Mika Omiya

Total colonoscopy with a transparent hood for trainees.

TO THE EDITOR: The article by Sawhney et al. compares the prevalence of advanced colonic neoplasm in nonanemic patients and anemic patients with a variety of ferritin levels (1). While the use of odds ratios in Table 3 is correctly interpreted in the text, the likelihood ratios in Table 4 are not clearly discussed in the text. Guyatt et al. define an odds ratio as “a ratio of the odds of an event in an exposed group to the odds of the same event in a group that is not exposed” (2). Using the data from Table 3, Sawhney et al. correctly state patients with ferritin <50 ng/mL and ferritin 51–100 ng/mL are almost 5 times more likely to have advanced colonic neoplasia than those patients with ferritin >100 ng/mL or nonanemic controls. A likelihood ratio is defined by Guyatt et al. as the relative likelihood that a given test would be expected in a patient with as opposed to one without a disorder of interest (2). Fletcher and Wagner state likelihood ratios reflect the probability of that test result in people with the disease divided by the probability of the result in people without disease (3). Therefore, using the likelihood ratios in Table 4, the authors should have clarified in the text that a ferritin >100 ng/mL is 0.27 times as likely to occur in a patient with colonic neoplasm compared to a patient without colonic neoplasm. If a statement defining likelihood ratios was included, readers could more readily interpret the results. The likelihood ratio and the odds ratio could then be expressed to a patient in terms that the patient could understand and weigh the risks and benefits.

Significant association of appendiceal neoplasms and ulcerative colitis rather than Crohn's disease.

To the Editor: We read with interest the article by Orta et al on a retrospective case– control study of primary appendiceal neoplasms in inflammatory bowel disease (IBD) patents. Although there was no significant difference in prevalence of appendiceal cystadenomas between IBD patients and non-IBD controls, appendiceal cystadenomas were 15-fold more prevalent among IBD patients with synchronous colorectal neoplasia compared with controls. Appendiceal carcinoids were equally prevalent in IBD patients and controls, and IBD patients with synchronous colorectal neoplasia and controls. Although they concluded that IBD patients with synchronous colorectal neoplasia is a risk factor for occurrence of appendiceal cystadenomas, we believe that further analysis might disclose the high prevalence of appendiceal neoplasms in ulcerative colitis (UC) patients rather than Crohn’s disease (CD) patients. IBD patients have an increased risk for the occurrence of colorectal neoplasia and cancer in affected areas, especially in patients with long-standing and extensive disease. IBD is characterized by repetitive cycles of acute inflammation, which may generate free radicals and other toxic metabolites, and cause DNA damages and mutation, leading to neoplasia and cancer. In CD patients, colorectal cancers and carcinoids also occur in unaffected areas, perhaps resulting from distant proinflammatory mediators rather than a local inflammatory effect from adjacent CD. The notion of a direct relationship between IBD and appendiceal neoplasia and cancer stems from the relatively high prevalence of chronic mucosal inflammation in the appendix of IBD patients. Appendiceal involvement of UC, ‘‘ulcerative appendicitis,’’ occurs in 48%–86% patients with UC, while CD appendicitis occurs in 40%– 52% patients with CD. Patients with ulcerative appendicitis experience a more aggressive and relapsing disease course compared with those without ulcerative appendicitis. In UC patients with ulcerative appendicitis, the inflamed appendiceal orifice may block excretion from the cavity, resulting in the occurrence of appendiceal neoplasia and cancer. Although many case–control studies suggest that previous appendectomy is rare in UC patients, appendectomy was associated with an early increased risk of CD occurrence. Patients with previous appendectomy have a delayed onset of UC, a reduced need for immunomodulators and proctocolectomy, and a reduced relapse rate and extent of UC. Moreover, we and several other investigators reported improvement of UC after appendectomy, especially in young patients with ulcerative appendicitis. We also disclosed that the proportion of early-but-not-mature-activated T cells is significantly increased in the appendix of UC patients, and suspect that the appendix may be a priming site in the occurrence of UC. We also analyzed the results of the study by Orta et al and recognized that appendiceal cystadenomas tended to occur in UC patients (6/377; 1.6%; P 1⁄4 0.0816), but not in CD patients (3/317; 0.9%; P 1⁄4 0.3826), compared with controls (2/498; 0.4%). Because of the high prevalence of appendiceal cystadenomas in UC patients, high appendiceal involvement of UC, and close association between the appendix and UC, rather than CD, we believe that further studies would disclose the significant high prevalence of appendiceal neoplasms in UC patients.

