Imagine facing the challenges of carrying twins, only to have preeclampsia throw a wrench into an already complex pregnancy. This serious condition, marked by high blood pressure and potential complications, can escalate quickly and put both mother and babies at grave risk. But here's where it gets intriguing—while we know preeclampsia affects about 5-7% of all pregnancies and is a leading cause of maternal and perinatal issues, its behavior in twin pregnancies remains a puzzle that's now being unlocked through new research. And this is the part most people miss: twin pregnancies aren't just twins; they're a unique scenario where preeclampsia might strike harder, demanding tailored predictions to protect moms and their little ones.
Published on November 18, 2025, in BMC Pregnancy and Childbirth (volume 25, article 1228), this open-access study delves into a prediction model for adverse maternal outcomes linked to preeclampsia in twin pregnancies. The research team, led by Shuisen Zheng and Wei Zheng, along with Yuting Gao, Xiaoling Chen, Na Lin, and Qing Han from Fujian Maternal and Child Health Hospital, aimed to pinpoint risk factors and craft a straightforward clinical tool to guide early interventions.
In the abstract, they summarize their findings: Preeclampsia, a pregnancy complication characterized by hypertension and proteinuria after 20 weeks, can lead to severe issues like eclampsia, placental abruption, and postpartum hemorrhage. This study focused on twin pregnancies to explore risk factors and build a simplified prediction model. From 2014 to 2021, they analyzed data from pregnant women hospitalized at Fujian Maternal and Child Health Hospital, including 2,570 singleton pregnancies and 459 twin pregnancies with preeclampsia. They divided the twin pregnancies into groups: those with and without adverse outcomes, based on criteria like maternal death, central nervous system problems, cardiorespiratory issues, renal or hepatic complications, hematological problems, or obstetric challenges such as HELLP syndrome (a dangerous mix of hemolysis, elevated liver enzymes, and low platelets). Using statistical methods like univariate analysis, least absolute shrinkage and selection operator (LASSO), and multivariate logistic regression, they identified key risk factors. The final model, evaluated with receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA), showed an area under the curve (AUC) of 0.798, indicating reliable predictive power. The model highlighted that gestational age at admission under 32 weeks, being multiparous, elevated peak systolic blood pressure, high creatinine levels, elevated neutrophil/high-density lipoprotein ratio (NHR), and low platelet counts were independent predictors of adverse outcomes in twin pregnancies. The calibration curve demonstrated that the model's predictions aligned well with actual outcomes, suggesting it could help clinicians intervene promptly and improve results in twin pregnancies complicated by preeclampsia.
The introduction sets the stage by explaining preeclampsia as a complex disorder with unclear origins, possibly stemming from various causes, that can deteriorate rapidly into life-threatening events. It affects 5-7% of pregnancies and contributes significantly to maternal and fetal morbidity. For instance, think of it as a storm brewing in the body where high blood pressure damages blood vessels, leading to organ stress. When it comes to twin pregnancies, rates have risen globally and in China due to advancements in assisted reproductive technologies like IVF. Studies show twin pregnancies carry a higher risk of adverse outcomes compared to singleton ones, making specialized tools essential. Existing models, such as the fullPIERS and miniPIERS developed by Peter von Dadelszen and colleagues, predict risks in preeclampsia but were based on mixed pregnancy types with few twins, raising doubts about their applicability here. This study fills that gap by focusing solely on twin pregnancies, conducting a cohort analysis to develop a tailored model that healthcare workers in primary or low-resource settings can use easily. By identifying high-risk women early, it could enable timely actions like magnesium sulfate for seizures, blood pressure control, or careful monitoring, potentially saving lives.
Diving into materials and methods, the researchers recruited women with preeclampsia from January 2014 to December 2021 at Fujian Maternal and Child Health Hospital, totaling 2,029 cases—2,570 singletons and 459 twins. They split the twins into adverse and non-adverse outcome groups, as shown in Figure 1. Adverse maternal outcomes included serious conditions like death, cardiovascular complications (e.g., heart therapy needs or pulmonary edema), central nervous system issues (e.g., eclampsia or stroke), liver dysfunction, acute kidney failure, hematological problems (e.g., platelet counts below 50 × 10^9/L or disseminated intravascular coagulation), or obstetric complications (e.g., HELLP syndrome, placental abruption, or severe postpartum hemorrhage). Preeclampsia was diagnosed if blood pressure hit 140/90 mmHg or higher on two occasions four hours apart, accompanied by proteinuria (over 300 mg/24 hours, protein/creatinine ratio above 0.3, or dipstick 2+), or without proteinuria if there were thrombocytopenia, renal dysfunction, liver issues, pulmonary edema, or symptoms like headaches or vision changes. DIC, for example, is a widespread clotting disorder triggered by conditions like preeclampsia, often leading to bleeding complications. HELLP syndrome is defined by hemolysis (with lactate dehydrogenase >600 U/L), elevated liver enzymes (aspartate aminotransferase ≥70 U/L), and low platelets (<100 × 10^9/L). Exclusion criteria removed cases with outcomes already present, incomplete data, abortions before 28 weeks, or early-onset preeclampsia.
