EXPEDITED REVIEW: COMMENTARY
High-Sensitivity C-Reactive Protein in Every Chart?The Use of Biomarkers in Individual Patients
Ori Ben-Yehuda, MD*
Department of Medicine, University of California, San Diego Medical Center, San Diego, California
Manuscript received March 6, 2007;
revised manuscript received March 20, 2007,
accepted March 28, 2007.
* Reprint requests and correspondence: Dr. Ori Ben-Yehuda, Department of Medicine, UCSD Medical Center, 200 West Arbor Drive, San Diego, California 92103. (Email: obyehuda{at}acc.org).
 |
Abstract
|
|---|
The use of high-sensitivity C-reactive protein (hsCRP) for risk stratification for cardiovascular disease is supported by epidemiologic evidence but remains controversial. The metabolic milieu in which hsCRP is likely to be elevated, namely abdominal obesity and insulin resistance, provides a framework for the understanding of the role of hsCRP as well its limitations. This commentary provides a critical assessment of the data in support of the use of hsCRP in clinical practice.
It has been over a decade since high-sensitivity C-reactive protein (hsCRP) was proposed as a risk marker for future cardiovascular events. The large body of evidence showing the incremental value of hsCRP when added to traditional risk factors has also been used to support the hypothesis that atherosclerosis is first and foremost an inflammatory disease. In this issue of the Journal, Dr. Paul Ridker, one of the leading researchers in this field, eloquently defends the use of hsCRP and advocates that a consensus can be reached for the use of this biomarker, particularly for patients at intermediate risk for cardiovascular events (1). Is this a closed and shut case, or is there still room for controversy regarding the utility of hsCRP? Should hsCRP be an integral part of cardiovascular risk assessment? The answer to this question has important implications for the inflammatory hypothesis in particular, as well as for the clinical use of biomarkers in general as prognostic markers.
Critics of the use of hsCRP can cite several studies, including recent reports from the Framingham Offspring Study (2), PROSPER (Prospective Study of Pravastatin in the Elderly at Risk) (3), and the multiethnic Dallas Heart Study (4), in which hsCRP provided only modest or no incremental information compared with traditional risk factors. The Dallas Heart Study is of particular interest because the assessment is of the atherosclerosis burden itself (detected by electron beam computed tomographic calcium score and magnetic resonance imaging-detected aortic plaque). How can we make sense of this literature, and is there a biological explanation for these divergent findings?
In the assessment of hsCRP (as with any biomarker) and its role in cardiovascular disease, we should first examine the biological context in which an elevation of this biomarker occurs. Proponents of the use of hsCRP emphasize that plaque can be a source of CRP as well as a target for its pathological effects. The main source of CRP in the body, however, is the liver (5). And although vascular inflammation may contribute to an elevation of CRP in the blood, the main association is with abdominal obesity and insulin resistance. Abdominal adipocytes are a rich source of various inflammatory cytokines, including interleukin-6 (6). Interleukin-6 in particular is a potent messenger for the liver to secrete CRP. Indeed, abdominal obesity and insulin resistance are predictors of an elevated hsCRP, and the presence of the metabolic syndrome correlates strongly with an elevated hsCRP level (5,7). Conversely, patients with 0, 1, or even 2 risk factors for the metabolic syndrome have low CRP levels (7).
Given the clear association between abdominal obesity, insulin resistance, and elevated hsCRP, we should examine the metabolic milieu in which elevated levels of hsCRP occur. The patient with metabolic syndrome and insulin resistance is likely to show a pattern B lipoprotein profile, with elevations in highly atherogenic apolipoprotein-B–containing particles, such as very-low-density lipoprotein (VLDL) remnants, intermediate-density lipoprotein, and chylomicron remnants, as well as an increase in small-dense low-density lipoprotein (LDL) and therefore LDL particle number (8). A fairly good assessment of the presence of this highly atherogenic profile can be obtained by calculating non–high-density lipoprotein (HDL) cholesterol, a simple measurement already available in our patients' charts. Thus, the risk attributed to hsCRP actually may be the result of elevations of non-LDL atherogenic particles rather than an independent contribution of inflammation. Indeed, most of the studies in support of the use of hsCRP did not fully control for the presence of the metabolic syndrome, abdominal obesity, insulin resistance, or the presence of a pattern B profile. One exception is the WHS (Women's Health Study), in which Ridker et al. (7) examined the effect of hsCRP in subjects with and without the metabolic syndrome. Most importantly, in the absence of the metabolic syndrome, an elevated hsCRP did not significantly predict increased risk significantly detracting from the argument that CRP is an independent and causally implicated mediator. And although hsCRP >3 mg/dl in subjects with the metabolic syndrome did predict even added risk, metabolic syndrome subjects with hsCRP <3 mg/dl were already at increased risk compared with those without the metabolic syndrome (relative risk of any cardiovascular event of 2.3 and 3.1 for coronary events) (7). For the patient with metabolic syndrome, the presence of an elevated hsCRP may further indicate the need for aggressive intervention. This is one group of patients in whom hsCRP may be useful as a barometer of the overall inflammatory state, further identifying the particularly high-risk patient.
