GASTROENTEROLOGY

i

na

npeg

Clinical Gastroenterology and Hepatology 2014;12:177186rifabutin-, and furazolidone-containing therapies. Finally,we provide recommendations for the efcient develop-ment (ie, identication and optimization) of new regimens,as well as how to prevent or minimize failures. The trial-and-error approach for identifying and testing regimensfrequently resulted in poor treatment success. Thedescribed approach allows outcome to be predicted andshould simplify treatment and drug development.

Keywords: Helicobacter pylori; Treatment; Quadruple Therapy;Review; Treatment Success; Concomitant Therapy; SequentialTherapy; Bismuth; Clarithromycin; Tetracycline; Metronida-zole; Amoxicillin; Proton Pump Inhibitors; Evidence Based.

Similar to other infectious diseases, the factorsresponsible for effective antimicrobial therapy of aHelicobacter pylori infection as well as those responsiblefor treatment failure are both straightforward and easily

practice. PP and MITT are the most useful for the devel-opment of new regimens, whereas for large multicenterrandomized comparisons most authorities prefer ITT.4

Considering that H pylori is a common infectiousdisease and 100% success is obtainable, outcome (eg, PPor ITT) also is scored in terms of efcacy (ie, as excellent,good, borderline acceptable, or unacceptable) becauseefcacy is the most important measure for patient care.For evaluating new therapies we score success (PP withsusceptible strains) as excellent (95% success), good(90 success), borderline acceptable (85%89% suc-cess), or unacceptable (

considered as the simultaneous administration of 2 tripletherapies plus a dual therapy (eg, a PPIamoxicillin-clar-ithromycin plus a PPIamoxicillin-metronidazole plus aPPI-amoxicillin dual therapy). Both triple regimens indi-vidually will reliably achieve 95% or greater success PPwith susceptible strains whereas the dual component willachieve approximately 50% success with clarithromycin-and metronidazole-resistant strains (ie, the strains areonly susceptible to amoxicillin). If resistance to clari-thromycin or metronidazole was not present, there wouldbe no indication to use the 4-drug regimen. However,when resistance results in unacceptably low treatmentsuccess rates when either is used empirically, the 4-drugcombination might be considered.

Unless there is an interaction between the antibiotics,the treatment population can be visualized as 4 separatesubgroups: one group with organisms susceptible to allantibiotics, one group with only clarithromycin-resistantorganisms, another group with only metronidazole-resistant organisms, and the nal group with organismsresistant to both (here, we assume an absence of resis-tance to amoxicillin). The subgroups without resistanceand those resistant to a single drug will each receive anoptimized triple therapy for their infection and most willbe cured, and the overall success thus will depend entirelyon the success of the PPI-amoxicillin therapy for those

178 Graham et al Clinical Gastroenterology and Hepatology Vol. 12, No. 2are the presence of organisms resistant to one or more ofthe antimicrobials used, poor compliance with therapy,or both. A number of studies have suggested a variety ofmiscellaneous factors that might be important includingage, presentation (eg, nonulcer dyspepsia vs duodenalulcer), and CagA status.57 However, these candidatestypically have been discovered in data-dredging studiesin which resistance was not assessed, and most of thestudies lacked biologic plausibility. Although some ofthese factors (eg, nonulcer dyspepsia vs duodenal ulcer)have proven to be surrogates for differences in theprevalence of resistant strains,8,9 none of the clinicalcorrelates other than resistance and compliance hasproven to be important in studies in which complianceand resistance have been assessed.

Therapy Choice

Similar to other infectious diseases, treatment resultsare best when reliably excellent regimens are used to treatpatients with organisms susceptible to the antimicrobialschosen. Pretreatment susceptibility testing, either byculture of the organism or indirectly by molecular testingof stools of infected patients or uorescent in-situ hy-bridization using parafn-embedded gastric biopsyspecimens, allows one to select a regimen tailored byantimicrobial susceptibility (ie, tailored therapy).3 How-ever, in many instances, one must choose therapyempirically and, in this instance, the best approach is touse regimens that have been proven to be reliably excel-lent locally.2 That choice should take advantage ofknowledge of resistance patterns obtained from local orregional antimicrobial surveillance programs and/orbased on local clinical experience with regard to whichregimens are effective locally. Finally, the history of thepatients prior antibiotic use and any prior therapies willhelp identify which antibiotics are likely to be successfuland those for which resistance is probable (Figure 1).

All other things being equal, data from any area or re-gion regarding the effects of resistance on outcome can beused reliably to predict outcome in any other area. Thus,strains with similar patterns of resistance in Italy, theUnited States, Iran, China, and so forth should be expectedto respond alike such that, if one knows the results withsusceptible and with resistant strains in one place, onereasonably can predict the outcome of therapy anywhere.

