A Meta-Analysis of the Effects of Gut Microbiota–Based Interventions on Gastrointestinal and Behavioral Symptoms in Children With Autism Spectrum Disorder

Abstract

Context

Despite an increasing body of research showing gut microbiota–based interventions can improve gastrointestinal (GI) symptoms and behavioral symptoms in both humans and animals, there are still disagreements about its impact on autism spectrum disorder (ASD) in children.

Objective

The goal of this systematic review and meta-analysis was to fully investigate the effects of gut microbiota–based interventions (eg, fecal microbiota transplantation, probiotics, prebiotics) on GI and behavioral symptoms in children with ASD.

Data Sources

The PubMed, Web of Science, the Cochrane Library, China National Knowledge Infrastructure, and Scopus databases were searched from inception to August 25, 2024.

Data Extraction

Data were extracted by 2 reviewers independently, and discrepancies in authors’ judgments were resolved by discussion or consulting a third author.

Data Analysis

The scale score of GI and behavioral symptoms before and after the intervention was extracted from included trials to evaluate the therapeutic effects of gut microbiota–based therapy in children with autism.

Results

A total of 5722 records were identified, of which 13 included in narrative synthesis and 8 studies included a meta-analysis. The nonsignificant overall effect size of gut microbiota–based intervention on GI symptoms (standardized mean difference [SMD] = –0.34 [95% CI, –0.76 to 0.07]; P = .11) and behavioral symptoms (SMD = –0.18 [95% CI, –0.37 to 0.02]; P = .08) was observed. Nevertheless, we observed a significant effect size on behavioral symptoms in the subgroup of the intervention duration (SMD = –0.26 [95% CI, –0.49 to –0.03]; P = .02).

Conclusions

In children with autism, the proof supporting the validity of gut microbiota–based intervention on GI and behavioral symptoms should be interpreted cautiously. More randomized controlled trials with rigorous methodological quality are required to precisely confirm the curative benefits of gut microbiota–based interventions on GI and behavioral symptoms in children with autism.

Systematic Review Registration

PROSPERO registration no. CRD42024583213.

INTRODUCTION

Autism is a lifelong neurologic impairment marked by difficulties with social communication and interaction, a limited variety of interests, and repeated, stereotyped behaviors.¹ Currently, there is no clear explanation for the etiology and pathophysiology of autism spectrum disorder (ASD),² which could be linked to the combined interaction of genetic and external environment variables.^3–5 According to statistics, the incidence of autism is rapidly growing, with the frequency among 8-year-old children in the United States increasing from 18.5 per 1000 in 2016 to 23 per 1000 in 2018.^6,7 Up to 2022, the prevalence of ASD in China is around 7 per 1000,⁸ with boys having a 4-fold higher prevalence than girls.⁹ Aside from the range of neurodevelopmental impairments in ASD, anxiety, depression, disrupted sleep, immunological dysfunction, and gastrointestinal (GI) disorders have all been described, which may influence the individual and family quality of life and place a significant burden on the expense of medical care.^10,11 By 2025, it is anticipated that in the United States, the yearly expense of caring for individuals with ASD will be $461 billion.¹² Although it is a serious public health issue, there are still no proven treatments for ASD-associated symptoms, such as GI (eg, diarrhea, constipation, abdominal pain) and behavioral (eg, irritability; repetitive, stereotyped behaviors; improper speech; social withdrawal) symptoms.

Conventional studies on the biology of ASD focus on the brain and aim to determine key brain circuits and regions as well as new genetic variants associated with ASD.¹³ Pharmacologic therapies that directly target the brain are used in clinical trials (eg, risperidone,¹⁴ vasopressin¹⁵). Beyond that, an expanding body of clinical research shows that behavior intervention (eg, applied behavior analysis¹⁶) and psychological intervention (eg, acceptance and commitment therapy,¹⁷ mindfulness-based interventions¹⁸) may potentially improve symptoms related to ASD in children. However, because of ASD’s complex etiology and wide range of symptoms,¹⁹ there are no standardized treatments. According to the most recent network meta-analysis of trials using pharmacological intervention, some medications can ameliorate the core symptoms of ASD to an extent. However, more high-quality trials are essential to verify the evidence for the security and benefits of drug intervention.²⁰ Although many clinical trials have been conducted to investigate the efficacy of behavior and psychological interventions, the intervention process always takes a long time, and there is no reasonable and reliable evidence to support their security and efficiency, due to the significant individual differences.²¹ As a result, the increasing frequency of autism and the paucity of efficient therapy for the alleviation of ASD-related symptoms necessitate an urgent search for new interventions.

