Contractile responses evoked by the 5-HT IB/D receptor agonists, dihydroergotamine, naratriptan and sumatriptan, were compared in canine isolated coronary artery rings before and after endothelial dysfunction as obtained by inhibition of nitric oxide synthase with Nw-nitro-L-arginine methyl ester (L-NAME; 10 microM). The three agonists contracted rings in the potency order of dihydroergotamine (geometric mean pD2 value with 95% confidence limits in parentheses: 6.9 [5.3-7.9] and 7.0 [5.4-7.3] in the absence and presence of nitric oxide synthase (NOS) inhibition [I], respectively) > or = naratriptan (6.8 [5.7-7.3] and 6.4 [5.7-6.6]) > sumatriptan (4.8 [3.6-5.6] and 5.0 [3.6-5.6]) independently of the presence or absence of L-NAME. In absence of L-NAME, efficacy, as assessed by the mean maximal contractile response (Emax), tended to be greater, although not significantly, for sumatriptan and naratriptan compared to dihydroergotamine. L-NAME per se markedly increased developed tension (43.0 +/- 4.6 mN; n = 50) and potentiated maximal responses (0.6 +/- 0.2 and 10.7 +/- 2.4 mN for dihydroergotamine in the absence and presence of L-NAME respectively; 1.7 +/- 0.6 and 18.7 +/- 3.7 mN for naratriptan; 2.5 +/- 0.6 and 21.3 +/- 3.8 mN for sumatriptan; P < 0.01 in each case). Emax values of sumatriptan and naratriptan were greater than those produced by dihydroergotamine in the presence of L-NAME but remained lower than the sub-maximal contractile responses evoked by the thromboxane A2 analogue, U-46619 (ie, 32.4 +/- 5.2 mN in the absence of L-NAME; n = 50), or L-NAME per se. In conclusion, 5-HT IB/D receptor agonist efficacies in contracting coronary arteries are relatively low under basal conditions and are potentiated in the presence of a dysfunctional endothelium, whereas agonist potencies remain unaffected.