Using Available Data to PredictTreatment Success

An optimized regimen is dened as one that reliablyachieves 95% or greater cures in patients with susceptibleorganisms. Although the effectiveness of any regimen canbe undermined by antimicrobial resistance, the effect ofresistance is not random and the effect of any particularlevel of resistance can be estimated based on studies with

that combination elsewhere, for example, use of theoptimized regimen (14-day concomitant therapy, con-sisting of a proton pump inhibitor [PPI], clarithromycin,metronidazole, and amoxicillin, given twice a day for 14days).10 The regimen contains 4 drugs, but for the purposeof understanding the effects of resistance can best be

Figure 1. Recommended approach to treatment of H pyloriinfections. Rx, treatment.with dual clarithromycin-metronidazole resistance.

In this example, both triple therapies achieve 97%treatment success and the dual therapy achieves 50%success (Table 1). One can calculate that treatment suc-cess will remain at or above 90% until dual resistanceexceeds 15%. That calculation is based on the followingformula: (% success with all-susceptible strains) (pro-portion with all-susceptible infections) (% successwith clari-susceptible strains) (proportion with clari-susceptible infections) (% success with met-susceptible strains) (proportion with met-susceptibleinfections) (% success with dual resistant strains)(proportion with dual resistant strains) 90%. Becausethe success with organisms susceptible to all antibioticsand single-drug resistances is the same, the 2 tripletherapies can be combined to simplify the calculation (eg,where X the proportion with dual resistance, the for-mulas is 0.97 (1 X) 0.5 X w0.90, and thus X 14.9%). Table 1 lists the approximate success rates witha number of common therapies.

clarithromycin resistance would need to exceed 50% to

February 2014 Rational H pylori Therapy 179Table 1. Approximate Treatment Success PP WithSusceptible Strains (Western Results)

Therapy Days Success rate

Clarithromycin triple therapy 7 94%Clarithromycin triple therapy 14 97%Sequential therapy 10 94%Sequential therapy 14 97%Hybrid therapy 14 97%Fluoroquinolone triple therapy 7

180 Graham et al Clinical Gastroenterology and Hepatology Vol. 12, No. 214-day dual therapy yields approximately 50% success,and results greater than 50% are uncommon when usingthe doses and durations typically used with commontherapies, and success decreases as the duration de-creases. The actual results will depend in part on theeffectiveness of the PPI in increasing the intragastric pHto high levels (eg, pH 6). PPI effectiveness depends inpart on the PPI used, its dose and frequency of admin-

Table 2. H pylori Therapies Recommended for EmpiricTherapy in Western Countries

For general use14-day concomitant therapy14-day bismuth quadruple therapy14-day hybrid sequential-concomitant therapy

Areas where there is clarithromycin-metronidazole dualresistance

overestimates the prevalence of metronidazole resis-tance such that resistance always should be conrmed(eg, by agar dilution) for an accurate estimation ofeffectiveness in the presence of resistance.27,28

Therapies Generally RecommendedOnly for Geographic Areas With a LowPrevalence of Resistance

Clarithromycin-Containing Triple Therapy

Despite the Maastricht IV recommendations,clarithromycin-containing triple therapy (PPI, amoxicillin1 g, clarithromycin 500 mg, all twice a day for 14 days) isan obsolete therapy whether given for 7, 10, or 14days.29 The Achilles heel is clarithromycin resistance,with success depending on clarithromycin resistance and

Metronidazole resistance undermines 10-day sequentialtherapy when it reaches 20% and 14-day sequentialtherapy at approximately 30% (Table 4). The regimens

Table 3. Achilles Heel of Individual Common Regimens

Optimized therapies Achilles heela

14-day clarithromycin triple therapy Clar >15%14-day metronidazole triple therapy Met >15%14-day concomitant therapy Clari-Met dual >15%14-day sequential therapy Clari-Met dual >5%14-day hybrid therapy Clari-Met dual >9%14-day uoroquinolone triple therapy Levo >13%14-day uoroquinolone bismuth

quadruple therapyLevo >25%b

14-day bismuth quadruple therapy Tetracycline resistance(rare), compliance

14-day bismuth-furazolidone therapy Furazolidone resistance(rare), compliance

5-day uoroquinolone sequentialtherapy

Levo w20%b

Clar, clarithromycin; Levo, levooxacin; Met, metronidazole.aThe resistance level () at which treatment success decreases to less than

ay Sequential and 14-Day Triple Therapies

14-day sequential 14-day triple

tronidazoletance, 20%;rithromycintance, 18%

Metronidazoleresistance, 30%;clarithromycinresistance, 6%

Metronidazoleresistance, N/A;clarithromycinresistance, 15%

s % n Success % n Success % n

65.6 65 99% 65.8 65 97% 85 82.514.4 12.6 88% 4.2 3.7 50% 15 7.516.4 12.3 75% 28.2 21 N/A - 3.6 0.5 15% 1.8 0.3 N/A -

90.4 90 90

and metronidazole susceptible (Ms) to dual resistance (Cr-Mr).

February 2014 Rational H pylori Therapy 181the duration of therapy (Tables 1 and 3). With 14-daytherapy the combination remains effective until clari-thromycin resistance exceeds approximately 15%,whereas 7-day therapy is compromised by clari-thromycin resistance exceeding 5%. Currently, there arefew regions in the world where clarithromycin resistanceis less than 15% (ie, the 14-day regimen is still useful insuch areas as Northern Europe and Thailand). Clari-thromycin triple therapy has been superceded by 14-dayconcomitant therapy, whose only Achilles heel is dualclarithromycin-metronidazole resistance.

