Author + information
- Received July 25, 2005
- Revision received October 6, 2005
- Accepted October 17, 2005
- Published online March 7, 2006.
- Mahmoud Suleiman, MD⁎,
- Rania Khatib⁎,
- Yoram Agmon, MD⁎,
- Riad Mahamid, MD⁎,
- Monther Boulos, MD†,
- Michael Kapeliovich, MD, PhD⁎,
- Yishai Levy, MD†,
- Rafael Beyar, MD, DSc⁎,
- Walter Markiewicz, MD⁎,
- Haim Hammerman, MD⁎ and
- Doron Aronson, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Doron Aronson, Department of Cardiology, Rambam Medical Center, PO Box 9602, Haifa 31096, Israel
Objectives We aimed to study the relationship between C-reactive-protein (CRP), obtained within 12 to 24 h of symptoms onset, and long-term risk of death and heart failure (HF) in survivors of acute myocardial infarction (MI).
Background A robust inflammatory response is an integral component of the response to tissue injury during MI. The magnitude of the early inflammatory response to ischemic injury might be an important determinant of long-term outcome.
Methods We prospectively studied 1,044 patients admitted with acute MI and discharged from hospital in stable condition.
Results During a median follow-up of 23 months (range, 6 to 42 months), 113 patients died and 112 developed HF. In a multivariable Cox regression model adjusting for clinical variables and predischarge ejection fraction, compared with patients in the first CRP quartile, the adjusted hazard ratios (HRs) for death progressively increased with higher quartiles of CRP (second quartile 1.4 [95% confidence interval (CI) 0.6 to 2.9]; third quartile 2.3 [95% CI 1.2 to 4.6]; fourth quartile 3.0 [95% CI 1.5 to 5.7]; for trend, p = 0.0002). Compared with patients in the first CRP quartile, the adjusted HRs for HF were: second quartile, 1.1 (95% CI 0.5 to 2.3); third quartile, 1.9 (95% CI 1.0 to 3.6); and fourth quartile, 2.1 (95% CI 1.2 to 3.9) (for trend, p = 0.005).
Conclusions C-reactive-protein is a marker of long-term development of HF and mortality in patients with acute MI and provides prognostic information beyond that provided by conventional risk factors and the degree of left ventricular systolic dysfunction.
A robust inflammatory response is an integral component of the response to tissue injury and plays a particularly active role after myocardial infarction (MI). Proinflammatory cytokines are elaborated soon after myocardial ischemic injury and mediate tissue repair and adaptation; however, there is evidence that an exaggerated inflammatory response might also promote tissue injury. Therefore, the degree of the inflammatory response might be an important determinant of outcome by leading to chronic cardiac dilatation, heart failure (HF), and death (1).
C-reactive protein (CRP) has emerged as a simple tool for detecting low-level systemic inflammation that portends coronary events in apparently healthy subjects (2,3) and in patients with a recent coronary event (4). In this setting, CRP levels might reflect inflammation in the vascular bed or an intensification of focal inflammatory processes that destabilize vulnerable plaques (5). By contrast, in the acute phase of MI, CRP levels are likely to be dominated by the inflammatory response to myocardial necrosis rather than by chronic vascular inflammation (6–8). The CRP levels in the early phase of MI might be a simple marker for the magnitude of the inflammatory response to myocardial necrosis, potentially providing prognostic information regarding the risk of death and HF (9,10).
Despite the pathophysiological link between inflammation and HF (11), previous studies that addressed the prognostic values of CRP after MI seldom included HF as an outcome, and most of the prognostic information was contributed by early in-hospital events (9,12,13); however, the end point of HF is important because of the contribution of MI survivors to the epidemic of HF (14,15) and because inflammation might antedate the development of HF in patients without prior MI (11). Therefore, the purpose of the present prospective study was to test the hypothesis that early determination of CRP, a potential marker of the intensity of inflammatory response to ischemic injury, might predict the long-term development of HF and death in patients who survived their index infarction.
All patients presenting to the intensive coronary care unit of the Rambam Medical Center with acute MI were eligible for entry into the study. The investigational review committee on human research approved the study protocol.
Myocardial infarction was diagnosed according to the Joint European Society of Cardiology/American College of Cardiology criteria (16). Exclusion criteria were known malignancy, known inflammatory disease, surgery, trauma, or MI in the previous month. Patients admitted at >24 h from symptoms onset were also excluded. Patients with ST-segment elevation of ≥0.1 mV in two contiguous electrocardiographic leads received thrombolytic therapy (tissue plasminogen activator or streptokinase) or underwent primary angioplasty.
