Helicobacter pylori : From bench to bedside

Can J Gastroenterol Vol 11 No 7 October 1997 589 N Chiba, A Matisko, P Sinclair, ABR Thomson. Helicobacter pylori: From bench to bedside. Can J Gastroenterol 1997;11(7):589-596. With the exponential increase in research in the field of Helicobacter pylori a paradigm shift has occurred. It is now recognized that H pylori is a chronic infection of the stomach causing inflammation. Some patients remain asymptomatic, while others may develop dyspepsia, duodenal or gastric ulcer, gastric cancer or a mucosa-associated lymphoid tissue lymphoma. However, the role of H pylori in contributing to nonulcer dyspepsia or nonsteroidal anti-inflammatory drug gastropathy remains controversial. An effective vaccine against H pylori is years away. Major interest has focused on the questions “who should be investigated and therefore treated” and “what is the latest gold standard for eradication of H pylori”? In Europe, guidelines have been developed to help the practitioner answer these important questions. Canadian guidelines will soon be available. For persons with known peptic ulcer disease there should be unequivocal acceptance that the good clinical practice of eradicating H pylori will result in substantial savings in health care expenses. The original ‘classical triple therapy’ (bismuth, metronidazole and tetracycline [BMT]) has now been surpassed by the combination of a proton pump inhibitor (PPI) plus two antibiotics (metronidazole plus clarithromycin; amoxicillin plus clarithromycin; or amoxicillin plus metronidazole), each given twice a day for one week. In Canada, the regimen of omeprazole plus one antibiotic (amoxicillin or clarithromycin) was approved recently but gives an eradication rate that is lower than the current target of 90%. According to the European (Mäastricht) recommendations, if a single treatment attempt with PPI plus two antibiotics fails, PPI plus BMT is recommended.

T he International Workshop on Gastroduodenal Pathol- ogy and Helicobacter pylori is an annual meeting organized by the European Helicobacter Pylori Study Group (EHPSG).The EHPSG was founded in Copenhagen in 1987, and the IXth International Workshop in Gastroduodenal Pathology and Helicobacter pylori was again held in Copenhagen, Denmark from October 16-19, 1996.Over the past 10 years there has been explosive progress in the field of H pylori.At this meeting, a broad scientific program was chaired by most of the leading people in the field of H pylori.The purpose of the meeting continues to be achieved with a high scientific standard -promoting and inspiring scientists to increase the progress of knowledge within the fields of H pylori and gastroduodenal pathology.
This review of the important advances reported at the meeting is organized into sections.Whenever possible, clinically relevant issues are raised and important questions answered.References to specific presentations are annotated using the abstract numbers published in GUT Supplement 2:39, 1996.These will not be relisted here, and the interested reader is encouraged to refer to this supplement.Information from presentations made at the plenary sessions of the IXth International Workshop are referenced by the speaker's name followed by EHPSG.
A European consensus (Mäastricht, September, 1996) has been developed for the management of patients with H pylori infection and was communicated (Malfertheiner, EHPSG, 1996) at this meeting.These forward-looking Mäastricht perspectives are mentioned throughout this review.

EPIDEMIOLOGY: THE NATURAL HISTORY OF H PYLORI INFECTION
The different prevalence of H pylori infection among newborns, children and young adults residing in the same geographic region may be due to ethnic origin, ie, German versus Turkish (Doppl, 4B:18).In some developed countries such as New Zealand, the seroprevalence of H pylori in a birth cohort of 21-year-olds may be very low (4.1%) (Fawcett, 4B:22).In developing countries, the seroconversion from H pylori-negative to -positive may be as high as 4% per year (Oliveira, 4B:25).However, much lower seroconversion rates of 0.33% per year (Menegatti, 4B:14) may occur, and one Danish study suggested that seroreversion occurs more frequently than seroconversion (Rosenstock 4B:06).
It is postulated that H pylori transmission occurs by the fecal-oral route.If this is the case, one might expect a close correlation between the prevalence of H pylori and that of hepatitis A. While the prevalence of both conditions increases with age, they are not necessarily present in the same person (Luzza, 4B:04).H pylori has been detected by sensitive polymerase chain reaction methods in 12% of municipal water supplies in Sweden, with even higher rates (38%) in well water (Hultén, 4B:32); however this finding is controversial.House flies are a possible vector for the spread of H pylori by the fecal-oral route (Grübel, 4B:33), but this interesting observation also needs to be confirmed.Lower socioeconomic status (Malaty, 4B:20), childhood living condi-tions and close person-to-person contact are all important in the transmission of H pylori (Rothenbacher, 4B:07; Martin de Argila, 4B:05).Children who have their own bedroom have a lower prevalence of H pylori infection than those sharing their bedrooms with others (Rothenbacher, 4B:07).Interestingly, molecular typing of H pylori indicates that different siblings in the same household can be infected with different strains (Rautelin, 2B:24).