A clinical report found 48.67 per 100 children with autism have GI symptoms, such as diarrhea, constipation, and abdominal pain.^22,23 As a result, numerous studies have begun to explore the association between gut microbiota, as important modulators of host physiology,²⁴ and the brain, based on the hypothesis of gut-brain axis. The gut-brain axis is a crucial mechanism for intercommunication across immunologic, endocrine, neural, and metabolic processes.²⁵ Researchers hypothesize that gut microbes will influence GI functions, and cognitive and behavioral manifestation by conveying information from the enteric nervous to the central nervous system.²⁶ A recent systematic review of research using probiotics, prebiotics, synbiotics, and other similar supplements reported that altering the gut-brain axis through gut microbiota–based interventions may have the potential to handle GI and behavioral symptoms in ASD. This suggests an association between the gut microbiota and the part of the brain responsible for ASD within the central nervous system.²⁷

As a result, there is great interest in the implications of gut microbiota–based intervention. When consumed in the right amount as a dietary ingredient or supplement, probiotics are live, nonpathogenic bacteria that positively influence the host’s health.²⁸ Prebiotics are described as a “substrate that grants a health advantage by being restrictively applied by host microorganisms.”²⁹ Fecal microbiota transplantation (FMT) is a procedure that conveys the GI microbiome from a typical donor to a recipient with GI symptoms, either via enema or orally in capsule form.³⁰

Despite research demonstrating that gut microbiota–based intervention can alleviate the behavioral symptoms and GI symptoms associated with ASD in children,^31–35 the evidence for efficacy to support that gut microbiota–based intervention has favorable therapeutic benefits for children with ASD is still insufficient.^36,37 Furthermore, a systematic literature review revealed a scarcity of evidence-based gut microbiota intervention for ASD in children and only 1 meta-analysis³⁸ that solely included probiotics as an intervention to explore the therapeutic effects on behavioral symptoms. This emphasizes the necessity of investigating the effectiveness of gut microbiota–based intervention targeting GI and behavioral symptoms in children with autism. In this regard, we aimed to objectively evaluate the obtainable data on the efficacy of gut microbiota–based intervention in improving GI and behavioral symptoms in children with autism.

METHODS

This study, registered in PROSPERO (CRD42024583213), was conducted under the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).

Search Strategy

Two authors (X.G. and Q.B.) independently conducted a comprehensive search of the databases of PubMed, the Cochrane Library, China National Knowledge Infrastructure, Scopus, and Web of Science from inception to August 25, 2024 for studies that investigated the association between gut microbiota–based intervention and GI and behavioral symptoms in children with autism. We searched the databases using the search terms “autistic disorder” or “autis* spectrum disorders” or “early infantile autism” or autis* or “autistic traits” or ASD, AND probiotic* or prebiotic* or synbiotic* or “fecal microbiota transplant*” or “Microbiota Transfer Therapy” or FMT or microbiota, without regard to language or status of the publication. A manual search also was carried out of the references in the included studies, the World Health Organization International Clinical Trials Registry Platform, and the US National Institutes of Health Trials Registry (ClinicalTrials.gov) (Table S1).

Inclusion and Exclusion Criteria

PICOS (population, intervention, comparator, outcomes, study design) criteria were used to select studies, as outlined in Table 1. Criteria used for inclusion were articles reporting on studies (1) of children and teenagers younger than 18 years with autism; (2) involving gut microbiota–based intervention (eg, probiotics, prebiotics, FMT) in the experiment group; (3) without limitations on comparison; (4) using appropriate measurement tools to appraise GI and behavioral symptoms; (5) that either were randomized controlled trials (RCTs) or quasi-randomized controlled trials (quasi-RCTs).

