Journal of the American College of Cardiology Vol. 36, No. 1, 2000 2000 by the American College of Cardiology ISSN 0735-1097/00/$20.00 Published by Elsevier Science Inc. PII S0735-1097(00)00680-X Lack of Association of C-Reactive Protein and Coronary Calcium by Electron Beam Computed Tomography in Postmenopausal Women: Implications for Coronary Artery Disease Screening Rita F. Redberg, MD, MSc, FACC*, Nader Rifai, PhD, Lauren Gee, MPH, Paul M. Ridker, MD, MPH, FACC San Francisco, California, and Boston, Massachusetts OBJECTIVES BACKGROUND METHODS RESULTS CONCLUSIONS We sought to test the hypothesis that C-reactive protein, a marker of inflammation, would correlate positively with coronary calcium, a marker of atherosclerosis, in postmenopausal women. High sensitivity testing for C-reactive protein (hscrp) has recently been shown in large population studies to predict cardiac events in asymptomatic postmenopausal women. Coronary calcification determined by electron beam computerized tomography (EBCT) has also been suggested to be predictive of cardiac events in women. We performed hscrp testing and determined calcium scores by EBCT in 172 asymptomatic postmenopausal women (mean age: 64.5 7.9 years) at risk for cardiac disease. Risk factors were determined by history, physical, electrocardiogram, exercise testing, and lipoprotein profiles. Calcium scores ranged from 0 to 2618. For analysis, calcium scores were divided into three groups; none (0 to 10), minimal ( 10 to 50), and significant ( 50). Overall, there was no significant positive relationship between hscrp level and calcium score. Specifically, the hscrp levels (mg/dl) were 0.24 0.43, 0.33 0.47 and 0.17 0.32 (medians 0.11, 0.15, and 0.06) for women with none, minimal, and significant coronary calcification, respectively. In subgroup analysis, a similar lack of positive association was observed after stratification by smoking status and by hormone replacement therapy use, two factors known to increase hscrp. In contrast to our a priori hypothesis, we found no evidence of a positive association between hscrp and calcium score by EBCT. These data thus raise the possibility that hscrp and EBCT calcium score reflect different pathologic processes, an issue with implications for coronary artery disease screening. (J Am Coll Cardiol 2000;36:39 43) 2000 by the American College of Cardiology Because lipid screening fails to identify almost half of all patients at risk for future myocardial infarction (MI), there is increased interest in nontraditional risk factors and their role in improving coronary risk prediction (1). One emerging risk factor is C-reactive protein, a sensitive marker for underlying vascular inflammation, which can be measured clinically with high sensitivity testing for C-reactive protein (hscrp). Several large-scale prospective studies demonstrate that baseline levels of hscrp predict risk of myocardial infarction, stroke (2,3) and peripheral arterial disease (4) in apparently healthy men as well as in higher risk groups such as smokers (5), the elderly (6), those with angina (7,8) and those with prior MI (9). Similarly, in asymptomatic postmenopausal women, baseline levels of From the *Division of Cardiology, University of California, San Francisco, California; Department of Pathology, Children s Hospital Medical Center, Boston, Massachusetts; Institute for Health Policy Studies, University of California, San Francisco, California; and Divisions of Cardiology and Preventive Medicine, Harvard Medical School, Brigham & Women s Hospital, Boston, Massachusetts. These studies were supported by grants from the National Institute of Health: RO1 HL57702 and RO1 HL58755. Dr. Ridker is also supported by an Established Investigator Award from the American Heart Association, Dallas, Texas. Manuscript received October 22, 1999; revised manuscript received January 17, 2000, accepted March 2, 2000. hscrp are an independent risk factor for future MI, stroke, revascularization or cardiovascular death (10). Assessment of hscrp also appears to add to the predictive value of total and high density lipoprotein (HDL) cholesterol screening (11). Thus, it has been hypothesized that hscrp may represent either a molecular marker for the extent of underlying preclinical atherosclerosis, or a marker for rupture prone and unstable plaque. Calcification of the coronary arteries is also highly correlated with atherosclerosis (12). In this regard, electron beam computerized tomography (EBCT) is highly sensitive in detecting coronary calcification and therefore represents a potential noninvasive screening modality for detection of subclinical coronary atherosclerosis (13). In contrast to hscrp, prospective data evaluating the predictive value of EBCT are currently limited in asymptomatic populations. However, one study suggests that coronary calcium may be predictive of cardiovascular events (14), and it has been speculated that calcium score by EBCT may be particularly suitable as a marker of atherosclerosis in women (15), particularly as other noninvasive modalities, such as stress testing, are less accurate in women (16).