In children born to H pylori-positive mothers, transplacental transfer of antibody was evidenced by high immunoglobulin G (IgG) titres at birth, which became negative by seven months and remained negative at 11 years of age (Ashorn, 2B:01).In developing countries, maternal seropositivity did not increase the risk of acquiring H pylori infection in childhood.Furthermore, in developing countries, transplacental anti-H pylori IgG does not provide protection against H pylori colonization in the first year of life.
Although 82% of spouses were positive for H pylori, 7.4% of cohabiting patients had a recurrent H pylori infection in the first year after initial eradication (Gisbert 4B:16).Thus, while the risk of reinfection is low, the overall rate of acquisition is higher than that seen in the general population (Menegatti, 4B:14).

HOST-BACTERIAL INTERACTIONS
The organism: The complete DNA sequence of the genome of a representative H pylori strain (KE26695) has been determined (Tomb, 3B:59).This is a remarkable achievement and will accelerate H pylori research.The high degree of H pylori heterogeneity may be due to point mutations in conserved genes; mosaicism in conserved genes (vacA genotypes) as well as nonconserved genes (cagA and cag pathogenicity island); or genomic rearrangement (map differences) of extragenetic elements (IS 605 and plasmids) (Blaser, EHPSG, 1996).Lewis X (LeX) is a complex human carbohydrate molecule that has been identified on the surface of H pylori.This antibody may act as a crosslinking agent to enhance bacterial binding to host epithelium (Taylor, 3B:75).The LeX antigen is involved in this diversity, as is the m1 and m2 dicotomy of the vacA products (at least in the United States, but to a lesser extent in Europe and to a much lesser extent in Asia).The s1/m1 vacA genotype may be seen more frequently in DU and gastric carcinoma patients than in patients with gastritis alone (Mendes, 3B:74); however, geographic differences may exist because this genotype has also been found in the same frequency in patients with DU and asymptomatic H pylori infection in the United States (Go, 3B:06).Almost no ulcer patients are infected with the s2 genotype.Understanding of genetic diversity is complicated by research using multilocus enzyme electrophoresis, which has demonstrated sufficient heterogeneity in the chromosomal gene to suggest that H pylori may not even be a single species (Hazell 3B:101).
The prevalence of the cagA gene is uniform across all age groups; however, there is wide variability (Crabtree, 4B:08) throughout the world (eg, 41.9% of H pylori strains are cagA-positive in Canada compared with 82.2% in Peru [Perez-Perez, 4B:11]).cagA positivity may be a 'pathogenicity island' associated with a greater risk of the H pylori carrier developing peptic ulcer disease (PUD) or gastric cancer (GC).This may be the result of a greater bacterial density, more H pylori nearer the epithelial surface than further out in the alkaline microenvironment mucus layer, greater interleukin-8 (IL-8) production, altered gastric acid secretion or more rapid development of atrophic gastritis.Interestingly, cagA-positive strains express the LeX antigen more frequently.In Japanese patients with diffuse-type gastric cancer, there is no association with cagA status.However, in intestinal-type GC there is an association with gastric atrophy and the presence of cagA (Kikuchi 3B:76).
The induced by contact with epithelium (ice) A gene (Peek, 3B:87) may be a marker for DU -iceA is seen in 100% of DU patients, whereas iceA2 is seen in only 29%.Also, iceA strains are associated with higher antral mucosal IL-8 levels than iceA2 (Peek 3B:87).In comparison, cagA is positive in 90% of iceA strains versus 52% of iceA2 strains.A lethal target gene (LtsH) from H pylori has been cloned and characterized (Ge, 3B:43).Gastric Th1 cells predominate in response to infection with H pylori, and these cells contribute to epithelial cell death through apoptosis, which may predispose the host to peptic ulceration (Crowe 3B:21).In contrast, the Th2 phenotype is associated with protection from or control of infection (Mohammardi, 1B:29).These markers distinguish among strains in terms of their virulence.There also appears to be DNase-sensitive (transformation) and DNase-resistant mechanisms that contribute to DNA transfer between H pylori cells (Kiopers, 1A:26).The development of this diversity is ongoing, with evolutionary change occurring in the stomach of individuals during their lifetimes.
It should be stressed that a person may be infected with more than one strain of H pylori.The existence of multiple infections and quasispecies indicates that analysis of a single H pylori isolate is not sufficient to define the genotype of H pylori strains that may be present in a given patient (Blaser, EHPSG, 1996).Changes in the epidemiology of H pylori from developing to developed countries may be due to the decline in the number of H pylori strains per infected person (multiplicity of infection).
