the athletic gut microbiota - journal of the international society … · 2020. 5. 12. · athletic...

REVIEW Open Access

The athletic gut microbiotaAlex E. Mohr1* , Ralf Jäger2, Katie C. Carpenter3, Chad M. Kerksick4, Martin Purpura2, Jeremy R. Townsend5,Nicholas P. West6, Katherine Black7, Michael Gleeson8, David B. Pyne9, Shawn D. Wells10, Shawn M. Arent11,Richard B. Kreider12, Bill I. Campbell13, Laurent Bannock14, Jonathan Scheiman15, Craig J. Wissent16, Marco Pane17,Douglas S. Kalman18, Jamie N. Pugh19, Carmen P. Ortega-Santos1, Jessica A. ter Haar20, Paul J. Arciero21 andJose Antonio22

Abstract

The microorganisms in the gastrointestinal tract play a significant role in nutrient uptake, vitamin synthesis, energyharvest, inflammatory modulation, and host immune response, collectively contributing to human health. Importantfactors such as age, birth method, antibiotic use, and diet have been established as formative factors that shape thegut microbiota. Yet, less described is the role that exercise plays, particularly how associated factors and stressors,such as sport/exercise-specific diet, environment, and their interactions, may influence the gut microbiota. Inparticular, high-level athletes offer remarkable physiology and metabolism (including muscular strength/power,aerobic capacity, energy expenditure, and heat production) compared to sedentary individuals, and provide uniqueinsight in gut microbiota research. In addition, the gut microbiota with its ability to harvest energy, modulate theimmune system, and influence gastrointestinal health, likely plays an important role in athlete health, wellbeing,and sports performance. Therefore, understanding the mechanisms in which the gut microbiota could play in therole of influencing athletic performance is of considerable interest to athletes who work to improve their results incompetition as well as reduce recovery time during training. Ultimately this research is expected to extend beyondathletics as understanding optimal fitness has applications for overall health and wellness in larger communities.Therefore, the purpose of this narrative review is to summarize current knowledge of the athletic gut microbiotaand the factors that shape it. Exercise, associated dietary factors, and the athletic classification promote a more“health-associated” gut microbiota. Such features include a higher abundance of health-promoting bacterial species,increased microbial diversity, functional metabolic capacity, and microbial-associated metabolites, stimulation ofbacterial abundance that can modulate mucosal immunity, and improved gastrointestinal barrier function.

Keywords: Athletes, Gut microbiome, Microbial ecology, Gut health, Sports nutrition, Sport performance, Exercise,Physical activity, Metagenome, Short-chain fatty acids

IntroductionThe human gut microbiota contains thousands of differ-ent bacterial taxa as well as various archaea, eukaryoticmicrobes and viruses, more than three million genes,and harbors an enormous metabolic capacity [1, 2]. Themicroorganisms in the gastrointestinal (GI) tract play arole in nutrient uptake, vitamin synthesis, energy

harvest, inflammatory modulation, and host immune re-sponse [3, 4]. In turn, numerous intrinsic and extrinsicfactors can affect the gut microbiota which results in acomplex gut ecosystem that is highly dynamic and indi-vidual [5, 6]. Important factors such as age, birth deliveryroute, antibiotic use, and diet can shape the gut micro-biota [7–10]. The role that exercise plays, in particularthe associated factors and stressors, such as sport/exer-cise-specific diet [11], environment [12], and their inter-actions, on the gut microbiota have been less well

© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you giveappropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate ifchanges were made. The images or other third party material in this article are included in the article's Creative Commonslicence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commonslicence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtainpermission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to thedata made available in this article, unless otherwise stated in a credit line to the data.

* Correspondence: [email protected] of Health Solutions, Arizona State University, Phoenix, AZ, USAFull list of author information is available at the end of the article

Mohr et al. Journal of the International Society of Sports Nutrition (2020) 17:24 https://doi.org/10.1186/s12970-020-00353-w

http://crossmark.crossref.org/dialog/?doi=10.1186/s12970-020-00353-w&domain=pdfhttp://orcid.org/0000-0001-5401-3702http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/publicdomain/zero/1.0/mailto:[email protected]

described. Athletes, although diverse as a populationgiven the wide variety of different types of exercise/fit-ness/athletic training, competition, dietary practices andattributes, exemplify these factors on a generally consist-ent and long-term basis.High-level athletes possess remarkable physiological

and metabolic adaptations (including muscular strength/power, aerobic capacity, energy expenditure, and heatproduction) and provide unique insight in gut micro-biome research. In addition, the gut microbiota with itsability to harvest energy, modulate the immune system,and influence mucosal and brain health, is likely to playa significant role in athlete health, wellbeing, and sportsperformance [13–15]. The microbiota has an indirectinfluence on various indices of exercise performance, re-covery, and patterns of illness, such as signaling throughmyokines and other cytokines, modulating activation ofthe hypothalamic–pituitary–adrenal axis, and affectingperformance-associated metabolic pathways [13, 16–18].Understanding the various roles the gut microbiota playsin relation to athletic performance is of great interest toathletes seeking to improve competition outcomes aswell as reduce recovery time from training. Such know-ledge may be of general benefit to further understandingof microbial contributions to human health and disease.Current research reports a higher abundance of health-promoting bacterial species and increased microbial di-versity in athletes [13, 18, 19].Given the increasing interest in exercise, associated

dietary factors, and athletes as a population in relationto the gut microbiota, the purpose of this narrative re-view is to summarize current knowledge of the athleticgut microbiota and the factors that shape it. While dif-ferences likely exist among how the gut microbiota is af-fected by the different types of sport/athlete/fitnesstraining regimens (e.g., resistance, interval, stretching/flexibility, endurance/aerobic, etc.), the primary aim is toprovide a “state-of-the-art research” statement. Keytopics covered include:

� How the athletic/exercise-associated gut microbiotadiffers in comparison to other populations.

� The effects of different types of exercise training onthe gut microbiota.

� The effects of an ‘athletic diet’ on the gutmicrobiota.

Influencing factors that shape the gut microbiotaNumerous factors such as age, genetics, drug use, stress,smoking, and diet can all affect the microbial compos-ition of the gut, influencing a complex ecosystem that ishighly dynamic and individual [5–7, 20]. For example,the manner in which we are born and raised can resultin substantial differences in the composition of the gut

microbiota. This outcome is related to the differences inexposure (or non-exposure) of bacteria in the birth canalduring vaginal birth [10], being bottle fed or breastfed[21], living with a dog, cat, or close to farm animals [22],the number of antibiotic treatments administered [8],and environmental toxin exposure [23]. From birth untilthe age of about 3 years, an individual assembles theircore of resident microbiota primarily dominated bygram-positive Firmicutes and gram-negative Bacteroi-detes phyla, and this subsequent make-up is as unique asa set of fingerprints [24–26]. The gut microbiota is alsoessential for processing dietary components and appearsto have a major role in shaping the immune system [27].Not surprisingly, the role of the gut microbiota in deter-mining host health and development of disease has beengaining clinical and community interest [9]. Altered gutmicrobiota composition and/or function is linked to agrowing number of conditions from metabolic disordersto some brain-related dysfunctions [28, 29]. Individualsin a known disease state can have a significantly differentgut microbiota composition compared with healthy indi-viduals [30]. A common observation is increased speciesdiversity and/or richness in the gut microbiota of healthyindividuals. Although new research suggests gene con-tent/diversity in the gut may be a better predictor ofphysiological states [31]. In addition, counterexamplesexist as recent work links high gut microbial diversitywith a longer colonic transit time and systemic circula-tion of potentially harmful protein degradation products[32]. On a compositional level, it could be that low di-versity indicates poorer health, while high diversity doesnot always guarantee improved health. Thus, informa-tion about compositional diversity alone is not sufficientto assess the health of the microbiota (and the host).Although, from an ecological perspective, functional di-versity may be a key factor in allowing an ecosystem tocontinue operating properly [33]. Resilience to both ex-ternal and internal changes (with the ability to rapidlyreturn to its baseline functional profile) is likely a keyfeature of a healthy gut microbiota’s ability to maintainitself [34].In individuals without disease, “health-associated”

microbiota is preferred to the term “healthy microbiota”,since gut microbial composition alone cannot predictany state of health or disease according to currentlyavailable research [30, 34]. It may turn out that manypossible states of microbial composition are associatedwith health or indeed that a “health-associated” micro-biota is more resilient and resistant to disruption [35]. Itis also important to keep in consideration that gutmicrobiome composition is quite stable over time [36,37]. For example, ~ 60% of microbiome composition wasfound to be stable up to a 5-year period in US adults[38]. In addition, while species composition varies

