Autism and the gut metabolome: Evidence for altered short-chain fatty acid profiles from a systematic review and meta-analysis

Abstract

Growing evidence implicates gut microbiome dysbiosis and its metabolites, particularly short-chain fatty acids (SCFAs), in the pathophysiology of autism spectrum disorder (ASD). However, individual study findings are inconsistent.

Objective

This systematic review and meta-analysis synthesized evidence comparing SCFA profiles between individuals with ASD and neurotypical controls.

Methods

We comprehensively searched PubMed, EMBASE, Cochrane Library, Web of Science, and CNKI databases from inception to December 1, 2024. Observational studies reporting quantitative SCFA measurements were included. Standardized mean differences (SMDs) with 95% confidence intervals (CIs) were calculated using random-effects models.

Results

Sixteen studies (473 ASD cases, 514 controls) were included. Meta-analysis revealed substantial heterogeneity across studies for most SCFAs. Despite this, levels of valeric acid (SMD = 0.76, 95% CI: 0.23 to 1.29) and hexanoic acid (SMD = 0.60, 95% CI: 0.05 to 1.16) were significantly elevated in individuals with ASD. Isobutyric acid also showed a positive association (SMD = 0.21, 95% CI: 0.03 to 0.46) with lower heterogeneity. No significant overall differences were found for acetic, propionic, or butyric acids, but subgroup analyses indicated crucial variations based on sample source (e.g., fecal, plasma, urine).

Conclusion

This meta-analysis provides evidence for altered SCFA profiles in ASD, specifically elevated valeric and hexanoic acids and a consistent signal for isobutyric acid, suggesting gut microbial dysbiosis involving distinct metabolic pathways. The significant heterogeneity and sample-source-dependent effects highlight the complexity of the gut-brain axis in ASD and underscore the need for future research with standardized protocols and longitudinal designs to clarify the role of SCFAs.

1. Introduction

Autism Spectrum Disorder (ASD) represents a complex neurodevelopmental condition characterized by persistent deficits in social communication and interaction, accompanied by restricted, repetitive patterns of behavior, interests, or activities (Lai et al., 2014; American Psychiatric Association, 2013). Epidemiological studies demonstrate a consistent upward trajectory in global ASD prevalence, primarily attributed to enhanced diagnostic criteria refinement and heightened clinical awareness (Maenner et al., 2020). This epidemiological trend presents substantial implications for healthcare systems and socioeconomic infrastructure, necessitating comprehensive investigation of underlying pathophysiological mechanisms (Buescher et al., 2014).

Contemporary research has established genetic factors as fundamental components in ASD pathogenesis (Sanders et al., 2015). However, emerging evidence demonstrates that environmental factors and their interactions with genetic predisposition significantly modulate disorder development and progression (Modabbernia et al., 2017). Within this context, alterations in the gut microbiome have emerged as a critical area of investigation (Vuong and Hsiao, 2017). The gut-brain axis facilitates bidirectional communication between the gastrointestinal system and central nervous system through multiple pathways, including immune, endocrine, and neural mechanisms, thereby influencing neurodevelopmental trajectories and behavioral manifestations (Cryan et al., 2019).

Short-chain fatty acid (SCFA), primarily comprising acetate, propionate, and butyrate, represent key metabolites produced through microbial fermentation of dietary fiber. These compounds serve essential physiological functions, including maintenance of intestinal barrier integrity, immunomodulation, and cellular energy metabolism (Koh et al., 2016). Accumulating evidence suggests that SCFA dysregulation may be implicated in ASD pathophysiology, particularly given the well-documented alterations in gut microbiota composition observed in individuals with ASD (Xu et al., 2019; Fattorusso et al., 2019a).

SCFA demonstrate significant influence on various aspects of brain function and neurodevelopment through multiple mechanistic pathways. Research indicates that butyrate exhibits neuroprotective properties through enhancement of neuroplasticity and cognitive function (Val-Laillet et al., 2018). Propionate demonstrates both beneficial and potentially deleterious effects on brain function, with outcomes dependent upon concentration and contextual factors (MacFabe, 2015). Additionally, acetate modulates appetite regulation and energy homeostasis mechanisms, potentially exerting indirect effects on neural function through metabolic pathways (Frost et al., 2014). These findings underscore the potential significance of SCFA metabolism in neurodevelopmental disorders, particularly ASD.

Despite extensive investigation of SCFA concentrations in ASD populations compared to neurotypical controls, current findings exhibit substantial heterogeneity and inconsistency. While certain studies report decreased butyrate concentrations in ASD individuals (Wang et al., 2012a), others demonstrate elevated SCFA levels (De Angelis et al., 2013) or fail to identify significant alterations (Gondalia et al., 2012). These discrepancies likely arise from methodological variations, limited sample sizes, and population heterogeneity. The absence of consensus presents significant barriers to understanding SCFA involvement in ASD pathogenesis and impedes development of microbiome-based therapeutic strategies.

To address these knowledge gaps, we conducted a systematic review and meta-analysis to synthesize the available evidence regarding SCFA levels in individuals with ASD compared to neurotypical controls. This investigation was designed with the following objectives: (a) To identify overall trends in SCFA alterations in ASD, with specific focus on quantifying differences in individual SCFAs (acetate, propionate, butyrate, valeric acid, and hexanoic acid) between ASD and control groups. (b) To examine potential factors contributing to heterogeneity in reported findings through subgroup analyses based on sample source, age groups, and geographical location. Our findings aim to enhance understanding of SCFA profiles in ASD and identify potential directions for future research in this rapidly evolving field.

2. Materials and methods

2.1. Search strategy and study selection

This systematic review and meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and was prospectively registered with PROSPERO (ID: CRD42024515666). The review question was formulated based on the PICO framework: Population (individuals with a formal ASD diagnosis), Intervention/Exposure (not applicable for observational synthesis), Comparison (neurotypical controls), and Outcome (quantitative measurements of SCFAs in biological samples).

A comprehensive literature search was conducted across multiple electronic databases: PubMed, EMBASE, Cochrane Library, Web of Science, and CNKI, from their inception through December 1, 2024. The search strategy incorporated validated combinations of Medical Subject Headings (MeSH) terms and keywords related to ASD and SCFAs (Supplementary File S1).

Eligibility criteria were as follows: Inclusion Criteria: ∗(1) Observational studies (cross-sectional or case-control) or controlled trials; (2) Participants of any age with a standardized diagnosis of ASD based on established tools (e.g., DSM-5, ADOS-2, ADI-R) (3) Quantitative measurement of at least one SCFA (e.g., acetate, propionate, butyrate, valerate, hexanoate); in human biological samples (feces, plasma, urine); (4) Inclusion of a neurotypical control group without ASD or major neurological/psychiatric conditions; (5) Original research published in English or Chinese with extractable data (means, standard deviations).∗ Exclusion Criteria: Case reports, reviews, meta-analyses, conference abstracts; studies on animal models; studies lacking a control group or quantitative SCFA data.

Two investigators (Z.Y. and X.C.) independently screened titles and abstracts of all retrieved records. The full text of potentially eligible articles was then assessed against the inclusion criteria. Discrepancies were resolved through discussion or by consultation with a third reviewer (X.H.). The reference lists of included studies and relevant reviews were manually examined to identify additional eligible studies.

Our initial search yielded 2805 records. After removing 477 duplicates, 2328 unique records were screened at the title/abstract level, of which 2272 were excluded (reasons included: animal study, wrong publication type, irrelevant outcome, or no ASD population). The remaining 56 full-text articles were assessed for eligibility, resulting in the inclusion of 16 studies in the final quantitative synthesis. A detailed flow diagram of the study selection process is presented in Fig. 1.

Fig. 1.

Flow chart of literature inclusion for PRISMA meta-analysis and systematic review.

Regarding ethical approvals and participant recruitment: As a systematic review of published data, this study did not involve direct interaction with human subjects or primary data collection. Therefore, ethical approval was not required. All participant recruitment methods, informed consent procedures, and ethical approvals were the responsibility of the original study authors, and we have extracted and reported these details where provided (Table 1).

Table 1.

Study characteristics.

