Articles
www.thelancet.com Published online November 16, 2012 http://dx.doi.org/10.1016/S0140-6736(12)61579-7 1
Articles
Published Online
November 16, 2012
http://dx.doi.org/10.1016/
S0140-6736(12)61579-7
See Online/Comment
http://dx.doi.org/10.1016/
S0140-6736(12)61849-2
Department of Internal
Medicine, National Taiwan
University Hospital, National
Taiwan University College of
Medicine, Taipei, Taiwan
(J-M Liou MD, C-C Chen MD,
M-J Chen MD, P-H Tseng MD,
H-P Wang MD, Prof J-T Lin MD,
Y-C Lee MD, Prof M-S Wu MD);
Graduate Institute of
Epidemiology and Preventive
Medicine, College of Public
Health, National Taiwan
University, Taipei, Taiwan
(J-M Liou, Y-C Lee); Department
of Internal Medicine, National
Taiwan University Hospital,
Yun-Lin County, National
Taiwan University College of
Medicine, Yun-Lin County,
Taiwan (C-C Chen, Y-J Fang MD,
J-Y Lee MD, S-J Hsu MD);
Department of Internal
Medicine, E-DA Hospital and
I-Shou University, Kaohsiung
County, Taiwan (C-Y Chang MD,
Y-C Hsu MD, C-H Tseng MD);
Department of Medicine,
National Yang-Ming University,
School of Medicine, and Taipei
Veterans General Hospital,
Taipei, Taiwan (J-C Luo MD);
Division of Gastroenterology,
Department of Internal
Medicine, Mackay Memorial
Hospital, Taipei, Taiwan
(W-H Chang MD); Institute of
Biomedical Informatics,
National Yang-Ming University,
School of Medicine, and Taipei
Veterans General Hospital,
Taipei, Taiwan
(Prof U-C Yang PhD); Department
of Pathology, National Taiwan
University Hospital, National
Taiwan University College of
Medicine, Taipei, Taiwan
(Prof C-T Shun MD); and Primary
Care Medicine, College of
Medicine, National Taiwan
University, Taipei, Taiwan
(Prof M-S Wu MD)
Sequential versus triple therapy for the fi rst-line treatment of
Helicobacter pylori: a multicentre, open-label, randomised trial
Jyh-Ming Liou, Chieh-Chang Chen, Mei-Jyh Chen, Chien-Chuan Chen, Chi-Yang Chang, Yu-Jen Fang, Ji–Yuh Lee, Shih-Jer Hsu, Jiing-Chyuan Luo,
Wen-Hsiung Chang, Yao-Chun Hsu, Cheng-Hao Tseng, Ping-Huei Tseng, Hsiu-Po Wang, Ueng-Cheng Yang, Chia-Tung Shun, Jaw-Town Lin,
Yi-Chia Lee, Ming-Shiang Wu, for the Taiwan Helicobacter Consortium
Summary
Background Whether sequential treatment can replace triple therapy as the standard treatment for Helicobacter pylori
infection is unknown. We compared the effi cacy of sequential treatment for 10 days and 14 days with triple therapy for
14 days in fi rst-line treatment.
Methods For this multicentre, open-label, randomised trial, we recruited patients (≥20 years of age) with H pylori
infection from six centres in Taiwan. Using a computer-generated randomisation sequence, we randomly allocated
patients (1:1:1; block sizes of six) to either sequential treatment (lansoprazole 30 mg and amoxicillin 1 g for the fi rst
7 days, followed by lansoprazole 30 mg, clarithromycin 500 mg, and metronidazole 500 mg for another 7 days; with
all drugs given twice daily) for either 10 days (S-10) or 14 days (S-14), of 14 days of triple therapy (T-14; lansoprazole
30 mg, amoxicillin 1 g, and clarithromycin 500 mg for 14 days; with all drugs given twice daily). Investigators were
masked to treatment allocation. Our primary outcome was the eradication rate in fi rst-line treatment by intention-to-
treat (ITT) and per-protocol (PP) analyses. This trial is registered with ClinicalTrials.gov, number NCT01042184.
