Diagnostic Utility of Endoscopy and Biopsy in Suspected Acute Gastrointestinal Graft-versus-Host Disease after Hematopoietic Progenitor Cell Transplantation.

Acute gastrointestinal graft-versus-host disease (GI-GVHD) after hematopoietic progenitor cell transplantation (HPCT) is a common and life-threatening complication. Endoscopic biopsy of the GI tract (GIT) is required for diagnosis. However, clear evidence to optimize this diagnostic approach is lacking, leading to variation in diagnostic sensitivity between institutions. We aimed to assess the clinical, endoscopic, and histologic findings of endoscopies performed for suspected acute GI-GVHD at our institution to better define the optimal use of this strategy. We performed a retrospective cohort study of adults who had undergone endoscopy for suspected acute GI-GVHD within 180 days after allogeneic HPCT for hematologic malignancy between 2011 and 2016. Details included symptoms at time of referral for endoscopy, type of procedure performed, macroscopic findings on endoscopy, and histologic findings after gut biopsy. Correlation was made with clinical GVHD severity scores. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated and compared for each procedure. Predictors of histologic GVHD and overall survival were also compared. Of the 123 patients included, acute GI-GVHD occurred in 59 (48%). Lower endoscopy demonstrated greater sensitivity than upper endoscopy (50% versus 39%). Single upper endoscopy for upper symptoms alone had the lowest yield of GI-GVHD (14%). Combination upper and lower endoscopy demonstrated strong histologic concordance between upper and lower procedures. The addition of upper endoscopy to lower endoscopy only identified an extra 2 (4%) cases of GVHD. Advanced age and the presence of lower GIT symptoms were the only pre-endoscopy predictors of histologic GVHD on multivariate analysis. Patients with isolated upper histologic GVHD showed similar survival to patients with negative biopsies. Endoscopy and biopsy only identified 74% of those ultimately requiring treatment for acute GI-GVHD. Acute GI-GVHD remains a clinical diagnosis supported by available histologic evidence. Isolated upper GI-GVHD is rare, and in the absence of lower GIT symptoms, routine upper endoscopy does not significantly improve diagnostic yield for histologic GVHD. Overall, endoscopy and biopsy underdiagnoses 26% of clinical GI-GVHD, highlighting a need for research into novel diagnostic strategies. Crown