Genetic and biological determinants of pulmonary embolism: Insights from Mendelian randomization studies

Abstract

Pulmonary embolism (PE) is a common and potentially fatal thromboembolic disease contributing to a major global public health burden. Its pathogenesis involves multiple hemodynamic, inflammatory, metabolic, and genetic factors. The multifactorial nature of PE makes it difficult to infer the direct effects of risk factors using conventional statistical approaches because of potential confounding and reverse causality. Mendelian randomization (MR) uses genetic variants as instrumental variables to strengthen causal inference. This narrative review synthesizes the available MR literature concerning potential causative factors of PE, with particular emphasis on genetic and biological pathways. MR evidence suggests that matrix metalloproteinases (MMP)-19 may be associated with increased PE susceptibility, whereas MMP-12 may be associated with decreased susceptibility. Impaired kidney function showed a positive association with PE risk, and reduced HLA-DR⁺ NK cell traits were linked to PE pathogenesis. Among gut microbiota, Clostridium innocuum was associated with increased PE risk, whereas Butyricicoccus and Actinobacteria showed protective associations. No consistent causal associations were identified for type 2 diabetes, atrial fibrillation, or epigenetic age acceleration. Larger multiethnic studies integrating multi-omics data are needed to clarify mechanisms underlying PE pathogenesis.

Key Words: Pulmonary embolism; Mendelian randomization; Matrix metalloproteinases; Immune dysregulation; Renal dysfunction; Genetic epidemiology; Thrombosis; Risk factors

Core Tip: Mendelian randomization evidence indicates that matrix metalloproteinases (MMP)-19 may be associated with increased risk of pulmonary embolism (PE), while MMP-12 may exhibit an inverse association. Reduced kidney function appears to have a potential genetic association with PE susceptibility. Immune pathways involving HLA-DR-positive natural killer cell traits and gut microbiota-related factors, such as Clostridium innocuum, Butyricicoccus, and Actinobacteria, may also contribute to PE susceptibility. In contrast, no consistent associations were observed for type 2 diabetes mellitus, atrial fibrillation, or epigenetic age acceleration. These findings underscore emerging biological pathways in PE; however, they should be interpreted cautiously, as they are hypothesis-generating and require further validation.

INTRODUCTION

Pulmonary embolism (PE) is a significant cause of morbidity and mortality worldwide. PE occurrence in the United States is estimated to be about 370000 cases annually and the mortality rate of PE is estimated to be between 60000-100000 deaths annually[1,2]. Due to the lack of specific symptoms, PE is a challenging disease to detect, and fewer than 10 percent of patients are ultimately diagnosed with PE, which implies that the actual burden might be even greater than the figures given[3].

The Virchow triad is a combination of three fundamental factors that contribute to thrombus development: Venous stasis, hypercoagulability, and endothelial damage. Venous thromboembolism (VTE) occurs due to the interaction of inherited genetic mutations and environmental factors acquired throughout life. These constitute non-modifiable risk factors such as aging, thrombophilias, and family history of VTE, and temporary risk factors including immobilization, cancer, the use of estrogen-containing contraceptives, pregnancy, and the postpartum state[3,4]. Despite epidemiological studies identifying risk factors that predispose to PE, a clear cause-and-effect relationship between the risk factors has not been established due to the multifactorial nature of these risk factors, the inability to determine a clear temporal relationship between the risk factors and the onset of PE, the presence of confounding variables, reverse causality, as well as the heterogeneity of the various factors contributing to it.

Mendelian randomization (MR) is a type of causal inference method that makes use of genetic data to answer questions regarding the impact of exposure on various outcomes. Genetic variations occur before the onset of the disease and hence MR can eliminate the issue of reverse causality, which is a major limitation in determining causal association in epidemiological studies[5]. Since the principles of MR are founded on the laws of Mendelian inheritance, where alleles are randomly distributed during meiosis, this method approximates the design of randomized controlled trials in which genetic variants are instrumental variables[6]. Genetic variants serve as instrumental variables that reduce confounding effects and improve the validity of causal inferences[7]. MR allows causal inferences in the face of unobserved confounding by eliminating bias due to confounding variables, thereby reducing the impact of unmeasured confounding[8].

