The potent opioid fentanyl, is commonly used as a general anesthetic for coronary artery bypass surgery. Experiments were designed to determine the direct effects of fentanyl on unstimulated coronary artery tissue. Isolated, endothelium denuded canine epicardial rings were suspended in physiologic tissue baths. Changes in tension were measured as the concentration of fentanyl was increased. Fentanyl caused increases in ring tension at concentrations of 10(-6)M-10(-4)M, then caused a decrease in tension at 10(-3) M. Calcium channel blockade by 10(-7)M nifedipine abolished all increases in contractile responses to fentanyl and prevented the relaxation in tension produced by fentanyl. The fentanyl dose-response curve was unchanged by opioid receptor blockade with 10(-6)M naloxone and by alpha and beta adrenoceptor blockade produced by 10(-6)M prazosin and 10(-6)M propranolol. Muscarinic blockade with 10(-6)M atropine and cyclooxygenase inhibition by 10(-6)M indomethacin attenuated the constrictor response to fentanyl. The opioids alfentanil, sufentanil, morphine, and naloxone all produced a dose-response similar to fentanyl that varied only in amplitude. These findings indicate that increasing concentration of the anesthetic opioid fentanyl can cause biphasic changes in basal canine epicardial coronary artery ring tension. These responses are calcium dependent and may be characteristics of other opioid agonists and antagonists.