This study was performed according to the Preferred Reporting for Systematic Reviews and Meta-analysis (PRISMA) (37). A systematic literature search was performed in PubMed, SCOPUS, and APA PsycInfo. Articles were included up to May 2021. Combinations of the following keywords were used: “probiotics” and “Alzheimer's disease.” These words were added to the search query together with synonyms or MeSH terms. Duplicates have been removed from the final study selection. After selecting relevant titles and abstracts selection, full-text articles were assessed. A PRISMA flow chart is used to graphically display the final selection of articles (see Figure 1).

Full-text articles were considered eligible if they met the following criteria: (1) the article is peer-reviewed, (2) the article is written in English, (3) a controlled study was conducted on probiotic supplementation, (4) the study uses an animal model of AD, and (5) at least one of the following outcome measures was addressed: cognitive functioning, gut microbiota composition, Aβ deposition, inflammatory markers, and SCFAs. Besides gut microbiota composition, gut permeability or other gut health markers were not included due to a lack of available data. The exclusion criterium was when a study uses mixtures of probiotics with other nutrients or interventions, so that the effect of probiotics cannot be distinguished.

The outcome measures of this study included gut microbiota composition, SCFAs, inflammatory markers, Aβ, and cognitive functioning. Gut microbiota composition encompassed species richness, microbial abundance, and microbiome diversity. Inflammatory markers comprised biomarkers that reflect the pro- or anti-inflammatory status, assessed either in blood or CSF. Finally, cognitive functioning was defined as mental abilities, including learning, thinking, reasoning, remembering, problem solving, decision making, and attention (38).

Data extraction was done on the following items: publication year, country, study sample characteristics of the intervention and control group (e.g., animal model; control group; sample size; sex), intervention characteristics (e.g., type, duration, and dosage of probiotics; type and duration and dosage of control condition), the operationalization of the outcome measures, and results. The extracted data were compared and analyzed for similarities and differences.

The risk of bias was assessed with SYRCLE (39), because this tool was specifically designed to assess the methodological quality of animal intervention studies. SYRCLE is based on the widely used Cochrane Collaboration Risk of Bias tool and was adapted to aspects of bias that specifically play a role in animal research (39).

Ten entries of potential biases in the included studies were reviewed, which are related to selection bias, performance bias, detection bias, attrition bias, reporting bias and other biases. Each potential bias was evaluated by answering SYRCLE's signaling questions with “yes” indicating a low risk of bias, “no” indicating a high risk of bias, and “no information” indicating an unclear risk of bias (39). A summary score was calculated based on the number of yeses compared to the total number of SYRCLE items.

The systematic literature search resulted in 633 studies from which 65 duplicates were removed. The title and abstract of 568 articles were screened for in- and exclusion criteria. In total, 24 articles were read in full-text. Out of these, six articles were excluded (40–45), because they were inaccessible, animals were supplemented with mixtures of probiotics with other nutrients, or had a different focus (glucose uptake, rough eye phenotypes, DNA, oxidation and SIRT1), leaving eighteen articles for analysis. The final PRISMA flow chart can be found in Figure 1.

In total, 18 articles were assessed for methodological quality by means of the SYRCLE tool (see Table 1). Although the SYRCLE score does not have an overall quality score, most articles scored a four, five, or six out of ten, which reflects a moderate quality. Two remarkable results can be distinguished: Ou et al. (46) had a relatively high methodological quality with a score of eight out of ten, whereas Lee et al. (62) had a relatively low methodological quality with a score of three out of ten. Most studies had incomplete descriptions of the methodology, resulting in an unclear risk of bias. The high frequency of an unclear risk of bias was especially true for items that measure selection bias, performance bias, detection bias, and attrition bias.

An overview of the study characteristics can be found in Table 2. All articles were published between 2017 and 2021. Most studies were conducted in mice models of AD (n = 14; mostly APP/PS1), and four studies used AD models in Wistar rats. Only two studies used female animals, compared to thirteen studies using male animals and three studies did not provide a description on the sex of the animals.

