Double-Blind Randomized Trial of Bismuth Subsalicylate and Clindamycin for Treatment of Helicobacter p y Zori Infection T. U. WESTBLOM, E. MADAN, M. A . SUBIK, D. E. DURIEX & B. R. MIDKIFF Dept. of Medicine, Division of Infectious Diseases, St. Louis University School of Medicine, St. Loui5, Missouri, and Dept. of Pathology and Dept. of Medicine, Sections of Gastroenterology and of Infectious Diseases, Marshall University School of Medicine, Huntington, West Virginia, USA
Scand J Gastroenterol Downloaded from informahealthcare.com by McMaster University on 12/29/14 For personal use only.
Westblom TU, Madan E, Subik MA, Duriex DE, Midkiff BR. Double-blind randomized trial of bismuth subsalicylate and clindamycin for treatment of Helicobacrerpylori infection. Scand J Gastroenterol 1992, 27, 249-252 We evaluated clindamycin and bismuth subsalicylate (Pepto-Bismol@) for treatment of Helicobacter pylori infection. Patients with culture or histology positive for H . pylori were randomized to receive two tablets of bismuth subsalicylate four times daily for 4 weeks or bismuth combined with 2 weeks of 300 mg clindamycin four times daily. Clinical symptoms were recorded before and after treatment by means of visual analog scales. Patients in both treatment arms showed improvement in clinical scores for abdominal pain, heartburn, and gas or bloating. Microbiologic cure was achieved in only 1 of 11 patients treated with bismuth alone and in none of 7 treated with bismuth/clindamycin. Successful eradication of H. pylori may require combination of multiple antibiotics, as recommended at the IXth World Congress of Gastroenterology, or pharmacokinetic modulators such as H2-blockers or omeprazole.
Key words: Antibiotics: bismuth; clindamycin; Helicobacterpylori; H,-receptor antagonists; omeprazole; pharrnacokinetics T. Ulf Westblom, M . D., Diuision of Infectious Diseases and Immunology, St. Louis Unioer.~itySchool of Medicine, 1402 S. Grand Blud., St. Louis, MO 63104, USA
Helicobacter pylori, formerly Campylobacter pylori (1). is strongly associated with gastritis (2,3) and peptic ulcer disease (4). The organism is susceptible to multiple antibiotics in vitro (5-7), but clinical trials using traditional antibiotic regimens have often been disappointing (8) unless combinations of multiple antibiotics are used (9). Part of the problems encountered may be due to the unique environment in the stomach. Few antibiotics retain their activity at a low p11. The pharrnacokinetics in gastric tissue are poorly understood, and antibiotic concentrations in gastric mucosa are generally not known. In a previous study we investigated gastric tissue concentrations of clindamycin in an animal model (10). Clindamycin is active at a low pH (ll), and we found that it also concentrates in gastric mucosa, with tissue levels exceeding known MICgoof H . pylori (10). The objectives of the current study were to evaluate clindamycin’s ability to eradicate H . p,ylori infection in a population of dyspeptic patients and to assess its effect on their clinical symptoms.
MATERIALS AND METHODS Population During a 12-month period we screened all patients referred for elective gastroduodenoscopy at the VA Medical Center, Huntington, West Virginia, for infection with H. pylori. All the patients were male and had a diagnosis of
dyspepsia. At the time of endoscopy we collected biopsy specimens for culture and Giemsa staining. Patients with positive cultures and/or H . pylori seen on histology were eligible for enrollment. All study patients had received a diagnosis of antral gastritis, and none were taking antibiotics at the time. We anticipated eradication rates for the combination treatment to be more than SO% and for bismuth alone to be around 20%. Original sample size was SO subjects, but a non-blinded observer was allowed to analyze the data and recommend earlier termination of the study if statistically justified. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Boards at Marshall University School of Medicine and the VA Medical Center. All patients gave informed consent before entering the study. Culture methods Biopsy specimens from the antrum were transported to the laboratory in a biphasic transport medium as previously described (12). They were ground in glass tissue grinders and spread on H . pylori-selective media (Brucella agar with 10% bovine blood and 1% IsoVitaleX; nalidixic acid, 20 mg/ I, vancomycin, 6 mg/l; amphotericin B, 8 mg/l; and cycloheximide, lOOmg/l). Agar plates were incubated in CampyPak jars (BBL Microbiology Systems, Cockeysville, Md., USA) at 37°C. We identified H . pylori on the basis of characteristic curved morphology and positive reactions for urease, catalase, and oxidase.
