Prothrombin G20210A and Factor V Leiden Variants Are Not Associated With Thrombotic Events in Congenital Heart Disease: An Observational Trial.
Mar 4, 2026DOI: 10.1161/jaha.125.046959Journal ArticleAHA
Written by: Feria A LadhaChristina VanderPluymEnrique Mondragon-EstradaPaul AvillachMartina BruecknerWendy K ChungJames F CnotaBruce D GelbMatthew LewisCong LiuAmy E RobertsChristine E SeidmanMartin Tristani-FirouziMichael WagnerSarah U MortonJane W NewburgerPediatric Cardiac Genomics Consortium Investigators
Advanced Heart Failure · Boston Children's Hospital
Methods
Deidentified individual participant data (including data dictionaries), study protocols, the statistical analysis plan, and the informed consent form will be made available upon publication to researchers who provide a methodologically sound proposal for use in achieving the goals of the approved proposal. Proposals should be submitted to b2bprogram@cchmc.org. Participants in the Pediatric Cardiac Genomics Consortium (PCGC)10 with exome sequencing and electronic medical record (EMR) data were identified. The PCGC is a multicenter, prospective cohort established through the National Heart, Lung, and Blood Institute Bench to Bassinet program and includes people of all ages with CHD recruited from multiple pediatric cardiac centers across the United States. To minimize bias, clinical data are derived from standardized chart abstraction and harmonized EMR fields, including cardiac diagnoses, surgical history, and relevant comorbid conditions. Detailed phenotyping is supplemented by exome sequencing and biospecimen collection, enabling integration of genomic and clinical data. Enrollment started in December 2012. Written informed consent was obtained from participants or their parents. The study was approved by the institutional review board at each study center and, after 2023, by a single institutional review board. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline. Thrombosis events were identified using PheCodes (phenotype codes; derived from International Classification of Diseases, Ninth Revision [ICD‐9] and Tenth Revision [ICD‐10] codes) 286.x, 411.x, 415.x, 433.x, 440.2, 444.x, 452.x, 586.3, 656.5, 671, 854, 856, 857, 858, and 859. PheCode 433.x was used to define acute ischemic stroke. Single‐ventricle heart disease was identified on the basis of Fyler codes1, 2 200, 210, 211, 213, 220, 230, 240, 250, 260, 270, 280, 300, 400, 411, 412, 413, 421, 422, 423, 431, 432, and 433. Antiplatelet agent (acetylsalicylic acid, clopidogrel, dipyridamole) or anticoagulation (apixaban, argatroban, bivalirudin, dabigatran etexilate, dalteparin, enoxaparin, fondaparinux, heparin, rivaroxaban, warfarin) use was identified using RxNorm codes. Central venous catheter use was identified using Current Procedural Terminology codes 36 555, 36 556, 36 558, 36 573, and 36 581. Exome data from PCGC and the Genome Aggregation Database11 version 4.1 were subset to F2 c.97G>A and F5 c.1601G>A (hg38 chr11:46739505 and chr1:169549811, respectively) and filtered for quality as previously described.3 Participants with missing data were excluded from analysis. Pearson's χ2 test was used for comparing categorical outcomes. Power analyses were performed using the noncarrier event rates as baseline (18.4% for , 18.3% for ) to calculate the minimum detectable effect size at 80% power (α=0.05), which corresponded to carrier event rates of ≥35% for (risk ratio, 1.9) and ≥29% for (risk ratio, 1.6).
