Practical use of GenoType® HelicoDR, a molecular test for Helicobacter pylori detection and susceptibility testing

Abstract

Compared to culture-based method, the sensitivity, specificity, positive, and negative predictive values of the GenoType^® HelicoDR for detecting Helicobacter pylori resistance were, respectively, 100, 86.2, 89.7%, and 100% to clarithromycin as well as 82.6, 95.1, 90.5%, and 90.7% to fluoroquinolones. This molecular assay detected a mixture of genotypes and could successfully analyze biopsies without transport/storage limitations.

Introduction

Helicobacter pylori colonizes the stomach of 30% to 50% of human beings. The infection is frequently subclinical but may evolve into severe gastritis, gastro-duodenal ulcer, gastric atrophy, gastric cancer, and MALT lymphoma (Chey and Wong, 2007, Malfertheiner et al., 2007). Since antimicrobial resistance is an important factor leading to eradication failure, reliable susceptibility tests are needed for adequate management of the infection. Culture allows testing for antibiotic susceptibility, but the results are available within 1 to 2 weeks and its accuracy depends on optimal transport and process conditions (Mégraud and Lehours, 2007). H. pylori resistance to amoxicillin and tetracycline is rare in Europe and in the USA (Boyanova and Mitov, 2010). Due to increasing resistance rate, it is strongly recommended to perform susceptibility testing for clarithromycin and fluoroquinolones. On the contrary, routine testing for metronidazole susceptibility is controversial due to incomplete clinicobacteriological correlation (Malfertheiner et al., 2007). H. pylori resistance originates essentially from point mutations which can be picked up with several molecular tests (Owen, 2002). Decrease of antibiotic accumulation has been also associated with tetracycline resistance (Wu et al., 2005).

We evaluated the performance and the usefulness of GenoType^® HelicoDR (Hain Life Science, Germany), a molecular test that combines polymerase chain reaction (PCR) and hybridization. It allows the molecular detection of H. pylori (specific region of the 23S-rRNA gene) and characterization of mechanisms of resistance for clarithromycin (mutations in 23S-rRNA gene in positions 2146 and 2147) and for fluoroquinolones (most significant mutations of the quinolone-resistance region of the gyrA gene, especially at codons 87 and 91) within 6 h.

From January to November 2009, gastric biopsies from the antrum (or [antrum and corpus] according to the referring centre) were obtained from 128 patients with various abdominal symptoms referred from several gastroenterology clinics (Miendje Deyi et al., 2011b) and included in this study: median age 42 years (range 5 to 92), 59 (46.5%) male patients, 40 (31.5%) previously treated for H. pylori infection.

Each biopsy was ground and an aliquot was frozen at −70°C for the molecular analysis, which was compared to phenotypic methods: culture and MICs.

The suspension was inoculated on in-house selective Columbia agar containing 10% horse blood. Plates were incubated for 3–7 days (extended to 10 days if known positive with rapid urease test [RUT]) at 37 °C under a microaerobic atmosphere (Miendje Deyi et al., 2010). MICs were determined using E-test^® (AB Biodisk, Solna, Sweden), and cut-off values of >1 μg were applied both for clarithromycin (NCCLS, 2000) as well as, arbitrarily, for levofloxacin as suggested previously by others (Bogaerts et al., 2006, Mégraud and Lehours, 2007).

H. pylori DNA was manually extracted using the QIAamp^® DNA minikit (QIAGEN Benelux, The Netherlands), and molecular analysis was performed and interpreted according to the manufacturer's instructions.

• The classic comparison involved 69 randomly selected gastric biopsy specimens with a known positive H. pylori culture and available MIC results (1 per patient, 10 children).

• The second group included 48 specimens (9 from children) from which the molecular test was requested by gastroenterologists based on several criteria:

  • discordance between antrum and corpus culture specimens: discordant susceptibility results — susceptible/resistant (n = 14, 29%) — and discordant culture results — positive/negative (n = 6, 12.5%),

  • negative culture in contrast to other diagnosis methods (n = 23, 48%),

  • unavailable phenotypic susceptibility tests (dead strains; n = 5, 10.5%).

• The robustness of the method was evaluated using 15 biopsies previously used for RUT and stored at room temperature during 10 days.

• The specificity of H. pylori detection was evaluated using 20 biopsy samples obtained from patients with well-documented negative H. pylori status (culture, histology, RUT, and serology results were negative).

