OBJECTIVE: To compare the frequency and outcome of upper gastrointestinal haemorrhage (UGH) patients who had undergone cardiac surgery with a control group of vascular surgery patients. PATIENTS: Patients who had undergone cardiac or vascular surgery from January 1999 to December 2000 were identified from departmental records. The inclusion criteria used were haematemesis and/or melaena in the post-operative period. RESULTS: Only 20 of the 2274 (0.9%) cardiac operations were complicated by UGH compared to eight of 708 (1.1%) vascular operations. Among those with UGH, 90% of the cardiac and 43% of the vascular patients were taking aspirin, warfarin or both. The mean interval between surgery and the UGH was 9.6 days (range 1-30) for the cardiac and 6 days (range 0-15) for the vascular patients. Duodenal and gastric ulcers were the most common cause of UGH (60%) in the cardiac group. Despite endoscopic intervention, more than one third of ulcer associated haemorrhages required surgical over-sewing, but none of the patients who had surgery died. The overall mortality on the cardiac surgery patients who experienced UGH was 15%, significantly higher than the 2.3% for the whole cardiac surgery group during the study period (P = 0.00075, OR = 8, 95% confidence interval 2.3-28). However, even this mortality is less than that of general inpatients who suffer UGH (33%). CONCLUSIONS: Cardiac and vascular surgical patients have similar low post-operative rate of UGH. Post-operative UGH is associated with increased mortality after primary surgery. Early surgical intervention appears to be life saving in those patients who are too ill to compensate for the haemodynamic disturbance of untreated UGH.
OBJECTIVE: To compare the frequency and outcome of upper gastrointestinal haemorrhage (UGH) patients who had undergone cardiac surgery with a control group of vascular surgery patients. PATIENTS: Patients who had undergone cardiac or vascular surgery from January 1999 to December 2000 were identified from departmental records. The inclusion criteria used were haematemesis and/or melaena in the post-operative period. RESULTS: Only 20 of the 2274 (0.9%) cardiac operations were complicated by UGH compared to eight of 708 (1.1%) vascular operations. Among those with UGH, 90% of the cardiac and 43% of the vascular patients were taking aspirin, warfarin or both. The mean interval between surgery and the UGH was 9.6 days (range 1-30) for the cardiac and 6 days (range 0-15) for the vascular patients. Duodenal and gastric ulcers were the most common cause of UGH (60%) in the cardiac group. Despite endoscopic intervention, more than one third of ulcer associated haemorrhages required surgical over-sewing, but none of the patients who had surgery died. The overall mortality on the cardiac surgery patients who experienced UGH was 15%, significantly higher than the 2.3% for the whole cardiac surgery group during the study period (P = 0.00075, OR = 8, 95% confidence interval 2.3-28). However, even this mortality is less than that of general inpatients who suffer UGH (33%). CONCLUSIONS: Cardiac and vascular surgical patients have similar low post-operative rate of UGH. Post-operative UGH is associated with increased mortality after primary surgery. Early surgical intervention appears to be life saving in those patients who are too ill to compensate for the haemodynamic disturbance of untreated UGH.