Amantadine is an antiviral agent effective against influenza A viruses. We investigated 1) the antiviral efficacy, 2) analytical detection, 3) bioavailability and disposition, 4) pharmacokinetic modelling and 5) adverse reactions of amantadine in the horse. In vitro, amantadine and its derivative rimantadine suppressed the replication of recent isolates of equine-2 influenza virus with effective doses (EDs) of less than 30 ng/ml. Rimantadine was more effective than amantadine against most viral isolates; we suggest a minimum plasma concentration of 300 ng/ml of amantadine for therapeutic efficacy. In vivo an i.v. dose of amantadine 15 mg/kg bwt produced mild, transient CNS signs which were no longer apparent after 30 min. Amantadine administered at a dose of 15 mg/kg bwt was established as the maximum safe single i.v. dose. However, if repeated i.v. administration of amantadine is required no more than 10 mg/kg bwt t.i.d. should be used. The maximal safe plasma concentration of amantadine was not evaluated but is probably greater than 2000 ng/ml and possibly greater than 4000 ng/ml. On the other hand, horses with lower seizure thresholds, or those on medications that lower seizure thresholds, may be at increased risk of amantadine-induced seizures, which show few premonitory signs and are rapidly fatal. After i.v. administration of amantadine 10 mg/kg bwt, the disposition kinetics were well fitted by a 2-compartment open model. The estimated peak plasma concentration after this dose was about 4500 ng/ml, the volume of distribution at steady-state (Vdss) was (mean +/- s.d.) 4.9 +/- 1.9 l/kg bwt and the beta phase half-life was 1.83 +/- 0.87 h. Computer projections of plasma amantadine concentrations after i.v. administration of amantadine at a dose of 10 mg/kg bwt t.i.d. at 8 h intervals suggest peak plasma concentrations of 4000-5000 ng/ml and troughs of less than 300 ng/ml will be achieved. Amantadine administered orally at 10 mg/kg bwt and 20 mg/kg bwt showed mean oral bioavailability of about 40-60% and a plasma half life of 3.4 +/- 1.4 h; however, there was substantial inter-animal variation in bioavailability. Projections based on the kinetics observed in individual animals suggest that some animals readily maintain effective plasma concentrations of amantadine after oral administration of 20 mg/kg bwt t.i.d. On the other hand, animals in which amantadine is poorly bioavailable may require up to a 6-fold (120 mg/kg bwt) increase in the oral dose to achieve effective blood concentrations. Withholding food for 15 h did not reduce these inter-animal differences in bioavailability. Our results showed that simple dosing with oral amantadine will not yield effective plasma concentrations in all animals. While i.v. administration yielded more reproducible plasma concentrations, care should be taken to see that the seizure threshold is not exceeded. In acute situations, i.v. administration (5 mg/kg bwt) every 4 h should maintain safe and effective plasma and respiratory tract concentrations of amantadine.