Urgent Computed Tomography for Determining the Optimal Timing of Colonoscopy in Patients with Acute Lower Gastrointestinal Bleeding

Transcription

1 ORIGINAL ARTICLE Urgent Computed Tomography for Determining the Optimal Timing of Colonoscopy in Patients with Acute Lower Gastrointestinal Bleeding Satoko Nakatsu, Hiroshi Yasuda, Tadateru Maehata, Masahito Nomoto, Nobuyuki Ohinata, Kosuke Hosoya, Shinya Ishigooka, Shunichiro Ozawa, Yoshiko Ikeda, Yoshinori Sato, Midori Suzuki, Hirofumi Kiyokawa, Hiroyuki Yamamoto and Fumio Itoh Abstract Objective We evaluated the diagnostic performance of computed tomography (CT) as an initial radiologic test for assessing the optimal timing of colonoscopy in patients with acute lower gastrointestinal bleeding (LGIB) and investigated the effectiveness of contrast-enhanced (CE) CT for detecting colonic diverticular bleeding. Methods This was a retrospective study of 1,604 consecutive patients who visited or were referred to St. Marianna University Hospital due to acute LGIB and underwent colonoscopy within three months after presentation between September 2004 and December The clinicopathological data of the subjects were obtained from their medical records. Results Among the 1,604 patients presenting with LGIB, 879 (55%) underwent a CT scan. Elective colonoscopy was considered in cases in which typical colonic wall thickening was observed on CT, suggesting colonic inflammation or malignancy (239 patients; 27%). The diagnoses in the elective cases included ischemic colitis (38%), infectious colitis (8%), inflammatory bowel disease (8%) and malignancy (5%). Urgent colonoscopy was performed after the CT examination in 640 cases (73%). The most common presumptive CT diagnosis was diverticulum (402/640; 63%). Of the 638 patients who underwent CE-CT, diverticula were observed in 346 cases, including 104 cases of extravasation indicating ongoing diverticular bleeding. Among these 104 patients, the site of bleeding was identified in 71 subjects (68%) during colonoscopy. The rate of detection of the bleeding source on colonoscopy was significantly higher in the patients with extravasation on CE-CT than in those without extravasation on CE-CT (68% vs. 20%, respectively; p<0.001). Conclusion Urgent CT is useful for determining the optimal timing of colonoscopy in cases of acute LGIB. CE-CT may be used to depict the presence and location of active hemorrhage and provides useful information for subsequent colonoscopy, especially in patients with diverticular bleeding. Key words: lower gastrointestinal bleeding, urgent colonoscopy, computed tomography, diverticular bleeding (Intern Med 54: , 2015) () Introduction Lower gastrointestinal bleeding (LGIB) is a common gastrointestinal emergency. The incidence of LGIB is increasing, especially among the elderly (1), and diverticula are the most common cause of acute LGIB (2). Other underlying lesions include ischemic colitis, vascular ectasia, hemorrhoids, inflammatory bowel disease, malignancy, infectious colitis, vascular ectasia and rectal ulcers. However, the optimal diagnostic method for detecting LGIB is uncertain. The American Society for Gastrointestinal Endoscopy guidelines Division of Gastroenterology and Hepatology, St. Marianna University School of Medicine, Japan Received for publication March 5, 2014; Accepted for publication July 22, 2014 Correspondence to Dr. Hiroshi Yasuda, hyasuda@marianna-u.ac.jp 553

2 recommend the use of colonoscopy for the early evaluation of severe acute LGIB, although these criteria do not provide a definition of early (3). Hence, determining the optimal timing for colonoscopy remains an area of controversy (4-8). There are a number of other potential management options, including computed tomography (CT) scanning, ultrasound, technetium 99m-red blood cell scintigraphy, mesenteric angiography and various combinations of these modalities (9-12). Of these options, helical, and now multidetector, CT scanning has emerged as the imaging modality of choice in patients with acute LGIB, including that involving various etiologies. Visualizing colonic diverticula with the extravasation of contrast material is a typical CT finding of diverticular bleeding (13, 14), and patients with such findings are frequently referred for emergent colonoscopy. Conversely, colonic wall thickening may indicate a significant pathology, such as ischemic colitis, inflammatory bowel disease or colorectal carcinoma (15, 16). Although ultrasonographic examinations are useful for assessing the degree of colonic thickening, it is difficult to diagnose diverticular and/or rectal bleeding with this modality (11). The aim of this study was therefore to determine the optimal timing for colonoscopy in patients with acute hematochezia