Findings Between Dec 28, 2009, and Sept 24, 2011, we enrolled 900 patients: 300 to each group. The eradication rate was
90·7% (95% CI 87·4–94·0; 272 of 300 patients) in the S-14 group, 87·0% (83·2–90·8; 261 of 300 patients) in the S-10
group, and 82·3% (78·0–86·6; 247 of 300 patients) in the T-14 group. Treatment effi cacy was better in the S-14 group
than it was in the T-14 group in both the ITT analysis (number needed to treat of 12·0 [95% CI 7·2–34·5]; p=0·003) and
PP analyses (13·7 [8·3–40], p=0·003). We recorded no signifi cant diff erence in the occurrence of adverse eff ects or in
compliance between the three groups.
Interpretation Our fi ndings lend support to the use of sequential treatment as the standard fi rst-line treatment for
H pylori infection.
Funding National Taiwan University Hospital and National Science Council.
Introduction
Helicobacter pylori is an important cause of peptic ulcer
disease and gastric cancer, but eradication rates with
standard triple therapy have decreased to less than 80% in
many countries.1–5 Several strategies have been proposed
to increase the eradication rate, including the extending of
treatment duration to 14 days, the use of a four-drug
regimen (quadruple, sequential, and concomi tant treat-
ments), and the use of novel antibiotics such as
levofl oxacin.6–13 Sequential treatment, which consists of a
proton-pump inhibitor and amoxicillin for the fi rst 5 days,
followed by a proton-pump inhibitor plus clarith romycin
and metronidazole (or tinidazole) for another 5 days, has
been shown to be more eff ective than triple therapy for
7 days or 10 days.11–13 The effi cacy of sequential treatment
seemed to be aff ected less by clarithromycin resistance
than is triple therapy and has the potential to become the
standard fi rst-line treatment for H pylori infection.15,16
However, several concerns need to be resolved before
sequential treatment can replace triple therapy as the
standard treatment.15,16 First, most of the studies did
not do susceptibility tests and their results cannot be
generalised to other countries where the prevalence of
antibiotic resistance is diff erent. Second, few studies
compared sequential treatment with triple therapy for
14 days, which is recommended by US guidelines.4,15 Two
studies from Latin America and South Korea that
compared sequential treatment for 10 days with triple
therapy for 14 days, however, showed contra dictory
results.9,17 The reasons behind the contradictory results
were unknown because susceptibility tests were not
done.9,17,18 Third, whether extending the duration of
sequential treatment from 10 days to 14 days would be
more eff ective than triple therapy for 14 days is unknown.
Fourth, despite the fact that knowing how to re-treat
patients who fail sequential treatment is important, few
studies addressed this issue.19 Finally, how to choose the
best regimen on the basis of the prevalence of antibiotic
resistance in diff erent geographical areas is unknown.
To address these issues, we did a randomised controlled
trial to compare the effi cacy of sequential treatment for
10 days and 14 days with triple therapy for 14 days in fi rst-
line treatment. We extensively assessed factors that
might aff ect eradication rates, such as anti biotic resis-
tance, host CYP2C19 polymorphisms, and bacterial
virulence factors (CagA and VacA). We also assessed the
Articles
2 www.thelancet.com Published online November 16, 2012 http://dx.doi.org/10.1016/S0140-6736(12)61579-7
effi cacy of the modifi ed sequential treat ment containing
levofl oxacin in patients who failed sequential treatment
and triple therapy.20 We constructed a decision model to
estimate the effi cacies of three regimens in the sensitivity
analysis according to the prevalence of antibiotic resis-
tance, aiming to solve the heterogeneity of treatment
effi cacies seen in previous studies.
Methods
Study design and participants
For this multicentre, open-label, randomised trial, we
recruited participants from gastroenterology clinics in
six medical centres in Taiwan. Study staff recruited
potential participants and explained to them the
purpose of the trial and eligibility requirements for
enrolment. Patients were eligible for recruitment if
they were aged 20 years or older and had documented
H pylori infection. Patients with any one of the following
criteria were excluded from the study: previous eradi-
cation treatment for H pylori, his tory of gas trectomy,
contraindication or previous allergic reactions to the
study drugs, pregnant or lactating women, use of
antibiotics within the previous 4 weeks, and severe
concurrent diseases or malignancy.
