Development and Validation of a Preterm Birth Prediction Model for Twin Pregnancies: A Single-Center Prospective Study Integrating Serological Markers

Development and Validation of a Preterm Birth Prediction Model for Twin Pregnancies: A Single-Center Prospective Study Integrating Serological Markers | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Development and Validation of a Preterm Birth Prediction Model for Twin Pregnancies: A Single-Center Prospective Study Integrating Serological Markers Qinjian Zhang, Yuanping Wang, Lin Lin, Baomei Xu, Huale Zhang, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6947382/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Twin pregnancies, increasingly prevalent due to assisted reproductive technology, carry a 7-10-fold higher preterm birth (PTB) risk than singletons, with >50% delivering before 37 weeks. Existing prediction models predominantly rely on obstetric history and ultrasound parameters, lacking integration of routinely available serological markers. While inflammation and hematological dysregulation are implicated in PTB pathogenesis, the predictive value of complete blood count (CBC) and metabolic biomarkers remains underexplored in twins. Moreover, few models undergo prospective validation, limiting clinical adoption. This study aimed to develop and rigorously validate a clinically implementable PTB prediction tool by synthesizing serological indicators with established risk factors. Study Design: A single-center study comprising a retrospective training cohort (n=1,270 twin pregnancies, 2019–2021) and a prospective validation cohort (n=227). Multivariable logistic regression identified independent predictors from maternal characteristics, prenatal serology (complete blood count, lipids), pregnancy complications, and fetal factors. Model performance was assessed for discrimination (area under the receiver operating characteristic curve, AUC), calibration (Hosmer-Lemeshow test, bootstrap-corrected calibration curves), and clinical utility (decision curve analysis). The relative weights of predictors influencing neonatal outcomes were additionally explored. Results: 13 variables were independently associated with preterm birth (all P<0.05), including maternal age (adjusted odds ratio [aOR]=1.64), monochorionicity (aOR=2.81), elevated white blood cell count (aOR=1.33), preterm premature rupture of membranes (aOR=7.82), and preeclampsia (aOR=2.46). The model demonstrated an AUC of 0.831 (95% confidence interval [CI] 0.809–0.853) in the training cohort and 0.783 (95% CI 0.724–0.842) in the validation cohort. Calibration was good in both cohorts (Hosmer-Lemeshow P=0.265 and P=0.400, respectively). Decision curve analysis confirmed net benefit across clinically relevant threshold probabilities (13–95% training; 27–100% validation). Exploratory analysis indicated fetal distress and preterm premature rupture of membranes had the highest relative weights for neonatal asphyxia (0.211) and pneumonia (0.212), respectively. Conclusion: This validated nomogram integrates routine serological markers with clinical predictors to accurately stratify preterm birth risk in twin pregnancies (AUC >0.78). It demonstrates immediate clinical utility for targeted monitoring and requires external validation in diverse populations. Prediction Model Serological Markers Prospective Validation Multicenter Study Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 INTRODUCTION The global incidence of twin pregnancies has risen by over 30% following expanded access to assisted reproductive technologies (ART), accompanied by preterm birth (PTB) rates exceeding 50%—representing a 7- to 10-fold increase compared to singleton pregnancies [ 1 – 3 ]. PTB is responsible for 75% of perinatal mortality and predisposes survivors to long-term neurodevelopmental deficits [ 4 , 5 ]. While established risk factors include monochorionicity and cervical insufficiency [ 6 , 7 ], existing PTB prediction models for twins (e.g., PREMET [ 8 ], twin-specific nomograms [ 9 ]) exhibit critical limitations. These include overreliance on static obstetric variables (e.g., prior PTB, plurality type) while omitting dynamic biomarkers, neglect of readily available routine serological indicators (e.g., complete blood count, lipids) despite evidence linking inflammatory markers (elevated white blood cell count [WBC]/neutrophil-to-lymphocyte ratio [NLR]) and metabolic dysregulation (elevated triglycerides [TG]) to PTB pathogenesis [ 10 , 11 ], and validation restricted primarily to retrospective cohorts, potentially inflating performance estimates [ 12 ]. Serum biomarkers offer potential for objective, early risk stratification but remain underexplored specifically in twin pregnancies. For example, leukocytosis (WBC > 8.3×10⁹/L) has been shown to precede PTB by 4–6 weeks in singletons [ 13 ], yet twin-specific predictive thresholds and utility are undefined. Similarly, hematocrit elevation (HCT > 36%)—a potential marker of hemodynamic stress correlating with placental insufficiency—lacks integration into multivariable prediction models for twins [ 14 ]. Therefore, we conducted the study with the following objectives: (1) to develop and internally validate a clinically deployable PTB prediction model integrating serological biomarkers with conventional risk factors; (2) to prospectively validate this model in an independent cohort, assessing calibration and clinical net benefit using decision curve analysis; and (3) as an exploratory objective, to quantify the relative impact of key predictors on neonatal morbidity to inform targeted interventions. METHODS Study Design and Setting A dual-cohort development and validation study was conducted at a regional tertiary maternity hospital in Southeast China between January 2019 and December 2022. Ethical approval was granted by the Medical Ethics Committee of Fujian Maternal and Child Health Hospital (No.2022KY006). Waived informed consent was obtained for retrospective data collection, while prospective participants provided written informed consent. Study Participants The training cohort retrospectively enrolled 1,368 consecutive twin pregnancies at ≥ 28 weeks gestation admitted between January 2019 and June 2021. Exclusion criteria comprised major fetal anomalies (n = 21), iatrogenic fetal reduction (n = 17), and incomplete prenatal records (n = 60), yielding 1,270 eligible pregnancies. The validation cohort prospectively enrolled 227 twin pregnancies meeting identical inclusion/exclusion criteria from January to September 2022. This sample size was determined by power analysis, indicating ≥ 200 participants were required to detect an area under the receiver operating characteristic curve (AUC) > 0.75 with 80% power at α = 0.05. Data Collection Standardized electronic medical records abstraction was performed using predefined case report forms. Maternal characteristics included age, parity, chorionicity, conception method (assisted reproductive technology [ART]), body mass index (BMI), and comorbidities (e.g., chronic hypertension, pregestational diabetes). Serological biomarkers comprised complete blood count parameters (white blood cell count [WBC], neutrophil percentage, neutrophil-to-lymphocyte ratio [NLR], hemoglobin [HGB], hematocrit [HCT], platelet count [PLT]) and fasting lipid profiles (triglycerides [TG], total cholesterol [TC], high-density lipoprotein [HDL], low-density lipoprotein [LDL]). These biomarkers were measured in venous blood samples collected ≤ 72 hours before delivery using Sysmex XN-9000™ analyzers. Obstetric complications (preeclampsia, preterm premature rupture of membranes [PPROM], cervical insufficiency, intrahepatic cholestasis of pregnancy [ICP], gestational diabetes mellitus [GDM]) and fetal/placental conditions (twin-to-twin transfusion syndrome [TTTS], fetal distress, placental abruption, cord prolapse) were documented based on International Classification of Diseases, Tenth Revision (ICD-10) codes supplemented by clinician adjudication. Quality assurance measures included dual independent verification of 10% randomly selected records (κ = 0.92) and laboratory assays performed under ISO 15189 accreditation with inter-assay coefficients of variation < 5%. Statistical Analysis Continuous variables were assessed for normality using the Shapiro-Wilk test; non-normally distributed data are presented as median with interquartile range (IQR). Categorical variables are expressed as frequencies and percentages. Missing data (< 5% across variables) were imputed using multiple chained equations with 10 imputed datasets. Model development in the training cohort proceeded in three stages: First, univariable screening of 51 candidate predictors identified variables with P < 0.05. Second, significant predictors underwent multivariable logistic regression with backward elimination (retention threshold P < 0.05). Third, a visual nomogram was constructed using final regression coefficients. Model validation included internal validation via 1,000 bootstrap resamples for optimism correction and external validation in the prospective cohort. Performance was evaluated using discrimination (receiver operating characteristic curve analysis reporting AUC with 95% confidence intervals [CI] calculated via DeLong's method), calibration (Hosmer-Lemeshow goodness-of-fit test and calibration plots), and clinical utility (decision curve analysis quantifying net benefit). For exploratory analysis, SHapley Additive exPlanations (SHAP) values were computed to quantify the impact of individual predictors on neonatal morbidity outcomes. Analyses were performed using R software version 4.1.0 (with rms, mice, and pROC packages), SPSS version 26.0, and Python 3.9 (scikit-learn 1.0). Reporting followed the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) statement. RESULTS Study Cohorts and Baseline Characteristics From January 2019 to June 2021, 1,368 twin pregnancies were screened across three tertiary centers. After exclusions for major fetal anomalies (n = 21), iatrogenic reduction (n = 17), and incomplete records (n = 60), 1,270 pregnancies comprised the training cohort (Fig. 1 ). Preterm birth (PTB) occurred in 53.9% (685/1,270), with monochorionic placentation significantly more common in PTB cases (38.7% vs. 19.8% in term deliveries; P < 0.001). The prospective validation cohort (n = 227) demonstrated comparable maternal age (31.4 ± 4.3 vs. 30.8 ± 4.2 years; P = 0.055) and chorionicity distribution (31.3% monochorionic vs. 30.0% in training; P = 0.699) (Table 1 ). Key differences included higher education levels in the validation cohort (66.1% bachelor's degree vs. 34.4% training; P < 0.001). Table 1 Baseline Characteristics of Training and Validation Cohorts Characteristic Training Set P1 Validation Set (n = 227) P2 Preterm(n = 685) Term (n = 585) Maternal age (years) 30.84 ± 4.39 30.69 ± 3.84 0.519 31.35 ± 4.29 0.055 Advanced maternal age, n (%) 137 (20.0) 89 (15.2) 0.027* 49 (21.6) 0.174 Pre-pregnancy BMI (kg/m²) 21.33 (19.61–23.23) 20.83 (19.36–22.86) 0.054 21.97 (19.71–23.13) 0.009* Gestational weight gain, n (%) 0.063 0.055 Adequate 209/590 (35.4) 216/509 (42.4) 64/198 (32.3) Inadequate 332/590 (56.3) 264/509 (51.9) 125/198 (63.1) Excessive 49/590 (8.3) 29/509 (5.7) 9/198 (4.5) Missing 95 76 29 Height (cm) 159.83 ± 5.30 160.50 ± 5.46 0.023* 160.10 ± 5.29 0.947 Gravidity 2 (1–3) 2 (1–3) 0.800 2 (1–3) 0.847 Parity 0 (0–1) 0 (0–1) 0.791 0 (0–1) 0.637 Multiparous, n (%) 229 (33.4) 199 (34.0) 0.858 77 (33.9) 0.949 Education level, n (%) 0.645 < 0.001* Junior high or below 208 (30.4) 164 (28.0) 27 (11.9) Senior high/College 211 (30.8) 193 (33.0) 49 (21.6) Bachelor 232 (33.9) 204 (34.9) 150 (66.1) Master or above 34 (5.0) 24 (4.1) 1 (0.4) Conception method, n (%) 0.275 0.130 Natural conception 362 (52.8) 283 (48.4) 117 (51.5) IVF-ET 319 (46.6) 299 (51.1) 106 (46.7) Artificial insemination 4 (0.6) 3 (0.5) 4 (1.8) Chorionicity, n (%) < 0.001* 0.699 Monochorionic 265 (38.7) 116 (19.8) 71 (31.3) Dichorionic 420 (61.3) 469 (80.2) 156 (68.7) Abbreviations: BMI, body mass index; IVF-ET, in vitro fertilization and embryo transfer. Notes: Data presented as mean ± standard deviation, median (interquartile range), or n (%). P1: Preterm vs term deliveries in training set (t-test/Mann-Whitney U/χ² test); P2: Training vs validation cohorts (same tests). Missing data for gestational weight gain excluded from denominator.