Exclusion criteria were as follows: (1) animal experiments and intervention studies; (2) unavailable or ambiguous data after contacting the authors of studies; (3) descriptive reviews or systematic reviews; (4) books, editorial articles, and letters from conferences; and (5) brief case reports.

Literature Screening and Data Extraction

After excluding duplicate records, 2 authors (X.G. and Q.B.) independently reviewed the retrieved studies based on their evaluations of the abstracts and titles. The 2 authors then went on to evaluate specific studies independently after perusing the entire text. Two authors (X.G. and N.F.), extracted data about each article’s first author and the study it reported on, including country, study design, age, sample size, intervention measures, comparison, duration, behavioral symptoms, and GI symptoms associated with ASD.

Study Quality Assessment

Two authors (X.G. and N.F.) independently evaluated the included studies using the Cochrane risk-of-bias instrument for randomized trials. The appraisal covers 7 domains: (1) random sequence generation; (2) allocation concealment; (3) participant and personnel blinding; (4) blinding of outcome assessment; (5) inadequate outcome data; (6) selective reporting; and (7) other bias.³⁹ Three categories were assigned to each bias: low risk, unknown risk, or high risk. Any disagreements among the authors’ conclusions were resolved through discussion or consulting the senior author (X.B.).

Statistical Methods

Review Manager (version 5.4; Cochrane, London, UK), which was used for all meta-analyses and subgroup analyses, was used to assess the bias risk for the RCTs and quasi-RCTs. We used Stata (version 18.0; StataCorp, College Station, TX) for meta-regression tests, publication bias, and sensitivity analyses. Because all the outcome indicators were continuous variables and multiple measurement instruments were used for the same intervention outcome indicator, the standardized mean difference (SMD) and its 95% CI were chosen as the effect size. The I² test and P value of Cochran’s Q test were used to estimate the heterogeneity among the included literature. The fixed-effects model was used for meta-analysis when P ≥ .1 and I² ≤ 50%; otherwise, the random effects model was used,⁴⁰ and the root of heterogeneity was analyzed.

Subgroup and meta-regression analyses were conducted to investigate the possible factors influencing effect size. Measurement instruments and the length of the intervention were moderators in the subgroup analysis. The country where the study was conducted, sample size, type, age, and control were the moderators in the meta-regression.

A shifting-effect model and leave-1-out sensitivity test were used to confirm the robustness and dependability of the meta-analysis results. Additionally, because the number of included articles was small (n = 8), we used Egger’s test⁴¹ rather than funnel plots to examine publication bias. P < 0.05 was deemed statistically significant.

RESULTS

Results of the Literature Search

Figure 1 shows the literature screening and selection procedure. The initial search yielded 5717 records from 5 electronic databases and 5 records through other sources. After eliminating duplicates and scrutinizing titles and abstracts, only 43 studies were read in full. Of those, 17 had incomplete text availability, and 5 had lower study quality. Eight of the studies were ineligible because of the participants’ age, intervention methods, and disease diagnosis. Of the remaining 13 studies included in the systematic review, 5 were eliminated from the final analysis because of problems with the study quality and unclear data, leaving only 8 studies in the meta-analyses.

Results of Characteristics of Included Literature

Table 2 lists the fundamental information of the included studies. Nine articles reported on RCTs,^{31,32,42–48} 2 were crossover controlled trials^49,50 with strict randomized design, and 2 were quasi-RCTs.^35,51 The systematic review included 596 children with autism, aged 1.5 to 17 years, from 6 countries (5 studies from China^{31,42–44,51}; 3 from Italy^32,48,49; 2 from the United Kingdom^45,50; and 1 each from the United States,³⁵ Switzerland,⁴⁷ and Australia⁴⁶).

