Chronic Inflammation and Impaired Coronary Vasoreactivity in Patients With Coronary Risk Factors
Background— The goal of this study was to examine a possible association between systemic microinflammation, as reflected by C-reactive protein (CRP) serum levels, and coronary vasomotion in patients with coronary risk factors but with angiographically normal coronary arteries.
Methods and Results— Coronary vasomotor function was studied in response to cold pressor testing (CPT) in 71 patients with normal angiograms. In all patients, CPT-induced changes in epicardial luminal area (LA; mm2) were assessed with quantitative angiography. Within 20 days, myocardial blood flow (MBF) responses to CPT were measured (mL · g−1 · min−1) noninvasively with 13N-ammonia and PET imaging. The CPT-induced mean changes in LA and in MBF in patients with elevated CRP (≥0.5 mg/dL) were significantly impaired compared with patients presenting with CRP levels within normal range (<0.5 mg/dL) (ΔLA, −1.09±0.86 versus 0.45±0.63 mm2; ΔMBF, 0.06±0.18 versus 0.44±0.31 mL · g−1 · min−1; P<0.0001, respectively). Coronary LA changes and MBF responses to CPT were inversely correlated with CRP serum levels (r=−0.84 and r=−0.63; P<0.0001). Lastly, regression analysis revealed a significant correlation between the changes in LA and MBF during CPT for patients with elevated CRP levels and those for patients with normal CRP levels (r=0.56 and r=0.66; P<0.001).
Conclusions— These findings suggest a direct association between systemic microinflammation and altered coronary vasomotor function of both the epicardial conductance and the arteriolar resistance vessels.
Received February 11, 2004; de novo received April 20, 2004; accepted May 20, 2004.
The vascular endothelium plays an active and pivotal role in regulating certain aspects of the integrity and metabolism of the vascular wall such as vascular structure and permeability, vasomotor tone, and hemostasis.1 Risk factors for atherosclerosis, however, have been shown to impair the activity of endothelium-derived nitric oxide, thereby affecting the neurohumoral dynamic balance of the vascular wall in favor of a coronary vasoconstrictor response.1,2 Impaired endothelium-dependent vascular function has been shown to precede and accompany the development and progression of atherosclerosis, with substantial diagnostic and prognostic implications.3–5 Recent observations suggest that the atherosclerotic process is characterized by a low-grade inflammation altering the endothelium of the coronary arteries and is associated with an increase in markers of inflammation such as C-reactive protein (CRP), levels of soluble endothelial cell adhesion molecules, and procoagulant activity.1,6 Indeed, recent studies have provided evidence that endothelial dysfunction of forearm vessels in patients with manifest coronary artery disease (CAD) or abnormal epicardial vasoreactivity in patients with various coronary risk factors correlated significantly with elevated serum levels of CRP.3,7 Because elevated serum levels of CRP are associated with a higher rate of cardiovascular events,3,8 it is intriguing to speculate that alterations in flow-mediated coronary vasomotion as a result of inflammatory responses may provide a link between systemic inflammation and progression of cardiovascular disease.8
In the present study, we examined a possible association between CRP serum levels and coronary vasomotor response to cold pressor testing (CPT) at the level of the epicardial conductance and the arteriolar resistance vessels in patients who had coronary risk factors but normal coronary angiograms.
Patient Population and Study Protocol
Seventy-one patients (44 men, 27 women; mean age, 58±8 years) consecutively referred for diagnostic catheterization for evaluation of chest pain and with normal coronary angiograms were studied. Each patient underwent quantitative coronary angiography (QCA) at baseline and during CPT for measurement of the epicardial luminal area (LA) and, within 20 days of that, PET measurements for myocardial blood flow (MBF) at baseline and during CPT.3,9 All patients revealed normal wall motion and ejection fraction on left ventriculography (n=49) or echocardiography (n=22). Each patient was screened by complete history, physical examination, and laboratory analyses. Excluded were patients with a history of acute coronary syndrome or myocardial infarction, hypertrophic obstructive and nonobstructive cardiomyopathy, congestive heart failure, malignant hypertension (defined as diastolic blood pressure >120 mm Hg with the presence of severe hypertensive retinopathy and organ damage), valvular heart disease, or endocrine, hepatic, renal, or inflammatory disease. In addition, patients with elevated white blood cell counts and/or erythrocyte sedimentation rate were not included in this study, thereby avoiding possibly confounding effects of occult infection or other systemic inflammatory disease on CRP plasma levels and coronary vasomotion. Patients did not qualify for the study if they were on cholesterol-lowering statin medication. Vasoactive medications, including calcium channel blockers, ACE inhibitors, nitrates, and β-blockers, were discontinued at least 24 hours before QCA and PET evaluation.
