Haemostatic markers and risk stratification

Clinical decision making requires appropriate tools to help in the diagnosis and to measure future risk in patients. Risk estimations of a patient having cardiovascular disease or of developing complications, such as myocardial infarction, sudden death or stroke, are necessary as they will determine which type of intervention to perform. The need for clinically useful markers, to help in the evaluation of risk, is, therefore, one of the primary targets of research. Biochemical parameters have been, and still are, sought to cover what epidemiological research has signalled as risk markers. Within the scope of this chapter, coagulation markers are noted which are good candidates to help with patient risk stratification and prognosis. Cardiac troponins have emerged recently as good tools for the detection of myocardial damage associated with thrombotic complications of the coronary atherosclerotic plaques.

Several markers of thrombus formation and lysis have been identified, both in stable and unstable patients [10]. Fibrinogen, activated factor VII, D-dimer, thrombin-antithrombin complexes, prothrombin fragments 1 and 2 and plasmin-a2-antiplasmin complexes are substrates and end-products of the coagulation and fibrinolysis cascades in ischaemic coronary patients [11-13]. However, little is known of their potential role in the triage and risk stratification of chest pain patients in the emergency room (ER).

80 r FVIIa

60 40 20

ANOVA, NS

10.0 r TAT

ANOVA, NS

ANOVA, NS

Nonl MI

Nonl MI

Nonl MI

1000 750

250 0

D-dimer

Nonl

Fibrinogen

ER PATIENTS' FINAL DIAGNOSIS 63% Nonl 23% UA 14% MI

Nonl

Figure 34.1 Prothrombotic markers in the diagnosis of emergency room patients with acute chest pain (<6 hours). (Nonl, nonischaemic chest pain; Ml, myocardial infarction; UA, unstable angina.)

Recently, the authors have performed a study in which plasma D-dimer levels were determined in order to assess their independent diagnostic value over conventional tests for the early diagnosis of acute coronary syndromes in the ER. Measurements of thrombin-antithrombin complex (TAT), prothrombin fragments 1 and 2 (F1+2), activated factor VII, fibrinogen and D-dimer were also determined and compared. This was a prospective study that enrolled 300 consecutive patients admitted to a teaching hospital ER for acute chest pain. All patients were 25 years or older and came to the ER with the principal complaint of central or left-sided chest pain. Only patients with an onset of symptoms <6 hours before admission were eligible for study [14].

D-dimer is the end-product of crosslinked fibrin breakdown and indicates active thrombus formation and lysis. In the past decade, D-dimer testing has been established as a useful aid in the diagnosis of deep-vein thrombosis of the lower limbs and pulmonary embolism [15, 16]. Few data exist on D-dimer values in coronary artery disease, and these data are controversial. Small studies show that D-dimer levels are higher in patients with myocardial infarction (MI) [10, 12, 17]. Francis et al. [12] provided the first demonstration of increased plasma concentration of crosslinked fibrin polymers in patients with MI using a laborious inhouse electro-phoretic technique, which was not suited for ER use. Lee et al. [17] showed that marked increases in circulating D-dimer are indicative of thrombotic complications in patients with MI, suggesting that D-dimer, besides being useful as a marker for early diagnosis, is also a risk factor for the development of MI complications. Other series, however, found normal D-dimer concentrations both in patients with MI and in healthy volunteers [18]. These studies determined D-dimer several hours after the onset of pain in hospitalized patients with a well-established diagnosis of MI.

In the authors' study, ischaemic patients had significantly higher plasma D-dimer concentrations (mean [95% CI] 656 |xg/l [484-865]) than nonischaemic patients (338 |xg/l [305-375]; p<0.01), even after adjustment for the covariates age, previous history of ischaemic heart disease and fibrinogen. Both MI (925 |xg/l [635-1349]) and unstable angina (UA) (484 ^g/l [398-589]) showed significantly higher levels than nonischaemic patients (p<0.001 and p<0.05, respectively) (Figure 34.1). Levels in MI were significantly higher than in UA (p<0.02). No statistically significant differences in D-dimer levels were found for Q-wave and nonQ-wave MI. Mean and 95% confidence interval for Q and nonQ-wave MI were 1047 ^g/l (562-1950 ^g/l) and 794 ^g/l (550-1122 ^g/l), respectively (not significant). Among the MI patients with D-dimer levels < 500 |xg/l, 58% of patients evolved to Q-wave MI and 42% of patients to nonQ-wave MI [14].

