Introduction

Attention Deficit/Hyperactivity Disorder is a neurodevelopmental condition with first symptoms appearing early in childhood. Its prevalence is estimated at 5% in children and 2.5% in adults, with a tendency to increase.1

According to the latest definition and diagnostic criteria of ADHD in DSM-V by the American Psychiatric Association (DSM-V, 2013), the main characteristics of the disorder involve a persistent pattern of inattention and/or hyperactivity/impulsivity that interfere with everyday actions and development. The three recognized subtypes of ADHD are: combined (ADHD-C), predominantly inattentive (ADHD-I) and predominantly hyperactive/impulsive (ADHD-H).1 In the International Classification of Diseases 10th edition (ICD-10) by the World Health Organization, ADHD is classified under “Hyperkinetic Disorders”.2

Clinically ADHD is a condition that appears early in life, before the age of 12 years, consolidates over time and shows great diversity.3 In adults ADHD demonstrates itself mainly with impulsivity, emotional dysregulation, excessive mind wandering and executive functions deficits, while hyperactivity is more subtle.4 In many cases, ADHD coexists with comorbidities, such as intellectual disability, communication and learning disorders. The etiology and pathophysiology of ADHD are unclear, but genetics and several environmental risk factors, such as gut microbiome, seem to play a role.3

The human gut microbiome is very diverse in comparison to other sites of the body, and there appear to be differences even among healthy individuals. Humans are considered to be born sterile and microbial colonization begins immediately at birth. Once established, the microbiome remains relatively stable, but it has been shown that environmental factors, such as nutrition, disease or antibiotics, can cause alterations.5

Different indices are used to describe microbial diversity in the human intestinal tract. Alpha-diversity refers to the microbial variation in a sample and it includes number of different species (richness) and equal balance of these species (diversity) in the sample, measured with specific indices. Simpson’s and Shannon’s indices account for both abundance and evenness of the species present. OTU (count of different species), Observed Species (count of unique OTUs in each sample), Chao1 index and ACE (abundance-based coverage estimators) are richness indices.6 Beta-diversity is the microbial variation between samples.7

When it comes down to the exact composition, the main phyla present in the gut microbiota are Actinobacteria, Proteobacteria, Firmicutes and Bacteroidetes with variations in their numbers throughout life. Fusobacteria, Cyanobacteria and Verrucomicrobia are present in smaller numbers.8 Their representatives in order, family, genus and especially species level can vary significantly among individuals.5,9 The analysis of the microbiota is performed by a number of methods such as sequencing of the 16 S rRNA-encoding gene and comparison to known databases.10 Another method is shotgun sequencing, which analyzes the entire microbial community.8 These two methods are the ones used by the studies included in this review.

The proposed relationship between ADHD and gut microbiome is based on the interplay between the intestinal tract and the central nervous system (CNS), also known as “gut-brain axis”, which is now well accepted.11 This communication interplay involves neural, endocrine and immunal pathways.12 The autonomic nervous system (ANS), vagus nerve, enteric nervous system (ENS), hypothalamus-pituitary axis (HPA), neurotransmitters, hormones, metabolites and, of course, the gut microflora are implicated in the “gut-brain axis” pathways.11 Recent studies have linked the gut microbiota with a variety of mental health and neurodevelopmental conditions, such as autism spectrum disorder (ADS).13,14,15,16,17 Other recent reviews, address the existence of alterations in the gut microbiome of people with ADHD.18,19 It is important to underline the fact that all studies targeting specifically the relationship between ADHD and the human gut microbiome are very recent and the corresponding reviews quite scarce.

Taking into account the above, the goal of this systematic review is to examine the associations between features of gut microbiome and ADHD risk or severity in children, adolescents and young adults.

Materials and methods

Search algorithm and eligibility criteria

This systematic review was performed following the PRISMA guidelines, in line with an a priori protocol agreed upon and signed by all authors. Eligible studies were identified in PubMed, whereas the first, most relevant 1,000 hits from Google Scholar were also screened. The following search algorithm was implemented: (microbiome OR microbiota OR gut OR microflora OR microbes OR microbe OR micro-organism OR micro-organisms) AND (“attention deficit” OR ADHD). There was no restriction regarding publication language; the end-of-search date was December 31, 2020. References of all eligible articles and relevant review articles were meticulously searched to identify additional studies, in a “snowball” procedure.

