C-Reactive Protein, Fibrin D-Dimer, and Incident Ischemic Heart Disease in the Speedwell Study
Are Inflammation and Fibrin Turnover Linked in Pathogenesis?
Abstract—Plasma levels of C-reactive protein (CRP, a marker of the reactant plasma protein component of the inflammatory response) and of fibrin D-dimer (a marker of cross-linked fibrin turnover) have each been associated in recent studies with the risk of future ischemic heart disease (IHD). Previous experimental studies have shown that fibrin degradation products, including D-dimer, have effects on inflammatory processes and acute-phase protein responses. In the Speedwell Prospective Study, we therefore measured CRP and D-dimer levels in stored plasma samples from 1690 men aged 49 to 67 years who were followed-up for incident IHD for an average of 75±4 months (mean±SD) and studied their associations with each other, with baseline and incident IHD, and with IHD risk factors. CRP and D-dimer levels were each associated with age, plasma fibrinogen, smoking habit, and baseline evidence of IHD. CRP was associated with D-dimer (r=0.21, P<0.00001). On univariate analyses, both CRP and D-dimer were associated with incident IHD. The incidence of IHD increased with CRP independently of the level of D-dimer (P=0.0002) and also increased with D-dimer independently of the level of CRP (P=0.048). In multivariate analyses, inclusion of D-dimer and conventional risk factors reduced the strength of the association between CRP and incident IHD; likewise, inclusion of CRP and conventional risk factors reduced the strength of the association between D-dimer and incident IHD. We conclude that although these respective markers of inflammation and fibrin turnover show modest association with each other in middle-aged men, they may have additive associations with risk of incident IHD. Further larger studies are required to test this hypothesis.
- Received June 27, 2000.
- Accepted September 29, 2000.
C-reactive protein (CRP), a marker of the reactant plasma protein component of the inflammatory response, has been associated with the risk of future ischemic heart disease (IHD), not only among patients with stable and unstable angina1 2 and high-risk subjects3 but also among population samples of apparently healthy middle-aged men4 5 and elderly men and women.6 A recent meta-analysis7 of these prospective studies found a combined risk ratio for IHD of 1.7 (95% confidence interval [CI], 1.4 to 2.1) for subjects in the top third of the distribution of CRP compared with the bottom third. There was no evidence of heterogeneity between these studies.
Fibrin D-dimer, a marker of cross-linked fibrin turnover, has also been shown in recent studies to be associated with the risk of future IHD in persons with and without baseline evidence of vascular disease.8 9 10 11 12 13 14 15 16 Local fibrin formation and lysis are part of the inflammatory response, and fibrin degradation products, including D-dimer, have been shown to have diverse effects on inflammatory processes and acute-phase responses, including neutrophil and monocyte activation; secretion of cytokines, including interleukin-6 and interleukin-1; and hepatic synthesis of acute-phase proteins, including fibrinogen and CRP.17 18 19 20
We therefore hypothesized that (1) plasma levels of CRP and D-dimer and their associations with incident IHD in the general population might be linked; (2) linkage might result from focal, vessel wall–related fibrin formation and lysis and an inflammatory response associated with unstable atherosclerotic plaque activity21 22 ; and (3) CRP and D-dimer might be related to IHD risk factors associated with thrombogenesis and inflammation, particularly cigarette smoking. We are unaware of previously published studies of CRP and D-dimer considered jointly. We tested these hypotheses by studying the mutual relationships of CRP, D-dimer, incident IHD, and risk factors (especially smoking habit) in the middle-aged men of the Speedwell population cohort.
The Speedwell Study began in 1979. It was based on all men who were registered with any of 16 general practitioners from 2 group practices in East Bristol and who were between 45 and 59 years old on September 1, 1978. A total of 2550 men were eligible for inclusion, and 2348 (92%) attended the first examination between 1979 and 1982. At that time their ages ranged from 45 to 63 years. The cohort was reexamined on 3 further occasions at intervals of 3 to 4 years. The current study uses as its baseline the second examination conducted between 1982 and 1985, at which time men were aged 49 to 67 years. Of the 2348 men seen at the first examination, 106 had died, 28 had moved out of the area or emigrated, and 9 were too ill to be seen. Of the 2205 eligible for the second examination, 2055 (93%) attended.