At least in mice, intranasal or intrarectal routes may be superior to the oral route for immunization with recombinant urease (Kleanthous, 3B:107).Urease immunization augments the ability of antimicrobials to eradicate Haemobartonella felis infection in mice (Kleanthous 3B:106).In other models, it may be possible to confer protective immunity to Rhesus monkeys through oral immunization with recombinant urease and Escherichia coli heat-labile enterotoxin (Dubois, 1B:20).Oral immunization with recombinant urease is at an early stage in asymptomatic H pylori-infected humans (Kreiss, 1B:01).
The host: The classical view of acid secretion holds that there are cephalic, gastric and intestinal phases of gastric hydrochloric acid ('acid') secretion.Food in the stomach stimulates the release of gastrin from antral G cells.Gastrin, in turn, stimulates acid secretion from the parietal cells of the gastric body.The release of gastrin is inhibited by somatostatin released from antral D cells, as well as by cholecytokinin (CCK) released when acid empties from the stomach and bathes the duodenal mucosa.CCK stimulates somatostatin release, which inhibits gastrin release, with a lowering of acid secretion.
The control of gastric acid secretion tested with gastrinreleasing polypeptide (GRP) is disrupted by H pylori infection, and is further altered by the presence of a DU (McColl, EHPSG, 1996).H pylori infection in healthy volunteers or in DU patients enhances both basal and stimulated gastrin release in response to GRP, which increases acid secretion, lowers somatostatin mRNA in the antral mucosa and impairs CCK inhibition of gastrin release.The result is marked impairment of the normal acid inhibition of gastrin release.
It is not known how H pylori impairs the somatostatinmediated acid inhibitory control of gastrin release.Speculated mechanisms include elevation of antral surface pH due to H pylori-associated secretion of ammonia (because gastric acid exerts trophic effects on these cells); atrophy of D cells produced by the bacterium's urease activity or a possible influence of inflammatory cytokines; or N-a-methyl histamines, which may suppress somatostatin synthesis via the histamine-3 receptor.Of interest, the cagA-positive strain of H pylori is more prevalent in persons who develop DU, and GRP-stimulated gastrin release is greater in healthy volunteers and in persons with nonulcer dispepsia (NUD) or DU who are infected with cagA-positive strains than in those with cagA-negative strains of H pylori. Somatostatin levels are lower in cagA-positive infected persons than in cagAnegative persons (Queiroz, 3B:37).
As a group, DU patients are known to have multiple abnormalities of acid secretion including impaired acid inhibition of gastrin release, increased basal and stimulated acid secretion, and a greater duodenal acid load.While these abnormalities may be due in part to the presence of H pylori infection (as already discussed above), there is evidence that there may be abnormalities in acid secretion that are not due to H pylori. Instead, these abnormalities are likely due to an inherent gastric defect in acid secretion.This is evidenced by a greater increase in acid secretion than expected from the same amount of hypergastrinemia in H pylori-positive DU patients than in H pylori-positive healthy volunteers.This can be expressed physiologically by comparing the maximum acid output (MAO) and the D50 of the curvilinear relationship between increasing doses of infused gastrin17 (pmol/kg/h) and acid output (mmol/h).The MAO is greater and the D50 is less in H pylori-positive DU patients than in H pylori-positive or -negative healthy volunteers.Furthermore, there is little change in the elevated MAO after eradication of H pylori.This suggests that the exaggerated response of the parietal cell to gastrin is related to the host and is not due soley to the presence of H pylori. Cigarette smoking is important in the pathogenesis of DU, and smoking is independently associated with increased parietal cell mass (McColl, EHPSG, 1996).
How important is this abnormality in acid secretion in the pathogenesis of DU?There is a linear relationship between peak acid output and duodenal-gastric metaplasia (DGM) -with more acid, DGM increases, and inhibition of acid secretion may or may not (Suriani, 3C:02) result in partial resolution of DGM.Therefore, acid may be a key part of the mechanism by which an H pylori infection produces an ulcer.

H PYLORI AND ITS DISEASES: BEYOND GASTRITIS AND ULCERS
GC and lymphoma: Gastric mucosa-associated lymphoid tissue (MALT) lymphoma represents clonal disease of postgerminal B cells (Thiede, 2A:08).Posteradication B cells in MALT lymphoma may still respond to antigen (Thiede, 2A:08).Patients with H pylori-associated MALT lymphoma may be treated successfully in specialty units by H pylori eradication, with regression occurring in more than 70% (Bayerdörffer, 2A:01).