Regarding the type of probiotics, lactobacilli and bifidobacteria were frequently used (i.e., L. plantarum, B. bifidum, and B. longum). Ten studies used a single-strain probiotic, compared to eight studies using a mixture. Two studies used the mixture VSL#3, which consists of L. plantarum, L. delbrueckii subsp. Bulgaricus, L. paracasei, L. acidophilus, B. breve B. longum, B. infantis, and Streptococcus salivarius subsp. thermophilus. One study used SLAB51, which is a probiotic mixture consisting of: Streptococcus thermophilus (DSM 32245), B. lactis (DSM 32246), B. lactis (DSM 32247), L. acidophilus (DSM 32241), L. helveticus (DSM 32242), L. paracasei (DSM 32243), L. plantarum (DSM 32244), and L. brevis (DSM 27961). Another study used the mixture Optibac, which consists of L. acidophilus and L. rhamnosus. An overview of the other mixtures can be seen in Table 2. Study durations ranged from 4 weeks to 6 months. In fifteen studies, probiotics were administered orally, while in three studies probiotics were administered intragastrically.

The majority of studies investigated the effect of probiotics on multiple outcome variables. Two studies looked at all the included outcome variables, whereas one study only looked at one outcome variable (see Table 3). In all studies, probiotics influenced at least one outcome variable of AD. Probiotic administration affected gut microbiota composition (n = 13 studies), SCFAs (n = 4 studies), inflammatory markers (n = 7 studies), and cognitive functioning (n = 12 studies). For Aβ, conflicting results were found, as some (n = 10 studies) found positive effects, whereas others (n = 5 studies) found no effect or no significant effect compared to AD control animals.

Thirteen studies looked at the effect of probiotics on gut microbiota composition, of which eight looked at microbiome richness and/or diversity (see Tables 3, 4). An improvement in both species richness (51, 57, 62) and bacterial diversity (57, 62) was found compared to control group animals. More specifically, a significant difference in α-diversity was found in four studies (49, 50, 57, 62), which was not replicated by Sun et al. (52) and Cecarini et al. (58). Similarly, two studies (51, 57) found a significant increase in β-diversity compared to AD control animals, whereas three others did not (49, 56, 58).

An increase in certain bacterial phyla was observed upon probiotic administration (i.e., Actinobacteria, Verrucomicrobia, Bacteroidetes) (50, 56). At the family level, an increase in Peptococcaceae, Ruminococcaceae, Prevotellaceae, S24-7, and Lactobacillaceae was found compared to control group animals (49, 52, 55–58, 62). An increase in Acetatifactor, Millionella, Alloprevotella, Parabacteroides, Bifidobacterium, Lactobacillus and Bacteroidales was observed compared to AD controls at the genus level (47, 51–53, 61). In contrast in other studies, a decrease at the phylum level was found for Firmicutes, Bacteroidetes, Proteobacteria, and Deferribacteres compared to control animals (49, 51, 52, 57). Also, at the family level, a decrease was observed for Odorbacteraceae, Lachnospiraceae, Helicobacteraceae, Ruminococaceae, and Pseudomonadaceae compared to AD controls (52, 56, 57, 61). A decrease compared to AD control animals was also found for Parabacteroides, Streptococcus, Desulfovibrio, Bacteroidales, Intestinimonas, Clostridium, and Helicobacter at the genus level (47, 49, 51, 52, 57). At the phylum level, no change in the overall Firmicutes/Bacteroidetes ratio was found (50).

When investigating the different probiotics, all included probiotics showed at least one effect on the gut microbiota composition (see Tables 3, 4).

Only four out of 18 studies addressed the effect of probiotics on SCFA levels, in which an improvement in plasma, serum, fecal, and hippocampal SCFA levels was observed compared to AD control animals (see Tables 3, 5) (47, 52, 56, 61).

A study by Kaur et al. (47) reported increased serum and hippocampal SCFAs compared to AD control animals. An increase in plasma acetate levels was found (56), as well as an increase in total fecal butyrate levels compared to AD control animals (52). Furthermore, an increase in fecal levels of acetic, propionic, and butyric acid was observed (61).

When looking at the different strains of probiotics, a significant increase in SCFA levels was observed upon administration with B. breve A1, the mixture VSL#3, the mixture SLAB51, and C. butyricum (see Tables 3, 5).

Eight studies investigated the effect of probiotics on relevant inflammatory markers for AD (50–52, 57, 58, 60–62) (see Tables 3, 6). Seven studies found a reduction in proinflammatory markers compared to control group animals. One study found a partial effect: a significant reduction in levels of proinflammatory cytokines in the ileum was found, as well as a significant improvement in serum eicosanoid levels compared to control group animals (50). In contrast, the same study found no significant effect on proinflammatory cytokines or Lipocalin-2 levels in the brain and no significant effect on protein levels in the ileum compared to AD controls (50).