Study	Country	Sample size		ASD		Control		Diagnostic criteria(ASD)	SCFAs Sample	Detection method	Confounding factors
		ASD	Control	Age (years)	Gender (M/F)	Age (years)	Gender (M/F)				Dietary Assessment	Probiotics andAntibiotics Usage Assessment
Afaf K El-Ansary (2011) (El-Ansary et al., 2011)	Saudi Arabia	26	26	4-12	NA	4-11	NA	ADI-R,ADOS,3DI	Plasma	GC	Yes	No
Dae-Wook Kang (2018) (Kang et al., 2018a)	USA	23	21	4-17	22:1	4-17	2.5:1	ATEC,PDD-BI	Feces	GC	Yes	Yes
James B Adams (2011) (Adams et al., 2011)	USA	58	39	2.5-18	6.25:1	2.5-18	0.86:1	Diagnosis by a psychiatrist or professional	Feces	GC	Yes	No
Jianquan He (2023) (He et al., 2023)	China	40	40	3-10	3:1	3-10	2:1	DSM-V,CARS,ABC,ICD-10	Feces	GC-MS	Yes	Yes
Jue Wang (2019) (Wang et al., 2019)	China	45	90	6-9	8:1	6-9	8:1	DSM-5,CARS	Feces	HPLC	Yes	No
Katarina Bojović (2020) (Bojović et al., 2020)	Serbia	36	28	2-9	3.5:1	2-11	2.1:1	ICD-10	Feces	HPLC	Yes	Yes
Kirsten Berding (2018) (Berding and Donovan, 2018)	USA	26	32	2-7	2.7:1	2-7	1.46:1	PDDBI-SV	Feces	GC	Yes	Yes
Kirsten Berding (2020) (Berding and Donovan, 2020)	USA	26	32	2-7	NA	2-7	NA	PDDBI-SV	Feces	GC	Yes	Yes
Lorena Coretti (2018) (Coretti et al., 2018)	Italy	11	1	2-4	4.5:1	2-4	1.3:1	DSM-5,ADOS 2,ADI-R	Feces	GC	Yes	No
Lv Wang (2020) (Wang et al., 2012b)	Australia	23	31	3-17.3	10.5:1	3.5-18.4	1:1	CARS,DSM-5	Feces	GC	Yes	Yes
Renshan Zhao (2017) (Zhao et al., 2017)	China	10	10	3-9	4:1	5-6	4:1	NA	Urine	GC-MS	No	No
Simeng Liu (2019) (Liu et al., 2019)	China	30	20	2.5-18	5:1	2.5-18	4:1	DSM-V,ICD-10	Feces	HPLC	Yes	Yes
Sungji Ha (2021) (Ha et al., 2021)	Korea	54	38	4-13	3.9:1	4-9	0.9:1	DSM-5,ADOS-2,ADI-R,SRS	Feces	HPLC	No	Yes
Wenlin Deng (2022) (Deng et al., 2022)	China	45	45	1-10	6.5:1	4-8	0.88:1	DSM-5	Feces	GC-MS	Yes	Yes
Ying Wang (2020) (Wang et al., 2020)	China	26	24	2-8	12:1	2-8	11:1	DSM-5,CCMD-3	Feces	GC-MS	Yes	Yes
Yu Zhang (2022) (Zhang et al., 2022)	China	25	24	3-12	2.57:1	2-14	1.4:1	DSM-5	Feces	GC	No	Yes

Note: 3DI: Developmental, dimensional, diagnostic interview; ICD-10: International Classification of Diseases, Tenth Revision; PDDBI-SV: Pervasive Developmental Disorder Behavior Inventory; DSM-5: Diagnostic and Statistical Manual of Mental Disorders, fifth edition; ADOS-2: Autism Diagnostic Observation Scale; ADI-R: Revised Autism Diagnostic Interview; SRS: Social Response Scale; CCMD-3: Chinese Classification and Diagnostic Criteria for Mental Disorders; ATEC: Autism Treatment Evaluation Checklist; PDD-BI: Pervasive Developmental Disorder Behavior Inventory; MSEL: Mullen Early Learning Scale; CARS: Childhood Autism Rating Scale; ABC: Autism Behavior Checklist; HPLC; High performance liquid chromatography; GC:Gas chromatography; GC-MS:Gas chromatography mass spectrometry; The specific SCFAs quantified in each study are listed in detail in Supplementary File S1.

2.2. Data extraction

Data extraction was performed independently by two investigators using a standardized electronic form. Extracted information encompassed: (1) study characteristics (author, year, country, design); (2) participant details (sample size, age, gender, diagnostic criteria — commonly including standardized instruments such as the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) for clinical diagnosis, the Autism Diagnostic Observation Schedule, 2nd edition (ADOS-2) for behavioral observation, and the Autism Diagnostic Interview-Revised (ADI-R) for caregiver interview); (3) SCFA assessment methods (sample source, types of SCFAs measured, analytical techniques); and (4) outcome data (mean SCFA concentrations and standard deviations). In cases where additional information was required or clarification needed, corresponding authors were contacted.

2.3. Quality assessment

The methodological quality of included studies underwent rigorous assessment using the Newcastle-Ottawa Scale (NOS), incorporating eight specific domains across three fundamental categories: selection, comparability, and outcome. Studies received quality classifications based on their cumulative NOS scores: high quality (≥7 points), moderate quality (4-6 points), or low quality (≤3 points). Two independent reviewers conducted quality assessments, with disagreements resolved through third-party arbitration.This assessment ensures that the studies included in the meta-analysis are methodologically sound and that the findings are robust.

2.4. Statistical analysis

The primary outcome for this meta-analysis was the standardized mean difference (SMD) in SCFA levels between ASD and control groups. For all key associations reported (i.e., for each individual SCFA), the pooled SMD is presented alongside its 95% confidence interval (CI).

To facilitate the interpretation of our meta-analytic findings for a broad readership, we provide a brief explanation of key statistical terms used. The effect size is expressed as the standardized mean difference (SMD), which quantifies the difference between groups in standard deviation units. Each SMD is reported with a 95% confidence interval (CI), indicating the precision of the estimate; a CI that does not cross zero suggests a statistically significant effect. Heterogeneity, or the degree of variability in effect sizes across studies, was assessed using the I² statistic, which describes the percentage of total variation that is due to between-study differences rather than chance. An I² value greater than 50% was considered to indicate substantial heterogeneity. Effect sizes were calculated as standardized mean differences (SMDs) with corresponding 95% confidence intervals (CIs) using random-effects models to account for anticipated heterogeneity. The I² statistic and Cochran's Q test quantified between-study heterogeneity, with I² >50% or p < 0.10 indicating substantial heterogeneity. Potential confounding was addressed through two strategies. First, we extracted data on whether individual studies adjusted for key confounders such as age, diet, or medication use (see Table 1) and considered this in the interpretation of findings. Second, we performed pre-specified subgroup analyses and meta-regression to investigate potential sources of heterogeneity that may reflect confounding factors at the study level (e.g., sample source, geographical location, dietary control). Subgroup analyses investigated potential sources of heterogeneity across: Geographical regions, Sample matrices (fecal, plasma, urine), Age stratifications, Analytical methodologies, Study quality metrics. Specifically, for the subgroup analyses of ‘Dietary control’ and ‘Probiotic and antibiotic control,’ studies were categorized dichotomously (Yes/No) based on whether the original publications explicitly reported implementing such control measures (e.g., using dietary records, or excluding participants who had recently used probiotics or antibiotics). This classification was performed during data extraction to investigate the potential influence of these methodological factors on SCFA levels and study heterogeneity. Meta-regression analyses examined the influence of continuous variables including mean age and quality assessment scores on SCFA outcomes. Sensitivity analyses evaluated result robustness through sequential study exclusion. Statistical analyses were conducted using RevMan 5.4 and Stata 16.0, with statistical significance set at p < 0.05 (two-tailed).

The sample size for this meta-analysis is determined by the total number of participants across all included studies. A key strength of meta-analysis is its ability to increase statistical power by pooling data from multiple smaller studies. Our comprehensive search strategy was designed to identify all available evidence to maximize the cumulative sample size and thus the precision and reliability of the pooled effect estimates. Formal power calculation for a meta-analysis is atypical, as the goal is to synthesize all existing data rather than to test a hypothesis with a pre-specified sample size.

3. Results

3.1. Literature search

A total of 2805 potentially relevant records were initially retrieved through a comprehensive search of electronic databases. After eliminating 477 duplicate entries, 2328 unique articles remained for preliminary evaluation. During the screening phase, 2272 studies were excluded based on established eligibility criteria: 330 were conference abstracts, case reports, or commentaries; 264 were review papers; 963 involved animal models; and 715 lacked either SCFA-related outcome measures or a valid ASD classification. Subsequently, 56 full-text articles underwent comprehensive evaluation against the inclusion criteria. Following detailed assessment, 16 studies met all eligibility requirements and were included in the final meta-analysis (El-Ansary et al., 2011; Kang et al., 2018a; Adams et al., 2011; He et al., 2023; Wang et al., 2012b, 2019, 2020; Bojović et al., 2020; Berding and Donovan, 2018, 2020; Coretti et al., 2018; Zhao et al., 2017; Liu et al., 2019; Ha et al., 2021; Deng et al., 2022; Zhang et al., 2022). The selected studies employed standardized diagnostic criteria for ASD classification and reported quantitative measurements of at least one SCFA. The systematic article selection process adhered to PRISMA guidelines, as illustrated in Fig. 1.

3.2. Characteristics of the included studies

A comprehensive analysis of the included studies revealed systematic patterns in methodology, participant demographics, and analytical approaches for SCFA assessment. The meta-analysis incorporated 16 studies conducted between 2011 and 2023, encompassing a total study population of 473 individuals with ASD and 514 age-matched neurotypical controls. Individual study sample sizes ranged from 6 to 58 participants. The geographical distribution of research sites comprised seven studies from China, four from the United States, and individual studies from Italy, Australia, Serbia, Saudi Arabia, and South Korea. Study participants' ages spanned from 1.75 to 18.4 years, with a predominant male-to-female ratio across all cohorts, consistent with established ASD prevalence patterns.

The methodological framework predominantly consisted of observational studies, including 11 cross-sectional and 4 case-control investigations. ASD diagnoses were established using standardized diagnostic instruments, primarily DSM-V, ADOS-2, and ADI-R protocols. Biological sample collection focused predominantly on fecal specimens, with one study specifically examining plasma matrices. The analytical scope encompassed comprehensive SCFA profiling, with acetate, propionate, and butyrate quantification reported consistently across studies. Subset analyses encompassed measurements of isobutyrate and isovalerate, whereas assessments of caproate were reported in a more limited subset of studies.

Gas chromatography (GC) was identified as the primary analytical technique, utilized in over half of the studies reviewed. In contrast, gas chromatography–mass spectrometry (GC-MS) was adopted in four investigations, offering superior sensitivity at the cost of higher operational complexity and resource demands. Approximately two-thirds of the studies incorporated adjustments for potential confounders, such as dietary habits and probiotic supplementation, recognizing their impact on microbial composition and subsequent SCFA biosynthesis.

Geographical analysis highlighted methodological divergence: studies conducted in China and the United States predominantly relied on GC-based approaches, whereas GC-MS was more commonly employed in smaller-scale investigations conducted in other regions. These differences likely reflect both regional disparities in analytical infrastructure and varying methodological preferences.