It is not fully understood which mechanisms specifically lead to PE. The present review aims to synthesize the available literature on MR studies concerning the causative factors of PE, with particular emphasis on the potential role of genetic and biological pathways.

REVIEW METHODS

We conducted a structured but non-systematic literature review using PubMed/MEDLINE, EMBASE, Scopus, and Google Scholar from database inception through July 2025. Studies evaluating genetic and biological determinants of PE using MR approaches were identified using the search terms “pulmonary embolism”, “venous thromboembolism”, “deep vein thrombosis”, “Mendelian randomization”, “genetics”, “biomarkers”, and “risk factors”, combined using Boolean operators (AND/OR). Reference lists of relevant articles were also screened to identify additional studies. Studies were included if they were original human investigations employing MR methodology to evaluate potential determinants of PE or closely related VTE outcomes. Studies published in languages other than English, as well as editorials, conference proceedings, duplicate publications, and studies not directly relevant to PE or not using MR-based approaches, were excluded. Given the narrative nature of this review, studies were selected to provide a focused synthesis of the literature. The factors discussed were identified based on recurring themes observed across eligible studies and were prioritized for their biological plausibility, consistency of findings, and potential clinical or translational relevance. Additional studies identified during the search are acknowledged separately but not discussed in equal detail to maintain clarity and avoid redundancy. Findings were synthesized qualitatively due to heterogeneity in exposures, instrumental variables, datasets, and analytic methods across studies. A PRISMA-style flow diagram was constructed to enhance transparency in study identification and selection; however, the review was conducted as a narrative (non-systematic) synthesis (Figure 1).

Figure 1 A PRISMA-style flow diagram summarizing the identification, screening, and selection of studies included in this narrative review. A total of 232 records were identified through database searching, of which 200 remained after duplicate removal. Following title and abstract screening, 52 full-text articles were assessed for eligibility. Of these, 41 articles were excluded based on predefined criteria, including lack of relevance to pulmonary embolism, non-mendelian randomization methodology, duplicate records, and non-original publications. Ultimately, 11 studies were included in the qualitative synthesis. MR: Mendelian randomization.

PATHOPHYSIOLOGY OF PE

PE arises from the separation and migration of thrombi from the venous system into the pulmonary arterial system, resulting in a combination of cardiovascular and respiratory impairment which may be fatal if untreated[9-11]. Virchow described three classic components of thrombus formation: Endothelial damage, blood flow stasis, and a predisposition to hypercoagulability[12,13]. Pathological changes in the endothelial lining caused by chemical, inflammatory, or traumatic injuries can damage the endothelial layer, depriving it of its antithrombotic functions, releasing tissue factor and activating the coagulation cascade and platelet aggregation[9]. Stasis of blood occurs during prolonged immobilization and conditions like heart failure and, in these cases, clotting factors accumulate and organize in regions of reduced circulation[12,14]. The MR risk factors reviewed in this paper can be conceptually mapped onto these three arms of Virchow’s triad (Figure 1).

Hypercoagulable states, which may be inherited or acquired, may predispose a person to thrombosis. These states can be due to specific gene mutations, malignancy, or physiological changes that take place during pregnancy[15]. There is growing evidence supporting the role of chronic inflammation and immune activation in the pathogenesis of PE. Endothelial injury may be aggravated, and the production of clot-forming substances may be promoted, by interleukin-6 and tumor necrosis factor-α[16,17]. Neutrophil extracellular traps (NETs) also provide a framework on which platelets may adhere and fibrin may be deposited, consequently connecting immune responses with thrombosis[18]. Chronic metabolic conditions, such as obesity and type 2 diabetes mellitus (T2DM), are associated with persistent inflammation and endothelial dysfunction and are linked to increased oxidative stress and altered signaling molecules such as adipokines[16,17].

