OBJECTIVES: Endoscopist fatigue potentially impacts colonoscopy. Fatigue is difficult to quantitate, but polyp detection rates between non-fatigued and fatigued time periods could represent a surrogate marker. We assessed whether timing variables impacted polyp detection rates at a busy tertiary care endoscopy suite. METHODS: Consecutive patients undergoing colonoscopy were retrospectively identified. Indications, clinical demographics, pre-procedural, and procedural variables were extracted from chart review; colonoscopy findings were determined from the procedure reports. Three separate timing variables were assessed as surrogate markers for endoscopist fatigue: morning vs. afternoon procedures, start times throughout the day, and queue position, a unique variable that takes into account the number of procedures performed before the colonoscopy of interest. Univariate and multivariate analyses were performed to determine whether timing variables and other clinical, pre-procedural, and procedural variables predicted polyp detection. RESULTS: During the 4-month study period, 1,083 outpatient colonoscopy procedures (57.5±0.5 years, 59.5% female) were identified, performed by 28 endoscopists (mean 38.7 procedures/endoscopist), with a mean polyp detection rate of 0.851/colonoscopy. At least, one adenoma was detected in 297 procedures (27.4%). A 12.4% reduction in mean detected polyps was detected between morning and afternoon procedures (0.90±0.06 vs. 0.76±0.06,P=0.15). Using start time on a continuous scale, however, each elapsed hour in the day was associated with a 4.6% reduction in polyp detection (P=0.005). When queue position was assessed, a 5.4% reduction in polyp detection was noted with each increase in queue position (P=0.016). These results remained significant when controlled for each individual endoscopist. CONCLUSIONS: Polyp detection rates decline as time passes during an endoscopist's schedule, potentially from endoscopist fatigue. Queue position may be a novel surrogate measure for operator fatigue.

This issue includes two studies focused on colonoscopy quality. Both examine endoscopist fatigue as a possible quality factor that influences polyp detection rate. Quality is an essential element of strategies to reduce mortality from colorectal cancer, and one possible factor influencing quality is endoscopist fatigue. In the first study, the investigators assessed whether the position of a patient in the endoscopy schedule impacted polyp detection rates. In this retrospective study, clinical demographics and preprocedural and procedural variables were extracted from consecutive patients undergoing colonoscopy. Three scheduling variables were assessed as surrogate markers for endoscopist fatigue: morning versus afternoon procedures, start times throughout the day, and queue position, a unique variable that accounted for the number of procedures performed before the colonoscopy of interest. During the 4-month study period, 1,083 outpatient colonoscopy procedures were identified, performed by 28 endoscopists with a mean polyp detection rate of 0.85 per colonoscopy. At least 1 adenoma was detected in 297 procedures (27%). A 12% reduction in the mean number of polyps detected was observed between morning and afternoon procedures. Moreover, for each hour that elapsed, there was a 4.6% reduction in polyp detection, and when queue position was assessed, a 5.4% reduction in polyp detection was noted with each increase in queue position. This study illustrates that polyp detection rates decline with time over the course of an endoscopist’s workday, possibly because of fatigue. See page 1457

Background & Aims: An inadequately cleansed colon can lead to missed lesions, repeat procedures, increased cost, and complications from colonoscopy. Because obesity, with its known link to colorectal neoplasia, might be associated with inadequate bowel cleansing, we investigated the impact of increased body mass index (BMI) on quality of bowel preparation at colonoscopy. Methods: All colonoscopy procedures performed at a tertiary referral center during a 4-month period were evaluated. Bowel preparation was assigned a unique composite outcome score that took into account a subjective bowel preparation score, earlier recommendation for follow-up colonoscopy as a result of inadequate bowel preparation, and the endoscopises confidence in adequate evaluation of the colon. Univariate and multivariate logistic regression analyses were performed to identify the role of BMI in predicting an inadequate bowel preparation. Results: During the study period, 1588 patients (59.1% female; mean age, 57.4 +/- 0.34 years) fulfilled inclusion criteria. An abnormal BMI (>= 25) was associated with an inadequate composite outcome score (P = .002). In multivariate logistic regression analyses, both BMI >= 25 (P = .04) and >= 30 (P = .006) were retained as independent predictors of inadequate bowel preparation. Each unit increase in BMI increased the likelihood of an inadequate composite outcome score by 2.1%. Additional independent predictors of inadequate preparation exponentially increased the likelihood of an inadequate composite outcome score; 7 additional risk factors identified 97.5% of overweight patients with an inadequate composite outcome score. Conclusions: Obesity is an independent predictor of inadequate bowel preparation at colonoscopy. The presence of additional risk factors further increases the likelihood of a poorly cleansed colon.