Metronidazole-Containing Triple Therapy

The Achilles heel of metronidazole-containing tripletherapy (PPI, amoxicillin 1 g, metronidazole/tinidazole500 mg, all twice a day for 14 days) is metronidazoleresistance, and metronidazole-containing triple therapynow rarely is used except as a tailored therapy or inJapan where the general use of metronidazole has been

Table 4. Effect of Metronidazole Resistance on 10- and 14-D

Treatment scenario patterna

10-day sequential

Metronidazoleresistance, >20%;clarithromycinresistance, 0

Meresisclaresis

Successb %c nd Succes

Cs-Ms 95% 80 76 99%Cr-Ms 80% 0 0 88%Cs-Mr 75% 20 15 75%Cr-Mr 10% 0 0 15%Overall success 91

N/A, not applicable.aResistance pattern in population ranging from clarithromycin susceptible (Cs)bPredicted treatment success for the pattern of resistance.cPercentage of the study population with that pattern of resistance.

dSuccess rate percentage of the resistant pattern group with successful therapy.strongly discouraged by the government because ofpossible genotoxicity.30,31 Overall success parallels theexperience with clarithromycin-containing triple therapyin relation to duration and the presence of resistance.

Sequential Therapy

Although 14-day sequential therapy (PPIamoxicillin1 g for 5 or 7 days followed by a PPIclarithromycin500 mgmetronidazole/tinidazole 500 mg all twice daily,for 5 or 7 days) provides better results than 10-daytherapy, both have the same Achilles heel (ie, dualresistance and metronidazole resistance)13 (Table 3).

90%.bThe number of subjects receiving these regimens is low, such that the esti-mate is only approximate.

90% PP and any level of clarithromycin resistance would

182 Graham et al Clinical Gastroenterology and Hepatology Vol. 12, No. 2than 5%, sequential therapy has had a poor showing inAsia and South and Central America13,32 (eg, in Taiwan,10-day sequential therapy achieved 78.9% successdespite no clarithromycin resistance).13

Fluoroquinolone-Containing Triple Therapy

Only 14-day therapy is successful with uo-roquinolone triple therapy (PPI twice daily, amoxicillin1 g twice daily, a uoroquinolone once a day such aslevooxacin, moxioxacin, or sitooxacin, for 14 days).However, success is restricted to areas with low uo-roquinolone resistance. Fluoroquinolone resistancecannot be overcome by increasing the dose or durationof triple therapy, which becomes ineffective whenresistance reaches 13%. Fluoroquinolone therapy is notrecommended for patients who have received any uo-roquinolone in the past or in areas where uo-roquinolone resistance exceeds 10%. Possibly betteruoroquinolone-containing regimens includeuoroquinolone-bismuth therapy and uoroquinoloneconcomitant therapy. Neither has been optimized ortested widely and generally they should be used astailored therapies (see later).

Fluoroquinolone Bismuth Therapy

Fluoroquinolone bismuth therapy (PPI twice daily,amoxicillin 1 g twice daily, bismuth twice daily, levo-oxacin 500 mg once daily for 14 days) is basically thecause it to decrease further. In contrast, despite 20%metronidazole resistance, success with 14-day sequentialtherapy remains greater than 90% until clarithromycinresistance exceeds 18%. There are instances when14-day triple therapy will be superior to sequentialtherapy because it is not affected by metronidazoleresistance and can withstand up to 15% clarithromycinresistance before decreasing to less than 90% success.The primary Achilles heel for sequential therapy ismetronidazole resistance (ie, the level of metronidazoleresistance determines the level of clarithromycin resis-tance required for success to decrease to

when given for 14 days. More recent studies have testedlower doses of amoxicillin and PPI (rifabutin 150 mgonce daily, amoxicillin 1 g twice daily, and esomeprazole40 mg twice daily for 12 days) with results as low as88.6%.38 Clearly, additional studies are needed to opti-mize the regimen in terms of dose and duration. Finally,a recent study from Western Australia evaluated thecombination of a PPI, bismuth, rifabutin, and ciproox-acin, and reported an eradication rate of 95.2% in sus-ceptible strains.39 The recent increase in uoroquinoloneresistance makes it unlikely that the combination willprove useful as other than a tailored regimen, but itbrings up the intriguing question regarding how muchimprovement, if any, would be obtained by the additionof bismuth to rifabutin triple therapy.40

variable is whether the infecting strain is susceptible. Inour experience, the same high success rates are obtainedwith the rst and the nth-line regimen, provided theorganisms are susceptible and good compliance is ob-tained. To the infection, all attempts with agents to whichit is susceptible are rst attempts. We do not knowwhether some patients may be more difcult to curethan others but, all things being equal, the effectivenessof rst-line agents, rst-line alternative agents, or evensalvage therapies is similar if the organisms are sus-ceptible. Repeated failures should prompt assessment ofcompliance and rare events such as the development ofamoxicillin resistance.

