Kabakambira et al.

BMC Gastroenterology (2018) 18:134

https://doi.org/10.1186/s12876-018-0863-2

RESEARCH ARTICLE Open Access

Efficacy of Helicobacter pylori eradication

regimens in Rwanda: a randomized
controlled trial
Jean Damascene Kabakambira1* , Celestin Hategeka2,3, Cameron Page4, Cyprien Ntirenganya5,
Vincent Dusabejambo1, Jules Ndoli5, Francois Ngabonziza1, DeVon Hale6, Claude Bayingana5 and Tim Walker5,7

Abstract
Background: Successful H. pylori treatment requires the knowledge of local antimicrobial resistance. Data on the
efficacy of H. pylori eradication regimens available in sub-Saharan Africa are scant, hence the optimal treatment is unknown.
Our goals were to determine the efficacy of available regimens in Rwanda as well as evaluate the effect of treatment on
health-related quality of life (HRQoL) in patients undergoing esophagogastroduodenoscopy.
Methods: This is a randomized controlled trial conducted from November 2015 to October 2016 at a tertiary
hospital in Rwanda. Enrollees were 299 patients (35% male, age 42 ± 16 years (mean ± SD)) who had a positive
modified rapid urease test on endoscopic biopsies. After a fecal antigen test (FAT) and HRQoL assessment by
the Short Form Nepean Dyspepsia Index (SF-NDI) questionnaire, patients were randomized 1:1:1:1 to either a
triple therapy combining omeprazole, amoxicillin and one of clarithromycin/ciprofloxacin/metronidazole or a
quadruple therapy combining omeprazole, amoxicillin, ciprofloxacin and doxycycline. All therapies were given
for a duration of 10 days. The outcome measures were the persistence of positive FAT (treatment failure) 4 to
6 weeks after treatment and change in HRQoL scores.
Results: The treatment success rate was 80% in the total population and 78% in patients with a history of prior
triple therapy. Significant improvement in HRQoL in the total group (HRQoL mean scores before and after treatment
respectively: 76 ± 11 and 32 ± 11, p < 0.001) and the group with functional dyspepsia (HRQoL mean scores before and
after treatment respectively: 73 ± 11 and 30 ± 9, P < 0.001) was observed across all treatment groups.
Using clarithromycin based triple therapy (standard of care) as a reference, the group treated with metronidazole had
worse HRQoL (p = 0.012) and had a trend towards worse treatment outcome (p = 0.086) compared to the ciprofloxacin
based combination therapies.
Conclusion: Clarithromycin and ciprofloxacin based combination therapies are effective and safe to use alternatively
for H. pylori eradication and improve HRQoL. Among the regimens studied, metronidazole based triple therapy is likely
to be clinically inferior.
Trial registration: The clinical trial was retrospectively registered (PACTR201804003257400) with the Pan African
Clinical Trial Registry database, on April 6th, 2018 in South Africa.
Keywords: H. pylori eradication, Dyspepsia, Clinical trial, Rwanda

* Correspondence: [email protected]
1
Kigali University Teaching Hospital (CHUK), Kigali, Rwanda
Full list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Kabakambira et al. BMC Gastroenterology (2018) 18:134 Page 2 of 9