250
T. 17. Westhlom et a1
Scand J Gastroenterol Downloaded from informahealthcare.com by McMaster University on 12/29/14 For personal use only.
Histologic examination Antral biopsy specimens were stained with a modified Giemsa technique (13). All slides were examined by the same pathologist, who was unaware of the results of culture. For identification purposes we defined H.pylori as spiralshaped bacteria with characteristic morphology (13). Questionnaire A t each visit we interviewed the patients about current symptoms of abdominal pains, heartburn, gas or bloating, nausea, vomiting, and diarrhea. These were recorded by means of visual analog scales (14, IS). The patients were asked to mark on a 100-mm line the extent to which they had each symptom. The ends of the line represented the two extremes, with 0 being the least (none at all) and 100 the most (all the time). On return visits they were not allowed to inspect their previous questionnaires. W e expressed differences between pretreatment and posttreatment answers as changes in millimeters along the 100-mm line. Patients were also interviewed about current medications, including the use of I f , antagonists in the preceding 2 weeks.
100
100
Therapy Patients were randomized to receive one of two regimens: two bismuth subsalicylate (Pepto-BismolB) tablets four times daily for 4 weeks with the addition of 300 mg clindamycin four times daily during the last 2 weeks, o r two bismuth subsalicylate tablets four times daily for 4 weeks with 2 weeks of placebo capsules four times daily. Randomization in groups of 10 was done by computer. W e did not perform posttreatment endoscopies until at least 4 weeks had elapsed since the end of therapy. Throughout the study both patients and investigators were blinded to which regimen had been used. Only the pharmacist and the non-blinded observer had access to the randomization code. RESULTS Although the study was double-blind, it soon became apparent that both regimens were going to be failures. After I8 patients had been evaluated, only 1 was cleared of ff. pylori. Eleven patients had been treated with bismuth subsalicylatc only, and seven with the combination that included clin-
100
100
90
90
90
90
80
80
80
80
:B 70
70
70
70
60
a0
a0
60
60 40
20
30
30
20
20
1
10
0
100
100
4
80
60 701
90
90
80
80
80
70
70
70
ILO
80
60
60
50
40
40
30
30‘*
20
20
c l:I
100
t?O 20
440 J’O 0
Fig. I . Average visual analog scores before (left y-axis) and after (right y-axis) treatment with bismuth/clindamycin (filled circles) and bismuth alone (open circlcs). Zero represents the least, and 100 the most discomfort from abdominal pain (A), heartburn (B), gas or bloating ( C ) ,and nausea or vomiting (D). Both treatment arms show a similar reduction in symptom scores, except for nausea or vomiting, which remained unchanged in both groups regardless of therapy.
Scand J Gastroenterol Downloaded from informahealthcare.com by McMaster University on 12/29/14 For personal use only.
Clindamycin and Bismuth for H. pylori
damycin. The mean ages in the two study groups were not significantly different (62.9 versus 60.4). The single patient with H . pylori eradication had been given only bismuth subsalicylate. Statistical analysis showed that the sample size in the clindamycin group was large enough to detect a difference in eradication rate of 0% versus 50% at a significance level of 0.05 and a type-I1 error of 0.2 (16). The observed frequency of eradication (0%) was significantly different from the hypothesized frequency (>50%) by Fisher’s exact test ( P = 0.03). All patients reported improvement in their symptom scores after treatment. Average scores before and after treatment are plotted in Fig. 1. For abdominal pain, heartburn, gas or bloating, all patients had a similar drop in scores regardless of which treatment arm they belonged to. The slope of the curve was almost identical for both groups (Fig. IA-C). Nausea and vomiting were rare in both treatment groups and remained unchanged regardless of the therapy given (Fig. 1D). When the patients were asked how much they had benefited from the treatment, the score was 80 in the bismuth group and 82 in the bismuth/clindamycin group. Neither group had any problems with diarrhea. Pre- and post-treatment diarrhea scores in the bismuth group were 7 and 4 and in the bismuth/clindamycin group 12 and 17, respectively. The patient who was cured of his infection did not differ from the other patients in terms of symptom relief except in the category of abdominal pain. He went from a score of 85 down to 4 after treatment, which was the largest improvement seen in any of the patients.
