Levels of intrinsic coagulation factors and the risk of myocardial infarction among men: opposite and synergistic effects of factors XI and XII

Myocardial infarction occurs when a thrombus evolves on a ruptured atherosclerotic plaque followed by vessel occlusion.¹  This exposes tissue factor to blood which triggers the extrinsic system of thrombin generation by the formation of tissue factor (TF)–factor VIIa complex and induces platelet adhesion, activation, and aggregation. The TF-factor VIIa complex activates factor X and factor IX, leading to the generation of thrombin and formation of fibrin.²  Once thrombin is formed, further thrombin is generated by activation of factor V, factor VIII, and factor XI.³  Thrombin generation also leads to activation of thrombin activatable fibrinolysis inhibitor (TAFI) resulting in down-regulation of fibrinolysis.^4,5  Inhibition of factor XI in an experimental thrombosis model in rabbits leads to enhanced thrombolysis, thereby showing an antifibrinolytic effect of factor XIa.⁶  Factor XI inhibition in baboons does not prevent thrombus initiation, but it does seem to reduce thrombus growth rate, overall thrombus mass, and thrombo-occlusion, thereby leading to more favorable outcomes.⁷  Additionally, factor XI-deficient mice were protective from an induced arterial occlusion, suggesting that therapeutic inhibition of factor XI may be a reasonable strategy for treating or preventing thrombus formation.⁸  Thus, factor XI acts as a procoagulant and an antifibrinolytic factor, and dysregulation may increase the risk of thrombosis. However, factor XI deficiency does not confer protection against myocardial infarction.⁹  Elevated levels may even increase the risk,¹⁰  similarly to the increased risk of venous thrombosis.¹¹ 

Factor XI is part of the intrinsic coagulation system and can be activated by factor XII (Hageman factor) independently from thrombin.¹²  The traditional view is that low levels of factor XII increase the risk of venous thrombosis. Indeed, in one study factor XII levels were lower in patients who had a venous thrombosis compared with healthy blood donors,¹³  although in a large casecontrol study we found no relation between factor XII levels and venous thrombosis.¹⁴  Several case reports pointed out the occurrence of myocardial infarction in persons with factor XII deficiency.^15,16  However, the few studies investigating the risk of myocardial infarction in persons with decreased levels gave conflicting results.^10,17  Although coagulation factors such as fibrinogen and factor VII have been investigated intensively, the role of the intrinsic system in the development of thrombosis, including myocardial infarction, remains unclear.

In the Study of Myocardial Infarctions Leiden (SMILE) we investigated factor XI and factor XII coagulant activity in 560 men with a nonfatal myocardial infarction and 646 control subjects from the Netherlands and calculated risks of myocardial infarction with increasing quintiles of activity. Additionally, the roles of factor VIII and factor IX have been investigated as well. We extensively assessed the relationship of factor XI and factor XII with other markers of risk and adjusted the risk of myocardial infarction for cardiovascular risk factors that were found to be associated with these factors.

Details of the Study of Myocardial Infarctions Leiden (SMILE) have been described elsewhere.¹⁸  Briefly, patients were men younger than age 70 with a first myocardial infarction that occurred between January 1990 and January 1996. The control group also consisted of men without a history of myocardial infarction and was frequency matched on age to the patients. Both patients and control subjects were born in the Netherlands. They completed a questionnaire and an interview took place prior to blood draw (see next paragraph). Questions referred to (former) smoking habits and alcohol use, diabetes, and current use of medications. In addition, for patients, diabetes and medication use prior to myocardial infarction was retrieved from discharge letters. A person was classified as having hypertension or hypercholesterolemia if he was taking prescription drugs for these conditions. Blood pressure was measured after a rest of at least 10 minutes with the person sitting in an upright position. The quetelet index was derived by dividing weight (in kilograms) by squared height (in meters), and obesity was defined as an index of 30 kg/m² or above. All participants gave informed consent, and the study protocol was approved by the Institutional Review Board of the Leiden University Medical Center.