In contrast with the rareness of gastric MALT lymphoma, GC must be viewed as a major public health problem.While the incidence of mortality from GC has declined steadily in countries such as Japan, and in North America, it remains a common cause of cancer mortality worldwide, especially in developing countries (Forman, EHPSG, 1996).In patients with GC, the five-year mortality rate varies widely, ranging from 6.9% to 23.3% from country to country.
How well established is the association between GC and H pylori? A meta-analysis of original data of eight prospective published and unpublished studies showed a pooled odds ratio of 2.0 for all cases of GC, and 2.8 for cases excluding GC of the cardia.In fact, the odds ratio for GC of the cardia was 0.9, almost low enough to speculate that H pylori infection is actually protective with respect to the development of cancer in this region.The strength of the association between GC and H pylori may become more evident with increasing levels of salt intake, a known risk factor for GC (1).
There appears to be an international association between the prevalence of infection with cagA-positive strains of H pylori and mortality from GC (Webb, 2A:07).The relative risk of GC with cagA-positive H pylori strains was 15.8 compared with a relative risk of 5.6 for H pylori per se.A high molecular weight cagA may be involved in the pathogenesis of H pylori-associated GC in Korea, or may simply be a marker for H pylori strains with a predilection to cause GC or multifocal gastritis with atrophy (Miehlke, 2A:11).GC is more likely to occur with cagA-positive than with cagA -negative strains of H pylori, particularly GC of the intestinal histotype form (Figura, 2A:13).Clearly, it is not appropriate to suggest that cagA-negative strains are harmless, but perhaps in time it will be possible to target which asymptomatic persons with H pylori should be treated with eradication therapy to prevent the development of GC.Presently, there is no evidence that screening for and eradicating H pylori reduces the risk of GC.
It would perhaps be more important to ask whether eradication of H pylori prevents development of GC.Most studies suggest that eradication of H pylori does not result in regression of intestinal metaplasia or atrophy in the stomach (Annibale, 1C:05; Fossati, 3C:08).Part of the problem with showing significant differences is the higher levels of interobserver disagreement in grading atrophy than in other histological parameters (Fossati, 3C:08).Nonrandomized data have shown that eradication of H pylori in Japanese patients treated by endoscopic resection of early GC prevents the development of new cases of GC in the following four years (none of 65 patients experienced recurrences with H pylori eradication, and six of 67 experienced recurrences without H pylori eradication) (2).This supports the recommendation of the Mäastricht meeting that H pylori eradication should be undertaken in persons treated for early GC.

Which extragastric diseases are linked to H pylori infection?
To determine whether an association exists between H pylori and conditions such as short stature or coronary artery disease (CAD), it is important to assess the consistency among studies, the strength of the association, whether there is a temporal relationship and the study design (case-control, cross-sectional cohort, prospective cohort or ideally a randomized, controlled study) (Veldhuyzen van Zanten, EHPSG, 1996).A growth disadvantage has been reported for H pylori-infected girls (3).This may be a marker for low socioeconomic status, although it is unclear why it would have such a strong sex bias (1.1 to 1.6 cm for girls, 0.2 cm for boys).
Likewise, a lower socioeconomic status may also be a risk factor for CAD, and the same argument can be made for the association between CAD and H pylori reported in a crosssectional study (4).Indeed, when a person's socioeconomic status is taken into account, the association between CAD and H pylori is lost (5).At this meeting, findings were divided with some showing an association of H pylori with CAD (Ossei Gerning, 1C:10; Caselli, 1C:12; Aromaa, 1C:13; Martin de Argila, 1C:30) and others finding no such association (Vakil, 1C:11; Strandberg, 1C:16; Maier, 1C:17).

DIAGNOSTIC ASPECTS OF H PYLORI INFECTIONS
The 'best' test to diagnose H pylori depends on the clinical setting, ie, serology for screening dyspeptic patients; histology and/or culture for confirmation of the diagnosis or in the already treated patient; and the 13G/14C urea breath test (UBT) for follow-up to confirm successful eradication (Megraud, EHPSG, 1996).Treatment with H 2 -receptor antagonists may reduce the sensitivity of urease tests (Lerang 4C:08).A single UBT performed one month post-treatment is noninvasive and reliable.However, the usefulness of the UBT in clinical practice depends on the availability of analytical facilities.This test overestimated eradication success by less than 5% compared with either two diagnostic tests at one month, or two serial UBTs at one and three months (Johnson, 4C:52).Whole blood 'office' antibody tests have a sensitivity and specificity of about 50% to 95% (Stone, 4C:15; Schrier, 4C:68; Crane, 4C:69) so there may be a problem of false-positive results, which would result in inappropriate anti-H pylori treatment of some dyspeptic patients.As in adults, the 13 C-UBT is useful to diagnose H pylori infection in children.A 2 h fast before testing is sufficient, and 30 mins is the optimal sampling time (Rowland, 2B:16; Bazzoli, 2B:05).
