Clarithromycin resistance of Helicobacter pylori strains isolated from children' gastric antrum and fundus as assessed by fluorescent in-situ hybridization and culture on four-sector agar plates

Abstract

Aim: To assess validity of culture on four-sector agar plates and fluorescent in-situ hybridization (FISH) test, and clarithromycin resistance rate in Helicobacter pylori strains isolated from children in the last 10 years.

Methods: In the last 5 years, gastric biopsy specimens from antrum and fundus were taken from 89 consecutive children (median age 9 years) with H. pylori gastritis and from 21 controls. Culture was performed on 176 gastric biopsies (89 from antrum, 87 from fundus) on four-sector agar plates, and FISH test with DNA ProbeMix. After its validity was evaluated, FISH test was applied on additional 119 biopsies from 68 children (68 from the antrum, 51 from the fundus) stored in the Pathology archive in the previous 5 years.

Results: Culture was positive in 157 of 176 biopsies (sensitivity: 89.2%, 95% confidence interval (CI) 85-94). In 33 of 89 children (37%) resistant strains were found in one or both gastric sites. FISH test was positive in 148 of 176 biopsies from infected children (sensitivity 84.1%, 95%CI 79-89) and in none of 42 biopsies from controls (specificity 100%). When applied on archive biopsies, FISH test was positive in 96 of 119 (80.7%, 95%CI 74-88). Total children harboring resistant strains in the last 10 years, as assessed by FISH test, were 66 of 157 (42%). Mixed infection with both sensitive and resistant strains were found in 40 children (25%) and in 12 of them resistant strains were in the fundus only.

Conclusions: Culture on four-sector agar plates and FISH test had a high sensitivity and specificity and showed co-presence of sensitive and resistant strains. In one-third of children with mixed infection, the resistant strains were in the fundus only. Clarithromycin resistance should be assessed in biopsies both from the antrum and the fundus, utilizing antral biopsies only can underestimate its prevalence.