Several studies found a reduction in proinflammatory cytokines in the brain (hippocampus), and blood (plasma) compared to AD control animals, such as IL-1α, IL-2, IL-6, IL-4, IL-12, IL-17, IL-1β, INF-γ, and TNF-α (51, 52, 57, 58, 60–62). Likewise, a decrease in serum eicosanoids, plasma lipopolysaccharide (LPS) levels, and plasma clusterin concentrations was found (50, 51, 57). Also, an increase in the anti-inflammatory cytokine IL-10 was observed in the brain compared to control group animals (58). Kaur et al. (50) did not find a decrease in proinflammatory proteins or protein levels of proinflammatory cytokines in the brain and ileum compared to AD control animals. When looking more specifically at gut inflammation, a reduction in TNF-α in the colon was reported (62). Cyclo-oxygenase 2 expression in the colon, and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-lb.) activation in the colon and in microglial BV-2 cells were found to be decreased compared to control group animals (57, 62).

Most included probiotics showed at least one significant effect on inflammatory markers (see Tables 3, 6). No significant results were found after administration with the mixture VSL#3. Significant improvements were found upon administration with: L. plantarum (ATCC8014), L. plantarum (C29), B. longum (NK46), B. longum (1714), the mixture SLAB51, L. lactis subsp. cremoris, and C. butyricum.

A total of 15 studies assessed the effect of probiotics on Aβ parameters (see Tables 3, 7). A reduction in the amount of Aβ plaques in the brain (46, 49, 51, 53), as well as the size of Aβ plaques in the brain (53), Aβ deposition in the brain (46, 52, 60, 61) and the Aβ expression in the hippocampus (57, 62) was found compared to control group animals. More specifically, a decrease in brain levels of Aβ(1-42) (52, 58, 61) and Aβ(1-40) (58) was observed in four studies compared to control group animals. Contrastingly, three studies did not find any effect in the brain (50) and hippocampus (52, 55), or no significant effect in the brain (32, 59) upon probiotic administration on Aβ. Similarly, another study also found no differences in the brain after probiotic administration compared to control group animals for Aβ(1-40) (61).

When looking at the different strains of probiotics, no effect regarding Aβ was found upon administration with L. plantarum (MTCC1325); the mixture VSL#3; L. acidophilus, B. bifidum & B. longum; and Optibac (see Tables 3, 7). Significant effects were observed upon administration with L. plantarum (ATCC8014); L. plantarum (C29); B. longum (NK46); B. longum (1714); B. lactis Probio-M8; Akkermansia municiphila; the mixture SLAB51; L. lactis subsp. cremoris; C. butyricum; and L. acidophilus, L. fermentum, B. lactis & B. longum.

Fifteen studies investigated the effect of probiotics on cognitive functioning, of which twelve studies found at least one improvement in cognitive functioning compared to the control groups (see Tables 3, 8) (32, 46, 48–54, 56–59, 61, 62). As shown in Table 3, not all tests performed found an improvement (e.g., only in one out of three tests). Six studies found an improvement for all tests, three of which used more than one test.

Improvements in spatial working memory, as assessed by the Morris Water Maze test, were found after probiotic supplementation compared to control group animals in all studies (32, 48, 51–53, 57, 59, 62). However, a probe trial test, part of the Morris Water Maze test that measures how long the test subject spends in the target quadrant, did not show significant differences between AD rats with probiotic supplementation and control rats (32). Learning and memory capacity, as measured by the Passive Avoidance Test, improved in two studies compared to control group animals (56, 57), whereas it did not improve in another study (61). Inconsistent results were also found for the Novel Object Recognition Test, which measures recognition memory: an improvement was observed in three studies (52, 57, 61), while four other studies did not find any differences compared to AD controls (48, 49, 51, 58). Improvements (46, 57), as well as no difference (56), were found for locomotor activity assessed via the Y-maze test compared to AD control animals. Similarly, the Open-Field test, which measures general locomotor activity, improved compared to AD controls in one study (46), which was not replicated in two other studies (48, 61). No differences compared to AD controls could be assessed for the Elevated Plus Maze Test, which measures anxiety responses (61).

When looking at different strains of probiotics, no effect on cognitive functioning was found upon administration with VSL#3, L. lactis subsp. cremoris, and L. plantarum & B. bifidum (see Tables 3, 8). Significant effects were observed upon administration with B. breve A1; L. plantarum (ATCC8014); L. plantarum (C29); L. plantarum (MTCC1325); B. longum (NK46); L. acidophilus, B. bifidum & B. longum; B. lactis Probio-M8; Akkermansia municiphila; SLAB51; Clostridium butyricum (C. butyricum); and L. acidophilus, L. fermentum, B. lactis & B. longum.