Risk factors, assessed at the time of coronary angiography, included hypertension, hypercholesterolemia, smoking, diabetes mellitus, and family history of premature coronary artery disease. Twenty-five patients had 1 and 29 patients had ≥2 coronary risk factors. Hypertension was defined as a well-established history of chronically elevated blood pressure (≥140/90 mm Hg) without any apparent underlying cause for ≥2 years that was treated with antihypertensive drugs by a primary physician. Hypercholesterolemia was defined as fasting total serum cholesterol values ≥240 mg/dL. Moreover, elevated LDL cholesterol and triglycerides were defined as ≥155 and ≥200 mg/dL, respectively; normal HDL cholesterol was defined as ≥35 mg/dL. Smokers were included if they presented with a history of smoking for ≥2 pack-years.3,10 The group of chronic smokers consisted of 12 men and 5 women (age, 47±12 years) with a history of 30±14 pack-years. All smokers refrained from smoking for ≥12 hours before angiography and the PET study of MBF. A positive family history for CAD was defined as evidence of CAD in a parent or a sibling who was <60 years of age at the time of diagnosis.4 Body mass index (BMI) was calculated as weight (kg) divided by height (m2). Premenopausal women who had regular menses and were not on birth control were studied at midcycle (day 11 to 14) of the menstrual period. Postmenopausal women (cessation of menses ≥1 year) were considered for study purposes if they were not on hormone replacement therapy. In each patient, CRP serum levels (turbidimetric test, Boehringer Mannheim) were determined from peripheral arm vein blood samples. On the basis of a predefined cutoff level for traditional CRP of 0.5 mg/dL,7,11 study patients were assigned to 2 groups: those with CRP levels within normal range (<0.5 mg/dL) and those with elevated CRP levels (≥0.5 mg/dL). In 22 of the 71 patients, CRP levels were also measured with an ultrasensitive CRP test (N Latex CRP Mono, Behring). The local ethics committee of the University of Freiburg approved the study, and written informed consent was obtained from all patients. The epicardial vasomotor responses of 25 of these patients were reported previously.3
Quantitative Coronary Angiography
All patients underwent routine diagnostic catheterization with a biplane, isocentric multidirectional digital angiographic system (BICOR-HICOR, Siemens).3 A prerequisite for inclusion in the study was the absence of angiographic evidence of CAD, as indicated by smooth coronary vessels without evidence of luminal wall irregularities or diffuse caliber reduction and stenosis. A relatively straight (4 to 8 mm long) nonbranching segment of the proximal LAD corresponding to vessel segment 12, in accordance with the American College of Cardiology/American Heart Association guidelines,12 was preselected, and a coronary angiogram was obtained at baseline. The patient then immersed the right hand in ice water for 90 seconds with immediate acquisition of a second coronary angiogram. End-diastolic images of coronary arteries on biplanar views at baseline and during CPT were analyzed quantitatively.2,3
PET Study Protocol
MBF was measured with intravenous 13N-ammonia and serial image acquisition by PET (951 ECAT- HR, Siemens).9 Measurements were performed first at baseline and then during CPT. For the CPT, study participants immersed the left hand in ice water for 60 seconds, and 13N-ammonia was injected again while CPT was continued for another 60 seconds. Beginning with each 13N-ammonia administration (15 to 20 mCi), serial transaxial emission images were acquired (12 frames of 10 seconds each, 2 frames of 30 seconds each, 1 frame of 60 seconds, and 1 frame of 900 seconds).9 Time-activity curves were used to calculate mean MBF (mL · g−1 · min−1) from the first 12 serially acquired, short-axis slices after assigning on 3 midventricular slices myocardial regions of interest to the myocardial territories of the left anterior descending artery (LAD), left circumflex coronary artery (LCx), and right coronary artery (RCA).9
Heart rate, blood pressure, and a 12-lead ECG were recorded continuously. From the average of heart rate and blood pressure during the first 2 minutes of each image acquisition, the rate-pressure product (RPP), the product of systolic blood pressure and heart rate, was determined as an index of cardiac work. The mean MBF, averaged over the 3 midventricular planes, was derived specifically for the coronary territory of the LAD for which measurements of the LA and its changes during CPT were obtained. MBFs were calculated for the 2 coronary territories subtended by the LCx and RCA not evaluated on quantitative angiography.
