Introduction

Atopic dermatitis (AD) is the most common chronic inflammatory skin disorder worldwide.1 Its pathogenesis is multifactorial,2,3 and different AD treatments purport to focus on one or more of the pathogenic mechanisms. The most common treatment target is the disordered immune system and resulting inflammation,2 but there is an emerging body of research suggesting that the skin microbiome can be manipulated for AD treatment.

A growing body of literature has reinforced the relationship between AD severity and alterations in the species distribution of the skin microbiome, specifically decreased biodiversity with predominance of Staphylococcus aureus (S. aureus) in AD.1,4–7 Whether these microorganisms are colonizers, true pathogens, or more nuanced pathobionts is still not fully understood. Some studies suggest that AD skin is permissive to S. aureus colonization and more susceptible to S. aureus virulence.1,5 In turn, S. aureus disrupts the epidermal barrier and induces Th2 response, both pathogenic mechanisms implicated in AD,5 but not all microbes have detrimental effects on the skin. Some coagulase-negative Staphylococcus species, which are deficient in lesional AD skin, produce antimicrobial peptides against S. aureus.8,9 Other studies have found that specific bacteria such as Roseomonas mucosa,10 Vitreoscilla filiformis,11 Lactobacillus johnsonii NCC 533,12 and Streptococcus thermophilus13 appear to reduce inflammation and may decrease AD severity as well.

Metabolic pathways provide another potential target in altering the skin microbiome for AD treatment. Metagenomic studies suggest that alterations in the skin microbiome’s metabolism may influence AD severity.14,15 Prebiotics are substrates in the metabolic pathway that can selectively increase or limit the population of beneficial and detrimental microbes, respectively.16 Postbiotics are the metabolites produced by the microbiome that can influence microbial diversity and inflammation.14,15,17 Many studies demonstrate the positive effect of oral prebiotics and postbiotics on the gut microbiome in a variety of conditions and found potential benefit towards AD,17–20 but few studies have investigated the effect of topical prebiotics on the skin microbiome of AD patients.

While the literature suggests that the skin microbiome plays an important role in the pathogenesis of AD, there have not yet been larger studies analyzing the effects of altering the metabolic pathways of the skin microbiome in the setting of AD. This systematic review aims to assess the current literature investigating the effects of prebiotics and metabolites of the skin microbiome on AD. The present study also evaluates the current literature’s quality to help guide investigators in designing and effectively conducting future studies such that protocols are optimized to reduce bias.

Materials and Methods

The selection process is outlined in Figure 1. This systematic review was conducted in accordance with PRISMA guidelines. Articles were retrieved from PubMed, Embase, and Web of Science database from the earliest records to March 2022 using the search formula: (“atopic dermatitis” OR eczema) AND ((metabolite* OR metabolom* OR metabonom* OR “metabolic profile” OR “fatty acid metabol*” OR ceramide* OR “fatty acid” OR sphingo* OR prebiotic) AND (“skin microbiome*” OR “skin flora”)) OR ((prebiotic OR oat OR glucan OR oligosaccharide) AND (topical OR emollient OR moisturizer OR lotion OR cream OR ointment)).

Figure 1
Figure 1.Flowchart depicting selection process of publications for this systematic review, number of publications (n)

Screening

Initial screening was conducted using Rayyan. Duplicates were deleted, and two independent reviewers assessed each article’s title and abstract to exclude articles that were reviews or published abstracts, not in English, not published in a peer-reviewed journal, did not use human subjects, not on AD, and not investigating metabolites of the skin microbiome or topical prebiotics. In the case of disagreement, a third reviewer made the decision to include or exclude a publication after reviewing the study.

Subsequently, two independent reviewers screened the remaining publications by evaluating the entire article. In the case of disagreement, a third reviewer made the decision to include or exclude a publication after reviewing the study. Studies were included if they met the following inclusion criteria: not a review or published abstract, published in a peer-reviewed journal, published in English, subjects with atopic dermatitis, and investigated the effect of metabolites of the skin microbiome or prebiotics on atopic dermatitis. An exclusion criterion based on type of study was not included due to the lack of research in this field.