Left ventricular ejection fraction (LVEF) was measured by echocardiography in 895 patients and by contrast left ventriculograms in the remaining 149 patients. Computer-assisted analysis of the traced left ventriculograms at end diastole and at end systole provided the LVEF.
Blood sampling procedure and CRP assay
Venous blood samples for CRP were obtained within 12 to 24 h of symptom onset. Blood samples were collected in citrate-treated tubes, centrifuged for at least 5 min, and stored at −70°C. High-sensitivity CRP was measured with latex-enhanced immunonephelometry on a Behring BN-ProSpec Nephelometer (Dade Behring, Marburg, Germany). The assay has a detection limit of 0.175 mg/l. The intra-assay coefficient of variation for CRP was 3.3% and the inter-assay coefficient of variation was 3.2%. The CRP levels were analyzed by laboratory personnel who were not involved in patient care. Test results were not available to the attending physicians.
Study end points
The primary end points of the study were: 1) all-cause mortality, and 2) development of HF, defined as readmission to hospital for the management of HF (defined by the presence of new symptoms of dyspnea or edema with one or more concurrent signs, including ventricular gallop rhythm, bilateral post-tussive rales in at least the lower third of the lung fields, elevated venous pressure, or pulmonary venous congestion on X-ray with interstitial or alveolar edema). The diagnosis of HF was confirmed with hospital records and discharge summaries. The secondary end point was the combined end point of death and HF. After hospital discharge, clinical end point information was acquired by reviewing the national death registry and by contacting each patient individually and independently reviewing the hospital course for major clinical events if the patient had been re-hospitalized.
Data are expressed as mean ± SD or median and interquartile range. Baseline characteristics of the groups were compared with analysis of variance (ANOVA) for continuous variables and by the chi-square statistic for noncontinuous variables. When continuous data was not normally distributed or had unequal variance, groups were compared with the nonparametric one-way ANOVA (Kruskal-Wallis test).
Event-free survival was estimated by the Kaplan–Meier method, and curves were compared with the log-rank test. Univariate and multivariate Cox proportional hazard analyses were performed to determine the relation between CRP and mortality. The following baseline clinical characteristics were considered in the multivariate procedure: age, gender, baseline creatinine, history of congestive HF, history of prior HF, history of diabetes, history of hypertension, smoking status, Killip class >I on admission, hypotension (systolic blood pressure <100 mm Hg) and tachycardia (heart rate >100 beats/min) on admission, anterior location of infarction, ST-segment elevation infarction, peak creatine kinase (CK), reperfusion therapy with either thrombolytics or primary angioplasty, and predischarge LVEF stratified as above or below 45%. Only variables with p < 0.2 in the univariate Cox regression analyses were used in the multivariate model.
Cox proportional hazards modeling was also used to determine the relationship between quartiles of CRP and admission for the treatment of HF. Known predictors of the development of HF in survivors of MI (15) were forced into the model (age, LVEF, baseline heart rate, Killip class at admission, history of hypertension and diabetes, previous infarction). In addition, other potential predictors were considered (gender, previous HF, anterior infarction, ST-segment elevation infarction, peak CK, and use of reperfusion therapy) if they demonstrated association with HF on univariate analysis (p < 0.2). Because recurrent MI increases the risk for HF (15) and is associated with greater degrees of HF, we performed two separate analyses for this end point: Model 1, including recurrent MI as a time-dependent covariate in the Cox model; and Model 2, in which only HF events occurring without a preceding recurrent MI were considered (i.e., HF that can be attributed to the index infarction).
We performed supplementary analyses for the combined end point of death and HF after dividing the study participants into 12 groups on the basis of CRP quartiles and three categories incorporating clinical evidence of HF during admission or reduced LV systolic function, as follows: 1) Killip class I on admission and preserved (≥45%) predischarge LVEF; 2) Killip class II to IV on admission or reduced (<45%) predischarge LVEF; and 3) Killip class II to IV on admission and reduced predischarge LVEF.
All models were also developed after excluding patients with previous history of HF (n = 48). These models were not appreciably different with regard to the relation between CRP and the primary end points. Therefore, only analyses based on the whole study population are presented in detail. Differences were considered statistically significant at the two-sided p < 0.05 level. All statistical analyses were performed with the SPSS statistical software (version 11.5, SPSS Inc., Chicago, Illinois).
Between July 2001 and June 2004, a total of 1,196 patients presenting with acute MI were enrolled. Of these, 1,044 survived their index event and were discharged from hospital in stable condition. Median CRP level of these patients was 13.6 mg/l (interquartile range 5.0 to 38.1 mg/l).