Fasting blood samples were drawn from the antecubital vein in Sarstedt Monovette tubes (Sarstedt, Nümbrecht, Germany) and obtained between July 1994 and February 1997. Blood in the first tube was allowed to clot, and the serum was used for measuring total cholesterol, high-density lipoprotein (HDL)–cholesterol, and triglyceride levels. Total cholesterol and triglyceride concentrations were measured using enzymatic assays adapted to a Hitachi 747 (Boehringer Mannheim, Mannheim, Germany), and HDL-cholesterol concentration was measured on a Hitachi 911 (Boehringer Mannheim). Blood taken in 0.109 mol/L trisodium citrate was centrifuged for 10 minutes at 3000g at room temperature. The citrated plasma was divided into aliquots in multiple tubes and immediately stored at –80°C. Factor XI activity (XIc), factor XII activity (XIIc), and factor IX activity (IXc) were measured in one-stage coagulation assays on a Behring Coagulation System (Dade Behring, Marburg, Germany) with protocols and reagents from the manufacturer. The factor XI and XII coagulant assays were standardized using commercial plasma (Standard Human Plasma from Dade Behring). The assigned values of the manufacturer were used, because at the time of analysis an international standard for either factor was not available. The factor IX coagulant assay was standardized with pooled plasma from more than 150 healthy hospital workers that was calibrated with 2 international standards (NIBSC 99/826 and NIBSC 01/618). Intra-assay variations for the coagulation assays were 6%, 6%, and 5% for factor XIc, XIIc, and IXc, respectively. In 6 patients and 6 control subjects factor IXc was not measured. For the analyses on factor IX levels, persons using oral anticoagulation medication were excluded (134 patients). Factor VIII (VIIIc) was measured in 2 dilutions by a one-stage clotting assay with factor VIII-deficient plasma and automated activated partial thromboplastin time (Organon Teknika, Boxtel, The Netherlands) on a STA (Diagnostic Stago; Boehringer Mannheim). Levels are expressed as IU/dL. Pooled normal plasma, calibrated against the WHO standard factor VIIIc (91/666) was used as a reference. In one control subject factor VIIIc was not measured. The median time between myocardial infarction and blood collection was 2.6 years, with a minimum of 6 months. To evaluate whether coagulation factors decreased after myocardial infarction in our study, we stratified the elapsed time since myocardial infarction into 1-year categories. Elapsed time since myocardial infarction did not relate to any coagulation factor (data not shown).

Means were calculated with standard deviations (SDs) and the appropriate 95% confidence intervals (CIs). Differences between mean values were tested by means of analysis of variance (ANOVA). Quintiles of coagulation factors were defined on the basis of the distribution among control subjects. The lowest quintile was used as a reference category for calculating odds ratios (ORs). Odds ratios were calculated as an approximation of relative risks. Unconditional logistic regression was used to adjust for age (ORadj) and other cardiovascular risk factors. The 95% confidence intervals (95% CIs) of odds ratios were calculated based on the standard errors from the logistic model. Quintiles of total cholesterol, HDL-cholesterol, and triglyceride levels were also defined on the basis of the distribution among control subjects. However, in the logistic regression model the 3 lipid levels were included as continuous variables. Triglyceride levels used in the model were 10log-transformed. SPSS for Windows version 12.0.1 (SPSS, Chicago, IL) was used for all statistical analyses.

Detailed characteristics of patients and control subjects have been described before.¹⁸  The mean age of patients was 56.2 years (SD, 9.0 years) and of control subjects 57.3 years (SD, 10.8 years). Risk factors as smoking, obesity, diabetes, hypertension, and hypercholesterolemia were more often found in patients than in control subjects, with the most striking contrasts in younger persons (Table 1).

Mean factor XIc among patients with myocardial infarction was 113.0% (CI, 111.5%-114.4%) and varied between 55% and 160%. For control subjects mean factor XIc was 109.8% (CI, 108.2%-111.3%), with a corresponding range between 56% and 194%. Factor XIc in patients was 3.2% (CI, 1.1%-5.4%) higher compared with control subjects. A 10% increase in factor XIc was associated with a 1.09-fold increased in risk of myocardial infarction (ORadj, 1.09; 95% CI, 1.02-1.15). The distribution of patients and control subjects by quintiles of factor XIc is shown in Table 2. The risk of myocardial infarction was increased for each quintile of factor XI compared with the lowest quintile (levels equal or below 94%). In other words, the risk was decreased for men in the lowest quintile compared with those with higher levels. The risk of myocardial infarction adjusted for age for men in the highest quintile compared with those in the lowest quintile was 1.8-fold increased (ORadj, 1.8; 95% CI, 1.2-2.7). For men younger than age 50 even higher risks were found, 3- to 4-fold increased compared with the lowest quintile. Further adjustment for cardiovascular risk factors and factor XIIc increased the odds ratios. Excluding persons who were using aspirin at time of blood draw (128 patients and 36 control subjects) in our overall analyses did not materially change the results, the difference in factor XIc between patients and control subjects remained the same (3.2% [CI, 0.9%-5.5%]). Excluding 134 patients who were using anticoagulants at the time of blood draw increased the difference between patients and control subjects to 5.6% (CI, 3.2%-7.9%). Consequently, the relative risk of myocardial infarction became even more pronounced for each quintile increase of factor XI compared with the lowest quintile and showed a dose-response relationship.