All studies used gut microbiota–based interventions. Of these, 8 studies^{31,32,42–44,48–50} used probiotics as the intervention, 2 studies used prebiotics,^45,46 2 used FMT,^35,51 and 1 used a Biological response modifier.⁴⁷ All studies, except for those by Grimaldi et al,⁴⁵ Palmer et al,⁴⁶ and Hrnciarova et al,⁴⁷ took the intervention doses into account. In 10 studies, the intervention dose varied between 3 × 10⁷ colony-forming units (CFU)³¹ and 2 × 10¹⁴ CFU.⁵¹ In greater than 50% of the trials (n = 6),^{32,42–44,50,51} the intervention doses were greater than 10¹⁰ CFU. In the study by Kang et al,³⁵ the participating children were separated into 2 dose groups according to the intervention measures and timing. At the 16th and 17th interventions, the children received an intervention dose of 2.5 × 10¹² CFU, and then they received oral FMT maintenance of 2.5 × 10⁹ CFU for 8 weeks. Rectal FMT was administered with an intervention dose of 2.5 × 10¹² at the 16th intervention, and after a week, the children received an oral dose of 2.5 × 10⁹ CFU lasting 7 weeks.³⁵

Ten studies used a single gut microbiota–based intervention,^{32,35,42,44,46–51} and 2 studies added an applied behavioral analysis (ABA) intervention^31,43 simultaneously. In their study, Grimaldi et al⁴⁵ divided children into 2 experimental groups based on diet: an exclusion diet plus Bimuno galactooligosaccharide (B-GOS) mixture and an unrestricted diet plus B-GOS group. For controls, 8 studies^{32,42,44,46–50} used a placebo, 2 studies used an ABA,^31,43 2 studies used no treatment.^35,51 The study by Grimaldi et al⁴⁵ was the only 1 in which children were divided into 2 groups with an exclusion diet and an unrestricted diet group. The therapeutic durations in the analyzed studies ranged from 4 weeks to 6 months. Of these, 7 studies conducted interventions lasting ≥12 weeks,^{31,32,44,47–50} whereas the other 6 studies were all completed in ≤10 weeks.^{35,42,43,45,46,51}

The Autism Treatment Evaluation Checklist (ATEC), which has 4 subscales (sociability, sensory/cognitive awareness, health/physical/behavior, and speech/language/communication)⁵² was used in 5 studies as a measurement tool to assess treatment effects. The impacts of medications and other treatments on children with severe developmental disabilities were evaluated in 3 studies using the Aberrant Behavior Checklist (ABC),⁵³ consisting of 5 subscales: irritation, hyperactivity/resistance, stupor/social disengagement, stereotyped behavior, and incorrect speech.⁵⁴ The Social Responsiveness Scale (SRS), standardized Childhood Autism Rating Scale, complete Autism Diagnostic Observation Schedule–calibrated severity score (ADOS-CSS), and Autistic Behavior Checklist–Taiwan version (ABC-T) were used in other investigations.

Behavioral issues in children with cerebral and developmental disorders were evaluated using a 47-item ABC-T questionnaire, which is broken down into 5 subscales: sensory, relating, body and object use, language, and social and self-help.⁴² The ADOS-CSS, which is a structured, revised severity score of the ADOS, was used to measure symptoms of autism.⁵⁵ The SRS is a norm-referenced, 65-item, parent report questionnaire with 3 subscales: social motivation, social communication, and mental state. It is intended for use with children ages 4 to 18 years.⁴⁸

Eight studies assessed GI symptoms, 3 of which assessed GI symptoms using the Gastrointestinal Symptoms Rating Scale,^35,48,51 and 3 studies using the 6-item Gastrointestinal Severity Index.^32,44,46 The GI history and Gastrointestinal Severity Index were used in the remaining investigations. Using a recall time of 6 weeks, the 6-item Gastrointestinal Severity Index questionnaire measures the most frequent GI symptoms reported in children with autism, such as stomach discomfort, diarrhea, constipation, flatulence, and bloating.⁵⁶ Abdominal pain, diarrhea syndrome, indigestion syndrome, and constipation syndrome⁵⁷ are among the various GI symptoms that are evaluated by the 15-item Gastrointestinal Symptoms Rating Scale questionnaire. Abdominal pain, gaseousness, diarrhea, constipation, vomiting, and blood in vomit⁵⁰ are among the symptoms for which the GI history has 10 Likert-scale items. Data entry sheets, also known as daily diaries, were given to parents of children with autism, who were instructed to keep a daily log of their child’s overall stool count, the quantity of liquid or soft stools, the average consistency of their stools, their average smell, any flatulence, stomach pain, unexplained daytime irritability, and any nighttime awakenings.⁵⁸

