Comparison of the Glasgow-Blatchford and AIMS65
The article: Yaka, E. et al., 2014. Comparison of the Glasgow-Blatchford and AIMS65 Scoring Systems for Risk Stratification in Upper Gastrointestinal Bleeding in the Emergency Department D. Mark Courtney, ed. Academic emergency medicine, 22(1), pp.22–30.
PICO: In patients presenting to the emergency department with upper GI bleeding, does AIM65 better than GBS identify high risk patients (defined as those that will require blood transfusion, endoscopic/surgical intervention, ICU admission, or experience rebleeding or 30day mortality)?
Design and Population:
Prospective observational study of 254 patients at a single academic center in Turkey
ED physicians aware of data collection and scoring, GI was not
Exclusions: treated at OSH for GI bleed, LGIB; AMA, or recent transfusion
Cohort events: Rebleeding in hospital 13%, endoscopy 83.1% (of which 64.2% in the ED!)–19% required endoscopy for hemostasis, surgery 2%, 7.1% mortality, 56.3% received blood transfusions, no 30day mortality or rebleeding after discharge
Three commonly cited scoring systems exist:
The Rockall score is based upon age, the presence of shock, comorbidity, diagnosis, and endoscopic stigmata of recent hemorrhage. Not practical for the ED given the need for endoscopy results to calculate score. Was appropriately not included in this Turkish study.
The Glasgow Blatchford score is based upon the blood urea nitrogen, hemoglobin, systolic blood pressure, pulse, and the presence of melena, syncope, hepatic disease, and/or cardiac failure. The score ranges from zero to 23 and the risk of requiring endoscopic intervention increases with increasing score. One meta-analysis found that a Blatchford score of zero was associated with a low likelihood of the need for urgent endoscopic intervention
AIMS65 is another scoring system that uses data available prior to endoscopy. Studies suggest it has high accuracy for predicting inpatient mortality among patients with upper GI bleeding. The score was derived using data from a database that contained information from 187 United States hospitals. The derivation cohort used data from 29,222 hospital admissions. The score was then validated using a separate data set containing information from 32,504 admissions. The study found that five factors were associated with increased inpatient mortality: Albumin less than 3.0 g/dL (30 g/L), INR greater than 1.5, Altered Mental status (Glasgow coma score less than 14, disorientation, lethargy, stupor, or coma), Systolic blood pressure of 90 mmHg or less, Age older than 65 years
A GBS of 0 was observed in 16 patients (6.3%) in the study group. Two of these were high-risk patients. A total of 101 (39.8%) patients had AIMS65 scores of 0. Thirty-four of these were high- risk patients.
A GBS of 0 had higher sensitivity than an AIMS65 score of 0 (98.68% vs. 77.6%). The negative predictive values of the GBS and AIMS65 of 0 were 87.5 and 66.3%, respectively.
The GBS and AIMS65 were similar with respect to predicting in-hospital mortality (AUCs of 0.85 vs. 0.81; p = 0.342). The GBS was superior to the AIMS65 in identifying high-risk patients, and also more accurate than the AIM65 in predicting the need for blood transfusions and interventions.
The study suggests BGS is more accurate in identifying low risk patients that might be suitable for discharge. However it is not 100% sensitive as it missed some high risk patients.
The BGS might also be more accurate in identifying high risk patients, in that it better predicted clinically relevant outcomes.
External Validity: The study was single center, and the frequency of endoscopic intervention (83% with 60% of those in the ED) , as well as frequency of blood transfusion (56% of the patients) does not match up well with our institution.
Secondary Endpoints: The use of blood transfusion and endoscopy as markers of high-risk patients is problematic as they are not purely clinical values. Multiple other factors may bias a physicians decision to resort to these two interventions. There was no discussion in the paper as to what the average Hgb was for these patients–it is hard to know whether all of these transfusions were indicated, particularly when the ED physicians making the decision to transfuse were not blinded to the GI bleeding scores. Similarly, there is no clear description of the decision to scope patients–it is a reasonable possibility that the mere decision to admit a patient, even if they had a low-risk score, could lead to endoscopic intervention, thus landing a previously “low-risk” patient into the “high-risk” group based on this papers definitions.
The AIMS65 score compared with the Glasgow-Blatchford score in predicting outcomes in upper GI bleeding. Gastrointest Endosc 2013 Apr 26;77(4):551-7. Epub 2013 Jan 26.The AIMS65 score is superior to the GBRS in predicting inpatient mortality from UGIB, whereas the GBRS is superior for predicting blood transfusion. Both scores are similar in predicting the composite clinical endpoint and other outcomes in clinical care and resource use.
This study suggest that BGS is better than AIMS65 at predicting low-risk and high-risk patients presenting to the ED with UGIB. It may be most useful in helping to predict which patients do not require emergent endoscopy and are safe to go home, BUT it is not 100% sensitive in that regard.