The general design and procedures have been described in detail elsewhere.23 24 In brief, at each examination the men were invited to attend a clinic where a detailed medical and lifestyle history was obtained; the London School of Hygiene and Tropical Medicine chest pain questionnaire25 was administered; a full 12-lead ECG was recorded; and height, weight, and blood pressure were measured. Current or last occupation was recorded, and from this information, social class was coded according to the Registrar General’s Classification.26 A personal history of diabetes and a family history of myocardial infarction among first-degree relatives were included in the medical history.
Blood Collection, Storage, and Analysis
At the first examination, the men returned after an overnight fast to an early morning clinic to give a blood sample. Standard methods were used for the estimation of lipids.27 At the second examination, a nonfasting blood sample was taken from 2016 (98%) of the men seen. Fibrinogen was measured by heat precipitation nephelometry.24 Lipids were not remeasured on this nonfasting sample, and hence, the fasting levels of total cholesterol from the first examination are used in this report.
Plasma samples from dipotassium edetate–anticoagulated blood at the nonfasting second examination were stored at −20°C. CRP and D-dimer levels were measured on these samples in 1997, when the samples had been stored for between 12 and 15 years. CRP was measured by a sensitive nephelometric assay (Behring) and D-dimer, with an ELISA (Biopool). One batch of samples was unavailable for the current analysis, so CRP was measured for 1690 men, who were a subset of 1719 men for whom D-dimer was measured first. Another batch containing 17% of the samples was thawed on 1 occasion due to freezer failure. Levels of CRP and D-dimer were both significantly higher (P<0.001) among these thawed samples. However, it is not certain that the measurements were affected by thawing of the samples. Fibrinogen, measured on the fresh, prestorage samples, and even fibrinogen and total cholesterol, as measured at the first examination, were all higher among the subjects whose stored, second-examination sample thawed. This suggests that the subjects whose samples had been thawed were, by chance, not a representative sample of the whole group. Further support for this explanation comes from the Caerphilly study.12 When D-dimer was measured there, 1 batch of samples had also previously been thawed, and D-dimer levels were slightly but not significantly lower among the thawed samples. It was thus decided not to exclude the results from the thawed samples but to include them in all analyses with an adjustment for the effect of thawing.
Incidence of IHD was measured between the second examination (the baseline for this report) and the fourth examination, which took place between 1988 and 1991. At that fourth examination, the men were seen in the same order as far as possible, and the average follow-up period was 75 months (mean±SD, 75±4). All men were entered on the National Health Service Central Registry, and all of the death certificates were coded by 1 of us (J.W.G.Y.). Death coded as 410 to 414 according to the 9th revision of the International Classification of Diseases (ICD) was used as the definition of fatal IHD. Questions about admission to hospital with severe chest pain and lists from hospital activity analysis of all men admitted with a diagnosis of ICD 410 to 414 were used as the basis for a search of hospital notes for events meeting standard World Health Organization criteria28 for acute myocardial infarction (MI). Finally, the appearance on any follow-up ECG of major or selected moderate Q waves (Minnesota codes 1.1 [any], 1.2.1 to 1.2.5 or 1.2.7) when there were no Q waves (1.1 [any], 1.2 [any]. or 1.3 [any]) on either the first or the baseline second-examination ECG was also taken as evidence that an MI had occurred during the follow-up period.
Under these definitions, there were 191 major IHD events of which 72 were fatal. The average annual incidence rate was 1.5%. Among the 1690 men with a measurement for CRP, there were 162 major IHD events, whereas among the 1719 for whom a D-dimer measurement was available, there were 165 such events.
The distributions of CRP and D-dimer both had a marked, positive skewness. In all analyses where they were used as continuous variables, they were transformed to (natural) logarithms. The transformations produced distributions that were close to gaussian, with back-transformed geometric means of 1.57 mg/L for CRP and 42.0 ng/mL for D-dimer.
Adjusted mean differences in CRP and D-dimer between men who developed IHD and those who did not were obtained by ANCOVA by using standard multiple regression techniques. The remainder of the analysis was performed by using multiple logistic regression with the occurrence or not of any of the 3 types of incident IHD as the dependent variable. Logistic regression takes no account of the duration of follow-up, but this factor is likely to be immaterial because follow-up was at a nearly constant interval of 75 months, with an SD of only 4 months. Furthermore, any model such as the Cox proportional-hazards model that involves the time to the event would face the problem that no time to event is available for the ECG-defined MIs. These would either have to be excluded or allocated an arbitrary time to event.