251
clinical scores suggests that most of our patients consumed their medications as prescribed. The design of this study allowed the patients to take any drugs their primary physician had ordered, unless it was an antibiotic. We made no attempts to influence the use of acidreducing drugs such as H2-receptor antagonists. We have since extended our original animal experiments to evaluate also the effects of cimetidine on antibiotic concentrations in gastric mucosa (23). H2 blockers can increase these levels up to tenfold through a change in the volume of distribution of the antibiotic. Unfortunately, our patients did not benefit from this information. Only two patients consumed any H2 blockers during the study, and neither of them had been randomized to the clindamycin arm. We conclude that our study is yet another testimony to the ineffectiveness of single or dual drug regimens in treating H . pylori infection. Other slow-growing organisms, such as mycobacteria, require combination treatments for clinical cure. We may have to accept that similar strategies are needed for H. pylori. The IXth World Congress of Gastroenterology issued a recommendation to use a combination of bismuth, tetracycline, and metronidazole as the therapy of choice (24). The ultimate regimen may, however, combine multiple antibiotics with acid-reducing agents such as Hz blockers or omeprazole. The animal model suggests that this leads to very favorable pharmacokinetics for the antibiotics (23), and a preliminary report by Unge et al. (25) has given further support to that approach. ACKNOWLEDGEMENTS
DISCUSSION On the basis of our own animal data (10) and reported synergy studies between clindamycin and bismuth (17) we anticipated most of our patients to be cured of their infections. Instead, both treatment arms were failures in terms of eradicating H . pylori. They only showed a therapeutic effect on clinical symptoms. All patients reported some improvement in their symptom scores. The identical slopes of the curves in Fig. 1 suggest that clindamycin contributed very little to this clinical improvement. Although we cannot rule out a placebo effect, we suspect that bismuth subsalicylate was primarily responsible. Clinical improvement by bismuth compounds in spite of failure to eradicate H . pylori has been noted previously by other investigators (18-2 1) . We can only speculate as to why clindamycin performed poorly in this trial. Antibiotic sensitivity data are available for four of the posttreatment isolates. Three of them were still sensitive to clindamycin after treatment. It therefore seems unlikely that development of resistance played any major role. Glupczynski et al. (22) have shown that compliance can be a problem when treating H. pylori gastritis for 4 weeks. We have no way of knowing whether our patients were compliant, but the general improvement in
This study was sponsored in part by the Upjohn Company. Parts of the study have previously been presented in preliminary form at the Second Meeting of the European Helicobacter pylori Study Group, Ulm, Germany, 12-14 October, 1989. REFERENCES 1. Goodwin CS, Armstrong JA, Chilvers T, et al. Transfer of Campylobacter pylori and Campylobacter mustelae to Helicobacter gen. nov. as Helicobacter pylori comb. nov. and Helicobacter mustelae comb. nov., respectively. Int J Syst Bacteriol 1989, 39, 397-405 pathogenetic 2. Wyatt JI, Dixon MF. Chronic gastritis-a approach. J Pathol 1988, 154, 113-124 3. Wyatt JI, Rathbone BJ, Dixon MF, Heatley RV. Campylobacter pyloridis and acid induced gastric metaplasia in the pathogenesis of duodenitis. J Clin Pathol 1987, 40, 841-848 4. Goodwin CS. Duodenal ulcer, Campylobacter pylori, and the ‘leaking roof‘ concept. Lancet 1988, 2, 1467-1469 5. Goodwin CS, Blake P, Blincow E. The minimum inhibitory and bactericidal concentrations of antibiotics and anti-ulcer agents against Campylobacter pyloridis. J Antimicrob Chemother 1986, 17, 309-314 6. Lambert T, Megraud F, Gerbaud G, Courvalin P. Susceptibility of Campylobacter pyloridis to 20 antimicrobial agents. Antimicrob Agents Chemother 1986, 30, 510-511 7. Westblom TU, Gudipati S , Midkiff BR. In vitro susceptibility of Helicobacter pylori to the new oral cephalosporins cefpodoxime, ceftibuten and cefixime. Eur J Clin Microbiol Infect Dis 1990, 9, 691-693
252
7: U. Wesiblorn el ul.