Mohr et al. Journal of the International Society of Sports Nutrition (2020) 17:24 Page 2 of 33

tremendously from person to person, there is substantialfunctional redundancy at the metabolic pathway level [1,2, 39]. Therefore, looking at metagenomic functions ra-ther than taxonomy alone may provide for a better ap-preciation of the true gut microbiome metabolic activityand the impact of microbial functions on human physi-ology [40, 41]. Physical activity has been an area ofgrowing interest in gut microbiota research and appearsto promote a health-associated microbial communityand increased metabolic functional potential. This workincludes identifying the impact that varying and diverseathletic and physical activity regimens exert on the gutmicrobiota.Physical activity focused gut microbiota research is

quite new and enabled by dramatic increases in scaleand scope due to advances in DNA-sequencing tech-nologies coupled with computational methods neededgiven the incredible information density of the micro-biota [42–44]. Data are obtained predominantly fromnext-generation sequencing in three forms: A) ribosomalRNA (rRNA) gene sequence surveys that provide a viewof microbiome membership, B) metagenomic data usedto portray functional potential, and, C) metatranscrip-tomic data to describe active gene expression (for a re-view, readers are directed to [42]). Currently, 16 rRNAgene surveys are the most commonly used as they aresubstantially more economical and therefore scale to lar-ger projects [44]. However, this technique is limited byshort read lengths obtained, sequencing errors, and dif-ferences arising from the different regions chosen (e.g.,hypervariable region V3 vs V4) [45]. 16 rRNA sequen-cing also has limited resolution and lower sensitivitycompared to whole-community shotgun metagenomicanalysis, such as characterization down to the genuslevel with minimal capability of species-level detection[46]. Therefore, shotgun metagenomics is displacing 16SrRNA amplicon analysis because of its expanded taxo-nomic range and strain-level resolution [42].Analyses used to interpret large data sets generated

from these high-throughput sequencing techniques com-monly include measures of biological diversity. Many ofthe studies included in this review measured alpha diver-sity which represents diversity within a sample. In calcu-lating alpha diversity, various metrics (e.g., Shannonindex, Chao1) consider the number of unique oper-ational taxonomic units (OTUs), termed ‘richness’, andtheir relative abundance, termed ‘evenness’. Also oftenused is beta-diversity, a measurement of how differentthe communities are between samples. Beta-diversitymetrics are quantitative (e.g., weighted UniFrac), whenconsidering samples phylogeny, and qualitative (e.g., uni-weighted UniFrac) when only evaluating the presence/absence of samples [47]. In addition, other ‘omic’ tech-niques, such as metabolomics, are being integrated with

these data to provide deeper insights into host metabol-ism and health. Metabolomics uses high throughputtechniques to characterize and quantify small moleculesin several biofluids (urine, serum, plasma, feces, saliva),revealing a unique metabolic signature [48]. As a com-plement to sequencing-based approaches, the use ofmetabolomics (particularly from feces) is encouraged asit offers a ‘functional’ readout of the microbiome provid-ing data on the metabolic interplay between the host,diet, exercise, and the gut microbiota [49, 50].While the aforementioned techniques have allowed for

a rapid increase in gut microbiota research, the variationin microbial analysis across different studies can makecomparing/contrasting study findings difficult. Indeed,variation in profiling techniques (e.g., sequencing strat-egy, platform, variable regions, sequencing depth, etc.)may act as a confounding variable resulting in divergentfindings due entirely to laboratory techniques ratherthan treatment [51]. Furthermore, many gut microbiotastudies may be underpowered, and scientists may not becontrolling for important confounding variables such asdiet, gender, ethnicity, GI problems, antibiotics, etc. Asthe investigation of the athletic microbiota is a newerfield of research, the intention of this review is to pro-vide a broad overview of the current state of the litera-ture. For future, more specific reviews, deeperdiscussions of methodological nuances are warranted.

The athlete/exercise-associated gut microbiotaEstablishing consistent relationships across studies inphysically active groups, such as athletes has been prob-lematic. Beyond the obvious methodological differencessuch as sample preparation techniques, DNA sequen-cing, bioinformatics tools, and reference databases [52–55], there is also large variation across exercise/athleticregimens. Moreover, confounding factors includingtraining history, level of physical fitness, training envir-onment, and dietary intake all have the potential toaffect study outcomes substantially, and make detectingdifferences due to exercise/athletic regimens on the gutmicrobiota difficult to ascertain. Therefore, when com-paring these individuals within or across various exer-cise/athletic disciplines and classifications, these factorsshould be accounted for and reported by investigators.While the current body of this comparative-type re-search is mixed and more limited (see Table 1), collect-ively it provides important insight and highlights keyareas of future study.

Athletes/physically active individuals vs other populationsSeveral studies have investigated the difference in com-positional gut microbiota between those physically active(including athletes) and a range of populations. Whilethe above confounding factors are still relevant when


Table 1 Athlete/exercise-associated gut microbiota in comparison to other populations or across athletic classification:Characteristics of included articles (by publication date)

Authors,year,country

Subjects Sex andage

Study design andgut microbiomeanalysis

Diet and/or exercise Key outcome(s)

1. Clarkeet al., 2014,Ireland [19]

Professional male rugbyathletes (n = 40) and healthysize, age and sex matchedcontrols (n = 46)

86Males23–35years

Cross-sectional,observational16S rRNA geneamplification of theV4 region

Diet recorded by a 187-food itemFFQa.Physical activity levels assessedby EPICb-Norfolk questionnaire.

• Athletes had a higher GMc

diversity compared to controls.• Athletes in low BMId (< 25 kg/m2)group had higher proportions ofgenus Akkermansia comparedwith high BMI (> 28 kg/m2) group.

2. Bressaet al., 2017,Spain [56]

Healthy premenopausalwomen; active defined byWHOe (n = 19) and sedentary(n = 21)

40Females18–40years

Cross-sectional,observational16S rRNA geneamplification of theV3 and V4 regions

Acceleration, energy expenditure,intensity of physical activity andbody position measured by Acti-Sleep V.3.4.2 accelerometer.Dietary pattern assessed by 97food item FFQ.

• Physical activity associated withincreased the abundance ofhealth-promoting bacteria (Bifido-bacterium species, Roseburia homi-nis, A. muciniphila andFaecalibacterium prausnitzii) in themicrobiota.

• Inverse association betweensedentary parameters andmicrobiota richness (number ofspecies, and Shannon andSimpson indices).

3. Mörklet al., 2017,Austria [57]

Anorexia nervosa patients(n = 18), athletes (n = 20),normal weight (n = 26),overweight (n = 22), andobese women (n = 20).

106Females18–40years


24-h recall dietary intakeActivity level assessed by IPAQf

score & METg/min

• Lower microbial richness in obeseand anorexic individualscompared to athletes.

4. Petersenet al., 2017,USA [18]

Professional (n = 22) andamateur (n = 11) levelcompetitive cyclists

22Males/11Females19–49years

Cross-sectional,observationalMetagenomic wholegenome shotgunsequencing andRNA sequencing

Exercise load (hours/week): 6–10;11–15; 16–20; and; 20 + .Diet: equal protein, fat,carbohydrate; vegetarian; highcomplex carbohydrate; paleo;gluten-free.

• No significant correlationsbetween taxonomic cluster andprofessional or amateur levelcyclists.

• High relative abundance ofPrevotella in cyclists training > 11h/week.

• Increased abundance ofMethanobrevibacter smithiitranscripts in professional cyclists.

5. Bartonet al., 2018,Ireland [13]

Professional male rugbyathletes (n = 40) and healthysize, age and gendermatched controls (n = 46)

86Males23–35years

Cross-sectional,observationalMetagenomic wholegenome shotgunsequencing andurine and fecalmetabolomics

Diet recorded by a 187-food itemFFQ.Physical activity levels assessedby EPIC-Norfolk questionnaire.Serum creatine kinase levels wereused as a proxy for level of phys-ical activity.

• The microbiota of athletes wasmore diverse than both the lowand high BMI control groups atthe functional level.

• Athletes had an enriched profileof SCFAsh and higher levels of themetabolite TMAOi.

6. Janget al., 2019,SouthKorea [11]

Healthy sedentary men (ascontrols; n = 15), bodybuilders(n = 15), and elite distancerunners (n = 15)

45Males19–28years


Physical activity level wasassessed using the IPAQ.Dietary intake was analyzed withthe computerized nutritionalevaluation program.

• Exercise type was associated withathlete diet patterns(bodybuilders: high-protein, high-fat, low-carbohydrate, and low-dietary fiber diet; distance runners:low-carbohydrate and low dietaryfiber diet). Athlete type was sig-nificantly associated with the rela-tive abundance of gut microbiotaat the genus and species level.

• Increased abundance ofFaecalibacterium, Sutterella,Clostridium, Haemophilus, andEisenbergiella in bodybuilders.

• Athlete type did not differ in gutmicrobiota alpha and betadiversity.

7.O’Donovanet al., 2019,Ireland [58]

Elite athletes from 16different sports (n = 37)

23Males/14Females27 ± 5

Cross-sectional,observationalMetagenomic wholegenome shotgunsequencing and

Diet recorded by FFQ. • Microbial diversity did not differbetween sport classification.

