affected by gastric lymphoma: in two of these, gastriclocalizations missed at other examinations were found.

For a time, it seemed that a tissue diagnosis of gastriclymphoma was difficiult to achieve endoscopically, but anumber of studies including this one have shown that apositive biopsy or cytology can be achieved in a majority ofcases. (Gastrointest Endosc 21:66-68,1974, and 21:123-125,1975). As in gastrointestinal adenocarcinoma, directed cy­tology adds to the diagnostic yield (Cancer 41:1456 -1461,1978), and infiltrative disease presents a greater problem inobtaining positive tissue than exophytic tumor masses ormalignant ulcers (Gastroenterology 69:1183 -1187, 1975). Ininfiltrative gastric lymphoma, some endoscopists have beenevaluating a "large particle" biopsy obtained with a two­channel endoscope and "lift and cut" technique in an effortto increase the diagnostic yield (Gastrointest Endosc 23:29­30, 1976), but the technique has not gained widespread usebecause of the apparent risks.

THE EDITOR

Value of colonoscopy in the detection of sigmoidmalignancy in patients with diverticular disease

WESDORP Ie, GLERUM ), AGENANT A, SCHRIJVER M,TYTGAT GN

Acta Chir Belg Nov-Dec 78(6):355-358, 1978Sixty patients with diverticular disease, referred be­cause a barium enema examination could not excludea co-existing malignancy, were studied in a retrospec­tive manner to evaluate the contribution of colonos­copy in the diagnosis of sigmoid carcinoma. All x-raystudies were blindly reviewed and divided in twocategories: (1) diverticular disease with malignancy orstrong suspicion for malignancy, and (2) diverticulardisease without suspicion for malignancy. The accu­racy of the endoscopic examination was evaluated bya follow-up study with a range of 3 months to 3 years.Colonoscopy appeared to be accurate in more thanthree-fourths of the patients, but was not helpfulwhen there was a severe stenosis and/or the diseasedsegment could not be reached for biopsy. The inci­dence diminished when a small caliber fiberendo­scope was used, practically always reaching or passingthe stenotic segment. There were no false-positive orfalse-negative endoscope results in this study. In asubstantial number of patients major surgical explo­ration was prevented. Colonoscopy is considered avaluable adjunct in detecting or eliminating cancer incolonic diverticular disease. However, the availabilityof small caliber fiberendoscopic instruments is a pre­requisite for reaching an acceptable success rate anddiagnostic accuracy.

Control of upper gastrointestinal hemorrhage byendoscopic spraying of dotting factors

LINSCHEER WG, FAZIO TLGastroenterology 77:642-646, 1979

Another new technique for the endoscopic control ofupper gastrointestinal hemorrhage is under develop­ment. It consists of a spraying device which delivers

206

atomized clotting factors (thrombin and fibrinogen)directly to a focal bleeding site. In vivo studies in astandardized dog model of acute gastric ulcer bleedingdemonstrated a significant reduction in bleeding timeby spraying thrombin and cryoprecipitate (mean ±standard error bleeding time = 223 ± 14 seconds) ascompared to a saline spray (mean = 705 ± 60 seconds)or to the untreated controls (mean = 498 ± 43 sec­onds), p < 0.001. Similar results were obtained in theheparinized dog model. The mean bleeding time was226 ± 11 seconds in the sprayed animals, while thecontrol ulcer bled continuously. The four lumen spraytube fits the biopsy lumen in commercial gastroscopes.The feasibility of the technique for treatment of oozingvenous bleeding sites in the upper gastrointestinaltract has been successfully tried in a clinical setting.The authors feel that the technique in its present formis ready for clinical testing in a controlled manner.

The presence of hydrochloric acid is known to adverselyaffect many stages in the coagulation system, and topicalacid has been shown to prolong bleeding time in the caninestomach (Am J Proctol Gastroenterol Colon Rectal Surg 30:33 -35, 1979). The addition of H 2-receptor blockade has atheoretical advantage in attempts to induce clotting in ooz­ing gastric lesions.

THE EDITOR

Acute right-sided hemorrhagic colitis associated withoral administration of ampicillin

SAKURAI Y, TSUCHIYA H, IKEGAMI F, FUNATOMI T, TAKASU S,

UCHIKOSHI T

Dig Dis Sci 24(12):910-915, 1979Among 56 cases who presented to Kanto-Teishin Hos­pital complaining of bloody diarrhea or considerablehematochezia of acute onset, eight cases (14.3%) wereconsidered due to colitis associated with oral ampicil­lin therapy. The bloody diarrhea, often with abdominalcramps, began 2 to 7 days after starting the treatment.The dosage of ampicillin taken ranged from 2.0 to 4.0g. Early total colonoscopy and biopsy revealed markedmucosal hemorrhage with minimal or no inflammatorychanges mainly in the right colon. Rectum and sigmoidcolon were completely normal except in one case.Symptoms rapidly resolved after the endoscopy. Atfollow-up colonoscopy, performed 4 to 12 days later,the mucosal changes had cleared completely. Therewas no evidence to support a hypersensitivity reactionof the colonic mucosa to ampicillin. The authors be­lieve that right-sided hemorrhagic colitis is one of thecommon forms of colitis associated with ampicillin.

Endoscopic duodenal biopsies in coeliac disease andduodenitis

HOLDSTOCK G, EADE GE, ISAACSON P, SMITH CL

Scand J Gastroenterol 14(6):717-720, 1979Endoscopic duodenal biopsies were taken from 27patients with suspected celiac disease and comparedwith intubation capsule jejunal biopsies. The speci­mens were reported without knowledge of the pa-

GASTROINTESTINAL ENDOSCOPY