Nomogram model for predicting spontaneous preterm birth in twin pregnancies: a case-control study

doi:10.21203/rs.3.rs-5021146/v1

Download PDF

Research Article

Nomogram model for predicting spontaneous preterm birth in twin pregnancies: a case-control study

https://doi.org/10.21203/rs.3.rs-5021146/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

This study aimed to analyze the correlation factors of spontaneous preterm birth in twin pregnancies and construct a predictive model, with the hope of providing clinical value for the prediction of spontaneous preterm delivery in twin pregnancies.

Methods

A total of 218 pregnant women with twin pregnancies at Jiaxing Women and Children's Hospital, Wenzhou Medical University between June 2021 and May 2024 were enrolled. One-way Analysis of Variance (One-way ANOVA) and multivariate logistic regression analysis were used to analyze the correlation factors, and a prediction model was constructed. The nomogram model was established using R, and evaluated by the area under the ROC curve, C-index, and decision curve analysis (DCA).

Results

One-way ANOVA showed that body mass index (BMI), length of cervical canal in the second trimester, presence or absence of cervical funnel, vaginitis during pregnancy, gestational diabetes mellitus (GDM) and intrauterine hemoglobin levels were associated with spontaneous preterm birth in twin pregnancies (P < 0.05). Multivariable logistic regression analysis showed that BMI, cervical length in the second trimester and gestational vaginitis are independent correlation factors for spontaneous preterm birth in twin pregnancies. After validation, the AUC value of the combined prediction of 3 correlation factors was the largest (0.852). Overall, a nomogram model with C-index of 0.838 was successfully constructed for predicting preterm birth in twin pregnancies.

Conclusion

BMI combined with cervical canal length in the second trimester and gestation vaginitis can improve the predictive value for the delivery outcome of spontaneous preterm birth in twin pregnancies.

Twin pregnancies

spontaneous preterm birth

delivery outcome

nomogram

risk factor

With the widespread application of assisted reproductive technologies, the incidence of twin or multiple pregnancies has increased significantly. It has been reported that the rate of twin pregnancies in China rose to 3.22% in 2020 [1]. Twin pregnancies elevate the risk of various maternal and fetal complications, particularly directly related to the occurrence of preterm birth [2, 3]. In 2019, the preterm birth rate among twin pregnancies was approximately 58.71% [4], with about 10.7% of the preterm births occurred at gestational age less than 32 weeks [5]. Preterm birth is classified into therapeutic preterm birth and spontaneous preterm birth based on their pathogenesis. Specifically, therapeutic preterm birth refers to the preterm termination of pregnancy for the safety of both the mother and baby. Spontaneous preterm birth is defined as preterm birth that occurs without evident induction factors or causes. A relevant study indicated that, regardless of whether vaginal delivery or cesarean section is employed, 40–45% are spontaneous preterm births [6]. It is noteworthy that preterm birth is the leading cause of neonatal death and various serious complications [7]. Thus, our study aims to identify high-risk individuals for preterm birth as early as possible and to implement timely intervention measures by exploring the clinical characteristics and associated risk factors of preterm birth in twin pregnancies and establishing an effective prediction model for preterm birth. This approach will not only reduce the economic burden on preterm birth families, but also contributes to the overall health of the entire birth population.

Study population

According to the inclusion and exclusion criteria described below, a total of 218 pregnant women with twin pregnancies from the Jiaxing Maternal and Child Health Hospital, Wenzhou Medical University between June 2021 and May 2024 were included in this study. This study was approved by the Ethics Committee of Jiaxing Maternal and Child Health Hospital, Wenzhou Medical University.

Inclusion criteria: a. Pregnant women with twin pregnancies who undergo regular prenatal check-ups at our hospital. b. Pregnant women with delivery time of ≥ 28 weeks and < 37 weeks and the newborn weight of ≥ 1000 g were included in the preterm birth group.

Exclusion criteria: a. Monochorionic monoamniotic twin pregnancies, b. Abortion at gestational age of < 28 weeks, c. Therapeutic preterm births: those who terminate pregnancy prematurely due to pregnancy-related diseases, complications, or other reasons such as eugenics or safety, d. Individuals who have a history of cervical surgery before pregnancy or underwent cervical cerclage surgery during pregnancy.