Quality Assessment

Two independent reviewers assessed the quality of each publication using the appropriate quality assessment tool. Randomized controlled trials (RCT), quasi-experimental studies, and case-control studies were assessed with the respective Joanna Briggs Institute (JBI) critical appraisal tool. To determine the risk of bias in an animal study, the SYRCLE’s risk of bias tool was used (Hooijmans). Reviewers judged if the publication had a low, moderate, or high risk of bias based on the criteria of each tool. In the event that there was a discrepancy in score between the two reviewers, a third reviewer made the final judgment for that publication.

Outcome Measurements

For the purpose of this study, treatment outcome was defined as any statistically significant change in AD severity reported by the authors. Treatment was considered clinically relevant if the mean change in SCORAD was 8.7 or higher.21 Adverse effects were identified and recorded.

Results

The initial search resulted in 1,124 studies. During the initial screening, 112 duplicate studies were removed, and 906 studies were excluded. The remaining 106 studies were assessed based on the full article, and 100 articles were excluded because they did not meet the inclusion criteria. Five articles were included in this systematic review.

Summary of Study Design

Four out of the five studies were clinical trials investigating the effect of topical prebiotic application on atopic dermatitis (Table 1). Of these studies, 50% (2/4) were double-blinded22,23 and 50% (2/4) were open-label.24,25 Only one study included a control.22 All four studies used SCORAD as the primary evaluation method of AD severity.

Table 1.Summary of patient demographics and treatment efficacy of topical prebiotics
Author Number of Subjects Age Treatment Prebiotics Treatment Regimen Treatment Length Clinically relevant decrease in AD severity1
Noll et al. (2021) 22 NR Synbiotic bath (n = 7)
Prebiotic bath (n = 8)
Placebo bath (n = 7)
Maltodextrin, inulin, apple pectin 10 mins daily 14 days Y
Rigoni et al. (2018) 26 1.5-45 years Prebiotic emollient Alpha-glucan-oligosaccharide 2-3 times daily 8 weeks N
Antao et al. (2017) 60 3-36 months Prebiotic face cream and prebiotic body wash (n = 30)
Prebiotic body cream (n = 30)
Alpha-glucan-oligosaccharide and inulin Cream: 2 times daily
Wash: daily
56 days Y
Seité et al. (2017) 60 6 months-63 years Prebiotic emollient (n = 26)
Placebo emollient (n = 27)
Mannose and thermal spring water 2 times daily 28 days N

Atopic dermatitis (AD), double-blinded (DB), no (N), not reported (NR), open label (OL), yes (Y)
1 Decrease in AD severity is considered clinically relevant if it is statistically significant (p < 0.05) and mean difference in SCORAD is 8.7

Table 2 summarizes the study design for trials investigating the utility of topical prebiotics. Exclusion criteria included: severe AD,24,25 psychological disease,24 immunodeficiencies,22 pregnancy,22 and simultaneous AD treatment.24,25 Two studies allowed concomitant treatment.24,25 Antao et al (2017) allowed topical corticosteroid use, but Rigoni et al (2018) was more stringent and did not allow any simultaneous topical or systemic treatment except for oral antihistamines.25 Seité et al (2017) only included patients with mild to moderate AD defined as SCORAD < 40 at the initial visit and excluded seven subjects if they improved less than 25% in the 15 days from screening to initial study visit.23

Table 2.Summary of study design for topical prebiotic trials
Author Study Design Exclusion Criteria Permitted AD Treatments Outcome Measures Statistical Analysis
Noll et al. (2021) DB Age < 5 years, pregnancy, immunodeficiency, ongoing antibiotic treatment None SCORAD, 10-point PRO for quality of life ITT
Rigoni et al. (2018) OL Age < 1 year, SCORAD ≥ 40, known allergy to ≥ 1 tested component, simultaneous administration of topical or systemic medications Oral antihistamines SCORAD, adjusted IGA ITT
Antao et al. (2017) OL SCORAD ≥ 35, psychological disease, simultaneous use of cosmetic creams Topical corticosteroids SCORAD; 5-point scale for xerosis, erythema, and edema PP
Seité et al. (2017) DB SCORAD ≥ 40, SCORAD improved by < 25% between screening and initial visit (15 days) None SCORAD PP

