Gary R. Zuckerman

AGA technical review on the evaluation and management of occult and obscure gastrointestinal bleeding.

This literature review and the recommendations therein were prepared for the American Gastroenterological Association Clinical Practice and Practice Economics committee. The paper was approved by the committee on May 16, 1999, and by the AGA governing board on July 18, 1999.

Prophylactic Therapy for Stress Ulcer Bleeding: A Reappraisal

The combined data from 16 prospective trials (2133 patients) appear to suggest that antacids prevent stress ulcer bleeding more effectively than does cimetidine. However, the use of occult blood detection methods to diagnose stress ulcer bleeding may have led to the recognition of clinically insignificant bleeding. When the data from these trials are categorized according to the criteria used for the diagnoses of bleeding (either occult blood detection or clinically overt bleeding), there was no significant difference between antacids and cimetidine in the prevention of overt bleeding (3.3% of 458 compared with 2.7% of 402 patients who bled, respectively; p = 0.69). In addition, both agents were more effective (p less than 0.001) than placebo (15% of 720 patients who bled) in the prevention of overt bleeding. Cimetidine and antacids are equal in preventing significant stress ulcer bleeding.

Diverticular Disease in Patients with Chronic Renal Failure Due to Polycystic Kidney Disease

Twelve patients with chronic renal failure and polycystic kidney disease represent 8% of the 151 hemodialysis patients followed up at the Chromalloy American Kidney Center, Washington University School of Medicine. Ten (83%) of these patients have diverticulosis, and four of these patients developed gross colonic perforation secondary to diverticulitis. Barium enemas on 31 chronic renal failure patients without polycystic kidney disease revealed diverticulosis in 10 (32%). None had diverticulitis. Barium enemas in 120 age-matched non-renal failure control patients revealed diverticulosis in 45 (38%). None had diverticulitis. These findings suggest that patients with chronic renal failure due to polycystic kidney disease have a high incidence of diverticulosis and diverticulitis, that diverticulosis occurs in patients with chronic renal failure without polycystic kidney disease at a rate similar to that in the general population, and that diverticulitis should be an initial consideration in the differential diagnosis of abdominal pain in patients with polycystic kidney disease.

Upper Gastrointestinal Bleeding in Patients with Chronic Renal Failure

Endoscopy to evaluate upper gastrointestinal bleeding was done for 482 patients over a 42-month period. Fifty-nine patients (12%) had chronic renal failure and upper gastrointestinal bleeding; the remaining 423 did not have renal failure. Angiodysplasia of the stomach or duodenum was the most frequent source of bleeding in patients with renal failure. Angiodysplasia (p less than 0.001) and erosive esophagitis (p less than 0.01) were significantly commoner causes of bleeding in the renal failure population than in the group without renal failure. Recurrent bleeding was also more frequent in patients with renal failure (25%) than in the other patients (11%). Angiodysplasia was the most frequent source of recurrent bleeding in patients with renal failure (53%) whereas peptic lesions were the most likely sources in those without renal failure (51%). These data show that the differential diagnoses of first and subsequent upper gastrointestinal bleeding sites differ for patients with and without chronic renal failure.

A prospective multicenter evaluation of new fecal occult blood tests in patients undergoing colonoscopy

