Despite frequent use of immunosuppressive drugs in patients with inflammatory bowel disease (IBD) and reports of cytomegalovirus (CMV) infection following post-transplant immunosuppression, data on the frequency and clinical significance of CMV in patients with IBD are scant. Sixty-three patients with IBD (61 ulcerative colitis and two Crohn's disease) were evaluated for CMV using serology (IgM antibody, ì-capture ELISA), PCR for CMV DNA in colonic biopsy and histological assessment of haematoxylin and eosin-stained colonic biopsy. Positive result in any test was considered as CMV infection. Various parameters associated with CMV infection were analysed using univariate and multivariate analysis. Ten of 63 (15 . 8 %) patients (age 36 . 0 AE 11 . 2 years, 31 female) were infected with CMV (DNA alone in four, IgM antibody alone in two and both in four, inclusion body in one). Patients with CMV infection were more often female (8/10 vs 23/53, P , 0 . 05), had pancolitis (10/10 vs 33/53, P , 0 . 05), histological activity (9/10 vs 17/53, P , 0 . 005) and used azathioprine (5/10 vs 7/53, P ¼ 0 . 04; Fisher exact test for all). On multivariate analysis, female gender, pancolitis and histological activity were the independent factors associated with infection. Patients with CMV infection more often required surgical treatment for IBD (4/10 vs 4/ 53, P ¼ 0 . 01) and had fatal outcome (3/10 vs 0/53, P ¼ 0 . 003). CMV infection in patients with IBD may be common and is associated with poor outcome. PCR of rectal biopsy was the most sensitive method of detection followed by IgM antibody for diagnosis.
INTRODUCTIONInfection with cytomegalovirus (CMV) is an important cause of morbidity and mortality after solid organ (kidney and liver) transplantation, as these patients receive multiple immunosuppressive drugs (Patel & Paya, 1997; Wiesner et al., 1993). Patients with inflammatory bowel disease (IBD), particularly those with severe, corticosteroid-refractory and -dependent states are frequently treated with immunosuppressive agents including corticosteroids, cyclosporine, azathioprine and methotrexate, either alone or in combination (Kho et al., 2001). Therefore, patients with IBD [ulcerative colitis (UC) and Crohn's disease (CD)] are expected to be at an increased risk of infection with CMV. However, data on CMV infection in patients with IBD is still scant and most studies used only histology and/or serology for diagnosis of CMV infection (Powell et al., 1961;Cooper et al., 1977;Berk et al., 1985;Eyre-Brook & Dundas, 1986;Vega et al., 1999;Cottone et al., 2001). Although histology is quite specific, it is of low sensitivity (Beaugerie et al., 1997). PCR has emerged as the most sensitive method for diagnosis of viral infection including that with CMV (Storch et al., 1994). However, only a few studies used PCR for diagnosis of CMV infection in IBD. With blood and buffy coat preparation of leukocytes as specimens, PCR failed to detect CMV in blood of patients with IBD in some of these studies (Adani et al., 2001). Demonstrati...