F2
F5
F2
F5
Results
Of 6371 PCGC participants with genotype data for thrombophilia variants prothrombin G20210A and factor V Leiden, 4008 also had EMR data (Table 1; Figure S1). Compared with the Genome Aggregation Database reference population, European participants in the PCGC cohort had a lower prevalence of heterozygous prothrombin G20210A and factor V Leiden variants (72/7634 [0.9%] versus 13 811/535449 [2.6%], P<0.001; and 126/7634 [1.7%] versus 29 147/589584 [5%], P<0.001, respectively). The frequency did not differ for participants with predominantly African (F2: 3/568 [0.53%] versus 165/36334 [0.45%], P=0.75; F5: 7/568 [1.2%] versus 282/37506 [0.75%], P=0.21), Admixed American (F2: 18/1421 [1.27%] versus 671/29247 [2.3%], P=0.02; F5: 20/1421 [1.4%] versus 460/30013 [1.5%], P=0.82), or South Asian (F2: 2/362 [0.55%] versus 173/43590 [0.4%], P=0.66; F5: 5/362 [1.4%] versus 1270/45534 [2.8%], P=0.14) genetic ancestry. In the overall cohort, the prevalence of homozygous variants did not differ significantly, with limited statistical power for comparison. Nine individuals (0.1%) were compound heterozygotes, carrying both F2 and F5 risk alleles.
Thrombophilia Genotype Prevalence
Prevalence of heterozygous F2 and F5 thrombophilia genotypes was compared for the following pairs: PCGC and Genome Aggregation Database, and with and without electronic medical record data among participants from PCGC. P value threshold of 0.0125 was used to define significance on the basis of Bonferroni correction for 4 test. OR indicates odds ratio; and PCGC, Pediatric Cardiac Genomics Consortium.
Among the 4008 PCGC participants with EMR data, 737 (18%) had PheCodes for a peripheral or central thrombotic event; these included acute ischemic stroke (n=93/737 [13%]) and other events such as peripheral venous thrombosis or cerebral embolism (n=644/737 [87%]). Central venous catheter exposure was present in 689 of the 4008 participants (17%), and its prevalence did not differ between those with and without thrombotic events (112/737 [15%] versus 577/3271 [18%], P=0.12). At least 1 antithrombotic medication, including antiplatelet agents and anticoagulants, was prescribed at some point to 1417 of the 4008 PCGC participants (37%).
Participants with single‐ventricle heart disease, compared with biventricular heart disease, had a higher frequency of thrombosis (166/520 [32%] versus 571/3488 [16%], P≤0.001) and were more likely to have received antithrombotic agents (Table 2). Thromboprophylaxis was administered for more participants with codes for thrombotic events than for those without such codes (384/459 [84%] versus 627/958 [65%], P≤0.001).
Thromboprophylaxis and Thrombosis Events
Prevalence of thromboprophylaxis in the overall cohort as well as by CHD type and thrombosis status. P value threshold of 0.025 was used to define significance on the basis of Bonferroni correction for 2 test. CHD indicates congenital heart disease; and OR, odds ratio.
PCGC participants with and without EMR data did not differ significantly in thrombophilia variant prevalence (Table 3). Among those with EMR data, the prevalence of prothrombin G20210A and factor V Leiden variants did not differ significantly according to univentricular versus biventricular status. Heterozygous thrombophilia genotypes were found in 30 (4%) of the 737 participants with codes for thrombotic events. The prevalence of a thrombophilia variant did not differ significantly according to presence or absence of codes for a thrombotic event or an acute ischemic stroke. Among the 93 participants with acute ischemic stroke, 2 were prothrombin G20210A heterozygous, and 1 was factor V Leiden heterozygous. Six of the 9 individuals who were heterozygotes for both F2 and F5 risk alleles had EMR data, and 2 (25%) had thrombotic events.
Thrombophilia Genotypes and Thrombosis
Sex balance and age at last electronic medical data are provided for the overall cohort and each subset. Prevalence of F2 and F5 risk genotypes was compared for the following pairs: single‐ventricle and biventricular CHD, with and without thrombosis, and with and without acute stroke among those with thrombosis. P value threshold of 0.0008 was used to define significance on the basis of Bonferroni correction for 4 tests. CHD indicates congenital heart disease; IQR, interquartile range; and OR, odds ratio.
