Stephen M. Schutz

Other clinical studyEndoscopic therapy of sclerosing cholangitis

The efficacy of endoscopic treatment in primary sclerosing cholangitis has not been clearly established. This report presents endoscopic intervention in 53 consecutive patients with this disorder. Pertinent data were abstracted from the GI-TRAC database, medical records, and cholangiograms, and clinical follow-up was obtained by telephone interview of the subjects. Assessed treatment outcomes were clinical symptom, liver function test, and cholangiographic appearance. Between 1986 and 1993, 85 patients with primary sclerosing cholangitis underwent successful ERCP, of which 36 men and 17 women underwent 100 therapeutic endoscopic procedures. Forty-three of 50 dilations, 37 of 38 stentings, 8 of 8 nasobiliary tube placements, and 11 of 17 stone extractions were technically successful. These treatments were complicated by cholangitis or pancreatitis in 15 patients. Clinical follow-up was obtained in 50 of 53 patients who had undergone 85 procedures (median follow-up of 31 months): 28 patients felt better, 21 felt the same, and 1 felt worse. Liver function tests obtained within 3 months of the endoscopic treatment were significantly improved compared with pretreatment values (P < .001). Cholangiograms showed improvement in 36% of the patients, no change in 51%, and the effect of therapy could not be assessed in 13%. Overall, 41 of 53 patients (77%) had improvements of their clinical symptoms, liver function tests, or cholangiograms.

Case of the month: "Air-in a view"

A 57-year-old female was referred to Duke University Medical Center (DUMC) for investigation of recurrent pancreatitis. Her first episode of pancreatitis, which had occurred 3 years earlier, manifested as epigastric pain and a markedly elevated serum amylase level which necessitated hospital admission for 9 days. Her history was negative for relevant pancreatic risk factors including alcohol abuse, medications, family history and hyperlipidaemia. Ultrasound (US) of the gallbladder showed no gallstones. CT of the abdomen showed a swollen head of pancreas with peri-pancreatic inflammatory change and fluid, consistent with pancreatitis. The biliary tree was not dilated. She made a complete recovery and was well until 8 months later when she had recurrent epigastric pain of a similar character with elevation of the serum amylase. She was admitted to hospital for 3 days. A repeat US scan showed no cholelithiasis. Following resolution of her symptoms, she underwent endoscopic retrograde cholangiopancreatography...

3642 YIELD OF COLONOSCOPY FOR ISOLATED PROXIMAL COLONIC DISEASE IN PATIENTS WITH RECTAL BLEEDING: A MULTICENTER STUDY OF 2,449 PATIENTS.

Introduction: Rectal bleeding is a common presenting symptom in gastroenterology clinics. However, the yield of colonoscopy for isolated proximal disease, especially in young patients, is unclear. Aim: To determine the yield of colonoscopy for significant proximal colonic lesions in the absence of distal disease. Methods: Data were retrieved from an endoscopic database (GI Trac) of 15,550 colonoscopies performed in 7 American endoscopy units. Patients with previous gastrointestinal surgery, recent colonoscopic or barium enema examinations or with known lower gastrointestinal disease were excluded. Proximal disease was defined as disease above the sigmoid colon, which might be missed on flexible sigmoidoscopy. Patients were categorized by age for analysis (18-39 years [n=300], 40-49 [n=471], 50-59 [n=598], 60-69 [n=568], 70-79 [n=375] and ≥80 [n=137]). Results: 2,449 patients (median age 57 years, range 18-100; 1048 female) undergoing colonoscopy for rectal bleeding were studied (15.7% of all colonoscopies). The incidence of hemorrhoids ranged from 49-66% between age groups (Chi-square test for trend, p=0.30) and incidence of small (

3719 LOW YIELD OF SIGNIFICANT PATHOLOGY WHEN COLONOSCOPY IS PERFORMED FOR ISOLATED ABDOMINAL PAIN: A MULTICENTER DATABASE ANALYSIS OF 15,550 PATIENTS.

Introduction: Abdominal pain is a common symptom, both in the general population and in gastroenterology clinics. The yield of colonoscopy for abdominal pain in the absence of other symptoms (rectal bleeding, diarrhea or anemia) is unclear. Aim: To determine the yield of colonoscopy in adults with isolated abdominal pain. Methods: All data were retrieved from an endoscopic database (GI Trac) of 15,550 colonoscopies performed in 7 American endoscopy units. Patients with rectal bleeding, diarrhea, anemia, previous gastrointestinal surgery, recent colonoscopic or barium enema examinations and those with known organic lower gastrointestinal disease were excluded from study. Significant disease was defined as cancer, polyps ≥ 1cm and inflammatory bowel disease (IBD). Results: 389 patients (mean age 54 years, range 18-89; 237 female) undergoing total colonoscopy for abdominal pain were studied. These constituted 2.5% of all colonoscopies. Significant pathology was detected in 16 patients (4.1%) (2 with cancer, 4 with IBD, 10 with polyps). 9 of these patients (56%) had disease in, or proximal to, the descending colon that might not have been detected by flexible sigmoidoscopy (2 with cancer [aged 68 and 81 years], 2 with IBD, 5 with polyps). In addition, isolated right-sided diverticula were found in 7 patients (1.8%) aged 41 to 68 years. One of 68 patients less than 40 years had Crohn s disease in both the proximal and distal colon, but no patient less than 40 had significant disease or diverticula that would have gone undetected by sigmoidoscopy. Male sex was the only predictor of significant disease beyond the reach of conventional flexible sigmoidoscopy (male, 7/152 (4.6%) v s female 2/237 (0.8%); Fisher s exact test, p=0.03). Conclusion: Colonoscopy for isolated abdominal pain rarely yields significant disease, especially in females and in patients less than 40 years.