Lastly, genetic background is also a significant factor in the pathophysiology of PE. Genome-wide association (GWAS) studies have identified a broad spectrum of PE risk loci, indicating that interactions between genetic predisposition and environmental or clinical triggers (such as surgery or hormonal therapy) contribute to the etiology of PE[15]. The application of polygenic risk scores to predict PE risk is becoming more common, with the analysis of multiple genetic variants, although such predictions require validation across diverse populations. MR studies also assist in determining causal associations since genetic variants are used as instruments to distinguish causal association from confounding. MR results indicate associations between PE and metabolic variables such as obesity, abnormal lipid profiles, and inflammatory markers[15]. These findings should be interpreted cautiously, as estimates can be affected by confounding and limited statistical power. Ongoing research is exploring targeted therapies that address inflammatory and metabolic pathways, such as immunomodulatory and personalized anticoagulation strategies[16,18].

ADVANTAGES OF MR OVER OBSERVATIONAL STUDIES

MR produces better results than observational research for identifying PE causal factors. The MR method employs genetic variants as instrumental variables to establish exposure-outcome causality between changeable factors and their effects, producing results similar to randomized controlled trials because alleles randomly distribute during meiosis. This research design helps minimize confounding factors and reverse causation, which are the main issues that affect observational study results[6-8].

The MR method uses SNPs from GWAS to develop exposure variables which help scientists study how these genetic markers affect the risk of developing PE. The research method of MR eliminates two major study limitations which traditional studies face because it removes the impact of lifestyle factors such as smoking and diet on study results and prevents reverse causation through its use of genetic variants which become established at birth. The technique has various restrictions. Results become biased when pleiotropy occurs because a single variant produces effects on multiple traits. Weak instrument bias occurs because genetic associations between variants and traits remain at low levels. Estimates become incorrect when two-sample MR studies share common participants. Detection of biases becomes possible through statistical methods which include inverse variance weighted, MR-Egger regression, weighted median MR, and MR-PRESSO[8]. Genetic data analysis through MR enables researchers to determine the actual causes of biological and metabolic factors which contribute to PE while delivering more reliable results than conventional observational studies.

GENETIC AND BIOCHEMICAL CONTRIBUTORS TO PE: INSIGHTS FROM MR ANALYSES

Recent MR investigations have highlighted the roles of genetic and biochemical elements in PE development. This section synthesizes evidence on some of key factors, which include matrix metalloproteinases (MMP), epigenetic age acceleration, gut microbial imbalances, kidney function decline, immune cell alterations, T2DM, atrial fibrillation (AF), and obesity-related metabolic features. Of note, MMPs were included in this review as part of the broader extracellular matrix remodeling and vascular inflammatory pathways that may contribute to endothelial injury and thromboinflammation in PE.

MMPs

MMP are a group of enzymes that need zinc as co-factor. They control the turnover of the extracellular matrix and are very important for keeping the structure and integrity of blood vessels[17,18]. These enzymes also play a role in angiogenesis and vascular remodeling-processes that maintain endothelial integrity[18]. Changes in MMP activity have been linked to arterial stiffening and aneurysmal formation. These vascular structural changes predispose to thrombus formation. A study done by Mastenbroek et al[18] revealed that different MMPs affect thrombogenesis at different stages, with MMP-1 and MMP-2 activating platelets and enhancing thrombus formation, while MMP-9 and MMP-14 inhibit these processes[19].

MR evidence suggests that MMP-19 may be associated with an increased risk of PE, whereas MMP-12 may show an inverse association[5]. These findings highlight a potential role of MMP pathways in PE pathophysiology. However, these associations should be considered hypothesis generating, and further MR and experimental studies are needed to clarify their biological relevance and potential clinical implications.

Epigenetic age acceleration

Epigenetic age acceleration quantifies the disparity between a person’s biological age, assessed through DNA methylation patterns, and their chronological age, indicating expedited cellular deterioration. With aging, oxidative damage and immune activation attenuates blood vessels, put more stress on the endothelium, and raise the risk of clotting. Studies support that aging leads to higher levels of clotting factors and that older people are more likely to develop deep vein thrombosis (DVT) and PE[20,21]. Nevertheless, MR analysis did not identify causal link between epigenetic age acceleration and PE[22]. This suggests that, notwithstanding possible intersections with PE risk factors such as obesity or vascular strain, epigenetic aging does not directly precipitate the onset of PE.

Gut microbiota dysbiosis

The gut microbiome helps with digestion and metabolism, makes antimicrobial substances, and does a lot of other things[23]. The gut microbiome influences systemic inflammation and coagulation by producing substances like short-chain fatty acids and trimethylamine N-oxide[24]. Dysbiosis, or an imbalance in microbial populations, has been linked to heart disease and disorders related to blood clots.

