Inflammatory markers and long-term risk of ischemic heart disease in men A 13-year follow-up of the Quebec Cardiovascular Study

Transcription

1 Atherosclerosis 182 (2005) Inflammatory markers and long-term risk of ischemic heart disease in men A 13-year follow-up of the Quebec Cardiovascular Study Annie C. St-Pierre a,b, Bernard Cantin c, Jean Bergeron b, Matteo Pirro b, Gilles R. Dagenais c, Jean-Pierre Després c, Benoît Lamarche a,b, a Department of Nutrition, Institute on Nutraceuticals and Functional Foods, Laval University, 2440 Hochelaga Blvd., Ste-Foy, Qué., Canada G1K 7P4 b Lipid Research Center, CHUL Research Center, Laval University, Qué., Canada c Québec Heart Institute, Laval Hospital, Laval University, Qué., Canada Received 19 November 2004; received in revised form 28 January 2005; accepted 9 February 2005 Available online 13 March 2005 Abstract We tested the hypothesis that elevated plasma interleukin-6 (IL-6), C-reactive protein (CRP) and fibrinogen concentrations are independent risk factors and interact in increasing the long-term risk of ischemic heart disease (IHD) in men. A total of 1982 IHD-free men from the Quebec Cardiovascular Study were followed over a period of 13 years during which 210 first fatal IHD events and non-fatal myocardial infarctions were recorded. Increased CRP levels (4th versus 1st quartile) were not associated with an increased risk of IHD after adjustment for non-lipid risk factors (age, body mass index, systolic blood pressure, diabetes, smoking and medication use at baseline), lipid risk factors (LDL and HDL cholesterol and triglyceride levels) and for IL-6 and fibrinogen (RR = 0.70, 95% CI = ). High plasma IL-6 levels (4th versus 1st quartile) were associated with a 70% greater risk of IHD independent of confounding risk factors and of the other 2 inflammatory markers (RR = 1.71, 95% CI = ). The relationship between high fibrinogen levels (4th versus 1st quartile) and IHD risk was borderline significant in multivariate analyses (RR = 1.53, 95% CI = ). An inflammation score based on plasma IL-6 and fibrinogen levels improved the IHD risk predictive value of a multivariate model of traditional risk factors (p = 0.03). Including plasma CRP levels into the inflammatory score provided no additional predictive value. In conclusion, elevated plasma IL-6 concentrations are more strongly related to IHD risk than CRP and fibrinogen. An inflammation score based on high plasma IL-6 and fibrinogen levels used in combination with traditional risk factors may improve our ability to adequately identify high risk individuals Elsevier Ireland Ltd. All rights reserved. Keywords: C-reactive protein; Interleukin-6; Fibrinogen; Inflammation; Ischemic heart disease 1. Introduction A large body of evidence now supports the role of inflammation in the development and progression of atherosclerosis [1]. Inflammatory processes may contribute to plaque instability and thrombosis, thus enhancing the risk of acute ischemic heart disease (IHD) events [2]. Several studies have shown that elevated plasma interleukin-6 (IL-6), C-reactive protein (CRP) and fibrinogen levels are associated with an increased risk of IHD and severity of atherosclerosis [3 5]. Corresponding author. Tel.: ; fax: address: benoit.lamarche@inaf.ulaval.ca (B. Lamarche). IL-6 is a cytokine with a broad range of humoral and cellular response related to infection, inflammation, host defense and tissue injury [6]. IL-6 is also a major initiator of the acute phase response and a primary determinant for the hepatic production of CRP, a non-specific acute phase reactant, and fibrinogen, a coagulation factor [7]. While plasma IL-6, CRP and fibrinogen levels are physiologically linked, it remains unclear whether or not there is inter-dependence between these pro-inflammatory cytokines in modulating the risk of IHD and whether they have synergistic impact on IHD risk. The aim of the present study was therefore to examine whether the IHD risk predictive value of plasma CRP, fibrinogen and IL-6 levels are independent of each other. We also /$ see front matter 2005 Elsevier Ireland Ltd. All rights reserved. doi: /j.atherosclerosis

2 316 A.C.St-Pierre et al./ Atherosclerosis 182 (2005) investigated the synergistic impact of these pro-inflammatory cytokines on IHD risk among men of the Québec Cardiovascular Study followed over a 13-year period. 