Association Between Gut Microbiota and Autism Spectrum Disorder: A Systematic Review and Meta-Analysis

Abstract

Autism spectrum disorder (ASD) is characterized by stereotyped behavior and deficits in communication and social interactions. Gastrointestinal (GI) dysfunction is an ASD-associated comorbidity, implying a potential role of the gut microbiota in ASD GI pathophysiology. Several recent studies found that autistic individuals harbor an altered bacterial gut microbiota. In some cases, remodeling the gut microbiota by antibiotic administration and microbiota transfer therapy reportedly alleviated the symptoms of ASD. However, there is little consensus on specific bacterial species that are similarly altered across individual studies. The aim of this study is to summarize previously published data and analyze the alteration of the relative abundance of bacterial genera in the gut microbiota in controls and individuals with ASD using meta-analysis. We analyzed nine studies, including 254 patients with ASD, and found that children with ASD had lower percentages of Akkermansia, Bacteroides, Bifidobacterium, and Parabacteroides and a higher percentage of Faecalibacterium in the total detected microflora compared to controls. In contrast, children with ASD had lower abundance of Enterococcus, Escherichia coli, Bacteroides, and Bifidobacterium and higher abundance of Lactobacillus. This meta-analysis suggests an association between ASD and alteration of microbiota composition and warrants additional prospective cohort studies to evaluate the association of bacterial changes with ASD symptoms, which would provide further evidence for the precise microbiological treatment of ASD.

Keywords: GI problems; autism spectrum disorder; children; gut microbiota; meta-analysis; microflora.

Figure 1

Flow diagram of study selection: article search strategy results.

Figure 2

Forest plot of percentages of Akkermansia, Bacteroides, Bifidobacterium, and Clostridium in ASD. (A–D) Percentages of Akkermansia, Bacteroides, Bifidobacterium, and Clostridium in the total detected microflora, respectively. Fixed-effects models were used to assess Akkermansia, Bacteroides, and Clostridium. A random-effects model was used to analyze Bifidobacterium, contributing to higher between-study heterogeneity (I ² > 50%). The pooled percentages of Akkermansia, Bacteroides, Bifidobacterium, and Clostridium from the included studies were 0.1%, 11.9%, 2.8%, and 6.4%, respectively.

Figure 3

Forest plot of percentages of Faecalibacterium, Parabacteroides, Ruminococcus, and Bacteroides in autism spectrum disorder (ASD). (A–C) Percentages of Faecalibacterium, Parabacteroides, and Ruminococcus in the total detected microflora, respectively. Fixed-effects models were used to assess Faecalibacterium, Parabacteroides, and Ruminococcus. The pooled percentages of Faecalibacterium, Parabacteroides, and Ruminococcus from the included studies were 8.8%, 0.4%, and 3.3%, respectively. (D) Relative abundance of Bacteroides. A random-effects model was used to analyze Bacteroides, contributing to higher between-study heterogeneity (I ² > 50%).

Figure 4

Forest plot of the relative abundance of Bifidobacterium, E. coli, Enterococcus, and Lactobacillus in ASD. (A–D) Relative abundance of Bifidobacterium, E. coli, Enterococcus, and Lactobacillus. Random-effects models were used to analyze Bifidobacterium, E. coli, and Lactobacillus, contributing to higher between-study heterogeneity (I ² > 50%), except Enterococcus.

Figure 5

Relative abundance of the included bacteria in the meta-analysis. (A) Ratio of the bacterial percentages in children with ASD and typically developing children. A value greater than 1 indicates higher abundance in children with ASD (Faecalibacterium, Ruminococcus, and Clostridium), whereas a value less than 1 indicates lower abundance in children with ASD compared to controls. (B) Relative abundance of the gut microbiota in children with ASD. A positive value indicates higher abundance in children with ASD (Lactobacillus), whereas a negative value indicates lower abundance in children with ASD.