3393 Benchmarking ercp performance with a prospective endoscopy database.

Benchmarking endoscopic activity is of increasing interest to professional societies, payers and patients, allowing identification of major variations in performance as a guide to improvement. Seven centers have used a single endoscopy reporting database (GI-Trac) for various periods since 1994, and have accumulated data on 8,094 ERCP procedures. Data are entered directly by the endoscopist, immediately after the procedure.We compared activities by center (not by individual endoscopists), for markers of overall activity, data quality, case complexity, and technical success. These data (table)show considerable variations in the spectrum of ERCP practice and quality of data entry. Technical success rates are consistently high, with some trend towards higher success with larger volumes. Further analyses by individual endoscopists and practice volume will lead to specific insights and recommendations. In the future, endoscopists and centers will be at a disadvantage if unable to provide similar performance data.

3372 Benchmarking colonoscopy performance using a common prospective multicenter endoscopic database.

Background: Measures of quality in endoscopy centers are demanded increasingly by patients and third party payers. Methods: Six academic medical centers (4 university, 1 VA, 1 Military) prospectively recorded demographic, clinical and endoscopic data on all colonoscopies performed over a median of 5 years. Common measures of quality were compared between centers. Results: A total of 18,913 colonoscopies were performed. Results are shown in Table 1. Conclusions: There were marked variations in practice patterns such as the method of polypectomy, but measures of quality such as complications, cecal intubation, and recovery of snared polyps were uniformly good across centers. Common endoscopic database systems allow for rapid assessment, and demonstration of quality measures. Dr. Wallace was supported by the American Digestive Health Foundation, Wilson Cook Award

4712 Seasonal variation in the incidence of upper gi bleeding.

BACKGROUND: Peptic ulcer disease, a common cause of upper GI bleeding, is believed to have a seasonal variation in incidence. Similar variation in two main etiological factors, H. Pylori infection and NSAID use, has not been reported. Variation in prevalence of endoscopically confirmed ulcers may be due to an increased number of procedures rather than pathology itself. However, since acute upper gastrointestinal bleeding warrants endoscopy in each case, it may better reflect the seasonal differences in the burden of disease. AIM: We analyzed the monthly incidence of upper GI bleeding (by the mean of number of endoscopies with that indication) and NSAID use (by the number of visits with NSAID prescribed).We then analyzed the correlation between these variables. METHODS: Using a common endoscopy database (GI-TRAC) we counted the number of upper GI endoscopies performed each month in a period of 3 years (July 1995 to June 1998) in 4 medical centers.We estimated use of NSAIDS by counting visits with an NSAID prescription reported in the National Hospital Ambulatory Medical Care Survey (NHAMCS) 1997.We evaluated correlation between the monthly incidence of upper GI bleeding and NSAID use by linear regression analysis. RESULTS: Incidence of upper GI bleeding was on average 28% higher during the six month period December-May than between June-November (p

4910 The yield and predictors of esophageal pathology when upper endoscopy is used to initially evaluate dysphagia: multivariable analysis of 1649 patients.

BACKGROUND: Dysphagia is a common presenting symptom with a wide range of etiologies. The yield and predictors for significant pathology when EGD is used as the initial evaluation is unclear. AIM: Determine the yield and predictive factors of significant pathology when EGD is performed to evaluate dysphagia. METHODS: Data were retrieved from an endoscopic database (GI Trac) of 29,568 EGDs performed at 6 endoscopy units. Patient/procedure data were included in analysis if EGD was performed to evaluate dysphagia. Exclusion criteria were: prior esophageal evaluation including barium swallow or manometry, known upper GI pathology, prior gastroesophageal surgery, and failed EGD. Major pathology was defined as cancer, benign or malignant stricture, Barretts esophagus, and erosive esophagitis. Multivariable logistic regression was performed to evaluate for any relationship between endoscopic findings and presenting clinical features. RESULTS: Data from 1649 patients with dysphagia (6% of total EGDs) (mean age 56.7 yrs, SD 16.4, M:F 3:2) were analyzed. Symptoms other than dysphagia included: abdominal pain (153), chest pain (44), hematemesis (52), heartburn (302), melena (4), odynophagia(178), vomiting (52), and weight loss (43). Anemia was present in 29 patients. Major pathology was encountered in 727 (44%) patients. Male gender (OR 2.37, p=0.0001), the presence of heartburn (OR 1.68, p=0.0001) and odynophagia (OR 1.63, p=0.003) significantly predicted the presence of major pathology. Cancer was found in 70 (4%) patients and was significantly predicted by male gender (OR 3.5, p=0.0002), age (OR 1.02 p=0.01) and weight loss (OR 2.7, p=0.04). The esophagus was normal in 483 (29%) patients and was predicted by female gender (p=0.0001), the absence of heartburn (p=0.0004) but not age. Other EGD findings included: benign stricture (6%), diverticulum (0.7%), foreign body (3%), hiatal hernia (26%), infection (5%), esophagitis (28%) and esophageal web (10%). CONCLUSIONS: 1] Major pathology is found when EGD is performed to evaluate dysphagia in 44% of patients. 2] Major pathology is more likely in males, and when heartburn or odynophagia are present. 3] EGD is normal in 29% of patients with dysphagia and is more likely in females and if heartburn is absent. Our findings support the use of EGD in the evaluation of dysphagia for structural disease in all age groups, particularly if associated with heartburn, weight loss, or odynophagia.