A Prothrombin Gene Mutation Is Significantly Associated With Venous Thrombosis
Abstract This case-control study examined the prevalence of a prothrombin gene mutation in the 3′-untranslated region (UTR) first reported by Poort et al in Dutch subjects with a history of venous thrombosis and in matched control subjects without a history of thrombosis. We tested the hypothesis that the presence of the 3′UTR prothrombin mutation would convey a higher risk of venous or arterial thrombosis and therefore would be found in a higher-than-normal percentage of subjects with a history of thrombosis. Our study included 100 subjects: 50 with a history of thrombosis (21 with venous thrombosis and 29 with arterial thrombosis, who had been recruited from an anticoagulation clinic) and 50 control subjects without a history of thrombosis. DNA from these subjects was analyzed by polymerase chain reaction and agarose gel electrophoresis. We found a statistically significant increase in the prevalence of the 3′UTR mutation in subjects with a history of venous thrombosis compared with subjects without thrombosis. The prevalence of the 3′UTR prothrombin mutation was 19% (4/21; 3 heterozygous and 1 homozygous) in subjects with a history of venous thrombosis, 0% (0/29) in subjects with a history of arterial thrombosis, and 2% (1/50) in control subjects (P<.0245, by Fisher’s exact test for comparison of subjects with versus those without a history of venous thrombosis). The G→A mutation at nucleotide 20 210 in the 3′UTR was confirmed by direct DNA sequencing. The similar increased prevalence of the 3′UTR mutation in subjects with venous thrombosis in our population and in the Dutch population studied by Poort et al suggests that this mutation is an important risk factor for venous thrombosis in the general white population.
- Received April 21, 1997.
- Accepted July 29, 1997.
Venous and arterial thromboses together account for a large proportion of the morbidity and mortality in developed countries. Underlying causes that predispose to thrombosis exert their effects by several mechanisms, some of which have a defined genetic basis.1 The most common hereditary risk factor for venous thrombosis known to date is factor V Leiden, a genetic variant of factor V that is resistant to inactivation by activated protein C.1 2 3 Genetic variants or deficiencies of protein C,4 protein S,5 antithrombin,6 and probably fibrinogen7 have also been shown to increase the risk of venous thrombosis and pulmonary embolism. Other disorders that predispose to venous thrombosis, such as hyperhomocysteinemia, may be due to a complex interaction of multiple genetic or acquired traits.1 Recently, a prothrombin gene mutation that increases the risk of venous thrombosis was discovered by Poort et al8 in a study in the Netherlands of subjects with a family history of venous thrombosis, defined as documented venous thrombosis in two or more family members and the proband.
The prothrombin gene on chromosome 11 consists of 21 kb, with 14 exons, 13 introns, a 5′UTR, and a 3′UTR.8 9 The mutation reported by Poort et al is a single-nucleotide G→A transition at position 20 210 in the sequence of the 3′UTR. No specific function has been defined for this 3′UTR, although some effect on transcriptional regulation was hypothesized.8 The mutation was found in 2% of the Dutch population, in 6% of subjects with a first episode of deep venous thrombosis, and in 18% of subjects with a personal and family history of venous thrombosis.8
We hypothesized that there would be an increased prevalence of the prothrombin gene mutation in subjects with a history of deep venous thrombosis in our clinic population. We chose to examine the prevalence of this 3′UTR prothrombin mutation in a case-control study of patients with a history of deep venous thrombosis and control subjects, who were recruited from the anticoagulation and hematology/oncology clinics at the University Medical Center at Stony Brook on Long Island. We also chose to test the secondary hypothesis that the presence of the 3′UTR prothrombin mutation increases the risk of arterial thrombosis. For this aim, we recruited subjects with a history of arterial thrombosis (coronary, cerebral, or peripheral vascular disease).
