Accessible Search Form           Advanced Search

Skip left side navigation and go to content


Epidemiology of C-Reactive Protein and Coronary Heart Disease: An Overview
John Danesh

C-reactive protein (CRP) is the classical acute-phase reactant produced by the liver, the levels of which can increase up to 10 000-fold in response to infection and tissue damage. Despite the potential for such substantial spikes in CRP levels during inflammatory states, the year-to-year variability of CRP within apparently healthy adults is relatively moderate, comparable to those observed for blood pressure or serum cholesterol values. Hence, as interest has increased in the possibility that inflammation may be relevant to the aetiology of cardiovascular diseases, many studies have investigated whether higher than average CRP levels over the long-term (eg, 2.4 mg/L vs 1 mg/L, corresponding approximately to the mean values in the top third vs bottom third of CRP levels in the population) are associated with subsequent risk of cardiovascular diseases. Moreover, interest in the possibility that CRP may play a direct causal role in the development of cardiovascular diseases (and, by implication, whether it might be an important therapeutic target) has been stimulated by reports in experimental studies that CRP exerts actions that might establish and promote atherosclerosis.

By 2006, more than 30 long-term prospective studies in essentially general populations had reported generally positive associations between circulating concentrations of C-reactive protein and the risk of coronary heart disease (CHD), involving a total of more than 10 000 incident CHD cases. By mid-2006, the large majority of these studies had shared data with the coordinating centre of the Emerging Risk Factors Collaboration, an initiative to pool individual participant data from each these studies in order to help characterize more reliably than previously possible the age- and sex-specific associations of CRP with CHD (and with other major vascular outcomes), after making corrections for within-person variability in levels of CRP and of possible confounding factors. This initiative should provide reliable information on the shape of the association between CRP levels and CHD and of its strength under different circumstances, such as by sex, at different ages, and at different levels of established risk factors and other emerging markers. Investigation of the combined dataset should also help to address uncertainties related to the predictive value of CRP measurement on top of measurement of known causal cardiovascular risk factors.

But, although the Emerging Risk Factors Collaboration will help to quantify associations between CRP levels and CHD more precisely and in greater detail than has been previously possible, inherent limitations in such observational studies may not enable even such data pooling to distinguish clearly whether CRP is largely a causal risk factor in CHD or mainly a marker of established cardiovascular risk factors to which it is correlated (such as obesity, blood pressure, and smoking) or mainly a marker of subclinical disease or some combination of these possibilities. Statistical adjustment in observational studies may not fully correct for potential confounding since not all relevant factors are measured (and even measured confounders are often measured with inaccuracy) and because it is difficult to make allowances for the potential effects of early or subclinical disease on CRP levels (ie, “reverse causation”). One approach to eliminate such residual biases is, of course, the conduct of randomised trials of selective CRP-lowering agents in long-term studies of CHD prevention (and such compounds are in the early stages of development); by contrast, use of existing agents known to lower CRP levels, such as statins, would not provide a specific test of causality because such agents also affect several other factors in lipid pathways.

A further complementary approach to help minimise such biases involves genetic epidemiology. “Mendelian randomization” experiments attempt to minimize confounding and avoid reverse association bias by measurement of common polymorphisms or haplotypes in regulatory regions of the CRP gene that have been reliably associated with differences in circulating CRP concentration (but not with any known change in CRP function). According to Mendel’s second law, the inheritance of genetic variants should be subject to the random assortment of maternal and paternal alleles at the time of gamete formation. So, if CRP levels actually increase the risk of CHD, then carriage of alleles (or haplotypes) that expose individuals to a long-term elevation of CRP should confer an increased risk of coronary events in proportion to the difference in CRP levels attributable to the allele. Because of the randomised allocation of alleles, potential confounders should be distributed evenly among the genotypic classes, and any bias due to reverse causation should be avoided because genotypes are fixed at conception and not prone to modification by the onset of disease. The CRP-Coronary Genetics Collaboration has been established to help generate and pool data on relevant CRP genetic variants in a total of about 15 000 CHD cases and about 15 000 controls to enable sufficiently powerful mendelian experiments to help test any causal relevance of CRP levels to CHD.


  1. 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;350:1387-97.
  2. Fibrinogen Studies Collaboration. Plasma fibrinogen level and the risk of major cardiovascular diseases and nonvascular mortality: an individual participant meta-analysis. JAMA 2005;294:1799-1809.
  3. Casas JP, Shah T, Cooper J et al. Insight into the nature of the CRP-coronary event association using Mendelian randomization. Int J Epidemiol 2006 Mar 24; [Epub ahead of print]
  4. Timpson NJ, Lawlor DA, Harbord RM et al. C-reactive protein and its role in metabolic syndrome: mendelian randomisation study. Lancet 2005;366:1954-59.
Back to Agenda
Twitter iconTwitterimage of external icon Facebook iconFacebookimage of external icon YouTube iconYouTubeimage of external icon Google+ iconGoogle+image of external icon