Persistent colonisation of gastric epithelium by Helicobacter pylori is one of the most important factors for the development of chronic gastritis and peptic ulcer disease. It is also a recognised risk factor for malignant diseases, including carcinoma and lymphoma of the mucosa-associated lymphoid tissue . The development of effective therapies for the treatment of patients with chronic colonisation by H. pylori has been an arduous process since the microorganism was first recovered in 1983. After a number of false starts, the so-called triple therapy (consisting of bismuth, metronidazole and tetracycline) was the first therapy to achieve a successful outcome of this infection in most patients . Subsequently, a number of therapeutic regimens have been developed, which reliably cure 80 % or more of infected patients. It is widely accepted that triple therapy consisting of a proton pump inhibitor (PPI) and two antibiotics out of amoxycillin, clarithromycin and metronidazole is the most effective therapy for this infection . As the efficacy of anti-H. pylori therapy improves, however, the mechanisms behind treatment failure become more difficult to study. Factors most frequently associated with therapeutic failure include poor patient compliance, inadequate drug delivery at the site of infection, low level of drug stability at the gastric location and antimicrobial resistance .
The issue of antibiotic resistance is of growing concern and has become an important factor leading to therapy failure. It is presently assumed that a major obstacle for an effective H. pylori eradication therapy is the presence of a resistance mechanism. However, a lack of correlation between in vitro resistance and clinical outcome has also been noted, leading to difficulty in interpreting clinical results when resistance is observed and effective eradication occurs. In the present paper, a review is made of the susceptibility profile of H. pylori, the resistance mechanisms and epidemiology, the microbiological factors involved in the development of resistance, and the clinical consequences of antimicrobial resistance in the eradication of H. pylori.
SUSCEPTIBILITY PROFILE AND METHODS FOR ANTIMICROBIAL SUSCEPTIBILITY TESTING OF H. PYLORI
H. pylori presents a uniform antimicrobial susceptibility profile modified only by the presence of acquired resistance mechanisms (Table 1). Since the first in vitro studies, it has been obvious that several antimicrobial agents inhibited H. pylori at low concentrations. However, this organism was considered intrinsically resistant to sulfonamides, trimethoprim, nalidixic acid, glycopeptides and polymyxins . Despite this uniform susceptibility profile, variations in minimum inhibitory concentrations (MIC) were often observed when susceptibility data from different laboratories were analysed. These differences were partially explained by the lack of a standardised antimicrobial susceptibility testing method for H. pylori, and so different microbiologists used different methods. Nowadays, there is lack of consensus and criteria for susceptibility testing are not yet clearly established.
Only recently has the National Committee for Clinical Laboratory Standards (NCCLS) included H. pylori in the susceptibility-testing document . This document establishes the agar dilution technique as the reference method. Mueller–Hinton agar supplemented with aged (at least 2 weeks old) sheep blood (5 % v/v) is the recommended culture medium. The NCCLS established a 3-day incubation period in a microaerophillic environment produced by a gas-generating system suitable for campylobacters. The inoculum size, one of the most important factors affecting susceptibility testing results, should be prepared in saline with a turbidity equivalent to the 2.0 McFarland standard (1 × 107–1 × 108 c.f.u./ml) using a 72-h-old subculture from a blood agar plate.
Chocolate agar or serum-enriched media, instead of blood supplemented media, have also been recommended and incubation has even been performed in CO2 at high humidity . Other methods, such as broth dilution, have been proposed although no common criteria are followed. The disk diffusion and E-test methods are simple and easy to perform, but the best correlation with the reference agar dilution technique has been obtained with the E-test method. Also crucial for susceptibility testing is the pH, as different pH values can affect susceptibility testing results. Aminopenicillins are 10-fold more active at neutral pH than at a more acidic pH. In contrast, only slight differences, if any, have been observed when susceptibilities to tetracycline or metronidazole were measured. However, several studies have pointed out problems posed by metronidazole susceptibility testing. A poor correlation between results offered by disk-diffusion and agar dilution method has been shown . The E-test is claimed to be more accurate, but variations have been observed depending on the inoculum size, culture medium, and, particularly with incubation conditions . In this regard, previous exposure of H. pylori to anaerobic conditions is sufficient to revert the metronidazole resistance phenotype in certain isolates.
Molecular-based methods can also be useful for identifying resistant strains obtained from biopsy or gastric juice specimens and have also been applied directly to these samples. However, they can be applied easily for detecting clarithromycin resistant strains but not for metronidazole resistance because of the high diversity within the genes involved .
ANTIBIOTIC RESISTANCE IN H. PYLORI
Despite high in vitro activity of most antibiotics against H. pylori (Table 1) treatment of infected patients has proven to be difficult and at least two antibiotics are needed to obtain a favourable outcome. Drug inactivation at low pH, inappropriate formulation, variable compliance, inability of the drug to accumulate in the gastric mucosa and high H. pylori inoculum in the gastric mucosa have been reported as responsible factors for this situation .
Lack of in vivo antibiotic uptake by H. pylori and efflux-related mechanisms can also be responsible for laboratory and clinical discrepancies. In vitro studies have demonstrated that H. pylori contains at least four porins which are weakly expressed compared with those in other bacteria [10,11], but no studies on in vivo porin modifications have been performed. On the other hand, it has been demonstrated recently that active efflux in H. pylori does not seem to play an important role in antimicrobial susceptibility. Genes related to restriction–nodulation-dependent efflux family systems, which are related to chloramphenicol and tetracycline resistance in other bacteria, are also present in H. pylori isolates. However, changes in these genes, and the corresponding changes in efflux properties, do not affect uptake of these antibiotics .