Data are presented as mean±SD for quantitative and absolute frequencies for qualitative variables. For comparison of differences, appropriate t tests for independent or paired samples were used (Statistical Analysis Software Institute). A comparison of CPT-induced changes in epicardial LA or MBF responses between the different groups was done by 2-way ANOVA, followed by Scheffé’s F test. Pearson’s correlation coefficient (r), assuming a linear regression, was calculated to investigate the associations between CPT-induced changes in epicardial LA or MBF responses and CRP levels. Multivariate analysis was performed with the logistic regression model. Statistical significance was assumed if the null hypothesis could be rejected at P=0.05.
Table 1 summarizes the characteristics of the 2 study groups and lists findings on quantitative angiography and on MBF at baseline and during CPT. Traditional CRP levels ranged from 0 to 4.5 mg/dL (mean, 1.6±1.3 mg/dL). Patients with elevated CRP levels had more coronary risk factors compared with patients with normal CRP levels. Total cholesterol, LDL cholesterol, and triglycerides were significantly higher in patients with elevated CRP levels than in those with normal CRP levels, whereas HDL cholesterol was significantly lower. In both groups, glucose levels were within normal range but were higher in patients with elevated CRP. In 22 patients, additionally measured high-sensitivity (hs) CRP levels ranged from 0.02 to 5.2 mg/dL (mean, 0.5±1.1 mg/dL). There was a significant correlation between traditional and hsCRP levels (r=0.80, P<0.0001; n=22). This correlation remained significant when these patients were grouped according to traditional CRP levels <0.5 mg/dL (n=12) and ≥0.5 mg/dL (n=10) (r=0.71 and r=0.81; P<0.01 and P<0.004, respectively). Finally, in 46 of the 71 patients (64%), significant ST-segment depression was noticed during treadmill exercise testing.
Hemodynamic parameters during angiographic assessment of epicardial coronary vasomotion and PET measurements of MBF at baseline and during CPT are listed in Table 2. Heart rate and blood pressure were similar at baseline and during CPT on the 2 study days (P=NS). During CPT, there was a significant increase in heart rate and systolic blood pressure (P<0.05). It is important to note that the RPPs at the time of both the angiographic and PET studies were similar at baseline and during CPT (P=NS). In addition, there were no significant differences in the percent increase in the RPP (%ΔRPP) as a result of CPT during coronary angiography and MBF evaluation with PET (P=NS).
Findings on Quantitative Angiography
Table 1 and Figure 1A summarize the findings on QCA. At baseline, the mean epicardial LA was similar for both study groups (P=NS). However, in patients with elevated CRP, the average LA of the LAD decreased by CPT (ΔLA, −1.09±0.86 mm2) compared with an increase in ΔLA in patients with normal CRP levels (0.45±0.63 mm2; P<0.0001). The group comparison of CPT-induced alterations in LA in patients with elevated CRP was significantly different compared with patients with normal CRP levels (P<0.001 by ANOVA). There was a inverse correlation between the combined analysis of traditional (n=49) and hsCRP (n=22) levels and LA change in response to sympathetic activation (r=0.78, P<0.0001; n=71). When those 22 patients for whom hsCRP levels were available were not taken into consideration, the correlation actually improved (r=0.84, P<0.0001; n=49). In the whole study population, the regression analysis of traditional CRP levels and the epicardial vasomotor response to CPT showed a significant correlation (n=71) (Figure 2A); higher levels of CRP were associated with a greater vasoconstrictor response to CPT. Moreover, this inverse correlation remained significant when the study population was grouped according to the mean CRP cutoff value of 1.6 mg/dL. Both groups with CRP levels <1.6 mg/dL (n=37) and ≥1.6 mg/dL (n=34) revealed a significant correlation with the CPT-induced alterations in the epicardial LA (r=0.61 and r=0.73; P<0.0001).
Determinants of Epicardial Vasoreactivity to CPT
On univariate analysis, CRP level (P≤0.0001), hypertension (P=0.0396), smoking (P≤0.0001), LDL cholesterol (P≤0.0001), HDL cholesterol (P=0.0037), and total cholesterol (P≤0.0001) were significantly associated with abnormal epicardial coronary vasoreactivity during CPT (Table 3), but age, sex, BMI, diabetes, and family history of CAD were not (P=NS).