Scand J Gastroenterol Downloaded from informahealthcare.com by McMaster University on 12/29/14 For personal use only.
8. Axon ATR. Carnpylobacter pylori-therapy review. Scand J Gastroenterol 1989. 24 Suppl 160, 35-38 9. Borody TJ, Cole P, Noonan S, et al. Recurrence of duodcnal
ulcer and Carnpylobactcr pylori infcction after eradication. Med J Aust 1989, 151, 431-435 10. Westhlom TLJ, Duriex DE, Madan E, Belshc RB. Guinea pig model for antibiotic transport across gastric mucosa: inhibitory tissue concentrations of clindamycin against Helicobacter pylori (Campylobacter pylori) following two separate dose regimens. Antimicrob Agent$ Chernother 1990, 34, 25-28 11. Klempner MS, Styrt B. Alkalinization of the intralysosornal pH by clindarnycin and its cffect on neutrophil function. J Antirnicrob Chernothcr 19x3, 12 Suppl C. 39-50 12. Wcstblom TU. Midkiff BR, Madan E, Subik MA. Transportation of gastric biopsies in a biphasic Cdrnpylohacter pylori media. J Clin Castroenterol 1989. 1 1 , 483-484 13. Madan E, Kemp J , Westblom TU, Subik M, Sexton S,Cook J . Evaluation of staining methodologics lor identifying Campylobactcr pylori. Am J Clin Pathol 1988. 90, 450-453 14. Aabakken I-, 1,iirsen S, Osnes M. Visual analogue scales lor endoscopic evaluation of nonsteroidal anti-inflammatory druginduced mucosal damage in the stomach and duodenum. Scand J Gastroenterol 1990, 25, 443-448 15. Czinn SJ, Bcrtrarn TA, Murray PD, Yang P. Relationship between gastric inflammatory response and symptoms in patients infected with Helicobacirrpylori. Scand J Gastroenterol 1991, 26 Suppl 181, 33-37 16. Casagrande JT. Pike MC, Smith PG. An improvcd approximate formula for calculating sample sizes for comparing two binomial distributions. Bionictrics 1978, 34, 483-486 17. Vogt K, Hahn H. Synergism between clindamycin and colloidal Received 12 August 1991 Acceptcd 23 October 1991
bismuth subcitrate against Helicobacter (Campylobactcr) pylori in vitro. Zentralbl Bakteriol 1990, 274, 246-249 18. McNulty CA, Gearty JC, Crump 8, et 81. Campylobacter pyloridis and associated gastritis: investigator blind, placebo controlled trial of bismuth salicylate and erythromycin ethylsuccinate. Br Med J 1986, 293, 645-649 19. Stanescu A , Malfertheiner P, M a y a D, Baczako K , Ditschuneit H. Bismuth aluminate in gastroenterology. Therapeutic effects in chronic erosive Campylobacter pylori-associated gastritis. Fortschr Med 1989, 107, 623-626 20. Dobronte Z, Lakatos F, Brittig F, et al. Has Campylobacter pylori infection any clinical relevance? Methodologic, epidemiologic and clinical studies. Orv I Ictil 19x9. 130. 2563-2568 21. Malfertheiner P, Stanescu A , Baczako K, Mayer D, Ditschuneit H. Efficacy of a bismuth cornhination preparation. Efficacy in the treatment of chronic active gastritis and non-ulcerous dyspepsia. Fortschr Med 1990, 108, 402-406 22. Glupczynski Y, Nyst JF, Burette A, Vanderlindcn MP, Dcltcnrc M. Lack of antibiotic compliance in patients treated for Campylobacter pylori-associated gastritis [Abstract 761. 'I'hc Vth International Workshop on Campylobacter Infections, Puerto Vallarta, Mexico, I989 23. Westblom TU, Duriex DE. Enhancement of antibiotic concentrations in gastric niucosa by H2 receptor antagonist: implications lor treatment of Helicohacter pylori infections. Dig Dis Sci 1991, 36, 25-28 24. Axon ATR. Hclicobacter pylori therapy: cffect on peptic ulcer disease. J Gastroenterol Hepatol 1991, 6 , 131-137 25. Unge P, Gad A, Gnarpe H, Olsson J . Does orneprazolt. improve antimicrobial therapy directed towards gastric Campylobacter pylori in patients with antral gastritis? Scand J Gastroentcrol 1989, 24 SUPPI 161, 49-54
Scand J Gastroenterol Downloaded from informahealthcare.com by McMaster University on 12/29/14 For personal use only.
