Frits R. Rosendaal, Department of Clinical Epidemiology, Leiden University Medical Center, PO Box 9600, 2300 RC Leiden, the Netherlands. Tel.: +31 71 5264037; fax: +31 71 5266994. E-mail: email@example.com
Summary. Background: The aim of this study was to examine whether genetic predisposition to high levels of coagulation factors influences the risk of developing fatal and non-fatal arterial cardiovascular events in men with a first myocardial infarction (MI). Methods: We performed a cohort study among 542 MI patients with a mean age of 56 years (range 32–70 years) at the time of the event. All of the men had a first MI between 1990 and 1996 and were followed until 1 September 2004. DNA was analyzed for polymorphisms of fibrinogen, prothrombin (factor II), factor V, factor VII and plasminogen activator inhibitor type 1, all of which are associated with gain of function of the protein. We collected information from hospital files and general practitioners on the occurrence of major arterial events. Results: In total, 254 major arterial cardiovascular events occurred during a median follow-up period of 11 years (range 0.2–15 years). The point estimates of the relative rates (RRs) of these events for the variant genotypes were all between 0.7 and 1.1 except for the prothrombin 20210A mutation: RR 1.8 (95% confidence interval 0.8–4.1). Conclusion: These findings suggest that there is no association between coagulation factor polymorphisms, previously associated with plasma levels, and the risk of recurrent cardiovascular events.
Most myocardial ischemic events occur through plaque ruptures that precipitate the formation of an occluding thrombus . The growth and stability of a thrombus may be affected by an increased activity of the coagulation system or a decreased activity of the fibrinolytic system . The association between plasma levels of coagulation and antifibrinolytic factors and the risk of first myocardial ischemic events has been studied extensively.
The Northwick Park Heart Study  was among the first studies to show that high levels of plasma fibrinogen, as well as of other coagulation factors, were associated with an increased risk of cardiovascular events. Several subsequent studies confirmed a positive association between plasma levels of coagulation factors and the risk of arterial cardiovascular diseases (CVDs) [4–9]. However, studies focussing on polymorphisms of these coagulation factors did not find an association with CVD [10–14]. Apparently, the observed associations between plasma levels and disease were mainly explained by other cardiovascular risk factors. Indeed, risks were reduced or disappeared after adjustment for these other risk factors, and the remaining risk may well be explained by residual confounding [5,15]. Coagulation factor polymorphisms, which are associated with increased plasma levels, are defined at birth and cannot be influenced by other factors, so the causal relationship between these factors and CVD can be studied in coagulation factor polymorphisms, without confounding by extraneous factors . Recent meta-analyses suggest that factor (F) V and prothrombin gene polymorphisms are associated with an increased risk for CVD [17,18]. In contrast, several multi-locus candidate gene polymorphisms for other coagulation factors appear not, or only weakly, associated with the risk of first myocardial infarction (MI) [19,20]. The causal role of elevated plasma levels in CVD remains unclear. Some have suggested that plasma levels of coagulation factors do increase the risk for recurrent events [9,21,22].
Our aim was to examine whether genetic predisposition to high levels of procoagulant and antifibrinolytic factors influences the risk of development of recurrent cardiovascular events among men.
We performed a cohort study among 560 men with a mean age of 56 years (range 32–70 years). All men were consecutive patients at the Leiden University Medical Center and the Diaconessen Hospital Leiden, the Netherlands, who had a first MI between 1990 and 1996 and were included in the Study of Myocardial Infarction Leiden (SMILE). Details of the study design have been reported elsewhere . At baseline, all patients completed a questionnaire concerning the presence of cardiovascular risk factors such as smoking habits. All questions referred to the period before the first MI. Body mass index (BMI) was calculated by dividing weight (kg) by squared height (m2). Medication use and history of diabetes were retrieved from discharge letters from the hospital. A person was classified as hypertensive or hypercholesterolemic when he was taking prescription drugs for these conditions prior to the MI. All patients were living in Leiden or in nearby communities. The study was approved by the Medical Ethics Committee of the Leiden University Medical Center, and all patients gave written informed consent.