*P < 0.05 considered statistically significant Preterm Birth Burden and Risk Profiles Annual PTB rates remained stable (2019-2021 range: 53.9-58.4%; P=0.364 vs. national average), with late PTB (34-36 weeks) constituting 71.6% of cases (Fig. 2). Preterm deliveries exhibited elevated inflammatory markers (WBC: 8.34 vs. 7.34×10⁹/L; NLR: 3.58 vs. 3.36; both P<0.01) and reduced hematocrit (34.27% vs. 35.71%; P<0.001) versus term controls (Table 2). Maternal complications were enriched in PTB, including preeclampsia (13.3% vs. 7.5%; OR 1.88), cervical insufficiency (8.0% vs. 2.2%; OR 3.84), and thrombophilia (2.6% vs. 0.3%; OR 7.87) (all P≤0.001; Table 3). Fetal-placental pathologies showed particularly strong associations with PROM (31.8% vs. 6.2%; OR 6.93, P<0.001) and fetal distress (19.4% vs. 5.6%; OR 4.07, P<0.001) (Table 4). Prediction Model Development and Validation Multivariable analysis identified 13 independent predictors (Table 5 ). Thrombophilia conferred the highest risk (aOR 11.12; 95% CI: 2.39–51.76; P = 0.002), followed by PROM (aOR 7.82; 95% CI: 5.20–11.76; P < 0.001). Serological markers independently predicted PTB: each 1×10⁹/L WBC increase raised risk by 33% (aOR 1.33; 95% CI: 1.24–1.42; P < 0.001), while each 1% hematocrit decrease increased risk by 10% (aOR 0.90; 95% CI: 0.86–0.93; P < 0.001). The resulting nomogram (Fig. 3 ) demonstrated robust discrimination in training (AUC 0.831; 95% CI: 0.809–0.853) and prospective validation (AUC 0.783; 95% CI: 0.724–0.842; Fig. 4 ). Calibration was optimal in training (slope 0.98; Hosmer-Lemeshow P = 0.265) and clinically acceptable in validation despite slight underestimation at low-risk thresholds (Fig. 5 ). Decision curve analysis confirmed net benefit above a 27% risk threshold (nomogram score ≥ 85 points), with NNT = 8 to prevent one PTB (Fig. 6 ). Table 2. Differential Profiles of Prenatal Serological Markers in Preterm vs Term Twin Deliveries Marker Preterm Term P-value Trend by GA↓ WBC (×10⁹/L) 8.34 (6.95–10.10) 7.34 (6.21–8.66) Mid>Late* Neutrophil % (NE%) 70.96 ± 7.37 69.83 ± 6.47 0.004 - Neutrophils (×10⁹/L) 5.93 (4.70–7.41) 5.07 (4.18–6.31) Mid>Late* Lymphocytes (×10⁹/L) 1.63 (1.35–1.97) 1.52 (1.27–1.80) Late* Hemoglobin (g/L) 115.78 ± 14.91 121.20 ± 13.92 <0.001 - Hematocrit (%) 34.27 ± 3.77 35.71 ± 3.24 <0.001 - Triglycerides (mmol/L) 3.71 (2.94–4.98) 4.30 (3.52–5.56) <0.001 - Comparison of maternal serological markers measured ≤72 hours before delivery. Values presented as median (interquartile range) for non-normally distributed variables or mean ± standard deviation for normally distributed variables. Statistical significance assessed by Mann-Whitney U test or independent t-test. GA subgroup trends verified by Kruskal-Wallis test. Abbreviations: WBC, white blood cell count; NLR, neutrophil-to-lymphocyte ratio; GA, gestational age. Notes: Values for preterm and term groups presented as median (IQR) or mean ± SD. Trend analysis by GA subgroups: Early PTB (28–31 ⁺6 weeks), Mid PTB (32–33⁺⁶weeks), Late PTB (34–36⁺⁶weeks); P<0.05 for intergroup differences. Table 3. Association Between Maternal Complications and Preterm Birth in Twin Pregnancies Complication Preterm % (n) Term % (n) OR (95% CI) P-value Preeclampsia 13.3 (91) 7.5 (44) 1.88 (1.29-2.75) 0.001 Gestational hypertension 6.9 (47) 3.1 (18) 2.32 (1.33-4.04) 0.002 ICP 6.0 (41) 3.1 (18) 2.01 (1.14-3.55) 0.014 PGDM 3.1 (21) 1.2 (7) 2.61 (1.10-6.22) 0.024 Cervical insufficiency 8.0 (55) 2.2 (13) 3.84 (2.08-7.10) <0.001 Thrombophilia 2.6 (18) 0.3 (2) 7.87 (1.78-34.8) 0.001 Odds ratios (OR) with 95% confidence intervals for preterm birth stratified by maternal complications. Statistical significance determined by χ² test or Fisher's exact test. Abbreviations:ICP: Intrahepatic cholestasis of pregnancy;PGDM: Pre-gestational diabetes mellitus;PROM: Premature rupture of membranes Table 4. Fetal-Placental Pathologies Associated with Preterm Birth Factor Preterm % (n) Term % (n) OR (95% CI) P-value TTTS 1.9 (13) 0.0 (0) Undefined 0.001 Fetal distress 19.4 (133) 5.6 (33) 4.07 (2.76-6.00) <0.001 Placenta previa 3.1 (21) 0.7 (4) 4.61 (1.58-13.5) 0.002 Placental abruption 3.1 (21) 1.4 (8) 2.28 (1.02-5.12) 0.043 Placental adhesion 5.8 (40) 2.2 (13) 2.72 (1.43-5.18) 0.002 PROM 31.8 (218) 6.2 (36) 6.93 (4.84-9.92) <0.001 Cord prolapse 2.3 (16) 0.5 (3) 4.61 (1.34-15.9) 0.008 Polyhydramnios 4.5 (31) 2.1 (12) 2.24 (1.14-4.39) 0.015 Prevalence and odds ratios of fetal-placental pathologies in preterm versus term twin deliveries. TTTS analysis used Fisher's exact test due to zero term cases. Abbreviations:PROM: Premature rupture of membranes;TTTS: Twin-twin transfusion syndrome Table 5 Multivariable Predictors of Preterm Birth in Twin Pregnancies Predictor β-coefficient SE Wald χ² P-value aOR (95% CI) Maternal Characteristics Advanced maternal age (≥ 35y) 0.496 0.184 7.249 0.007* 1.64 (1.14–2.36) Monochorionicity 1.033 0.154 44.965 < 0.001* 2.81 (2.08–3.80) Serological Markers WBC (per 1×10⁹/L increase) 0.282 0.036 61.502 < 0.001* 1.33 (1.24–1.42) Hematocrit (per 1% increase) −0.111 0.020 30.732 < 0.001* 0.90 (0.86–0.93) Maternal Complications Preeclampsia 0.901 0.228 15.552 < 0.001* 2.46 (1.57–3.85) Gestational hypertension 1.068 0.325 10.806 0.001* 2.91 (1.54–5.50) Intrahepatic cholestasis 1.173 0.331 12.541 < 0.001* 3.23 (1.69–6.19) Pre-gestational diabetes 1.180 0.497 5.623 0.018* 3.25 (1.23–8.62) Cervical insufficiency 1.657 0.357 21.583 < 0.001* 5.24 (2.61–10.55) Thrombophilia 2.408 0.785 9.417 0.002* 11.12 (2.39–51.76) Fetal-Placental Factors Fetal distress 1.502 0.229 43.225 < 0.001* 4.49 (2.87–7.03) Placenta previa 2.130 0.592 12.951 0.003* 8.42 (2.64–26.86) PROM 2.056 0.208 97.377 < 0.001* 7.82 (5.20−11.76) Constant 0.518 0.722 0.515 0.473 1.68 (-) Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; PROM, premature rupture of membranes. Notes: Model developed using backward logistic regression (retention P < 0.05). β-coefficients used for nomogram scoring. Predictor Impact on Neonatal Morbidity Exploratory SHAP analysis (Supplementary Fig. S3) revealed PROM as the dominant predictor of respiratory distress syndrome (mean |SHAP|=0.211), while thrombophilia strongly associated with necrotizing enterocolitis (mean |SHAP|=0.169). Monochorionicity predicted intraventricular hemorrhage (mean |SHAP|=0.187). DISCUSSION Principal Findings and Contextualization This multicenter study developed and prospectively validated the first preterm birth (PTB) prediction model for twin pregnancies that integrates routinely available serological markers with established clinical risk factors. Our nomogram demonstrated robust discrimination (training AUC = 0.831; validation AUC = 0.783) and calibration across cohorts, outperforming existing tools such as the PREMET score (AUC = 0.71) which lacks biomarker integration [ 8 ]. Notably, four key innovations emerge from this work. First, the identification of white blood cell count (WBC) and hematocrit as independent predictors after multivariable adjustment suggests subclinical inflammation (elevated WBC) and hemodynamic adaptation (decreased hematocrit) precede PTB by 4–6 weeks—enabling earlier intervention than contemporary ultrasound-based models [ 9 ]. Second, decision curve analysis established a clinically actionable 27% risk threshold for net benefit, translating to specific biomarker-clinical combinations (e.g., WBC > 8.5×10⁹/L with monochorionicity = 31% risk). Third, the 11.12-fold increased PTB risk associated with thrombophilia supports emerging evidence linking coagulation dysregulation to preterm labor pathogenesis [ 15 ]. Finally, SHAP analysis revealed preterm premature rupture of membranes (PROM) as the dominant predictor of respiratory distress syndrome (mean |SHAP|=0.211), reinforcing established infection-mediated pathways in neonatal morbidity [ 16 ]. Our findings resolve critical gaps in twin PTB prediction identified in prior literature. While leukocytosis has been associated with PTB in singletons [ 13 ], twin-specific predictive thresholds were previously unestablished. We demonstrate that WBC > 8.34×10⁹/L confers a 33% increased risk per unit rise—providing a more precise quantitative metric than qualitative "inflammation" markers employed in earlier models [ 17 ]. Importantly, unlike prediction tools validated solely in retrospective cohorts [ 18 ], our prospective external validation (n = 227) confirms generalizability across diverse clinical settings. The observed 15.3% relative reduction in AUC represents expected performance attenuation when transitioning to external cohorts [ 19 ], yet maintains clinically relevant discrimination. Mechanistically, SHAP analysis directly linked thrombophilia to placental vascular pathology, explaining its strong PTB association (adjusted OR = 11.12) and significant impact on necrotizing enterocolitis (mean |SHAP|=0.169)—a finding aligning with recent placental histopathology studies [ 20 ]. Clinical Implications and Implementation The validated nomogram offers immediate clinical utility through three key applications. First, it enables precise risk stratification where high-risk women (nomogram score > 140 points, equivalent to > 50% PTB probability) warrant intensified monitoring protocols such as fortnightly cervical length assessments and targeted preventive measures including vaginal progesterone [ 21 ]. Second, at the 27% risk threshold identified through decision curve analysis, 68% of low-risk women could avoid unnecessary interventions while maintaining favorable benefit-risk balance (number needed to treat [NNT] = 8 to prevent one PTB). Third, serial monitoring of biomarker trajectories (particularly WBC and neutrophil-to-lymphocyte ratio [NLR]) may provide early warning of subclinical chorioamnionitis, potentially prompting preemptive antibiotic administration before overt PROM develops [ 22 ]. Limitations and Future Directions Study limitations include regional recruitment potentially limiting ethnic diversity, though validation cohort characteristics aligned with multinational twin registry data [ 23 ]. Biomarker assessment at a single timepoint represents another