2. Methods 2.1. Study population and follow-up The Québec Cardiovascular Study population has been previously reported [8,9]. Briefly, a random sample of 4635 Canadian men (aged years), among whom 99% were from French decent and representing 65.5% of the population screened from seven towns in the Québec City metropolitan area, were randomly selected from provincial electoral lists in 1973 for the study of cardiovascular disease risk factors, namely plasma cholesterol levels and hypertension. This cohort was re-evaluated in 1975, 1980 and 1985, and data collected in 1985 on 2552 (55.1%) of the 4635 participants evaluated in 1974 were used as the baseline characteristics for the present prospective analyses. Electrocardiograms (ECG) and plasma lipid and lipoprotein levels were obtained in the 1985 participants. Among the 2552 participants, 102 volunteers were not part of the original random sample and 265 patients had a previous history of IHD prior to These men were excluded from the present prospective analyses. Diagnosis of diabetes was considered in men who self-reported the disease. Use of hypolipidemic drugs, mainly clofibrate and cholestyramine in 1985, was limited to 1% of men both with and without IHD. In [8,9] and in 1998, participants were contacted by mail and invited to complete a short questionnaire, which provided history of cardiovascular diseases and type 2 diabetes. For those who reported such diseases and those who died, hospital charts were reviewed by cardiologists of the study who were blinded to the participants risk profile. Over the 13-year follow-up period and among the 2185 subjects eligible for follow-up, only 8 subjects (0.4%) were lost to follow-up. Thirty-five subjects with available 5-year follow-up data could not be retraced in These participants were not excluded since we were able to use 5-year or time-to-event censoring in the survival analysis. Among the 2177 eligible men, plasma was no longer available in 195 subjects and measures of the various pro-inflammatory markers could not be performed. Thus, analyses were conducted in a sample of 1982 men with a full metabolic profile including inflammatory markers Definitions of IHD events The occurrence of a first IHD event, which included coronary death and non-fatal myocardial infarction, was diagnosed as reported previously [9,10]. Among the cohort of 1982 middle-aged men initially free of clinical manifestations of IHD in 1985, a total of 210 first cases of IHD were recorded over the 13-year follow-up: 155 men had a first nonfatal myocardial infarction, and 55 a fatal coronary events Laboratory analyses Twelve-hour fasting blood samples were obtained at the 1985 baseline evaluation and immediately used for all lipid and apolipoprotein measurements using methods that have been detailed in previous publications [9,11]. Cytokines, insulin and LDL size were measured on plasma obtained at the 1985 baseline evaluation and stored at 80 C. Plasma CRP levels were measured using the Behring Latex-Enhanced highly sensitive CRP assay on a Behring Nephelometer BN- 100 (Behring Diagnostic, Westwood, MA) and the calibrators (N Rheumatology Standards SL) provided by the manufacturer [12]. Plasma IL-6 levels were measured using the commercially available Quantikine HS Immunoassay ELISA kit (R&D Systems Inc., Minneapolis, MN) and calibrators (Diluent HD6F). Fibrinogen concentrations were determined by thrombin clotting time assay as described previously [13]. Fasting insulin concentrations were measured using a commercial kit as described previously [14] and LDL peak particle size and the proportion of small LDL (LDL% <255 Å ) were measured by polyacrylamide gradient gel electrophoresis [8] Statistical analyses Mean baseline characteristics of incident IHD cases and of IHD-free men during follow-up were compared by Student s t-test for parametric variables and by the Wilcoxon test for non-parametric variables. Differences in frequency data were tested by Chi-square analysis. The non-lipid and lipid risk variables in men classified according to number of inflammatory markers were compared using a general linear model (GLM) with the Tukey post hoc test to locate subgroup differences. An inflammatory score ranging from 0 to 2 was computed by attributing one point for a value greater than the median of the cohort samples for IL-6 and fibrinogen levels. Duration of follow-up was calculated in personyears by using the follow-up of each participant from the 1985 baseline evaluation until death, onset of IHD or the last contact. Cox proportional-hazards models were used to estimate risks of IHD events. For all Cox models, the proportional hazards assumptions were formally tested. Age, body mass index, systolic blood pressure, type 2 diabetes (presence versus absence), smoking habits (smokers of more than 20 cigarettes per day versus others), medication use (presence versus absence) at baseline, plasma triglyceride (natural logtransformed), LDL and HDL cholesterol levels were included as potential confounders where indicated. Receiver operating characteristic (ROC) curves were used to examine the additional predictive value of including individual inflammatory cytokines or an inflammatory score to a traditional risk factor model in discriminating subjects who did or did not suffer a first IHD event during follow-up. The areas under the ROC curve were the primary end point for these analyses and were compared using the likelihood ratio test. Statistical analyses were performed on SAS (SAS Institute, Cary, NC).

