AIM: To identify specific colonoscopic findings in patients with ulcerative colitis (UC) complicated by cytomegalovirus (CMV) infection.

METHODS: Among UC patients who were hospitalized due to exacerbation of symptoms, colonoscopic findings were compared between 15 CMV-positive patients and 58 CMV-negative patients. CMV infection was determined by blood test for CMV antigenemia. Five aspects of mucosal changes were analyzed (loss of vascular pattern, erythema, mucosal edema, easy bleeding, and mucinous exudates) as well as five aspects of ulcerative change (wide mucosal defect, punched-out ulceration, longitudinal ulceration, irregular ulceration, and cobblestone-like appearance). Sensitivity, specificity, positive predictive value, and negative predictive value of each finding for CMV positivity were determined.

RESULTS: The sensitivity of irregular ulceration for positive CMV was 100%. The specificity of wide mucosal defect was 95%. Punched-out ulceration and longitudinal ulceration exhibited relatively high sensitivity and specificity (more than 70% for each).

CONCLUSION: Specific colonoscopic findings in patients with UC complicated by CMV infection were identified. These findings may facilitate the early diagnosis of CMV infection in UC patients.

Ulcerative colitis (UC) is a worldwide, chronic, idiopathic, inflammatory disease of the rectal and colonic mucosa. Patients with UC suffer from abdominal pain, diarrhea with blood, tenesmus, and urgency to defecate. In Japan, the number of patients with UC was reported to be only 4400 in 1980. However, the number has increased sharply in recent years, and was estimated to be approximately 100 000 persons in 2008. This number is thought to increase by about 5000 persons each year.

Cytomegalovirus (CMV) belongs to the herpes virus group and is prevalent in adults. Primary CMV infection in immunocompetent people is usually asymptomatic, but the virus remains in the white blood cells throughout the life of the host. Clinically significant diseases occur in patients with suppressed cellular immunity, especially in transplant patients, acquired immunodeficiency syndrome (AIDS) patients, and patients undergoing chemotherapy. The disease is usually the result of reactivation of the latent virus rather than reinfection with the virus. Significant CMV disease may occur in various organs such as the retina, lung, and gastrointestinal tract, and the target organ may be related to the etiology of immunosuppression. In the gastrointestinal tract, CMV disease can occur in all locations from the mouth to the rectum, and usually forms ulcers in the mucosa, often accompanied by hemorrhage[1].

CMV infection has been described as a cause of relapse of inflammatory bowel disease[2-5]. In particular, CMV infection was observed in UC patients, especially those receiving high-dose corticosteroid therapy[4-7]. Moreover, CMV infection can exacerbate the disease[2-5]. Retrospective studies of UC patients have found CMV infection in 5%-21% of surgically resected specimens[2-4,8]. Steroid resistance is also reported to be one of the characteristics of CMV infection in UC[6,8]. In recent prospective studies, 25%-81% of patients with steroid-refractory UC were found to harbor the virus[5-7]. Thus, CMV infection in UC patients can have a severe clinical course and may cause death if appropriate treatment is not given. However, with the development of ganciclovir (GCV) therapy, outcomes have greatly improved[9].

Therefore, it is necessary to make an early diagnosis of CMV infection in UC patients. Although several different methods have been developed to examine CMV infection, such as histology including immunohistochemistry[10,11], serology[10-12], CMV culture[13,14], polymerase chain reaction (PCR) for CMV genome[14-16], and CMV antigenemia[14-16], it is necessary, with regard to clinical aspects, to suspect that UC patients also have CMV infection. In general, the symptoms of UC alone are not sufficient to distinguish exacerbation of UC due to CMV infection from exacerbation of UC unrelated to CMV infection. Specific colonoscopic findings for CMV infection, if any, may facilitate the early diagnosis of CMV infection.

Several reports have examined colonoscopic findings related to CMV infection. However, most of these reports focused on colonoscopic findings in patients with AIDS or in transplant patients[17,18]. The spectrum of colonoscopic findings in those patients was variable and ranged from patchy erythema, exudates, and microerosions to diffusely edematous mucosa, multiple mucosal erosions, deep ulcers and pseudotumors[1,13,17-22]. In contrast, colonoscopic findings of UC complicated by CMV infection have rarely been reported. Only a few studies have described findings without confirmation by statistical analysis[20].

Accordingly, the aim of this study was to identify specific colonoscopic findings in patients with UC complicated by CMV infection. To achieve this, we compared colonoscopic findings in CMV-positive and -negative patients, and determined the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of each specific finding for CMV positivity.

Colonoscopy is generally performed in UC patients who undergo flare-ups. Direct observation of the colon provides detailed information on the disease status and is useful for comprehending the severity of disease, and thus it is an important tool for formulating treatments. However, although CMV infection can cause poor prognosis in UC patients (i.e. a high chance of colectomy), reports on the specific colonoscopic findings of UC with CMV infection are scarce. Therefore, in this study we analyzed the colonoscopic findings of UC patients with and without CMV infection, and showed that there were specific findings that predicted CMV infection in UC patients.

