Active bleeding in benign gastro-duodenal ulcers:Predictors of failure of endoscopic injection hemostasis

Abstract

SUMMARY
The aim of this study was to define factors associated with
failed endoscopic therapy in patients with active (spurting
or oozing) upper gastrointestinal bleeding. We evaluated
clinical and endoscopic parameters related to failure of
injection hemostasis in 286 peptic ulcer patients with either
active spurting or oozing bleeding. Endoscopic injection
hemostasis was permanently successful in 218 (76.2%) and
failed in 68 patients (23.8%) who needed surgical
hemostasis. The overall mortality was 4.9 % (fourteen
patients). In univariate analysis, therapeutic failure was
significantly related to the presence of: 1) shock on
admission (p<0,0001), hematemesis (p=012), spurting
bleeding (p<0.001), low hemoglobin on admission
(p=0,005), duodenal ulcers (p = 0.04), and stomal ulcers
(p = 0.001). Previous peptic ulcer bleeding (p=0,013), or
non-use of NSAIDs (p=0,001) were negative predictors for
the outcome of endoscopic hemostasis. In multivariate
analysis, only shock, low hemoglobin concentration, nonuse
of NSAIDs and spurting bleeding were negative
predictors. It is possible, with specific characteristics, to
define a subgroup of high-risk patients for continuing
bleeding or rebleeding in spite of therapeutic endoscopy
and, thus, candidates for complimentary endoscopic method
of hemostasis or emergency surgical intervention.
Key words: endoscopic hemostasis, active bleeding,
adrenaline injection, failure, peptic ulcer blee