Genetic risk assessment of pregnancy-associated venous thromboembolism: dual utility of maternal peripheral blood in NIPT and VTE evaluation

Abstract

Objective

Venous thromboembolism (VTE) is a leading cause of maternal mortality, with pregnancy significantly increasing VTE risk due to physiological hypercoagulability. Current risk assessment methods, such as VTE scoring systems and D-dimer testing, have limitations in identifying high-risk individuals, highlighting the need for improved stratification.

Methods

Whole genome sequencing (WGS) was performed on peripheral blood samples collected from 29 pregnant women with clinically diagnosed VTE during routine non-invasive prenatal testin (NIPT). The analysis focused on a curated list of 162 thrombosis-related genes (18 high-risk, 144 moderate/low-risk), incorporating detection of pathogenic variants, copy number variations (CNVs). Genetic risk factors were compared against conventional VTE risk scores and D-dimer levels. Additionally, a control group of 74 healthy pregnant women was included to enable allele frequency analyses.

Results

Pathogenic/likely pathogenic variants were identified in 58.6% of cases (17/29), with TUBB1 and vWF as key contributors. 17 pathogenic CNVs were detected in 41.4% (12/29), involving PRSS1, C4A, etc. Allele frequency analysis highlighted 9 loci across 4 genes, indluding HLA-B, PRSS1, ACE, C4A linked to VTE susceptibility. Traditional VTE risk scores and D-dimer levels showed limited predictive ability, particularly in cases with low clinical risk scores but high genetic risk. Notably, among the 11 women with a pre-delivery VTE score of 0, 7 had genetic predispositions. Similarly, among the 15 women with low pre-delivery D-dimer levels, 9 had genetic risk, and among the 5 women with low D-dimer levels at 24 h postpartum, 3 had genetic risk. These findings collectively highlight the inability of traditional markers to capture hidden genetic risk in pregnancy-associated VTE.

Conclusions

The dual use of NIPT samples for VTE genetic assessment reduces invasive procedures and costs, offering a novel approach to optimize risk stratification, particularly for individuals with low traditional risk scores but high genetic susceptibility. These findings support integrating genetic screening into prenatal care to enable personalized prevention.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-025-08212-9.

Introduction

As maternal mortality from other causes during pregnancy has declined, venous thromboembolism (VTE), which includes deep vein thrombosis (DVT) and pulmonary embolism (PE), has become the leading cause of maternal death [1, 2]. Notably, approximately 80% of PEs originate from DVTs in the lower limbs, and many cases of DVT are asymptomatic [3]. The clinical implications are serious as PE has an alarming in-hospital mortality rate of 14%, and this rate escalates further to a substantial 30-day mortality rate in the context of perioperative PE occurrences. Consequently, there is an urgent need to proactively identify DVT in the preoperative phase to facilitate timely intervention and prevent the progression or onset of VTE, thereby reducing associated maternal risks.

The risk of thrombosis increases throughout pregnancy and is highest postpartum [4, 5]. The main reason for the increased risk of VTE in pregnant women is the hypercoagulable state, which may have evolved to protect women from bleeding due to miscarriage and childbirth. The risk of thromboembolism in women during pregnancy and the postpartum period is 4–5 times higher than when they are not pregnant [6]. 80% of thromboembolic events in pregnancy are venous, with rates ranging from 0.49 to 1.72 per 1000 pregnancies [7, 8]. Risk factors for thrombosis in pregnancy include genetic and acquired thrombophilia, a history of thrombosis, maternal age over 35 years, certain medical conditions, and various complications. Despite a low absolute risk, VTE during pregnancy is a major cause of maternal morbidity and mortality. Diagnosing, preventing, and treating pregnancy-associated VTE is particularly challenging because both fetal and maternal well-being need to be considered.

Thromboembolism is a complex genetic trait determined by multiple genetic and environmental factors. Genetic factors play a major role in VTE risk, especially in patients with unprovoked VTE. Identified genetic risk factors include protein C (PC), protein S (PS), antithrombin (AT) deficiency, factor V G1691A (FV), and prothrombin G20210A (FII) polymorphisms [9–11]. These five genetic defects, known as major risk factors for thrombosis, involve either coagulation or anticoagulation genes. Hereditary thrombosis has been shown to increase the risk of first VTE events. The identified genes are SERPINC1, PROC, PROS1, FII, FGG, FV, and ABO. In three independent case-control studies, two VTE-associated loci, TSPAN15 and SLC44A2, were identified [12]. A previous meta-analysis of 12 genome-wide association studies (GWAS) on venous thromboembolism detected 6,751,884 SNPs in its analysis, with nine loci reaching genome-wide significance-including six known loci, ABO, F2, F5, F11, FGG, PROCR, and three unknown loci [13]. De Haan et al.’s targeted sequencing of 734 hemostasis-related genes identified 62 DVT-associated variants, primarily in F5, ABO, and FGG. This study informed our selection of 162 thrombosis-related genes, validating the relevance of these loci in thrombotic risk stratification [14]. Lindström conducted a meta-analysis of multiple GWAS datasets and identified 11 novel VTE-associated genes, including C1orf198, PLEK, OSMRAS1, NUGGC/SCARA5, GRK5, MPHOSPH9, ARID4A, PLCG2, SMG6, EIF5A, and STX10. Furthermore, an additional five genes, SH2B3, SPSB1, RP11-747H7.3, RP4-737E23.2, and ERAP1 were identified via TWAS [15]. These VTE-related genes are associated with coagulation, anticoagulation, platelet function, red cell biology, and vascular regulation, with well-established biological links to thrombosis [15–18]. Therefore, genotyping known (and unknown) genetic factors for thrombosis risk in women contributes to understanding coagulopathy in pregnant women and enhances the evidence-based tools available for clinicians practicing personalized medicine.

