Introduction

Attention-deficit hyperactivity disorder (ADHD) is a neurodevelopmental disorder affecting 7.2% children below 18 years old globally [1]. The prevalence keeps increasing for the last 20 years [2]. Although people with ADHD generally have good quality of life, some of them may experience difficulties in navigating daily life. Without proper treatment, ADHD may lead to some serious consequences, such as academic failure, social disruption, unwanted accidents, strained family relationships, and disorganized career. Furthermore, ADHD is a disorder that affects many aspects of life, including financial, psychosocial, and cultural aspects. The financial burden of ADHD ranges from $503 to $1.343 annually, which comprises hospitalization expenses, psychiatry consultations, and medicines [3]. In addition, it is reported that parents tend to resign from their jobs to provide more attention to their ADHD children. Children with ADHD also need extra help from their teachers at school, but such assistance is not always optimally provided due to the workload of said teachers [4].

Gut–brain axis (GBA) is a system widely studied right now because of the novel understanding that the gut microbiota environment can affect brain activity and vice versa. The bidirectional communication involves the central nervous system, brain and spinal cord, autonomic nervous system, enteric nervous system, and hypothalamic pituitary adrenal (HPA) axis [5]. Unbalanced microbiota composition, known as dysbiosis, is caused by the increase in inflammatory microbes that may impair gut permeability. This in turn can cause microbial translocation, which leads to systemic inflammation. A systemic inflammation may trigger the disruption of blood–brain barrier and increase the level of pro-inflammatory cytokines such as IL-6 and IFN-γ. In addition, dysbiosis may produce oxidative stress that affects neuron cell and neurotransmitters related to ADHD [6].

A study conducted in animal models has found that several taxa differed significantly between mice that were colonized by ADHD microbiota and healthy control mice. Fourteen genera increased, while 17 genera were found to be more abundant in control. Mice with transplanted microbiota displayed abnormalities in their brain such as decreased integrity in both white and gray matter regions. Other than that, the MRI result showed a decreased resting-state connectivity between right motor and visual cortices [7].

Finally, several studies have compared the gut microbiota in ADHD with healthy individuals. We systematically reviewed the scientific literature of case–control studies focusing on gut microbiota composition in ADHD.

Methods

Literature search for gut microbiota studies in ADHD

The population of this systematic review is ADHD patient, the intervention is the profiles of gut microbiota, the comparator is healthy control, and the outcome of this study is to find any differences in gut microbiota profiles between ADHD and healthy controls. The literature search was conducted using Google Scholar, PubMed, and Science Direct with keywords “ADHD”, “gut microbiota”, “stool”, “gut”, and “microbiota”. The studies were then selected based on the inclusion and exclusion criterias. The inclusion criteria include: (1) all studies must be in English and discuss gut microbiota in ADHD; (2) the samples in the studies were diagnosed based on Diagnostic and Statistical Manual of Mental Disorders (DSM) 4th or 5th edition or ICD-10 with the code F90.9; (3) the studies conducted in human and the gut microbiota were identified using fecal samples. We excluded studies with samples below ten person(s) and studies that focus on other condition besides ADHD, such as interventional study.

Results

Literature search

The studies selection was conducted using Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [8]. From the database search, we found 146 studies with the keywords. We removed duplication and found 131 studies. After screening through the titles and abstracts, we found 122 potential studies to be reviewed. From that number we narrowed it down to 9 eligible studies to be reviewed, but 3 must be removed (one was an animal study, one was an interventional study, and another was in Spanish). Finally, we found six case–control studies ranging from 2017 until 2020, involving 407 research participants, 172 ADHD, 15 subthreshold ADHD, and 220 healthy controls (see Fig. 1).

Fig. 1
figure 1

PRISMA

Characteristics of studies (see Table 2 )

Samples in studies

Three out of six studies were conducted in Asia 9,10,11] (Beijing, Taiwan, and Zhejiang). Two other studies were conducted in the Netherlands [12,13] and the other one was conducted in Germany [14]. The samples for each study varied between 14 and 77 participants.

