Exploration of Serum lipid levels during twin pregnancy

Abstract

Objective: This study aims to characterize changes in serum lipid levels throughout twin pregnancies and explore the relationship between lipid levels and gestational diabetes mellitus (GDM) and hypertensive disorders complicating pregnancy (HDCP).

Methods: We retrospectively studied 297 twin pregnancies of women who received regular prenatal care and delivered at the Beijing Obstetrics and Gynecology Hospital over a period of two years. Demographic and medical data of the participants were collected by questionnaires and medical records review. Serum lipid levels were measured in the first trimester (6–13 weeks), second trimester (24–28 weeks), and third trimester (34–37 weeks). A multivariate regression model was constructed to examine the association between lipid levels and pregnancy complications. A decision tree was used to explore the relationship between early serum lipid glucose levels and GDM and HDCP in twin pregnancies.

Results: Triglyceride (TG), total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) levels increased significantly from the first trimester to the third trimester, with the exception of high-density lipoprotein cholesterol (HDL-C), which decreased in the third trimester in twin pregnancies (p < 0.001). The levels of TC in the GDM and HDCP group were significantly elevated compared to those in the normal group in early pregnancies (p < 0.05, p < 0.05). In the second trimester, TG in the HDCP group was substantially higher than that in the normal group (p = 0.01). In the third trimester, LDL-C and HDL-C levels in the GDM group are significantly lower than that in the normal group (p < 0.05, p < 0.05). After adjusting for confounders, body mass index (BMI) is independently associated with GDM (odds ratio [OR] = 1.129, 95% confidence interval [CI]: 1.007–1.266) and HDCP(odds ratio [OR] = 1.170, 95% confidence interval [CI]: 1.031–1.329). The variation amplitude of HDL-C in the third trimester is related to the occurrence of GDM and HDCP(GDM:OR = 0.271, 95%CI: 0.095–0.778; HDCP: OR =0.249, 95% CI: 0.075–0.823). TG and TC levels in DCDA twins were significantly higher than that in MCDA twins in the first trimester(TG: p < 0.05, TC: p < 0.05). In the decision tree model for GDM, fasting blood glucose in the first trimester (FBG), TC, and pre-pregnancy BMI were identified as important nodes, while in the HDCP model, pre-pregnancy BMI, TC, and TG were key nodes.

Conclusion: Serum lipid levels in twin pregnancies increase gradually during pregnancy. BMI is independently associated with the occurrence of GDM and HDCP. HDL-C may serve as a protective factor for GDM and HDCP. The predictive effect of early blood lipid on GDM and HDCP in twin pregnancy needs further study.

Keywords:

• Twin pregnancies
• lipid level
• complications

BACKGROUND

During pregnancy, the mother experiences a series of physiological changes, including an increase in blood lipid levels that occurs as part of normal pregnancy physiology. The lipoprotein lipid physiology during pregnancy has important implications for fetal growth and development, childbirth preparation, and breastfeeding [1], however, dyslipidemia occurring during pregnancy is associated with adverse outcomes for both mother and child in singleton [2–4]. To optimize healthy birth outcomes, reducing the number of embryos transferred and increasing the use of single embryo transfer is considered, rates of twins plateaued in recent years after rising in the past few decades [5–7], but they still account for 3.3% of pregnancy in China [8]. Twin pregnancies are associated with a higher incidence of complications, including gestational diabetes mellitus (GDM), hypertensive disorder complicating pregnancy (HDCP) and premature birth, which compromise the safety of both the mother and fetuses [9,10], blood glucose and lipid level during pregnancy have a lasting impact on lipid metabolism for up to 10–14 years postpartum [11]. How to mitigate maternal complications in multiple pregnancies ranks among the top 10 research inquiries identified by the Global Twins and Multiples Priority Setting Partnership (PSP) [12]. Several studies have focused on predicting preeclampsia and GDM in twin pregnancies, the models used by these studies include a number of maternal factors such as age, body mass index (BMI), ethnicity, history of preeclampsia, hematological parameters and plasma levels of malondialdehyde and nitrates, however, none of these models considered lipids and some biochemical indicators that are not easily obtainable in clinical practice [13–15]. The Chinese Medical Association lists HDL-C < 1mmol/L and/or TG > 2.8 mmol/L as high risk factors for diabetes in the guidelines for the diagnosis and treatment of hyperglycemia during pregnancy, suggesting that pregnant women that present these high risk factors should be supported with health education and lifestyle management [16]. Energy intake increases 700 Kcal from the first to the second and third trimesters to support gestational weight gain and the rise in energy expenditure of Dichorionic Diamniotic Twins (DCDA) [17], the lipid levels of twin pregnancy are higher than those in singleton [18], recent research has found that total cholesterol (TC) and low-density lipoprotein-cholesterol (LDL-C) during early pregnancy are associated with excessive catch-up growth of dichorionic twin offspring [19]. Limited studies have been conducted on alterations in maternal total cholesterol (TC) and triglyceride levels during twin pregnancies [18,20]. However, our current comprehension and appreciation of the full scope and implications of dyslipidemia in twin pregnancies on both maternal and fetal outcomes remains incomplete. Therefore, we conducted a longitudinal lipid profile study in women carrying twins to investigate the alterations in serum lipid levels during twin pregnancies as well as their correlation with GDM and HDCP.

