The application of soluble CD40 ligand (sCD40L) as a biomarker has garnered great scientific and clinical interest. However, there are many uncertainties with regard to the biology of sCD40L. Although presumed to be a marker of platelet activation, relative levels in plasma, serum, and platelet expression are unknown, as is the optimal method for its measurement. We measured CD40L from serum, platelet-poor plasma, and platelet surface in adults who had stable cardiovascular disease (CVD) and those who had unstable CVD (n = 40). Plasma sCD40L did not differ significantly between groups. Serum sCD40L was significantly lower (1.4 +/- 1.3 vs 5.2 +/- 3.7 ng/ml, p <0.001) and platelet membrane CD40L expression was higher (1.4 +/- 0.7% vs 0.9 +/- 0.6%, p = 0.03) in unstable compared with stable CVD. When the 2 groups were considered together, there was a significant correlation between plasma and serum sCD40L levels (rho = 0.4, p = 0.02) and negative correlations between plasma (rho = -0.3, p = 0.04) and serum (rho = -0.4, p = 0.01) sCD40L levels with platelet membrane CD40L expression. In unstable CVD, the correlation between sCD40L measurements was poor. Consistent with enhanced platelet activation, there was a positive correlation between platelet aggregation and surface CD40L expression (rho = 0.5, p = 0.02) and between platelet expression of CD40L and P-selectin (rho = 0.4, p = 0.05) in unstable CVD. There was no correlation between CD40L and platelet count or C-reactive protein. Only surface expression of CD40L compared with platelet-derived (plasma) or total (serum) CD40L level proved a reliable marker of platelet function in patients who had stable CVD and those who had unstable CVD. In conclusion, our data demonstrate the complex nature of CD40L and highlight the distinct processes of expression, shedding, and clearance of this ligand in patient populations.