NHLBI Workshop on C-Reactive Protein: Basic and Clinical Research Needs
July 10-11, 2006

CRP Measurement: Methods, Accuracy, and Variability
Russell Tracy

This discussion will focus on assays of CRP used clinically. This topic has been reviewed recently (1). CRP, a serum protein produced primarily in the liver in response to activation of the innate immune system, is generally measured using an immunoassay. Of the liver proteins and cytokines which seem to in some way reflect this activation, CRP has been used most widely, followed by fibrinogen. Epidemiologically, these two proteins yield approximately the same predictive power (2,3). Technically, there are features of the CRP assay which make it more readily adapted to a clinical laboratory. Whether there are metabolic features of CRP as a biomarker that make it superior, remains unclear.

Many clinical laboratories offer CRP measurements using relatively insensitive nephelometric assays, yielding actual values only when CRP is above ~10 mg/L, as a consequence of it’s use as an acute phase reactant. For research purposes, an immunoradiometric assay had been developed which yielded values across the healthy reference range (4), and as clinical research increased ELISA methods followed (5,6). Currently, CRP is predominantly assayed under the term “high-sensitivity CRP”, using semi- or fully automated ELISA or nephelometric methods calibrated to yield values at most concentrations of CRP(e.g., (7)). By some methods, those samples with the lowest CRP levels may not yield a value.

Standardization (accuracy) of clinical assays is reasonably simple using established reference materials (8), and precision appears to be driven predominantly by the effort put into assay design and implementation, with analytical coefficients of variation (CVs) under 5% common. In our experience immunoassays such as the CRP ELISA exhibit the greatest variance at the lowest and highest values.

Assay variance may be stratified into three types: analytical, within subject, and between subject (9). Using cholesterol as a comparison, most CRP assays exhibit slightly greater within subject variance, as might be expected (6). From an epidemiological perspective this is well tolerated, since the between subject variance is also greater than for cholesterol, making a single measurement roughly equivalent in it’s ability to accurately rank the members of a population. Clinically, although repeat measures of CRP at different times seem to classify people into general categories reasonably consistentlyn (7,10), the levels of analytical and within subject variance need to be considered if the use of CRP is widespread. There are current recommendations from a CDC/AHA consensus panel which seem reasonable (1):

  1. Of the inflammatory markers identified, C-reactive protein (CRP) has the analyte and assay characteristics that are the most conducive for use in practice.
  2. To obtain a CRP concentration in metabolically stable patients, 2 measurements, fasting or nonfasting, should be made (optimally 2 weeks apart) and the results averaged. If the CRP level is >10 mg/L, then the test should be repeated and the patient examined for sources of infection or inflammation.
  3. CRP results should be expressed only as milligrams per liter and expressed to 1 decimal point.
  4. Risk assessment should be modeled after the lipids approach via 3 risk categories: low risk, average risk, and high risk. On the basis of the CRP population distributions, the following tertiles are recommended for categorizing patients: low risk, <1.0 mg/L; average risk, 1.0 to 3.0 mg/L; and high risk, >3.0 mg/L. It should be recognized that other acute inflammatory conditions may result in mildly to moderately increased CRP levels, such as inflammatory bowel disease, rheumatoid arthritis, and long-term alcoholism.
  5. Performance goals for CRP measurement, similar to those developed for total cholesterol, HDL and LDL cholesterol, and triglycerides, need to be developed with a view toward better characterization of the total allowable error required to measure CRP reliably.


  1. Myers GL, Rifai N, Tracy RP, Roberts WL, Alexander RW, Biasucci LM, Catravas JD, Cole TG, Cooper GR, Khan BV, Kimberly MM, Stein EA, Taubert KA, Warnick GR, Waymack PP. CDC/AHA Workshop on Markers of Inflammation and Cardiovascular Disease: Application to Clinical and Public Health Practice: report from the laboratory science discussion group. Circulation. 2004;110:e545-9.
  2. Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A, Lowe GD, Pepys MB, Gudnason V. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med. 2004;350:1387-97.
  3. Danesh J, et al. Plasma fibrinogen level and the risk of major cardiovascular diseases and nonvascular mortality: an individual participant meta-analysis. Jama. 2005;294:1799-809.
  4. Shine B, de Beer F, Pepys M. Solid phase radioimmunoassay for human C-reactive protein. Clin. Chim. Acta. 1981;117.
  5. Wilkins J, Gallimore JR, Moore EG, Pepys MB. Rapid automated high sensitivity enzyme immunoassay of C-reactive protein. Clin Chem. 1998;44:1358-61.
  6. Macy E, Hayes T, Tracy R. Variability in the measurement of C-reactive protein in healthy subjects: implications for reference interval and epidemiological applications. Clin. Chem. 1997;43:52-58.
  7. Rifai N, Tracy RP, Ridker PM. Clinical efficacy of an automated high-sensitivity C-reactive protein assay. Clin Chem. 1999;45:2136-41.
  8. Kimberly MM, Vesper HW, Caudill SP, Cooper GR, Rifai N, Dati F, Myers GL. Standardization of immunoassays for measurement of high-sensitivity C-reactive protein. Phase I: evaluation of secondary reference materials. Clin Chem. 2003;49:611-6.
  9. Fraser CG, Harris EK. Generation and application of data on biological variation in clinical chemistry. Crit Rev Clin Lab Sci. 1989;27:409-37.
  10. Ockene IS, Matthews CE, Rifai N, Ridker PM, Reed G, Stanek E. Variability and classification accuracy of serial high-sensitivity C- reactive protein measurements in healthy adults. Clin Chem. 2001;47:444-50.

Back to Agenda
Skip footer links and go to content
Twitter iconTwitterExternal link Disclaimer         Facebook iconFacebookimage of external link icon         YouTube iconYouTubeimage of external link icon         Google+ iconGoogle+image of external link icon