To test whether CRP serum levels were independent predictors of CPT-induced responses of the epicardial LA, a multivariate analysis was performed. As shown in Table 3, in addition to smoking and LDL cholesterol, CRP serum levels remained a statistically significant independent predictor of CPT-induced changes in epicardial LA. In contrast, a positive family history of CAD, age, diabetes, BMI, hypertension, HDL cholesterol, and sex did not have a significant effect in the multivariate analysis. Thus, independently of classic coronary risk factors, elevated CRP levels are associated with an impaired epicardial coronary vasomotor response to CPT.
MBF Responses to CPT
PET-measured blood flows in the myocardial territory supplied by the epicardial vessel analyzed on angiography (LAD) were similar at baseline for both study groups (P=NS) (Table 1). In patients with elevated CRP levels, MBF failed to increase during CPT compared with a significant increase (ΔMBF) in patients with normal CRP levels (0.06±0.18 versus 0.44±0.31 mL · g−1 · min−1; P<0.0001; Figure 1B). The increase in MBF during CPT in patients with normal CRP levels differed significantly from that in patients with elevated CRP levels (P<0.001 by ANOVA). Furthermore, in patients with elevated and normal CRP levels, average MBFs in the 2 territories supplied by the LCx and RCA were similar to MBFs in the territory of the LAD submitted to QCA (0.75±0.20 to 0.80±0.24 mL · g−1 · min−1 versus 0.74±0.19 to 0.80±0.26 mL · g−1 · min−1 versus 0.74±0.20 to 1.22±0.35 mL · g−1 · min−1 versus 0.75±0.22 to 1.19±0.36 mL · g−1 · min−1; P=NS).
There was an inverse correlation between the combined analysis of traditional and hsCRP levels and MBF alterations in response to sympathetic activation (r=0.61, P<0.0001). After the exclusion of patients with hsCRP levels, the correlation increased (r=0.74, P<0.0001). In the whole study population, regression analysis revealed an inverse correlation between traditional CRP levels and the MBF response to CPT (Figure 2B); higher levels of CRP were associated with a significantly impaired MBF response to CPT. This inverse correlation between CRP levels and CPT-induced MBF changes was also observed for both groups with CRP <1.6 and ≥1.6 mg/dL (r=0.57 and r=0.34, P<0.0003 and P<0.04, respectively).
Determinants of MBF Responses to CPT
On univariate analysis, CRP level (P≤0.0001), hypertension (P=0.0007), smoking (P=0.0004), LDL cholesterol (P≤0.0001), HDL cholesterol (P=0.0067), and total cholesterol (P=0.0004) were significantly associated with abnormal MBF changes with CPT (Table 4). Age, sex, BMI, and family history of CAD were not associated with an altered MBF response to CPT. However, multivariate analysis to assess the predictive value of CRP serum levels of CPT-induced MBF responses, as shown in Table 4, did not show CRP serum levels as a statistically significant independent predictor of CPT-induced changes in MBF. In addition, age, sex, BMI, HDL cholesterol, diabetes, and family history of CAD did not have a significant effect in the multivariate analysis. Conversely, independent predictors of altered MBF responses to CPT were hypertension, smoking, and LDL cholesterol. Thus, elevated CRP serum levels are not independent from coronary risk factors associated with an impaired MBF response to CPT.
Correlation Between Responses to CPT of Epicardial Coronary Artery and MBF
The regression analysis between the percent changes of the LA of the epicardial vessel and MBF during CPT for patients with elevated CRP and with normal CRP levels revealed a statistically significant correlation (r=0.56 and r=0.66, P<0.001), suggesting that the CPT-induced change in the epicardial LA was associated with a corresponding change in MBF.
The present study is the first to demonstrate that in patients with normal coronary angiograms but with coronary risk factors, elevated CRP levels, which are indicative of systemic microinflammation, are associated with abnormal coronary vasomotion of the epicardial conduit and arteriolar resistance vessels to sympathetic stimulation. These observations are in keeping with recent results7 on endothelium-dependent forearm flow responses to acetylcholine in patients with manifest CAD, but they extend the findings to the coronary circulation.