MR evidence suggests that elevated levels of Clostridium innocuum and Coprococcus 1 are associated with an increased risk of PE, whereas Butyricicoccus, Bacteroides, and the phylum Actinobacteria are identified as factors that potentially have protective association[24,25]. Additional studies indicated that the genus Butyricicoccus reduced the risk of both lower extremity DVT associated with PE, whereas the genus Clostridium innocuum increased the risk of both conditions[25].

Although no direct causal link has been established between Butyricicoccus and PE, emerging research suggests potential protective effects. Studies indicate that Butyricicoccus is negatively correlated with obesity, dyslipidemia, and diabetes, which are known risk factors for PE[26,27]. Additionally, Butyricicoccus may reduce inflammatory responses and lipopolysaccharide levels, both of which are implicated in PE pathogenesis[28]. These findings support the hypothesis that Butyricicoccus could contribute to a decreased likelihood of PE, although further research is needed to substantiate this relationship.

Renal dysfunction (estimated glomerular filtration rate decline)

MR analyses demonstrate a significant causal association between declining estimated glomerular filtration rate and increased PE risk[28]. This is in line with what we already know about the link between chronic kidney disease and a higher risk of VTE[29,30]. The research conducted by Wattanakit et al[29] indicated that individuals with stage III and IV chronic kidney disease had a 28% elevated risk of VTE in comparison to those with normal kidney function.

There are a few biological processes that could explain this link. In individuals with diminished renal function, pre-coagulation is initiated, which may expedite the onset of VTE. Reduced kidney function is associated with elevated levels of coagulation factors, such as D-dimer, fibrinogen, and factors VII, VIII, and von Willebrand factor[31]. On the other hand, endogenous anticoagulants like antithrombin are reduced because they are lost in urine more than they are made[32].

Circulating blood cells and immune dysregulation

The immune system and hemostasis are regulated by circulating blood cells. Local hemostasis occurs when platelets, leukocytes, and endothelial cells activate coagulation factors at the site of injury. Pathological thrombosis occurs when there is inappropriate coagulation in the absence of vascular injury or uncontrolled coagulation at sites of injury.

In the pulmonary vasculature, thrombi may develop locally or as a result of extension of a clot from the deep veins. Neutrophils make an important contribution to venous thrombosis by releasing NET, which are direct triggers of coagulation. This process connects immunity and hemostasis and is known as immunothrombosis[33]. Host defense mechanisms include NETs, which participate in combating infection, but when they are dysregulated, they enhance sterile inflammation, autoimmunity, and increase the risk of developing arterial and venous thrombosis[34].

MR and reverse MR using single-variable analyses suggests that neutrophil levels and PE risk are negatively associated, but multivariate MR studies have demonstrated no significant impact[35]. Variability in analytical approaches may explain these observed differences, rather than a true underlying biological association. Although the extent of the actions of monocytes in coagulation is still being studied, their best-established functions are the initiation of coagulation through the expression of tissue factor and the amplification of thromboinflammation through activation of inflammasomes[36]. MR studies indicate that reduction in a subclass of lymphocytes known as HLA-DR⁺ NK cells is linked with an increased risk of PE, which suggests a potential role in protecting against excessive clot formation[35].

HLA-DR⁺ NK cells are a distinct hybrid population of natural killer cells and dendritic cells and are important in the regulation of immune responses[37]. To date, no studies have directly investigated the role of HLA-DR⁺ NK cells in PE or their relationship with thrombosis, despite their recognized immunological significance. The way in which HLA-DR⁺ NK cells could contribute to the development of PE represents an important potential area for future investigation.

T2DM

T2DM is characterized by hyperglycemia, insulin resistance, and metabolic syndrome. The effects of hyperglycemia in T2DM have been studied extensively and are known to promote increased coagulation activity and impaired fibrinolysis, particularly in patients with poor glycemic control. Previous studies have shown that patients with T2DM exhibit increased thrombin generation in platelet-rich plasma, as well as elevated thrombin levels[38]. In another cohort of inadequately managed T2DM, a notable increase in the concentration of plasminogen activator inhibitor-1 (PAI-1) was observed, suggesting impaired fibrinolysis[39]. PAI-1 levels were also reduced following treatment with glucose-lowering medications, indicating that the change is probably related to glycemic control rather than the specific pharmacological intervention[40].

