This study employed a convenience sampling method from January 2024 to December 2024, selecting 26 children diagnosed with GDD attending the Longgang District Maternal and Child HealthCare Hospital of Shenzhen City as the disease group(GDD). Simultaneously, 59 age-matched healthy children who came for routine check-ups at the hospital were selected as the healthy group(HC). Inclusion criteria: age ≤ 5 years; meeting the diagnostic criteria for GDD, as defined by a developmental quotient (DQ) of < 70 in at least two domains of the Gesell Developmental Schedules (GDS) (National Center for Children's Health, et al., 2024); guardians provided informed consent and were willing to cooperate. Exclusion criteria: ①chronic gastrointestinal diseases such as inflammatory bowel disease and ulcerative colitis; ②severe liver and kidney diseases; ③use of antibiotics or probiotics within the 4 weeks prior to enrollment. This study was approved by the hospital’s Ethics Committee (Ethics Approval No: LGFYKYXMLL-2024-103), and informed consent was obtained from all guardians.

This study adopts a sample size calculation method based on effect size, and the detailed calculation formula is as follows (Equation 1). The mean value of Chao1 in the HC group is 200(SD = 50), and the mean value of the GDD group is 150(SD = 50). Therefore, Δ=50, σ=50. By substituting these values, it is calculated that each group requires at least 17 cases(with a total sample size of 34 cases). This study selected 26 children with comprehensive developmental delay as the disease and 59 healthy children of the same age who underwent physical examination in the pediatric health department as the healthy group. The final sample size of 26 GDD and 59 HC participants exceeded the minimum requirement of 17 per group, thus satisfying the statistical power criteria.

The sample size was calculated using the following formula for comparing two means:

https://media/image2.jpeg{width=“2.6in” height=“0.7666666666666667in”}.

Where:

n is the required sample size per group.

Zα is the Z-value for the Type I error rate (α); we set α = 0.05 (two-tailed), hence Zα = 1.96.

Zβ is the Z-value for the Type II error rate (β) or statistical power (1-β); we set β = 0.20 (i.e., 80% power), hence Zβ = 0.84.

Δ is the expected mean difference in the Chao1 index between groups (Δ = 50).

σ is the common standard deviation (σ = 50).

The baseline data of children were processed using SPSS 26.0 software. Count data was described using frequency, rate, and composition ratio, and between group comparisons were performed using chisquare test, continuous variables that conformed to a normal distribution are presented as mean ± standard deviation. Differences in continuous baseline characteristics (e.g., age, BMI) between the two groups (GDD vs. HC) were assessed using an independent samples t-test. Differences in alpha diversity indices (e.g., Chao1) between the two groups (GDD vs. HC) were assessed using independent samples t-tests. For beta diversity, permutational multivariate analysis of variance (PERMANOVA) was applied based on unweighted UniFrac distances. If the ANOVA result was significant (P < 0.05), post-hoc pairwise comparisons were conducted using Tukey’s Honest Significant Difference (HSD) test to identify which specific groups differed. Principal component analysis (PCA) was performed using the ade4 package in R (v3.3.3) based on the composition and relative abundance of bacterial genera. The overall bacterial distribution of the two groups was visualized. The Wilcoxon method was used to test for differential species between the two groups at the Phylum, Genus, and Species levels, with P<0.05 considered statistically significant. Functionality of the gut microbiome was inferred using KEGG (Kyoto Encyclopedia of Genes and Genomes) database, and functional differences between the two groups were analyzed.

There was no statistically significant differences in the demographic characteristics and disease-related information between the two groups (P>0.05), indicating age, BMI value, gender,delivery mode, family residence, feeding type and dietary patterns etc (Table 1). (Note: Although detailed dietary records were not collected, the use of nutritional supplements (e.g., vitamins, minerals) between the GDD and HC groups showed no statistically significant difference (P = 0.638, Table 1). This suggests that gross differences in supplement intake are unlikely to be the primary driver of the observed microbiota disparities, though the influence of unmeasured specific dietary components cannot be ruled out.).

The analysis results indicated that: 1) An independent samples t-test revealed a significant difference in microbial alpha diversity (Chao1 index) between the two groups of children (P = 0.007) (Figure 1A). 2) Principal coordinate analysis (PCoA) was used to perform dimensionality reduction on the gut microbiota data of both groups. The β-diversity of gut microbiota composition showed a clear separation between the two groups (Figure 1B).