Colonoscopy with a transparent hood: simple technique for improved quality of colonoscopy.

To the Editor: We read with interest the article by Radaelli et al. (1) on the factors that influence the technical performance of colonoscopy. In nationwide quality improvement programs of colonoscopy in Italy, researchers set out to identify the factors in clinical practice. They concluded that the sedation/analgesic use, bowel preparation quality, endoscopist experience, and colonoscopy volume of centers influenced the quality of colonoscopy. We believe that a simple technique can improve the quality of colonoscopy.

Much colonic surface visualization by a standard colonoscope with a transparent hood.

TO THE EDITOR: We read with interest the article by East et al. (1) on the colonic surface visualization by computerized tomography (CT) colonography-simulated colonoscopy. With CT colonography simulation, they examined how much colonic surface is visualized by a standard optical colonoscope (field of view 140◦) and a wide-angle (170◦) colonoscope with or without a retrograde viewing auxiliary imaging device (RVAID;135◦). The percentage of colonic surface visualized by a simulated 140◦ and 170◦ colonoscope was 86.6% and 92.2%, respectively, and that in addition with an RVAID was 98.7% and 98.9%, respectively. Although a simulated 140◦ colonoscope with an RVAID looks promising, with almost complete colonic surface visualization, we believe that the simulated method cannot be applicable to routine clinical settings because a retrograde viewing in the entire colon is infeasible. Due to the rapid progress of colonoscopic technology, optical colonoscopy is becoming increasingly popular for the management of colorectal diseases and is being considered the gold standard for the diagnosis of mucosal lesions (2). While the value of colonoscopy depends largely on the ability of endoscopists, incomplete colonoscopy by inexperienced hands and the risk of complications have been considered (2). Although colonoscopic polypectomy has been shown to reduce the incidence of subsequent colorectal cancers, some polyps, and even cancers, may be missed during colonoscopy because they lie outside the field of view, either hidden behind the semilunar folds, at the flexures, or in the lower rectum (3). Previous studies, including our study (2, 4–6), showed that an auxiliary device, a transparent hood, attached to the tip of a standard optical colonoscope was effective for total colonoscopy. Moreover, the technique requires less experience of endoscopists (2, 6). By depressing the semilunar folds, colonoscopy with the hood ensures good visual fields and easy recognition of the luminal continuity through the transparent wall of the hood at bends (2). On randomized controlled trials of colonoscopy with or without the hood (2, 6), more polyps are detected during colonoscopy with the hood. Although East et al. (1) did not address the issue of polyp detection in their study, focusing instead on colonic surface visualization as a surrogate, we believe that colonoscopy with the hood is a more practical and effective candidate. Because colonoscopy with the hood is simple, inexpensive, and readily available in most institutions, the technique should be performed in routine clinical practices for much colonic surface visualization, especially for more polyp detection.

Improved polyp detection: narrow-band imaging colonoscopy with a transparent retractable extension device.

Improved Polyp Detection: Narrow-Band Imaging Colonoscopy With a Transparent Retractable Extension Device

Complete endoscopic resection of large sessile or flat colorectal polyps with high-magnification chromoendoscopy or endoscopic submucosal dissection.

