fluid reduction rate, preoperative cervical canal length, and the distance between the Trocar insertion point and the internal cervical os during surgery are independent influencing factors for the risk of preterm premature rupture of membranes (PPROM) following fetoscopic treatment for twin transfusion syndrome (TTS). Among these factors, postoperative amniotic fluid leakage, operation duration, and the intraoperative amniotic fluid reduction rate serve as risk factors for postoperative PPROM, whereas preoperative cervical canal length and the distance between the Trocar insertion point and the internal cervical os act was associated with reduced risk. These variables were incorporated into a regression analysis to construct a predictive model. The efficacy of this model was rigorously validated: the area under the curve (AUC) of the model was 0.802, indicating its high accuracy in predicting PPROM occurrence after surgery. Calibration curve results revealed that both the working curve and the deviation-corrected curve exhibited trends consistent with the ideal curve and remained relatively close, with an average absolute difference of merely 0.043, thus confirming the excellent calibration of the nomogram. Additionally, the Hosmer–Lemeshow test yielded a P value of 0.165, further indicating that the goodness-of-fit of this prediction model is notably high and statistically indistinguishable from the ideal model. Moreover, clinical decision curve analysis demonstrated that, under identical threshold probabilities, predictions generated using the nomogram in this study provided greater clinical benefit compared to extreme curve strategies, thereby affirming the clinical applicability of this model. Collectively, these validations confirm the establishment of a robust risk prediction model for preterm PPROM following fetoscopic TTS treatment. In the future, software development for this model could enable clinicians to perform individualized patient evaluations via web-based platforms or mini-applications, facilitating timely and precise assessments.
Limitations of this study: (1) Given the retrospective nature of this study, attention should be paid to the reliability of data collection and the limitations of selection bias. Our findings are based on a retrospective cohort study; thus, it is necessary to conduct prospective studies to verify the clinical effectiveness of our findings. (2) Although cases potentially associated with premature rupture of membranes due to infection or other factors were excluded, it remains challenging to strictly differentiate between iatrogenic and spontaneous preterm premature rupture of membranes (PPROM). This distinction requires further clarification in future studies. (3) Given the short-term nature of this study, maternal–fetal outcomes were not followed up systematically. Long-term follow-up of pregnant women and their offspring is essential for comprehensive evaluation. However, due to significant population mobility in China in recent years, a substantial number of participants were lost to follow-up, precluding an assessment of treatment safety. Consequently, it is impossible to determine whether parturients experienced symptom recurrence or whether offspring developed short-term or long-term complications. (4) Although this study was conducted as a multicenter retrospective analysis, most participating research units were concentrated in the eastern plain region of China, with only one medical institution from the western area involved. This geographical distribution limits the persuasiveness regarding regional differences. The small variations in region, ethnicity, and dietary structure hindered the ability to conduct a thorough analysis of ethnic and racial disparities. A larger-scale multicenter trial encompassing diverse regions is warranted for validation. (5) In this study, some fetal endoscopic surgeries were performed by colleagues from external hospitals. Researchers relied solely on complete surgical records provided by these institutions for quantifying evaluation indicators, making it impossible to assess the impact of the specific fetal endoscopy technique on outcomes. (6) None of the surgeries included in this study utilized tissue sealants. While there have been reports suggesting the use of tissue sealants to seal Trocar insertion points, their benefits remain unconfirmed in broader studies [31]. Due to strict adherence to current medical guidelines in China, excessive use of tissue sealants was not observed, rendering the evaluation of this factor infeasible. (7) The predictive model has not undergone external validation, which will limit its generalizability. Before clinical application, any new predictive model must undergo rigorous evaluation. Future efforts will focus on optimizing the model through prospective external validation and incorporating additional factors to correct biases, thereby enhancing its applicability and accuracy. In addition, high-quality clinical trials with larger sample sizes and longer follow-up periods are needed to validate the model proposed in this study.