The Puregene DNA isolation kit was obtained from Gentra Systems, Inc. This kit included RBC lysis solution, cell lysis solution, protein precipitation solution, DNA hydration solution, and RNase A solution. GeneAmp PCR core reagents were purchased from Perkin-Elmer. PCR core reagents consisted of Amplitaq DNA polymerase, dATP, dCTP, dGTP, dTTP, 10× PCR buffer II, and 25 mmol/L MgCl2solution. The primers used for PCR were identical to those used by Poort et al8 and were synthesized by Life Technologies. Primers used for DNA sequencing (see below) were also from Life Technologies. HindIII restriction enzyme was purchased from Life Technologies. Metaphor agarose and NuSieve 3:1 agarose were obtained from FMC BioProducts. The QIAEX II gel extraction kit, containing QIAEX II suspension, buffer QX1 (with a pH indicator), and buffer PE, was purchased from QIAGEN, Inc. Tris-acetate-EDTA buffer was obtained from Sigma Chemical Co. The ABI prism dye terminator cycle sequencing ready reaction kit with Amplitaq DNA polymerase was purchased from Perkin-Elmer.
A case-control design was selected with recruitment designed to yield a control group that would be comparable to and twice the size of each thrombosis group. A sample size estimate of ≈25 subjects per thrombosis group and 50 control subjects was determined by calculating a sample size based on the prevalence of the prothrombin mutation in the report of Poort et al. This estimate was generated by assuming a value of P<.05 and a power of 80% using a two-tailed estimate. To confirm the familial nature of abnormal PCR results in subjects with thrombosis, we also recruited available family members for PCR testing. Measurement of plasma prothrombin activity was planned in the subjects who were not on anticoagulant therapy with warfarin.
Written informed consent was obtained from all subjects. The study was approved by the Committee on Research in Human Subjects at the University at Stony Brook, following the principles of the Declaration of Helsinki. Both men and women, aged 20 to 77 years (mean age, 55 years; 68 men and 32 women), were enrolled on a volunteer basis from the patients who were being followed up by the anticoagulation and hematology/oncology clinics and from personnel at the University Medical Center at Stony Brook. There were 21 patients in the venous thrombosis group (12 men and 9 women; mean age, 50 years) and 29 patients in the arterial thrombosis group (22 men and 7 women; mean age, 59 years). Inclusion criteria for subjects required that they had a documented history of thrombosis with one or more episodes of deep venous thrombosis or pulmonary embolism or a documented history of coronary artery disease, ischemic cerebrovascular disease, or peripheral ischemic arterial disease. Subjects on anticoagulation therapy for cardiac valve prostheses, cardiomyopathy of nonischemic etiology, or atrial fibrillation without documented coronary artery disease were excluded. All patients eligible by these criteria agreed to participate. The criterion for recruitment of control subjects was the lack of any history of a thrombotic disorder (the reverse of the inclusion criteria for thrombosis subjects). Control subjects were matched for age and sex and ranged from 22 to 79 years old (mean age, 55 years), including 34 men (mean age, 57 years) and 16 women (mean age, 51 years). The control group included 18 healthy subjects and 32 subjects with a variety of nonthrombotic hematologic and oncologic diagnoses, including various cytopenias, leukemia, lymphoma, and solid tumors, with no more than 4 subjects having the same diagnosis. None of these 32 subjects had a history of venous or arterial thrombosis.
Blood Collection and DNA Isolation
Blood samples were collected from each subject after written informed consent was obtained. A total of 5 mL of blood was collected from each patient by venipuncture into a sterile evacuated tube containing EDTA and into tubes containing 3.8% sodium citrate. Plasma was separated by centrifugation at 3000g for 15 minutes at room temperature and then stored in 0.5-mL aliquots at −80°C. Genomic DNA was isolated from white blood cells in the whole-blood EDTA samples by using the Puregene DNA isolation kit and stored at 4°C in 0.1-mL aliquots.
