Present role of intraoperative enteroscopy in small bowel bleeding: A tertiary center experience

The aim of this study was to assess the diagnostic yield and outcomes of intraoperative enteroscopy (IOE) in patients with overt small bowel bleeding in the era of balloon‐assisted enteroscopy and capsule endoscopy. We retrospectively reviewed the medical records of patients with small bowel bleeding who underwent IOE from January 2005 to April 2016 in a tertiary medical center. A total of 18 patients with overt small bowel bleeding who underwent a total 18 IOE procedures were included. The mean amount of blood transfusion was 35.7 units (SD, 21.4 units), and 11 patients had hypovolemic shock. The diagnostic yield of the IOE procedures was 94.4% (17/18). The most common pathologic diagnosis was ulcer. The rebleeding rate was 44.4% (8/18), and the overall survival rate was 61.1% (11/18). IOE remains a valuable tool for diagnosing overt small bowel bleeding. The usage of IOE might potentially be limited due to critical clinical conditions in the era of deep enteroscopy. A specific caution should be taken in performing IOE due to the high morbidity of the procedure.


| INTRODUCTION
Small bowel lesion has accounted for approximately 5% to 10% of unexplained gastrointestinal bleeding cases. 1 Intraoperative enteroscopy (IOE) has been practiced for surveying small bowel lesions, which was first developed in the mid-1970s. 2IOE was the first choice of clinicians to explore bleeding lesions within the entire small bowel.With the advance of technology, wireless capsule endoscopy (CE) and deep enteroscopy (DE) were invented in 2000 and 2001, respectively.DE techniques have therapeutic and diagnostic ability.Presently available DE methods include single-balloon enteroscopy (SBE), spiral enteroscopy (SE), and double-balloon enteroscopy (DBE).These methods are belonging to device-assisted enteroscopy (DAE), including balloon-assisted enteroscopy (BAE).4][5] The diagnostic yield of CE in overt small bowel bleeding has been reported to range from 73.3% to 92.3%, [6][7][8][9] and the superiority and utility of DBE for the diagnosis and treatment of small-bowel disorders has been demonstrated in several studies with diagnostic yields varying between 43% and 80%. 10 Moreover, BAE enables not only observation of the entire small bowel but also therapeutic intervention at most locations within the small bowel, so it is now considered the standard technique for endoscopic management of small bowel lesions.
In the past decades, CE and DE have been increasingly employed for diagnosing small bowel disorders with satisfactory outcomes.Despite of the advance in enteroscopy, patients with previous surgeries and bowel adhesions are not considered for CE or BAE. 1 In those cases, IOE could be of use to detect small bowel lesions along the whole lumen, allowing for operation and/or endoscopic treatment.The aim of this study, therefore, was to evaluate the diagnostic yield and outcomes of IOE in patients with overt small bowel bleeding in the era of BAE.

| Patients
We retrospectively reviewed the medical records of all patients who underwent IOE from January 2005 to April 2016 at Chang Gung Memorial Hospital.Of 19 patients who underwent IOE, 1 patient with anemia without overt small bowel bleeding was excluded.A total of 18 patients with overt small bowel bleeding and wanted to receive IOE for immediate management were thus included.The Chang Gung Medical Foundation Institutional Review Board approved the study (IRB number: 20150399B0) and waived the requirement for informed consent due to the study's retrospective design.The patients' characteristics, including age, gender, baseline laboratory data prior to IOE, American Society of Anesthesiologists (ASA) classification, co-morbidities, drug use, hypovolemic shock, amount of blood transfusion prior to IOE and follow-up time were analyzed.The preoperative diagnostic procedures for the patients including esophagogastroduodenoscopy (EGD), colonoscopy, angiography, 99m Tc-labeled RBC scintigraphy, computed tomography, CE, and BAE for the patients were also documented (Table 1).The surgical specimens from the patients were subjected to pathological examination.

| Definition
Overt small bowel bleeding was defined as patients suffered from either melena or hematochezia, or hematemesis, with the bleeding focus in the small intestine.Obscure gastrointestinal bowel bleeding was defined as patients were not noted to have a source of bleeding after performance of standard upper and lower endoscopic examination, small bowel evaluation with video capsule endoscopy and/or enteroscopy, and radiographic testing. 1 The diagnostic yield for IOE was documented on the basis of positive IOE.Rebleeding was defined as the recurrence of hematochezia or melena or a declining hemoglobin count from baseline of more than 2 g/dL and/or the transfusion of one or more units of packed red blood cells with overt bleeding after IOE. 11

| Procedures
All of the IOE procedures were conducted in collaboration of surgeons and gastroenterologists.The techniques used, including transoral approach, transanal approach or enterotomy approach, were chosen according to the findings of preoperative exams.If the bleeding focus was detected, surgical management was performed for hemostasis depending on the given patient's clinical condition.

