This study evaluated the validity of measuring fentanyl concentrations in equine plasma using radioimmunoassay (RIA) by comparing it to the established technique of liquid chromatography-mass spectrometry (LC-MS). Equine plasma samples were analyzed using a solid-phase Coat-A-Count fentanyl RIA and a validated LC-MS method. The fentanyl concentrations derived by both methods were compared by linear regression and pharmacokinetic analysis. The cross-reactivity of the primary equine fentanyl metabolite, N-[1-(2-phenethyl-4-piperidinyl)]maloanilinic acid (PMA), with the RIA was determined. The binding potency of fentanyl and PMA were compared at three opioid receptor subtypes in equine cerebral cortex using a radioligand binding technique. Fentanyl concentrations determined by RIA and LC-MS correlated, but the RIA overestimated low fentanyl concentrations and underestimated high fentanyl concentrations. The overestimation of low fentanyl concentrations is most likely due to the 29% cross-reactivity of PMA with the RIA. As a result, pharmacokinetic variables determined from an intravenous fentanyl bolus to four anesthetized horses differed depending on the analytical method. Although fentanyl bound with nanomolar potency to the three receptor subtypes, PMA exhibited no binding activity even at micromolar concentrations. In conclusion, when compared with LC-MS, fentanyl concentrations determined by RIA in equine plasma are misleading, especially for the calculation of fentanyl pharmacokinetics.