Recently, septic shock has become the main cause of death in intensive care units, surpassing mortality caused by myocardial infarction, stroke and trauma. The mortality rate is extremely high, and the main cause of death is organ failure and refractory hypotension. To cause excessive vasorelaxation and hypotension, the endothelial lining may release vasodilators. The most significant of these are nitric oxide (NO), prostacyclin (PGI2), and the elusive endothelial-derived hyperpolarizing factor (EDHF). Although NO is critical in controlling vascular tone, inhibiting NOS in septic shock does not improve outcome, on the contrary. The term reactive oxygen species (ROS) collectively describes a group of oxygen derivatives that includes superoxide radicals (O2.-), hydrogen peroxide (H2O2) and hydroxyl radicals (.OH). Generally, ROS are believed to cause chronic endothelial dysfunction, and to play an important role in cardiovascular diseases such as hypertension and diabetic vasculopathy. However, many studies also support the notion that endothelial cells may produce a vasodilating ROS. As these relaxations are mostly sensitive to catalase, a role for H2O2 in vasodilatation was suggested, and H2O2 was even proposed as a candidate for EDHF. We previously reported that TNF combined with the caspase inhibitor zVAD causes hyperacute shock in mice (Cauwels et al., 2003). Here we show that this shock does not depend on NO, prostaglandins, epoxyeicosatrienoic acids (EETs) or H2O2, and that calcium-dependent small-conductance K+ channels (SK) play a prominent role. Tempol, a potent O2.- and .OH scavenger, and a superoxide dismutase (SOD) mimetic, prevented the abrupt hypotension, whereas NOS inhibition did not. To understand the mechanism of action of tempol, we compared its effects with SOD and cell-permeable SOD. Surprisingly, only tempol could fully protect, indicating that .OH, and not O2.-, are responsible for the hyperacute shock. In addition, treatment with catalase pointed towards a possible detrimental effect of H2O2 removal. Indeed, catalase aggravated hyperacute zVAD+TNF toxicity, and when combined with tempol, it even undid the long-term protection provided by tempol. Together, our data thus indicate that H2O2 has a protective role in TNF-induced shock, and that part of the protective capacity of tempol resides in its ability to cause H2O2 accumulation. Importantly, tempol also completely protected against TNF- or LPS-induced shock. In conclusion, our study indicates a shock-inducing effect of .OH radicals, as well as a shock-antagonizing effect of H2O2 accumulation, and provides a rationale for the therapeutic use of tempol (as .OH scavenger and H2O2 inducer), or analogous compounds, in (septic) shock.