In conclusion, this study developed a concise yet precise nomogram scoring system for predicting the risk of preterm premature rupture of membranes (PPROM) following fetoscopic treatment for twin-to-twin transfusion syndrome (TTTS). This predictive model is specifically tailored for Han women residing in the plain regions of China and serves as a valuable tool for fetal medicine specialists to identify high-risk populations susceptible to PPROM post-fetoscopic intervention for TTTS. By providing a robust foundation for clinical decision-making, the model facilitates the formulation of targeted intervention strategies and empowers clinicians to conduct individualized assessments of treatment efficacy while enabling timely adjustments to therapeutic approaches.
Acknowledgements
The study would like to thank each of the women who participated in the study, who have maintained contact with our hospital over the years, so that we can observe the safety and effectiveness of the treatment.
Author contributions
Xi Ye is the first author, responsible for research design and paper writing; Yulin Wu is the second author, responsible for data collection; Tian Le is the third author, responsible for the smooth progress of the research procedures; Zhiqing Song is the fourth author, responsible for contacting the funding project; Yanjie Cao is the fifth author, responsible for statistics; Yanyue Zhang is the sixth author, responsible for statistical analysis; Xuanxuan Hong is the seventh author, responsible for data classification; Le Yu is the eighth author, responsible for data summary; Liehong Wang is the corresponding author, responsible for the calibration and review of the paper. All authors reviewed the manuscript.
Funding
This research received funding from the clinical medical research center for obstetrics and gynecology diseases in Qinghai Province (Grant No. 2024-SF-L03).
Availability of data and materials
All data generated or analyzed during this study are included in this published article.
Declarations
Conflict of interest
The authors state that there are no conflicts of interest to disclose.
Consent for publication
Written informed consent was obtained from the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.
Human ethics approval and consent to participate
This study was approved by from the Medical Ethics Committee of Hefei Maternal and Child Health Hospital (YYLL20240130-YNXM-LL-01-2.4) and performed in accordance with the tenets of the Declaration of Helsinki. The requirement for written informed consent was waived by the Institutional Review Board of the Anhui Medical University and Qinghai University Clinical School of Medicine, because data were retrospectively analyzed.
Consent to participate
Written informed consent was obtained from all participating patients.
Consent for publication
Not applicable.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Jena LM (2021) Twin to twin transfusion syndrome. Transl Pediatr 10(5):1518–1529. 10.21037/tp-20-264 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Christian B, Kurt H (2019) Update on twin-to-twin transfusion syndrome. Best Pract Res Clin Obstet Gynaecol 58(7):55–65. 10.1016/j.bpobgyn.2018.12.011 [DOI] [PubMed] [Google Scholar]
- 3.Nicholas B, Henry LG (2016) Twin-twin transfusion and laser therapy. Curr Opin Obstet Gynecol 28(2):79–85. 10.1097/GCO.0000000000000247 [DOI] [PubMed] [Google Scholar]
- 4.Tollenaar L, Slaghekke F, Middeldorp JM, Lopriore E (2023) Fetal anemia in monochorionic twins: a review on diagnosis, management, and outcome. Expert Rev Hematol 16(1):9–16. 10.1080/17474086.2023.2166921 [DOI] [PubMed] [Google Scholar]
- 5.Lauren N, Rebecca F, Stephanie EM, Roshni W (2021) Review of international clinical guidelines related to prenatal screening during monochorionic pregnancies. J Clin Med 10(5):1128. 10.3390/jcm10051128 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Nicholas B, Henry LG (2021) Fetal growth in multiple gestations: evaluation and management. Obstet Gynecol Clin North Am 48(2):401–417. 10.1016/j.ogc.2021.02.009 [DOI] [PubMed] [Google Scholar]
