Circulating Endothelial Cell Count as a Diagnostic Marker for Non–ST-Elevation Acute Coronary Syndromes
Background— Shedding of endothelial cells from damaged endothelium into the blood occurs in a variety of vascular disorders. The purpose of this study was to evaluate the utility of circulating endothelial cell (CEC) count as a diagnostic marker of non–ST-elevation acute coronary syndromes (ACSs).
Methods and Results— CEC counts were determined immediately (H0), 4 hours (H4), and 8 hours (H8) after admission in 60 patients with documented non–ST-elevation ACS and 40 control patients with no evidence of coronary artery disease. A total of 32 patients in the ACS group had elevated CEC counts (>3 cells/mL) in relation to early admission and single-episode chest pain. Patients from the control group had normal CEC counts. The interval between the chest pain episode and elevation was significantly shorter for CEC than troponin I. No correlation was found between the 2 markers. Interestingly, a subgroup of ACS patients with initially normal troponin I levels had high CEC counts, thus allowing early diagnosis in 30% more cases. At H0, the mean area under the receiver operating characteristic curve was significantly higher with the CEC count than with the troponin I level. At H4 and H8, the combined use of CEC and troponin was significantly better as a marker of ACS than CEC alone or troponin I alone.
Conclusions— This study demonstrates that CEC count can be used as an early, specific, independent diagnostic marker for non–ST-elevation ACS. A combined strategy using CEC count and troponin I level could provide an effective diagnostic tool.
Received September 18, 2003; de novo received February 24, 2004; accepted May 20, 2004.
Assessment of acute chest pain is a frequent, time-consuming clinical challenge. Approximately 30% of cases involve acute coronary syndromes (ACSs).1 Early diagnosis and appropriate treatment of patients are pivotal to avoid serious complications and death. On the basis of the ECG and clinical data, many non–ST-elevation ACSs are missed.2,3 An estimated 2% to 10% of ACS patients presenting in emergency rooms are mistakenly discharged,4 with an increased risk of progression to ST-elevation myocardial infarction or death.4
Various laboratory markers of myocardial cell necrosis have been proposed to assist diagnostic and therapeutic decision making. These biomarkers include creatine phosphokinase of muscle band (CPK-MB) and troponins for myocardial necrosis,2,5 brain natriuretic peptide for left ventricular overload,6 C-reactive protein (CRP) for inflammation,7 and pregnancy-associated plasma protein8 (PAPP-A), a protein present in atherosclerotic plaques; however, these markers have several shortcomings. The diagnostic specificity of brain natriuretic peptide and CRP is low. Troponin levels remain low in many cases of non–ST-elevation ACS, and initial findings may be negative in some high-risk patients depending on the time of sampling, assay sensitivity, marker release, and clearance kinetics.9 These problems can delay proper diagnosis and care. The availability of a specific and early marker of non–ST-elevation ACS, the level of which becomes elevated before or in the absence of an elevation of troponin, might facilitate diagnosis and improve therapeutic decision making.
ACSs usually result in coronary plaque rupture or erosion with endothelial disruption. Measurement of circulating endothelial cells (CECs) shed from damaged endothelium might provide a simple, noninvasive method to improve current diagnostic strategies. Whereas CEC counts do not exceed 3 cells/mL in normal individuals,10 elevated counts have been documented in a variety of vascular disorders.11–15 We previously reported that CEC levels were high in patients with ACS and normal in subjects with effort angina or noncoronary chest pain.16 CECs from ACS patients were mature endothelial cells of macrovascular origin; however, to the best of our knowledge, the diagnostic value of the CEC count in ACS has not been evaluated.
We hypothesized that elevation of CECs may identify a subgroup of ACS patients with an initially normal troponin level. The aim of the present study was to enumerate CECs in a population of patients with non–ST-elevation ACS associated with an identifiable culprit lesion and to evaluate the contribution of CEC count to the current diagnostic strategies.
CEC counts were measured in 2 age-matched groups of patients admitted to the coronary care unit of the Timone Hospital in Marseilles, France. The control population included 40 non-ACS patients admitted for workup before valve surgery. None of them presented clinical or angiographic evidence of significant coronary atherosclerosis, and we made certain that they presented CEC levels comparable to healthy subjects.
The ACS population included 60 patients with non–ST-elevation ACS. The preliminary inclusion criterion was resting chest pain of suspected ischemic origin within 6 hours before admission (Braunwald class IIIB). Myocardial ischemia was confirmed by ECG changes consisting of ST-segment depression of at least 0.1mV, T-wave inversion of at least 0.3 mV, or transient (<30 minutes) ST-segment elevation of at least 0.1 mV in at least 2 contiguous leads. Patients with prolonged (>30 minutes) chest pain with persistent ST-segment elevation, secondary angina, left bundle-block branch, a pacemaker, or documented myocardial infarction within the previous 30 days and those with serious systemic disease, sepsis, serum creatinine level >2.5 mg/dL, or pregnancy were excluded. All patients underwent antithrombotic and anti-ischemic treatment according to current guidelines17 and had coronary angiography within 48 hours of admission. Patients in whom no culprit lesion was identified were excluded.