40 Redberg et al. JACC Vol. 36, No. 1, 2000 C-Reactive Protein and Calcium in Postmenopausal Women July 2000:39 43 Abbreviations and Acronyms BMI body mass index CAD coronary artery disease CI confidence interval EBCT electron beam computerized tomography FLASH Females, Lipids, Activity and Sex Hormone Study HDL high density lipoprotein HRT hormone replacement therapy hscrp high sensitivity testing for C-reactive protein MI myocardial infarction OR odds ratio Based on these observations, we hypothesized that, because hscrp levels are consistently associated with increased risk of future coronary events, hscrp levels would be significantly higher among women with coronary calcification than among women without calcification. We tested this hypothesis in a group of asymptomatic postmenopausal women. METHODS Subjects. The study population consisted of 172 participants in the Females, Lipids, Activity and Sex Hormone (FLASH) Study, a cross-sectional evaluation of cardiac risk factors in women conducted at the University of California Medical Center (San Francisco), an academic tertiary referral center. Women were recruited by posting flyers in the community and at local medical practices. Eligible participants were postmenopausal (one-year postmenses) with at least one additional risk factor for coronary artery disease (CAD) and without known CAD. All study participants gave written informed consent. The protocol was approved by the Committee on Human Research at the University of California, San Francisco, California. We assessed cardiac risk factors, hscrp levels, and calcium scores by EBCT per research protocol. Age, hormone replacement therapy (HRT), and smoking status were determined by interview, blood pressure, and body mass index (BMI) by examination, lipoprotein values by laboratory testing, and exercise time by treadmill testing. High sensitivity testing for C-reactive protein was performed according to the methods described by the manufacturer (Dade Behring, Newark, Delaware) (17). All EBCT scans were performed on an Imatron C-150 scanner (South San Francisco, California) and scored using a previously described protocol (18). We evaluated the association between levels of hscrp and calcium score using the Spearman rank correlation in the entire group and in subgroups stratified by HRT (19,20) or smoking status (21), as previous studies have shown CRP levels to be affected by these factors. Because neither parameter is normally distributed, we used a proportional odds model on ordered quintiles of hscrp and three ordered categories of calcium (0 to 10 none ; 11 to 50 minimal ; 50 significant ) to assess the strength of the association, adjusting for age, BMI, smoking status, amount of exercise, and HRT. We checked for violations of the proportional odds assumption using the Score test. Differences in the distribution of hscrp among the three calcium score groups were assessed with the Kruskal-Wallis test. Summary results are expressed as mean SD, unless otherwise stated. RESULTS Demographics and risk factor profiles for the study participants and subgroups of nonsmokers and non-hrt users were similar (Table 1). Average Framingham risk score was 14, which would predict a 4% event rate in five years (22). Overall, no positive relationship was seen between hscrp and calcium scores. Specifically, hscrp levels (mg/dl) were 0.24 0.43, 0.33 0.47 and 0.17 0.32, (medians 0.11, 0.15, and 0.06 and ranges 0.02 to 3.10, 0.02 to 1.85, 0.02 to 2.07) for women with none, minimal, and significant coronary calcification, respectively (Fig. 1). However, Spearman rank correlation provided no evidence of a positive association between hscrp and calcium scores Table 1. Cardiac Risk Factors Sample Total (n 172) Non-HRT Users (n 91) Noncurrent Smokers (n 161) Age (yrs) 64.5 7.9 65.5 8.1 64.6 8.0 Smokers (%) Current 5.2 2.1 Never 51.2 50.5 53.4 Past 43.6 47.4 46.6 HRT users (%) 47.1 46.6 BMI (kg/m 2 ) 24.4 4.1 24.4 4.4 24.3 4.1 TC (mg/dl) 256 54 264 59 256 54 HDL (mg/dl) 60.5 18.0 59.4 1.53 60.9 18.0 LDL (mg/dl) 164 49 175 54 165 50 Diabetes (%) 4.7 3.3 5 FH of CAD (%) 44.8 44 44.7 BMI body mass index; CAD coronary artery disease; FH family history; HDL high density lipoprotein; HRT hormone replacement therapy; LDL low density lipoprotein; TC total cholesterol.