A comprehensive overview of study features—including sample sizes, age and sex distributions, diagnostic frameworks, SCFA profiling techniques, and analytical methodologies—is systematically presented in Table 1.

3.3. Quality assessment and risk of bias in included studies

The methodological quality evaluation, conducted using the NOS, revealed that among the 16 studies included in this meta-analysis, 14 demonstrated high methodological quality (NOS scores ≥7), while 2 exhibited moderate quality (scores between 4 and 6) (Table 2). No investigations were classified as low quality (scores ≤3).

Table 2.

Quality assessment of the included studies.

Author (Year)	Selection (0–4)	Comparability (0–2)	Outcome (0–3)	Total Score (0–9)	Quality Level
Afaf K El-Ansary(2011)	4	0	2	6	Middle
Dae-Wook Kang(2018)	4	1	3	8	High
James B Adams(2011)	4	1	2	7	High
Jianquan He(2023)	4	1	2	7	High
Jue Wang(2019)	4	1	2	7	High
Katarina Bojović(2020)	4	1	2	7	High
Kirsten Berding(2018)	4	2	3	9	High
Kirsten Berding(2020)	4	2	2	8	High
Lorena Coretti(2018)	4	2	3	9	High
Lv Wang(2020)	3	1	2	6	Middle
Renshan Zhao(2017)	4	2	2	8	High
Simeng Liu(2019)	4	1	2	7	High
Sungji Ha(2021)	4	1	2	7	High
Wenlin Deng(2022)	4	1	2	7	High
Ying Wang(2020)	4	1	3	8	High
Yu Zhang(2022)	4	1	2	7	High

The majority of investigations demonstrated robust case selection methodology, implementing clearly delineated inclusion criteria and standardized diagnostic instruments, including DSM-5, ADOS-2, and ADI-R. However, potential selection bias was identified in several studies where sample representativeness was inadequately documented, particularly regarding control group recruitment protocols. Comparability between ASD and control cohorts was appropriately addressed in most investigations, with statistical adjustments for confounding variables encompassing age, sex, and dietary factors. Nevertheless, a subset of studies demonstrated insufficient control for critical confounding variables, notably the influence of probiotic or antibiotic administration, potentially impacting observed SCFA concentrations.

Measurement bias was predominantly minimal, with most investigations employing validated analytical methodologies, including gas chromatography (GC) or high-performance liquid chromatography (HPLC) for SCFA quantification. However, several studies exhibited inadequate documentation regarding the consistency of measurement protocols across case and control groups, potentially introducing methodological variability.

Response rates and attrition data demonstrated inconsistent reporting patterns, with several studies omitting comprehensive follow-up documentation, potentially contributing to attrition bias.

3.4. Acetic acid

The meta-analysis for acetic acid included 15 studies comparing individuals with ASD to neurotypical controls. The pooled standardized mean difference (SMD) was 0.28 (95% CI: −0.15 to 0.70, p = 0.21), with substantial heterogeneity (I² = 90%, p < 0.00001) (Fig. 2a). Subgroup analyses revealed no significant differences based on geographical location, detection methods, or age groups, with high heterogeneity persisting in all subgroups(Table 3, Supplementary File S2). However, Subgroup analysis demonstrates that dietary control emerged as a significant contributor to heterogeneity. Studies without dietary control reported a significant association (SMD = 1.12, 95% CI: 0.78 to 1.45, p < 0.00001, I² = 0%), while studies with dietary control showed no significant effect (SMD = 0.09, p = 0.70, I² = 89%) (Fig. 2b). These findings highlight the importance of dietary factors in influencing acetic acid levels and suggest that methodological variability, particularly in dietary control, may contribute to the observed heterogeneity.

Fig. 2.

Acetic acid levels in ASD: meta-analysis and dietary control subgroup. (a) Pooled analysis comparing acetic acid levels between individuals with ASD and controls.; (b) Subgroup analysis stratified by dietary control, evaluating its potential moderating effect on acetic acid levels in ASD.

Table 3.

Summary table of the subgroup analysis results for acetic acid.

Subgroup	Category	SMD (95% CI)	I2	P
Overall	-	0.28 (−0.15 to 0.70)	90%	0.21
Geographical location	Asian	0.46 (−0.10 to 1.02)	90%	0.31
	Western	−0.00 (−0.69 to 0.68)	89%
Detection method	GC	0.52 (−0.25 to 1.30)	93%	0.62
	GC-MS	0.07 (−0.60 to 0.74)	83%
	HPLC	0.04 (−0.83 to 0.92)	91%
Age group	Young children (<12 years)	0.14 (−0.31 to 0.59)	89%	0.31
	Pre-adolescents (12-18 years)	0.85 (−0.45 to 2.14)	93%
Study quality	High (NOS ≥7)	0.11 (−0.31 to 0.54)	88%	0.14
	Moddle (NOS 4-6)	1.37 (0.24 to 2.99)	92%
Sample source	Fecal	0.10 (−0.31 to 0.51)	88%	<0.00001∗
	Plasma	2.22 (1.51 to 2.92)	N/A
	Urine	0.91 (0.02 to 1.84)	N/A
Dietary control	Yes	0.22 (−0.20 to 0.63)	84%	0.67
	No	0.51 (−0.78 to 1.81)	96%
Probiotic and antibiotic control	Yes	0.09 (−0.37 to 0.55)	89%	0.0004∗
	No	1.12 (0.78 to 1.45)	0%

Note: SMD: Standardized Mean Difference; CI: Confidence Interval; GC: Gas Chromatography; GC-MS: Gas Chromatography-Mass Spectrometry; HPLC: High-Performance Liquid Chromatography; NOS: Newcastle-Ottawa Scale. Statistically significant (p < 0.05). (∗) denote statistically significant results.

3.5. Propionic acid

The meta-analysis for propionic acid, including 16 studies, yielded no significant overall difference between ASD and control groups (SMD = 0.02, 95% CI: −0.30 to 0.35, p = 0.89), with substantial heterogeneity (I² = 83%, p < 0.00001) (Fig. 3a). Subgroup analyses revealed no significant differences based on geographical location, detection methods, or age groups(Table 4). However, sample source emerged as a significant factor (p < 0.00001). Plasma samples showed significantly lower levels in ASD (SMD = −2.20, 95% CI: −2.90 to −1.50, p < 0.00001), while urine samples indicated higher levels (SMD = 1.13, 95% CI: 0.17 to 2.09, p = 0.02) (Fig. 3b). However, this non-significant pooled result masked profound and opposing differences based on biological matrix. Subgroup analysis by sample source revealed a dramatic dichotomy: propionic acid levels were significantly lower in plasma (SMD = −2.20, 95% CI: −2.90 to −1.50) but significantly higher in urine (SMD = 1.13, 95% CI: 0.17 to 2.09, p = 0.02) of individuals with ASD compared to controls (p for subgroup difference <0.00001) (Fig. 3b). This pattern suggests complex alterations in the systemic handling of propionic acid in ASD beyond its gut microbial production.

Fig. 3.

Meta-analysis of propionic acid levels in ASD and subgroup analyses. (a) Pooled analysis comparing propionic acid levels between individuals with ASD and controls.; (b) Subgroup analysis examining the influence of sample source (e.g., feces, plasma, urine) on propionic acid levels in ASD.; (c) Subgroup analysis investigating the effect of dietary control on propionic acid levels in ASD.

Table 4.

Summary table of the subgroup analysis results for Propionic acid.

Subgroup	Category	SMD (95% CI)	I2	P
Overall	-	−0.02 (−0.30 to 0.35)	83%	0.89
Geographical location	Asian	0.02 (−0.50 to 0.53)	89%	0.99
	Western	0.02 (−0.34 to 0.38)	62%
Detection method	GC	−0.09 (−0.70 to 0.53)	89%	0.79
	GC-MS	0.20 (−0.41 to 0.82)	80%
	HPLC	0.11 (−0.19 to 0.41)	31%
Age group	Young children (<12 years)	0.14 (−0.14 to 0.42)	72%	0.40
	Pre-adolescents (12-18 years)	−0.56 (−2.13 to 1.02)	95%
Study quality	High (NOS ≥7)	0.12 (−0.15 to 0.38)	71%	0.54
	Moddle (NOS 4-6)	−0.77 (−3.56 to 2.02)	97%
Sample source	Fecal	0.11 (−0.14 to 0.36)	70%	<0.00001∗
	Plasma	−2.20 (−2.90 to −1.50)	N/A
	Urine	1.13 (0.17 to 2.09)	N/A
Dietary control	Yes	−0.15 (−0.50 to 0.20)	82%	0.008∗
	No	0.79 (0.19 to 1.39)	63%
Probiotic and antibiotic control	Yes	0.17 (−0.10 to 0.45)	65%	0.34
	No	−0.29 (−1.19 to 0.61)	92%

Note: SMD: Standardized Mean Difference; CI: Confidence Interval; GC: Gas Chromatography; GC-MS: Gas Chromatography-Mass Spectrometry; HPLC: High-Performance Liquid Chromatography; NOS: Newcastle-Ottawa Scale. Statistically significant (p < 0.05). (∗) denote statistically significant results.

Dietary control also significantly influenced the results (p = 0.008), with studies lacking dietary control reporting higher levels in ASD (SMD = 0.79, 95% CI: 0.19 to 1.39, p = 0.009) (Fig. 3c). These results emphasize the importance of both sample source and dietary control in interpreting propionic acid levels, highlighting the need for rigorous methodological standardization in future research.