Local ethics committee approval was obtained, and all patients gave informed consent.
CEC Counts and Plasma Protein Assays
Blood samples were collected from an antecubital vein into EDTA for determination of CEC counts and into sodium citrate (0.129 mol/L) for measurement of troponin I, high-sensitivity (hs)-CRP, and soluble adhesion molecule levels. Determination of CEC counts or troponin I levels was performed immediately (H0) and 4 (H4), and 8 (H8) hours after admission, but only on admission (H0) for soluble adhesion molecules and hs-CRP.
CEC counts were performed by an independent operator unaware of clinical features and coronary angiogram. To avoid contamination by endothelial cells from the punctured vessel wall, the first 2 mL of blood drawn was discarded. Immunocapture of CECs from whole blood was performed at 4°C according to a method previously developed in our laboratory11 with magnetic beads (Dynabeads M-450, Dynal) coated with S-Endo 1 (Biocytex), a monoclonal antibody raised against the endothelial CD146 antigen. To avoid nonspecific binding of leukocytes to CD146-coated beads, the cell suspensions were flushed vigorously through the pipette tip during the washing steps and then suspended in acridine orange (10 μg/mL in PBS; Sigma) for counting under an optical fluorescence microscope (λexc=490 nm). CECs were identified according to size, number of beads bound, and morphological criteria, ie, >10 beads bound to 20- to 50-μm cells, or cells with <10 beads but with a well-preserved and recognizable morphology (clear nucleus in a well-delineated cytoplasm and a size concordant with endothelial cells). For aggregates, the number of cells was determined from the number of nuclei or from the number of spherical rosette features detected in the aggregate. The endothelial nature of isolated cells was confirmed by the expression of von Willebrand factor and lack of expression of the leukocyte antigen CD45.
Troponin I was measured with a sandwich ELISA immunoassay (Dade Behring OPUS troponin I). Patients with troponin I levels higher than 0.4 ng/L were considered to be troponin positive. The coefficient of variation was <5%.
Hs-CRP levels were measured by a high-sensitivity, latex-enhanced immunonephelometric assay (Dade Behring BN II analyzer). Levels of soluble intracellular adhesion molecule-1 (sICAM-1), soluble vascular cell adhesion molecule-1 (sVCAM-1), soluble E-selectin, and soluble P-selectin were analyzed with a sandwich enzyme-linked immunosorbent method (R&D System Europe). Intra-assay coefficients of variation for the ELISAs were <5%, and interassay coefficients of variation were <10%. Normal values found in our laboratory on a population of healthy subjects were, respectively, 228±72, 413±130, 36±17, and 43±18 ng/mL for sICAM-1, sVCAM-1, soluble E-selectin, and soluble P-selectin levels.
Demographic and clinical characteristics of the study and control populations were compared with the χ2 test with Yates correction, Student t test, and ANOVA whenever possible. The Fischer exact test was used in some cases.
Because results showed nongaussian distribution, data were summarized with medians and interquartile range (IQR). Median comparison was performed with the Kruskal-Wallis test followed by a Mann-Whitney test in case of significant differences. Correlation between results was evaluated with the Spearman rank correlation coefficient. Sensitivity and specificity were evaluated with receiver operating characteristic curves. The area under each curve was determined and comparisons were performed by a nonparametric approach.
CEC Levels in Study and Control Populations
Sixty patients with non–ST-elevation ACS and 40 control patients were prospectively included. The highest of the 3 counts measured at H0, H4, and H8 was used (Figure 1). Counts in all control patients were normal, ie, <3 cells/mL. In ACS patients, 28 cases had normal CEC levels (CEC−; median 1 cell/mL), and 32 cases exhibited high CEC counts (CEC+; median 25.5 cells/mL, IQR 8 to 35.3 cells/mL).
Analysis of changes in CEC counts at the 3 times (H0, H4, and H8) in the CEC+ subgroup of ACS patients indicated that the highest counts were measured at H0 (median 8.5 cells/mL, IQR 1.5 to 25.5 cells/mL) and then decreased with time (median 6.5 cells/mL, IQR 3.5 to 27.5 cells/mL at H4; median 2 cells/mL, IQR 0 to 10 cells/mL at H8).
