Methods Sixty old patients with coronary heart disease scheduled for upper abdominal surgery were randomly allocated into one of the two groups: GEA group and GA group.