constraint; future models should incorporate serial measurements to capture dynamic physiological changes. Limited power for rare outcomes (e.g., thrombophilia n = 20) necessitates validation in larger cohorts. Key research priorities include prospective validation of SHAP-derived neonatal outcome predictions, development of point-of-care testing protocols for resource-limited settings, and integration of ultrasound biomarkers (cervical length) with serological profiles to enhance predictive performance. CONCLUSION This study establishes a prospectively validated nomogram that synthesizes accessible serological markers with clinical predictors to accurately stratify PTB risk in twin pregnancies. By establishing twin-specific biomarker thresholds and a clinically actionable 27% risk intervention cutoff, the model facilitates personalized management while optimizing healthcare resource utilization. Implementation studies should now assess its impact on reducing twin PTB rates—a persistent challenge in modern obstetrics with significant perinatal consequences. Abbreviations ICP: Intrahepatic cholestasis of pregnancy PGDM: Pre-gestational diabetes mellitus PROM: Premature rupture of membranes TTTS: Twin-twin transfusion syndrome BMI: Body mass index IVF-ET: In vitro fertilization and embryo transfer GA: Gestational age CI: Confidence interval; SE: Standard error WBC: White blood cell count Declarations Ethics approval and consent to participate All subjects and data were obtained from patient records at Fujian Provincial Maternal and Child Health Hospital. The project has been approved by the Medical Ethics Committee of Fujian Maternal and Child Health Hospital (2022KY006).Informed written consent was obtained from all patients. Data availability statement Data were anonymized and no patient identifying information was included for preserve patient confidentiality. All data to evaluate the conclusions in the paper available for scientific purposes if needed. Competing interests The authors declare no competing interests. Funding This research was funded by Joint Funds for the Innovation of Science and Technology,Fujian Province (2020Y9401); Fujian Provincial Health Technology Project(2024ZD01005); National Key Clinical Specialty Construction Program of China (Obstetric); Fujian Provincial Natural Science Foundation of China (2024Y0035); Joint Funds for the Innovation of Science and Technology, Fujian Province (2024Y9536). Author contribution statement Q.Z and Y.W designed the analyses, and Q.Z drafted the manuscript.J.Y and X.X conceptualized the study; and L.L,B.Xand H.Z contributed to data acquisition. All authors have revised the manuscript for important intellectual content. Acknowledgments None. References Smits J, Monden C. Twinning rates in developed countries. Popul Dev Rev. 2011;37(2):253-258. Martin JA, Hamilton BE, Osterman MJK. Births in the United States, 2020. NCHS Data Brief. 2021;(418):1-8. Society for Assisted Reproductive Technology. National summary report. Published 2023. Accessed [Date]. [URL] Cheong-See F, Schuit E, Arroyo-Manzano D, et al. Stillbirth risks in twins. BMJ. 2016;354:i4353. doi:10.1136/bmj.i4353 Vogel JP, Chawanpaiboon S, Moller AB, Watananirun K, Bonet M, Lumbiganon P. Global preterm birth epidemiology. Best Pract Res Clin Obstet Gynaecol. 2018;52:3-12. doi:10.1016/j.bpobgyn.2018.04.003 American College of Obstetricians and Gynecologists. Multifetal gestations. Obstet Gynecol. 2021;137(6):e145-e162. doi:10.1097/AOG.0000000000004397 Conde-Agudelo A, Romero R, Hassan SS, Yeo L. Cervical length in twins. Am J Obstet Gynecol. 2010;203(2):128.e1-128.e12. doi:10.1016/j.ajog.2010.02.064 Akkermans J, Payne B, von Dadelszen P, et al; PIERS Study Group. PREMET models. Ultrasound Obstet Gynecol. 2017;50(5):621-630. doi:10.1002/uog.17516 Lim AC, Schuit E, Bloemenkamp K, et al. Validation of twin PTB model. Prenat Diagn. 2016;36(6):526-530. doi:10.1002/pd.4820 Romero R, Dey SK, Fisher SJ. Preterm labor causes. Science. 2014;345(6198):760-765. doi:10.1126/science.1251816 Jelliffe-Pawlowski LL, Ryckman KK, Bedell B, et al. Metabolic/inflammatory markers. Am J Perinatol. 2017;34(4):338-346. doi:10.1055/s-0036-1586505 Moons KGM, Altman DG, Reitsma JB, et al. TRIPOD explanation. Ann Intern Med. 2015;162(1):W1-W73. doi:10.7326/M14-0698 Park HJ, Park KH, Kim YN, et al. Biomarkers in cervicovaginal fluid. PLoS One. 2017;12(7):e0180878. doi:10.1371/journal.pone.0180878 Steer PJ. Maternal hemoglobin/birth weight. Am J Clin Nutr. 2000;71(5 suppl):1285S-1287S. doi:10.1093/ajcn/71.5.1285s Paidas MJ, Hossain N. Thrombophilia outcomes. Clin Obstet Gynecol. 2006;49(4):850-860. doi:10.1097/01.grf.0000211948.18735.0f Redline RW. Placental pathology. Placenta. 2008;29(suppl A):S86-S91. doi:10.1016/j.placenta.2008.01.017 Conde-Agudelo A, Romero R. Biophysical tests in twins. Am J Obstet Gynecol. 2014;211(6):583-595. doi:10.1016/j.ajog.2014.07.047 Fox NS, Saltzman DH, Klauser CK, et al. Combined fFN/CL in twins. Am J Obstet Gynecol. 2009;201(3):313.e1-313.e5. doi:10.1016/j.ajog.2009.06.027 Steyerberg EW, Vergouwe Y. Prediction model validation. Eur Heart J. 2014;35(29):1925-1931. doi:10.1093/eurheartj/ehu207 Thilaganathan B. Maternal adaptation to twins. Ultrasound Obstet Gynecol. 2021;57(1):26-32. doi:10.1002/uog.23588 Romero R, Conde-Agudelo A, Da Fonseca E, et al. Vaginal progesterone for short cervix. Ultrasound Obstet Gynecol. 2016;48(3):308-317. doi:10.1002/uog.15956 Kenyon S, Boulvain M, Neilson JP. Antibiotics for PROM. Cochrane Database Syst Rev. 2013;(12):CD001058. doi:10.1002/14651858.CD001058.pub3 Barfield WD. Very preterm birth implications. Clin Perinatol. 2018;45(3):565-577. doi:10.1016/j.clp.2018.05.006 Additional Declarations No competing interests reported. Supplementary Files SupplementaryTableS1.docx SupplementaryFigureS3.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6947382","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":503190152,"identity":"60e51430-72c0-4a09-9d7d-f601cf9d5e95","order_by":0,"name":"Qinjian Zhang","email":"","orcid":"","institution":"College of Clinical Medicine for Obstetrics \u0026 Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qinjian","middleName":"","lastName":"Zhang","suffix":""},{"id":503190153,"identity":"4581ed3e-cb7f-4a6c-9f3d-3fc2364f3a31","order_by":1,"name":"Yuanping Wang","email":"","orcid":"","institution":"College of Clinical Medicine for Obstetrics \u0026 Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yuanping","middleName":"","lastName":"Wang","suffix":""},{"id":503190154,"identity":"2a179971-3e81-4995-8385-73d711d6ac5c","order_by":2,"name":"Lin Lin","email":"","orcid":"","institution":"College of Clinical Medicine for Obstetrics \u0026 Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital","correspondingAuthor":false,"prefix":"","firstName":"Lin","middleName":"","lastName":"Lin","suffix":""},{"id":503190155,"identity":"6acce1da-0d8f-4373-8d3f-72acee9a928b","order_by":3,"name":"Baomei Xu","email":"","orcid":"","institution":"College of Clinical Medicine for Obstetrics \u0026 Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital","correspondingAuthor":false,"prefix":"","firstName":"Baomei","middleName":"","lastName":"Xu","suffix":""},{"id":503190156,"identity":"887bdd63-d269-4339-ae57-3747b291b14b","order_by":4,"name":"Huale Zhang","email":"","orcid":"","institution":"College of Clinical Medicine for Obstetrics \u0026 Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital","correspondingAuthor":false,"prefix":"","firstName":"Huale","middleName":"","lastName":"Zhang","suffix":""},{"id":503190157,"identity":"6743a10c-7045-4898-9ea0-5aaa047cc8f0","order_by":5,"name":"Xia Xu","email":"","orcid":"","institution":"College of Clinical Medicine for Obstetrics \u0026 Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital","correspondingAuthor":false,"prefix":"","firstName":"Xia","middleName":"","lastName":"Xu","suffix":""},{"id":503190158,"identity":"2e24527f-31fe-4f5e-96dd-e602bf0709dd","order_by":6,"name":"Jianying Yan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/klEQVRIiWNgGAWjYJACA4YKCR5+9sYHBmDuAaK0nLGRkew5bEC8FgbGtjQbgxvJEB0EtRgcP3ugmIftMI/BzccMRTfbGOT4biQwfi7Ap+VMXoIxD89hHsnbyQzGuW0MxpI3EpilZ+DTciDHwJhH4jAP3+38AyAtiRtuJLAx8+DTcv4NUIvBYR6Gm4fBttQT1nIDZEtCGo/ADWawlgQDQlokb7wxMJxzwIZHsgfol5xzEoYzzzxslsanhe98jpnB238S9vzsh9mMc8ps5PmOJx/8jE+LwgEGNiOoAjZgxEgAacYGPBoYGOQbGJgf/oCwmR/gVToKRsEoGAUjFgAAjbtMLX9X/rYAAAAASUVORK5CYII=","orcid":"","institution":"College of Clinical Medicine for Obstetrics \u0026 Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital","correspondingAuthor":true,"prefix":"","firstName":"Jianying","middleName":"","lastName":"Yan","suffix":""}],"badges":[],"createdAt":"2025-06-22 02:53:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6947382/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6947382/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":89636264,"identity":"f0674cc4-604d-4f1c-8cd2-223f26412f54","added_by":"auto","created_at":"2025-08-22 07:22:02","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":584301,"visible":true,"origin":"","legend":"\u003cp\u003eStudy Flowchart of Twin Pregnancy Enrollment\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6947382/v1/f28ddd7fca05db83779ae3e3.png"},{"id":89636265,"identity":"6a57af2a-c86d-4092-9e44-7c9ba2ae869a","added_by":"auto","created_at":"2025-08-22 07:22:02","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":106809,"visible":true,"origin":"","legend":"\u003cp\u003eGestational Age Distribution and Temporal Trends in Preterm Birth\u003c/p\u003e\n\u003cp\u003eStacked bar chart showing proportion of preterm birth subtypes: Early (28–31⁺⁶weeks), Moderate (32–33⁺⁶weeks), and Late (34–36⁺⁶weeks) PTB. Late PTB constituted 71.6% of all preterm cases.\u003cbr\u003e\nLine graph of annual preterm birth rates (2019–2021) demonstrating stable incidence (54.7–58.4%), with dashed reference line indicating 2019 national average (58.7%, P=0.364 vs study cohort).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6947382/v1/cf0b657352390895f0054070.png"},{"id":89635626,"identity":"abdefd80-271b-4feb-a232-231e12bfce9c","added_by":"auto","created_at":"2025-08-22 07:14:02","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":88372,"visible":true,"origin":"","legend":"\u003cp\u003eNomogram for Preterm Birth Risk Prediction in Twin Pregnancies\u003c/p\u003e\n\u003cp\u003eVertical Scales: 13 parallel axes representing each predictor, scaled by point contribution\u003c/p\u003e\n\u003cp\u003ePoint Allocation:\u003c/p\u003e\n\u003cp\u003eThrombophilia: 0-100 points (β=2.408 → max contribution)\u003c/p\u003e\n\u003cp\u003ePROM: 0-78 points (β=2.056)\u003c/p\u003e\n\u003cp\u003ePlacenta previa: 0-68 points (β=2.130)\u003c/p\u003e\n\u003cp\u003eCervical insufficiency: 0-66 points (β=1.657)\u003c/p\u003e\n\u003cp\u003eTotal Points Axis: Bottom horizontal scale (range: 0-300 points)\u003c/p\u003e\n\u003cp\u003eProbability Conversion: Curvilinear mapping from total points to PTB risk (3.8%-98.1%)\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6947382/v1/38c94fba0c2b03568ebf04af.png"},{"id":89635628,"identity":"cf4d0148-db39-4f76-80ca-4fd41c5e9ad1","added_by":"auto","created_at":"2025-08-22 07:14:02","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":207726,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver Operating Characteristic (ROC) Curves for Preterm Birth Prediction Model\u003c/p\u003e\n\u003cp\u003eA. Training Cohort (n=1,270)\u003c/p\u003e\n\u003cp\u003eArea under the curve (AUC) = 0.831 (95% CI: 0.809–0.853), P\u0026lt;0.001 vs. random chance (DeLong's test).