Comparison of ACS Patients With Normal Versus High CEC Counts
In the ACS patient population, the CEC− and CEC+ subgroups were compared with regard to demographic, clinical, ECG, laboratory, and angiographic findings and medical treatment (Table). The only statistically significant differences were the interval between admission and the last episode of chest pain (3.1±1.8 hours for CEC+ versus 5.0±1.2 hours for CEC−, P<0.0001), interval between admission and first episode of chest pain in patients with recurrent angina (4.9±4.6 hours for CEC+ versus 45.4±37.4 hours for CEC−, P<0.0001), and number of episodes of chest pain (1 episode in 69% of CEC+ versus 18% of CEC−). All other parameters were comparable, which suggests that CEC count could be an early marker of non–ST-elevation ACS.
Comparison of CEC Count and Troponin I Level
The interval between chest pain episodes and elevation of CEC count or troponin I level was compared. This interval was significantly shorter for CEC count in relation to the first episode (median 5.5 hours, IQR 2.5 to 6 hours versus median 10.5 hours, IQR 6 to 48 hours; P<0.0001). Similarly, this interval was significantly shorter for CEC count in relation to the last episode (median 4 hours, IQR 2 to 6 hours versus median 6 hours, IQR 5 to 8 hours; P<0.0001). CEC count and troponin I level varied independently, as shown by the lack of significant correlation at all 3 sample times (Spearman coefficients: 0.16 [P=0.22] at H0, 0.03 [P=0.80] at H4, and 0.11 [P=0.38] at H8).
The sensitivity of CEC counts for the diagnosis of ACS was lower than that of troponin I level: 53.3% versus 61.7%, respectively. The specificity of both was 100%. In accordance with these high-specificity and limited-sensitivity values, the positive and negative predictive values of CEC count were 100% and 58.8%, respectively, and those of troponin I level were 100% and 63.5%, respectively.
Combined Use of CEC Count and Troponin I Level for Diagnosis of ACS
To evaluate the combined use of CEC count and troponin I level, patients were classified into 4 groups, ie, CEC−/troponin+, CEC+/troponin+, CEC+/troponin−, and CEC−/troponin− at each sample time. Interestingly, a subgroup of patients with non–ST-elevation ACS had positive CEC counts and initially negative troponin tests (Figure 2). From a practical standpoint, CEC count led to a 30%, 20%, and 8% increase in the number of patients diagnosed at H0, H4, and H8, respectively.
The diagnostic value of combined use of CEC count and troponin I level was confirmed by comparing the mean area under the receiver operating characteristic curves for CEC count alone, troponin I level alone, and combined CEC count and troponin I level at each sample time (Figure 3). At H0, the area under the curve was significantly higher for CEC count alone than for troponin I level alone (0.7879 versus 0.6750, P=0.0241). The combined use of CEC count and troponin I level was significantly better than troponin I level alone (0.8250 versus 0.6750, P<0.0001) but not than CEC alone (0.8250 versus 0.7879, P=0.3265). At H4, the areas under the curve for CEC count alone and troponin I level alone were similar (0.7667, P=1.000), but the area under the curve for combined CEC count and troponin I level remained significantly higher than those of CEC count alone (0.8667 versus 0.7667, P=0.0105) or troponin I level alone (0.8667 versus 0.7667, P=0.0001). In contrast, at H8, the area under the curve was significantly lower for CEC count alone than troponin I level alone (0.6729 versus 0.8083; P=0.0038), but the area under the curve for combined CEC count and troponin level remained higher than troponin I level alone (0.8500 versus 0.8083, P=0.0206). These results indicate that CEC count alone (H0), or combined with troponin I level (H4, H8), improves the diagnostic accuracy of ACS from H0 to H8.
The present study shows that CEC count is a novel diagnostic marker of non–ST-elevation ACS. Compared with troponin level, CEC count was an earlier and independent diagnostic marker. The combined use of CEC count and troponin level led to a 30% increase in the number of patients diagnosed within the first hours after admission.
From a practical standpoint, determination of the CEC count is noninvasive and simple to perform. Results can be obtained in less than 1 hour. The limited sensitivity of the CEC count is probably linked to the clearance time. The time-course distribution of CEC levels indicated that the highest CEC counts were detected on admission and decreased with time. Indeed, from observations of CEC circulation kinetics after coronary angioplasty,11 the CEC count remains elevated for 4 to 6 hours after cell detachment. Normal values obtained in some patients may be related to timing or the effect of multiple episodes and do not exclude the possibility of initially positive CEC counts. The window for detection of elevated CECs is limited to the first hours after the initial ischemic episode. In the present study, 70% of patients with elevated CECs had only 1 episode of chest pain, and the mean interval between the episode and blood sampling was 3.1 hours. The period of CEC detectability coincides with the clinical “twilight zone” in which the attending physician has no clinical findings on which to base his decision making, because the ECG has returned to normal and standard biomarkers are still negative.
