Antibiotic susceptibilities of Helicobacter pylori strains isolated in the Province of Alberta

Helicobacter pylori is associated with chronic gastritis, duodenal and gastric ulcers, and gastric cancer, and is, therefore, a major public health concern (1). The United States National Institutes of Health Consensus Conference on the role of H pylori in peptic ulcers (2) recommends treatment of H pylori infection for all patients with acute or recurrent duodenal or gastric ulcers in whom H pylori infection is present. Similar guidelines were recently published for Can-


MATERIALS AND METHODS
H pylori strains were isolated from gastric biopsies obtained from patients examined at the University of Alberta Hospital endoscopy unit (Edmonton, Alberta) using methods described previously (5)(6)(7).All patients had endoscopically proven gastritis or additional symptoms of duodenal or gastric ulcer, or nonulcer dyspepsia.None of the patients was treated with antibiotic therapy for eradication of H pylori before biopsy.Information on treatment with antibiotics for other conditions was not available.Strains of H pylori were isolated from biopsies obtained between 1992 and 1995, and 31 were randomly selected for antimicrobial susceptibility testing.H pylori strain HP439, which is resistant to metronidazole (minimal inhibitory concentration [MIC] 64 mg/ mL), was used as positive control in tests of susceptibility to metronidazole.An additional 11 randomly selected strains were tested for resistance to metronidazole.
MICs of antibiotics were determined using an agar dilu-tion method for most tests in this study.Serial twofold dilutions of antibiotics were made with sterile water as follows: amoxicillin (Sigma) 0.0125 to 1 mg/mL; clarithromycin (Bayer, Leverkusen, Germany) 0.00625 to 32 mg/ mL; erythromycin (Sigma) 0.0125 to 64 mg/mL; metronidazole (Sigma) 0.5 to 64 mg/ml; and tetracycline (Sigma) 0.5 to 8 mg/mL.For each plate, 1 mL each of diluted antibiotic was mixed with 24 mL of supplemented BHI-YE media (animal serum 5%, Mores).Plates were dried briefly before use.A 48 h culture of each strain was suspended in BHI broth and was adjusted to a concentration equal to McFarland standard #3.A 5 mL drop of each such cell suspension was immediately placed on the agar plate.For each concentration tested, MIC was determined on three separate occasions.Metronidazole resistance was also tested using the E test method (Epsilometer gradient agar diffusion test strips, AB Biodisk, Solna, Sweden).BHI-YE supplemented agar plates were spread evenly with a swab soaked in the same cell suspension used for the agar dilution method.An E test strip was placed in the centre of each plate.All plates were incubated under microaerobic conditions as described previously (5).Results were recorded after 72 h incubation.
Confidence intervals were calculated using the normal approximation of the binomial distribution (8).

RESULTS
Susceptibility test results as determined by the agar dilution method are shown in Table 1.The majority of the H pylori strains examined were highly susceptible to amoxicillin (MIC 0.0125 mg/mL or less).Only one strain of 31 tested had a greater MIC (0.1 mg/mL).The antibiotic history of the individual from whom this H pylori strain was taken was not available.Likewise, susceptibility to tetracycline was observed in 31 strains of H pylori tested with MIC 90 1 mg/mL.Two H pylori strains were slightly more resistant with tetracycline MICs of 2 mg/mL.
Almost all H pylori strains of the 31 tested were susceptible to macrolide antibiotics clarithromycin (MIC 90 0.0125 mg/mL or less) and erythromycin (MIC 90 0.125 mg/ mL).One strain, UA1182, was found to be resistant to clarithromycin (MIC 8 mg/mL) and to erythromycin (16 mg/mL).
Metronidazole resistance defined as 8 mg/mL or greater was noted in five of 42 H pylori strains (12%, 95% CI 4% to 26%) isolated from the Alberta population during 1992 to 1995, as determined by the agar dilution method.An H pylori control culture with a known metronidazole MIC of 64 mg/mL (HP439) was tested at the same time.The E test was also used to determine MICs of metronidazole for 42 strains.With the use of the E test, 6 strains (14%, 95% CI 5% to 29%) were identified as resistant with MICs of 8 mg/mL or greater.These strains included the five identified as resistant to metronidazole by agar dilution testing plus an additional strain.Therefore, the difference between prevalence of resistance as determined by agar dilution testing and the E test is not considered statistically significant.However,

296
Can J Gastroenterol Vol 12 No 4 May/June 1998  when MICs of at least 32 mg/mL were used as the cut-off with the agar dilution method, one of 42 strains was resistant (2%, 95% CI 0.1% to 3%).In contrast, the E test identified four of 42 strains as resistant (9.5%, 95% CI 3% to 23%).