OBJECTIVE:Guaiac-based fecal occult blood (FOB) tests, in particular, Hemoccult II (HO), are commonly used to detect colorectal neoplasia. Because the sensitivity and specificity of these tests are critical to cost-effective screening programs, we aimed to investigate the improved performance characteristics of new FOB tests for known colonic lesions.METHODS:Nine centers worldwide performed FOB testing with guaiac-based tests (Hemoccult II [HO] and Hemoccult II SENSA [SENSA]) and immunochemical tests (HemeSelect [HS] and FlexSure OBT [FS]) on 554 patients referred for colonoscopy for predetermined indications. A combination testing strategy consisting of SENSA followed by HS or FS (which was considered positive only when both tests were positive) was also evaluated. Results of FOB tests were compared to findings on colonoscopy.RESULTS:Cancers were identified in 2.9% of subjects, whereas adenomas ≥10 mm were found in 39 patients. Small adenomas, colitis, and other lesions were identified in 141 patients. The positivity rate of HO for adenomas ≥10 mm was less than for SENSA (20.5% vs 35.9%, p < 0.05), whereas the positivity rate of HO, SENSA, FS, HS, or the combination tests for cancers was not statistically different. The overall positivity rates were significantly greater for FS (15.9%, p = 0.0002) and significantly lower using the combination tests (SENSA/FS 6.0%, p = 0.01; SENSA/HS 6.2%, p = 0.02) compared to HO (9.4%). In this study population, the relative specificity (i.e., true-negative tests/true-negatives + false-positives in patients without adenomas ≥10 mm or cancers) of HO (93.9%; 95% CI, 91.7–96.1) was similar to that of SENSA (92.8%; 95% CI, 90.4–95.2) and HS (90.1%; 95% CI, 87.4–92.8), and greater than FS (88.0%; 95% CI, 85.1–90.9, p < 0.001). When considering adenomas ≥10 mm, cancers alone or cancers and adenomas combined, the combination test using SENSA/FS was associated with significantly fewer false-positive tests than any of the individual tests.CONCLUSIONS:Compared to single tests, the combination test with the highly sensitive SENSA and an immunochemical test had slightly reduced sensitivity but significantly fewer false-positive tests than any single test. These data raise the possibility that a combination test (i.e., highly sensitive guaiac plus immunochemical) could reduce the costs of screening for colon cancer, and suggest that further study of combination test strategies is warranted.

Impact of Obesity on Bowel Preparation for Colonoscopy

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 endoscopists 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 (> or =25) was associated with an inadequate composite outcome score (P = .002). In multivariate logistic regression analyses, both BMI > or =25 (P = .04) and > or =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.

Triage considerations for patients with acute gastrointestinal hemorrhage admitted to a medical intensive care unit.

OBJECTIVE To determine whether previously identified clinical criteria, available at the time of triage, can predict clinical outcomes for patients with acute gastrointestinal (GI) hemorrhage. DESIGN An inception cohort study. SETTING Barnes Hospital, an academic tertiary care center. PATIENTS One hundred eight consecutive hospital admissions (103 patients) triaged to intensive care for GI hemorrhage. INTERVENTIONS Prospective patient surveillance, data collection, and risk stratification using preselected clinical criteria and outcomes assessment. MEASUREMENTS AND MAIN RESULTS Using clinical data available at the time of triage, 28 (25.9%) intensive care unit admissions were classified as low risk for having poor outcomes. There was no difference in the distribution of upper and lower GI tract sources of hemorrhage for the two risk groups (p = .310). Stigmata of recent hemorrhage were endoscopically identified for six (21.4%) of the low-risk patient admissions and for 16 (20.0%) of the high-risk patient admissions (p = .872). Patient admissions identified as low risk had significantly lower rates of recurrent GI hemorrhage (3.6% vs. 22.5%; p = .022), less acquired organ system derangements (1.0 +/- 0.3 vs. 1.5 +/- 1.0 organs; p < .001), shorter lengths of hospitalization (4.9 +/- 3.5 vs. 8.8 +/- 7.4 days; p < .001), required transfusion with fewer units of packed red blood cells (1.3 +/- 1.2 vs. 6.2 +/- 4.7 units; p < .001), and had a lower overall hospital mortality rate (0.0% vs. 21.3%; p = .008) compared with patient admissions identified as being high risk. CONCLUSION These data suggest that objective clinical criteria, available at the time of triage determination, can be utilized to identify a low-risk group of patients with acute GI hemorrhage, having favorable outcomes and potentially no need for intensive care unit services.

An objective measure of stool color for differentiating upper from lower gastrointestinal bleeding

Subjective reporting of the color of blood passed per rectum has been used to predict the location of gastrointestinal bleeding, but the validity of this clinical approach has never been evaluated systematically. In this study we determined the spectrum of patient and physician descriptors used to characterize the color of blood passed per rectum and evaluated prospectively if an objective test of stool color would correlate with or improve upon subjective descriptions in predicting bleeding locations. The objective test employed was a card containing five numbered colors that typify the spectrum of stool blood colors. One hundred twenty patients used 23 different descriptors or terms to verbalize the color of blood they passed per rectum, and in 22% of cases there was a seeming discrepancy between their verbalized color and the color they pointed to on the test card. Patients pointing to card color 4 (the black color) resulted in a closer matching to an upper bleeding source than physicians using terminology such as melena or tarry stools. Likewise, patients picking card colors 1 and 2 (the brightest red colors) resulted in closer matching to a coloanorectal bleeding source than physicians using the terms hematochezia or bright red blood per rectum (P<0.02 for each comparison). The positive predictive value of card color 4 for an upper bleeding source was very high both when patients pointed to this color or when it was determined from the available stool (0.95 and 0.98, respectively). The positive predictive value of card color 1 for lower bleeding was greater for patients selecting this color than for a direct stool comparison (1.00 vs 0.83). This study revealed marked variability and surprising inconsistency in subjective color reporting for both patients and doctors and the superiority of several card colors for separating upper from lower bleeding sources. This simple objective test should improve upon terms such as melena and hematochezia and help direct the initial diagnostic evaluation of gastrointestinal bleeding.