February 2014 Rational H pylori Therapy 183Second or Subsequent Treatments forTreatment Failures

Generally, clinicians should have 2 preferred rst-lineregimens known to be effective locally, with the choicebetween them based on the patients history of prior druguse and exposure (Figure 2). The regimen with thehighest predicted successful outcome always should beused rst.41 Treatment success always should beconrmed, generally using a noninvasive test for activeinfection such as the stool antigen or urea breath test.42

Conrmation of cure also provides the clinician with anearly warning of the development of increasing resis-tance in the community.

H pylori is naturally resistant to many antimicrobialsand rapidly has become resistant to others. The use ofagents to which the organism is resistant either naturallyor by acquired resistance has no effect on the outcome oftherapy with agents to which the strain is susceptible.Prior use of an antibiotic for another infection often re-sults in the H pylori becoming resistant (a bystandereffect), and clarithromycin and other macrolides, uo-roquinolones, and rifabutin should not be used again.Generally, amoxicillin and tetracycline can be re-usedbecause resistance rarely develops. The key outcome

Figure 2. Example ofchoices of clarithromycin-containing regimens foran individual patient basedon predicted resistanceto clarithomycin and

metronidazole.Compliance and Adherence

Poor compliance with a regimen and antimicrobialresistance are the primary reasons for failure of what isotherwise a reliably excellent regimen. Large multicenterclinical trials have shown that although side effectsrelated to the antibiotics used are common, in the ma-jority of trials the drop-out rates because of side effectsare low (eg, in the range of 5%). Although there isconsiderable literature regarding compliance withmedication use, treatment of H pylori has not been apopular area of such research. The fact that H pyloritherapy often involves multiple drugs and multipledosing intervals makes patient education extremelyimportant. Emphasizing the importance of taking all thedrugs, as generally performed in large multicenterstudies, repeatedly has shown that this is associated witha high degree of compliance despite the complexity ofsome regimens. When tested in a randomized trial, pa-tient counseling and follow-up evaluation have beenshown to improve the outcome and compliance ofH pylori therapy and is recommended.43 It is worthwhileto consider direct counseling regarding the regimen andthe need to be compliant as well as to give handoutsregarding the objectives and the details of the regimen.Although it is important to try to keep patients ontherapy despite side effects, it also is important to test

to conrm that, at least in that population, sequential4749

Helicobacter 2011;11(Suppl 1):136.

14. Furuta T, Graham DY. Pharmacologic aspects of eradication

184 Graham et al Clinical Gastroenterology and Hepatology Vol. 12, No. 2outcome variables are determined by resistance andcompliance. Figure 2 shows how suspected resistancemarkedly inuences the choice for an individual subject(eg, previous use of a macrolide or metronidazole wouldtherapy was superior to triple therapy. Importantly,the detrimental effect of clarithromycin resistance wasnoted, but the critical effect of metronidazole resistanceon outcome remained unrecognized.32 When sequentialtherapy was tested in Southern Italy and other pop-ulations with higher metronidazole resistance, it generallyfailed to achieve its prior success.32,50,51 The process tookapproximately 10 years, during which thousands of sub-jects were randomized to triple therapy, which repeatedlyhad been proven to provide unacceptable results1,52;many meta-analyses were performed but the severe lim-itations of the regimen remained unrecognized. It nallywas optimized in 2013.13 Additional details are availablein the Supplementary Material.

Summary Recommendations

Sufcient data from treatment trials in which theoutcomes in relation to susceptibility resistance havebeen provided to allow an evidence-based approach tochoosing antiH pylori therapies. We now can add to theadmonition to use what works locally by being able toreliably identify which regimens have the greatestchance of working. Figure 1 outlines a general schema,with therapy chosen based on pretreatment susceptibil-ity testing, or, if unavailable, based on a combination oflocal experience and information obtained from the pa-tients and the patients records. The earlier-describeddata and discussions generally focused on therapeuticchoices for a population. Whether one considers an in-dividual patient (ie, population n of 1) or a group, thefor cure even if patients were unable to complete theregimen because even a short course of therapy will curea proportion of patients.

Recommendations RegardingDeveloping New Regimens

The trial-and-error approach to the development ofH pylori therapies has proven to be inefcient and toprovide misleading results. The history of sequentialtherapy is a good example. Originally, 10-day sequentialtherapy was devised in response to failure of tripletherapy in Italy44,45 and it proved to be successful andsuperior to triple therapy.46 Unfortunately, it was pre-sumed to be optimal and no further attempts were madeto optimize it or to systematically examine its limitations.Rather, sequential therapy was repeated in the samepopulation to prove its superiority to triple therapy. Thesemultiple samplings then were combined in meta-analysesmake triple or sequential therapy poor choices) such thattherapy for Helicobacter pylori infection. Gastroenterol ClinNorth Am 2010;39:465480.

15. Essa AS, Kramer JR, Graham DY, et al. Meta-analysis: four-12. Megraud F, Coenen S, Versporten A, et al. Helicobacter pyloriresistance to antibiotics in Europe and its relationship to anti-biotic consumption. Gut 2013;62:3442.