Polymerase chain reaction is a new but expensive diagnostic test that is highly sensitive but may be associated with false positive results and, hence, will likely have limited practical use (Hyde, 4C:60).In phase contrast microscopy, a gastric biopsy is smeared on a slide, a drop of saline is added and covered with a cover slip, and the endoscopist looks for the organism through a microscope.After eradication treatment, phase contrast microscopy combined with urease testing detects more H pylori-positive cases than histology alone (Daskalopoulos 4C:37).

WHO TO TREAT?
At the EHPSG this year, the major area of consideration was not how to diagnose H pylori but rather who needs to be or should be tested, and therefore treated?Patterns, expectations and guidelines: H pylori is a chronic infection of the stomach that causes chronic active gastritis.The infection has a wide spectrum of disease manifestation, remaining asymptomatic in many persons, causing gastric ulcers and DUs in others, and possibly causing GC or lymphoma in still other infected individuals.While it may be speculated that 'good' strains of H pylori exist that confer some benefit on the infected host (Blaser, EHPSG, 1996), there is no evidence for this possibility.And so the provocative question is, "Is there still a reason for withholding H pylori eradication treatment in those known to be infected?".The answer is probably not (Marshall, EHPSG, 1996); knowledge of positive H pylori status usually obliges the physician to treat the infection, so it is recommended to test for H pylori only in those patients a doctor plans to treat.
What is the attitude of physicians in the United States with respect to eradication of a known H pylori infection in a person with dyspepsia and suspected versus proven peptic ulceration?It appears that only about half of family physicians or internists treat symptomatic H pylori patients without a firm ulcer diagnosis, and an even lower proportion of gastroenterologists use this approach (Table 1) (Breuer, 3A:03).Quite appropriately, most family physicians, internists or gastroenterologists use eradication treatment for H pylori in the patient with a proven DU, whether on first or recurrent presentations.This difference in the approach to the H pylori-positive dyspeptic patient is important when considering the development of treatment algorithms.
The United States National Institutes of Health (NIH) consensus (1994) recommended H pylori treatment for the ulcer patient on first presentation or recurrence, or in ulcer patients on maintenance therapy with acid inhibitory drugs.The Mäastricht consensus (1996) confirmed and extended the NIH approach to include the H pylori-positive person with an nonsteroidal anti-inflammatory drug-associated ulcer and the H pylori-positive person with a bleeding ulcer (Misiewicz, EHPSG, 1996).
Of three approaches to treat the patient with a known DU -antisecretory therapy for a short interval; antisecretory therapy initially followed by prolonged maintenance with a lower dose of an antisecretory drug; or H pylori eradicationthe latter results in fewer ulcer recurrences (approximately 75%, 25% and 5% recurrence rates per year, respectively) and fewer episodes of rebleeding (approximately 25% versus less than 5% per year for maintenance versus eradication therapy).Early eradication of H pylori in ulcer patients is a superior cost-effective strategy compared with traditional antisecretory therapy.For example, in Belgium the ratio of annual costs for various treatment strategies is 1:3:4.5 for eradication, episodic H 2 receptor antagonist/PPI, maintenance H 2 receptor antagonist and maintenance PPI, respectively (Deltenre, EHPSG, 1996).The costs of endoscopy and medication in Belgium are close to those of Canada, and it is estimated that eradicating H pylori in dyspeptic persons saves the Belgian health care system (ie, direct costs) a million dollars American per year per million inhabitants.It is possible that the Mäastricht approach to the dyspeptic patient (screen for H pylori, then treat) will be cost effective in Canada, where the cost of endoscopy (EGD) is much lower than in the United States.However, the availability of this diagnostic test is perhaps less than ideal in some Canadian communities (6).The empirical approach to eradication therapy (neither endoscopy nor H pylori testing performed) saves costs only when EGD costs more than US$500 (7).The physician's fee for EGD is much less in Canada, but there are considerable hospital-related acquisition, maintenance and staffing costs.Therefore, screening for H pylori and treating without EGD may be ethically and economically acceptable.In the United States, treatment of acute DU with omeprazole plus clarithromycin, compared with omeprazole alone or ranitidine alone reduces by about 50% the number of needed EGDs, ulcer-related clinic visits, ulcer-related days lost from work, all hospitalizations and all hospital days, and reduces to zero ulcer-related hospitalizations and therefore hospital days (Sonnenberg, 3A:07).With more effective treatment with PPI and two antibiotics, even greater cost savings are be expected.The cost effectiveness of H pylori status-based management strategies for dyspeptic patients is increasingly recognized (Bazalle, EGPSG, 1996; 8).H pylori eradication is effective in some patients,  3A:25).One of the central questions is "does eradication of H pylori before beginning NSAIDs reduce the prevalence of later complications?".In a preliminary report from Hong Kong, eradicating H pylori before a course of naproxen resulted in fewer ulcers as well as less pain and bleeding (Lee, 3A:24).In another important study, patients with NSAIDinduced ulcers and erosions were treated with either H 2 receptor antagonists or anti-H pylori eradication treatment.In the latter group, the healing rate of the mucosal lesions was much greater even though the NSAIDs were continued (Kordecki, 3A:29).