Cohort, case-control and cross-sectional studies on humans, comparing gut microbiome features (diversity, phylum, class, order, family, genus and species) in subjects with ADHD versus those without ADHD, or along ADHD severity, were deemed eligible. In addition, interventional studies on probiotic administration to enrich microbiome were also retained but were reported separately. Studies on animals, case reports and case series were excluded. In case of overlapping publications, the largest study and the most recent results were retained. Two authors (DG, KM), working independently to each other, performed the selection of studies and, in case of disagreement, consultation with a third author (TNS) and team consensus was reached.

Data abstraction and evaluation of risk of bias

The following data were abstracted from eligible studies: first author, year of publication, journal name, country in which the study was conducted, study design, study period, sample size, number of subjects with ADHD, number of participants free from ADHD, mean age, age range, selection criteria for the study population, method of ADHD diagnosis, ADHD severity, technique of studying gut microbiome, details about intervention (if applicable), and results pertaining to gut microbiome features (diversity, phylum, class, order, family, genus and species). Data was abstracted into previously piloted forms by two reviewers working blindly to each other (DG, KM); in case of disagreement, consultation with a third author (TNS) and team consensus was reached. The assessment of risk of bias for included studies was performed using the Newcastle-Ottawa scale. Regarding the relevant items pertaining to cohort studies, all available studies have been assessed, whereas attrition rate of 10% or less was deemed adequate.

Results

Selection and description of studies

The search in PubMed and Google Scholar identified a total of 1197 items; after the removal of duplicates, 1085 items were screened in title and abstract Fig. 1. Of these 1043 were excluded as irrelevant from title and abstract. Forty-two were evaluated in full-text; of them, 31 were excluded for reasons presented in Supplementary Table 1. The total number of eligible studies was 11, ten of which were case-control studies performed in children, adolescents and young adults diagnosed with ADHD versus healthy controls. Two of these studies were available online only in pre-print format, prior to completion of peer review.20,21 One interventional study investigated the development of ADHD and the differences of gut microbiome after the administration of probiotics to a cohort of children, versus a placebo group.22 Description of eligible studies is provided in Table 1.

Fig. 1: Prisma Flow Chart.
figure 1

The steps regarding the selection of eligible studies are presented in the following chart, according to Prisma guidelines.

Table 1 Characteristics of included studies.

Alpha-diversity and beta-diversity: ADHD cases versus controls

Results of individual studies are presented in detail in Table 2. Alpha-diversity was examined in eight studies; in five studies21,23,24,25,26 no significant difference was found in ADHD cases versus controls. However, Li et al. (2020)21 revealed lower gene numbers in the ADHD group compared to HCs, but no difference in alpha diversity. Prehn-Kristensen et al.27 found one index (Shannon) lower in ADHD cases, but the other indexes (Chao1, observed species) showed no difference. In the study by Wang et al. (2019)28 two indexes (Shannon, Chao1) were higher in ADHD subjects, one index (Simpson) was lower and one (ACE) was similar in both groups.

Table 2 Results of included studies.

Fan et al. (2019)20 showed that three indexes (ACE, Shannon, Chao1) were lower in the ADHD-I group while Simpson Index was higher in the ADHD-I and the ADHD-C groups, compared to healthy controls.

Beta-diversity was examined in eight studies; of them, four23,24,26,28 showed no significant difference whereas the remaining four20,22,25,27 showed compositional differences between the two groups. Szopinska-Tokov (2020)25 specified that the ADHD group had a smaller variation in the gut microbiota composition, which means a higher taxonomic similarity (within the group) compared to healthy controls; however, the remaining three studies did not provide the relevant details.20,21,27

Microbial phylum, class, order and family differences in ADHD versus controls

The microbiota composition at the phylum level was evaluated in eight studies. Of them, four24,25,27,29 found no difference. Aarts et al. (2017)23 showed an increase in Actinobacteria and a decrease in Firmicutes in ADHD subjects versus controls. Li et al. (2020)21 showed that healthy controls had higher numbers of Fusobacteria, while Wang et al. (2019)28 revealed higher numbers of Fusobacteria in the ADHD group. When it comes to ADHD subgroups, only Fan et al. (2019)21 suggested decreased numbers of Verrucomicrobia in ADHD-I.

None of the eligible studies gave information about the difference in abundance of microbial classes between ADHD and healthy controls.

Information about the microflora composition at the order level was investigated in four studies. Aarts et al. (2017)23 found Clostridiales, within the phylum Firmicutes, to be decreased in ADHD and Bifidobacteriales increased. On the other hand, Cheng et al (2019)28 found Clostridiales increased in ADHD. Akram et al. (2017)29 reported differences in Burkholderiales, Alcaligenaceae and Erysipelotrichaceae without any further details. Li et al. (2020)21 mentioned lower numbers of Fusobacteriales, Flavobacteriales, Rhodocyclales and Sphigomonadales in ADHD and especially in ADHD-I lower Rhizobiales, versus controls.