Mean factor XIIc among patients with myocardial infarction was lower than in control subjects, respectively, 93.0% (CI, 91.3%-94.6%) and 98.6% (CI, 97.0%-100.2%), with a difference of 5.6% (CI, 3.3%-7.9%). The range among patients varied from 23% to 139%, whereas the range among control subjects varied from 37% to 164%. A 10% increase in factor XIIc was associated with a 0.87-fold decreased risk of myocardial infarction (ORadj, 0.87; 95% CI, 0.82-0.92). The risk of myocardial infarction was decreased for men with factor XIIc levels in the upper 2 quintiles each with levels above 107% (Table 2). The odds ratio of myocardial infarction for men in the highest quintile versus those in the lowest quintile was 0.4 (ORadj, 0.4; 95% CI, 0.2-0.5). Among men younger than age 50 the risk seemed to be decreased only in those with factor XIIc of 119% or more (ORadj, 0.7; 95% CI, 0.4-1.5). Further adjustment for cardiovascular risk factors and factor XIc slightly changed the odds ratios. Excluding persons who were using aspirin at the time of blood draw in our overall analyses did not materially change the results, the difference in factor XIc between patients and control subjects remained the same (5.5 [CI, 3.0-8.1]). However, excluding anticoagulant users in our overall analyses decreased the difference in factor XIIc between patients and control subjects to 3.5% (CI 0.9%-6.0%). Persons with factor XIIc in the highest quintile remained at a decreased risk of myocardial infarction (ORadj, 0.5; 95% CI, 0.4-0.8).

Mean factor VIIIc among patients with myocardial infarction was 126.3 IU/dL (CI, 123.6-129.1 IU/dL) and varied between 51 and 277 IU/dL. For control subjects mean factor VIIIc was 123.4 IU/dL (CI, 120.8-126.0 IU/dL), with a corresponding range between 53 and 273 IU/dL. Factor VIIIc in patients was 2.9 IU/dL (CI, –0.9 to 6.7 IU/dL) higher compared with control subjects. The distribution of patients and control subjects by quintiles of factor VIIIc is shown in Table 3. The risk of myocardial infarction appeared to be increased for most quintiles compared with the lowest quintile, and overall increased slightly after adjustment for cardiovascular risk factors and factor IXc. Mean factor IXc among patients with myocardial infarction was higher than in control subjects, respectively, 96.6% (CI, 95.5%-97.7%) and 93.2% (CI, 92.2%-94.2%), with a difference of 3.4% (CI, 1.9%-4.9%). The range among patients varied from 35% to 135%, whereas the range among control subjects varied from 32% to 134%. The risk of myocardial infarction was increased for men in each quintile compared with the lowest quintile, and even more increased for men younger than age 50 (Table 3). Overall, adjustment for cardiovascular risk factors and factor VIIIc slightly changed the odds ratios.

In control subjects no associations were found between factor XIc and smoking, alcohol use, or diabetes (Table 4). Obese persons with a quetelet index of 30 kg/m² or above had higher levels of factor XIc compared with lean persons with a quetelet index below 20 kg/m². No associations were found between factor XIIc and any of these cardiovascular risk factors. Systolic and diastolic blood pressures were not associated with factor XIc or factor XIIc. Hypertension as determined by medication use was not associated with clotting factor levels; the difference in factor XIc between 107 men with hypertension versus 539 without hypertension was –2.2% (CI, –6.4% to 1.9%) and in factor XIIc –1.9% (CI, –6.4% to 2.5%). Factor XIc was associated with total cholesterol levels and with triglyceride levels in control subjects (Table 5). Factor XIc increased with every increase in quintile of total cholesterol and triglyceride level. No associations were found with HDL-cholesterol levels. Similar results were found for factor XIIc. Hypercholesterolemia as determined by medication use was not associated with levels; the difference in factor XIc between 11 men with hypercholesterolemia versus 635 men without hypercholesterolemia was 0.2% (CI, –11.7% to 12.2%) and in factor XIIc 8.2% (CI, –4.6% to 20.9%).

