Polymorphisms of the Factor VII Gene and Circulating FVII:C Levels in Relation to Acute Cerebrovascular Disease and Poststroke Mortality
Background and Purpose FVII:C has been shown to be an independent risk factor for myocardial infarction and is related to environmental and genetic factors. This study sought to investigate FVII:C levels and factor VII (FVII) gene polymorphisms in relation to stroke and disease outcome.
Methods To examine the association of FVII:C and the Msp I and promoter insertion polymorphisms of the FVII gene in acute stroke, 317 patients and 198 age-matched control subjects were studied.
Results FVII:C levels were significantly lower in patients at onset than 3 months later (119% versus 135%, respectively; P<.0005). Levels were significantly lower in patients at onset than in control subjects (124% [95% confidence interval, 120% to 129%] versus 141% [95% confidence interval, 135% to 148%], respectively; P<.0005) but were not significantly different at 3 months (135% [95% confidence interval, 128% to 141%] versus 141% [95% confidence interval, 135% to 148%], respectively). We found no difference in genotype distribution for either polymorphism between patients and control subjects, no difference in FVII:C level or genotype distribution between pathological types of stroke, and no relationship with poststroke mortality. Both polymorphisms were significantly associated with FVII:C levels in patients and control subjects. In a multiple regression model for patients, Msp I genotype, cholesterol, and smoking remained as independent predictors of FVII:C levels, accounting for 32% of interindividual variation.
Conclusions These results suggest that neither FVII:C levels nor FVII gene polymorphisms are associated with cerebrovascular disease. There were no genotype-specific correlations of environmental factors with FVII:C, but there was evidence of an acute-phase or consumptive fall in FVII:C levels at the time of stroke, whereas levels increased to those similar for healthy age-matched control subjects by 3 months, when the acute phase had presumably subsided.
Factor VII is a serine protease, synthesized principally in the liver. It is a single-chain glycoprotein with a relative molecular mass of 48 000 and is activated by a number of other proteases in the presence of tissue factor to the two-chain form (FVII:A).1 FVII:A then initiates activation of factors IX and X in the presence of Ca2+ and phospholipid.2 FVII is the first enzyme involved in the extrinsic pathway and has a key role in the coagulation cascade. By contributing to a prothrombotic state, elevated levels of FVII:C may provide a mechanism for increased vascular risk.
CVD shares many of the risk factors associated with IHD,3 and components of the hemostatic system shown to be risk factors for IHD4 5 6 may also have a role in the pathogenesis of this disorder. The Northwick Park Heart Study showed elevated FVII:C levels to be a strong, independent risk factor for IHD, particularly fatal events.4 Extensive work has been performed to determine the relationship between FVII and risk of IHD.4 6 7 8 Few studies have investigated the relationship between FVII and risk of CVD, although small case-control studies have suggested an association between FVII, IHD, and stroke risk.9
Genetic elements have been shown to regulate FVII:C levels. A common coding polymorphism detected by Msp I endonuclease has been identified in exon 8 of the FVII gene.10 A single base substitution of guanine (G) to adenine (A) at codon 353 results in the replacement of arginine (allele M1) to glutamine (allele M2) in the protein product. Several studies have observed higher levels of FVII:C in M1 homozygotes than in carriers of the M2 allele10 11 12 13 14 15 16 ; possibly the amino acid change causes an alteration in the protein conformation of FVII, resulting in reduced secretion.17 It has been postulated that carriers of the M2 allele may have a reduced risk of vascular disease compared with M1 homozygotes as a result of reduced FVII:C levels.
A decanucleotide insertion polymorphism of unknown functional significance at position −323 in the promoter18 has also been shown to relate to FVII:C levels16 19 and shows linkage disequilibrium with the Msp I restriction fragment length polymorphism.16 19 FVII gene polymorphisms have not previously been studied in relation to CVD.
FVII:C has also been shown to correlate with environmental factors, particularly blood lipids20 21 22 and age.23 Some studies have demonstrated a more pronounced association of FVII:C with blood lipids in M1 homozygotes, although others dispute this finding.12 13 14 15 16
It is unknown whether FVII:C levels are elevated in stroke, although a small study found a trend toward reduction in FVII and factor X together.24
In this study we sought to determine the relationship between FVII:C levels and FVII gene polymorphisms with CVD and disease outcome by studying stroke subtypes and poststroke mortality.
