Plasma von Willebrand factor (vWf) displays a complex pattern of repeating multimers, whose heterogeneous size distribution seems to depend on the proteolytic cleavage of the constituent vWf subunit. Smaller vWf multimers are thought to derive by proteolytic cleavage of the larger forms. To clarify the relationship between large multimer representation and the structure of small vWf oligomers, DDAVP was infused in patients with type-2A and -2B von Willebrand disease (vWd) variants which lack circulating high vWf forms. Before infusion, high-resolution multimer analysis demonstrated a more pronounced representation of the satellite bands of each oligomer, mainly concerning fast-moving components, especially in type 2B vWd. After DDAVP, in type-2A vWd each oligomer displayed a different organization depending on whether restoration of large vWf multimers occurred. The lack of large vWf multimer restoration, as shown in citrated samples, was associated with the fast band being significantly more represented than the slower, and almost similar to the central component. In contrast, when the high-molecular-weight vWf forms were restored, as occurred in the samples collected in the presence of protease inhibitors, the relative representation of the fast- and slow-moving bands was similar to that of normal samples. In type-2B vWd, regardless of the anticoagulant used, DDAVP infusion did not restore large vWf multimers, and each oligomer displayed a significant increase in both the central band and fast-moving satellite, the fast being even more highly represented. These findings suggest that, regardless of the origin, the disappearance of large circulating multimers in type-2A and -2B vWd induces an increased representation of the fast-moving satellite of the low-molecular-weight multimers. Moreover, the time course of large and low/intermediate multimer decrease and increase provides a further demonstration that low vWf multimers derive from the larger ones, and that mainly the fast-moving band of the oligomer is involved.