Specific ultrastructural modifications in platelets from patients with severe COVID-19. PRP from healthy donors (n = 6), septic patients without COVID-19 (bacterial sepsis, n = 4), and patients with severe COVID-19 (n = 10) were fixed and analyzed by TEM (A). Representative electron micrographs at different magnification are shown. Different parameters and cellular elements were quantified using electron micrographs of platelet complete cross-sections (186 from n = 6 healthy donors [31 for each], 160 from n = 4 septic patients without COVID-19 [40 for each], and 240 from n = 10 patients with severe COVID-19 [24 for each]). The percentage of platelets exhibiting enlarged vesicles (vesicle surface/platelet cross section surface ratio > 6.1, which is the upper quartile for control platelets) was quantified (B). **P < .01; ***P < .001 according to the nonparametric Mann-Whitney test. To check whether the vesicle-like structures were part of the OCS, platelet-rich plsma from 4 healthy donors and 6 patients with severe COVID-19 was fixed and stained with tannic acid before analysis by TEM (C). Representative transmission electron micrographs are shown (the top two from healthy controls and the bottom ones from patients with COVID-19). Magnifications of selected areas (×10 000) show that the vesicles were not labeled with tannic acid, excluding their belonging to the OCS (see also supplemental Figure 11). The cross-sectional surface (D), the number of vesicle-like structures per cross-section (E), the surface of vesicle-like structures per cross-section (F), and the vesicle-like surface/platelet cross-section surface ratio (G) were quantified. ****P < .0001 according to the nonparametric Mann-Whitney test. The presence or absence of microtubules (H) and pseudopodia (I) per platelet cross-section was also quantified as parameters of platelet activation. Fischer’s test was used for statistical analysis.