Double blind (DB), intention to treat (ITT), open label (OL), per protocol (PP), SCORing Atopic Dermatitis (SCORAD)

The total number of subjects in each study ranged from 22-60. Two studies enrolled subjects from a wide range of ages, including both children and adults,23,25 but one study focused on children from 3-36 months old.24 One study did not report the age of their subjects.22

The final included study was a translational study. Gel-patch and swab were used to collect skin samples from 19 subjects with AD and 19 controls and analyzed for the concentration of indole-3-aldehyde (IAId), which is produced by the skin microbiome after metabolizing tryptophan (Trp). To study the effects of topical IAId application, a mouse model for AD was used, and MC903 was painted on mouse ears for 14 days to induce an AD-like dermatitis before treatment with IAId.26

Treatment Effect on AD

Prebiotics

Different modalities of treatment were studied including emollients,23–25 baths,22 and body washes.24 One study investigated the use of a prebiotic face cream and prebiotic body wash in comparison to a prebiotic body cream alone.24 The most commonly used prebiotics were alpha-glucan oligosaccharide24,25 and inulin,22,24 which were included in 50% (2/4) of the studies. Other prebiotics studied include maltodextrin,22 apple pectin,22 mannose,23 and thermal spring water.23

Out of the four studies, 75% (3/4) of the studies found a statistically significant decrease in AD severity and reported decreased dermatitis symptoms such as erythema and lichenification,22,24,25 and 33% (2/3) of the studies found a clinically relevant decrease in AD severity. Although Seité et al (2017) reported a decrease in AD severity with topical prebiotic treatment, it was not statistically significant compared to treatment with placebo. However, the authors noted that the SCORAD variation (calculated as SCORAD at study end minus SCORAD at first visit) between the treatment and placebo group was significant.23

Metabolites of the Skin Microbiome

Yu et al (2019) found a higher concentration of IAId in the skin samples of controls and non-lesional skin samples of subjects with AD. In their mouse model, mice treated with topical IAId had less visible erythema and scale, and histopathology demonstrated decreased epidermal thickness and less inflammatory infiltrates, suggesting that IAId decreased the inflammatory process associated with AD.26

Adverse Effects

Out of 168 total subjects across four studies, only one subject discontinued the use of a face cream because of an unpleasantly strong fragrance.24 Otherwise, none of the studies reported an adverse effect, and Rigoni et al (2018) actually reported that 77% (20/26) and 33% (6/26) of subjects rated the prebiotic emollient as very good and good, respectively.25

Risk of Bias

The studies demonstrated varying levels of bias (Figure 2). One study was judged to not have high risk of bias in any criteria but did have moderate bias in two criteria.22 All other studies had high risk of bias in 8% (1/12) to 25% (5/20) of the graded criteria in their respective quality assessment tools.24

Figure 2
Figure 2.Risk of bias in each included study presented as percentages

Table 3 summarizes the risk of bias for each criterion. The JBI checklist for RCT was used to assess two publications.22,23 Both demonstrated moderate risk of selection bias in randomization and allocation concealment. Of the two publications, one also had high risk of attrition bias because they chose to analyze patients because they did not include every randomized subject in their analysis.23

The JBI checklist for quasi-experimental studies was used to assess two publications.24,25 Both lacked a control group and were open label. Furthermore, one study demonstrated moderate risk of selection bias in the baseline characteristics of their subjects.

Yu et al (2019) required evaluation based on both the JBI checklist for case-control and SYRCLE’s risk of bias tool because their publication included multiple study designs.26 There was moderate risk of selection bias in subject recruitment, but they appropriately matched subjects in both groups. In the animal study, there was high risk of selection bias in randomization and allocation concealment, high risk of performance bias, and some moderate risk of attrition and reporting bias.