Dietary pattern

Two studies [12,13] did not explain about the dietary pattern in their samples, but the healthy controls from study by Aarts et al. were selected from the healthy siblings of the ADHD patients. Therefore, the comparison between gut microbiome in ADHD patients and healthy controls might be more accurate, as dietary variations can be eliminated from the factors influencing gut microbiome, further homogenizing samples. One study [14] recorded the dietary pattern only for fast food, meat/sausage/cold cuts, fruits/vegetables, and yoghurt/dairy products. Two studies used questionnaires for their samples’ dietary log; one used a questionnaire filled by patients’ parents [9] and the other one used Food Frequency Questionnaire (FFQ) [10]. The last study [11] stated that their samples were asked to maintain their regular dietary pattern for 1 week and record their food intake using food diary.

ADHD medication and consumption of probiotic/prebiotic/antibiotic

Out of all studies, only two studies [13,14] included individuals with ADHD medication. Two other studies [9,10] excluded those who have a history of ADHD medication, while the rest [11, 12] did not mention this criteria in their articles. Other than that, only one study [9] excluded prebiotic/antibiotic use within 2 months prior to samples collection. One study [10] also excluded prebiotic/antibiotic use, but did not explain the duration of consumption. On the other hand, one study [11] specifically excluded the use of probiotic within 1 month prior to samples collection. Finally, the last study [13] stated that they could not evaluate the use of probiotic/prebiotic/antibiotic in their samples.

Psychiatric condition, allergic history, and other medical conditions

Two studies [9, 11] excluded depressive and anxiety symptoms in their samples along with atopic history (asthma, eczema, and allergic rhinitis). In addition, they also excluded samples with a history of digestive or chronic diseases. Only one study [10] stated that they excluded all neuropsychiatric conditions except ADHD and other major diseases. The other studies did not give any information about this criterion. Other than that, the study by Jiang et al. [9] excluded samples with obesity and Wan et al. [11] excluded samples with body mass index (BMI) below 20 kg/m2.

Methodology

Clinical assessments

The diagnostic tools included K-SADS (Kiddie Schedule for Affective Disorders and Schizophrenia) based on DSM-IV, DSM-IV-TR, and DSM-5. Three studies [11,12,13] used K-SADS to diagnose their samples. Two other studies [9, 14] used K-SADS-PL, the present and lifetime version which is used for affective and psychotic disorders screening, as well as other disorders, such as major depression disorder (MDD), maniac, bipolar disorder, schizophrenia, schizoaffective disorder, generalized anxiety disorder (GAD), obsessive–compulsive disorder (OCD), ADHD, conduction disorder, anorexia nervosa, bulimia, and post-trauma stress disorder (PTSD). The last study [10] used K-SADS-E, the epidemiological version of K-SADS (see Table 1).

Table 1 Study population characteristics

Microbiota analysis

Four out of six studies used 16s rRNA sequencing, but in different regions. Three studies [9, 10, 12] analyzed region V3–V4 and the other study [13] analyzed region primer 27F-DegS(5′GTTYGATYMTGGTCAG)-338RI-II (5′GCWGCC [T/A] CCCGTAGG [A/T] GT). In addition, the study by Prehn-Kristensen et al. (2018) used DNA amplification on primer 27F-338R and the last study by Wan et al. used shotgun metagenomic sequencing (see Table 2).

Table 2 Study population methodology

Diversity analysis (see Table 3)

Table 3 Microbial diversity in each study

Five studies [9,10,11,12, 14] identified α-diversity using Shannon and Chao1 index. The study by Wan et al. also used Simpson index, while Jiang et al. also analyzed the Abundance-based Coverage (ACE) and Simpson index. Only one study [13] used Shannon–Wiener index and Faiths’ phylogenetic diversity.

Two studies [11, 12] did not analyze the β-diversity of their microbiota profiles. Three [9, 10, 13] out of four studies that assessed β-diversity used unweighted and weighted UniFrac and principal coordinates analysis (PCoA). The study by Szopinska-Tokov et al. also assessed ADONIS index and betadisper, while Prehn-Kristensen et al. additionally used ADONIS, ADONIS, and betadisper for their study.

