2. Stolte M, Wiesser V, Schaffner 0, Koch H. Vascularisation of the papilla of Vateri and bleeding risk of papillotomy.

Leber Magen Darm 1980;10:293-301. 3. Dunham F, Bourgeois N, Gelin M, Jeanmart J, Toussaint J, Cremer M. Retroperitoneal perforation following endoscopic sphincterotomy; clinical course and management. Endoscopy 1982;14:92-6.

Upper gastrointestinal endoscopy in Turkey: a review of 5,000 cases

To the Editor: The superiority of fiberoptic endoscopy over the barium swallow technique has been well established in diagnosing gastrointestinal disease and the site of bleeding. This letter reviews our experience in Turkey with endoscopy and compares it with radiological and surgical findings. Five thousand upper gastrointestinal endoscopies were performed at Hacettepe University Hospital, Ankara, from January 1980 to December 1984. The patients' ages ranged from 17 to 81 and included 3,222 males and 1,778 females. 184 out of 5,000 patients had an emergency endoscopic examination for acute upper gastrointestinal bleeding (UGIB). Examinations were performed by trained endoscopists. The forward viewing panendoscope, Olympus model GIF D3, was used to diagnose acute UGIB within 24 hours of initial bleeding episodes. Upper gastrointestinal series was performed on all patients except those without UGIB, and the results were compared with endoscopic examination. A total of 1,685 (34%) patients had a normal endoscopic examination. Three hundred eleven (7%) had hiatal hernia, which was the most common esophageal lesion seen, and 224 (68%) had coexisting esophagitis; 11 patients had candidial esophagitis. Varices were diagnosed in 5% of patients and graded according to severity. Gastritis was diagnosed in 489 (10%) patients. This was the most common type of gastric lesion. One hundred eightythree (4%) had gastric carcinoma, and 2% of the patients with gastric carcinoma had coexisting intestinal metaplasia. In this series, 151 (3%) presented with gastric ulcers and 55 (1%) had gastric polyps. The most common overall lesion was duodenal ulcers (714 or 16%). Five hundred forty (11%) patients had duodenitis, 50 (1 %) had pyloric stenosis, 15 had duodenal polyps, and 4 had duodenal diverticula. Pyloric stenosis and duodenal polyps were seen more commonly in men than in women; the male/female ratio was 17.1/1 and 14/1, respectively. Hiatal hernias were seen more commonly in our series than in Europel and Africa. 2 The frequency of gastric ulcer was higher in our series than in Sudan.3 Our ratio of duoTable 1. Endoscopic findings in postgastrectomy patients
No. of cases Gastritis Anastomotic ulcer Esophagitis Intestinal metaplasia Hiatal hernia Gastric carcinoma Normal Total
94 44 6 4 2 2 134 286