3 A.C.St-Pierre et al./ Atherosclerosis 182 (2005) Table 1 Baseline characteristics of the 1772 men without IHD and the 210 men who had a first IHD event during the 13-year follow-up Variables Ischemic heart disease % Difference p Without (N = 1772) With (N = 210) Age (years) 56.2 ± ± <0.001 Body mass index (kg/m 2 ) 26.1 ± ± Systolic blood pressure (mmhg) 129 ± ± <0.001 Type 2 diabetes mellitus (% (N)) 4.4% (78) 9.1% (19) Smokers (% (N)) 22.4% (397) 29.1% (61) Total cholesterol (mmol/l) 5.7 ± ± <0.001 LDL cholesterol (mmol/l) 3.9 ± ± <0.001 HDL cholesterol (mmol/l) 1.03 ± ± Cholesterol/HDL cholesterol 5.8 ± ± <0.001 Triglycerides (mmol/l) 1.6 ± ± LDL % <255 Å (%) 39.5 ± ± C-reactive protein (mg/l) 1.8 ± ± Interleukin-6 (mg/l) 1.0 ± ± <0.001 Fibrinogen (g/l) 4.0 ± ± <0.001 LDL, low density lipoprotein; HDL, high density lipoprotein; LDL % <255 Å, proportion of total LDL with a diameter <255 Å. Values are mean ± S.D. Plasma triglyceride, C-reactive protein, interleukin-6 and fibrinogen levels are presented as geometric means. 3. Results Table 1 shows the risk profile of incident IHD cases and IHD-free men. Men with IHD had a higher plasma CRP, IL-6 and fibrinogen levels compared to men without IHD. Plasma IL-6 and fibrinogen levels correlated with plasma CRP concentrations (r = 0.54 and r = 0.51, respectively, p < 0.001) while the correlation between IL-6 and fibrinogen levels was of a lower magnitude (r = 0.37, p < 0.001). These strong but relatively low correlation coefficients indicated that there was no collinearity between each of these three markers of inflammation. Fig. 1 shows the risk ratio (RR) of IHD with their 95% confidence intervals (CI) computed for each quartile of CRP, IL-6 and fibrinogen taken individually, using the first quartile as reference (RR = 1.0). Increased plasma CRP levels (4th versus 1st quartile) were associated with a 1.8-fold increase in the long-term risk of IHD (RR = 1.78, 95% CI = , p = 0.003). However, elevated CRP levels did not predict an increased risk of IHD after adjustment for non-lipid and lipid risk factors (RR = 0.98, 95% CI = , p = 0.93, not shown in Fig. 1). The association between CRP levels and the risk of IHD was also attenuated to a point of non-significance after adjustment for plasma IL-6 and fibrinogen concentrations (RR = 1.13, 95% CI = , p = 0.61) or after concomitant adjustment for non-lipid and lipid risk factors and IL-6 and fibrinogen levels (RR = 0.70, 95% CI = , p = 0.14). Virtually identical results were obtained when the 1.0 and 3.0 mg/l CRP cutpoints rather than quartiles were used in these analyses (not shown). Elevated plasma IL-6 levels (4th versus 1st quartile) were associated with a 2.3- fold greater risk of IHD (RR = 2.28, 95% CI = , p < 0.001) in univariate analysis and this association remained significant after adjustment for non-lipid and lipid risk factors (p = 0.02, not shown), for plasma CRP and fibrinogen concentrations (RR = 1.83, 95% CI = , p = 0.01) and for all risk factors simultaneously (RR = 1.71, 95% CI = , p = 0.03). Finally, increased plasma fib- Fig. 1. Comparison of the multivariate risk ratio (RR) of ischemic heart disease (IHD) associated with quartiles of plasma CRP, IL-6 and fibrinogen levels in men. The RR of IHD were computed using men in the first quartile as reference group (RR = 1.0). Black circles represent the unadjusted RR. White circles represent RR after adjustment for the two other inflammatory markers. Gray squares represent RR after simultaneous adjustment for non-lipid risk factors (age, body mass index, systolic blood pressure, diabetes, smoking and medication use at baseline), lipid risk factors (LDL and HDL cholesterol and log-transformed triglyceride levels) and the two other inflammatory markers. Number of cases from quartiles 1 to 4 was 43, 37, 60 and 70 for CRP, 34, 45, 58 and 73 for IL-6 and 33, 47, 53 and 77 for fibrinogen.

