Arterial hypoxaemia in cirrhosis: fact or fiction?