High-sensitivity CRP also has been proposed as a predictor of clinical response to statin therapy. It has been claimed that the benefit of statin therapy may be related to the so-called pleiotropic effects of statins, particularly their anti-inflammatory and hsCRP-lowering effects. As Ridker et al. (9) point out, patients in the PROVE-IT–TIMI-22 (Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis In Myocardial Infarction-22) trial who had LDL <70 mg/dl as well as hsCRP <2 mg/dl had lower event rates compared with those who did not have both of these target goals achieved. This analysis, however, did not take into account baseline LDL cholesterol levels or baseline hsCRP levels. Similarly, the finding that patients with elevated hsCRP levels are particularly likely to benefit from statin therapy in terms of cardiovascular event reduction is confounded by the fact that it is the low-HDL, high-triglyceride patient (who also is likely to be the metabolic syndrome insulin-resistant patient and therefore to have a high hsCRP) who is the one most likely to benefit from statin therapy (10).
Statin therapy has been shown to reduce hsCRP levels. What is the mechanism of hsCRP reduction seen with statins? Is it a direct anti-inflammatory effect of the statins, or are other mechanisms possible? It also has been claimed that LDL reductions do not correlate with hsCRP reduction. These analyses are limited by not taking into account the baseline hsCRP level, because a significant reduction is not expected in subjects without a baseline elevation. Additionally, a meta-analysis of various statin and statin-ezetimibe combination trials has found that the majority of the CRP reduction is indeed related to the degree of LDL lowering (11). The data for ezetimibe are particularly interesting and relevant in this regard. As monotherapy ezetimibe does not reduce hsCRP levels, is only a modest LDL cholesterol-lowering drug (with an average LDL reduction of 18% compared with 30% to 40% with starting doses of statins), and also leads to up-regulation of cholesterol synthesis. In combination with statin therapy, however, the combined inhibition of intestinal and hepatic cholesterol pathways results in both a significant additive LDL cholesterol-lowering effect and a greater reduction in hsCRP compared with statin monotherapy. The ezetimibe/statin data therefore suggest that the CRP reductions seen with statin therapy may be related to the reduction in liver cell cholesterol content rather than a direct anti-inflammatory effect in the vasculature.
An important distinction needs to be made between the predictive value of a risk factor in a population versus that in an individual (12). Beyond the issue of whether hsCRP is independently predictive beyond its metabolic milieu, there is significant overlap in hsCRP levels between those with and without events. Furthermore, elevated hsCRP levels are very prevalent. Indeed, in the Dallas Heart Study, about one-half of black subjects between the ages of 30 and 65 years had CRP >3 mg/dl, with 63% of black women having levels above 3 mg/dl (13). Most positive studies on hsCRP have compared risk based not on hsCRP as a linear variable but on hsCRP quartiles. Coupled with the biological linkage of hsCRP and other metabolic risk factors, it is therefore not surprising that the sensitivity and specificity of the test may be insufficient to adequately predict risk in an individual. In this regard, the receiver-operating characteristic curve and its attendant C-statistic is an appropriate test for assessing the predictive value of a test in an individual patient.