Both current and acquired factors during pregnancy mean that risk is individual and changes over time. The accumulation of these risk factors makes it possible to forecast and rate the risk of VTE. Some authors have developed risk scoring systems to estimate the risk of thrombosis in pregnancy on an individualized basis, generally incorporating various risk factors such as personal and family medical history, surgical history, prolonged bed rest, obesity, and hormone use, with the aim of identifying individuals at high risk of developing VTE [19, 20]. However, even if a pregnant woman has a low-risk assessment result using the pregnancy thrombosis scoring system, thrombotic events can still occur. Therefore, we need to further improve the assessment of the risks of thrombosis during pregnancy.

Notably, maternal peripheral blood, a readily accessible biological sample in obstetric practice, has gained increasing attention for its dual potential in prenatal care. Among its applications, non-invasive prenatal testing (NIPT), a routine procedure involving analysis of cell-free fetal DNA in maternal peripheral blood for prenatal screening, has become a standard tool for detecting fetal chromosomal abnormalities. Beyond its established role in NIPT, maternal peripheral blood also holds promise for evaluating maternal VTE risk.

In the present study, 29 pregnant women with clinically confirmed VTE were identified from 11,235 pregnancies. Using residual peripheral blood samples originally collected for routine NIPT, we performed whole-genome sequencing to retrospectively investigate hereditary factors associated with pregnancy-related thrombosis in these cases. Our primary aim was to compare genetic risk assessment with conventional tools in identifying pregnancy-associated VTE risk.

Materials and methods

Subjects

From January to December 2022, a total of 11,253 Chinese pregnant women underwent routine VTE risk scoring and clinical blood laboratory tests at the International Peace Maternal and Child Health Hospital (Shanghai, China). The hospital-specific VTE scoring criteria are detailed in Table I. Pregnant women with suspected VTE underwent further diagnostic evaluations, including ultrasound or computed tomography angiography (CTA). Cases of clinically diagnosed VTE were confirmed based on established clinical diagnostic criteria, corroborative imaging findings, and laboratory results.

We excluded pregnant women with pre-existing conditions that could affect coagulation and those taking medications influencing hemostasis prior to VTE diagnosis. Additional exclusions included cases with incomplete follow-up due to sample collection/storage failure, missing key data, participant withdrawal, or testing errors. A total of 29 pregnant women with VTE were enrolled, with strict adherence to these criteria to minimize confounding variables and ensure comparability with the control group. The control group comprised 74 pregnant women matched for gestational age (12–16 weeks) and ethnicity, with no history of VTE or coagulation-related comorbidities.

Peripheral blood samples were collected from the pregnant women at the International Peace Maternal and Child Health Hospital between gestational weeks 12 and 16. A 5 mL sample was obtained by venipuncture into ethylenediaminetetraacetic acid (EDTA) tubes and subsequently stored at −70 °C for analysis.

Whole genome sequencing

Genomic DNA was extracted from blood samples (80–200 ng). DNA fragments were processed with AMPure XP beads (Life Sciences, Indianapolis, IN, USA) to obtain 100–300 bp fragments. Library construction included end repair, A-tailing, adapter ligation, and PCR amplification. The PCR products were thermally denatured to form single-stranded DNA, which was then recycled with DNA ligase, and the remaining linear molecules were digested with exonuclease. After constructing DNA nanospheres, 100 bp paired-end sequencing was performed on the MGISEQ-T7 platform with a minimum read depth of 40-fold, as described previously [21].

Genetic analysis and variant identification

Based on the OMIM database and pertinent literature references [9, 12, 14–18, 22–35], we compiled a list of 162 thrombosis-associated genes, categorized into 18 major high-risk genes and 144 genes associated with moderate and low risk of thrombosis. After initial variant annotation, candidate variants were identified through stringent screening. Variants were classified according to the American College of Medical Genetics and Genomics (ACMG) criteria and guidelines as pathogenic, likely pathogenic, or variants of uncertain significance (VUS) [36]. This included established pathogenic variants, novel missense changes at the same codon as known pathogenic variants, and genes implicated in thrombosis that are expected to affect RNA splicing.

Copy number variation analysis

CNVs were analyzed from whole genome sequencing data using a standardized pipeline. Raw sequencing reads were first quality-filtered to remove adapters and low-quality bases, then aligned to the human reference genome (GRCh37/hg37) using BWA-MEM. Duplicate reads were removed, and base quality recalibration was performed to optimize alignment accuracy. CNVs were called using a combination of read depth-based (CNVnator) and structural variant-focused (Delly) algorithms to ensure robust detection [37, 38]. Raw calls were filtered to retain high-confidence variants (length ≥ 1 kb, quality score ≥ 50, minor allele frequency < 1% in gnomAD/1000 Genomes Project) and annotated using ANNOVAR to identify overlaps with 162 thrombosis-related genes.

Statistical analysis

Univariate analysis was performed using the Fisher exact test, with a p-value threshold of less than 0.01 indicating statistical significance.