Westblom TU, Madan E, Subik MA, Duriex DE, Midkiff BR. Double-blind randomized trial of bismuth subsalicylate and clindamycin for treatment of Helicobacrerpylori infection. Scand J Gastroenterol 1992, 27, 249-252 We evaluated clindamycin and bismuth subsalicylate (Pepto-Bismol@) for treatment of Helicobacter pylori infection. Patients with culture or histology positive for H . pylori were randomized to receive two tablets of bismuth subsalicylate four times daily for 4 weeks or bismuth combined with 2 weeks of 300 mg clindamycin four times daily. Clinical symptoms were recorded before and after treatment by means of visual analog scales. Patients in both treatment arms showed improvement in clinical scores for abdominal pain, heartburn, and gas or bloating. Microbiologic cure was achieved in only 1 of 11 patients treated with bismuth alone and in none of 7 treated with bismuth/clindamycin. Successful eradication of H. pylori may require combination of multiple antibiotics, as recommended at the IXth World Congress of Gastroenterology, or pharmacokinetic modulators such as H2-blockers or omeprazole.
Key words: Antibiotics: bismuth; clindamycin; Helicobacterpylori; H,-receptor antagonists; omeprazole; pharrnacokinetics T. Ulf Westblom, M . D., Diuision of Infectious Diseases and Immunology, St. Louis Unioer.~itySchool of Medicine, 1402 S. Grand Blud., St. Louis, MO 63104, USA
Helicobacter pylori, formerly Campylobacter pylori (1). is strongly associated with gastritis (2,3) and peptic ulcer disease (4). The organism is susceptible to multiple antibiotics in vitro (5-7), but clinical trials using traditional antibiotic regimens have often been disappointing (8) unless combinations of multiple antibiotics are used (9). Part of the problems encountered may be due to the unique environment in the stomach. Few antibiotics retain their activity at a low p11. The pharrnacokinetics in gastric tissue are poorly understood, and antibiotic concentrations in gastric mucosa are generally not known. In a previous study we investigated gastric tissue concentrations of clindamycin in an animal model (10). Clindamycin is active at a low pH (ll), and we found that it also concentrates in gastric mucosa, with tissue levels exceeding known MICgoof H . pylori (10). The objectives of the current study were to evaluate clindamycin’s ability to eradicate H . p,ylori infection in a population of dyspeptic patients and to assess its effect on their clinical symptoms.
MATERIALS AND METHODS Population During a 12-month period we screened all patients referred for elective gastroduodenoscopy at the VA Medical Center, Huntington, West Virginia, for infection with H. pylori. All the patients were male and had a diagnosis of
dyspepsia. At the time of endoscopy we collected biopsy specimens for culture and Giemsa staining. Patients with positive cultures and/or H . pylori seen on histology were eligible for enrollment. All study patients had received a diagnosis of antral gastritis, and none were taking antibiotics at the time. We anticipated eradication rates for the combination treatment to be more than SO% and for bismuth alone to be around 20%. Original sample size was SO subjects, but a non-blinded observer was allowed to analyze the data and recommend earlier termination of the study if statistically justified. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Boards at Marshall University School of Medicine and the VA Medical Center. All patients gave informed consent before entering the study. Culture methods Biopsy specimens from the antrum were transported to the laboratory in a biphasic transport medium as previously described (12). They were ground in glass tissue grinders and spread on H . pylori-selective media (Brucella agar with 10% bovine blood and 1% IsoVitaleX; nalidixic acid, 20 mg/ I, vancomycin, 6 mg/l; amphotericin B, 8 mg/l; and cycloheximide, lOOmg/l). Agar plates were incubated in CampyPak jars (BBL Microbiology Systems, Cockeysville, Md., USA) at 37°C. We identified H . pylori on the basis of characteristic curved morphology and positive reactions for urease, catalase, and oxidase.