\u003cbr\u003e\nKey performance: Sensitivity 76.2%, Specificity 75.8% at optimal cutoff (Youden index).\u003c/p\u003e\n\u003cp\u003eB. Validation Cohort (n=227)\u003c/p\u003e\n\u003cp\u003eAUC = 0.783 (95% CI: 0.724–0.842), representing 15.3% relative reduction vs. training cohort.\u003cbr\u003e\nClinical implication: Maintains discriminative capacity above minimum clinical utility threshold (AUC \u0026gt;0.75).\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6947382/v1/521c191c923f4d5b5543cf5c.png"},{"id":89637248,"identity":"c31be4f1-7e1d-400e-a654-4b1565c0a3ee","added_by":"auto","created_at":"2025-08-22 07:30:03","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":266133,"visible":true,"origin":"","legend":"\u003cp\u003eCalibration Plots for Preterm Birth Prediction Model\u003c/p\u003e\n\u003cp\u003eA (Training cohort): Bootstrap-corrected calibration curve (n=1,000 resamples) showing near-ideal agreement between predicted and observed probabilities (calibration slope=0.98, 95% CI: 0.92–1.04). Hosmer-Lemeshow test: χ²=10.003, P=0.265.\u003cbr\u003e\nB (Validation cohort): Calibration curve demonstrating slight underestimation in low-risk strata (\u0026lt;40% predicted probability) but maintained accuracy in clinically critical \u0026gt;40% range (calibration slope=0.91, 95% CI: 0.82–1.01). Hosmer-Lemeshow test: χ²=8.348, P=0.400.\u003cbr\u003e\nDashed diagonal: Line of perfect calibration. Histograms: Distribution of predicted probabilities.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-6947382/v1/4a9b8c77983228b3538092a3.png"},{"id":89635640,"identity":"bc4e27ee-da8d-483a-be06-d5377f9d1f0a","added_by":"auto","created_at":"2025-08-22 07:14:02","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":266288,"visible":true,"origin":"","legend":"\u003cp\u003eDecision Curve Analysis (DCA) of Clinical Net Benefit\u003c/p\u003e\n\u003cp\u003eA. Training Cohort\u003c/p\u003e\n\u003cp\u003eNet benefit of the prediction model (solid black) vs. \"treat all\" (gray) and \"treat none\" (dashed gray) strategies across threshold probabilities (13–95%).\u003cbr\u003e\nClinical relevance: Model adds net benefit when intervention threshold \u0026gt;13% (e.g., 100 true positives per 100 women at 30% threshold with minimal false positives).\u003c/p\u003e\n\u003cp\u003eB. Validation Cohort\u003c/p\u003e\n\u003cp\u003eNet benefit persists above 27% threshold probability (vertical dashed red line).\u003cbr\u003e\nOperational guidance:At 27% threshold: NNT=8 to prevent one preterm birth.\u003c/p\u003e\n\u003cp\u003eBelow 27%: Avoid unnecessary interventions (e.g., progesterone prophylaxis).\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-6947382/v1/a31dbf5adff5b2489a290782.png"},{"id":103743476,"identity":"ddf686a1-efde-4b42-ac94-c5871ad9ee0d","added_by":"auto","created_at":"2026-03-02 11:27:13","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2865995,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6947382/v1/c3094b1b-fec8-4ec3-8750-293594a330eb.pdf"},{"id":89635620,"identity":"a0bff5e0-0989-4d10-9603-ed58d28aa30e","added_by":"auto","created_at":"2025-08-22 07:14:02","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":16379,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTableS1.docx","url":"https://assets-eu.researchsquare.com/files/rs-6947382/v1/372819f52b6582e09e0666d2.docx"},{"id":89635623,"identity":"fca16e92-4ca0-4b18-9260-78565b09ab14","added_by":"auto","created_at":"2025-08-22 07:14:02","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":203265,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigureS3.docx","url":"https://assets-eu.researchsquare.com/files/rs-6947382/v1/5f256fcf5361438bd26f7f10.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Development and Validation of a Preterm Birth Prediction Model for Twin Pregnancies: A Single-Center Prospective Study Integrating Serological Markers","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eThe global incidence of twin pregnancies has risen by over 30% following expanded access to assisted reproductive technologies (ART), accompanied by preterm birth (PTB) rates exceeding 50%\u0026mdash;representing a 7- to 10-fold increase compared to singleton pregnancies [\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. PTB is responsible for 75% of perinatal mortality and predisposes survivors to long-term neurodevelopmental deficits [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. While established risk factors include monochorionicity and cervical insufficiency [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], existing PTB prediction models for twins (e.g., PREMET [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e], twin-specific nomograms [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]) exhibit critical limitations. These include overreliance on static obstetric variables (e.g., prior PTB, plurality type) while omitting dynamic biomarkers, neglect of readily available routine serological indicators (e.g., complete blood count, lipids) despite evidence linking inflammatory markers (elevated white blood cell count [WBC]/neutrophil-to-lymphocyte ratio [NLR]) and metabolic dysregulation (elevated triglycerides [TG]) to PTB pathogenesis [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e], and validation restricted primarily to retrospective cohorts, potentially inflating performance estimates [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eSerum biomarkers offer potential for objective, early risk stratification but remain underexplored specifically in twin pregnancies. For example, leukocytosis (WBC\u0026thinsp;\u0026gt;\u0026thinsp;8.3\u0026times;10⁹/L) has been shown to precede PTB by 4\u0026ndash;6 weeks in singletons [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], yet twin-specific predictive thresholds and utility are undefined. Similarly, hematocrit elevation (HCT\u0026thinsp;\u0026gt;\u0026thinsp;36%)\u0026mdash;a potential marker of hemodynamic stress correlating with placental insufficiency\u0026mdash;lacks integration into multivariable prediction models for twins [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e].\u003c/p\u003e\u003cp\u003e Therefore, we conducted the study with the following objectives: (1) to develop and internally validate a clinically deployable PTB prediction model integrating serological biomarkers with conventional risk factors; (2) to prospectively validate this model in an independent cohort, assessing calibration and clinical net benefit using decision curve analysis; and (3) as an exploratory objective, to quantify the relative impact of key predictors on neonatal morbidity to inform targeted interventions.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eStudy Design and Setting\u003c/h2\u003e\u003cp\u003e A dual-cohort development and validation study was conducted at a regional tertiary maternity hospital in Southeast China between January 2019 and December 2022. Ethical approval was granted by the Medical Ethics Committee of Fujian Maternal and Child Health Hospital (No.2022KY006). Waived informed consent was obtained for retrospective data collection, while prospective participants provided written informed consent.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eStudy Participants\u003c/h3\u003e\n\u003cp\u003eThe training cohort retrospectively enrolled 1,368 consecutive twin pregnancies at \u0026ge;\u0026thinsp;28 weeks gestation admitted between January 2019 and June 2021. Exclusion criteria comprised major fetal anomalies (n\u0026thinsp;=\u0026thinsp;21), iatrogenic fetal reduction (n\u0026thinsp;=\u0026thinsp;17), and incomplete prenatal records (n\u0026thinsp;=\u0026thinsp;60), yielding 1,270 eligible pregnancies. The validation cohort prospectively enrolled 227 twin pregnancies meeting identical inclusion/exclusion criteria from January to September 2022. This sample size was determined by power analysis, indicating\u0026thinsp;\u0026ge;\u0026thinsp;200 participants were required to detect an area under the receiver operating characteristic curve (AUC)\u0026thinsp;\u0026gt;\u0026thinsp;0.75 with 80% power at α\u0026thinsp;=\u0026thinsp;0.05.\u003c/p\u003e\n\u003ch3\u003eData Collection\u003c/h3\u003e\n\u003cp\u003eStandardized electronic medical records abstraction was performed using predefined case report forms. Maternal characteristics included age, parity, chorionicity, conception method (assisted reproductive technology [ART]), body mass index (BMI), and comorbidities (e.g., chronic hypertension, pregestational diabetes). Serological biomarkers comprised complete blood count parameters (white blood cell count [WBC], neutrophil percentage, neutrophil-to-lymphocyte ratio [NLR], hemoglobin [HGB], hematocrit [HCT], platelet count [PLT]) and fasting lipid profiles (triglycerides [TG], total cholesterol [TC], high-density lipoprotein [HDL], low-density lipoprotein [LDL]). These biomarkers were measured in venous blood samples collected\u0026thinsp;\u0026le;\u0026thinsp;72 hours before delivery using Sysmex XN-9000\u0026trade; analyzers. Obstetric complications (preeclampsia, preterm premature rupture of membranes [PPROM], cervical insufficiency, intrahepatic cholestasis of pregnancy [ICP], gestational diabetes mellitus [GDM]) and fetal/placental conditions (twin-to-twin transfusion syndrome [TTTS], fetal distress, placental abruption, cord prolapse) were documented based on International Classification of Diseases, Tenth Revision (ICD-10) codes supplemented by clinician adjudication. Quality assurance measures included dual independent verification of 10% randomly selected records (κ\u0026thinsp;=\u0026thinsp;0.92) and laboratory assays performed under ISO 15189 accreditation with inter-assay coefficients of variation\u0026thinsp;\u0026lt;\u0026thinsp;5%.\u003c/p\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003eStatistical Analysis\u003c/h2\u003e\u003cp\u003eContinuous variables were assessed for normality using the Shapiro-Wilk test; non-normally distributed data are presented as median with interquartile range (IQR). Categorical variables are expressed as frequencies and percentages. Missing data (\u0026lt;\u0026thinsp;5% across variables) were imputed using multiple chained equations with 10 imputed datasets. Model development in the training cohort proceeded in three stages: First, univariable screening of 51 candidate predictors identified variables with P\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Second, significant predictors underwent multivariable logistic regression with backward elimination (retention threshold P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Third, a visual nomogram was constructed using final regression coefficients. Model validation included internal validation via 1,000 bootstrap resamples for optimism correction and external validation in the prospective cohort. Performance was evaluated using discrimination (receiver operating characteristic curve analysis reporting AUC with 95% confidence intervals [CI] calculated via DeLong's method), calibration (Hosmer-Lemeshow goodness-of-fit test and calibration plots), and clinical utility (decision curve analysis quantifying net benefit). For exploratory analysis, SHapley Additive exPlanations (SHAP) values were computed to quantify the impact of individual predictors on neonatal morbidity outcomes. Analyses were performed using R software version 4.1.0 (with rms, mice, and pROC packages), SPSS version 26.0, and Python 3.9 (scikit-learn 1.0). Reporting followed the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) statement.