Probiotics in 14-day triple therapy for Asian pediatric patients with Helicobacter pylori infection: a network meta-analysis

Although 14-day triple therapy has been widely administrated for eradicating Helicobacter pylori (H. pylori) in Asia, its antibiotic-associated side effects restrict the effectivity of eradication therapy in pediatric patients. Therefore, a network meta-analysis (NMA) was conducted to compare the efficacy and safety of probiotics supplemented in 14-day triple therapy in Asian pediatric patients. Materials and Methods Randomized controlled trials (RCTs) were retrieved comprehensively in electronic databases (such as PubMed, Cochrane library, Embase, CNKI, Wan fang database, VIP database and CBM) until April 2017. Additional references were obtained from reviewed articles. NMA was performed using a random-effects model under a frequentist framework. Results Seventeen RCTs were included. NMA indicated that Bifidobacterium infantis+Clostridium butyricum was most beneficial for H. pylori eradication rates (P-score = 0.82) and Bacillus mesentericus+Clostridium butyricum+Streptococcus faecalis for total side effects (P-score = 0.77). Taken together, Bacillus mesentericus+Clostridium butyricum+Streptococcus faecalis was the best one to supplement in 14-day triple therapy due to its efficacy (P-score = 0.72) and safety (P-score = 0.77). Additionally, pairwise meta-analysis indicated that probiotics supplemented 14-day triple therapy significantly increased H. pylori eradication rates (RR: 1.16, 95%CI: 1.07–1.26) and reduced the incidence of total side effects (RR: 0.40, 95%CI: 0.34–0.48) compared with placebo. Conclusions Bacillus mesentericus+Clostridium butyricum+Streptococcus faecalis is the optimal probiotic regime of reducing total side effects and improving eradication rates when supplemented 14-day triple therapy. Further direct evidence is needed to warrant it.


INTRODUCTION
In contrast with developed countries, Helicobacter pylori (H. pylori) infections occur earlier in life and with a higher prevalence in the developing world [1,2]. In parts of the developing world, especially in Asian countries, the prevalence of H. pylori infection increased with age and exceeded 80% by 19 years of age [2]. For instance, in Bangladesh, the prevalence of H. pylori was already 58% by 4 years of age and quickly rose to 82% by 9 years of age [2]. Similar findings were reported from pediatric studies in other Asian countries including India, Sri Lanka and China [1,3,4]. Children infected with H. pylori could develop many complications and symptoms like recurrent

Meta-Analysis
abdominal pain, dyspepsia, iron deficiency anemia, chronic gastritis, peptic ulcers and gastric cancer [5]. H. pylori infection has also been reported to cause a wide variety of extra-digestive manifestations, such as sudden infant death syndrome and short stature [6]. H. pylori eradication not only can cure peptic ulcers temporarily, but also reduce the lifetime risk of gastric cancer in long term due to increasing longevity in developing countries [7].
Because H. pylori is a bacterium, antibiotics certainly as a therapeutic agent will be administrated to eradicate it, like recent triple therapy. Nevertheless, recently accepted triple therapies still demonstrate failure rates as high as 20% to 30% due to increasing clarithromycin and metronidazole resistances as well as poor tolerability and reduced compliance [7][8][9]. A longer treatment duration may increase eradication rates that suggests an increase to 14 days instead of 7 days [7]. There is a strong recommendation by Joint ESPGHAN/ NASPGHAN Guidelines for 14-day triple therapy as firstline therapy for pediatric patients with H. pylori infection where clarithromycin and metronidazole resistances are presenting commonly in developing countries [10]. In addition, supplementary probiotics can reduce side effects and improve efficacy of H. pylori eradication therapies [11][12][13][14][15][16].
Although these studies suggest that there may be a benefit for reduction of side effects and improved efficacy when added to H. pylori eradication therapies, the evidence from Asia is limited and the same meaningful conclusions amongst these different probiotic strains seemed not be obtained as well as for 14-day triple therapy. Therefore, network meta-analysis (NMA) was performed to compare the efficacy and safety of specific probiotic supplementations in 14-day triple therapy for Asian pediatric population.

MATERIALS AND METHODS
We performed and reported the results of this pairwise and NMA in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statements [17,18].

Literature research
We conducted a comprehensive search on databases including PubMed, Cochrane library, Embase and 4 Chinese Database [China National Knowledge Infrastructure (CNKI), database of Wan fang, VIP database and the Chinese Biomedical Database(CBM)] until April 2017. Literature were identified with keywords including (Helicobacter pylori OR H. pylori OR HP), probiotics, (pediatrics OR children), (randomized controlled trial OR RCT OR clinical trial) and (triple therapy OR eradication therapy) (search strategy seen in Supplementary Files). Additionally, manual search for the references cited in reviews and meta-analyses was performed as well. There were no language restrictions.