13. Liou JM, Chen CC, Chen MJ, et al. Sequential versus tripletherapy for the rst-line treatment of Helicobacter pylori: amulticentre, open-label, randomised trial. Lancet 2013;381:205213.high prevalence of gastric cancer such as Nicaragua.what might be recommended for a population oftendiffers when individualized to a single patient.

Supplementary Material

Note: To access the supplementary material accom-panying this article, visit the online version of ClinicalGastroenterology and Hepatology at www.cghjournal.org,and at http://dx.doi.org/10.1016/j.cgh.2013.05.028.

References1. Graham DY. Helicobacter pylori eradication therapy research:

ethical issues and description of results. Clin GastroenterolHepatol 2010;8:10321036.

2. Graham DY, Fischbach LA. Empiric therapies for Helicobacterpylori infections. CMAJ 2011;183:E506E508.

3. Rimbara E, Fischbach LA, Graham DY. Optimal therapy forHelicobacter pylori infections. Nat Rev Gastroenterol Hepatol2011;8:7988.

4. Sheiner LB, Rubin DB. Intention-to-treat analysis and the goalsof clinical trials. Clin Pharmacol Ther 1995;57:615.

5. Broutet N, Marais A, Lamouliatte H, et al. cagA Status anderadication treatment outcome of anti-Helicobacter pylori tripletherapies in patients with nonulcer dyspepsia. J Clin Microbiol2001;39:13191322.

6. Scaccianoce G, Hassan C, Panarese A, et al. Helicobacter pylorieradication with either 7-day or 10-day triple therapies, and witha 10-day sequential regimen. Can J Gastroenterol 2006;20:113117.

7. Suzuki T, Matsuo K, Ito H, et al. Smoking increases the treat-ment failure for Helicobacter pylori eradication. Am J Med 2006;119:217224.

8. Taneike I, Nami A, OConnor A, et al. Analysis of drug resistanceand virulence-factor genotype of Irish Helicobacter pyloristrains: is there any relationship between resistance to metro-nidazole and cagA status? Aliment Pharmacol Ther 2009;30:784790.

9. Zullo A, Perna F, Hassan C, et al. Primary antibiotic resistance inHelicobacter pylori strains isolated in northern and central Italy.Aliment Pharmacol Ther 2007;25:14291434.

10. Molina-Infante J, Romano M, Fernandez-Bermejo M, et al.Optimized non-bismuth quadruple therapies cure most patientswith Helicobacter pylori infection in populations with high ratesof antibiotic resistance. Gastroenterology 2013;145:121128.

11. Graham DY, Gonzalez C, Palacios C, et al. Importance ofdetermining the pattern of H. pylori resistance in countries with adrug, three-antibiotic, non-bismuth-containing concomitant

February 2014 Rational H pylori Therapy 185therapy versus triple therapy for Helicobacter pylori eradication.Helicobacter 2009;14:109118.

16. Gisbert JP, Calvet X. Review article: non-bismuth quadruple(concomitant) therapy for eradication of Helicobacter pylori.Aliment Pharmacol Ther 2011;34:604617.

17. Kongchayanun C, Vilaichone RK, Pornthisarn B, et al. Pilotstudies to identify the optimum duration of concomitant Heli-cobacter pylori eradication therapy in Thailand. Helicobacter2012;17:282285.

18. Greenberg ER, Anderson GL, Morgan DR, et al. 14-day triple,5-day concomitant, and 10-day sequential therapies forHelicobacter pylori infection in seven Latin American sites: arandomised trial. Lancet 2011;378:507514.

19. Sardarian H, Fakheri H, Hosseini V, et al. Comparison of hybridand sequential therapies for Helicobacter pylori eradication inIran: a prospective randomized trial. Helicobacter 2013;18:129134.

20. Hsu PI, Wu DC, Wu JY, et al. Modied sequential Helicobacterpylori therapy: proton pump inhibitor and amoxicillin for 14 dayswith clarithromycin and metronidazole added as a quadruple(Hybrid) therapy for the nal 7 days. Helicobacter 2011;16:139145.

21. Salazar CO, Cardenas VM, Reddy RK, et al. Greater than 95%success with 14-day bismuth quadruple anti-Helicobacter pyloritherapy: a pilot study in US Hispanics. Helicobacter 2012;17:382389.

22. Liang X, Xu X, Zheng Q, et al. Efcacy of bismuth-containingquadruple therapies for clarithromycin-, metronidazole-, anduoroquinolone-resistant Helicobacter pylori infections in aprospective study. Clin Gastroenterol Hepatol 2013;11:802807.

23. Fischbach L, Evans EL. Meta-analysis: the effect of antibioticresistance status on the efcacy of triple and quadruple rst-linetherapies for Helicobacter pylori. Aliment Pharmacol Ther 2007;26:343357.

24. Dore MP, Farina V, Cuccu M, et al. Twice-a-day bismuth-containing quadruple therapy for Helicobacter pylori eradica-tion: a randomized trial of 10 and 14 days. Helicobacter 2011;16:295300.