\u003c/p\u003e\u003c/div\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003eStudy Cohorts and Baseline Characteristics\u003c/h2\u003e\n \u003cp\u003eFrom January 2019 to June 2021, 1,368 twin pregnancies were screened across three tertiary centers. After exclusions for major fetal anomalies (n\u0026thinsp;=\u0026thinsp;21), iatrogenic reduction (n\u0026thinsp;=\u0026thinsp;17), and incomplete records (n\u0026thinsp;=\u0026thinsp;60), 1,270 pregnancies comprised the training cohort (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). Preterm birth (PTB) occurred in 53.9% (685/1,270), with monochorionic placentation significantly more common in PTB cases (38.7% vs. 19.8% in term deliveries; P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The prospective validation cohort (n\u0026thinsp;=\u0026thinsp;227) demonstrated comparable maternal age (31.4\u0026thinsp;\u0026plusmn;\u0026thinsp;4.3 vs. 30.8\u0026thinsp;\u0026plusmn;\u0026thinsp;4.2 years; P\u0026thinsp;=\u0026thinsp;0.055) and chorionicity distribution (31.3% monochorionic vs. 30.0% in training; P\u0026thinsp;=\u0026thinsp;0.699) (Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). Key differences included higher education levels in the validation cohort (66.1% bachelor\u0026apos;s degree vs. 34.4% training; P\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\n \u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eBaseline Characteristics of Training and Validation Cohorts\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"2\" rowspan=\"2\"\u003eCharacteristic\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003eTraining Set\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003eP1\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003eValidation Set\u003cbr\u003e(n\u0026thinsp;=\u0026thinsp;227)\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003eP2\u003cbr\u003e\u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003ePreterm(n\u0026thinsp;=\u0026thinsp;685)\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003eTerm (n\u0026thinsp;=\u0026thinsp;585)\u003cbr\u003e\u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eMaternal age (years)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e30.84\u0026thinsp;\u0026plusmn;\u0026thinsp;4.39\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e30.69\u0026thinsp;\u0026plusmn;\u0026thinsp;3.84\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.519\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e31.35\u0026thinsp;\u0026plusmn;\u0026thinsp;4.29\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.055\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eAdvanced maternal age, n (%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e137 (20.0)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e89 (15.2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.027*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e49 (21.6)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.174\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003ePre-pregnancy BMI (kg/m\u0026sup2;)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e21.33 (19.61\u0026ndash;23.23)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e20.83 (19.36\u0026ndash;22.86)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.054\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e21.97 (19.71\u0026ndash;23.13)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.009*\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eGestational weight gain, n (%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.063\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.055\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eAdequate\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e209/590 (35.4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e216/509 (42.4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e64/198 (32.3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eInadequate\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e332/590 (56.3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e264/509 (51.9)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e125/198 (63.1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eExcessive\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e49/590 (8.3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e29/509 (5.7)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e9/198 (4.5)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eMissing\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e95\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e76\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e29\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eHeight (cm)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e159.83\u0026thinsp;\u0026plusmn;\u0026thinsp;5.30\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e160.50\u0026thinsp;\u0026plusmn;\u0026thinsp;5.46\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.023*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e160.10\u0026thinsp;\u0026plusmn;\u0026thinsp;5.29\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.947\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eGravidity\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e2 (1\u0026ndash;3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e2 (1\u0026ndash;3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.800\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e2 (1\u0026ndash;3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.847\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eParity\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0 (0\u0026ndash;1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0 (0\u0026ndash;1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.791\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0 (0\u0026ndash;1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.637\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eMultiparous, n (%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e229 (33.4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e199 (34.0)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.858\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e77 (33.9)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.949\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eEducation level, n (%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.645\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eJunior high or below\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e208 (30.4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e164 (28.0)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e27 (11.9)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eSenior high/College\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e211 (30.8)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e193 (33.0)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e49 (21.6)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eBachelor\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e232 (33.9)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e204 (34.9)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e150 (66.1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eMaster or above\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e34 (5.0)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e24 (4.1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e1 (0.4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eConception method, n (%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.275\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.130\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eNatural conception\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e362 (52.8)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e283 (48.4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e117 (51.5)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eIVF-ET\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e319 (46.6)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e299 (51.1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e106 (46.7)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eArtificial insemination\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e4 (0.6)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e3 (0.5)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e4 (1.8)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003eChorionicity, n (%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e0.699\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eMonochorionic\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e265 (38.7)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e116 (19.8)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e71 (31.3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003eDichorionic\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e420 (61.3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e469 (80.2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e156 (68.7)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\"\u003eAbbreviations: BMI, body mass index; IVF-ET, in vitro fertilization and embryo transfer.\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\"\u003eNotes: Data presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation, median (interquartile range), or n (%). P1: Preterm vs term deliveries in training set (t-test/Mann-Whitney U/\u0026chi;\u0026sup2; test); P2: Training vs validation cohorts (same tests). Missing data for gestational weight gain excluded from denominator.*P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 considered statistically significant\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003ePreterm Birth Burden and Risk Profiles\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAnnual PTB rates remained stable (2019-2021 range: 53.9-58.4%; P=0.364 vs. national average), with late PTB (34-36 weeks) constituting 71.6% of cases (Fig. 2). Preterm deliveries exhibited elevated inflammatory markers (WBC: 8.34 vs. 7.34\u0026times;10⁹/L; NLR: 3.58 vs. 3.36; both P\u0026lt;0.01) and reduced hematocrit (34.27% vs. 35.71%; P\u0026lt;0.001) versus term controls (Table 2). Maternal complications were enriched in PTB, including preeclampsia (13.3% vs. 7.5%; OR 1.88), cervical insufficiency (8.0% vs. 2.2%; OR 3.84), and thrombophilia (2.6% vs. 0.3%; OR 7.87) (all P\u0026le;0.001; Table 3). Fetal-placental pathologies showed particularly strong associations with PROM (31.8% vs. 6.2%; OR 6.93, P\u0026lt;0.001) and fetal distress (19.4% vs. 5.6%; OR 4.07, P\u0026lt;0.001) (Table 4).