4 318 A.C.St-Pierre et al./ Atherosclerosis 182 (2005) Fig. 2. Risk ratio (RR) of ischemic heart disease IHD and P levels according to IL-6 (above or below median of 0.94 mg/l) and fibrinogen (median 4.05 g/l) stratified for the median CRP levels (1.76 mg/l). The group with low CRP, IL-6 and fibrinogen concentrations was used as reference (RR = 1.0). RR were adjusted for age, body mass index, systolic blood pressure, type 2 diabetes, medication use at baseline and smoking habits, LDL and HDL cholesterol and log-transformed triglyceride levels. rinogen levels (4th versus 1st quartile) were associated with an univariate 2.3-fold increased risk of IHD (RR = 2.29, 95% CI = , p < 0.001). This relationship remained significant after adjustment for the other 2 pro-inflammatory markers (RR = 1.89, 95% CI = , p = 0.006) but became only borderline significant after multivariate adjustment for non-lipid and lipid risk factors (p = 0.06, not shown) or for all risk factors simultaneously (RR = 1.53, 95% CI = , p = 0.07). Similar results were obtained while using CRP, IL- 6 or fibrinogen as continuous variables (data not shown). Fig. 2 illustrates the combined impact of concomitant variations in CRP, IL-6 and fibrinogen levels on the long-term risk of IHD. Neither increased CRP levels nor elevated fibrinogen concentrations were associated with an increased RR of IHD if they were not combined with elevated plasma IL-6 levels. On the other hand, increased levels of IL-6 (above ver- Fig. 3. Comparison of the multivariate risk ratio (RR) of ischemic heart disease (IHD) associated with increasing number of arbitrarily elevated inflammatory markers in men. The RR of IHD were computed using men with levels below the median for each marker of inflammation as reference group (RR = 1.0). Black circles represent the unadjusted RR. White circles represent RR after adjustment for non-lipid risk factors (age, body mass index, systolic blood pressure, diabetes, smoking, medication use at baseline). Grey squares represent RR after adjustment for non-lipid and lipid risk factors (LDL and HDL cholesterol and log-transformed triglyceride levels). sus below the median) were associated with a significantly elevated IHD risk even when plasma levels of CRP and fibrinogen were arbitrarily low (below their respective median) (RR = 2.38, 95% CI = , p = 0.003). As shown in Table 2, individuals having increased levels of any 2 or 3 inflammatory markers, i.e. above the median of the cohort distribution for IL-6 ( 0.94 mg/l), CRP ( 1.76 mg/l) and/or fibrinogen ( 4.05 g/l), had a globally more deteriorated risk profile compared to individuals with low levels of all three pro-inflammatory markers. The risk of IHD associated with the number of inflammatory markers was dependant upon cumulative number of inflammatory markers (Fig. 3) and was independent of concomitant adjustment for non-lipid and lipid risk factors. However, the risk of IHD in men with relative elevations in 2 or in all 3 of the inflammatory markers appeared as being virtually identical. Table 2 Risk profile of the 1982 men according to the number of arbitrarily elevated inflammatory markers Variables Cumulative number of inflammatory markers * 0(N = 488) 1 (N = 506) 2 (N = 488) 3 (N = 500) Age (years) 53.9 ± ± 6.5 * 57.0 ± 7.0 *, 59.1 ± 7.4 *,, Body mass index (kg/m 2 ) 25.2 ± ± 3.5 * 26.6 ± 3.8 *, 26.7 ± 4.2 *, Systolic blood pressure (mmhg) 126 ± ± 15 * 132 ± 17 *, 135 ± 19 *,, Total cholesterol (mmol/l) 5.7 ± ± ± ± 0.9 LDL cholesterol (mmol/l) 3.8 ± ± ± ± 0.9 HDL cholesterol (mmol/l) 1.09 ± ± 0.26 * 1.01 ± 0.24 *, 0.97 ± 0.25 *,, Cholesterol/HDL cholesterol 5.5 ± ± 1.7 * 6.0 ± 1.8 *, 6.2 ± 1.8 *, Triglycerides (mmol/l) 1.5 ± ± 1.5 * 1.7 ± 1.5 * 1.7 ± 1.5 *, C-reactive protein (mg/l) 0.64 ± ± 2.2 * 2.6 ± 2.1 *, 5.9 ± 2.2 *,, Interleukin-6 (mg/l) 0.57 ± ± 1.77 * 1.1 ± 1.8 *, 1.9 ± 1.7 *,, Fibrinogen (g/l) 29.6 ± ± 1.9 * 65.3 ± 2.0 *, ± 2.1 *,, Insulin (pmol/l) 61.8 ± ± 1.5 * 69.2 ± 1.5 * 74.6 ± 1.6 *,, LDL, low density lipoprotein; HDL, high density lipoprotein; LDL % <255 Å, proportion of total LDL with a diameter <255 Å. Values are mean ± S.D. Inflammatory marker levels were dichotomized (low or high) using the median of the distribution for CRP (1.76 mg/l), IL-6 (0.94 mg/l) and fibrinogen (4.05 g/l). Plasma triglyceride, C-reactive protein, interleukin-6, fibrinogen and insulin levels are presented as geometric mean. * Significantly different from 0. Significantly different from 1. Significantly different from 2.

