The annual incidence of peptic ulcer disease in developed

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2003;1: Pretreatment Antimicrobial Susceptibility Testing Is Cost Saving in the Eradication of Helicobacter pylori MARCO ROMANO,* RICCARDO MARMO, ANTONIO CUOMO,* TERESA DE SIMONE,* CATERINA MUCHERINO,* MARIA ROSARIA IOVENE, FORTUNATO MONTELLA, MARIA ANTONIETTA TUFANO, CAMILLO DEL VECCHIO BLANCO,* and GERARDO NARDONE *Dipartimento di Internistica Clinica e Sperimentale, Cattedre di Gastroenterologia, and Microbiologia Clinica, C.I.R.A.N.A.D., Seconda Università di Napoli, Napoli; Cattedra di Gastroenterologia, Università Federico II, Napoli; and Servizio di Gastroenterologia, Ospedale di Polla (SA), Polla, Italy Background & Aims: The major obstacle to 100% effective eradication of Helicobacter pylori infection is represented by antimicrobial-resistant H. pylori strains. This randomized study was designed to evaluate whether regimens based on pretreatment susceptibility testing were more effective and cost saving compared with standard nonsusceptibility testing based therapy in the eradication of H. pylori infection. Methods: We studied 150 consecutive H. pylori infected dyspeptic subjects. Patients were randomly assigned to omeprazole 20 mg twice daily, clarithromycin 500 mg twice daily, and metronidazole 500 mg twice daily for 7 days or to omeprazole 20 mg twice daily and 2 antimicrobials chosen based on susceptibility testing. H. pylori status was reevaluated 12 weeks after the end of treatment by the 13 C-urea breath test. Results: Susceptibility testing based regimens led to the following results. (1) Eradication rates were 97.3% (95% confidence interval [CI ], 91.2% 99.5%) (71 of 73) and 94.6% (95% CI, 87.6% 98.3%) (71 of 75) in the per-protocol and intentionto-treat analysis, respectively. These were significantly higher (P < 0.005) than eradication rates obtained without susceptibility testing, that is, 79.4% (95% CI, 69.1% 87.6%) (58 of 73) and 77.3% (95% CI, 66.9% 85.7%) (58 of 75) in the per-protocol and intention-to-treat analyses, respectively. (2) There were savings of approximately $5 U.S. per patient compared with standard triple therapy. Conclusions: Pretreatment antimicrobial susceptibility testing is more effective and cost saving and, in health systems that confirm cost advantage, microbial susceptibility testing should be routinely used for eradication of H. pylori infection. The annual incidence of peptic ulcer disease in developed countries is approximately 1 3 per 1000 inhabitants. 1 Since the introduction of inhibitors of acid secretion, the indirect costs of this disease have steadily decreased, although direct costs have been increasing. Helicobacter pylori, one of the most frequent infections worldwide, is the major causative agent of chronic gastritis and peptic ulcer disease in humans. 2 The possibility of healing the patient and curing the disease through the eradication of H. pylori infection has the potential for a huge economic impact, considering the long-term cost/ benefit ratio. However, antibiotic-resistant H. pylori strains are becoming increasingly prevalent, 3 and this jeopardizes the success of therapeutic regimens aimed at the eradication of the infection. 4 Several studies in Europe have shown that the prevalence of resistance to metronidazole or clarithromycin (the most used antimicrobials in association with proton pump inhibitors) ranges from 10% to 50% or from 0% to 15%, respectively. 5,6 On the other hand, resistance to amoxicillin is uncommon. 5,6 Currently, most physicians treat H. pylori infection without relying on antimicrobial susceptibility testing to choose the best regimen. In a preliminary report in a different series of patients, we previously showed that specific therapy achieved a significantly lower treatment failure rate than that obtained with standard triple therapy in the eradication of H. pylori infection. 7 This study was therefore designed to (1) evaluate the eradication rate of 2 different therapeutic regimens, one based on susceptibility testing and the other not, in 150 consecutive H. pylori infected dyspeptic patients and (2) assess whether this approach to H. pylori eradication (i.e., choosing antimicrobials based on susceptibility testing) was cost saving. We found that therapy based on pretreatment antimicrobial susceptibility testing was more effective and resulted in savings of approximately $5 U.S. per patient compared with standard triple therapy. Abbreviation used in this paper: CI, confidence interval by the American Gastroenterological Association /03/$30.00 doi: /s (03)