Increased prothrombotic activity, including tissue factor procoagulant activity, plasma factor VII, factor VIII, and thrombin-antithrombin complexes, has also been associated with hyperinsulinemia. According to a study conducted by Boden et al[40], insulin levels increased in parallel with TF-PCA, and the rise was even stronger in patients with both hyperglycemia and hyperinsulinemia[41]. These findings support the concept that patients with T2DM are more likely to experience thrombotic events due to the potential synergistic effects of these metabolic disturbances.

The relationship between T2DM and VTE, such as PE, remains less conclusive. According to existing evidence, T2DM patients may have a higher risk of developing PE or VTE, although other studies have not consistently confirmed this association[42]. Furthermore, T2DM or glycemic traits did not identify a consistent causal association with PE in MR studies by Yang et al[42].

AF

The relationship between PE and AF has been of great interest due to the possibility of a bidirectional interaction involving right-sided pressure overload and the release of inflammatory cytokines. PE may lead to AF by placing strain on the right heart or by provoking systemic inflammatory responses, while AF may, conversely, predispose to PE through facilitating thrombus formation in the right atrial appendage[43,44].

Similar observations have been reported in proinflammatory states such as sepsis, where AF is found in 8%-23% of patients and VTE occurs in approximately one-third of patients, suggesting a potential overlap in underlying inflammatory mechanisms[45,46]. Other inflammatory conditions, including heart failure, autoimmune diseases, and metabolic disorders such as obesity and metabolic syndrome, have shown similar trends with increased incidences of both AF and PE[44].

Observational studies have reported a higher risk of PE in the first six months following AF diagnosis, with a crude incidence of 18.5 cases per 1000 person-years[47]. However, an MR study using GWAS data did not support a causal association between AF and PE[48]. Thus, despite observational data indicating an association, genetic evidence does not support AF as a direct causal factor in the development of PE.

Obesity and metabolic traits (body mass index, blood metabolites)

Obesity is defined as excess body fat normally determined through body mass index (BMI). Central obesity is closely related to coronary heart disease and metabolic diseases, including dyslipidemia and glucose intolerance[49].

There has been a large amount of research demonstrating the strong correlation between obesity and idiopathic VTE, such as DVT and PE, independent of other known risk factors[42]. The impact of obesity on thrombosis and the coagulation cascade is very pronounced. PAI-1 levels are higher in obese patients than in their non-obese counterparts, and they drop considerably in obese patients who undergo weight loss[50]. Patients with VTE also have high levels of PAI-1, indicating that the increase in PAI-1 due to obesity may place individuals at risk of microvascular and macrovascular thrombosis.

Increased concentrations of prothrombotic factors such as tissue factor, factor VII, factor VIII, prothrombin, and fibrinogen are also linked to obesity[51]. It is a prothrombotic state characterized by elevated levels of proinflammatory cytokines such as interleukin-6 and tumor necrosis factor-alpha, and reductions in these markers are observed following weight loss[52,53]. This inflammatory environment is further exacerbated by elevated oxidative stress, driven by increased production of reactive oxygen species, including F2-isoprostanes[54].

This oxidative stress promotes endothelial dysfunction, platelet aggregation, and thrombus formation and impairs antioxidant defenses in obese individuals[51]. Besides, when excess adipose tissue is present, there is an increase in non-esterified fatty acids, increasing triglycerides, very-low-density lipoprotein, and small, dense low-density lipoprotein alongside a reduced high-density lipoprotein, indicating that obesity is a state characterized by increased atherogenicity[51].

Dyslipidemia results in platelet hyperactivity, hypercoagulability, and hypofibrinolysis, which facilitate thrombosis[51]. MR evidence suggests a possible association between BMI and thrombotic risk, although results are not consistent across reported studies[55-58]. These findings reflect the complex role of obesity in thrombotic risk through inflammatory, oxidative, and lipid-mediated pathways. Figure 2 summarizes these mechanisms within the framework of Virchow’s triad.