Background Korea, Byoungrak et al. investigated antibiotic resistance

Helicobacter pylori is a known successful human patho- in isolates in two cohorts in 2009–2010 and 2011–2012.
gen living on the luminal surface of the gastric epithe- Resistance to metronidazole was found to be 45.1% and
lium, responsible for various gastro-intestinal 56.3% respectively in those two cohorts [10]. In Brazil,
pathologies. The prevalence of infection is greatest in Eisig et al. detected H. pylori resistance to metronidazole
countries of the developing world. The most recent and clarithromycin of 51% and 8% respectively [11]. In
Rwandan H. pylori prevalence study in 2012 found a Africa, resistance to metronidazole is common. In
prevalence of 75% in an endoscopy population in Cameroon, H. pylori resistance was found to be as high
Southern Rwanda [1]. The prevalence was the same as 93% for metronidazole, 85% for amoxicillin, 45% for
three decades earlier in an endoscopy population in clarithromycin and 44% for tetracycline in 2008 [12]. In
Kigali City, Rwanda [2]. For successful eradication pol- South Africa, marked resistance (96%) to metronidazole
icies in any country, there is need for accurate diagnostic was observed. However marked susceptibility to
tests and treatment tailored to local antibiotic resistance ciprofloxacin (100%), amoxicillin (98%), clarithromycin
patterns. (80%) and gentamicin (73%) was observed in the same
study [13].
H. pylori diagnosis No prior studies have assessed H. pylori antibiotic re-
There is no single gold standard diagnostic test for H. sistance in Rwanda. However, data suggest that more
pylori. The choice of a diagnostic test is influenced by than half of patients presenting for endoscopy with a
the pretest probability of infection, cost, availability, history of prior triple therapy remain positive for H.
population prevalence of infection and factors such as pylori infection, raising concern about potentially high
the previous use of proton pump inhibitors (PPIs) and resistance rates to current triple therapies used for H.
antibiotics that may alter test results. Several diagnostic pylori eradication [1]. According to the Rwandan
tests rely on testing of endoscopy biopsy samples, serum Internal Medicine Clinical Treatment Guidelines, the
or stools. The urea breath test uses a different technique first choice for H. pylori eradication regimen is a triple
based on the principle that H. pylori hydrolyses urea to therapy combining omeprazole 20 mg twice daily, clari-
produce carbon dioxide and ammonia but it is a nuclear thromycin 500 mg twice daily and amoxicillin 1 g twice
medical technique that requires the use of carbon 13 daily for a 10–14 day duration [14]. While the suscepti-
and 14 isotopes that are not currently available in bility of H. pylori to clarithromycin varies widely from
Rwanda. Major diagnostic challenges exist when it 55 to 94% across the world, the efficacy of the
comes to checking eradication after treatment in clarithromycin-based triple therapy in Rwanda is un-
resource-poor settings. The urea breath test which is the known. Furthermore, the cost of clarithromycin-based
best option to document eradication is not available in regimen is approximately US$18, making it less
most resource limited countries. The Maastricht V/Flor- affordable in a country with a GDP per capita of US$703
ence Consensus also suggests bacterial culture for anti- and an annual total health spending per capita of US$
biotic resistance testing in those who fail first line 55 [15]. The present study assessed the efficacy of other
therapy [3]. However, H. pylori has historically been dif- cheaper H. pylori eradication regimens, pragmatically
ficult to culture although techniques are improving. It is constructed using antibiotics cheaply available in Rwanda.
also known that H. pylori antibiotic sensitivity in vitro
may not always predict response in vivo [4]. No labora- Health related quality of life
tory in Rwanda cultures H. pylori reliably at present, and Functional dyspepsia is defined as the presence of one or
resistance testing is also not available. more of the following: epigastric pain or burning, post-
Fecal antigen assays have been reported in the litera- prandial fullness and early satiety without evidence of
ture to have sensitivity and specificity above 90% [4–6]. structural disease (including at upper endoscopy) that
can explain the symptoms [16]. Whether H. pylori im-
H. pylori treatment proves quality of life among patients with functional dys-
Several studies have been conducted to evaluate efficacy pepsia, remains controversial. A study by Lane et al. in
of available H. pylori infection combination treatments 2006 failed to detect an improvement in HRQoL after H.
[7–9]. The common practice is a combination of a PPI pylori eradication [17]. A large meta-analysis by Li-Jun
and two antibiotics (triple therapy) or the addition of revealed contradictory results across 25 randomized
bismuth salts to these three drug agents (quadruple ther- controlled trials and recommends individual assessment
apy). The choice of antibiotics should be region specific, for clinicians desiring to eradicate H. pylori in patients
based on local H. pylori resistance to those antibiotics. with functional dyspepsia [18]. The American College of
