Patterns and Grading of Gastrointestinal Graft-Versus-Host Disease: A Clinicopathologic Correlation Study

The present study investigated gastrointestinal involvement patterns of acute graft-versus-host disease and assessed the correlation of pathologic severity with clinical grading.

Pathology reports of gastrointestinal (GI) endoscopic biopsies taken from 164 post-hematopoietic stem cell transplant patients with at least 1 endoscopic gastrointestinal biopsy diagnosed as "consistent with acute graft-versus-host disease" between 2005 and 2019 were retrieved from the automated hospital database. Endoscopic, pathologic and clinical gradings were performed using Freiburg criteria, Lerner and modified Seattle-Glucksberg grading systems, respectively.

The majority of the patients (n = 140, 85.4%) were investigated with more than one biopsy from various gastrointestinal sites with a total of 479 biopsies: 44 (9.2%) esophagus, 90 (18.8%) stomach, 91 (19.0%) duodenum, 20 (4.2%) terminal ileum, 32 (6.7%) right colon, 87 (18.2%) left colon and, 115 (23.9%) rectum. Overall, lower gastrointestinal (n = 118/126, 93.6%) and upper gastrointestinal (n = 91/97, 93.8%) involvements were similar (P = .3). While the most severely affected site was duodenum (P = .021) in upper gastrointestinal, pathologic grades were similar in lower gastrointestinal sites, though more severe than upper gastrointestinal (P = .003). Pathologic grading had a low positive correlation with both clinical (r = 0.308, P = .001) and endoscopic grading (coefficient: 0.261, P = .003).

Considering the similar graft-versus-host disease frequency of upper and lower gastrointestinal tract, distal colon evaluation with rectosigmoidoscopy seems to be a practical approach in patients with suspected gastrointestinal graft-versus-host disease. As it was positively correlated with both endoscopic and clinical grade, pathologic grading should be performed in these patients to assess gastrointestinal involvement patterns.