Plasmid DNA material and bacteriophages have been observed in H. pylori isolates, but so far they have not been linked to antimicrobial resistance . In general, the emergence of antimicrobial resistance in H. pylori appears to be originated by mutations. Clarithromycin resistance is due to mutation in the 23S rRNA gene , which encodes ribosomal target proteins for clarithromycin. In addition, metronidazole resistance occurs mainly by mutations in the frxA gene, which encodes an essential enzyme that reduces metronidazole to active metabolites . Mutation phenomenon occurs in natural populations and is normally observed at extremely low frequencies (< 10−8). However, under antimicrobial pressure, susceptible bacteria can be eliminated and mutated resistant bacteria can be selected, particularly when antibiotic appropriate concentrations at the site of infection are not reached by treatment. Recently, organisms in chronic infections have been postulated to augment the mutation rate due to alterations in DNA repair and error-avoidance genes, and consequently, the emergence of resistance can be enhanced . It is well recognised that H. pylori colonisation is a chronic process. Moreover, the H. pylori niche does not allow the appropriate diffusion of antibiotics and the effect of antibiotics is limited by the short time that these drugs are in the stomach. Consequently, mutated (resistant) bacterial population can easily escape the antimicrobial action. In addition, the spontaneous mutation rate obtained in vitro by culturing H. pylori under subinhibitory concentrations of antibiotics reaches high values, particularly with macrolides (2 × 10−5 to 3.5 × 10−8) and metronidazole (10−5 to 10−6) . This high frequency of variation is also observed in phenotypic characteristics: spontaneous phenotypic variants with smooth colony morphology occur with a high frequency (10−2 to 10−3) .
Another factor affecting bacterial resistance is the easily transformable nature of H. pylori  and the presence of insertion sequences. The former property can explain the integration of resistance genes and exchange of exogenous DNA. Recombination phenomena can be forced by the presumed deficiency in mismatch repair systems. On the other hand, the frequent presence of insertion sequences in H. pylori-associated with transposon-based spontaneous mutations and consequently, interruption of gene sequences may lead to resistance to antimicrobial agents. Insertion sequences contribute to drug resistance as a consequence of bacterial adaptive changes. In that sense, redxA gene interruption due to mini-IS 605, which determines metronidazole resistance, has been identified in H. pylori .
ANTIMICROBIAL RESISTANCE MECHANISMS IN H. PYLORI
As in Mycobacterium tuberculosis, H. pylori resistance to antibiotics can be primary (pre-treatment) or can develop during the eradication treatment course (secondary resistance). Primary resistance to clarithromycin or metronidazole does adversely affect treatment success and may represent a concern for the clinical management of H. pylori- associated disease. Data on secondary resistance indicate that resistance rate is at least 50 % for both metronidazole and clarithromycin in patients in whom eradication has failed, when these drugs have been used for the eradication [21,22]. This relevant distinction is not affected by the mechanisms by which resistance occurs . However, primary resistance in H. pylori clinical isolates is thought to be a stable phenomenon, whereas resistance that appears during therapy is not always stable , probably due to the emergence of mixed populations. Table 2 shows the resistance mechanisms to antimicrobial drugs described in H. pylori.
With the exception of cefsulodine, β-lactam antibiotics possess a high in vitro activity against H. pylori isolates, with MIC values much lower than concentrations obtained in gastric mucosa when they are administered regularly (Table 1). In clinical practice, amoxycillin is the only β-lactam antibiotic used. It is highly stable at acidic pH, more so than ampicillin, and its optimum pH value is almost that obtained in gastric mucosa when patients are treated with omeprazole or other PPI. In general, cephalosporins display a lower intrinsic activity than those obtained with amoxycillin or ampicillin. Clinical data with cephalosporins are scarce and results obtained have been disappointing.
β-lactams are bactericidal antibiotics that inhibit bacterial cell wall synthesis. Amoxycillin and carbapenems are bactericidal in vitro (≥ 99.9 % killing) against H. pylori. In contrast, ampicillin, piperacillin and several cephalosporins, including cefuroxime, cefixime and ceftazidime act as bacteriostatic agents . Bactericidal activity of amoxycillin in H. pylori seems to be concentrationdependent  and amoxycillin concentrations reached in the gastric mucosa are probably sufficient to obtain a good bactericidal effect . It has been speculated that amoxycillin resistance  or other mechanisms, including the possible intracellular location of H. pylori , explain treatment failure with regimens containing this antibiotic.
Until very recently, stable amoxicillin resistance in H. pylori had not been reported . β-lactamase production, and subsequent inactivation of the β-lactam antibiotic, is the most common mechanism involved in Gram-negative β-lactam resistance. Until now, no β-lactamase production has been demonstrated in H. pylori isolates and amoxycillin resistance appears to be related to the absence of a penicillin-binding protein (PBP) of 30–32 kDa  (Table 2).
As in other organisms, PBPs are the targets for β-lactam antibiotics. PBPs are a set of enzymes involved in the synthesis of the peptidoglycan layer in the bacterial cell wall involving transpeptidase, transglycosylase, endopeptidase and carboxypeptidase activities. At least four major and two minor PBPs have been identified in H. pylori, but different ratios of these proteins are observed when studied in helical- (log-phase) or coccoid-form (stationary-phase) cultures . Coccoid forms have been associated with slight increases in β-lactam MIC values. These forms arise after in vitro exposure of H. pylori isolates to amoxycillin . The presence of coccoids, which are not actively dividing and have a PBP profile different from that of helical forms, can be responsible for failure of amoxycillin treatment . Moreover, induction of spheroplasts and a change in morphological shape have also been observed after 21 h exposure of H. pylori to amoxycillin. This exposure also causes a low-level increase in the number of amoxycillin-resistant populations . Recently, this antibiotic has been shown to bind to a major PBP protein of 72 kDa .
Amoxycillin resistance in H. pylori has been defined as amoxycillin MIC values equal to or greater than 8 μg/ml. However, displacements of MIC distribution to higher values has been noted in isolates obtained from biopsy specimens from peptic ulcer patients from the USA and dyspeptic patients from Italy [27,32]. A detailed analysis of MIC values of these isolates revealed two different groups. The first displayed amoxycillin MIC values of 2 μg/ml whereas the second one displayed MIC values of > 32 μg/ml. The modal MIC value of amoxycillin for sensitive isolates in this study was 0.016 μg/ml. Moreover, amoxycillin resistance was associated with a minimal bactericidal concentration : MIC ratio of ≥ 32 μg/ml, indicating that these resistant strains are tolerant to this antibiotic . Resistance was lost after storage at −80°C but was restored by consecutive transfers onto amoxycillin gradient plates.