2 274 ROMANO ET AL. CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 1, No. 4 Patients and Methods Patients and Study Protocol We studied 150 consecutive H. pylori infected patients with dyspeptic symptoms and without previous H. pylori eradication treatment. These patients were referred to our endoscopy unit from general practitioners because of a positive serology and/or 13 C-urea breath test for H. pylori. Patients were randomized according to a predetermined randomization schedule to receive standard triple therapy or susceptibility testing based therapy. Patients were randomly assigned with equal allocation to the treatment groups as determined by the treatment number. Treatment numbers were assigned in ascending order. Exclusion criteria included age younger than 15 years or older than 80 years, prior eradication treatment, treatment with inhibitors of acid secretion and/or antibiotics during the 6 weeks before the study, gastrointestinal malignancy, previous gastric surgery, severe concomitant diseases, history of allergy to any of the drugs used in the study, pregnancy or lactation, alcohol abuse, drug addiction, and long-term use of corticosteroids or nonsteroidal anti-inflammatory drugs. For histology, we took 2 biopsy specimens from the antrum, 2 from the body, and one from the angulus; diagnosis and grading of gastritis was made according to the updated Sydney system. 8 One biopsy specimen was taken from the antrum for a rapid urease test. In those patients who, according to the randomization protocol, were included in the pretreatment antimicrobial susceptibility testing based therapy group, an additional biopsy specimen was taken from the antrum for culture. Because the sensitivity of bacterial culture is not 100%, 9 in case of unsuccessful culture and testing, we decided, with patient consent, to perform a new endoscopy with antral and body biopsy to repeat culture and antimicrobial sensitivity testing. This was taken into consideration in the analysis of costs. Diagnosis of H. pylori infection was confirmed based on positivity to both histology and the rapid urease test. We did not perform susceptibility testing in those patients who had been randomized to standard triple therapy, because, in the case of in vitro resistance, it was not considered ethical to give patients drugs that based on in vitro testing might prove to be ineffective in vivo. Following randomization, 75 subjects (52 men and 23 women; age range, years; median age, 32 years) were treated with standard proton pump inhibitor based triple therapy and 75 subjects (49 men and 26 women; age range, years; median age, 34 years) were treated based on antimicrobial susceptibility testing results. Twelve weeks after completion of therapy, H. pylori status was evaluated by 13 C- urea breath test. Patients gave their written informed consent to enter the study, which was approved by the ethical committee of the Second University and the Federico II University of Naples, Italy. Bacterial Culture and Antimicrobial Susceptibility Testing All isolates identified as H. pylori were subcultured in Helicobacter-selective agar at 37 C in a microaerophilic atmosphere for 3 6 days. From each subculture, a suspension in Brucella broth (2 ml) supplemented with fetal bovine serum (2%), yielding a viable count of 10 8 to 10 9 colony-forming units/ml, equivalent to 3 4 McFarland turbidity standard, was prepared. From each suspension, 100 L was transferred onto the surface of a Mueller-Hinton agar plate with 5% sheep blood and streaked with a cotton swab. Subsequently, 2 antibiotic