- 7.Patricia K, Enrico L, Femke S, Jeanine K (2022) Long-term follow-up of complicated monochorionic twin pregnancies: focus on neurodevelopment. Best Pract Res Clin Obstet Gynaecol 84(11):166–178. 10.1016/j.bpobgyn.2022.03.014 [DOI] [PubMed] [Google Scholar]
- 8.Zhila N, Kamran S, Negar A, Najmieh S (2024) Fetoscopic laser versus amnioreduction, septostomy, and expected management for the treatment of twin-twin transfusion syndrome (TTTS): an economic evaluation analysis in Iran. Cost Effect Res Alloc 22(1):39. 10.1186/s12962-024-00551-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Alireza AS, Ramen HC, Julien S, Mounira AH, Anthony J, Kamran H, Shayan M, Ahmed N, Roopali D, Magdalena C, Jimmy E, Eyal K, Michael B (2023) Solomon versus selective fetoscopic laser photocoagulation for twin-twin transfusion syndrome: a systematic review and meta-analysis. Prenat Diagn 43(1):72–83. 10.1002/pd.6246 [DOI] [PubMed] [Google Scholar]
- 10.D’Antonio F, Herrera M, Oronzii L, Khalil A (2022) Solomon technique vs selective fetoscopic laser photocoagulation for twin-twin transfusion syndrome: systematic review and meta-analysis of maternal and perinatal outcomes. Ultrasound Obstet Gynecol 60(6):731–738. 10.1002/uog.26095 [DOI] [PubMed] [Google Scholar]
- 11.D’Antonio F, Marinceu D, Eltaweel N, Smriti P, Asma K (2025) Survival rates in pregnancies complicated by twin-to-twin transfusion syndrome undergoing laser therapy: a systematic review and meta-analysis. Am J Obstet Gynecol MFM 7(5):101503. 10.1016/j.ajogmf.2024.101503 [DOI] [PubMed] [Google Scholar]
- 12.Mustafa HJ, Aghajani F, Patrick E, Maryam MB, Arias-Sánchez P, Khalil A (2024) Perinatal outcomes following fetoscopic laser surgery for early twin-to-twin transfusion syndrome: systematic review and meta-analysis. Acta Obstet Gynecol Scand 103(5):824–831. 10.1111/aogs.14806 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Stirnemann J, Djaafri F, Kim A, Mediouni I, Bussieres L, Spaggiari E, Veluppillai C, Lapillonne A, Kermorvant E, Magny JF, Colmant C, Ville Y (2018) Preterm premature rupture of membranes is a collateral effect of improvement in perinatal outcomes following fetoscopic coagulation of chorionic vessels for twin-twin transfusion syndrome: a retrospective observational study of 1092 cases. BJOG 125(9):1154–1162. 10.1111/1471-0528.15147 [DOI] [PubMed] [Google Scholar]
- 14.Braxton F, Mounira H (2020) Unique considerations: preterm prelabor rupture of membranes in the setting of fetal surgery and higher order pregnancies. Obstet Gynecol Clin North Am 47(4):653–669. 10.1016/j.ogc.2020.08.008 [DOI] [PubMed] [Google Scholar]
- 15.Nassr A, Hessami K, Shazly A, Meshinchi N, Corroenne R, Espinoza J, Donepudi R, Sanz Cortes M, Belfort M, Shamshirsaz A (2021) Perinatal outcomes of iatrogenic chorioamniotic separation following fetoscopic surgery: systematic review and meta-analysis. Ultrasound Obstet Gynecol 58(3):347–353. 10.1002/uog.23588 [DOI] [PubMed] [Google Scholar]
- 16.Serjosha R, Raoul H, Costanza G, Lukas M, Alexandra D, Magdalena SC, Martin E, Nicole O, Jan D, Edoardo M (2024) A new ex vivo model system to analyze factors affecting the integrity of fetal membranes in fetoscopic surgery. J Mech Behav Biomed Mater 160(12):106764. 10.1016/j.jmbbm.2024.106764 [DOI] [PubMed] [Google Scholar]
- 17.Deepak K, Robert M, Brian M, Joseph M, John M (2020) Mechanism of human fetal membrane biomechanical weakening, rupture and potential targets for therapeutic intervention. Obstet Gynecol Clin North Am 47(4):523–544. 10.1016/j.ogc.2020.08.010 [DOI] [PubMed] [Google Scholar]
