Factor V Leiden Gene Mutation and Thrombin Generation in Relation to the Development of Acute Stroke
Abstract To determine the prevalence of the factor V Leiden gene mutation in relation to the phenotypes of cerebral infarction and cerebral hemorrhage, we studied 386 randomly selected cases of acute stroke and 247 control subjects. Factor V genotype was determined by amplification of a 267-bp sequence of exon/intron 10 of the factor V gene. Levels of prothrombin fragment F1+2, a marker of thrombin generation, were determined in both acute and convalescent stroke and related to factor V genotype. Prothrombin fragment F1+2 was assessed by using an enzyme-linked immunosorbent assay. Sixteen stroke cases (4.1%) were identified as having the mutation compared with 14 (5.6%) control subjects. Prothrombin fragment F1+2 levels were estimated in 191 cases and found to be elevated both acutely and after 3 months, but they were not related to factor V genotype. Prothrombin fragment F1+2 is elevated in acute stroke and requires further evaluation in relation to cerebrovascular disease. These results suggest that the factor V Leiden gene mutation is not a risk factor for arterial thrombosis causing stroke.
- Received March 6, 1995.
- Accepted March 8, 1995.
Activated protein C (APC) resistance is associated with a missense mutation in codon 506 of the factor V Leiden gene in which adenine is substituted for guanine at nucleotide position 1691.1 Heterozygotes (G/A) for this mutation produce wild-type and variant factor V, the latter of which, when activated, is not properly inactivated by APC, leading to increased thrombin generation. The resulting prothrombotic tendency is a major risk factor for development of venous thrombosis.2 Familial transmission of the mutation has been reported.2 3
The mutation is common in the normal Dutch1 and Swedish populations (2% to 7%). It is currently unknown whether the factor V Leiden gene mutation is a risk factor for the development of arterial, as opposed to venous, thrombosis.4
We present the largest population study to date of the prevalence of the Leiden mutation in acute cerebrovascular disease. We have also related the presence of the mutation to a marker of total thrombin generation, prothrombin fragment F1+2.
Three hundred eighty-six patients who satisfied the World Health Organization definition for the diagnosis of acute stroke were prospectively recruited from wards of four acute-care hospitals in Leeds. Each patient gave informed consent to participate in the study, which was approved by the local Ethics Committee, in accordance with institutional guidelines. Cranial computed-tomography scanning was undertaken in each case to establish stroke phenotype (cerebral infarction or intracerebral hemorrhage). Subjects provided venous blood samples for genotyping initially and for F1+2 levels both at presentation and after 3 months. A total of 247 control subjects from the Leeds Blood Transfusion Service and general practitioners (GPs) were genotyped. Thirty-three volunteers gave blood samples for F1+2 estimation only.
Venous blood samples were collected in 10-mL EDTA anticoagulant tubes. Genomic DNA was extracted from leukocytes by a detergent and salt exchange method.5 Samples for F1+2 estimation were immediately taken into 0.1 mol/L cold citrate and centrifuged at 4°C for 30 minutes at 2000g. Samples were snap-frozen in liquid nitrogen and stored at −40°C prior to assay.
Factor V genotype was determined as described by Bertina et al.1 The digestion products were separated by gel electrophoresis in 2.5% agarose gel. The presence of the mutant allele was indicated by a 200-bp product and the normal allele by a 163-bp product, the heterozygotes having both. Fig 1⇓ shows exon/intron 10 amplification and Mnl I restriction digestion.
The mutation in the detected heterozygotes was confirmed as the FV506 Leiden mutation by direct sequencing of the 267-bp product. After attachment of Streptavidin magnetic beads (Dynal) to the biotinylated polymerase chain reaction product and sodium hydroxide treatment to obtain a single-stranded template, samples were sequenced by using a Sequenase 7-deaza-dGTP system (United States Biochemical Corp). The sequencing ladder was visualized by using autoradiography to detect incorporated [α-35S]dCTP. The direct sequencing of exon/intron 10 is shown in Fig 2⇓. Prothrombin fragment F1+2 estimation was performed with an enzyme-linked immunosorbent assay.6
Genotype frequencies were compared by using the χ2 test. F1+2 levels were compared between groups by using a two-sample t test on log-transformed values or the Wilcoxon two-sample test. Differences between initial and 3-month F1+2 levels were compared by a paired t test.
Patient and control subject genotype frequencies are shown in the Table⇓. Results are presented as median ages (interquartile range). Median age in the stroke group was 74 (65 to 80) years, for the GP control group, 76 (73 to 79) years, and for the volunteer group (F1+2 samples only), 67 (60 to 71) years.