strips (E-test) were applied onto the surface of each dried agar plate. After incubation, in a microaerophilic atmosphere at 37 C for 3 6 days, the concentration shown in the E-test strip that was closest to the intersection point with growth on the plate (i.e., the lowest concentration of drug inhibiting visible bacterial growth) was defined as minimal inhibitory concentration. E-test strips containing amoxicillin, tetracycline, clarithromycin, and metronidazole were used. Therapeutic Regimens Patients treated with standard triple therapy received omeprazole 20 mg twice daily, clarithromycin 500 mg twice daily, and metronidazole 500 mg twice daily for 7 days. Patients whose eradication regimen was chosen based on the results of pretreatment antimicrobial susceptibility testing received omeprazole 20 mg twice daily, clarithromycin 500 mg twice daily, and metronidazole 500 mg twice daily for 7 days if H. pylori strain was sensitive to both clarithromycin and metronidazole. In the case of resistance to clarithromycin or metronidazole, we used amoxicillin 1 g twice daily plus metronidazole 500 mg twice daily or amoxicillin 1 g twice daily plus clarithromycin 500 mg twice daily, respectively. In the case of resistance to both clarithromycin and metronidazole, we used amoxicillin 1 g twice daily, tetracycline 500 mg 4 times daily, and bismuth subcitrate 125 mg 4 times daily for 14 days. In those patients who failed to eradicate the infection, we used as second-line therapy omeprazole 20 mg twice daily, tetracycline 500 mg 4 times daily, metronidazole 500 mg 3 times daily, and bismuth subcitrate 125 mg 4 times daily for 14 days. 10 Cost Analysis and Statistics Both intention-to-treat and per-protocol analyses were used for the assessment of the eradication rates of H. pylori infection in the 2 groups. One-way analysis of variance was used to compare continuous variables; dichotomous variables and proportions were evaluated by 2 test. The significance level was set at P The 95% confidence intervals (CI) were constructed by normal approximation. Cost analysis (i.e., standard error) was performed using the ICERconf.EXE software applying the natural logarithm of the cost variable, and new conversion, with exponential, was applied. The smearing retransformation effect 11 was considered, and the result was then recalculated to describe it in U.S. dollars. The

3 July 2003 TESTING FOR ERADICATION OF H. PYLORI 275 cost-effectiveness ratio and incremental cost-effectiveness ratio 12 were calculated. Costs are the charge estimates for office visits, endoscopy plus biopsy, rapid urease test, histology, culture and antimicrobial susceptibility testing, 13 C-urea breath test, and cost per tablet of omeprazole, amoxicillin, tetracycline, clarithromycin, metronidazole, and bismuth subcitrate. Cost estimates for procedures and physician fees were based on the Italian National Health System reimbursement scheme. Cost estimates for drugs were derived by averaging the list prices of the available brand names for each drug used. Data were calculated in Euros and transformed to U.S. dollars, considering that 1 Euro is roughly equal to $1 U.S. Results A total of 150 patients were enrolled into the randomized parallel-group study, with 75 patients in the susceptibility test group and 75 patients in the standard triple therapy group. Two patients in each group discontinued therapy because of side effects (diarrhea, nausea, and vomiting) and were therefore excluded from analysis. The rate of culture and testing for resistance was 72 of 75 (96%). A second endoscopy was performed in 3 patients, and the success rate for culture and testing was 3 of 3 (100%). Table 1. Endoscopic, Histologic, and Serologic Characteristics of Patients in the Pretreatment Antimicrobial Susceptibility Testing Group and in the Standard Triple Therapy Group Susceptibility testing group (%) Standard triple therapy group (%) Endoscopy Erosions 56 (74.6) 58 (77.3) Erythema 15 (20.0) 14 (18.6) Duodenal ulcer 4 (5.3) 3 (4.0) Histology Chronic gastritis Mild 43 (57.3) 42 (56.0) Moderate 19 (25.3) 21 (28.0) Severe 13 (17.3) 12 (16.0) Activity Mild 48 (64.0) 49 (65.3) Moderate 13 (17.3) 15 (20.0) Severe 13 (17.3) 11 (14.6) Atrophy Mild 8 (10.6) 5 (6.6) Moderate 3 (4.0) 2 (2.6) Severe 4 (5.3) 2 (2.6) Metaplasia Mild 7 (9.3) 6 (8.0) Moderate 3 (4.0) 2 (2.6) Severe 0 (0.0) 1 (1.3) H. pylori density Mild 29 (38.6) 33 (44.0) Moderate 33 (44.0) 28 (37.3) Severe 13 (17.3) 14 (18.6) Anti-CagA positive 55 (73.3) 52 (69.3) Table 2. Gradient and Break Point of Antibiotics Used for E-Test Antibiotic gradient ( g) Antibiotic break point ( g/ml) Metronidazole Amoxicillin Clarithromycin Tetracycline The endoscopic, histologic, and serologic findings in all 150 patients are summarized in Table 1. Both groups showed comparable (i.e., not statistically different) prevalence of endoscopic and histologic alterations. Also, the prevalence of anti-caga positivity was not statistically different in the 2 groups of patients (Table 1). The gradient and the break point of resistance for the antimicrobials used in the study are shown in Table 2. The prevalence of resistance of the H. pylori isolates obtained from the 75 patients who received a specific therapy based on the results of susceptibility testing is indicated in Table 3. No resistance to amoxicillin or tetracycline was found in any of the H. pylori isolates. The prevalence of resistance to clarithromycin and metronidazole was 12.5% and 22.5%, respectively. Two H. pylori isolates (i.e., 5%) were resistant to both clarithromycin and metronidazole. Pretreatment antimicrobial susceptibility testing was associated with 97.3% (95% CI, 91.2% 99.5%) (71 of 73) and 94.6% (95% CI, 87.6% 98.3%) (71 of 75) eradication rates in the per-protocol and intention-totreat analysis, respectively, which were significantly higher (P 0.005) than those obtained with standard nonsusceptibility testing based triple therapy, that is, 79.4% (95% CI, 69.1% 87.6%) (58 of 73) and 77.3% (95% CI, 66.9% 85.7%) (58 of 75) in the per-protocol and intention-to-treat analyses, respectively (Table 4). Two patients in the susceptibility testing group and 15 patients in the standard triple therapy group were not eradicated of the infection following the first-line therapy and were therefore retreated with a standard second-line therapy 10 with omeprazole, metronidazole, tetracycline, and bismuth subcitrate for 2 weeks. Both patients in the Table 3. Prevalence of Antimicrobial Resistance in the 75 H. pylori Isolates From Subjects in the Susceptibility Testing Group Antimicrobial Resistance (%) Amoxicillin 0/75 Tetracycline 0/75 Clarithromycin 8/75 (12.5) Metronidazole 16/75 (22.5) Clarithromycin plus metronidazole 3/75 (5)

4 276 ROMANO ET AL. CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 1, No. 4 Table 4. H. pylori Eradication in the Standard Therapy Group and in the Susceptibility Testing Group After First-Line Treatment Eradication rate Per protocol (95% CI) Intention to treat (95% CI) Susceptibility testing 71/ % a (91.2% 99.5%) 71/ % a (87.6% 98.3%) Standard therapy 58/ % (69.1% 87.6%) 58/ % (66.9% 85.7%) a P vs. standard therapy. susceptibility testing group and 11 of 15 patients in the standard triple therapy group were eradicated of the infection at the end of the second cycle of treatment. Therefore, the overall eradication rates (i.e., at the end of first- plus second-line therapies) were 100% (95% CI, 95.0% 100%) versus 94.5% (95% CI, 86.5% 98.5%) in the per-protocol analysis and 97.0% (95% CI, 90.3% 100%) versus 92.0% (95% CI, 83.0% 98.0%) in the intention-to-treat analysis in the susceptibility testing group and in the standard therapy group, respectively (Table 5). However, this difference did not reach statistical significance. We then evaluated whether pretreatment antimicrobial susceptibility testing, besides being more effective in the eradication of H. pylori infection, was also cost saving. Taking into account a number of variables, including cost estimates for office visits, endoscopy plus biopsy, rapid urease test, histology, culture and antimicrobial susceptibility testing, 13 C-urea breath test, and cost per tablet of