The endpoint in the present study was recurrent major arterial cardiovascular events. We defined recurrent events as MI, unstable angina pectoris (UAP), percutaneous transluminal coronary angiography (PTCA), coronary artery bypass graft (CABG) surgery and death due to coronary events. We collected follow-up information from hospital files. When follow-up information was missing from hospital files, information was obtained from the patient’s general practitioner or by a questionnaire sent to the patient. September 1st 2004 was considered the end date of follow-up for those without an event. Vital status was verified in hospital files, in municipal registries, or by contacting general practitioners. For all deceased patients, death certificates were requested at the National Central Bureau of Statistics in order to obtain causes of death. All calculations concerning causes of death are based on this database.
MI was characterized by at least three of the following criteria: chest pain, ischemic electrocardiogram, elevated creatine kinase (CK; >170 units L−1) and CK-MB (>10% of CK), and a positive troponin test. UAP was characterized by at least two of the following criteria: hospitalization due to chest pain occurring in rest, normal cardiac enzymes, and negative troponin test. Revascularization therapies (PTCA and CABG) were considered as a recurrent event when the procedure took place at least 31 days after the first MI. Causes of death were classified into three groups according to ICD-10-CM codes: coronary artery disease (CAD) death, codes I20–I25; other cardiac death, codes I00–I02, I05–I09, I10–I15 or I30–I53; and non-cardiac causes of death including all other ICD-10-CM codes. Death from a recurrent event was defined as dying from CAD after the first event and when no other endpoint had occurred first.
Blood sample analysis
A morning fasting blood sample was collected at baseline from each participant. The median time between first MI and blood sampling was 2.6 years (range 0.2–6.0 years). DNA was extracted and polymorphisms were genotyped after amplification of relevant DNA regions as described earlier . Polymorphisms in six coagulation factor genes encoding fibrinogen α and β, prothrombin (FII), FV, FVII, and plasminogen activator inhibitor type 1 (PAI-1) were determined. Restriction enzyme digestion was used for the detection of the α-fibrinogen gene (FGA) polymorphism TaqI (position 59855 in GB:AC107385.4; common allele coded as T1 and rare allele as T2), the β-fibrinogen gene (FGB) polymorphisms BclI (position 48980 in GB:AC107385.4; common allele coded as B1 and rare allele as B2) and HaeIII [-455G/A, reference SNP cluster identifier (rs) number 1800790; common allele coded as H1 and rare allele as H2], and the PAI-1 promoter 4G5G polymorphism (rs1799768). This method was also used for the detection of the FVII polymorphism R353Q (rs6046), R coding for arginine (353Arg) and the rare allele Q coding for glutamine (353Gln), and for the FV Leiden mutation (G1691A, rs6025). The detection of the G20210A allele (rs1799963) of the prothrombin gene was performed as described previously . The PAI-1 4G5G polymorphism was analyzed in the first 331 patients who gave a blood sample.
Descriptive data are expressed as frequencies or means. Incidence rates were calculated by dividing the number of recurrent events (the endpoint) by the sum of the total number of person-years each patient contributed to the follow-up period. Person-years were calculated from the date of diagnosis of first MI until the first recurrent event or the last date of follow-up or end-date of study, whichever occurred first. Cumulative survival was estimated according to the method of Kaplan and Meier. Cox proportional hazard models were used to determine the rate ratio for occurrence of recurrent events according to the different genotypes. In an additional Cox proportional hazard model we adjusted for age, hypertension, hypercholesterolemia and diabetes, BMI, and smoking habit at the start of follow-up. The reference groups are homogeneous wild-type carriers (common alleles) and the genotype associated with low plasma levels (the FGB polymorphism BclI). SPSS for Windows version 12.0 (SPSS Inc., Chicago, IL, USA) was used for all statistical analysis.
A total of 560 patients were included. Five patients did not consent to participate in the follow-up study, and another 13 patients were found to have had an ischemic event before the initial MI. All of these patients were excluded from the present study. Follow-up data were collected for the remaining 542 patients and only four patients were lost to follow-up: three moved out of the country, and vital status could not be obtained for one. Patients had a mean age of 56 years (range 32–70 years) and a median follow-up of 11 years (range 0.2–15 years). Baseline characteristics of the patients are shown in Table 1.