Discussion

Historically, there has been a debate concerning whether AD follows an “outside-in” hypothesis or “inside-out” hypothesis.27 That is to say, whether immune dysregulation causes skin barrier dysfunction or that skin barrier dysfunction is primary with resultant secondary inflammation. Regardless of which aspect may come first, it is now abundantly clear that AD is characterized by both intrinsic and extrinsic factors, and that the skin microbiome is intimately entwined with these processes. An emerging area of research focuses on the metabolic pathways of the skin microbiome and demonstrates potential treatment avenues.

Therapeutic Potential for Treatment Alternatives

Prebiotics

All studies found that AD severity decreased with topical prebiotic treatment, but the difference between the treatment and placebo group at endpoint was not clinically meaningful in two studies.23,25 Rigoni et al (2018) reported a statistically significant change in SCORAD, and Seite et al (2017) found a significant SCORAD variation between the two groups such that the calculated decrease in SCORAD from initial to final treatment was significantly different between patients using topical prebiotics and those on placebo. These results suggest that topical prebiotics + emollients may provide more clinical improvement than emollients alone.23

Although all four studies suggest some therapeutic benefit from topical prebiotic use, the results may not be generalizable to the entire population of AD patients because three out of four studies included only patients with mild to moderate AD. In fact, one study had patients treated for 15 days between screening and baseline visit and disqualified any patients whose SCORAD did not improve at least 25%, excluding patients that may be treatment resistant.23 The results of these current studies indicate that topical prebiotics may be efficacious in reducing AD symptoms, but more studies are required to conclude if topical prebiotics provide any benefit for patients with more severe or hard-to-treat AD.

The vehicle of a topical prebiotic may or may not inherently affect its efficacy, given the results of a recent study that reported no difference in efficacy between lotions, creams, gels, and ointments.28 One study comparing prebiotic body wash to body cream found no significant difference in the AD severity of the two treatment groups.24 However, the actual ingredients within a preparation clearly matter. A recent meta-analysis found that combining emollients with a topical active treatment provides a clinically significant improvement in comparison to regular emollients alone.29 Topical prebiotics are promising alternative active ingredients that may make emollients more efficacious in treating AD. Studies comparing topical prebiotics to vehicle controls suggest that topical prebiotics may help decrease AD severity more than other emollients,22,23 but more randomized, controlled trials with larger sample sizes are required.

The studies did not report any severe adverse effects with topical prebiotic use. Because the results of these comparative studies suggest that topical prebiotics may have some additional benefit over emollients, this is clearly an approach worthy of more investigation.

Metabolites of the Skin Microbiome

IAId is a tryptophan metabolite produced by the skin microbiome that was shown to decrease the inflammatory response and subsequent barrier dysfunction related to AD.26 It was able to decrease epidermal thickness, suggesting that it may even be efficacious in patients suffering from chronic AD with lichenification.26 IAId acts by modulating expression of thymic stromal lymphopoietin (TSLP) in keratinocytes, which is a cytokine implicated in Th2-mediated inflammation in AD.26 These novel findings suggest that modulating microbial metabolites may be another potential method of treating AD, including direct administration of microbial metabolites in topical or systemic treatment, or altering the skin microbiome to promote production of specific beneficial metabolites.

Risk of Bias in the Current Literature

Although the current literature does not generally suffer from high risk of bias, several components of their study design have the potential to introduce bias. Based on the quality assessment performed, there was a risk of bias in the selection process, determination of treatment efficacy, and statistical analyses.

Table 3.Summary of risk of bias in each criterion for included studies. Red denotes high risk, blue denotes moderate risk, and green denotes low risk of bias.
JBI Checklist for Randomized Controlled Trials
Noll et al. (2021) Seite et al. (2017)
Selection bias – randomization
Selection bias – allocation concealment
Selection bias – baseline characteristics
Performance bias – blinding
Performance bias – treatment
Detection bias – blinding
Detection bias – outcome measurement
Attrition bias – differences in follow-up
Attrition bias – intention-to-treat
Reporting bias
Other bias – statistical analysis
JBI Checklist for Quasi-Experimental Studies
Rigoni et al. (2018) Antao et al. (2017)
Selection bias – baseline characteristics
Performance bias – control group
Performance bias – treatment
Detection bias
Attrition bias
Other bias – statistical analysis
Other bias – cause vs. effect
JBI Checklist for Case-Control Studies
Yu et al. (2019)
Selection bias – baseline characteristics
Selection bias – matched cases and controls
Selection bias – recruitment methods
Detection bias – outcome measurement
Detection bias – confounding factors
Other bias – statistical analysis
SYRCLE’s Risk of Bias Tool
Yu et al. (2019)
Selection bias – randomization
Selection bias – allocation concealment
Selection bias – baseline characteristics
Performance bias – random housing
Performance bias – blinding
Detection bias – random outcome assessment
Detection bias – blinding
Attrition bias
Reporting bias
Other bias