Erythromycin, the first macrolide used in clinical practice, has no role in therapy for the eradication of H. pylori. This antibiotic is highly unstable at low pH. On the contrary, new macrolides, including clarithromycin, roxithromycin, dirithromycin, azitromycin, josamycin and midekamycin are more stable to pH variations. These macrolides, alone or in combination with other antibiotics and a PPI, have been used in eradicating regimens, but the best results have been obtained with clarithromycin. Nevertheless, some studies suggest that roxithromycin may be used effectively, while others suggest that azithromycin is also appropriate. Clarithromycin is a 14-atom ring macrolide, which shows bactericidal activity against H. pylori, even higher than that of amoxycillin. Moreover, a 14-OH metabolite is generated in vivo that prolongs clarithromycin bactericidal activity. Interestingly, MIC values of clarithromycin are extremely low (modal MIC values of 0.03 μg/ml) and are increased only moderately at low pH values.
Macrolides act by binding to ribosomes and resistance is due to modification in ribosomes that enable this interaction. Macrolide resistance is due mainly to the synthesis of an enzyme that methylates an adenine residue of the 23S rRNA. This resistance mechanism, called MLS, which affects macrolides, lincosamides and streptogramin B compounds, has not been identified in H. pylori. On the other hand, analysis of clarythromycin-resistant strains revealed that point mutation in the peptidyl transferase domain of the 23S rRNA is the mechanism of resistance to macrolides (Table 2). Versalovic et al.  identified adenine-to-guanine transition mutations at either position 2142 or position 2143. More recently, mutations at positions 2142 and 2143 have also been identified . Wang and Taylor  established that two major types of clarithromycin resistance were correlated with specific point mutations in the 23S rRNA gene (Table 3). The first phenotype (type I) determines high level resistance (≥8 μg/ml) to all macrolides, lincosamides and streptogramin B. The mutation responsible for this phenotype was a guanine instead of an adenine at position 2142 (A2142G). In the second phenotype (type II), an intermediate resistance to clarithromycin and clindamycin was observed, but no resistance to streptogramin B was detected. This phenotype was linked to A2143G mutation. Moreover, A2142C and A2142T substitutions also determine cross-resistance to all MLS antibiotics (type I), whereas the corresponding A2143C and A2143T were associated with phenotype II. Both phenotypes were observed in strains selected after direct mutagenesis and natural transformation and in wild H. pylori isolates. It is worth noting that in both types of clarithromycin resistance, as well as in clarithromycin-sensitive isolates, streptogramin A and streptogramin B demonstrated a synergistic effect. To our knowledge, no streptogramin regimens, either oral with pristinamycin or virginiamycin or even parenteral with quinupristin/dalfopristin have been investigated.
Discrepancies in susceptibility to erythromycin and clarithromycin have been demonstrated in some H. pylori isolates. Mutation A2142G was consistently associated with a clarithromycin MIC of > 256 μg/ml, whereas mutants carrying A2143G had MIC ranging from ≤ 0.06 to > 256 μg/ml, suggesting that additional factors may be responsible for the observed multiple level resistance to clarithromycin  (Fig. 1). On the other hand, A to G mutants are observed predominantly among clarithromycin-resistant clinical isolates probably due to growth advantage over other mutant types .
H. pylori isolates possess two copies of the 23S rRNA  and mutations can be observed in a single copy or in both copies. However, most clarithromycin-resistant H. pylori isolates are homozygous mutants. Finally, the simultaneous colonisation of H. pylori with and without mutations in the 23S rRNA gene has been identified in patients with no history of previous clarithromycin exposure ; this could easily explain the eventual emergence of secondary resistance.
Metronidazole, tinidazole and ornidazole have similar activity against H. pylori and cross-resistance is observed when the activity of one of them is affected. As in other bacteria, for metronidazole to act against H. pylori it must be taken up by an energised membrane and then reduced to a toxic metabolite . For this reduction several nitroreductases are present; however, in H. pylori it has been suggested that an oxygen-insensitive NADPH nitroreductase encoded by the rdxA gene is the most important of these. NADPH nitroreductase reduces metronidazole to active metabolites that are directly toxic to H. pylori. Nitroimidazoles are commonly used in the treatment of anaerobic and parasitic infections. The extensive use of these drugs for these purposes, particularly in developing countries, has been associated with increased metronidazole resistance in H. pylori. In the laboratory, in vitro serial passages of metronidazole-sensitive strains on plates containing subinhibitory concentrations of metronidazole increased the MIC values of this antibiotic. Moreover, the stability of metronidazole-resistant isolates emerging after one or two passages on metronidazole-containing media and the stability of resistance in strains after three passages are statistically significant .
Several mechanisms of metronidazole resistance have been proposed, including scavenging of toxic oxygen radicals by an altered catalase or super-oxide dismutase, a more efficient DNA damage repair mechanism, and loss of function of a critical reductase . Undoubtedly, mutational inactivation of the rdxA gene, which encodes an oxygen-insensitive NAPDH nitroreductase, is responsible for metronidazole resistance in H. pylori . Moreover, deletions and insertions of transposable elements within the rdxA gene and mutations in the promoter region also affect the expression of this gene and increase metronidazol MIC values to as much as > 32μg/ml [20,40].
Despite the straightforward recombination characteristics of H. pylori , metronidazole resistance arises by mutations and appears not to be due to the co-existence of sensitive and resistant unrelated strains or to horizontal transfer between unrelated isolates. It is worth noting that metronidazole heteroresistance is observed frequently in H. pylori populations, particularly after metronidazole exposure. Mixed populations of susceptible and resistant bacteria genetically indistinguishable by random amplified polymorphic DNA (RAPD) analysis may co-exist even in metronidazole-resistant primary cultures . It has been remarked that nitroimidazoles are mutagenic agents and that nitroimidazole exposure can increase mutation frequency in all genes, including the rdxA gene. For this reason, the rapid development of metronidazole resistance is not surprising after the use of a nitroimidazole regimen. Metronidazole heteroresistance has important microbiological implications for the accuracy of in vitro susceptibility testing of H. pylori and it is important to test multiple colonies.
The rdxA gene-dependent mechanisms appear to be the most frequent resistance mechanism. However, mutations in the frxA gene, encoding an NAD(P)H flavin reductase with 25 % homology with the rdxA product, also affect metronidazole susceptibility. Moreover, diminished intracelullar accumulation of metronidazole by reduced uptake or active efflux and overexpression of RecA leading to increased DNA repair has also been proposed as resistance mechanism in metronidazole-resistant H. pylori isolates . In this case, metronidazole resistance is expressed to a lesser extent and MIC values tend to be lower than those observed in isolates with rdxA gene-mediated resistance.