Five studies—2 using FMT,^35,51 1 using probiotics,⁴⁹ and 2 using the combination of probiotics and ABA as the major intervention^31,43—reported that gut microbiota–based interventions significantly improved GI and behavioral symptoms in children with autism. Three studies only indicated slight improvement in the GI symptoms,^32,44,46 and 4 studies demonstrated that gut microbiota–based interventions only improved behavioral symptoms connected to ASD.^42,45,47,48 Nevertheless, neither GI alterations nor behavior improvement in children with autism were noted in the Emmanuel et al study.⁵⁰

Results of Quality Assessment

The RCTs and quasi-RCTs were assessed using 7 dimensions, following PRISMA standards (Table S7). All studies (n = 13) had a low risk of reporting bias and other biases; however, all but 1 had a low risk of attrition bias. However, more than half of the studies fared poorly on measurement bias. Of the total articles, only 4 reported on studies that were of excellent quality^32,46,48,50; the quality of 2 studies^35,51 was low. The reviewed literature was generally of mediocre quality (Figure S2).

Results of Meta-Analysis in Behavioral Symptoms

The meta-analysis included 8 studies comprising 350 children aged 1.5 to 16 years from 5 countries: China (n = 3), Italy (n = 2), the United Kingdom (n = 1), Australia (n = 1), and Switzerland (n = 1), all of which were published within a 5-year period. Furthermore, 204 children were placed in the comparison group and 210 children in the experiment group. The features of the 8 studies are reported in Table 3. The fixed-effects model was applied because there was minimal heterogeneity (P > .05; I² = 0%) among the included studies. The overall improvement of behavioral symptoms associated with autism did not differ significantly between the experiment and the comparison groups (pooled SMD = –0.18 [95% CI, –0.37 to 0.02]; Z = 1.77; P = .08). The pooled results are displayed in the forest plots (Figure 2A).

Results of Meta-Analysis of GI Symptoms

The meta-analysis comprised 4 studies totaling 183 children ages 1.5 to 16 years from 3 nations: China (n = 1), Italy (n = 2), and the United Kingdom (n = 1) that were published over approximately 4 years. Combined, 124 children in these studies were placed in the control group and 125 children in the intervention group. Table 4 presents the characteristics of the 4 studies. The random effects model indicated medium heterogeneity (P > .05; I² = 51%) among the included studies. When compared with the control group, no significant difference in the overall amelioration of GI symptoms related to autism was observed (pooled SMD = –0.34 [95% CI, –0.76 to 0.07]; Z = 1.61; P = .11). The pooled analysis results are displayed in Figure 3A.

Results of Subgroup Analysis

Because of the factors influencing the overall effect size, we looked at the intervention duration and the measurement instruments in a subgroup analysis. The experiment group significantly outperformed the control group in behavioral symptoms in the subgroup with intervention duration ≥12 weeks (pooled SMD = –0.26 [95% CI, –0.49 to –0.03]; P = .02) and in the subgroup with intervention duration >24 weeks. The general difference in the amelioration of GI scores almost reached statistical significance (pooled SMD = –0.38 [95% CI, –0.77 to –0.00]; P = .05). Figures 2B and 3B illustrate that, overall, no statistically significant variation was found in the amelioration of GI symptoms and behavioral disorders in children with autism between the experiment and comparison groups in the subgroup of intervention length ≤12 weeks (pooled SMD P = 0.05 [95% CI, –0.32 to 0.42]; P = .78), and ≤24 weeks (pooled SMD P = –0.80 [95% CI, –2.56 to 0.97]; P = .38).

Different groups were also established on the basis of measurement instruments. As seen in Figures 2C and 3C, no statistically significant distinction was found in the improvement of GI and behavioral symptoms associated with ASD between the comparison group and the experiment group in any of the subgroup analyses.