The analysis of dominant bacterial phyla in both groups is shown in Figure 2. At the phylum level, the relative abundance of Actinobacteria was significantly higher in the GDD compared to the HC (P< 0.01), while Bacteroidetes was significantly lower in the GDD than in the HC(P< 0.05).

At the genus level, we selected the top six dominant bacterial genera in both groups and calculated their relative abundances (Figures 3). The analysis results revealed the following:Genera significantly more abundant in the GDD compared to the HC: Lachnoclostridium, Blautia,Hungatella and Erysipelatoclostridium (all P< 0.001).Genera significantly less abundant in the GDD compared to the HC: Faecalibacterium, Phascolarctobacterium, Alistipes and Sutterella (all P < 0.001).

Compared with the HC, functional prediction based on 16S rRNA data suggested potential upregulation of gut microbiota functions in the GDD group, including mRNA surveillance pathway, Proteasome pathway, and Atrazine degradation pathway. Conversely, significantly downregulated functions included Polyketide sugar unit biosynthesis, Chloroalkane and chloroalkene degradation, and Styrene degradation (Figures 4).

The microbiota–gut–brain axis has emerged as a key focus in understanding neurodevelopmental disorders, with growing evidence supporting its role in modulating brain function and behavior via immune, metabolic, and neural pathways (Cryan and Dinan, 2012). Alterations in gut microbial composition have been consistently reported in conditions such as ASD, CP, and ADHD. For instance, individuals with ASD often exhibit a reduced ratio of Bacteroidetes to Firmicutes, decreased Prevotella, and overgrowth of Clostridium—patterns that correlate with behavioral symptoms (Ye et al., 2021). Similarly, children with CP demonstrate diminished microbial diversity and increased pathogenic bacteria, which may contribute to motor deficits and chronic inflammation (Peng et al., 2023). In ADHD, reduced levels of Lactobacillus and Bifidobacterium have been implicated in disrupted dopamine metabolism, potentially exacerbating core symptoms (Wan et al., 2020). Despite these advances, the gut microbiota profile of children with GDD—a common early manifestation of various neurodevelopmental disorders—remains poorly characterized. GDD often represents a prodromal or transitional stage toward more specific diagnoses such as ASD, suggesting that early microbial disturbances may play a formative role in disease progression. Supporting this, microbiota-directed interventions in ASD, including probiotic supplementation and fecal microbiota transplantation, have shown potential to alleviate certain neurobehavioral abnormalities (Zhang et al., 2022), underscoring the therapeutic relevance of targeting the gut–brain axis during early development.

This study systematically compared the characteristics of the gut microbiota between children in the GDD group and the HC group. First, it was found that there was a significant difference in the overall community structure of gut microbiota between children with GDD and the HC group (P<0.01), which was specifically manifested as a change in α diversity. This finding is consistent with the study by Fu et al. (Fu et al, 2025), mainly because the abnormal development of the central nervous system in children with GDD can affect intestinal motility, secretory function, and intestinal barrier integrity through pathways such as the autonomic nerve and the hypothalamic-pituitary-adrenal axis (Lou et al., 2022). Second, at the classification level, the gut microbiota of children with GDD showed obvious loss of balance at both the phylum level and the genus level. At the phylum level, the relative abundance of Actinobacteria in the GDD group was significantly increased (P<0.01), while the relative abundance of Bacteroidetes was significantly decreased (P<0.05).This finding contrasts with the conclusion of a study by Borghi et al. on children with CDKL5 deficiency disorder (Borghi et al., 2024). Children with GDD have a monotonous diet or nutritional malabsorption, which may inhibit the colonization of Bacteroidetes (Nistal et al., 2012). In addition, GDD studies may include mucosal biopsy samples, which are more likely to test for mucosa-associated Bacteroidetes.

