C-Reactive Protein and ST-Segment Elevation Myocardial Infarction Discordance

• high-sensitivity C-reactive protein
• myocardial infarction
• ST-segment elevation myocardial infarction

In this issue of the Journal, Cristell et al. (1) performed a novel study to test the prevalence of elevated circulating high-sensitivity C-reactive protein (CRP) immediately prior to ST-segment elevation myocardial infarction (STEMI). This was accomplished in 887 patients with STEMI as well as 887 control patients from urban areas in 3 regions: Italy, Scotland, and China.

The investigators indicate that the diagnosis of infarction was documented unequivocally in all by the presence of ST-segment elevation on presentation and by subsequent peak creatine phosphokinase elevation. Many STEMI patients successfully underwent reperfusion. Blood samples were obtained up to 6 h from symptom onset prior to reperfusion. The methods of collection of blood seemed appropriate, and all samples were shipped in dry ice to a core laboratory in Milan, Italy.

Symptom time was the least in Scottish patients, and troponin I measurements were also less in the Scottish patients than in the Italian and Chinese patients.

Most of the patients in this study had not received any antiangina, antilipid, or antiplatelet therapy prior to entry into the study.

The investigators restricted inclusion criteria in order to try to select a phenotypically homogeneous group. However, this was not the case, because results indicate a large scatter of CRP values among the patients studied. The overlap of CRP values with those of control subjects was great and was not significantly related to the risk factor levels examined. Median CRP values were significantly higher in patients than in control subjects, but the overlap was such that 37% of control subjects had values greater than 2 mg/l, and there was a wide overlap of values as shown in the Cristell et al. Figure 2 (1) for both high-sensitivity CRP as well as interleukin-6.

The investigators make the point that their observations are indicative of pre-infarction levels for CRP. Because others have reported (2) that CRP levels begin to increase about 6 h after infarction onset, it seems to me that this would be about the same time as other cardiac markers such as creatine kinase-myocardial band and troponin I, both of which would rise as well.

The investigators found no difference in the average CRP levels in patients sampled at <2 h, 2 to 4 h, and 4 to 6 h from the onset of symptoms. In this regard, it is important to point out that the timing of symptom onset in the human may not necessarily be the same as the timing of the infarction onset. The first indication for a STEMI comes generally with the electrocardiogram (ECG) taken in the emergency department or in the emergency vehicle and not at the time of symptom onset.

Perhaps one can identify patients whose myocardial infarction occurred recently: on or just prior to admission versus those who infarcted somewhat earlier by assessing the troponin I level. If the troponin I was abnormal on admission, the infarction probably would have occurred at the time of or near the time of the onset of symptoms, and if the troponin I was normal on admission, it could have occurred at a time closer to the initial ECG with ST-segment elevation. The Cristell et al. Table 2 (1) compares symptom time in minutes to troponin I in ng/ml and peak creatine phosphokinase units/l. Although these are average values, they do not reflect the initial troponin on admission to hospital or entry into the study, nor do they reflect initial creatine phosphokinase values. It might be instructive to have a scatter plot as in Figure 2 in Cristell et al. (1), in which all the individual data points for high-sensitivity CRP taken in the various countries are plotted versus the troponin I measured at the time of ECG diagnosis of STEMI.

In my opinion, troponin I defines the timing of the acute myocardial infarction. For example, if it is normal, the STEMI may have been very recent, whereas if the troponin I was elevated at the time of the abnormal ECG, the infarct was not recent. What is not known, in my view, is what the relationship is of the CRP under these circumstances. For example, was high-sensitivity CRP <2 seen more often in patients with normal troponin I on admission compared to high-sensitivity CRP >2?

Average CRP levels were similar in patients with average troponin I <0.2 and >0.2. Although 52% of patients had an initial admission troponin I of <0.2 ng/ml, 48% had elevation of the troponin I to an average of 2.09 ng/ml. CRP levels were similar in patients with troponin I <0.2 and >0.2. Because of the wide scatter of the data, measurements of CRP prior to STEMI do not adhere to the concept of “one size fits all.” This was not a homogeneous population. Although statistical significance was found in this study, and one can prognosticate in large populations, this may not have any clinical relevance for the individual because of the overlap of data.

The investigators indicate that at present we have no plausible explanation for the finding of high-sensitivity CRP <2 mg/l in 41% of STEMI cases for all 3 ethnic groups considered. Is it possible that the myocardial infarction just occurred at the time of arrival of the patient in the emergency department, rather than at the time of the onset of symptoms? In other words, in the emergency department, ECG changes of ST-segment elevation do not mean myocardial infarction occurred at that time.

My take on this paper is that each patient is an individual and thus clinical trials cannot be applied to every patient because clinical trials provide average results and, as the investigators point out, “one size does not fit all.” As a clinician, it is obvious to me that patient-specific therapy still requires clinical judgment, and clinical judgment is based on not just 1 factor, but on several factors, making it a composite of clinical trial results, experience, intuition, and common sense.

To summarize, in the individual patient, prognostication is an educated guess, nothing more. Because prognostic information on medical conditions is based on population studies, for any given condition, a certain percentage will have a major adverse cardiac event and others will be free of major adverse cardiac events. Thus, when it comes to the individual patient, there are no percentages other than 0 or 100.

Clinical trials, epidemiologic studies, and registries are good but not perfect. They do not explain all the clinical outcomes (i.e., the outliers), and thus there is a great deal to learn from these outliers and unusual clinical cases.

The critical question that needs to be answered in the future is: “What percentage of patients develop an acute myocardial infarction in spite of normal high-sensitivity CRP levels?” These investigators are planning to take a very hard look at the “outliers,” that is, those control subjects who had elevated high-sensitivity CRP and those STEMI patients who had normal high-sensitivity CRP.

Future studies will involve biologic and genetic characterization of these 2 extreme groups. From what I gather of this investigation, the investigators will eventually focus on the reasons for the outcomes in the patients who are outliers—a difficult challenge, and I wish them well because I believe these types of observations will advance our understanding of prognosis of the individual. Overall, the authors are to be congratulated for taking a different look at data collection while making the point that average data reported in any clinical trial do not represent many of the outliers.

• American College of Cardiology Foundation