DNA samples were analyzed for genetic variations in the prothrombin gene by PCR, as reported by Poort et al.8 The 3′UTR of the prothrombin gene was amplified by a Tempcycler (COY Lab Products, Inc) using the primer 5′ TCTAGAAACAGTTGCCTGGC-3′ (nucleotides 19 889 to 19 908) and a mutagenic primer 5′-ATAGCACTGGGAGCATTGAA*GC-3′ (nucleotides 20 233 to 20 212). This mutagenic oligonucleotide introduced a novel HindIII restriction site into the amplified gene fragment in the presence of the nucleotide 20210 A allele. PCR was carried out for 35 cycles consisting of 94°C ×1.25 minutes, 55°C ×1 minute, and 72°C ×3 minutes. PCR products were digested with the HindIII restriction enzyme. The DNA fragments of the restriction digest were fractionated by size by agarose gel electrophoresis on 3% MetaPhor Agarose gel and stained with ethidium bromide. Water blanks (no DNA) were included in each PCR run.
PCR products of those subjects who were heterozygous or homozygous for the mutation were extracted from 4% NuSieve 3:1 agarose gels and purified with the QIAEX II gel extraction kit. Purified PCR gene fragments from 5 subjects were sequenced using the ABI prism dye terminator cycle sequencing ready reaction kit and 370A Applied Biosystems Inc DNA sequencer (Perkin-Elmer). Primers for the sequencing were 5′-GATCAGTTTGGAGAGTAGGGG-3′ (nucleotides 20 096 to 20 116) and 5′-TGGTGGATTCTTAAGTCTTCT-3′ (nucleotides 20 304 to 20 284) based on the sequence data of Degen and Davie.9
Prothrombin Activity Assay
Prothrombin clotting activity was measured in the 50 control subjects and 6 offspring of 2 subjects with the 3′UTR prothrombin mutation. The method of assay and the materials used were identical to those used in a previous report from this laboratory.10 Results were expressed as a percentage of the amount of prothrombin in pooled normal plasma, which was arbitrarily designated as 100%.
The data were analyzed using Fisher’s exact test11 for comparison of the prevalence of the prothrombin gene mutation in subjects with or without a history of venous thrombosis and for comparison of subjects with or without a history of arterial thrombosis. This test was preferable to the χ2 test owing to the small values of the cells in the 2×2 tables seen in Tables 1⇓ and 2⇓. A value of P<.05 was considered statistically significant. An OR was calculated for the comparison between the two groups.12 The Wilcoxon two-sample rank sum test was used for comparison of prothrombin activity in subjects with or without the prothrombin gene variant.13
PCR amplification of the 3′UTR prothrombin gene yielded products of three different sizes when digested with HindIII: a 345-bp fragment, a 322-bp fragment, and a 23-bp fragment, as described by Poort et al. The normal 345-bp product generates fragments of 322 bp and 23 bp in the presence of the mutation. Thus, for the heterozygous subjects, bands of 345 and 322 bp were found (representing the normal and mutant alleles, respectively), and in the homozygous subject, a single band at 322 bp was found (Fig 1⇓). The 22-bp fragment was rarely seen due to the small size of this fragment.
As shown in Table 1⇑, we found a statistically significant increase in the prevalence of the 3′UTR prothrombin mutation in those subjects with a history of venous thrombosis (4/21, or 19%) in comparison with subjects without a history of thrombosis (1/50, or 2%) (P<.0245 by Fisher’s exact test). The OR for this comparison revealed that the presence of the mutation significantly increased the risk of thrombosis (OR, 11.5; confidence interval, 1.2 to 110.5). The prevalence of the 3′UTR prothrombin gene mutation in subjects with a history of arterial thrombosis (0/29, or 0%) was not different from that in the control subjects without a history of thrombosis (1/50, or 2%) as shown in Table 2⇑. The allele frequency of the 20 210 A allele was 11.9% in subjects with a history of venous thrombosis and 1.0% in subjects without a history of thrombosis. One subject with a history of venous thrombosis was homozygous for the 3′UTR prothrombin gene mutation, while the 3 other subjects with venous thrombosis were heterozygous.