These antibiotics are not included in current eradication regimens. In general, fluoroquinolones have high intrinsic activity with MIC lower than 0.1 μg/ml. However, these lower values do not predict a high in vivo efficacy. Clinical trials with fluoroquinolones have demonstrated poor eradication results. These results could be due to decreased quinolone activity at gastric pH and to the rapid emergence of quinolone-resistance following ciprofloxacin treatment regimens . Nevertheless, in vitro frequency of selection of resistant mutants is lower than that observed with metronidazole or macrolides .
The resistance mechanism to fluorquinolones in H. pylori is similar to that observed in Escherichia coli and other organisms. H. pylori ciprofloxacin-resistant mutants present single or double mutations in the gyrA gene, which occur at the homologous quinolone-resistant determining region of the gyrA gene of E. coli . The gyrA gene encodes a topoisomerase which acts by relaxing the supercoiled DNA to allow its replication. Single mutations in H. pylori affect amino acids 87, 88 and 91 whereas double mutations are frequent at amino acids 91 and 97. Moreover, quinolone resistance is observed when susceptible strains are transformed with the amplified gyrA fragment from the resistant strains. This fluoroquinolone-resistance mechanism is probably not exclusive of H. pylori because in one study one out of 11 isolates was resistant to quinolones but no gyrA mutations were identified . In addition, it is possible to select in vitro fluoroquinolone-resistant isolates that are also resistant to metronidazole. This resistance could be due to altered permeability or efflux-based mechanisms.
Tetracyclines act in inhibiting protein synthesis by interacting with the 50S subunit ribosome. Despite their bacteriostatic activity, these antibiotics are part of the classical eradication triple therapy. Minocycline and doxyciline have also been used in different regimens, although the intrinsic activity is lower than that observed with tetracycline. In contrast with fluoroquinolones and macrolides, these compounds are stable at gastric pH. Resistance to tetracyclines was first reported in 1996 in a patient who received standard classical triple therapy for H. pylori. No eradication was observed and the MIC for tetracycline in the recovered strain was > 256 μg/ml . In addition, tetracycline-resistant H. pylori isolates have also been reported in multicentre surveillance studies . The resistance mechanism has not been identified, but clinical failure could be associated with resistance .
Rifampicin and other derivatives, including rifabutin, inhibit H. pylori at low concentration (MIC values 0.002–0.25 μg/ml) and have been used in combination with amoxicillin and a PPI after failure of other eradication therapies. Resistance to rifampicin and rifabutin has been observed both in vitro and in vivo . This resistance is caused, as rifampicin resistance in E. coli and M. tuberculosis, by amino acid exchanges in the β-subunit of the DNA-directed RNA polymerase (RpoB) . Mutagenesis of the rpoB gene showed different levels of resistance, depending on the replaced amino acid .
H. pylori susceptibility to bismuth compounds, associated with two antibiotics in classical triple therapy, is related to the expression of metal transporter proteins in the cytoplasmic membrane and disruption of glycocalyx cell wall [49,50]. Bacteria lacking Hpn, a 60-amino acid histidine-rich protein that avidly binds nickel and zinc ions, are more susceptible to bismuth. On the contrary, disruption of the nixA gene encoding a high-affinity metal transporter does not affect susceptibility to this compound.
EPIDEMIOLOGY OF RESISTANCE
Several multi-centre and non-multi-centre studies have evaluated the incidence of H. pylori-resistant isolates. Overall, resistance to clarithromycin ranges from 0 % to 15 %, whereas resistance to metronidazole has increased from 20 % to 95 %. In a recent large European multi-centre trial (the MACH2 study) carried out on 485 isolates obtained from gastric biopsy specimens from 516 patients, resistance rates to clarithromycin and metronidazole were 3 % (> 1 μg/ml) and 27 % (> 8 μg/ml), respectively . Although clarithromycin resistance was low, secondary resistance to this antibiotic occurred in strains from 12 of 105 patients (11.4 %) after failure of a clarithromycin-based regimen. A2142G and A2143G mutations were responsible for clarithromycin resistance in all but one strain. Interestingly, no resistance to amoxycillin was found. The MIC distribution of amoxycillin was unimodal, with the highest MIC at 0.5 μg/ml. Moreover, the comparison of clarithromycin resistance results in the MACH2 study with those published previously in the MACH1 study revealed no trends in resistance rates . Resistance to metronidazole in the MACH2 study, 27 % (range, 16–42 %), was lower than percentages reported previously from many European countries, probably due to different susceptibility testing methodologies.
Resistance to clarithromycin and metronidazole may vary from country to country and even in different areas . In general, in Europe there is a north-to-south gradient of clarithromycin resistance that may be related with the differences in macrolide consumption. In the Mediterranean area, the level of clarithromycin resistance reaches 12–15 %, whereas in Scandinavian countries it is less than 3 %. On the other hand, an increase in metronidazole resistance has been noted particularly during the last 5 years. In the USA, several studies have been performed. Resistance to clarithromycin and metronidazole was 6.1 % and 37.4 %, respectively in a large study in a metropolitan hospital  and the frequency of H. pylori with resistance to both antibiotics was 3 %. These values are slightly higher than those found in The Netherlands: resistance to clarithromycin and metronidazole was 1.7 % and 21.2 %, respectively . The reverse situation was recently observed in Bulgaria: clarithromycin resistance increased up to 12.5 % during 1997 and 1998 and was often associated with metronidazole resistance . Figure 2 shows the evolution of metronidazole and erythromycin resistance observed in H. pylori isolates obtained from untreated patients with eradication therapies at our hospital.
Metronidazole resistance rates vary according to the populations studied. Resistance is higher in developing than in developed countries, reaching 80–90 % in Africa , and it is generally more frequently in women (38.5 %) than in men (24.4 %) . This higher prevalence of metronidazole-resistant H. pylori in women in some communities has been correlated with the use of metronidazole to treat Trichomonas vaginalis or other protozoan infections . It has also been noted that metronidazole resistance in H. pylori isolated from children is lower than that in adults .
Interestingly, clinical conditions associated with H. pylori could also affect resistance results. In a randomised French multi-centre study, clarithromycin-resistance rates among strains isolated from patients with ulcers were significantly lower than those found in isolates recovered from patients with other H. pylori- associated diseases .