Purpose: A poorly prepared colon can lead to missed lesions and repeated procedures that increase cost and complication risk. The aim of the study was to identify predictive factors of poor colon prep, including obesity, which increases risk of neoplasia. Methods: Retrospective review of 1815 reports of inpatient (IP) and outpatient (OP) colonoscopies performed from June to October 2007 at a tertiary referral center. Prior colectomy, repeat colonoscopy, incomplete reports were excluded and 1588 reports analyzed. The endoscopist's inability to adequately evaluate the entire colonic mucosa and request for earlier than expected repeat colonoscopy (ACG, ASGE guidelines) were used to determine a poor or inadequate prep (PIP). Results: Good to excellent prep quality was reported in 400 (25.2%) patients and PIP in 627 (39.5%) patients. Patients with PIP were older (59.57 ± 13.5 yrs) and had higher weight (86.19 ± 23.9 kg) and BMI (29.74 ± 7.84 kg/m2) (P < .01). In univariate analysis, male gender, inpatient status (IP), Polyethylene glycol based lavage (PEG), smoking, lack of regular alcohol consumption, diabetes mellitus (DM), hypertension, coronary artery disease (CAD) increased BMI, depression, mental retardation (MR) and use of anti-depressant medications or narcotics were predictive of PIP. However, in multivariate logistic regression analysis, only male gender (OR = 0.75, P= 0.02), IP status (OR = 1.92, P= 0.002), smoking (OR = 1.62, P= 0.003), lack of regular alcohol consumption (OR = 0.68, P= 0.0003), DM (OR = 1.64, P= 0.0006), MR (OR = 3.54, P= 0.0004), narcotic use (OR = 3.04, P < 0.0001), and PEG (OR = 0.53, P= 0.0004) were predictors of poor bowel preparation. Modeling for composite outcome using univariate analysis: older age, male gender, IP status, smoking, lack of regular alcohol consumption, DM, hypertension, CAD, increased BMI, depression, obstructive sleep apnea, MR and consumption of anti-depressant medications or narcotics were predictive of PIP. However, in multivariate logistic regression analysis, only older age (OR = 1.009, P= 0.01), male gender (OR = 0.71, P= 0.002), IP status (OR = 1.55, P= 0.008), smoking (OR = 1.35, P= 0.01), lack of regular alcohol consumption (OR = 0.7, P= 0.01), increased BMI (OR = 1.02, P= 0.004), MR (OR = 2.17, P= 0.03), use of antidepressant medication (OR = 1.69, P= 0.001), and narcotic use (OR = 2.1, P= 0.001) were predictors of PIP. Conclusion: Poorly prepped colons reduce diagnostic yield. Aborted and repeat procedures greatly increase the cost of colonoscopy. Prior identification and aggressive bowel cleansing of patients with predictive factors of PIP can lead to improved diagnostic yield and cost savings. This needs to be verified in a prospective study.