Table 1. Baseline characteristics of 542 men with a first myocardial infarction
*Obesity is defined as a body mass index exceeding 30 kg m−2. †A person was classified as having hypertension or hypercholesterolemia if he was taking specific prescription drugs.
Age, years (range)
Current smokers (%)
Overall, 254 men out of 542 (47%) developed at least one recurrent event. Of these 254 men, 89 (35%) developed non-fatal recurrent MI, 60 (24%) suffered from UAP, 87 (34%) received revascularization therapy more than 31 days after the first MI, and 18 (7%) died from a recurrent CAD. A total of 99 patients died during follow-up (Table 2).
Table 2. Recurrent events
Recurrent major arterial cardiovascular event
Non-fatal myocardial infarction
Unstable angina pectoris
Fatal myocardial infarction
Death during total follow-up
Acute coronary event
Other cardiac causes
The β-fibrinogen polymorphisms HaeIII and BcllI, as well as the α-fibrinogen polymorphism TaqI, were not associated with recurrent events – HaeIII: H1H2 relative rate (RR) 1.1 [95% confidence interval (95% CI) 0.8–1.4], H2H2 RR 1.1 (95% CI 0.6–2.2); BclI: B1B2 RR 0.7 (95% CI 0.4–1.5), B2B2 RR 0.8 (95% CI 0.4–1.6); TaqI: T1T2 RR 0.8 (95% CI 0.6–1.0), T2T2 RR 1.0 (95% CI 0.6–1.6) (Table 3).
Table 3. Incidence and relative rate (RR) for recurrent major arterial cardiovascular events by genotype
Incidence (1000 person-years−1)
RR (95% CI)
RR (95% CI) adjusted*
*Adjusted for hypertension, hypercholesterolemia and diabetes before first myocardial infarction, body mass index, age at start of follow-up and smoking. †Reference category; ‡n (1691AA) = 1, n (1691AG) = 15; §n = 331. 95% CI, 95% confidence interval; FGA, α-fibrinogen gene; FGB, β-fibrinogen gene; PAI-1, plasminogen activator inhibitor type 1.
The prothrombin 20210A mutation appeared to be associated with recurrent events [RR 1.8 (95% CI 0.8–4.1)] for carriers of the A-allele compared with the homozygous wild-type genotype. Two out of six recurrent events were due to a recurrent MI. Carriers of the FV Leiden gene polymorphism did not have an increased risk for recurrent events [RR 1.0 (95% CI 0.6–1.6)]. FVII R353Q polymorphism was not associated with the risk for recurrent events [353RQ RR 1.0 (95% CI 0.7–1.3); 353QQ RR 1.0 (95% CI 0.3–4.2)], nor was carriership of the PAI-1 4G5G variant [4G5G RR 1.0 (95% CI 0.6–1.4), 4G4G RR 1.0 (95% CI 0.6–1.6)]. Adjustment for traditional risk factors of arterial CVD did not alter these findings.
Figure 1 and Table 4 show the survival curves of patients with increasing numbers of high-risk coagulation factor polymorphisms, divided into three groups. For each patient the number of high-risk polymorphisms was calculated: the heterozygous high-risk polymorphism counted as 1, the homozygous high-risk polymorphism counted as 2. All patients had two or more high-risk polymorphisms. For comparison, the three groups were based on equal numbers of patients per group. The first group (n = 158) consisted of patients with zero to four high-risk polymorphisms, the second group (n = 187) consisted of all patients with five high-risk polymorphisms, and the third group (n = 197) consisted of all patients with six to nine high-risk polymorphisms. No association was found between the number of high-risk polymorphisms in these patients and the occurrence of recurrent events.