Subjects and Methods
We recruited 317 white subjects with a clinical diagnosis of acute stroke from four hospitals in Leeds and 198 age-matched healthy control subjects, clinically free of vascular disease, from large employers in the Leeds area and from Family Health Services Authority general practice registers, as described previously.25 26 Clinical characteristics of the subjects are presented in Table 1⇓. There were fewer control subjects than patients because of a lack of similarly aged healthy control subjects and a bias of more women than men available to study. Informed consent was obtained for all subjects, and the study was approved by the United Leeds Teaching Hospitals Trust Research Ethics Committee.
We recorded current drug therapy, blood pressure, and BMI, calculated by weight in kilograms divided by the square of height in meters. Subjects were classified as smokers if they currently smoked or had previously smoked within the last 5 years. Hypertension was defined as present if the mean premorbid blood pressure was greater than 160 mm Hg systolic or 90 mm diastolic or the subject was previously described as hypertensive and receiving treatment.
The pathological type of stroke was confirmed by noncontrast cranial CT scan within 10 days (mean, 2 days) of the acute event and classified as intracerebral hemorrhage or intracerebral infarction. Clinical subtype of cerebral infarction was then further classified according to the Oxfordshire Community Stroke Project criteria.27 Subjects presenting lacunar syndromes (pure sensory, pure motor, sensorimotor stroke, or ataxic hemiparesis) whose CT scan showed an appropriate small deep infarct or no relevant lesion (LACI) were considered to have probable small-vessel disease. Those presenting syndromes that were suggestive of TACI or PACI were considered to have probable large-vessel infarction, and those presenting with syndromes referable to the posterior (vertebrobasilar) circulation were considered likely to have mixed small- and large-vessel pathology. This classification predicts the volume of parenchymal damage, which is a function of the diameter of the occluded vessel (in descending order: TACI, PACI/POCI, LACI). Patients were also “flagged” with the Office of Population Censuses and Surveys for notification of death.
Free-flowing, venous blood samples were obtained within 10 days of the development of stroke. Ten milliliters of blood was placed into 0.9% citrate at room temperature for assay of FVII:C levels, separated by centrifugation at 3000g for 20 minutes at room temperature (20°C), snap-frozen in liquid nitrogen, and stored in aliquots at −40°C until assay. Plasma samples were thawed at 37°C and used immediately. Ten milliliters of blood was collected into EDTA for DNA extraction and 5 mL of blood into lithium heparin for plasma lipid analysis. An additional blood sample was taken at least 3 months after the acute event for analysis of FVII:C and other hemostatic factors only. Control subjects gave a single sample for determination of FVII:C levels, plasma lipid analysis, and DNA extraction.
Measurements of cholesterol and triglycerides were made with a Hitachi 747 autoanalyzer (Boehringer Mannheim).
Plasma FVII:C was assayed by the Northwick Park assay, a one-stage biological assay with increased responsiveness to activated FVII.28 29 The assay was performed using an H Amelung KC 10 coagulometer (Brownes Ltd), with an in-house FVII-deficient plasma and rabbit brain thromboplastin with an International Sensitivity Index of 1.4 (Diagen, Thame) as reagents. The results were calibrated with a plasma pool that was calibrated against a standard (NIBS & C). Since the thrombotic response to tissue factor release may be determined by circulating levels of activated FVII, this assay may be preferable to other FVII:C assays for estimating increased FVII:C levels.29
Genomic DNA was extracted by a salt-detergent method.30
Amplification of DNA was performed by PCR in a final volume of 25 μL containing 100 ng genomic DNA, 50 pmol of each primer, 1.5 mmol/L MgCl2, 3.2 mmol/L dNTPs, 0.0025% W-1, 1× Taq buffer, and 1.25 U Taq polymerase (GIBCO).
The nucleotide sequences of the primers for the Msp I polymorphism (5′ through 3′) were GGG AGA CTC CCC AAA TAT CAC and ACG CAG CCT TGG CTT TCT CTC, as described previously.10 The PCR reaction consisted of denaturation for 1 minute at 94°C, annealing for 1 minute at 59°C, and extension for 1 minute at 72°C and continued for 34 cycles. The PCR product was digested with 10 U Msp I (Northumbria Biologicals Ltd), and genotype patterns were photographed after separation on a 2% agarose gel. Msp I digestion yielded a constant band of 40 bp, two bands of 205 and 67 bp in the presence of the cutting site (allele M1), and only one band of 272 bp in the absence of the cutting site (allele M2).