There have been reports on the endoscopic findings of CMV colitis in AIDS or transplant patients. These findings include patchy erythema, exudates, microerosions, diffusely edematous mucosa, multiple mucosal erosions, or deep ulcers and pseudotumors[1,13,17-22]. In UC patients, however, the findings would be more complicated, because inflammatory changes due to UC itself in colonic mucosa exists prior to CMV infection. Therefore, in previous studies, no specific endoscopic findings were observed in UC patients with concomitant CMV infection[6,26]. In contrast, in our study, specific colonoscopic findings of UC with CMV infection were observed; the sensitivity of irregular ulceration was 100%, the specificity of wide mucosal defect was 95%, and the sensitivity and specificity of punched-out ulceration and longitudinal ulceration were both greater than 70%. Therefore, in cases where these colonoscopic findings are seen in UC patients, examination to confirm CMV infection should be performed as soon as possible.

It is known that UC patients with CMV infection, especially those who are compromised by corticosteroid therapy, experience severe symptoms. These patients are likely to undergo colectomy, and CMV infection has been confirmed in 4.6%-25% of patients who have undergone colectomy[2,3,7,8]. Recently the efficacy of GCV in CMV infection has become widely recognized. Pfau et al[9] considers GCV a beneficial treatment that significantly decreases the mortality rate and the need for surgery. Wada et al[6] found that GCV was effective in 8 (67%) of 12 patients with CMV infection. They also reported that steroid withdrawal improved steroid-resistant UC. Matsuoka et al[27] reported that most CMV reactivation-positive patients responded to conventional immunosuppressive therapies, however, high values of CMV antigenemia may be an indicator for antiviral treatment regardless of histology. Thus, treatment for UC patients with CMV infection should be different from that for UC patients without CMV infection. Therefore, our findings that predict CMV infection should be useful in therapeutic decision-making for UC patients.

In addition, there are several studies on the prophylactic/pre-emptive therapy of CMV diseases in patients with solid organ transplants, hematopoietic stem cell transplants, and human immunodeficiency virus disease[27-29]. These reports suggested the importance of early detection of CMV reactivation and early administration of anti-viral drugs before the development of CMV disease. Although no concept of prophylactic therapy has been established for UC patients, early detection and early antiviral therapy may be considered because of the high rate of colectomy in these patients. In this context, early prediction of CMV infection in UC patients by colonoscopy would seem helpful in clinical settings.

There are several methods of detecting CMV infection: histology including immunohistochemistry[10,11], serology[10-12], CMV culture[13,14], PCR for CMV genome[14-16], and CMV antigenemia[14-16]. Each method has advantages and disadvantages for the precise diagnosis of CMV infection. For example, histological examination is a relatively easy method, but the sensitivity is not so high (10%-87%)[10-12]. Meanwhile, PCR for the CMV genome is very sensitive, but the false-positive rate is relatively high. In our analysis, CMV antigenemia was adopted to detect CMV infection, because examination of CMV antigenemia is relatively sensitive (60%-100%) and easy to measure within a short period[10,14,16,30]. In addition, this method has been used for monitoring CMV infection in heart transplant recipients, and for the early diagnosis of CMV infection in renal transplant recipients[29]. Moreover, it has been reported that positive or negative results of CMV antigenemia are a good indication for antiviral therapy[28,31]. In fact, in our study, 15 (83%) of 18 patients who were positive for CMV antigenemia required antiviral therapy with GCV, and all of these patients entered remission without colectomy. In general, immunohistochemistry is considered the gold standard in the diagnosis of CMV infection. In our cohort, however, among 6 patients with positive CMV antigenemia whose biopsy specimens were immunostained with antibody for CMV, only 3 exhibited positive immunostaining. In addition, although we performed immunostaining in 16 of 58 patients with negative CMV antigenemia, none of them exhibited positive staining. Therefore, the use of CMV antigenemia for CMV detection in this study was appropriate.

Our study has several limitations. First, CMV infection was determined by CMV antigenemia alone. As described above, there are several methods to detect CMV infection. A combination of these methods may have increased the CMV detection rate. Second, of the 18 CMV antigenemia-positive patients, 15 patients who were given GCV were analyzed in our study. However, it is not certain that all 15 patients really required GCV therapy. Some of these patients may have improved without GCV similar to the remaining 3 patients. Third, this is a retrospective study that relied on the examination of recorded colonoscopic images. Therefore, bias due to patient collection and missed recording of colonoscopic findings may inevitably exist. Finally, this is a relatively small scale study in Japan, and it is possible that the results might differ in patients of other ethnicities from other parts of the world.

In the case of a severe flare-up of UC, the possibility of a concurrent CMV infection causing or worsening colitis should be considered, especially when patients are on immunosuppressive medications that may make them more susceptible to this viral infection. The prognosis of UC cases with CMV infection is considered poor. Thus, CMV infection should be closely investigated in such cases, and anti-CMV treatment (GCV) should be given immediately to reduce the rate of colectomy. The specific colonoscopic findings of patients with UC complicated by CMV infection observed in this study may facilitate early diagnosis and treatment of CMV infection.