A systematic search was performed in Embase, Ovid Medline, Cochrane Library and LILACS using the following strategy: type of study: randomized clinical trials (RCTs) assessing the effectiveness and safety of using probiotics as adjuvant therapy in combination with triple or quadruple therapy for the eradication of H. pylori in adults; publication period: January 2010 to May 2020; language of publication: English and Spanish; search terms (alone and with different combinations and their variations depending on the database): "Helicobacter pylori", "H. pylori, eradication therapy", "triple therapy", "quadruple therapy, probiotics", "lactobacillus reuter", "Saccharomyces boulardii", "Lactobacillus plantarum", "Pediococcus acidilactici boulardii", "Lactobacillus rhamnosus", "Lactobacillus strains", "Lactobacillus bulgarius", "Streptococcus faecium", "Streptococcus thermophilus", and "Bifidobacterium longum", which were combined with the Boolean connectors "OR" and "AND" to establish the search equations.

Additionally, a gray literature search was performed in Google Scholar and the database search was supplemented by a snowball search.

Studies in adult patients (>18 years) in which the intervention group was treated with standard eradication therapy plus probiotics and the control group with the same eradication scheme with or without placebo were included. On the other hand, studies with pregnant women, trials using adjuvant eradication treatments other than probiotics, and papers whose full text was not available were excluded.

Two investigators (GJT and WAOR) independently performed the initial screening of the studies retrieved in the initial searches by reading titles and abstracts; subsequently, they verified compliance with the inclusion criteria by reading the full text. Disagreements regarding the final selection of studies were resolved by consensus.

The methodological quality assessment of the selected studies was carried out based on the Cochrane Handbook for Systematic Reviews of Interventions,¹⁹ which evaluates the following domains: random sequence generation (selection bias), allocation concealment (allocation bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), and selective reporting (reporting bias). The risk of publication bias was assessed using funnel plots. The GRADE approach was used to assess the degree of certainty of the evidence supporting the interventions.²¹

Two reviewers extracted the data from the studies included in the review (GJT and KEO). During this process, general information about the studies was obtained, such as author/ authors, year and place of publication, design, and objectives. Likewise, the following information was extracted: characteristics and number of participants; types of standard therapy used (triple or quadruple); type of probiotics used; dose, frequency, and route of administration of probiotics; and quantitative data on the effectiveness and safety outcomes of adjuvant probiotics in combination with standard therapy for the eradication of H. pylori.

Data are presented descriptively using standardized tables, while information extracted from the studies included in the review is presented in summary tables. The pooled effect (odds ratio (OR) or relative risk (RR)) was estimated for the effectiveness (eradication) and safety (adverse events) outcomes, with their respective confidence intervals (95%CI) depending on the type of outcome. Absolute risk reduction and relative risk reduction of the assessed interventions versus placebo or no intervention were also calculated.

The statistical model to determine the pooled effect using OR and RR was selected based on the number of included studies and their heterogeneity, which was tested during the analysis using the chi-square test (x²) and the index of inconsistency (I²), where heterogeneity values >30% measured with I2 but with p-values >0.05 obtained using x2 showed that the differences in effect can be explained by chance. Given the foregoing, a fixed effects model for the meta-analysis and the Z value for the statistical significance test were performed to estimate the pooled measure of effect.

Subgroup analyses were also performed considering the type of eradication therapy (triple or quadruple) and the type of probiotic used. Moreover, clinical, statistical, and methodological sources of heterogeneity between studies were explored by means of a sensitivity analysis when heterogeneity according to P was >60%. When it was not possible to explain the high heterogeneity (P>60%) or when it was not possible to adjust by presenting the evidence by subgroups, the range of effects was presented per type of outcome based on the estimator reported in the studies, together with their respective 95%CI. Statistical analyses were performed in RevMan software.

The initial search yielded 1 023 results, 823 from the databases and 200 from Google Scholar, of which 367 were removed because they were duplicates and 578 after reviewing the title and abstract (504 and 74, respectively). Of the 78 references selected for full-text review, 21 were excluded because the full text was not available, 11 because they were written in languages other than English or Spanish, 15 because they were not RCTs, and 16 because they did not meet the inclusion criteria; there were also 3 duplicate studies (details of the excluded studies are presented in Annex 1). Thus, 12 studies were included for full analysis (Figure 1).