Figure 2 Virchow’s triad as a conceptual framework for Mendelian randomization-supported evidence and related biological pathways in pulmonary embolism. Each circle represents one arm of Virchow’s triad. Risk factors identified in Mendelian randomization (MR) studies, as summarized in Table 1, are mapped to their predominant mechanistic domain: Hypercoagulability includes epidermal growth factor receptor decline[29] and glycemic traits[43]; endothelial injury includes matrix metalloproteinases (MMP)-19/MMP-12[5] and epigenetic age acceleration[22]; venous stasis includes atrial fibrillation[48] and HLA-DR⁺ NK cells[35]; gut dysbiosis[25] spans hypercoagulability and endothelial injury; and obesity/body mass index[55] spans all three arms, reflecting its multifactorial prothrombotic effects. Notably, MR analyses did not support a causal association for type 2 diabetes[43], atrial fibrillation[48], or epigenetic age acceleration[22] with pulmonary embolism. BMI: Body mass index; MMP: Matrix metalloproteinases; eGFR: Estimated glomerular filtration rate.

Table 1 summarizes MR-derived genetic, metabolic, and immunologic determinants associated with PE risk. Table 2 outlines established and emerging pathophysiological mechanisms of PE and their potential clinical implications, with emphasis on mechanisms supported or informed by MR evidence, as summarized in Table 1.

Table 1 Summary of causative factors for pulmonary embolism: Insights from Mendelian randomization studies.

Ref.	Population	Sample size	Data sources	Exposure(s)	Methods	Results (effect estimates)	Main interpretation
Gong et al[5], 2025	Predominantly European ancestry GWAS (United Kingdom biobank-based; includes icelandic cohort)	500000	United Kingdom Biobank, GWAS summary statistics	MMP	Bidirectional two-sample MR (IVW, MR-PRESSO, sensitivity analyses)	MMP19 associated with increased PE risk (OR = 1.0009, P = 0.041); MMP12 associated with decreased PE risk (OR = 0.9992, P = 0.038)	MMP19 may be associated with increased PE risk; MMP12 may show an inverse association
Tong et al[21], 2024	Predominantly European ancestry GWAS (EAA n = 34467; PE 3940 cases, 480658 controls)	515000	Large-scale GWAS datasets	EAA	Bidirectional two-sample MR (IVW, MR-egger, weighted mode)	No consistent association between EAA and PE	EAA does not appear to be associated with PE risk
Xu et al[24], 2024	Predominantly European ancestry GWAS (MiBioGen, IEU Open GWAS)	378000	MiBioGen, IEU Open GWAS	Gut microbiota traits	Bidirectional two-sample MR (IVW, MR-Egger, weighted median)	Butyricicoccus associated with decreased PE risk; Clostridium innocuum associated with increased PE risk	Gut microbiota traits may be associated with PE susceptibility
Lyu et al[28], 2024	Predominantly European ancestry cohorts (CURES, CHARLS, FinnGen)	17547	CURES, CHARLS, FinnGen GWAS	Renal function (eGFR)	Nested case-control study with MR analysis	Reduced eGFR associated with increased PE risk (OR = 4.26, P < 0.001)	Renal dysfunction may be associated with increased PE risk
Jiang et al[34], 2024	Predominantly European ancestry GWAS (blood cell consortium, FinnGen)	500000	Blood cell consortium, FinnGen	Blood cell traits	Single-variable and multivariable MR (IVW, MR-egger)	Lower lymphocyte count associated with increased PE risk (OR = 0.84, P = 0.0139)	Immune cell traits may be associated with PE risk
Yang et al[42], 2024	Predominantly European ancestry GWAS (IEU Open GWAS database)	600000	IEU Open GWAS	Glycemic traits (T2DM, FG, FI, GH)	Two-sample MR (IVW, MR-PRESSO, sensitivity analyses)	No consistent association between glycemic traits and PE	No strong genetic evidence linking glycemic traits with PE
Liu et al[47], 2024	Predominantly European ancestry GWAS (AF and PE datasets)	1000000	Large European GWAS datasets	AF	Two-sample MR (IVW, MR-egger, MR-PRESSO)	No causal association between AF and PE	AF is unlikely to be a causal factor for PE
Wei et al[54], 2023	Predominantly European ancestry GWAS	1500000	Multiple European GWAS sources	BMI, smoking, HF, alcohol intake, IBD	Multivariable MR (IVW, Wald ratios, Cochran’s Q test)	BMI shows weak association with PE risk (OR = 1.002, P = 0.039); no consistent association for other exposures	BMI may have a weak association with PE risk; overall evidence is inconsistent
Feng et al[56], 2022	Predominantly European ancestry GWAS (United Kingdom biobank-based)	500000	United Kingdom Biobank GWAS, GEO	Blood metabolites	LDSC, MR, transcriptomic analysis	Hydroxytryptophan associated with increased PE risk; LIPC and NAT2 genes linked to PE	Metabolic pathways may be associated with PE risk
Cen et al[57], 2025	Predominantly European ancestry GWAS (IEU Open GWAS; MiBioGen)	360000 participants	IEU Open GWAS, MiBioGen	Gut microbiota and PE	Two-sample MR (IVW, pleiotropy/heterogeneity tests)	Several genera (e.g., Bacteroidetes, Oscillospira) associated with reduced PE risk	Gut microbiota composition may influence PE susceptibility
Zhang et al[58], 2023	Han Chinese GWAS cohorts	18000 participants	GWAS discovery + replication cohorts	Genetic susceptibility loci and PE	GWAS + MR + PRS analysis	FABP2 Locus identified; LDL-C and total cholesterol causally linked to PE	Lipid metabolism pathways may contribute to PE risk