Formalin-fixed, paraffin-embedded tissue sections were stained with Giemsa to detect H. pylori and associated gastritis. RUT consisted of inoculating a gastric biopsy onto a commercially available test containing urea and pH-sensitive dye (Hut-test^®, AstraZeneca, Wedel, Germany). If H. pylori is present, its urease converts the urea to ammonia, increasing the pH and changing the color of the dye. Serologic testing consisted of a commercially available enzyme-linked immunosorbent assay (ELISA), the anti–Helicobacter pylori ELISA (IgG) (EUROIMMUN, Luebeck, Germany) used to detect the presence of IgG anti–H. pylori in patients' sera.

In case of discrepancy in susceptibility testing, a control was processed for the 2 methods by using a common H. pylori isolate.

• Among the 69 biopsies, 5 were discarded due to noninterpretable PCR results (probably related to a technical incident). Molecular analysis of clarithromycin susceptibility showed in 46 (72%) of 64 samples concordance with MIC determination; a mixture of genotypes (presence of both susceptible and resistance profiles) was observed in 18 (28%) of 64 biopsies. Fluoroquinolone susceptibility testing yielded concordant results between PCR and MIC in 48 (75%) of 64 cases, a mixture of genotypes in 11 cases (17%), and a real discordance in 5 (8%) specimens (1 resistance only with PCR [MIC 0.004 μg/mL] and 4 only with MICs: 2, 3, and 2 times >32 μg/mL). Thus, compared to MIC, the sensitivity, specificity, positive, and negative predictive values of HelicoDR for detecting resistance were, respectively, 100, 86.2, 89.7%, and 100% to clarithromycin as well as 82.6, 95.1, 90.5%, and 90.7% to fluoroquinolones.

• a) HelicoDR analysis revealed that the phenotypic difference between antrum and corpus results was mostly (6/7 patients) related to mixed H. pylori population or lack of sensitivity of the culture since biopsies related to the 3 negative cultures revealed the presence of H. pylori with PCR. b) Twenty-two (96%) of 23 biopsies with negative culture revealed a positive genotyping result as was the case with other diagnostic methods and, consecutively, the susceptibility results became available. c) The 5 biopsies with unavailable MIC, owing to loss of viability or poor culture, probably related to lack of nutrients (Kusters et al., 1997), yielded valuable susceptibility results with PCR.

These results suggest a higher sensitivity of the molecular method for detecting H. pylori.

• Fourteen (93%) of the 15 samples from positive RUT revealed the presence of H. pylori, and, moreover, genotyping revealed a mixed population in 5 (36%) of them. This finding seems very useful since it could encourage gastroenterologists to easily send the biopsy specimens at room temperature and therefore to treat the patient according to susceptibility testing results as suggested, for example, in populations with high clarithromycin resistance rate (Raymond et al., 2010). However, the number of tested samples is low to draw any conclusions and make formal recommendations.

• We found a specificity of 100% for molecular detection of H. pylori since the bacteria were not detected in the 20 negative control samples.

Respectively, 30 (75%) of 40 and 16 (40%) of patients previously treated for H. pylori infection yielded strains harboring clarithromycin and levofloxacin resistance.

The overall frequency of mixed H. pylori populations combining both susceptible and resistant isolates (33%) is comparable to the results of a previous French study (Cambau et al., 2009). It could reflect coinfected cases or selection of resistant mutants.

As shown in Table 1, molecular analysis showed either single mutation or various combination profiles (12% and 25% of double mutations for clarithromycin and fluoroquinolones, respectively) as previously described (Bogaerts et al., 2006, vanDoorn et al., 2001). The distribution of clarithromycin resistance genotypes was comparable to the results of our previous study (Miendje Deyi et al., 2011a). Moreover, HelicoDR allowed differentiation between the 3 mutations showing 78% of A2147G (formerly A2143G), followed by 31% of A2146G (formerly A2142G) and 1.7% of A2146C (formerly A2142C).

In conclusion, owing to the restricted use of culture, GenoType^® HelicoDR is a very promising molecular test for diagnosis and detection of clarithromycin and fluoroquinolones resistance in H. pylori directly from gastric biopsy specimens or in isolates obtained from culture. The test advantages are its robustness, its rapidity, its excellent sensitivity, and its ability to reveal the presence of infection with multiple strains harboring different susceptibility profiles. However, the necessity of adapting the probes to the local prevalence of mutations in order to optimize the search and to detect the majority of circulating mutated strains remains a limitation of the method. A gyrA sequencing of a high number of fluoroquinolone-resistant strains could be helpful in detecting new mutations that are not yet included in the kit.