The PCR conditions for the promoter polymorphism were as for the Msp I polymorphism, and the primer sequences (5′ through 3′) were GAG CGG ACG GTT TTG TTG CCA GCG and GGC CTG GTC TGG AGG CTC TCT TC (Fiona Green, personal communication, 1994). The reaction proceeded for 30 cycles, each with a denaturation step of 1 minute at 94°C, annealing for 1 minute at 71°C, and extension for 1 minute at 72°C. DNA fragments were visualized as for Msp I, yielding a band of 204 bp (A) in the absence of the insertion and 214 bp (a) in its presence.
Age was non-normally distributed and therefore was treated as a nonparametric variable. Values for triglyceride, cholesterol, and BMI exhibited a log-normal distribution, and these variables were log-transformed to permit use of parametric methods.
The Mann-Whitney U test was used to compare median ages between patients and control subjects. The unpaired t test was used to compare levels of the various metabolic parameters between sexes and between patients and control subjects and to compare FVII:C levels between smokers and nonsmokers, between hypertensive and normotensive patients, between nonsurvivors and survivors, and between intracerebral hemorrhage and infarction stroke types. The paired t test was used to compare levels of FVII:C between patients’ initial and second visits. Bivariate correlations were analyzed according to Pearson’s method, except when age was a variate, when Spearman’s method was used.
The χ2 test was used to compare genotype distributions for the polymorphisms to that expected if the alleles were in Hardy-Weinberg equilibrium; it was also used to compare frequencies of genotype distributions and smoking between sexes and genotype distributions between pathological types of stroke and between survivors and nonsurvivors. Linkage disequilibrium was assessed with a method developed by Hill31 and exemplified by Baumann and Henschen.32 One-way ANOVA was used to compare FVII:C levels between genotype groups and between pathological types of stroke, and multifactorial ANOVA was performed with age, sex, BMI, triglyceride, and cholesterol as covariates.
Triglyceride, cholesterol, and BMI in relation to FVII:C levels in each genotype group were analyzed with regression analysis, and logistic regression analysis was performed to further determine the relationship between FVII genotype and stroke risk in patients versus control subjects.
In view of the infrequency of genotype M2M2, the genotypes M2M2 and M1M2 were combined for analysis. The relationship of the various metabolic features and genotypes to FVII:C levels was then assessed by multiple linear regression analysis. Genotypes were entered as a dummy variable, where M2M2 and M1M2=0, M1M1=1, AA=1, and aa and Aa=0. All computations were performed with the use of the SPSS for Windows (version 6.1) statistical package.
Table 1⇑ shows the clinical characteristics of patients at the initial visit and control subjects. Cholesterol levels were significantly higher in control subjects than in patients, more patients than control subjects had a history of hypertension, and there was a trend toward higher triglyceride levels in patients than in control subjects. All other parameters were similar in both groups.
Mean levels of FVII:C were significantly higher in control subjects than in patients at the initial visit: control subjects, 141% (95% CI, 135% to 148%); patients, 124% (95% CI, 120% to 129%) (P<.0005). Results were similar when levels were adjusted for other correlating factors (age, cholesterol, genotype, sex, and triglyceride): control subjects, 144%; patients, 111% (P<.0005). Mean levels of FVII:C were not significantly different between patients at the second visit compared with control subjects: control subjects, 141% (95% CI, 135% to 148%); patients, 135% (95% CI, 128% to 141%). Mean FVII:C levels at the initial visit were significantly lower than those at the second visit (initial visit, 119%; second visit, 135%; r=.56, P<.0005). Analysis of levels by genotype separately showed that these differences were attributable mainly to FVII:C levels between M1 homozygote groups, as shown in Fig 1⇓.
The genotype distributions for the promoter and Msp I polymorphisms were not different from those predicted by Hardy-Weinberg equilibrium in either group (for Msp I patients: M2M2, n=5; M1M2, n=59; M1M1, n=235; for Msp I control subjects: M2M2, n=2; M1M2, n=42; M1M1, n=134; for promoter patients: aa, n=3; Aa, n=55; AA, n=241; for promoter control subjects: aa, n=1; Aa, n=38; AA, n=136). There was also no difference in allele frequencies between groups (for Msp I patients: M2, 0.13; M1, 0.87; for Msp I control subjects: M2, 0.12; M1, 0.88; for promoter patients: a, 0.11; A, 0.89; for promoter control subjects: a, 0.3; A, 0.9). Percent distributions of genotypes (and 95% CIs) are shown in Table 2⇓. The Msp I and insertion genotypes were in linkage disequilibrium (±D′=0.96, P<.00005).