omeprazole, amoxicillin, tetracycline, clarithromycin, metronidazole, and bismuth subcitrate (Table 6), eradication regimens based on the results of antimicrobial susceptibility testing saved approximately $5 U.S. per patient. This, corrected by the effectiveness, led to a cost-effectiveness ratio of approximately $15 U.S. lower than that achieved by standard therapy. Moreover, the incremental cost-effectiveness ratio (i.e., the ratio defined as the difference in average per-patient costs divided by the corresponding difference in effectiveness fractions) was between $87 and $120 U.S. in the per-protocol and intention-to-treat analyses, respectively (i.e., H. pylori eradication with standard triple therapy costs an additional $87 $120 U.S. compared with that achieved through a regimen based on pretreatment antimicrobial susceptibility testing) (Table 7). Discussion H. pylori infection is a chronic, transmissible infectious disease that causes gastritis and peptic ulcer and is implicated in the development of carcinoma of the distal stomach. 13,14 Successful therapy requires a combination of drugs that prevents the emergence of resistance and reaches the bacteria within its various niches. The ideal eradication regimen would be one that is simple (e.g., few medications and short duration) and essentially 100% effective. Regimens that have been proven effective consist of combinations of proton pump inhibitor or ranitidine bismuth citrate and 2 antimicrobials (amoxicillin/clarithromycin, metronidazole/clarithromycin, amoxicillin/metronidazole) for 7 14 days (e.g., proton pump inhibitor or ranitidine bismuth citrate based triple therapy). The overall cure rates in clinical practice using these regimens range from 60% to 86%, with the major obstacle to 100% effective therapy represented by antimicrobial-resistant H. pylori strains, the prevalence of which is increasing. In fact, the prevalence of H. pylori strains with primary resistance to clarithromycin or metronidazole is approximately 10% and 25%, respectively. 3 6 The utility of culture (with consequent antimicrobial susceptibility testing) by providing a knowledge of the organism s antimicrobial susceptibility might aid the selection of the therapy regimen. We hypothesized that choosing the treatment regimen based on pretreatment antimicrobial susceptibility testing might significantly decrease the number of treatment Table 5. H. pylori Eradication in the Standard Therapy Group and in the Susceptibility Testing Group After Second-Line Treatment Eradication rate Per protocol (95% CI) Intention to treat (95% CI) Susceptibility testing 73/73 100% (95.0% 100%) 73/ % (90.3% 100%) Standard therapy 69/ % (86.5% 98.5%) 69/ % (83.0% 98.0%)

5 July 2003 TESTING FOR ERADICATION OF H. PYLORI 277 Table 6. Costs Used in Analysis Cost item Average charge ($) First office visit Subsequent office visits Endoscopy plus biopsy Omeprazole 20 mg 1.03 Amoxicillin 1 g 0.55 Clarithromycin 500 mg 3.3 Metronidazole 250 mg 0.13 Tetracycline 250 mg 0.19 Bismuth subcitrate 125 mg 0.28 Culture and susceptibility testing Histology C-urea breath test Rapid urease test 5.8 failures. Moreover, we hypothesized that, even with the supplementary cost of culture and testing, this approach might be cost saving and decrease the need for retreatment. We therefore prospectively recruited 150 H. pylori infected dyspeptic patients who were randomly assigned to a standard nonsusceptibility testing based triple therapy or to a specific therapy based on the results of susceptibility testing. Our study shows that, both in the per-protocol and intention-to-treat analyses, pretreatment antimicrobial susceptibility testing led to a significantly higher eradication rate compared with that obtained with standard triple therapy. In fact, only 2 patients (who were infected with an H. pylori strain resistant to clarithromycin) in the susceptibility testing group as opposed to 15 in the standard triple therapy group did not have eradication of the infection after the first-line treatment. This result is in agreement with previous reports by our group in a different series of patients 7 and by Toracchio et al., who found eradication rates comparable to those found in this study when using a personalized therapy prescribed on the basis of the results of susceptibility testing. 