• Overall, samples dominated byspecies from one or acombination of five species:


interpreting these studies, such research provides an im-portant comparator to the gut microbiota of these indi-viduals. In an observational study, the gut microbiotas ofsedentary and physically active premenopausal womenwere compared [56]. Physical exercise was not related todifferences in the microbiota diversity or richness (thetotal number of OTUs or species recorded); however,sedentary parameters (i.e., sedentary time and breaks)correlated negatively with microbiota richness. Further,quantitative polymerase chain reaction analysis revealedhigher abundance of health-promoting bacterial speciesin active women, including Faecalibacterium prausnitzii,Roseburia hominis and Akkermansia muciniphila. In an-other cross-sectional study in females, Mörkl et al. [57]compared the gut microbiota of anorexia nervosa inpa-tients to recreational athletes from a range of sports andoverweight, obese, and normal weight controls. Micro-biota diversity was markedly lower in anorexia nervosapatients and obese participants compared to othergroups, while athletes showed the highest alpha diversity(species richness). Interestingly, total fat mass, serumlipids, C-reactive protein, depression scales, and smokingstatus were negatively associated (R2 = − 0.012 to −0.256, P < 0.05) with microbiota diversity. It is importantto note that these associations are likely driven by life-style. While caution should be used in interpreting thisdata given the cross-sectional nature and inability to ac-count for other potential influences, investigation of gutmicrobiota composition, diversity, and function should

be useful in characterizing key elements of a healthylifestyle.Clarke et al. [19] reported the gut microbiota of pro-

fessional male rugby players was more diverse thanhealthy, non-athlete subjects matched for body massindex (BMI), age, and gender. Given the physical size ofmodern rugby players, two control groups were assessed;one matched for athlete size with a comparable (ele-vated) BMI (> 28 kg/m2) and a second reflecting thebackground age-matched and sex-matched population(lower BMI of < 25 kg/m2). Importantly, the alpha diver-sity of the elite athlete’s microbiota was higher than thatof both control groups. Further, the athletes and thosein the low BMI control group had higher proportions ofthe genus Akkermansia than the high BMI controlgroup. Moreover, protein consumption was correlatedpositively (R = 0.24–0.43) with microbial diversity acrossall groups, indicating that greater protein intake waslinked to higher levels of microbial diversity. There is apossibility that the increased diversity of the athlete’s gutmicrobiota was due, in part, to their higher protein in-take. Barton et al. [13] re-examined the microbiota inthese participants using whole metagenome shotgun se-quencing to provide deeper insight into taxonomic com-position and metabolic potential. Differences in fecalmicrobiota between athletes and sedentary controlsshowed even greater separation at the metagenomic andmetabolomic level than at the gut microbiota compos-itional level. Relative to controls athletes appear to have

Table 1 Athlete/exercise-associated gut microbiota in comparison to other populations or across athletic classification:Characteristics of included articles (by publication date) (Continued)


Subjects Sex andage

Study design andgut microbiomeanalysis

Diet and/or exercise Key outcome(s)

years urine and fecalmetabolomics

Eubacterium rectale,Polynucleobacter necessarius,Faecalibacterium prausnitzii,Bacteroides vulgatus andGordonibacter massiliensis.

• Athletes with high dynamicj

component were associated withgreater abundance ofBifidobacterium animalis,Lactobacillus acidophilus, Prevotellaintermedia and F. prausnitzii.

• Athletes with both a highdynamic and statick componentwere associated with greaterabundance of Bacteroides caccae.

aFFQ Food frequency questionnairebEPIC European Prospective Investigation of CancercGM Gut microbiotadBMI Body mass indexeWHO World Health OrganizationfIPAQ International Physical Activity QuestionnairegMET Metabolic equivalenthSCFA Short-chain fatty acidiTMAO Trimethylamine N-oxidejDynamic Classified by estimated VO2maxkStatic Classified by maximal voluntary contraction


increases in metabolic pathways (e.g., amino acid andantibiotic biosynthesis and carbohydrate metabolism)and fecal metabolites (e.g., microbial produced short-chain fatty acids (SCFAs) including acetate, propionate,and butyrate) associated with enhanced fitness and over-all health when compared to control groups [13].

Athletes across different classifications and disciplinesTo explore possible differences between levels of ath-letes, Petersen et al. [18] investigated the gut microbiotasof 22 professional and 11 amateur competitive cyclists.Using metagenomic whole genome shotgun sequencing,no significant correlations were evident between thetaxonomic clusters in professional or amateur cyclingstatus. However, the amount of exercise (time reportedexercising during the average week) was correlated posi-tively with a greater abundance of the genus Prevotella(≥2.5%). Increased abundance of Prevotella, common innon-Western populations and associated with plant/fiberrich diets, was further positively correlated with severalamino acid and carbohydrate metabolism pathways, in-cluding branched-chain amino acid (BCAA) metabolism.Using metatranscriptomic sequencing, there was in-creased abundance of Methanobrevibacter smithii tran-scripts in a number of the professional cyclists incomparison to amateur cyclists. Moreover, this archaeonhad upregulation of genes involved in the production ofmethane and when methane metabolism was upregu-lated, there was similar upregulation of energy andcarbohydrate metabolism pathways. Similar to Clarkeet al. [19], there was low abundance of Bacteroides inthe athletes. In addition, Akkermansia was present in 30out of 33 cyclists, with seven cyclists having relativeabundances > 2% of this microbe in their metagenomiccommunity. These outcomes may have been a reflectionof higher fitness/metabolism of these athletes as in-creased proportions of Akkermansia are generally associ-ated with a healthier metabolic profile [59].To examine the gut microbiota and metabolome across

sport classification, O’Donovan and colleagues [58] col-lected fecal and urine samples from 37 elite Irish athletesacross 16 different sports (many of whom were participat-ing in the 2016 Summer Olympics). To gain an under-standing of the impact of dynamic (classified by estimatedVO2max) versus static (classified by maximal voluntarycontraction) components of exercise, each sport was clas-sified into a broader sports classification group. Fecal sam-ples were prepared for shotgun metagenomic sequencingand fecal and urine samples underwent metabolomic pro-filing. Athletes participating in sports with a high dynamiccomponent were the most distinct compositionally(greater differences in proportions of species; see Table 1),while athletes participating in sports with both a high dy-namic and static component were the most functionally

distinct (greater differences in functional potential). Fecaland urine derived metabolites also varied between classifi-cation including increased lactate (urine) in sports with astatic component and creatinine (feces) in sports with ahigh dynamic and low static component. While increasedlactate in the more static-based sports was not surprising,increased creatinine in more dynamic-based sports was. Itmay have been that the dynamic exercises in this athleticcohort involved substantial muscle turnover, potentiallyresulting in the increase in the production of creatinine. Itis unclear what the causative factors of this finding andother reported differences in the gut microbiota and me-tabolome were due to the sampling variation, sample sizeand cross-sectional nature of this study. However, thesedifferences were observed despite no significant variationin dietary intakes across athletes from different classifica-tions; suggesting that variations in training loads and com-petition requirements contributed to these microbiomeand metabolome-related patterns.To investigate the long-term effects of a specific exercise

type and athletes’ diets on gut microbiota, Jang et al. [11]compared fecal microbiota characteristics, dietary intake,and body composition in 15 healthy sedentary men (ascontrols), 15 bodybuilders, and 15 distance runners. Exer-cise type was associated with athlete diet patterns (i.e.,bodybuilders: high-protein, high-fat, and low-carbohydrate/dietary fiber diet; distance runners: low-carbohydrate and low-dietary fiber diet). While athletetype did not differ regarding gut microbiota alpha andbeta diversity, it was significantly associated with the rela-tive abundance of several bacteria. For instance, at thegenus level, Faecalibacterium, Sutterella, Clostridium,Haemophilus, and Eisenbergiella were the highest, whileBifidobacterium and Parasutterella were the lowest inbodybuilders. At the species level, intestinal beneficial bac-teria widely used as probiotics (Bifidobacterium adolescen-tis, Bifidobacterium longum, Lactobacillus sakei) andthose producing short chain fatty acids (Blautia wexlerae,Eubacterium hallii) were the lowest in bodybuilders andthe highest in controls. In distance runners, protein intakewas negatively correlated with diversity (Shannon indexR = − 0.63; P = 0.01) and in bodybuilders, fat intake wasnegatively correlated with Bifidobacteria (R = − 0.52; P =0.05). These differences may relate to the nutrition statusof athletes in the study (i.e. insufficient carbohydrate anddietary fiber; higher fat).

Summary of the athlete/exercise-associated gut microbiotaFrom the limited evidence, it appears that athletes har-bor an increased abundance of functional pathwayswithin the microbiome that are exploited by the host forpotential health benefits, as well as carbohydrate degrad-ation and secondary metabolite metabolism compared tocontrol groups [60]. While difficult to isolate, the


influence of dietary intake on many of these differencesis likely and needs to be considered in the context of thisresearch. Furthermore, athletes have an enriched profileof SCFAs, previously associated with numerous healthbenefits and a lean phenotype [61–63]. This profilealong with an increased number of detected SCFA path-ways in athletes is conducive to an enhanced rate ofSCFA production [64]. While evidence is currently lim-ited, results from Clarke et al. [19] and Petersen et al.[18] suggest A. muciniphilia may be more present inathletes than non-athletes. A. muciniphilia is a mucin-degrading bacterium that resides in the nutrient-richmucus layer of the gut that appears to be associated withpositive metabolic function [59]. A. muciniphilia levelsare decreased in obese and type 2 diabetic mice, andtreatment with these bacteria reversed high-fat diet-induced metabolic disorders, including fat mass gain,metabolic endotoxemia, adipose tissue inflammation,and insulin resistance [65]. A. muciniphilia can controlmucus production by the host and restore mucus layerthickness in mice with high-fat diet-induced obesity,thereby reducing gut permeability. This outcome led tothe hypothesis that A. muciniphilia engages in cross-talkwith the intestinal epithelium to control inflammationand gut barrier function [65]. A. muciniphilia has alsobeen found to be enriched in mice fed a ketogenic diet,exhibiting gut-brain functions including conferring seiz-ure protection in two preclinical models for refractoryepilepsy [66]. While the role of A. muciniphila is lesscertain in humans, it is depleted in individuals with sev-eral metabolic and inflammatory disorders. For example,of subjects undergoing dietary calorie restriction treat-ment for obesity, those with higher levels of these bac-teria exhibited the best metabolic status and clinicaloutcomes [59]. Future research in athletes should con-tinue to investigate the role A. muciniphilia plays in thegut microbiota and its functional impact on metabolism.In relation to obesity, some athletes who may be de-