In the logistic regression analyses, CRP and D-dimer were first divided into 5 equally sized groups by using 4 cutpoints: 0.6, 1.1, 2.0, and 4.2 mg/L for CRP and 26, 37, 47, and 63 ng/mL for D-dimer. Results were then presented as the odds of IHD in each group relative to the odds in the 20% of men with the lowest levels. Tests for trend were obtained by entering CRP or D-dimer as logarithmically transformed continuous variables, and the trends were summarized by standardized relative odds (SROs): the odds associated with a 1-SD increase in the logarithm of CRP or D-dimer.
Evidence of ischemia at baseline was assessed from the chest pain questionnaire and the ECG. Three categories, namely angina; history of at least 1 episode of prolonged, severe chest pain; and ECG ischemia were defined in a standard manner.29 Among the 2055 men, 486 (24%) had some evidence of ischemia at baseline. This prevalence is slightly lower than that found by the British Regional Heart Study30 for men of similar age. These men were not excluded from the analysis. Exclusion of such a large group, among whom 42% of the incident events occurred, does not seem satisfactory. Neither does the usual practice of excluding just a very small percentage (<5%) of men for whom there is good evidence of a previous MI. Instead, we have chosen to include all men and to adjust for the presence of ischemia at baseline by including the 3 standard measures of confounders in the logistic regression analyses.12
The baseline characteristics of the 191 men who developed major IHD are compared in Table 1⇓ with those of the 1864 men who did not. Those who developed IHD were slightly older (P<0.001) and had higher total cholesterol (P=0.001), body mass index (P=0.029), and systolic (P<0.001) and diastolic (P=0.002) blood pressures. They were more likely to be smokers (P=0.001), to be diabetic (P=0.003), and to have a family history of MI (P=0.024). The proportion from the manual social classes was similar in the 2 groups.
Univariate Analyses of CRP, D-Dimer, and Incident IHD
In Table 2⇓, the data show that mean CRP and D-dimer levels were each higher among the men who developed IHD. The age- and thawing-adjusted mean difference between the 2 groups was highly significant for CRP (P=0.00008) and significant for D-dimer (P=0.017).
In Table 3⇓, the data show that the incidence of IHD increased steadily from 5.7% in the 20% of men with the lowest levels of CRP to 15.5% among the 20% with the highest levels. Unadjusted, the corresponding relative odds increased steadily to 3.07 (95% CI, 1.80 to 5.22) in the top 20% of the distribution. This trend was highly statistically significant (P<0.00001), and the SROs, the relative odds associated with a 1-SD increase in CRP, were 1.48. Adjusting for age slightly reduced the relative odds in the top 20% to 2.73 (95% CI, 1.60 to 4.67) and the SRO to 1.41, but the trend was still highly significant (P=0.00005). Further adjustment for whether or not the sample had been thawed during the freezer failure had no material effect on these relative odds (Table 3⇓).
The incidence of IHD increased from 7.0% in the 20% of men with the lowest levels of D-dimer to 12.0% and 11.2% in the fourth and fifth highest quintile groups, respectively (Table 3⇑). Unadjusted, the corresponding relative odds increased steadily to 1.68 (95% CI, 1.00 to 2.82) in the top 20% of the distribution. This trend was highly statistically significant (P=0.0018), and the SROs were 1.27. Adjusting for age slightly reduced the relative odds in the top 20% to 1.44 (95% CI, 0.85 to 2.45) and the SRO to 1.22, but the trend was still significant (P=0.012). Adjustment for thawing had only a minor effect (Table 3⇑).
Associations With Cardiovascular Risk Factors
The data in Table 4⇓ show how CRP and D-dimer varied with cigarette smoking habit and with evidence of IHD at baseline. Current smokers had geometric mean levels of CRP that were nearly double those among men who had never smoked. Among those current smokers, the lowest geometric mean level (1.87 mg/L) was found among the lightest smokers, and even this lowest level was significantly (P<0.001) higher than that found among the men who had never smoked. There was no clear dose response among current smokers, but among ex-smokers there was a clear trend with the length of time since quitting. Even those who gave up more than 10 years ago had a geometric mean CRP of 1.36 mg/L, which was still higher (P=0.037) than the geometric mean of 1.13 mg/L among the men who had never smoked. The average length of time since quitting among these men was nearly 23 years, with a range from 10 to 48 years. Current smokers also had higher D-dimer levels than men who had never smoked (P=0.046), with no dose response. Again, there was a clear trend with the length of time since quitting; 5 to 9 years after quitting, their D-dimer levels returned to the levels observed in nonsmokers.