The microbial families represented in the gut microbiome were reported in five case-control studies. Aarts et al. (2017)23 found that Rikenellaceae, Porphyromonadaceae and Bifidobacteriaceae were more abundant in ADHD, while Ruminococcaceae and Lachnospiraceae were increased in healthy controls. Jiang et al. (2018)24 found a lower percentage of Alcaligenaceae and higher of Peptostreptococcaceae, Moraxellaceae, Xanthomonadaceae, and Peptococcaceae in the ADHD group. Prehn-Kristensen et al. (2018)27 found elevated levels of Prevotellaceae, Catabacteriaceae, and Porphyromonadaceae in healthy controls, but Neisseriaceae and Bacteroidaceae in ADHD children. In the study by Li et al. (2020),21 Fusobacteriaceae, Flavobacteriaceae, Rhodocyclaceae, Sphigomonadaceae and Bacillales noname were less abundant in the ADHD patients than in healthy controls, while Prevotellaceae was more abundant in the ADHD patients. In regard to ADHD subgroups, the study concluded that Oscillospiraceae was more frequently encountered in healthy controls and ADHD-I compared to ADHD-C, whereas Listeriaceae, Prevotellaceae and Veillonellaceae were more frequently encountered in ADHD-C compared with ADHD-I or HC. According to Fan et al. (2019)20 Ruminococcaceae were decreased in ADHD-I and ADHD-H; Lachnospiraceae, Verrucomicrobiaceae and Rikenellaceae were decreased in the ADHD-I group and Prevotellaceae were increased in ADHD-C and ADHD-H, versus controls.

Microbial genus and species in ADHD cases versus controls

Regarding the genus level, all eligible case-control studies reported the relevant results. According to the findings by Aarts et al. (2017),23 Bifidobacterium, Eggerthella, Alistipes, Parabacteroides and Odoribacter were increased in ADHD, whereas Subdoligranulum, Ruminococcus and Coprococcus were more abundantly encountered in healthy controls. Akram et al. (2017)29 reported differences in Phascolarctobacterium, Paraprevotella, Veillonella and Odoribacter, without further specifications. According to Cheng (2019),30 Desulfovibrio was more abundant in ADHD. According to Prehn-Kristensen et al. (2018),27 Prevotella and Parabacteroides were detected as markers for the control group and Neisseria for the ADHD group. In the study by Szopinska-Tokov et al. (2020),25 Ruminoclostridium 9, Ruminococcus 2, Clostridiales g, Ruminococcaceae NK4A214 group, Ruminococcaceae UCG 003, Ruminococcaceae UCG 004, Ruminococcaceae UCG 005, Ruminococcaceae g uncultured, Family XIII AD3011 group were more frequent in ADHD and Haemophilus in HC. According to the findings by Wan et al. (2020),26 Faecalibacterium and Veillonella were significantly reduced in the ADHD group, while Odoribacter and Enterococcus were significantly more abundant. Wang et al. (2019)28 concluded that Fusobacterium was elevated in the ADHD group, while the relative abundance of Lactobacillus was enriched in healthy controls. Finally, Jiang et al. (2018)24 reported decreased numbers of Faecalibacterium, Lachnoclostridium, Sutterella and Dialister in the ADHD group.

Fan et al. (2019)20 found greater abundance of Megamonas, Coprococcus 2 and Paraprevotella in the ADHD-C, relatively to healthy controls. The aforementioned study also mentioned lower percentage of Faecalibacterium and higher of Marvinbryantia, Intestinimonas, Prevotella 9 and Eggerthella in ADHD-H subjects compared to healthy controls, but reduced Akkermansia, Ruminococcaceae UCG002, Lachnospiraceae NK4A136 group, Eubacterium coprostanoligenes group, Christensenellaceae R-7 group, Ruminococcaceae UCG014, Ruminococcaceae UCG005, Unclassified f Lachnospiraceae, Anaerotruncus, Coprococcus 1, Ruminoclostridium 5, Alistipes, norank f Bacteroidales S24-7 group in ADHD-I. Li et al. (2020)21 found Prevotella and Scardovia more frequent in ADHD, whereas Subdoligranulum, Phascolarctobacterium, Adlercreutzia, Fusobacterium, Gemella, Methyloversatilis and Brevundimonas were more frequent in healthy controls. Regarding ADHD subgroups, the study found that Prevotella and Listeria were more frequent in ADHD-C, while Bifidobacterium, Subdoligranulum, Bilophila, Oscillibacter and Acidaminococcus were more frequent in ADHD-I.