A second important question is "does eradication of H pylori decrease the risk of bleeding from NSAIDs?".Eradication of H pylori in patients on NSAIDs who have established gastric ulcer or DU may be considered, but this should not be substituted for treatment of such ulcers on their own merits with acid suppressive drugs or with mucosal protection prostaglandin analogues (Misiewicz, EHPSG, 1996).

HOW TO TREAT What and at what costs?
The Mäastricht consensus recommended, based on currently available data, that the first line of therapy for H pylori eradication is 'PPI plus two' (proton pump inhibitor plus two antibiotics).No distinction was made about which PPI should be used, but clearly the foundation to establish the efficacy of this combination must be based on randomized, controlled clinical trials and the majority of data to date are based on omeprazole-based combination therapies.Studies with newer data on lansoprazole triples (Catalano, 4A:08; Ho, 4A:09; Burette, 4A:13; Misiewica, 4A:14) have reported comparable efficacy with omeprazole triples, but one study found disappointing results (Lamouliatte, 4A:20).Pantoprazole was used in one triple therapy study with high efficacy (Adamek, 4A:32).The previous gold standard triple therapy (bismuth, metronidazole and tetracycline [BMT]) has lost its glitter.The 'three PPI plus two' regimens give a 85% to 90% success rate of eradication with a simple, well-tolerated, twice-a-day, one-week program combining two of metronidazole, clarithromycin or amoxicillin.A suggested sequence of therapeutic choices is as follows.
• First choice: PPI + metronidazole In an American survey, 77% of gastroenterologists used the most effective triple or quadruple therapies, while only 15% used outdated dual therapy.Among family practitioners, the corresponding figures were 60% and 13%, respectively, and a further 26% used some other ineffective regimen.These data confirm that education of colleagues about the most effective regimens is still a high priority (Breuer, 3A:06).A survey of Dutch gastroenterologists indicated that their treatment choices were already in keeping with the Mäastricht recommendations: PPI triples were used by 40%, PPI quadruple therapy by 26%, and only 14% still used the classical BMT (Boekema, 3A:09).
Whether an antisecretory agent is needed at all was answered by a randomized trial in which NUD patients were treated with either clarithromycin alone, omeprazole plus clarithromycin, clarithromycin plus tinidazole, or omeprazole plus clarithromycin plus tinidazole.An eradication rate of 93.8% was achieved with omeprazole plus clarithromycin plus tinidazole, 59.4% with clarithromycin plus tinidazole, 31.3% with omeprazole plus clarithromycin and 6.3% with clarithromycin (Bazolli, 4A:12).Clearly a PPI is needed to achieve an acceptable rate of eradication.
While it appears that a PPI is the necessary antisecretory drug, the issue is not settled, particularly because H 2 receptor antagonists are cheaper.In one cohort study of active DU patients, triple therapy with clarithromycin, metronidazole and either ranitidine 150 mg bid or omeprazole 20 mg once 594 Can J Gastroenterol Vol 11 No 7 October 1997 daily was given.Eradication rates in both arms were similar, with success in about 80% (Sacca, 4A:18).However, this is a lower eradication rate than is usually seen with omeprazole, clarithromycin and metronidazole combinations.Further, direct comparative studies are needed, with particular reference to antibiotic resistance, which was not assessed in this study.
In PPI-based triple therapies, it is still unclear whether the PPI is needed once or twice daily.One study with lansoprazole 30 mg once daily or bid in combination with amoxicillin and clarithromycin found a lower eradication rate (70%) with the once daily PPI arm than with the twice daily treatment arm (89.7%) (Catalano, 4A:08).However, the issue is far from settled, with other studies finding no difference between PPI given once or twice daily (Chiba, 4A:28; Burette, 4A:13).