TO THE EDITOR: We read with interest the article by Arebi et al. [1] on endoscopic mucosal resection (EMR) of large sessile or flat colorectal polyps. Because all the polyps were 20 mm or larger in size, all resections were performed in a piecemeal fashion. During follow-up, 60 (40.3%) of the 149 polyps recurred, and 7 of the 60 polyps required surgery because of persistent recurrence of 4 benign polyps or incomplete resection of 3 invasive cancers. The recurrence was significantly related to the larger polyp size. Although the investigators concluded that a stricter follow-up is necessary for larger polyps because of a higher risk of recurrence, we believe that the recurrence rate could be reduced with the use of a novel technique. Large sessile or flat colorectal polyps have a greater malignancy potential [2,3]. Although, traditionally, these polyps are treated surgically, endoscopic resection is preferable to surgery because of its lower cost and lower morbidity [2]. EMR is a useful therapeutic technique because it permits en bloc resection [4,5], thereby providing a complete specimen for histological evaluation [4,6]. Based on the size and the location of the polyps, endoscopic resection can be performed en bloc or piecemeal [5]. The piecemeal technique is used most frequently in large sessile or flat polyps [5] as in the study by Arebi et al. [1]. In our previous study of EMR for gastric tumors, piecemeal resection directly influenced incomplete therapy [7]. The piecemeal resection of large colorectal polyps is also more likely to result in incomplete therapy compared with en bloc resection [4], and may be associated with a higher rate of recurrence [2,6]. This is presumably because of the residual adenomatous tissue that is not recognized with conventional colonoscopy [2,8]. The delay in providing adequate treatment also brings the risk of malignant changes in the residual polyps. Moreover, histological definition of the tumor margin is difficult on the multiple tissue specimens [2]. Argon plasma coagulation (APC) is suitable for treating a large mucosal area with a limited and predictable depth of tissue coagulation [2]. Whereas application of APC to the margins treated after piecemeal resection is reported to reduce the recurrence rate of large sessile or flat colorectal polyps [2], another study has shown that the recurrence rate is similar regardless of whether APC is added to complete the EMR [6]. The intense submucosal desmoplasia after APC also makes further resection difficult. Although Arebi et al. [1] applied APC to the residual adenomatous tissue, 40.3% of the polyps recurred after piecemeal resection. Endoscopic submucosal dissection (ESD) enables en bloc resection, regardless of tumor size, with a higher rate of radical cure than EMR [5]. Before ESD had been positively applied for early gastric cancers in Japan, we had performed endoscopic resection of gastric tumors after circumferential incision of the normal mucosa surrounding the tumors in order to avoid incomplete resection [7]. Although ESD is not a commonly used technique for large colorectal tumors because of the technical

Pathophysiology of the Appendix in Ulcerative Colitis

To the Editor: We read with interest the article by Myrelid et al. ( 1 ) on a national cohort study of appendectomy and the risk of colectomy in ulcerative colitis (UC). Th ey assessed the longitudinal relationship between appendectomy, appendicitis, and disease course in UC patients. Th ey concluded that appendectomy early in life and before developing UC was associated with a lower risk of colectomy as well as UC-related hospital admissions. Appendectomy for appendicitis aft er established UC appeared to be associated with a worse disease course, with an increased rate of subsequent colectomy. We would like to emphasize the pathophysiology of the appendix in UC patients based on our previous investigations ( 2,3 ). Although UC is characterized by continuous and diff use colonic infl ammation extending proximally from the rectum, appendiceal involvement of UC, ulcerative appendicitis, has been increasingly recognized in 48–86% patients with distal UC ( 2–4 ). Patients with ulcerative appendicitis experience a more aggressive and relapsing disease course compared with those without ulcerative appendicitis. Although the human appendix is considered a vestigial remnant, many case–control studies suggest that UC patients with previous appendectomy is rare, and have a delayed onset of UC, a reduced need for immunomodulators and proctocolectomy, and a reduced relapse rate and extent of UC ( 1–4 ). Th eir eff ects are recognized in young patients with appendectomy for an 1 Third Department of Internal Medicine, Kansai Medical University , Osaka , Japan . Correspondence: Mitsunobu Matsushita, MD, Third Department of Internal Medicine, Kansai Medical University , 10-15 Fumizono-cho , Moriguchi , Osaka 570-8506 , Japan . E-mail: matsumit@mc.kmu.ac.jp