Four studies presented information about the species composition. Aarts et al. (2017)23 concluded that Coprococcus eutactus was increased in HC, Bacteroides (vulgatus, ovatus, uniformis), and Bifidobacterium (longum, adolescentis, pseudocatenulatum) were increased in ADHD. In the study by Wan et al. (2020)26 Faecalibacterium prausnitzii, Lachnospiraceae bacterium, and Ruminococcus gnavus numbers were significantly decreased in the ADHD group, while Bacteroides caccae, Odoribacter splanchnicus, Paraprevotella xylaniphila, and Veillonella parvula were significantly increased. Wang et al. (2019)28 found that Bacteroides coprocola in the ADHD group was significantly less frequent than in the control group, while the relative abundance of Bacteroides uniformis, Bacteroides ovatus, and Sutterella stercoricanis in the ADHD group were significantly more frequent than in the control group.

According to Li et al. (2020),21 Bacteroides (ovatus, fragilis, thetaiotaomicron, intestinalis, cellulosilyticus, salyersiae, fluxus, nordii) were more abundant in healthy controls, whereas Bifidobacterium (breve, bifidum) and Prevotella (amnii, buccae, copri) were more abundant in ADHD. Especially regarding ADHD subgroups, Bacteroides (cellulosilyticus, fluxus, nordii, ovatus), Lachnospiraceae bacterium, Bilophila wadsworthia, Oscillibacter unclassified and Subdoligranulum unclassified were retrieved in higher numbers in healthy controls and I-ADHD compared to C-ADHD, whereas Listeria marthii was more abundant in C-ADHD compared to I-ADHD or healthy controls.

Gut microbiome and ADHD severity

Associations with the severity of ADHD were reported in four studies. Jiang et al. (2018)24 concluded that Faecalibacterium was negatively associated with ADHD severity and hyperactivity. According to Li et al. (2020),21 the abundant species Prevotella buccae, Bifidobacterium breve and Bifidobacterium bifidum in ADHD-C compared with healthy controls, were enriched and positively associated with the results regarding ADHD of both total Conners Parent Rating Scales (CPRS) and DSM. Increased relative abundances of Bacteroides nordii, Bacteroides cellulosilyticus and Bacteroides intestinalis were associated with fewer symptoms in both hyperactivity/impulsivity and inattention, while Bacteroides thetaiotaomicron and Bacteroides ovatus were negatively associated only with inattention scores. In the study by Prehn-Kristensen et al. (2018),27 levels of hyperactivity significantly correlated with a change in alpha-diversity, whereas Bacteroides species levels correlated with levels of hyperactivity and impulsivity. According to Szopinska-Tokov et al. (2020),25 the variation in beta-diversity was explained by disorder status and inattention level. Also Ruminococcus 2 and Ruminococcaceae UCG 004 were associated with higher inattention levels.

Randomized trial results

In regard to the randomized, double-blind,placebo-controlled prospective follow-up study by Pärtty et al. (2015),22 at the age of 13 years, three children out of 75 total were diagnosed with ADHD (4%) and two children with both ADHD and ASD (2.7%); all these children were in the placebo group. At the age of 3 months, significantly lower median numbers of Bifidobacterium longum were noted among the children with neuropsychiatric disorders versus healthy children. At the age of 6 months, when the probiotic intervention was completed, the genus Bifidobacterium was significantly less frequent among children with neuropsychiatric disorders than in those without. At the age of 18 months, Bacteroides and Lactobacillus-Enterococcus group bacteria were less frequent among children with ADHD versus healthy children. At the age of 24 months, the Clostridium histolyticum group was less abundant among children with ADHD. At the age of 13 years, there were no statistically significant differences in gut microbiota composition between children with or without neuropsychiatric disorders. The randomized trial did not present data about alpha-, beta-diversity and composition at class, order and family level.

Risk of bias

The assessment of risk of bias is presented in Supplementary Tables 2 and 3. As far as selection of ADHD patients and healthy controls (or exposed / unexposed cohort) is concerned, all studies complied with low risk of bias in the Newcastle—Ottawa Quality Assessment Scale (for case—control and cohort studies respectively). Some problems occurred in the comparability of the two groups, as only in two studies they were matched for all confounding factors. Finally, none of the studies gave information on non-responder rates.