Coronary Vasomotion and Serum CRP
In patients with elevated CRP levels, the flow-mediated responses of both epicardial coronary arteries and arteriolar resistance vessels to sympathetic stimulation were significantly impaired compared with patients with normal CRP levels. Changes in the LA of the epicardial artery were correlated with the MBF responses to CPT in patients with both elevated and normal CRP levels. The correlation argues in favor of a functional association between flow-dependent regulatory mechanisms at the level of the epicardial conduit as determined by angiography and of the coronary resistance vessels as measured with PET. Functional alterations associated with elevated CRP levels are therefore likely to affect both the conduit and resistance vessels. However, there is a well-described relationship between coronary risk factors and abnormal endothelium-dependent vasomotion of the coronary circulation.1–3,13 The exact mechanisms underlying abnormal coronary vasomotor function remain to be elucidated but have been attributed in part to an imbalance of the redox equilibrium between nitric oxide and reactive oxygen species toward oxidative stress, which, apart from reducing the bioavailability of endothelium-derived nitric oxide associated with impaired endothelium-dependent vasomotor function, may lead to the induction of a whole array of inflammatory genes involved in the pathogenesis of the atherosclerotic process.6 The inflammatory activation at the site of the vessels is associated with the production of interleukin-6, which increases systemic markers of inflammation such as CRP.6 Because, in keeping with the present study population, elevated plasma CRP levels are commonly found in patients with coronary risk factors,1,6,8 there is very little information to establish elevated CRP levels also as an alternate mediator to coronary vascular disease rather than a phenomenon of no pathological consequence. Recently, CRP was shown to potentially mediate proatherosclerotic effects, implying the expression of endothelial cell adhesion molecules, the production of chemoattractant chemokines, the macrophage LDL uptake, and the downregulation of endothelial nitric oxide synthase expression.6,14,15 Results of the present study provide evidence that the CPT-induced abnormal vasomotor response of the epicardial coronary arteries related to elevated CRP levels contributes to the effects of coronary risk factors in patients without apparent CAD. In addition, altered MBF responses associated with elevated CRP levels do appear to exert synergistically adverse effects with coronary risk factors on arteriolar vessels. It is intriguing to speculate that an absent or attenuated flow-mediated epicardial vasodilation most likely contributed to a marked impairment of MBF increases during CPT in patients with elevated CRP levels.2,16,17 This could also explain, at least in part, why CRP levels correlated independently with changes in epicardial luminal diameter but not with alterations in MBF in response to sympathetic stimulation. However, results of the present study suggest that the pathophysiological consequences of elevated CRP levels do extend into the human coronary microcirculation as measured by PET.
Previous observations on forearm vessels yielded contradictory results with regard to the association of elevated CRP levels and endothelium-dependent vasomotion.7,18 In patients with familial hypercholesterolemia and without evidence of clinically manifest coronary atherosclerosis, no relationship between flow-mediated endothelial dysfunction and low-grade inflammatory response, as reflected by levels of CRP, was observed.18 This contrasts with recent observations7,11 in patients with an acute coronary syndrome that demonstrate not only a significant relationship between levels of CRP and endothelial dysfunction of forearm vessels but also enhanced epicardial artery vasoreactivity in response to sympathetic activation. These contradictory findings7,11,18 are in keeping with findings in patients with normal coronary angiograms. Elevated CRP levels correlated significantly with severe paradoxical vasoconstrictor responses beyond 15% reduction in epicardial LA to sympathetic activation, but not for less severe decreases in epicardial LA between 0% and 15%.3 Although the reason for this discordant observation remains unclear,3,7,11,18 it may emphasize the complex nature of the underlying mechanisms of abnormal vasomotion with both genetic and environmental determinants,1 whereby impaired coronary vasomotion and inflammation may be related in more severe disease stages of the development of the atherosclerotic process, when markers of systemic microinflammation such as CRP may predominate.8,15
Interestingly, a large body of evidence suggests the value of elevated CRP levels as a mediator of future cardiovascular events not only in patients with stable or unstable CAD but also in apparently healthy men.6,8,19 Similarly, recent trials have demonstrated a significant association between abnormal endothelial function of the coronary circulation in response to various stimuli and the occurrence of cardiovascular events.3–5 Thus, the assessment of elevated CRP levels as an independent predictor of abnormal coronary vasomotion may provide an important mechanistic link between systemic markers of inflammation and the development and progression of atherosclerotic disease.
