Preeclampsia is a unique complication of pregnancy and occurs in 5–7% of all pregnancies. It is the leading cause of maternal and fetal perinatal morbidity and mortality (Rana et al., 2019). The etiology and pathogenesis of preeclampsia remain unclear (Ives et al., 2020). Besides, there is no effective treatment for preeclampsia in addition to the termination of pregnancy. Therefore, it is important to strengthen the screening and prediction of high-risk groups before delivery. With the development of assisted reproductive technology, the incidence of twin pregnancy is increasing yearly (American College of Obstetricians and Gynecologists, 2021). Clinical studies have shown that the incidence of preeclampsia is two to three times higher in twin than in singleton pregnancies (Baha et al., 2000; Wang et al., 2021). However, in recent years, research on the risk prediction of preeclampsia is mostly limited to singleton pregnancies. Maternal serum makers such as coagulation function and liver function are considered to be closely related to preeclampsia (Duan et al., 2020). In addition, many studies have shown that primiparas and pregnant women with a high pregestational body mass index are more likely to develop preeclampsia (Yang et al., 2021). But it is unclear whether these predictors in singletons equally apply to twin pregnancies. Therefore, we conducted a retrospective cohort study to explore the high-risk factors associated with preeclampsia in twin pregnancies and built a clinical prediction model to guide early interventions and improve pregnancy outcomes.

We found that the probability of twin pregnancy being complicated with preeclampsia was 13.16%, slightly higher than in previous reports (Francisco et al., 2017). We performed a more comprehensive screening of high-risk factors for preeclampsia in twin pregnancies through the acquisition of clinical medical records of patients. The results of logistic multivariate analysis showed that primiparity, regular obstetric visits, pre-pregnancy BMI, CREm, URIC, MPV, HDL, LDH, Fib, primipara, pre-pregnancy BMI, and regular prenatal were independently associated with preeclampsia in twin pregnancies. The nomogram prediction model (concordance index 0.821) was further constructed with good accuracy and conformity, and could make individualized predictions for pregnant women with twin pregnancies complicated by preeclampsia, to allow for timely clinical interventions and thus reduce the number of adverse outcomes in mothers and children.

The expert consensus published by the American College of Obstetricians and Gynecologists showed that primiparity, obesity, advanced age, and assisted reproductive technology were high-risk factors for pregnancies complicated by preeclampsia in singletons [American College of Obstetricians and Gynecologists (ACOG), 2020]. Taguchi et al. (2014) found that pre-pregnancy BMI and primiparity were independent risk factors for preeclampsia in twin pregnancies. Similarly, in another prospective cohort study, Chen et al. (2020) reported that pre-pregnancy BMI was significantly associated with the risk of preeclampsia in twin pregnancies, which is consistent with our findings. In addition, regular antenatal check-ups, strengthening the monitoring of risk factors, and timely intervention can reduce the incidence of preeclampsia in twin pregnancies.

Various biochemical factors in maternal serum play an important role in the pathophysiological development of preeclampsia, and the composition, source, and mechanism are different for each of these factors. At present, the serum markers used to screen for preeclampsia in twin pregnancies mainly include pregnancy-associated plasma protein A, placental protein 13, placental growth factor (Maymon et al., 2017), inhibin A, and unconjugated estriol (Kim et al., 2019); however, widespread roll-out of these indicators is difficult to develop and promote in underdeveloped areas. Therefore, this study measures peripheral blood parameters, screens out predictive indicators with high sensitivity and specificity, and provides a basis for the development of economical and efficient screening programs for twin pregnancies with preeclampsia.

Previous studies on the relationship between MPV and preeclampsia were based on pregnant women with singleton pregnancies, while there were fewer studies involving those with twin pregnancies. In a longitudinal study, Mayer-Pickel et al. (2021) found that MPV is significantly elevated even in early pregnancy in women who develop preeclampsia. A cross-sectional comparative study by Tesfay et al. (2019) showed that MPV had a good predictive value for the development of preeclampsia. Jakobsen et al. (2019) found that women with preeclampsia had significantly higher MPV than in women without preeclampsia in a systematic review and meta-analysis and they believed that MPV is the most promising biomarker for evaluating platelet function in preeclampsia. However, these studies were based on singleton pregnancies; we found that an increase in peripheral blood MPV was associated with an increased risk of preeclampsia in twin pregnancies as well. The increase in MPV reflects the enhancement of platelet activation. In patients with preeclampsia, because of the increase in platelet consumption and destruction, the bone marrow produces and releases a large quantity of platelets, which leads to an increase in MPV (Kaito et al., 2005). Conversely, preeclampsia may cause complex diseases in the endogenous blood coagulation pathway and consumes Fib, resulting in a decrease in Fib (Han et al., 2014). We found an increase in the risk of preeclampsia in twin pregnancies with decreasing Fib, which is similar to the results reported in related studies (Liu et al., 2020).