| Statistical analysis
The patients' demographic and baseline clinical data, endoscopic findings, and surgical management were summarized by descriptive statistics.Continuous variables are presented as means and standard deviations and categorical variables are presented as percentages.The diagnostic yield was calculated as follows: the number of patients with a positive finding of IOE was divided by the total number of patients who underwent IOE.All statistical tests were two-sided and were performed using IBM Statistical Product and Service Solutions, version 22 (IBM, Armonk, New York, USA).

| Patients
Eighteen patients (11 males; mean age: 62.9 ± 20.5 years old) underwent 18 IOE for the management of overt small bowel bleeding during the study period.Among the 18 procedures, 17 were performed by enterotomy approach, and 1 were conducted by enterotomy and transoral approach.The demographic and baseline clinical characteristics of the patients are presented in Tables 2  and 3.The mean hemoglobin level at the initial episode of small bowel bleeding was 8.2 g/dL (SD, 1.6 g/dL).The use of drugs was reported as follows: 3 patients had longterm aspirin or clopidogrel use, 7 were taking nonsteroidal anti-inflammatory drugs (NSAIDs), and 6 were taking steroid.Hypovolemic shock occurred in 11 patients.The mean amount of blood transfusions was 35.7 units (SD, 21.4 units).

| Enteroscopic findings and pathologic results
The IOE findings, the pathologic results and surgical management are shown as in Table 4.The diagnostic yield was 94.4% (17 out of 18 patients).Small bowel resections were performed in all the patients.In four patients presenting fresh blood, the active bleeding region was specifically localized by IOE (at the jejunum in two patients and at the ileum in two patients), and no episode of rebleeding occurred after the operations.In one patient, a gross jejunal diverticulum was found by the surgeon and IOE was also performed to detect any other bleeding sources.The pathological findings were as follows: ulcer (n = 10; one was diagnosed as small intestinal T-cell lymphoma; Figures 1 and 2), angiodysplasia (n = 5), diverticulum (n = 3), and mucosal hemorrhage (n = 1).The IOE findings were different than the pathologic results in two of the patients.IOE findings in one 72-year-old male with liver cirrhosis and renal insufficiency is angiodysplasia, but pathological result is ulcer.Another 77-year-old female with coronary artery disease and the IOE findings is ulcer, but pathological result is angiodysplasia.In addition, BAE was performed successfully in two patients, but one peroral BAE was failed to insert into the jejunum in the patient with small intestinal lymphoma due to prior operation for a bowel perforation.Between the two patients with endoscopic diagnosis by DBE, one received argon plasma coagulation (APC) for angiodysplasia and the other received a clip marking on the small bowel segment with fresh blood followed by operation.Only one patient underwent CE.The diagnostic impression in three patients who underwent CE or successful BAE were confirmed by IOE.

| Rebleeding rate and mortality rate
Eight patients (44.4%) experienced rebleeding after operations during follow-up period in an average of 15.6 months.Of the eight patients, the most common pathologic results in the patients with rebleeding were small bowel ulcer (62.5%, 5/8), followed by multiple angiodysplasia (25%, 2/8) and diverticulum (12.5%, 1/8).Among the patients with small bowel ulcer and rebleeding, two patients had bowel leak, one was still under extracorporeal membrane oxygenation until 9 days after the operation, one was diagnosed with small intestinal T-cell lymphoma, and one had multiple ulcers treated by a secondary operation, after which there was no rebleeding.The overall survival rate was 61.1% (11 out of 18 patients).Four patients died of multiple organ failure due to postoperative pneumonia, and three patients died of progressive intra-abdominal infection.

