The use of cyclooxygenase inhibitors has been seen to reduce the efficacy of many antihypertensive drugs. However, cyclooxygenase inhibitors are normally non-selective because they affect both vascular tissue, where the endothelial prostanoids exert principally a vasodilatory action, and the kidneys, where they also play an important role in regulating hydroelectrolytic metabolism by redistribution of intraparenchymal flow. To evaluate the relative importance of vascular district in the hypertensive patient, we administered ibuprofen - a drug acting with only a minimal antagonist activity. A group of 20 male hypertensives were randomly allocated, according to a single-blind protocol, to treatment with amlodipine (A, 10 mg/day) or lisinopril (L, 20 mg/day). Blood pressure was significantly reduced after 30 days, with a mean difference of -21.75 mmHg for systolic blood pressure (SBP) (95% confidence interval (Cl): -27.46 to -16.04; P< 0.0001) and -14.15 mmHg for diastolic blood pressure (DBP) (95% Cl: -17.13 to -11.17; P< 0.0001). Brachial artery compliance showed a mean increase of 1.657 x 10(-7) dyn-1 cm(4) (95% Cl: 1.188 to 2.126; P<0.001), and forearm resistances showed a mean decrease of -41.973 mmHg ml(-1)s (95% Cl: -75.479 to -8.467; P = 0.017). Changes in compliance were significantly related to those in SBP (r= -0.546; P= 0.013). The administration of ibuprofen (400 mg, three times a day for 3 days) was accompanied by a slight but significant increase in SBP, but not in brachial artery compliance or forearm resistances. Only SBP was affected, showing a mean increase of 4.25 mmHg (95% Cl: 1.26 to 7.24; P = 0.008). There was also reduced urinary excretion of PGI(2) and TXA(2) metabolites. The mean change in 6-keto-PGF(1 alpha) and 2,3-dinor-6-keto-PGF(1 alpha) was 45.71 ng per g urinary creatinine (uCr) (95% Cl: -0.16 to-91.25; P= 0.049) and -73.17 ng (g uCr)(-1) (95% Cl: -38.81 to -107.53; P<0.001), respectively. The mean decrease in TXA(2) catabolites was highly significant: -39.2 ng (g uCr)(-1) (95% Cl: -18.17 to-60.22; P< 0.001) and -102.87 ng (g uCr)(-1) (95% Cl: -61.86 to -143.88; P< 0.001) for TXB(2) and 2,3-dinor-TXB(2), respectively. Our study highlighted an inverse correlation between changes in blood pressure and those in urinary 2,3-dinor-6-keto-PGF(1alpha) excretion, irrespective of antihypertensive regimen. This suggests that, in the hypertensive patient treated with NSAIDs, inhibition of vascular prostanoid synthesis may play an important role in countering the efficacy of an important vascular tone regulatory mechanism.