However, raising antimicrobial therapy resistance to H. Gastroenterology acknowledges the lack of sufficient
pylori poses great management challenge worldwide. In evidence to show the benefit of treating H. pylori in
Kabakambira et al. BMC Gastroenterology (2018) 18:134 Page 3 of 9

patients with functional dyspepsia but currently them. This nurse was not allowed to discuss this infor-
recommends testing for and treating H. pylori given dur- mation or treatment allocations with the treating clini-
able benefit documented in some patients in previous cians or study staff; neither was she allowed to complete
studies [19–21]. patient assessments at the second visit.
The current study examines the efficacy of ciprofloxa- At visit 2 scheduled at 4–6 weeks after treatment com-
cin and metronidazole based H. pylori eradication ther- pletion, patients were clinically evaluated by a study clin-
apies compared with clarithromycin based triple therapy, ician and completed the HRQoL questionnaire again.
and the observed impact on HRQoL, among patients Patients with an initially positive fecal antigen test (FAT)
referred for endoscopy. also provided a stool sample for a post-treatment FAT.

Methods Investigations
Patients H. pylori status was determined by urease activity on 4
The study enrolled patients attending the University (2 antral and 2 fundal) biopsies. MRU test was under-
Teaching Hospital of Butare (CHUB) for esophagogas- taken by exposing gastric antral/body biopsies to a solu-
troduodenoscopy (EGD). CHUB is one of the 4 tertiary tion of 1 ml of 10% urea in water to which a drop of 1%
level hospitals in Rwanda with approximately 380 beds, phenol red has been added. When H. pylori is present,
located in the Southern Province of Rwanda. Patients the bacterial urease catalyzes urea to ammonia and car-
are referred for EGD at CHUB by clinicians working in bon dioxide which can be detected by the typical red
inpatient and outpatient facilities of CHUB, as well as color change in the solution [22].
satellite district hospitals and private health facilities in A positive reaction, manifested by a color change
the town. Endoscopies are undertaken by two gastroen- within 3 h, was necessary for patients to be eligible for
terologists or by a trainee physician under the supervi- the study.
sion of the gastroenterologists. Patients were also required to provide stool samples
The study enrolled patients who were 21 years and for FAT. Patients who were unable to provide samples
older, had a positive modified rapid urease test on endo- the same day were instructed to return samples the fol-
scopic biopsies and were willing to come back for lowing morning before starting medication. FAT was
follow-up. Patients were excluded from the study if they: performed with a rapid antigen test (HEALGEN;-
had used PPIs or a histamine H2 receptor antagonist or ORIENT GENE;DS; catalog number GCHP-602).
antimicrobial therapy in the previous 4 weeks; were al-
lergic to any of the study drugs (omeprazole, amoxicillin, Treatment regimens
clarithromycin, ciprofloxacin, metronidazole and doxy- Enrollees were randomized to one of the four following
cycline); or had endoscopic or clinical evidence of gastric regimens:
malignancy. Female patients were not breastfeeding and
had a negative pregnancy test prior to randomization. Group 1: omeprazole 20 mg twice daily + amoxicillin
1 g twice daily + clarithromycin 500 mg twice daily for
Randomization and study process 10 days (CLARITHRO).
The study required two visits to CHUB, in Huye District Group 2: omeprazole 20 mg twice daily + amoxicillin
in the Southern Province of Rwanda. 1 g twice daily + ciprofloxacin 500 mg twice daily for
At visit 1, patients underwent endoscopy. Biopsies 10 days (CIPRO).
were tested for H. pylori infection by the MRU test. Group 3: omeprazole 20 mg twice daily + amoxicillin
Patients provided a stool sample for fecal antigen test 1 g twice daily + metronidazole 500 mg three times a
and completed a questionnaire about medical history as day for 10 days (METRO).
well as the HRQoL questionnaire. Randomization was Group 4: omeprazole 20 mg twice daily +amoxicillin
done by picking a folded piece of paper under the obser- 1 g twice daily + ciprofloxacin 500 mg twice daily+
vation of a study nurse, from a basket containing thor- doxycycline 100 mg twice daily for 10 days (CIPRO-Plus).
oughly mixed pieces of paper labeled with numbers
corresponding to one of the 4 arms of treatment, each in Health related quality of life (HRQoL)
equal quantity. During the first and second visits, the Short Form Ne-
Patients and the treating clinicians were blinded to pean Dyspepsia Index (SF-NDI) questionnaire was com-
treatment. However, given that a high number of pa- pleted to assess HRQoL before and after treatment.
tients were unable to read and comprehend drug dosage The SF-NDI questionnaire has been translated and
instructions in English, a different nurse not associated validated for use in a Kinyarwanda speaking population
with study analysis, opened envelopes containing medi- [23]. SF-NDI is a questionnaire with 10 items measured
cations together with the patients to explain how to take on six-point Likert scales. The instrument assesses five
Kabakambira et al. BMC Gastroenterology (2018) 18:134 Page 4 of 9

