High levels of prostacyclin (PGI2; measured as 6-keto-PGF1alpha) have been reported in patients under septic shock. Because this was at variance with our previous findings of nitration and inhibition of PGI2 synthase by endotoxin (LPS) in the endothelium, we examined the role of vascular smooth muscle as an alternative source of PGI2. Bovine aortic smooth muscle cells (SMC) in passage 1 contained high levels of PGI2 synthase but no activity and no detectable levels of COX-1 or COX-2. LPS exposure for 3 h caused COX-2 mRNA and protein levels to rise during 8 h together with a large increase in PGI2 synthase activity. In contrast, cytokines lead to only a moderate increase of both PGI2 and PGE2. Specific COX-2 inhibitors completely blocked PGI2 formation but PGE2 synthesis only partially. Unexpectedly, *NO formation remained low over 6-8 h, which may be a reason for the lack of nitration and inhibition of prostacyclin synthase in LPS exposed SMC. Our results can explain the clinical observation of severe hypotension in progressive stages of septic shock as a mechanism to compensate endothelial dysfunction. According to our data, the use of COX-2-specific inhibitors may not be advisable in septic patients. In contrast, administration of COX-1-specific blockers could prevent platelet aggregation during progressed stages of endotoxic shock.