Objective:To investigate the effect of epidural anesthesia combined with inhalation or intravenous anesthesia on intrapulmonary shunt and oxygenation in patients undergoing long term single lung ventilation.Methods:Eighty patients, aged 35-75, American Society of Anesthesiology (ASA) classification of I-III, undergoing thoracic surgery with one lung ventilation more than three hour, were randomly divided into propofol group (group Pro), propofol combined with epidural anesthesia group (group Pro+Epi), isoflurane group (group Iso) and isoflurane combined with epidural anesthesia group (group Iso+ Epi), 20 patients in each group. Arterial blood and mixed venous blood were taken for blood gas analysis, and hemodynamic data were recorded at following time points: before induction in supine position (T1), 30min after bilateral lung ventilation (T2), 15min after one lung ventilation (T3), 30min after one lung ventilation (T4), 60min after one lung ventilation (T5), 180min after one lung ventilation (T6), intrapulmonary shunt (Qs/Qt) was calculated according to the correlation formula.Results:Qs/Qt values at T2-6 in four groups were significantly higher than that of T1, and Qs/Qt values at T3-6 was significantly higher than that of T2 (P< 0.05); PaO2 at T2-6 were significantly higher than that of T1, with PaO2 at T3-6 were significantly lower than T2 (P< 0.05). Between groups, Qs/Qt values in group Iso were significantly higher than that of group Pro, Pro+Epi and Iso+Epi at T3-5 (P< 0.05). There was no significant difference in PaO2 between groups (P> 0.05). CI at T3-6 in group Iso and Iso+Epi were significantly higher than that of T1 (P<0.05), and were significantly higher than that of propofol group (P<0.05). MAP at T3-6 in group Pro+Epi and Iso+Epi were significantly lower than that at T1 (P <0.05). Heart rate at T4-6 in group Iso were significantly higher than T1, and higher than group Pro and group Iso+Epi (P <0.05).Conclusion:One lung ventilation may predispose to increase of intrapulmonary shunt and decrease in arterial partial pressure of oxygen; isoflurane inhalation anesthesia is more likely to cause intrapulmonary shunt, but no changes in arterial partial pressure of oxygen.