Allele frequency case/control association test

An allele-based association analysis was conducted to investigate the correlation between genetic heritability and the occurrence of VTE during pregnancy. We used PLINK v1.9 and SPSS 24.0 software for statistical analyses, including association studies, Hardy-Weinberg equilibrium (HWE) tests, and determination of genotype and allele frequencies among VTE patients and controls, consistent with previous studies [39].

Venous thromboembolism scoring system

To assess venous thromboembolism (VTE) risk in pregnant women, we synthesized established VTE scoring systems [40, 41] to develop a hospital-tailored tool. This system categorizes risk factors into prenatal factors and temporary factors. Each factor is assigned a weighted score, with total scores stratified as low (< 3 points), moderate (3–5 points), or high (> 5 points) risk. For detailed factor definitions and scoring criteria, refer to Table S1. Pre-delivery VTE scores of participants were collected and analyzed for risk stratification.

Clinical D-dimer assay

• dimer testing is routinely integrated into prenatal care for pregnant women, with data collected retrospectively from clinical records as part of this standard monitoring. For the experimental analysis. Blood samples were collected from pregnant women prior to delivery via venipuncture into anticoagulant tubes, centrifuged at 3000 rpm for 10 min to separate plasma, which was then stored at −20 °C. Using a commercial assay kit, 50 µL of thawed plasma was added to precoated microtiter plate wells, incubated at 37 °C for 30 min, washed three times, incubated with secondary antibody-conjugated horseradish peroxidase for 15 min, washed again, and substrate was added for color change measured at 450 nm. Quality control was ensured with provided positive and negative controls, and data were recorded and analyzed. The results of D-dimer measurements before and 24 h after delivery were selected for analysis.

Protein structure analysis

We used the AlphaFold protein structure database (https://alphafold.ebi.ac.uk/) to predict the structural and functional impacts of genetic variations on protein folding and function. We obtained the predicted three-dimensional structure of the wild-type protein from the AlphaFold database, which provides confidence scores for each residue to assess the reliability of the model. For each identified genetic variation, we analyzed its potential impact on protein folding by evaluating changes in key structural features, including residue conservation, solvent accessibility, and interactions with adjacent amino acids. These evaluations were based on the structure model built by AlphaFold.

Results

Patients

We analyzed 29 pregnant women with venous thromboembolism (VTE, cases) and 74 healthy pregnant women (controls) (Fig. 1). As summarized in Table 1, cases had a median age of 32.7 ± 4.2 years (controls: 32.4 ± 3.8 years), a BMI of 24.1 ± 2.3 kg/m² (controls: 23.8 ± 2.1 kg/m²), and comparable gestational age at sampling (14.2 ± 1.5 vs. 14.2 ± 1.4 weeks). Among cases, 13 had pulmonary embolism (PE) and 16 had deep vein thrombosis (DVT); controls had no VTE.

Fig. 1.

Flowchart of the genetic risk analysis process in pregnant women with thrombosis via WGS. This diagram outlines the systematic approach taken in our study, which included a cohort of 29 pregnant women diagnosed with VTE through ultrasound and venography. Subsequently, these participants underwent whole genome sequencing to identify and assess the underlying genetic risks associated with thrombosis

Table 1.

Characteristics of the patient cohort

Variable	Case	Control
Number of pregnant women	29	74
Maternal age	32.7 ± 4.2 years	32.4 ± 3.8 years
BMI	24.1 ± 2.3 kg/m²	23.8 ± 2.1 kg/m²
Gestational age at sampling	14.2 ± 1.5 weeks	14.2 ± 1.4 weeks
VTE sub-type distribution
PE	13	/
DVT	16	/

Probable disease-causing variants identified by WGS

The analysis focused on 162 thrombo-associated genes, encompassing 18 major high-risk and 144 medium- to low-risk genes (Fig. 2) [9, 12, 14–18, 22–35]. This approach detected 15 pathogenic or likely pathogenic variants in 13 pregnant women with VTE (44.8%), distributed across 10 distinct genes (Table 2). Notably, TUBB1 and vWF variants were most frequent, strongly suggesting their role in pregnancy-associated VTE. TUBB1 variants may disrupt endothelial microtubule stability, impeding blood flow and promoting deep vein thrombosis, while vWF variants likely alter platelet adhesion and coagulation cascade functions, increasing VTE risk. Other genes, CYP4V2, F5, NBEAL2, THPO, ANO6, STXBP2, F7, MMACHC, each carried variants in 3.4% of cases. These findings highlight how specific genetic factors contribute to the heterogeneous clinical manifestations of VTE during pregnancy.

Fig. 2.

Representation of genes associated with thromboembolism and their corresponding functions. A total of 162 genes are depicted, among which 18 major high-risk thrombo-associated genes are prominently highlighted in red color, while the remaining 144 are medium- and low-risk thrombus genes

Table 2.

