Endoscopic injection therapy for non-variceal upper gastrointestinal bleeding at Auckland Hospital

Aim The aim of this study was to audit the efficacy of endoscopic injection therapy for non-variceal upper gastrointestinal (GI) bleeding at Auckland Hospital. Methods The medical records of 183 patients who had undergone endoscopic injection of adrenaline for non-variceal upper GI bleeding at Auckland Hospital between November 1996 and May 1999 were reviewed. Results Primary haemostasis was achieved in 177 (96.7%) patients and six (3.3%) patients had early surgery due to failure of injection therapy. The overall rate of rebleeding was 18.6%. The rate of rebleeding according to the stigmata seen at the time of endoscopy was 36.8% for patients showing spurting vessel; 24% for active ooze; 18.2% for visible vessel; and 3.2% for adherent clot. Active spurting vessels seen on endoscopy were associated with higher risk of repeat injection (p <0.01) and death (p <0.001). Sixteen (9%) patients had repeat injection. Haemostasis was achieved in all, and no deaths or surgical interventions were observed in this subgroup. Conclusions Endoscopic injection therapy is an effective, simple and cheap first-line therapy for non-variceal upper GI bleeding. The repeat injection also appears to be an effective treatment for patients rebleeding after the initial injection therapy. Acute non-variceal upper gastrointestinal (GI) bleeding is a common cause of hospitalisation, with an approximate incidence of 100/100 000 adults per year. 1 Mortality rates between 6% and 14% have been reported. Mortality appears to be higher among hospitalised and elderly patients due to the high rate of comorbidity. 2–4 Patients aged 60 and below without malignancy or organ failure have mortality as low as 0.5–1%. 5,6 Approximately 20% of patients with bleeding ulcers present with melaena, 30% with haematemesis, 50% with both, and 5% with haematochesia. 7 Approximately 70% of acute non-variceal bleeding stops spontaneously, 10% bleeds continuously, and up to 20% rebleeds in the first 24–72 hours. 5 Rebleeding rates can be predicted by the stigmata of active or recent bleeding seen at the time of endoscopy. The rates of spontaneous rebleeding are approximately 90% for spurting vessel; 30% for active oozing; 50% for non-bleeding visible vessel; and 20% for adherent clot at the base of the ulcer. 6,8 Medical therapy (including H2-receptor antagonists, vasopressin, secretin, prostaglandins and somatostatin) has been found to be ineffective in both control of the bleeding as well as prevention of rebleeding. 9 Omeprazole may be effective for prevention of rebleeding as an adjunct to endoscopic therapy, but is not proven as first-line treatment of high-risk peptic ulcers. 10–12 Endoscopic therapy has been shown to control bleeding, reduce the rates of rebleeding, mortality, emergency surgery, transfusion requirements and hospital