JACC Vol. 36, No. 1, 2000 July 2000:39 43 Redberg et al. C-Reactive Protein and Calcium in Postmenopausal Women 41 Figure 1. Box plots illustrating the distribution of levels of C-reactive protein for women with calcium scores in three groups (0 to 10 none ; 11 to 50 minimal ; 50 significant ). Each box spans the 25th through 75th percentile; and the median is shown as a horizontal line. (r 0.10, 95%CI [confidence interval] 0.25 to 0.05, p 0.18). Calcium scores ranged from 0 to 2618, with 77 women (45%) having a calcium score of zero. The average hscrp in this zero calcium group was 0.23 mg/liter, compared to 0.24 mg/liter in the women with calcium present (calcium score 1). In contrast to our a priori hypotheses, we observed an inverse relationship between CRP quintiles and calcium scores in proportional odds model. Specifically, women in the none calcium group were 2.0 times more likely (95% CI 1.1 to 3.7, p 0.028), and women in the minimal group were 3.2 times more likely (95% CI 1.4 to 7.4, p 0.006) to have higher hscrp than women with significant calcium. Controlling for risk factors such as age, smoking status, BMI, exercise time, and HRT increased the strength of this unexpected inverse relationship, so that adjusted odds ratios [OR] for the none and minimal calcium groups were 2.9 (95% CI 1.4 to 5.7, p 0.003) and 5.0 (95% CI 2.0 to 12.4, p 0.001), respectively. Subgroup analyses among noncurrent smokers and non- HRT users likewise showed no positive association between hscrp and calcium scores (Fig. 2), with Spearman rank correlations of r 0.09 for nonsmokers (95% CI 0.24 to 0.07, p 0.27) and r 0.047 for non-hrt users (95% CI 0.25 to 0.61, p 0.66), respectively. DISCUSSION On an a priori basis, we had hypothesized that women with EBCT-documented coronary calcium would have significantly higher plasma concentrations of hscrp than women without such calcification. However, in this cross-sectional study of 172 postmenopausal women, we found no evidence of a positive relationship between hscrp and calcium scores. Indeed, if anything, these data suggest an inverse relationship between hscrp levels and coronary calcium Figure 2. Box plots illustrating the distribution of levels of C-reactive protein for noncurrent smokers (A) and nonhormone replacement therapy (HRT) users (B) with calcium scores in three groups (0 to 10 none ; 11 to 50 minimal ; 50 significant ). Each box spans the 25th through 75th percentile; and the median is shown as a horizontal line. measured by EBCT. This possible inverse relationship is weak and may be due to chance, as the direction of association is not compatible with the a priori hypothesis based on proposed biological mechanisms. Nonetheless, we believe the lack of a positive association between hscrp and coronary calcium score deserves careful consideration. The FLASH participants were similar in demographics and risk factor profile to the postmenopausal women enrolled in the Women s Health Study (10) where hscrp was shown to be an independent predictor of future cardiovascular events. Thus, these data suggest that the EBCT score, while a marker for atherosclerotic burden, may not be as useful as hypothesized in the prediction of actual clinical events in postmenopausal women. These data also raise the possibility that hscrp and EBCT may fundamentally measure different pathophysiologic processes. In this regard, hscrp has been shown to correlate with plasma concentrations of ICAM-1 and IL-6, and thus appears to reflect the extent of low-grade systemic vascular inflammation. Recent evidence suggests that plasma concentrations of IL-6 vary among those with and without a propensity to
42 Redberg et al. JACC Vol. 36, No. 1, 2000 C-Reactive Protein and Calcium in Postmenopausal Women July 2000:39 43 plaque rupture and subsequent coronary events (23). Thus, hscrp, in addition to being a marker for atherosclerotic burden, may also reflect an underlying propensity to plaque instability. In contrast, the lack of correlation in the current data between EBCT score and hscrp suggests that calcification may be less likely to reflect inflammation per se; EBCTdetected calcification may predominantly be a marker for mature and hence stable atherosclerotic plaque, and thus only be an indirect marker for the presence of uncalcified rupture-prone lesions, which may be more likely markers for future cardiac