Experiments were performed to test the hypothesis that subarachnoid hemorrhage (SAH) causes functionally relevant perturbations of cyclooxygenase activity in cerebral arteries. Four groups of rabbits were formed: (I) controls; (II) sham injected animals (2 mi physiological solution in the cisterna magna); (III) SAH group (2 mi blood in cisterna magna); (IV) indomethacin group (4 mg/kg i.v. 30 min before sacrifice). Animals of groups II and III were used 3 days after injection. The basilar arteries (BAs) were removed and perfused at a constant flow rate (after electrocoagulation of all branches) in vitro in a 2-ml bath at 37 degrees C. After 45 min equilibration, the arteries were subjected to a fixed protocol: first, in Krebs solution, contraction to increasing extraluminal concentrations of histamine (HA), followed by a single maximal extraluminal concentration of acetylcholine (ACh); then, after 30 min rest, the same tests were repeated in oxyhemoglobin (oxyHb) solution (extraluminal, 10(-4) M). Perfusion pressure changes reflected changes in artery resistance. Although oxyHb alone increased pressure, indicating contraction of the arteries, its most important effect was to increase contraction to HA (in groups II, III, and IV but not controls) and to strongly inhibit ACh-induced relaxation in the SAH (-66.3%) and indomethacin (-46.9%) groups (III and IV) but not the control (-27.6%) group. The latter result suggests that a relaxing factor released by ACh in oxyHb solution in the control group was not present in groups III and IV. In conjunction with the results on HA, which is known to normally release prostacyclin (PGI(2)) from the endothelium, it is concluded that PGI, was not or Little released from arteries of the SAH group when they bathed in oxyHb solution. Alternatively, in the SAH group constrictor prostaglandins were released in response to HA and ACh in place of PGI(2). (C) 1998 Elsevier Science B.V.