3.6. Butyric acid

Meta-analysis of butyric acid levels across 16 studies (473 ASD cases and 514 controls) revealed no significant overall difference between ASD and control groups (SMD = 0.00, 95% CI: −0.28 to 0.28, p = 0.99) (Fig. 4a). However, substantial heterogeneity was observed (I² = 77%, p < 0.00001), warranting further investigation through comprehensive subgroup analyses.

Fig. 4.

Meta-analysis of butyric acid levels in ASD and subgroup analysis. (a) Pooled analysis comparing butyric acid levels between individuals with ASD and controls. (b) Subgroup analysis examining the impact of sample source (feces, plasma, urine) on butyric acid levels in ASD.

Stratification by sample source revealed distinct patterns: fecal samples showed minimal difference (SMD = −0.02, 95% CI: −0.30 to 0.26, I² = 75%), while plasma samples indicated significantly lower butyric acid levels in ASD participants (SMD = −0.62, 95% CI: −1.18 to −0.06, p = 0.03). Notably, urinary butyric acid levels were markedly elevated in ASD (SMD = 1.45, 95% CI: 0.44 to 2.46, p = 0.005), suggesting potential differences in butyric acid metabolism or excretion patterns(Fig. 4b).

Geographic analysis revealed comparable effects between Asian (SMD = −0.11, 95% CI: −0.54 to 0.32, I² = 84%) and Western populations (SMD = 0.15, 95% CI: −0.18 to 0.48, I² = 55%). Age-stratified analysis showed no significant differences between young children (SMD = 0.03, 95% CI: −0.29 to 0.34, I² = 78%) and pre-adolescents (SMD = −0.10, 95% CI: −0.81 to 0.62, I² = 79%) (Supplementary File S2).

Analytical method comparison demonstrated consistency across platforms: GC (SMD = −0.07, 95% CI: −0.50 to 0.36, I² = 79%), GC-MS (SMD = 0.22, 95% CI: −0.42 to 0.86, I² = 81%), and HPLC (SMD = −0.05, 95% CI: −0.61 to 0.51, I² = 78%). Study quality assessment revealed similar results between high-quality (SMD = 0.00, 95% CI: −0.29 to 0.30, I² = 77%) and moderate-quality studies (SMD = −0.02, 95% CI: −1.19 to 1.16, I² = 89%) (Supplementary File S2).

Importantly, sensitivity analyses demonstrated the robustness of these findings, with no single study significantly altering the overall effect estimate. The discordant findings across different sample types suggest complex alterations in butyric acid metabolism in ASD, potentially involving altered gut absorption, hepatic metabolism, or renal handling of this important SCFA.

3.7. Isobutyric acid

Meta-analysis of isobutyric acid levels encompassed 8 studies with 221 ASD participants and 283 controls, revealing a significant positive association between ASD and isobutyric acid levels (SMD = 0.21, 95% CI: 0.03 to 0.46, p = 0.09). Notably, the analysis demonstrated moderate heterogeneity (I² = 44%, p = 0.09), substantially lower than observed for other SCFAs in this meta-analysis(Fig. 5a).

Fig. 5.

Meta-analysis and sensitivity analysis of isobutyric acid levels in ASD. (a) Pooled analysis comparing isobutyric acid levels between individuals with ASD and controls. (b) Sensitivity analysis assessing the robustness of the results for isobutyric acid levels in ASD.

Sensitivity analysis identified one influential study (Wang et al., 2019) that impacted the overall effect estimate. Upon its exclusion, the association strengthened (SMD = 0.29, 95% CI: 0.08 to 0.50) with remarkably low heterogeneity (I² = 0%) (Fig. 5b). This finding suggests a robust relationship between isobutyric acid levels and ASD status when accounting for study variability.

The observed pattern differs markedly from other SCFAs analyzed in this study, particularly in two aspects: (a)The consistently lower heterogeneity (I² = 44% vs. >70% for other SCFAs). (b)The stable positive association across most included studies.

Individual studies showed varying effect sizes, ranging from small (SMD = 0.03) (He et al., 2023) to large (SMD = 1.14) (Zhao et al., 2017), with most studies reporting modest positive associations. This consistency in direction, despite varying magnitudes, reinforces the reliability of the observed relationship.

These findings suggest that isobutyric acid metabolism might represent a more specific marker of gut microbiota alterations in ASD compared to other SCFAs, warranting further investigation into its potential role in ASD pathophysiology and its utility as a potential biomarker.

3.8. Valeric acid

Our meta-analysis examining valeric acid levels encompassed 12 studies with 380 ASD participants and 413 controls. The findings revealed significantly elevated valeric acid levels in individuals with ASD compared to controls (SMD = 0.76, 95% CI: 0.23 to 1.29, p = 0.005) (Fig. 6). The substantial heterogeneity observed (I² = 92%, p < 0.00001) prompted further investigation through detailed subgroup analyses to understand the underlying patterns.

Fig. 6.

Pooled analysis comparing valeric acid levels between individuals with ASD and controls.

Geographic analysis uncovered intriguing regional variations in valeric acid profiles. Studies from Asian populations demonstrated more pronounced elevations in valeric acid levels (SMD = 1.07, 95% CI: 0.32 to 1.83) compared to Western populations (SMD = 0.19, 95% CI: −0.46 to 0.84) (Table 5 and Supplementary File S2). This regional disparity might reflect differences in dietary patterns, environmental factors, or genetic variations influencing valeric acid metabolism (see Table 6).

Table 5.

Summary table of the subgroup analysis results for Valeric acid.

Subgroup	Category	SMD (95% CI)	I2	P
Geographical location	Asian	1.07 (0.32, 1.83)	93%	0.005
	Western	0.19 (−0.46, 0.84)	85%	0.56
Detection method	GC	0.87 (0.03, 1.70)	94%	0.06
	GC-MS	0.30 (−0.18, 0.79)	68%	0.22
	HPLC	1.24 (−1.54, 4.03)	98%	0.38
Age group	Pre-Adolescents (<12 years)	1.54 (−0.29, 3.37)	94%	0.1
	Young Children (12-18 years)	0.60 (0.06, 1.14)	91%	0.03∗
Study quality	High	0.60 (0.06, 1.14)	91%	0.03∗
	Middle	1.54 (−0.29, 3.37)	94%	0.1
Sample source	Feces	0.51 (0.01, 1.02)	90%	0.04∗
	Plasma	2.49 (1.75, 3.23)	N/A	<0.00001∗
	Urine	1.69 (0.64, 2.75)	N/A	0.002∗
Dietary control	Yes	0.76 (0.21, 1.31)	88%	0.007∗
	No	0.78 (−0.49, 2.05)	95%	0.23
Probiotic and antibiotic control	Yes	0.56 (0.03, 1.08)	91%	0.04∗
	No	1.95 (1.36, 2.53)	0%	<0.00001∗

Table 6.

Summary of key subgroup analyses results.

Geographic Location
SCFA Type	Asian Studies SMD (95% CI)	Western Studies SMD (95% CI)	P
Acetic Acid	0.31 (−0.22 to 0.84)	0.24 (−0.38 to 0.86)	0.86
Propionic Acid	0.02 (−0.50 to 0.53)	0.02 (−0.34 to 0.38)	0.97
Butyric Acid	0.05 (−0.35 to 0.45)	−0.08 (−0.52 to 0.36)	0.65
Valeric Acid	1.07 (0.32 to 1.83)∗	0.19 (−0.46 to 0.84)	0.08∗

Sample Source
SCFA Type	Fecal SMD (95% CI)	Plasma SMD (95% CI)	Urine SMD (95% CI)	P
Propionic Acid	0.11 (−0.14 to 0.36)	−2.20 (−2.90 to −1.50)∗	1.13 (0.17 to 2.09)∗	<0.00001∗
Butyric Acid	−0.02 (−0.31 to 0.27)	−0.62 (−1.18 to −0.06)∗	1.45 (0.49 to 2.41)∗	0.002∗

Age Groups
SCFA Type	Young Children SMD (95% CI)	Pre-adolescents SMD (95% CI)	P
Propionic Acid	0.14 (−0.10to 0.42)	−0.56 (−2.13 to 1.02)	0.14
Butyric Acid	0.08 (−0.16 to 0.32)	−0.38 (−1.95 to 1.19)	0.22

Note: SMD: Standardized Mean Difference; CI: Confidence Interval. ∗Statistically significant (p < 0.05).

The examination of different biological matrices revealed compelling patterns. While fecal samples showed moderate elevation (SMD = 0.51, 95% CI: 0.01 to 1.02), plasma samples exhibited markedly higher levels (SMD = 2.49, 95% CI: 1.75 to 3.23), and urinary samples also demonstrated significant increases (SMD = 1.69, 95% CI: 0.64 to 2.75) (Table 5 and Supplementary File S2). This consistent elevation across multiple biological compartments suggests systemic alterations in valeric acid metabolism rather than localized changes.

Age-related analyses(Table 5 and Supplementary File S2) revealed stronger associations in pre-adolescents (SMD = 1.54, 95% CI: 0.29 to 3.37) compared to young children (SMD = 0.60, 95% CI: 0.06 to 1.14), suggesting potential developmental influences on valeric acid metabolism in ASD. This finding raises interesting questions about the role of age-related changes in gut microbiota composition and metabolic patterns.

The methodological robustness of these findings was supported by consistent results across different analytical platforms. While GC analysis showed the strongest association (SMD = 0.87, 95% CI: 0.03 to 1.76), both GC-MS and HPLC methods corroborated the overall trend, albeit with varying effect sizes(Table 5 and Supplementary File S2). This technical consistency strengthens the reliability of our findings.

These results suggest that altered valeric acid metabolism may represent a significant metabolic feature of ASD, potentially reflecting broader perturbations in gut microbiota function and fermentation patterns. The observed variations across geographical regions and age groups provide novel insights into the complex relationship between valeric acid metabolism and ASD pathophysiology. These findings may have implications for understanding disease mechanisms and developing potential biomarkers for ASD.