Table 4. Number at risk and number of recurrent events in the three different groups in 14 years of follow-up
We investigated seven polymorphisms in coagulation and antifibrinolytic factors and their association with recurrent cardiovascular events. Except for prothrombin 20210A, none of the polymorphisms was associated with the risk of recurrent major arterial cardiovascular events. Because these polymorphisms are all associated with a gain of function (either in plasma levels or in stability), this suggests that slightly increased plasma concentrations of these coagulation and antifibrinolytic factors do not increase the risk of recurrent arterial thrombosis. Increased prothrombin levels, however, may be associated with recurrent events.
The two β-fibrinogen gene polymorphisms BclI and HaeIII are known to be associated with increased plasma fibrinogen [24,25]. High fibrinogen levels have repeatedly been associated with recurrent cardiovascular events . Previous studies found that carriership of the FV Leiden mutation [18,27,28], the FVII 353RQ polymorphism [11,29,30], the PAI-1 4G5G polymorphism  and PAI-1 plasma levels  may be associated with an increased risk of a first MI. Another study found an increased risk for cardiovascular death among carriers of the PAI-1 4G5G polymorphism . Recently, Moss et al.  found, in a large cohort of 1008 carriers of high-risk genotypes for CVD, a risk reduction of 11% for the development of recurrent coronary events. Except for fibrinogen, all associations with a first myocardial event remain controversial, and are weak at best. In our study we did not find an association with recurrent cardiovascular events, which occurred at a high frequency of 47% in all men with a first MI over a median follow-up period of 11 years.
The prothrombin 20210A mutation is a major determinant of plasma levels of prothrombin . A meta-analysis including 40 studies found that the prothrombin 20210A mutation was associated with a 1.3-fold increased risk of a first MI . We found a similar increased risk for a recurrent event. A previous prospective cohort study also suggested that carriers of the prothrombin 20210A mutation may have a worse prognosis than non-carriers after a first acute coronary syndrome . Our findings confirm this notion; however, the findings are based on very small numbers of patients.
The effects of single polymorphisms on the risk of recurrent cardiovascular events may be too small to be detected in a study with a relatively small sample size like ours. However, most of the point estimates were close to the null effect, suggesting that an association was truly absent. We saw no effect of multiple carriership of high-risk coagulation factor polymorphisms on the risk of recurrent CVD. No difference was found in the survival of the three groups, suggesting that even carriership of several of these high-risk polymorphisms does not influence the prognosis of these patients.
Some may argue that we should have restricted the endpoint to MI. We feel that all cardiovascular events in which atherothrombosis is involved should be classified as endpoints. Otherwise, misclassification of these events to the no-event group would lead to the underestimation of an existing effect. Additionally, our study population had a very long median follow-up of 11 years, in which only four patients were lost to follow-up, suggesting that our results were not affected by loss to follow-up. Furthermore, this study was restricted to men below the age of 70 years. As the median time between the first MI and blood sampling was 2.6 years, several men dying shortly after the first event were not included.
Anticoagulation therapy combined with aspirin after first ischemic events decreases the risk for a second event . Thus, the lowering of vitamin K-dependent coagulation factor levels influences the prognosis. Our baseline study population consisted of patients who had suffered a previous MI. As a result most, if not all, patients were treated with medication to minimize the risk of a recurrent event. It is plausible that this treatment modified a possible effect of plasma coagulation factor levels on the risk of recurrent cardiovascular events. If this is the case, the clinical relevance of determining coagulation and antifibrinolytic polymorphisms after the first ischemic event is low.
In conclusion, our findings suggest that polymorphisms associated with lifelong slightly increased coagulation and antifibrinolytic plasma levels are not associated with the risk of recurrent cardiovascular events among men who survive a first MI.
We would like to thank the cardiologists of the department of cardiology, Leiden University Medical Center, and the general hospital Diaconessenhuis Leiden and the general practitioners for their kind cooperation. We thank T. Visser for drawing blood samples and for performing the laboratory measurements, H. L. Vos for his contribution to the nomenclature of the polymorphisms and I. de Jonge for her administrative support. We also express our gratitude to all of the individuals who participated in the Study of Myocardial Infarction Leiden. This study was supported by the Netherlands Heart Foundation (grant no. 2003B248 and grant no. 92.345).
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.