Duodenal Biopsies Are Correlated with Diagnostic Endoscopy in Gastrointestinal Graft vs Host Disease

Introduction: Hematopoietic stem cell transplant is standard in many hematological diseases and malignancies. Gastrointestinal (GI) symptoms after transplantation can be challenging to differentiate from drug side effects, toxicity, infections, or graft vs host disease (GvHD). GvHD commonly affects the skin, gut, and liver, presenting with rash, jaundice, anorexia, diarrhea, and nausea. Diagnosis requires biopsy revealing crypt cell apoptosis and loss. Incidence of GI GvHD can reach 50%, yet there are no guidelines for biopsy protocol. Methods: This was a retrospective study of 224 patients 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: In our study, 110 patients had diagnostic endoscopy for GI GvHD. The highest yield for diagnosis was in patients that had duodenal biopsies (OR 2.15, P< 0.0312). Diagnostic yield was independent of findings on endoscopy. All endoscopy types were significant, however flexible sigmoidoscopy had the highest odds ratio at 5.5 (p < 0.0001). Significant presenting symptoms included diarrhea (OR 11.6) and nausea (OR 6.8). There was no correlation between demographics, transplant indication, or conditioning intensity with evidence of GVHD (Table 1). Conclusion: Timely diagnosis of GI GVHD is crucial because acute disease carries significant morbidity and mortality. Several studies have examined the yield of biopsy sites and protocols with varying results. Most concur that a combination of upper and lower endoscopy has the highest yield results, although some report that upper and lower tract biopsy yield was approximately the same. Based on our large cohort of GvHD patients, both upper and lower endoscopy should be pursued, regardless of presenting symptom, and biopsies should be obtained in a standardized format regardless of findings on endoscopy, especially including duodenal biopsies. Further research is warranted to optimize the diagnostic algorithm. Table 1. - Analysis of Presenting Symptoms and Endoscopic Parameters and Findings in Patients Undergoing Diagnostic Endoscopy for GI GvHD Odds Ratio (95% CI) P- value Presenting Symptom Diarrhea 11.67 (4.43-37.0) < 0.0001 Rash 1.24 (0.60-2.58) 0.56 Nausea 6.81 (2.63-20.1) 0.0002 Anorexia 7.47 (1.11-149) 0.076 Dysphagia 4.89 (0.450- 108) 0.20 Abdominal pain 59,100,000 (0.32-1.8) 0.99 Endoscopy Type EGD 3.41 (1.25-9.94) 0.02 Colonoscopy 2.74 (1.07-7.09) 0.04 Flexible Sigmoidoscopy 5.50(2.86- 10.9) < 0.0001 Endoscopic Findings Edema 0.62 (0.26-1.23) 0.27 Friable 2.53 (0.68-12.26) 0.19 Ulcer 0.85 (0.61-2.34) 0.69 Gastritis 0.94 (0.34-2.60) 0.91 Duodenitis 1.10 (0.36-3.55 ) 0.86 Colitis 2.22 (0.51-12.0) 0.31 Esophagitis 0.66 (0.24-1.79) 0.41 Normal 0.41 (0.15-1.08) 0.07 Biopsy location Esophagus 0.38 (0.14-1.00) 0.050 Stomach 17100517 (< 0.0001-NA) 0.98 Duodenum 5.46 (1.20-28.7) 0.03 Cecum 0.51 (0.04-6.23) 0.58 Ascending colon 2.54 (0.46-18.2) 0.31 Transverse colon 0.20 (0.03-1.19) 0.09 Descending colon 0.56 (0.18-1.78) 0.32 Sigmoid 0.84 (0.19-3.43) 0.82 Rectum 0.77 (0.21-2.52) 0.67 Composite location Upper GI tract < 0.0001 (NA-7.58) 0.99 Lower GI tract 4.31 (0.80-25.7) 0.09