CRP for the Clinician J. C. Kaski, D.Sc., M.D., D.M (Hons), F.E.S.C., F.R.C.P., F.A.C.C. F.A.H.A Professor of Cardiovascular Science Head, Cardiovascular Sciences Research Centre St George s, University of London, London, UK Conflict of Interest: None
Inflammation and Atherosclerosis Image courtesy of Roche Inflammation plays a central role in all phases of atherosclerosis CRP is a good marker of inflammation and an excellent research tool
Stratification of Cardiovascular Risk Current Challenges Traditional risk factors have been shown to predict coronary heart disease and cardiovascular disease (CVD) in a large number of prospective observational studies (Wilson PW et al. Circulation. 1998;97:1837 1847) High v. Low risk issues There is interest in developing more effective strategies to identify the true risk of persons considered to be at an intermediate level of risk for CHD based on conventional assessment. Intermediate risk is an arbitrary/dynamic category that is likely to change as risk assessment becomes more efficient CRP may be used to refine estimates of cardiovascular risk stratification (Ridker PM et al. NEJM 1987; 336:973-979; Ridker PM et al. N Engl J Med. 2000;342:836 843)
Characteristics of Clinically Useful Markers of Cardiovascular Risk Relate to pathogenesis of CV disease Provide independent information over and above that afforded by conventional markers of risk Cost-effective, practical and reproducible Validated in real life settings, with a large area under ROC curve Useful for patient risk reclassification
What Clinicians Want to Know Before Using a New Biomarker of Cardiovascular Risk Useful for patient management? Helps with therapeutic decisions/primary prevention? Changes clinical practice as a result of its use for screening, diagnosis, prognosis, or treatment
CRP and Cardiovascular Risk Marker of Risk Associated with the disease but need not be causally linked; it may be a measure of the disease process itself Risk Factor Associated with a disease because it is in the causal pathway leading to the disease Pathogenic Factor A cause of the disease
C-Reactive Protein is a Marker of Cardiovascular Risk In apparently healthy subjects, individuals with conventional risk factors for CAD and in patients with established cardiovascular disease (i.e. stable and unstable CAD), increased hs-crp levels have been shown to be an independent marker of risk No controversial issues here!
EVIDENCE AGAINST A PATHOGENIC ROLE FOR CIRCULATING CRP IN ATHEROSCLEROSIS Animal and in vitro studies Human genetic studies Hirschfield GM et al. Transgenic human C-reactive protein is not proatherogenic in apolipoprotein E-deficient mice. Proc Natl Acad Sci 2005 Koike T et al. Human CRP does not promote atherosclerosis in transgenic rabbits. Circulation 2009; 120:2088-94. Taylor KE et al. C-reactive proteininduced in vitro endothelial cell activation is an artefact caused by azide and lipopolysaccharide. Arterioscler Thromb Vasc Biol 2005;25:1225-30.
MENDELIAN RANDOMISATION STUDIES The CRP /CHD Genetics Collaboration (BMJ 2011; 342:d548) confirmed in >46 000 CAD patients and >147 000 controls that: 1.Genetically raised concentrations of CRP are unrelated to conventional risk factors and risk of CAD 2.CRP concentration itself is unlikely to be even a modest causal factor in CAD Associations between CRP and CAD reported in prospective studies are likely to arise from reverse causality or residual confounding
Predictive Role of CRP in ACS CRP predicts inhospital patient outcome (n=34) Liuzzo. NEJM 1994, death at 15 d - Morrow [TIMI 11 A; n= 437] JACC 1998 and Montalescot [ESSENCE (n=68) ] (1999) death+mi+ua (n=105,15 mg/l) Ferreiros Circulation 1999 and 30-day mortality GUSTO-IV ACS Study (>7000 pts) James et al. JACC 2003 hs-crp level assessed at hospital admission in 1128 ACS patients was a modest predictor for 30-day mortality Combined with the GRACE risk score, CRP (cut off 22 mg/l) improved risk classification Schiele F et al. Eur Heart J 2010;31:290-297
Clinical Usefulness of hs-crp for Risk Prediction in Patients with ACS Nested case-cohort study within OPUS-TIMI 16 to evaluate hscrp in ACS. hscrp measured in 3225 ACS patients and compared in patients who suffered an adverse cardiac outcome within 10 months of study entry and those who did not. Increased baseline concentrations of hscrp were strongly associated with mortality and heart failure across the ACS spectrum. Index diagnosis-specific CRP cut-points were identified and predictive value of the marker differed according to timing of CRP measurement BM Scirica et al. Clinical Application of C-Reactive Protein Across the Spectrum of Acute Coronary Syndromes Clin Chem. 2007;53:1800-1807.