The clinical characteristics of patients according to the quartiles of CRP are shown in Table 1.Patients with CRP in the upper quartiles were older and had higher baseline creatinine and more history of previous HF; they presented with higher heart rates and higher Killip class and had lower EF. They were less likely to receive reperfusion therapy. At discharge, use of statins was less frequent in patients with elevated CRP.
The median duration of follow-up after hospital discharge was 23 months (range, 6 to 42 months). During the follow-up period 113 patients (10.8%) died. Median CRP levels were significantly higher among patients who died than among patients who survived (31.6 mg/l interquartile range [13.8 to 96.7] vs. 12.6 mg/l interquartile range [4.7 to 32.8], p < 0.0001). Kaplan-Meier analysis showed an increased probability of death during follow-up with increasing quartiles of CRP (Fig. 1).The Kaplan-Meier survival curves for CRP quartiles separated early, yielding a statistically significant difference in mortality at 30 days after hospital discharge (log-rank p = 0.005). The separation of the curves continued throughout the follow-up period (log-rank p < 0.0001), with mortality rates of 4.4% (12 patients), 6.1% (16 patients), 12.3% (32 patients), and 20.3% (53 patients) in the respective quartiles.
In a multivariable Cox proportional hazards model adjusting for other potential clinical predictors of mortality and for EF, increasing quartiles of CRP independently contributed to the prediction of post-discharge death (Table 2).Adjusting for LVEF as a continuous variable gave similar results. Compared with patients in the first CRP quartile, the adjusted HRs for death were 1.3 for patients in the second quartile (95% confidence interval [CI], 0.6 to 2.7; p = 0.7); 1.9 for patients in the third quartile (95% CI, 1.0 to 3.8; p = 0.05); and 2.6 for patients in the fourth quartile (95% CI, 1.4 to 5.1; p = 0.004). Additional adjustments for concomitant cardiovascular medications at discharge had no discernible effect on these estimates.
During follow-up 112 patients (10.7%) were admitted for the treatment of HF. Median CRP levels were significantly higher among patients who developed HF than among patients who did not develop HF (27.0 mg/l, interquartile range [13.3 to 75.2] vs. 12.7 mg/l interquartile range [4.8 to 33.8], p < 0.0001). Recurrent infarction occurred more frequently in patients who developed HF than in patients who did not develop HF (15.2% vs. 6.0%, p < 0.0001). The risk of subsequent death was significantly greater in patients with HF developing during follow-up compared with patients without HF (27.7% vs. 8.8%, p < 0.0001).
Kaplan-Meier analysis showed a graded increased probability for HF during follow-up with increasing quartiles of CRP (Fig. 2).Table 3displays the results of multivariate analyses examining the relationship between CRP levels and risk of HF. C-reactive protein added prognostic information above and beyond that provided by known clinical predictors of HF, pre-discharge EF, and recurrent infarction during follow-up (Table 3, Model 1). Other significant predictors of HF included age (HR 1.4 per 10 years; 95% CI 1.2 to 1.7, p < 0.0001), Killip class above I at admission (HR 2.0; 95% CI 1.3 to 3.0, p = 0.002), reduced LVEF (HR 2.4; 95% CI 1.5 to 3.7, p < 0.0001), peak CK (HR 1.3 per quartile, 95% CI 1.1 to 1.6, p = 0.003) and recurrent MI (HR 2.5; 95% CI 1.3 to 4.8, p = 0.008).
The relationship between CRP and HF was also examined in patients in whom development of HF could be attributed to the index infarction, by excluding HF events that occurred in patients with recurrent infarction (n = 17). The effect of increasing CRP was particularly striking after excluding these events, with an adjusted HR for HF of 2.8 in the upper CRP quartile (Table 3, Model 2). Adjusting for LVEF as a continuous variable resulted in an adjusted HR of 2.8 in patients in the fourth CRP quartile compared with patients in the first CRP quartile (95% CI 1.3 to 5.9; p = 0.007). No significant differences were found when the analyses in Table 3were repeated after exclusion of patients (n = 48) with prior HF.
C-reactive protein levels were not predictive of the development of recurrent MI (16, 21, 20, and 20 recurrent infarctions occurring in patients in the first, second, third, and fourth CRP quartile, respectively; p = 0.84).