All manifestations of IHD at baseline were associated with higher geometric mean levels of CRP. Men with angina from the Rose chest pain questionnaire had levels of CRP nearly double those of men without angina. CRP was raised by ≈50% among men with a history of prolonged, severe chest pain or with evidence of ischemia on ECG. All of these differences were highly statistically significant (P<0.001). D-dimer levels compared with CRP showed less elevation in men with evidence of ischemia, but these elevations were still statistically significant for men with a history of severe chest pain (P=0.014) or with evidence of ischemia on ECG (P<0.001). CRP was also raised by 50% among the small number (2.8%) of diabetics. Among the much larger proportion (26%) of men with a first-degree relative with a history of MI, CRP was raised by 17% (P=0.011). There was no association between CRP and social class. D-dimer was not significantly associated with diabetes, family history of MI, or social class.
CRP increased with age (r=0.15) and showed positive associations with other conventional cardiovascular risk factors such as total cholesterol (r=0.08), diastolic blood pressure (r=0.07), and body mass index (r=0.14). All of these were statistically significant (P<0.01) but modest in size. There was a much stronger association with fibrinogen (r=0.42). D-dimer also increased with age (r=0.15, P<0.01) and fibrinogen (r=0.16, P<0.01) but not with total cholesterol (r=0.01), diastolic blood pressure (r=0.03), or body mass index (r=0.00).
There was a positive correlation between CRP and D-dimer (r=0.21, P<0.00001). This association did not arise simply because both were positively associated with age, smoking habit, baseline evidence of IHD, and fibrinogen. On adjusting for all of these factors, the partial correlation declined to only 0.17 and remained statistically significant (t=5.29, P<0.00001).
Incidence of IHD With Increasing CRP and D-Dimer
The data in Table 5⇓ show that the incidence of IHD increased with CRP at each level of D-dimer and that it also increased with D-dimer at every level of CRP. The trend for incidence of IHD to increase with CRP independently of the level of D-dimer was statistically significant (P=0.00015), as also was the trend for incidence to rise with D-dimer independently of the level of CRP (P=0.048). There was no evidence that the association between IHD and CRP was different at different levels or D-dimer or that the association between IHD and D-dimer differed with level of CRP (test for interaction, χ2 (2 df)=3.54, P=0.17).
Multivariate Analysis for CRP
In Table 6⇓ are shown the results of a series of multiple logistic regression analyses that examined the change in association between CRP and incident IHD as groups of cardiovascular risk factors were successively added to the regression model. In the first model, which adjusted only for age and the thawing (or not) of the sample, the relative odds of IHD rose to 2.49 (95% CI, 1.44 to 4.30) among men in the top 20% of the distribution of CRP, and the SROs were 1.37 (P=0.0002). These figures differ from those in Table 3⇑ only because those in Table 3⇑ were based on all 1690 men with a measurement of CRP, whereas the figures in Table 6⇓ were based on the 1595 men who had a complete set of data for all of the variables included in the models of Table 6⇓.
Addition of a set of conventional cardiovascular risk factors (smoking habit, body mass index, diastolic blood pressure, total cholesterol, and evidence of ischemia at baseline) reduced the relative odds in the top 20% of the distribution of CRP to 1.60 (95% CI, 0.90 to 2.83) and the SROs to 1.20. The test for trend was still just significant at the conventional 5% level. The further addition of fibrinogen and then D-dimer to the model reduced the relative odds in the top 20% of the CRP distribution to 1.45 (95% CI, 0.79 to 2.66) and the SROs to 1.15. The test for trend (P=0.16) was no longer statistically significant.
A stepwise multiple logistic regression analysis was then performed for CRP (Table 7⇓). The variables considered for inclusion in the stepwise regression were all those cardiovascular risk factors appearing in the models of Table 6⇑. The base model again consisted of age and whether or not the sample had been thawed, as well as CRP. At each stage of the stepwise procedure, the cardiovascular risk factor that produced the largest reduction in the SROs of IHD was added to the model. First to be so added was evidence of ischemia at baseline, which caused the SROs to decline from 1.37 to 1.27, but the test for trend remained significant (P=0.008). At the next stage the addition of smoking habit caused the largest further reduction to 1.21 (P=0.033). Thereafter the addition of fibrinogen and then D-dimer reduced the SROs first to 1.19 (P=0.084) and then to 1.16 (P=0.13). No individual risk factor then caused any further substantial reduction in SROs. The joint addition of body mass index, diastolic blood pressure and total cholesterol only reduced the SROs to 1.15 (P=0.16) as shown in Table 6⇑. When family history of MI instead of social class and whether or not the subject was a diabetic were added jointly to the model, the SROs declined only from 1.16 (P=0.13) to 1.14 (P=0.17).