3.9. Isovaleric acid

Our meta-analysis of isovaleric acid levels incorporated data from 8 studies, including 221 individuals with ASD and 283 controls. The analysis revealed no significant difference in isovaleric acid levels between ASD and control groups (SMD = 0.15, 95% CI: −0.13 to 0.43, p = 0.29). The analysis demonstrated moderate heterogeneity (I² = 56%, p = 0.03) (Fig. 7a), notably lower than observed for other SCFAs in this study.

Fig. 7.

Meta-analysis and sensitivity analysis of isovaleric acid levels in ASD. (a) Pooled analysis comparing isovaleric acid levels between individuals with ASD and controls. (b) Sensitivity analysis evaluating the robustness of the isovaleric acid levels in ASD.

Initial analysis of the 8 included studies showed varied individual effect sizes, ranging from slightly negative (SMD = −0.33) (Wang et al., 2019) to strongly positive associations (SMD = 1.53) (Zhao et al., 2017). A sensitivity analysis was conducted by sequentially removing individual studies to assess their impact on the overall effect estimate. The removal of the Renshan Zhao et al., 2017 study reduced heterogeneity to I² = 23% while maintaining the non-significant overall effect (SMD = 0.05, 95% CI: −0.16 to 0.26) (Fig. 7b), suggesting this study's influential role in the observed heterogeneity.

The relative consistency of findings across most studies, despite varying analytical methods and geographical locations, strengthens the reliability of these results. The moderate heterogeneity observed for isovaleric acid contrasts markedly with the high heterogeneity seen in other SCFA analyses, suggesting potentially more stable metabolic patterns for this particular compound.

This finding adds an important dimension to our understanding of SCFA profiles in ASD, suggesting that while some SCFAs show marked differences between ASD and control groups, isovaleric acid levels remain relatively stable. This stability might reflect distinct metabolic pathways or regulatory mechanisms for isovaleric acid that are less affected by the pathophysiological processes associated with ASD.

3.10. Hexanoic acid

Our investigation of hexanoic acid levels encompassed 6 studies with 170 ASD participants and 221 controls, revealing significant alterations in this metabolite across different biological matrices. The meta-analysis demonstrated moderately elevated hexanoic acid levels in individuals with ASD (SMD = 0.60, 95% CI: 0.05 to 1.16, p = 0.03), with substantial heterogeneity across studies (I² = 84%, p < 0.00001) (Fig. 8a).

Fig. 8.

Meta-analysis and subgroup analyses of hexanoic acid levels in ASD. (a) Pooled analysis comparing hexanoic acid levels between individuals with ASD and controls. (b) Subgroup analysis by sample source (feces, plasma, urine). (c) Subgroup analysis by age-stratification (pre-adolescents, young children). (d) Subgroup analysis by geography (Asian countries, Western countries). (e) Sensitivity analysis evaluating the robustness of hexanoic acid levels in ASD.

Further analysis by sample source revealed distinct patterns across biological matrices. Plasma samples showed the most pronounced elevation (SMD = 1.99, 95% CI: 1.32 to 2.66, p < 0.00001), while fecal samples demonstrated a more modest increase (SMD = 0.16, 95% CI: −0.09 to 0.40, I² = 14%). Urinary hexanoic acid levels also showed significant elevation (SMD = 1.10, 95% CI: 0.14 to 2.05, p = 0.02) (Fig. 8b). This differential distribution across biological compartments suggests complex alterations in hexanoic acid metabolism and transport in ASD.

Age-stratified analysis revealed interesting developmental patterns, with pre-adolescents showing stronger associations (SMD = 1.27, 95% CI: 0.12 to 2.65, I² = 90%) compared to young children (SMD = 0.18, 95% CI: −0.15 to 0.51, I² = 39%) (Fig. 8c). This age-dependent variation might reflect developmental changes in gut microbiota composition or metabolic processes.

Geographic analysis indicated comparable effects between Asian (SMD = 0.62, 95% CI: −0.05 to 1.29, I² = 87%) and Western populations (SMD = 0.58, 95% CI: 0.02 to 1.13), suggesting the robustness of these findings across different populations and dietary patterns(Fig. 8d). Sensitivity analysis identified one influential study (El-Ansary et al., 2011) whose removal moderated the overall effect while reducing heterogeneity, though the general trend remained consistent(Fig. 8e).

These findings suggest that hexanoic acid metabolism is significantly altered in ASD, with particularly pronounced changes in plasma levels. The observed variations across biological matrices and age groups provide novel insights into the complex metabolic perturbations associated with ASD, potentially reflecting broader alterations in fatty acid metabolism and gut microbiota function.

3.11. Total SCFAs

Our meta-analysis examining total SCFA concentrations incorporated data from 5 studies, encompassing 185 individuals with ASD and 226 neurotypical controls. The comprehensive analysis revealed no significant difference in total SCFA levels between ASD and control groups (SMD = −0.01, 95% CI: −0.60 to 0.58, p = 0.97), with notable heterogeneity among studies (I² = 86%, p < 0.00001) (Fig. 9a).

Fig. 9.

Meta-analysis and sensitivity analysis of total SCFA levels in ASD. (a) Pooled analysis comparing total SCFA levels between individuals with ASD and controls. (b) Sensitivity analysis assessing the robustness of total SCFA level differences in ASD.

A sensitivity analysis excluding the James 2011 study yielded interesting results, with the effect size shifting toward a positive direction (SMD = 0.29, 95% CI: −0.06 to 0.64) and heterogeneity substantially decreasing (I² = 46%, p = 0.14) (Fig. 9b). This marked reduction in heterogeneity suggests that methodological or population differences in this study might have contributed to the initial high heterogeneity observed.

The individual studies showed varying effect sizes, ranging from moderately negative (SMD = −0.91) (Adams et al., 2011) to moderately positive (SMD = 0.65) (Wang et al., 2012b). This variability might reflect differences in study populations, dietary patterns, or analytical methodologies. Notably, three of the five studies reported positive effect sizes, although the magnitude and precision of these estimates varied considerably.

The divergent findings across studies, coupled with the moderate to high heterogeneity, suggest that total SCFA levels alone may not serve as a reliable marker for ASD. This observation is particularly significant when contrasted with the results for individual SCFAs, which often demonstrated more consistent and discernible trends. These findings underscore the metabolic complexity of the gut microbiota in ASD and suggest that profiling individual SCFAs may yield greater insight into metabolic disturbances than assessments based solely on total SCFA concentrations.

By illustrating that individual SCFAs may exhibit specific patterns of dysregulation while the cumulative SCFA pool remains relatively unchanged, these results provide important contributions to our understanding of gut microbial metabolism in ASD. The relative stability in total SCFA levels, despite notable variation in individual components, may reflect compensatory microbial mechanisms or adaptive physiological responses by the host aimed at maintaining metabolic homeostasis.

4. Discussion

The present findings contribute to a deeper understanding of the complex interplay between gut microbiota, SCFAs, and ASD, offering insights into the potential mechanisms underlying ASD pathogenesis. Furthermore, this meta-analysis identifies candidate biomarkers and therapeutic targets that may support earlier diagnosis or intervention strategies. In parallel, it underscores critical knowledge gaps and emphasizes the need for further investigation to delineate the precise role of SCFAs in ASD, thereby informing the design of future studies in this rapidly advancing domain.

4.1. SCFAs associated changes in autism

This meta-analysis reveals a distinct and reproducible signature of gut microbial dysbiosis in ASD, characterized not by a global shift in total SCFA output but by specific elevations in valeric and hexanoic acids, alongside a more consistent signal for isobutyric acid. The significant elevation of these particular metabolites, alongside a more consistent signal for isobutyric acid, suggests a targeted perturbation in microbial metabolic pathways rather than a blanket reduction in fermentative capacity (Macfabe, 2012a). Notably, the association with isobutyric acid was characterized by substantially lower heterogeneity (I² = 44%), which decreased to 0% upon sensitivity analysis, indicating a particularly stable and reproducible alteration that may implicate shifts in protein-fermenting microbial communities. This pattern contrasts with previous findings that reported widespread reductions in acetate, propionate, or butyrate (Wang et al., 2012a; Adams et al., 2011), and may instead reflect compensatory shifts in overall microbial fermentation activity or adaptive metabolic responses in the host, even in the presence of imbalances among individual SCFA species.

These findings both corroborate and extend previous investigations of SCFA profiles in ASD. The observed elevations in valeric and hexanoic acids align with established literature documenting altered gut microbial fermentation patterns in ASD. For instance, Wang et al. (2019) previously documented elevated valeric acid concentrations in fecal samples from children with ASD, concordant with our meta-analytic findings. However, our analysis of acetic, propionic, and butyric acids presents a more nuanced perspective than previous investigations. While Adams et al. (2011) reported decreased concentrations of these SCFAs in ASD, our meta-analysis reveals substantial heterogeneity without significant overall differences, emphasizing the importance of methodological standardization and confounding factor consideration in interpreting SCFA profiles.

This meta-analysis contributes several novel insights to the field. First, we identified differential patterns across biological matrices, with plasma samples frequently demonstrating more pronounced alterations compared to fecal or urinary specimens. This finding, particularly evident in propionic and hexanoic acid analyses, suggests complex alterations in SCFA metabolism and transport mechanisms extending beyond gut production. Second, our age-stratified analyses revealed potential developmental influences on SCFA profiles, with pre-adolescents frequently demonstrating stronger associations than younger children. This age-dependent variation, particularly notable for valeric and hexanoic acids, provides novel insights into the developmental trajectory of metabolic alterations in ASD. Finally, our analysis of isobutyric acid, characterized by markedly lower heterogeneity compared to other SCFAs, suggests its potential utility as a more stable indicator of gut microbiota alterations in ASD.

