Susceptibility of Helicobacter pylori to metronidazole.

OBJECTIVES

The reliability of the Epsilometer-test (E-Test) and the disk diffusion (DD) method in the assessment of susceptibility of Helicobacter pylori (H. pylori) to metronidazole has recently been questioned, with possible clinical implications for the management of patients undergoing H. pylori eradication. The aims of this study were: 1) to compare the E-Test and disk diffusion methods to the agar dilution method for determining the susceptibility of H. pylori to metronidazole; and 2) to investigate whether potential discrepancies could be caused by the simultaneous presence of metronidazole susceptible and metronidazole resistant bacterial subpopulations.

METHODS

A total of 109 H. pylori strains from 121 consecutive patients were examined. All tests were carried out at the same time starting from primary plates. Agar dilution was performed according to National Committee for Clinical Laboratory Standard (NCCLS) standards, the E-Test according to the manufacturer's guidelines, and disk diffusion according to standard procedure using 5-mug metronidazole disks. Isolates were considered to be metronidazole resistant if the minimal inhibitory concentration was >8 mug/ml for the agar dilution and the E-Test, or if the inhibition zone around the disk was <20 mm for disk diffusion. Of 109 isolates, 43 were also investigated to detect mixed infection. Quantities of 100 mul of bacterial suspensions of each strain were seeded onto plain agar plates and plates containing 8 mug/ml of metronidazole. Cultures were considered to be mixed if the number of colonies on agar plates exceeded by at least 30% those on the metronidazole plates.

RESULTS

According to agar dilution, 57 strains (52.3%, 95% CI = 43-61.4) were metronidazole resistant. E-Test misdiagnosed two strains that were considered sensitive to metronidazole, but according to the agar dilution test they were resistant. Disk diffusion misdiagnosed three strains. Two of these strains (the same as the E-Test) were sensitive, but according to agar dilution they were metronidazole resistant; the third strain was resistant, but according to agar dilution it was sensitive. The percentages of discordance were 1.9 (95% CI = 0.5-6.6) and 2.8 (95% CI = 0.9-7.8), respectively, when the E-Test and disk diffusion were compared to agar dilution. Intertest variability among agar dilution and the E-Test showed that 39.4% (95% CI = 30.8-48.8) of minimal inhibitory concentrations were equivalent (within +/-1 log(2)), 60.6% (95% CI = 51.2-69.2) were major errors (more than +/-1 log(2)), and 3% (95% CI = 0.8-10.4) were very major errors (change in susceptibility pattern). Mixed infection was found in six of the 43 cases examined (13.9%). In four cases, metronidazole resistant strains were 1 log(10) less numerous than those that were metronidazole susceptible. In the remaining two cases, the metronidazole resistant strains were 2-3 log(10) less numerous, which caused the two misdiagnoses.

CONCLUSIONS

The E-Test and disk diffusion method are very good alternatives to agar dilution. Mixed infections are a possible cause of the discrepancies between these tests and the reference method.