At the genus level, the GDD group exhibited significantly higher relative abundances of Lachnoclostridium, Blautia, Hungatella, and Erysipelatoclostridium (all P < 0.001), while the relative abundances of Faecalibacterium, Phascolarctobacterium, Alistipes, and Sutterella were significantly lower compared to the HC group (all P < 0.001). The observed increase in Blautia and Lachnoclostridium, both known to be acetate-producers (Vacca et al., 2020), might suggest a shift in short-chain fatty acid (SCFA) production profiles, although the net functional impact requires further investigation. Notably, the depletion of Faecalibacterium—a genus renowned for its anti-inflammatory properties, production of the SCFA butyrate, and role in maintaining intestinal barrier integrity (Lopez-Siles et al., 2017)—likely contributes to the low-grade systemic inflammation and disruption of the gut environment observed in GDD patients. The reduction of Alistipes and Sutterella, which have been implicated in tryptophan metabolism and mucosal immunity, may further disrupt gut-brain communication (Parker et al., 2020; Kaakoush, 2020). These findings partially align with and extend those reported by Huang et al. (2024), highlighting a conserved gut microbiota signature in neurodevelopmental disorders. Furthermore, accumulating evidence supports the notion that gut microbiota dysbiosis can promote the pathogenesis of GDD through multiple mechanisms, including increased intestinal permeability, activation of glial cells, amplification of inflammatory responses, and induction of oxidative stress (Huang et al., 2024; Munoz-Pinto et al., 2024).

In this study, we employed functional prediction to generate hypotheses regarding the potential metabolic consequences of the observed microbiota dysbiosis. The analysis suggested that the gut microbiome of children with GDD might have an altered functional potential, particularly in three key pathways: 1) mRNA surveillance pathway: Certain gut microbiota may activate localized mRNA translation through abnormal metabolites (such as GABA) (Huang et al, 2022), which interferes with neuronal dendritic development and synaptic connection; 2) Proteasomal pathway: The decrease in butyric acid caused by the absence of faecal bacteria may weaken proteasomal function (Davidson and Pickering, 2023), exacerbate erroneous protein aggregation and neuroinflammation; 3) Atrazine degradation: Up-regulation of the Atrazine degradation pathway suggests that exposure to environmental toxin may accelerate neural stem cell senescence through microbial metabolites(Genovese et al., 2021; Chen et al., 2024), but the exact mechanism needs to be further verified in the context of toxin exposure history.

Compared with the HC, functional prediction based on 16S rRNA data suggested potential upregulation of gut microbiota functions in the GDD group, including mRNA surveillance pathway, Proteasome pathway, and Atrazine degradation pathway. Conversely, significantly downregulated functions included Polyketide sugar unit biosynthesis, Chloroalkane and chloroalkene degradation, and Styrene degradation (Figure 4). The downregulation of polyketide sugar unit biosynthesis may reflect a reduction in the production of secondary metabolites, which have been implicated in maintaining gut-brain homeostasis (Kim et al., 2024). Similarly, impaired degradation pathways for environmental toxins such as chloroalkanes and styrene could indicate a diminished capacity for detoxification, potentially exacerbating neurotoxic effects (Zhou et al., 2024). These functional shifts, while derived from predictive analyses, suggest a broader disruption in microbial metabolic support and environmental resilience in GDD.

The functional predictions (e.g., upregulation of mRNA surveillance and proteasome pathways) were derived from 16S rRNA data using PICRUSt2, which infers metagenomic potential rather than measuring actual gene expression or metabolite levels. Consequently, these findings represent hypotheses about functional consequences rather than direct evidence. The predicted upregulation in the atrazine degradation pathway, for instance, might indicate a microbial response to environmental exposures, but its direct impact on neural stem cells remains speculative without measuring serum or fecal metabolites. It should be noted that these functional predictions are inferred from 16S rRNA sequencing data using PICRUSt2 and have not been validated by metagenomic or metabolomic analyses. Therefore, they represent hypothetical functional shifts rather than confirmed metabolic activities.

As a single-center study conducted in Shenzhen, China, our findings should be interpreted with consideration of regional and institutional factors. The dietary habits, environmental exposures, and genetic backgrounds of the participating children may reflect the characteristics of an urban population in Southern China, which could limit the direct applicability of our results to other geographic or cultural settings. For instance, dietary patterns rich in certain fermented foods or exposure to local environmental toxins (e.g., pesticides) may influence gut microbiota composition. Future multi-center studies involving diverse populations are needed to validate these findings and enhance their generalizability.