The present study was performed to examine the mechanism by which cyclooxygenase blockade produces vasoconstriction in well-developed coronary collateral vessels. Eight dogs were studied 4 to 6 months after occlusion of the left anterior descending coronary artery (LAD) had been performed to stimulate collateral vessel growth. At the time of study, the LAD was cannulated at the site of occlusion for measurement of retrograde blood flow as an index of collateral blood flow. Levels of 6-ketoprostaglandin F,a were 32卤13%higher in blood diverted from the collateral-dependent LAD than in aortic blood (P<.05); the increase in this stable product of prostacyclin metabolism indicated production of prostacyclin across the coronary collateral system. Administration of arachidonic acid into the left main coronary artery to reach collateral vessels entering the LAD resulted in a 21卤6% increase in retrograde flow (P<.01), demonstrating cyclooxygenase activity with production of vasodilator prostaglandins in the collateral system. Ibuprofen (10 mg/kg IV) caused a 55卤7% decrease in retrograde flow (P<.03) cyclooxygenase blockade inhibited tonic production of vasodilator prostaglandins in the collateral system. In contrast, neither thromboxane synthase inhibition with dazmegrel nor thromboxane receptor blockade with SQ 30741 caused a significant change in collateral flow, thus failing to support thromboxane-induced collateral constriction. After cyclooxygenase blockade, prostacyclin infused into the left main coronary artery was able to restore retrograde flow to the preibuprofen level. In contrast, lipoxygenase blockade with BW A4C (3.4 mg/kg IV) did not increase retrograde flow, indicating that the decreased collateral flow after cyclooxygenase blockade was not the result of increased production of vasoconstrictor leukotrienes. These findings support the concept that collateral vasoconstriction produced by cyclooxygenase blockade is the result of inhibition of the tonic production of vasodilator prostaglandins within the collateral system. (Circ Res. 1994;74:310-317.) Key Words * coronary blood flow * ibuprofenthromboxane A2 * lipoxygenase * prostaglandins the response to arachidonic acid infusion. To examine whether arachidonic acid may have vasodilating properties independent of prostaglandin production, the infusion was repeated after cyclooxygenase blockade. To determine whether prostacyclin production was enhanced in the collateral vessels, 6-ketoprostaglandin F,a, the stable product of prostacyclin metabolism, was measured in blood diverted from the collateral system. We hypothesized that if perivascular inflammation resulted in augmented production of vasodilator prostaglandins, then local production of vasoconstrictor eicosanoids, including thromboxane A2, might also be increased. To test this hypothesis, thromboxane synthase and thromboxane A2 receptor blockade were produced with dazmegrel and SQ 30741, respectively, to determine whether these interventions would increase collateral blood flow.710 To dete...