OBJECTIVE--To determine the changes of flow in the hepatic artery and portal vein of patients undergoing tricuspid valve repair together with mitral valve replacement. DESIGN--Non-randomised parallel group study. SETTING--Tertiary referral centre. PATIENTS--18 patients with mitral valve disease with or without tricuspid regurgitation (age range 46-73; six men) and six patients with coronary artery disease and normal left ventricular function (women age range 39-78; four men). INTERVENTIONS--Mitral valve replacement with or without modified De Vega repair of the tricuspid valve or coronary artery bypass surgery. MAIN OUTCOME MEASURES--Flow in the hepatic artery and portal vein as measured by duplex Doppler flowmetry before and two to four weeks after operation. RESULTS--Preoperative hepatic artery flow was less in patients with mitral valve disease than in patients with coronary artery disease (162.7(13.1) ml.min-1 v 242.5(6.6) ml.min-1, p < 0.05). Portal vein perfusion was significantly reduced only in patients with associated tricuspid regurgitation compared with (844(83) ml.min-1 v 1422(64) ml.min-1 p < 0.05). Hepatic perfusion was unaltered after operation in patients undergoing coronary artery bypass surgery (p > 0.05). Flow in the hepatic artery and portal vein was improved only in patients undergoing mitral valve replacement with associated tricuspid valve repair (p < 0.05). CONCLUSION--Abnormalities of hepatic perfusion can be measured non-invasively in all patients with mitral valve disease but especially in those with associated tricuspid regurgitation needing valve repair. These abnormalities improved two to four weeks after mitral valve surgery and tricuspid valve repair, suggesting that measurement of portal flow might be a useful means of assessing the severity of tricuspid regurgitation.