Utility of CMV PCR in the Evaluation of Allograft Recipients Presenting with Diarrhea

Introduction.  Graft Versus Host Disease (GVHD) and immunosuppression in allogeneic hematopoietic stem cell transplant (alloHSCT) patients can contribute to cytomegalovirus (CMV) disease.  CMV disease of the gut is associated with intestinal necrosis and ulceration and can lead to debilitating diarrhea.  Diagnostic evaluation includes assessment for CMV viremia and intestinal biopsy to confirm diagnosis.  Adequate research supporting intestinal biopsy in serum negative patients does not exist. In an effort to potentially minimize use of invasive endoscopic procedures to rule out CMV as a cause for diarrhea, we evaluated the diagnostic yield of intestinal biopsies in the work-up of allograft recipients presenting with diarrhea.

Methods.  This retrospective study evaluated a total of 485 patients that underwent alloHSCT after 2006 and were admitted to the inpatient BMT service between January 1^st, 2008 and April 30^th, 2013.  A subset of patients was identified that completed esophagogastroduodenoscopy (EGD) or colonoscopy for work-up of diarrhea.  Comparisons were made between serum CMV PCR (108 bp primer directed at CMV Immediate Early antigen product) and gastrointestinal biopsy (with morphologic evaluation for cytopathic effect and immunohistochemistry for immediate early non-structural antigen).  Pathologic evaluation confirmed gut-associated CMV disease.

Results.  CMV viremia was evident the day of intestinal biopsy 25% (99 total biopsies) of the time.  Nine biopsies (9%) in 7 different patients were positive for CMV and confirmed CMV gut disease.  Of these, 6 patients had corresponding CMV viremia.  One patient (cord blood recipient) was diagnosed with CMV gut disease by biopsy alone.  Significant association (p=0.003) and agreement (p=0.006) between CMV viremia and CMV gut disease were observed in this cohort, although we do note discordances of interest.  No apparent association between lymphocyte count and the presence of CMV intestinal disease was observed (p=0.23).  In patients with biopsy-proven disease, 43% had lymphocytosis.  100% of patients with CMV viremia and no CMV gut disease were lymphopenic.  There did not appear to be any association between CMV PCR copy number and absolute lymphocyte count (r=0.016; p=0.49).  There was no apparent pattern or relationship between CMV PCR copy number and degree of cytopathic changes on intestinal biopsy.

Conclusion.  We hypothesized that all patients with gut-associated CMV disease would have corresponding CMV viremia.  One patient with biopsy-proven disease had negative CMV PCR, so we are evaluating for potential sources of false positivity because gut-confined CMV (CMV disease without CMV viremia) is relatively uncommon in the HSCT setting.  Our data support that a lack of CMV viremia does not rule out gut-associated CMV disease and that intestinal biopsy by EGD or colonoscopy is warranted in this patient population.