Discrepancies between E-Test and Agar Dilution Methods for Testing Metronidazole Susceptibility of Helicobacter pylori

We read with interest the recently published paper by Piccolomini et al. (7). They evaluated three different methods to study the susceptibility of Helicobacter pylori to 20 antimicrobial agents. They found a very good correlation between E-test and agar dilution methods, as well as between broth microdilution and agar dilution. They recommended the E test as a reliable and alternative method for testing susceptibility of H. pylori to a wide range of antimicrobial agents in clinical practice. However, they found six major errors (6 of 71, 8.5%) and two very major errors (2 of 71, 2.8%) with the metronidazole E tests when they compared the results with those obtained by reference methods. They consider that chance is the explanation for why these errors were observed only with metronidazole. Among the 20 antibiotics studied by Piccolomini et al., only some, metronidazole, tetracycline, amoxicillin, and clarithromycin, have been widely used in clinical practice. It is important to consider that for precisely 11.3% of the isolates, discrepancies between the E-test and agar dilution methods are with metronidazole. We studied the in vitro activity of metronidazole against 36 H. pylori clinical isolates by E-test and agar dilution methods, and we found two major errors (5.5%) and three very major errors (8.3%). We found no discrepancies when amoxicillin was studied. Although the E test is much less laborious and is easier to perform than the agar dilution method, especially for routine purposes, the results obtained with metronidazole should be confirmed by agar dilution. Some investigators have reported an excellent correlation between E-test results and those obtained by standards methods, with no major or very major errors (2, 4). However, Von Recklinghausen et al. reported a 13.3% discrepancy, major or very major, for metronidazole when the E test was compared with agar dilution (8). On the other hand, Piccolomini et al. studied 71 isolates of H. pylori and found MICs at which 90% of the isolates were inhibited (MIC90s) by agar dilution of 0.25 mg/liter for azithromycin, 1 mg/liter for clarithromycin, 0.5 mg/liter for erythromycin, and 0.125 mg/liter for roxithromycin. They found 0% resistance to azithromycin and roxithromycin, 6% resistance to clarithromycin, and 8% resistance to erythromycin. We have studied Spanish H. pylori clinical isolates and found that clarithromycin was the most active among the five macrolides tested (erythromycin, clarithromycin, azithromycin, roxithromycin, and midecamycin) (1). The results are shown in Table 1. Other authors have also studied the in vitro activities of macrolides. Hardy et al. found that clarithromycin (MIC90, 0.03 mg/liter) is the most active of the macrolides tested, clarithromycin being 4 to 32 times more active than other macrolides (3). The MIC90 for erythromycin, roxithromycin, and azithromycin was 0.25, while for clarithromycin it was 0.03 mg/liter (3). Clarithromycin (MIC90, 0.03 mg/liter) was also found to be significantly more active than either erythromycin (MIC90, 0.125 mg/liter) or azithromycin (MIC90, 0.25 mg/liter) in the study of Malanoski et al. (6). Among the most frequently used antibiotics for H. pylori infection, metronidazole and clarithromycin show different percentages of resistance, depending on the population studied (5); therefore, H. pylori susceptibility should be determined. It is important that different laboratories perform H. pylori susceptibility tests under the same conditions, and standardization is strongly recommended so that results from different laboratories can be compared.