Although it is generally agreed that one week of treatment is adequate, one study reported that efficacy improved with 12 days (84.3%) of omeprazole, clarithromycin, ameprazole triple therapy, compared with six days (61.6%) (Hermida, 4A:11).
A variety of new treatments were also reported.zole 20 mg bid, clarithromycin 500 mg bid and Denol (not available in Canada) 120 mg qid for one week resulted in 89.5% eradication (Georgopoulous, 4A:17).Ranitidine bismuth citrate 400 mg bid for one month plus clarithromycin 500 mg bid for two weeks (dual therapy) resulted in an eradication rate of 83%, equivalent to that of triple therapy with ranitidine bismuth citrate, clarithromycin and metronidazole 400 mg bid for two weeks (92%) (Bardhan, 4A:24).Further work with these new regimens is anticipated with considerable interest.
Although the Mäastricht conference suggested that classical triple therapy with BMT has been superseded, a metaanalysis showed that overall BMT is quite effective (81.8% eradication, seven to 14 days) (Chiba, 4A:27).However, H pylori metronidazole resistance significantly reduces eradication success from 89.4% in sensitive strains to 50.6%.Greater acid suppression with the addition of omeprazole to BMT (OBMT) achieves 94.8% eradication after only seven days.This quadruple therapy was evaluated in nine study arms and was very consistent (95% CI, 90.8-98.8)(Chiba 4A:27).Thus, while compliance with taking the large number of pills taken daily may be an issue, the infrequent treatment failures may make it a useful regimen in the future.After one week of PPI triple or quadruple therapy, without continuing with further antisecretory drugs, DU healing of about 89% to 98% was reported (Gisbert, 3A:27; Kung, 4A:06; Misiewicz, 4A:15).Similar healing rates were seen with 10-day PPI triple therapy (Wurzer, 4A:33) and with two-week dual therapy (Gisbert, 3A:27, Wurzer, 4A:33).This is an important finding because antisecretory agents are only needed for the duration of anti-H pylori treatment in uncomplicated DU patients, which can result in significant cost savings.
Although initial treatments are highly effective, some patients will have persistent H pylori infection after treatment.
In one small study of treatment failures given 10 days of omeprazole, clarithromycin and amoxicillin, each given bid for one week, H pylori eradication was seen in 100% (18 of 18 patients), and of these, 17 of 18 had metronidazole-resistant H pylori (Lerang, 4A:26).This suggests that the nonmetronidazole-containing regimen may be effective in treatment failures or in areas of high metronidazole resistance.It must be stressed that the currently approved dual therapy regimens of H pylori eradication in Canada -omeprazole plus either amoxicillin or clarithromycin -achieve unacceptably low rates of eradication (about 60% to 75%), in light of the high rates achieved using combinations such as PPI plus two.
However, despite eradication of H pylori and a marked reduction in ulcer recurrence, some DU patients (36% to 52%), for unknown reasons, suffer heartburn, recurrent dyspepsia or de novo development of esophagitis.In a prospective study, 14.2% (24 of 169) of peptic ulcer patients developed reflux esophagitis six months after H pylori eradication (Sacca, 1C:09).The prevalence of H pylori in patients with esophagitis is lower than in patients without esophagitis (Loffeld, 1C:01; Mihara, 1C:27; Lee 3A:35).Thus, it is speculated that eradication of H pylori may actually put the patient at risk for the development of esophagitis.After eradication of H pylori, the ability to suppress gastric acid is reduced for both PPIs (Labenz, 2C:06) and the H 2 -receptor antagonists (Tillenburg, 2C:07).Thus, one explanation for the new development of esophagitis is that there is significantly increased acid production that results as a consequence of H pylori eradication.H pylori-positive patients on long term therapy for gastroesophageal reflux disease may have exacerbations of fundic gastritis because of migration of H pylori to that region of the stomach (Misiewicz, EHPSG, 1996), or their chronic active gastritis may progress to chronic atrophic gastritis (10).However, a recent Food and Drug Administration (FDA) review of this hypothesis identified shortcomings of this paper (separate cohorts and different age groups), and the FDA decided that evidence does not support a causal relationship between long term treatment with PPI and development of atrophic gastritis.