Data collection

The data of the study subjects were reviewed by professionally trained and qualified investigators. Clinical data related to this study were collected, including age, prenatal body mass index (BMI) (kg/m²), gestational age, prenatal hemoglobin level, gravidity, parity (primipara = 0, multipara = 1), number of abortions, conception mode (natural conception, assisted reproductive technology, conception by ovulation induction), history of full-term births (yes = 1, no = 0), history of preterm births (yes = 1, no = 0), type of twin pregnancies (monochorionic diamniotic twins = 1, dichorionic diamniotic twins = 0), scarred uterus (yes = 1, no = 0, cervical length during the second trimester (20–24 weeks), and cervical funneling (yes = 1, no = 0). Additionally, gestational vaginitis, gestational thyroid dysfunction (including hyperthyroidism or hypothyroidism), gestational diabetes, and pregnancy with hypertension (or preeclampsia) were recorded.

Statistical analysis

All data analyses were conducted using SPSS version 25.0 software, quantitative data were presented as means ± standard deviation (SD). An independent sample t-test was used for inter-group comparison. Count data were presented as numbers (%) and compared with chi-square test or rank sum test as appropriate. The correlates of spontaneous preterm birth in twin pregnancies were analyzed using logistic regression model, and a predictive model was established. A nomogram model of risk factors was constructed using R software (R3.6.3) to predict the risk of spontaneous preterm birth in twin pregnancies. Furthermore, the predictive value of risk factors was evaluated by receiver operating characteristic (ROC) curve. P-value < 0.05 indicates a statistically significant difference.

Clinical data

A total of 218 pregnant women with twin pregnancies who met the inclusion criteria were enrolled in this study (Fig. 1). According to gestational weeks of delivery, 102 cases (47.00%) were included in the preterm birth group (28 ~ 36^+ 6 weeks) and 116 cases (53.00%) were included into full-term birth group (≥ 37 weeks). As shown in Table 1, there were no significant differences between the two groups in terms of age, gravidity, parity, number of abortions, conception mode, previous preterm births, full-term births, type of twin pregnancies, scarred uterus, pregnancy with hypertension, and pregnancy with thyroid dysfunction (P > 0.05). In contrast, the BMI, prenatal hemoglobin level, and cervical length during the second trimester in women of the preterm group was significantly lower than that in the full-term group. Moreover, the preterm birth group exhibited a lower proportion of women with a cervical funnel and gestational diabetes, but a higher proportion of women with gestational vaginitis.

Table 1

Basic data analysis of twin pregnancies pregnant women
Factors	Variables	Full-term birth group (n = 116)	Preterm birth group (n = 102)	χ²/t/Z	P-Value
Age (years)		30.09 ± 3.94	29.07 ± 4.69	1.152	0.083
BMI (kg/m²)		28.29 (26.58, 30.54)	25.61 (23.37, 29.08)	4.570	0.000
Prenatal hemoglobin level (g/L)		114.96 ± 12.53	108.91 ± 15.12	2.126	0.001
Gravidity (%)	1	50 (51.50%)	47 (48.50%)	0.401	0.815
	2	32 (52.50%)	29 (47.50%)
	≥ 3	34 (56.70%)	26 (43.30%)
Parity (%)	Primipara	77 (50.70%)	75 (49.30%)	1.314	0.252
	Multipara	39 (59.10%)	27 (40.90%)
Abortion times (%)	0	73 (55.30%)	59 (44.70%)	1.766	0.414
	1	21 (44.70%)	26 (55.30%)
	≥ 2	22 (56.40%)	17 (43.60%)
Conception method	Natural	58 (49.20%)	60 (50.80%)	1.702	0.192
	IVF	58 (58.00%)	42 (42.00%)
History of full-term birth	No	77 (59.70%)	78 (50.30%)	3.282	0.194
	Yes	39 (61.30%)	24 (38.70%)
History of preterm birth	No	111 (54.70%)	92 (45.30%)	2.556	0.110
	Yes	5 (33.30%)	10 (66.70%)
Type of twin pregnancies	Monochorionic diamniotic twins	29 (45.30%)	35 (54.70%)	2.270	0.132
	Dichorionic diamniotic twins	87 (56.50%)	67 (43.50%)
Scarred uterus (%)	0	102 (54.30%)	86 (45.70%)	0.598	0.439
	≥ 1	14 (46.70%)	16 (53.30%)
Cervical length during the second trimester (mm)		35 (31, 38)	25 (15, 33)	7.560	0.000
Cervical funnel (%)	No	102 (62.90%)	64 (37.10%)	36.740	0.000
	Yes	14 (10.00%)	36 (90.00%)
Gestational vaginitis	No	79 (66.40%)	40 (33.60%)	18.271	0.000
	Yes	37 (37.40%)	62 (62.60%)
Gestational thyroid dysfunction	No	103 (51.50%)	97 (48.50%)	2.848	0.091
	Yes	13 (72.20%)	5 (27.80%)
Gestational diabetes mellitus	No	80 (49.10%)	83 (50.90%)	4.429	0.035
	Yes	36 (65.10%)	19 (34.50%)
Pregnancy induced hypertension	No	107 (53.20%)	94 (46.80%)	0.010	0.981
	Yes	9 (52.90%)	8 (47.80%)
BMI, body mass index. P < 0.05 indicates a statistically significant difference.

