Introduction
Graft versus host disease (GvHD) is the most common complication of bone marrow transplant (BMT) and affects the gastrointestinal system in 50% of cases. Symptoms can be difficult to separate from immunosuppression or chemo side effects. The most common parts of the GI tract involved are stomach, small intestine, and rectum. Diagnosis is made by tissue biopsy with characteristic findings of crypt cell apoptosis and loss. Endoscopic appearance can include normal mucosa to erythema, erosions, ulcerations, or bleeding. There is little guidance on patient selection and biopsy location in regarding the diagnostic value of Endoscopy in these patients. This study examines patient characteristics and endoscopy that has highest yield for GI GvHD. Biopsy decisions are typically made based on patient symptoms and appearance on endoscopy.
Methods
This was a retrospective study of 455 patients identified across three Mayo Clinic sites that have undergone allogeneic bone marrow transplant that had subsequent endoscopy for GI symptoms. Collected data included demographics, medical comorbidities, diagnosis and transplant history, GvHD prophylaxis and diagnosis, endoscopy biopsy sites and findings, and GvHD treatment.
Results
Four-hundred and fifty-four patients met our inclusion criteria. Of those, 201 patients were diagnosed with graft vs host disease. 257 were male. Mean age was 61.2 years old (SD 14.4) (see table 1). The most common indications in diagnostic endoscopies were acute myeloid leukemia (AML) (15.9%), myelodysplastic syndrome (MDS) (9.03%), acute lymphoblastic leukemia (ALL) (6.17%), and chronic myelogenous lymphoma (CML) (3.96%). In patients with CML (N=20), 90% had GvHD found on endoscopy. Conversely, in diffuse large B cell lymphoma, cutaneous t cell lymphoma, and follicular lymphoma, many more patients did not have diagnostic endoscopy. Average GvHD grade at diagnosis was 2. Biopsy locations that were statistically significant were the duodenum (OR 1.6, p 0.048), sigmoid colon (OR 2.60, p 0.002), and rectum. All endoscopy types were statistically significant. Diarrhea was a statistically significant symptom (OR p 0.013) (see table 2).
Discussion
This is the largest study exploring endoscopy in diagnosis of GI GvHD. Identifying the optimal candidates for post-transplant endoscopy can lead to earlier diagnosis and treatment. The most common indications for transplant corresponded with high rates of diagnostic endoscopy. Notable, patients with specific transplant indications including AML, MDS, ALL, and CML, warrant heightened consideration for endoscopy. Both upper and lower endoscopy should be performed, and biopsies should be strongly considered in the duodenum, rectum, and sigmoid colon.