CLINICAL IMPLICATIONS OF ANTIMICROBIAL RESISTANCE
Current regimens for patients infected with sensitive strains are highly effective. However, in patients with H. pylori strains resistant to nitroimidazoles and/or clarithromycin, therapies are likely to fail and subsequent successful treatment will be more difficult. Identification of patients infected with resistant strains, either by knowledge of the prevalence of resistance in a particular population or by the direct assessment of individuals, allows the use of alternative treatments. Unfortunately, several problems inherent to antimicrobial susceptibility testing of H. pylori have led many clinicians to underestimate the relevance of such testing . Determination of antimicrobial susceptibility of H. pylori does not always solve the problems. The correlation between in vitro sensitivity and in vivo effectiveness may be incomplete.
The clinical effect of clarithromycin resistance is essentially the complete loss of any clarithromycin anti-H. pylori effect, and outcome of therapy can generally be predicted on the basis of what would be expected if only the other antimicrobials in the regimen were used [60,61]. Thus, clarithromycin resistance has the effect of antibiotic being replaced by placebo. The clinical relevance of H. pylori resistance to nitroimidazoles detected in vitro has been controversial in different studies. In vitro resistance does not always predict in vivo results, as combination therapies seem to partially overcome in vitro resistance to metronidazole . Several factors might have contributed to this phenomenon, including different methods used for diagnosing metronidazole resistance [63,64], different cut-off values for resistance [2,65], and probably different locations where biopsy samples were taken . However, a continuous spectrum of metronidazole MIC, in contrast with the bimodal clarithromycin MIC distribution, is probably the major determinant of this phenomenon .
The clinical relevance of measuring nitroimidazole resistance in vitro is still unclear. The goal of culturing and susceptibility testing is to detect clinically relevant antimicrobial resistance. That means that, via an in vitro assay, one hopes to predict the likelihood of successful therapy with a particular antimicrobial regimen. Neither in vitro data nor animal studies can predict the clinical efficacy of anti-H. pylori regimens containing nitroimidazoles. Therefore, clinical studies are essential to compare the eradication rates with the MIC of the antibiotic . Furthermore, the secondary bacterial resistance to metronidazole or clarithromycin increases significantly in patients in whom previous eradication treatments have failed [69,70]. This suggests that previous treatment failure is a significant contributing factor to acquired bacterial resistance for both metronidazole and clarithromycin .
Dosage of antimicrobials is also an important factor. Data from recent studies suggest an increase in secondary clarithromycin resistance when this antibiotic, in combination with omeprazole and metronidazole, is used at low dose . Nevertheless, the importance of development of secondary resistance in H. pylori infection in eradication regimens remains unclear and has to be investigated further.
TREATMENT ALTERNATIVES IN H. PYLORI RESISTANT TO ANTIMICROBIALS
As mentioned previously the prevalence of metronidazole-resistant H. pylori varies between different areas, different populations and different groups within these populations. Prevalence is usually lower in developed countries than in developing countries, although certain populations in developed countries may also have high rates of metronidazole-resistance due to early acquisition of resistant strains in their country of origin . World-wide, the prevalence of clarithromycin-resistant H. pylori is usually much lower than metronidazole-resistance; nevertheless, it is steadily increasing and, at present, ranges from 7 % to 14 % [53,60].
As antimicrobial effectiveness is not always predicted in vitro and given the difficulties inherent to H. pylori susceptibility testing, the current guidelines of the Digestive Health Initiative International Update Conference , the European Maastricht Consensus Report  and other local consensus conferences do not recommend performing susceptibility tests before primary eradication treatment. Antibiotic susceptibility testing is recommended only for obtaining epidemiological data in order to adapt eradication treatments to the resistance situation. Therefore, susceptibility results are rarely available and clinicians are almost exclusively interested in the overall effectiveness of therapy. In that sense, different predictive models based on large clinical trials have been published to establish the overall effectiveness of a regimen containing a particular antibiotic .
Once H. pylori eradication therapy has failed, the optimal treatment needs to be established. Although current therapies achieve high eradication rates, approximately 10–20 % of patients still remains infected after a first eradication trial. These infected patients continue to have a high risk of ulcer recurrence and complications, and are at an obvious disadvantage when considering the enormous benefits derived from H. pylori eradication, at least in gastro-duodenal ulcer disease, such as increased ulcer healing, less ulcer recurrence and lower ulcer complications. It is still controversial and debatable, particularly in the field of gastroenterology, when to perform culture and antibiotic susceptibility testing after first eradication failure. The knowledge of the organism's antibiotic susceptibility profile can be a great aid in selecting a second-line therapy regimen (frequently named ‘rescue’ treatment). Ideally, follow-up treatment of these patients should be guided by susceptibility data. Unfortunately, susceptibility results are often unavailable as endoscopy is not always performed, or when biopsy specimens are obtained they are not always sent to the Microbiology Laboratory. Resistance is generally suspected when an appropriate eradication treatment fails. In this case, the second-line treatment is established depending on the treatment used initially. As a general rule, any of the antibiotics against which H. pylori has probably become resistant should not be re-administered: if a clarithromycin-based regimen was first used, a metronidazole-based regimen should be used afterwards, and vice versa (Fig. 3). It is unwise to use the same antibiotic twice . When a second eradication attempt has failed, treatment options depend on antibiotic resistance, and in that circumstance, there is a major consensus on the benefits of antimicrobial susceptibility testing before a new treatment against H. pylori is commenced.
Local (geographical) prevalence of antimicrobial resistance will determine if it is better to start with a regimen based on clarithromycin or with one based on metronidazole [75,76]. Given that clarithromycin resistance is still low in most populations, treatment should usually be started with a triple regimen based on this antibiotic. Regimens based on metronidazole should be used first if clarithromycin resistance is more than 15 % (Fig. 4) . Given that triple regimens using metronidazole are less effective in primary metronidazole-resistant strains and in view of the high prevalence of metronidazole resistance worldwide, a quadruple regimen can be used as first-line treatment when clarithromycin resistance is high. Metronidazole-based triple regimens can, however, be used safely in areas with a low prevalence of metronidazole resistance . It is not advisable to start with regimens containing both clarithromycin and metronidazole: if this treatment is used and then fails, empirical treatment is not possible and any ‘rescue’ regimen requires prior culture and antimicrobial susceptibility testing.