The 4 subjects with a history of venous thrombosis and the 3′UTR prothrombin gene mutation were further investigated by chart review and by additional interviews and testing. All 4 subjects had been placed on long-term warfarin therapy by their referring physicians because of recurrent deep venous thrombosis or chronic venous insufficiency after an episode of severe proximal deep venous thrombosis; their ages at the first episode of thrombosis ranged from 43 to 66 years. None of the 4 subjects was found to have the factor V Leiden variant, protein C deficiency, protein S deficiency, or antithrombin deficiency. Two of the 4 subjects had 1 family member with deep venous thrombosis or pulmonary embolus. Three of the 4 subjects are entirely of Italian ancestry, including the homozygous subject, and 1 is entirely of Irish ancestry, as determined by a detailed family history. The homozygous subject had 1 parent with a history of pulmonary embolus and no family history of consanguinity, with paternal and maternal ancestry from two separate regions of Italy. He developed central retinal vein occlusion at the age of 66, deep venous thrombosis and pulmonary embolus at age 69, and recurrent deep venous thrombosis at age 70, after warfarin therapy was stopped for 3 months because of a gastrointestinal hemorrhage. He is again being treated with long-term warfarin therapy; hyperhomocysteinemia was also identified and responded to folic acid and vitamin B6 therapy.
Two of the 4 subjects had family members who consented to testing for the prothrombin gene mutation and for plasma prothrombin activity. Each of these 2 subjects was found to have at least 1 child who was heterozygous for the mutation. The prothrombin activity in these two heterozygous adult children was 127% and 114% respectively, compared with 88%, 88%, 85%, and 80% for 3 children and 1 grandchild, respectively, who did not have the mutation. The prothrombin activity in the 1 subject with the mutation from the control group was 139%; this subject had 1 uncle with a history of deep venous thrombosis and family ancestry that was entirely Swedish. The mean prothrombin activity of 127% in 3 individuals with the mutation (2 family members and 1 control subject) who were not taking warfarin was significantly elevated compared with the mean of 81% in 50 individuals without the mutation who were also not on warfarin (P<.02 by Wilcoxon rank sum test).
The family pedigree is shown for one of these families (Fig 2⇓). The proposita and her daughter, who also had a history of deep venous thrombosis, were both heterozygous for the 20 210 A allele, while the grandson was homozygous for the normal allele. Evaluation of the proposita had been negative for the presence of factor V Leiden, antithrombin deficiency, protein C deficiency, or protein S deficiency. The medical history of the Italian parents of the proposita could not be obtained.
DNA sequencing of the 3′UTR of the prothrombin gene from 4 subjects identified by PCR as positive for the mutation and from 1 normal control subject confirmed the presence of the 20 210 A allele in the 4 subjects with an abnormal PCR and also confirmed the homozygosity of 1 of these subjects for this allele. Sequence analysis failed to identify any other nucleotide variations.
Our results indicate that the 3′UTR prothrombin gene mutation is a common variant in an anticoagulation clinic population with a history of venous thrombosis. Our finding of a 19% prevalence of the mutation in subjects with a history of venous thrombosis compared with 2% in control subjects is strikingly similar to the 18% prevalence in subjects with a personal and family history of venous thrombosis compared with 2% in the general population in the Netherlands,8 even though our subjects with venous thrombosis were not selected on the basis of family history. Indeed, none of our 4 subjects with the gene mutation and venous thrombosis had a family history that met the inclusion criteria for the Dutch study (2 family members with thrombosis in addition to the proband), although 2 of our 4 subjects with venous thrombosis and the prothrombin gene mutation had 1 family member with a history of venous thrombosis. Our study population may be influenced by selection bias, since subjects are referred to the anticoagulation clinic by their primary physicians for a variety of reasons.
