Published in the April 1, 2004 issue of the New England Journal of Medicine and previously reported by heartwire, the results of the study were immediately disputed by experts who accused the investigators of "stirring up a controversy" where none should exist. Dr Paul Ridker (Brigham and Women's Hospital, Boston MA) told heartwire at the time that Danesh et al's interpretation of the data would set "cardiovascular prevention backward at the very moment in time when we finally have hard evidence that we can do a better job."

In their controversial study, Danesh and colleagues analyzed data from the Reykjavik Study, a large prospective study involving 2400 patients who had a nonfatal MI or who died from CHD, and found that after adjusting for risk factors such as smoking status, blood pressure, body-mass index, and total cholesterol levels, patients with a CRP value in the top third (cutoff value, 2.0 mg/L) had a relative risk of CHD of 1.45, as compared with patients whose values were in the bottom third. The investigators pointed out that this risk was much lower than odds ratios from previous studies (which have given estimates of around 2.0).

An updated meta-analysis of 21 studies, which added 5115 cases of CHD from a further 12 studies compared with the last meta-analysis, showed a result similar to their main study, giving an odds ratio of about 1.5 for CRP. In addition, after multivariate analysis, the authors reported that the predictive value of CRP measurement adds relatively little to that provided by assessments of traditional risk factors.

In the July 15, 2004 issue of the Journal, Ridker and colleagues, in addition to several other high-profile experts, continue to dispute the investigators' interpretation.2 Ridker, Drs Wolfgang Koenig (University of Ulm Medical Center, Ulm, Germany), and Valentin Fuster (Mount Sinai Cardiovascular Institute, New York, NY) write that the 10-year-risk estimate reported by Danesh et al for CRPwith an adjusted odds ratio of 1.84is fully in line with data from the Women's Health Study, the MONICA study, and the ARIC study, further supporting the role of CRP as a predictor of CHD risk.

We believe, however, that, if anything, there should be a broadening, rather than a curtailing, of the use of C-reactive protein in screening selected populations.

They also stress that the primary findingan odds ratio of 1.5is the same as the risk associated with hypertension and is statistically similar to that associated with smoking. According to this analysis, if CRP adds little clinical value for risk prediction, the investigators must believe hypertension and smoking are also unimportant predictors, argue Ridker and colleagues.

A second letter, written by Drs Peter Libby (Brigham and Women's Hospital, Boston, MA), Eugene Braunwald (Brigham and Women's Hospital), and James T Willerson (Texas Heart Institute, Houston), takes issue with the cutoff point used by the investigators.3 By using a cutoff value of 2.0 mg/L, rather than 3.0 mg/L, as recommended by the American Heart Association (AHA) and the Centers for Disease Control and Prevention (CDC), Danesh et al may have underestimated the risk associated with CRP, they write.

The group also contends that adjusting for diabetes, triglycerides, obesity, pulmonary function, social status, and other inflammatory markersas the investigators didreduces the predictive value of CRP, as does including study end points that failed to meet confirmation criteria for CV events.

In response, Danesh, Drs Mark Pepys (Royal Free and University College Medical School, London, UK), and Vilmundur Gudnason (Icelandic Heart Association, Kopavogur, Iceland) contend the main analysis was adjusted appropriately for possible confounders and that further adjustments for socioeconomic analysis and other inflammatory markers were included only in subsidiary analysis.4 They add that 92% of the cases in the study were confirmed nonfatal MI or death from coronary causes.

Still, Libby and colleagues argue that with 22 prospective studies supporting CRP as a predictor of risk in CHD, including the Reykjavik Study, the evidence continues to support its use in clinical practice.

"Given these findings, we concur that recommendations regarding the use of C-reactive protein may need to be reviewed," write Libby et al. "We believe, however, that, if anything, there should be a broadening, rather than a curtailing, of the use of C-reactive protein in screening selected populations."

In addition to taking issue with the CRP cutoff values, Drs Joanne Foody (Yale University School of Medicine, New Haven, CT), Antonio Gotto (Weill Medical College of Cornell University, New York, NY), and Nanette Wenger (Emory University School of Medicine, Atlanta, GA) argue that even the authors' own conservative estimate of the increment in risk prediction50%translates into a meaningful risk assessment at the population level.5

"Given the high societal costs of cardiovascular disease and the current suboptimal screening paradigms, high-sensitivity C-reactive protein is an inexpensive, valid tool to predict the risk of this disease," they write. "We strongly support the AHA-CDC recommendations with respect to C-reactive protein and believe that high-sensitivity C-reactive protein should be more fully incorporated into clinical practice, rather than restricted, as Danesh et al suggest."

In a fourth and final letter, Drs Robert J Glynn and Nancy R Cook (Harvard Medical School, Boston, MA) state that a more "statistically principled consideration of the case-control data would lead to the conclusion that CRP is an important predictor of risk." Like Ridker, they also note the odds ratio for the 10-year riskthe basis of most risk models, including the Framingham Heart Studyis in line with previous investigations.6

In response, Danesh and colleagues reiterate their findings, writing that the current data indicate that established coronary risk factors are stronger predictors than CRP values. In their study, the odds ratio for elevated total cholesterol and smoking were higherwith relative risks of 2.4 and 1.9, respectivelythan CRP, and they contend that CRP added only marginally to the predictive value of established risk factors.

With regard to the CRP cutoff values, they write that the use of any particular cutoff point, such as 3.0 mg/L, consistently applied among studies, would also yield similar contrasts with previous estimates, provided the shape of this relationship between CRP and coronary risk were approximately log-linear. To characterize this relationship and investigate the use of CRP for risk prediction in specific subgroups will require pooling individual records from prospective studies. This work is now in progress, they add.