CHARLS: China Health and Retirement Longitudinal Study; CURES: Cardiovascular and Metabolic Diseases Etiology Research Center; FABP2: Fatty acid binding protein 2; FG: Fasting glucose; FI: Fasting insulin; GEO: Gene expression omnibus; GH: Glycated hemoglobin; BMI: Body mass index; HF: Heart failure; IBD: Inflammatory bowel disease; IEU Open GWAS: Integrative Epidemiology Unit Open Genome-Wide Association Studies database; IVW: Inverse variance weighted; LDSC: Linkage disequilibrium score regression; MiBioGen: Human Microbiome Consortium Genome-Wide Association Study; MR: Mendelian randomization; MR-egger: Mendelian randomization-egger regression; MR-PRESSO: Mendelian randomization pleiotropy RESidual sum and outlier; PRS: Polygenic risk score; LDL-C: Low-density lipoprotein cholesterol; PE: Pulmonary embolism; EAA: Epigenetic age acceleration; eGFR: Estimated glomerular filtration rate; AF: Atrial fibrillation; T2DM: Type 2 diabetes mellitus.

Table 2 Pathophysiological mechanisms of pulmonary embolism in relation to Mendelian randomization evidence.

Mechanism	Description	Clinical relevance	MR-informed associations
Thrombus formation	Formation of clots in deep veins driven by Virchow’s triad: Endothelial injury, hypercoagulability, and venous stasis	Risk of embolization to the pulmonary circulation; prevention primarily via anticoagulation	MR-specific associations not directly evaluated
Pulmonary vascular obstruction	Occlusion of pulmonary arteries leading to increased vascular resistance and right ventricular afterload	May result in right ventricular dysfunction and hemodynamic instability requiring urgent intervention	Renal dysfunction (reduced eGFR) associated with increased PE susceptibility
Gas exchange impairment	Ventilation-perfusion mismatch resulting in hypoxemia due to increased dead space ventilation	Manifests as dyspnea and hypoxemia; supportive oxygen therapy required	Obesity-related metabolic traits show variable MR associations
Inflammatory response	Cytokine release, NET formation, and oxidative stress contribute to thromboinflammation	Potential role of anti-inflammatory pathways in disease modulation	Reduced HLA-DR-positive NK cell traits associated with increased PE risk
Endothelial dysfunction	Endothelial injury promotes activation of coagulation pathways and thrombus formation	May contribute to chronic thromboembolic pulmonary hypertension	MMP19 associated with increased PE risk; MMP12 shows inverse association
Metabolic dysregulation	Obesity and metabolic imbalance promote a pro-inflammatory and prothrombotic state	May influence risk stratification and long-term outcomes	BMI and gut microbiota traits show variable MR associations
Coagulation cascade activation	Activation of clotting pathways through tissue factor exposure and thrombin generation	Guides use of anticoagulant therapies	Not directly evaluated in MR studies included in this review

NET: Neutrophil extracellular traps; eGFR: Estimated glomerular filtration rate; PE: Pulmonary embolism; MR: Mendelian randomization; MMP: Matrix metalloproteinases.