METHODS

Study population

We retrospectively analyzed 297 twin pregnancies in women who received regular prenatal care and delivered at the Beijing Obstetrics and Gynecology Hospital, Capital Medical University between January 2018 and December 2019. The inclusion criteria were: twin pregnancy with two viable fetuses delivered at more than 28 weeks of gestation. Exclusion criteria were: preexisting diabetes, hypertension, thyroid disease, twin pregnancies complicated by twin-to-twin transfusion syndrome and congenital fetal anomalies, monochorionic monoamniotic twin pregnancies, immune system disorders and missing data on pre-pregnancy weight, height, 75 g oral glucose tolerance test results, or lipid data. The study was approved by the Ethics Committee of the Beijing Obstetrics and Gynecology Hospital (Reference number: 2020-KY-005-02). All participants provided written informed consent.

Data collection

Baseline characteristics of patients were obtained from medical records, including maternal age, education level, height, and pre-pregnancy weight, as well as mode of conception. Medical data such as lipid concentrations during pregnancy, gestational age, neonatal birth weight, and pregnancy complications, were also extracted from the same records. Gestational age was calculated from the date of the last menstrual period (LMP) and adjusted using ultrasound measurements. Ultrasound examination was performed when the patients were confirmed as pregnant for the first time (at 6–10 weeks according to the LMP). The second ultrasound examination was performed at 11–13 6/7 weeks of gestation. If pregnancy resulted from assisted reproductive technology (ART), the ART-derived GA was used to estimate the due date. All investigators in our team underwent training prior to administering surveys to patients, and each completed questionnaire was subsequently reviewed by an inspector for verification. All compiled data were independently entered into the record system by two individuals and subsequently verified by a third party.

Clinical determination of serum lipids

Serum lipid levels were assessed in maternal brachial venous blood (after 8 h fasting) during the first trimester (6–13 weeks), second trimester (24–28 weeks), and third trimester (34–37 weeks). All samples were assayed in the institutional clinical laboratory. TC, TG, HDL-C, and LDL-C were assayed on the fully automated ARCHITECT ci16200 Integrated System Chemistry/Immunology Analyzer (Abbott Park, IL, USA) using a cholesterol assay kit (H05119R02, Abbott Park, IL, USA), triglyceride assay kit (G07893R02, Abbott Park, IL, USA), LDL assay kit (G69452R13, Abbott Park, IL, USA) and HDL assay kit ANNALS OF MEDICINE 1633(G05251R03, Abbott Park, IL, USA).