250
T. 17. Westhlom et a1
Scand J Gastroenterol Downloaded from informahealthcare.com by McMaster University on 12/29/14 For personal use only.
Histologic examination Antral biopsy specimens were stained with a modified Giemsa technique (13). All slides were examined by the same pathologist, who was unaware of the results of culture. For identification purposes we defined H.pylori as spiralshaped bacteria with characteristic morphology (13). Questionnaire A t each visit we interviewed the patients about current symptoms of abdominal pains, heartburn, gas or bloating, nausea, vomiting, and diarrhea. These were recorded by means of visual analog scales (14, IS). The patients were asked to mark on a 100-mm line the extent to which they had each symptom. The ends of the line represented the two extremes, with 0 being the least (none at all) and 100 the most (all the time). On return visits they were not allowed to inspect their previous questionnaires. W e expressed differences between pretreatment and posttreatment answers as changes in millimeters along the 100-mm line. Patients were also interviewed about current medications, including the use of I f , antagonists in the preceding 2 weeks.
100
100
Therapy Patients were randomized to receive one of two regimens: two bismuth subsalicylate (Pepto-BismolB) tablets four times daily for 4 weeks with the addition of 300 mg clindamycin four times daily during the last 2 weeks, o r two bismuth subsalicylate tablets four times daily for 4 weeks with 2 weeks of placebo capsules four times daily. Randomization in groups of 10 was done by computer. W e did not perform posttreatment endoscopies until at least 4 weeks had elapsed since the end of therapy. Throughout the study both patients and investigators were blinded to which regimen had been used. Only the pharmacist and the non-blinded observer had access to the randomization code. RESULTS Although the study was double-blind, it soon became apparent that both regimens were going to be failures. After I8 patients had been evaluated, only 1 was cleared of ff. pylori. Eleven patients had been treated with bismuth subsalicylatc only, and seven with the combination that included clin-
100
100
90
90
90
90
80
80
80
80
:B 70
70
70
70
60
a0
a0
60
60 40
20
30
30
20
20
1
10
0
100
100
4
80
60 701
90
90
80
80
80
70
70
70
ILO
80
60
60
50
40
40
30
30‘*
20
20
c l:I
100
t?O 20
440 J’O 0
Fig. I . Average visual analog scores before (left y-axis) and after (right y-axis) treatment with bismuth/clindamycin (filled circles) and bismuth alone (open circlcs). Zero represents the least, and 100 the most discomfort from abdominal pain (A), heartburn (B), gas or bloating ( C ) ,and nausea or vomiting (D). Both treatment arms show a similar reduction in symptom scores, except for nausea or vomiting, which remained unchanged in both groups regardless of therapy.
Scand J Gastroenterol Downloaded from informahealthcare.com by McMaster University on 12/29/14 For personal use only.
Clindamycin and Bismuth for H. pylori
damycin. The mean ages in the two study groups were not significantly different (62.9 versus 60.4). The single patient with H . pylori eradication had been given only bismuth subsalicylate. Statistical analysis showed that the sample size in the clindamycin group was large enough to detect a difference in eradication rate of 0% versus 50% at a significance level of 0.05 and a type-I1 error of 0.2 (16). The observed frequency of eradication (0%) was significantly different from the hypothesized frequency (>50%) by Fisher’s exact test ( P = 0.03). All patients reported improvement in their symptom scores after treatment. Average scores before and after treatment are plotted in Fig. 1. For abdominal pain, heartburn, gas or bloating, all patients had a similar drop in scores regardless of which treatment arm they belonged to. The slope of the curve was almost identical for both groups (Fig. IA-C). Nausea and vomiting were rare in both treatment groups and remained unchanged regardless of the therapy given (Fig. 1D). When the patients were asked how much they had benefited from the treatment, the score was 80 in the bismuth group and 82 in the bismuth/clindamycin group. Neither group had any problems with diarrhea. Pre- and post-treatment diarrhea scores in the bismuth group were 7 and 4 and in the bismuth/clindamycin group 12 and 17, respectively. The patient who was cured of his infection did not differ from the other patients in terms of symptom relief except in the category of abdominal pain. He went from a score of 85 down to 4 after treatment, which was the largest improvement seen in any of the patients.