Purpose: CT colonoscopy is approved as an alternative method of screening for colon polyps and is used frequently as a salvage method after an incomplete colonoscopy. The effectiveness of this approach has not been well investigated. We evaluated the utility of CT colonoscopy (CTC) as a salvage screening method after failed colonoscopy at a tertiary care center with easy access to radiology. Methods: Data from all imaging studies ordered following colonoscopy was collected from June to October 2007. Median follow up period was 310 days. Results of the imaging studies, outcome, effect on clinical care and alternative findings were retrieved and followed by data analysis. Results: From a total of 1628 colonoscopy reports, 50 patients had imaging studies requested after the cecum was not reached (3% of total); 46 after their first and 4 after their second colonoscopy. Technical difficulty in reaching the cecum (79%) was the most common reason for requesting an imaging study. Proximal colon was reached in 82.3% of failed colonoscopies that led to a request for CT colonoscopy (CTC). CTC was ordered in 29 pts; in 21 a combination of abdominal CT, barium enema and MR enterography comprised the remaining requests. Of the total procedures (50) requested, 14 (28%) were not performed, including 41% of requested CTC. Imaging studies in 36 patients revealed a total of 37 luminal and 42 non-luminal findings. No polyps were reported by CTC. Only one colon polyp was identified by MRI that on repeat colonoscopy was removed and found to be inflammatory. In multivariate logistic regression analysis, while controlling for other variables, female gender (OR = 3.04, P= 0.0064), and indication for colonoscopy (bleeding vs. screening) (OR = 2.73, P= 0.025) were independently predictive of request for an imaging study after failed colonoscopy. Conclusion: When CTC was requested following an incomplete colonoscopy, it was often not performed. Following a failed colonoscopy, the yield of CTC or other imaging studies in detecting neoplasia was poor. Large population based studies would be necessary to evaluate the effectiveness of this approach in the general population. The high rate of unperformed CTC and its poor yield questions the value of this procedure in the setting of an incomplete colonoscopy.

We describe an exceedingly rare case of severe gastritis that was temporally associated with primary Epstein-Barr virus (EBV) infection. The patient was a 59-year-old immunocompetent man who presented with intermittent fever of unknown origin and epigastric pain for 18 days. A computed tomographic scan of the abdomen showed diffuse thickening of the gastric wall and esophagogastroduodenoscopy revealed numerous ulcers in the stomach. Histologic examination of gastric biopsies showed a dense and diffuse atypical lymphoid infiltrate in the lamina propria with erosions and focal lymphoepithelial lesions. No lymphoid follicles or Helicobacter microorganisms were identified. Immunohistochemical studies demonstrated the lymphoid infiltrate to consist of mixed T and B cells. Immunoglobulin heavy chain gene arrangement analysis showed a polyclonal pattern. The plasma cells present in the biopsies exhibited no light chain restriction as determined by in situ hybridization. Concurrent clinical work-up revealed peripheral lymphocytosis with atypical lymphocytes and positive serum IgM antibody to EBV capsid antigen in the absence of IgG antibody. These findings indicated that the gastric abnormalities were related to primary EBV infection as the predominant manifestation of infectious mononucleosis. This was further confirmed by subsequent in situ hybridization showing numerous EBV-positive lymphocytes in the gastric mucosa. The patient's symptoms were spontaneously resolved with only supportive treatment. A follow-up endoscopy 2 months later showed completely normal gastric mucosa and he remained well with no gastrointestinal complaints for 2 and a half years. This case illustrates the importance of a high index of suspicion to avoid misdiagnosis of gastric lymphoma that requires more aggressive therapies.

BACKGROUND & AIMS:Amyloidosis is characterized by the pathologic deposition of specific proteins throughout the body. Gastrointestinal involvement with amyloid associated with plasma cell dyscrasias (AL type amyloidosis) is common, but systematic description of the condition is lacking. The aim of this investigation was to characterize the clinical presentation, endoscopic findings, and histopathologic correlates in a series of patients with systemic AL amyloidosis of the luminal gastrointestinal tract.METHODS:Eligible patients were identified by interrogating the histopathology database of our institution during a 14-year time period. Medical record, histopathologic, and laboratory data were collected, analyzed, and correlated with endoscopic findings.RESULTS:Nineteen patients with systemic AL amyloidosis of the luminal gastrointestinal tract were identified. Gastrointestinal symptoms or signs related to amyloid involvement were noted in 95% of patients; abdominal pain, change in bowel habits, overt gastrointestinal bleeding, and complaints related to altered motility were the predominant presentations. Endoscopic abnormalities were found in nearly three fourths of patients, including ulcerations and submucosal hematomas. When gastrointestinal bleeding was the presenting symptom, submucosal hematomas were a common finding during endoscopic evaluation.CONCLUSIONS:AL type amyloidosis of the luminal gastrointestinal tract is a rare disease that presents with common, nonspecific complaints. The endoscopic detection of a submucosal hematoma in the setting of gastrointestinal bleeding in patients with plasma cell dyscrasias should raise suspicion for the disease.

