Exercise-induced hypervolemia is normal in racehorses. Stress failure of the blood-gas barrier causes EIPH and occurs when the transmural pulmonary capillary (Pcap)-alveolar pressure difference (Ptm) exceeds the barrier's stress failure threshold. Why Pcap increases is incompletely understood. We hypothesized that alterations in blood volume (BV) affect left ventricular (LV) and pulmonary arterial wedge (PAW) pressures, and Pcap, and correspondingly affect EIPH severity. Six thoroughbreds with EIPH underwent treadmill exercise at the same speed (≈11.9m/s [11.1, 12.2]; median [IQR]) before (≈119%V̇O2max; B), 2hr after 14L depletion of blood (≈132%V̇O2max; D) and 2hr after reinfusing the blood (≈111%V̇O2max; R). LV, pulmonary arterial (PAP), PAW and intrapleural (Ppl) pressures were measured throughout exercise. Pcap = (PAP+PAW)/2 and Ptm = Pcap - Ppl. EIPH severity was assessed 60min post-exercise by tracheoendoscopy (EIPHgrade) and bronchoalveolar lavage erythrocyte count (BALRBC). A mixed-effect model and Tukey post-hoc test analyzed effects of BV changes on LV, PAW, Pcap, Ppl, Ptm and EIPH. p≤0.05 was significant. EIPH severity was affected by BV (EIPHgrade: p=0.01, BALRBC: p=0.003). Peak Ptm was not different between B (146mmHg [140; 151]) and R (151mmHg [137; 160]) but was lower for D (128mmHg [127; 130]; B: p=0.005, R: p=0.02). LV end diastolic pressure (LVED) was correlated with maxPcap (r2=0.52, p=0.001). Ppl was unaffected by changes in BV. Vascular pressures and Ppl fluctuated constantly during exercise and independently influenced Ptm. Circulating BV but not exercise intensity had a major effect on LVED, Pcap and Ptm, and was correlated with EIPH severity in thoroughbred racehorses (r2=0.48, p=0.001).