CRITICAL APPRAISAL OF MR EVIDENCE

The studies of the relationship between biological factors and disease outcomes can be useful when conducted as MR studies. Nevertheless, the interpretation of MR studies in PE should be undertaken with caution and in the proper epidemiological context. MR is based on the premise that the instrumental variable (genetic variant) is related to the exposure (etiological factor) of interest, that it is not in a state of linkage disequilibrium with other variables that may confound the association, and that it does not affect the outcome by alternative biological mechanisms. Horizontal pleiotropy is another issue. Although several sensitivity analyses (MR-egger regression, weighted median, or MR-PRESSO) can be conducted to determine the robustness of results, residual bias cannot be completely ruled out.

In addition to these issues, most MR studies are based on GWAS data in which the outcome is defined as composite VTE rather than PE. Genetic variants utilized as instruments in MR may not significantly affect the outcome of interest, and it has been proposed that the available genetic instruments may explain only a small fraction of the variance of the phenotype of interest, which can lower statistical power and increase the possibility of weak-instrument bias. Lastly, GWAS studies have predominantly been conducted in people of European descent, thereby limiting the generalizability of findings to other populations.

Each of the above considerations indicates that great care is required in interpreting MR results in humans, since MR can only show the impact of the mutation, and the impact of the mutation may not always be reflected in relevant changes in physiological functioning as determined by direct functional studies, prospective cohort studies, or translational studies. Improving understanding of the implications of the identified pathways in risk stratification and therapeutic approaches to PE requires the integration of genetic epidemiology with biological and clinical research.

In addition, several limitations should be considered. First, given the narrative design of this review focusing on MR studies, the included exposures represent a limited set for which MR evidence exists. While multiple established biological and clinical risk factors for PE have been described in literature, the findings of this review should be interpreted within the scope of available genetic evidence, and not as comprehensive evaluation of all causative factors.

CLINICAL IMPLICATIONS

MR may be used to clarify the causal association between risk factors and disease[8]. In this review, we evaluated evidence on the possible causal association between PE and various risk factors. Genetic studies have revealed that the levels of MMP-19 and HLA-DR⁺ NK cell phenotypes are higher than normal in individuals with PE outside traditional risk groups[5,35]. This supports consideration of targeted PE screening in high-risk profile patients through prior risk evaluation. There are various risk factors that patients who develop this condition may have, and they include T2DM, renal dysfunction, and obesity. Measurement of MMP-19 concentrations in combination with HLA-DR⁺ NK cell features should be considered in research to obtain meaningful findings in high-risk groups[5,35]. The relationship between the HLA-DR⁺ NK cell characteristics and the development of PE is an issue which needs thorough research because it may help scientists develop effective prevention strategies[35]. The therapeutic targeting of MMP pathways shows promise, but scientists have yet to determine the precise cause–effect relationship before such methods can be incorporated into medical therapies[5]. Further research should be done to determine the role of HLA-DR⁺ NK cells in PE development and other immunological processes that influence PE risk[35].

CONCLUSION

PE develops due to a complex interaction of genetic and biological factors. These include changes in MMP activity, immune system disturbances, kidney dysfunction, and shifts in microbiota pathways. Recent MR studies have suggested possible causal links between such factors and PE, but these results should be seen as preliminary until confirmed by further research. Although conditions like AF, obesity, and T2DM share some inflammatory and metabolic features with PE, genetic studies have not consistently shown them to be direct causes. Most findings from MR have not yet been confirmed in clinical trials, which limits their practical use. In addition, most data come from people of European descent, so the results may not apply broadly. Future progress will depend on integrating multi-omics data, confirming findings in the lab, and studying larger, more diverse groups. This will help clarify key mechanisms, improve risk prediction, and support the development of more precise prevention and treatment strategies for PE.