Single nucleotide variants (SNV) causing disease and associated with thromboembolism

Sample ID	Gene	Transcript	Variant class	Inheritance	Genotype Type	dbSNP	Accession	Variant physical location	Gene-Phenotype Relationships
Pregnant woman 17	F5	NM_000130.5	P	AD, AR	Het	rs118203912	VCV000000645.1	c.439G > T (p.Glu147Ter)	Thrombophilia, susceptibility to, due to factor V Leiden
	CYP4V2	NM_207352.4	P	AR	Het	rs199476186	VCV000039260.10	c.253 C > T (p.Arg85Cys)	Bietti crystalline corneoretinal dystrophy
			LP		Het	-	VCV003338887.2	c.434del (p.Gly145fs)
	NBEAL2	NM_015175.3	LP	AR	Hom	rs2035871619	VCV001013472.30	c.250del (p.Leu84fs)	Gray platelet syndrome
Pregnant woman 18	TUBB1	NM_030773.4	LP	AD	Het	rs890185415	VCV001684388.1	c.726 C > G (p.Phe242Leu)	Macrothrombocytopenia, isolated, 1, autosomal dominant
Pregnant woman 21	THPO	NM_001290028.1	LP	AD	Het	-	VCV003256569.1	c.−64_−63dup	Thrombocythemia 1, Thrombocytopenia 9
Pregnant woman 27	vWF	NM_000552.5	LP	AD, AR	Het	rs1256082707	VCV001065291.3	c.6973T > A (p.Cys2325Ser)	von Willebrand disease, type 1, type 3, types 2 A, 2B, 2 M, and 2 N
	ANO6	NM_001025356.3	P	AR	Het	rs549442808	VCV002626770.2	c.1255 C > T (p.Arg419Ter)	Scott syndrome
			LP		Het	rs2547462651	VCV002798210.2	c.634-1G > A
Pregnant woman 6	STXBP2	NM_006949.4	P	AR	Het	rs2512691738	VCV002746710.2	c.297 C > G (p.Tyr99Ter)	Hemophagocytic lymphohistiocytosis, familial, 5, with or without microvillus inclusion disease
			LP		Het	rs992368305	VCV003240554.1	c.253 C > T (p.Gln85Ter)
Pregnant woman 7	vWF	NM_000552.5	LP	AD, AR	Het	rs2136385288	VCV001065285.3	c.6551G > C (p.Cys2184Ser)	von Willebrand disease, type 1, type 3, types 2 A, 2B, 2 M, and 2 N
Pregnant woman 8	F7	NR_051961.2	LP	AR	Het	rs368272420	VCV002506137.1	c.64 + 1005G > A	Myocardial infarction, decreased susceptibility to
			LP		Het	rs2503041151	VCV002683435.1	c.385T > C (p.Cys129Arg)
Pregnant woman 9	vWF	NM_000552.5	LP	AD, AR	Het	rs2136385288	VCV001065285.3	c.6551G > C (p.Cys2184Ser)	von Willebrand disease, type 1, type 3, types 2 A, 2B, 2 M, and 2 N
Pregnant woman 11	MMACHC	NM_015506.3	LP	AR	Het	rs770084300	VCV002445735.2	c.685 C > T (p.Gln229Ter)	Methylmalonic aciduria and homocystinuria, cblC type
			LP		Het	rs1553162943	VCV000556363.2	c.574del (p.Leu192fs)
Pregnant woman 13	TUBB1	NM_000552.5	LP	AD	Het	rs890185415	VCV001684388.1	c.726 C > G (p.Phe242Leu)	Macrothrombocytopenia, isolated, 1, autosomal dominant
Pregnant woman 1	TUBB1	NM_000552.5	LP	AD	Het	rs2515915876	VCV001709749.2	c.166 + 1G > C	Macrothrombocytopenia, isolated, 1, autosomal dominant

P Pathogenic, LP Likely Pathogenic, VUS Variant of Uncertain Significance

Probable disease-causing copy number variations (CNVs)

Our analysis revealed pathogenic or likely pathogenic CNVs in 12 of 29 VTE cases (41.4%), with 17 disease-associated CNVs implicating 9 thrombosis-related genes (Table 3). The most prominent CNV, detected in 13.8% of cases, involved the q34 region encompassing PRSS1, a gene encoding proteases critical for coagulation and fibrinolysis regulation, suggesting copy number alterations here disrupt hemostatic balance. The second most frequent CNV (10.3% of cases) affected the p21.33 region containing C4A, a key complement system component; given known crosstalk between complement activation and coagulation, C4A copy number changes may influence thrombotic risk via complement-mediated effects. CNVs in hematopoietic regulators SPI1 (p11.2 region, 10.3%) and GATA1 (p11.23 region, 6.9%), both involved in megakaryocyte differentiation and platelet production-further highlight multi-pathway contributions. Additional CNVs in ENG, JAK2, KLKB1, KNG1, and STXBP2 (each 3.4%) underscore the complex genetic architecture of pregnancy-associated VTE, where gene dosage effects disrupt procoagulant-anticoagulant balance. These results emphasize the importance of CNV screening in thrombophilia evaluation.

Table 3.