15 A second-line treatment with omeprazole, metronidazole, tetracycline, and bismuth subcitrate for 14 days eradicated the infection in the 2 patients in the susceptibility testing group and in 11 of 15 patients in the standard triple therapy group. The overall eradication rate at the end of 2 treatments was approximately 6% greater in the susceptibility testing group compared with the standard triple therapy group, even though this did not reach statistical significance. A cost analysis, which was performed while taking into account the cost estimates for a number of variables, including office visits, endoscopy plus biopsy, rapid urease test, histology, culture and antimicrobial susceptibility testing, 13 C-urea breath test, and cost per tablet of omeprazole, amoxicillin, tetracycline, clarithromycin, metronidazole, and bismuth subcitrate, showed that eradication regimens based on the results of antimicrobial susceptibility testing saved up to $5 U.S. per patient. The $5 U.S. per patient savings in the susceptibility testing group was accounted for by the decreased number of visits and the reduced costs related to drugs and the 13 C-urea breath test. When this was normalized to the effectiveness of the different treatments, H. pylori eradication obtained through pretreatment susceptibility testing based therapy resulted in a cost-effectiveness ratio of approximately $15 U.S. per patient lower compared with that achieved by standard triple therapy. In addition, by calculating the incremental cost-effectiveness ratio, it appeared that H. pylori eradication based on pretreatment antimicrobial susceptibility testing was approximately $87 $120 U.S. less expensive than nonsusceptibility-based therapy. The results of our randomized, prospective study are comparable to that obtained by Breuer and Graham who, by using a decision model, found that endoscopy plus biopsy followed by antimicrobial susceptibility testing and tailored antibiotic treatment would save approximately $37 U.S. per patient compared with endoscopy plus biopsy followed by empiric antibiotic treatment of H. pylori infected patients with ulcer. 16 H. pylori resistance to antimicrobials can be either primary (i.e., existing before therapy) or secondary (i.e., developing as the result of failed therapy). The incidence of secondary resistance has been dependent on the type of primary eradication therapy. Clarithromycin resistance may develop in those who used clarithromycin, 3 whereas metronidazole resistance may develop in those who used metronidazole. 17 Recently, Pilotto et al. found that the incidence of secondary H. pylori resistance to metronidazole and/or clarithromycin in treatment failures after Table 7. Results of Base-case Cost-effectiveness Analysis Cost (mean SE) ($) Effectiveness (%) Cost-effectiveness ($) ICER ($) Strategy PP ITT PP ITT PP ITT PP ITT Susceptibility testing Standard therapy PP, per protocol; ITT, intention to treat; ICER, incremental cost-effectiveness ratio.

6 278 ROMANO ET AL. CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 1, No. 4 1-week proton pump inhibitor based triple therapies was approximately 70%. 18 Therefore, one of the main issues today in H. pylori therapy is to decrease the number of eradication failures as much as possible. 19 We postulate that the significantly lower rate of treatment failures observed following first-line therapy in the susceptibility testing group compared with standard triple therapy might contribute to prevention of the emergence of bacterial secondary resistance. In conclusion, regimens based on pretreatment antimicrobial susceptibility testing were associated with a significantly higher eradication rate in the first-line treatment of H. pylori infected dyspeptic patients. More importantly, at the end of first- and second-line treatment, specific therapy based on susceptibility testing was cost saving. We postulate that, because antimicrobial-resistant strains are becoming increasingly prevalent, therapeutic regimens based on susceptibility testing should always be used as first-line therapy, especially in areas with high prevalence of resistance to clarithromycin and metronidazole. 