[Timing and concepts of surgical treatment of upper gastrointestinal haemorrhage]

Upper gastrointestinal bleeding is a frequently occurring clinical scenario with a potentially serious prognosis. In spite of excellent endoscopic results, the mortality rate after an insufficient endoscopic treatment is exception-ally high (12.5-36 %). It is crucial to recognise factors in which endoscopy reaches its limitations. Until now, no uniform guidelines and concepts concerning diagnosis and treatment as well as timing of surgical interventions, in particular, have been defined. The main goal of this study is to lower the morbidity and mortality rates after upper gastrointestinal bleeding, with potential risk stratification according to the literature and our own data. PATIENTS / MATERIAL AND METHODS: In a retrospectively designed study 220 patients were evaluated with upper gastrointestinal haemorrhage, who were hospitalised as emergencies from 1999 to 2002. Only those patients were accepted in the study who were examined within 48 hours endoscopically by oesophagogastroduodenoscopy. In order to exclude bleeding complications of a preceding endoscopic therapy, those patients were excluded who had been investigated by endoscopy in the past than 8 days.

After endoscopic evaluation of the bleed-ing activity of 33 Forrest I a / I b bleedings 5 patients and of 52 Forrest II a / II b / II c bleedings 6 patients had to undergo surgery. The haemoglobin content of conventionally treated patients was on average 10.3 mg / dL as compared to 8.4 mg / dL for the operated patients. The conventionally treated patients received an average of 3 red cell concentrates whereas the operated patients had 11 blood transfusions. The source of haemorrhage in the operated patients was located in bulbus duodeni (n = 7), cardia and fundus (n = 2) and the corpus (n = 2).

The evaluation of our own patient data including the experiences of other authors shows that a risk stratification is possible and meaningful. The indication for surgery thereby -depends on different factors: the comorbidity of the patient, the haemodynamic in- / stability, the number of necessary blood transfusions and the localisation of the bleeding source.