Copy number variations (CNV) causing disease and linked to thromboembolism

Sample	Chr	Cytoband	Type	Start	End	Length	Pathogenicity	Including VTE related genes	Gene-Phenotype Relationships
Pregnant woman 1	6	p21.33	DEL	31,994,301	32,008,600	14,300	LP	C4A	C4a deficiency
Pregnant woman 6	7	q34	DUP	142,749,201	142,754,000	4800	P	PRSS1	Pancreatitis, hereditary
Pregnant woman 7	7	q34	DUP	142,746,301	142,751,500	5200	P	PRSS1	Pancreatitis, hereditary
	11	p11.2	DEL	47,355,401	47,357,500	2100	LP	SPI1	Agammaglobulinemia 10, autosomal dominant
Pregnant woman 8	X	p11.23	DEL	48,790,401	48,791,200	800	LP	GATA1	Thrombocytopenia with or without dyserythropoietic anemia
Pregnant woman 9	9	q34.11	DUP	127,786,301	127,843,600	57,300	P	ENG	Telangiectasia, hereditary hemorrhagic, type 1
Pregnant woman 10	19	p13.2	DEL	7,643,201	7,644,400	1200	LP	STXBP2	Hemophagocytic lymph histiocytosis, familial, 5, with or without microvillus inclusion disease
Pregnant woman 17	7	q34	DUP	142,749,801	142,755,500	5700	P	PRSS1	Pancreatitis, hereditary
	X	p11.23	DEL	48,786,801	48,791,200	4400	LP	GATA1	Thrombocytopenia with or without dyserythropoietic anemia
Pregnant woman 20	11	p11.2	DEL	47,355,001	47,358,700	3700	LP	SPI1	Agammaglobulinemia 10, autosomal dominant
Pregnant woman 21	6	p21.33	DEL	31,984,701	31,994,500	9800	LP	C4A	C4a deficiency
Pregnant woman 19	7	q34	DUP	142,749,301	142,755,500	6200	P	PRSS1	Pancreatitis, hereditary
	3	q27.3	DEL	186,729,801	186,731,600	1800	LP	KNG1	Angioedema, hereditary, 6
Pregnant woman 24	11	p11.2	DEL	47,355,401	47,358,500	3100	LP	SPI1	Agammaglobulinemia 10, autosomal dominant
	4	q35.2	DEL	186,238,301	186,240,500	2200	LP	KLKB1	Fletcher factor (prekallikrein) deficiency
	9	p24.1	DEL	5,123,201	5,133,300	10,100	LP	JAK2	Thrombocythemia 3
Pregnant woman 13	6	p21.33	DEL	31,997,801	32,012,400	14,600	P	C4A	C4a deficiency

Genetic VUS variants and their impact on protein structures and functions

Our study identified missense variants in ACE, PROCR, vWF, and KNG1 exons, plus a CREB1 intronic splice variant (Table 4). The ACE V423M variant, classified as VUS by ClinVar, affects an α-helix residue (Val423) that forms hydrogen bonds with Ala419 and Leu426; this mutation may destabilize the α-helix, increase β-sheet content, and compromise protein function (Fig. 3A). In PROCR, the P145L variant disrupts hydrogen bonds between Pro145 (on a β-sheet) and Ala148/Arg179, undermining β-sheet integrity, altering protein conformation, and impairing anticoagulant properties-potentially disrupting coagulation regulation (Fig. 3B). For KNG1, the SV621-622X mutation affects a β-sheet segment (Ser621-Val622), where Ser621 bonds with Ser623/Glu624 and Val622 bonds with Gln92/Ile625; this alteration disrupts β-sheet structure, likely impairing vascular signal transduction or vasoactive molecule binding, and influencing thrombosis-related processes (Fig. 2C).

Table 4.

Variants of uncertain significance (VUS) associated with thromboembolism

Sample ID	Gene	Transcript	Variation	Pathogenicity	Classification of variation	Gene-Phenotype Relationships
Pregnant woman 1	ACE	NM_001382701.1	c.1267G > A(p.Val423Met)	VUS	Missense variant	Benign serum increases in angiotensin i converting enzyme, microvascular complications of diabetes, susceptibility to myocardial infarction, progression of SARS, hemorrhagic stroke, and renal tubular dysplasia
Pregnant woman 19	PROCR	NM_006404.5	c.433 C > T(p.Pro145Leu)	VUS	Missense variant	Endothelial protein C receptors are involved in a variety of biological processes, binding proteins that, as co-receptors, affect T cell function and immune responses as part of a co-suppressor gene program
Pregnant woman 8	vWF	NM_000552.5	c.3793G > T(p.Pro1265Gln)	VUS	Missense variant	von Willebrand disease, type 1, type 3, types 2 A, 2B, 2 M, and 2 N
Pregnant woman 13	KNG1	NM_001102416.3	c.1861_1864delCAGT(p.Ser621fs*1)	VUS	Frameshift variant	Angioedema, hereditary, 6
Pregnant woman 14	CREB1	NR_163947.1	c.100G > A	VUS	Splice donor variant	Histiocytoma, angiomatoid fibrous, somatic

Fig. 3.

Comparative analysis of genetic risk assessment for thrombosis in relation to VTE score assessment and laboratory D-dimer detection. A This figure illustrates the differences and associations among the three assessment methods, offering a comprehensive view of how genetic risk evaluation contrasts with and correlates to traditional VTE scoring and D-dimer detection approaches. PW indicates a pregnant woman. The white five-pointed star highlights instances where a pre-delivery VTE score of 0 and low D-dimer levels suggest a low risk of VTE, yet genetic risk factors for VTE are identified. The white triangle indicates cases where a postpartum VTE score of 0 and low D-dimer levels suggest a low risk of VTE, with genetic risk factors for VTE also detected. B Distribution of VTE risk stratification by pre- and post-delivery clinical assessments, contextualized with genetic thrombophilia findings. This bar chart compares four traditional risk assessment dimensions in 29 pregnant women with VTE: post-delivery D-dimer levels, pre-delivery D-dimer levels, post-delivery VTE scores, and pre-delivery VTE scores

