Author + information
We read with great interest the article by James et al. (1)in a recent issue of the Journal.
We were puzzled by the finding that C-reactive protein (CRP) was related to 30-day mortality, but not to the occurrence of myocardial infarction (MI). Indeed, as shown by Table 3, not only was CRP unrelated to MI, but patients with a CRP value ≥1.84 mg/l had a paradoxically lower probability of MI than did patients with values <1.84 mg/l (odds ratio 0.76, 95% confidence interval 0.59 to 0.98, p = 0.03). We believe that these findings can be explained by considering the potential relation (not addressed in the report) between high CRP values and impaired left ventricular (LV) function. Indeed, patients in the fourth CRP quartile had a higher rate of presentation with heart failure than did patients included in the other quartiles (p < 0.001). Moreover, as shown in Table 2 of the study, a strict relation existed between median troponin T values and CRP (0.04, 0.08, 0.1, and 0.3 μg/l being, respectively, the troponin T values in the four CRP quartiles, p < 0.001). It is likely that high CRP values would reflect large infarcts with impairment of LV function, an important predictor of early death, not necessarily related to the occurrence of a subsequent MI.
Unfortunately, the investigators (1)did not include in the multivariable analysis any index of LV function; therefore, the independent prognostic predictivity of CRP remains questionable. Moreover, the researchers state that “the levels of troponin and CRP provide important, different and complementary prognostic information… . The combination of both markers allows the best prediction of mortality.” We do not share such enthusiastic comment. Table 4 in the James et al. (1)study (“Mortality at 30 Days in Relation to Quartiles of Troponin T and CRP”) shows that patients with the highest values of troponin T (>0.47 μg/l) and CRP (>9.62 mg/l) had a 9.1% death rate, which is not very different, on clinical grounds, from the 7.4% death rate of all patients included in the highest troponin T quartile (the clinical relevance of the 3.6% mortality in the first CRP quartile being affected by the small number of patients in that subgroup, n = 250). In contrast, in the lowest troponin T quartile (<0.01 μg/l), the death rate was 1.4% in patients with the highest CRP values (>9.62 mg/l), not dissimilar from the 1.1% death rate of all patients with <0.01 μg/l troponin levels. From these data it seems that CRP does not provide any additional important prognostic information to the simple knowledge of troponin T values.
- American College of Cardiology Foundation