Primary resistance to antibiotics and its clinical impact on the efficacy of Helicobacter pylori lansoprazole-based triple therapies


Dr W. C. Wang, Department of Life Science, National Tsing Hua University, Hsinchu, Taiwan 300. E-mail:



To evaluate Helicobacter pylori primary resistance and its clinical impact on the efficacy of two lansoprazole-based eradication triple therapies.


H. pylori-positive patients (n=228) were randomized to receive one of the 1-week regimens: lansoprazole 30 mg, clarithromycin 500 mg and amoxicillin 1 g (LAC), or lansoprazole 30 mg, clarithromycin 500 mg and metronidazole 500 mg (LMC), each given twice daily. H. pylori status was assessed by 13C-urea breath test and culture at diagnosis and by 13C-urea breath test 6 weeks after therapy. Antibiotic susceptibility was determined by E-test (n=98).


The eradication rates with per protocol/ intention-to-treat analyses were: LAC (n=95/114) 83%/69% and LMC (n=96/114) 85%/72%. Primary resistance was 11% for clarithromycin, 41% for metronidazole and 0% for amoxicillin. Eradication in metronidazole-susceptible/-resistant strains was 85%/82% in LAC and 83%/63% in LMC. Significantly lower cure rates were observed in clarithromycin-resistant patients treated with LAC (95% vs. 0%, P < 0.001) and LMC (86% vs. 0%, P < 0.001).


One-week LAC and LMC are similarly effective therapies. Clarithromycin resistance significantly affected H. pylori eradication in both regimens.


Helicobacter pylori is a Gram-negative bacterium that chronically colonizes the stomach mucosa of more than half of the human population.1 It is a major aetiologic agent for gastritis and peptic ulcer and is also implicated in the development of gastric cancer of infected patients.2, 3 Accumulated evidence from world-wide studies shows a significant reduction in the recurrence of peptic ulcers after eradication of H. pylori infection.4, 5 The current recommended eradication therapies comprise a proton pump inhibitor, clarithromycin and amoxicillin or metronidazole, twice daily for a 1-week period.6–8 By use of these treatments, > 90% cure rates are reported in some studies.9–12 However, an eradication failure ranging from 20% to 40% is found, mainly because of antimicrobial resistance.13–17

The reported prevalence of metronidazole resistance is common and varies from 10% to 50% among different geographical regions.18–20 Acquired resistance to clarithromycin is also described, but with a lower rate (usually from 5% to 10%).18, 19 Amoxicillin resistance is barely reported until recently.21–23 Because growing numbers of patients have been treated with these agents in the past few years, an increase in antibiotic resistance is also expected and is now being recognized as an essential problem in medicine.15–17 Close monitoring of H. pylori resistance in a given region is thus required for a better choice of an appropriate therapy for patients. The aims of this study were: (i) to evaluate the H. pylori eradication rate after treatment regimens consisting of lansoprazole and clarithromycin, combined with either amoxicillin or metronidazole, and (ii) to assess the impact of primary resistance of H. pylori to clarithromycin, metronidazole and amoxicillin on their efficacy in Taiwan.



A total of 228 patients (163 male and 65 female) undergoing upper digestive endoscopy for the evaluation of dyspeptic symptoms at Taichung Veterans General Hospital, Taichung, Taiwan, between April 1998 and April 2001 were enrolled in this study. The subjects ranged in age from 21 to 83 years (mean ± standard deviation, 52.8 ± 12.2). All these patients were H. pylori-positive on the basis of 13C-urea breath test and bacterial culture performed on biopsies. Exclusion criteria included prior antibiotic therapy for H. pylori and a history of allergy to any of the study medications.

Treatment regimen

Patients were randomized to receive one of the two regimens. Patients in the first group were treated with lansoprazole 30 mg b.d., clarithromycin 500 mg b.d. and amoxicillin 1 g b.d. (LAC) for 1 week. Patients in the second group were treated with lansoprazole 30 mg b.d., clarithromycin 500 mg b.d. and metronidazole 500 mg b.d. (LMC) for 1 week.

Assessment of H. pylori infection

The assessment of H. pylori status was performed by 13C-urea breath test and bacterial culture at diagnosis and by 13C-urea breath test at least 6 weeks after the end of therapy. At least two biopsy specimens were obtained from the stomach of each patient for H. pylori culture. Patients that had positive responses for 13C-urea breath test and had positive bacterial culture at the initial examination were considered as H. pylori-positive. Successful eradication was defined as a negative 13C-urea breath test in the follow-up visit after therapy.