Allele frequency case/control association test

Rare genetic variants affecting protein function may contribute to thrombotic risk. We conducted a case-control study of 29 VTE cases and 74 controls, validating Hardy-Weinberg equilibrium (HWE) in controls using PLINK v1.9 (significance threshold p < 0.05). All 9 candidate SNPs (across HLA-B, PRSS1, ACE, C4A, included in final analyses were in HWE (p > 0.05), supporting control population validity and minimizing stratification bias. Association analysis identified 56,895 alleles/genotypes with significant differences between cases and controls (p < 0.01); after filtering low/intermediate-risk variants, 9 high-risk variants (frameshift, missense, stop gain, splice) across 4 genes remained (Table 5). HLA-B missense variants (rs1065386, rs1050570, rs41546313, rs151341293, rs77665001), suggest immune-mediated thrombotic pathways. PRSS1 variants (rs796173487, rs747228052; splice/missense) may affect vascular/pancreatic functions, while the ACE splice/intron variant (rs772819042) could impact vascular regulation. The C4A frameshift variant (rs17874654) may disrupt complement-coagulation crosstalk. These variants, linked to immune function and vascular regulation, likely play critical roles in the genetic network underlying pregnancy-associated VTE.

Table 5.

Association test of allele frequencies between case and control groups

Gene	Chr	SNP	REF	ALT	P-value	Position	Allele Freq (Cases)	Allele Freq (Controls)	Consequence	Disease
HLA-B	6	rs1065386	G	C	0.008084	31,356,770	`0.310	0.101	Missense variant	Susceptibility to a variety of diseases (including abacavir hypersensitivity, drug-induced liver injury, spondyloarthropathy, Stevens-Johnson syndrome, synovitis, toxic epidermal necrolysis).
		rs1050570	T	C	0.005431	31,356,772	0.280	0.100	Missense variant
		rs41546313	G	C	0.002273	31,356,753	0.220	0.080	Missense variant
		rs151341293	T	A	0.002222	31,356,259	0.200	0.070	Missense variant
		rs77665001	G	A	0.00002452	31,355,485	0.350	0.050	Missense variant
PRSS1	7	rs796173487	A	T	0.001149	142,749,531	0.207	0.047	Splice variant & intron variant	Pancreatitis, hereditary
		rs747228052	C	G	0.001149	142,749,524	0.180	0.040	Missense variant
ACE	17	rs772819042	TCGCCAA	TTTTTGAGACGGAGTCTCGCTCTGTCGCCCAGGCGCCAA	0.00808	63,488,552	0.155	0.027	Splice variant & intron variant	Benign serum increases in angiotensin i converting enzyme, microvascular complications of diabetes, susceptibility to myocardial infarction, progression of SARS, hemorrhagic stroke, and renal tubular dysplasia
C4A	6	rs17874654	A	G	0.003575	31,996,451	0.250	0.090	Frameshift variant	C4a deficiency

Genetic risk factors bersus conventional clinical markers in DVT and PE subgroups

Among the 29 pregnant women with VTE, 16 were diagnosed with deep vein thrombosis (DVT) and 13 with pulmonary embolism (PE) (Fig. 4A). Genetic analysis revealed distinct patterns in genetic variants and copy number variations (CNVs) between the two subgroups. In the DVT group, pathogenic or likely pathogenic variants were predominantly detected in genes such as vWF, STXBP2, and ACE, with a detection rate of 56.3% (9/16). CNVs were mainly identified in PRSS1 (18.8%, 3/16) and C4A (12.5%, 2/16), involving pathways related to coagulation and complement system regulation. For the PE group, variants in TUBB1, PROCR, and F5 were more frequent, with an overall variant detection rate of 61.5% (8/13). CNVs in this subgroup were primarily found in PRSS1 (23.1%, 3/13) and GATA1 (15.4%, 2/13), which are associated with megakaryocyte differentiation and platelet function. Notably, vWF variants were more common in DVT patients (25.0%, 4/16) compared to PE patients (7.7%, 1/13), while TUBB1 variants showed a higher prevalence in PE cases (15.4%, 2/13) than in DVT cases (6.3%, 1/16). These differences suggest potential genotype-phenotype correlations between specific genetic markers and VTE subtypes in pregnant women.

Fig. 4.

Structural characteristics of specific protein domains and distribution of key amino acid residues. A This panel depicts the three-dimensional structure of ACE. The cartoon model illustrates its helical conformation, with key residue Val423 and its significant interactions highlighted in a colored stick model. These residues are distributed across the active site, which is crucial for its catalytic function. B This panel shows the three-dimensional structure model of PROCR, with key residue Pro145 and its associated critical connections prominently displayed in a colored stick model. These residues are distributed across the active site, which is essential for protein C binding. C This panel presents the three-dimensional structure of KNG1, represented with a ribbon model that reveals its loose yet orderly folding. Key amino acid residues Ser621-Val622 are highlighted using a ball-and-stick model, which is involved in enzymatic activation and receptor binding, forming specific functional sites

All 29 VTE cases underwent pre- and post-delivery VTE scoring (based on Table S1) and D-dimer testing (Fig. 4A), with genetic testing complementing these traditional risk stratification methods across antenatal and postpartum stages. For VTE scoring, 11 cases (37.9%) had pre-delivery scores of 0 (traditionally low risk), yet 7 of these harbored genetic risk factors; post-delivery, 2 cases (6.9%) maintained scores of 0, with 1 showing genetic predisposition. Maternal D-dimer levels typically rise with gestation (first trimester: 505 mg/L [IQR 292–963]; second: 730 mg/L [505–1260]; third: 1120 mg/L [818–1718]) [42]. In our cohort, 15 cases (51.7%) had pre-delivery D-dimer < 1.5 mg/L (traditionally low risk), with 9 (60%) exhibiting genetic risk, further demonstrating genetics’ role in identifying hidden hereditary risk (Fig. 4B). Conversely, among postpartum cases with traditional D-dimer positivity, genetic testing often revealed no relevant variants, indicating such risks were likely driven by transient physiological changes rather than genetic susceptibility.