by retrospectively assessing the diagnostic performance of CT as the initial diagnostic examination in cases of acute LGIB. In addition, we evaluated the effectiveness of contrast-enhanced (CE) CT for detecting colonic diverticular bleeding. Materials and Methods This was a retrospective study of 2,398 consecutive patients who visited or were referred to St. Marianna University Hospital due to acute hematochezia between September 2004 and December We excluded patients with postpolypectomy bleeding. Of these patients, 1,604 underwent colonoscopy within three months after presenting with hematochezia. Of the 1,604 subjects, 627 required admission and 166 presented with hematochezia during hospitalization for other diseases. The 1,604 patients included 843 men and 761 women, with a mean age of 62 ± 18.1 years. Urgent CT was performed in 879 patients and urgent colonoscopy was performed in 844 patients. Cases in which CT or colonoscopy were performed within 24 hours of presentation with hematochezia were defined as urgent. All other cases were defined as elective. Elective colonoscopy was performed 16.1±16.0 days after the presentation of hematochezia. The subjects clinicopathological data were obtained from their medical records. We analyzed concomitant diseases, antithrombotic drugs, the usefulness of the CT scan findings and the endoscopic results. We followed the guidelines for the use of data in research issued by the Nationwide Registration System for Intractable Diseases and the ethical guidelines for clinical studies endorsed by the Japanese government. The study protocol was approved by the institutional review board of St. Marianna University School of Medicine. For CE-CT, a total of 90 ml iopamidol was power injected intravenously at a rate of 1.5 ml/s. The patients were scanned using either 16- or 64-detector CT. The presence of diverticular bleeding on CE-CT was defined as follows: the visualization of colonic diverticula and extravasation of contrast material into the bowel lumen on arterial phase images and/or abnormal colonic wall enhancement (14). Ischemic colitis was considered presumptive if the following typical findings were present: moderate continuous circumferential thickening of the left colonic wall with moderately long segmental involvement including loss of haustrations without the involvement of the rectum coupled with an appropriate history, such as the sudden onset of abdominal pain followed by diarrhea and hematochezia (15). The colonoscopy procedures were performed by a gastroenterologist on the endoscopy service during normal business hours and by the gastroenterologist on call in cases performed outside of normal hours. A standard video colonoscope (Olympus, Tokyo, Japan) was used for all procedures. Urgent colonoscopy was performed after preparation with colonic lavage [1-2 L polyethylene glycol (PEG)- based lavage solution] over two to four hours. In patients with suspected rectal or anal bleeding, colonoscopy was performed following a glycerin enema or without bowel preparation. A transparent cap was routinely attached on the tip of the endoscope, and intradiverticular blood clots were removed by injecting water using a non-traumatic injection tube to effectively detect exposed vessels (Fig. 1). The diverticula were considered to be the definitive source of hematochezia if active bleeding or the presence of stigmata due to recent hemorrhage were noted during colonoscopy. Diverticula were considered presumptive if diverticulosis was present without any other potential bleeding sites (7). Ischemic colitis was considered to be the definitive source of hematochezia if classic colonoscopic features, including focal submucosal hemorrhage, red-purple blebs, superficial ulceration, necrotic ulceration or dusky purple mucosa, were observed. Hemorrhoids were considered to be the cause of bleeding in the absence of other potential sources of bleeding in cases involving active bleeding or stigmata of hemorrhage. The final diagnosis of the type of colitis was made based on the clinical and laboratory data and colonoscopic and biopsy findings. Elective upper esophagogastroduodenoscopy was performed in some cases in order to rule out an upper source of bleeding. Consequently, eight of the 1,604 patients had an upper GI source of bleeding despite exhibiting a lack of symptoms of upper GI bleeding. Written informed consent for the CE-CT and colonoscopy procedures was obtained from all patients. For the statistical analysis, the χ 2 test was used to compare the data. Differences of p<0.05 were considered to be statistically significant. 554

