The impact of ulcerogenic drugs on predictability of the Glasgow-Blatchford, Rockall and Full Rockall risk scoring systems

Aims Several scoring systems have been designed to assist with the risk stratification of nonvariceal upper gastrointestinal bleeding (NVUGIB) [1]. These include the Glasgow Blatchford Score (GBS), the Rockall Score (RS) and Full Rockall score (FRS). Data are lacking on the impact of ulcerogenic drugs use in prediction of validated scores for specific outcomes of NVUGIB. Our primary objective was to assess the diagnostic accuracy of the bleeding risk prediction scores utilizing receiver operating characteristic (ROC) curve and to examine whether the use of ulcerogenic drugs impact their diagnostic accuracy.

Methods The study was conducted retrospectively, in patients who were admitted to our emergency department due to NVUGIB. We assessed the use of following medications: acetyl salicylic acid (ASA), NSAID, corticosteroids, oral anticoagulants (OAC and NOAC), antiplatelets drugs (APD). Subjects were dichotomized into ulcerogenic drug-users group (n=222, 73.8%) and non-ulcerogenic drug users group (26.2%). The following adverse events were followed: the need for transfusions of deplasmatised erythrocytes (DPE) and the need for endoscopic intervention(EI),while severe adverse events (SAE) were rebleeding (RB), the need for surgical intervention(SI) and lethal outcome (D). GBS, RS and FRS were calculated for each patient.

Results Our cohort consisted of 301 patients with NVUGIB, of which 58.8% were male (n=177). Peptic ulcer disease (60.8%) was the most common finding. The use of OAC or NOAC (p=0.024) as well as the use of APD (p=0.003) were significantly associated with mortality. In the whole study group, we found that GBS (AUC 0.804, [95% CI, 0.754—0.847]) outperformed RS and FRS in predicting the need for DPE (p<0.001). FRS outperformed RS (p<0.001) in predicting the need for EI and SI (p<0.001), and rebleeding (p=0.031).The AUC for all three scores were high (GBS=0.689, RS=0.710, FRS=0.778) in predicting SAE, and FRS outperformed RS (p=0.012). In ulcerogenic drugs-users GBS outperformed RS and FRS (p<0.001) in predicting the need of DPE. FRS demonstrated better performance than GBS in predicting the need for EI (p=0.039), and compared to RS in predicting the need for SI (p<0.001). FRS (AUC 0.762, 95% CI, 0.701-0.817) outperformed RS in predicting SAE (p=0.016). In the control group, GBS showed better performance in predicting the need for DPE(p<0.001). FRS outperformed GBS (p=0.035) and RS(p<0.001) in predicting the need for EI. Among patients who used APD, GBS (AUC 0.833) was superior to the RS (AUC 0.778) and FRS (AUC 0.722) in predicting SAE. Among ASA, OAC or APD-users, FRS was significantly less reliable in predicting SAE, compared to the controls. In patients using APD, FRS was inferior compared to GBS and RS for SAE and lethal outcome.

Conclusions The use of ulcerogenic drugs demonstrated an impact on GBS,RS and FRS reliability in patients with NVUGIB.In ASA and APD users,FRS was less reliable in predicting SAE. However,further prospective studies are needed with larger sample size and longer follow-up.

Publication History

Article published online:
27 March 2025

• References

• 1 Gralnek IM, Stanley AJ, Morris AJ. et al. Endoscopic diagnosis and management of nonvariceal upper gastrointestinal hemorrhage (NVUGIH): European Society of Gastrointestinal Endoscopy (ESGE) Guideline – Update 2021. Endoscopy 2021; 53 (3): 300-332 Epub 2021 Feb 10