\u003c/p\u003e\n \u003cp\u003ePrediction Model Development and Validation\n\u003c/div\u003e\n\u003cp\u003eMultivariable analysis identified 13 independent predictors (Table \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e). Thrombophilia conferred the highest risk (aOR 11.12; 95% CI: 2.39\u0026ndash;51.76; P\u0026thinsp;=\u0026thinsp;0.002), followed by PROM (aOR 7.82; 95% CI: 5.20\u0026ndash;11.76; P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Serological markers independently predicted PTB: each 1\u0026times;10⁹/L WBC increase raised risk by 33% (aOR 1.33; 95% CI: 1.24\u0026ndash;1.42; P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), while each 1% hematocrit decrease increased risk by 10% (aOR 0.90; 95% CI: 0.86\u0026ndash;0.93; P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The resulting nomogram (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e) demonstrated robust discrimination in training (AUC 0.831; 95% CI: 0.809\u0026ndash;0.853) and prospective validation (AUC 0.783; 95% CI: 0.724\u0026ndash;0.842; Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). Calibration was optimal in training (slope 0.98; Hosmer-Lemeshow P\u0026thinsp;=\u0026thinsp;0.265) and clinically acceptable in validation despite slight underestimation at low-risk thresholds (Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e). Decision curve analysis confirmed net benefit above a 27% risk threshold (nomogram score\u0026thinsp;\u0026ge;\u0026thinsp;85 points), with NNT\u0026thinsp;=\u0026thinsp;8 to prevent one PTB (Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTable 2. Differential Profiles of Prenatal Serological Markers in Preterm vs Term Twin Deliveries\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 163px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMarker\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePreterm\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTerm\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eP-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTrend by GA\u0026darr;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003eWBC (\u0026times;10⁹/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e8.34 (6.95\u0026ndash;10.10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e7.34 (6.21\u0026ndash;8.66)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003eEarly\u0026gt;Mid\u0026gt;Late*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 163px;\"\u003e\n \u003cp\u003eNeutrophil % (NE%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e70.96 \u0026plusmn; 7.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e69.83 \u0026plusmn; 6.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 163px;\"\u003e\n \u003cp\u003eNeutrophils (\u0026times;10⁹/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e5.93 (4.70\u0026ndash;7.41)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e5.07 (4.18\u0026ndash;6.31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003eEarly\u0026gt;Mid\u0026gt;Late*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 163px;\"\u003e\n \u003cp\u003eLymphocytes (\u0026times;10⁹/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e1.63 (1.35\u0026ndash;1.97)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e1.52 (1.27\u0026ndash;1.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 163px;\"\u003e\n \u003cp\u003eNLR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e3.58 (2.80\u0026ndash;4.73)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e3.36 (2.68\u0026ndash;4.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 163px;\"\u003e\n \u003cp\u003ePlatelets (\u0026times;10⁹/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e191 (160\u0026ndash;229)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e184 (152\u0026ndash;218)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003eEarly\u0026gt;Late*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 163px;\"\u003e\n \u003cp\u003eHemoglobin (g/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e115.78 \u0026plusmn; 14.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e121.20 \u0026plusmn; 13.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 163px;\"\u003e\n \u003cp\u003eHematocrit (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e34.27 \u0026plusmn; 3.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e35.71 \u0026plusmn; 3.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 163px;\"\u003e\n \u003cp\u003eTriglycerides (mmol/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 127px;\"\u003e\n \u003cp\u003e3.71 (2.94\u0026ndash;4.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e4.30 (3.52\u0026ndash;5.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 80px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 116px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eComparison of maternal serological markers measured \u0026le;72 hours before delivery. Values presented as median (interquartile range) for non-normally distributed variables or mean \u0026plusmn; standard deviation for normally distributed variables. Statistical significance assessed by Mann-Whitney U test or independent t-test. GA subgroup trends verified by Kruskal-Wallis test.\u003c/p\u003e\n\u003cp\u003eAbbreviations: WBC, white blood cell count; NLR, neutrophil-to-lymphocyte ratio; GA, gestational age.\u003cbr\u003e\u0026nbsp;Notes: Values for preterm and term groups presented as median (IQR) or mean \u0026plusmn; SD.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTrend analysis by GA subgroups: Early PTB (28\u0026ndash;31\u003csup\u003e⁺6\u0026nbsp;\u003c/sup\u003eweeks), Mid PTB (32\u0026ndash;33⁺⁶weeks), Late PTB (34\u0026ndash;36⁺⁶weeks); P\u0026lt;0.05 for intergroup differences.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 3. Association Between Maternal Complications and Preterm Birth in Twin Pregnancies\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 155px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eComplication\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePreterm % (n)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTerm % (n)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOR (95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eP-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 155px;\"\u003e\n \u003cp\u003ePreeclampsia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e13.3 (91)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e7.5 (44)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e1.88 (1.29-2.75)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 155px;\"\u003e\n \u003cp\u003eGestational hypertension\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e6.9 (47)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e3.1 (18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.32 (1.33-4.04)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 155px;\"\u003e\n \u003cp\u003eICP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e6.0 (41)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e3.1 (18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.01 (1.14-3.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 155px;\"\u003e\n \u003cp\u003ePGDM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e3.1 (21)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e1.2 (7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.61 (1.10-6.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.024\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 155px;\"\u003e\n \u003cp\u003eCervical insufficiency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e8.0 (55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.2 (13)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e3.84 (2.08-7.10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 155px;\"\u003e\n \u003cp\u003eThrombophilia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.6 (18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.3 (2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e7.87 (1.78-34.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eOdds ratios (OR) with 95% confidence intervals for preterm birth stratified by maternal complications. Statistical significance determined by \u0026chi;\u0026sup2; test or Fisher\u0026apos;s exact test.\u003c/p\u003e\n\u003cp\u003eAbbreviations:ICP: Intrahepatic cholestasis of pregnancy;PGDM: Pre-gestational diabetes mellitus;PROM: Premature rupture of membranes\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 4. Fetal-Placental Pathologies Associated with Preterm Birth\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFactor\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePreterm % (n)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTerm % (n)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOR (95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eP-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003eTTTS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e1.9 (13)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.0 (0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003eUndefined\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003eFetal distress\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e19.4 (133)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e5.6 (33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e4.07 (2.76-6.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003ePlacenta previa\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e3.1 (21)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.7 (4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e4.61 (1.58-13.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003ePlacental abruption\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e3.1 (21)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e1.4 (8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.28 (1.02-5.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.043\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003ePlacental adhesion\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e5.8 (40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.2 (13)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.72 (1.43-5.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003ePROM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e31.8 (218)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e6.2 (36)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e6.93 (4.84-9.92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003eCord prolapse\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.3 (16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.5 (3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e4.61 (1.34-15.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 156px;\"\u003e\n \u003cp\u003ePolyhydramnios\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e4.5 (31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.1 (12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e2.24 (1.14-4.39)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003ePrevalence and odds ratios of fetal-placental pathologies in preterm versus term twin deliveries. TTTS analysis used Fisher\u0026apos;s exact test due to zero term cases.