5 A.C.St-Pierre et al./ Atherosclerosis 182 (2005) Table 3 Area under the receiver operating characteristics (AUROC) curve Base a Model 2 Model 3 Model 4 Model 5 Base + C-reactive protein b Base + interleukin-6 Base + fibrinogen Base + score AUROC c 70.5% 70.6% (p = 0.5) d 71.1% (p = 0.07) 71.0% (p = 0.1) 71.3% (p = 0.03) a Model 1 includes age, BMI, systolic blood pressure, smoking, type 2 diabetes, medication use at baseline, LDL-C, HDL-C, log-transformed triglycerides. b The inflammatory markers are dichotomized (low or high) using the median of the distribution for each parameter: C-reactive protein (1.76 mg/l), interleukin- 6 (0.94 mg/l), and fibrinogen (4.05 g/l). The score ranging from 0 to 2 indicates low-grade systemic inflammation. It was computed by attributing one point for a value greater than the median of the cohort samples for IL-6 and fibrinogen levels. c AUROC corresponds to the area under the sensitivity to 1-specificity curve, in %. d This p-value reflects the incremental predictive value of adding one inflammatory marker or the score to the traditional model of risk factors (model 1) in discriminating IHD cases from non-cases. It is obtained using the likelihood ratio test in the logistic regression model. The incremental benefit of considering inflammatory markers in addition to the series of traditional non-lipid and lipid risk factors to discriminate incident IHD cases from non-cases was investigated using ROC curves obtained using logistic regression analysis (Table 3). The area under the ROC curve (AUROC) based on the combination of traditional risk factors (age, body mass index, systolic blood pressure, type 2 diabetes, smoking habits, medication use at baseline, LDL and HDL cholesterol and log-transformed triglyceride levels) was 70.5% (model 1). CRP (AUROC = 70.6%), IL- 6 (AUROC = 71.1%) and fibrinogen (AUROC = 71.0%), defined as dichotomic variables using the median of the distribution, added virtually no discriminating power to the model of traditional risk factors (p > 0.05). When an inflammatory score based on plasma IL-6 and fibrinogen levels was added to a multivariate model of traditional risk factors, the ability to discriminate incident IHD cases from non-cases was significantly increased (AUROC = 71.3, p = 0.03). Including plasma CRP levels in that inflammatory score provided no added value in predicting IHD risk (not shown). An inflammatory score based on CRP and fibrinogen levels also did not add to the predictive value of a model based on traditional risk factors only. Similar results were obtained when the inflammatory score was defined based on the 3rd quartile of the cohort rather than the median, or when we used the 1.0 and 3 mg/l cutpoints for CRP (not shown). 4. Discussion Several previous studies have evaluated how proinflammatory markers such as plasma CRP, IL-6 and fibrinogen levels modulate the risk of IHD [3 5,15]. However, very few studies have examined the inter-dependence of several pro-inflammatory cytokines in modulating the risk of IHD as well as the extent to which they may have a synergistic impact on IHD risk. In that context, the Quebec Cardiovascular Study is one of only a few long-term prospective populationbased studies to compare the ability of a series of inflammatory cytokines, examined individually or combined as a score, to discriminate the population at high versus low risk for IHD. The present 13-year prospective, population-based study of men without IHD at baseline provides new evidence indicating that elevated plasma IL-6 levels, not CRP, may be independently associated with long-term IHD risk after simultaneous adjustment for non-lipid and lipid risk factors, while the association between elevated fibrinogen levels and long-term IHD risk failed to reach statistical significance. Our data