Sign and shape: correlation of clinical findings and clot ultrastructure in arterial thrombi

Background: Thrombus architecture is an important determinant of thrombus stability and affects the outcome of preventive and therapeutic interventions in acute myocardial infarction, stroke and peripheral arterial disease, but it is hardly accessible for evaluation in the everyday clinical practice. Here we address the potential correlations between routinely available clinical data and structure of thrombi removed with percutaneous coronary intervention (PCI) or thrombendarterectomy of large arteries.

Methods: Thrombus samples removed by PCI-thrombus aspiration following acute myocardial infarction (n = 101) or surgical open repair (n = 50) in a heterogeneous group of patients (age range 36–98 years, male-female ratio 6:4) were processed in two parallel ways: glutaraldehyde-fixation for scanning electron microscopy or freezing at 80 °C for cryosections and indirect immunostaining for fibrin and platelet receptor GpIIb/IIIa. Ten to fifteen images were taken of each thrombus with both microscopic techniques, and then analyzed morphometrically to determine fibrin fiber diameter, relative occupancy by fibrin, platelet, red blood cells (RBC), white blood cells (WBC). The correlation between the measured ultrastructural characteristics and selected clinical parameters (age, sex, location of vascular lesion, blood cell counts, haematocrit, C-reactive protein (CRP) in plasma, ECG findings, anti-platelet medication, accompanying diseases) was assessed using multiple hypothesis testing and regression analysis.

Results: Fibrin content of peripheral thrombi showed positive correlation with CRP and male sex (P = 0.014 and P = 0.04, respectively), but no such dependence was observed in coronary thrombi. Platelet content of thrombi correlated stronger with the hematocrit (P = 3 x 10^-12, R² = 0.75 coronary; P = 0.02, R² = 0.238 peripheral) than with the platelet count in blood (P = 2 x 10^-4, R² = 0.20). Aspirin premedication reduced the role of local factors seen as increased dependence of thrombus platelet content on systemic platelet count (P = 4 x 10^-7, R² = 0.54) and stronger dependence of fibrin structure on RBC count in blood. Fiber diameter of peripheral thrombi decreased at higher RBC counts (P = 0.009, R² = 0.16) and the dependence was significantly stronger in the aspirin-treated group (P = 0.003, R² = 0.29). No such effect was found for clopidogrel. Sorting thrombi by their vessel of origin revealed a marked difference in fibrin-platelet ratio with lower values in the coronaries than in the iliofemoro-popliteal arterial region (P < 0.023 by Kuiper’s test for various combinations of subgroups). In line with this observation, platelet content was significantly higher in left anterior descending coronary thrombi than in the ilio-femoral subgroup (P = 0.037). In terms of platelet content and fibrin-platelet ratio coronary thrombi were similar to those of aortic origin. Neither the registered ECG findings, nor the accompanying diseases proved to be important determinants of thrombus structure on their own, but complex regression models revealed effects of combination of factors. For example, age at operation and CRP value had an additive effect on WBC content of thrombi in patients with atherosclerosis (P = 0.003, R² = 1) or hypertonia (P = 0.03, R² = 0.95).

Conclusion: The fibrin and platelet content of arterial thrombi as well as their fibrin structure are differentially affected at different vascular locations and by systemic blood cell counts. Conventional anti-platelet drugs differ in their impact on thrombus structure. Improved individually tailored strategies for prevention of acute thrombotic events could be developed on the basis of these findings.