Results of Meta-Regression Analysis

We conducted a meta-regression analysis to identify the potential factors that might affect the intervention’s results. Results indicate that the research country (P = .894), year (P = .864), study type (P = .820), sample size (P = .724), control measure (P = .646), and other factors did not significantly affect the improvement of behavioral symptom scores associated with ASD between the experiment and the comparison groups (Figure S3). Similarly, no significant difference was discovered in the improvement of GI symptoms based on study nation (P = .253), year (P = .203), type (P = .191), or sample size (P = .267) (Figure S4).

Results of Publication Bias and Sensitivity Analysis

Instead of using funnel plots to assess publication bias, Begg and Egger tests were used, because there were fewer studies to conduct a meta-analysis. The outcome indicated a reduced risk of publication bias (P > .05). We performed sensitivity analyses on the included studies to validate the robustness and dependability of the meta-analysis findings. After adjusting the effects model and excluding any particular study, the overall effect size remained relatively unchanged. Consequently, the meta-analysis’s findings were largely reliable and robust (Figures S5 and S6).

DISCUSSION

Our meta-analysis complemented a prior one³⁸ that selected only probiotics as an intervention to investigate whether gut microbiota–based interventions might alleviate GI and behavioral symptoms in children with autism. By including more studies that used different gut microbiota–based interventions, such as FMT, probiotics, prebiotics, and biological response modifiers, we expanded the breadth of the literature review. However, only 13 relevant studies (n = 9 RCTs, 2 quasi-RCTs, and 2 crossover-controlled trials) were selected for the analysis. There were just 4 high-quality studies within the whole body of research, the rest being of moderate quality. The intervention lasted anywhere from 4 to 24 weeks, and a variety of very variable assessments were used to measure GI and behavioral symptoms associated with ASD. Thus, rigorous studies confirming the therapeutic effects of gut-microbiota intervention on GI and behavioral symptoms in children with autism remain scarce.

In the meta-analysis, the pooled SMD showed gut microbiota–based interventions did not improve the ASD-related GI and behavioral symptoms, which is consistent with findings of earlier studies.^27,59 However, some studies suggested gut microbiota–based interventions could improve GI and behavioral symptoms associated with ASD in animals and humans.^27,59,60,61 The same outcomes were also shown by subgroup analysis in measuring tools and meta-regression analysis by, for example, study nation, year, and sample size. Interestingly, however, the subgroup analysis of the intervention length revealed that ≥12 weeks’ intervention duration notably improved the related behavioral symptoms in children with autism. Additionally, among children with autism, the ≥12 weeks intervention period demonstrated a more notable amelioration in the related behavioral problems than did the < 12 weeks intervention time. Unexpectedly, the subgroup analysis of therapy duration ≥24 weeks showed a pooled SMD that almost approached statistical significance. However, the subgroup of intervention duration <24 weeks did not show a significant improvement in GI symptoms in children with autism. In children, gut microbiota–based interventions may be therapeutic for GI disorders and behavioral problems associated with ASD, according to the pooled SMD of the intervention period.

After 30 days of treatment, the intervention group’s ATEC scale scores did not substantially improve (P > .05) in the Wang et al⁴⁴ study. However, following 60 and 108 days of intervention, there was a significant (P < .05) improvement in the ATEC scale scores compared with baseline. The control group, on the other hand, showed no appreciable gains 30, 60, or 108 days after taking a placebo (P > .05). This could be explained by the fact that ASD is a neurodevelopmental condition that encompasses a wider range of symptoms and that both hereditary and environmental variables play a role in its pathogenesis. Furthermore, a longer intervention may provide a better analysis of the effects, because gut microbiota may continue to colonize the host for a long time.⁶² Thus, more investigation is required to ascertain the therapeutic effectiveness of a gut microbiota–based intervention in children with autism.