3 Table 1. Baseline Characteristics n=1,604 mean ages 62±18.1 M/F 843/761 Concomitant disease 528 cardiovascular disease 266 diabetes mellitus 186 celebrovascular disease 110 renal disease 87 rheumatoid disease 62 liver disease 39 antithrombotic drugs/nsaids 327 aspirin 133 warfarin 73 NSAIDs 53 multiple use 94 others 27 Figure 1. Endoscopic view of an exposed vessel in a colonic diverticulum of the ascending colon in a 78-year-old woman with rheumatoid arthritis. A transparent cap was attached to the tip of the endoscope. Removing the intradiverticular blood clot by injecting water using a non-traumatic injection tube was effective for detecting the exposed vessel. Results Baseline characteristics of the patients The demographic and clinical characteristics of the 1,604 patients presenting with hematochezia are shown in Table 1. Overall, 528 patients (33%) had concomitant disease, the most common of which was cardiovascular-related disease. Other concomitant diseases included diabetes, cerebrovascular disease and renal disease. Three hundred and twentyseven patients (20%) were currently taking antithrombotic drugs, including 86 patients taking more than two agents. In addition, 53 patients (3%) were currently taking nonsteroidal anti-inflammatory drugs. Urgent CT for acute LGIB to evaluate the optimal timing of colonoscopy Of the 1,604 patients presenting with LGIB, 879 (55%) underwent CT based on their clinical and laboratory data (638 patients underwent CE-CT; 241 patients underwent non-contrast CT). First, in order to determine which patients were considered for urgent CT, we compared the bleeding sources of LGIB diagnosed on colonoscopy between the subjects with or without a history of prior CT evaluations (Table 2). The most common colonoscopic diagnosis of LGIB after the CT examination was colonic diverticula (423 cases). In contrast, the two most common colonoscopic diagnoses of LGIB among the patients without a prior history of CT examinations were anorectal lesions (170 cases) and inflammatory bowel disease (74 cases). As the initial examination strongly suggested diverticular bleeding in these cases, such as hematochezia without abdominal pain, urgent CT was preferentially considered at our institution. We then evaluated the diagnostic performance of CT as the initial diagnostic examination in the patients with acute LGIB. The presumptive CT diagnoses of LGIB are shown in Table 3. In cases involving typical colonic wall thickening, suggesting ischemic colitis, inflammatory bowel disease or malignancy, but not ongoing bleeding, emergent colonoscopy was not considered in many patients. In addition, elective colonoscopy was considered if the bleeding was thought to be inactive during bowel preparation in cases of colonic diverticula. After the initial CT evaluations, elective colonoscopy was performed in 239 patients (27%), among whom a presumptive CT diagnosis was obtained in 200 cases. Of these 200 patients, the presumptive source of bleeding was confirmed during elective colonoscopy in 71% (141/200) of cases. The most common diagnosis in the elective cases was ischemic colitis. Other diagnoses included infectious colitis, ulcerative colitis and malignancy. Of the 1,604 patients presenting with LGIB, urgent colonoscopy was performed after the initial CT evaluation in 640 cases. The most common presumptive CT diagnosis in these cases was diverticulum (Table 3; 402/640; 63%). Endoscopic hemostasis was performed during urgent colonoscopy in 243 cases (29%), including 154 patients (63%) with diverticular bleeding and 50 patients (20%) with rectal ulcer bleeding. Five cases of diverticular bleeding (2%) were refractory to endoscopic hemostasis, and surgical intervention (four cases) or interventional radiology (one case) was performed in these patients in order to obtain further hemostasis. Accordingly, elective colonoscopy was considered if colonic wall thickening was observed on urgent CT. In contrast, urgent colonoscopy was considered if colonic diverticula were observed on urgent CT in many cases. Contrast-enhanced CT to detect colonic diverticular bleeding As mentioned above, diverticula are the most common cause of acute LGIB. We therefore focused on the effectiveness of CE-CT for detecting colonic diverticular bleeding. Of the 638 patients who underwent CE-CT, diverticula were observed in 346 cases, including extravasation indicating ongoing hemorrhage in 104 cases (30%). Of these 104 patients, the site of bleeding was identified during colonoscopy 555