AD pathology has been associated with alterations in the gut microbiota composition and GI inflammation (13, 50, 63, 64). It can be argued that increased gut permeability allows increased concentrations of LPS in the gut and in the circulation, which in turn can trigger amyloid secretion in the gut, while amyloid secretion in the gut can exacerbate the intestinal permeability (46, 65, 66). This process results in increased production and translocation of cytokines and pro-inflammatory components of GI origin into the body (46, 65, 66). These inflammatory compounds could not only increase systemic inflammation, but could also cross the BBB and may induce neuroinflammation, amyloid secretion, neuronal injury, dysfunction of specific brain regions, the development of insulin resistance and ultimately lead to neuronal death in AD (46, 65, 67–70). Some animal studies reported a significant decrease in GI inflammation and attenuated intestinal permeability upon probiotic administration in AD model animals (50, 62). Another study argued that probiotic administration could ameliorate cognitive decline by means of a reduction of gut microbiota LPS production and the regulation of microbiota LPS-mediated NF-κB activation in BV-2 cells, a type of microglial cells (57). From a clinical perspective, a recent meta-analysis summarizing human RCTs found a significant improvement in cognition, as well as a significant reduction in high-sensitivity C-reactive protein levels in the probiotics group compared to the control group (35), supporting the idea that probiotics can reduce systemic inflammation. Overall, these findings suggest that probiotic supplementation suppresses the downward spiral of GI inflammation, altered gut microbiota composition, increased gut permeability, translocation of pro-inflammatory compounds through the BBB, and potential adverse inflammatory and metabolic processes in the brain.

Another underlying mechanism might be SCFA production of the metabolites SCFAs by the gut microbiome. Some metabolites, like SCFAs, can pass through the BBB directly, induce a decrease in pro-inflammatory cytokines in the brain, and modulate the maturation of microglia (32, 52, 56, 71). Additionally, SCFAs are found to interfere with protein-protein interactions, which are necessary for the formation of toxic soluble Aβ aggregates/converse Aβ peptides into Aβ neurotoxic aggregates (72). This interference capability is especially true for valeric acid, butyric acid, and propionic acid (72). This way, SCFAs may help to alleviate elements of the pathophysiological processes of AD.

This review found that probiotic administration in AD rodent models increased plasma, fecal, and hippocampal SCFA levels, reduced inflammatory markers in the blood and brain, and improved cognitive functioning in multiple studies, which is consistent with the hypotheses mentioned above.

B. longum (NK46), C. butyricum, and the mixture SLAB51 seem to be the most promising probiotics for AD, as they have shown the most positive outcomes. B. longum is an anaerobic, non-halophilic, gram-positive bacterium that is naturally present in the human GI tract (73). B. longum is considered safe by the EFSA (74). Although multiple health benefits have been found upon B. longum administration, such as diarrhea prevention in antibiotic treated patients, and immune stimulation (73), it is unclear why this probiotic is effective for AD hallmarks in animal studies. A possible rationale is that B. longum (NK46) may stimulate the production of butyrate. Another hypothesis is that B. longum is associated with reduced intestinal inflammation and improved epithelial barrier integrity, and therefore decreases the passage of pro-inflammatory compounds through the BBB and reduces other potential adverse metabolic processes (50). B. longum is able to decrease gut microbiota LPS production as well as regulate LPS-induced NF-κB activation in microglial BV-2 cells in AD mice (57).