- 18.Julie P, Mancy T, Paulomi A, Ja Y, Mary P, Gil M, Vikki A (2020) Viral infection dampens human fetal membrane type I interferon responses triggered by bacterial LPS. J Reprod Immunol 140(8):103126. 10.1016/j.jri.2020.103126 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Benjamin JA, Ryan JH, Karyn AR, Kelly JC, Stuart BH, Philip D (2021) Why do the fetal membranes rupture early after fetoscopy? A review. Fetal Diagn Ther 48(7):493–503. 10.1159/000517151 [DOI] [PubMed] [Google Scholar]
- 20.Papanna R, Molina S, Moise KY, Moise KJ Jr, Johnson A (2010) Chorioamnion plugging and the risk of preterm premature rupture of membranes after laser surgery in twin-twin transfusion syndrome. Ultrasound Obstet Gynecol 35(3):337–343. 10.1002/uog.7476 [DOI] [PubMed] [Google Scholar]
- 21.César M, Delfina C, Horacio A, Gustavo L, Santiago M, Lucas O (2021) Update on prenatal diagnosis and fetal surgery for myelomeningocele. Arch Argent Pediatr 119(3):e215–e228. 10.5546/aap.2021.eng.e215 [DOI] [PubMed] [Google Scholar]
- 22.Anouk S, Esther S, Marc S, Frank V (2022) Computer-assisted fetal laser surgery in the treatment of twin-to-twin transfusion syndrome: recent trends and prospects. Prenat Diagn 42(10):1225–1234. 10.1002/pd.6225 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Juliana G, Nahla K, Matthew W, Edward O (2024) Monochorionic twin complications and fetoscopic interventions. Magn Reson Imaging Clin N Am 32(3):513–528. 10.1016/j.mric.2024.02.010 [DOI] [PubMed] [Google Scholar]
- 24.Roxana B, Ana C, Vlad D, Lonita D, Corina G, Valentin V, Florentina F (2021) Expectant management of PPROM improves neonatal outcome-a retrospective study of 562 patients. J Clin Med 11(1):214. 10.3390/jcm11010214 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Alberto R, Mackenzie N, Andreina C, Elena L, Andrei R, Nathan F (2022) The association between sonographic cervical length components and preterm birth in women with ultrasound- or exam-indicated cerclage. J Matern-Fetal Neonat Med 35(25):5703–5708. 10.1080/14767058.2021.1892061 [DOI] [PubMed] [Google Scholar]
- 26.Claudio S, Faezeh A, Ali J, Eyal K, Yinka O, Ramesha P, Ramen C, Alireza S (2025) Vasa previa: prenatal diagnosis and the rationale behind using a 5 cm distance from internal Os. J Clin Med 14(3):1009. 10.3390/jcm14031009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Enkhtuya R, Rheanna UG, Nathan E, Mariana S, Richard P, Arum H, Ramkumar M (2022) Generation and characterization of human fetal membrane and decidual cell lines for reproductive biology experiments. Biol Reprod 106(3):568–582. 10.1093/biolre/ioab231 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Sungjin K, Po L, Lauren R, Ramkumar M, Arum H (2024) A dynamic flow fetal membrane organ-on-a-chip system for modeling the effects of amniotic fluid motion. Biomed Microdevice 26(3):32. 10.1007/s10544-024-00714-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Mascio D, Khalil A, D’Amico A, Buca D, Benedetti P, Flacco E, Manzoli L, Liberati M, Nappi L, Berghella V, D’Antonio F (2020) Outcome of twin-twin transfusion syndrome according to Quintero stage of disease: systematic review and meta-analysis. Ultrasound Obstet Gynecol 56(6):811–820. 10.1002/uog.22054 [DOI] [PubMed] [Google Scholar]
- 30.Florian S, Andreas M, Michael S, Sven K, Frederik S, Sophia A, Adriana T, Constanza P, ChristianB MB, Florian F (2023) Cerebroplacental versus umbilicocerebral ratio-analyzing the predictive value regarding adverse perinatal outcomes in low- and high-risk fetuses at term. Medicina 59(8):1385. 10.3390/medicina59081385 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Eva AR, Frauke S, Yannick D, Josep R, Nele S, Miriam W, Phillip M, Ladina V, Ueli M, Martin E, Nicole OK (2024) In vivo sealing of fetoscopy-induced fetal membrane defects by mussel glue. Fetal Diagn Ther 49(11–12):518–527. 10.1159/000528473 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
All data generated or analyzed during this study are included in this published article.