Finally, it is important to consider that failure of H. pylori eradication therapy is not always a consequence of antimicrobial resistance. Many other reasons, including patient, bacterial and treatment-related factors, may play an important role. It is well known that H. pylori eradication rate declines when patient compliance rates decrease . Results from recent studies also indicate that different genotypes of H. pylori may result in different clinical outcomes and may respond differently to antibiotic therapy. As an example, the cagA +/ vacA s1 genotype is associated with higher eradication rates . Moreover, dosing frequency and galenic formulations may also play an important role in H. pylori eradication .
Although clinical significance still remains unclear, there are prevalent digestive and extra-digestive diseases in which H. pylori infection may play an aetiological role . Consequently, the applicability of eradication therapy to these situations could enhance the relevance of H. pylori resistance.
1. Vandenplas Y. Helicobacter pylori
infection. Clin Microbiol Infect 1999, 5: 1–11.
2. Borody TJ, Cole P, Noonan S. et al
.: Recurrence of duodenal ulcer and Campylobacter pylori
infection after eradication. Med J Aust 1989, 151: 431–435.
3. De Boer W, Tytgat GNJ. Treatment of Helicobacter pylori
infection. BMJ 2000, 320: 31–34.
4. Malferthainer P. Compliance, adverse events and antibiotic resistance in Helicobacter pylori
treatment. Scand J Gastroenterol 1993, 28 (Suppl 196): 34–37.
5. Cantón R, García Arata I, Baquero F, de Rafael L. Aspectos microbiologícos de Helicobacter pylori
(in Spanish). In Infección por
Helicobacter pylori. Dónde está el límite?
Edited by Boixeda D, Gisber JP, Martín de Argila C. Barcelona: Prous Science; 1996: 13–33.
6. NCCLS:Methods for Dilution. Antimicrobial Susceptibility Test for Bacteria that Grow Aerobically. Approved Standard. 5th Edn.
Wayne, Pennsylvania: NCCLS document M7-A5; 1999.
7. Mégraud F. Resistance of Helicobacter pylori
to antibiotics. Aliment Pharmacol Ther 1997, 11 (Suppl 1): 43–53.
8. Chatsuwan T, Amyes GB. Setting the standard for determining the in vitro
susceptibility of Helicobacter pylori
to metronidazole. J Antimicrob Chemother 1999, 44: 291–293.
9. Marais A, Monteiro A, Occhiliani A, Pina M, Lamouliatte, Mégraud F. Direct detection of Helicobacter pylori
resistance to macrolides by polymerase chain reaction/DNA enzyme inmmunoassay in gastric biopsy specimens. Gut 1999, 44: 463–467.
10. Exner MM, Doig P, Trust TJ, Hancock REW. Isolation and characterization of a family of porin proteins from Helicobacter pylori
. Infect Immunol 1995, 63: 1567–1572.
11. DeLoney CR, Schiller NL. Competition of various β-lactam antibiotics for the major penicillin-binding proteins of Helicobacter pylori
: antibacterial activity and effects on bacterial morphology. Antimicrob Agents Chemother 1999, 43: 2702–2709.
12. Bina JE, Alm RA, Uria-Nickelsen M, Thomas SR, Trust TJ, Hancock REW. Helicobacter pylori
uptake and efflux: basis for intrinsic susceptibility to antibiotics in vitro
. Antimicrob Agents Chemother 2000, 44: 248–254.
13. Dunn BE, Cohen H, Blaser MJ. Helicobacter pylori
. Clin Microbiol Rev 1997, 10: 720–741.
14. Versalovic J, Shotridge D, Kliber K. et al
.: Mutations in 23S rRNA are associated with clarithromycin resistance in Helicobacter pylori
. Antimicrob Agents Chemother 1997, 41: 2550–2553.
15. Goodwin A, Kersulyte D, Sisson G, Veldhuyzen van Zanten SJ, Berg DE, Hoffman PS. Metronidazole resistance in H. pylori
is due to null mutations in a gene (rdxA
) that encodes an oxygen-insensitive NADPH nitroreductase. Mol Microbiol 1998, 28: 383–393.
16. Oliver A, Cantón R, Campo P, Baquero F, Blázquez R. High frequency of hypermutable Pseudomonas aeruginosa
in cystic fibrosis lung infection. Science 2000, 228: 1125–1253.
17. Haas CE, Nix DE, Schentag JJ. In vitro
selection of resistant H. pylori
. Antimicrob Agents Chemother 1990, 34: 1637–1641.
18. Bertram-Drogatz PA, Sobek-Klocke I, Moller C. et al
.: Growth characteristics and influence of antibiotics on rough/smooth phenotypic variants of Helicobacter pylori
. Eur J Clin Microbiol Infect Dis 1999, 18: 490–495.
19. Suerbaum S, Achtman M. Evolution of Helicobacter pylori:
the role of recombination. Trends Microbiol 1999, 7: 182–184.
20. Debets-Ossenkopp YJ, Pot RG, van-Westerloo DJ. et al
.: Insertion of mini-IS605 and deletion of adjacent sequences in the nitroreductase (rdxA
) gene cause metronidazole resistance in Helicobacter pylori
. Antimicrob Agents Chemother 1999, 43: 2657–2662.
21. Glupczynski Y, Burette A. Drug therapy for Helicobacter pylori
infection: problems and pitfalls. Am J Gastroenterol 1990, 85: 1545–1551.
22. Bazzoli F, Berretti D, De Luca L. et al
.: What can we learn from the new data about antibiotic resistance? Are there any practical clinical consequences of Helicobacter pylori
antibiotic resistance? Eur J Clin Gastroenterol Hepatol 1999, 11 (Suppl 2): S39–S42.
23. van Zwet AA, Thijs JC, Schievink de Vries W, Schiphuis J, Snijder JA. In vitro studies on stability and development of metronidazole resistance in Helicobacter pylori
. Antimicrob Agents Chemother 1994, 38: 360–362.
24. Hassan IJ, Stark RM, Greenman J, Millar MR. Activities of b-lactams and macrolides against Helicobacter pylori
. Antimicrob Agents Chemother 1999, 43: 1387–1392.
25. Berry V, Jennings K, Woodnutt G. Bactericidal and morphological effects of amoxicillin on Helicobacter pylori
. Antimicrob Agents Chemother 1995, 39: 1859–1861.
26. Lozniewski A, Duprez A, Renault C. et al
.: Gastric penetration of amoxicillin in a human Helicobacter pylori
-infected xenograft model. Antimicrob Agents Chemother 2000, 43: 1909–1113.