Figure 1 Source: Own elaboration.

The 12 studies included in the meta-analysis are RCTs, and of these, 10 were double-blind,^22-311 was triple-blind,³² and in the other no placebo was used.³³ Triple therapy was used in 9 of these 12 studies^{22-25,27-30,33}with a total of 824 participants, while quadruple therapy was used in 2,^31,32with a total of 256 participants; similarly, one of these studies evaluated the use of both triple therapy and quadruple therapy.²⁶

Of the 9 studies that used triple therapy, 2 used the drugs clarithromycin, amoxicillin, and lansoprazole.;^24,28 6 used a triple therapy with different drugs: omeprazole plus clarithromycin and amoxicillin,²⁷ rabeprazole plus clarithromycin and tinidazole,²² esomeprazole plus levofloxacin and amoxicillin,³³ lansoprazole plus tetracycline and furazolidone,²⁹clarithromycin plus amoxicillin and rabeprazole,³⁰ and esomeprazole plus amoxicillin and clarithromycin;²³ finally, a single study used clarithromycin and amoxicillin and did not specify the PPI.²⁵

The duration of triple therapy was variable: in 4 studies, probiotics were administered for 2 weeks;^24,27,30,33in 4 studies, they were administered for 4 weeks;^22,23,28,29and in 1 study, they were administered for 13 weeks.²⁵

Both studies that analyzed quadruple therapy^31,32used bismuth, amoxicillin, clarithromycin, and omeprazole for 2 weeks.^31,32

The study that evaluated the use of both therapies used omeprazole, clarithromycin and amoxicillin for 10 days in the triple therapy and bismuth, amoxicillin, clarithromycin and omeprazole, also for 10 days, in the quadruple therapy.²⁶

Of the 12 studies included in the analysis, 4 used Lactobacillus reuteri as an adjuvant probiotic,^22,25,27,33 5 used a mixture of probiotics,^{26,28,29,31,32}1 used Saccharomyces, ²⁴ another used Bacillus clausi, ³⁰ and another used Bifidobacterium longum. ²³

In 11 studies, a comparison versus placebo plus triple or quadruple therapy was made, except in the study by Ojetti et al., ³³ where the comparison was made only with triple therapy. Likewise, 10 studies specified that the placebo was equal to the investigational product in terms of form, presentation, smell and taste,^22-31while the other did not specify anything in this regard.³² Details of the included studies are presented in Tables 1 and 2.

Table 1

Study	Methods	Participants	Type of therapy	Intervention/dose/ duration	Control	Outcomes
Armuzzi et al. 22 2001 Italy	Double-blind, prospective, randomized, placebo-controlled trial	Healthy asymptomatic patients over 18 years of age, positive for Helicobacter pylori and under treatment	Triple therapy: rabeprazole, clarithromycin and tinidazole	Lactobacillus GG for 14 days	Placebo	- Adverse events - Eradication
Chitapanarux et al. 23 2015 Thailand	Double-blind, prospective, randomized, placebo-controlled trial	Patients with H. Pylori infection on antibiotic treatment	Triple therapy: esomeprazole, amoxicillin and clarithromycin	Bifidobacterium longum for 4 weeks	Placebo	- Adverse events - Eradication
Cindoruk et al. 24 2007 Turkey	Prospective, randomized, placebo-controlled trial	Adult patients with H. pylori infection confirmed on upper digestive tract biopsy and dyspepsia symptoms	Triple therapy: clarithromycin, amoxicillin and lansoprazole	Saccharomyces boulardii for 14 days	Placebo	- Adverse events - Eradication
Francavilla et al. 25 2014 Italy	Double-blind, prospective, randomized, placebo-controlled trial	Adult patients (over 18 years of age) with dyspepsia who had never received H. pylori eradication therapy	Triple therapy: proton pump inhibitor, clarithromycin and amoxicillin	Lactobacillus reuteri (mixture of L. reuteri DSM 17938 and L. reuteri ATCC PTA 6475) for 28 days	Placebo	- Adverse events - Eradication
Emara et al. 27 2014 Egypt	Double-blind, prospective, randomized, placebo-controlled trial	Adult patients (18-60 years old) with dyspepsia and confirmed H. pylori infection and no previous treatment	Triple therapy: omeprazole, amoxicillin and clarithromycin	L reuteri (mixture of L. reuteri DSM 17938 and L. reuteri ATCC PTA 6475) for 4 weeks	Placebo	- Adverse events - Eradication
Myllyluoma et al. 28 2005 Finland	Double-blind, randomized, placebo-controlled trial	Adult patients (18-70 years of age) with H. pylori infection confirmed by rapid blood test	Triple therapy: lansoprazole, clarithromycin and amoxicillin	Lactobacillus rhamnosus GG, L. rhamnosus LC705, Bifidobacterium breve Bb99, and Propionibacterium freudenreichii ssp. shermanii JS for 3 weeks	Placebo	- Adverse events - Eradication
Navarro-Rodríguez et al. 29 2013 Brazil	Double-blind, randomized, placebo-controlled trial	Adult patients (over 18 years of age) with previously untreated H. pylori infection, diagnosed with peptic ulcer or functional dyspepsia, without decompensated chronic disease, and no use of antiinflammatory drugs or antibiotics in the last 4 weeks	Triple therapy: lansoprazole, tetracycline and furazolidone	Lactobacilos acidophilus, L. rhamnosus, Bifidobacterium bifidum, and Streptococcus faecium for 30 days	Placebo	- Adverse events - Eradication
Nista et al. 30 2004 Italy	Single-center, double-blind, prospective, randomized, placebo- controlled trial	Adult patients (18-65 years old) without gastrointestinal symptoms in the last 3 months and with H. pylori infection confirmed by breath test	Triple therapy: rabeprazole, clarithromycin, amoxicillin	Bacillus clausii and Enterogermina for 14 days	Placebo	- Adverse events - Eradication
Ojetti et al. 33 2012 Italy	Single-center, controlled, prospective, randomized trial	Adult patients (18-65 years old) with H. pylori infection confirmed by breath test	Triple therapy: esomeprazole, levofloxacin and amoxicillin	L. reuteri for 14 days	None	- Adverse events - Eradication
Shafaghi et al. 31 2016 Iran	Double-blind, prospective, randomized, placebo-controlled trial	Patients with H. pylori infection confirmed by histopathological examination	Quadruple therapy: bismuth, clarithromycin, amoxicillin and omeprazole	Lactobacillus casei, L. rhamnosus, Streptococcus thermophilus, Bifidobacterium breve, L. acidophilus, B. longum, and Lactobacillus bulgarius for 14 days	Placebo	- Adverse events - Eradication
Shavakhi et al. 32 2013 Iran	Triple-blind, randomized, placebo-controlled trial	Adult patients with peptic ulcer disease and H. pylori infection confirmed by rapid urease test or histological studies	Quadruple therapy: omeprazole, bismuth, amoxicillin and clarithromycin.	Seven bacterial species including Lactobacillus strains (L. casei, L. rhamnosus, L. acidophilus and L. bulgaricus), Bifidobacterium strains (B. breve and B. longum), and Streptococcus thermophiles	Placebo	- Adverse events - Eradication
McNicholl et al. 26 2018 Spain	Double-blind, randomized, placebo-controlled trial	Adult patients (18-70 years of age) with confirmed H. pylori infection	Triple therapy: omeprazole, clarithromycin and amoxicillin Quadruple therapy: clarithromycin, bismuth, amoxicillin and omeprazole	Lactobacillus plantarum CETC7879 and Pediococcus acidilactici CETC7880 for 10 days	Placebo	- Adverse events - Eradication

ATCC: American Type Culture Collection, Manassas, VA, USA. DSM: DSMZ-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany.

Table 2

Study	Number of patients	Probiotics	Control	Age (mean)	Female sex and probiotics	Female sex control (n)	Male sex and probiotics	Male sex control (n)	Eradication with probiotics	Eradication control (n)
Armuzzi et al. 22 2001	60	30	30	40 years old	NR	NR	NR	NR	25	24
Chitapanarux et al. 23 2015	63	31	32	NR	17	18	14	14	28	22
Cindoruk et al. 24 2007	124	62	62	48 years old	36	44	26	18	44	37
Francavilla et al. 25 2014	86	43	43	NR	32	29	18	21	33	29
McNicholl et al. 26 2018	68	34	34	NR	NR	NR	NR	NR	32	31
Emara et al. 21 2014	70	35	35	NR	22	24	13	11	26	23
Myllyluoma et al. 28 2005	47	24	23	55.6 years old	13	16	10	8	21	19
Navarro-Rodríguez et al. 29 2013	107	52	55	NR	34	33	21	19	45	40
Nista et al. 30 2004	120	60	60	NR	27	35	33	25	39	37
Ojetti et al. 33 2012	90	45	45	NR	NR	NR	NR	NR	36	27
Shafaghi et al. 31 2016	76	38	38	43.5 years old	20	21	18	17	35	24
Shavakhi et al. 32 2013	180	90	90	NR	41	30	49	60	73	69

NR: not reported; n: absolute number of participants.

The risk of random sequence generation (selection bias) was unclear in 3 studies.^22,30,31In the allocation concealment domain, 2 studies had high risk^30,33and the risk was unclear in 7.^{22-24,26-28,31}in the domains blinding of participants and personnel, and blinding of outcome assessment, 1 study showed high risk³³ and in another the risk was unclear.²⁷ In the incomplete outcome data domain, 1 study showed high risk²³ and the risk was unclear in 2 studies.^30,31Finally, in the selective reporting domain, the risk was not clear in 2 studies.^27,29No other types of bias were observed in the studies (Figure 2).

Figure 2 Source: Own elaboration.

The exploration of publication bias in the funnel plot is reported in Annex 2.

The use of probiotics as adjuvant therapy in triple therapy for H. pylori eradication showed a benefit over the use of placebo (OR=1.48; 95%CI: 1.05-2.09; I²=0%), as illustrated in Figure 3, which also reports the x² and I² heterogeneity values and the Z value of the statistical significance test for the pooled measure of effect.

Figure 3 Source: Own elaboration.

Regarding the use of adjuvant probiotics in quadruple therapy, it was not possible to estimate a pooled effect measure due to the high heterogeneity (P=78%); however, in the three studies identified, there was no evidence of differences in the effect found between the use of probiotics and placebo as adjuvant therapy: (OR=7, 95%CI: 0.35-138.14),²⁶ (OR=6.81, 95%CI: 1.76-26.27),³¹ and (OR=0.77, 95%CI: 0.37-1.57).³² Moreover, an absolute and relative increase in H. pylori eradication success with probiotics versus placebo of 8.3% and 11%, respectively, was found. Likewise, adjuvant therapy with the probiotic L. reuteri plus quadruple therapy showed a significant benefit compared to placebo (RR=1.75, 95%CI: 1.05-2.92; P=0 %), with an absolute eradication rate of the probiotic group exceeding the eradication treatment by 5.51% and a relative increase in eradication rate success of 14.16% (Annex 3).

In the relative risk analysis, it was observed that 2 studies using a probiotic mixture did not show a significant benefit in terms of effectiveness compared to the use of placebo: (RR=0.969, 95%CI: 0.847-1.108)²⁶ and (RR=0.945, 95%CI: 0.812-1.100)³² (Table 3).

Table 3

Study	Relative risk reported in the original study	95%CI	Risk difference	95%CI
Probiotic mixture
McNicholl et al. 26 2018	0.969	0.847-1.108	-0.029	(-0.153)-0.094
Myllyluoma et al. 28 2005	1.105	0.857-1.426	0.083	(-0.127)-0.293
Navarro-Rodríguez et al. 29 2013	1.064	0.876-1.292	0.049	(-0.104)-0.202
Shafaghi et al. 31 2016	1.458	1.124-1.891	0.289	0.114-0.465
Shavakhi et al. 32 2013	0.945	0.812-1.100	-0.044	(-0.163)-0.075
Saccharomyces
Cindoruk et al. 24 2007	1.189	0.918-1.541	0.113	(-0.053)-0.279
Bacillus clausii
Nista et al. 30 2004	1.015	0.799-1.290	0.010	(-0.161)-0.182
Bifidobacterium longum
Chitapanarux et al. 23 2015	1.314	1.013-1.705	0.216	0.024-0.407

CI: confidence interval.

This publication is derived from the thesis entitled Efectividad y seguridad del uso de probióticos en la erradicación de Helicobacter pylori: revisión sistemática de la literaturay meta-análisis (Effectiveness and safety of the use of probiotics in the eradication of Helicobacter pylori: systematic literature review and meta-analysis),³⁶ developed by the authors of this manuscript for obtaining a degree in gastroenterology.