Definitions and timing of planned delivery

Gestational diabetes mellitus is diagnosed when any of the following values was exceeded in the 75-g oral glucose tolerance test (OGTT) performed between 24 and 28 weeks as recommended by the American Diabetes Association criteria: 5.1 mmol/L at 0 h (fasting); 10.0 mmol/L at 1 h; and 8.5 mmol/L at 2 h. HDCP includes gestational hypertension, preeclampsia, and eclampsia diagnosed according to the American College of Obstetricians and Gynecologists for HDCP [21]. Gestational hypertension is defined as systolic blood pressure > = 140 mm Hg or/and diastolic blood pressure > = 90 mm Hg, measured on two occasions at least 4 h apart after 20 weeks of gestation, in a woman with previously normal blood pressure. Preeclampsia is diagnosed when urine protein (> = 300 mg per 24 h urine collection, Protein/creatinine ratio > = 0.3 mg/dL or Dipstick reading of 2+) occurs on the basis of gestational hypertension or, in the absence of proteinuria, any of the following conditions: thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema or cerebral or visual symptoms. Ultrasonography is utilized for determining chorionicity in the first trimester [22], with postnatal confirmation of diagnosis. Dichorionic diamniotic twins were identified based on two separate gestational sacs at 6–10 weeks or the fetuses having different sexes, the presence of two separate placentas, or the presence of the ‘twin peak’ or ‘lambda (λ) sign, which denotes a triangular projection of tissue extending from a fused dichorionic placenta between layers of the intertwined membrane. Monochorionic diamniotic twins were identified based on one separate gestational sac at 6–10 weeks or the fetuses having the same sex, the presence of a single placenta, and the presence of an intertwin membrane with the ‘T sign, a thin intertwined membrane consisting of two amnions as they take off at an angle of 90° from the placenta. The delivery schedule utilized in our study adhered to the Chinese guidelines on the management of twin pregnancy. Uncomplicated monochorionic diamniotic twin gestations underwent delivery under close monitoring, with the aim of achieving delivery by 37 weeks of gestation whenever possible. For patients with uncomplicated dichorionic diamniotic twin gestations, we recommend termination of pregnancy after 37 weeks of gestation. In cases where patients have other complications, such as HDCP or intrahepatic cholestasis of pregnancy (ICP), the timing of planned delivery is adjusted in accordance with these comorbidities.

Data analysis

All statistical analyses were conducted using SPSS version 21.0 (IBM Corp., Armonk, NY). Numerical values are expressed as mean and standard deviation(SD) for variables with normal distribution and as the median and percentiles for nonnormally distributed data. Categorical variables are reported as percentages. Comparisons between any two groups were performed using Student’s unpaired t-test for parametric data and the Mann-Whitney rank sum test for non-parametric data. The differences between more than two groups were determined by the analysis of variance (ANOVA) for parametric data and the Kruskal-Wallis rank sum test for nonparametric data. If the ANOVA test or the Kruskal-Wallis rank sum test demonstrated a significant difference among groups, post-hoc analyses were performed by multiple comparisons. p < 0.05 was considered statistically significant. Decision trees were used to predict the risk of GDM and HDCP based on lipid levels and fasting blood glucose (FBG) in early pregnancy, age, and pre-pregnancy BMI. The rpart (Recursive Partitioning and Regression Trees) package of R (https://cran.r-project.org) was used to carry out decision tree analysis through the CART algorithm based on gini index splitting criteria. The decision tree was constructed using all the default parameters of the rpart package. Because the positive and negative samples of the original data were not uniform in size (GDM = 48, Normal = 193, total samples = 241; HDCP = 37, Normal = 193, total samples = 230), we used the ROSE (Randomly Over Sampling Examples) package in R to correct this class imbalance [23,24]. We under-sampled multiple data points and oversampled fewer data point while keeping the new sample size the same as the original sample size. We set the parameter p = 0.5 of the resample function in ROSE to keep the proportion of positive samples approximately equal to 0.5 (GDM = 175, Normal = 191; HDCP = 175, normal = 191), therefore achieving a balanced number of samples in the positive and negative groups. The modified data was used for further analyses.

Results

This study comprised a cohort of 297 gravid women carrying twin fetuses. Serum lipid profiles of all participants were completed in the first, second, and third trimester. 67 women (22.6%) were diagnosed with GDM and 56 cases (18.9%) were diagnosed with HDCP. 19 women were diagnosed both with GDM and HDCP. Clinical characteristics of the population studied are presented in Table 1 grouped by different diseases(GDM or HDCP). The BMI of the disease group (GDM or HDCP) was significantly higher than that of the non-disease group (GDM: 22.4 vs Normal: 21.5, p = 0.011; HDCP:22.4 vs Normal: 21.5, p = 0.014). Additionally, the neonatal weight differed significantly between the GDM and normal groups (p = 0.038). Preterm birth rates exhibited significant differences between the HDCP and normal groups (p = 0.047). The incidence rate of premature rupture of membranes (PROM) is lower in the GDM group compared to the normal group(p = 0.009). The early fasting glucose level, OGTT fasting, 1 h, 2 h glucose levels were significantly different between the GDM group and the normal group (p < 0.001). There were no significant differences in maternal age, education level, mode of conception [in vitro fertilization-embryo transfer (IVF-ET) or natural pregnancy (including ovulation induction)], and intrahepatic cholestasis of pregnancy (ICP).