Combined end point of death and HF
For the combined end point of death and HF, 23 (8.8%), 27 (10.3%), 58 (22.2%), and 86 (33.0%) events occurred in the first, second, third, and fourth CRP quartile, respectively (p < 0.0001). Compared with patients in the first CRP quartile, the adjusted HRs for the combined end point of death and HF were as follows: second quartile, 1.2 (95% CI, 0.6 to 2.1; p = 0.60); third quartile, 1.9 (95% CI, 1.1 to 3.2; p = 0.02); and fourth quartile, 2.5 (95% CI, 1.5 to 4.1; p < 0.0001).
We further estimated the HRs for the combined end point of death or HF after dividing the study participants into 12 groups on the basis of CRP quartiles and clinical evidence of HF during admission or reduced LV systolic function. As shown in Figure 3,after adjustment to other clinical risk factors, CRP added prognostic information with regard to the risk of death and HF even in the low-risk group of patients with no evidence of HF during admission and preserved predischarge EF (adjusted HR 2.7; 95% CI 1.2 to 6.4, p = 0.02 for patients in the fourth compared with the first CRP quartile). Similarly, CRP added prognostic information for post-discharge death of HF in the high-risk subgroup of patients with both clinical evidence of HF during admission and reduced LV systolic function (adjusted HR 2.8; 95% CI 1.1 to 7.4, p = 0.04 for patients in the fourth compared with the first CRP quartile).
The present prospective study demonstrates a strong, positive, graded association between CRP obtained within 12 to 24 h from symptoms onset and long-term development of HF and death in survivors of acute MI. The association between CRP and the development of HF was maintained after adjusting for the occurrence of recurrent MI during follow-up. The relationship between elevated CRP and adverse outcome remained robust even when assessed in apparently low-risk patients (preserved predischarge LVEF and Killip class I during admission) as well as in high-risk patients (reduced EF and clinical evidence for HF at admission).
CRP and outcomes in acute coronary syndromes
The mechanism for CRP elevation and its predictive value for different outcomes in patients with acute coronary syndromes depends critically on the clinical setting and time of measurement. The low-grade CRP elevation in the chronic phase of atherosclerosis is associated with future coronary events among apparently healthy individuals (2,3). C-reactive protein levels might further increase in patients with unstable angina, owing to widespread vascular inflammation (17), and are associated with recurrent clinical instability and recurrent infarctions (18); however, in unstable coronary disease the rise in CRP and other markers of inflammation such as interleukin-6 occurs almost exclusively in patients with evidence of myocardial necrosis (7,8). Thus myocardial necrosis seems to elicit a more pronounced activation of CRP that is superimposed on the low-grade, vascular-mediated inflammation (6,8). Because the extent of CRP elevation in patients with MI is dominated by the inflammatory response to myocardial injury rather than vascular inflammation, it has been associated with mortality but not with recurrent infarctions (9,12,19,20).
The finding that elevated CRP in the early phase of MI predicts long-term development of HF also supports the concept that at this time point CRP levels mainly reflect ongoing myocardial inflammation. These results are in keeping with previous studies showing that elevated peak CRP is related to early mechanical complications such as cardiac rupture, ventricular aneurysmal formation, and thrombus formation (21).
The acute phase response shown by CRP during the acute phase of MI is highly dynamic, with levels beginning to rise within a few hours after symptoms onset, peaking at 2 to 4 days (6,8,12). Importantly, the relationship between CRP elevation and HF or death is present only if CRP levels are obtained early in the course of MI (6–8). The acute inflammatory response associated with myocardial necrosis subsides within 4 to 6 weeks, with CRP levels returning to baseline values (8,12). At this time, elevated CRP levels again reflect the degree of vascular inflammation and are predictive of future coronary events (4).
CRP and HF
The development of late HF in patients after MI is particularly ominous because these patients have a several-fold increase in the risk of death when compared with other MI survivors (14,15,22). In a recent study, MI survivors who developed HF had a 10-fold risk of death compared with patients who did not develop HF (15). Our data indicates that inflammatory markers might be useful in identifying patients at high risk for HF despite apparent clinical stability. Late development of HF in MI survivors is probably related to several mechanisms, including progressive remodeling (23,24), recurrent infarction, and subclinical ischemia; however, CRP was predictive for HF events after excluding patients who developed HF in the context of recurrent infarction. Taken together, these findings suggest that CRP in the early phase of infarction might be indicative of an intense inflammatory state that contributes to progressive postinfarction ventricular remodeling. Importantly, CRP in not only a marker of inflammation but appears to actively participate in myocardial damage via local complement activation (25,26).
The mechanisms relating elevated CRP to future development of HF are unclear, although accumulating evidence suggests that inflammatory responses, primarily mediated by cytokines, play a key role in the process of healing and remodeling after MI. During infarction, cytokines are elaborated by immune, vascular, and interstitial tissue to regulate important biological processes of cell growth, migration, repair, and fibrosis (1).