Participants provided written informed consent before
enrolment. This trial was approved by the Institutional
Review Board of each hospital.
Randomisation and masking
Using a permuted block randomisation with a block size
of six, we randomly allocated eligible patients to receive
one of the following regimens (1:1:1): sequential treat-
ment for 14 days (S-14), sequential treatment for 10 days
(S-10), or triple therapy for 14 days (T-14; all given twice
daily). An independent research assistant at the National
Taiwan University Hospital generated the computerised
random number sequence. The sequence was concealed
in an opaque envelope until the inter vention was
assigned. Envelopes were kept at the National Taiwan
University Hospital. After the written informed consents
were obtained from eligible patients, the independent
research assistant telephoned study staff to give them
each patient’s treatment allo cation. All investigators were
masked to the random isation sequence. Patients who
Correspondence to:
Dr Yi-Chia Lee,
7 Chung-Shan S Road,
Taipei 10002, Taiwan
yichialee@ntu.edu.tw
and
Dr Ming-Shiang Wu,
7 Chung-Shan S Road, Taipei
10002 Taiwan
mingshiang@ntu.edu.tw
Figure 1: Trial profi le
ITT=intention-to-treat. PP=per-protocol. MS-14=modifi ed sequential treatment containing levofl oxacin. S-10=sequential treatment for 10 days. S-14=sequential
treatment for 14 days. T14=triple therapy for 14 days.
Fi
rs
t-
lin
e t
re
at
m
en
t
Se
co
nd
-li
ne
tr
ea
tm
en
t
2553 screened for eligibility
1653 excluded
1406 Helicobacter pylori negative
98 refused
149 not eligible
900 randomly allocated
300 assigned to S-10 (ITT population) 300 assigned to T-14 (ITT population)
10 lost to follow-up and took
<80% of drugs
2 lost to follow-up but took
≥80% of drugs
3 took <80% drugs
8 lost to follow-up and took
<80% of drugs
6 lost to follow-up but took
≥80% of drugs
7 took <80% drugs
285 completed follow-up (PP analysis) 279 completed follow-up (PP analysis)
39 S-14 failed first-line treatment 53 S-14 failed first-line treatment
12 lost to follow-up
2 refused further treatment
14 lost to follow-up
4 refused further treatment
25 given MS-14 35 given MS-14
6 lost to follow-up
300 assigned to S-14 (ITT population)
4 lost to follow-up and took
<80% of drugs
5 lost to follow-up but took
≥80% of drugs
6 took <80% drugs
285 completed follow-up (PP analysis)
28 S-14 failed first-line treatment
9 lost to follow-up
3 refused further treatment
16 given MS-14
1 lost to follow-up
15 completed follow-up 25 completed follow-up 29 completed follow-up
Articles
www.thelancet.com Published online November 16, 2012 http://dx.doi.org/10.1016/S0140-6736(12)61579-7 3
remained positive for H pylori after the initial treatment
were retreated with modifi ed sequential treatment for
14 days (MS-14).
Procedures
Study treatment regimens were all given twice a day
and contained the following: S-14 (lansoprazole 30 mg
and amoxicillin 1 g for the fi rst 7 days, followed by
lansoprazole 30 mg, clarithromycin 500 mg, and metro-
nidazole 500 mg for another 7 days), S-10 (lansoprazole
30 mg and amoxicillin 1 g for the fi rst 5 days, followed by
lansoprazole 30 mg, clarithromycin 500 mg, and metro-
nidazole 500 mg for another 5 days), T-14 (lansoprazole
30 mg, amoxicillin 1 g, and clarithromycin 500 mg for
14 days). MS-14 was also given twice a day and contained
lansoprazole 30 mg and amoxicillin 1 g for the fi rst
7 days, followed by lansoprazole 30 mg, metronidazole
500 mg, and levofl oxacin 250 mg for another 7 days.