25. Dore MP, Maragkoudakis E, Pironti A, et al. Twice-a-dayquadruple therapy for eradication of Helicobacter pylori in theelderly. Helicobacter 2006;11:5255.

26. Graham DY, Belson G, Abudayyeh S, et al. Twice daily (mid-dayand evening) quadruple therapy for H. pylori infection in theUnited States. Dig Liver Dis 2004;36:384387.

27. Osato MS, Reddy R, Reddy SG, et al. Comparison of the Etestand the NCCLS-approved agar dilution method to detectmetronidazole and clarithromycin resistant Helicobacter pylori.Int J Antimicrob Agents 2001;17:3944.

28. Osato MS, Graham DY. Etest for metronidazole susceptibility inH. pylori: use of the wrong standard may have led to the wrongconclusion. Am J Gastroenterol 2004;99:769.

29. Malfertheiner P, Megraud F, OMorain CA, et al. Management ofHelicobacter pylori infectionthe Maastricht IV/ FlorenceConsensus Report. Gut 2012;61:646664.

30. Hori K, Miwa H, Matsumoto T. Efcacy of 2-week, second-lineHelicobacter pylori eradication therapy using rabeprazole,amoxicillin, and metronidazole for the Japanese population.Helicobacter 2011;16:234240.

31. Graham DY, Rimbara E. Helicobacter pylori therapy in the west.Japanese J Helicobacter Res 2012;13:49.32. Gisbert JP, Calvet X, Oconnor A, et al. Sequential therapy forHelicobacter pylori eradication: a critical review. J Clin Gastro-enterol 2010;44:313325.

33. Graham DY, Shiotani A. Which therapy for Helicobacter pyloriinfection? Gastroenterology 2012;143:1012.

34. Federico A, Nardone G, Gravina AG, et al. Efcacy of 5-daylevooxacin-containing concomitant therapy in eradication ofHelicobacter pylori infection. Gastroenterology 2012;143:5561.

35. Graham DY, Lu H. Furazolidone in Helicobacter pylori therapy:misunderstood and often unfairly maligned drug told in a storyof French bread. Saudi J Gastroenterol 2012;18:12.

36. Gisbert JP, Calvet X. Review article: rifabutin in the treatment ofrefractory Helicobacter pylori infection. Aliment Pharmacol Ther2012;35:209221.

37. Borody TJ, Pang G, Wettstein AR, et al. Efcacy and safety ofrifabutin-containing rescue therapy for resistant Helicobacterpylori infection. Aliment Pharmacol Ther 2006;23:481488.

38. Fiorini G, Vakil N, Zullo A, et al. Culture-based selection therapy forpatientswho did not respond to previous treatment forHelicobacterpylori infection. Clin Gastroenterol Hepatol 2013;11:507511.

39. Tay CY, Windsor HM, Thirriot F, et al. Helicobacter pylori erad-ication in Western Australia using novel quadruple therapycombinations. Aliment Pharmacol Ther 2012;36:10761083.

40. Graham DY, Gisbert JP. Helicobacter pylori: tailored therapywith novel sequential quadruple therapies. Nat Rev Gastro-enterol Hepatol 2013;10:68.

41. Graham DY, Calvet X. Guide regarding choice of second-linetherapy to obtain a high cumulative cure rate. Helicobacter2012;17:243245.

42. Attumi TA, Graham DY. Follow-up testing after treatment ofHelicobacter pylori infections: cautions, caveats, and recom-mendations. Clin Gastroenterol Hepatol 2011;9:373375.

43. Al-Eidan FA, McElnay JC, Scott MG, et al. Management of Heli-cobacter pylori eradicationthe inuence of structured counsel-ling and follow-up. Br J Clin Pharmacol 2002;53:163171.

44. Buzas GM. Gastric tubes as vectors of Helicobacter pyloritransmission. Med Hypotheses 2010;75:4749.

45. Laheij RJ, Rossum LG, Jansen JB, et al. Evaluation of treatmentregimens to cure Helicobacter pylori infectiona meta-analysis.Aliment Pharmacol Ther 1999;13:857864.

46. Zullo A, Rinaldi V, Winn S, et al. A new highly effective short-term therapy schedule for Helicobacter pylori eradication.Aliment Pharmacol Ther 2000;14:715718.

47. Jafri NS, Hornung CA, Howden CW. Meta-analysis: sequentialtherapy appears superior to standard therapy for Helicobacterpylori infection in patients naive to treatment. Ann Intern Med2008;148:923931.

48. Gatta L, Vakil N, Leandro G, et al. Sequential therapy or tripletherapy for Helicobacter pylori infection: systematic review andmeta-analysis of randomized controlled trials in adults andchildren. Am J Gastroenterol 2009;104:30693079.

49. Zullo A, De FV, Hassan C, et al. The sequential therapy regimenfor Helicobacter pylori eradication: a pooled-data analysis. Gut2007;56:13531357.

50. Iovene MR, Romano M, Pilloni AP, et al. Prevalence of antimi-crobial resistance in eighty clinical isolates of Helicobacterpylori. Chemotherapy 1999;45:814.