Notably, varied results may stem from discrepancies in the assessment tools and outcomes evaluated, even though in neither of the 2 subgroups identified by the assessment instruments did we discover significant improvements in the GI and behavioral symptoms connected to children with autism. For example, although other tools, like the ATEC⁶³ and ABC,⁶⁴ encompass larger constructs of physical health, impatience, and hyperactivity, the SRS focuses on features of social responsiveness and repetitive behaviors. According to the ADOS-2 outcome measure used in the Mazzone et al study,⁴⁸  Lactobacillus reuteri (now Limosilactobacillus reuteri) did not lower the severity of co-occurring behavioral disorders, restricted and repetitive activities, or mental health difficulties. However, L. reuteri administration resulted in significantly increased social functioning as measured by the SRS and improved adaptive social functioning as measured by the social adaptive composite score of Adaptive Behavior Assessment System, second edition, when compared with a placebo.⁴⁸ Therefore, when selecting assessment instruments to evaluate the impact of interventions, it will be critical in future research to focus on specific, well-defined clinical subdomains rather than overly broad global functional composite scores.

Individual differences in GI symptoms at baseline may also influence the therapeutic outcomes of the intervention, even though there were no substantial therapeutic effects on GI symptoms in children with autism. According to several studies, children with autism have different gut microbiota profiles related to abnormalities, compared with neurotypical children.⁶⁵ Moreover, the microbial composition of the gut microbiota could be changed by restoring the normal components of gut microbiota,⁶⁶ potentially delaying the onset or reducing the phenotypic manifestation of neuropsychiatric and neurologic disorders.⁶⁷ Briefly, children with autism who do not show signs of GI problems are more impressionable to gut microbiota therapy because they have less aberrant intestinal flora than children who do have GI symptoms. According to research by Santocchi et al,³² children with autism with and without GI symptoms can include 2 different groups, and probiotic treatments might have different effects, probably because they target different parts of the microbiota. Therefore, it is still important to conduct research that stratifies children with autism based on the severity of GI symptoms to examine the potential therapeutic effects of gut microbiota–based intervention on GI symptoms.

Strengths and Limitations

In addition to conducting more-thorough subgroup analyses, sensitivity analyses, and publication bias, this meta-analysis included a larger body of literature. Consequently, our findings could lead to a greater understanding of the therapeutic advantages of gut microbiota–based intervention for children with ASD. However, our study still had several potential limitations. First, because only 5 databases were searched, the number of studies included might be lower than the actual number of eligible studies. Second, an exhaustive examination of the origin of heterogeneity was not possible for the restricted quantity of studies in this meta-analysis. Third, the limited number of studies from 5 nations and the small sample sizes in each study may have reduced the validity of our results. Fourth, the reliability of the results was undoubtedly affected by the estimation of some measures based on the suggested approach of the Cochrane Handbook in cases where original data for certain studies could not be obtained. Finally, it was challenging for us to offer a reliable assessment of the safety of gut microbiota–based intervention for children with autism because most of the included studies did not reveal adverse events.

CONCLUSION

In summary, this meta-analysis did not find any evidence of statistically significant effects of gut microbiota–based intervention in improving GI and behavioral symptoms in children with autism. Because of the small sample size, shortened intervention length, use of multiple measures, and poor research quality, there is insufficient evidence to demonstrate the effectiveness of gut microbiota–based interventions on GI symptoms and behavioral disorders in children with autism. Subgroup analysis showed that longer intervention periods were therapeutically beneficial for children with autism, and studies using different assessment instruments may reveal different results of gut microbiota–based interventions in improving GI and behavioral symptoms related to ASD in children. Additionally, the therapeutic outcomes of the intervention may be affected by children who have GI symptom disparities at baseline. To evaluate the therapeutic effects of gut microbiota–based interventions on children with autism, randomized, double-blind, placebo-controlled studies must be conducted with strict adherence to trial standards. When conducting pertinent studies, researchers should also consider the diverse GI symptoms of children with autism, assessment instruments, and intervention length to produce more accurate and trustworthy evidence.

Acknowledgments

 

Author Contributions

X.B. and X.G. designed the study; X.G. and N.F. conducted the statistical studies and analysis; Q.B. and N.F. contributed to the program; X.G. and N.F. curated the data; X.G. and Q.B. analyzed the data; X.G., Q.B., and N.F. wrote the first draft of the manuscript; X.B., N.F., and X.G. modified the draft as needed; and each author reviewed and approved the version of the manuscript submitted for publication.

Supplementary Material

Supplementary Material is available at Nutrition Reviews online.

Funding

None declared.

Conflicts of Interest

None declared.

References

Author notes