The A −844G Polymorphism in the PAI-1 Gene Is Associated With a Higher Risk of Venous Thrombosis in Factor V Leiden Carriers

Patient Selection Criteria

We studied the genotype distribution of the polymorphisms A −844G and −675 4G/5G located in the promoter region of the PAI-1 gene in 382 individuals heterozygous for the FV Leiden mutation. Individuals were divided into 3 groups (Table 1⇓). Groups 1 and 2 corresponded to all consecutive heterozygous FV Leiden carriers (n=308) from 220 unrelated families who were diagnosed in 1995 to 1998 at the Hematology Laboratory of Timone (Marseille, France) and who were without antithrombin, protein C, or protein S deficiency and lupus anticoagulant. Group 1 included patients with a personal history of VT. Group 2 included asymptomatic relatives of patients from group 1 or individuals with a familial but no personal history of VT. Group 3 consisted of 74 unrelated asymptomatic FV Leiden carriers diagnosed through systematic screening for the FV Leiden mutation on the occasion of a health checkup examination at 3 different health centers (Marseille, Paris, and Lille).

All of the study subjects were interviewed by a physician about their medical history, which emphasized manifestations of deep venous thrombosis (DVT), pulmonary embolism (PE), and superficial thrombophlebitis (STP). The date of occurrence and site of every episode of VT and the presence of precipitating factors (such as surgery, trauma, prolonged immobilization, pregnancy or puerperium, and oral contraceptive intake) were collected. None of the patients had any evidence of neoplastic disease. The term DVT refers to deep leg VT as well as thrombosis at unusual locations such as the axillary, mesenteric, and cerebral veins. VT refers to DVT, PE, or STP. Venous thromboembolism (VTE) refers to DVT complicated or not by PE. The thrombotic events were documented by venography, Doppler ultrasound, angiography, and/or ventilation/perfusion lung scan. Written, informed consent were obtained from each individual.

DNA Analysis

Genomic DNA was prepared by standard salting-out techniques.¹⁸ Detection of FV Leiden (G+1691A), factor II (G20210A), and genotypes of the PAI-1 −675 4G/5G polymorphism was performed by using an allele-specific polymerase chain reaction. A polymerase chain reaction was performed for each allele determination according to the following described conditions. Amplification was carried out in 25 μL in a Perkin Elmer Thermocycler 9600 (Applied Biosystems). Each sample contained 62 ng of genomic DNA in 1× Taq polymerase buffer (2.5 mmol/L MgCl₂), 0.77 mmol/L dNTPs, 5 pmol of each primer (forward and reverse primers in each case plus the allele-specific primer that corresponded to the analyzed genotype), and 0.38 U Taq polymerase (Biotaq, Quantum Bioprobe). A first denaturation at 95°C for 2 minutes was followed by 40 cycles of the following conditions: 1 minute at annealing temperature (determined for each reaction); 72°C for 1 minute, 30 seconds (extension); 95°C for 45 seconds (denaturation); and then 72°C for 5 minutes. All primer sequences and annealing temperature are described in Table 2⇓. Five microliters of each amplification sample was loaded onto a 2% agarose gel that was stained with ethidium bromide. The genotypes resulting from the G-to-A substitution at position −844 were assessed by Xho1 endonuclease digestion as previously described.¹⁰

Statistical Methods

Because individuals within 1 family are not independent, conventional statistical procedures could not be used. Statistical analyses were carried out by using the estimating equations technique proposed by Liang and Zeger.¹⁹ We developed an application of this technique for studying associations between genetic markers and phenotypes by using family data.²⁰ Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated by this technique. The linkage disequilibrium coefficient (D′) between the 2 PAI-1 polymorphisms was estimated by using log-linear model analysis.²¹ Statistical significance was taken at P<0.05.

Characteristics of the Study Population

Table 1⇑ summarizes the main characteristics of the population. No significant difference in sex distribution was observed among the 3 groups (P=0.8). Whereas the mean age at which blood was drawn was significantly lower in patients from group 2 compared with those from group 1 (P<0.001), no significant difference was observed between the mean age at onset of the first thrombosis and the mean age of individuals from group 2 (P=0.7). The mean age at the time of blood drawing was similar in patients with and without VT and in those in whom a fortuitous discovery of the FV Leiden mutation was made (group 3).

The 20210 G→A mutation of the prothrombin gene was present in 11 (7%) patients from group 1, 8 (6%) from group 2, and none of the subjects from group 3. When patients with VT (group 1) were compared with subjects without VT (groups 2 and 3), the prevalence of the A allele carriers was not significantly higher in VT (7%) patients than in asymptomatic individuals (4%, P=0.21). There was a slight but nonsignificant difference in the mean age of the first VT episode between patients with the 20210 G→A mutation and those without (30 versus 37 years, respectively; P=0.2).

Distribution of A −844G and −675 4G/5G Genotypes in Patients With and Without VT

Results are presented in Table 3⇓. The A −844G allele frequency significantly differed among the 3 groups (P=0.048), the A allele being more frequent in patients who suffered from VT (61%) than in subjects without VT (52%, P=0.015). No significant difference was observed in allele frequency between asymptomatic individuals with a familial history of VT and asymptomatic individuals in whom a fortuitous discovery of FV Leiden was made (A allele frequency of 0.52 and 0.51, respectively).

When we compared the −675 4G/5G allele frequency, a difference, though nonsignificant, was observed in allele frequency among the 3 groups, the 4G allele frequency being 0.54 in symptomatic and 0.45 in asymptomatic individuals with a fortuitous discovery of FV Leiden. The 4G allele frequency in asymptomatic individuals with a familial history of VT was intermediate (0.49).

After excluding subjects with the 20210A allele of the prothrombin gene, the distribution of A −844G and −675 4G/5G genotypes was not at all modified in symptomatic and asymptomatic individuals (data not shown). The 2 polymorphisms of the PAI-1 gene were found to be in nearly complete linkage disequilibrium (D′=0.99, P<0.001), the G allele being almost always associated with the 5G allele.

Relative Risk for VT in −844 AA Carriers

The prevalence of the −844 AA genotype was significantly higher in symptomatic than in asymptomatic individuals (38% versus 26%, P=0.015), leading to an OR of 1.74 (95% CI, 1.11 to 2.71; Table 4⇓). This prevalence was even higher in patients with DVT but without a history of PE (42%, P=0.003, compared with asymptomatic individuals), leading to an OR of 2.08 (95% CI, 1.28 to 3.40). When we reexamined the data for the group of patients with at least 1 objectively verified episode of PE, the prevalence of the AA genotype was not statistically different from that observed in asymptomatic individuals (33% versus 21%). Once again, exclusion of individuals with the 20210 G→A mutation did not modify the results, the prevalence of the −844 AA genotype being 39% in patients with VT and 27% in individuals without VT (P=0.015, OR=1.8). There was no difference in the mean age of the first episode of VT between patients with the −844 AA genotype and those with the AG or GG genotype (37 versus 38 years, respectively; P=0.6).