Nationwide Studies of Hematological Malignancies in Inflammatory Bowel Disease

infl ammatory bowel disease: a nationwide study in Taiwan”. Th ey identifi ed 3,348 infl ammatory bowel disease (IBD) patients without previous cancer, including 685 with Crohn’s disease (CD) and 2,663 with ulcerative colitis (UC), and analyzed the incidence of cancer in IBD patients. Although the risk of colorectal cancer did not increase signifi cantly in either CD ( P =0.7) or UC ( P =0.22) patients, CD ( P <0.01) and UC ( P =0.02) patients had a higher risk of hematological malignancies (HM). Th ey found no signifi cant association between thiopurine use and HM. We would like to emphasize the fact based on our previous nationwide study in Japan ( 2 ). IBD, CD and UC, is a chronic intestinal disorder of unknown cause ( 1,2 ). Th e incidence of IBD has increased rapidly in recent decades. An abnormal mucosal immune response has a major role in the pathogenesis of IBD ( 3 ). Immunemediated diseases oft en cause chronic infl ammation, which may promote cancer development ( 1,4 ). Th e infl ammatory cells may generate free radicals and other toxic metabolites that could cause DNA damage and mutation ( 3 ). Th e risk of HM is also worth noting because several studies have found the higher risk of HM in patients with specifi c immune-mediated diseases ( 1,2,4 ). A previous study also showed that thiopurine use is associated with an increased risk of HM in IBD patients ( 5 ). To disclose the development of HM in IBD patients in Japan, we conducted a multi-center nationwide survey ( 2 ). Questionnaires for the development of HM in IBD patients were sent to 88 facilities in the fi rst survey, followed by the second survey with a more detailed questionnaire, sent to the 27 facilities where HM developed. Out of 36,939 IBD patients in the fi rst survey, 28 cases of HM related with IBD [12 of 10,500 UC patients (0.11%), 16 of 6,310 CD patients (0.25%)] in the second survey were analyzed. In our study of Japanese IBD patients, thiopurine use did not act as a risk factor of HM In a recent French population-based cohort, Cheddani et al. ( 4 ) identifi ed 844 elderly onset IBD patients, including 370 CD and 474 UC. Although there was no signifi cant increased risk of colorectal cancer in IBD patients, either CD or REFERENCES 1. Pellicano R . Is viral hepatitis decreasing in the United States? Am J Gastroenterol 2017 ; 112 : 1170 (this issue) . 2. Kabbany MN , Conjeevaram Selvakumar PK , Watt K et al. Prevalence of nonalcoholic steatohepatitis-associated cirrhosis in the United States: an analysis of National Health and Nutrition Examination Survey data . Am J Gastroenterol 2017 ; 112 : 581 – 7 . 3. Ditah I , Ditah F , Devaki P et al. Th e changing epidemiology of hepatitis C virus infection in the United States: National Health and Nutrition Examination Survey 2001 through 2010 . J Hepatol 2014 ; 60 : 691 – 8 . 4. Doycheva I , Watt KD , Rifai G et al. Increasing burden of chronic liver disease among adolescents and young adults in the USA: A silent epidemic . Dig Dis Sci 2017 ; 62 : 1373 – 80 . 5. Roberts H , Kruszon-Moran D , Ly KN et al. Prevalence of chronic hepatitis B virus (HBV) infection in U.S. households: National Health and Nutrition Examination Survey (NHANES), 1988-2012 . Hepatology 2016 ; 63 : 388 – 97 . 6. Center for Health Statistics N . National Health and Nutrition Examination Survey: Analytic Guidelines 2011 . Available at: https://www. cdc.gov/nchs/data/nhanes/analytic_guidelines_11_12.pdf .