251
clinical scores suggests that most of our patients consumed their medications as prescribed. The design of this study allowed the patients to take any drugs their primary physician had ordered, unless it was an antibiotic. We made no attempts to influence the use of acidreducing drugs such as H2-receptor antagonists. We have since extended our original animal experiments to evaluate also the effects of cimetidine on antibiotic concentrations in gastric mucosa (23). H2 blockers can increase these levels up to tenfold through a change in the volume of distribution of the antibiotic. Unfortunately, our patients did not benefit from this information. Only two patients consumed any H2 blockers during the study, and neither of them had been randomized to the clindamycin arm. We conclude that our study is yet another testimony to the ineffectiveness of single or dual drug regimens in treating H . pylori infection. Other slow-growing organisms, such as mycobacteria, require combination treatments for clinical cure. We may have to accept that similar strategies are needed for H. pylori. The IXth World Congress of Gastroenterology issued a recommendation to use a combination of bismuth, tetracycline, and metronidazole as the therapy of choice (24). The ultimate regimen may, however, combine multiple antibiotics with acid-reducing agents such as Hz blockers or omeprazole. The animal model suggests that this leads to very favorable pharmacokinetics for the antibiotics (23), and a preliminary report by Unge et al. (25) has given further support to that approach. ACKNOWLEDGEMENTS
DISCUSSION On the basis of our own animal data (10) and reported synergy studies between clindamycin and bismuth (17) we anticipated most of our patients to be cured of their infections. Instead, both treatment arms were failures in terms of eradicating H . pylori. They only showed a therapeutic effect on clinical symptoms. All patients reported some improvement in their symptom scores. The identical slopes of the curves in Fig. 1 suggest that clindamycin contributed very little to this clinical improvement. Although we cannot rule out a placebo effect, we suspect that bismuth subsalicylate was primarily responsible. Clinical improvement by bismuth compounds in spite of failure to eradicate H . pylori has been noted previously by other investigators (18-2 1) . We can only speculate as to why clindamycin performed poorly in this trial. Antibiotic sensitivity data are available for four of the posttreatment isolates. Three of them were still sensitive to clindamycin after treatment. It therefore seems unlikely that development of resistance played any major role. Glupczynski et al. (22) have shown that compliance can be a problem when treating H. pylori gastritis for 4 weeks. We have no way of knowing whether our patients were compliant, but the general improvement in
This study was sponsored in part by the Upjohn Company. Parts of the study have previously been presented in preliminary form at the Second Meeting of the European Helicobacter pylori Study Group, Ulm, Germany, 12-14 October, 1989. REFERENCES 1. Goodwin CS, Armstrong JA, Chilvers T, et al. Transfer of Campylobacter pylori and Campylobacter mustelae to Helicobacter gen. nov. as Helicobacter pylori comb. nov. and Helicobacter mustelae comb. nov., respectively. Int J Syst Bacteriol 1989, 39, 397-405 pathogenetic 2. Wyatt JI, Dixon MF. Chronic gastritis-a approach. J Pathol 1988, 154, 113-124 3. Wyatt JI, Rathbone BJ, Dixon MF, Heatley RV. Campylobacter pyloridis and acid induced gastric metaplasia in the pathogenesis of duodenitis. J Clin Pathol 1987, 40, 841-848 4. Goodwin CS. Duodenal ulcer, Campylobacter pylori, and the ‘leaking roof‘ concept. Lancet 1988, 2, 1467-1469 5. Goodwin CS, Blake P, Blincow E. The minimum inhibitory and bactericidal concentrations of antibiotics and anti-ulcer agents against Campylobacter pyloridis. J Antimicrob Chemother 1986, 17, 309-314 6. Lambert T, Megraud F, Gerbaud G, Courvalin P. Susceptibility of Campylobacter pyloridis to 20 antimicrobial agents. Antimicrob Agents Chemother 1986, 30, 510-511 7. Westblom TU, Gudipati S , Midkiff BR. In vitro susceptibility of Helicobacter pylori to the new oral cephalosporins cefpodoxime, ceftibuten and cefixime. Eur J Clin Microbiol Infect Dis 1990, 9, 691-693
252
7: U. Wesiblorn el ul.
Scand J Gastroenterol Downloaded from informahealthcare.com by McMaster University on 12/29/14 For personal use only.