In rodent models of MI, a robust up-regulation of intra-myocardial cytokines occurs within the first hours to 1 day after the infarct. The messenger ribonucleic acid expression of cytokines such as tumor necrosis factor-alpha, interleukin-6, and interleukin-1-beta increases dramatically in the infarct area (up to 50-fold). Cytokine expression is not confined to the infarct zone, because its expression is also up-regulated in cardiac myocytes in the noninfarcted myocardium (up to 15-fold) (1,27,28). On the basis of experimental studies, it has been postulated that the early inflammatory response might have long-term consequences in terms of ventricular function and remodeling (1).
Cytokines originating from the myocardium or infiltrating inflammatory cells (particularly interleukin-6, the chief stimulator of the production of CRP in the liver) contribute to the raised CRP concentrations observed in patients with acute MI (9,21). Thus, in patients with acute MI, CRP might be a simple marker for the intensity of the inflammatory response within the myocardium.
The inflammatory reaction to ischemic injury entails the elaboration of multiple inflammatory mediators (1,13,27,28). We measured only one global marker of inflammation but other markers might prove to be more useful for risk stratification. The inflammatory reaction to ischemic injury begins within several hours and peaks after several days. We measured CRP levels at one time point within 12 to 24 h from symptoms onset; however, CRP levels at other time points, particularly peak CRP levels, might better reflect the magnitude of the inflammatory response to myocardial necrosis.
Cardiac natriuretic peptides reflect ventricular impairment and the severity of hemodynamic decompensation in heart disease. Elevated plasma B-type natriuretic peptide (BNP) levels are associated with increased risk of adverse events in patient with acute coronary syndromes (29). Pertinent to the present study is the recent demonstration that BNP taken early after the onset of symptoms improves risk stratification for mortality and readmission to hospital for HF, beyond clinical risk factors and LVEF (30). Thus, the inclusion of both CRP and BNP in our analysis might have improved risk prediction, because these biomarkers have been shown to provide additive prognostic information in patients with acute coronary syndromes (31).
Finally, diagnosis of HF was defined as admission for the management of HF. Thus, milder forms of HF that could be managed on an outpatient basis were not considered.
The results of the present study indicate that CRP level obtained early in the acute phase of acute MI is a powerful independent marker for long-term mortality and the development of HF. Beyond the use of CRP for prognostic purposes, these results support the concept that the magnitude of the early inflammatory response to infarction affects long-term outcome.
- Abbreviations and Acronyms
- B-type natriuretic peptide
- confidence interval
- creatine kinase
- C-reactive protein
- ejection fraction
- heart failure
- hazard ratio
- left ventricle/ventricular
- myocardial infarction
- Received July 25, 2005.
- Revision received October 6, 2005.
- Accepted October 17, 2005.
- American College of Cardiology Foundation
- Nian M.,
- Lee P.,
- Khaper N.,
- Liu P.
- Morrow D.A.,
- Braunwald E.
- Cusack M.R.,
- Marber M.S.,
- Lambiase P.D.,
- Bucknall C.A.,
- Redwood S.R.
- Vasan R.S.,
- Sullivan L.M.,
- Roubenoff R.,
- et al.
- Anzai T.,
- Yoshikawa T.,
- Shiraki H.,
- et al.
- Valgimigli M.,
- Ceconi C.,
- Malagutti P.,
- et al.
- Lewis E.F.,
- Moye L.A.,
- Rouleau J.L.,
- et al.
- Alpert J.S.,
- Thygesen K.,
- Antman E.,
- Bassand J.P.
- James S.K.,
- Armstrong P.,
- Barnathan E.,
- et al.
- Toss H.,
- Lindahl B.,
- Siegbahn A.,
- Wallentin L.,
- FRISC Study Group
- Cohn J.N.,
- Ferrari R.,
- Sharpe N.
- Pfeffer M.A.,
- Braunwald E.
- Lagrand W.K.,
- Visser C.A.,
- Hermens W.T.,
- et al.
- Lagrand W.K.,
- Niessen H.W.,
- Wolbink G.J.,
- et al.
- Deten A.,
- Volz H.C.,
- Briest W.,
- Zimmer H.G.
- Irwin M.W.,
- Mak S.,
- Mann D.L.,
- et al.
- Richards A.M.,
- Nicholls M.G.,
- Espiner E.A.,
- et al.
- Sabatine M.S.,
- Morrow D.A.,
- de Lemos J.A.,
- et al.