Effect of recombinant ADAMTS13 on VWF-mediated and activation-independent platelet thrombus formation under flow. Plasma-free washed blood cells were suspended in a HEPES/Tyrode buffer (10 mM HEPES, 135 mM NaCl, 2.7 mM KCl, 0.4 mM NaH₂PO₄, and 2.8 mM dextrose), pH 7.4, containing 10 μM PG E₁, 10 μM mepacrine, 20 μg/mL purified human VWF multimers, 2 mM Ca²⁺, and 0.5 mM Mg²⁺. The blood cell suspension (platelet count, 180-390 × 10⁹/L; hematocrit, 0.38-0.43) was perfused over a surface coated with immobilized VWF without or with the addition of recombinant ADAMTS13, as indicated. The protease activity added corresponded to 17.5% of that present in normal blood. The images shown are single frames from a real time recording (Video S3). The timeline (black horizontal bar) shows at what moment during the experiment a given image was taken, as well as the wall shear rate during different periods of perfusion, as explained in the Figure 1 legend (A, 30 seconds, 5500 s^-1; B, 170 seconds, 13 000 s^-1; and C, 370 seconds, 13 000 s^-1). The bar graph shows the average count, length, and area (mean ± SEM) of all the platelet aggregates measured in 5 separate fields of view after 10 minutes of perfusion in the absence or presence of recombinant ADAMTS13 during each of 3 experiments performed with blood cells from different donors. Differences are statistically significant for count (P < .009) and area (P < .04), but not for length (P > .05).