domains namely: tension/anxiety, interference with daily Results

activities, knowledge/control, eating/drinking and work/ In a period of 1 year, from November 2015 to October
study. Individual items in each sub-scale are aggregated 2016, a total of 866 patients underwent EGD. Of the 866
to obtain a score range from 0 (best HRQoL score) to patients, 308 had a negative modified rapid urease
100 (worst HRQoL score) as defined by the question- (MRU) test for H. pylori while 96 patients did not
naire developers [24].Questionnaires were available in undergo MRU testing.
English and Kinyarwanda. Adapted questionnaire was Therefore, 462 patients had a positive MRU test but
availed for illiterate patients to be completed with the 99 patients did not meet inclusion criteria while 134 pa-
help of a literate support person. tients declined to participate to the study. Thus, 229 pa-
tients (35% male, age 42 ± 16 (mean ± SD), range
21-81 years) were randomized (Fig.1).
H. pylori eradication
Patients with initially positive FAT had a second FAT
performed at visit 2. Patients were considered H. pylori Baseline characteristics
negative if the second FAT was negative. The patient characteristics of the overall study group and
each assigned treatment group are presented in Table 1.
Safety considerations Access to EGD was generally rapid (mean number of days
Patients were instructed to report any side effects between endoscopy request by treating physician and en-
through a phone hotline to the research group. Symp- doscopy day: mean 2.2 ± 3.1).The population had a female
toms were classified as mild, moderate or severe. Pa- preponderance at 65%. All patients carried a health insur-
tients with severe symptoms were instructed to cease ance: 88% possessed community health insurance (Mutuel
medication and were assessed by a gastroenterologist de Santé) while 12% had private health insurance. Previ-
who decided on a further treatment regimen. At visit 2, ous medication use was very common among the study
a follow-up with a gastroenterologist was arranged for population: 77% had used PPIs or histamine (H2) receptor
all patients who were still symptomatic. antagonists, 17% had taken triple therapy before and 14%
reported using antibiotics for other illnesses.
The most common presenting symptoms were epigas-
Study outcomes tric pain (96%) and vomiting (30%).
Primary outcome The most common endoscopic diagnoses were gastritis
(56%), duodenal ulcer (30%) and gastric ulcers (11%).
 Positive fecal antigen test 4 to 6 weeks after There was no lesion seen on endoscopy in 37 (16%) pa-
treatment (among those with an initially positive tients. These patients with functional dyspepsia had lower
fecal antigen test). baseline HRQoL score compared to the patients with le-
sions on endoscopy (73 ± 12 and 77 ± 11, p = 0.027).
Therefore, a post-hoc subgroup analysis was performed
Secondary outcome
after treatment with regards to HRQoL. H. pylori testing
at endoscopy was positive in 60% of patients who under-
 Health related quality of life score change by SF-
went the modified rapid urease test. Only 37% (85/229)
NDI questionnaire.
patients had a positive FAT. The mean HRQoL score was
76 ± 11 and there was no sex difference in all subdomains,
Statistical analyses thus men and women were analyzed together after treat-
An intention-to-treat analysis was planned and con- ment. The previous use of PPIs or H2 blockers was nei-
ducted. Data are presented as mean ± standard deviation ther associated with negativity of the FAT (OR = 1.1, p =
(SD) for normally distributed continuous variables and 0.847 CI [0.5,2.2]) nor difference in HRQoL among users
frequencies (percentage) for categorical variables. To versus non-users (76 ± 11 and 77 ± 12, p = 0.354). Simi-
compare the efficacy of each treatment arm vs standard larly, a previous exposure to triple therapy was neither as-
of care (i.e., CLARITHRO), we performed independent sociated with negativity of the FAT (OR = 1.1, p = 0.738 CI
samples t-test for continuous outcome (i.e., mean change [0.5, 2.7]) nor difference in HRQoL among users versus
in HRQoL score) and Fisher exact test for categorical non-users (75 ± 12 and 77 ± 11, p = 0.549).
outcomes. A paired t-test was used to compare pre- and Overall, there was no difference in baseline character-
post-treatment HRQoL scores. Statistical analyses were istics with regards to prior exposure to triple therapy ex-
two-tailed and p values of < 0.05 were considered to cept for the presence of duodenal ulcers that were more
show statistical significance. Analyses were performed common among the non-exposed than in the exposed
with STATA (v 15, College Station, Texas). group (70% vs 30%, p = 0.025). Therefore, a subgroup
Kabakambira et al. BMC Gastroenterology (2018) 18:134 Page 5 of 9