also indicated that considering an inflammatory score defined as a combination of high plasma IL-6 and fibrinogen levels, in combination with a series of traditional risk factor currently used in clinical practice, may better discriminate incident IHD cases from IHD-free individuals, thus potentially contributing to a more adequate prediction of long-term IHD risk. In the present study, plasma CRP levels were not an independent predictor of long-term IHD risk. This is in disagreement with a series of previous reports showing that increased plasma CRP levels were associated with an increased risk of developing IHD independent of other non-lipid and lipid risk factors [16]. It must be stressed that the magnitude of the increase in the risk ratio obtained in the present study when using tertiles of plasma CRP levels was very similar to that reported by Danesh et al. in a recent meta-analysis (not shown) [17]. The majority of the previous studies on CRP also did not adjust their multivariate risk model for concomitant variations in other pro-inflammatory cytokines [18]. Such adjustment is of relevance since the association between CRP and IHD risk may partly reflect mutual association with other inflammatory markers such as IL-6 and fibrinogen. In the Caerphilly Prospective Heart Disease Study in men, Mendall et al. observed that the impact of increased plasma CRP levels on the 5-year risk of IHD was abolished after adjustment for established inflammatory factors such as fibrinogen, which is concordant with our observations [19]. Prospective data on IL-6 and IHD risk are more limited. Ridker et al. have demonstrated that men in the highest quartile of IL-6 had a risk ratio 2.3 times higher than those in the lowest quartile of IL-6 over a 6-year follow-up period [20]. They also reported that IL-6 remained significantly associated with IHD risk after adjustment for CRP. Unfortunately, the inverse analysis, i.e. the extent to which plasma CRP levels remained significantly associated with IHD risk after adjustment for IL-6 concentrations, was not presented in that report. Nevertheless, these data are consistent with our own long-term follow-up data, which suggested that the re-

6 320 A.C.St-Pierre et al./ Atherosclerosis 182 (2005) lationship between IL-6 levels and the long-term risk of IHD persisted after adjustment for non-lipid and lipid risk factors as well as for plasma CRP and fibrinogen levels. It is possible that lesser biological variability may contribute to this better performance of IL-6 compare to CRP. Indeed, Browning et al. have suggested out that IL-6 levels have a better discrimination ratio (ratio of between-subject to within-subject) relative to CRP levels [21]. Plasma fibrinogen levels, which have been recognized as an independent IHD risk factor in previous studies [22], are now considered not only as a coagulation factor but also as a marker of systemic inflammation [23]. Data in a large sample of middle-aged men and women with or without baseline IHD from the Scottish Heart Health Study indicated that elevated plasma fibrinogen levels predicted an increased risk of fatal and non-fatal IHD over an 8-year follow-up period [24]. Concordant with our own short-term data [25], the association between plasma fibrinogen levels and incident IHD in the Scottish Heart Health Study remained strong even after multivariate adjustment for non-lipid and lipid risk factors [26]. However, data presented in the present report suggest that the impact of elevated fibrinogen levels on the long-term risk of IHD may be attenuated by simultaneous adjustment for other risk factors. In these previous analyses, risk ratios obtained through Cox s proportional hazard regression were used to estimate relationships between pro-inflammatory risk variables and incident IHD. In a further analysis, receiver operating characteristics analysis and the corresponding c statistic were employed to compare the ability of various models incorporating the different pro-inflammatory markers to accurately and adequately discriminate IHD-free individuals from incident IHD cases. Since improvement in predictive value, not increase in