fined as physically active may not necessarily be health-ier based on BMI [67]. For example, Rugby or Americanfootball players commonly have large amounts of leanmass, and many will have relatively healthy percent bodyfat levels, typically 12–20%, so BMI is considered to be apoor measure of obesity status in these athletic cohorts.While Clarke et al. [19] and Barton et al. [13] comparedthe gut microbiome of athletes to matched controls con-sidered overweight by BMI, future work should also in-vestigate this comparison at the obese classification.Findings from such research could provide importantdata in connection with the pathogenesis of obesity andthe gut microbiota. One leading theory on the pathogen-esis of obesity emphasizes a close link between the meta-bolic and immune systems via the gut microbiota [68].This body of work suggests that increased intestinal

permeability from high-fat / high sugar diets allows bac-terial lipopolysaccharide (LPS), an outer membranecomponent of gram-negative bacteria linked with induc-tion of inflammation, from the gut microbiota to trans-locate into the systemic circulation, resulting in systemicendotoxemia. Activation of pro-inflammatory cytokinesis observed, leading to the chronic low-grade inflamma-tion often implicated in obesity [69]. In contrast, athleteshave been noted to have lower levels of circulating LPScompared to sedentary individuals [70]. These findingssupport of the notion that other factors beyond BMIlevels should be considered when assessing any relation-ship between metabolic health, obesity and the micro-biome status of competitive athletes.Barton et al. [13] speculated that the athlete gut

microbiome may possess a functional capacity primedfor tissue repair and a greater ability to harness energyfrom the diet with increased capacity for carbohydratemetabolism, cell structure, and nucleotide biosynthesis.This assertion reflects the significant energy demandsand tissue adaptation that occurs during intense exerciseand elite sport. It appears that being physically active isanother important factor in the relationship between themicrobiota and host metabolism. Intervention-basedstudies to delineate this relationship will be importantand provide further insights into optimal therapies to in-fluence the gut microbiota, and its relationship withhealth and disease as well as athletic performance.

Key Points 1 – Athlete/exercise-associated microbiota

• Although limited in evidence, active individual’s microbiota display ahigher abundance of health-promoting bacterial species such as A.muciniphila and increased diversity.

• Body composition and physical activity are positively correlated withseveral bacterial populations.

• Investigating metagenomic functions rather than taxonomy aloneprovides a more meaningful understanding of gut microbiota and theimpact of microbial metabolic functions on human physiology.

• Athletes have more fecal metabolites (e.g., microbial SCFAs includingacetate, propionate, and butyrate) associated with enhanced muscleturnover (fitness) and overall health than less active individuals. Thesedifferences are likely driven by the effects of exercise training and/ordietary intake.

• The athlete/exercise-associated gut microbiome may possess afunctional capacity primed for tissue repair and a greater ability toharness energy from the diet with increased capacity for carbohydratemetabolism, cell structure, and nucleotide biosynthesis.

The effect of exercise on the gut microbiotaAnimal researchFew studies have focused on the impact that voluntaryexercise has on gut microbiota and, to date, all but sevenof these experimental studies utilized murine models[17, 71]. These preliminary studies indicate that exerciseinfluences the composition of the gut microbiota


community. Matsumoto et al. [72] reported that regularrunning exercise in rats was related to an increase inbutyrate-producing bacteria in the microbiota compos-ition along with an increase of butyrate concentration.Other animal studies demonstrated that daily wheel run-ning exercise may improve some aspects of unhealthystates, such as diet-induced obesity, diabetes, and tox-icity, by impacting the gut microbial composition inmice [73–75]. These effects included altering the ratiobetween the dominant phyla Firmicutes and Bacteroi-detes, which has been found to be correlated with obes-ity and other diseases [76–78]. Animal and human datahave reported that the Firmicutes/Bacteroidetes ratio ishigher (i.e., increased Firmicutes and/or decreased Bac-teroidetes) in obese people compared to lean people[79–82]. However, this is not always the same betweenstudies [83, 84] and may be an oversimplification ofphylum-level patterns in relation to host health. Further,mechanisms by which specific members of the micro-biota, such as Firmicutes and Bacteroidetes, can affecthuman phenotypes remain to be fully elucidated [85].Therefore, changes in phyla ratios should be interpretedwith caution.Amongst the exercise studies in animals there was

little agreement regarding what taxa are influenced bychronic exercise. Other than a positive correlationbetween exercise and Lactobacillus [73, 86, 87], thereare no other taxa that consistently increase in relativeabundance in regularly exercised mice or rats. Inparticular, Choi et al. [73] reported that, in comparisonto a sedentary control, mice with running wheel exercisehad higher phylum Firmicutes but fewer phylaTenericutes and Bacteroidetes, which attenuated changesin gut microbiota induced by oral exposure to theenvironmental toxin, polychlorinated biphenyls. Lambertet al. [75] described that exercised mice presented agreater abundance of Firmicutes species and lowerBacteroides/Prevotella genera compared with sedentarymice. In contrast, Evans et al. [74] concluded thatexercise increased Bacteroidetes, while it decreasedFirmicutes in mice, implying that exercise plays animportant role in prevention of diet-induced obesityproducing a microbial composition similar to lean mice.The changes in the Firmicutes/Bacteroidetes ratioexerted by exercise were inversely proportional to thedistance traveled by animals [74]. Campbell et al. [88]asserted that exercised mice have bacteria related to Fae-calibacterium prausnitzii which may protect the digest-ive tract by producing butyrate and lowering the oxygentension in the lumen by a flavin/thiol electron shuttle.Briefly, the health-related effects of butyrate are associ-ated with anti-inflammatory properties, direct feeding ofcolonocytes, and an impact on satiety [64]. Notably, bu-tyrate, along with propionate and acetate, provides ~

10% of the daily caloric requirements in humans whoconsume high (~ 60 g/day) amounts of dietary fiber [64].The bacterial abundance of Clostridiaceae and Bacteroi-daceae families and Ruminoccocus genus were negativelyassociated with blood lactate levels in exercised animals,whereas a positive association was reported for Oscillos-pira genus [86].Moreover, it seems that the changes exerted by

physical exercise depend on the physiological state ofthe individual. For example, regular forced exercisedifferentially affected microbiota richness whether theywere obese-hypertensive or normal rats [86]. Alterationsto the microbiota exerted by exercise in rats following ahigh-fat diet were different to rats on a normal diet [89],as well as the alterations produced in diabetic mice weredifferent to their control counterparts [75]. Collectively,these findings indicate that modulation of the microbiotaby chronic exercise depends not only on the physio-logical state of the individual, but also on the diet. More-over, another significant difference between forced vs.voluntary exercise in animals is exercise volume. This isrecapitulated in human cyclist data [18] and requiresfurther investigation in animal models. Finally, it hasbeen observed that exercise induces more effectivechanges in the microbiota in juvenile rats than in adultrats [90]. A common finding in these murine studiesexamining the effects of exercise training on the gutmicrobiome, is an increase in alpha diversity [17]. Sev-eral other studies using murine-based models also dem-onstrated increased alpha diversity in animals thatexercised vs. those that were sedentary [73, 74, 86, 87].

Cross-sectional research in humansIn healthy individuals, Estaki et al. [14] reported highercardiorespiratory fitness (as measured by peak ormaximum oxygen uptake, VO2peak or VO2max)correlated with increases in both microbial diversity andfecal butyrate (see Table 2). Also identified was a coreset of gene related functions rather than a core set ofbacterial taxa in individuals with high levels of fitness[14]. Further, ~ 20% of the variation in gut bacterialalpha diversity could be explained by VO2peak alone; infact, VO2peak stood as the only variable that contributedsignificantly to increased alpha diversity. The primaryfindings indicate that cardiorespiratory fitness is a goodpredictor of gut microbial diversity in healthy humans,outperforming several other variables including sex, age,BMI, and multiple dietary components. Additionally,enhanced bacterial diversity was correlated positivelywith certain microbial metabolic functions includingchemotaxis, motility, and fatty acid biosynthesis. AsVO2peak was not significantly associated with variationin community composition, this result suggests functionmay be a better predictor than species richness, as noted


Table 2 Effect of exercise and/or athletic diet on the gut microbiota: Characteristics of included articles (by publication date)


Subjects characteristics Study design and gutmicrobiome analysis

Diet and/or exercise Duration Key outcome(s)

1. Clarkeet al.,2014,Ireland[19]

Professional male rugbyplayers, n = 40, 29 ± 4 years;Healthy matched controls,n = 46, 29 ± 6 years

Cross-sectional16S rRNA geneamplification of the V4region

Observational Cross-sectional

• Athletes had a higherdiversity of gut micro-organisms, representing 22distinct phyla.

• GMa diversity indices werepositively correlated withprotein intake and serumcreatine kinase in athletes.

2. Estakiet al.,2016,Canada[14]

Healthy young males andfemales, n = 39, stratified bycardiorespiratory fitness, Low:25.5 ± 3.3 years; Average:24.3 ± 3.7 years; High: 26.2 ±5.5 years

Cross-sectional16S rRNA geneamplification of the V3 andV4 regions


• VO2peakb, independent of

diet, positively correlatedwith increased GM diversity.

• VO2peak explainedsignificant variation in GMpredicted metagenomicfunctions, aligning positivelywith genes related tobacterial chemotaxis,motility, and fatty acidbiosynthesis.