| DISCUSSION
In this study, IOE had a high diagnostic yield of 94.4% in patients with overt small bowel bleeding; however, the mortality and morbidity associated with the procedure should be of considerable concern due to the procedure's invasive nature.We also found that the main diagnosis underlying small bowel bleeding in this study was ulcer.The variation in diagnosis between this study and previous reports probably implies that the group of patients selected to undergo IOE will gradually come to consist primarily of those patients with more critical small bowel ulcer bleeding since BAE has been available in our hospital since 2003. 12Meanwhile, the associated outcomes of IOE, such as rebleeding, morbidity and mortality should be investigated in more large and well-designed studies.
The diagnostic yield of IOE has been shown various ranging from 58.3% to 100%, [13][14][15][16][17][18] with two most common findings namely small bowel angiodysplasia and small bowel ulcer, respectively. 19Our finding of diagnostic yield of 94.4% with IOE in patients with small bowel bleeding was comparable to that of previous studies.In our study, the major diagnosis in this study was benign small bowel ulcer (50%, 9/18) while the second most common diagnosis was small bowel angiodysplasia (27.8%, 5/18).Two review articles 19,20 reported small bowel ulcer accounting for about 14% of cases, while one recent study 21 found small bowel ulcer in 10.4% of patients with small bowel bleeding and anemia.Of the patients with small bowel ulcer in present study, eight patients were treated with antiplatelet agents, anticoagulants, steroids, and NSAIDs, and five patients were under mechanical ventilation for more than 2 days.An explanation for the disparity in diagnosis is that medications prescribed stand a high risk of inducing small bowel ulcers, whereas stress ulceration occurred in the four patients who had serious illness at initial admission, including decompensated heart failure, neutropenic fever, traumatic pneumothorax due to a traffic accident, and respiratory failure caused by influenza viral pneumonia.
We found that the patients treated by surgical resection without IOE procedures had poor outcomes.Hartmann et al. presented a study that enrolled 47 patients who underwent IOE 22 ; endoscopic treatment using APC was performed in 17 patients with angioectatic lesions.Jakobs et al. published results for 81 patients with obscure gastrointestinal bleeding who underwent IOE. 16In that study, 20 patients with angiodysplasia were also managed by APC.Kop ačov a et al. diagnosed arteriovenous malformations in 12 patients with obscure gastrointestinal bleeding, and those lesions were managed by electrocautery with or without surgical resection. 23In a series reported by Douard et al. 22.2% of patients (2/18) with angiodysplasia and Dieulafoy's lesion underwent endoscopic treatment. 15To sum up briefly, the specific etiology, such as angiodysplasia, was usually treated by intraoperative endoscopic management.On the other hand, a study of 100 patients with small bowel lesions investigated by DBE disclosed that endoscopic clipping was performed in 2 of 51 patients with small bowel ulcer, while the others were managed by medication, surgery, transarterial embolization or observation. 24In addition, the predominant etiology of small bowel bleeding in our study was ulcer, and this might be one reason there was no endoscopic intervention by IOE in our study.
6][27] Our finding of rebleeding rate after IOE was similar to those of previous studies.Most of the patients who experienced episodes of rebleeding had an outcome of death.The causes of high mortality rate are postulated to be multifactorial.The patients with major co-morbidities were likely to cause result in death after IOE, including episodes of hypovolemic shock and rebleeding, heavy use of drugs, and greater amounts of blood transfusions (Table 2).It is suggested that these patients had higher preoperative risk and greater likelihood to progress to multiple organ failure.Multiple organ failure due to recurrent bleeding or sepsis was also reported as the most common cause of mortality. 19herefore, patients should be selected carefully for IOE and the risk of the procedure should be clearly explained before the procedures.
This study had several limitations.Two major limitations of the study were its retrospective nature and underpowered sample size.There was thus a high possibility that confounders and selection bias could have interfered with the outcomes of this study.Secondary concern was that CE and/or BAE was performed prior to IOE in only a small number of cases, an approach which departed from the current guideline recommendations.The reason for this could be classified into three groups: (a) massive bleeding not allowing sufficient time for CE; (b) safety issues relating to patient history, such as pacemaker use or severe adhesion due to a previous operation; and (c) the cost of CE and BAE not being covered by national health insurance until 2017.Even so, this study presents some important characteristics and outcomes regarding the changing role of IOE.
The results suggest that IOE represents a valuable diagnostic tool in patients with moderate to severe overt small bowel bleeding with a substantial role in guiding surgical management.Taken high mortality, surgeons and endoscopists must evaluate patients carefully and proceed with caution to avoid using this high-risk procedure unnecessarily, such that IOE will gradually be performed only on more seriously ill patients.
T A B L E 1Abbreviations: CT, computer tomography; RBC, red blood cells.HUANG ET AL.
Baseline characteristics of the patients with overt small bowel bleeding.Individual patient demographics.
T A B L E 2Note: Data are presented as mean ± SD or number (%).The categorical variables were assessed via Fisher's exact test.Continuous variables were analyzed with nonparametric Mann-Whitney U-test.All statistical tests were two-sided.A p < .05 was considered statistically significant.Abbreviations: ASA, American Society of Anesthesiologists; BUN, blood urea nitrogen; CVA, cerebrovascular accident; CKD/ESRD, chronic kidney disease/ end stage renal disease; INR, internaktional normaized ratio; IOE, intraoperative enteroscopy; NSAIDs, nonsteroidal anti-inflammatory drugs.T A B L E 3Abbreviations: CT, computed tomography; CTA, computed tomography angiography; EGD, esophagogastroduodenoscopy; RBC, red blood cells.aMortality.bHypovolemicshock.T A B L E 4 Finding, pathologic report, and surgical management of 18 patients subjected to 18 intraoperative enteroscopy procedures.Note: Data are presented as number.