FVII:C levels were related to both polymorphisms in patients initially and at 3 months and in control subjects as well (Figs 1⇑ and 2⇓); these correlations remained when adjustments were made for age, sex, cholesterol, and triglyceride.
In the patient group, FVII:C correlated significantly with cholesterol and triglyceride, and there was a trend toward correlation with BMI (Table 3⇓). Mean FVII:C levels were significantly higher in women than in men: women, 128% (95% CI, 120% to 135%); men, 115% (95% CI, 109% to 120%) (P=.006). In the control subjects, FVII:C levels correlated significantly with triglycerides and age (Table 3⇓) and were higher in women than in men: women, 148% (95% CI, 140% to 156%); men, 131% (95% CI, 122% to 141%) (P<.005). There was no difference in mean FVII:C levels between patients with a history of hypertension compared with those without: history of hypertension, 121% (95% CI, 118% to 130%); no history of hypertension, 127% (95% CI, 120% to 133%).
In a linear regression model containing BMI, triglyceride, cholesterol, stroke type, sex, age, hypertension, smoking, and genotype, the factors that remained as independent and significant predictors of FVII:C levels were genotype, cholesterol, and smoking. Such a regression model containing Msp I genotype gave a greater predictive value of FVII:C than when replaced by the promoter polymorphism, with R2 values of .32 compared with .28. Msp I genotype accounted for 18% interindividual variation, cholesterol an additional 11%, and smoking an additional 3%. In the control subjects, Msp I genotype also gave a stronger predictive value for FVII:C levels, contributing 10% variation, with age accounting for 7% and triglyceride an additional 6%.
There was no difference in the relationship between FVII:C and BMI, cholesterol, or triglyceride when analyzed by genotype.
When FVII genotypes were entered into a logistic regression model containing other conventional risk factors for stroke (eg, smoking, hypertension, cholesterol), neither emerged as a risk factor for stroke: for Msp I, odds ratio=1.02 (95% CI, 0.43 to 1.63) (P=.6); for insertion polymorphism, odds ratio=1.44 (95% CI, 0.73 to 2.85) (P=.3). In contrast, age, hypertension, smoking, and previous history of stroke or IHD did emerge as risk factors.
When intracerebral infarction and hemorrhage stroke types were compared, there were no differences in FVII:C levels: hemorrhage, 115% (95% CI, 102% to 120%) (n=31); infarction, 121% (95% CI, 116% to 126%) (n=286). There were also no differences in genotype distribution: hemorrhage, M2, 0.11; M1, 0.89; infarction, M2, 0.12; M1, 0.88).
Table 4⇓ shows the number of patients in each group according to pathological type of stroke. FVII:C levels were similar in each group, and there were no significant differences in genotype distributions between groups. When patients who had died within 3 months were compared with survivors, we found no significant differences in FVII:C levels (alive, 124% [95% CI, 119% to 129%] [n=276]; dead, 118% [95% CI, 106% to 131%] [n=41]), nor did we find any significant differences in genotype distributions (alive, M2, 0.11; M1, 0.89; dead, M2, 0.13; M1, 0.87). There were also no differences when we compared those who had died within 1 month with those who had not (alive, 123% [95% CI, 119% to 128%] [n=300]; dead, 128% [95% CI, 105% to 152%] [n=17]).
CVD is a major cause of morbidity and mortality in the Western world. Epidemiological evidence suggests that CVD shares many of the risk factors for IHD,3 such as age,33 hypertension,34 and smoking.35 It is also possible that features of the hemostatic system may influence the development of CVD, as has been suggested for IHD. Neither FVII:C levels nor FVII gene polymorphisms have previously been studied in stroke.
This study has demonstrated that despite observations that FVII:C is associated with genetic factors, there was no association of FVII:C or FVII gene polymorphisms with pathological type of stroke or with poststroke mortality. FVII:C was associated with environmental factors in patients with CVD, as described previously for other populations, and levels of FVII:C were observed to fall at the time of stroke.
In this stroke patient group, FVII:C correlated with cholesterol and triglyceride, as observed in most previously studied populations,12 13 14 15 16 22 and was higher in women than men, as previously observed in healthy individuals.20 21 22 FVII:C was not associated with age or with BMI, as previously reported,23 although a relationship with age may be masked by the small range of ages present in this group and a relationship with BMI may be masked by a relatively small distribution of BMI and little evidence of obesity.