Controlled trial of medical therapy for active upper gastrointestinal bleeding and prevention of rebleeding

UNLABELLED This multicentered, placebo-controlled trial evaluated the efficacy of medical therapy to stop bleeding in 285 patients with active upper gastrointestinal bleeding (bleeding phase) and 194 patients who had ceased gastrointestinal bleeding and in whom therapy was instituted to prevent rebleeding during the same hospitalization (prevention phase). Patients in the bleeding phase were given cimetidine (300 mg every six hours) or intravenous placebo. There was no significant overall difference between intravenous cimetidine (71 percent) and placebo (77 percent) in stopping acute upper gastrointestinal bleeding. There was also no significant difference noted between intravenous cimetidine and placebo when specific bleeding lesions were evaluated. Once gastrointestinal bleeding had stopped, recurrence of bleeding while receiving prevention therapy (cimetidine tablets 300 mg one three times a day and at bedtime, or Mylanta II liquid 30 ml every hour, or cimetidine plus hourly antacids, or placebo) was evaluated in 194 of the patients in the bleeding phase. Twenty-four percent (12 of 51 patients) rebled while receiving cimetidine, 13 percent (five of 39 patients) rebled while receiving hourly antacids, 11 percent (six of 54 patients) rebled while receiving cimetidine plus hourly antacids, and 26 percent (13 of 50 patients) rebled while receiving placebo. None of these prevention regimens reached statistical significance (p = 0.13). Evaluation of specific bleeding lesions within this group also failed to show any significant value of prevention therapy. IN CONCLUSION (1) intravenous cimetidine offers no advantage over placebo in stopping active upper gastrointestinal bleeding; (2) the occurrence of rebleeding during the same hospitalization does not appear to be significantly affected by any of the medical regimens used for prevention. These findings would suggest that the cessation of active bleeding and the prevention of recurrent upper gastrointestinal bleeding during a single hospitalization appear to be unaffected by therapy directed at acid neutralization or reduction.

Enteroscopy-enteroclysis: experience with a combined endoscopic-radiographic technique

BACKGROUND Video enteroscopy provides high-quality diagnostic and therapeutic capabilities in the proximal small bowel. Enteroclysis remains an essential diagnostic technique in the distal small bowel. We report our experience with the combination of these techniques. METHODS Seventy-one patients with obscure gastrointestinal bleeding (group A, 54 patients) or abnormal radiologic studies (group B, 17 patients) were evaluated with enteroscopy. Enteroclysis via a tube inserted on withdrawal of the enteroscope was performed in all patients with nondiagnostic enteroscopy. RESULTS Enteroscopy identified bleeding sites in 29 of 54 (54%) group A patients (12 angiodysplasia, 10 ulcers, 7 gastric erosions, 1 vessel, 1 aortoenteric fistula), and lesions in 11 of 17 (65%) group B patients (7 ulcers, 3 benign strictures, 2 radiation enteritis, 1 mass). In group A, 13 (24%) patients had findings detectable by standard esophagogastroduodenoscopy. Enteroclysis identified masses in 2 of 24 (8%) group A patients, and lesions in 5 of 10 (50%) group B patients (3 strictures, 1 mass, 1 large diverticulum). No complications occurred. CONCLUSIONS The combination of enteroscopy and enteroclysis is safe and offers quality small bowel examinations in more comfortable and convenient single diagnostic sittings. This combination detected bleeding sources in 57% and lesions in 70% of patients. Though enteroclysis identified bleeding sources in only 8% of patients, this study excluded lesions other than angiodysplasia.