Before enrolment, the status of H pylori infection in
patients with upper gastrointestinal symptoms was
deter mined by rapid urease test, histology, culture, and
serology. Patients with positive results in at least two of
these tests were eligible for enrolment. Asymptomatic
individuals who underwent cancer screening were also
eligible for enrolment if they had a positive ¹³C urea
breath test (¹³C-UBT). Post-treatment H pylori status was
assessed by ¹³C-UBT at least 6 weeks after completion of
treatment. All patients were asked to stop treatment with
proton-pump inhibitor and histamine-2 blocker for at
least 2 weeks before their ¹³C-UBT. The urea kit (which
contained 75 mg ¹³C-urea) was dissolved in water and
mixed with orange juice. Baseline and 30 min breath
samples were assayed with an infrared spectrometer
that produced computer-generated results in the Taipei
Institute of Pathology (Taipei City, Taiwan). Positive and
negative results were defi ned according to results of our
previous validation study21 as a Δ value of 4 units or
higher for positive and less than 2·5 units for negative.
Patients with inconclusive results received another
¹³C-UBT at least 2 weeks after the inconclusive test until
the results were conclusive.
The primary endpoint of the study was H pylori
eradication rates in fi rst-line treatment. The secondary
endpoints were the frequency of adverse events and
treatment compliance. The patients were informed of the
common side-eff ects from the study drugs before
treatment and were asked to record these symptoms
during treatment in provided diaries. A standardised
interview was also arranged at the end of treatment to
assess the adverse events and compliance. Compliance
was recorded as low when less than 80% of pills
were taken.
The biopsy specimens were cultured on plates con-
taining Brucella chocolate agar with 7% sheep blood
and incubated for 7 days under microaerobic conditions.
The minimum inhibitory concentrations were assessed
by agar dilution test in the central laboratory in National
Taiwan University Hospital. We defi ned resistance
breakpoints for every antibiotic (amoxicillin ≥0·5 mg/L,
S-14 group (N=300) S-10 group (N=300) T-14 group (N=300)
Men 165 (55%) 159 (53%) 167 (56%)
Mean age in years (SD) 53·7 (12·5) 52·8 (13·8) 53·3 (14·1)
Cigarette smoking 59 (20%) 68 (23%) 67 (22%)
Alcohol drinking (>40mL) 77 (26%) 71 (24%) 74 (25%)
Peptic ulcer disease 193 (64%) 209 (70%) 197 (66%)
Body-mass index of 27 or greater 63 (21%) 52 (17%) 66 (22%)
CYP2C19 (poor metaboliser) 43/286 (15%) 27/286 (9%) 38/287 (13%)
CagA-positive 147/177 (83%) 157/191 (82%) 144/183 (79%)
23S rRNA mutation 15/178 (8%) 15/192 (8%) 21/183 (11%)
GyrA mutation 16/172 (9%) 23/190 (12%) 17/179 (9%)
Clarithromycin resistance 16/177 (9%) 18/192 (9%) 21/183 (11%)
Metronidazole resistance 39/177 (22%) 46/192 (24%) 48/183 (26%)
Amoxicillin resistance 4/177 (2%) 4/192 (2%) 5/183 (3%)
Levofl oxacin resistance 17/177 (10%) 22/192 (11%) 22/183 (12%)
Helicobacter pylori test
Serology 293/298 (98%) 292/295 (99%) 291/298 (98%)
Rapid urease test 235/253 (93%) 237/255 (93%) 239/252 (95%)
Histology 252/263 (96%) 254/264 (96%) 248/262 (95%)
Culture 181/234 (77%) 195/235 (83%) 184/230 (80%)
Urea breath test 66/66 (100%) 65/65 (100%) 70/70 (100%)
Data are number of patients (%) or patients positive/patients tested (%), unless otherwise stated. S-10=sequential
treatment for 10 days. S-14=sequential treatment for 14 days. T-14=triple therapy for 14 days.