DISCUSSION
Treatment of patients with various antimicrobial agents in combination therapies for eradication of H pylori (1) has resulted in an increase in resistance to some antibiotics.Although resistance to amoxicillin has rarely been reported, a few H pylori strains resistant to tetracycline have appeared recently (9,10).Resistance to neither antibiotic was observed in this study.
The prevalence of clarithromycin-resistant H pylori is considered to be low (less than 5%) in most countries, but as high as 10% in France and Belgium, as reported in the review by Goddard and Logan (11).In a study done in The Netherlands, the incidence of clarithromycin resistance in pretreatment isolates was less than 1.5%, whereas in Italy resistance was reported to be 6% (10).In Canada, clarithromycin resistance was reported to be 1% in patents from Montreal, Quebec (12) and 1.8% in patients from Halifax, Nova Scotia (13).
In our study, one strain of H pylori, UA1182, exhibited cross-resistance to clarithromycin and erythromycin with MICs of 8 and 16 mg/mL, respectively.Our study indicates that 3% (95% CI 0.1% to 17%) of the H pylori strains isolated from gastric biopsies in Alberta are clarithromycinresistant, but our sample size of the H pylori isolate tested was small (n=31).It is possible that the prevalence of clarithromycin resistance in Alberta strains could be lower than 3% if a larger sample of H pylori strains were tested.
One difficulty encountered in comparing incidence of antibiotic resistance in H pylori is the great variability in the MICs taken as breakpoints of resistance.Table 2 (14)(15)(16)(17)(18)(19)(20)(21)(22) lists some examples of MIC breakpoints used to define resistance to metronidazole and clarithromycin taken from several recent studies.Regarding clarithromycin, the majority of our strains, 30, were susceptible based on the most stringent criteria defined in these previous studies.
Resistance to clarithromycin, associated with crossresistance to another macrolide, erythromycin, has been shown to depend on chromosomal mutation (7,14,23).Both of the two rRNA gene copies encoding 23S rRNA are usually mutated.Mutations occur in the H pylori 23S rRNA at coordinates (adenosine) A2042 or A2043, replacing the nucleoside with guanosine (G).We observed that mutations of A2042®G resulted in high level resistance to clarithromycin with MIC of 32 mg/mL and erythromycin cross-resistance at 256 mg/mL.In contrast, intermediate resistance -MIC 0.5 to 1 mg/mL and MIC 64 to 128 mg/mL for clarithromycin and erythromycin, respectively -were associated with A2043®G substitutions (7).In the present study, we observed moderate level resistance to clarithromycin and erythromycin in one strain of H pylori.The mutation present in the resistant isolate has not yet been determined and will be the subject of a separate study.The patient from whom this H pylori strain was obtained had not apparently been treated with a regimen containing clarithromycin for eradication of H pylori before isolation of the strain, but reported visiting the Middle East on a regular basis and may have acquired the strain during foreign travel.
Wide geographic variation in the prevalence of metronidazole-resistant H pylori has been observed, ranging from 20% to 80% in Western Europe, to approximately 100% in Africa (15,24).In Canada, the prevalence of resistance has been reported to vary, being 32% in a combined study of the cities of Montreal, Toronto and Edmonton (18) using agar dilution testing.A study in Montreal reported 20% in Quebec using agar dilution (12) and another in Halifax reported 38% in Nova Scotia using the E test (17).
As is clear from Table 2, the decision on what metronidazole MIC defines resistance has a bearing on the level of resistance.Taking at least 8 mg/mL to define resistance, we obtained an incidence of resistance of 12% in Alberta using agar dilution and 14% using the E test results.If however, a MIC 32 mg/mL or greater is used to define metronidazole resistance, only one of 42 strains (2%) would be considered metronidazole-resistant.The E test identified four of 42 strains (9.5%) as resistant to metronidazole.The Sanford Guide to Antimicrobial Therapy (25) gives peak serum level achievable for metronidazole as 2.5 to 13 mg/mL with the  An MIC breakpoint of 2 mg/ mL was regarded as intermediate resistance mean as 6.2 mg/mL for a 250 mg oral dose.Therefore, 8 mg/mL metronidazole appears to be the most appropriate cut-off value in relation to treatment response.In contrast, for clarithromycin the peak serum level achievable for an oral dose of 500 mg is 2 to 3 mg/mL (25).A cut-off value of 2 or 4 mg/mL to define resistance may be most appropriate for clarithromycin.
The E test is rapid and convenient, and some investigators have reported excellent correlation of E test results with those obtained by standard susceptibility testing methods for H pylori (9,15,26).Nevertheless, the E test for metronidazole has been shown to yield somewhat greater numbers of resistant strains than either agar or broth dilution methods, especially when the E test MIC values fall between 8 and 32 mg/mL (10,27).Our results confirm this finding.The explanation for the discrepancies between metronidazole MICs for the E test and other methods were not determined in this or previous studies (10,27).
Our study highlights that standardization of MIC breakpoints in reporting studies of antimicrobial resistance in H pylori would be an important advance.

CONCLUSIONS
H pylori isolates from biopsies obtained from individuals in Alberta were highly susceptible to amoxicillin and tetracycline.Resistance to clarithromycin and erythromycin was identified in one of 31 H pylori isolates (3%).At present, amoxicillin, clarithromycin and possibly tetracycline remain useful as part of combination therapies for eradication of H pylori. Resistance to metronidazole was also found to be fairly low at 12%, where resistance is defined as 8 mg/mL or greater, and 2%, where resistance is defined as 32 mg/mL or greater.The level of resistance of H pylori in Alberta should continue to be monitored for the emergence of resistant isolates so that eradication therapy can be chosen wisely.