Correlative factors of spontaneous preterm birth in twin pregnancies

To further explore the correlates of spontaneous preterm birth in twin pregnancies, a multivariable logistic regression analysis was conducted on the six factors that exhibited significant differences between the two groups. The results of multivariable logistic regression analysis are summarized in Table 2, which indicated that three factors are statistically independent risk factors for spontaneous preterm birth in twin pregnancies, including BMI (Exp (B) = 0.887, 95% CI: 0.806 ~ 0.977), cervical length in the second trimester (Exp (B) = 0.886, 95% CI: 0.842 ~ 0.932), gestational vaginitis (Exp (B) = 0.344, 95% CI: 0.168 ~ 0.703). Whereas, cervical funnel, gestational diabetes mellitus, and prenatal hemoglobin level were not statistically significant risk factors for spontaneous preterm birth in twin pregnancies.

Table 2

Correlative factors of spontaneous preterm birth in twin pregnancies
Factors	B Value	S.E.	Wald Value	P Value	Exp(B)	95% CI
BMI	− 0.12	0.049	5.911	0.015	0.887	0.806 ~ 0.977
Cervical funnel	− 0.679	0.691	0.965	0.326	0.507	0.131 ~ 1.966
Cervical length in the second trimester	− 0.121	0.026	21.984	0.000	0.886	0.842 ~ 0.932
Gestational vaginitis	1.068	0.365	8.574	0.003	0.344	0.168 ~ 0.703
Gestational diabetes mellitus	0.818	0.430	3.619	0.057	2.266	0.976 ~ 5.263
Prenatal hemoglobin level	− 0.020	0.012	2.675	0.102	0.980	0.956 ~ 1.004
BMI, body mass index. P < 0.05 indicates a statistically significant difference.

Validation of the prediction model

A nomogram model was constructed to visualize the prediction model, and the results showed that the corresponding probability of the sum of gestational diabetes mellitus, BMI, and cervical length in the second trimester represented the risk of spontaneous preterm birth (Fig. 2A). To validate the nomogram model, the calibration curve was generated (Fig. 2B). The concordance index (C-index) for the nomogram was 0.838, the AUC value was 0.852 (95% CI: 0.799 ~ 0.905) (Fig. 2C), demonstrating that the predictive value of the combined prediction of the three correlating factors was significantly superior to that of BMI (AUC = 0.680), cervical length in the second trimester (AUC = 0.796) and gestational vaginitis (AUC = 0.644) (Table 3). Furthermore, to provide a reference for clinical decision-making, the decision curve analysis (DCA) was conducted, which suggested that the validity of the model increased when the probability of spontaneous preterm birth was ranged from 12–100% (Fig. 2D).

Table 3

Related parameters of ROC curves
Factors	AUC	Specificity	Sensitivity	SD	P Value	95% CI
BMI	0.644	0.520	0.836	0.037	0.000	0.248 ~ 0.393
Cervical length in the second trimester	0.796	0.667	0.845	0.033	0.000	0.143 ~ 0.265
Gestational vaginitis	0.644	0.681	0.608	0.038	0.000	0.571 ~ 0.718
Combined predictor	0.852	0.905	0.706	0.027	0.000	0.799 ~ 0.904
BMI, body mass index. P < 0.05 indicates a statistically significant difference.

Spontaneous preterm birth is a global issue that is closely related to maternal and infant health, especially in twin pregnancies [8]. With the enhanced capacity of our country to rescue newborns, the treatment rate of preterm infants has been increasing, and more and more preterm infants are able to survive into adulthood. However, preterm infants, particularly extremely preterm infants, are associated with a variety of long-term risks, including hearing impairment, cerebral palsy, visual impairment, and intellectual disability, which have a significant and profound effects on their social functioning and their families. Reducing the incidence of preterm birth in twin pregnancies or predicting preterm birth provides greater time and opportunities for early intervention to improve adverse pregnancy outcomes. However, the occurrence of preterm birth is not caused by a single condition. Therefore, our study conducted a retrospective analysis of clinical cases to identify the related influencing factors of spontaneous preterm birth in twin pregnancies and established a predictive model to provide a theoretical basis for the clinical prevention and treatment of twin pregnancies.