16 This approach to H. pylori infection is more effective and cost saving and might help prevent the emergence of secondary resistance. References 1. Deltenre MA. Economics of Helicobacter pylori eradication therapy. Eur J Gastroenterol Hepatol 1997;9(suppl 1):S27 S Zarrilli R, Ricci V, Romano M. Molecular response of gastric epithelial cells to Helicobacter pylori-induced cell damage. Cell Microbiol 1999;1: Megraud F. Epidemiology and mechanism of antibiotic resistance in Helicobacter pylori. Gastroenterology 1998;115: Graham DY. Antibiotic resistance in Helicobacter pylori: implications for therapy. Gastroenterology 1998;115: Iovene MR, Romano M, Pilloni AO, Giordano B, Montella F, Caliendo S, Tufano MA. Prevalence of antimicrobial resistance in eighty clinical isolates of Helicobacter pylori. Chemotherapy 1999;45: Debets-Ossenkopp YJ, Herscheid AJ, Pot RG, Kuipers EJ, Kusters JG, Vanderbroucke-Grauls CM. Prevalence of Helicobacter pylori resistance to metronidazole, clarithromycin, amoxicillin, tetracycline and trovafloxacin in The Netherlands. J Antimicrob Chemother 1999;43: Romano M, Iovene MR, Montella F, Vitale LM, De Simone T, Del Vecchio Blanco C. Pretreatment antimicrobial-susceptibility testing in the eradication of Helicobacter pylori infection (letter). Am J Gastroenterol 2000;95: Dixon MF, Genta RM, Yardley JH, Correa P. Participants in the International Workshop on the Histopathology of Gastritis, Houston 1994: classification and grading of gastritis. The updated Sidney system. Am J Surg Pathol 1996;20: Working Party of the European Helicobacter pylori Study Group. Technical annex: tests used to assess Helicobacter pylori infection. Gut 1997;41(Suppl 2):S10 S Malfertheiner P, Megraud F, O Morain C. Current concepts in the management of Helicobacter pylori infection. The Maastricht Consensus Report. Aliment Pharmacol Ther 2002;16: Duan N. Smearing estimate: a nonparametric retransformation method. J Am Stat Assoc 1983;78: Black WC. The CE plane: a graphic representation of cost-effectiveness. Med Decis Making 1990;10: Ricci V, Zarrilli R, Romano M. Voyage of Helicobacter pylori in human stomach: odyssey of a bacterium. Dig Liver Dis 2002;34: Blaser MJ. Hypothesis: the changing relationships of Helicobacter pylori and humans: implications for health and disease. J Infect Dis 1999;179: Toracchio S, Cellini L, Di Campli E, Cappello G, Malatesta MG, Ferri A, Ciccaglione AF, Grossi L, Marzio L. Role of antimicrobial susceptibility testing on efficacy of triple therapy in Helicobacter pylori eradication. Aliment Pharmacol Ther 2000;14: Breuer T, Graham DY. Costs of diagnosis and treatment of Helicobacter pylori infection: when does choosing the treatment regimen based on susceptibility testing become cost-effective? Am J Gastroenterol 1999;94: van der Wouden EJ, Thijs JC, van Zwet AA, Kleibeuker JH. Review article: nitroimidazole resistance in Helicobacter pylori. Aliment Pharmacol Ther 2000;14: Pilotto A, Franceschi M, Rassu M, Leandro G, Bozzola L, Furlan F, Di Mario F. Incidence of secondary Helicobacter pylori resistance to antibiotics in treatment failures after 1-week proton pump inhibitor-based triple therapies: a prospective study. Dig Liver Dis 2000;32: De Boer WA, Borody TJ. Treatment failures and secondary resistance to antibiotics. A growing concern in Helicobacter pylori therapy. Dig Liver Dis 2000;32: Address requests for reprints to: Marco Romano, M.D., Dipartimento di Internistica Clinica e Sperimentale-Cattedra di Gastroenterologia, C/o II Policlinico, Edificio 3, Secondo piano, Via Pansini 5, Napoli. marco.romano@unina2.it; fax: (39) Supported in part by grants from Ministero per la Istruzione, Università e Ricerca, Consiglio Nazionale delle Ricerche, and C.I.R.A.N.A.D., Seconda Università di Napoli. The authors thank Mr. A. Piccolo (endoscopy nurse) for assistance.