• Increased production offecal butyrate abundancesof butyrate-producing taxa(Clostridiales, Roseburia,Lachnospiraceae, and Erysi-pelotrichaceae) amongstphysically fit participants.

3. Yanget al.,2017,Finland[94]

Premenopausal females withlow, moderate and high-cardiorespiratory fitness(VO2max), n = 71, LowVO2max: 40.4 ± 3.5 years;Moderate VO2max: 39.7 ± 4.2years;High VO2max: 30.6 ± 5 years

Cross-sectional16S rRNA hybridization,DNA-staining, and flowcytometry


• Decreased Bacteroides,increased Eubacteriumrectale-Clostridium coccoidesin low VO2max participantsvs high VO2maxparticipants.

• VO2max inversely associatedwith Eubacterium rectale-C.coccoides but not withother bacteria.

4. Allenet al.,2018, USA[98]

Sedentary lean males andfemales, n = 18, 25.1 ± 6.52years; Obese males andfemales, n = 14, 31.14 ± 8.57years

Longitudinal design16S rRNA geneamplification of theV4 region

30–60 min moderate-to-vigorous intensity (60–75% ofHRRc) aerobic exercise, 3x perweek

6 weeks • Exercise induced shifts inSCFAd-producing taxa(Faecalibacterium andLachnospira species) andgenetic machinery (BCoATe)were more substantial inlean versus obeseparticipants.

• A return to sedentaryactivity for 6 weeks led to aBMIf-dependent reversion ingut microbiomecomposition.

5. Bartonet al.,2018,Ireland[13]

Professional male rugbyplayers, n = 40, 29 ± 4 years;Healthy matched controls,n = 46, 29 ± 6 years

Cross-sectionalMetagenomic wholegenome shotgunsequencing and urine andfecal metabolomics


• Relative increase inpathways (amino acid andantibiotic biosynthesis andcarbohydrate metabolism)in athletes vs control.

• Increase in fecal metabolites(acetate, propionate andbutyrate) in athletes vscontrol.

6. Croninet al.,2018,Ireland[199]

Sedentary overweight/obesemales and females, n = 90,18–40 years

Randomized controlledtrial,parallel group designMetagenomic wholegenome shotgun

Groups:• Protein-only; 30 g protein(24 g whey)

• Exercise-only; Combinedaerobic and resistance

8 weeks • Increase in alpha diversity inExercise + Protein group vsProtein group.

• Decrease in beta diversityof the gut virome in


Table 2 Effect of exercise and/or athletic diet on the gut microbiota: Characteristics of included articles (by publication date)(Continued)




sequencing and urine andfecal metabolomics

training (aerobic training ofmoderate intensity and re-sistance training of 3 sets of8 repetitions on 7 differentresistance machines) 3 x perweek

• Exercise + protein

participants consumingprotein.

7.Moreno-Pérezet al.,2018,Spain[196]

Endurance-trainedmales, n = 18,35.38 ± 9.0 years (controlgroup), 34.90 ± 9.49 years(protein group)

Randomized controlledtrial,parallel group design16S rRNA geneamplification of the V3 andV4 regions

Groups:• Control: maltodextrin• Protein: blend of wheyisolate (10 g) and beefhydrolysate (10 g)

Minimum training frequencyof 5 endurance trainingsessions per week; ≥240minper week

10 weeks • Increase in Bacteroidetes inprotein group.

• Decrease in Roseburia,Blautia, and Bifidobacteriumlongum in protein group.

8.Taniguchiet al.,2018,Japan[101]

Healthy elderly Japanesemales, n = 33, 62–76 years

Randomized crossover trial16S rRNA geneamplification of the V3 andV4 regions

5-week control period OR 5-week supervised, progressiveaerobic exercise program.Cycle ergometer 3x per week.60% of pre exercise VO2peakduring week 1, 70% duringweeks 2 and 3, and 75% dur-ing weeks 4 and 5. Durationwas 30 min for weeks 1 and2, and 45min for weeks 3–5.

5 weeks • Short-term endurance exer-cise did not appreciably in-fluence diversity andcomposition of gut micro-biota. Minor changes in thegut microbiota were associ-ated with cardiometabolicrisk factors.

• Decreased relativeabundance of C. difficilesignificantly decreased,whereas Oscillospirasignificantly increasedduring exercise ascompared to the controlperiod.

9. Durket al.,2019, USA[93]

Healthy young males andfemales, n = 37, 25.7 ± 2.2years

Cross-sectionalQuantitative PolymeraseChain Reaction (qPCR) thatspecifically measured thequantity of a target gene(16 s RNA) found inFirmicutes and Bacteroidetes


• Firmicutes/Bacteroidetes ratiopositively correlated toVO2max.

• VO2max explained ~ 22% ofthe variance of anindividual’s relative gutbacteria as determined byFirmicutes/Bacteroidetesratio.

10.Keohaneet al.,2019,Ireland[99]

Ultra-endurance maleathletes, n = 4,26.5 ± 1.3 years

A prospective, repeated-measures, within-subjectreportMetagenomic wholegenome shotgunsequencing

Observational 33-dayevent; 3-monthfollow-up

• Increased alpha diversitythroughout event.

• Increased abundance ofbutyrate producing species(Roseburia hominis andmembers of the genusSubdoligranulum) andspecies associated withimproved metabolic health(Dorea longicatena).

• Many of the adaptions inGM community structureand metaproteomicspersisted at 3 monthsfollow up.

11. Kernet al.,2019,Denmark[108]

Overweight/obese males andfemales; n = 88; 20–45 years

Randomized controlledtrial,parallel group design16S rRNA geneamplification of the V4region

Exercise groups:• Habitual living (CON),• Active commuting by bike(BIKE)

• Leisure-time exercise ofmoderate intensity (MOD)

6 months • Increase alpha diversityindex in VIG at 3 monthscompared with CON.

• Beta diversity changed in allexercise groups comparedwith CON; VIG decreased






• Vigorous intensity exercise(VIG)

heterogeneity.• Increased inferredfunctional potential ofmicrobiota in the exercisegroups, primarily at 3months and in MOD.

12. Moritaet al.,2019,Japan[103]

Healthy sedentary elderlyfemales; n = 32; ≥ 65 years

Non-randomizedcomparative trial16S rRNA terminalrestriction fragment lengthpolymorphismanalyses

Exercise groups:• Trunk muscle training group1 h per week.

• Aerobic exercise traininggroup, brisk walking at ≥3METSg, 1 h daily.

12 weeks • Increased Bacteroidesrelative abundance inaerobic exercise group.

• Increased Bacteroidesfollowing exerciseintervention associated withincreased 6-min walk test.

13.Motianiet al.,2019,Finland[109]

Obese sedentary,prediabetic/type 2 diabeticmales and females; n = 26(n = 9, n = 17 respectfully);49 ± 4 years

Parallel group design16S rRNA geneamplification of the V3 andV4 regions

Exercise groups:• Sprint interval traininggroup; 30 s exercise bouts(4–6) of all out cyclingefforts with 4 min ofrecovery, 3x a week.

• Moderate intensitycontinuous training group;40–60 min cycling at 60% ofVO2peak intensity, 3x aweek.

2 weeks • Increased Bacteroidetes inboth groups.

• Decreased Firmicutes:Bacteroidetes ratio in bothgroups.

• Decreased Clostridiumgenus and Blautia.

• Colonic glucose uptakepositively associated withBacteroidetes and inverselywith Firmicutes phylum,Firmicutes: Bacteroidetesratio and Blautia genus.

14.Murtazaet al.,2019,Australia[230]

Elite male endurance racewalkers; n = 21; 20–35 years

Non-randomizedcomparative trial16S rRNA geneamplification of the V6-V8regions

Diet groups:• High-Carbohydrate• Periodized Carbohydrate• ketogenic LCHFh

Consumed during anintensified training program.

3 weeks • Microbiota profiles atbaseline could be separatedby Prevotella or Bacteroidesdominated enterotype.

• LCHF diet resulted inincreased relativeabundance of Bacteroidesand Dorea and decreasedFaecalibacterium.

• Negative correlationsbetween Bacteroides and fatoxidation, and betweenDorea and economy testfollowing LCHFintervention.

15.Scheimanet al.,2019, USA[16]

Experiment 1:• Athletes from the 2015Boston Marathon (n = 15),sedentary controls (n = 10)

Experiment 2:• Ultramarathoners andOlympic trial rowers (n = 87)

Observational16S rRNA geneamplification of the V4region for experiment 1Metagenomic wholegenome shotgunsequencing for experiment2

Experiment 1:• Marathon eventExperiment 2:• Exercise bout

Experiment1:• Fecalsamplescollected 7days beforeand afterMarathonevent

Experiment2:• Pre/postexercise

Experiment 1:• Increased Veillonella relativeabundance in marathonrunners post marathon.

• Veillonella was moreprevalent among runner’svs non-runners, althoughthis was not statisticallysignificant.

Experiment 2:• Increased Veillonellaabundance post exercise.

• Veillonella methylmalonyl-CoA pathway was overrep-resented in athlete metage-nomic samples postexercise.

