Prevalence of graft versus host disease and cytomegalovirus infection in patients post-haematopoietic cell transplantation presenting with gastrointestinal symptoms.

BACKGROUND: There is lack of consensus regarding whether both upper and lower endoscopic examinations are required for diagnosis of gastrointestinal acute graft versus host disease (GI-AGVHD).

AIM: To evaluate the impact of endoscopic procedures on the diagnosis of GI-AGVHD.

METHODS: We performed a retrospective case-control study of recipients of allogeneic haematopoetic cell transplant (HCT) from 2000 to 2011, who presented with GI symptoms between 20 and 125 days post-HCT. GI-AGVHD status was based on the National Institutes of Health (NIH) clinical grading system.

RESULTS: One hundred and twenty-nine clinical GI-AGVHD cases and 184 controls underwent endoscopic examinations. Diarrhoea was present in 73% of cases and 38% of controls (P < 0.0001); 99% of patients with nausea ± vomiting and diarrhoea underwent bidirectional endoscopy. Histology had a sensitivity of 92% and specificity of 91% compared to the clinical criteria. The sensitivity for GI-AGVHD was 80% or greater when upper endoscopy (EGD) was performed with either sigmoidoscopy or colonoscopy, or if lower endoscopic examinations were performed alone. The sensitivity of EGD alone was only 48% (P = 0.003). Sensitivity was highest with biopsy of the terminal ileum (79%), followed by the ascending (74%), transverse/descending (73%) and sigmoid colons (69%). Diagnostic yield for cytomegalovirus (CMV) infection was equivalent for biopsies from both upper and lower GI tracts. Patients found to have concurrent GI-AGVHD and CMV infection (N = 18) had a poorer overall prognosis.

CONCLUSION: In patients post-HCT with GI symptoms, sigmoidoscopy alone had equivalent diagnostic yield for GI-AGVHD and CMV infection, compared with the addition of EGD or performance of full colonoscopy.