Given the previously detailed limitations for hsCRP, is there any role for its routine measurement? As often is the case in medicine, the answer is: "it depends." Atherosclerosis does not occur in an isolated state, and the inflammatory milieu can be identified by a multitude of signs. Indeed, similar to hsCRP, other markers of an inflammatory state, such as white blood cell count and particularly the neutrophil count, have been shown to predict coronary disease (14). High-sensitivity CRP is easy to measure and is a relatively inexpensive test. Its strengths as well as its limitations are a product of its relatively nonspecific nature—elevated in many and reflective of an underlying metabolic milieu. Some will therefore find this "global" test of inflammation useful, particularly in educating the individual patient on the harmful effects of their metabolic perturbations. Others will find it predictable and confirmatory.
How can we improve risk stratification given the limitations of hsCRP and other biomarkers (including LDL cholesterol, which is, however, not only a biomarker but also a target for therapy)? The underlying disease is ultimately atherosclerosis, an anatomically defined entity. Imaging modalities for atherosclerosis (15) may provide better predictive tools, and as these improve and become more affordable, they are likely to become the ideal partners to our traditional risk assessment tools.
|
References
|
|---|
- Ridker P. C-reactive protein and the prediction of cardiovascular events among those at intermediate risk: moving an inflammatory hypothesis toward consensus. J Am Coll Cardiol 2007;49:xxxx–xxxx.
- Wang TJ, Gona P, Larson MG, et al. Multiple biomarkers for the prediction of first major cardiovascular events and death N Engl J Med 2006;355:2631-2639.[Abstract/Free Full Text]
- Sattar N, Murray HM, McConnachie A, et al. C-reactive protein and prediction of coronary heart disease and global vascular events in the prospective study of pravastatin in the elderly at risk (PROSPER) Circulation 2007;115:981-989.[Abstract/Free Full Text]
- Khera A, deLemos JA, Peshock RM. Relationship between C-reactive protein and subclinical atherosclerosisThe Dallas Heart Study. Circulation 2006;113:38-43.[Abstract/Free Full Text]
- Lemieux I, Pascot A, Prud'homme D, et al. Elevated C-reactive protein: another component of the atherothrombotic profile of abdominal obesity Arterioscler Thromb Vasc Biol 2001;21:961-967.[Abstract/Free Full Text]
- Bajaj M, Ben-Yehuda O. A big fat wedding: association of adiponectin with coronary vascular lesions J Am Coll Cardiol 2006;48:1163-1165.[Free Full Text]
- Ridker PM, Buring JE, Cook NR, et al. C-reactive protein, the metabolic syndrome, and risk of incident cardiovascular events an 8-year follow-up of 14,719 initially healthy American women Circulation 2003;107:391-397.[Abstract/Free Full Text]
- Krauss RM. Dietary and genetic probes of atherogenic dyslipidemia Arterioscler Thromb Vasc Biol 2005;25:2265-2272.[Abstract/Free Full Text]
- Ridker PM, Morrow DA, Rose LM, et al. Relative efficacy of atorvastatin 80 mg and pravastatin 40 mg in achieving the dual goals of low-density lipoprotein cholesterol <70 mg/dl and C-reactive protein <2 mg/l: an analysis of the PROVE-IT TIMI-22 trial J Am Coll Cardiol 2005;45:1644-1648.[Abstract/Free Full Text]
- Ballantyne CM, Olsson AG, Cook TJ. Influence of low high-density lipoprotein cholesterol and elevated triglyceride on coronary heart disease events and response to simvastatin therapy in 4S Circulation 2001;104:3046-3051.[Abstract/Free Full Text]
- Kilnay S. Low-density lipoprotein-dependent and -independent effects of cholesterol-lowering therapies on C-reactive protein: a meta-analysis. J Am Coll Cardiol 2007;49:2003–9.
- Ware JH. The limitations of risk factors as prognostic tools N Engl J Med 2006;355:2615-2617.[Free Full Text]
- Khera A, McGuire DK, Murphy SA, et al. Race and gender differences in C-reactive protein levels J Am Coll Cardiol 2005;46:464-469.[Abstract/Free Full Text]
- Horne BD, Anderson JL, John JM, et al. Which white blood cell subtypes predict increased cardiovascular risk? J Am Coll Cardiol 2005;45:1638-1643.[Abstract/Free Full Text]
- Taylor AJ, Bindeman J, Feuerstein I, et al. Coronary calcium independently predicts incident premature coronary heart disease over measured cardiovascular risk factors J Am Coll Cardiol 2005;46:807-814.[Abstract/Free Full Text]
|
Top
Abstract
References