A Comparative Study Of Multiple Inflammatory Markers For Risk Prediction In Acute Coronary Syndrome SIESTA Study J.C. Kaski et al. Atherosclerosis 2010 Prospective cohort study of NSTE-ACS patients admitted to 25 hospitals in Spain 610 patients (73% men) 36% unstable angina and 64% NSTEMI 360-day follow-up AIMS To compare the relative prognostic ability of different inflammatory biomarkers To assess whether inflammatory biomarkers contribute additional independent information to that of clinical risk scores
CRP DOES NOT IMPROVE TIMI RISK SCORE S C-STATISTIC In the SIESTA study, the area under the ROC curve for TIMI risk score was 0.58, and was not improved by the addition of CRP as a new variable (p= 0.92) JC Kaski et al. Atherosclerosis 2008 Eggers KM et al Am Heart J 2010 - An increase of the C-statistic was achieved only after the addition of GDF-15 (0.81, P =.003) and cystatin C (0.81, P =.035) CRP did not improve prognostic discrimination in ACS patients- CRP modified TIMI risk score TIMI risk score Reference curve
Normal hs-crp levels at the onset of first STEMI in 41% of cases (a multiethnic case-control study) In over 40% of 887 STEMI patients admitted within 6 h from symptoms onset, hscrp serum levels were <2 mg/l. The number of patients with low CRP prior to the AMI could be even larger considering that hscrp is likely to have increased before sampling as a result of the acute process in at least some patients Using a 2 mg/l cut point, the measurement of hscrp prior to the event would not have predicted 41% of MIs in these patients without evidence of previous CAD Cristell N et al. JACC 2011.
Use of CRP in the setting of ACS - ESC Guidelines for the Management of NSTE-ACS Eur Heart J 2011 The committee acknowledges the predictive role of CRP in ACS, as suggested in numerous clinical studies No recommendation is made regarding the use of CRP for diagnosis of ACS or patient risk stratification in this setting
CRP Levels and Clinical Outcome in Stable Angina PEACE (Prevention of Events with Angiotensin Converting Enzyme inhibition) Trial 3771 stable CAD patients Trandolapril vs placebo Elevated hs-crp >1 mg/l is a predictor of CV events (death, AMI and stroke) High CRP predicts new onset heart failure High CRP patients did not benefit from ACEI CUMULATIVE INCIDENCE OF DEATH, AMI OR STROKE (%) P<0.001 CRP >3 mg/l CRP 1-3 mg/l CRP <1 mg/l MONTHS Sabatine MS et al. Circulation 2007; 115:1528-36
Evaluating the Quality of Research into a Single Prognostic Biomarker: A Systematic Review and Meta-analysis of CRP in Stable CAD Hemingway H et al. PLoS Med 2010; 7(6): e1000286. doi:10.1371 83 studies, 61,684 patients and 6,485 outcome events. Pooled relative risk (RR) for the top versus bottom third of CRP distribution: 1.97 (95% CI 1.78 2.17) (heterogeneity +) Only 13 studies adjusted for conventional risk factors and these showed a RR of 1.65 (95% CI 1.39 1.96). Adjusting for publication bias (for which there was strong evidence; p =0.001) reduced the RR to 1.19 (95% CI 1.13 1.25). Conclusion: Reporting and publication bias make any independent association between CRP and prognosis among patients with stable CAD sufficiently uncertain that no clinical practice recommendations can be made