Discussion

Our study advances the genetic assessment of pregnancy-associated VTE by employing WGS, a more comprehensive approach than traditional methods PCR-based methods limited to genes like FV, FII, MTHFR, and PAI-1 [43, 44]. The advent of next-generation sequencing has greatly expanded our understanding of thrombosis genetics. For instance, Eun-Ju Lee’s whole-exome sequencing study identified probable pathogenic variants in 60.9% of 64 VTE patients in America [45]. Similarly, Mohammad Athar et al. detected 19 variants in 10 genes among 29 idiopathic VTE cases, with a mutation detection rate of 76% [46]. Our WGS analysis, by contrast, identified genetic risks in 58.6% of 29 pregnant women with VTE, including 19 pathogenic or likely pathogenic variants in 9 genes and 17 pathogenic or likely pathogenic CNVs involving 10 genes. This comprehensive approach captures both single nucleotide and copy number variations, enriching our understanding of the genetic landscape underlying VTE in pregnancy.

Our findings highlight unique genetic patterns in pregnant women:TUBB1 and vWF variants with higher frequencies, may interact with pregnancy-specific physiological changes to amplify thrombotic risk. TUBB1 variants by disrupting endothelial microtubule stability, and vWF variants by altering platelet adhesion and coagulation cascades. CNV analysis revealed high frequencies of copy number alterations in key thrombosis-related genes, with PRSS1 (13.8%) and C4A (10.3%). Additionally, the identification of CNVs in SPI1 and GATA1, which regulate megakaryocyte differentiation and platelet production, underscores the multifactorial nature of pregnancy-associated VTE. These findings support enhanced coagulation monitoring for pregnant women with these variants, and integrating CNV screening into thrombophilia evaluation to refine risk stratification.

Certain genes, such as F2 and F5, are well-known for their dual associations with prohemorrhagic and prothrombotic conditions. The F5 Leiden mutation (rs6025) is a well-documented predisposing allele for thrombosis in Western populations. Its allele frequency ranges from 3% to 8% in the Western populations, was absent in our cohort, consistent with its low prevalence in Asian populations (~ 0.45%) [47, 48]. However, we identified a rare pathogenic variant in F5 (c.4949G > A, p.Ala1650Thr) in one pregnant woman, potentially a functionally significant mutation in the Chinese population [49]. Similarly, the F2 g.20210G > A variant, which is relatively common in Europeans (approximately 3%), appears to be exceedingly rare in Asian and African populations [50, 51]. These findings highlight ethnic differences in thrombotic genetic risk factors, emphasizing the need for population-specific screening: while Western-centric variants are critical in their populations, other factors may drive risk in Chinese pregnant women.

The genetic mechanisms of pregnancy-associated VTE involve complex pathways and gene interactions. Previous studies have linked HLA polymorphisms to thrombotic disorders, showing HLA-A, HLA-B, HLA-DPB1, and HLA-DRB1 influence susceptibility to chronic thromboembolic pulmonary hypertension [52]. In our study, the allele frequency test identified HLA-B as a key gene with 5 rare missense variants, suggesting its potential role in VTE susceptibility in pregnant women. The identified HLA-B missense variants (e.g., rs1065386, rs77665001) exhibited significant p-values, implicating immune-mediated thrombotic pathways.

Beyond HLA-B, we identified variants in PRSS1, ACE, and C4A that may contribute to thrombotic risk. PRSS1 variants (e.g., rs796173487) may disrupt the coagulation and fibrinolysis balance, ACE variant (rs772819042) may impact vascular regulation, and the C4A frameshift variant (rs17874654) may alter complement-coagulation crosstalk. These findings highlighte roles for immune regulation, vascular function, and complement activation. Beyond HLA-B, PRSS1 variants (e.g., rs796173487) may disrupt coagulation-fibrinolysis balance, ACE variants (rs772819042) could impact vascular regulation, and the C4A frameshift variant (rs17874654) may alter complement-coagulation crosstalk-collectively highlighting roles for immune regulation, vascular function, and complement activation.

Traditional VTE risk assessment tools have limitations in pregnancy. D-dimer levels rise physiologically with gestation, reducing specificity: while > 0.5 mg/mL indicates need for imaging in non-pregnant populations, pregnancy-associated elevations weaken this utility. Our study reinforces these limitations: 11 women with pre-delivery VTE score = 0 (traditionally low risk) included 7 with genetic thrombotic variants, and 15 with pre-delivery D-dimer < 1.5 mg/L included 9 with genetic predispositions. Postpartum, 5 women with D-dimer < 1.5 mg/L included 3 with genetic risk, underscoring genetics’ role in identifying hidden vulnerabilities. This supports integrating genetic information into risk assessment for comprehensive stratification.