To improve the limited sensitivity of the CEC count, we combined it with the troponin level. The additional information provided by this multimarker approach has already been demonstrated in patients with chest pain through myoglobin combined with troponin or brain natriuretic peptide combined with CRP.9,18 In the present study, inclusion of CEC counts in a multimarker strategy appeared to result in improved sensitivity and earlier detection, extending the scope of laboratory markers from the consequences of ischemia to a marker of initial endothelial damage. However, a large prospective study that includes patients presenting in the emergency department with recent resting chest pain of suspected ischemic origin will be needed to confirm the diagnostic importance of the CEC counts.
Given current concepts on the pathophysiology of ACS, an elevated CEC count may be considered as a marker of endothelial injury associated with plaque rupture and/or erosion.19 The likelihood that the CECs originate from plaque lesions remains speculative but is supported by analysis of atherosclerotic plaques in coronary arteries.19,20 Once plaque formation is established, endothelial cells are lost in focal areas, thus exposing the subendothelial matrix and allowing adherence of platelets. Histological observations suggest that endothelial denudation is not an initial event in atherogenesis but develops later, when lesions begin to destabilize as a prelude to coronary thrombosis. In addition, demonstration that the CECs did not express CD36 is consistent with their macrovascular origin, because CD36 is associated with the microvessels.21 The present data indicate that levels of the soluble adhesion molecules sICAM-1, sVCAM-1, and sP-selectin, as well as the endothelium-specific sE-selectin, were not correlated with high CEC count. This lack of correlation is consistent with the fact that soluble adhesion molecules reflect systemic activity, whereas the CEC count reflects local injury.
At present, the mechanisms of EC detachment remain unclear. However, the inflammatory process at the site of plaque rupture or erosion probably plays a key role.19,22 In particular, inflammatory cytokines and cell interactions trigger endothelial cell apoptosis and production of metalloproteinases that break down the attachment between endothelial cells and underlying intima.23–25
The findings of the present study show that CEC count is an early, specific, independent diagnostic marker for non–ST-elevation ACS. A multimarker strategy combining CEC count and troponin I level may not only assist diagnostic and therapeutic decision making but may also improve our understanding of the pathophysiology underlying ACS.
The authors are grateful to the Biocytex Company for providing SEndo-1 antibody and to M. Revest and A. Boyer for technical assistance. They also thank the staff of the cardiology unit for its cooperation in this work and V. Gurewich for careful reading of the manuscript.
Erhardt L, Herlitz J, Bossaert L, et al. Task force on the management of chest pain. Eur Heart J. 2002; 23: 1153–1176.
Ridker PM. Clinical application of C-reactive protein for cardiovascular disease detection and prevention. Circulation. 2003; 107: 363–369.
Newby LK, Storrow AB, Gibler WB, et al. Bedside multimarker testing for risk stratification in chest pain units the chest pain evaluation by creatine kinase-MB, myoglobin, and troponin I (CHECKMATE) study. Circulation. 2001; 103: 1832–1837.
George F, Brouqui P, Boffa MC, et al. Demonstration of Rickettsia conorii–induced endothelial injury in vivo by measuring circulating endothelial cells, thrombomodulin, and von Willebrand factor in patients with Mediterranean spotted fever. Blood. 1993; 82: 2109–2116.
Mutin M, Canavy I, Blann A, et al. Direct evidence of endothelial injury in acute myocardial infarction and unstable angina by demonstration of circulating endothelial cells. Blood. 1999; 93: 2951–2958.
Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA guidelines for the management of patients with unstable angina and non–ST-segment elevation myocardial infarction: executive summary and recommendations: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on the Management of Patients With Unstable Angina). Circulation. 2000; 102: 1193–1209.
Sabatine MS, Morrow DA, De Lemos JA, et al. Multimarker approach to risk stratification in non–ST-elevation acute coronary syndromes: simultaneous assessment of troponin I, C-reactive protein, and B-type natriuretic peptide. Circulation. 2002; 105: 1760–1763.
Swerlick AR, Lee KH, Wick TM, et al. Human dermal microvascular endothelial but not human umbilical vein endothelial cells express CD36 in vivo and in vitro. J Immunol. 1992; 148: 78–83.
Van der Wal AC, Becker AE, Van der Loos CM, et al. Site of intimal rupture or erosion of thrombosed coronary atherosclerotic plaques is characterized by an inflammatory process irrespective of the dominant plaque morphology. Circulation. 1994; 89: 36–44.
Slowik MR, Min W, Ardito T, et al. Evidence that tumor necrosis factor triggers apoptosis in human endothelial cells by interleukin-1–converting enzyme-like protease-dependent and independent pathways. Lab Invest. 1997; 7: 257–267.
Rajavashisth TB, Liao JK, Galis ZS, et al. Inflammatory cytokines and oxidized low density lipoproteins increase endothelial cell expression of membrane type 1–matrix metalloproteinase. J Biol Chem. 1999; 274: 11924–11929.