O.M.G.E. International Upper Gastro-intestinal Bleeding Survey 1978-1982.

This presentation describes a further survey conducted under the auspices of the O.M.G.E. Research Committee; an international survey of over 4000 patients presenting to 185 clinicians in 44 centres in 21 countries. This multinational survey has been the first in this area of medicine to use specifically designed proformata and pre-agreed definitions of terminology, alongside automated data analysis. A total of 4073 patients' data were forwarded from the 44 centres, the vast majority (4010, 98.5%) containing acceptable data from subsequent analysis. As regards diagnosis, 36% of patients had peptic ulceration, 13% had varices and 6% gastric erosions as sole cause of their bleeding: 22% had multiple pathology (the most common combination being hiatus hernia and peptic ulceration). The diagnostic breakdown varied however from centre to centre. Detailed computer-aided studies show it is currently impossible to determine the source of bleeding without recourse to high technology such as radiological or endoscopic intervention. If identification of the source of the bleeding is desirable clinically, then these high technology investigations must be employed. As regards prognosis, a computer program has enabled categorisation of patients into risk categories as regards further or continued bleeding or death. This program has been tested on 2623 patients with considerable accuracy. In the computer's 'high risk' category, 60% of patients re-bled and 32% died. Conversely in the 'very low' risk category, only 4% re-bled and no patient died. These findings have enabled the construction of a simple prognostic system for use in remote areas, Initial testing (in Sikkim) indicates such a system may be of real value. Finally, immediate treatment and outcome were studied in 4010 patients. As regards management, roughly four fifths (82.7%) were transfused, 85% came to endoscopy (65.3% coming to endoscopy within 36 h of their admission to hospital) and 8.5% came to surgery. The overall death rate was low (8.3%) but amongst patients who bled further, the mortality was 30.7%--these factors being influenced by diagnosis rather than centre. The high mortality rate on conservative treatment and relatively low post-operative mortality, perhaps argue a case for more aggressive investigation and treatment.