Aims Upper GI bleeding (UGIB) remains a common emergency with significant mortality. Scores
help in triaging patients, but which score should be used in the different decision
moment to identify patients at high or low death risk remains unclear. We aimed to
compare the overall performances of the more validated scores and their cut off performances
in different time-point. Secondary outcomes were to compare the scores’ performance
for predicting the need for therapeutic endoscopy, transfusion(s), rebleeding and
surgery/interventional radiology.
Methods We conducted a prospective multicenter cohort study, including UGIB patients admitted
to 50 Italian hospitals. We collected demographic, clinical data, and information
to calculate and compare performances of Rockall, PNED, AIMS65, Glasgow-Blatchford
(GBS) and ABC scores.
Results Data on 2.307 outpatients (mean age 67.5y, male 69 %) were included. For mortality,
ABC and PNED scores are the more useful (ROC>80) to classify high-risk patients, while
GBS has the best overall performance (ROC 0.74) for low risk (p<0.001). For a cut
off 0/1, GBS and ABC scores provide the higher PPV (100 %) for low-risk patients.
For high-risk patients, the highest sensitivity was shown by the AIMS65 (80.7 %),
and the highest specificity by the PNED score (94.5 %). The latter also showed a higher
PPV for mortality (28 %), followed by the ABC score (20 %) (p<0.08). The PNED PPV
for mortality was clinically and statistically higher than GB, Rockall, and AIMS65
scores (p<0.001). For secondary outcomes, no scores were clinically relevant, except
for PNED score that had a higher performance (ROC>0.87) in identify rebleeding patients,
leading to a modification of risk assessment during hospitalization.
Conclusions At admission, GBS and ABC score identify low-risk patients suitable for outpatient
management, while PNED and ABC scores can be used to identify high-risk patients.
During hospitalization, PNED score should be used to re-assess the mortality risk
if a modification of clinical status occurs.
Citation: Marmo R, Soncini M, Bucci C et al. OP151 COMPARISON OF ASSESSMENT TOOLS IN ACUTE
UPPER GASTROINTESTINAL BLEEDING: WHICH ONE FOR WHICH DECISION. Endoscopy 2021; 53:
S62.