Table 1. Maternal and pregnancy characteristics in different groups.

	GDM(n = 48)	HDCP (n = 37)	Normal (n = 193)	p*-value	p**-value
Maternal age(years)	33(31-36)	33(31-34)	32(30-35)	0.055	0.514
Pre-pregnancy BMI	22.40(20.55-24.95)	22.40(20.70-24.20)	21.50(19.30-23.05)	0.011	0.014
Education level				0.343	0.817
Postgraduate	7(14.58%)	6(16.22%)	32(16.67%)
University	29(60.42%)	22(59.46%)	93(48.44%)
Junior college	6(12.5%)	6(16.22%)	44(22.92%)
High school and under	6(12.5%)	3(8.11%)	21(11.05%)
Mode of conception
Natural pregnancy	21(43.75%)	10(27.03%)	81(41.97%)	0.871	0.101
IVF-ET	27(56.25%)	27(72.97%)	112(58.03%)
Gestational age	37.00(35.75-37.00)	36.00(35.00-37.00)	37.00(35.00-37.00)	0.569	0.223
Chorionicity
DCDA	35(72.92%)	32(86.49%)	146(75.65%)	0.867	0.198
MCDA	13(27.08%)	5(13.51%)	47(24.35%)
Neonatal weight 1	2747.5(2427.5-2967.5)	2440(2230-2625)	2605(2310-2780)	0.025	0.066
Neonatal weight 2	2582.5(2387.5-2845.0)	2480(2030-2730)	2490(2225-2730)	0.099	0.366
Premature birth	20(41.67%)	24(64.86%)	88(45.6%)	0.746	0.047
PROM	2(4.17%)	6(16.22%)	37(19.17%)	0.009	0.820
ICP	1(2.08%)	1(2.7%)	5(2.6%)	0.796	1.000
OGTT(fasting)	5.02(4.57-5.45)	4.32(4.04-4.54)	4.36(4.17-4.62)	<0.001	0.297
OGTT 1hour	10.06(9.35-10.52)	7.80(7.03-8.63)	7.53(6.64-8.27)	<0.001	0.295
OGTT 2hour	8.79(7.09-9.45)	6.49(6.13-6.99)	6.31(5.75-7.05)	<0.001	0.320
FBG^1st	4.90(4.56-5.16)	4.53(4.41-5.05)	4.61(4.38-4.89)	<0.001	0.590
Weight gain at 24weeks	7.50(5.50-10.38)	8.00(6.10-11.05)	9.00(6.88-11.05)	0.088	0.548
Weight gain throughout pregnancy	15.00(13.00-17.00)	19.00(14.75-21.25)	18.00(14.50-22.00)	0.002	0.488

*:GDM vs normal, **:HDCP vs normal; BMI: body mass index; IVF-ET: in vitro fertilization-embryo transfer; DCDA: dichorionic diamniotic; MCDA: monochorionic diamniotic; PROM: premature rupture of membranes; ICP: intrahepatic cholestasis of pregnancy , OGTT: oral glucose tolerance test; FBG^1st: fasting blood glucose in the first trimester.

Table 2. Serum lipids levels during pregnancy.

	First trimester	Second trimester	Third trimester	p-value
TC	4.37(3.94-4.89)	6.40(5.64-7.17)	6.71(6.00-7.62)	<0.001*
TG	1.26(0.94-1.72)	2.76(2.14-3.58)	3.93(3.06-5.07)	<0.001*
LDL-C	2.11(1.83-2.50)	3.14(2.50-3.78)	3.26(2.52-4.06)	<0.001*
HDL-C	1.69(1.47-1.97)	1.89(1.61-2.16)	1.65(1.40-1.89)	<0.001*

TC: cholesterol; TG: triglyceride; LDL: low-density lipoprotein; HDL: high-density lipoprotein.