Among control subjects factor XIc was associated with factor XIIc with every 1% increase in factor XIIc factor XIc increased with 0.4% (95% CI, 0.3%-0.5%). To disentangle the effect of factor XIIc from factor XIc, we looked at the effect of each factor adjusted for the other. Because univariate effects were opposite, and levels were associated in the same direction, adjustment led to higher risk estimates, that is, 3-fold increased risk (OR, 3.2; 95% CI, 2.1-4.9) for the highest quintile of factor XI, and 80% reduced risk (OR, 0.2; 95% CI, 0.1-0.4) for the highest quintile of factor XII (Table 2). Further adjustment for total cholesterol, HDL-cholesterol, and triglyceride levels and quetelet index did not much affect the odds ratios, neither did adjustment for factor IXc or factor VIIIc.

Tertiles were calculated based on the distribution among control subjects. Overall, men in the highest tertile of factor XIc had a 1.5-fold increased risk of myocardial infarction compared with men in the lowest tertile (ORadj, 1.5; 95% CI, 1.1-2.0). In contrast, men in the highest tertile of factor XIIc had a 60% decreased risk (ORadj, 0.4; 95% CI, 0.3-0.6) compared with those in the lowest tertile. The highest risk of myocardial infarction was found among men in the highest tertile of factor XIc and the lowest tertile of factor XIIc (ORadj, 6.4; 95% CI, 2.2-18.0) compared with men with the lowest factor XIc and highest factor XIIc tertiles (Table 6). Further adjustment for cardiovascular risk factors increased the odds ratios.

Although positively associated, levels of factors XIc and XIIc had opposite and synergistic effects on the risk of myocardial infarction. Higher factor XIc increased the risk, whereas higher factor XIIc decreased the risk (ie, low factor XIIc increased the risk of myocardial infarction). Additionally, factor VIIIc and factor IXc increase the risk of myocardial infarction in men.

In our Study of Myocardial Infarctions Leiden (SMILE) high factor XIc increased the risk of myocardial infarction, which is in accordance with a Swiss case-control study that included 200 patients with a history of myocardial infarction.¹⁰  Factor XIc has procoagulant and antifibrinolytic effects; thus, an increased risk of myocardial infarction in the presence of high factor XIc was to be expected and is in line with studies on venous thrombosis.¹¹  However, the results of a recently published case-control study that included young women did not find a difference in level between 200 women with a first myocardial infarction and control subjects.¹⁹ 

High factor XIIc decreased the risk of myocardial infarction in our study. In the second Northwick Park Heart Study among 1153 men aged 50 to 61 years in which only 104 coronary heart disease events occurred during the 7.8 years of follow-up, plasma factor XIIa was an independent risk factor. Men in the upper third had a higher risk compared with those in the middle tertile, but also those in the lower third seemed to have a higher risk.²⁰  In a second analysis that included 1745 men high incidence rates were only found for those with high factor XIIa, with similar rates in the first and second third.²¹  In contrast, in the West of Scotland Coronary Prevention Study factor XIIa antigen did not predict the risk of a coronary event,²²  neither did factor XIIc and antigen in other smaller studies.^10,23  Our results could also be interpreted such that low factor XIIc increased the risk of myocardial infarction, which is in correspondence to the lower levels found in patients with venous thrombosis compared with healthy blood donors,¹³  although some studies had different results.¹⁴ 

Factor VIIIc increased the risk of myocardial infarction about 1.4-fold or more, although not dose dependently. In the Risk of Arterial Thrombosis In relation with Oral Contraceptive use (RATIO) study, in which factor VIIIc was measured in the exact same way, the risk was more clearly increased among young women. In that particular study the risk decreased after adjustment for von Willebrand factor (VWF).¹⁹  Follow-up studies have also found associations between factor VIII and ischemic heart disease in men^24-26  and a risk which reduced after adjustment for VWF.²⁶  Factor IXc increased the risk of myocardial infarction, even more strongly among young men. This finding is in line with the results of the RATIO study among young women.¹⁹ 

The risk of myocardial infarction was highest for men with both high factor XIc and low XIIc. This would suggest that the effect of factor XII on risk is not by activation of factor XI, but by an alternative mechanism. Previous studies have suggested that plasma levels of factor XIIa are related to plasma levels of cholesterol and triglyceride,^23,27,28  but not HDL cholesterol.²²  The present study corroborates these findings. However, we did not find any associations among factor XIIc and obesity, smoking, alcohol use, diabetes, or blood pressure, which is in contrast to a study among 2464 men aged 51 to 62 years in which positive associations were found between factor XIIa with body mass index, smoking, and blood pressure.²⁷  Even though associations between factor XIIc and lipid levels clearly exist, the effect of factor XIIc on the risk of myocardial infarction remained present for men in the highest quintiles after adjusting for lipid levels and other cardiovascular risk factors, indicating that factor XIIc itself seems to have an effect on myocardial infarction.