risk ratio, should be considered the most important endpoint when evaluating various models for risk prediction, our results suggest that using individual measures of proinflammatory markers such as CRP, IL-6 and fibrinogen may not be justified to predict long-term IHD risk, since each of these inflammatory markers do not globally better discriminate IHD cases from non-cases after consideration for other more traditional risk factors. However, we hypothesized that an integrated inflammatory score may improve the predictive value of a traditional model of IHD risk. Our data indicated that when an inflammatory score based on concomitant elevations in plasma IL-6 and fibrinogen levels (using the median or quartiles as arbitrary cut-off points) was added to a model of traditional risk factors, the ability to discriminate incident IHD cases from non-cases was significantly increased. We hypothesize that IL-6 and fibrinogen may trigger the atherosclerotic process through distinct pathways and therefore could better discriminate incident IHD cases from IHD-free individuals when combined together. It must be stressed that inclusion of plasma CRP into the inflammatory score already combining information from plasma IL-6 and fibrinogen levels did not improve its risk predictive value. In conclusion, our data suggested that the individual use of pro-inflammatory markers such as CRP, IL-6 and fibrinogen levels do not improve our ability to predict long-term IHD risk in men. However, the use of an integrated inflammatory score based on a combination of IL-6 and fibrinogen levels did provide incremental predictive value in discriminating incident IHD cases from healthy men after consideration for traditional risk factors. While the use of an inflammatory score based on plasma IL-6 and fibrinogen levels appears to provide additional information on risk from a statistical standpoint, future cost-effectiveness analyses will have to be undertaken to demonstrate if this small improvement in predictive value justifies the increased laboratory work and ensuing costs. The extent to which our findings can be extended to other populations such as women or other ethnic groups will also have to be determined in future studies. Unfortunately, the extent to which the evolution of statin use over the course of the follow-up period may have confounded the assessment of risk in men with elevated inflammatory marker levels remains unknown. Recent data suggested that other inflammatory markers not related to the IL-6 metabolic cascade may also be regarded as potentially independent risk factors for IHD [27]. Data from Packard et al. have suggested that elevated lipoprotein-associated phospholipase A 2 (Lp-PLA 2 ) levels were associated with an increased risk of IHD independent of traditional risk factors, CRP and fibrinogen levels while Ballantyne et al. have suggested that increased levels of Lp-PLA 2 and CRP may be complementary in identifying individuals at high risk for IHD [28]. Acknowledgments This study was supported in part by an operating grant from the Canadian Institute for Health Research (MOP 14475) and by an unrestricted Medical School Grant from Merck-Frosst. Benoît Lamarche is the recipient of a Canada Research Chair in Nutrition, Functional Foods and Cardiovascular Health. Annie St-Pierre is the recipient of a training fellowship from the Heart and Stroke Foundation of Canada. References [1] Ross R. Atherosclerosis-an inflammatory disease. N Engl J Med 1999;340: [2] Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med 2000;342: [3] de Maat MP, Pietersma A, Kofflard M, Sluiter W, Kluft C. Association of plasma fibrinogen levels with coronary artery disease, smoking and inflammatory markers. Atherosclerosis 1996;121: [4] Thompson SG, Kienast J, Pyke SD, Haverkate F, van de Loo JC. Hemostatic factors and the risk of myocardial infarction or sudden death in patients with angina pectoris. European Concerted Action on Thrombosis and Disabilities Angina Pectoris Study Group. N Engl J Med 1995;332: [5] Biasucci LM, Vitelli A, Liuzzo G, et al. Elevated levels of interleukin-6 in unstable angina. Circulation 1996;94:874 7.

