LETTER

According to a recent consensus statement concerning Helicobacter pylori infection in children (3), upper gastrointestinal endoscopy with biopsies is the preferred method for establishing an etiologic diagnosis of infection. The treatment recently recommended for children combines a gastric acid inhibitor, usually a proton pump inhibitor (PPI), with two antimicrobial agents, an antimicrobial agent plus a bismuth salt, or two antimicrobial agents (4). In France, where the use of bismuth salts is not authorized, treatment consists of amoxicillin plus either clarithromycin or metronidazole. The consensus statement (3) also proposed a strategy for reevaluating those who remain infected after an initial course of therapy: repeat endoscopy with culture and antimicrobial susceptibility testing, with secondary treatment based on susceptibility test results.

We have already shown that the rates of resistance to metronidazole and clarithromycin among isolates of H. pylori from children are high before treatment (i.e., 43 and 21%, respectively) (5). No data are available concerning the antimicrobial susceptibility patterns of strains obtained from cultures from children who had failed an initial course of therapy. The aim of this study was to assess antimicrobial resistance rates in H. pylori strains after an initial course of treatment.

During the period from 1993 to 2000, 15 girls and 8 boys (mean age ± standard deviation, 10.9 +/ 4.8 years; range, 1.4 to 17 years) with culture-confirmed H. pylori gastritis failed to respond to an initial course of therapy (1 week of treatment with a PPI and amoxicillin together with either clarithromycin [n = 14] or metronidazole [n = 9]). Six weeks after the end of treatment, the [¹³C]urea breath test was performed on all patients. If a positive result was obtained, a second endoscopy was performed and gastric biopsy samples were obtained for culture and antimicrobial susceptibility testing. Biopsies were performed prior to initiating a second course of treatment.

Clarithromycin-resistant strains of H. pylori were recovered from 8 children (34.7%) prior to the initial course of treatment and from 12 children (52.1%) after treatment (difference not significant); metronidazole-resistant isolates were obtained from 13 children before treatment (56.5%) and from 12 children after treatment (52.1%) (difference not significant); and isolates resistant to both clarithromycin and metronidazole were obtained from 4 children before treatment (17.3%) and from 7 children after treatment (30.4%) (difference not significant). All strains remained susceptible to amoxicillin (Table 1). Resistance to clarithromycin was noted among posttreatment isolates of H. pylori in 3 of the 14 children (21.4%) whose triple therapy included clarithromycin. In those 14, one isolate lost metronidazole resistance. Metronidazole resistance was not recognized among any of the posttreatment isolates from 9 children whose triple therapy included metronidazole. However, one posttreatment isolate from this group was noted to be resistant to clarithromycin (Table 1).

In a previous study of adults, Pilotto et al. (8) reported a high secondary resistance to metronidazole and clarithromycin (42.1 and 52.6%, respectively) or to both (26.7%) after failure of a 1-week PPI-based triple therapy including metronidazole or clarithromycin. The present preliminary study was devoted to children and indicates lower rates of secondary resistance. Secondary resistance implies that some strains are able to develop resistance in vivo (1). However, the apparent loss of resistance to metronidazole observed for one isolate may be explained by any one of several phenomena, such as contamination with a new strain or the unreliable results of antimicrobial susceptibility testing under microaerophilic conditions, as previously described (10). Colonization with a clarithromycin-resistant strain of H. pylori in patients with no history of clarithromycin exposure has been previously reported (7). In addition, for Wang et al. (11), in vitro spontaneous mutations in H. pylori conferred resistance to clarithromycin or metronidazole. In vivo emergence of antibiotic resistance may be attributed to the combined effects of spontaneous mutation and recombination. At least, one cannot exclude infection of children with multiple strains (2).

In conclusion, secondary clarithromycin-resistance may develop in H. pylori strains following first-course treatment and account for failure to eradicate the bacterium. Further studies including more children will precisely define the parts played on the one hand by acquired resistance and on the other hand by infection with multiple strains in determining which antibiotics select the resistant strain.