A 61-yr-old liver transplant recipient presented with abdominal cramping and nonbloody diarrhea resulting in orthostasis. Multiple ulcerations throughout the colon were seen during endoscopy, and biopsies from the ulcer edges revealed histoplasmosis. Treatment with a course of itraconazole improved the diarrhea. The patient later presented with pericarditis and symptomatic pleural effusions, the latter of which was confirmed to be a result of disseminated histoplasmosis. Treatment with amphotericin B led to resolution. Histoplasmosis should be considered in liver transplant patients with diarrhea and large ulcers in the colon. The presence of disseminated histoplasmosis should be ruled out once colonic histoplasmosis has been diagnosed.

Despite advances in management, recurrent bleeding and mortality remain high after variceal upper gastrointestinal bleeding (UGIB). We reviewed our data to determine predictors of a poor outcome in patients treated with initial EVL. Methods: All cirrhotic patients with definitive UGIB and endoscopic evidence of varices were eligible for inclusion. Recurrence of bleeding within 30 days was considered rebleeding, while bleeding presentation after 30 days was counted as a distinct episode. Presentation, clinical course, laboratory and endoscopic data were extracted from review of inpatient charts and endoscopy reports. Results: The study group consisted of 136 patients (53±1 yr, 40F/96M) with 156 distinct EVL sessions for acute variceal UGIB over a 3-year period (esophageal varices=153, gastric varices=3). EVL was performed when active variceal bleeding was encountered (35 instances), stigmata of recent bleeding were identified (114 instances), or large varices were seen in the absence of other bleeding sources (7 instances). Initial EVL success rate was 96%, but 38 patients subsequently rebled. Of these, 2 died before endoscopy could be attempted, and 36 underwent repeat endoscopy (EVL related ulcers requiring no endoscopic therapy=9, repeat EVL=13, TIPS=11, sclerotherapy=3). When attempted, repeat EVL controlled bleeding in 87% (p=0.15 compared to initial EVL success). No clinical or laboratory predictors for rebleeding were identified. There were 12 deaths (9%) due to uncontrolled bleeding (initial failure of bleeding control=6, rebleeding=6); another 17 patients died of causes not directly related to bleeding within 30 days, for an overall mortality rate of 21%. Patients who died from uncontrolled bleeding had significantly higher CPT scores, MELD scores and INR values compared to patients who had successful bleeding control (p<0.01 for each comparison); on multivariate analysis, only INR predicted mortality (p<0.001). The median INR value in patients who died of uncontrolled bleeding was 2.5 (interquartile range 1.7-3.5), compared to 1.5 (interquartile range 1.3-1.8) in patients who survived the bleeding episode (p<0.0001). Conclusions: EVL is similarly successful in achieving hemostasis in initial as well as recurrent variceal bleeding. Coagulopathy manifesting as a high INR value is identified as the only predictor of a poor outcome. This may suggest a role for aggressive correction of coagulopathy in patients presenting with acute variceal UGIB.

Purpose: Wireless capsule endoscopy (Given Imaging, Yoqneam, Israel) is a novel imaging technique for evaluating the small bowel, with a high diagnostic yield in obscure GI bleeding. We addressed the impact of WCE in managing patients hospitalized with obscure GI bleeding. Methods: Inpatients undergoing WCE for obscure GI bleeding over a 3 year period were identified from retrospective chart and endoscopy database review. All patients had undergone conventional endoscopy prior to WCE. Clinical characteristics and hospital course were analyzed to determine if WCE led to changes in management and outcome. Results: During the study period, 56 successful WCE studies were performed on 55 inpatients (age 67.82 ± 2 yrs, 25F/30M) with obscure GI bleeding, accounting for 30% of WCE studies in the same period. Two-thirds of the studies were performed for obscure-overt bleeding. Abnormalities were seen in 36 patients (65%). The frequency of finding abnormalities trended higher in obscure-overt compared to obscure-occult bleeding (70% vs. 56%, p = ns). Findings consisted of vascular abnormalities including angiodysplasia (21 patients, 38%), fresh blood (7 patients, 13%), mucosal ulceration (4), and other potential bleeding lesions (4). Findings were in the small bowel in 29 (52%), while 7 (13%) had bleeding lesions in the stomach/esophagus (4) or colon (3). Intervention was recommended after definitive localization of the bleeding source in 15 (27%) patients, of which two-thirds had obscure-overt bleeding. Recommendations included push-enteroscopy (4 patients), surgery (5 patients), and other endoscopy (upper endoscopy 2, colonoscopy 3, endoscopic ultrasound 1 patient). Interventions were performed in the small bowel in 8 patients, stomach in 2 patients, colon in 2 patients, and one patient underwent angiography; 2 other patients recommended intraoperative enteroscopy declined. Potentially curative interventions (ablation, surgery) were performed in 5 patients (9%). WCE was misleading in two cases: a) angiodysplasia were seen, but the patient subsequently was found to have bled from an aortoenteric fistula, and b) a patient rebled after small bowel resection directed by WCE. Conclusions: WCE appears to be of benefit in the inpatient setting, where definitive localization may direct intervention in about a quarter of inpatients presenting with obscure GI bleeding. Inpatient WCE may be more useful in obscure-overt bleeding. WCE findings can be misleading.