27. Dore M, Piana A, Carta M. et al
.: Amoxycillin resistance is one reason for failure of amoxycillin-omeprazole treatment of Helicobacter pylori
infection. Aliment Pharmacol Ther 1998, 12: 635–639.
28. van Zwet AA, Vanderbroucke-Grauls CMJE, ven der Wouden EJ. Stable amoxicillin resistance in Helicobacter pylori
. Lancet 1998, 352: 1595.1595.
29. Dore MP, Graham DY, Sepulveda AR. Different penicillin-binding profiles in amoxicillin-resistant Helicobacter pylori
. J Antimicrob Chemother 1999, 43: 47–54.
30. Soberg M, Hanberger H, Nilsson M, Bjrkman A, Nilsson LE. Risk of development of in vitro
resistance to amoxicillin, clarithromycin, and metronidazole in Helicobacter pylori
. Antimicrob Agents Chemother 1998, 42: 1222–1228.
31. Harris AG, Hazell SL, Netting AG. Use of digoxigenin-labelled ampicillin in the identification of penicillin-binding proteins in Helicobacter pylori
. J Antimicrob Chemother 2000, 45: 591–598.
32. Dore MP, Osato MS, Realdi G, Mura I, Graham DY, Sepulveda AR. Amoxicillin tolerance in Helicobacter pylori
. J Antimicrob Chemother 1999, 43: 47–54.
33. Taylor DE, Zhongmig GE, Purch D, Lo T, Hiratsuka K. Cloning and sequences analysis of two copies of a 23S rRNA gene from Helicobacter pylori
and association of clarithromycin resistance with 23S rRNA mutations. Antimicrob Agents Chemother 1997, 41: 2621–2628.
34. Wang GE, Taylor DE. Site-specific mutations in the 23S rRNA gene of Helicobacter pylori
confer two types of resistance to macrolides–lincosamide–streptogramin B antibiotics. Antimicrob Agents Chemother 1998, 42: 1952–1958.
35. García-Arata MI, Baquero F, de Rafael L. et al
.: Mutations in 23S RNA in Helicobacter pylori
conferring resistance to erythromycin do not always confer resitance to clarithromycin. Antimicrob Agents Chemother 1999, 43: 374–376.
36. Wang G, Rahman MS, Humayun MZ, Taylor DE. Multiplex sequence analysis demonstrated the competitive growth advantage of the G-to-A mutants of clarithromycin-resistant Helicobacter pylori
. Antimicrob Agents Chemother 1999, 43: 683–685.
37. Matsuoka M, Yoshida Y, Hayakawa K, Fukuchi S, Sugano K. Simultaneous colonisation of Helicobacter pylori
with and without mutations in the 23S rRNA gene in patients with no history of clarithromycin exposure. Gut 1999, 45: 503–507.
38. Moore RA, Beckthold B, Bryan LE. Metronidazole uptake in Helicobacter pylori
. Can J Microbiol 1995, 41: 746–749.
39. van Zewt AA, Thijs JC, Schievink de Vries W, Snijder JA. In vitro
studies on stability and development of metronidazole resistance in Helicobacter pylori
. Antimicrob Agents Chemother 1994, 38: 360–362.
40. Tankovic J, Lamarque D, Delchier JC, Soussy C-J, Labigne A, Jenks PJ. Frequent association between alteration of the rdxA
gene and metronidazole resitance in French and north African isolates of Helicobacter pylori
. Antimicrob Agents Chemother 2000, 44: 608–613.
41. Chang KC, Ho SW, Yang JC, Wang JT. Isolation of a genetic locus associated with metronidazole resistance in Helicobacter pylori
. Biochem Biophys Res Commun 1997, 236: 785–788.
42. Cantón R, de Rafael L, Boixeda D, Serrano MT, García-Gonzalez M, Baquero F. In vivo
acquired resistance to fluoroquinolones in Helicobacter pylori
during therapy with ciprofloxacin or ciprofloxacin plus amoxicillin/clavulanate. In Helicobacter pylori and Gastroduodenal Pathology.
Edited by Pajare MJ, Malfertheiner P. Berlin: Springer-Verlag; 1993: 305–311.
43. Moore RA, Beckthold B, Wong B, Kureishi A, Bryan L. Nucleotide sequence of the gyrA
gene and characterization of ciprofloxacin-resistant mutant of Helicobacter pylori
. Antimicrob Agents Chemother 1995, 39: 107–101.
44. Midolo PD, Korman MG, Turnbridge JD, Lambert JR. Helicobacter pylori
resistance to tetracycline. Lancet 1996, 347: 1194–1195.
45. Karim QN, Logan RPH. The epidemiology of Helicobacter pylori
antimicrobial resistace – preliminary findings of a national survey. Gut 1995, 37 (Suppl 2): A6.A6.
46. Hepp M, Beck D, Bayerdörffer E, Lehn N. Rifampin and rifabutin resistance mechanism in Helicobacter pylori
. Antimicrob Agents Chemother 1999, 43: 1497–1499.
47. Hepp M, Rieger U, Beck D, Lehn N. Mutations in the beginning of the rpoB
gene can induce resistance to rifamycins in both Helicobacter pylori
and Mycobacterium tuberculosis
. Antimicrob Agents Chemother 2000, 44: 1075–1077.
48. Hepp M, Odenbreit S, Beck D. et al
.: Mutations at four distinct regions of the rpoB
gene can reduce the susceptibility of Helicobacter pylori
to rifamycins. Antimicrob Agents Chemother 2000, 44: 1713–1715.
49. Stratton CW, Warner RR, Coudron PE, Lilly NA. Bismuth-mediated disruption of the glycocalyx-cell wall Helicobacter pylori
: ultrastructural evidence for a mechanism of action for bismuth salts. J Antimicrob Chemother 1999, 43: 659–666.
50. Mobley HL, Garner RM, Chippendale GR, Gilbert JV, Kane AV, Plaut AG. Role of Hpn and NixA of Helicobacter pylori
in susceptibility and resistance to bismuth and other metal ions. Helicobacter 1999, 4: 162–169.
51. Megraud F, Lehn N, Lind T. et al
.: Antimicrobial susceptibility testing of Helicobacter pylori
in a large multicenter trial: the MACH study. Antimicrob Agents Chemother 1999, 43: 2747–2752.