There are limitations worthy of consideration in any interpretation of our data. First, because we did not perform intravascular ultrasound to assess the vascular wall structure, the possible presence of early atherosclerosis in the study population may have remained unobserved.12 Second, it can be observed that the 2 groups present a rather asymmetrical distribution of coronary risk factors, which certainly explains the differences observed between the groups in CPT-induced alterations of coronary vasomotor function; yet we were able to show in a multivariate analysis adjusting for the effect of coronary risk factors that elevated CRP levels were independent of coronary risk factors associated with impaired epicardial vasoreactivity to sympathetic stimulation, but not at the site of the arteriolar vessels. Third, additionally measured hsCRP levels were available in only a small number of patients, and after those patients with hsCRP levels were eliminated, the correlation coefficient between CRP levels and coronary vasoreactivity increased to r=−0.84 and to r=−0.74 at the site of the epicardial and arteriolar vessels, respectively, indicating that the relation was not determined primarily by the hsCRP levels. The reason for the latter slight and nonsignificant increase in correlation coefficients is not clear and may be related to the smaller number of CRP values and coronary vasomotor responses evaluated. Fourth, because we did not assess MBF increases to pharmacologically induced vasodilation in these patients, further studies are warranted to explore a possible association between CRP levels and hyperemic flow increases.
These findings imply a direct association between elevated serum CRP levels and altered coronary vasomotor function on both the epicardial conductance and arteriolar vessels. Thus, systemic microinflammation might reflect, at least in part, the presence of impaired coronary vasomotion, providing an important link between inflammation and the development of CAD.
This work was supported by grants from the German Research Foundation (So 241/2-2), Baden-Württemberg (Projekt: Sch-A1/A2), and the National Heart, Lung and Blood Institute, Bethesda, Md (HL-331777).
Ganz P, Vita JA. Testing endothelial vasomotor function: nitric oxide, a multipotent molecule. Circulation. 2003; 108: 2049–2053.
Schindler TH, Hornig B, Buser PT, et al. Prognostic value of abnormal vasoreactivity of epicardial coronary arteries to sympathetic stimulation in patients with normal coronary angiograms. Arterioscler Thromb Vasc Biol. 2003; 23: 495–501.
Schachinger V, Britten MB, Zeiher AM. Prognostic impact of coronary vasodilator dysfunction on adverse long-term outcome of coronary heart disease. Circulation. 2000; 101: 1899–1906.
Halcox JP, Schenke WH, Zalos G, et al. Prognostic value of coronary vascular endothelial dysfunction. Circulation. 2002; 106: 653–658.
Szmitko PE, Wang CH, Weisel RD, et al. New markers of inflammation and endothelial cell activation, part I. Circulation. 2003; 108: 1917–1923.
Fichtlscherer S, Rosenberger G, Walter DH, et al. Elevated C-reactive protein levels and impaired endothelial vasoreactivity in patients with coronary artery disease. Circulation. 2000; 102: 1000–1006.
Blake GJ, Ridker PM. C-reactive protein, subclinical atherosclerosis, and risk of cardiovascular events. Arterioscler Thromb Vasc Biol. 2002; 22: 1512–1513.
Nitzsche EU, Choi Y, Czernin J, et al. Noninvasive quantification of myocardial blood flow in humans: a direct comparison of the [13N]ammonia and the [15O]water techniques. Circulation. 1996; 93: 2000–2006.
Campisi R, Czernin J, Schoder H, et al. l-Arginine normalizes coronary vasomotion in long-term smokers. Circulation. 1999; 99: 491–497.
Tomai F, Crea F, Gaspardone A, et al. Unstable angina and elevated C-reactive protein levels predict enhanced vasoreactivity of the culprit lesion. Circulation. 2001; 104: 1471–1476.
Calabro P, Willerson JT, Yeh ET. Inflammatory cytokines stimulated C-reactive protein production by human coronary artery smooth muscle cells. Circulation. 2003; 108: 1930–1932.
Yeh ET, Anderson HV, Pasceri V, et al. C-reactive protein: linking inflammation to cardiovascular complications. Circulation. 2001; 104: 974–975.
Zeiher AM, Drexler H, Wollschlager H, et al. Modulation of coronary vasomotor tone in humans: progressive endothelial dysfunction with different early stages of coronary atherosclerosis. Circulation. 1991; 83: 391–401.
Hernandez-Pampaloni M, Keng FY, Kudo T, et al. Abnormal longitudinal, base-to-apex myocardial perfusion gradient by quantitative blood flow measurements in patients with coronary risk factors. Circulation. 2001; 104: 527–532.