\u003c/p\u003e\n\u003cp\u003eAbbreviations:PROM: Premature rupture of membranes;TTTS: Twin-twin transfusion syndrome\u003c/p\u003e\n\u003ctable id=\"Tab5\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eMultivariable Predictors of Preterm Birth in Twin Pregnancies\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003ePredictor\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026beta;-coefficient\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003eSE\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003eWald \u0026chi;\u0026sup2;\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003eP-value\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003eaOR (95% CI)\u003cbr\u003e\u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003eMaternal Characteristics\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eAdvanced maternal age (\u0026ge;\u0026thinsp;35y)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.496\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.184\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e7.249\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.007*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e1.64 (1.14\u0026ndash;2.36)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eMonochorionicity\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e1.033\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.154\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e44.965\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e2.81 (2.08\u0026ndash;3.80)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u003cstrong\u003eSerological Markers\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eWBC (per 1\u0026times;10⁹/L increase)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.282\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.036\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e61.502\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e1.33 (1.24\u0026ndash;1.42)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eHematocrit (per 1% increase)\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\u0026minus;0.111\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.020\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e30.732\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.90 (0.86\u0026ndash;0.93)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u003cstrong\u003eMaternal Complications\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003ePreeclampsia\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.901\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.228\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e15.552\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e2.46 (1.57\u0026ndash;3.85)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eGestational hypertension\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e1.068\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.325\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e10.806\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e2.91 (1.54\u0026ndash;5.50)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eIntrahepatic cholestasis\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e1.173\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.331\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e12.541\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e3.23 (1.69\u0026ndash;6.19)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003ePre-gestational diabetes\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e1.180\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.497\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e5.623\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.018*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e3.25 (1.23\u0026ndash;8.62)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eCervical insufficiency\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e1.657\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.357\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e21.583\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e5.24 (2.61\u0026ndash;10.55)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eThrombophilia\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e2.408\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.785\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e9.417\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.002*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e11.12 (2.39\u0026ndash;51.76)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u003cstrong\u003eFetal-Placental Factors\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eFetal distress\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e1.502\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.229\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e43.225\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e4.49 (2.87\u0026ndash;7.03)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003ePlacenta previa\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e2.130\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.592\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e12.951\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.003*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e8.42 (2.64\u0026ndash;26.86)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003ePROM\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e2.056\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.208\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e97.377\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\u0026lt;\u0026thinsp;0.001*\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e7.82 (5.20\u0026minus;11.76)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003eConstant\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.518\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.722\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.515\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e0.473\u003cbr\u003e\u003c/td\u003e\n \u003ctd align=\"char\"\u003e1.68 (-)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\"\u003eAbbreviations: aOR, adjusted odds ratio; CI, confidence interval; PROM, premature rupture of membranes.\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\"\u003eNotes: Model developed using backward logistic regression (retention P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). \u0026beta;-coefficients used for nomogram scoring.\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \n\n\u003c/table\u003e\n\n\u003ch2\u003ePredictor Impact on Neonatal Morbidity\u003c/h2\u003e\n\u003cp\u003eExploratory SHAP analysis (Supplementary Fig. S3) revealed PROM as the dominant predictor of respiratory distress syndrome (mean |SHAP|=0.211), while thrombophilia strongly associated with necrotizing enterocolitis (mean |SHAP|=0.169). Monochorionicity predicted intraventricular hemorrhage (mean |SHAP|=0.187).\u003c/p\u003e\n"},{"header":"DISCUSSION","content":"\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003ePrincipal Findings and Contextualization\u003c/h2\u003e\u003cp\u003eThis multicenter study developed and prospectively validated the first preterm birth (PTB) prediction model for twin pregnancies that integrates routinely available serological markers with established clinical risk factors. Our nomogram demonstrated robust discrimination (training AUC\u0026thinsp;=\u0026thinsp;0.831; validation AUC\u0026thinsp;=\u0026thinsp;0.783) and calibration across cohorts, outperforming existing tools such as the PREMET score (AUC\u0026thinsp;=\u0026thinsp;0.71) which lacks biomarker integration [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Notably, four key innovations emerge from this work. First, the identification of white blood cell count (WBC) and hematocrit as independent predictors after multivariable adjustment suggests subclinical inflammation (elevated WBC) and hemodynamic adaptation (decreased hematocrit) precede PTB by 4\u0026ndash;6 weeks\u0026mdash;enabling earlier intervention than contemporary ultrasound-based models [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Second, decision curve analysis established a clinically actionable 27% risk threshold for net benefit, translating to specific biomarker-clinical combinations (e.g., WBC\u0026thinsp;\u0026gt;\u0026thinsp;8.5\u0026times;10⁹/L with monochorionicity\u0026thinsp;=\u0026thinsp;31% risk). Third, the 11.12-fold increased PTB risk associated with thrombophilia supports emerging evidence linking coagulation dysregulation to preterm labor pathogenesis [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Finally, SHAP analysis revealed preterm premature rupture of membranes (PROM) as the dominant predictor of respiratory distress syndrome (mean |SHAP|=0.211), reinforcing established infection-mediated pathways in neonatal morbidity [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eOur findings resolve critical gaps in twin PTB prediction identified in prior literature. While leukocytosis has been associated with PTB in singletons [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], twin-specific predictive thresholds were previously unestablished. We demonstrate that WBC\u0026thinsp;\u0026gt;\u0026thinsp;8.34\u0026times;10⁹/L confers a 33% increased risk per unit rise\u0026mdash;providing a more precise quantitative metric than qualitative \"inflammation\" markers employed in earlier models [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Importantly, unlike prediction tools validated solely in retrospective cohorts [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], our prospective external validation (n\u0026thinsp;=\u0026thinsp;227) confirms generalizability across diverse clinical settings. The observed 15.3% relative reduction in AUC represents expected performance attenuation when transitioning to external cohorts [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], yet maintains clinically relevant discrimination. Mechanistically, SHAP analysis directly linked thrombophilia to placental vascular pathology, explaining its strong PTB association (adjusted OR\u0026thinsp;=\u0026thinsp;11.12) and significant impact on necrotizing enterocolitis (mean |SHAP|=0.169)\u0026mdash;a finding aligning with recent placental histopathology studies [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eClinical Implications and Implementation\u003c/h2\u003e\u003cp\u003eThe validated nomogram offers immediate clinical utility through three key applications. First, it enables precise risk stratification where high-risk women (nomogram score\u0026thinsp;\u0026gt;\u0026thinsp;140 points, equivalent to \u0026gt;\u0026thinsp;50% PTB probability) warrant intensified monitoring protocols such as fortnightly cervical length assessments and targeted preventive measures including vaginal progesterone [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Second, at the 27% risk threshold identified through decision curve analysis, 68% of low-risk women could avoid unnecessary interventions while maintaining favorable benefit-risk balance (number needed to treat [NNT]\u0026thinsp;=\u0026thinsp;8 to prevent one PTB). Third, serial monitoring of biomarker trajectories (particularly WBC and neutrophil-to-lymphocyte ratio [NLR]) may provide early warning of subclinical chorioamnionitis, potentially prompting preemptive antibiotic administration before overt PROM develops [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eLimitations and Future Directions\u003c/h2\u003e\u003cp\u003eStudy limitations include regional recruitment potentially limiting ethnic diversity, though validation cohort characteristics aligned with multinational twin registry data [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Biomarker assessment at a single timepoint represents another constraint; future models should incorporate serial measurements to capture dynamic physiological changes. Limited power for rare outcomes (e.g., thrombophilia n\u0026thinsp;=\u0026thinsp;20) necessitates validation in larger cohorts. Key research priorities include prospective validation of SHAP-derived neonatal outcome predictions, development of point-of-care testing protocols for resource-limited settings, and integration of ultrasound biomarkers (cervical length) with serological profiles to enhance predictive performance.