In multiple regression models, cholesterol, Msp I genotype, and smoking remained significant predictors of FVII:C levels in patients. Msp I genotype emerged as a stronger predictor of FVII:C levels than the promoter polymorphism, which suggests that the Msp I polymorphism is functional, as previously suggested.17
A genotype-specific correlation between FVII:C and triglyceride, with a stronger association in the M1M1 homozygotes, has been described in some although not all studies.12 13 14 15 16 No genotype-specific correlations were found for any parameter in this study. Possible explanations for this discrepancy include ethnic differences,13 15 the fact that subjects were not always in a fasting state,10 11 12 13 15 and differences in assay for FVII:C used. These latter explanations, however, seem unlikely for this study because subjects were not fasting and the same FVII:C assay was used as that used for studies previously demonstrating genotype-environmental correlations. Our findings therefore suggest that FVII genotype influences FVII:C levels directly and not through other confounding factors.
We have shown that FVII:C levels are influenced by the Msp I and promoter genotypes in stroke, as demonstrated in patients with IHD, patients with non–insulin-dependent diabetes mellitus, and healthy individuals,10 11 12 13 14 15 16 and that this association remains both at the time of stroke and at 3 months after the acute event. These associations are independent of confounding factors and are not sex specific, contrary to the finding in a previous study of Asian Indians.13
Since FVII has been suggested to be a risk factor for vascular disease and levels of FVII are strongly influenced by genetic parameters, it has been suggested that these genetic polymorphisms may also be risk factors for disease. However, a recent study of FVII:C, Msp I genotype, and IHD found no association of FVII or genotype and IHD,36 and a similar study of IHD patients in our department also found no association.37 The Northwick Park Heart Study found FVII to be a risk factor for fatal events only,4 which could suggest that FVII is a risk factor for myocardial infarction independent of the extent of atherosclerosis. The lack of an association of FVII:C with mortality from stroke or with stroke type, along with recent studies of IHD and FVII:C,36 37 seems to suggest that FVII:C is not the powerful risk factor for vascular disease it was first thought to be. However, all of these studies have been retrospective, with older patients who certainly had some degree of vascular disease, and early mortality from stroke may have led to exclusion of subjects with the most severe strokes. Additionally, case-control studies and hospital-based studies have limitations in that many cases will inevitably be missed, severely ill patients may die before they reach the hospital, and volunteer control subjects may not be true reflections of the general population. A larger study may have greater statistical power to detect a significant result, but in light of our result in a relatively large study population and other similar studies,36 37 it seems unlikely that FVII genotype has a true biological role in vascular disease. A prospective study would therefore be needed to determine whether FVII:C or FVII genotypes were associated with an increased risk of stroke.
It has also been suggested that FVII may be released in response to vascular damage; however, a highly significant reduction in FVII levels was observed in this study at the time of stroke compared with 3 months later. FVII:C levels were significantly lower at the time of stroke than in healthy control subjects but not 3 months later. Analysis of each genotype group separately showed that these differences occurred at a significant level only in the M1 homozygote and A homozygote groups. This could suggest that since M2 carriers have reduced secretion of FVII compared with M1 homozygotes, the fall in FVII levels is only obviously seen in the M1 homozygotes. An acute-phase reduction in cholesterol has been observed in stroke,38 and therefore the finding of a reduction in FVII may relate to the fall in cholesterol also seen in this patient group. Indeed, an association of FVII with stroke may be masked by correlations of FVII:C with hypercholesterolemia, hypertriglyceridemia, and increased BMI. Likewise, the clustering of all these factors may govern development of CVD rather than individual factors.
In conclusion, this study suggests that neither FVII:C levels nor FVII genotypes are independently involved in the pathogenesis or outcome of stroke and that FVII is in fact reduced at the time of stroke compared with 3 months after stroke and compared with healthy control subjects. A prospective study may determine the state of FVII:C levels before stroke and show whether FVII:C levels or FVII genotypes are associated with a further increased risk of stroke.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|IHD||=||ischemic heart disease|
|PACI||=||partial anterior circulation infarction|
|PCR||=||polymerase chain reaction|
|POCI||=||posterior circulation infarction|
|TACI||=||total anterior circulation infarction|
This study was supported by the United Leeds Hospitals Special Trustees and the Stroke Association. We are grateful to Professor Tom Meade for assaying the FVII levels.