Fig. 1 Flow chart for patient selection and randomization

analysis was undertaken with regards to prior exposure Compared to the treatment success in the overall group,
to triple therapy status. there was a trend towards higher treatment failure in the
METRO group although this did not reach statistical sig-
Efficacy of H. pylori eradication regimens nificance (36%, p = 0.086) (Table 2). Analysis stratified to
Treatment failure the group without prior exposure to triple therapy
Of the total 85 patients who had an initially positive showed a similar trend in treatment failure with the
FAT, 20% (17/85) had a positive test after treatment. highest proportion in the METRO group (39%, p =
Table 1 Baseline population characteristics according to arm of treatment
Total (n = 229) CLARITHRO CIPRO METRO CIPRO-Plus
Characteristics n = 6127% n = 5624% n = 5725% n = 5524%
Demographic characteristics Male (%) 35 43 43 35 35
Age, years mean ± SD 42 ± 16 40 ± 15 44 ± 16 41 ± 16 41 ± 17
Married (%) 57 52 68 60 49
CommunityHealth Insurance (%) 88 85 88 93 87
Access to endoscopy (days) 2.2 ± 3.1 1.9 ± 0.6 2.8 ± 6 2.3 ± 1.7 1.9 ± 0.6
Medical history PPI or H2 blocker before (%) 77 74 79 79 78
Triple therapy before (%) 17 13 23 16 18
Antibiotics before (%) 14 13 13 13 18
Symptoms Epigastric pain (%) 96 97 95 95 96
Vomiting (%) 30 27 38 19 35
Hematemesis (%) 8 10 9 4 9
Melena (%) 2 2 4 2 2
Endoscopy finding Normal endoscopy (%) 16 18 14 21 11
Gastritis (%) 56 41 55 61 67
Gastric ulcer (%) 11 10 14 14 7
Duodenal ulcer (%) 30 41 36 25 16
Initially positive FAT (%) 37 39 39 39 31
Baseline HRQoLtotal group(mean ± SD score) 76 ± 11 77 ± 13 76 ± 12 76 ± 8 78 ± 12
Baseline HRQoLin functional dyspepsia (n = 37) (mean ± SD score) 73 ± 12 71 ± 9 70 ± 17 74 ± 10 76 ± 14
Baseline characteristics of the intention to treat population were not significantly different, except for gastritis (CIPRO Plus and METRO were significantly different
from CLARITHRO, with p = 0.005 and p = 0.02, respectively) and for duodenal ulcer (CIPRO Plus was significantly different from CLARITHRO, p =
0.004).Abbreviations: SD Standard deviation, FAT Fecal Antigen Test, PP Proton Pump Inhibitors
Kabakambira et al. BMC Gastroenterology (2018) 18:134 Page 6 of 9

Table 2 Study outcomes

Outcome Total CLARITHRO CIPRO METRO CIPRO-Plus p-
N = 229 n = 6127% n = 5624% n = 5725% n = 5524% value1
Treatment failure (Positive FAT) (%) Total group (n = 85) 20 13 18 36 12 0.191
No prior exposure to triple therapy (n = 22 14 21 39 14 0.265
73)
HRQoLin total group (mean ± SD) Score after treatment 32 ± 11 31 ± 10 31 ± 10 36 ± 12** 31 ± 9 0.023
Mean difference 44 ± 14 46 ± 15 44 ± 15 40 ± 13* 47 ± 13 0.032
HRQoL in functional dyspepsia (n = 37) Score after treatment 30 ± 9 33 ± 13 28 ± 7 32 ± 6 24 ± 5 0.229
(mean ± SD)
Mean difference 42 ± 15 38 ± 17 42 ± 17 42 ± 10 51 ± 16 0.392
Persistence of symptoms (%) Total group (n = 229) 22 18 23 26 20 0.715
No prior exposure to triple therapy (n = 23 19 26 29 18 0.492
189)
P-value by multiple regression for comparison to the reference group (CLARITHRO), *P ≤ 0.05, **P ≤ 0.01
*
1
Student t-test or chi square test as appropriate