Table 1: Baseline characteristics
S-14 group S-10 group T-14 group p value
Eradication after fi rst-line treatment
ITT analysis (n/N [%; 95% CI]) 272/300 (90·7%; 87·4–94·0)* 261/300 (87·0%; 83·2–90·8) 247/300 (82·3%; 78·0–86·6)* 0·011
PP analysis (n/N [%; 95% CI]) 269/285 (94·4%; 91·7–97·1)†‡ 258/285 (90·5%; 87·1–93·9)‡ 243/279 (87·1%; 83·2–91·0)†‡ 0·012
Eradication after second-line treatment
ITT analysis (n/N [%; 95% CI]) 283/300 (94·3%; 91·7–97·0) 283/300 (94·3%; 91·7–97·0) 275/300 (91·7%; 88·6–94·8) 0·31
PP analysis (n/N [%; 95% CI]) 280/284 (98·6%; 96·4–99·4) 280/285 (98·2%; 96·0–99·2) 271/273 (99·3%; 97·4–100) 0·56
ITT=intention-to-treat. PP=per-protocol. S-10=sequential treatment for 10 days. S-14=sequential treatment for 14 days. T-14=triple therapy for 14 days. *p=0·003 for S-14
vs T-14. †p=0·003 for S-14 vs T-14. ‡In the six patients in the S-14 group, three patients in the S-10 group, and seven patients in the T-14 group who took less than 80% of
the study drugs, H pylori eradication was successfully achieved in three patients, three patients, and four patients, respectively.
Table 2: Helicobacter pylori eradication in fi rst-line and second-line treatments
Articles
4 www.thelancet.com Published online November 16, 2012 http://dx.doi.org/10.1016/S0140-6736(12)61579-7
clarithromycin ≥1 mg/L, levofl oxacin ≥1 mg/L, and
metronidazole ≥8 mg/L).8,22 The genotypes of gyrA and
23S rRNA were established by PCR followed by direct
sequencing with the automatic sequencer (ABI PRISM
3100 Genetic Analyzer; Applied Biosystems, Foster City,
CA, USA).22 The CagA gene and the VacA signal region
(signal region 1 and 2) and midregion (midregion 1 and 2)
mosaics were determined by PCR as described
previously.23 Genotyping for the CYP2C19 polymorphism
was done with the SEQUENOM MassARRAY System
(Sequenom, Sand Diego, CA, USA) in the Taiwan
National Genotyping Centre.24
Statistical analysis
On the basis of fi ndings from a previous meta-analysis,15
we hypothesised that there would be about a 10%
diff erence in the eradication rates between the three
study regimens. Findings from a previous study sug-
gested that the eradication rate with T-14 would be about
85%,25 so our original sample size estimation was for at
least 155 individuals in each group, giving a power of
80% and a 0·05 two-sided type 1 error, assuming 10%
loss to follow-up. After an interim report, we decided to
increase the sample size to a conservative estimate of
300 individuals in each group, which would give a power
of 90% in rejecting the null hypothesis and to adjust the
type 1 error for multiple comparisons with Bonferroni
correction. We made this decision to increase the
precision of our study and to ensure an overall nominal
signifi cance level of 0·05, assuming 15% loss to follow-
up. We did intention-to-treat (ITT) and per-protocol (PP)
analyses in the assess ment of the primary endpoint. All
randomised patients were included in the ITT analysis.
All individuals who violated the study protocol, such as
patients not taking at least 80% of treatment drugs, or
with unknown post-treatment H pylori status were
excluded from the PP analysis. Patients who did not
return for a follow-up ¹³C-UBT were recorded as
treatment failures. We compared categorical data using
the χ² test or Fisher’s exact test, as appropriate. We
compared continuous data with the Student’s t test and
gave results as mean (SD). For the primary endpoint,
we adjusted for multiple comparisons by setting a
Bonferroni-corrected α level of 0·01. For secondary
variables, we did exploratory ana lyses by setting an α level
of 0·05 without adjustment for multiple comparisons.
We used SPSS (version 12.0 for Microsift Windows) for
all statistical analyses.
To assess factors aff ecting eradication rates, we did a
multiple logistic regression analyses with the following
predictors of interest: clarithromycin resistance, metro-
nidazole resistance, amoxicillin resistance, age, sex,
peptic ulcer disease, and smoking. We did not include
patients with missing data in the regression analyses.
After identifi cation of factors associated with treat-
ment failure, we constructed a decision model (not
described in the protocol) to elucidate the heterogeneity
Figure 2: Effi cacies of fi rst-line and second-line Helicobacter pylori treatments
ITT=intention to treat. PP=per protocol. MS-14=modifi ed sequential treatment
containing levofl oxacin. S-10=sequential treatment for 10 days. S-14=sequential
treatment for 14 days. T14=triple therapy for 14 days. Absolute diff erences in the
effi cacy of fi rst-line and secon