from vascular malformations. It is a reasonable option in re- fractory cases and may obviate the need for surgical interven- tion. Further clinical trials to better define the role of Thalid- omide in angiodysplasia are warranted. A pilot study proposed by our group in collaboration with celgene is underway. Muhammad Shurafa, M.D. Ginny Kamboj, M.D. Division of Hematology-Oncology Henry Ford Hospital Detroit, Michigan REFERENCES 1. Jaikishen JP, Falterman JB, Stueben ET, et al. Persistent and massive gastrointestinal hemorrhage. Hosp Pract 1995;30(4): 109, 113– 4. 2. Junquera F, Saperas E, de Torres I, et al. Increased expression of angiogenic factors in human colonic angiodysplasia. Am J Gastroenterol 1999;94:1070. 3. Tordjman R, Delaire S, Plouet J, et al. Erythroblasts are a source of angiogenic factors. Blood 2001;97(7):1968 –74. 4. Keyhani A, Jendiroba DB, Freireich EJ. Angiogenesis and leukemia. Leuk Res 2001;25:639 –45. 5. Bauditz J, Wedel S, Loch H, et al. Thalidomide for treatment of intestinal bleeding: Correlation to VEGF. Gastroenterology 2002;122(suppl):A-194 (abstract). Reprint requests and correspondence: Muhammad Shurafa, M.D., Henry Ford Hospital, Division of Hematology-Oncology, 2799 West Grand Boulevard, Detroit, MI 48202. Received Aug. 1, 2002; accepted Aug. 12, 2002. Diagnostic Accuracy of 13 C-Urea Breath Test for Turkish Children With Helicobacter pylori Infection TO THE EDITOR: We read the Kato et al’s. (1) article with great interest. To evaluate diagnostic accuracy of the 13 C-urea breath test (UBT) for children and to determine its optimum cutoff value, they studied 220 Japanese children aged 2–16 yr (mean 11.9) who underwent upper gastroin- testinal endoscopy for gastrointestinal symptoms and gastric biopsies for Helicobacter pylori (H. pylori) infection. En- doscopic diagnosis included gastritis, gastric ulcer, duode- nal ulcer, and combined ulcer. Antral biopsy specimens were obtained for histology, urease test, and culture. When culture was positive or when both histology and urease test were positive, the patient was considered to be infected with H. pylori (biopsy test). At the same time, the UBT test was performed on all patients. Breath samples were obtained at baseline and at 20 min after ingestion of 13 C-urea without a test meal and were analyzed by isotope ratio mass spec- trometry. Based on biopsy tests, an optimum cutoff value of 3.5%0 at 20 min was determined using a receiver operating characteristic curve. According to the biopsy test, 89 (40%) of 220 patients were infected with H. pylori. Endoscopic diagnoses were gastritis in 45 patients, gastric ulcer in eight, duodenal ulcer in 35, and combined ulcer in one patient with H. pylori infection. The overall sensitivity and specificity of the UBT test at a cutoff value of 3.5%0 were 97.8% and 98.5%, respectively. The high sensitivity and specificity of the UBT test for diagnosis of H. pylori infection were demonstrated in all age groups. In this study, we evaluated the diagnostic accuracy of the UBT test in Turkish children with H. pylori infection. During a period of 18 months (February, 2001–July, 2002), 181 patients (93 boys, 51.4%) who underwent endoscopy for dyspeptic symptoms or other reasons were enrolled into the study. The mean age of the patients was 11.0  3.2 yr (range 3–18, median 11). During endoscopy, antral biopsies for histological examination, ure- ase test, and culture were obtained. H. pylori infection was diagnosed when culture was positive or when both histology and urease test were positive. At the same time, the UBT test was performed on all patients. After a fasting period of at least 4 h, the child drank 150 ml of orange juice, and thereafter 50 (30 kg) or 75 (30 kg) mg of 13 C-labeled urea. Breath samples were obtained at baseline and at 25 min and were analyzed by isotope ratio mass spectrometry (IRIS IEC 601-1, Ayka, Germany). The cutoff value was 4%0. H. pylori infection was found in 160 (88.4%) patients. Abnormal endoscopic findings were antral nodularity in 112 patients, antral hyperemia in 14, duodenal ulcer in four, and duodenogastric reflux in three patients. For a cutoff value of 4%0, the sensitivity and specificity of the UBT test were 99.4% and 100%, respectively. The pos- itive predictive value was 100%, and the negative predictive value was 92.3%. Only one boy, aged 10 yr, had a negative UBT result (3.5%0), whereas culture was positive. Similar to the study of Kato et al. (1), we demonstrated high sensitivity and specificity of the UBT test in Turkish children with H. pylori infection. Kato et al. (1) defined the best cutoff value as 3.5%0. We agreed with them that with a cutoff value of 3.5%0, our sensitivity and specificity would be 100%. Although, we performed the UBT test by using orange juice (test meal) and we obtained a second breath sample at 25 min, the sensitivity and specificity of the UBT test was high. It seems that these parameters do not make any difference on UBT results. In conclusion, the 13 C-UBT test is accurate in the diagnosis of H. pylori infection in children. I ˙ nci Nur Saltık, M.D. Hu ¨lya Demir, M.D. Nurten Koc ¸ak, M.D. Hasan O ¨ zen, M.D. Figen Gu ¨rakan, M.D. Aysel Yu ¨ce, M.D. Hacettepe University Faculty of Medicine Department of Pediatrics Section of Gastroenterology Ankara, Turkey 222 Letters to the Editor AJG – Vol. 98, No. 1, 2003