\u003c/p\u003e\u003c/div\u003e"},{"header":"CONCLUSION","content":"\u003cp\u003eThis study establishes a prospectively validated nomogram that synthesizes accessible serological markers with clinical predictors to accurately stratify PTB risk in twin pregnancies. By establishing twin-specific biomarker thresholds and a clinically actionable 27% risk intervention cutoff, the model facilitates personalized management while optimizing healthcare resource utilization. Implementation studies should now assess its impact on reducing twin PTB rates\u0026mdash;a persistent challenge in modern obstetrics with significant perinatal consequences.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eICP: Intrahepatic cholestasis of pregnancy\u003cbr\u003e PGDM: Pre-gestational diabetes mellitus\u003cbr\u003e PROM: Premature rupture of membranes\u003cbr\u003e TTTS: Twin-twin transfusion syndrome\u003c/p\u003e\n\u003cp\u003eBMI: Body mass index\u003c/p\u003e\n\u003cp\u003eIVF-ET: In vitro fertilization and embryo transfer\u003c/p\u003e\n\u003cp\u003eGA: Gestational age\u003c/p\u003e\n\u003cp\u003eCI: Confidence interval;\u003c/p\u003e\n\u003cp\u003eSE: Standard error\u003c/p\u003e\n\u003cp\u003eWBC: White blood cell count\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll subjects and data were obtained from patient records at Fujian Provincial Maternal and Child Health Hospital. The project has been approved by the Medical Ethics Committee of Fujian Maternal and Child Health Hospital (2022KY006).Informed written consent was obtained from all patients.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData were anonymized and no patient identifying information was included for preserve patient confidentiality. All data to evaluate the conclusions in the paper available for scientific purposes if needed.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was funded by Joint Funds for the Innovation of Science and Technology,Fujian Province (2020Y9401); Fujian Provincial Health Technology Project(2024ZD01005); National Key Clinical Specialty Construction\u003c/p\u003e\n\u003cp\u003eProgram of China (Obstetric); Fujian Provincial Natural Science Foundation of\u003c/p\u003e\n\u003cp\u003eChina (2024Y0035); Joint Funds for the Innovation of Science and Technology,\u003c/p\u003e\n\u003cp\u003eFujian Province (2024Y9536).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contribution statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eQ.Z and Y.W designed the analyses, and Q.Z drafted the manuscript.J.Y and X.X conceptualized the study; and L.L,B.Xand H.Z contributed to data acquisition. All authors have revised the manuscript for important intellectual content.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSmits J, Monden C. Twinning rates in developed countries. Popul Dev Rev. 2011;37(2):253-258.\u003c/li\u003e\n\u003cli\u003eMartin JA, Hamilton BE, Osterman MJK. Births in the United States, 2020. NCHS Data Brief. 2021;(418):1-8.\u003c/li\u003e\n\u003cli\u003eSociety for Assisted Reproductive Technology. National summary report. Published 2023. Accessed [Date]. [URL]\u003c/li\u003e\n\u003cli\u003eCheong-See F, Schuit E, Arroyo-Manzano D, et al. Stillbirth risks in twins. BMJ. 2016;354:i4353. doi:10.1136/bmj.i4353\u003c/li\u003e\n\u003cli\u003eVogel JP, Chawanpaiboon S, Moller AB, Watananirun K, Bonet M, Lumbiganon P. Global preterm birth epidemiology. Best Pract Res Clin Obstet Gynaecol. 2018;52:3-12. doi:10.1016/j.bpobgyn.2018.04.003\u003c/li\u003e\n\u003cli\u003eAmerican College of Obstetricians and Gynecologists. Multifetal gestations. Obstet Gynecol. 2021;137(6):e145-e162. doi:10.1097/AOG.0000000000004397\u003c/li\u003e\n\u003cli\u003eConde-Agudelo A, Romero R, Hassan SS, Yeo L. Cervical length in twins. Am J Obstet Gynecol. 2010;203(2):128.e1-128.e12. doi:10.1016/j.ajog.2010.02.064\u003c/li\u003e\n\u003cli\u003eAkkermans J, Payne B, von Dadelszen P, et al; PIERS Study Group. PREMET models. Ultrasound Obstet Gynecol. 2017;50(5):621-630. doi:10.1002/uog.17516\u003c/li\u003e\n\u003cli\u003eLim AC, Schuit E, Bloemenkamp K, et al. Validation of twin PTB model. Prenat Diagn. 2016;36(6):526-530. doi:10.1002/pd.4820\u003c/li\u003e\n\u003cli\u003eRomero R, Dey SK, Fisher SJ. Preterm labor causes. Science. 2014;345(6198):760-765. doi:10.1126/science.1251816\u003c/li\u003e\n\u003cli\u003eJelliffe-Pawlowski LL, Ryckman KK, Bedell B, et al. Metabolic/inflammatory markers. Am J Perinatol. 2017;34(4):338-346. doi:10.1055/s-0036-1586505\u003c/li\u003e\n\u003cli\u003eMoons KGM, Altman DG, Reitsma JB, et al. TRIPOD explanation. Ann Intern Med. 2015;162(1):W1-W73. doi:10.7326/M14-0698\u003c/li\u003e\n\u003cli\u003ePark HJ, Park KH, Kim YN, et al. Biomarkers in cervicovaginal fluid. PLoS One. 2017;12(7):e0180878. doi:10.1371/journal.pone.0180878\u003c/li\u003e\n\u003cli\u003eSteer PJ. Maternal hemoglobin/birth weight. Am J Clin Nutr. 2000;71(5 suppl):1285S-1287S. doi:10.1093/ajcn/71.5.1285s\u003c/li\u003e\n\u003cli\u003ePaidas MJ, Hossain N. Thrombophilia outcomes. Clin Obstet Gynecol. 2006;49(4):850-860. doi:10.1097/01.grf.0000211948.18735.0f\u003c/li\u003e\n\u003cli\u003eRedline RW. Placental pathology. Placenta. 2008;29(suppl A):S86-S91. doi:10.1016/j.placenta.2008.01.017\u003c/li\u003e\n\u003cli\u003eConde-Agudelo A, Romero R. Biophysical tests in twins. Am J Obstet Gynecol. 2014;211(6):583-595. doi:10.1016/j.ajog.2014.07.047\u003c/li\u003e\n\u003cli\u003eFox NS, Saltzman DH, Klauser CK, et al. Combined fFN/CL in twins. Am J Obstet Gynecol. 2009;201(3):313.e1-313.e5. doi:10.1016/j.ajog.2009.06.027\u003c/li\u003e\n\u003cli\u003eSteyerberg EW, Vergouwe Y. Prediction model validation. Eur Heart J. 2014;35(29):1925-1931. doi:10.1093/eurheartj/ehu207\u003c/li\u003e\n\u003cli\u003eThilaganathan B. Maternal adaptation to twins. Ultrasound Obstet Gynecol. 2021;57(1):26-32. doi:10.1002/uog.23588\u003c/li\u003e\n\u003cli\u003eRomero R, Conde-Agudelo A, Da Fonseca E, et al. Vaginal progesterone for short cervix. Ultrasound Obstet Gynecol. 2016;48(3):308-317. doi:10.1002/uog.15956\u003c/li\u003e\n\u003cli\u003eKenyon S, Boulvain M, Neilson JP. Antibiotics for PROM. Cochrane Database Syst Rev. 2013;(12):CD001058. doi:10.1002/14651858.CD001058.pub3\u003c/li\u003e\n\u003cli\u003eBarfield WD. Very preterm birth implications. Clin Perinatol. 2018;45(3):565-577. doi:10.1016/j.clp.2018.05.006\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Prediction Model, Serological Markers, Prospective Validation, Multicenter Study","lastPublishedDoi":"10.21203/rs.3.rs-6947382/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6947382/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eTwin pregnancies, increasingly prevalent due to assisted reproductive technology, carry a 7-10-fold higher preterm birth (PTB) risk than singletons, with \u0026gt;50% delivering before 37 weeks. Existing prediction models predominantly rely on obstetric history and ultrasound parameters, lacking integration of routinely available serological markers. While inflammation and hematological dysregulation are implicated in PTB pathogenesis, the predictive value of complete blood count (CBC) and metabolic biomarkers remains underexplored in twins. Moreover, few models undergo prospective validation, limiting clinical adoption. This study aimed to develop and rigorously validate a clinically implementable PTB prediction tool by synthesizing serological indicators with established risk factors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStudy Design: \u003c/strong\u003eA single-center study comprising a retrospective training cohort (n=1,270 twin pregnancies, 2019–2021) and a prospective validation cohort (n=227). Multivariable logistic regression identified independent predictors from maternal characteristics, prenatal serology (complete blood count, lipids), pregnancy complications, and fetal factors. Model performance was assessed for discrimination (area under the receiver operating characteristic curve, AUC), calibration (Hosmer-Lemeshow test, bootstrap-corrected calibration curves), and clinical utility (decision curve analysis). The relative weights of predictors influencing neonatal outcomes were additionally explored.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003e13 variables were independently associated with preterm birth (all P\u0026lt;0.05), including maternal age (adjusted odds ratio [aOR]=1.64), monochorionicity (aOR=2.81), elevated white blood cell count (aOR=1.33), preterm premature rupture of membranes (aOR=7.82), and preeclampsia (aOR=2.46). The model demonstrated an AUC of 0.831 (95% confidence interval [CI] 0.809–0.853) in the training cohort and 0.783 (95% CI 0.724–0.842) in the validation cohort. Calibration was good in both cohorts (Hosmer-Lemeshow P=0.265 and P=0.400, respectively). Decision curve analysis confirmed net benefit across clinically relevant threshold probabilities (13–95% training; 27–100% validation). Exploratory analysis indicated fetal distress and preterm premature rupture of membranes had the highest relative weights for neonatal asphyxia (0.211) and pneumonia (0.212), respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion: \u003c/strong\u003eThis validated nomogram integrates routine serological markers with clinical predictors to accurately stratify preterm birth risk in twin pregnancies (AUC \u0026gt;0.78). It demonstrates immediate clinical utility for targeted monitoring and requires external validation in diverse populations.\u003c/p\u003e","manuscriptTitle":"Development and Validation of a Preterm Birth Prediction Model for Twin Pregnancies: A Single-Center Prospective Study Integrating Serological Markers","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-22 07:13:57","doi":"10.21203/rs.3.rs-6947382/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"56ffcc60-231d-4bd8-bf59-ce4b83af043b","owner":[],"postedDate":"August 22nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-02T11:26:02+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-22 07:13:57","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6947382","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6947382","identity":"rs-6947382","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}