What can the patient reasonably expect in the management of their PUD in 1997?Short, simple, safe, cheap, long lasting treatment (Bardhan, EHPSG, 1996).H pylori eradication accelerates ulcer healing and symptomatic relief with significantly fewer recurrences (11)  After failed H pylori eradication therapy, the prevalence of secondary or acquired resistance is 19% for metronidazole, 6% for clarithromycin and 0% for amoxicillin (Bouchard, 1A:05).These are in vitro figures, and it is unclear how this applies to in vivo resistance of H pylori, especially when treatment regimens with two antibiotics are used.Metronidazole resistance may reduce eradication success.With omeprazole plus clarithromycin plus metronidazole, eradication was 100% if H pylori was metronidazole-sensitive, and 82% if H pylori was metronidazole-resistant (P=0.002).Therefore, eradication is significantly lower, but still highly effective, if H pylori is metronidazole-resistant, suggesting that clarithromycin may help overcome metronidazole resistance (Peitz, 1A:03).Another study using lansoprazole, clarithromycin and metronidazole showed 95% eradication if H pylori was metronidazole-sensitive and 76% if H pylori was metronidazole-resistant (Misiewicz, EHPSC, 1996).However, a third study (Moayyedi, 1A:07) showed no difference whether H pylori was metronidazole-resistant (metronidazole-sensitive 90%, metronidazole resistance 93%).Thus, for PPI triple therapy with clarithromycin, the eradication rate seems to be only slightly affected by metronidazole resistance.For lansoprazole plus amoxicillin plus metronidazole, the eradication rate was lowered from 91% to 46%.Thus, amoxicillin-containing regimens are less effective than clarithromycin-containing PPI triple regimens in treating metronidazole resistant strains of H pylori.
Mechanisms for the development of metronidazole and clarithromycin resistance are the subject of intense study.A single genomic locus may be responsible for metronidazole resistance, and through recombination, a locus required for metronidazole toxicity may be inactivated (Goodwin, 1A:18).Activation of metronidazole by reduction of the nitrogen moiety of this antibiotic may be crucial for initiating activity against H pylori rather than via futile cycling (under microaerophilic conditions oxygen may convert reduced metronidazole back to the parent compound), and metronidazole resistance may occur via mechanisms that prevent reduction of metronidazole (Jorgensen, 1A:22).Increased intragastric pH caused by omeprazole leads to an increase in nonionized metroniadazole to which H pylori is exposed (Kareemi, 1A:28).Clarithromycin binds to bacterial ribosomes, and resistance may occur by a point mutation.Gastric mucosal flux of aminopenicillins (eg, amoxicillin) was greater than that of monobasic penicillins (eg, penicillin V) due to lower concentrations of protein binding within the tissue, thus providing a possible explanation of differences in efficacy (Goddard, 1A:3).Clarithromycin resistance commonly occurs because of a single base substitution (Stone, 1A:16, Ver Salovic, 1A:19; Debets-Ossenkapp, 1A:17; Occhialini, 1A:20) that causes decreased binding macrolides to ribosomes.

CONCLUSION
H pylori research encompasses many different areas of science and medicine, permitting the perfect milieu for extensive collaboration among disciplines.While considerable progress has been made, particularly at a molecular level, all the answers are not yet available.Meetings such as these are 'guilty' of providing more questions than answers, a fortunate plight for those interested in hearing about progress at next year's meeting.The explosion of information continues.

TABLE 1 Proportion of physicians treating symptomatic Helico- bacter pylori infections without a firm ulcer diagnosis
(9)ylori.The prevalence of in vitro metronidazole resistance of H pylori in Canada varies from 18% of H pylori isolates in Montreal(9)to 38% in Halifax (Best, 1A:39).The Hong Kong experience demonstrates that the prevalence of metronidazole resistance increases over time.Clarithromycin resistance varies from a high prevalence in France and Spain (10% to 13%), to intermediate values in Germany (3%) and to low rates in Canada (1%) (Best, 1A:39).Fortunately, to date, resistance of H pylori to amoxicillin has not been reported.The clinical relevance of this in vitro resistance to antimicrobials will be discussed in a later section.H pylori and NSAIDs: H pylori and the use of NSAIDs represent independent risk factors for the development of PUD, and their effects may be additive.Both H pylori and NSAIDs appear to be independent risk factors for bleeding ulcers (Kohl, 3A:22).The risk of bleeding from gastric ulcers is 63.7% for NSAIDs alone, 44% for H pylori and 71.8% when both NSAIDs and H pylori are present as risk factors (Ng, compared with maintenance therapy.Remission of DU and symptoms is prolonged, and rebleeding rates are markedly suppressed.With the proper selection of therapy, acceptable rates of H pylori eradication can be achieved, meeting today's reality and expectations.Problems with the assessment of metronidazole resitance include differences in the tests used, eg, E-test versus broth dilution versus agar dilution, and the respective cutoff used to define resistance (Megraud, EHSG 1996).Further problems arise because metronidazole-sensitive and -resistant H pylori strains can coexist in the same patient (Alarcon, 1A:36).This is a high rate of acquired resistance, which is another reason not to recommend dual therapy (PPI plus one antibiotic).