16. Liuet al.,2019,China

Overweight/obeseprediabetic males; n = 39;20–60 years

Randomized controlledtrial,parallel group designMetagenomic whole

Groups:• Control; No exercise.• High-intensity exercise; 70min combined aerobic and

12 weeks • Microbiota profiles weredifferentially altered inexercise responders (n = 14)vs non-responders (n = 6).


previously. This study also confirmed results byMatsumoto and colleagues [72] who initially reportedthat exercising rats exhibited a positive correlationbetween high-cardiorespiratory fitness and an increasein the SCFA, butyrate. Increases in fecal butyrate wereobserved when relative abundances of Clostridiales,Roseburia, Lachnospiraceae, and Erysipelotrichaceaewere increased [14]. These SCFAs are derived from fer-mentation of undigested plant fiber by the microbiota inthe large intestine.Functional categories most strongly correlated with

VO2peak were related to bacterial motility (bacterialmotility proteins, flagella assembly, and bacterialchemotaxis), sporulation, and to a lesser extent the two-component system known to enable bacterial communi-ties to sense and respond to environmental factors [14].One possible mechanism behind these associations maybe derived from the observation that butyrate, whichwas more abundant among participants with higher car-diorespiratory fitness, can modulate neutrophil chemo-taxis [91]. VO2peak was inversely correlated with LPSbiosynthesis and LPS biosynthesis proteins which wereelevated among less fit participants. LPS is a major com-ponent of the cell wall of gram-negative bacteria andconsidered an endotoxin when present in the blood. Bybinding to extracellular toll-like receptor 4 located on

many cell types, LPS elicits strong inflammatoryresponses that may be detrimental to the host. Continu-ous low-level translocation of LPS into circulation caninduce chronic low-level inflammatory states associatedwith development of obesity and other metabolic syn-dromes [92]. These inflammatory states are thought tobe derived to some extent from inflammatory responsesto blood LPS which is elevated in sedentary humans[70]. Exercise training attenuates inflammation in partby reducing elevated blood LPS [70]. The inverse rela-tionship between VO2peak and LPS biosynthesis path-ways implies a beneficial consequence of increasedphysical activity which may result in decreased LPSbiosynthesis.Durk et al. [93] also explored the relationship between

cardiorespiratory fitness and relative gut microbiotacomposition in healthy young adults showing thatFirmicutes/Bacteroidetes ratio was significantly positivelycorrelated to VO2max. While no other relationshipsbetween the gut microbiota and fitness, nutritionalintake, or anthropometric variables were found, VO2maxaccounted for ~ 22% of the variance of an individual’srelative gut bacteria (as determined by the Firmicutes/Bacteroidetes ratio). In a cross-sectional study in pre-menopausal women, cardiorespiratory fitness was associ-ated with gut microbiota composition, independent of





[106] genome shotgunsequencing and fecalmetabolomics

resistance interval training,3x a week.

• The microbiome ofresponders had increasedfunctional capacity for SCFAbiosynthesis and BCAAi

catabolism.• Exercise-induced gut micro-biota changes were posi-tively correlated withimprovements in glucosehomeostasis and insulinsensitivity.

• Baseline microbiomefeatures accuratelypredicted personalizedexercise responses.

• Fecal microbiotatransplantation fromresponders conferred themetabolic benefits ofexercise in mice.

aGM Gut microbiotabVO2 Volume of oxygen utilizationcHRR Heart rate reservedSCFA Short-chain fatty acideBCoAT Butyryl-CoA:acetate CoA-transferase, a butyrate-regulating genefBMI Body mass indexgMETS Metabolic equivalentshLCHF Low-Carbohydrate High-FatiBCAA Branched-chain amino acid


age and carbohydrate or fat intake [94]. Participants withlow VO2max had lower Bacteroides, but higher Eubac-terium rectale-Clostridium coccoides than the highVO2max group. Aerobic capacity was inversely associ-ated with Eubacterium rectale-Clostridium coccoides butnot with other bacteria. After adjusting for age and diet-ary intake, all significant associations remained.In professional rugby players, who had a unique

dietary pattern (higher energy intake and quantities ofprotein, fat, carbohydrates, sugar and saturated fat perday, with protein accounting for considerably more ofthe total energy) and a high level of physical activity,there was a higher diversity of gut microbiota comparedwith controls [19]. However, it was unclear if this effectwas due to exercise, a high-protein diet, a combinationof these two factors, or other factors [19].

Acute exercise effects on the human gut microbiotaTo investigate the effect of an acute exercise bout inathletes, Zhao et al. [20] examined the fecal metabolitesand microbiota in 20 amateur runners before and after ahalf-marathon race using metabolomics and 16S rRNAsequencing analysis. According to the alpha diversityanalysis, there were few differences in diversity, never-theless, abundances of certain microbiota membersshowed differences before and after running. At thephylum level, Lentisphaerae and Acidobacteria, whosefunctions in the human gut are unknown, were detectedafter running. At the species level there was a markedincrease in the families Coriobacteriaceae and Succinivi-brionaceae. Coriobacteriaceae (phylum Actinobacteria) isinvolved in metabolism of bile salts and steroid hor-mones as well as activation of dietary polyphenols in thehuman gut [95]. Coriobacteriaceae was correlated posi-tively with 15 metabolites, indicating that metabolism ofCoriobacteriaceae may be a potential mechanism under-lying the role of exercise in preventing disease and im-proving health outcomes. These increased metabolitesindicate a microbiota-derived metabolism was promotedby running. At the genus level, half-marathon runningreduced the abundance of fecal Ezakiella, Romboutsia,and Actinobacillus, but increased the abundance ofCoprococcus and Ruminococcus bicirculans. Actinobacil-lus species are purportedly responsible for several dis-tinct animal diseases, such as actinomycosis in cattle,potent septicemia in the neonatal foal, and human peri-odontal disease [96]. Thus, inhibition of this potentialpathogen indicated an anti-inflammatory effect of exer-cise. Interestingly, the pentose phosphate pathway, ametabolic pathway parallel to glycolysis and involvingthe oxidation of glucose, was the most enriched pathwayafter a half-marathon run. These findings highlight amicrobiota-derived mechanism for the health-promotingbenefits of exercise.

In a unique study by Scheiman et al. [16], athletes whowere to run the Boston Marathon were recruited, alongwith a set of sedentary controls to identify gut bacteriaassociated with athletic performance and recovery states.16S ribosomal DNA sequencing was conducted on dailyfecal samples collected up to 1 week before and 1 weekafter the marathon event. The relative abundance of thebacterial genus Veillonella increased after the marathonand was the most differentially abundant microbialfeature between the pre- and post-exercise states. Add-itionally, Veillonella was more prevalent among runnerscompared to non-runners. Veillonella species metabolizelactate into the SCFA acetate and propionate via themethylmalonyl-CoA pathway [97]. To replicate these re-sults in a second experiment and an additional cohort ofhuman athletes, Scheiman and colleagues [16] per-formed shotgun metagenomic sequencing on stool sam-ples from ultramarathoners and Olympic trial rowersboth before and after an exercise bout. Similar findingswere reproduced as relative taxonomic abundances Veil-lonella were increased post-exercise. In addition, theVeillonella methylmalonyl-CoA pathway was overrepre-sented in the metagenomic samples post-exercise acrossthe cohort. Given the limited prevalence of themethylmalonyl-CoA pathway across lactate-utilizingmicrobes, this enrichment post-exercise may implicateVeillonella in causing functional changes in the meta-bolic repertoire of the gut microbiota. It seems that thegenus Veillonella is enriched in athletes after exerciseand the metabolic pathway that Veillonella speciesutilize for lactate metabolism is also enriched. Gutcolonization of Veillonella may augment the Cori cycleby providing an alternative lactate-processing methodwhereby systemic lactate is converted into SCFAs thatre-enter the circulation. Higher levels of lactic acid inathletes’ GI tract favor the growth of this genus andthese bacteria may in turn produce a compound thatcould aid performance.In a third set of experiments, Scheiman et al. [16]

isolated a strain of Veillonella atypica from a stoolsample of one of the aforementioned marathonrunners and inoculated mice. In a pre-clinical cross-over trial, Veillonella inoculated mice had a 13% im-provement in time to exhaustion running tests as wellas significant reductions in inflammatory cytokinespost exercise compared to control. They also more ef-fectively converted lactate to the SCFAs propionate.Importantly, Scheiman et al. [16] found systemic lac-tate in these animals was able to cross the gut barrierinto the lumen, making it available as a substrate formicrobial SCFA conversion. Taken together, these ex-periments revealed that V. atypica improved run timevia its metabolic conversion of exercise-induced lac-tate into propionate, thereby identifying a natural,


microbiome-encoded enzymatic process that enhancesathletic performance. While other studies reportedbutyrate as a prominent feature of the athletic micro-biota [13, 14, 72, 88, 98, 99], this research implicatedpropionate as another significant beneficial SCFA.Moreover, these preclinical, proof-of-concept experi-ments displayed for the first time that beneficial mi-crobes from athletes can be effectively transferred toenhance performance. While performed in animalmodel, this research is notable as it took an import-ant step from correlative to causal function, implicat-ing athlete microbiomes for broader human healthand performance applications.