As a polypeptide hormone, corticotropin-releasing hormone (CRH) exhibits an exponential increased during pregnancy, and the transmission mechanism is initiated by the elevation of CRH, which originates in the trophoblast cells of the placenta. The placental volume in twin pregnancies is larger than that of single pregnancy, and the secretion of CRH is correspondingly increased [9]. This may be one of the significant contributors to preterm birth in twin pregnancies. Additionally, twin pregnancies result in increased uterine size, overexpansion of uterine smooth muscle, and heightened production of prostaglandins and oxytocin, all of which can easily lead to preterm birth. In this study, both monochorionic diamniotic twins and dichorionic diamniotic twins were included, and cervical length in the second trimester was identified as an independent risk factor in the preterm birth group by comparing the clinical data of spontaneous preterm and full-term birth group, consistent with the findings reported by Conde et al [10]. Our study found that for every 1-unit decrease in cervical length during the second trimester, the risk of preterm birth increases by a factor of 0.886. The ROC curve analysis indicated that the AUC value was 0.796. The Jordan index was 0.512, and the corresponding critical value was 30.50 mm. This differs from the recommendation for screening at 14 weeks in the 2017 International Society of Obstetrics and Gynecology Ultrasound Guidelines [11], which states that the cutoff value of cervical length measured by transvaginal ultrasound is 25 mm. This may be related to gestational age, as the study population consisted of twin pregnancies, given that cervical forces and uterine cavity pressures during the second trimester of twin pregnancies are greater than those in single pregnancy. Our study demonstrated that monitoring cervical length during the second trimester can effectively predict the occurrence of spontaneous preterm birth in twin pregnancies, which may provide a convenient and efficient approach in clinical practice.

In this study, gestational vaginitis comprises abnormal routine examinations of vaginal secretions, as well as abnormal leukorrhea culture, including Chlamydia and Neisseria gonorrhoeae infections. Our results showed that the proportion of gestational vaginitis in the spontaneous preterm birth group was 62.60%, significantly higher than 37.40% in the full-term birth group. Multivariable logistic analysis revealed that the incidence of spontaneous preterm birth among pregnant women with gestational vaginitis was 1.344 times higher than that in the full-term birth group. Therefore, it can be inferred that the risk of spontaneous preterm birth is higher among pregnant women with gestational vaginitis, consistent with a previous study [12]. However, there is currently no clear research elucidating the mechanism that causes vaginitis. Furthermore, the AUC value of the ROC curve for predicting spontaneous preterm birth in twin pregnancies combined with gestational vaginitis is 0.644, indicating moderate predictive efficacy. Thus, comprehensive analysis should be conducted in combination with other high-risk factor in clinical practice.

High BMI is reported to be an important risk factor for infertility, abortion, pregnancy complications, and fetal and maternal mortality [13]. However, Goldstein et al. [14] indicated that the risk of preterm birth in women with insufficient weight gain during pregnancy was 1.70 times that of those with appropriate weight increase, consistent with our results. Multivariable logistic analysis revealed that BMI (OR = 0.887, AUC = 0.644, P < 0.05) is a protective factor for spontaneous preterm birth in twin pregnancies, with a critical risk of BMI threshold of 25.02 kg/m². Strengthening weight management during pregnancy can enhance adverse perinatal outcomes by improving patient feeding behaviors, guiding rational dietary practices and nutrition, and promoting appropriate weight gain. Many scholars have suggested that low prenatal BMI and inadequate weight gain during pregnancy directly impact preterm birth [13], however, there are few studies on the BMI of pregnant women with twin pregnancies. Our study provides valuable guidance for clinical practice.

The AUC of the nomogram model in our study is 0.852 and the sensitivity and specificity of the optimal cut-off values at 70.59% and 90.52%, respectively. This indicated that the nomogram model has good predictive value. It was found that when the threshold probability was 12 ~ 100%, the validity of the model was increased, suggesting that this model is clinically useful for predicting spontaneous preterm birth in twin pregnancies. Therefore, this predictive model can assist medical staff in pregnancy management, focusing more on high-risk pregnant women, and predicting the risk of spontaneous preterm birth individually based on the circumstances of different pregnant women. Timely, targeted measures should be implemented based on the predicted results, which can effectively reduce the incidence of spontaneous preterm birth. Although this predictive model offers advantages in convenience, feasibility, and high operability, numerous correlated factors contribute to the occurrence of spontaneous preterm birth in twin pregnancies, and any single risk factor has limitations in the predictive effect.

