Incidence, Risk Factors and Outcome Of Cytomegalovirus (CMV) Viremia and CMV Gastroenteritis In Patients With Gastrointestinal Graft Versus Host Disease (GI GVHD)

GI GVHD is one of the most common causes of morbidity and mortality after allogeneic transplant. Since the response to treatment requires resolution of symptoms, a lack of an appropriate response requires that other very often other causes need to be excluded to guide therapy. One of the most challenging treatment decisions involves differentiating between GI GVHD and CMV gastroenteritis.

We retrospectively analyzed charts of patients who developed GI GVHD Between Jan 2005 to Dec 2011 to evaluate the incidence, risk factors and outcomes of CMV viremia and CMV gastroenteritis in these patients.

Between Jan 2005 to Dec 2011, 780 pts underwent Allo-SCT at our institution for various hematological malignancies. Of these, 252 pts developed GI GVHD at a median of 27 days (range 5-238 days) post transplant. Among these 252 patients, median patient age at the time of SCT was 51 years (range 20-70 years), M/F 131/121, 89/252 (35%) were related and 163/252 (65%) were unrelated donor transplants. Seventy five patients (29%) patients received reduced intensity conditioning. The distribution of patients according to the CMV IgG sero-status of donor and recipient was as follows: D+/R+ 31%, D-/R+ 26%, D+/R- 13.4%, D-/R- 30%. The diagnosis of GI GVHD with based on symptoms and was confirmed by biopsy in 239 (94%) patients. The median time from stem cell transplant to first GI biopsy was 33 days (range 12-292 days). Two patients had evidence of CMV gastroenteritis on the first biopsy and both of them had evidence of CMV viremia prior to diagnosis of GVHD.

A total of 114/252 (45%) patients with GI GVHD had evidence of CMV viremia post transplant. The median time to development of CMV viremia post transplant was 34 days (range 14-236 days). Of these 114 patients, 18 (15%) developed CMV viremia prior to the diagnosis of gut GVHD, 8 (7%) developed both concomitantly, while the rest 88 (78%) developed it after the diagnosis of gut GVHD. Treating death as a competing risk, recipient CMV IgG sero-status was significantly associated with CMV viremia with a p-value <0.001. The incidence of CMV viremia in the four sero-status subgroups was: D+/R+ 71%, D-/R+ 67%, D+/R- 18%, and D-/R- zero.

A second GI biopsy was done in 116/252 (46%) patients because of persistent, worsening or recurrent GI symptoms at a median duration of 44 days (range 8-292 days) from the first biopsy. It showed presence of CMV gastroenteritis in 29/116 (25%), normal biopsy in 17/116 (15%) and persistent GVHD in 70/116 (60%). CMV viremia was detected prior to development of CMV gastroenteritis in 28 out of 31 patients who developed CMV gastroenteritis. In the univariate analysis, five variables were associated with development of CMV gastroenteritis in a competing risks model formulation: recipient CMV IgG sero-status (+ve vs. –ve p= 0.006), development of CMV viremia (p<0.001), race (Caucasian vs. non-Caucasian p=0.019), transplant type (unrelated vs. related p=0.033) and log of CMV qPCR peak (p<0.001). Multivariable analysis yielded only recipient CMV IgG sero-status and preceding CMV viremia as being significant predictors for development of CMV gastroenteritis. A higher peak of CMV qPCR increased the risk of development of CMV gastroenteritis. The incidence of CMV gastroenteritis among patients who underwent second GI biopsies (n=116) in the four sero-status subgroups was: D+/R+ 19%, D-/R+ 27%, D+/R- 11%, D-/R- zero. The difference in the incidence of CMV gastroenteritis was not statistically significant among the D+/R+ and D-/R+ groups.

Overall survival of the 252 patients at a median follow-up of 35.5 (23.9, 45.7) months was 38% (95/252). Non-relapse mortality in this group of patients was 49 %. Pts who developed CMV gastroenteritis had an overall mortality rate of 77%. Increasing clinical grade of GVHD (p<0.001) and development of CMV gastroenteritis (p=0.008) were associated with poor survival. Development of CMV viremia had no effect on survival.

The presence of CMV viremia and GI GVHD predicts for a high incidence of CMV gastroenteritis with an ensuing high mortality. Identifying patient populations with high risk features and initiating prompt treatment may have a beneficial effect at decreasing mortality. We recommend doing second GI biopsies in patients with inadequate response to treatment of GVHD to rule out CMV gastroenteritis.

No relevant conflicts of interest to declare.