Crucially, the interpretation of our findings must be tempered by the substantial heterogeneity observed. For acetic, propionic, and butyric acids, the extremely high between-study variability (I² >75%) means the pooled SMDs, while indicating no statistically significant overall difference, are unreliable as definitive estimates of a true effect. This high heterogeneity precludes drawing firm conclusions about the direction or consistency of change for these major SCFAs in the ASD population as a whole. Even for valeric acid, where a significant pooled effect was found, the notably high heterogeneity (I² = 92%) suggests that the magnitude of elevation varies considerably across studies and populations.

These findings underscore the importance of considering multiple SCFAs, diverse biological matrices, and developmental stages in future gut-brain axis research in ASD. The observed patterns suggest that complex metabolic alterations may contribute to ASD pathophysiology through multiple mechanistic pathways.

4.2. SCFA metabolism and gut microbiota dysbiosis

The altered SCFA profiles identified in this meta-analysis, particularly the significant elevations in valeric and hexanoic acids, indicate substantial perturbations in gut microbial ecology among individuals with ASD. These metabolic alterations correlate with documented shifts in microbial community composition, notably the increased abundance of specific bacterial consortia, including certain Clostridium clusters (e.g., clusters XIVa and IV) and Desulfovibrio species, which are recognized producers of valeric acid and other longer-chain fatty acids (He et al., 2023; Macfabe, 2012a). Furthermore, the observed elevation in isobutyric acid, a product primarily derived from protein fermentation, reinforces the hypothesis of an altered microbial metabolic balance in ASD, possibly favoring proteolytic over saccharolytic pathways (De Maria et al., 2013).

The maintenance of total SCFA concentrations despite significant alterations in individual components (e.g., elevated valeric and hexanoic acids) points towards a proportional imbalance within the SCFA pool rather than a blanket increase or decrease in microbial fermentative output. This metabolic profile suggests a reconfiguration of fermentative pathways within the gut microbial ecosystem. Specific bacterial taxa may expand their niche, leading to overproduction of certain metabolites (like valerate from Clostridium clusters) while the output of others (like acetate from Bifidobacterium) may be proportionally reduced, keeping the total fermentative flux relatively constant. This scenario aligns with the concept of functional redundancy in the gut microbiome. Furthermore, the stability of total SCFAs could also reflect host physiological adaptations, such as modulated intestinal absorption or hepatic metabolism, aimed at maintaining systemic metabolic homeostasis despite an altered microbial metabolite profile. Critically, this proportional shift—particularly the increased share of branched-chain and longer-chain fatty acids—may have distinct biological consequences compared to a change in total SCFA load, potentially influencing neuroimmune signaling through specific receptors (e.g., FFAR2/3) with different affinities for various SCFAs (Kang et al., 2018b). The complex interactions between dietary composition, microbial metabolism, and SCFA production in ASD warrant further investigation, particularly given the significant impact of dietary variables on our meta-analytic outcomes.

4.3. Impact on intestinal barrier and inflammation

The findings of this meta-analysis indicate that individuals with ASD frequently exhibit alterations in the composition of gut microbiota, which may be closely associated with intestinal barrier dysfunction and enhanced inflammatory responses. Although previous studies have demonstrated a correlation between gastrointestinal (GI) symptoms and ASD, the underlying mechanisms—particularly those involving intestinal barrier integrity and immune activation—are complex and multifactorial.

Accumulating evidence suggests that microbial dysbiosis in ASD contributes to increased intestinal permeability, commonly referred to as “leaky gut” (Fiorentino et al., 2016). This phenomenon is characterized by the disruption of tight junctions between intestinal epithelial cells, resulting in the translocation of luminal antigens, microbial metabolites, and endotoxins such as lipopolysaccharide (LPS) into the systemic circulation. Notably, elevated levels of circulating LPS and zonulin—a physiological modulator of tight junctions—have been observed in children with ASD, further indicating compromised gut barrier function (Fiorentino et al., 2016; Esnafoglu et al., 2017).

A disrupted intestinal barrier facilitates systemic immune activation and low-grade chronic inflammation, which may in turn aggravate neuroinflammatory processes via the gut-brain axis. The gut microbiota modulates host immunity by influencing T-helper cell differentiation, especially the balance between regulatory T cells (Tregs) and pro-inflammatory Th17 cells (Lee, 2018). In ASD, this dysregulated immune balance is reflected by elevated concentrations of pro-inflammatory cytokines such as interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α), and interleukin-17A (IL-17A) in both peripheral circulation and intestinal mucosa (Ashwood et al., 2011; Rose et al., 2018a). These cytokines are capable of crossing the blood-brain barrier and may contribute to microglial activation and neuroinflammatory responses, ultimately impairing neuronal development and synaptic function.

Several bacterial genera associated with ASD, including Clostridium, Desulfovibrio, and Bacteroides, are known to produce SCFAs, which exhibit concentration-dependent effects on gut epithelial integrity and immune modulation. Rose et al. (2018b) recently demonstrated that SCFAs directly regulate the expression of tight junction proteins in intestinal epithelial cells. Among these, butyrate is particularly critical for maintaining epithelial barrier function and exerting anti-inflammatory effects (Parada Venegas et al., 2019). In contrast, elevated levels of propionate have been linked to behavioral abnormalities and increased inflammatory responses in animal models of autism (Macfabe, 2012; Parada Venegas et al., 2019; Liu et al., 2015).

Moreover, our meta-analysis identified increased concentrations of valeric and caproic acids, which may exert additional regulatory effects on intestinal barrier function and inflammatory processes. A recent study by Zeng X et al. (Zeng et al., 2023) reported that valeric acid can activate free fatty acid receptor 2 (FFAR2), leading to enhanced secretion of the pro-inflammatory cytokine IL-17.

Collectively, these alterations in SCFA profiles may significantly influence intestinal barrier integrity and immune regulation in ASD, further underscoring the central role of the gut microbiota–immune–brain axis in the pathophysiology of the disorder.

4.4. Systemic metabolic implications and potential mechanisms underlying plasma-urine disparities

A pivotal finding of this meta-analysis is the stark contrast in SCFA levels across different biological compartments. The divergent patterns—such as significantly lower plasma but higher urinary propionate and butyrate in ASD—cannot be explained by gut production alone and underscore complex systemic alterations in SCFA handling. These compartment-specific profiles likely arise from the interplay of several potential mechanisms:

Increased intestinal permeability (‘leaky gut’), frequently reported in ASD (Fiorentino et al., 2016; Esnafoglu et al., 2017), could facilitate the paracellular passage of SCFAs and other microbial metabolites into the portal circulation. While this might increase systemic exposure, the concomitant triggering of systemic immune responses and potential changes in enterocyte function could also affect the active transcellular transport mechanisms for SCFAs, leading to a net shift in absorption dynamics that differs among individual SCFAs.

The liver plays a crucial role in clearing portal vein SCFAs via first-pass metabolism, primarily through β-oxidation or conjugation. Altered hepatic function or metabolic capacity in ASD could modify this clearance rate. For instance, reduced hepatic extraction of propionate or butyrate would lead to higher peripheral plasma levels, which is not consistently observed. Conversely, enhanced hepatic metabolism might lower plasma concentrations, while potentially increasing the production and subsequent urinary excretion of water-soluble metabolites derived from SCFA catabolism.

The kidney actively reabsorbs SCFAs from the glomerular filtrate, primarily in the proximal tubule. A dysfunction in these renal transport systems (e.g., downregulation of monocarboxylate transporters) could reduce reabsorption, leading to increased urinary excretion of SCFAs even in the context of normal or low plasma levels. This mechanism could directly explain elevated urinary SCFAs alongside decreased or unchanged plasma concentrations.

These mechanisms are not mutually exclusive and may operate in concert. For example, altered gut permeability could change the profile and quantity of SCFAs reaching the liver, which in turn processes them differently, ultimately presenting a modified substrate load to the kidneys. Our findings shift the focus from a solely ‘gut-centric’ view to one that acknowledges the integrated host-microbe metabolic interplay along the gut-liver-kidney axis. Elucidating the dominant mechanism(s) requires future studies that concurrently measure SCFAs in multiple biofluids (portal/mesenteric blood, peripheral plasma, urine) alongside assessments of barrier function, hepatic enzymes, and renal clearance in well-phenotyped ASD cohorts.”

4.5. Neurodevelopmental implications: age-related patterns

Age-stratified analyses revealed distinct developmental trajectories in SCFA profiles, with stronger associations between SCFAs and ASD observed in pre-adolescent individuals compared to younger children. These findings carry important implications for understanding the neurodevelopmental course of ASD. The amplified associations in pre-adolescents may be driven by several interconnected factors. First, the onset of puberty involves significant hormonal changes (e.g., in sex steroids and growth hormone) that can modulate gut permeability, immune function, and potentially the gut microbiota itself, possibly altering the host's response to microbial metabolites like SCFAs. Second, dietary patterns often shift during childhood towards greater independence and potentially lower fiber intake, which could further shape the gut microbial community and its metabolic output over time, making dysbiosis-related signatures more pronounced. Third, the stronger association may represent a cumulative effect of prolonged exposure to an altered gut microbial environment and its metabolic products, with pathophysiological impacts accruing across development. Specific SCFAs, notably butyrate and propionate, have been implicated in key neurobiological processes, including epigenetic modulation and neurotransmitter synthesis (Silva et al., 2020). The observed age-dependent variations, particularly in valeric and hexanoic acid levels, indicate that gut microbial metabolites may exert stage-specific effects during critical periods of neurodevelopment.