Receiver operating characteristic (ROC) curve analysis was performed at different D-dimer levels ranging from 400 |xg/l (the upper normal range in our control group) to 700 |xg/l. The discriminate power of the ROC curve analysis (combined sensitivity and specificity) for MI was best at levels of 500 |xg/l (area under the curve, 0.81; standard error, 0.04), allowing the study of MI patients with D-dimer values above 500 |xg/l (DD > 500 |xg/l). The sensitivity and specificity of D-dimer values >500 |xg/l to identify MI were 65% and 80%, respectively, and the positive and negative predictive values were 36% and 93%, respectively. Moreover, D-dimer levels over 1000 |xg/l were found in 46% of MI, in 17.2% of UA patients and in 3.7% of nonischaemic patients (p<0.001) [14].

Plasma levels of activated factor VII were not significantly different when the three patient groups (nonischaemic, MI, UA) were compared. TAT was significantly higher in MI patients than in the other two groups, and F1+2 was higher in MI patients than in nonischaemic patients before adjusting for age and previous ischaemic heart disease. However, after adjustment for these variables, the significance for both markers was lost (p=0.185 and p=0.456, respectively). Fibrinogen levels were higher among patients with MI and UA than in nonischaemic patients (p=0.02) (Figure 34.1).

Logistic regression modelling identified four variables containing independent prognostic information for MI: D-dimer >500 |xg/l, clinical assessment, positive ECG and plasma CK levels >180 U/l. Using a stepwise approach similar to that currently used in ER evaluation, a model that included clinical assessment, positive ECG and first CK>180 U/l had a diagnostic sensitivity for MI of 73% and a specificity of 98%. The addition of D-dimer >500 ^g/l to the analysis resulted in a 19% increase of the diagnostic sensitivity for MI, from 73% to 92%, without losing specificity (97%). None of the eliminated variables improved this model significantly when included one at a time (all p>0.2). The multivariate analysis did not identify D-dimer levels as independent predictors for UA, although they were significantly raised in UA.

Many studies suggest that 2-4% of patients discharged from the ER may develop MI within 48 hours [19]. In the authors' series, five patients who ultimately had MI confirmed arrived to the ER with nonspecific ECG and initial cardiac enzymes in the normal range. D-dimer levels were raised (>500 ^g/l) in 60% of these patients. Troponin assays are highly selective for myocardial injury, but, for early triage, have the limitation of being measurable only after 3-4 hours after the onset of pain [20]. D-dimer levels rise earlier than cardiac injury markers (including myoglobin) in acute ischaemic events, since they represent an earlier stage in the pathophysiology of MI and UA. The most appealing potential of D-dimer relates to detecting ongoing thrombus formation/dissolution in patients with acute ischaemic syndromes that are undetectable by conventional methods. This could result in a more cost-effective use of hospital facilities and, thereby, reduce overall costs. It is pos-

sible that elevated D-dimer levels can identify those patients most suitable for lytic and/or antiplatelet therapy, while those without such elevations are more suitable for interventional therapies.

In addition to the diagnostic utility of D-dimer for MI, this marker may also be of potential prognostic utility. D-dimer prognostic information may be particularly useful in UA patients with less elevated basal D-dimer levels. Some reports have found higher D-dimer concentrations in atherosclerotic patients than in control subjects [21], and the Physicians' Health Study [22] found that raised D-dimer levels were a marker for future MI among healthy men. Likewise, a recent prospective study shows that D-dimer is a predictor of recurrent coronary events [23], with implications for the prevention of secondary coronary events. These data fit with the concept of a hypercoagulable state that has been hypothesized to precede clinical events [24]. These results and others [17] suggest that D-dimer increases probably reflect enhanced physiological fibrinolysis as a result of ongoing thrombosis or a more pronounced fibrinolytic response to coronary thrombosis.

In another study, the authors also studied hospitalized patients with UA, subclas-sified into three groups: group A included patients with postMI angina (from days 3 to 30 after MI); group B comprised patients with new-onset angina (<2 months from onset, without previous MI during the last month) and no history of coronary heart disease (CHD); and group C included patients with crescendo angina (>2 months from onset, with no history of MI during the last month) and a history of CHD [13]. TAT, F1+2 and plasminogen activator inhibitor (PAI-1) were not different among these groups of patients. Only D-dimer and plasmin-antiplasmin (PAP) showed significant changes and were higher than levels in healthy controls. Increased levels of PAP and D-dimer were found, particularly in the group with postinfarct angina, and both parameters were significantly correlated (Figures 34.2 and 34.3). This relationship is not found in new-onset angina or progressive angina. However, in this series, PAP levels did not predict an uneventful outcome, either in the acute phase or long term (6 months) [13].

Was this article helpful?

0 0

Post a comment