events. Deposition of calcium in atherosclerotic lesions has been shown to be an active process analogous to the formation of bone spicules (24,25). Furthermore, it appears to involve cells of special embryonic lineage. Thus, coronary calcification may not merely be a direct consequence of atherogenesis but rather may depend upon the presence of specific determinants independent of the central processes involved in plaque formation. Such determinants may be more directly related to coronary calcium scores than inflammatory markers, which are associated with cardiac events. In our study, 56% of women had calcium scores close to or actually at a level of zero while exhibiting a full range of hscrp levels, suggesting that there was evidence of underlying vascular inflammation even without coronary calcification. These data may help explain the observation in some studies that individuals with normal EBCT scores may nonetheless suffer MI (26). These findings are consistent with the American Heart Association Writing Group s Scientific Statement, which found that only a small proportion of individuals with atherosclerosis and detectable coronary calcium will develop clinical coronary events (13). Our observation also supports the possibility that these markers detect different components of the atherothrombotic process. Furthermore, these data are consistent with the observation that hscrp predicts future vascular risk in subgroups of women who otherwise appear to be at low coronary risk by traditional methods for risk detection (10). Study limitations. There are several limitations to our study. The sample size is modest and we had relatively few women with high calcium scores. However, in these data, the power to detect a difference in hscrp levels across calcium score groups is nonetheless high. Further, because the negative predictive value of EBCT is critical to evaluating its utility as a screening method, we believe our population to be an appropriate cohort for this analysis. However, as our data came from women without known coronary events, generalizing to a secondary prevention setting or to men must be done with caution. Second, our data are cross-sectional and thus cannot determine whether hscrp or EBCT scores predict event rates. However, women in the FLASH study were similar to those in the Women s Health Study where hscrp did provide a strong independent marker of risk for future cardiovascular events. In contrast, Secci and colleagues (27) have reported that calcium score is a weak predictor of coronary death and infarction, but is most accurate for predicting revascularization procedures. Understandably, as most studies of EBCT to date have been unblinded, it is difficult to assess whether the predictive value of EBCT for revascularization reflects progression of underlying atherosclerosis or physician bias. Recent data from a study of asymptomatic high-risk men with 41-month follow-up found that calcium score by EBCT did not accurately predict cardiac events (28). In sum, this cross-sectional study found no evidence of a positive association between hscrp and coronary calcification as detected by EBCT, despite the fact that large-scale prospective studies have demonstrated that hscrp predicts risk of future coronary events. These data highlight the importance of careful prospective clinical evaluation of emerging technologies such as EBCT in the detection of coronary disease. Such studies, if performed in a blinded fashion such that physician and patient interventions are not biased by knowledge of EBCT score, will provide a critical evaluation of this imaging technology. Reprint requests and correspondence: Dr. Rita F. Redberg, The Division of Cardiology, University of California, San Francisco, 505 Parnassus Avenue, San Francisco, California 94143-0124. E-mail: redberg@medicine.ucsf.edu. REFERENCES 1. Ridker PM. Evaluating novel cardiovascular risk factors; can we better predict heart attacks? Ann Intern Med 1999;130:933 7. 2. Ridker PM, Cushman M, Stampfer MJ, et al. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997;336:973 9. 3. Koenig W, Sund M, Fröhlich M, et al. C-Reactive protein, a sensitive marker of inflammation, predicts future risk of coronary heart disease in initially healthy middle-aged men: results from the MONICA (Monitoring Trends and Determinants in Cardiovascular Disease) Augsburg Cohort Study, 1984 to 1992. Circulation 1999;99:237 42. 