These results are consistent with prior work by Borre et al., which identified sensitive developmental windows during which gut-derived metabolites may have heightened impact on brain function (Borre et al., 2014a). Such evidence highlights the need for longitudinal studies to better characterize the temporal dynamics of SCFA-mediated gut–brain signaling in ASD. Moreover, the amplified associations observed in pre-adolescents raise the possibility of cumulative effects of altered SCFA profiles over time and point to the potential utility of age-specific therapeutic strategies targeting the gut–brain axis in ASD.

4.6. Biomarker potential: implications of SCFA alterations in ASD

The results of this meta-analysis suggest that specific SCFAs --- particularly valeric acid, hexanoic acid, and isobutyric acid --- exhibit alterations in individuals with ASD, highlighting them as promising candidates for future biomarker research. Biomarkers, by definition, must be biologically relevant, quantifiable, reproducible across populations, and ideally linked to disease mechanisms. The observed SCFA patterns in ASD populations, especially when evaluated across different biological matrices, meet several of these essential criteria.

Elevated plasma levels of valeric acid (SMD = 2.49, 95% CI: 1.75–3.23) and hexanoic acid (SMD = 1.99, 95% CI: 1.32–2.66) are particularly compelling, given that systemic SCFA concentrations are less influenced by transient gastrointestinal fluctuations compared to fecal measurements and more likely to reflect persistent metabolic dysregulation. SCFAs in plasma are a downstream reflection of intestinal production combined with gut barrier integrity and hepatic metabolism, thus providing a composite signal that integrates gut microbial activity and host metabolic processing (De Filipe et al., 2014). However, the substantial heterogeneity observed indicates that the magnitude of these alterations varies considerably across studies, reflecting influences from methodological and population differences. This heterogeneity currently precludes firm conclusions about their consistency as universal biomarkers and underscores that their potential utility is contingent upon resolving these sources of variation. For instance, the overrepresentation of valeric acid and hexanoic acid in plasma implies either overproduction by gut microbes, impaired intestinal absorption dynamics, or altered hepatic clearance — processes which may be pathophysiologically relevant in ASD.

From a mechanistic standpoint, SCFAs act as signaling molecules by binding to G-protein-coupled receptors such as GPR41 and GPR43 (also known as FFAR3 and FFAR2, respectively), which modulate immune responses, hormonal signaling, and energy metabolism (Brown et al., 2003). Aberrant activation of these pathways, driven by elevated valeric and hexanoic acids, could theoretically contribute to the neurodevelopmental and immunological abnormalities frequently observed in ASD. Specifically, valeric acid has been reported to stimulate pro-inflammatory cytokine production through FFAR2 signaling, promoting interleukin-17 (IL-17) secretion from Th17 cells, which are implicated in neuroimmune dysregulation (Zeng et al., 2023).

Furthermore, the lower heterogeneity associated with isobutyric acid across studies (I² = 44%) and its stability upon sensitivity analysis highlights its potential as a particularly robust candidate biomarker. Isobutyric acid is produced predominantly from protein fermentation rather than carbohydrate metabolism, and its systemic elevation may indicate a distinct gut microbiota composition with a higher prevalence of proteolytic bacteria — a feature increasingly documented in ASD-related dysbiosis (De Maria et al., 2013). This metabolite's methodological consistency across geographically diverse populations suggests that it may be less sensitive to external dietary variables compared to other SCFAs, favoring its reliability for further investigation.

It is also noteworthy that the biological matrix plays a crucial role in biomarker evaluation. Plasma and urine-derived SCFAs tend to reflect systemic exposure and metabolic clearance, whereas fecal SCFAs represent luminal concentrations influenced by short-term dietary intake, microbial fermentation, and gut transit time. The strong elevation of valeric and hexanoic acids in plasma and urine, rather than feces alone, supports the argument that these SCFAs are not merely epiphenomena of altered gut microbiota but may participate directly in ASD pathophysiology through systemic metabolic or neuroimmune pathways (West et al., 2014).

Recent studies have begun to explore this diagnostic potential using advanced modelling. Although not focused solely on SCFAs, integrative multi-omics approaches combining metabolomic and microbiome data from 2022 to 2023 have shown promising accuracy in distinguishing ASD from controls (Deng et al., 2022; Yap et al., 2010), underscoring the value of incorporating SCFA profiles into broader biomarker panels.

However, before the promising associations identified here can be translated into clinically useful biomarkers, several critical steps are necessary: (1) The development and application of rigorous, standardized protocols across studies to minimize methodological heterogeneity; (2) validation in large, multicenter, longitudinal cohorts; and (3) integration with clinical, behavioral, and neuroimaging endpoints to confirm biological relevance. Future research should also explore the dynamic range of SCFAs across developmental stages and symptom severity, which would enable their application as state or trait biomarkers rather than mere correlates of gut microbiota imbalance.

In summary, this meta-analysis highlights valeric acid, hexanoic acid, and isobutyric acid as promising candidates for ASD biomarker development. Their consistent elevation across biological matrices and plausible mechanistic links to neuroimmune regulation and host metabolism strengthen their potential as objective biochemical markers for ASD diagnosis and stratification. Nonetheless, further mechanistic studies and longitudinal validations are warranted to fully establish their diagnostic and predictive value.

4.7. Therapeutic implications and personalized medicine approaches

The distinct SCFA dysregulation patterns observed in this meta-analysis, particularly elevated valeric and hexanoic acids, suggest targeted modulation of gut microbiota composition as a potential therapeutic strategy. Probiotic interventions tailored to suppress taxa associated with elevated branched-chain fatty acids (e.g., Clostridium spp.) or prebiotics promoting beneficial SCFA-producing genera (e.g., Bifidobacterium) warrant further investigation (Sanctuary et al., 2019). However, the substantial heterogeneity in acetic, propionic, and butyric acid profiles underscores the need for personalized therapeutic approaches. Precision nutrition strategies, informed by individual microbiome and metabolomic profiles, may optimize intervention efficacy (Berding and Donovan, 2018). The age-dependent variations in SCFA associations, particularly stronger effects in pre-adolescents, suggest critical neurodevelopmental windows for intervention, consistent with hypotheses positing time-sensitive gut-brain axis modulation (Eshraghi et al., 2018). While butyrate supplementation remains a plausible neuroprotective strategy (Rose et al., 2018c), the complex interplay between individual SCFAs necessitates caution to avoid unintended metabolic perturbations. Future clinical trials should stratify participants by baseline SCFA profiles, age, and dietary patterns to elucidate subgroup-specific therapeutic responses.

However, a critical step in validating their clinical utility remains absent: the direct correlation between plasma or fecal levels of these SCFAs and standardized, quantitative measures of ASD symptom severity. None of the studies included in this meta-analysis provided such an integrated analysis.

4.8. Strengths of meta-analysis

This study offers several methodological and conceptual advances to ASD research. First, the inclusion of 16 studies with 987 participants (473 ASD, 514 controls) provides robust statistical power to detect subtle SCFA alterations, addressing limitations of underpowered individual studies. Second, the analysis extends beyond conventional SCFAs (acetate, propionate, butyrate) to incorporate less-studied metabolites (e.g., valeric, hexanoic acids), offering a comprehensive perspective on gut microbial metabolism in ASD. Third, stratified analyses by sample type, age, and geography elucidate context-dependent variations in SCFA profiles, aligning with the recognition of ASD as a heterogeneous disorder requiring multi-factorial investigation (Jeste et al., 2014; Kang et al., 2020; Masi et al., 2017). The integration of sensitivity analyses and random-effects models strengthens the reliability of findings despite substantial heterogeneity, adhering to rigorous meta-analytic standards (Gu et al., 2020; Borenstein et al., 2021). These methodological rigor and breadth of analysis enhance the translational relevance of findings for both mechanistic research and clinical applications.

4.9. Study limitations

Despite the strengths of our meta-analysis, several limitations warrant consideration. Firstly, the cross-sectional design of virtually all included studies precludes any inference about causality or the temporal direction of the observed associations between SCFA profiles and ASD status. Secondly, the high heterogeneity observed across studies for most SCFAs, particularly acetic, propionic, and butyric acids, limits the generalizability of our findings. This heterogeneity likely reflects the complex interplay of factors influencing SCFA levels, including diet, genetics, and environmental factors, as highlighted by Kushak et al. in their study of gastrointestinal issues in ASD (Kushak et al., 2017). Thirdly, variability in study quality and methodological rigor may have influenced the effect estimates. For instance, some subgroup analyses suggested differences in effect sizes between studies with higher versus moderate NOS scores, indicating that methodological limitations such as less rigorous control for confounders could contribute to heterogeneity and potentially inflate certain associations. Fourthly, the variability in sample collection methods and analytical techniques across studies may have introduced methodological bias. For instance, the use of different GC columns or derivatization methods can affect SCFA quantification, as noted by De Baere et al. in their review of SCFA analysis techniques (De Baere et al., 2013). Additionally, the lack of standardized fasting protocols or dietary records in some studies may have confounded results, given the significant impact of recent food intake on SCFA levels (Cummings et al., 1987). Our meta-analysis was also limited by the predominantly cross-sectional nature of included studies, precluding conclusions about causality or temporal changes in SCFA profiles. the marked underrepresentation of female participants across the included studies, reflecting the historical recruitment bias in ASD research, has resulted in a critical knowledge gap regarding potential sex-specific differences in gut microbial metabolism and SCFA profiles. This limits our ability to explore potential sex-specific differences in SCFA metabolism, an important consideration raised by Werling and Geschwind in their review of sex differences in ASD (Werling and Geschwind, 2013).