4 Table 2. Colonoscopic Diagnosis in Patients with or without Prior CT Evaluation With Urgent CT evalution (n=879) Without CT evaluation (n=725) Colonic diverticula Ano-rectal lesion Ischemic colitis Inflammatory bowel disease Malignancy Colitis Others Non-diagnositic Table 3. Presumptive CT Diagnosis in Patients with Acute LGIB (n=638) Considered elective CS (n=239) Consided urgent CS (n=640) Ischemic colitis 95 (85) 31 (30) Infectious colitis 31 (16) 28 (10) Malignancy 6 (6) 16 (16) Inflammatory bowel disease 8 (8) 5 (5) Diverticula without active bleeding 49 (23) 300 (266) Diverticula with extravasation 2 (0) 102 (98) Ano-rectal lesion 0 39 (36) abnormal colonic wall enhancement 7 (1) 26 (12) Extravasation without diverticula 1 (1) 1 (1) Others 1 (1) 9 (5) Non-diagnostic 39 (19) 83 (24) The number that the diagnosis was confirmed during colonoscopy is shown in (). in 71 cases (68%), among which the bleeding points noted on CE-CT and colonoscopy were identical in 66 cases (93%, Fig. 2). The rate of detection of the bleeding source on subsequent colonoscopy was significantly higher among the patients with extravasation on CE-CT than among those without extravasation on CE-CT (68% vs. 20%, respectively; p<0.001; Table 4). These results indicate that the CE- CT examinations were effective in identifying bleeding diverticula prior to colonoscopy. Discussion In the present study, we found CT to be a useful initial radiological test for determining the optimal timing of colonoscopy in patients with acute LGIB. Emergent colonoscopy should be considered in cases of acute LGIB associated with extravasation of contrast material into the bowel lumen from the colonic diverticula, as the rate of detection of the bleeding source on urgent colonoscopy was very high in such cases in the present study. In contrast, elective colonoscopy should be considered in patients with acute LGIB presenting with colonic wall thickening, suggesting colonic inflammation or neoplasms. The optimal timing for colonoscopy in cases of acute LGIB remains an area of controversy. In a trial by Green et al. (6) in which the authors assessed the efficacy of colonoscopy following bowel preparation performed within 12 hours vs. elective colonoscopy conducted within 74 hours, a definitive bleeding source was identified in 42% of the patients undergoing urgent colonoscopy, compared to 22% of those undergoing elective colonoscopy. Conversely, in a trial of colonoscopy performed within 12 hours vs. that performed within hours, Laine et al. (7) found no significant differences in outcomes, including the diagnosis. However, these reports merely compared the effectiveness of colonoscopy between urgent or elective timing in patients treated without an initial CT evaluation. In contrast, in the present study, elective colonoscopy was considered after a CT examination in 27% of the patients with acute LGIB. Therefore, the CT findings were used to direct the clinical management of these patients. Urgent CE-CT is reportedly a useful diagnostic method for assessing diverticular disease (13, 14, 17). In the present study, extravasation of contrast medium was observed in 30% of the patients on CE-CT, and bleeding diverticula were identified and treated with colonoscopy in 68% of the patients with diverticular bleeding. The ability to detect bleeding diverticula often requires the careful removal of clotted blood from each diverticulum throughout the colon during the colonoscopy procedure. In the present study, the rate of detection of the bleeding source on colonoscopy was significantly higher among the patients with extravasation on CE-CT than among those without extravasation on CE-CT. Hence, CE-CT provides useful information for subsequent endoscopy, especially in patients with diverticular bleeding. In addition, there are several reports regarding the usefulness of CE-CT for diagnosing various other origins of gastrointestinal bleeding (18). Therefore, CE-CT should be consid- 556

5 a b Figure 2. (a) Extravasation of contrast medium (arrow) in the ascending colon lumen. (b) Endoscopic view of active bleeding from a colonic diverticulum of the ascending colon in a 63-year-old woman. Table 4. Colonic Diverticula Diagnosed on CE-CT in Patients with LGIB (n=346) Diverticula with extravasation 104 Bleeding point identified during colonoscopy 71 (68%) Diverticula without active bleeding 242 Bleeding point identified during colonoscopy 49 (20%) p<0.001 ered after clinical assessments have been conducted and stabilization has been achieved in patients with acute LGIB, unless there are contraindications to the use of intravenous contrast materials. Urgent colonoscopy provides not only a diagnostic method, but also a therapeutic maneuver for treatment (19). In the present study, 63% of the endoscopic hemostatic cases involved diverticular bleeding. Our experience with endoscopic treatment for diverticular bleeding was associated with a treatment success rate of 98%, as only five patients required further surgical or radiological intervention. Diverticular bleeding is present in 3-15% of individuals with diverticular disease (20). Such bleeding is thought to be due to the rupture of altered arteriosclerotic diverticular vessels. Although diverticular hemorrhage ceases spontaneously in many cases, delayed recurrent bleeding may occur in 16-38% of patients. As shown in a previous report (21), patients with diverticular bleeding often exhibit various comorbidities and have a history of antithrombotic drug use; thus, surgical intervention sometimes carries a very high risk. Therefore, identifying the source of bleeding prior to surgery using CE-CT may result in a less invasive urgent colonoscopy procedure and more effective hemostasis. Compared with elective colonoscopy, early colonoscopy has been shown to have particular promise in patients with diverticular bleeding in several reports. For example, in a study by Jensen et al., patients undergoing endoscopic hemostasis for diverticular stigmata demonstrated significantly lower rates of re-bleeding and surgery compared to that of historical controls who did not receive endoscopic treatment (5). In summary, performing urgent CE-CT in patients with acute LGIB is feasible and may be used to depict the presence and location of active hemorrhage. Colonoscopic therapy, especially for acute diverticular hemorrhage, is also safe and effective and should be considered after conducting a proper CT evaluation. The authors state that they have no Conflict of Interest (COI). References 1. Lanas A, Garcia-Rodriguez LA, Polo-Tomas M, et al. Time trends and impact of upper and lower gastrointestinal bleeding and perforation in clinical practice. Am J Gastroenterol 104: , Longstreth GF. Epidemiology and outcome of patients hospitalized with acute lower gastrointestinal hemorrhage: A population based study. Am J Gastroenterol 92: , Davila RE, Rajan E, Adler DG, et al. ASGE Guideline: the role of endoscopy in the patient with lower-gi bleeding. Gastrointest Endosc 62: , Chaudhry V, Hyser MJ, Gracias VH, Gau FC. Colonoscopy: the initial test for acute lower gastrointestinal bleeding. Am Surg 64: , Jensen DM, Machicado GA, Jutabha R, Kovacs TOG. Urgent colonoscopy for the diagnosis of severe diverticular hemorrhage. 557

6 N Engl J Med 342: 78-82, Green BT, Rockey DC, Portwood G, et al. Urgent colonoscopy for evaluation and management of acute lower gastrointestinal hemorrhage: a randomized controlled trial. Am J Gastroenterol 100: , Laine L, Shah A. Randomized trial of urgent vs. elective colonoscopy in patients hospitalized with lower GI bleeding. Am J Gastroenterol 105: , El-Tawil AM. Pros and cons of colonoscopy in management of acute lower gastrointestinal bleeding. World J Gastroenterol 18: , Yoon W, Jeong YY, Shin SS, et al. Acute massive gastrointestinal bleeding: detection and localization with arterial phase multidetector row helical CT. Radiology 239: , Martí M, Artigas JM, Garzón G, Álvarez-Sala R, Soto JA. Acute lower intestinal bleeding: feasibility and diagnostic performance of CT angiography. Radiology 262: , Yamaguchi T, Manabe N, Hata J, Tanaka S, Haruma K, Chayama K. The usefulness of transabdominal ultrasound for the diagnosis of lower gastrointestinal bleeding. Aliment Pharmacol Ther 23: , Zink SI, Ohki SK, Stein B, et al. Noninvasive evaluation of active lower gastrointestinal bleeding: comparison between contrastenhanced MDCT and 99mTc-labeled RBC scintigraphy. Am J Roentgenol 191: , Kominami Y, Ohe H, Kobayashi S, et al. Dynamic computed tomography is useful for the diagnosis and colonoscopic treatment of colonic diverticular bleeding. Nihon Shokakibyo Gakkai Zasshi (Japanese Journal of Gastroenterology) 108: , 2011 (in Japanese, Abstract in English). 14. Obana T, Fujita N, Sugita R, et al. Prospective evaluation of contrast-enhanced computed tomography for the detection of colonic diverticular bleeding. Dig Dis Sci 258: , Thoeni RF, Cello JP. CT imaging of colitis. Radiology 240: , Wolff JH, Rubin A, Potter JD, et al. Clinical significance of colonoscopic findings associated with colonic thickening on computed tomography. Is colonoscopy warranted when thickening is detected? J Clin Gastroenterol 42: , DeStigter KK, Keating DP. Imaging update: acute colonic diverticulitis. Clin Colon Rectal Surg 22: , Lee S, Welman CJ, Ramsay D. Investigation of acute lower gastrointestinal bleeding with 16- and 64-slice multidetector CT. J Med Imaging Radiat Oncol 53: 56-63, Hokama A, Kishimoto K, Kinjo F, Fujita J. Endoscopic clipping in the lower gastrointestinal tract. World J Gastrointest Endosc 1: 7-11, McGuire HH. Bleeding colonic diverticula. A reappraisal of natural history and management. Ann Surg 220: , Yamada A, Sugimoto T, Kondo S, et al. Assessment of the risk factors for colonic diverticular hemorrhage. Dis Colon Rectum 51: , The Japanese Society of Internal Medicine 558