Serum lipid levels increase significantly during pregnancy, except for a slight decrease in HDL-C levels during the third trimester (Table 2 and Figure 1). Serum levels of TC, TG and LDL-C increase significantly from early to middle to late pregnancy, with the greatest increase in TG. HDL-C increases from early to middle pregnancy, but decreases in late pregnancy. Table 3 shows that TG and TC levels in DCDA twins were significantly higher compared to those in MCDA twins during the first trimester (TC: p = 0.012; TG: p < 0.001).

Table 3. Comparison of Serum lipids levels between DCDA and MCDA.

		DCDA	MCDA	p-value
First trimester	TC	4.40(3.98-4.92)	4.20(3.77-4.63)	0.012*
	TG	1.31(1.05-1.79)	1.04(0.77-1.49)	<0.001*
	HDL-C	1.69(1.45-1.99)	1.70(1.49-1.89)	0.098
	LDL-C	2.13(1.85-2.58)	2.05(1.67-2.33)	0.132
Second trimester	TC	6.42(5.62-7.24)	6.37(5.76-6.85)	0.305
	TG	2.76(2.14-3.61)	2.73(2.13-3.33)	0.487
	HDL-C	1.90(1.61-2.16)	1.81(1.58-2.11)	0.381
	LDL-C	3.18(2.48-3.82)	3.09(2.69-3.63)	0.601
Third trimester	TC	6.77(6.09-7.68)	6.68(5.85-7.20)	0.090
	TG	3.99(3.07-5.14)	3.84(3.06-4.70)	0.368
	HDL-C	1.65(1.41-1.93)	1.64(1.32-1.82)	0.237
	LDL-C	3.33(2.50-4.08)	3.19(2.64-3.74)	0.502
Pre-pregnancy BMI		21.9(19.7-23.7)	21.7(19.1-23.9)	0.503
Maternal age		33(31-35)	31(29-34)	0.104
Mode of conception				<0.001*
	Natural pregnancy	49(21.3%)	60(89.6%)
	IVF-ET	181(78.7%)	7(10.4%)

The levels of TC in the GDM and HDCP group were significantly higher than that in the normal group in early gestation (p < 0.05, p < 0.05). During the second trimester, TG in the HDCP group was substantially higher than that in the normal group (p = 0.01). In the third trimester, LDL-C and HDL-C levels in the GDM group is significantly lower than that in the normal group (p < 0.05, p < 0.05) (Table 4 and Figure 2). After adjusting for confounders, body mass index (BMI) is independently associated with GDM (odds ratio [OR] = 1.129, 95% confidence interval [CI]: 1.007–1.266) and HDCP(odds ratio [OR] = 1.170, 95% confidence interval [CI]: 1.031–1.329). The variation amplitude of HDL-C in the third trimester is related to the occurrence of GDM and HDCP(GDM: OR = 0.271, 95%CI: 0.095–0.778; HDCP: OR =0.249, 95% CI: 0.075–0.823) (Table 5–7).

Figure 1. Box plots of lipid levels in three trimesters.

Figure 2. Comparison of Serum lipid levels in GDM and normal groups in three trimesters; Comparison of Serum lipid levels in HDCP and normal groups in three trimesters.

Table 4. Relationship between pregnancy complications and blood lipid levels (univariate analysis).