51. Romano M, Cuomo A, Gravina AG, et al. Empirical levooxacin-containing versus clarithromycin-containing sequential therapyfor Helicobacter pylori eradication: a randomised trial. Gut 2010;59:14651470.

52. Pounder RE, Talley NJ. Letter: the ethics of using inferior regi-mens in H. pylori randomised trialseditors reply. AlimentPharmacol Ther 2012;35:858.

Reprint requestsAddress requests for reprints to: David Y. Graham, MD, Michael E. DeBakeyVeterans Affairs Medical Center, RM 3A-318B (111D), 2002 HolcombeBoulevard, Houston, Texas 77030. e-mail: dgraham@bcm.edu; fax: (713)790-1040.

Conicts of interestThis author discloses the following: David Graham is an unpaid consultant forNovartis in relation to vaccine development for the treatment or prevention ofHelicobacter pylori infection, a paid consultant for RedHill Biopharma regardingnovel H pylori therapies and for Otsuka Pharmaceuticals regarding diagnostic

testing; and has received royalties from Baylor College of Medicine patentscovering materials related to the 13C-urea breath test. The remaining authorsdisclose no conicts.

FundingSupported in part by the Ofce of Research and Development of the MedicalResearch Service Department of Veterans Affairs, a Public Health Service grant(DK56338) that funds the Texas Medical Center Digestive Diseases Center(DK067366 and CA116845 to D.Y.G.). The contents are solely the responsibilityof the authors and do not necessarily represent the ofcial views of theDepartment of Veterans Affairs or the National Institutes of Health. Also sup-ported by research grants from the National Science Council of Taiwan forresearch into the prevention of gastric cancer and the pathogenesis of Heli-cobacter pylori (Y.-C.L. and M.-S.W.), and by grants from the National Centerof Excellence for Clinical Trial and Research in the National Taiwan UniversityHospital for the foundation of the Taiwan Helicobacter Consortium (M.-S.W.).

186 Graham et al Clinical Gastroenterology and Hepatology Vol. 12, No. 2

Supplementary Material

Recommendations for New Regimens

The goal of the rst experiment is to provide deni-tive data (ie, allow a go or no go decision). One shouldnever need to ask whether the results would have beenbetter if we had increased the duration, dose, dosinginterval, and so forth. To prevent this, the initial studymust use full doses and the maximum reasonable dura-tion. Thus, the initial study should use full doses and theduration should be 14 days (ie, 14 days generally isconsidered the upper limit for duration and there arenumerous examples in which 7- or 10-day regimensproved inferior to 14-day therapy). For example, hun-dreds of patients received uoroquinolone, PPI, amoxi-cillin triple therapy before it was recognized that 95%cure rates could be achieved with 14-day therapy.1

The initial experiments should be a pilot study withpredened success (eg, at least 90% or 95% success PP)and include stopping rules to limit the exposure of sub-jects to unnecessary risks. The effect of resistance onoutcome must be able to be assessed. This can be per-formed as a tailored regimen and only those with knownpretreatment susceptible strains receive therapy, orbiospy samples can be obtained and stored for after-the-fact assessment. Once good to excellent regimens areidentied it is critical to know the effects of resistance,which can be tested directly in patients with knownresistant strains, or this information can be obtained aspart of empiric use trials. The lack of susceptibility datalargely have been responsible for much of the confusionabout the usefulness of both new and old therapies, andunits that are unable to assess susceptibility should beprecluded from performing exploratory clinical trials.

Optimizing Regimens

Optimization can include cost effectiveness, conve-nience, minimization of side effects, higher success in thepresence of resistance to one or more components, andso forth. Only successful regimens should undergo opti-mization and optimization must not compromiseoutcome (eg, the original excellent success rate ismaintained despite simplifying the regimen in terms ofdosing, dosing interval, or duration). As noted earlier,optimization includes a study of the effects of changes instudy parameters (eg, duration) on the results withresistant infections. For example, 7-day PPI bismuthquadruple therapy provides excellent results in theabsence of metronidazole resistance but produces un-acceptable low outcomes in the presence of metronida-zole resistance. Thus, in some countries, 7-day therapymay be optimal whereas in other countries 14-daytherapy is the minimum for an optimal result.

Avoiding Common Errors

Hopefully, the days in which trial and error is theprimary approach used to identify new antiH pyloriregimens are nearing their end and will be replaced byapproaches that are both efcient and logical.2 Much ofthe current confusion related to how to best treatH pylori has resulted from poorly designed studies andattempts to understand failures while ignoring the crit-ical role of resistance. For example, a recent largemulticenter study of H pylori therapy in a South andCentral America treatment trial did not take advantage ofpublished data showing that metronidazole resistancewas extremely high in the region and that prior studieswith clarithromycin triple therapy typically had pro-duced unacceptably low cure rates. The outcome of thetrial was predictable (ie, Western regimens using clari-thromycin and metronidazole would produce unaccept-ably low treatment success with the rank order favoringnonmetronidazole-containing regimens).3,4 Compara-tive trials also have remained a problem because thecomparisons chosen generally have not been of good orexcellent regimens, but primarily have been comparisonsof ineffective regimes or good therapies with knownineffective regimens, the latter of which are, by deni-tion, unethical.5,6 As a result, meta-analyses generallyhave not provided useful information regarding whichtreatments to use and where. Another aw of meta-analysis has been that comparisons of treatments thatcontain the same drugshave differed markedly in termsof details such as doses, durations, and adjuvants, and inrelation to the populations examined, which oftendiffered greatly in terms of the prevalence of resis-tance.7,8 The results of meta-analysis have declared un-acceptably low outcome regimens as equivalent or evenas one superior to another despite the fact that bothregimens produced unacceptably low outcomes.7,9,10