We did not find the 3′UTR mutation in our subjects with arterial thrombosis, suggesting that it is not an important risk factor for arterial thrombosis. However, few of our subjects with arterial thrombosis were <50 years old, and most had coronary artery disease. Our study was not designed with the power to detect a possible relation of the 3′UTR mutation to the risk of premature arterial thrombosis or to the risk of thrombosis in various subsets of patients, such as young adults with ischemic stroke.
An elevation in plasma prothrombin activity has been shown by Poort et al to accompany the 3′UTR prothrombin gene mutation.8 Our measurements of prothrombin activity in subjects not on warfarin therapy were highly consistent with those findings. Because our subjects with thrombosis were on long-term warfarin therapy, we did not measure the prothrombin activity in these subjects, as it would have primarily reflected the influence of warfarin rather than the effect of the mutation. It is likely that the elevation in plasma prothrombin activity leads to greater thrombin generation when coagulation is activated, thereby increasing the risk of overt clinical thrombosis, but this hypothesis remains to be directly tested.
We did not specifically exclude subjects with known genetic defects, such as protein S deficiency or factor V Leiden, from our study population, and yet none of the subjects in whom we found the 3′UTR prothrombin gene mutation had a second genetic risk factor identified. Individuals who have more than one genetic risk factor for venous thrombosis, such as the prothrombin mutation and factor V Leiden, appear to have a higher risk for thrombosis than those with only one risk factor identified.8 It was fortuitous that each of our subjects found to have the 3′UTR prothrombin gene mutation could precisely identify his or her European ancestry to be entirely from a single country (Italy, Ireland, or Sweden). Although we cannot prove that this gene mutation was inherited from a parent in each case rather than arising by de novo mutation, the former possibility is more likely; these data suggest that the mutation is widely distributed in northern and southern Europe and may be present in the general white population. Our finding of a 19% prevalence of the prothrombin gene mutation in venous thrombosis patients attending an anticoagulation clinic supports the utility and probable importance of testing for this mutation in patients with a history of venous thrombosis.
Selected Abbreviations and Acronyms
|PCR||=||polymerase chain reaction|
This study was supported by vascular diseases academic award HL-02821 (to M.B.H.) from the National Institutes of Health, Bethesda, Md. We would like to thank Dr Roger Grimson for his assistance in the study design and statistical analysis, Jim Kelly for his assistance in the DNA analysis, Shirley Murray for her secretarial support, and the nursing and support staff—Arlene Neuroth, Sharon Azzato, Linda Levy, Marcella Zimmer, and Dorothy Boll—for their assistance with subject recruitment, scheduling, and blood collection.
De Stefano V, Finazzi G, Mannucci P. Inherited thrombophilia: pathogenesis, clinical syndromes, and management. Blood. 1996;87:3531-3544.
Dahlbäck B, Carlsson M, Svensson PJ. Familial thrombophilia due to a previously unrecognized mechanism characterized by poor anticoagulant response to activated protein C: prediction of a cofactor to activated protein C. Proc Natl Acad Sci U S A. 1993;90:1004-1008.
Griffin JH, Evatt B, Zimmerman TS, Kleiss AJ, Wideman C. Deficiency of protein C in congenital thrombotic disease. J Clin Invest. 1981;68:1370-1373.
Comp PC, Nixon RR, Cooper MR, Esmon CT. Familial protein S deficiency is associated with recurrent thrombosis. J Clin Invest. 1984;74:2082-2088.
Poort SR, Rosendaal FR, Reitsma PH, Bertina RM. A common genetic variation in the 3′-untranslated region of the prothrombin gene is associated with elevated plasma prothrombin levels and an increase in venous thrombosis. Blood. 1996;88:3698-3703.
Schefler WC. Statistics for Health Professionals. Reading, Mass: Addison-Wesley Publishing Co; 1984:230-238.
Forthofer RN, Lee ES. Introduction to Biostatistics: A Guide to Design, Analysis, and Discovery. San Diego, Calif: Academic Press Inc; 1995:298-303.
Schefler WC. Statistics for Health Professionals. Reading, Mass: Addison-Wesley Publishing Co; 1984:166-169.