Findings (see Table 4 )

Table 4 Microbiota profile found significantly different in ADHD

The total of 49 bacterium taxa were found to be significantly (p < 0.05) different between ADHD and healthy controls. The result of each study varied that it was difficult to point out which bacteria taxa differed the most in ADHD. Study from Aarts et al. stated that Clostridiales order, Rikenellaceae and Porphyromonadaceae families, and Bifidobacterium and Eggerthella genera might be the potential markers for ADHD. The change in Bifidobacterium abundance is related to phenylalanine pathway for carbohydrate-deficient transferrin. Genus Dialister decreased in the study by Jiang et al. [9], but increased in medicated samples in study by Szopinska-Tokov et al. [13]. Genus Faecalibacterium decreased in the study by Jiang et al. [9] and Wan et al. [11], but it is known that the amount of this genus does not correlate with CPRS (Conners Parent Rating Scale) score and the hyperactivity index score [9].

At OTU level, study by Prehn-Kristensen et al. [14] found an increased level of Bacteroides OTU_7 and Bacteroides OTU_577. The abundance of this genus varied between studies. A study by Wang et al. [10] found increased levels of B. uniformis and B. ovatus, but decreased level of B. coprocola. In another study by Wan et al. [11], they found that the B. caccae level was increased. The abundance of B. uniformis, B. ovatus, and S. sterrcoricanis were found to be strongly correlated with the consumption of fat and carbohydrate, and also with ADHD symptoms. In addition, S. sterrcoricanis is also associated with specific food intake, including milk, nuts, ferritin, and magnesium.

Family of Ruminococcaceae was found to be increased in the study by Szopinska-Tokov et al. [13]. The Ruminococcaceae_UCG_004 genus especially is associated with inattention score, but not with hyperactivity/impulsivity score. Interestingly, ADHD medication does not affect the abundance of this genus. In this study, it is also found that Phascolarctobacterium is decreased in ADHD-medicated samples.

Discussion

Main findings

The results are varied between each study and all studies have the distinct taxa findings between ADHD and healthy control groups. Thus, there was minimal consensus that examined which bacterial taxa correlated most to ADHD.

Risk of bias

Differences in sample

Out of six studies, only one study by Wan et al. [11] displayed their age information in median, thus explaining that the distribution of their samples is abnormal. It has been discussed that age is related to the composition of gut microbiota [15]. Other than that, a study by Szopinska-Tokov et al. [11] also included samples from subthreshold ADHD. Study by Prehn-Kristensen et al. [14] is the only study with all-male samples. This actually made the samples more homogenous as gender is also associated with gut microbiota composition. In mice model study, it is found that Allobaculum, Anaeroplasma, and Erwinia genera are more abundant in male mice, while SMB53 genus from Clostridiaceae family and Dorea, Coprococcus, and Ruminococcus genera are more abundant in female mice [16].

Dietary patterns in sample

A study in Korea found that high consumption of fast foods, soft drink, and instant noodles is highly associated with increased K-ARS (Korean version of ADHD Rating Scale) score [17]. Vegetarian diet pattern is known for its effect in increasing the amount of protective species and decreasing the pathologic ones like Enterobacteriaceae. In addition, vegetarian diet pattern may reduce inflammation caused by the increased level of Bacteroides fragilis and Clostridium species, leading to decreased level of intestinal lipocalin-2 and short-chain fatty acids. Lipocalin-2 is a biomarker for inflammation [18]. In ADHD, it is known that IL-6 level is increased because of the gut microbiota dysbiosis [19]. By excluding vegetarian diet pattern, the study may better represent the normal majority of population.

Body mass index effect on samples

BMI can affect the composition of gut microbiota, especially those who are overweight or obese. There was an increased ratio of Firmicutes:Bacteroides in obese children. In addition, the concentration of B. vulgatus decreased while Lactobacillus spp. concentration increased. S. staphylococcus was also known to correlate with energy intake [20]. BMI should be included as a variable in conducting this type of study.