No large ulcer: A predictor of latent cytomegalovirus infection in ulcerative colitis

We read with interest the clinical review article by Lawlor and Moss on the role of cytomegalovirus (CMV) in exacerbations of inflammatory bowel disease (IBD). They mentioned that colonic polymerase chain reaction (PCR) for CMV DNA appears to have a high ‘‘false-positive’’ rate, due to the detection of viral DNA in the absence of histological signs of disease. They also described that there are insufficient data to determine whether antiviral therapy actually impacts outcomes such as colectomy and remission rates, except in cases of severe systemic reactivation when CMV reactivation has been detected in refractory patients. We believe a simple endoscopic feature predicts latent CMV infection in ulcerative colitis (UC) patients with positive mucosal viral assay in whom antiviral therapy is not required. Many techniques have been developed for detecting CMV infection, such as histopathology, serology, endoscopy, CMV antigenemia assay, and CMV DNA assay. CMV antigenemia assay has been widely used to diagnose and monitor CMV infection, but antigenemia levels in UC patients are lower than those observed in transplant recipients with CMV infection. Because of its high sensitivity, the CMV DNA assay with the use of colonic tissue is more reliable than blood tests for detection of CMV reactivation in UC patients at an early stage. However, it is still a serious problem that the detection of CMV DNA may not always indicate active CMV infection because it is also possible to detect a latent CMV infection that is not associated with the exacerbation of UC. Although the role of CMV in the exacerbation of UC has been much debated, the clinical significance of CMV reactivation complicating UC patients remains unclear. Whereas detection of CMV infection at an early stage and start of appropriate treatment are important for UC patients, CMV is frequently reactivated in active UC patients, but disappears with clinical improvement without antiviral therapy. Some investigators suggested that CMV behaves in the intestine as a nonpathogenic bystander and innocent observer, and others reported that CMV has a crucial role in triggering the onset of inflammation, resulting in its complication. There is no valid method to distinguish CMV infection requiring therapy from that disappearing without therapy, even with the use of a combination of several modalities. Recently, we evaluated a simple endoscopic feature, no large ulcer, as a predictor of latent CMV infection in active UC patients with positive mucosal viral assay. Patients with welldemarcated large ulcer (ulcerated group) were treated with antiviral therapy and UC therapy. Patients without large ulcer (nonulcerated group) were treated only with UC therapy. In the ulcerated group (n 1⁄4 10), three patients retained active disease at 2 months and underwent colectomy. Although the other seven patients became in remission at 2 months, two of the three patients with CMV DNApositive and two of the four patients with CMV DNA-negative were flareup. Although all patients in the nonulcerated group (n 1⁄4 10) became in remission and maintained remission, two patients were positive for the assay. Although there is no consensus about how to manage active UC patients complicated by CMV reactivation, we believe that patients without large ulcer develop remission without antiviral therapy.

Effective hemostasis with multiple hemoclips and endoloops for GI bleeding

patients, evaluated the size data with at least 3 tests, did not correct for multiple testing, and performed no multivariate analysis to adjust for confounders. The authors themselves state in their ‘‘Methods’’ section that these P values ‘‘should be taken as descriptive only.’’ The O’Neil et al abstract, in fact, does not present any data on these predictors of yield. In the Al-Haddad and Dewitt editorial, they state that they found positive associations among size, number of FNA passes, and yield, but they present no supporting data or statistical analysis. We suggest resolving the perceived ‘‘discrepancies’’ among these reports as follows: The cytological yield of EUS-FNA is 80% or more. The immunohistochemical yield exceeds 50%. The 3 reports disagree as to the importance of size, with our study showing no association with yield, the Sepe et al study showing a negative association, and the Al-Haddad and Dewitt study claiming a positive association. We conclude that size alone should not be used to decide on tissue sampling. The number of FNA passes to take is unclear but needs to be adequate, a determination that should be based upon the expertise of the endoscopist and, whenever possible, feedback from in-room cytology.