Some studies have shown that, as pregnancy progresses, serum total cholesterol, TG, and LDL concentrations increase, and this abnormal dyslipidemia during pregnancy is related to preeclampsia and other adverse pregnancy outcomes (Jin et al., 2016). The pathophysiological basis may be that the increase in circulating lipid levels leads to an accumulation of lipids in endothelial cells, which reduces the release of prostacyclin and leads to oxidative stress (Ghio et al., 2011). A meta-analysis by Spracklen et al. (2014) showed that low levels of HDL are significantly associated with the risk of preeclampsia, and our study found that women with twin pregnancies who have low HDL also have this risk.

The main energy supply pathway to the placenta occurs through glycolysis. LDH is an intracellular enzyme required for glycolysis. Under hypoxic conditions, LDH is activated to promote glycolysis and produce large amounts of lactic acid. Therefore, elevated LDH levels often indicate cell damage and dysfunction. Studies have shown that peripheral blood LDH levels in patients with preeclampsia are significantly increased (Wu et al., 2021). Our study found this feature in twin pregnancies with preeclampsia. In addition, some scholars have found that high levels of LDH are significantly related to the occurrence of adverse perinatal outcomes in patients with preeclampsia (Dave et al., 2016). It is necessary to explore in future studies whether there is a dose-response relationship between high levels of LDH and the occurrence of adverse preeclampsia outcomes.

The kidney is an important organ involved in preeclampsia. Damage to glomerular endothelial cells and destruction of the relationship between endothelial cells and podocytes are the main underlying pathogenetic mechanisms of preeclampsia (Moghaddas Sani et al., 2019). Patients with preeclampsia are prone to renal dysfunction caused by extensive renal arteriolar spasm. This causes glomerular swelling, decreased glomerular filtration rate, and decreased renal blood flow, which leads to obstruction of the excretion and clearance of renal metabolites such as URIC and CREm; this results in an increase in serum URIC and CREm. Zhao et al. (2021) found that elevated serum uric acid in Chinese Han women with gestational hypertension indicated an increased risk of their progression to preeclampsia. In a prospective case-control study, Enaruna et al. (2014) found that the URIC level of patients with preeclampsia was significantly increased and that higher levels reflected disease severity. Jhee et al. (2019) developed models using machine learning to predict late-onset preeclampsia and found that CREm levels were one of the most influential variables included in the prediction models. However, Few studies focused on the relationship between these indicators and preeclampsia in twin pregnancy. Our study make up for this and demonstrating that URIC and CREm levels were independent risk factors for preeclampsia in twin pregnancies.

We are aware of the limitations of our study. Our study is a single-center retrospective cohort study, which may result in a risk of overestimating model performance. However, as a tertiary hospital in Southeast China, our number of cases is still representative, and we believe the model would be useful in pregnancy supervision. Further prospective cohort and multi-center joint research will help to further expand the research results, improve the prediction accuracy of the model, and confirm the conclusions.

In summary, the prediction model of preeclampsia in twin pregnancies constructed in this study has good accuracy and high clinical application value, and can be used as a reference for obstetricians. We should pay attention to twin pregnancies with high-risk factors in clinical practice and perform careful screening, early intervention, and active treatment to reduce the occurrence of adverse pregnancy outcomes.

This study performed a more comprehensive screening of high-risk factors for preeclampsia in twin pregnancies through the acquisition of clinical medical records of patients. The model for the prediction of preeclampsia in twin pregnancies has high accuracy and specificity. It can be used to assess the risk of preeclampsia in twin pregnancies.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

The studies involving human participants were reviewed and approved by Ethics Committee of Fujian Maternity and Child Health Hospital. The patients/participants provided their written informed consent to participate in this study.

QH and JY: conception and design of the research. SZ and RC: acquisition of data. HZ, SZ, and RC: analysis and interpretation of the data and statistical analysis. QH, JY, and HZ: funding acquisition. QH and SZ: writing of the manuscript. JY, HZ, and QH: critical revision of the manuscript for intellectual content. All authors read and approved the final draft.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.