Most commonly, studies lacked controls against selection bias or did not adequately describe their protocol for selection. In two studies, the authors explicitly stated that their study was double-blind and randomized, but they did not detail how subjects were randomized.22,23 As a result, it is not possible to determine whether the subjects were randomized and allocated into groups appropriately to prevent bias. In a third study, it is unclear if the mice were randomized at all.26

Yu et al (2019) also did not mention if the investigators were blinded to whether the mice in the AD model were given treatment or not, which could lead to performance and detection bias.26 However, the authors attempted to decrease detection bias by preparing more objective outcome measurements including quantitative measurements of ear and epidermal thickness.

Some studies potentially had attrition bias due to their chosen method of statistical analysis. Seite et al (2017) and Antao et al (2017) followed a per protocol (PP) rather than intention-to-treat (ITT) protocol.23,24 With this protocol, data analysis only includes subjects who completed the treatment according to the protocol and disregards subjects who were non-adherent or lost to follow-up. Although it is acceptable to use a PP principle for statistical analysis, it does have the potential to overestimate treatment effect.

Limitations

This study did not use any criteria that excluded studies based on study design as research in this field of study is relatively new and there were few studies found. Consequently, a meta-analysis was not conducted because of the variations in study design, including open label versus double-blind studies along with differences in statistical analysis methods.

Conclusion

The studies included in this systematic review suggest that topical prebiotics may be a valuable therapeutic modality for patients with mild-to-moderate AD. They have demonstrated potential in modifying the skin microbiome to decrease AD severity, but many questions remain about the specific prebiotic or prebiotics, dosing, administration, and optimal patient selection. Current results highlighting the influence of metabolites of the skin microbiome provide the basis for developing multiple new methods of AD treatment. However, both areas of study require more investigation. Future studies on prebiotics should aim to perform more randomized, controlled trials with larger patient populations and double blinding, and further investigation is required on the skin microbiome’s metabolism before potential treatments can be developed.


Corresponding Author

Sheshanna Phan
Department of Clinical Sciences
College of Osteopathic Medicine
Touro University Nevada
874 American Pacific Dr
Henderson, NV, USA
Phone: (657) 222-5078
Email: sphan3@student.touro.edu

Conflicts of Interest

Dr. Lio reports research grants/funding from AOBiome, Regeneron/Sanofi Genzyme, and AbbVie; is on the speaker’s bureau for Regeneron/Sanofi Genzyme, Pfizer, Incyte, Eli Lilly, LEO, Galderma, and L’Oreal; reports consulting/advisory boards for Almirall, ASLAN Pharmaceuticals, Bristol-Meyers, Concerto Biosciences (Stock Options), UCB, Dermavant, Regeneron/Sanofi Genzyme, Merck, Pfizer, LEO Pharmaceuticals, AbbVie, Eli Lilly, Micreos, L’Oreal, Pierre-Fabre, Johnson & Johnson, Level Ex, KPAway (Stock), Unilever, Menlo Therapeutics, Theraplex, IntraDerm, Exeltis, AOBiome, Realm Therapeutics, Altus Labs, Galderma, Verrica, Arbonne, Amyris, Bodewell, Burt’s Bees, My-Or Diagnostics, Sibel Health, and Kimberly-Clark. In addition, Dr. Lio has a patent pending for a Theraplex product with royalties paid and is a Board member and Scientific Advisory Committee Member of the National Eczema Association and an investor at LearnSkin. The other authors declare no conflicts of interest.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.