0.140).The sample size of the group with prior exposure revealed that improvement in HRQoL score was not dif-
to triple therapy was too small to establish a statistically ferent in all the four arms of treatment compared to the
sound comparison. standard of care (Table 2).
An assessment of clinical evolution after treatment re-
Health-related quality of life vealed a persistence of symptoms in 22% of the total
There was a dramatic change in HRQoL scores from 76 population. The METRO group had the highest rate of
± 11 to 32 ± 11 after treatment (p = 0.032) in the total persistence of symptoms (26%) followed by CIPRO
group. A paired t-test showed a significant improvement (23%), CIPRO-Plus (20%) and CLARITHRO (18%) but
in HRQoL across all the four arms of treatment (p < no regimen was statistically different from the standard
0.001) but a group comparison to the standard of care of care.
showed a significant difference only with the METRO Patients with persistence of symptoms after treatment
group (Fig. 2). Infact, the METRO group maintained a had a higher HRQoL than patients who experienced
higher post-treatment score than the standard of care symptom resolution (35 ± 11 vs 31 ± 10, p = 0.018) sug-
(36 ± 12 vs 31 ± 10, p = 0.008) and registered a lower gesting that persistence of symptoms was associated
mean score change than the standard of care (40 ± 13 vs with worse HRQoL. Stratified analysis within the group
46 ± 15, p = 0.012). In the group with functional dyspep- of patients with functional dyspepsia (n = 37) showed a
sia, a paired t-test showed significant improvements significant change in HRQoL scores overall (mean differ-
across all the four arms of treatment. Further analysis ence score: 42 ± 15, p < 0.001) but group comparison to

Fig. 2 Health-Related Quality of Life scores before (blue) and 4 to 6 weeks after (red) treatment. Scores were measured by the use of the SF-NDI
questionnaire on a range from 0 (best HRQoL) to 100 (worst HRQoL). Arms of treatment: 1: CLARITHRO, 2: CIPRO, 3: METRO, 4: CIPRO-
Plus. *p-value by multiple regression for comparison to the reference (REF) group (CLARITHRO), *p ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001
Kabakambira et al. BMC Gastroenterology (2018) 18:134 Page 7 of 9

the standard of care revealed no difference in any arm of On the other hand, ciprofloxacin based triple therapy
treatment. The greatest improvement in HRQoL was ob- (omeprazole + amoxicillin+ ciprofloxacin) and quadru-
served in “interference with daily activities” (p = 0.019) ple therapy (omeprazole + amoxicillin+ ciprofloxacin+
and “eating/drinking” (p = 0.050) subdomains. doxycycline) were not inferior to clarithromycin based
triple therapy and presented a very good safety profile.
This finding aligns nicely with studies conducted in
Drug safety and specific treatment related side effects
Nigeria and South Africa which didn’t detect any resist-
A total of 34% (79/229) patients reported treatment re-
ance to ciprofloxacin, suggesting that a fluoroquinolone
lated adverse effects. The adverse effects were reported
based regimen may be of utility in Africa [12, 27]. Pend-
to the investigator and were classified as mild, moderate
ing more rigorous diagnostic tests for eradication, this
or severe. The most commonly reported side effects
finding offers hope that the combination of omeprazole,
were taste perversion (27%), nausea (18%), dizziness
amoxicillin, ciprofloxacin and doxycycline could be used
(14%) and vomiting (9%). Some patients had more than
as a salvage therapy, particularly since bismuth combina-
one adverse effect. Only two patients in the METRO
tions are still unaffordable in Rwanda.
group had symptoms severe enough to stop medication.
This study shows that H. pylori was positive in 60% of
These two patients were called back to see a gastroenter-
the eligible population. This prevalence is lower than
ologist for further management but they were analyzed
other prevalences reported in Africa possibly due to the
in the METRO group. Overall, 99% of patients com-
fact that a portion of our population had previously been
pleted treatment and no patient was lost to follow-up.
treated for H. pylori [28].
Although reported to be of high accuracy for initial
Discussion and post-treatment diagnosis, FAT was only able to de-
This study explored the efficacy of pragmatic H. pylori tect 37% of patients with H. pylori.
eradication regimens available in Rwanda. Our results The diagnostic performance of FAT shows large varia-
indicate that treatment success rate was 80% in the total tions across the world. Studies assessing the diagnostic
group and 78% in the group without prior exposure to accuracy of FAT have concluded to sensitivities and
triple therapy. Significant improvement in HRQoL, specificities above 90% in Europe and Taiwan [4–6]. A
expressed by decrease in SF-NDI scores, from baseline study conducted in Uganda, a neighboring country with
was observed across all the 4 arms of treatment. Our Rwanda, found a contrasting sensitivity and specificity of
study results align with eradication success rates found 56% and 74% respectively [29]. In Nigeria, Olufemi et al.
in other studies around the world. The Maastricht IV/ reported H. pylori prevalence of 68.7% using a serology
Florence Consensus Report reported that the widely test but it was only 20.2% using FAT [30]. The poor per-
used triple therapy regimen cured 70% of patients [25]. formance of FAT in this study as well as the two studies
Although there was no statistically significant differ- in Uganda and Nigeria raise concern about the utility of
ence between any of the arms of treatment and the FAT as a diagnostic test in Rwanda and in Africa in gen-
standard of care, the metronidazole (METRO) based eral. Evidence shows that prior exposure to PPIs inter-
triple therapy showed some signals to suggest metro- feres with diagnostic accuracy of FAT but we had
nidazole based triple therapy might be inferior to other attempted to control this problem by excluding patients
regimens. Infact, metronidazole based triple therapy reg- who used PPIs or histamine receptor antagonists in the
istered the highest failure rate (36%) followed by CIPRO past 4 weeks [31]. The explanation of the diversity in
(18%), CLARITHRO (13%) and CIPRO-Plus (12%). FAT accuracy as a test which uses antigens is most likely
Similarly, HRQoL was improved in all the treatment to be linked with the diversity of genome and virulence
groups but improvement was much less in the metro- of H. pylori strains across the world [32]. Emerging data
nidazole based triple therapy than in the standard of from studies predominantly conducted in Asian popula-
care. Although the metronidazole based triple therapy tions unequivocally show large geographical variations in
was inferior to CLARITHRO group in improving the distribution of H. pylori strains [33, 34]. Although
HRQoL, the difference in treatment failure did not reach extensive work has been done to elucidate how genetic
the level of statistical significance, probably due to lower diversity is related to human cancer, little is known
than expected failure rates overall, and the concomitant about the effect of genetic diversity on the performance
reduction in study power. No susceptibility study has of stool antigen test for the diagnosis of H. pylori [35].
ever been conducted particularly in Rwanda but the re- Our study found significant change in HRQoL scores
sistance of H. pylori to metronidazole is notoriously high from baseline across all the 4 arms of treatment. The
(90–100%) in Africa, which may explain the modest per- findings were even true in the sub-group of patients who
formance of metronidazole based triple therapy that was had normal endoscopy before treatment (functional
found [26–28]. dyspepsia).
Kabakambira et al. BMC Gastroenterology (2018) 18:134 Page 8 of 9