52. Lind T, van Zanten SV, Unge P. et al
.: Eradication of Helicobacter pylori
using one week triple therapies combining omeprazole with two antimicrobials: the MACH 1 study. Helicobacter 1996, 1: 138–144.
53. Alarcon T, Domingo D, López-Brea M. Antibiotic resistance problems with Helicobacer pylori
. Int J Antimicrob Agents 1999, 12: 19–26.
54. Osato MS, Reddy R, Graham DY. Metronidazole and clarithromycin resistance amongst Helicobacter pylori
isolates from a large metropolitan hospital in the United States. Int J Antimicrob Agents 1999, 12: 341–347.
55. van Zwet AA, Deboer WA, Schneeberger PM, Weel J, Jansz AR, Thijs JC. Prevalence of primary resistance to metronidazole and clarithromycin in The Netherlands. Eur J Clin Microbiol Infect Dis 1996, 15: 861–864.
56. Boyanova L, Spassova Z, Krastev Z. et al
.: Characteristics and trends in macrolide resistance among Helicobacter pylori
strains isolated in Bulgaria over four years. Diagn Microbiol Infect Dis 1999, 34: 309–313.
57. Wolle K, Nilius M, Leodolter A, Muller WA, Malfertheiner P, Koning W. Prevalence of Helicobacter pylori
resistance to several antimicrobial agents in a region of Germany. Eur J Clin Microbiol Infect Dis 1998, 17: 519–521.
58. Broutet N, Guillon F, Sauty E, Letuaire D, Megraud F. Survey of the in vitro
susceptibility of Helicobacter pylori
to antibiotics in France. Gastroenterology 1998, 114: A81.A81.
59. Goddard AF, Logan RPH. Antimicrobial resistance and Helicobacter pylori
. J Antimicrob Chemother 1996, 37: 639–643.
60. Graham DY. Antibiotic resistance in Helicobacter pylori
: implications for therapy. Gastroenterology 1998, 115: 1272–1277.
61. Ducóns JA, Santolaria S, Guirao R, Ferrero M, Montoro M, Gomollón F. Impact of clarithomycin resistance on the effectiveness of a regimen for Helicobacter pylori
: a prospective study of 1-week lansoprazole, amoxycillin and clarithromycin in active peptic ulcer. Aliment Pharmacol Ther 1999, 13: 775–780.
62. Mégraud F, Doermann HP. Clinical relevance of resistant strains of Helicobacter pylori
: a review of current data. Gut 1998, 43 (Suppl 1): S61–S65.
63. Moayyedi P, Ragunathan PI, Mapastone N. et al
.: Relevance of antibiotic sensitivities in predicting failure of omeprazole, clarithromycin, and tinidazole to eradicate Helicobacter pylori
. J Gastroenterol 1998, 33 (Suppl 10): 62–65.
64. Hachem CY, Clarridge JE, Reddy R. et al
.: Antimicrobial susceptibility testing of Helicobacter pylori
. Comparison of E-test, broth microdilution, and disk diffusion for ampicillin, clarithromycin, and metronidazole. Diagn Microbiol Infect Dis 1996, 24: 37–41.
65. Xia H, Keane CT, Beattie S. et al
.: Standardization of disk diffusion test and its clinical significance for susceptibility testing of metronidazole against Helicobacter pylori
. Antimicrob Agents Chemother 1994, 38: 2357–2361.
66. Xia HN, Kalantar J, Talley NJ. Metronidazole- and clarithromycin-resistant Helicobacter pylori
in dyspeptic patients in western Sydney as determined by testing multiple isolates from different gastric sites. J Gastroenterol Hepatol 1998, 13: 1044–1049.
67. Midolo PD, Turnidge J, Lambert JR, Bell JM. Validation of a modified Kirby–Bauer disk difussion method for metronidazole susceptibility testing of Helicobacter pylori
. Diagn Microbiol Infect Dis 1995, 21: 135–140.
68. Houben MHMG, van de Beek D, Hensen EF, de Craen AJM, Rauws EAJ, Tytgat GNJ. A systematic review of Helicobacter pylori
eradication therapy – the impact of antimicrobial resistance on eradication rates. Aliment Pharmacol Ther 1999, 13: 1047–1055.
69. Bucley MJ, Xia HX, Hyde DM. et al
.: Metronidazole resistance reduces efficacy of triple therapy and leads to secondary clarithromycin resistance. Dig Dis Sci 1997, 42: 2111–2115.
70. Miyaji H, Azuma T, Ito S. et al
.: Susceptibility of Helicobacter pylori
isolates to metronidazole, clarithromycin and amoxicillin in vitro
and in clinical treatment in Japan. Aliment Pharmacol Ther 1997, 11: 1131–1136.
71. Huang JQ, Hunt RH. Treatment after failure: the problem of ‘non-responders'. Gut 1999, 45 (Suppl 1): I40–I44.
72. Huang JQ, Hunt RH. The importance of clarithromycin dose in the management of Helicobacter pylori
infection: a meta-analysis of triple therapies with a proton pump inhibitor, clarithromycin and amoxicillin or metronidazole. Aliment Pharmacol Ther 1999, 13: 719–729.
73. The Report of the Digestive Health Intiative International Update Conference on Helicobacter pylori
74. Current European concepts in the management of Helicobacter pylori
infection. The Maastricht Consensus Report. Gut
75. Graham DY, de Boer WA, Tytgat GNJ. Choosing the best anti-Helicobacter pylori
therapy: effect of antimicrobial resistance. Am J Gastroenterol 1996, 91: 1072–1076.
76. De Boer WA, Tytgat GNJ. How to treat H. pylori
infection. Should treatment strategies be based on bacterial susceptibility? A personal viewpoint. Eur J Gastroenterol Hepatol 1996, 8: 709–716.
77. Wermeille J, Dederding JP, Cunningham M. et al
.: Failure of Helicobacter pylori
eradication in an ambulatory population: is poor compliance the main cause? Digestion 1998, 59 (Suppl 3): 437.437.
78. Van Doorn LJ, Schneeberger PM, Nouhan N, Plaisier AP, Quint WGV, de Boer WA. Importance of Helicobacter pylori cagA
status for the efficacy of antibiotic treatment. Gut 2000, 46: 321–326.
79. Martín de Argila C, Boixeda D. Extra-digestive manifestations of Helicobacter pylori
infection. Science or fiction? Med Clin (Barc) 2000, 114: 308–317.