Discussion

This systematic review highlighted differences in gut microbiome features in association with ADHD occurrence and severity. Convergence of results between studies was observed while studying microbial genera. Aarts et al. (2017)23 and Wan et al. (2020)26 both agreed upon an increased abundance of Odoribacter in the ADHD group versus controls; Akram et al. (2017)29 also reported a difference in the abundance of Odoribacter. An effect on neurotransmitter production and dopamine metabolism has been attributed to Odoribacter31 which could potentially contribute to the occurrence of ADHD. Odoribacter contributes to the production of short-chain fatty acid (SCFA).32 These acids have neuroactive and anti-inflammatory effects and in high levels have been shown to worsen ASD symptoms.33,34 Similar effects have been hypothesized for ADHD.33,35 Higher numbers of Odoribacter have also been associated with other neurodevelopmental disorders such as autism spectrum disorders.34 On the other hand, in individuals with ADHD, a decrease of activity of the dopamine reward system has been established.36

Increased abundance of Eggerthella was found in ADHD according to two studies20,23 and this bacterial genus has been linked to dopamine metabolism as well.37 Faecalibacterium was found by three studies in decreased abundance in ADHD.20,24,26 Low numbers of Faecalibacterium have been linked to inflammation through the release of cytokines, which have been reported to be higher in ADHD children38,39 indicating another possible pathway to ADHD. Faecalibacterium has been also found decreased in adults suffering from psychiatric conditions, such as bipolar disorder and depression.40,41 According to Jiang et al. (2018),22 Faecalibacterium was negatively associated with ADHD severity and hyperactivity.

At the species level Wan et al. (2020)26 remained consistent with the findings pertaining to bacterial phyla, reporting lower abundance of Faecalibacterium prausnitzii and higher of Odoribacter splanchnicus and thus strengthening the hypotheses mentioned above.

Another bacterial species found in greater abundance in the ADHD group, according to two studies23,28 was Bacteroides uniformis. This particular species has been suspected of having a role in the development of parts of the brain, such as the frontal lobe and the hippocampal region.42 In individuals with ADHD certain regions of the brain, such as prefrontal cortex, basal ganglia and parietal lobe show changes when compared to healthy controls.43 Similar changes are observed in the connectivity between these regions in ADHD patients.44 Wang et al.(2019)28 suggested that increased frequency of B.uniformis could be related to ADHD susceptibility and Prehn-Kristensen (2018)27 concluded that Bacteroides species levels correlated with levels of hyperactivity and impulsivity.

Regarding alpha diversity, no consistent differences arose between ADHD patients and healthy controls. The findings of two studies20,21 indicated though that more information should be obtained; afterall according to Prehn-Kristensen et al. (2018)27 levels of hyperactivity were significantly correlated with changes in alpha diversity. The results regarding b-diversity were also inconclusive, whereas only half the studies came up with results. Studies that examined21,23,28 the microbial phyla did not seem to agree, so no safe conclusion can be derived at this point. Similarly, the results about differences in the gut microbiome at the order and family level were contradictory when comparing ADHD subjects versus healthy controls; nevertheless, two studies agreed upon an increase in abundance of the Prevotellaceae family in ADHD-C cases.20,21

Regarding the randomized trial by Pärtty et al. (2015)22 it is interesting to mention that, although differences were noted up to the age of 24 months, by the time of the last examination of participants at the age of 13 years, there was no significant difference in the consistency of the gut microbiome between the children that were diagnosed with ADHD versus those that were not. (Table 3).

Table 3 Table showing the most important findings on gut microbiome composition, between ADHD cases and controls.

As deduced from the above mentioned results, a number of findings of the included studies do not seem to correlate or overlap. A few reasons could be suggested for this, such as different geographic regions, variations in age and different dietary habits of the participants.

Various limitations of this systematic review can be addressed. First, no quantitative synthesis (meta-analysis) was attempted, in view of the small number of eligible studies and marked heterogeneity in the reporting or results and indices across the individual published reports. In addition, the vast majority of the evidence stemmed from case-control studies; longitudinal cohort studies are needed to further validate the relevant hypotheses. Concerning the external validity of results, various world regions, such as Latin America, Australia and Africa were not represented in the body of evidence. Variability in the age groups and dietary habits should also be addressed by future studies as potential modifiers of the association between gut microbiome and ADHD.45

In conclusion, even correlations between gut microbiome features and manifestation of ADHD symptoms seem to emerge, with plausible mechanistic explanations, additional studies are needed to further investigate the interplay between intestinal microflora, ADHD occurrence and severity.

The datasets analyzed during the current study are available from the corresponding author on reasonable request.