- Received August 2, 1996.
- Revision received December 19, 1996.
- Accepted December 21, 1996.
- Copyright © 1997 by American Heart Association
Broze GJ Jr, Majerus PW. Purification and properties of human coagulation factor VII. J Biol Chem. 1980;255:1242-1247.
Grant PJ, Prentice CRM. Coronary artery disease. In: Bloom AL, Forbes CD, Thomas DP, Tuddenam EDG, eds. Haemostasis, Thrombosis. 3rd ed. Edinburgh, Scotland: Churchill Livingstone; 1994:1231-1254.
Heinrich J, Balleisen L, Schulte H, Assmann G, Van de Loo J. Fibrinogen and factor VII in the prediction of coronary risk: results from the PROCAM study in healthy men. Arterioscler Thromb. 1994;14:54-59.
Balleisen L, Schulte H, Assman G, Epping PH, Van de Loo J. Coagulation factors and the progress of coronary heart disease. Lancet. 1987;2:461.
Green F, Kellener C, Wilkes H, Temple A, Meade T, Humphries S. A common genetic polymorphism associated with lower coagulation factor VII levels in healthy individuals. Arterioscler Thromb. 1991;11:540-546.
Lane A, Cruickshank JK, Mitchell J, Henderson A, Humphries S, Green F. Genetic and environmental determinants of factor VII coagulant activity in ethnic groups at differing risk of coronary heart diseases. Atherosclerosis. 1992;94:44-50.
Humphries SE, Lane A, Green FR, Cooper J, Miller GJ. Factor VII coagulant activity and antigen levels in healthy men are determined by interaction between factor VII genotype and plasma triglyceride concentration. Arterioscler Thromb. 1994;4:193-198.
Saha N, Liu Y, Heng CK, Hong S, Low PS, Tay JSH. Association with factor VII genotype with plasma factor VII activity and antigen levels in healthy Indian adults and interaction with triglycerides. Arterioscler Thromb. 1994;14:1923-1927.
Arbini AA, Bauer KA. Reduced plasma factor VII coagulant activity due to the Arg353Gln polymorphism in the factor VII gene results from defective secretion. Blood. 1994;86:86. Abstract.
Mariani G, Marchetti G, Arcieri P, Bernardi F. The role of factor VII gene polymorphisms in determining FVII activity and antigen plasma level variation. Blood. 1994;86:86. Abstract.
Catto A, Carter AM, Barrett J, Stickland M, Bamford J, Davies A, Grant PJ. Angiotensin-converting enzyme insertion/deletion polymorphism and cerebrovascular disease. Stroke. 1996;27:435-440.
Catto A, Carter A, Ireland H, Bayston TA, Philippou H, Barrett J, Lane DA, Grant PJ. Factor V leiden gene mutation and thrombin generation in relation to the development of acute stroke. Arterioscler Thromb Vasc Biol. 1995;15:783-785.
Miller GJ, Stirling Y, Esnouf MP, Heinrich J, van de Loo J, Kienast J, Wu KK, Morrissey JH, Meade TW, Martin JC, Cooper JA, Finch A. Factor VII-deficient substrate plasmas depleted of protein C raise the sensitivity of the factor VII bio-assay to activated factor VII: an international study. Thromb Haemost. 1994;71:38-48.
Dennis M, Bamford JM, Sandercock P, Molyneux A, Warlow C. Incidence of transient ischemic attacks in Oxfordshire, England. Stroke. 1989;20:789-794.
Shinton R, Beevers G. Meta analysis between cigarette smoking and stroke. Br Med J. 1989;298:789-794.
Lane A, Green F, Scarabin PY, Nicaud V, Bara L, Humphries S, Evans A, Luc G, Cambou JP, Arveiler D, Cambien F. Factor VII Arg/Gln353 polymorphism determines factor VII coagulant activity in patients with myocardial infarction (MI) and control subjects in Belfast and France but is not a strong indicator of MI risk in the ECTIM study. Atherosclerosis. 1996;119:119-127.
Heywood DM, Ossei-Gerning N, Grant PJ. Association of factor VII:C levels with environmental and genetic factors in patients with ischaemic heart disease and coronary atheroma characterised by angiography. Thromb Haemost. 1996;2:161-165.
Butterworth RJ, Bath PMW, Marshall WJ. Measurements of lipid profiles. Cerebrovasc Dis. 1995;5:230. Abstract.