Acknowledgements

Not applicable.

Authors’ contributions

All authors contributed to the study conception and design. Wei-Na Xu and Ling Ai designed research scheme and collected the data, Xiao-Yan Zhang and Jian-Guo Wang analyzed the data, Yi-Min Huang wrote the main draft. This manuscript was examined by all the authors.

Funding

This work was supported by Science and Technology Bureau of Jiaxing City (Grant numbers [2023AY31023]).

Availability of data and materials

The dataset generated and analyzed during the study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Jiaxing Women and Children's Hospital, Wenzhou Medical University (2024-Y-43). Written informed consent was obtained from all women enrolled in the study.

Consent for publication

Not applicable.

Competing interests

The authors have no relevant financial or non-financial interests to disclose.

Chen P, Li M, Mu Y, et al. Temporal trends and adverse perinatal outcomes of twin pregnancies at differing gestational ages: an observational study from China between 2012-2020. BMC Pregnancy Childb. 2022; 22(1): 467.
Benn P, Rebarber A. Non-invasive prenatal testing in the management of twin pregnancies. Prenat Diagn. 2021; 41(10):1233-1240.
D’Antonio F, Berghella V, Di Mascio D, et al. Role of progesterone, cerclage and pessary in preventing preterm birth in twin pregnancies: A systematic review and network meta-analysis. Eur J Obstet Gynecol Reprod Biol, 2021; 261: 166-177.
Osterman M, Hamilton B, Martin J A, et al. Births: Final Data for 2020. Natl Vital Stat Rep, 2021; 70(17): 1-50.
Roman A, Ramirez A, Fox N S. Screening for preterm birth in twin pregnancies. Am J Obstet Gynecol MFM, 2022; 4(2S): 100531.
Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet, 2008; 371(9606): 75-84.
McCormick MC. The contribution of low birth weight to infant mortality and childhood morbidity. N Engl J Med, 1985; 312(2): 82-90.
Ono T, Katsura D, Yamada K, et al. Use of ultrasound shear-wave elastography to evaluate change in cervical stiffness during pregnancy. J Obstet Gynaecol Res, 2017; 43(9): 1405-1410.
Herrera CL, Maiti K, Smith R. Preterm birth and corticotrophin-releasing hormone as a placental clock. Endocrinology, 2023; 164(2): bqac206.
Conde AA, Romero R. Predictive accuracy of changes in transvaginal sonographic cervical length over time for preterm birth: a systematic review and meta-analysis. Am J Obstet Gynecol, 2015; 213(6): 789-801．
Khalil A, Rodgers M, Baschat A, et al．ISUOG practice guidelines: role of ultrasound in twin pregnancies. Ultrasound Obstet Gynecol, 2016; 47(2): 247-263．
Fettweis JM, Serrano MG, Brooks JP, et al. The vaginal microbiome and preterm birth. Nat Med, 2019; 25(6): 1012-1021.
Langley-Evans SC, Pearce J, Ellis S. Overweight, obesity and excessive weight gain in pregnancy as risk factors for adverse pregnancy outcomes: A narrative review. J Hum Nutr Diet, 2022; 35(2): 250-264.
Goldstein RF, Abell SK, Ranasinha S, et al. Association of gestational weight gain with maternal and infant outcomes: a systematic review and meta-analysis. JAMA, 2017; 317(21): 2207-2225.

No competing interests reported.

Download PDF

Reviewers invited by journal
09 Sep, 2024
Editor invited by journal
04 Sep, 2024
Editor assigned by journal
03 Sep, 2024
Submission checks completed at journal
03 Sep, 2024
First submitted to journal
02 Sep, 2024

You are reading this latest preprint version

Nomogram model for predicting spontaneous preterm birth in twin pregnancies: a case-control study

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

INTRODUCTION

METHODS

Study population

Data collection

Statistical analysis

RESULTS

Clinical data

Correlative factors of spontaneous preterm birth in twin pregnancies

Validation of the prediction model

DISCUSSION

Declarations

References

Additional Declarations

Status:

Version 1