C. butyricum is an anaerobic, gram-positive, spore-forming bacteria that is common in the human colon (75). It has various implications for human health, ranging from pathogenic to beneficial, such as inducing botulism in infants and helping to overcome antibiotic-associated diarrhea in children (76, 77). Furthermore, C. butyricum is known for its ability to produce large amounts of SCFAs, such as butyrate and acetate (75). It is hypothesized that this high production of SCFAs is the reason for the beneficial health effects of C. butyricum in rodent studies. In Asia, C. butyricum is frequently used as a probiotic (75). However, C. butyricum is not yet on the QPS safety list of the EFSA (74). A recent animal study showed a neuroprotective effect of C. butyricum in mouse models of traumatic brain injury, partially due to increased secretion of glucagon-like peptide 1 (GLP-1), a 30-amino acid peptide hormone, through the GBA (78). Stoeva et al. (79) hypothesize that C. butyricum stimulates the secretion of GLP-1, which protects the BBB, potentially via the modulation of tight junctions. Likewise, C. butyricum is found to prevent brain endothelial barrier dysfunction, as demonstrated by decreased brain water content and the restoration of normal levels of tight junction protein expression (79). Furthermore, significant improvements in neurological dysfunction, brain edema, neurodegeneration, and BBB impairment were observed after C. butyricum administration (78). Moreover, C. butyricum decreased plasma-d lactate and colonic IL-6, whilst protecting the intestinal barrier integrity and upregulating the expression of occludin (78). Another animal study, which looked at vascular dementia in mice, found that C. butyricum significantly ameliorated cognitive dysfunction and histopathological changes, increased brain-derived neurotrophic factor (BDNF) and Bcl-2 levels, which are cell survival proteins that inhibit apoptosis (80), decreased Bax levels, which is part of the Bcl-2 family (81), decreased p-Akt levels, which is phospholyrated protein kinase B, and reduced neuronal apoptosis (79, 82). C. butyricum was also found to restore butyrate levels in feces and brain and regulate the gut microbiota in mouse models of vascular dementia (82). When looking at mouse models of AD, administration of C. butyricum prevented cognitive impairment, Aβ deposits, microglia activation, and production of TNF-α and IL-1β in the brain of AD mice (52). Also, C. butyricum treatment reversed abnormal gut microbiota and butyrate (52). More specifically, butyrate treatment was found to reduce CD11b and Cyclo-oxygenase 2 levels, and suppress the phosphorylation of NF-κB p65 in the Aβ-induced BV2 microglia (52). However, in this research, only one study investigated the effect of C. butyricum on AD mice, meaning that no firm conclusions can be drawn upon it yet.

Like mentioned before, SLAB51 is a probiotic mixture consisting of: Streptococcus thermophilus (DSM 32245), B. lactis (DSM 32246), B. lactis (DSM 32247), L. acidophilus (DSM 32241), L. helveticus (DSM 32242), L. paracasei (DSM 32243), L. plantarum (DSM 32244), and L. brevis (DSM 27961). All probiotics, except B. lactis, are considered safe by the EFSA (74). An animal study in AD mice found that the SLAB51 mixture significantly decreased oxidative stress in AD mice brains by activating Sirtuin 1 (SIRT1)-dependent mechanisms (45). SIRT1 is a protein family that is used during the cellular response to inflammatory, metabolic, and oxidative stressors (83). These proteins play a role in NAD^* dependent deacetylation of histones, as well as in neuronal plasticity, cognitive function, and neuronal degeneration (84). SIRT1 levels were found to be lower in serum samples of patients with AD or MCI compared to age matched controls (85). Additionally, beneficial antioxidant effects were found in the brain of AD mice after SLAB51 administration (45). Another study in AD mice found that the mixture SLAB51 ameliorated the impaired glucose metabolism in AD by restoring the brain levels of glucose transporters GLUT3 and GLUT1, ameliorating brain glucose homeostasis, reducing tau phosphorylation by modulating pAMPK and pAkt, and decreasing advanced glycation end products (43). Because of this, the authors argue that amelioration of the impaired glucose metabolism in AD is able to delay AD progression through gut microbiota manipulation with SLAB51 (43). From a broader perspective, the SLAB51 mixture is also considered a promising candidate for the prevention or (coadjuvant) treatment of Parkinson's disease, as the mixture was able to protect dopaminergic neurons and improve behavioral impairments in in-vivo studies (86). Also, the mixture SLAB51 counteracted neuroinflammation and oxidative stress in both in-vivo and in-vitro studies (86). Moreover, the mixture SLAB51 modulated the BDNF pathway, increased neuroprotective protein levels, and decreased neuronal death proteins in the in-vitro studies (86).

No clinical trials have been found that investigated the effect of B. longum (NK46), C. butyricum, or the mixture SLAB51 in AD or MCI patients yet. From a broader perspective, two clinical trials investigated a mixture containing bifidobacteria and lactobacilli in AD patients and found significant improvements in cognitive functioning and some metabolic parameters (87, 88). One of these clinical trials used co-supplementation with selenium. In contrast, a recent meta-analysis, that included these two studies and an RCT in patients with severe AD (89), found no effect on cognitive function in AD patients upon probiotic supplementation, which all consisted of lactobacilli and bifidobacteria (36).

