The syndrome of "viral hemorrhagic fever" in man caused by certain viruses, such as Ebola, Lassa, Dengue, and Crimean-Congo hemorrhagic fever viruses, is often associated with a shock syndrome of undetermined pathogenesis. However, the vascular system, particularly the vascular endothelium, seems to be directly and indirectly targeted by all these viruses. Here we briefly summarize the current knowledge on Marburg and Ebola virus infections, the prototype viral hemorrhagic fever agents, and formulate a working hypothesis for the pathogenesis of viral hemorrhagic fever. Infections with filoviruses show lethality up to 89% and in severe cases lead to a shock syndrome associated with hypotension, coagulation disorders and an imbalance of fluid distribution between the intravascular and extravascular tissue space. The primary target cells for filoviruses are mononuclear phagocytotic cells which are activated upon infection and release certain cytokines and chemokines. These mediators indirectly target the endothelium and are thought to play a key role in the pathogenesis of filoviral hemorrhagic fever. In addition, direct infection and subsequent destruction of endothelial cells might contribute to the pathogenesis. Filoviruses, particularly Ebola virus, encode nonstructural glycoproteins which are released from infected host cells. Their function as potential determinants in pathogenicity remains to be investigated.