	GDM			HDCP
	GDM(n = 48)	Normal (n = 193)	p-value	HDCP (n = 37)	Normal (n = 193)	p-value
TC
First	4.56(4.28-4.90)	4.31(3.84-4.89)	0.017	4.58(4.24-5.11)	4.31(3.84-4.89)	0.042
Second	6.53(5.61-6.94)	6.40(5.73-7.21)	0.551	6.36(5.81-7.31)	6.40(5.73-7.21)	0.848
Third	6.65(5.68-7.47)	6.79(6.10-7.72)	0.150	6.70(6.03-7.58)	6.79(6.10-7.72)	0.748
TG
First	1.44(1.01-1.87)	1.23(0.88-1.57)	0.066	1.40(1.10-2.05)	1.23(0.88-1.57)	0.058
Second	3.02(2.19-4.06)	2.63(2.11-3.22)	0.050	3.02(2.65-3.82)	2.63(2.11-3.22)	0.010
Third	4.21(3.10-5.39)	3.84(3.03-4.81)	0.210	4.17(3.16-5.03)	3.84(3.03-4.81)	0.257
LDL-C
First	2.27(1.86-2.61)	2.07(1.81-2.40)	0.100	2.16(1.87-2.51)	2.07(1.81-2.40)	0.302
Second	2.96(2.34-3.73)	3.15(2.68-3.80)	0.174	3.23(2.31-3.70)	3.15(2.68-3.80)	0.796
Third	2.96(2.16-3.85)	3.33(2.73-4.23)	0.036	3.49(2.56-4.01)	3.33(2.73-4.23)	0.826
HDL-C
First	1.77(1.60-2.04)	1.65(1.42-1.94)	0.115	1.79(1.53-2.10)	1.65(1.42-1.94)	0.184
Second	1.81(1.58-2.08)	1.91(1.64-2.22)	0.271	1.95(1.75-2.10)	1.91(1.64-2.22)	0.893
Third	1.53(1.30-1.89)	1.67(1.46-1.93)	0.039	1.51(1.31-1.84)	1.67(1.46-1.93)	0.062

Table 5. Multivariate analysis of early blood lipids with GDM and HDCP.

	GDM				HDCP
	Beta	OR	95% CI of OR	p-value	Beta	OR	95%CI of OR	p-value
Age	0.073	1.075	(0.973-1.188)	0.154	0.028	1.028	(0.917-1.153)	0.637
BMI	0.122	1.129	(1.007-1.266)	0.037	0.157	1.170	(1.031-1.329)	0.015
Chorionicity	0.135	1.144	(0.520-2.520)	0.738	−0.572	0.564	(0.190-1.677)	0.303
TC	0.046	1.047	(0.262-4.176)	0.948	−0.708	0.492	(0.102-2.387)	0.379
TG	0.171	1.187	(0.599-2.350)	0.624	0.675	1.963	(0.951-4.052)	0.068
LDL-C	0.259	1.295	(0.274-6.132)	0.744	0.789	2.202	(0.357-13.592)	0.395
HDL-C	0.494	1.638	(0.368-7.299)	0.517	1.473	4.361	(0.751-25.321)	0.101

Table 6. Multivariate analysis of lipid changes with GDM and HDCP.

	GDM				HDCP
	Beta	OR	95% CI of OR	p-value	Beta	OR	95% CI of OR	p-value
Age	0.077	1.080	(0.975-1.195)	0.139	0.019	1.019	(0.907-1.146)	0.748
BMI	0.120	1.128	(1.006-1.265)	0.039	0.203	1.225	(1.073-1.400)	0.003
Chorionicity	−0.013	0.987	(0.442-2.203)	0.975	−0.707	0.493	(0.168-1.449)	0.199
△^aTC	−0.238	0.788	(0.333-1.867)	0.588	−0.783	0.457	(0.149-1.403)	0.171
△^aTG	0.088	1.092	(0.639-1.867)	0.748	0.343	1.409	(0.780-2.543)	0.255
△^aLDL-C	0.085	1.089	(0.421-2.814)	0.861	0.918	2.505	(0.789-7.948)	0.119
△^aHDL-C	−0.955	0.385	(0.116-1.277)	0.119	−0.473	0.623	(0.156-2.489)	0.503

△^a: difference between lipid levels in the first and in the second trimester.

Table 7. Multivariate analysis of lipid changes with GDM and HDCP.

	GDM				HDCP
	Beta	OR	95% CI of OR	p-value	Beta	OR	95% CI of OR	p-value
Age	0.065	1.067	(0.962-1.184)	0.220	0.013	1.013	(0.901-1.146)	0.831
BMI	0.119	1.127	(1.004-1.265)	0.043	0.207	1.229	(1.079-1.400)	0.002
Chorionicity	−0.173	0.841	(0.375-1.888)	0.675	−0.865	0.421	(0.144-1.449)	0.113
△^bTC	−0.185	0.831	(0.426-1.624)	0.589	−0.143	0.867	(0.394-1.403)	0.723
△^bTG	−0.065	0.937	(0.630-1.392)	0.746	−0.088	0.915	(0.554-2.543)	0.730
△^bLDL-C	−0.179	0.836	(0.429-1.630)	0.600	0.200	1.222	(0.538-7.948)	0.632
△^bHDL-C	−1.305	0.271	(0.095-0.778)	0.015	−1.391	0.249	(0.075-0.823)	0.023

△^b: difference between lipid levels in the first and in the third trimester.