ADHD medication effect on samples

There was a decreased α-diversity in ADHD-medicated samples in study by Prehn-Kristensen et al. [14]. In other studies, Szopinska-Tokov et al. [13] found increased level of Dialister genus in ADHD-medicated samples, while Jiang et al. [9] found decreased level of this genus as they only included samples who have never been on ADHD medication. The genus Dialister is known for its potential in modulating gamma-aminobutyric acid (GABA) neurotransmitter, which may play a role in the pathogenesis of ADHD [21]. We suggest that this genus may be the biomarker for ADHD-medicated patients. There should be a further study just to compare those who have never taken any ADHD medication and those who have taken ADHD medication to find if there are any differences between two groups.

Prebiotic/probiotic/antibiotic consumption

Consumption of prebiotic/probiotic/antibiotic will affect the gut microbiota composition. For example, Bifidobacterium infantis can decrease pro-inflammatory cytokine in animal model study with irritable bowel syndrome (IBS). Moreover, it is known for its anti-inflammatory effect when consumed with α-linoleic [22]. Other effects of prebiotic/probiotic are widely studied in other disorders such as autism spectrum disorder (ASD), but further study is still warranted to elaborate its benefits in neuropsychiatry disorder [23].

Antibiotics such as penicillins, quinolones, macrolides, sulfonamides, and anti-tuberculosis agents are claimed to induce psychotic behavior. These antibiotics may disrupt the microbiota’s metabolism, therefore affecting the neurotransmitter and SCFA (short-chain fatty acids) which will lead to brain disturbance through GBA [24]. In addition, antibiotics use may also relieve anxiety symptoms and may be helping those with ADHD achieve a calmer mental state. This in turn can possibly affect the levels of gut microbiota in ADHD patients [25]. Furthermore, in an animal model study, an antibiotic regimen of bacitracin, neomycin, and primaricin were given to mice for 7 days and resulted in an increased population of Lactobacilli and Actinobacteria. This change indicates that antibiotics consumption can alter normal gut microbiota composition.

Neuropsychiatric disorder and other diseases effects on samples

Other neuropsychiatric disorders such as depression, schizophrenia, bipolar, Alzheimer’s and Parkinson’s diseases are extensively studied about their correlation with gut microbiota. Several genera are known to induce depression through HPA axis, GABA, SCFAs, immune system and gut barrier [26]. In patient with bipolar disorder, Actinobacteria phylum and Coriobacteria class were found more abundant compared to healthy individuals [27]. In MDD, the amount Lactobacillus (P = 0.067) and Bifidobacterium (p = 0.012) has been found to be lower [28]; whereas in schizophrenia, it is found that the abundance of Proteobacteria was increased. Six genera were found increased (Succinivibrio, Megasphaera, Collinsella, Clostridium, Klebsiella, and Methanobrevibacter), while three other genera were found decreased (Blautia, Coprococcus, Roseburia) [29]. In addition, increased abundance of Escherichia/Shigella was found in Alzheimer’s disease and increased Lactobacillaceae was found in Parkinson’s disease [30]. Therefore, by excluding other neuropsychiatric disorders, the study may be more accurate. Other diseases, like infection, may affect or be affected by gut microbiota. In human immunodeficiency virus type-1 (HIV-1), for example, it is said that gut microbiota may predict the immune status of the patient. The genera composition in HIV-1 was heavily disturbed and antiretroviral (ARV) medication was implicated with the decrease of Prevotella genus [31]. A study in 2018 found that 17 common diseases were associated with at least one microbiota marker (false discovery rate (FDR) < 0.05). Given the example, decreased Ruminococcaceae is associated with irritable bowel syndrome [32]. Other than that, type 2 diabetes mellitus (T2DM) was also found to correlate with a smaller amount of phyla Firmicutes and Clostridia [33].

Possible link between gut microbiota and ADHD

GABA-producer microbiota role in pathogenesis of ADHD

The role of GABA in the pathogenesis of ADHD is still unknown. A study in 2012 found decreased GABA level in ADHD [34]. On the contrary, a study in 2015 stated that adults with ADHD have increased concentration of GABA+, but not in children. This finding also explain that age is positively correlated with GABA+ concentration and ADHD [35]. Another study in 2016 said that there was a negative correlation between GABA level with impulsivity and aggression score [36]. We found several interesting findings in the studies reviewed in this paper. Bifidobacterium [12] genus and Peptostreptococcaceae [9] family were found increased in ADHD. Furthermore, four species from Bacteroides spp. were found increased, while one species was found decreased [10, 11]. Those taxa were known as the GABA producer, especially Bifidobacterium. This genus is known to be the most efficient GABA producer [37]. Thus, we contend that GABA may play a role in the pathogenesis of ADHD, but further study is warranted to learn the exact mechanism of this link.