In the development of VTE risk assessment tools for pregnancy, Lindqvist et al. pioneered the proposal of a pregnancy-specific VTE risk score [53]. However, its limited validation has hindered widespread adoption in clinical decision-making. Subsequently, Chau et al. introduced a novel scoring system, validated in a cohort of 1,000 patients, which applied during the first prenatal visit markedly enhanced the likelihood of high-risk individuals receiving appropriate preventive interventions. Despite this progress, these scoring systems remain primarily rooted in clinical factors and fail to adequately integrate genetic risk considerations [54]. Furthermore, fluctuations in postpartum D-dimer levels underscore the intricacies of postpartum VTE risk assessment. In our study, 5 women exhibited D-dimer levels below 1.5 mg/L at 24 h postpartum, yet 3 of these individuals carried thrombophilia-associated genetic variants. This observation indicates that even in the absence of significantly elevated D-dimer levels, genetic factors can substantially heighten postpartum VTE risk. Consequently, postpartum VTE risk assessment should incorporate genetic information rather than relying exclusively on D-dimer levels.

While WGS currently faces cost and data complexity barriers in clinical use, its ability to simultaneously detect SNVs, CNVs, and structural variants makes it invaluable for research and complex case diagnosis, offering new insights into pregnancy-associated VTE genetics. With advancing sequencing technologies driving cost reductions, its clinical prospects are improving. With advancing sequencing technologies driving cost reductions, its clinical prospects are improving. Integrating AI and machine learning will enhance analysis efficiency and accuracy, lowering interpretation costs, while its multifunctionality strengthens cost-effectiveness. Despite current limitations, WGS’s comprehensiveness, technological potential, and versatility position it as a promising tool for future practice. Future work should focus on cost reduction, workflow optimization, and developing WGS-based clinical decision systems to broaden its use in managing pregnancy-associated VTE and other complex diseases.

The dual use of maternal peripheral blood for NIPT and VTE genetic risk assessment holds unique clinical promise, though translating it to routine practice requires addressing practical, ethical, and counseling aspects. Reusing residual NIPT samples avoids additional venipuncture and reduces sequencing costs by 30–40% compared to separate testing, enhancing access in resource-limited settings. Leveraging existing WGS workflows lowers marginal costs for thrombosis gene analysis, making large-scale implementation feasible. However, it is important to acknowledge that the current study has not validated the identified genetic markers against specific clinical outcomes, which limits the definitive assessment of their predictive utility in clinical practice. Our study’s secondary use of NIPT samples followed ethical approval, with safeguards including no participant burden, strict de-identification, and research-only findings. Future clinical use will require updated consent processes explicitly including VTE screening, ensuring participants understand the scope and implications of additional findings, alongside dedicated studies to validate genetic risk associations with clinical outcomes to strengthen the translational value of this approach.

Conclusions

In summary, this study advances our understanding of the genetic underpinnings of VTE during pregnancy and highlights the potential of genetic assessment to complement traditional risk evaluation methods. Our findings suggest that genetic screening may add value to VTE risk stratification in pregnancy, particularly in identifying high-risk individuals who might be missed by conventional scoring systems or D-dimer testing. However, the clinical applicability of such genetic screening requires further validation in larger, prospective cohorts to confirm its predictive utility and establish standardized protocols for integration into routine care.

Abbreviations

VTE

Venous Thromboembolism

DVT

Deep Vein Thrombosis

PE

Pulmonary Embolism

NIPT

Non-invasive Prenatal Testing

WGS

Whole Genome Sequencing

CNVs

Copy Number Variations

SNVs

Single Nucleotide Variants

Authors’ contributions

Y.W., J.Z., and R.H. conceived and designed the experiments. X.H., L.Y., and L.C. performed the experiments. R.H., X.H., W.X., J.Z., and L.Z. collected the data. R.H. and Z.Y. analyzed the data and wrote the manuscript. All authors read and approved the manuscript.

Funding

This work is supported by the National Natural Science Foundation of China (81971401), and (No.22Y11902300).

Data availability

Data generated and analyzed during this study can be found within the published article. Additional data are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was submitted to and approved by the Ethics Committee of the International Peace Maternal and Child Health Hospital (IPMCH), China Academy of Welfare, with the reference number: (GKLW) 2021-19.

This research involved a retrospective analysis utilizing blood samples that had been previously collected for NIPT. Given that the original NIPT samples were obtained with informed consent from pregnant women for the specific purpose of prenatal screening, and considering that this study represents a secondary analysis focused on assessing genetic risk factors for thrombosis in these same pregnant women, no additional consent was required from the women or their guardians for this particular research project.

This study was conducted in strict adherence to the principles outlined in the World Medical Association’s Declaration of Helsinki (https://www.wma.net/policies-post/wma-declaration-of-helsinki/) for medical research involving human subjects and human data.

Consent for publication

Not Applicable. This manuscript only includes de-identified clinical test indicators of participants (without any personal identifiers or details that may compromise anonymity). Therefore, consent for publication from participants is not required.

All authors consent to the publication of this manuscript. Written informed consent was obtained from all participants prior to their inclusion in the study. The authors confirm that this work is original, has not been published elsewhere, and is not under consideration by another journal. If accepted, the article will not be published elsewhere in the same form without the copyright holder’s written consent.

Competing interests

The authors declare no competing interests.