We utilized a decision tree for the analysis and prediction of GDM and HDCP. We have done a 10-fold cross-validation. The predictive accuracy, sensitivity and specificity for GDM is 0.756(95% CI: 0.706–0.806), 0.725(95% CI: 0.638–0.811), 0.786(95%CI: 0.691–0.880), respectively. The predictive accuracy, sensitivity and specificity for HDCP is 0.768(95% CI: 0.717–0.819), 0.847(95%CI: 0.785–0.909), 0.696(95%CI: 0.618–0.773) respectively.

In the prediction model for GDM, when early FBG < 5.13 mmol/L, TC< 4.24 mmol/L and pre-pregnancy BMI < 20.95Kg/m², the prediction incidence of GDM is 0. For HDCP, when pre-pregnancy BMI > = 20.15Kg/m², early TC > =4.31 mmol/L, early FBG > =5mmol/L, the incidence of HDCP was as high as 91.67% (Figures 3 and 4).

Figure 3. Decision tree : Prediction of GDM.

Figure 4. Decision tree : Prediction of HDCP.

Discussion

This study portrayed distribution of blood lipids during different stages of twin pregnancy and explored the association between early lipid levels and their change with GDM and HDCP, thereby providing clinical evidence for the management of twin pregnancies. In the present study, we found that TG, TC and LDL-C levels significantly increase from the first to the third trimester, which is consistent with previous research on lipid profiles in both singleton and twin pregnancies [18]. Interestingly, we observed that HDL levels decreased in the third trimester. We could not confirm if this is a pattern in twin pregnancies and more research is needed to validate this result. A new finding is that TG and TC levels in DCDA twins were elevated during the first trimester compared to those in MCDA twins, while no significant differences were observed during the second and third trimesters. Further analysis revealed a higher incidence of IVF-ET in DCDA compared to MCDA, yet the underlying mechanism remains elusive. Luteal support is routinely given after IVF-ET, and the elevation of TG and TC in the early stage may be associated with luteum support, however, no relevant theoretical basis has been identified to date. This deserves further investigation.

GDM is currently the most common medical complication of pregnancy [25]. The prevalence of GDM depends on the screening methods, population characteristics, and diagnostic criteria. The prevalence of gestational diabetes mellitus (GDM) in our study was found to be 22.6%, which indicates a prevalence of twin pregnancies in the Chinese population that is consistent with previous reports [26,27]. In singleton, previous studies reported that early blood lipid is associated with the occurrence of GDM [2,28]. Hypertriglyceridemia at the start of pregnancy is closely related to impaired insulin action and β-cell function, increasing the risk of developing GDM at 24 to 28 weeks[29]. Our study on twin pregnancies revealed that the levels of early pregnancy cholesterol in the GDM group were higher than that in the normal group and HDL-C in the third trimester was lower in the GDM group. But we did not observe an association between early lipid levels and GDM in multivariate analysis. This suggests that the correlation between early blood lipid levels and GDM in twin pregnancies may not be as obvious as in single pregnancies. However, due to our small sample size, the relationship between early blood lipids and GDM in twins deserves further study. Our study also indicates that early blood glucose levels and OGTT blood glucose values in the GDM group were significantly elevated compared to those in the normal group, this suggests that consideration of risk factors other than maternal BMI are crucial for the detection and management of GDM in early pregnancy.