Purpose: Acute variceal upper gastrointestinal bleeding (UGIB) carries a high mortality in the absence of definitive management. We reviewed our data to determine the outcome of therapy of recurrent bleeding after successful EVL. Methods: All patients with recurrent variceal UGIB within 30 days of successful initial EVL were eligible. Recurrent bleeding was diagnosed if patients with successful hemostasis during the initial endoscopic procedure developed recurrent hematemesis or melena requiring repeat endoscopy for bleeding control. Inpatient charts were reviewed to confirm the source of recurrent UGIB and determine clinical outcome. Results: Over a 3-year period, 149 distinct episodes of acute variceal UGIB were treated with EVL in 131 patients. Recurrent bleeding occurred in 38 instances in 36 patients (29%), 82% during the initial hospital stay. Two patients died before further intervention. The study group consisted of 34 patients (age 56.0 ± 2.3 years, 11F/23M) who underwent endoscopy for 36 episodes of recurrent bleeding after a mean of 6.4 ± 0.8 days after EVL. On endoscopy, ulcers at band ligation sites were identified as the cause of recurrent bleeding in 9 instances (25%), managed conservatively without further recurrence. Further therapy consisted of repeat EVL in 13 of the remaining 27 instances (48%), sclerotherapy in 3 instances (11%), and transjugular intrahepatic portosystemic shunt (TIPS) placement in 11 instances (41%). A third EVL was performed in 2 instances. TIPS was performed in one instance each for failed EVL and failed sclerotherapy. EVL was overall successful in 73% (as compared to 81% for initial bleeds, p = ns), and comparable numbers were obtained for TIPS (Table). Overall mortality after recurrent bleeding was 33%, significantly higher compared to patients without rebleeding after initial EVL (12%, p = 0.009). Coagulopathy (INR > 1.7, platelets <75,000) demonstrated a trend (p = 0.08) for higher mortality in patients with recurrent bleeding on univariate analysis, while MELD score, CPT score and variceal size did not predict mortality.Table: Comparison of EVL and TIPS for recurrent variceal bleedingConclusions: Repeat EVL for recurrent bleeding results in bleeding control comparable to EVL for initial variceal bleeding. Rebleeding appears to be a marker for higher mortality, and since coagulopathy may predict higher mortality, aggressive measures to correct coagulopathy could be important in recurrent variceal bleeding.