Bacterial culture and antibiotic resistance

Isolation, identification and storage of H. pylori strains were performed essentially as described elsewhere.24, 25 Strains were recovered from frozen stocks on Brucella broth plates (Oxoid, UK) with 10% sheep blood, 1% IsoVitalex, 6 μg/mL of vancomycin, and 2 μg/mL of amphotericin B in a microaerophilic atmosphere for 2–4 days at 37 °C. The H. pylori strains were tested for metronidazole, clarithromycin and amoxicillin susceptibility using the E-test (AB Biodisk, Sweden). The minimal inhibitory concentration (MIC) was defined as the concentration on the E-test strip closest to the point of intersection with growth on the plate. Metronidazole resistance was defined as an MIC of > 8 μg/mL, clarithromycin resistance was defined as an MIC of > 2 μg/mL and amoxicillin resistance was defined as an MIC of  > 8 μg/mL.

Statistical analysis

The comparison of treatment efficacy was performed using both intention-to-treat and per protocol analyses. Intention-to-treat analysis included all patients who were H. pylori-positive before treatment and had received at least one dose of the drug. The per protocol analysis comprised only patients who had taken at least 80% of study medications. The second objective of the study was to evaluate the impact of primary antibiotic resistance on the outcome of triple therapies. Statistical analysis was done by χ2 test or by Fisher’s exact test. Differences at P < 0.05 were considered statistically significant. The 95% confidence intervals were calculated with respect to the cure rates.


Patient population

From April 1998 to April 2000, a total of 228 patients were included in the study: 114 were randomized to receive the LAC therapy and 114 were randomized to receive the LMC therapy. There were 37 patients (19 in LAC and 18 in LMC) excluded from the per protocol analysis: 11 had adverse effects resulting in discontinuation of medications (six in LAC and five in LMC) and 26 were lost for follow-up examination. The properties of patients are shown in Table 1. No statistical significance was found in endoscopic diagnosis, age or gender between these two groups.

Table 1.   Demographic and clinical characteristics of the two study groups Thumbnail image of

H. pylori eradication

By per protocol and intention-to-treat analyses, H. pylori infection was eradicated from 83% (79/95) and 69% (79/114), respectively, of patients in the LAC group and from 85% (82/96) and 72% (82/114), respectively, of patients in the LMC group. There was no significant difference in eradication rates between the two regimens by both the per protocol and intention-to-treat analyses. No significant difference was found in yearly eradication rates from 1998 to 2000 in either of the two regimens. Sex, smoking and drinking had no effect on the outcome of eradication.

Antimicrobial resistance and eradication

A big earthquake on 21 September 1999 near Taichung caused freezer failure, therefore only a small portion of strains collected from April 1998 to September 1999 were recovered. Results of antibiotic resistance were thus available in only 98 out of 228 (43%) patients. Resistance to metronidazole was detected in 40 patients (41%) and to clarithromycin in 11 patients (11%). No resistance to amoxicillin was found. Resistance to both metronidazole and to clarithromycin was found in six patients (6%). No difference in metronidazole and clarithromycin resistance rate was found between the LAC and LMC groups. Infection with metronidazole-resistant strains was found in 20 (61%) of 33 female patients but in 20 (31%) of 65 male patients (P < 0.005), possibly due to the use of metronidazole for the treatment of gynaecological infections in females. Age, year of endoscopy and clarithromycin resistance was not significantly associated with metronidazole resistance. Clarithromycin resistance was not associated with sex, age and year of endoscopy. No significant difference in metronidazole or clarithromycin resistant rate was seen between the LMC and LAC groups. Table 2 shows the intention- to-treat and per protocol analyses with respect to clarithromycin and metronidazole susceptibility in the LAC and the LMC groups. A high eradication rate was seen in both clarithromycin-susceptible and metronidazole-susceptible cases in LAC (96%) and LMC (94%). In the LAC group, > 80% was found for either metronidazole-sensitive or metronidazole-resistant cases (85% vs. 82%). A lower cure rate, nevertheless, was found in the metronidazole-resistant cases (63% vs. 83%, P=0.169) of the LMC group, although metronidazole resistance was not significantly associated with treatment failure. A comparison between LAC and LMC indicated that there was a higher eradication rate in LAC (82%) than in LMC (63%) for the metronidazole-resistant cases (P=0.201), but with no statistical significance. On the other hand, similar eradication was seen for the metronidazole-susceptible cases (85% in LAC and 83% in LMC). Eradication in primary clarithromycin-susceptible cases was 95% (37/39) as compared with 0% (0/5) in the clarithromycin-resistant cases in the LAC group (P < 0.001). Similar results were also found in LMC; there was 86% (25/29) cure rate in clarithromycin-susceptible cases but 0% (0/5) (P < 0.001) in resistant cases. Of 11 clarithromycin-resistant cases, 10 failed and one was with unknown result, suggesting that clarithromycin resistance was a high risk factor affecting the clinical success of clarithromycin-containing triple therapies.