This finding adds to the currently accumulating litera- urea breath test to better characterize H. pylori eradica-
ture in favor of improvement of HRQoL by treating H. tion in sub-Saharan Africa. Further work needs to be
pylori in patients with functional dyspepsia. However, we done examining other alternatives, including high dose
had a small number of patients with functional dyspep- dual therapy, if treatment recommendations are to be
sia and we are not statistically powered enough to draw optimized.
a firm conclusion. Further studies with larger sample
sizes are still needed to confirm this finding. Conclusion
Lastly, although all study participants had health insur- Given the balance of cost, efficacy and safety profile doc-
ance, most (88%) had community health insurance, umented in this study; clinicians should feel confident to
which is the cheapest and usually the main option ac- use clarithromycin and ciprofloxacin based combination
cessible to Rwandans with limited financial resources. therapies for H. pylori eradication in Rwanda. Our find-
Community health insurance only gives access to health ings suggest that metronidazole based triple therapy is
care and medicine available in public health care facil- likely to be clinically inferior, and make it the worst
ities. Due to the high cost and low availability in many choice among the four regimens we studied.
public facilities, clarithromycin is the least affordable
Abbreviations
medication for the great majority of patients. It is not CHUB: Butare University Teaching Hospital; CHUK: Kigali University Teaching
available at all in many rural health care facilities. Be- Hospital; CIPRO: Omeprazole 20 mg twice daily + amoxicillin 1 g twice daily
cause of this, clinicians struggle to select appropriate ini- + ciprofloxacin 500 mg twice daily for 10 days; CIPRO-Plus: Omeprazole
20 mg twice daily + amoxicillin 1 g twice daily+ ciprofloxacin 500 mg twice
tial and salvage regimens for H. pylori eradication. The daily+ doxycycline 100 mg twice daily for 10 days; CLARITHRO: Omeprazole
treatment success rate trend and safety profile from this 20 mg twice daily + amoxicillin 1 g twice daily + clarithromycin 500 mg
study make the ciprofloxacin based combination therap- twice daily for 10 days; EGD: Esophagogastroduodenoscopy; FAT: Fecal
Antigen Test; GDP: Gross domestic product; HRQoL: Health-related quality of
ies a strong and cost-effective alternative to clarithromy- life; METRO: Omeprazole 20 mg twice daily + amoxicillin 1 g twice daily +
cin based therapy. metronidazole 500 mg three times a day for 10 days; MRU: Modified rapid
urease; PPI: Proton Pump Inhibitors; SD: Standard deviation; SF-NDI: Short
Form Nepean Dyspepsia Index
Strengths and limitations
This is the first ever H. pylori eradication clinical trial Acknowledgements
conducted in Rwanda. By using a multi-arm trial, we We thank the clinical staff in the Gastroenterology Unit at CHUB: Dr. Nicaise
Nsabimana, Steven Ndayisaba, Leonie Niragire, Fiacre Ngarambe and Samuel
were able to study efficacy of various H. pylori eradica- Bizimana. Dr. Aimee Nyiramahirwe, Fanny Giraneza and Gladys Uwera have
tion regimens simultaneously using the same control greatly contributed in data collection.
group, thus saving time and resources to provide locally
Funding
appropriate evidence to guide clinical practice. Findings The study was funded by Butare University Teaching Hospital (CHUB),
are important because they offer chance to clinicians to located in the Southern Province of Rwanda, Huye District. The funding
prescribe affordable treatment regimens with confidence. covered study drugs and cost for endoscopy. The funder of the study had
no role in study design, data collection, data analysis, data interpretation, or
Due to financial constraints, this study was conducted writing of the report.
in one center and the outcome measurement was lim-
ited to the use of FAT, which turned out to be a poorly Availability of data and materials
The datasets used and/or analyzed during the current study are available
sensitive diagnostic test in our study population, thus
from the corresponding author on reasonable request.
limiting the study’s power.
The treatment duration was 10 days across all the four Authors’ contribution
arms of treatment in our study. Infact, earlier studies did JDK, TW, CB, JN and DH were involved in the conception and design of the
study. DH and CB provided laboratory expertise. JDK, TW, CN and FN
not report major differences between short duration acquired the data. JDK and CH conducted data analysis. JDK, TW, CP, VD and
treatments (7 days) and long duration treatments CH interpreted the data. JDK drafted the manuscript. JDK, TW, CH, CB, JN,
(14 days) [36]. Our choice of a 10 day duration was in- DH, VD, CN, CP and FN critically revised the manuscript for intellectual
content, approved the final version and agreed to be accountable for all
spired by the Maastricht IV/ Florence Consensus Report aspects of the work.
which suggested that extending therapies to 10–14 days
improves eradication success by 5% [25]. We chose the Authors’ information
JDK: Attending physician at CHUK. Currently undertaking a Research
lower end of the optimal duration due to financial and Fellowship at the National Institutes of Health, USA. CH: Vanier Scholar and
adverse drug effect considerations. It is important to PhD Candidate, School of Population and Public Health, Faculty of Medicine,
mention that the weight of recent literature now advo- University of British Columbia, Vancouver, BC, Canada, CP: Associate Professor
of Medicine, University Hospital of Brooklyn, New York, USA, CN: Attending
cates a long course of treatment (14 days) to optimize physician at CHUB, VD: Attending physician at CHUK, Academic Head of
outcomes [21, 37]. Medicine, University of Rwanda, JN: Senior anesthesiologist at CHUB, FN:
This study raises awareness in policy makers and paves Gastroenterologist at CHUK, DH: Emeritus Professor of Medicine at University
of Utah, School of Medicine, CB: Associate Professor of Medical Microbiology,
a path for subsequent studies that will apply more rigor- University of Rwanda. TW: Former Chief of Gastroenterology Unit at CHUB,
ous diagnostic methods such as bacterial culture and Clinical Dean, Calvary Mater Newcastle, Australia.
Kabakambira et al. BMC Gastroenterology (2018) 18:134 Page 9 of 9