Discussion
In this multicenter CHD cohort, we found no significant association between heterozygous prothrombin G20210A or factor V Leiden variants and thrombotic outcomes in CHD. Only a small proportion of participants with thrombotic events carried 1 of these variants, and no enrichment was observed compared with those without events. These data suggest that, in the era of routine prophylactic antithrombotic therapy, common thrombophilia variants contribute little to thrombotic risk in CHD. Instead, hemodynamic burden, intracardiac shunting, surgical interventions, and flow disturbances likely play a greater role in driving outcomes. Central venous catheter exposure, another known risk factor for thrombotic events, was not more frequent among those with thrombotic events in our cohort, although catheter‐related risk may have been obscured by concurrent hemodynamic and procedural factors, as well as antithrombotic prophylaxis. This aligns with current clinical practice, which does not recommend routine screening for F2 or F5 variants when determining prophylactic therapy.
We also observed a lower prevalence of heterozygous prothrombin G20210A and factor V Leiden variants compared with the general population in people of European genetic ancestry. Prior studies reporting higher frequencies were often single‐center or selectively recruited, differing demographically from ours and sometimes including patients without CHD or perioperative or high‐risk subsets.12, 13 These findings suggest survivorship bias. Thrombophilia variants including factor V Leiden and prothrombin G20210A have been linked to placental thrombosis, fetal loss, and adverse perinatal outcomes, and carriers with concomitant CHD may face even higher risks of not surviving to enrollment.14, 15 Thus, the lower prevalence in our CHD cohort may reflect attrition beginning as early as fetal life. Larger, ancestrally diverse cohorts will be required to determine whether survivorship bias underlies the reduced prevalence of thrombophilia variants in CHD.
We relied on EMR and medication data from cardiac centers, which may contain coding errors or omissions, and thrombotic events occurring in outpatient or nonparticipating inpatient settings may not have been captured. The timing of antithrombotic therapy relative to events could not be determined. Even in hospitalized patients, diagnostic codes likely underestimate event frequency, as many thrombotic events, particularly cerebral infarcts, are clinically silent. Additionally, although prior studies report that many patients with single‐ventricle CHD experience thrombosis by early childhood, we could not determine the precise timing of thrombotic events in our cohort due to limitations in EMR documentation, preventing direct comparison with age‐specific rates reported in the literature. For example, brain magnetic resonance imaging studies after the Fontan procedure demonstrate that a substantial proportion of strokes go undiagnosed clinically.16 Due to the low prevalence of prothrombin G20210A or factor V Leiden variants, the study was powered only to detect large effects on thrombotic event rates (≥1.9‐fold for F2 and ≥1.6‐fold for F5), and smaller relative increases could not be reliably excluded. Calculations of the minimum detectable effect size showed that only large relative risks could be detected and that additional sensitivity or subgroup analyses would have been underpowered and statistically unstable. Thrombophilia genotypes beyond prothrombin G20210A and factor V Leiden were not assessed, and our study was underpowered to evaluate contributions from rare homozygous or compound genotypes.
Taken together, our findings clarify 2 points: (1) that common thrombophilia variants appear reduced in prevalence in CHD, possibly reflecting survivorship bias; and (2) that these variants are not significantly associated with thrombotic outcomes in this population. Future work incorporating ancestry‐stratified analyses; larger and more diverse cohorts; and models integrating clinical, procedural, and genomic factors will be essential to refine thrombotic risk prediction and optimize prophylaxis for individuals with CHD.
Sources of Funding
This project was supported by the National Center for Advancing Translational Sciences (UL1‐TR000003 and UL1‐TR002541), the National Institutes of Health (U01‐HL098188, U01‐HL098147, U01‐HL098153, U01‐HL098163, U01‐HL098123, U01‐HL098162, U01‐HL131003, 5R01‐HL162356, 5U54‐HG006504 [C.E.S. and S.U.M] and K08‐HL157653 [S.U.M.]), Boston Children's Hospital Office of Faculty Development (Career Development Award, S.U.M.) and American Heart Association (Career Development Award, S.U.M.), the Farb Family Fund (J.W.N.), and the Kostin Family Innovation Fund (J.W.N.). The sponsors had no role in the design and conduct of the study.