7 A.C.St-Pierre et al./ Atherosclerosis 182 (2005) [6] Papanicolaou DA, Wilder RL, Manolagas SC, Chrousos GP. The pathophysiologic roles of interleukin-6 in human disease. Ann Intern Med 1998;128: [7] Heinrich PC, Castell JV, Andus T. Interleukin-6 and the acute phase response. Biochem J 1990;265: [8] St-Pierre AC, Ruel IL, Cantin B, Dagenais GR, Després JP, Lamarche B. Comparison of various electrophoretic characteristics of LDL particles and their relationship to the risk of ischemic heart disease. Circulation 2001;104: [9] Lamarche B, Despres JP, Moorjani S, et al. Prevalence of dyslipidemic phenotypes in ischemic heart disease (prospective results from the Quebec Cardiovascular Study). Am J Cardiol 1995;75: [10] Dagenais GR, Robitaille NM, Lupien PJ, et al. First coronary heart disease event rates in relation to major risk factors: Québec Cardiovascular Study. Can J Cardiol 1990;6: [11] Moorjani S, Dupont A, Labrie F, et al. Increase in plasma high density lipoprotein concentration following complete androgen blockage in men with prostatic carcinoma. Metabolism 1987;36: [12] Pirro M, Bergeron J, Dagenais GR, et al. Age and duration of followup as modulators of the risk of ischemic heart disease associated with high plasma C-reactive protein levels in men. Arch Intern Med 2001;161: [13] Clauss A. Rapid physiological coagulation method in determination of fibrinogen. Acta Haematol 1957;17: [14] Despres JP, Lamarche B, Mauriege P, et al. Hyperinsulinemia as an independent risk factor for ischemic heart disease. N Engl J Med 1996;334: [15] Berk BC, Weintraub WS, Alexander RW. Elevation of C-reactive protein in active coronary artery disease. Am J Cardiol 1990; 65: [16] Ridker PM, Rifai N, Pfeffer MA, et al. Inflammation, pravastatin, and the risk of coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events (CARE) Investigators. Circulation 1998;98: [17] Danesh J, Wheeler JG, Hirschfield GM, et al. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med 2004;350: [18] Cesari M, Penninx BW, Newman AB, et al. Inflammatory markers and cardiovascular disease (The Health Aging and Body Composition [Health ABC] Study). Am J Cardiol 2003;92: [19] Mendall MA, Strachan DP, Butland BK, et al. C-reactive protein: relation to total mortality, cardiovascular mortality and cardiovascular risk factors in men. Eur Heart J 2000;21: [20] Ridker PM, Rifai N, Stampfer MJ, Hennekens CH. Plasma concentration of interleukin-6 and the risk of future myocardial infarction among apparently healthy men. Circulation 2000;101: [21] Browning LM, Krebs JD, Jebb SA. Discrimination ratio analysis of inflammatory markers: implications for the study of inflammation in chronic disease. Metabolism 2004;53: [22] Kannel WB, Wolf PA, Castelli WP, D Agostino RB. Fibrinogen and risk of cardiovascular disease. The Framingham Study. JAMA 1987;258: [23] Tracy RP. Inflammation markers and coronary heart disease. Curr Opin Lipidol 1999;10: [24] Woodward M, Lowe GD, Rumley A, Tunstall-Pedoe H. Fibrinogen as a risk factor for coronary heart disease and mortality in middleaged men and women The Scottish Heart Health Study. Eur Heart J 1998;19: [25] Cantin B, Despres JP, Lamarche B, et al. Association of fibrinogen and lipoprotein(a) as a coronary heart disease risk factor in men (The Quebec Cardiovascular Study). Am J Cardiol 2002;89: [26] Woodward M, Rumley A, Lowe GD, Tunstall-Pedoe H. C-reactive protein: associations with haematological variables, cardiovascular risk factors and prevalent cardiovascular disease. Br J Haematol 2003;122: [27] Packard CJ, O Reilly DS, Caslake MJ, et al. Lipoprotein-associated phospholipase A2 as an independent predictor of coronary heart disease. West of Scotland Coronary Prevention Study Group. N Engl J Med 2000;343: [28] Ballantyne CM, Hoogeveen RC, Bang H, et al. Lipoproteinassociated phospholipase A2, high-sensitivity C-reactive protein, and risk for incident coronary heart disease in middle-aged men and women in the Atherosclerosis Risk in Communities (ARIC) Study. Circulation 2004;109: