Abstract

Background:

The role of oral bacterial lysates (OBLs) as adjuvant immunomodulatory therapy in pediatric asthma requires clarification. This systematic review and meta-analysis evaluates their efficacy and safety.

Methods:

We searched eight databases for randomized controlled trials (RCTs) in children comparing standard asthma therapy plus OBLs versus standard therapy/placebo. Primary outcomes were clinical (wheezing/infection frequency, symptom improvement time, treatment efficacy) and lung function parameters. Secondary outcomes included immune biomarkers and adverse events. Random-effects meta-analyses were performed.

Results:

Twenty-eight RCTs (n = 2,893) were included. Adjunctive OBLs therapy significantly reduced wheezing/exacerbation frequency (Mean Difference (MD) = −3.00,95% confidence intervals (CI): 4.07 to −1.93), shortened symptom improvement time (MD = −3.13 days, 95%CI: 4.10 to −2.15), reduced Respiratory Tract Infection (RTI) frequency (MD = −2.43,95%CI: 3.62 to −1.23) and increased overall treatment efficacy rate (relative rates (RR) = 1.17, 95%CI: 1.13–1.21). Improvements occurred in lung function (Forced Expiratory Volume in 1 s [FEV1], Forced Vital Capacity [FVC], Peak Expiratory Flow [PEF]) and immune parameters (increased the level of T-lymphocyte subsets (CD3+,CD4+,CD4+/CD8+) and salivary secretory immunoglobulin A (sIgA), decreased peripheral eosinophil (EOS) count and the level of interleukin-4(IL-4), eosinophil cationic protein (ECP), fractional exhaled nitric oxide (FeNO). Adverse events did not increase significantly (RR = 1.26, 95%CI: 0.93–1.70). Subgroup analyses showed consistent benefits across follow-up duration, age, and sample size, with background inhaled corticosteroids (ICS) therapy being a potential effect modifier (P = 0.06).

Conclusion:

Adjunctive OBLs therapy is associated with improved clinical outcomes, lung function, and immune modulation in pediatric asthma, with a favorable safety profile. However, methodological limitations, substantial heterogeneity, and potential biases warrant caution. While promising, more rigorous and long-term trials are needed to define its precise therapeutic role and target population.

Systematic Review Registration:

Identifier CRD420261281796.

1 Introduction

Asthma remains the most prevalent chronic respiratory disease in children worldwide, constituting a significant global public health challenge characterized by recurrent episodes of wheezing, breathlessness, chest tightness, and cough (Bateman et al., 2008). These symptoms stem from a complex interplay of chronic airway inflammation, bronchial hyperresponsiveness, and variable airflow obstruction (Barnes, 2008). The global burden is substantial, with an estimated 5%–10% of pediatric populations affected, leading to frequent healthcare utilization, diminished quality of life, impaired school performance, and considerable economic costs for families and healthcare systems (Asher and Pearce, 2014).

The cornerstone of pediatric asthma management involves controller medications aimed at suppressing underlying inflammation and preventing exacerbations. Inhaled corticosteroids (ICS) are the first-line and most effective anti-inflammatory controller therapy (Sanjiv-Sin gh Rawat. Global Initiative for Asthma GINA, 2025). Leukotriene receptor antagonists (LTRA) and long-acting beta2-agonists (LABA), often in combination with ICS, serve as additional options (Muireann et al., 2009). While these therapies are effective for many, clinical challenges persist. A subset of children continues to experience breakthrough symptoms, acute exacerbations often triggered by viral respiratory infections, and a phenomenon described as the “recurrent wheeze-infection cycle” (Holt and Sly, 2012; Jartti and Gern, 2017). Furthermore, concerns regarding the potential long-term effects of ICS, though generally considered safe, and suboptimal adherence to inhalation devices can limit real-world effectiveness (Petrisko et al., 2008; Marjolein et al., 2014). These limitations highlight the need for complementary therapeutic strategies that target different aspects of asthma pathophysiology, particularly the immune dysregulation and susceptibility to infections that drive disease morbidity.

The pathogenesis of asthma, especially in children, involves a dysregulated immune response. A shift towards T-helper 2 (Th2) cell-dominant inflammation is common, leading to elevated levels of cytokines like interleukin-4 (IL-4), IL-5, and IL-13, which promote eosinophilic inflammation, immunoglobulin E (IgE) production, and airway remodeling (Gans Melissa and Gavrilova, 2019). Concurrently, impaired innate immune responses and mucosal defense mechanisms are believed to increase susceptibility to respiratory pathogens, which are the most common triggers of acute exacerbations in children (Busse et al., 2010). This intersection between allergic inflammation and infection susceptibility provides a rationale for immunomodulatory interventions.

Early life is crucial for establishing immune tolerance, which is important for maintaining the stability and normal physiological functions of innate and adaptive immune responses. Stable immune function in the airways depends on appropriate interactions among the microbiota colonizing the mucosa, the host immune response, and environmental microorganisms (Bloomfield Sally et al., 2016). Urbanization has significantly improved hygiene conditions and changed daily life patterns. The increased rate of cesarean sections, the decline in breastfeeding rates, and the overuse of antibiotics may significantly reduce microbial diversity, leading to immune tolerance dysfunction and promoting the occurrence of allergic diseases such as asthma (Adnan et al., 2012). Based on these theoretical backgrounds, the impact of environmental microbial components on immune system function has received increasing attention. Subsequently, a large number of animal and clinical trials have been conducted to study microorganisms and their lysates or metabolites. Bacterial lysates are oral immunomodulators prepared from extracts of inactivated bacteria commonly involved in respiratory tract infections (e.g., Haemophilus influenzae, Streptococcus pneumoniae, Klebsiella pneumoniae, etc.). Products like OM-85 (Broncho-Vaxom) have been extensively studied. Their proposed mechanism of action involves stimulating both innate and adaptive immunity via the gut-associated lymphoid tissue (Vadim et al., 2021; Adriana and Chorostowska-Wynimko, 2008). They are thought to enhance mucosal defense by increasing sIgA production, promote a balanced Th1/Th2 response by modulating cytokine production (e.g., increasing interferon-γ (IFN-γ), decreasing IL-4), and potentially reduce the frequency and severity of respiratory infections (Agnieszka et al., 2022; Elpida et al., 2025). By potentially breaking the “infection-wheeze” cycle and modulating the underlying inflammatory milieu, bacterial lysates present a promising non-steroidal adjuvant therapy for pediatric asthma.

Over the past 2 decades, numerous randomized controlled trials (RCTs), particularly from regions like China and Europe, have investigated the add-on effect of bacterial lysates to standard asthma therapy in children. These studies have reported outcomes across diverse domains: clinical (e.g., exacerbation rate, symptom recovery time), functional (lung function), immunological (T-cell subsets, cytokines), and safety. However, the results have been inconsistent. Some trials report significant benefits in reducing wheezing episodes and improving immune parameters, while others show more modest or non-significant effects (Lan et al., 2022; Razi et al., 2010; Zhuang-Gui et al., 2007; Emeryk et al., 2018; Siri et al., 2020). Furthermore, existing systematic reviews on this topic are either limited in scope, outdated, or focused on specific subsets of patients (e.g., those with recurrent infections), leaving a gap in the evidence synthesis (Castro-Rodriguez et al., 2024).

A comprehensive and up-to-date quantitative synthesis is therefore urgently needed to clarify the overall value of this adjunctive therapy. Specifically, there is a need to concurrently evaluate its core clinical benefits (on exacerbations, symptoms, and infections), its objective physiological impact (on lung function), its immunomodulatory effects, and its safety profile within a single analytical framework. Such an analysis would provide clinicians and guideline developers with a clearer, evidence-based perspective on the role of bacterial lysates in the management of pediatric asthma.

Therefore, we conducted this systematic review and meta-analysis of RCTs to critically appraise and synthesize the existing evidence regarding the efficacy and safety of bacterial lysates as an adjuvant therapy in children with asthma. Our primary objectives were to determine their effects on: 1) key clinical outcomes (wheezing/exacerbation frequency, time to symptom improvement, RTI frequency, overall treatment efficacy rate); 2) lung function (FEV1, FVC, PEF). Secondary objectives were to assess their impact on: 3) immunological and inflammatory biomarkers (levels of T-lymphocyte subsets, serum cytokine levels, immunoglobulin levels in serum, FeNO levels, peripheral EOS numbers, ECP levels in serum); and 4) the incidence of adverse events.

2 Materials and methods2.1 Protocol and registration

This systematic review and meta-analysis was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021). A pre-defined protocol for this review was registered, the registration number is CRD420261281796.

2.2 Eligibility criteria2.2.1 Types of studies

We included all published randomized controlled trials (RCTs) that evaluated the adjuvant use of bacterial lysates in children with asthma. RCTs were eligible regardless of their blinding status (open-label, single-blind, or double-blind) or publication language (Chinese or English). Studies with non-randomized designs (e.g., retrospective studies, cohort studies), review articles, meta-analyses, animal studies, case reports, and editorials were excluded.

2.2.2 Types of participants

The population of interest comprised children and adolescents under 18 years of age with a physician-diagnosis of bronchial asthma, based on the diagnostic criteria applied in each original study. Studies involving participants with other primary respiratory conditions (e.g., isolated allergic rhinitis, cystic fibrosis) were excluded.

2.2.3 Types of interventions

The intervention group was required to receive standard asthma therapy plus an adjunctive oral bacterial lysate. The primary bacterial lysates of interest were OM-85 (Broncho-Vaxom, also known as “FanFuShu”) and Polyvalent Mechanical Bacterial Lysates (PMBL, Ismigen). The control group received standard asthma therapy alone or an identical placebo. Standard asthma therapy was defined according to contemporary guidelines and could include, but was not limited to: inhaled corticosteroids (ICS), short or long-acting beta-2 agonists (SABA/LABA), leukotriene receptor antagonists (e.g., montelukast), systemic corticosteroids, or combinations thereof.

2.2.4 Types of outcome measures

Outcomes were categorized as primary or secondary.

Primary Outcomes focused on clinical efficacy and lung function:

Wheezing/Exacerbation Frequency: Mean number of wheezing episodes or asthma exacerbations during treatment and a specified follow-up period.

Time to Symptom Improvement: Mean duration (in days) for the improvement of cough and wheezing from treatment initiation.

Respiratory Tract Infection (RTI) Frequency: Mean number of acute RTI episodes during follow-up.

Overall Treatment Efficacy Rate: The proportion of patients categorized as “effective” or “markedly effective” based on composite clinical criteria defined in each original study (a dichotomous outcome).

Lung Function Parameters: Absolute post-treatment values of Forced Expiratory Volume in 1 s (FEV1, L), Forced Vital Capacity (FVC, L), and Peak Expiratory Flow (PEF, L/s).

Secondary Outcomes included immunological, inflammatory, and safety parameters:

Immunological Biomarkers: Peripheral blood levels of T-lymphocyte subsets (CD3+%, CD4+%, CD4+/CD8+ ratio), serum cytokines (Interferon-γ [IFN-γ], Interleukin-4 [IL-4], Interleukin-10 [IL-10]), serum eosinophil cationic protein (ECP), and peripheral eosinophil (EOS) count.

Immunoglobulin Levels: Serum levels of Immunoglobulin M (IgM) and G (IgG), salivary secretory Immunoglobulin A (sIgA), and total serum Immunoglobulin E (IgE).

Airway Inflammation Biomarker: Fractional exhaled Nitric Oxide (FeNO, ppb).

Safety: Incidence of all reported adverse events (AEs), such as gastrointestinal symptoms (abdominal pain, nausea, vomiting, diarrhea), drowsiness, dizziness, and skin rash.

2.2.5 Exclusion criteria

Studies were excluded if they met any of the following criteria: (1) published in languages other than Chinese or English; (2) duplicate publications (the most complete version was retained); (3) lacked relevant primary or secondary outcome data as defined above; (4) had a total sample size of less than 60 participants; (5) presented incomplete, missing, or non-extractable data for meta-analysis; (6) the full-text article could not be obtained.

2.3 Information sources and search strategy

A comprehensive and systematic literature search was performed across eight electronic databases from their inception to 28 February 2025: PubMed, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science Core Collection, China National Knowledge Infrastructure (CNKI), WanFang Data, Chinese Biomedical Literature Database (CBM), and VIP Database for Chinese Technical Periodicals (VIP). The search strategy employed a combination of controlled vocabulary (e.g., MeSH in PubMed, Emtree in EMBASE) and free-text keywords related to the population (“asthma”, “child”) and the intervention (“bacterial lysates”, “OM-85″, “Broncho-Vaxom”). The search syntax was adapted for each database. No filters for date or language were applied. The reference lists of all included studies and relevant review articles were manually screened to identify any additional eligible publications. An example of the full search strategy for PubMed is provided in Supplementary Material.

2.4 Study selection process

Records retrieved from all databases were imported into EndNote X9 (Clarivate Analytics) for deduplication. The study selection process was conducted independently by two reviewers. Initially, titles and abstracts were screened against the eligibility criteria. Subsequently, the full texts of potentially relevant articles were obtained and assessed in detail. Any disagreements between the two reviewers at any stage of the selection process were resolved through discussion or, if necessary, by arbitration from a third senior reviewer. The study selection process, including the number of records identified, excluded, and included, was documented using a PRISMA flow diagram.

2.5 Data extraction and management

A standardized, pre-piloted data extraction form was developed in Microsoft Excel. Two reviewers independently extracted the following data from each included study:

Study Identification and Characteristics: First author, year of publication, journal, country where the study was conducted, study design.

Participant Characteristics: Diagnostic criteria for asthma, total sample size, number of participants randomized to intervention and control groups, age (mean ± standard deviation [SD]), gender distribution.

Intervention Details: Type, dosage, and regimen of the bacterial lysate; detailed description of the concomitant standard therapy in both groups; treatment duration and follow-up period.

Outcome Data: For continuous outcomes, the post-treatment mean, SD, and sample size for both groups were extracted. For dichotomous outcomes, the number of events and the total number of participants in each group were extracted. When data were presented only graphically, WebPlotDigitizer (Version 4.6) was used to extract numerical values. For outcomes reported as median and interquartile range (IQR), estimates of mean and SD were derived using established conversion formulae (Xiang et al., 2014).

Information for Risk of Bias Assessment: Two reviewers independently assessed the methodological quality of each included study. Details pertinent to judging the risk of bias were extracted according to the Cochrane Risk of Bias tool, covering domains such as random sequence generation and allocation concealment. Any discrepancies were resolved through discussion or by a third reviewer.

Any discrepancies in extracted data were resolved by consensus after re-checking the original article. Corresponding authors were not contacted for missing data.

2.6 Data synthesis and statistical analysis

Statistical analyses were performed using Review Manager (RevMan, version 5.4, The Cochrane Collaboration) and Stata (version MP 16.0, StataCorp). A two-tailed P-value <0.05 was considered statistically significant for all tests except where otherwise specified for heterogeneity.

Measures of Treatment Effect: For dichotomous outcomes (e.g., treatment efficacy, adverse events), the pooled treatment effect was expressed as a Risk Ratio (RR) with a 95% Confidence Interval (CI). For continuous outcomes measured on the same scale (e.g., FEV1 in liters), the Mean Difference (MD) with 95% CI was calculated. For continuous outcomes measured with different scales or units (e.g., cytokine levels), the Standardized Mean Difference (SMD) with 95% CI was used.

Assessment of Heterogeneity: Statistical heterogeneity across studies was assessed using the Cochran’s Q chi-squared test (with a significance level of P < 0.10 indicating significant heterogeneity) and quantified using the I2 statistic (Higgins, 2011). I2 values of approximately 25%, 50%, and 75% were interpreted as indicating low, moderate, and high heterogeneity, respectively (Higgins Julian et al., 2003).

Data Synthesis and Model Selection: Meta-analysis was performed only when three or more studies reported the same outcome. If no substantial heterogeneity was detected (I2 < 50% and P for Q test >0.10), a fixed-effect model (Mantel-Haenszel method for dichotomous data, inverse-variance method for continuous data) was applied. Otherwise, a random-effects model (DerSimonian and Laird method) was used.

Subgroup and Sensitivity Analyses: Pre-specified subgroup analyses were conducted for the primary outcomes of wheezing frequency and time to symptom improvement to explore potential sources of heterogeneity. Subgroups were defined by: (1) use of ICS in the control regimen (Yes vs. No/Placebo), (2) follow-up duration (<12 months vs. ≥12 months), (3) participant age (≤5 years vs. >5 years), and (4) study sample size (<100 vs. ≥100). Sensitivity analysis was performed using the “leave-one-out” method, sequentially removing each study to assess its impact on the overall pooled estimate and heterogeneity.

Assessment of Reporting Biases: Publication bias was assessed using a combination of graphical and statistical methods. First, funnel plots were visually inspected for asymmetry. Then, for outcomes pooled from ten or more studies, statistical tests were performed, including Begg’s rank correlation test and Egger’s linear regression test (Jartti and Gern, 2017; Sterne et al., 2011). A P-value <0.10 in these tests was considered suggestive of statistically significant asymmetry, which could indicate the presence of publication bias or other small-study effects.

2.7 Risk of bias assessment in individual studies

The methodological quality of each included RCT was independently assessed by two reviewers using the original Cochrane Collaboration’s ‘Risk of Bias’ tool (RoB 1.0) (Higgins Julian et al., 2011). This tool evaluates six domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, and selective reporting. Each domain was judged as having a “low,” “high,” or “unclear” risk of bias. Disagreements were resolved by consensus. The overall risk of bias for each study was summarized. The results of the assessment were presented graphically using the robvis package in R software.

3 Results3.1 Study selection

A total of 686 records were initially identified through database searching: CBM (n = 89), CNKI (n = 93), VIP (n = 16), WanFang (n = 110), PubMed (n = 31), EMBASE (n = 237), Web of Science (n = 81), and the Cochrane Library (n = 29). No additional records were identified from other sources. After removing 248 duplicates, 438 records remained for title and abstract screening. Of these, 373 records were excluded for the following reasons: irrelevant research content (n = 255), review/systematic review (n = 102), non-Chinese/non-English publication (n = 10), and animal studies (n = 6). The full texts of the remaining 65 articles were assessed for eligibility. Subsequently, 37 articles were excluded due to: inability to obtain full text (n = 0), sample size <60 (n = 12), absence of relevant outcome measures (n = 8), incomplete or non-extractable data (n = 12), and non-RCT study design (n = 5). Ultimately, 28 randomized controlled trials (RCTs) were included in the qualitative and quantitative synthesis (meta-analysis). The detailed study selection process is illustrated in the PRISMA flow diagram (Figure 1).

PRISMA flow diagram of study selection process. Flow diagram illustrating the study selection process according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A total of 686 records were identified from eight databases (CBM, CNKI, VIP, WanFang, PubMed, EMBASE, Web of Science, Cochrane Library). After removing 248 duplicates, 438 records were screened by title and abstract. Of these, 373 records were excluded (irrelevant content: n = 255; review articles: n = 102; non-Chinese/non-English publications: n = 10; animal studies: n = 6). The remaining 65 full-text articles were assessed for eligibility. A further 37 articles were excluded for the following reasons: non-RCT design (n = 5); incomplete or non-extractable data (n = 12); absence of relevant outcome measures (n = 8); sample size <60 participants (n = 12). Ultimately, 28 randomized controlled trials (RCTs) were included in the qualitative synthesis and quantitative meta-analysis.

3.2 Characteristics of included studies

The 28 included RCTs were published between 2010 and 2024. The sample sizes ranged from 61 to 200 participants, with a total of 2,893 pediatric asthma patients (1,484 in the intervention groups and 1,409 in the control groups). The age of participants spanned from infancy to adolescence, with mean/median ages between 2.04 and 10.89 years across studies.

All trials employed a parallel-group, add-on design. The intervention group in all studies received standard asthma therapy (as per the control group) plus an oral bacterial lysate. The specific product was OM-85 (Broncho-Vaxom) in 27 studies (typical dose: 3.5 mg/day; one study used 7.0 mg/day), and Polyvalent Mechanical Bacterial Lysate (PMBL, Ismigen) in one study (7 mg/day). The control group received standard asthma therapy alone (n = 24) or a placebo (n = 4). Standard therapies varied and included inhaled corticosteroids (ICS) like budesonide or fluticasone, leukotriene receptor antagonists (e.g., montelukast, zafirlukast), long-acting beta-agonists (LABA) in combination, or short-acting bronchodilators.

Treatment duration varied from 14 to 180 days, with the majority of studies (n = 19) employing a 90-day intervention period. Follow-up duration for outcome assessment ranged from the end of treatment up to 12 months post-intervention.

The reported outcome measures were comprehensive, encompassing clinical efficacy (wheezing/exacerbation frequency, time to symptom improvement, respiratory infection rate, overall treatment efficacy rate), lung function (FEV1, FVC, PEF), immunological parameters (T-cell subsets, cytokines, immunoglobulins), inflammatory biomarkers (FeNO, ECP, EOS count), and adverse events. The specific outcomes reported by each study are summarized in Table 1.

Study (year)CountrySample size (T/C)Age, years (T/C)Intervention (T)Control (C)Duration (days)Reported outcomes*Lin et al. (2024)China40/4010.02 ± 1.01/9.98 ± 0.98Control therapy + OM-85 (3.5 mg/d)Montelukast + Budesonide suspension inhalation90①,②,③,④,⑧,⑨,⑩Lu et al. (2023)China42/428.39 ± 1.28/8.35 ± 1.24Control therapy + OM-85 (3.5 mg/d)Budesonide suspension + Montelukast90④,⑤,⑥,⑨,⑩Lin et al. (2023)China100/1007.92 ± 1.53/7.95 ± 1.44Control therapy + OM-85 (3.5 mg/d)Budesonide/Formoterol (80 μg + 4.5 μg)90④,⑥,⑨,⑩Shen et al. (2023)China53/538.82 ± 2.14/8.25 ± 2.05Control therapy + OM-85 (3.5 mg/d)Budesonide aerosol30④,⑤,⑨,⑩Luo et al. (2023)China45/453.49 ± 1.10/3.57 ± 1.05Control therapy + OM-85 (3.5 mg/d)Fluticasone propionate inhalation aerosol90⑥,⑦,⑨,⑩Qi et al. (2022)China55/558.55 ± 1.13/8.83 ± 1.26Control therapy + OM-85 (3.5 mg/d)Zafirlukast + Budesonide aerosol14④,⑤,⑥,⑨Wang (2022)China31/303.22 ± 1.12/3.21 ± 1.01Control therapy + OM-85 (3.5 mg/d)Budesonide suspension + Montelukast, Salbutamol + Ipratropium90②,⑤,⑥,⑨,⑩Xu and Xin (2022)China47/473.71 ± 0.32/3.85 ± 0.37Control therapy + OM-85 (7.0 mg/d)Beclomethasone dipropionate suspension21②,⑥Zhu (2021)China46/457.85 ± 0.68/8.03 ± 0.56Control therapy + OM-85 (3.5 mg/d)Budesonide suspension60④,⑨,⑩Jin and Li (2021)China45/447.84 ± 1.57/7.41 ± 1.25Control therapy + OM-85 (3.5 mg/d)Salmeterol/Fluticasone90④,⑤,⑥,⑨Zhang (2021)China50/405.53 ± 1.06/4.17 ± 1.32Control therapy + OM-85 (3.5 mg/d)Budesonide nebulization30④,⑥Yang et al. (2020a)China60/603.5 ± 0.7/3.5 ± 0.5Control therapy + OM-85 (3.5 mg/d)Fluticasone propionate aerosol/Budesonide suspension90②,⑤,⑩Yang (2020)China42/4210.18 ± 0.96/10.52 ± 1.03Control therapy + OM-85 (3.5 mg/d)Salmeterol/Fluticasone90④,⑤,⑥,⑨Gao et al. (2010)China87/863.67/3.67Control therapy + OM-85 (3.5 mg/d)Montelukast90①,②,③,⑤,⑦,⑩Zhang and Ding (2019)China44/446.73 ± 0.82/6.45 ± 0.74Control therapy + OM-85 (3.5 mg/d)Budesonide nebulization30④,⑨Tang et al. (2017)China44/437.8 ± 2.0/7.6 ± 1.9Control therapy + OM-85 (3.5 mg/d)Beclomethasone dipropionate aerosol120④,⑥,⑨Wu (2019)China49/497.43 ± 2.62/7.31 ± 2.71Control therapy + OM-85 (3.5 mg/d)Budesonide suspension90④,⑥,⑨Liu (2020)China44/446.62 ± 0.76/6.41 ± 0.63Control therapy + OM-85 (3.5 mg/d)Zafirlukast + Budesonide aerosol60⑤,⑨,⑩Zhang et al. (2014)China31/336.6 ± 2.1/6.7 ± 2.7Control therapy + OM-85 (3.5 mg/d)ICS (unspecified)90④,⑥Hu et al. (2011)China47/467.78 ± 2.29/8.04 ± 1.84Control therapy + OM-85 (3.5 mg/d)Fluticasone propionate inhalation aerosol90①,②,③,⑥,⑦,⑩Yang (2017)China43/435.30 ± 2.06/5.02 ± 1.82Control therapy + OM-85 (3.5 mg/d)Budesonide suspension21⑤,⑨Chen et al. (2015)China67/545.2 ± 2.2/5.4 ± 2.1Control therapy + OM-85 (3.5 mg/d)Budesonide suspension21⑤,⑥,⑦,⑨,⑩Yang et al. (2020b)China68/686.16 ± 2.57/6.59 ± 2.37Control therapy + OM-85 (3.5 mg/d)Salmeterol/Fluticasone90①,⑤,⑦,⑧,⑩Lu et al. (2015)China24/368.9 ± 2.8/8.7 ± 2.7Control therapy + OM-85 (3.5 mg/d)ICS (unspecified)180①,②,③,⑤,⑥,⑩Hu and Wen (2022)China66/6610.89 ± 0.31/10.23 ± 0.24Control therapy + OM-85 (3.5 mg/d)Fluticasone/Salmeterol90①,②,④,⑤,⑦,⑩Emeryk et al. (2018)Poland74/769.3/9.8Placebo + PMBL (7 mg/d)Placebo90①,②,③,⑩Han et al. (2016)China74/622.3 ± 0.6/2.2 ± 0.4ICS + OM-85 (3.5 mg/d)ICS (unspecified)90①,②,⑥Razi et al. (2010)Turkey35/4026 months (16–37)/24.5 months (14–45)Placebo + OM-85 (3.5 mg/d)Placebo90①,②,③,⑩

Characteristics of included randomized controlled trials.

T, Treatment group; C, Control group.

Outcome codes: ① Wheezing frequency; ② Time to symptom improvement (days); ③ Respiratory tract infection frequency; ④ Lung function (FEV1, FVC, PEF); ⑤ T-lymphocyte subsets (CD3+, CD4+, CD4+/CD8+); ⑥ Serum cytokines/ECP/EOS levels; ⑦ Salivary sIgA, serum IgM, IgG, IgE; ⑧ FeNO level; ⑨ Overall treatment effective rate; ⑩ Adverse events.

3.3 Risk of bias assessment

The methodological quality of the included studies, assessed using the Cochrane Risk of Bias Tool (RoB 1.0), is summarized in Figure 2. The overall risk of bias was judged as moderate to high, primarily due to inadequate reporting of methodological details.

Risk of bias assessment for the 28 included randomized controlled trials using the Cochrane Collaboration's Risk of Bias Tool (RoB 1.0). Each row represents a risk of bias domain, and each column represents an individual study. Color coding indicates the judgment: Green (Low risk) = adequate protection against bias; Yellow (Unclear risk) = insufficient information to permit judgment; Red (High risk) = significant methodological flaws likely to introduce bias. Random sequence generation: 11 studies (39.3%) low risk, 15 (53.6%) unclear, 2 (7.1%) high risk. Allocation concealment: All 28 studies (100%) unclear risk due to no reporting of allocation concealment methods. Blinding of participants and personnel: 26 studies (92.9%) unclear risk, 1 low risk (double-blind), 1 high risk. Blinding of outcome assessment: 26 studies (92.9%) unclear risk, 2 studies (7.1%) low risk. Incomplete outcome data: All 28 studies (100%) low risk. Selective reporting: All 28 studies (100%) low risk. Other bias: All 28 studies (100%) low risk. Overall, the methodological quality was judged as moderate, primarily due to inadequate reporting of allocation concealment and blinding procedures.

Random Sequence Generation: Eleven studies (39.3%) were judged as having a low risk of bias, clearly describing appropriate random sequence generation methods (e.g., random number table). For fifteen studies (53.6%), the risk was unclear due to insufficient description (e.g., stated as “randomized” without details). Two studies (7.1%) were judged as high risk.

Allocation Concealment: The risk of bias was unclear for all 28 studies (100%), as none reported the method used to conceal the allocation sequence.

Blinding of Participants and Personnel: Twenty-six studies (92.9%) were rated as having an unclear risk, as blinding of participants and healthcare providers was not mentioned or described. One study (3.6%) was low risk (double-blind), and one was high risk.

Blinding of Outcome Assessment: Similar to performance blinding, the risk was unclear for 26 studies (92.9%). Two studies (7.1%) explicitly reported blinded outcome assessment and were rated low risk.

Incomplete Outcome Data: All 28 studies (100%) were judged as low risk, as they reported complete outcome data or provided acceptable reasons for attrition with balanced groups.

Selective Reporting: All studies (100%) were judged as low risk, as all pre-specified outcomes in the methods were reported in the results sections.

Other Bias: All studies (100%) were rated as low risk for other potential sources of bias.

3.4 Results of meta analysis3.4.1 Clinical efficacy

The clinical efficacy of bacterial lysates was assessed based on the following outcome measures: wheezing/exacerbation frequency, time to symptom improvement, respiratory tract infection (RTI) frequency, and overall treatment efficacy rate.

As shown in Figure 3A, 9 RCTs involving a total of 1,035 participants (515 in the bacterial lysates group and 520 in the control group) reported data on wheezing/exacerbation frequency. The results demonstrated a significant reduction in wheezing/exacerbation frequency in the bacterial lysates group compared with the control group (Mean Difference [MD] = −3.00, 95% Confidence Interval [CI]: 4.07 to −1.93; Z = 5.50, P < 0.001). Heterogeneity was substantial (I2 = 89%, P < 0.001), indicating variability across studies in patient characteristics, follow-up duration, and background therapy.