Although in this review we looked exclusively to interventions with only probiotics, in the future it may be wise to look broader than probiotics alone, since the effects on AD hallmarks appear to be stronger when probiotic supplementation is combined with another intervention. This insight may imply that probiotics work synergistically with other interventions. For instance, a stronger effect was found upon probiotic administration together with memantine (1 mg/ml), which is an AD drug classified as an N-Methyl-D-aspartic acid receptor antagonist, exercise, L. plantarum-fermented soybean, and p62-transformed L. lactis compared to probiotics alone (51, 54, 58, 62). Taking this chain of thought further, this review focused exclusively on probiotics due to the lack of animal studies on prebiotics. Prebiotics are substrates (often carbohydrates) that are selectively utilized by host microorganisms that confer a health benefit. There are multiple types of prebiotics that are suggested to benefit the microbiota (90–92). Prebiotics metabolized by the gut microbiota to SCFAs are, like probiotics, suggested to slow down AD progression due to their effect on the intestinal microbiota (33, 72). One animal AD study, investigating the prebiotic effects of fructooligosaccharides, found that these specific fructooligosaccharides could, among other things, improve oxidative stress and inflammation, regulate the synthesis and secretion of neurotransmitters, positively affect the diversity and stability of the microbiome of AD rats, and down-regulate the expression of both tau and Aβ1-42 (93). From a broader perspective, healthy dietary patterns characterized by high levels of prebiotics and probiotics, in association with other nutrients, are found to delay cognitive decline and decrease the risk of AD (66, 94).

This review is conducted from an animal perspective. Therefore, results from this study cannot be one-on-one translated into the clinic. This is not only due to the dissimilarity of AD animal models compared to AD patients, but also due to confounders in animal experiments, such as environmental factors and host genetic background (95). This review is also based on studies that used a variety of AD models, which could have introduced some bias by potentially affecting the composition of the gut microbiota differently (95). In addition, both the human microbiome (96) and AD characteristics (97) are sex-specific, and thus, women and men may benefit differently from probiotic augmentation. Most animal studies used male rodents and separate studies are needed to investigate females.

In this review, four studies used AD rat models (32, 53, 54, 59). Three of these studies used a mixture of bifidobacteria and lactobacilli strains, while one study used L. plantarum. Interestingly, the findings are relatively in line with the results of mice studies in the sense that significant evidence can be found for improvement on both cognitive functioning and gut microbiota composition, whereas more ambiguous results can be found for Aβ levels.

To go from animal to clinical trials, information on preliminary efficacy, toxicity, safety, and pharmacokinetics is needed. Research should look into the optimal duration of probiotic supplementation in AD, as well as the ability of probiotics to survive passage through the human GI tract. For future clinical trials, it is recommended to use oral administration (i.e., in the form of a sachet or pill), as this is minimally invasive for patients.

Some limitations of this review must be mentioned. First, the included studies used a variety of animal models, as well as a variety of cognitive tests. These cognitive tests assess a slightly different part of cognitive functioning, which negatively influences the internal validity of this review. Second, as assessed in Chapter 3.2, only one study (46) has a high methodological quality. Therefore, the general poor methodological quality may have negatively influenced the results of this study. These findings support the statement that the reporting of experimental details on animals, methods and materials in animal studies is often poor (98). Third, the included studies used relatively small intervention groups and mostly male mice, which both negatively affect the external validity of the results. This is unfortunate as sex and gender differences exist both in the human form of AD and in the human microbiome. Furthermore, most of the included studies were conducted over a period of 2 months. Therefore, long-term conclusions cannot be drawn. Also, specific probiotics were only tested in a single study, and no direct comparisons between probiotics have been made so far. Lastly, publication bias may have caused an underreporting of studies where no effects were found. A strength of this review is that this review provides an in-depth overview of all animal studies on AD on probiotic strain level, whilst taking multiple outcome variables into account.

Future research should investigate the clinical effects of probiotic supplementation on AD symptoms and hallmarks. B. longum (NK46), C. butyricum and the mixture SLAB51 are promising types of probiotics to test in clinical trials. B. longum is considered safe for human consumption by the EFSA (74). It is advised to use oral administration. It is also argued that AD drugs can potentially have a negative effect on the gut microbiota whilst temporarily alleviating the symptoms of AD (e.g., cognitive improvement, reduced inflammation, or a reduction of Aβ and tau proteins) (94). Others argue that the combined administration of prebiotics, probiotics, and treatment could prevent or alleviate gut problems, which may potentially strengthen the efficacy of AD drugs by eliminating a possible factor that sustains the disease (94). This combination may be especially relevant to investigate in AD patients who are currently unresponsive to pharmacological treatment (94).