Meta-analyses of suboptimal regimens, those that failedto account for resistance patterns, or compared entirelydifferent doses and durations probably should neither beperformed nor published. We can and should do better.

References1. Miehlke S, Krasz S, Schneider-Brachert W, et al. Randomized

trial on 14 versus 7 days of esomeprazole, moxioxacin, andamoxicillin for second-line or rescue treatment of Helicobacterpylori infection. Helicobacter 2011;16:420426.

2. Graham DY. Efcient identication and evaluation of effectiveHelicobacter pylori therapies. Clin Gastroenterol Hepatol 2009;7:145148.

3. Greenberg ER, Anderson GL, Morgan DR, et al. 14-day triple,5-day concomitant, and 10-day sequential therapies forHelicobacter pylori infection in seven Latin American sites: arandomised trial. Lancet 2011;378:507514.

February 2014 Rational H pylori Therapy 186.e1

4. Graham DY, Trespalacios AA. Treatment of Helicobacter pyloriin Latin America. Lancet 2012;379:408409.

5. Graham DY. Helicobacter pylori eradication therapy research:ethical issues and description of results. Clin GastroenterolHepatol 2010;8:10321036.

6. Graham DY, Rimbara E. Understanding and appreciatingsequential therapy for Helicobacter pylori eradication. J ClinGastroenterol 2011;45:309313.

7. Luther J, Higgins PD, Schoenfeld PS, et al. Empiric quadruplevs. triple therapy for primary treatment of Helicobacter pyloriinfection: systematic review and meta-analysis of efcacy andtolerability. Am J Gastroenterol 2010;105:6573.

8. Gatta L, Vakil N, Leandro G, et al. Sequential therapy or tripletherapy for Helicobacter pylori infection: systematic review andmeta-analysis of randomized controlled trials in adults andchildren. Am J Gastroenterol 2009;104:30693079.

9. Li Y, Huang X, Yao L. e al. Advantages of moxioxacin andlevooxacin-based triple therapy for second-line treatments ofpersistent Helicobacter pylori infection: a meta analysis. WienKlin Wochenschr 2010;122:413422.

10. Fuccio L, Minardi ME, Zagari RM, et al. Meta-analysis: dura-tion of rst-line proton-pump inhibitor based triple therapy forHelicobacter pylori eradication. Ann Intern Med 2007;147:553562.

186.e2 Graham et al Clinical Gastroenterology and Hepatology Vol. 12, No. 2

Clarithromycin resistance (%)0 20 40 60 80

0

5

10

15

20

25

30

35

10%*20%*

40%*80%*

Dua

l res

ista

nce

(%)

*=Met resistance

Supplementary Figure 1. The effect of increasing metroni-dazole and clarithromycin resistance and the prevalence ofdual clarithromycin-metronidazole resistance. When dualresistance equals 15% (dotted line), then 14-day concomitanttherapy will cure less than 90% of patients. The proportion ofpatients with dual resistance is shown at different preva-lences of metronidazole resistance, with the proportion withdual resistance plotted against the prevalence of clari-thromycin resistance. As long as metronidazole resistancelevels are less than 40%, the regimen is very resistant totypical increases in clarithromycin resistance. However, athigh levels of resistance of either, only modest levels ofresistance of the other will result in dual resistance exceedingthe threshold level for treatment failure.

February 2014 Rational H pylori Therapy 186.e3

Rational Helicobacter pylori Therapy: Evidence-Based Medicine Rather Than Medicine-Based EvidenceTherapy ChoiceUsing Available Data to Predict Treatment SuccessResistance EffectsCurrent Recommended RegimensCaveatConcomitant TherapyHybrid TherapyBismuth Quadruple Therapy

Therapies Generally Recommended Only for Geographic Areas With a Low Prevalence of ResistanceClarithromycin-Containing Triple TherapyMetronidazole-Containing Triple TherapySequential TherapyFluoroquinolone-Containing Triple TherapyFluoroquinolone Bismuth TherapyFluoroquinolone Concomitant Therapy

Salvage Therapies: After at Least 2 Treatment Failures With Different RegimensFurazolidone Bismuth Quadruple TherapyRifabutin-Containing Regimens

Second or Subsequent Treatments for Treatment FailuresCompliance and AdherenceRecommendations Regarding Developing New RegimensSummary RecommendationsSupplementary MaterialReferencesSupplementary MaterialRecommendations for New RegimensOptimizing RegimensAvoiding Common Errors

References