Inclusion and exclusion criteria
We included the RCTs meeting following eligibility criteria: (i) clinical trials from Asia; (ii) pediatric patients with H. pylori infection; (iii) clinical trials comprising of 14-day triple therapy (proton pump inhibitor and two antibiotics) plus placebo or no extra intervention and the same triple therapy plus probiotics; (iv) H. pylori eradication was confirmed at least 4 weeks after therapy termination. Articles were excluded if (i) reviews or comments; (ii) articles without data of H. pylori eradication rates and/or side effects.

Study selection
The articles were screened and selected by two researchers (Pengfei Chen and Lirong Zeng) independently. Initially, records of duplicates were removed using Endnote 7.4 software (Thomson Reuters). Remaining abstracts and full texts were reviewing based on inclusion and exclusion criteria. Disagreements were resolved by discussions.

Data extraction and quality assessment
Two reviewers (Juanjuan Wen and Pailan Peng) independently extracted the relevant information from each eligible study depending on a pre-prepared data abstraction sheet. Data were including location and study design of trials, clinical characteristics and number of patients, initial and rechecking methods for conformation of H. pylori infection, strains and period of probiotic regimens, eradication regimens and outcome data. The quality of included studies was assessed using Jadad scale, including three items such as randomized (2 points), double-blinding (2 points) and withdrawals and drop outs (1 point) [19]. A Jadad score of 3 or higher was considered as high quality. Disagreements were resolved by discussions.
The primary outcomes were eradication rates of H. pylori and incidence of total side effects, and secondary outcomes included incidence of specific side effects. The specific side effects of interest were the common symptoms that appeared during treatment, including diarrhea, nausea or vomiting, taste disturbance, loss of appetite, bloating, headache, constipation, and abdominal pain.

Data synthesis and analysis
Conventional pairwise meta-analysis was initially carried out for the comparison between probiotic group and placebo group. The estimated effects using relative ratios (RRs) with 95% confidence intervals (CIs) based on the random effects model were analyzed by STATA 11.0 software (Stata Corporation, College Station, TX, USA). Secondary, NMA based on the random effects model was performed by netmeta R package with a frequentist framework [20,21]. The P-scores that were based solely on the point estimates and standard errors of the network estimates were calculated to rank all probiotic regimens and the higher P-score value was correspond to better result for respective probiotic regimen [22].
Publication bias was evaluated using the Egger's and Begg's tests [23,24]. If a P value < 0.05, a trim and fill method was further conducted and recalculated the pooled estimate iteratively, until the funnel plot was symmetrical around the (new) pooled estimate [25]. Additionally, heterogeneity was measured with Q statistic and I 2 statistic [26,27]. If an I 2 value of > 50% or a P value < 0.05, sensitivity analysis was conducted by excluding one individual study each time to assess the influence of each individual study on the pooled RRs.
The included trials involving 1932 pediatric patients has investigated 10 probiotic regimens, such as (1

Network meta-analysis
There were 9 and 8 probiotic regimens supplemented 14-day triple therapy reporting data of eradication rates and total side effects, respectively ( Figure 4). In term of improving eradication rates of H. pylori, B. infantis+C. butyricum was the best to add in 14-day triple therapy (P-score = 0.82). B. mesentericus+C. butyricum+S. faecalis was most beneficial for total side effects (P-score = 0.77). Taken together, B. mesentericus+C. butyricum+S. faecalis was the best one to supplement in 14-day triple therapy due to its efficacy (P-score = 0.72) and safety (P-score = 0.77) ( Figure 5). The relative effectiveness of all possible pairs of probiotics was presented in league tables (Supplementary Tables 3-4).

Publication bias
Publication bias was assessed using data of H. pylori eradication rates. Visual inspection of the Begg's funnel plot found a slightly asymmetrical distribution ( Figure 7A), which indicated having potential publication bias (P = 0.08). By trimming and inputting 5 studies using the trim and fill method ( Figure 7B), the recalculated pooled RR for H. pylori eradication rates was 1.20 (95% CI: 1.12-1.29) (Supplementary Figures 1-2), which was not significantly changed from the initial estimate (RR: 1.16, 95% CI: 1.07-1.26).

DISCUSSION
This complexed meta-analysis included RCTs all from Asian countries, where appears high prevalence rate of H. pylori infection [7]. Our meta-analysis showed that probiotic supplementation can statistically benefit 14-day triple therapy with both eradication rates and safety. Amongst included probiotics, B. infantis+C. butyricum was best for eradication rates of H. pylori and B. mesentericus+C. butyricum+S. faecalis best for reducing the incidence of total side effects. Additionally, it further found that beneficial effects of probiotics are strain specific. For example, as an addition in 14-day triple therapy, B. bifidum+B. infantis+L. acidophilus+L. bulgaricus+L. casei+L. reuteri+ Streptococcus could reduce the incidence of diarrhea, which is the most common side effect associated with antibiotics. Given the results from this study, probiotics as a supplementation to 14-day triple therapy are considered as an alternative treatment regimen. Moreover, specific probiotics are administrated for corresponding specific side effect.
Probiotics, namely probiotic bacteria, have defined as live microorganisms, which most distribute in anterointernal mucosa and skin, confer a health benefit for the host. In infectious disease, probiotics act in the gastrointestinal luminal environment by reducing luminal pH, inhibiting adhesion of other harmful microorganisms and antagonizing pathogenic bacteria by secreting bacteriocins [45]. Kabir AM and colleagues [46] found that L. salivarius was against and inhibited colonization of H. pylori, when supplemented in eradication therapy may be enhance the eradication of H. pylori. L. salivarius inhibited both the attachment and IL-8 release of H. pylori in vitro. H. pylori could not colonize the stomach of L. salivarius infected gnotobiotic BALB/c mice, but colonized in large numbers and subsequently caused active gastritis in germ free mice. In addition, L. salivarius given after H. pylori implantation could eliminate colonization of H. pylori [46]. The conventional meta-analysis by Li S et al. in children demonstrate that probiotics supplementation has efficacy for H. pylori eradication therapy [14], which was in line with our conventional meta-analysis; but without more information on specific probiotics. Lately, two network meta-analyses have revealed the comparative effectiveness on probiotics supplementation for H. pylori eradication therapy [47,48]. One network meta-analysis suggested that L. acidopilus was a slightly better choice for triple therapy of 7 days and 14 days in adult patients, and S. boulardii was more applicable for 10-day triple therapy [48]. Another network meta-analysis of pediatric population suggested that, to supplemented triple therapy, L. casei was the best for H. pylori eradication rates, and multi-strain probiotics (L. acidophilus plus L. rhamnosus) for total side effects [47]. These results of above network meta-analyses were different from our study, which might be greatly associated with different population and different area. Additionally, these two network meta-analysis also included some studies from non-Asian counties, which generally used the probiotics of Lactobacillus species. However, less studies of Lactobacillus species were included in our network metaanalysis, which maybe weaken the strength of evidence on Lactobacillus species when ranking the effectiveness among different probiotics.
In 2017, the Joint ESPGHAN/NASPGHAN Guidelines has recommended 14-day triple therapy as first-line therapy for pediatric patients with H. pylori infection because of clarithromycin and metronidazole resistances in developing countries [10]. In our study, probiotics, especially B. infantis+C. butyricum, as addition could be a new choice for enhancing 14-day triple therapy to eradicate H. pylori in children.
Apart from the efficacy of eradication therapy, antibiotic-associated side effects and bacterial resistance cannot be neglected because these factors could lead to the poor tolerability of drugs and hinder patients' compliance, and consequently cause anti-H. pylori treatment failure in clinical practice [49]. In our study,     the occurrence of antibiotic-associated total side effects was significantly reduced by the addition of probiotics as well as specific probiotic stains like multi-strain probiotics (such as B. longum, L. bulgaricus and S. thermophilus), S. boulardii and L. acidophilus compared with the placebo (Supplementary Table 2).
This study has several strengths and weaknesses. We concentrated on clinical trials that all investigated14-day triple therapy in Asia where H. pylori infection have great impact on children healthy. So, such clinical conclusions for mongoloid race, a main population in Asia, can be firstly obtained. However, the methodologies of all RCTs were not rigorously carried out, which might affect the conclusions. Additionally, strengths of this complex metaanalysis were that the different strains or regimens of probiotics were separated for assessing efficacy and safety. Therefore, probiotics of most effectivity can be guided to add in 14-day triple therapy for the specific purpose, especially to prevent certain side effect. Nevertheless, most trials did not take antibiotic-associated side effects as key outcomes, and not documented more information (such as antibiotic resistant of H. pylori, initial treatment or second-line strategy for experiencing a failure of eradicating H. pylori, etc.).
All in all, our meta-analyses found that probiotics are beneficial for 14-day triple therapy in Asian pediatric patients with H. pylori infection. For obtaining greater benefit of improving eradication rates and reducing antibiotic-associated side effects, the mixture of B. mesentericus+C. butyricum+S. faecalis may be recommended to add in 14-day triple therapy. Further direct evidence is needed to warrant it.