Pregnancy-induced hypertension, and especially preeclampsia, can damage multiple systems and seriously threaten the health of the mother and fetus. The pathogenesis of preeclampsia remains unclear, however, it is generally recognized that the inadequate remodeling of spiral arteries and maternal vascular endothelial dysfunction are the main pathophysiological changes [30]. Dyslipidemia during pregnancy is a risk factor for the development of preeclampsia [31], but the precise role of dyslipidemia in the pathogenesis of preeclampsia remains to be established. The accumulation of TG in endothelial cells lining the uterine spiral arteries that decrease the production of prostacyclin and result in endothelial dysfunction and increased oxidative stress is a proposed role of TG in preeclampsia [32]. In our study the incidence of hypertension among twin pregnancies was 18.9%, which is consistent with the findings reported by Yanqin(18.8%) [18], serum lipid levels in twin preeclampsia were significantly higher than those in single pregnancies [20], this may partially account for the increased prevalence of hypertensive disorders in twin pregnancies. We found higher levels of TC in the first trimester and higher TG in the second trimester in the HDCP group compared to the normal group, multivariate analysis showed a correlation between HDL changes from the first trimester to the third trimester and the occurrence of HDCP. It has been reported that the placenta is more prone to aging in the first trimester when the level of HDL decreases [33]. Apolipoprotein A1, a lipoprotein with anti-inflammatory response, is the main component of HDL. In animal models of preeclampsia, apolipoprotein A1 has been shown to reverse the inflammatory response and the placental changes induced by tumor necrosis factor α (TNF-α)[34]. In the current investigation, we did not find early differences in HDL between the HDCP and normal groups, but multivariate analysis showed a correlation between HDL changes from the first trimester to the third trimester and the occurrence of HDCP indicating that HDL-C may be a protective factor for HDCP.

Twin pregnancies are an important risk factor for GDM and HDCP, especially preeclampsia which is associated with a significant risk of maternal and perinatal morbidity. Given the burden of GDM and HDCP, early identification of high-risk women is imperative to facilitate preventive strategies and close monitoring. This study explored the potential of early maternal correlation factors in predicting GDM and HDCP in twin pregnancies by decision trees. Decision trees offer a more intuitive and greatly convenient approach for clinicians to provide clinical consultation. Previous studies have suggested that BMI was associated with the development of GDM and HDCP [35,36], our findings confirmed this, especially in the decision tree model for HDCP, BMI is the most important root node for predicting HDCP. Lipids are low-cost, routinely collected biochemistry markers. Clinical data support the association between early lipid levels and GDM, while animal experiments have also demonstrated that abnormal lipid levels precede the development of GDM [37], adding lipid parameters to the prediction model can further analyze the risk of GDM and HDCP in pregnant women with the same BMI level.

Strengths of the study are as follows. At present, there are few studies about the lipid levels in twin pregnancy women and even fewer studies on the sequential study of early, middle, and late trimesters of pregnancy. We described blood lipids at all three stages of twin pregnancy, especially in the early stages. Additionally, we used the intuitive statistical method of decision tree to predict the occurrence of GDM and HDCP in twin pregnancies. We acknowledge that our study has certain limitations. Firstly, a comprehensive classification of gestational hypertension disorders is lacking. Secondly, this is a retrospective study and the sample size was relatively small. The relationship between lipid levels and gestational metabolic diseases in twin pregnancies needs to be examined further on a larger cohort. Our next step involves conducting prospective studies with expanded sample sizes.

Conclusions

To the best of our knowledge, there is still no consensus regarding the normal range of blood lipids during pregnancy. The most common comorbidities during pregnancy are HDCP and GDM, while the incidence of complications is higher in twin pregnancies. We studied the characteristics and levels of blood lipids during twin pregnancies. Doctors carrying out consultation and predicting early pregnancy complications based on BMI, age, and early blood lipid levels, would be crucial to prevent the occurrence of pregnancy complications [38,39]. By establishing a decision tree model, we can preliminarily understand the probability of gestational comorbidities. The model we developed can help clinicians better manage patients during pregnancy. In the future, we will conduct a prospective study with an increased sample size to minimize bias and stabilize the decision tree, in order to verify the association between early lipids and gestational diabetes mellitus and gestational hypertension.

Acknowledgements

The authors thank Wenbin Xu for his statistical assistance and advice.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author, [Xiaokui Yang Email: [email protected]], upon reasonable request.

Additional information

Funding

This work was supported by the Beijing Municipal Administration of Hospitals Incubating Program [grant number PX2020057]; Beijing Municipal Administration of Hospitals Clinical Medicine Development(grant number ZYLX201830); Beijing Hospitals Authority’ Ascent Plan [grant number DFL20191401].