This prospective study enrolled 286 consecutive patients that were referred to the hospital between July 2001 and February 2002 for acute coronary syndromes. The diagnosis of AMI was established by the presence of chest pain lasting >20 minutes associated with electrocardiographic and enzymatic changes (Peak CK ≥ 300 U/l). Electrocardiographic changes included: an ST-segment elevation of at least 0.1 mV in two or more limb leads, an elevation of at least 2 mV in two or more contiguous precordial leads on the surface electrocardiogram, or a new-onset left bundle branch block. Unstable Angina was defined as a crescendo pattern of chest pain at rest with documented non-specific electrocardiographic changes including: a transient ST segment depression or elevation of at least 0.1 mV in at least 2 contiguous electrocardiographic leads or a T wave inversion. Patients with interfering non-cardiac diseases and malignancies were not included in the study. On admission, blood was drawn from all subjects under standardized conditions and plasma samples were stored at -80°C until analysis. All patients underwent urgent coronary angiography for percutaneous coronary intervention (PCI). The study protocol was approved by the institutional ethics committee and informed consent was obtained from all subjects.

To analyze the sources of circulating TF activity, an additional group of 36 consecutive patients with uAP and AMI were recruited between December 2005 and November 2006. A control group comprised 28 patients with stable angina (sAP) undergoing elective PCI.

Plasma TF activity was measured with an FXa generation assay (Actichrome TF, American Diagnostica; Detection treshold 2 pM, intra-assay variabilities <10%; inter-assay variabilities <10%). To inhibit endogenous TFPI TF activity was measured using an FXa generation assay in the absence and presence of 100 μg/ml rabbit anti-TFPI-1 IgG (kindly provided by Dr. W. Ruf) or nonimmune rabbit IgG (Sigma). TF activity in the presence of anti-TFPI mAbs represents the total TF activity in the plasma. TF activity associated with circulating MP was measured by first isolating MP by antibody capture and then analyzing FXa generation using a commercial available assay (Actichrome^®Microparticle activity #817, American diagnostic; Detection treshold ≤ 0,05 nM, intra-assay variabilities 3–8%; inter-assay variabilities 5–10%). Plasma TF antigen was measured by immunoassay (Imubind Tissue Factor ELISA kit, American Diagnostica; Detection treshold 1,4 pM, intra-assay variabilities 4,5%; inter-assay variabilities 7,5%).

Serum concentrations of creatine kinase, C-reactive protein (CRP), fibrinogen, and brain natriuretic protein (BNP) (Roche NT-pro-BNP sandwich electrochemiluminescent assay, Roche Diagnostics, Mannheim, Germany) were determined in routine clinical chemistry laboratory analyses.

For analysis of circulating MP, platelet free plasma (PFP) was immediately snap frozen in liquid nitrogen and stored at -80°C until use. To separate MP from PFP, aliquots of 250 μl were centrifuged for 60 min at 100,000 × g; then 225 μl of the supernatant was decanted and stored for further analysis. MP were resuspended in the remaining 25 μl. Flowcytometric analyses of MP suspensions were performed according to the manufacturer's instructions using annexin V Alexa Fluor^® 568 in an appropriate buffer (Boehringer Mannheim, Germany) and either specific monoclonal antibodies (mAb) or isotype-matched control antibodies. Antibodies were labeled with fluoresceinisothiocyonate (FITC), peridinin-chlorophyll-protein complex (PerCP), or Allophycocyanin (APC) and included: FITC-conjugated anti-TF antibody (American diagnostic), CD41 PerCP (Coulter), and CD 66 APC (BectonDickinson). Before FACS analysis was performed, a defined number of latex beads (LB30, Sigma) were added to the samples in order to calculate the absolute amount of MPs. Analysis was stopped when 5000 beads were counted.

All enrolled patients were contacted for follow-up with a mean of 3.26 years after the percutaneous coronary intervention. The primary outcome measures of the study were: cardiovascular mortality, which included a) fatal myocardial infarction and b) ventricular arrhythmias with heart failure as a consequence of MI, and all cause mortality.

Discrete variables were expressed as counts (%) and compared by chi-square or Fisher's exact test. Continuous variables were expressed as mean ± S.D. and compared by the Wilcoxon's rank sum test. The Youden-index was used to determine statistically reliable cut-off value of TF by maximizing the sum of sensitivity and specificity in reference to observed survival state. In this term, area under the ROC-curve (AUC) was reported to assess overall prognostic performance of plasma TF activity. Further, the corresponding 95% confidence interval of AUC was reported indicating statistically significant predictive capability if the critical value of 0.5 is not included. Spearman-Rho correlation coefficient was used to identify correlates of TF activity. Survival analysis was performed by applying the Kaplan-Meier method. Differences in survival were assessed with log rank test. Cox's proportional hazard model was used to assess the independent association of TF activity with mortality. All statistical analyses were performed using SPSS 15.0. P-values < 0.05 were considered statistically significant.