CRP and CV Risk in Population Studies The Emerging Risk Factors Collaboration Study RR CAD 1.63 (1.51-1.76) 1.37 (1.27-1.48) 1.23 (1.07-1.42) RR Stroke 1.44 (1.32-1.57) 1.27 (1.15-1.40) 1.32 (1.18-1.49) RR Vascular mortality Lancet 2010; 375: 132 40 An Individual Participant Meta-Analysis Individual records of 160,309 Adjustment people without Age a history & sex of Risk CVD (27,769 factors fatal Fibrinogen or non-fatal outcomes) from 54 prospective studies. Within-study regression analyses adjusted for within-person variation in risk factor (RF) levels Results CRP concentration was linearly associated with conventional RFs and other inflammatory markers and nearly log-linearly with the risk of ischaemic vascular disease and non-vascular mortality (i.e. cancer, trauma and respiratory) NON-SPECIFIC RISK MARKER 1.71 (1.53-1.91) 1.55 (1.37-1.76) 1.34 (1.18-1.52) Adjustment for conventional RFs and plasma fibrinogen considerably weakened the associations RR Non-vascular 1.55 (1.41-1.69) 1.54 (1.40-1.68) 1.34 (1.20-1.50) of CRP mortality with risk of CAD, ischaemic stroke and non-vascular death
No Incremental Coronary Risk Prediction Using CRP and Other Novel Markers ARIC (AR Folsom et al. Arch Intern Med 2006) Case-cohort study, n=15 792 (45-74 yrs) Events: Death, AMI, PCI/CABG @ DEATH 5 yr f-up Conventional RFs vs 19 novel biomarkers (including CRP) - AUC (ROC) to assess additional contribution beyond RFs for CAD (AUC 0.8) CRP did not add significantly to the AUC, and neither did most other novel risk factors MAJOR CV EVENTS Routine measurement of novel biomarkers is not warranted for risk assessment Conventional RFs 10 biomarkers measured in RFs + Log CRP Probability of Event Probability of Event Men TJ Wang et al. NEJM 2006 3209 subjects (general unselected population) attending a routine examination cycle of the Framingham Heart Study. CRP, BNP, NT pro ANP, Women renin, aldosterone, fibrinogen, PAI-1, D- dimer, homocysteine, etc. Multiple biomarkers DO NOT contribute substantial additional information for the prediction of Risk Decile major CV events and death
Better Methodologies Required for Assessment? Descriptive statistics such as the relative risk of a new factor that is added to a multivariable prediction and the c statistic can be inadequate to show how such new factors may mediate risk in a population setting. New methods may help to better to quantify such risk Improved patient risk reclassification using the Reynolds Risk Score (CRP and Parental history)- PM Ridker et al. JAMA. 2007;297:611-619 C-Reactive Protein resulted in moderate improvement in reclassification of risk Cardiovascular Risk in the Framingham Heart Study Wilson PW et al. Circ Cardiovasc Qual Outcomes. 2008;1:92-97
Novel and Conventional Biomarkers for Prediction of Incident Cardiovascular Events in the Community Melander O et al. JAMA. 2009;302(1):49-57 5067 participants (age of 58) from the population-based, prospective epidemiologic Malmo Diet and Cancer study, without CVD at baseline (1991-1994). Contemporary set of biomarkers, 12.8 yrs f-up, and stateof-the-art statistical approach For cardiovascular event prediction, CRP and N-BNP together improved the conventional risk factor C statistic of 0.758 by only 0.007 (P=0.04). The whole set of significant biomarkers together increased the C statistics by ~0.01 RECLASSIFICATION The biomarkers changed the risk classification, based on conventional factors, for only 8% of patients for CV risk and 5% for coronary risk. Net improvement in reclassification was not significant for CV events Correct reclassification was confined to down-classification of individuals without events rather than up-classification of those with events.