Multivariate Analysis for D-Dimer
The data in Table 6⇑ show that addition of the set of cardiovascular risk factors reduced the relative odds in the top 20% of the distribution of D-dimer to 1.15 (95% CI, 0.65 to 2.05) and the SROs to 1.17. The test for trend after this adjustment was nonsignificant and was reduced further by adjustment for fibrinogen and CRP. Stepwise multiple logistic regression showed that the confounding factor that most reduced the association of D-dimer with IHD was CRP, followed by evidence of ischemia at baseline and fibrinogen (Table 7⇑).
To our knowledge, this is the first report to compare CRP (a marker of the reactant plasma protein component of the inflammatory response) with D-dimer (a marker of fibrin turnover) in the prediction of incident IHD in a population cohort. The inflammatory and the thrombotic components of coronary atherosclerosis and IHD are of current interest in pathophysiology,21 22 31 and there is some experimental evidence that they may be linked.17 18 19 20 Because both CRP and D-dimer are easily measured in stored plasma (or serum) samples, their potential use in risk stratification for IHD merits a comparison.
We observed that in this population cohort of men aged 49 to 67 years, plasma CRP and D-dimer levels showed a moderate correlation (r=0.21, P<0.00001). On adjusting for potential confounders (age, smoking, baseline evidence of IHD, and fibrinogen), the association was reduced but remained highly statistically significant (r=0.17, P<0.00001). This correlation supports our hypothesis that there may be a link between these respective markers of inflammation and fibrin turnover in middle-aged men, which may be due in part to their associations with asymptomatic and symptomatic arterial lesions. However, the correlation is not strong, which suggests that other factors have different effects on plasma levels of these 2 variables.
As expected, we found that cigarette smoking habit had important, reversible effects on both CRP7 32 33 and D-dimer.34 35 The elevations of CRP and D-dimer in current cigarette smokers were not dose dependent but were reversible after quitting (Table 4⇑). Mean plasma CRP was approximately doubled in current smokers compared with never-smokers and may partly reflect elevations in smokers of interleukin-6, which is a major regulator of the reactant plasma protein component of the inflammatory response.36 Although plasma CRP levels fell progressively with time since quitting smoking, they remained significantly elevated 10 years after quitting compared with those in never-smokers (Table 4⇑). In contrast, the elevation in mean plasma D-dimer in current smokers was only ≈10% and fell to levels seen in nonsmokers, 5 to 9 years after quitting (Table 4⇑). These data suggest that the “inflammatory” effect of cigarette smoking is both larger in magnitude (10-fold) and longer-lasting than its effect on cross-linked fibrin turnover. The relationships of these observations to underlying smoking-related pathology in the arteries, respiratory tract, and other organs and to the time course of reduction of IHD risk in smokers who quit merit future study.