Older literature suggests that 53% of patients with known varices presenting with acute UGIB have a nonvariceal source; bleeding peptic ulcers and erosions account for almost 40% (Dagradi, Am J Gastroenterol 1970). These proportions may have changed with advances in endoscopic evaluation and overall management. Methods: All patients presenting with definitive acute UGIB and varices on upper endoscopy were eligible for inclusion. Inpatient charts were reviewed to confirm the source of acute UGIB and corroborate the diagnosis of portal hypertension. Variceal bleeding was diagnosed if active bleeding or red wale sign were visualized. Nonvariceal etiologies required the presence of stigmata of bleeding (active bleeding or adherent clot). Multiple bleeding presentations within a 30 day period were considered part of one bleeding episode. Results: Over a 2-year period, varices were visualized on endoscopy in 143 consecutive patients (55±1.1 yr, 50F:93M) presenting with 173 distinct episodes of acute UGIB. Bleeding presentation included hematemesis (bloody or coffee ground emesis) in 65 instances (37%), melena in 41 (24%) and both in 67 (39%). The source of UGIB was conclusively localized in 137 bleeding episodes (79%). A variceal source was identified in 115 instances (66%), 108 from esophageal varices, and 7 from gastric varices. Nonvariceal bleeding accounted for 22 bleeding episodes (13%), and included peptic ulcers and erosions (9 episodes, 5%), angioectasia including gastric antral vascular ectasia (9, 5%), Mallory-Weiss tear (1, 0.6%), oozing portal hypertensive gastropathy (1, 0.6%), and oro- and nasopharyngeal sources (2, 1%). The bleeding source could not be definitively identified in 36 instances (21%); endoscopic findings included Grade II or larger varices in 10 (4 underwent band ligation), peptic ulcers and erosions in 9, nonbleeding angioectasia in 4 (3 treated), esophagitis in 4 and Mallory Weiss tears in 2. All patients with angioectasia had melena but no hematemesis, and all patients with gastric variceal bleeding had bloody emesis. However, bleeding presentation did not predict source of UGIB (p=NS). Conclusions: Using strict criteria to establish the bleeding source, variceal bleeding accounts for as many as two-thirds of patients with known varices presenting with UGIB - a significant change from older reports. Peptic ulcers are uncommon sources of bleeding in this population in the present day. Bleeding presentation cannot be used to predict the most likely source of UGIB.

Background Aorto-enteric fistula is rare but can result in exsanguination without timely surgery or endovascular stent placement. Methods Four cases of aorto-enteric fistula were reviewed in which the presentation was unusual and diagnosis difficult. Observations The first patient had an aorto-sigmoid fistula in the setting of an aorto-bi-femoral graft. Two patients had a primary aorto-enteric fistula, one to the stomach from a suprarenal aortic aneurysm, and the other, to the duodenum in the setting of retroperitoneal spread of renal cancer. The aortoduodenal fistula recurred in the 4th patient within 3 months of surgical repair; this patient is the only one who survived long term. Conclusions When presentation is atypical, the diagnosis of aorto-enteric fistula can be extremely difficult. Because investigative studies are not consistently useful in making a definitive pre-operative diagnosis, a strong index of clinical suspicion and a willingness to consider surgical exploration are essential for timely and successful management.

BACKGROUND:Historically, acute lower intestinal bleeding has incorporated small bowel with colonic sources. This potentially obscures the unique characteristics of small bowel bleeding, which are eclipsed by the attributes of the much more common colonic bleeding. Separating acute lower intestinal bleeding into small bowel and colonic sources may delineate characteristics of each, thereby making it possible to determine whether clinical outcomes vary by anatomic level of bleeding.METHODS:A total of 29 consecutive patients (15 women, 14 men; age 68.6 +/-2.4 years) with acute small bowel bleeding were compared with two other groups, each with 29 consecutive patients, with either acute colonic bleeding or acute upper GI bleeding. Clinical presentation, outcomes, and resource utilization for small bowel bleeding were compared with similar parameters for acute colonic bleeding and upper GI bleeding.RESULTS:Although the clinical presentation did not always distinguish the 3 groups, resource utilization was significantly higher in the small bowel bleeding group. The latter group required a higher number of diagnostic procedures (p < 0.001) and blood transfusions (p < 0.001), remained in hospital longer (p < 0.05), and had a higher cost of hospitalization (p < 0.001) compared with the colonic bleeding and upper GI bleeding groups. The mortality rate for patients with small bowel bleeding was 10%. Although none of the patients with upper GI bleeding and only 14% of those with colonic bleeding required greater than 3 diagnostic procedures, 79% of patients with small bowel bleeding required 4 procedures for diagnostic localization (p < 0.0001).CONCLUSIONS:Small bowel bleeding ("mid-intestinal bleeding") is a distinct clinical entity with significantly worse outcomes compared with colonic bleeding and upper GI bleeding. The focus of the investigation should be directed to the small bowel, with enteroscopy or capsule endoscopy, when 3 investigative procedures fail to localize recurrent overt GI bleeding.