Table 2. H. pylori eradication rates in relation to treatment groups and primary antibiotic resistance Thumbnail image of

Compliance and specific treatment-related side-effects

Table 3 shows the most commonly reported adverse events. Serious side-effects that led to a noncompliance to the treatment regimen were reported for six cases in the LAC regimen and for five in the LMC regimen. The rest of study subjects took at least 95% of their prescribed medications. No clinically significant changes in haematology or serum chemistry values or its vital signs were observed during the study. No significant difference was seen in adverse events between these regimens.

Table 3.   The most commonly reported adverse events Thumbnail image of


We report a single-centre study to eradicate H. pylori infection by two lansoprazole-based triple therapies. Most patients studied (203/228) had an active peptic ulcer associated with the infection. Approximately 80% H. pylori eradication efficacy on a per protocol basis was found for both LAC and LMC therapies, similar to that in many studies but lower than the high eradication rate (> 95%) seen in an earlier report.6, 14, 26 Given that all 191 patients included for the per protocol analysis took > 95% of medications, the limited success was not related to compliance. No clinical factor was linked to treatment failure. The main cause thus might be antimicrobial resistance.13–17 Despite no statistical significance, we did observe that the cure rate dropped from 83% to 63% for the metronidazole-containing therapy in the presence of primary metronidazole resistance, similar to that (21%, P < 0.0001) in a systematic review.17 An even more significant drop (from 81% to 45%) was seen for metronidazole resistant cases in a recent report from Hong Kong.27 A comparison between these studies shows that there is indeed a higher microbial resistance in Hong Kong than in Taiwan (43% vs. 41% for metronidazole and 19% vs. 11% for clarithromycin). Apart from this, our patients were treated with higher dose of clarithromycin (500 mg vs. 250 mg) and metronidazole (500 mg vs. 400 mg), which might increase the therapy efficacy.28 These data thus collectively suggest that metronidazole resistance affects the success of metronidazole-containing triple therapies and that high dose of antibiotics in triple therapies may be related to a better efficacy.

The most striking finding in this study is the strong negative impact on the efficacy of both regimens in cases of primary clarithromycin resistance. Nearly all patients with resistant strains (five in LAC and six in LMC) failed in H. pylori eradication; 10 failed and one was with unknown result. In contrast, for patients with clarithromycin-sensitive strains, 37 out of 39 in LAC and 25 out of 29 in LMC were cured. The prevalence of clarithromycin resistance is usually considered to be low (below 10%) in most countries. Resistant cases are thus scant and studies have been conducted with relatively small subject numbers in the past.15, 16 There is, however, > 10% clarithromycin resistance reported not only in Taiwan as seen here, but also recently among various geographical regions.27, 29–35 An alarming 19% resistance has been recently observed in Hong Kong and in France, respectively.27, 29 Consistent with this high rate, only 70% (59/84) and 68% (69/102), respectively, were cured using clarithromycin-containing triple therapies.27, 29 Moreover, 15 out of 17 resistant subjects in Tankovic et al. failed in the proton pump inhibitorndash;amoxicillin–clarithromycin triple therapies.29 Two recent studies with the same regimen for a 7–10-day treatment also have a similar trend; the cure rates for resistant cases are 20% (2/10) and 29% (4/14), respectively.36, 37 In respect of the proton pump inhibitor–metronidazole–clarithromycin regimen, there is also low cure rate in the presence of clarithromycin resistance; for instance, there is only 27% (3/11) in Wong et al. and 40% (2/5) in Katelaris et al.27, 38 A significant drop (a mean drop in efficacy of 56%) is also seen in a systematic review using similar triple therapies.17 Taken together, these results suggest that successful treatment of H. pylori infection using the proton pump inhibitor-based triple therapies is largely compromised by emerging resistance to clarithromycin. Our data, although limited by small numbers, also support the critical role of clarithromycin in triple therapies.

In conclusion, our results show a similar efficacy for LMC and LAC. Even though there was high resistance for metronidazole (41%) and for clarithromycin (11%), > 80% eradication could be achieved, possibly because of the high doses of antibiotics (500 mg of clarithromycin plus 500 mg of metronidazole or 1 g of amoxicillin) used in therapies. In vitro primary clarithromycin resistance appears to well predict the failure of triple proton pump inhibitor-based therapies to eradicate H. pylori infection. If confirmed in other studies, our results suggest that in areas with high primary clarithromycin resistance, drug susceptibility tests need to be performed prior to treatment of H. pylori infection.


S.-K. Poon and C.-S. Chang contributed equally to this work. We deeply thank C.-D. Lu for her excellent secretarial assistance. This work was supported by Veterans General Hospital–National Tsing Hua University–National Yang Ming University Joint Research Program (VTY89-P4-27), Medical Research Advancement Foundation in Memory of Dr C.-S. Tsou, Taiwan, Republic of China and by National Science Council (NSC 89-2320-B-007-007), Taiwan.