As with smoking, the relationships of CRP to both baseline IHD (Table 4⇑) and incident IHD (Tables 2and 3) were stronger than those of D-dimer to baseline and incident IHD. These relationships to IHD are consistent with the literature for CRP7 and D-dimer.8 9 10 11 12 13 14 15 16 We are not aware of previous studies directly comparing the predictive value of CRP and D-dimer for IHD. The present study suggests that measurement of both variables may be useful in risk stratification (Table 5⇑). The incidence of IHD increased with CRP independently of the level of D-dimer and vice versa. The risk of IHD over ≈6 years follow-up was ≈6% (ie, 1% per year) in middle-aged men with levels in the lower third of both CRP and D-dimer compared with almost 18% (ie, almost 3% per year) in those in the upper third of both CRP and D-dimer. These data, combined with the practical issue that both CRP and D-dimer are easily measured in stored plasma (or serum) samples, suggest the need for further evaluation of both variables in risk stratification for IHD. This suggestion has a plausible pathophysiological basis: both inflammation and thrombosis are important in the pathogenesis of IHD.21 22 31
In the present study, we observed that inclusion of conventional risk factors as well as CRP or D-dimer in multiple logistic regression analyses of the relationships of the other variable to incident IHD reduced the strength of the association (Tables 5⇑, 6⇑, and 7⇑). With regard to the relationship between CRP and incident IHD, the inclusion of fibrinogen in the model reduced the relationship to below the conventional level of statistical significance (reduction of SROs from 1.21, P=0.033 to 1.19, P=0.084); however, because CRP and fibrinogen are both measures of the reactant plasma protein component of inflammation, the validity of this adjustment is debatable. The addition of D-dimer to the model further reduced the relationship (SROs of 1.16, P=0.13), which suggests that the relationship between CRP and incident IHD is partly confounded by their mutual relationships to D-dimer. Conversely, with regard to the relationship between D-dimer and incident IHD, the inclusion of CRP in the model reduced the relationship from an SRO of 1.21 (P=0.023) to 1.15 (P=0.11). However, the limited number of major IHD events in this study (191) results in wide confidence intervals for estimates of the mutual relationships between CRP, D-dimer, conventional risk factors, and incident IHD. Hence, further prospective cohort studies and collaborative meta-analyses are required to define these with greater precision.
Assays of CRP and D-dimer were performed after storage at −20°C for 12 to 15 years. Although both of these proteins appear resistant to proteolysis during storage and the assays used in the current study appear sufficiently robust to withstand the effects of long-term storage (and thawing), it is also important to perform further prospective cohort studies in which these assays are performed together in samples stored at lower temperatures (−70°C or below). The effect of storage at different temperatures on CRP and D-dimer levels should also be studied prospectively by serial analyses over time. However, comparison of the distributions of CRP and D-dimer values in the present study to those obtained in previous studies in which samples were stored at −70°C or below1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 (including the Caerphilly Study, whose subjects and methods were otherwise similar to those of the Speedwell Study12 ) shows close agreement, which supports the validity of our conclusions. Furthermore, the results of the present study are very similar to the overall results in meta-analyses of CRP and D-dimer (J. Danesh, personal communication, 2000).
In conclusion, our data suggest that in a population cohort of middle-aged men, markers of inflammation (CRP) and of fibrin turnover (D-dimer) are related to each other, smoking, age, plasma fibrinogen, and baseline (as well as incident) IHD. These findings may be related to the association of inflammation and fibrin turnover in arterial lesions and at other body sites. However, measurement of both CRP and D-dimer may be a logical and practical enhancement of current risk stratification for IHD. Further studies are required to test these hypotheses.
We thank the Medical Research Council (UK) for financial support. P.M.S. is currently supported by the British Heart Foundation.
Toss H, Lindahl B, Siegbahn A, Wallentin L. Prognostic influence of increased fibrinogen and C-reactive protein levels in unstable coronary artery disease. Circulation. 1997;96:4204–4210.
Kuller LH, Tracy RP, Shaten J, Meilahn EN. Relation of C-reactive protein and coronary heart disease in the MRFIT nested case-control study. Am J Epidemiol. 1996;144:537–547.
Koenig W, Frohlich F, Sund M, Doering A, Fischer HG, Loewel H, Hutchinson WL, Pepys MB. C-reactive protein (CRP) predicts risks of coronary heart disease (CHD) in healthy middle-aged men: results from the MONICA-Augsburg cohort study, 1984/85–1992. Circulation. 1997;96(suppl I):I-1–I-99.
Tracy RP, Lemaitre RN, Psaty BM, Ives DG, Evans RW, Cushman M, Meilahn EN, Kuller LH. Relation of C-reactive protein to risk of cardiovascular disease in the elderly: results from the Cardiovascular Health Study and the Rural Health Promotion Project. Arterioscler Thromb Vasc Biol. 1997;17:1121–1127.
Cortellaro M, Confrancesco E, Boschetti C, Mussoni L, Donati MB, Cardillo M, Catalano M, Gabrielli L, Lombardi B, Specchia G. Increased fibrin turnover and high PAI-1 activity as predictors of ischemic events in atherosclerotic patients: a case-control study. Arterioscler Thromb Vasc Biol. 1993;13:1412–1417.
Ridker PM, Hennekens CH, Cerskus A, Stampfer MH. Plasma concentration of cross-linked fibrin degradation products (D-dimer) and the risk of future myocardial infarction among apparently healthy men. Circulation. 1994;90:2236–2240.