Norepinephrine and dopamine role in pathogenesis of ADHD

Though the exact pathogenesis of ADHD is still unclear, if we deduce the mechanism of action of methylphenidate, we can assume that norepinephrine and dopamine play a big role in ADHD. Methylphenidate works by inhibiting the reuptake of dopamine and norepinephrine. D1 dopamine receptor activation (DRD1) regulates NOD-like receptor protein 3 (NLRP3) through cyclic adenosine monophosphate (cAMP), thus destroying NLRP3 in all sites helped by E3 ubiquitin MARCH7. Dopamine and DRD1 may reduce neurotoxin-induced neuroinflammation, lipopolysaccharide (LPS)-induced systemic inflammation, and monosodium urate crystal (MSU)-induced peritoneal inflammation [38]. Norepinephrine, on the other hand, may trigger the growth of protective bacteria like Escherichia coli [5]. Unfortunately, from all the studies we reviewed, we cannot find a specific gut microbiota that has the ability to modulate or produce dopamine nor norepinephrine. It is still unknown how methylphenidate can affect the composition of gut microbiota of ADHD patients. As stated before, we suggest that the Dialister genus may be the biomarker for those with ADHD medication, but further research is needed to identify the mechanism between methylphenidate and gut microbiota.

Micronutrient supplementation in ADHD and its effect on gut microbiota

A study in 2019 tried to compare the gut microbiota composition between ADHD children who were given micronutrient treatment and those with placebo. It was found that Bifidobacterium genus is decreased in children who were given the micronutrient treatment. The decreased level of Bifidobacterium was in line with a decreased in ADHD-IV-RS score [39]. A study by Jiang et al. [9] found an increased abundance of Bifidobacterium in ADHD patients. This could hint that micronutrients may restore the balance of Bifidobacterium, also that this genus may play a big role in the pathogenesis of ADHD. As explained before, Bifidobacterium is one of the many bacteria that can produce GABA. The role of GABA in ADHD still needs to be explored and further study regarding the connection between Bifidobacterium and its effect on ADHD symptoms still needs to be analyzed.

Effect of pro-inflammatory cytokines modulated by gut microbiota to ADHD pathogenesis

Two studies found a decreased amount of Faecalibacterium genus [9, 11]. This genus is known for its anti-inflammatory factors, and a decreased level of this genus may lead to overproduction of pro-inflammatory cytokines [11], thus causing systemic inflammation contributing to ADHD pathogenesis. A study reviewed the possible link between pro-inflammatory cytokines as well as the gene modulating them with ADHD and found increased IL-6 and IL-10 in cytokine protein level [40]. Microglia is activated by pro-inflammatory cytokines, while at the same time also promoting its production. These pro-inflammatory cytokines will lead to a neuroinflammation that may contribute to the pathogenesis of ADHD [41]. Moreover, synaptic plasticity and neurogenesis are also affected by pro-inflammatory cytokines that may alter cognitive processes, such as working memory and reaction time [42]. Further study is still needed to learn more about pro-inflammatory cytokines which are regulated by gut microbiota. This is particularly important as it is a known fact that inflammation process is associated with ADHD pathogenesis.

Conclusion

Studies comparing the gut microbiota condition between ADHD and healthy individual are still limited. So far there is no common agreement on which bacterial taxa is most relevant to the incidence of ADHD, thus the link between gut microbiota and ADHD remains unclear. Numerous criteria, such as sample size, gender, BMI, dietary pattern, use of prebiotic/probiotic/antibiotic, and history of ADHD medication, should be taken into consideration in conducting this study in the future. In addition, further studies regarding neurotransmitter-modulated gut microbiota are needed, as there are many bacteria whose function remains undiscovered.