4. Ridker PM, Cushman M, Stampfer MJ, et al. Plasma concentration of C-reactive protein and risk of developing peripheral vascular disease (see comments). Circulation 1998;97:425 8. 5. Kuller LH, Tracy RP, Shaten J, et al. Relation of C-reactive protein and coronary heart disease in the MRFIT nested case-control study. Multiple Risk Factor Intervention Trial. Am J Epidemiol 1996;144: 537 47. 6. Tracy RP, Lemaitre RN, Psaty BM, et al. Relationship 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 7. 7. Liuzzo G, Biasucci LM, Gallimore JR, et al. The prognostic value of C-reactive protein and serum amyloid A protein in severe unstable angina. N Engl J Med 1994;331:417 24. 8. Haverkate F, Thompson SG, Pyke SD, et al. Production of C-reactive protein and risk of coronary events in stable and unstable angina. European Concerted Action on Thrombosis and Disabilities Angina Pectoris Study Group. Lancet 1997;349:462 6. 9. 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;98:839 44. 10. Ridker PM, Buring JE, Shih J, et al. Prospective study of C-reactive protein and the risk of future cardiovascular events among apparently healthy women. Circulation 1998;98:731 3. 11. Ridker PM, Glynn RJ, Hennekens CH. C-reactive protein adds to the
JACC Vol. 36, No. 1, 2000 July 2000:39 43 Redberg et al. C-Reactive Protein and Calcium in Postmenopausal Women 43 predictive value of total and HDL cholesterol in determining risk of first myocardial infarction. Circulation 1998;97:2007 11. 12. Blankenhorn D. Coronary arterial calcification: a review. Am J Med Sci 1961;2424:41 9. 13. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcification: pathophysiology, epidemiology, imaging methods, and clinical implications. A statement for health professionals from the American Heart Association. Writing Group. Circulation 1996;94:1175 92. 14. Arad Y, Spadaro LA, Goodman K, et al. Predictive value of electron beam computed tomography of the coronary arteries. 19-month follow-up of 1173 asymptomatic subjects. Circulation 1996;93: 1951 3. 15. Janowitz WR, Agatston AS, Kaplan G, et al. Differences in prevalence and extent of coronary artery calcium detected by ultrafast computed tomography in asymptomatic men and women. Am J Cardiol 1993; 72:247 54. 16. Kwok Y, Kim C, Grady D, et al. Meta-analysis of exercise testing to detect coronary artery disease in women. Am J Cardiol 1999;83: 660 6. 17. Rifai N, Tracy RP, Ridker PM. Clinical efficacy of an automated high-sensitivity C-reactive protein assay. Clin Chem 1999;45:2136 41. 18. Agatston AS, Janowitz WR, Hildner FJ, et al. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990;15:827 32. 19. Cushman M, Meilahn E, Psaty B, et al. Hormone replacement therapy, inflammation, and hemostasis in elderly women. Arterioscler Thromb Vasc Biol 1999;19:893 9. 20. Ridker PM, Hennekens CH, Rifai N, et al. Hormone replacement therapy and increased plasma concentration of C-reactive protein. Circulation 1999;100:713 6. 21. Tracy RP, Psaty BM, Macy E, et al. 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 76. 22. Anderson KM, Wilson PW, Odell PM, et al. An updated coronary risk profile. A statement for health professionals. Circulation 1991;83: 356 62. 23. Biasucci LM, Vitelli A, Liuzzo G, et al. Elevated levels of interleukin-6 in unstable angina. Circulation 1996;94:874 7. 24. Parhami F, Demer LL. Arterial calcification in face of osteoporosis in ageing: can we blame oxidized lipids? Curr Opin Lipidol 1997;8: 312 4. 25. Parhami F, Morrow AD, Balucan J, et al. Lipid oxidation products have opposite effects on calcifying vascular cell and bone cell differentiation. A possible explanation for the paradox of arterial calcification in osteoporotic patients. Arterioscler Thromb Vasc Biol 1997;17: 680 7. 26. Detrano RC, Wong ND, Tang W, et al. Prognostic significance of cardiac cinefluoroscopy for coronary calcific deposits in asymptomatic high risk subjects. J Am Coll Cardiol 1994;24:354 8. 27. Secci A, Wong N, Tang W, et al. Electron beam computed tomographic coronary calcium as a predictor of coronary events: comparison of two protocols. Circulation 1997;96:1122 9. 28. Detrano RC, Wong ND, Doherty TM, et al. Coronary calcium does not accurately predict near-term future coronary events in high-risk adults (see comments). Circulation 1999;99:2633 8.