Objective: To compare pulmonary function and gas exchange in ponies during maintenance of anaesthesia with isoflurane or by a total intravenous anaesthesia (TIVA) technique. Study design: Experimental, cross-over study. Animals: Six healthy ponies weighing mean 286 (range 233-388)SD 61kg, age 13 (9-16) 3years. Methods: The ponies were anaesthetized twice, a minimum of twoweeks apart. Following sedation with romifidine [80gkg(-1) intravenously (IV)], anaesthesia was induced IV with midazolam (0.06mgkg(-1)) and ketamine (2.5mgkg(-1)), then maintained either with inhaled isoflurane (FeIso=1.1 vol%) (T-ISO) or an IV infusion of romifidine (120gkg(-1)hour(-1)), midazolam (0.09mgkg(-1)hour(-1) IV) and ketamine (3.3mgkg(-1)hour(-1)) (T-TIVA). Ponies were placed in lateral recumbency. Breathing was spontaneous and FiO(2) 60%. After an instrumentation/stabilisation period of 30minutes, arterial and mixed venous blood samples were taken simultaneously every 10minutes for 60minutes and analysed immediately. Oxygen extraction ratio (O2ER) and venous admixture were calculated. Tidal volume (TV), minute volume (MV), respiratory rate (f(R)), packed cell volume (PCV), arterial blood pressure and heart rate (HR) were measured and recorded. Data were analysed with mixed model anova (=0.05). Treatments were compared overall and at two selected time points (T30 and T60) using Bonferroni correction. Results: Arterial and mixed venous partial pressures of O-2 and CO2, and TV were significantly lower and MV and f(R) were higher in T-TIVA compared to T-ISO. Venous admixture did not differ between treatments. O2ER was significantly higher in T-TIVA. Mean arterial pressure was higher and HR was lower in T-TIVA compared to T-ISO. Conclusions and clinical relevance: Whilst arterial CO2 was within an acceptable range during both protocols, the impairment of oxygenation was more pronounced with the T-TIVA evidenced by lower arterial and venous oxygen partial pressures.