To the Editor: Sung et al. (Jan. 19 issue)1 reported that inpatients with Helicobacter pylori infection and gastric ulcersunrelated to the use of nonsteroidal antiinflammatory drugs,"one week of antibacterial therapy without acid suppressionheals the ulcers as well as omeprazole and reduces the rateof their recurrence." The authors imply that they have provedthat therapy directed against H. pylori, and not against ulcers,cures ulcers. There are a number of ways to accelerate ulcerhealing without using antisecretory drugs (such as sucralfateor bismuth subcitrate). In our view, Sung et al. actually showedthat two regimens that had previously been shown to accelerateulcer healing were approximately equally effective. Their descriptionof bismuth subcitrate as an antibacterial agent is very misleading,because it is an extremely effective antiulcer agent independentlyof its antimicrobial activity. Colloidal bismuth subcitrateis an effective antiulcer agent whose effectiveness is not limitedto ulcers associated with H. pylori.2,3 Other bismuth preparationsdo not appear to have similar behavior with respect to experimentalulcers. For example, histochemical staining has been used tocompare the ability of bismuth subnitrate, bismuth subcarbonate,bismuth subsalicylate, and colloidal bismuth subcitrate to coatexperimental gastric ulcers in rats.3 When colloidal bismuthsubcitrate was administered, bismuth was deposited in a uniformlayer covering the ulcer base. The histochemical staining ofthe ulcer base was negative with the other bismuth salts andremained negative even when the concentrations of bismuth subcarbonateor bismuth subnitrate were increased by a factor of 5. Subsequentstudies have demonstrated important differences between differentformulations of bismuth subcitrate; colloidal bismuth subcitrateis effective in preventing ulceration in Shay rats, whereasnoncolloidal bismuth subcitrate is not.4
David Y. Graham, M.D. Mae F. Go, M.D. Veterans Affairs MedicalCenter Houston, TX 77030
References
Sung JJY, Chung SCS, Ling TKW, et al. Antibacterial treatment of gastric ulcers associated with Helicobacter pylori. N Engl J Med 1995;332:139-142. [Free Full Text]
Hall DW. Review of the modes of action of colloidal bismuth subcitrate. Scand J Gastroenterol Suppl 1989;157:3-6. [Medline]
Koo J, Ho J, Lam SK, Wong J, Ong GB. Selective coating of gastric ulcer by tripotassium dicitrato bismuthate in the rat. Gastroenterology 1982;82:864-870. [Medline]
Lavy UI, Koekkoek PH, Jaitly KD. Anti-ulcer activity of colloidal bismuth subcitrate in Shay-rats. Arch Int Pharmacodyn Ther 1976;224:291-298. [Medline]
To the Editor: Sung et al. randomly assigned patients to receiveeither omeprazole or a course of bismuth subcitrate, tetracycline,metronidazole, and antacid tablets (Mylanta). The amount ofantacid taken by the patients is not reported, nor is it clearwhether they were allowed to take their own antacids. Bismuthcompounds and antacids both have ulcer-healing properties equivalentto those of ranitidine.1,2 It is possible that the combinationof bismuth and antacids may have resulted in substantial healingof the relatively small (average size, <1 cm), uncomplicatedulcers in this study, independently of any antibacterial effects.Study of a group of patients treated with bismuth and antacidsalone would be required to rule out this possibility.
The authors do not explain how they determined that H. pylorihad been eradicated. In their inclusion criteria, a positiveCampylobacter-Like Organism test alone was not considered definitiveevidence of infection; a confirmatory smear or culture was required.In the United States, the gold standard for the diagnosis ofH. pylori infection includes antral biopsy for histologic analysisafter the application of special stains.3
It is also not clear how the one-year follow-up evaluation wasconducted. did all patients undergo endoscopy again, or onlythose with symptoms? Was the endoscopist blinded to the patients'original treatment groups? This information is important forthe interpretation of the data.
Douglas O. Faigel, M.D. David C. Metz, M.D. University of PennsylvaniaMedical Center Philadelphia, PA 19104
References
Wagner S, Gebel M, Haruma K, et al. Bismuth subsalicylate in the treatment of H2 blocker resistant duodenal ulcers: role of Helicobacter pylori. Gut 1992;33:179-183. [Free Full Text]
Hunter JO, Walker RJ, Crowe J, et al. Double-blind randomized multicenter study comparing Maalox TC tablets and ranitidine in healing of duodenal ulcers. Dig Dis Sci 1991;36:911-916. [Medline]
Brown KE, Peura DA. Diagnosis of Helicobacter pylori infection. Gastroenterol Clin North Am 1993;22:105-115. [Medline]
To the Editor: How many patients with gastric ulcer did Sunget al. exclude before they identified 100 patients with ulcersassociated with H. pylori for the study? What was meant by thestatement that "patients were excluded if they . . . had receivedantibacterial therapy in the past"? Does this mean they wereexcluded if they had received antibacterial therapy for ulcerdisease or antibacterial therapy for any reason within the precedingsix months, or is there some other explanation?
Harry E. Salyards, M.D. 606 N. Minnesota Ave. Hastings, NE 68901-5297
To the Editor: Of the 40 patients treated with omeprazole alone,H. pylori was eradicated in 5. This somewhat surprising outcomemay represent a false negative result due to the redistributionof H. pylori from the antrum to the body and fundus of the stomachthat occurs during omeprazole therapy.1 This effect of omeprazoleconsiderably lessens the sensitivity of antral biopsy for detectingH. pylori in patients who are taking the drug.
