Frusemide reduces pulmonary vascular pressures in resting horses and attenuates exercise-induced increases in these pressures in exercising horses. The mechanism underlying these effects of frusemide is unclear. We tested the hypothesis that the haemodynamic effects of frusemide are dependent on diuresis by examining the effect of frusemide in anaesthetised horses in which diuresis was prevented by ligation of ureters. Twenty four horses were assigned randomly to one of 4 treatments: 1) frusemide (1 mg/kg bwt i.v.) and intact ureters; 2) frusemide and ligated ureters; 3) saline placebo and ligated ureters; and 4) frusemide and phenylbutazone (4.4 mg/kg bwt i.v. 12 h and 15 min before frusemide) and ligated ureters. Frusemide administration to anaesthetised horses with intact ureters increased plasma total protein concentration and reduced mean right atrial, pulmonary artery and aortic pressures. There was no significant effect of frusemide administration on haemodynamic variables or plasma total protein concentration in horses with ligated ureters. The combination of frusemide and phenylbutazone increased mean right atrial, pulmonary artery and aortic pressures in horses with ligated ureters. This study demonstrates that, in anaesthetised horses, the haemodynamic effect of frusemide is dependent upon diuresis. We interpret these results as providing further evidence that the haemodynamic effect of frusemide in horses is attributable to a reduction in plasma and blood volume.