Data collection pulled from electronic records, covering demographics (age, pre-pregnancy weight/BMI, education, delivery mode, parity, IVF use), symptoms (dizziness, headache, bleeding, nausea), treatments (magnesium sulfate, antihypertensives), lab tests (blood counts, liver/renal function, coagulation), complications (gestational diabetes, scarred uterus, premature rupture of membranes, cholestasis of pregnancy, placenta issues), and inflammatory markers like systemic inflammation response index (SIRI: neutrophil count × monocyte count / lymphocyte count), systemic immune inflammation index (SII: platelet count × neutrophil count / lymphocyte count), neutrophil/high-density lipoprotein ratio (NHR: neutrophils / HDL-C), monocyte/high-density lipoprotein ratio (MHR: monocytes / HDL-C), lymphocyte/high-density lipoprotein ratio (LHR: lymphocytes / HDL-C), and platelet/high-density lipoprotein ratio (PHR: platelets / HDL-C). Statistical analysis used Excel and R software, imputing missing data under 30% with multiple imputation. They ran univariate analysis, LASSO for variable selection with 10-fold cross-validation, and multivariate logistic regression. The model was evaluated with ROC AUC, calibration plots, and DCA. Ethical approval came from the hospital's committee, with consent waived due to retrospective design.
Results revealed differences between singleton and twin pregnancies with preeclampsia. Twins showed higher rates of adverse outcomes (17.4% vs. 13.6% in singletons), with no differences in education levels, but twins had younger ages, more primiparous women, higher height, admission weight/BMI, lower pre-pregnancy weight/BMI, and earlier delivery weeks. Specific complications like severe postpartum hemorrhage and HELLP were more common in twins, though placental abruption, posterior reversible encephalopathy, and eclampsia didn't differ significantly. Comparing adverse vs. non-adverse twin groups, no differences existed in age, height, pre-pregnancy BMI, admission BMI, gravidity, education, gestational age at admission/delivery, advanced age, or prenatal exams. Adverse outcome women had more placenta previa, but no differences in comorbidities like gestational diabetes, premature rupture, scarred uterus, preterm birth, IVF, oligohydramnios, hydramnios, cholestasis, hypothyroidism, hyperthyroidism, autoimmune diseases, or early preeclampsia. They showed higher systolic blood pressure, more MgSO4 use, but no symptom differences. Lab-wise, adverse groups had lower albumin, direct bilirubin, lactate dehydrogenase; higher urea nitrogen, creatinine, uric acid; no differences in proteinuria. Hematologically, no differences in hematocrit, hemoglobin, mean platelet volume, or distribution width, but differences in platelet hematocrit, large cell ratio, and count. Coagulation showed higher FDP and thrombin time, lower fibrinogen. In inflammatory markers, NHR and PHR differed. LASSO analysis selected variables like gestational age <32 weeks, multiparity, peak systolic blood pressure, AST, creatinine, NHR, and platelet count for multivariate regression. The model identified gestational age <32 weeks (OR=3.028), multiparity (OR=2.030), peak systolic blood pressure (OR=1.021), NHR (OR=1.238), creatinine (OR=1.034) as risk factors, and platelet count (OR=0.986) as protective. The nomogram (Figure 3) visualized the model. Evaluation showed a 0.798 AUC, good calibration (Figure 4A), and DCA indicating better net benefit than treating all or none (Figure 4B-C).
The discussion emphasizes that preeclampsia heightens adverse outcome risks, and twins face even greater challenges due to complex physiology. Studies support higher risks in twins, warranting specialized models. Risk factors like early gestational admission, high systolic pressure, and multiparity align with prior research—early admission often means more severe disease, and multiparity's role needs further study. Serological markers, such as platelets and creatinine, are crucial; low platelets signal clotting issues, and high creatinine indicates kidney stress. NHR, a new inflammatory marker, reflects inflammation and lipid profiles, linked to preeclampsia's onset. The model's AUC of 0.798 is promising, comparable to others but tailored for twins. Unlike fullPIERS or miniPIERS with limited twin data, this study focuses exclusively on twins, enhancing relevance. However, it raises questions: Should we intervene aggressively in twin pregnancies, even if it means early delivery that could affect neonatal health? This is where controversy sparks—balancing maternal safety with baby well-being requires careful judgment. But here's a counterpoint: Some experts argue that over-intervention might lead to unnecessary risks, like preterm complications, so personalized risk assessments are key. Future multicentre studies could expand this, but for now, clinicians can use this model to stratify risks and guide care.
Limitations include the retrospective design, potential selection bias, small sample, and single-centre focus, which might limit generalizability. Missing data and excluded variables could have impacted accuracy, and long-term follow-ups are absent. Still, this model offers a practical tool for low-resource settings.
In conclusion, this study pinpoints key risk factors for adverse outcomes in preeclamptic twin pregnancies and validates a prediction model to aid timely interventions, ultimately benefiting patients. Data availability is upon request to the corresponding authors.
Peer review reports are available, and the article lists references from 1 to 34, detailing studies on preeclampsia epidemiology, models, and markers.
Funding came from Fujian Province joint funds and natural science foundation. Authors contributed as noted, with correspondence to Na Lin or Qing Han.
Ethics approval was granted, with consent waived.
What do you think—does this model change how we approach twin pregnancies with preeclampsia, or should we debate the ethics of early interventions? Share your views in the comments; do you agree that inflammation markers like NHR could revolutionize risk assessment, or is there a better way? Let's discuss!