Of the 286 patients enrolled in the study, 112 presented with AMI and 174 were classified as uAP. During the follow up 16 cardiovascular deaths and 8 non-cardiovascular deaths were reported among patients with AMI; similarly, 15 cardiovascular deaths and 9 non-cardiovascular deaths were reported among patients with uAP. In AMI the majority of cardiovascular death occurred in the first year caused by cardiogenic shock or terminal heart failure, whereas in uAP only 3 deaths occurred during this time. The majority of cardiovascular death in uAP occurred after 1 year or later due to reinfarction, terminal heart failure and sudden cardiac death. According to a receiver operating characteristic (ROC) analysis patients were categorized into 2 groups: the high-TF activity group included patients with TF values >24 pmol/L, the low TF activity group included patients with TF values ≤ 24 pmol/L. ROC – analysis revealed a substantial prognostic capability of TF with an AUC of 0.66 (95%CI: 0.51 to 0.81). The clinical characteristics of each patient group and the incidence of cardiovascular death are shown in Table 1 and 2. As expected, patients with AMI that died from cardiovascular causes were older and had higher peak CK and NT-proBNP levels at baseline than those that survived (data not shown). Also, on admission they had a reduced left ventricular ejection fraction and presented with a higher Killip class. Patients with uAP that died had a higher prevalence of hypertension and preclinical resuscitation than those that survived. No other characteristics were different between patients that died and those that survived. The majority of AMI patients received abciximab, and the majority of uAP patients received only heparin.

In patients with AMI the TF activity was 24.9+2.78 pmol/L (mean+SEM) in survivors and 40.9+7.96 pmol/L in nonsurvivors (P = 0.024). Kaplan-Meier estimates of survival regarding TF activity in AMI above or below 24 pmol were 81.3% and 92.2% with an odds ratio of 3.02 (95% CI [1.05–8.79], P = 0.03; Figure 1). In uAP no differences in TF activity were observed in survivors and nonsurvivors (25.7 ± 2.14 pmol/L versus 25.0 ± 8.04 pmol/L; P = 0.59). The Cox proportional hazards model adjusting for correlates of mortality showed that plasma TF activity was a prognostic factor of mortality in patients with AMI (Table 3).

Among patients with AMI, TF activity correlated with parameters of myocardial necrosis (peak CK: R = 0.22, P = 0.025; peak CK-MB: R = 0.23; P = 0.019; CK on admission: R = 0.23, P = 0.037), Killip class (R = 0.22, P = 0.022), and incidence of cardiopulmonary resuscitation on admission (R = 0.20, P = 0.037). No association was found between TF activity and markers of inflammation. In contrast, among patients with uAP, TF activity was only correlated with parameters of inflammation (CRP: R = 0.17, P = 0.03; fibrinogen: R = 0.23, P = 0.005, and NT-proBNP: R = 0.16, P = 0.035) but not with parameters of myocardial necrosis, Killip class, or incidence of cardiopulmonary resuscitation.

To evaluate the functional significance and the underlying sources of circulating TF, we performed a second study comparing 36 patients with uAP and AMI and 28 control patients with sAP. In contrast to patients with ongoing ACS, almost no TF activity was detected in the circulation of sAP patients (0.0 ± 0.0 pmol vs 0.5 ± 0.16 pmol, P = 0.008, Table 4), even after abrogating the endogenous inhibition by TFPI-1 (0.98 ± 0.4 pmol vs 12.63 ± 1.82 pmol, P < 0.001).

Currently there are two known sources that contribute to circulating TF: alternatively spliced soluble TF (asTF) and MP bearing membrane-bound TF. In order to analyze both fractions we applied a differential centrifugation procedure, resulting in pellet and supernatant fractions. FACS analysis showed the pellet was enriched in MP, while no viable cell particles were found in the supernatant (Figure 2A, B, F, G). Nevertheless, both fractions contained comparable amounts of TF antigen (Figure 2C, H). We reasoned that the pellet fraction most likely represented circulating MPs bearing TF, and the supernatant fraction contained a soluble form of TF. The pellet fraction exhibited a clear TF activity, which further increased after inhibition of endogenous TFPI. However, the supernatant had only little TF activity, devoid of any regulation by TFPI (Figure 2D, I). No TF activity was detected in the supernatant using the assay for MP-bound TF activity (Figure 2E, J). In the pellet, TF activity measured by the chromogenic and the MP assays were comparable, indicating that MPs were a main component of plasma TF activity. In contrast, in the supernatant, soluble forms of TF contributed to circulating TF antigen, but possessed no TF activity and were not regulated by TFPI. Flow cytometric characterization of the MPs of patients with ACS revealed that circulating MPs stained double-positive for platelet (CD41) and granulocyte (CD66b) markers (Table 5). Among the MPs that were Annexin+TF+, the percentages that were also CD41+CD66b+ correlated with the basal TF activity. Thus, in the plasma of patients with ACS, the TF activity was a consequence of circulating MPs.