Gastrointestinal Bleeding in Hospitalized covid‐19 Patients: a propensity Score Matched Cohort Study

Accepted Article

---

This article is protected by copyright. All rights reserved

The analysis for this paper was generated using SAS Studio version 3.8 (SAS Institute Inc., Cary, NC, 

USA), and a p-value <0.05 was considered statistically significant.

RESULTS

Patient characteristics 

After manual chart review, 339 patients with COVID-19 were identified as having a GI bleed yielding 

a point prevalence of 3.04% (339/11,158) among all patients hospitalized with COVID-19. Of these 

patients, 25 patients were excluded from the case control analysis (1 patient was missing Charlson 

Comorbidity Index needed for matching and 24 were still admitted at the time of data collection), for 

sample size of 314 patients with a GI bleed. GI bleed characteristics can be found in Table 1.  The 

majority of patients who developed GI bleeding had upper GI bleeds (68%, n=214/314), and had a 

mean hemoglobin of 8.4 g/dL. Of these patients, 144/314 received a blood transfusion (46%) with the 

median number of units being transfused was 2 (1-3).  Of the GI bleeders, 20/314 patients (6%) 

underwent endoscopy of which 85% (17/20) were upper endoscopies. The most common etiology for 

bleeding found on endoscopy was gastroduodenal ulceration.  

One hundred and four COVID-19 patients with a GI bleed on admission were matched to 104 

COVID-19 patients without GI bleeding based on propensity scores. The absolute standard difference 

for each matched characteristics (Table 2) was acceptable for each characteristics (<0.25).  The mean 

age of patients was approximately 70 years, with a nearly equal distribution across men and women. 

Two hundred and ten COVID-19 hospitalized patients who developed a GI bleed during their 

hospitalization were matched to 210 COVID-19 patients without a GI bleed based on propensity 

score. The absolute standard difference for each matched characteristics (Table 3) was acceptable for 

each characteristics (<0.25). 

The analysis for medications as risk factors for GI bleeding is summarized in Table 4.  Antiplatelet, 

anticoagulation, steroids, or NSAIDs were not associated with developing a GI bleed during the