Finally, related to methodological transparency, our analysis is constrained by the variable reporting of analytical details in the primary studies. While we extracted and reported the primary detection methods (e.g., GC, HPLC) in Table 1, critical technical parameters that affect reproducibility and cross-study comparability—such as detailed calibration procedures, limits of detection and quantification, and intra- or inter-assay coefficients of variation—were seldom reported. This lack of standardized reporting is a common challenge in metabolomic synthesis and may contribute to the unexplained heterogeneity observed in our analyses. Future studies in this field would benefit from adhering to consensus reporting guidelines for metabolomic data to facilitate more robust meta-analyses.

4.10. Knowledge gaps identified through analysis

Our meta-analysis has illuminated several crucial knowledge gaps in the field of SCFA research in ASD. Firstly, the paucity of longitudinal studies examining SCFA profiles across different developmental stages of ASD represents a significant research gap. Given our findings of age-related differences in SCFA associations, longitudinal investigations are vital to understand how gut microbial metabolism evolves throughout the course of ASD. This aligns with calls by Borre et al. for more developmental studies of the gut-brain axis in neurodevelopmental disorders (Borre et al., 2014b). Secondly, the limited exploration of less common fatty acids, such as heptanoic and octanoic acids, in the context of ASD represents a missed opportunity. Recent work by Kikuchi et al. suggests that these longer-chain fatty acids may have unique neuroactive properties (Kikuchi et al., 2021), warranting their inclusion in future ASD studies. Additionally, the lack of studies examining the relationship between SCFA levels and specific ASD symptoms or severity represents a critical gap. Such investigations could provide valuable insights into the potential role of SCFAs in ASD heterogeneity, as suggested by the dimensional approach to ASD proposed by Lombardo et al. (2019). Lastly, our analysis highlights the need for more mechanistic studies exploring how altered SCFA profiles influence neurodevelopment and behavior in ASD. While animal studies have provided some insights, translational research linking SCFA alterations to specific neurobiological changes in humans with ASD remains limited, a gap noted by Needham et al. in their review of gut-brain axis research in ASD (Needham et al., 2020).

Furthermore, as noted in the review process, a critical mechanistic gap concerns the specific enzymatic pathways involved. No studies to date have investigated whether alterations in host or microbial enzymes (e.g., acyl-CoA synthetases, dehydrogenases) or their isoforms are responsible for the observed metabolic shifts in SCFAs, such as the elevated valeric and hexanoic acids. Future research employing metabolomic flux analyses and investigating genetic polymorphisms in relevant metabolic pathways is needed to move from correlation to mechanistic causation.

4.11. Research priorities and recommendations

Our meta-analysis highlights several key research priorities for advancing our understanding of SCFAs in ASD. Future studies must implement and report strict protocols to control for key confounders identified here, notably diet and medication/probiotic use. The analysis and reporting of SCFA data should be performed and presented separately by biological matrix (feces, plasma, urine) to advance our understanding of compartment-specific biology. Firstly, longitudinal studies examining SCFA profiles from early childhood through adolescence are crucial to elucidate the developmental trajectory of gut microbial metabolism in ASD. Such studies could reveal critical windows for intervention, as suggested by Borre et al. in their review of microbiota and neurodevelopmental windows (Borre et al., 2014c). Secondly, integrative multi-omics approaches combining SCFA profiling with metagenomics, metabolomics, and neuroimaging could provide a more comprehensive picture of gut-brain interactions in ASD. This aligns with recent work by Vargason et al., who demonstrated the power of integrating metabolomic and clinical data in ASD research (Vargason et al., 2019). Thirdly, investigating the relationship between SCFA levels and specific ASD symptoms or endophenotypes could yield valuable insights into the heterogeneity of ASD. This approach is supported by the Research Domain Criteria (RDoC) framework proposed by Insel et al. for studying mental disorders (Insel et al., 2010). Additionally, mechanistic studies exploring how SCFAs influence neurodevelopment and behavior in ASD are needed. In vitro and in vivo models, such as those developed by Rose et al. to study butyrate's effects on mitochondrial function in ASD (Rose et al., 2018c), could be valuable in this regard. Thirdly, intervention studies targeting the gut microbiome and SCFA production in ASD should be prioritized, building on promising preliminary work such as that by Kang et al. on microbiota transfer therapy (Kang et al., 2019). Finally, future observational studies must systematically correlate SCFA profiles with detailed clinical phenotyping, including ASD severity scores (e.g., from ADOS-2, CARS) and co-occurring conditions (e.g., GI distress, anxiety). This is essential to determine if SCFAs are trait markers, state markers that fluctuate with symptom severity, or predictors of specific behavioral domains.

4.12. Methodological improvements needed

To address the limitations identified in our meta-analysis, several methodological enhancements are essential for advancing SCFA research in ASD. Standardization of sample collection, processing, and analysis protocols is critical to reducing heterogeneity and enhancing comparability across studies. This includes the implementation of consistent fasting protocols and dietary record-keeping, as recommended by Reade et al. in their analysis of the fecal metabolome (Reade et al., 2019). The adoption of advanced analytical techniques, such as the use of stable isotope-labeled internal standards and tandem mass spectrometry, could significantly improve the accuracy and reproducibility of SCFA quantification, as demonstrated by Han et al. (2015). Additionally, the inclusion of larger sample sizes and more diverse participant cohorts—particularly with better representation of females and different age groups—is necessary to enhance the generalizability of findings. Incorporating sibling controls and longitudinal study designs could help to disentangle genetic and environmental contributions to SCFA profiles in ASD, an approach successfully applied by Fattorusso et al. in their examination of gut microbiota in ASD families (Fattorusso et al., 2019b). Finally, the establishment and validation of standardized SCFA reference ranges across different age groups and populations would greatly aid in the interpretation of results across studies. Such reference ranges have proven valuable in other domains of ASD research, as demonstrated by Shen et al. in their development of normative models for brain development in ASD (Shen et al., 2017).

In summary, the present study advances the current understanding of ASD and gut metabolism in several key respects. First, it goes beyond documenting generalized gut dysbiosis to identify SCFA metabolites—valeric acid and hexanoic acid—that are consistently elevated in individuals with ASD. Second, by synthesizing data across fecal, plasma, and urine specimens, this study demonstrates the systemic nature of these metabolic alterations, indicating substantial engagement of host metabolic pathways rather than confinement to the intestinal lumen. Third, our subgroup analyses uncover previously underexplored modifiers of the SCFA-ASD association, including age and geographical region, thereby providing a mechanistic framework for interpreting heterogeneity in prior findings. Collectively, these insights address a critical knowledge gap by delineating a more precise ASD-associated metabolic signature and paving the way for future mechanistic investigations and translational research focused on targeted pathways and personalized therapeutic approaches.

5. Conclusions

This meta-analysis provides compelling yet nuanced evidence for dysregulated SCFA profiles in individuals with ASD. It is crucial to acknowledge the substantial methodological and biological heterogeneity characterizing the existing literature. Specifically, the high degree of inconsistency across studies (I² >75%) for acetate, propionate, and butyrate precludes reliable conclusions regarding their overall directional change in ASD. The most robust and consistent findings are the significant elevations of valeric acid and hexanoic acid across biological matrices, pointing to a specific gut microbial metabolic signature in ASD.

Furthermore, our analysis reveals that compartment-specific dysregulation—such as the opposing directions of change for propionic acid in plasma versus urine—may be as informative as absolute levels. This highlights likely alterations in host metabolic or excretory pathways in ASD, extending the pathophysiology beyond gut microbial production alone. Subgroup analyses indicated that the magnitude of SCFA perturbation is influenced by age, sample type, and geographic region, with notably larger effect sizes in pre-adolescents, suggesting developmental windows of sensitivity to microbial metabolites. The consistent pattern of isobutyric acid across studies supports its potential utility as a stable biomarker candidate.

Mechanistically, elevated SCFAs, particularly valeric and hexanoic acids, may contribute to ASD pathophysiology via modulation of immune signaling, neurotransmitter pathways, and host-microbiota interactions. Collectively, these findings advance the understanding of the gut–brain–metabolite axis in ASD by moving beyond generalized dysbiosis to identify specific metabolic perturbations and systemic host responses. They underscore the urgent need for longitudinal, multicenter studies employing standardized protocols to resolve existing heterogeneity, validate these SCFAs as candidate biomarkers, and elucidate their potential causal roles in ASD pathogenesis.

CRediT authorship contribution statement

Zhaoming Yang: Writing – original draft, Visualization, Validation, Software, Methodology, Formal analysis, Data curation, Conceptualization. Xuemei Chen: Writing – original draft, Supervision, Software, Methodology, Formal analysis, Data curation. Yinhua Li: Visualization, Software, Methodology, Formal analysis, Data curation. Yanqing Feng: Writing – review & editing, Supervision, Resources, Methodology, Investigation, Formal analysis, Conceptualization. Liheng Lin: Visualization, Validation, Methodology. Xiaohui Hou: Writing – review & editing, Supervision, Resources, Conceptualization.

Informed consent statement

Not applicable.

Institutional review board statement

Not applicable.

Funding

This research was funded by the Guangdong Provincial Philosophy and Social Sciences Planning 2022 Youth Program (No. GD22YTY06).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

AbbreviationsThe following abbreviations are used in this manuscript:

ASD	Autism Spectrum Disorder
SCFA	Short chain fatty acid
CI	Confidence Interval
SMD	Standardized Mean Difference
DSM-5	Diagnostic and Statistical Manual of Mental Disorders, fifth edition
ADOS-2	Autism Diagnostic Observation Scale
ADI-R	Revised Autism Diagnostic Interview
GC	Gas chromatography
GC-MS	Gas chromatography mass spectrometry
HPLC	High-Performance Liquid Chromatography
NOS	Newcastle-Ottawa Scale

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data availability

Data will be made available on request.