denal to gastric ulcers (7/1) was less than in Sudan (31/1)3 but higher than in the United States (3/1): Diagnostic accuracy of radiology was 68% compared to endoscopy, whereas the accuracy of endoscopy was 93% compared to surgery. Comparison of endoscopy with surgery at Mayo Clinic demonstrated that endoscopy missed fewer lesions.s Most of the missed lesions were found high in the gastric fundus, between large gastric folds or in a scarred duodenum. In our study, the majority of diagnostic errors by x-ray examination occurred primarily with gastritis. Upper endoscopic examination in 152 (53%) postsurgical patients found definite abnormalities as indicated in Table 1. In comparing x-ray and endoscopic examination, 77 patients (27%) were accurately diagnosed by both procedures. Endoscopic examination was established to be most helpful in diagnosing postsurgical patients since radiology showed the correct diagnosis in only 27% of the patients and these were predominantly marginal ulcers. Hirschowitz and Luketic6 found radiologic results to be positive in only 36% of 111 patients with marginal ulcers. The result of our study clearly indicates that an adequate evaluation of the postoperative stomach cannot be considered complete until fiberoptic examination has been carried out. Endoscopy was successfully performed on 184 patients with acute UGIB, 220 lesions being identified in 171 patients, with a diagnostic accuracy of 93% proven by surgery. In 13 patients no bleeding site was found. The duodenum was the most common site of bleeding, with duodenal ulceration in 68 (37%) patients. Fifty-one patients had erosive gastritis, the stomach being the second most common bleeding site. Esophageal varices accounted for 29 (16%), gastric ulcers for 9 (5%), and gastric carcinoma for 8 (4%). Two gastric hemangiomas and 2 gastric polyps presented in this series. The ratio of duodenal to gastric ulcers was 7.6/1, which was higher than the 0.87/1 1 and 6.22/1 7 found in reports from Great Britain. Serious endoscopic complications occurred in 6 patients: 2 cases of pseudo-acute abdomen, 1 of deep hematoma of the neck, and 1 of mandibular dislocation. One patient died of cardiac arrest during endoscopy; another patient with thrombocytopenia had a deep neck infection with hematoma as a major complication and died 6 days later. These complications represented 0.12% of 5,000 endoscopies. The experience with gastrointestinal endoscopy in Turkey is similar to that reported from Europe and the United States. However, the incidence of duodenal ulcer appears to be higher and that of esophageal varices lower, perhaps reflecting cultural influences.
Halis Simsek, MD Hasan Telatar, MD Sukran Karacadag, MD Burhan Kayhan, MD Figen Batman, MD Section of Gastroenterology
Hacettepe University Hospital Ankara, Turkey

Percent
35 15 2 1 1 1 45 100

REFERENCES
1. Cotton PB. Fiberoptic endoscopy and the barium meal. Results and implications. Br Med J 1973;2:161-5. 2. Hoare AM. Comparative study between endoscopy and radiology in upper gastrointestinal hemorrhage. Br Med J 1975;1:27. 3. Fedail SS, Homeida MM, Aneba BM, Ghandour ZM. Upper

68

GASTROINTESTINAL ENDOSCOPY

4. 5. 6. 7.

gastrointestinal fiberoptic endoscopy experience in the Sudan. Lancet 1985;15:2:897-9, McGuigan JE. Peptic ulcer in Harrison's principle of internal medicine. New York: McGraw-Hill, 1987:1242. Cameron B, Ott BJ. The value of gastroscopy in clinical diagnosis: a computer-assisted study. Mayo Clin Proc 1977;52:8068. Hirschowitz BI, Luketic BC. Endoscopy in the postgastrectomy patient, an analysis of 580 patients. Gastrointest Endosc 1971;18:27. Jonston SJ, Jones PF, Kyle J, Needham CD. Epidemiology and course of gastrointestinal hemorrhage in northeast Scotland. Br Med J 1973;3:655-60.

Transillumination of light in the right lower quadrant during total colonoscopy

To the Editor: Transillumination of light in the right lower quadrant is one of the landmarks traditionally used to denote the presence of the endoscope tip being in the cecum during total colonoscopy.l-4 Although transillumination has long been used by colonoscopists to confirm reaching the cecum, the incidence of its occurrence during total colonoscopy has never been previously studied. We recently determined the incidence of transillumination and compared the finding of light in the right lower quadrant between a standard fiberoptic colonoscope and video colonoscope. 275 consecutive office patients having total colonoscopy were entered into the study. The cecal location was visually confirmed in each case by identification of cecal anatomic landmarks (appendiceal orifice, ileocecal valve, or junction of the teniae coli). All instruments were manufactured by the Olympus Corporation of America, and the CFLB3W fiberendoscope was compared with the colon video endoscope VIOL. The LB3W used the CLV-10 light source and a video camera (OTV) attached to the eye piece transmitted the image to a television monitor. The light setting was on automatic, set at the minimal illumination level necessary for adequate operator visualization of intraluminal events. The video endoscope was at standard settings. 169 patients were examined with the fiberendoscope, and 106 patients were examined with the video endoscope. Light transillumination could be seen in the right lower quadrant when the colonoscope tip was in the cecum in 152 patients examined with the fiberendoscope (90%) and in 57 patients examined with the video endoscope (54%). Whenever the tip of the endoscope was in the cecum and transilluminated light could not be readily detected in the right lower quadrant, the room lights were dimmed and an attempt was made to palpate the abdomen in various areas while observing the abdominal wall. Frequently, moving the instrument from the cecal pole to an area just above the ileocecal valve resulted in transillumination that had not been possible because the endoscope tip was deep in the pelvis. When transillumination was present using the video colonoscope, only a faint glow was seen as opposed to the more intense light from the fiberoptic instrument. There are several endoscopic landmarks that are useful for cecal identification. These include a change in the color of the intraluminal fecal effluent in the right colon, a notch or an indentation on a prominent fold in the right colon denoting the superior lip of the ileocecal valve, the appenVOLUME 34, NO.1, 1988