Ethics approval and consent to participate 14. “Internal-Medicine-Clinical-Treatment-Guidelines-9-10-2012-1.pdf.” [Online].

The University of Rwanda’s Institutional Review Board approved the study Available: http://www.moh.gov.rw/fileadmin/templates/Clinical/Internal-
(CMHS/IRB/65/2015) and the trial was registered with The Pan African Clinical Medicine-Clinical-Treatment-Guidelines-9-10-2012-1.pdf. [Accessed: 24-Sep-
Trials Registry (Clinical Trials: PACTR201804003257400). All patients gave 2015].
written informed consent after counseling about the study and its potential 15. Farmer PE, et al. Reduced premature mortality in Rwanda: lessons from
risks and benefits. success. Bmj. 2013;346:f65.
16. Tack J, et al. Functional Gastroduodenal disorders. Gastroenterology. 2006;
Consent for publication 130(5):1466–79.
Nota applicable. 17. Lane JA, et al. Impact of helicobacter pylori eradication on dyspepsia, health
resource use, and quality of life in the Bristol helicobacter project:
Competing interests randomised controlled trial. BMJ 2006;332:199.
The authors declare that they have no competing interests. 18. Du L-J, Chen B-R, Kim JJ, Kim S, Shen J-H, Dai N. Helicobacter pylori
eradication therapy for functional dyspepsia: systematic review and meta-
analysis. World J. Gastroenterol. 2016;22(12):3486.
Publisher’s Note 19. P. Moayyedi et al., “Eradication of helicobacter pylori for non-ulcer
Springer Nature remains neutral with regard to jurisdictional claims in dyspepsia,” Cochrane Libr., 2006.
published maps and institutional affiliations. 20. Suzuki H, Moayyedi P. Helicobacter pylori infection in functional dyspepsia,
Nat. Rev. Gastroenterol. Hepatol. 10(3):168, 2013.
Author details 21. Chey WD, Leontiadis GI, Howden CW, Moss SF. ACG clinical guideline:
1
Kigali University Teaching Hospital (CHUK), Kigali, Rwanda. 2Centre for treatment of Helicobacter pylori infection. Am. J. Gastroenterol. 2017;
Health Services and Policy Research, School of Population and Public Health, 112(2):212.
Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada. 22. Katelaris PH, Lowe DG, Norbu P, Farthing MJG. Field evaluation of a rapid,
3
Collaboration for Outcomes Research and Evaluation, Faculty of simple and inexpensive urease test for the detection of helicobacter pylori.
Pharmaceutical Sciences, University of British Columbia, Vancouver, BC, J Gastroenterol Hepatol. 1992;7(6):569–71.
Canada. 4Department of Medicine, University Hospital of Brooklyn, New York, 23. Nkurunziza A, Dusabejambo V, Everhart K, Bensen S, Walker T. Validation of
USA. 5Butare University Teaching Hospital (CHUB), Huye, Rwanda. the Kinyarwanda-version Short-Form Leeds Dyspepsia Questionnaire and
6
Department of Medicine, University of Utah School of Medicine, Salt Lake Short-Form Nepean Dyspepsia Index to assess dyspepsia prevalence and
City, UT, USA. 7School of Medicine and Public Health, Faculty of Health and quality-of-life impact in Rwanda. BMJ Open. 2016;6(6):e011018.
Medicine, University of Newcastle, Newcastle, Australia. 24. Talley NJ, Verlinden M, Jones M. Quality of life in functional dyspepsia:
responsiveness of the Nepean dyspepsia index and development of a new
Received: 12 July 2018 Accepted: 20 August 2018 10-item short form. Aliment Pharmacol Ther. 2001;15(2):207–16.
25. Malfertheiner P, et al. Management of Helicobacter pylori infection—the
Maastricht IV/Florence consensus report. Gut. 2012;61(5):646–64.
References 26. Secka O, et al. Antimicrobial susceptibility and resistance patterns among
1. Walker TD, Karemera M, Ngabonziza F, Kyamanywa P. Helicobacter pylori helicobacter pylori strains from the Gambia, West Africa. Antimicrob Agents
status and associated gastroscopic diagnoses in a tertiary hospital Chemother. 2013;57(3):1231–7.
endoscopy population in Rwanda. Trans. R. Soc. Trop. Med. Hyg. 2014; 27. Oyedeji KS, Smith SI, Coker AO, Arigbabu AO. Antibiotic susceptibility
https://doi.org/10.1093/trstmh/tru029. patterns in helicobacter pylori strains from patients with upper
2. Rouvroy D, Bogaerts J, Nsengiumwa O, Omar M, Versailles L, Haot J. gastrointestinal pathology in western Nigeria. Br J Biomed Sci. 2009;
Campylobacter pylori, gastritis, and peptic ulcer disease in central Africa. Br. 66(1):10–3.
Med. J. Clin. Res. Ed. 1987;295(6607):1174. 28. Smith SI, Seriki A, Ndip R, Pellicano R. Helicobacter pylori infection in Africa:
3. Malfertheiner P, et al. Management of Helicobacter pylori infection—the 2018 literature update. Minerva Gastroenterol. Dietol. 2018;64(3):222-34.
Maastricht V/Florence consensus report. Gut, p. gutjnl–2016. 2016; 29. Segamwenge IL, Kagimu M, Ocama P, Opio K. The utility of the helicobacter
4. Vaira D, et al. Diagnosis of helicobacter pylori infection with a new non-in pylori stool antigen test in managing dyspepsia: an experience from a low
vasive antigen-based assay. Lancet. 1999;354(9172):30–3. resource setting. Afr Health Sci. 2014;14(4):829–34.
5. Wu DC, et al. Comparison of stool enzyme immunoassay and 30. Olufemi FO, Quadri R, Akinduti PA, Bamiro SA. Potential risk Fcators and
immunochromatographic method for detecting< i> helicobacter pylori</i> prevalence of infection of helicobacter pylori in Nigeria. J Sci Res Rep. 2015;
antigens before and after eradication. Diagn Microbiol Infect Dis. 2006; 7(1):42–8.
56(4):373–8. 31. Manes G, et al. Accuracy of the stool antigen test in the diagnosis of
6. Veijola L, Myllyluoma E, Korpela R, Rautelin H. Stool antigen tests in the helicobacter pylori infection before treatment and in patients on
diagnosis of Helicobacter pylori infection before and after eradication omeprazole therapy. Aliment Pharmacol Ther. Jan. 2001;15(1):73–9.
therapy. World J. Gastroenterol. 2005;11(46):7340. 32. Thorell K, Lehours P, Vale FF. Genomics of Helicobacter pylori. Helicobacter.
7. Gatta L, et al. A 10-day levofloxacin-based triple therapy in patients who 2017;22(Suppl. 1):e12409.
have failed two eradication courses. Aliment Pharmacol Ther. 2005;22(1):45–9. 33. Gorrell RJ, Zwickel N, Reynolds J, Bulach D, Kwok T. Helicobacter pylori CagL
8. Gisbert JP, et al. Proton pump inhibitor, clarithromycin and either Hypervariable Motif: A Global Analysis of Geographical Diversity and
amoxycillin or nitroimidazole: a meta-analysis of eradication of helicobacter Association With Gastric Cancer. J. Infect. Dis. 2016;213(12):1927–31.
pylori. Aliment Pharmacol Ther. 2000;14(10):1319–28. 34. Phan TN, et al. Genotyping of helicobacter pylori shows high diversity of
9. Fischbach LA, Zanten SV, Dickason J. Meta-analysis: the efficacy, adverse strains circulating in Central Vietnam. Infect Genet Evol J Mol Epidemiol Evol
events, and adherence related to first-line anti-helicobacter pylori quadruple Genet Infect Dis. 2017;52:19–25.
therapies. Aliment Pharmacol Ther. 2004;20(10):1071–82. 35. Hatakeyama M. Anthropological and clinical implications for the structural
10. An B, et al. Antibiotic resistance in helicobacter pylori strains and its effect diversity of the helicobacter pylori CagA oncoprotein. Cancer Sci. Jan. 2011;
on H. Pylori eradication rates in a single center in Korea. Ann Lab Med. 102(1):36–43.
2013;33(6):415–9. 36. Kim BG, et al. Comparison of 7-day and 14-day proton pump inhibitor-
11. Eisig JN, Silva FM, Barbuti RC, Navarro-Rodriguez T, Moraes-Filho JPP, containing triple therapy for helicobacter pylori eradication: neither
Pedrazzoli J Jr. Helicobacter pylori antibiotic resistance in Brazil: treatment duration provides acceptable eradication rate in Korea.
clarithromycin is still a good option. Arq Gastroenterol. 2011;48(4):261–4. Helicobacter. 2007;12(1):31–5.
12. Tanih NF, et al. An African perspective on helicobacter pylori: prevalence of 37. GRAHAM DY. Treating Helicobacter pylori effectively while minimizing
human infection, drug resistance, and alternative approaches to treatment. misuse of antibiotics. Cleve. Clin. J. Med. 2017;84(4):311.
Ann Trop Med Parasitol. 2009;103(3):189–204.
13. Tanih NF, et al. Marked susceptibility of south African helicobacter pylori
strains to ciprofloxacin and amoxicillin: clinical implications. S Afr Med J.
2010;100(1):49–52.