8. Axon ATR. Carnpylobacter pylori-therapy review. Scand J Gastroenterol 1989. 24 Suppl 160, 35-38 9. Borody TJ, Cole P, Noonan S, et al. Recurrence of duodcnal
ulcer and Carnpylobactcr pylori infcction after eradication. Med J Aust 1989, 151, 431-435 10. Westhlom TLJ, Duriex DE, Madan E, Belshc RB. Guinea pig model for antibiotic transport across gastric mucosa: inhibitory tissue concentrations of clindamycin against Helicobacter pylori (Campylobacter pylori) following two separate dose regimens. Antimicrob Agent$ Chernother 1990, 34, 25-28 11. Klempner MS, Styrt B. Alkalinization of the intralysosornal pH by clindarnycin and its cffect on neutrophil function. J Antirnicrob Chernothcr 19x3, 12 Suppl C. 39-50 12. Wcstblom TU. Midkiff BR, Madan E, Subik MA. Transportation of gastric biopsies in a biphasic Cdrnpylohacter pylori media. J Clin Castroenterol 1989. 1 1 , 483-484 13. Madan E, Kemp J , Westblom TU, Subik M, Sexton S,Cook J . Evaluation of staining methodologics lor identifying Campylobactcr pylori. Am J Clin Pathol 1988. 90, 450-453 14. Aabakken I-, 1,iirsen S, Osnes M. Visual analogue scales lor endoscopic evaluation of nonsteroidal anti-inflammatory druginduced mucosal damage in the stomach and duodenum. Scand J Gastroenterol 1990, 25, 443-448 15. Czinn SJ, Bcrtrarn TA, Murray PD, Yang P. Relationship between gastric inflammatory response and symptoms in patients infected with Helicobacirrpylori. Scand J Gastroenterol 1991, 26 Suppl 181, 33-37 16. Casagrande JT. Pike MC, Smith PG. An improvcd approximate formula for calculating sample sizes for comparing two binomial distributions. Bionictrics 1978, 34, 483-486 17. Vogt K, Hahn H. Synergism between clindamycin and colloidal Received 12 August 1991 Acceptcd 23 October 1991
bismuth subcitrate against Helicobacter (Campylobactcr) pylori in vitro. Zentralbl Bakteriol 1990, 274, 246-249 18. McNulty CA, Gearty JC, Crump 8, et 81. Campylobacter pyloridis and associated gastritis: investigator blind, placebo controlled trial of bismuth salicylate and erythromycin ethylsuccinate. Br Med J 1986, 293, 645-649 19. Stanescu A , Malfertheiner P, M a y a D, Baczako K , Ditschuneit H. Bismuth aluminate in gastroenterology. Therapeutic effects in chronic erosive Campylobacter pylori-associated gastritis. Fortschr Med 1989, 107, 623-626 20. Dobronte Z, Lakatos F, Brittig F, et al. Has Campylobacter pylori infection any clinical relevance? Methodologic, epidemiologic and clinical studies. Orv I Ictil 19x9. 130. 2563-2568 21. Malfertheiner P, Stanescu A , Baczako K, Mayer D, Ditschuneit H. Efficacy of a bismuth cornhination preparation. Efficacy in the treatment of chronic active gastritis and non-ulcerous dyspepsia. Fortschr Med 1990, 108, 402-406 22. Glupczynski Y, Nyst JF, Burette A, Vanderlindcn MP, Dcltcnrc M. Lack of antibiotic compliance in patients treated for Campylobacter pylori-associated gastritis [Abstract 761. 'I'hc Vth International Workshop on Campylobacter Infections, Puerto Vallarta, Mexico, I989 23. Westblom TU, Duriex DE. Enhancement of antibiotic concentrations in gastric niucosa by H2 receptor antagonist: implications lor treatment of Helicohacter pylori infections. Dig Dis Sci 1991, 36, 25-28 24. Axon ATR. Hclicobacter pylori therapy: cffect on peptic ulcer disease. J Gastroenterol Hepatol 1991, 6 , 131-137 25. Unge P, Gad A, Gnarpe H, Olsson J . Does orneprazolt. improve antimicrobial therapy directed towards gastric Campylobacter pylori in patients with antral gastritis? Scand J Gastroentcrol 1989, 24 SUPPI 161, 49-54