Among the stroke group, only 4.1% (16/386) were heterozygous for the mutation compared with 6.3% in the Blood Transfusion Service group and 2.6% in the GP control group. Testing showed no evidence of a difference in these proportions (χ2=1.762 on 2 df, P=.41).
Median F1+2 levels in the stroke group (n=191) at the time of presentation were 55.5 ng/mL (interquartile range, 38.4 to 73.8 ng/mL). Mean log-transformed F1+2 levels were compared by genotype between 185 homozygotes (G/G) and 6 heterozygotes (G/A). The ratio of geometric means of the heterozygotes compared with homozygotes was 1.15, with a wide 95% confidence interval (CI: 0.73, 1.82). F1+2 levels were also compared between 16 cases of cerebral hemorrhage and 175 cases of cerebral infarction. There was no difference in the means (t test, P=.31). The ratio of geometric means between the two groups was 1.16 (95% CI: 0.87, 1.55). F1+2 levels both at the time of presentation and after 3 months were available in 68 cases. The median change was 2.6 ng/mL, with the interquartile range from a decrease of 16.7 to an increase of 11.5 ng/mL. This change was not significantly different from zero (P=.47).
The median level in the volunteer group was 41.9 ng/mL (interquartile range, 36.6 to 54.2 ng/mL) compared with 55.5 ng/mL in the stroke group (interquartile range, 38.4 to 73.8 ng/mL). F1+2 levels were significantly higher in the stroke group (Wilcoxon, P=.003).
APC is a protease formed from its zymogen by the action of thrombin and thrombomodulin, which inactivate factors Va and VIIIa in the presence of nonenzymatic protein S. Initial work by Dahlback et al7 suggested the presence of a novel anticoagulant in the protein C pathway that was functionally described as APC cofactor 2. APC cofactor 2 deficiency is associated with APC resistance. APC resistance has been linked to a sevenfold increase in the risk of deep vein thrombosis.2
Bertina and coworkers1 investigated a large family with APC resistance plus a venous thrombotic tendency by means of linkage studies. Segregation of microsatellite markers for a number of loci in the 1q21 through 25 region were identified. Mutations in the regions that contained the putative APC binding site and APC cleavage site of factor V (Arg506) were sought. Polymerase chain reaction fragment sequencing revealed two patients homozygous for a guanine→adenine substitution at nucleotide position 1691. The occurrence of this mutation predicted the replacement of Arg506 (CGA) by glutamine (CAA), or factor V Leiden. The APC cleavage site mutation accounted for the observed thrombotic tendency, ie, the resulting factor V molecules were resistant to APC yet retained normal factor V procoagulant activity.
Voorberg and colleagues8 assessed 27 consecutive patients with recurrent thromboembolism for both APC resistance and the presence of the factor V mutation. Their results indicate that APC resistance was linked to a single mutation at the putative APC cleavage site Arg506 in factor V. Unlike cases of venous thrombosis, in which the mutation is common,1 the frequency of the mutation in cases of arterial thrombosis has not hitherto been described.
We have determined the incidence of the mutation in a large group of elderly patients with cerebral hemorrhage and cerebral infarction as determined by cranial computed-tomography scanning. For the whole stroke group, the G/A heterozygote frequency was 4.1%; this was not statistically different from the control population G/A frequencies (6.3% and 2.6%), which are comparable with other studies.1 There was no difference in proportions of the mutation in those cases with cerebral infarction (15/348) compared with those who developed cerebral hemorrhage (1/38).
The low incidence of the mutation among stroke patients is in contrast to that observed in cases of venous thrombosis.8 However, most cases of stroke result from arterial as opposed to venous thrombosis, suggesting that different pathological processes are operating. In comparison with cases of deep vein thrombosis, we suggest that the routine screening of elderly patients with stroke for this mutation is unlikely to be of value in clinical practice, although it remains to be established if the factor V mutation is a risk factor for juvenile stroke.
Since F1+2 levels are a measure of total thrombin generation6 and a further marker of activation of coagulation, cases with APC resistance might be expected to have higher F1+2 levels than those individuals not bearing the mutation. This was not the case in the small sample of heterozygotes tested. Compared with the levels in the control population, patients’ F1+2 levels were elevated both initially and 3 months after stroke. These results suggest that enhanced thrombin generation occurs when the acute phase response is waning, but it appears unrelated to the factor V mutation. The role of F1+2 in relation to stroke outcome warrants further study.
In summary, although the factor V Leiden mutation is an important risk factor for venous thrombosis, we have demonstrated that it does not have an important role in arterial thrombosis causing stroke.
This work was supported by the Stroke Association, the Leeds General Infirmary Locally Organised Research and Special Trustees, The Wellcome Trust, and Westminster Medical School Research Trust. We thank Dr John Bamford, consultant neurologist and physician in cerebrovascular medicine, for his continued support and advice.