diceal orifice, and the junction of the teniae coli, which may produce a crow's foot appearance as they conjoin at the base of the cecum. 1- 5 Transillumination of light in the right lower quadrant is one of the traditional landmarks used for cecal identification. In many instances, light may be seen in the right lower quadrant, although the tip of the instrument is not actually in the cecum. 4 5 Reliance on transillumination can result in spurious endoscope tip localization since right lower quadrant light may be seen when the tip of the endoscope is in the sigmoid colon (pushed to the right) or in the midportion of the transverse colon, which has been pushed by the advancing endoscope toward the right. Although light in the right lower quadrant is not as useful nor as reliable a parameter of total colonoscopic intubation as the confirmatory endoscopic landmarks, this study was performed to ascertain the frequency with which transillumination can be seen in the right lower quadrant utilizing conventional fiberoptic and the newer video instruments. The light can be seen in 90% of cases when total colonoscopic intubation has been achieved using fiberendoscopes. In 10% of cases, either due to anatomic position, obesity, or other factors, light cannot be transilluminated in the right lower quadrant when the instrument tip is in the right lower quadrant (in the cecal pole or just above the ileocecal valve). In comparison, the video endoscope with its less brilliant illumination results in transillumination through the abdominal wall in only 54% of patients when the tip of the endoscope is in the cecum.
Jerome D. Waye, MD Mary Ann E. Atchison, RN, GIA Maria C. Talbott, GIA Blair S. Lewis, MD
New York, New York

REFERENCES
1. Sugawa C, Schuman B. Primer of gastrointestinal endoscopy. Boston: Little, Brown, and Co, 1981:116-7. 2. Cotton P, Williams C. Practical gastrointestinal endoscopy. London: Blackwell Scientific Publications, 1980:118-9. 3. Shinya H. Colonoscopy: diagnosis and treatment of colonic diseases. New York: Igaku-Shoin, 1982:67-8. 4. Waye J. Colonoscopy intubation techniques without fluoroscopy. In: Hunt R, Waye J, eds. Colonoscopy. London: Chapman and Hall, 1981:174-6. 5. Whalen J, Riemenschneider P. An analysis of the normal anatomic relationships of the colon as applied to roentgenographic observations. Am J Roentgenol Radiat Ther Nucl Med 1967;99:55-61.

Cecal perforation following flexible sigmoidoscopy

To the Editor: Flexible fiberoptic sigmoidoscopy is a routine study and, although it is a relatively innocuous procedure, hazardous complications can occur. Cecal perforation following flexible sigmoidoscopy is an unusual complication. 1 An 80-year-old woman underwent flexible sigmoidoscopic examination to ascertain the etiology of blood in her stool. The procedure was performed to 25 em and revealed two polyps and spasm of the sigmoid that did not relax after intravenous glucagon. The procedure was terminated, and no biopsies or polypectomies were attempted. Soon after,
69