JUPITER Study Ridker et al. NEJM 2008 (Justification for the Use of statins in Primary prevention: an International Trial Evaluating Rosuvastatin) A randomized, double-blind, placebo controlled, multicenter trial conducted at 1315 sites in 26 countries 17,802 apparently healthy men and women with LDLcholesterol levels <130 mg per deciliter (3.4 mmol/l) and hs- CRP levels 2.0 mg per liter Patients were randomized to rosuvastatin, 20 mg daily, or placebo An event-driven trial designed to continue until 520 confirmed primary end points had been documented. The trial was stopped after a median follow-up of 1.9 years
CAN CRP GUIDE THERAPY? - JUPITER Study Statin therapy prevented first-ever events in apparently healthy subjects with LDL-C <130 mg/dl but elevated hscrp (2mg/L) NNT to prevent 1 end-point= 25 Canadian guidelines for the diagnosis and treatment of dyslipidemia and prevention of cardiovascular disease in the adult 2009 recommendations J Genest et al. Can J Cardiol 2009;25:567-579. Ridker PM et al. N Engl J Med 2008;359:2195207 Relative hazard reductions in the rosuvastatin group were similar for women (46%) and men (42%) and were observed in every subgroup evaluated (age, ethnic group, traditional risk factors, and Framingham risk score)
Problems with JUPITER METHODOLOGICAL Lack of a low-crp arm Rosuvastatin lowers both LDL-C and hs-crp ~40% of subjects had metabolic syndrome Early termination why? 240 serious events (truncated trials overestimate effects) Questionable interpretation of mortality data. Curves converging at end of trial CLINICAL AND EPIDEMIOLOGICAL ISSUES Lack of detailed CV mortality data - No SCD reported- and identical CV mortality in placebo and rosuvastatin groups Low fatal to non-fatal MI ratio and very low case fatality rate (<9%) mainly in the placebo group (rosuvastatin 29%)
JUPITER Study Findings not Endorsed by Results of the Heart Protection Study In 69 UK hospitals, 20 536 men and women 40 80 yrs at high risk of CV events were randomised to simvastatin 40 mg/d or placebo for 5 yrs. Patients were categorised into six baseline CRP groups. Primary endpoint for subgroup analyses: major vascular events, i.e. coronary death, MI, stroke,pci/cabg Effect of simvastatin allocation on vascular events by baseline concentration of CRP Simvastatin treatment significantly reduced the incidence of first major vascular events, 24% (95% CI 19 28) There were significant risk reductions in each of the baseline CRP groups, including participants with CRP <1.25 mg/l (29% risk reduction, 99% CI 12 43; p<0.0001). Participants with high baseline CRP had higher mean BMI, LDL cholesterol, and TG concentration, and lower mean HDL-cholesterol concentration Lancet 2011; 377: 469 76
Peter S. Sever*, Neil R Poulter, Choon L Chang, Aroon Hingorani, Simon McG Thom, Alun D Hughes, Paul Welsh, Naveed Sattar, on behalf of the ASCOT Investigators *Imperial College London Evaluation of CRP, prior to and on-treatment, as a predictor of benefit from atorvastatin: Nested case control study from ASCOT 485 cases (355 CHD & 130 stroke) matched with 1,367 controls - CRPmeasured at baseline and 6 months Investigator-led, multinational randomised controlled trial conducted in 19 342 hypertensive patients, 40-79 yrs, with no prior history of CHD, but with 3 additional cardiovascular risk factors Summary - Inclusion of CRP into a modified Framingham risk model in the whole cohort very modestly improved risk prediction Baseline CRP was not an indicator of the effect of atorvastatin on CV outcome Lowering LDL-C at 6 months but not CRP- correlated with a reduction in CV events CONCLUSION -The results do not support current proposals to measure CRP in the clinical setting to assign statins to individuals on the basis of an elevated CRP alone, or to monitor CRP levels as an indicator of the efficacy of statin treatment
USA Guidelines for hscrp testing SHAPE (Society for Heart Attack Prevention and Eradication) The new guidelines recommend CRP testing for all intermediate- and high-risk individuals and elevate the patient's risk category if CRP >3 mg/l. The ACCF/AHA writing committee finds that CRP generally adds little prognostic information to traditional risk factors and does not recommend CRP testing for high- or low-risk individuals: Class IIa (It is reasonable to perform the procedure) In men 50 years of age or older or women 60 years of age or older with LDL-C<130 mg/dl; not on lipid-lowering, HRT, or immunosuppressant therapy; without clinical CHD, diabetes, CKD, severe inflammatory conditions, or contraindications to statins, measurement of CRP can be useful in the selection of patients for statin therapy. Class IIb (Procedure may be considered) In asymptomatic intermediate-risk men aged <50 y or women >60 years of age Class III (No Benefit) In asymptomatic high-risk adults and in low-risk men <50 years of age or women <60 years of age
CRP for the Clinician Summary/Conclusions CRP is a marker of inflammation and a predictor of CV risk, however, it provides modest additional information over and above that afforded by conventional clinical markers Screening of the entire adult population is not recommended Measurement of hs-crp in high risk subjects is unlikely to influence patient evaluation or therapy In some patients at intermediate risk for CHD (10 to 20% at 10 years by FRS) hs-crp may help direct further evaluation and therapy for primary prevention. A high CRP might be treated as an additional risk factor when determining BP or LDL-cholesterol control targets