Chronic exercise effects on the human gut microbiotaIn a longitudinal research design, Allen et al. [98]reported for the first time that exercise training canmodulate the composition and metabolic capacity ofthe human gut microbiota in previously sedentaryindividuals. Lean and obese subjects underwent 6weeks of endurance training with diet controlled,followed by a 6-week washout period [98]. Exercise-induced modulations of the gut microbiota and SCFAwere strongly associated with changes in body com-position in lean participants and VO2max in obeseparticipants, independent of diet. After the washoutperiod, exercise-induced changes in the microbiotawere largely reversed once exercise training ceased.This study supports the notion that gut microbiotacomposition is linked to exercise status and thatexercised-induced changes may be temporary and re-quire continual stimulus [100]. Of note, the exerciseinduced shifts in SCFA-producing taxa (Faecalibacter-ium species [spp.] and Lachnospira spp.) and geneticmachinery (butyrate-regulating gene, BCoAT) weremore substantial in lean versus obese participants.This connection is also supported by the observedshifts in metabolic capacity of the gut microbiotawhich may be transient and likely dependent on re-peated exercise stimuli.Taniguchi et al. [101] evaluated whether endurance

exercise modulates the gut microbiota in elderlysubjects, and whether these changes are associatedwith host cardiometabolic phenotypes. In arandomized crossover trial, 33 elderly Japanese menparticipated in a 5-week endurance exercise program.16S rRNA gene-based metagenomic analyses revealedthat the effect of endurance exercise on gut micro-biota diversity was not greater than interindividualdifferences, whereas changes in alpha diversity indicesduring intervention were negatively correlated withchanges in systolic and diastolic blood pressure, espe-cially during exercise. Microbial composition analysesshowed that the relative abundance of Clostridioides

difficile decreased, whereas that of Oscillospira in-creased during exercise compared to the controlperiod. The changes in these taxa were correlatedwith the changes in several cardiometabolic risk fac-tors. These findings indicate that the changes in gutmicrobiota were associated with cardiometabolic riskfactors, such as systolic and diastolic blood pressure.It appears microbial SCFAs influence blood pressureby interacting with host SCFA receptors [102]. In an-other study with elderly subjects, Morita et al. [103]examined the effect of a 12-week exercise interven-tion on the composition of intestinal microbiota inhealthy women. The relative abundance of intestinalBacteroides increased in subjects that completed aer-obic exercise. In addition, the increases in Bacteroidesfollowing the exercise intervention were positively as-sociated with increases in a 6-min walk test. It iswidely accepted that lower levels of Bacteroides areassociated with the higher prevalence of obesity andmetabolic syndrome and that Bacteroides species mayhelp in suppressing metabolic dysfunction [79, 104].However, Bacteroides sometimes correlate with higherBMI and a Westernized diet [105]; therefore, the in-creased relative abundance observed after the 12-weekexercise intervention may be due to its ability to shiftsubstrates compared to other taxa that more reprodu-cibly decline with a high-fat diet (i.e., Bifidobacteria).More recently, Liu et al. [106] conducted a well-

controlled exercise intervention in 39 prediabetic,medication-naive overweight men with a comprehen-sive metagenomics and metabolomics analysis. Sub-jects maintained their normal dietary routine andwere randomized into either a control group (no ex-ercise; n = 19) or a supervised high-intensity exerciseprogram (n = 20) consisting of 70 min combined aer-obic and resistance interval training 3 times per week.Despite the overall metabolic benefits of the interven-tion, ~ 30% of subjects responded poorly to exercisein terms of improvement in glycemic control andinsulin sensitivity. Those that did respond showed aremarkable decrease in fasting insulin and Homeo-static Model Assessment of Insulin Resistance indexvalues (− 42.70% and − 49.60%, respectively). Upon in-vestigating the gut microbiota there was a clear segre-gation in compositional and functional changesbetween exercise responders and non-responders, ac-companied by distinct alterations in microbial metab-olites. Specifically. responders displayed greater geneexpression of functional pathways generating SCFAsand breaking down BCAAs. These may have been re-lated to the positive findings in glucose metabolismas the rise of BCAA has been associated with insulinresistance [107]. There was also a trend in changes ofthe metabolites of amino-acid catabolism (BCAAs and


aromatic amino acids) and carbohydrate fermentationthat was consistent with the altered patterns of genesencoding for associated metabolic enzymes. Con-versely the microbial profiles of non-responders afterthe exercise intervention shared more similarity withthose of the sedentary controls, suggesting a maladap-tation of gut microbiota in these individuals. Similarto the findings from Barton et al. [13] in professionalrugby athletes, pathways involved in DNA replicationand amino acid metabolism were preferentially en-hanced in responders. Moreover, genes related toglycan biosynthesis and lipid metabolism were ele-vated in responders. Although there was no obviousdifference in baseline microbial structures between re-sponders and non-responders, Liu and colleagueswere able to establish a model based on a machine-learning algorithm from baseline microbiome signa-tures which accurately predicted the exercise out-comes with respect to glycemic control and insulinsensitivity. This raised the possibility of screening forindividuals with high likelihood of exercise resistanceusing gut microbiota, so that personalized adjust-ments can be implemented in time to maximize theefficacy of exercise intervention.To examine the potential causal relationship

between the differentially shaped microbiotas, Liuet al. [106] then transplanted conventional antibiotic-treated mice with the microbiota from two respondersand two non-responders from the above experiment.A substantial improvement in mice gavaged withmicrobiota from responders was mimicked in gly-cemic control and insulin sensitivity, in contrast tothe lack of change in mice colonized with microbiotafrom non-responders. Together, the results from boththe human and animal studies indicate that exercisemay impose a differential impact on the compositionand function of the gut microbiota across individuals.While future research is warranted, this study raisedthe possibility that the makeup of the gut microbiotamay be a determiner for the efficacy of exercise (i.e.,responders vs non-responders) and that targeting thegut microbiota could maximize the benefit of exercise.It may have been that exercise amplified subtle differ-ences of the gut microbiota at baseline by remodelingthe intestinal microenvironment (such as inflamma-tory and oxidative status and local immunity) criticalfor microbial growth and interaction, which ultimatelylead to a divergent response of glycemic control toexercise intervention. Finally, these findings furtherreinforce the notion that the functional capacity ofgut microbiota, as assessed by metagenomics andmetabolomics, can be significantly altered withoutmajor shifts in its community structure, and thatchanges in host phenotype may be more dependent

on the metabolic capacity and metabolites of themicrobiota, instead of the composition per se.In the longest exercise intervention to date, Kern

et al. [108] investigated the effects of regular aerobictraining of different intensities and modalities withsimilar exercise energy expenditure on gut microbiotaover a 6-month period. A total of 88 sedentary over-weight/obese subjects were randomized into fourarms, including habitual living (control), active com-muting by non-motorized bicycle, leisure-time exer-cise of moderate intensity, or vigorous intensityexercise. Beta diversity changed in all exercise groupscompared to control, with participants in the vigorousintensity group showing decreased heterogeneity. Fur-ther, the vigorous exercise group experienced agreater increase in alpha diversity at 3 months com-pared to control. More intense exercise may beneeded to induce change in the gut microbiota insedentary, overweight/obese subjects. In a study ofacute exercise, both high intensity interval and mod-erate continuous training affected the gut microbiotain insulin resistant, sedentary individuals following a2-week exercise intervention [109]. Specifically, Bac-teroidetes increased and the Firmicutes/Bacteroidetesratio decreased. This outcome has relevance to ath-letes as the increase in Bacteroidetes plays an essentialrole in the metabolic conversions of complex sugarpolymers and degradation of proteins [110]. Therewas also a decrease in Clostridium and Blautia gen-era. Clostridium plays an important role in whole-body immune responses, while Blautia purportedlyincreases the release of proinflammatory cytokines[111]. Interestingly, colonic glucose concentrations as-sociated positively with Bacteroidetes and inverselywith Firmicutes phylum, the Firmicutes/Bacteroidetesratio, and Blautia genus. In addition, lower abun-dance of Blautia genus was associated with betterwhole-body insulin sensitivity. These results highlightthe importance of intestinal substrate uptake on thewhole-body and changes, especially in glucose uptake,might have a positive effect on the gut microbiota.Finally, in an observational study, Keohane et al. [99]

explored the gut microbiota response of four well-trained ultra-endurance male athletes to prolonged, highintensity trans-oceanic rowing, describing changes in mi-crobial diversity, abundance, and metabolic capacity.Serial stool samples were obtained from the athletes formetagenomic whole-genome shotgun sequencing to rec-ord microbial community structure and relevant func-tional gene profiles pre-race, mid-race, race-finish, and3months post-race after a continuous, unsupported 33-day, 5000-km transoceanic rowing event. Alpha diversityincreased throughout the ultra-endurance event and wasevident as early as day 17 in the race. This increase


occurred independent of any change in cardiorespiratoryfitness, with VO2max similar pre- and post-race. Varia-tions in taxonomic composition included increasedabundance of butyrate producing species and species as-sociated with improved metabolic health and improvedinsulin sensitivity.The functional potential of bacterial species involved

in specific amino and fatty acid biosynthesis alsoincreased. Specifically, the gene expression of functionalmetabolic pathways involved in L-isoleucine and L-lysine production increased, which play an importantrole in reducing muscular fatigue and damage duringstrenuous exercise [112]. Microbial-derived lysine mayalso contribute to the body protein pool in humans[113]. Changes in essential amino acid availability influ-ence hematopoiesis, which in turn may increase oxygencarrying capacity and cardiorespiratory fitness [112].Many of the adaptations in microbial communitystructure and metaproteomics persisted at 3 monthsfollow up.