Smith FB, Lee AJ, Fowkes FGR, Rumley A, Lowe GDO. Hemostatic factors as predictors of ischemic heart disease and stroke in the Edinburgh Artery Study. Arterioscler Thromb Vasc Biol. 1997;17:3321–3325.
Moss AJ, Goldstein RE, Marder VJ, Sparks CE, Oakes D, Greenberg H, Weiss HJ, Zareba W, Brown MW, Liang CS, Lichstein E, Little WC, Gillespie JA, Van Voorhees L, Krone RJ, Bodenheimer MM, Hochman J, Dwyer EM Jr, Arora R, Marcus FI, Watelet LF, Case RB. Thrombogenic factors and recurrent coronary events. Circulation. 1999;99:2517–2522.
Cushman M, Lemaitre RN, Kuller LH, Psaty BM, Macy EM, Sharrett AR, Tracy RP. Fibrinolytic activation markers predict myocardial infarction in the elderly: the Cardiovascular Health Study. Arterioscler Thromb Vasc Biol. 1999;19:493–498.
Ritchie DG, Levy BA, Adams MA, Fuller GM. Regulation of fibrinogen and fibrin: an indirect feedback pathway. Proc Natl Acad Sci U S A. 1982;79:1530–1534.
Edgington TS, Curtiss LK, Plow EG. A linkage between the haemostatic and immune systems embodied in the fibrinolytic release of lymphocyte suppressive peptides. J Immunol. 1985;134:471–477.
Gauldie J, Northemann W, Fey GHO. IL-6 functions as an exocrine hormone in inflammation: hepatocytes undergoing acute phase responses require exogenous IL-6. J Immunol. 1990;144:3804–3808.
Davies MJ. A macro and micro view of coronary vascular insult in ischemic heart disease. Circulation. 1990;82(suppl II):II-38–II-46.
Bainton D, Sweetnam P, Baker I, Elwood P. Peripheral vascular disease: consequence for survival and association with risk factors in the Speedwell prospective heart disease study. Br Heart J. 1994;72:128–132.
Yarnell JWG, Baker IA, Sweetnam PM, Bainton D, O’Brien JR, Whitehead PJ, Elwood PC. Fibrinogen, viscosity and white blood cell count are major risk factors for ischemic heart disease: the Caerphilly and Speedwell Collaborative Heart Disease Studies. Circulation. 1991;83:836–844.
Office of Population Censuses and Surveys. Classification of Occupations. London, UK: HMSO; 1980.
Bainton D, Miller NE, Bolton CH, Yarnell JWG, Sweetnam PM, Baker IA, Lewis B, Elwood PC. Plasma triglyceride and high density lipoprotein cholesterol as predictors of ischaemic heart disease in British men: the Caerphilly and Speedwell Collaborative Heart Disease Studies. Br Heart J. 1992;68:60–66.
World Health Organisation Regional Office for Europe. Myocardial Infarction Community Registers. Copenhagen, Denmark: WHO; Public Health in Europe No. 5, 1976.
Bainton D, Baker IA, Sweetnam PM, Yarnell JWG, Elwood PC. Prevalence of ischaemic heart disease: the Caerphilly and Speedwell surveys. Br Heart J. 1988;59:201–206.
Shaper AG, Cook DG, Walker M, MacFarlane PW. Prevalence of ischaemic heart disease in British men. Br Heart J. 1984;51:595–605.
Tracy RP, Psaty BM, Macy E, Bovill EG, Cushman M, Cornell ES, Kuller LH. Lifetime smoking exposure affects the association of C-reactive protein with cardiovascular disease risk factors and subclinical disease in healthy elderly subjects. Arterioscler Thromb Vasc Biol. 1997;17:2167–2176.
Danesh J, Muir J, Wong Y-K, Ward M, Gallimore JR, Pepys MB. Risk factors for coronary heart disease and acute-phase proteins: a population-based study. Eur Heart J. 1999;20:954–959.
Lee AJ, Fowkes FGR, Lowe GDO, Rumley A. Determinants of fibrin D-dimer in the Edinburgh Artery Study. Arterioscler Thromb Vasc Biol. 1995;15:1094–1097.
Yarnell JWG, Sweetnam PM, Rumley A, Lowe GDO. Lifestyle and hemostatic risk factors for ischaemic heart disease: the Caerphilly Study Arterioscler Thromb Vasc Biol. 2000;20:271–279.