Using a one-week regimen of bismuth subcitrate, tetracycline,and metronidazole, Sung et al. found in an intention-to-treatanalysis that H. pylori was eradicated in 41 of 48 patients(85.4 percent). We believe the intention-to-treat analysis shouldinclude all 51 patients who were randomized and fulfilled theentry criteria. The inclusion of these patients would decreasethe rate of eradication of H. pylori to 80.4 percent.
Adam Harris, M.R.C.P. J.J. Misiewicz, F.R.C.P. Central MiddlesexHospital London NW10 7NS, United Kingdom
References
Logan RPH, Walker MM, Misiewicz JJ, Gummett PA, Karim QN, Baron JH. Changes in the intragastric distribution of Helicobacter pylori during treatment with omeprazole. Gut 1995;36:12-16. [Free Full Text]
To the Editor: The 95 percent confidence intervals in the articleby Sung et al. are expressed as accurate to 0.1 percent, whenin fact they vary from the exact 95 percent confidence intervalsby 1.0 to 4.1 percent.
Terry Fagan, M.D. Wilkes-Barre Veterans Affairs Medical Center Plains,PA 18711
The authors reply:
To the Editor: We excluded patients who had used nonsteroidalantiinflammatory drugs, those with ulcer bleeding in the previousfour weeks, and those who had received antibacterial therapyfor H. pylori infection in the past. During the two-year studyperiod, 1233 patients with gastric ulcers were seen in our unit.Of these, 920 presented with bleeding gastric ulcers. The remaining313 patients had nonbleeding gastric ulcers, including 32 whohad taken nonsteroidal antiinflammatory drugs, 72 with a negativeurease test, 58 with concomitant duodenal ulcers, 44 who wereolder than 70 years and had premorbid conditions, 4 who hadpreviously received anti-helicobacter therapy, and 3 who wereoffered surgery for gastric-outlet obstruction. These patientswere excluded from the study. Thus, 100 patients were studied.
Infection with H. pylori was considered to have been eradicatedif the results of the Campylobacter-Like Organism test, smears,and cultures were negative five weeks after randomization (i.e.,four weeks after the completion of antibacterial therapy). Afterulcer healing was confirmed by endoscopy, the patients wereasked to return for a one-year follow-up evaluation. All patientswere offered a repeated endoscopy by an endoscopist blindedto their previous treatment.
We acknowledge that bismuth subcitrate and antacids are effectiveulcer-healing agents that might have caused healing of the ulcersin the triple-therapy group (bismuth subcitrate, tetracycline,and metronidazole). Previous data have shown that 70 to 90 percentof ulcers heal when treated with bismuth for four to eight weeks.1,2,3Multiple mechanisms are involved in the ulcer-healing actionof bismuth, but suppression of H. pylori is one of the mostimportant.4 The relatively low relapse rate after treatmentwith bismuth as compared with H2 antagonists supports this hypothesis.1,2In our study, bismuth subcitrate was given for seven days aspart of the triple therapy. One week of bismuth subcitrate aloneis unlikely to achieve the high healing rate — 84.4 percent— noted in the antibacterial-treatment group in our study.
Antacid (Mylanta) was given freely to our patients for symptomaticrelief. We did not count the tablets consumed.
As Dr. Fagan observed, a number of the 95 percent confidenceintervals were not accurate. We have recalculated them usingthe exact method rather than one involving approximation (Table 1).5
Table 1. Corrected 95 Percent Confidence Intervals for the Results of Triple Therapy and Omeprazole Therapy, According to a Standard Analysis and an Intention-to-Treat Analysis.
Joseph J.Y. Sung, M.D. S.C. Sydney Chung, M.D. Man Yee Yung,B.N. Prince of Wales Hospital Shatin, N.T., Hong Kong
References
Bianchi Porro G, Petrillo M, De Nicola C, Lazzaroni M. A double-blind endoscopic study with De-Nol tablets and cimetidine for duodenal ulcer. Scand J Gastroenterol 1984;19:905-908. [Medline]
Martin DF, Hollanders D, May SJ, Ravenscroft MM, Tweedle DE, Miller JP. Differences in relapse rates of duodenal ulcer after healing with cimetidine or tripotassium dicitrato bismuthate. Lancet 1981;1:7-10. [CrossRef][Medline]
Ward M, Halliday C, Cowen AE. A comparison of colloidal bismuth subcitrate tablets and ranitidine in the treatment of chronic duodenal ulcers. Digestion 1986;34:173-177. [Medline]
Marshall BJ, Armstrong JA, Grancis GJ, Nokes NT, Wee SH. Antibacterial action of bismuth in relation to Campylobacter pyloridis colonization and gastritis. Digestion 1987;37:Suppl 2:16-30.
Fleiss JL. Statistical methods for rates and proportions. 2nd ed. New York: John Wiley, 1981.
Jeong, J.-Y., Berg, D. E.
(2000). Mouse-Colonizing Helicobacter pylori SS1 Is Unusually Susceptible to Metronidazole Due to Two Complementary Reductase Activities. Antimicrob. Agents Chemother.
44: 3127-3132
[Abstract][Full Text]
Correction
to
Sung et al., 
Previous