Summary of the effect of exercise on the gut microbiotaOverall, the mechanisms by which physical activity maypromote a rich bacterial community and increasedfunctional pathways have not been fully elucidated butlikely involve a combination of intrinsic and extrinsicfactors. For example, physically active individuals aremore likely to be exposed to their environmentalbiosphere (e.g., time spent outdoors) and follow anoverall healthy lifestyle and, consequently, harbor aricher microbiota. Simultaneously, intrinsic adaptationsto endurance training, such as decreased blood flow,tissue hypoxia, and increased transit and absorptivecapacity can lead to changes in the GI tract [114, 115].Changes in GI transit time have been reported to affectthe pH within the colonic lumen which could lead toalterations in the composition of the gut microbiota.For instance, longer colonic transit time is associatedwith decreased gut microbiota diversity, which isparalleled by an increase in pH during transit from theproximal to the distal colon [116, 117]. Repeated boutsof aerobic exercise can increase GI transit time inhealthy individuals and middle-aged patients withchronic constipation [118–120]. However, at higher in-tensities (e.g., above 70% VO2max), gastric emptyingappears to be delayed [121–124]. Aerobic exercise alsoincreases fecal SCFA concentration which can decreasepH in the colonic lumen [125]. Furthermore, metabo-lites that are a by-product of exercise and circulatethroughout the body (e.g., lactate) may filter throughthe gut and serve as an energy source for certain bac-terial taxa (e.g., Veillonella). There is expected competi-tion for nutrients and resources in every ecosystem,including the gut microbiota. Therefore, many of these

microbial characteristics may be a result of ‘form fitsfunction’, as communities in the gut are shaped byavailable resources, as determined by the physiology oftheir host. These and other potential adaptive mecha-nisms, such as a change in gut pH, may create an envir-onmental setting that allows for richer communitydiversity and metabolic functions. Anaerobic capacityand resistance exercise training may also influencecommunity composition, though to date, no work hasexamined these parameters in relation to gutmicrobiota.A single acute bout of prolonged excessive exercise

can have a deleterious influence on intestinalfunction. Intense exercise redistributes blood from thesplanchnic circulation to actively respiring tissues[126]. Prolonged intestinal hypo-perfusion impairsmucosal homeostasis and causes enterocyte injury. In-testinal ischemia may result, particularly in the settingof dehydration, manifesting as abdominal cramps,diarrhea, or occasionally bloody diarrhea [127]. Thisadverse effect is particularly the case in endurancesports [128]. As a result, increased intestinal perme-ability ensues, thought to be driven by the phosphor-ylation of several tight junction proteins [129]. Theseevents render the gut mucosa susceptible to endo-toxin translocation [130]. Moderate endurance exer-cise in mice has been associated with a lesser degreeof intestinal permeability, preservation of mucousthickness, and lower rates of bacterial translocationalong with up-regulated anti-microbial protein pro-duction and gene expression in small intestinal tissue(α-defensin, β-defensin, Reg IIIb and Reg IIIc) [131].Together these changes might help mitigate the ef-fects of stress-induced intestinal barrier dysfunction.In humans, physical activity can improve gastrointes-tinal symptoms in subjects with irritable bowel syn-drome [132]. Collectively, these outcomes areevidence of a differential and dose-response effect ofexercise on gut health, with the underlying mecha-nisms yet to be fully explored in healthy humans.The current body of research supports the role of

exercise as an important behavioral factor that canaffect qualitative and quantitative changes in the gutmicrobial composition and function with benefits tothe host. Although these changes may not occur in asimilar fashion across individuals and may alsodepend on baseline characteristics of both themicrobiota and host. However, based on the currentbody of research, exercise appears to enrichmicrobiota diversity, stimulate the proliferation ofbacteria which can modulate mucosal immunity,improve barrier functions, and stimulate bacteria andfunctional pathways capable of producing substancesthat protect against gastrointestinal disorders and


improve performance (i.e., SCFAs). Indeed, exercisemay be an important intervention to alter gutmicrobiota composition and restore gut symbiosis[100]. However, excessive and/or prolonged high-intensity exercise may not impart these effects. Not-ably, certain taxa may be enriched in athletes such asthe lean phenotype-associated A. muciniphila, andpropionate producing Veillonella (via metabolism oflactate). In addition, higher diversity of microbiotacomposition was associated with lean phenotypescompared to that of obese individuals. It is likely thatthe diverse, metabolically favorable intestinal micro-biota evident in the elite athlete is the cumulativemanifestation of many years of high nutrient intakeand high degrees of physical activity and trainingthroughout youth, adolescence, and during adult par-ticipation in high-level sports [133]. Future areas ofgut microbiota research in relation to athletes and ex-ercise is presented in Table 3.

Key Points 2 – Exercise and gut microbiota

• Higher cardiorespiratory fitness (as measured by VO2peak) appears tobe correlated positively with increases in microbial diversity andmetabolic function, and increases in the SCFA butyrate.

• Exercise training can modulate the composition and metaboliccapacity of the human gut microbiota in sedentary individuals.

• Changes in host phenotype may be more dependent on themetabolic capacity and metabolites of the microbiota, instead ofstrictly microbial composition.

• Changes in the gut microbiota exerted by exercise seem to dependon the physiological state of the individual.

• Gene content/diversity in the gut may be a better predictor ofphysiological states compared to species richness.

• Exercise may promote a rich bacterial community-induced shift inSCFA-producing taxa by providing selective advantage for thecolonization of certain microbes, including A. muciniphila andVeillonella.

• Exercise induced shifts in metabolic capacity of the gut microbiotamay be transient and likely dependent on repeated exercise stimuli.

• Prolonged excessive exercise has a deleterious influence on intestinalfunction, including increased intestinal permeability.

• Nearly all studies included in this review have shown positivecorrelations between gut taxa and exercise. Overall exercise appears toenrich microbiota diversity, stimulate the proliferation of bacteriawhich can modulate mucosal immunity, improve barrier functions, andfunctional pathways capable of producing substances (e.g., butyrateand propionate) that can increase performance and health.

The effect of athletic diet on the gut microbiotaIn researching the human gut microbiota, it is difficultto examine exercise and diet separately. Thisrelationship is compounded by changes in dietaryintakes often associated with physical activity (e.g.,increased protein intake in resistance trained athletes orcarbohydrate intake in endurance athletes and increased

total energy and nutrient intake in general). Athletesoften consume a diet that differs from the generalpopulation with implications on the composition of thegut microbiome.Diet is an established modulator of gut microbiota

composition, with significant alterations reported within24 h of a dietary change [134]. This ability to rapidlychange has implications in research design for thetiming of measurements in exercise studies, as doesdietary composition. Indeed, various food components,dietary patterns, and nutrients all have the potential tosubstantially alter the growth of different gut microbialpopulations. Medication and diet are principalenvironmental factors that influence gut microbiotacomposition according to large-cohort studies [135,136]. The gut microbiota is an important factor thatshapes both energy harvest and storage through metab-olism of proteins and production of several metabolitesincluding SCFAs, ammonia, sulfur-containing metabo-lites such as hydrogen sulfide and methanethiol, andneuroactive compounds such as tryptamine, serotonin,phenethylamine, tryptophan, and histamine [137, 138].Moreover, the gut microbiota can also synthesize denovo amino acids and is involved in the utilization andcatabolism of several amino acids originating from bothalimentary and endogenous proteins. These amino acidscan serve as precursors for the synthesis of other metab-olites produced by the microbiota including SCFAs[139]. Animal studies have revealed communication be-tween the gut microbiota and muscle, in which gutmicrobiota can affect muscle energy homeostasis byinterfering with fat deposition, and lipid and glucose me-tabolism through various metabolites including SCFAsand secondary bile salts [17]. Broadly, athletes consumehigher energy diets compared to sedentary individualsand are often encouraged to consume a diet high incarbohydrate and protein and lower in fat [140]. Duringtraining and competition, fiber intake may be reduced toavoid potential GI issues including gas and distension[141]. Importantly athletes’ dietary plans often accountfor macro- and micronutrient needs, hydration, the tim-ing of nutrients, and dietary supplements, but rarely isthe health of the gut microbiota considered [140]. Herewe describe the influence of total energy intake and theprincipal macronutrient classes (protein, carbohydrate,and fat) on the gut microbiota.

Energy intakeThe GI tract represents the interface between ingestednutrients and the host where energy is effectivelyextracted. In healthy adults, ~ 85% of carbohydrates, 65–95% of proteins, and nearly all fats are absorbed beforeentering the large intestine [142]. Consequently,indigestible carbohydrates and proteins that enter the


Table 3 Future areas of gut microbiota research in relation to athletes and exercise

Current gaps in knowledge:a Recommendation(s) for future studies to address:

1. Athletic-focused studies have assessed the taxonomic composition ofthe gut microbiota, while placing less emphasis on functional capacity.

• The functional capacity of gut microbiota, as assessed by metagenomicsand other omic-driven techniques (e.g., metabolomics), can be signifi-cantly altered without major shifts in community structure. Effects onthe host may be more dependent on the metabolic capacity and me-tabolites of the microbiota, instead of the composition per se. Therefore,investigators should incorporate functional assessment of the gutmicrobiota.

2. At present it is unclear how exercise affects the gut microbiota overlong durations (i.e., > 6 months).

• Assess the effect of exercise on the gut microbiota, as well as the gutmicrobiota in athletes over longer periods of time.

3. Little research is available directly comparing the potential differencesin the gut microbiota between athletic disciplines, exercise routines,training loads, and physical fitness status.

• Examine the potential differences in the gut microbiota between thesefactors in both cross-sectional and longitudinal investigations.

4. There is limited research on how environmental factors affect the gutmicrobiota in athlet

the athletic gut microbiota - journal of the international society … · 2020. 5. 12. · athletic...

Documents