Our results are in line with the current concept of coagulation and that in contrast to factor XI, factor XII is not involved in physiologic thrombin and fibrin formation. Patients deficient for factor XII do not bleed, and, although factor XII is essential for contact activation and the activated partial thromboplastin time, it has been assumed that factor XII does not contribute to clot formation. In contrast, factor XI is involved in normal thrombin and fibrin formation, and elevated levels of factor XI have previously been found to be associated with an increased risk of venous thrombosis.¹¹  Also, factor XI-deficient mice were protected in models of arterial thrombosis.⁸ 

However, a recent report showed that factor XII–deficient mice were protected in a model of lethal pulmonary embolism and had a defective arterial thrombosis formation, suggesting that factor XII also contributes to thrombin and fibrin formation.²⁹  These results are in contradiction with our epidemiologic findings, in which we found an increased risk of myocardial infarction with low factor XIIc. One of the explanations might be that humans have evolved to a different mechanism of coagulation than mice. The presence of atherosclerosis plays an important role in the occurrence of myocardial infarction in humans, whereas vessels of the mice used in the models clearly differ in this aspect. An alternative explanation is that the potential functions of factor XII only become apparent at certain concentrations of factor XII. In our epidemiologic study, the concentration of factor XIIc in the subjects was not below 23%. In the mouse study, only homozygote-deficient mice were protected against arterial thrombosis. The heterozygote mice showed similar results compared with mice containing normal levels of factor XII, with nearly all vessels containing thrombi, a slightly lower time to occlusion and slightly more platelets per surface area.²⁹  Because the used models were not designed or sensitive for enhanced thrombus formation, it can therefore not be excluded that heterozygote mice had enhanced thrombus formation. Taken together and on the basis of our findings, we can then speculate that lower levels of factor XII result in enhanced thrombus formation by decreased fibrinolysis or effects on inflammation and complement activation.^30,31 

By necessity we studied patients who survived the myocardial infarction. It cannot be excluded that patients who died during the acute phase of the myocardial infarction had higher or lower coagulation levels. We think this is unlikely, however, because survival of a person after myocardial infarction is influenced by other factors, such as patient-induced delay and delay in providing effective assistance, which affect the timeframe from onset of symptoms to the start of interventions such as thrombolytic therapy.³²  Several other factors influencing 30-day mortality are the level of systolic blood pressure, heart rate, Killip class, and localization of myocardial infarction.³³  It does not seem likely that these 4 coagulation factors would play a role in this.

In conclusion, our results show that factors XIc and XIIc have an opposite and synergistic effect on the risk of myocardial infarction. If factor XIIc may be a new target for antithrombotic therapies as suggested before²⁹  remains to be established. Both factor VIIIc and factor IXc increase the risk of myocardial infarction among men.

Contribution: C.J.M.D. designed the overall study, performed the data collection and statistical analyses, and drafted the manuscript; J.C.M.M. designed the present study and was responsible for execution of the assays and participated in the writing of the manuscript; and F.R.R. designed the overall study and critically reviewed the analyses and the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

We thank Dr T. Renné, University of Würzburg, Germany, for sharing his data on the factor XII–deficient mice before publication. We thank the cardiologists of the departments of cardiology, Leiden University Medical Center; the general hospital Diaconessenhuis Leiden and F. J. M. van der Meer, head of the Leiden Anticoagulant Clinic, for their kind cooperation; and J. H. M. Souverijn for his assistance with the lipid measurements. We thank T. Visser for drawing blood samples and the personnel of the Department of Experimental Vascular Medicine of the Academic Medical Center for performing the laboratory measurements. We thank J. J. Schreijer and I. de Jonge for their secretarial and administrative support. We also express our gratitude to all individuals who participated in the Study of Myocardial Infarctions Leiden.

This work was supported by the Netherlands Heart Foundation (grant no. 92.345).