Stadlbauer et al.

BMC Geriatrics (2020) 20:248

https://doi.org/10.1186/s12877-020-01644-2

RESEARCH ARTICLE Open Access

Dysbiosis, gut barrier dysfunction and

inflammation in dementia: a pilot study
Vanessa Stadlbauer1,2*† , Lara Engertsberger1†, Irina Komarova1, Nicole Feldbacher1,2, Bettina Leber3,
Gerald Pichler4, Nicole Fink4, Monika Scarpatetti4, Walter Schippinger4, Reinhold Schmidt5 and Angela Horvath1,2

Abstract
Background: Dementia is an increasing public health threat worldwide. The pathogenesis of dementia has not been
fully elucidated yet. Inflammatory processes are hypothesized to play an important role as a driver for cognitive decline
but the origin of inflammation is not clear. We hypothesize that disturbances in gut microbiome composition, gut barrier
dysfunction, bacterial translocation and resulting inflammation are associated with cognitive dysfunction in dementia.
Methods: To test this hypothesis, a cohort of 23 patients with dementia and 18 age and sex matched controls without
cognitive impairments were studied. Gut microbiome composition, gut barrier dysfunction, bacterial translocation and
inflammation were assessed from stool and serum samples. Malnutrition was assessed by Mini Nutritional Assessment
Short Form (MNA-SF), detailed information on drug use was collected. Microbiome composition was assessed by 16S
rRNA sequencing, QIIME 2 and Calypso 7.14 tools.
Results: Dementia was associated with dysbiosis characterized by differences in beta diversity and changes in taxonomic
composition. Gut permeability was increased as evidenced by increased serum diamine oxidase (DAO) levels and
systemic inflammation was confirmed by increased soluble cluster of differentiation 14 levels (sCD14). BMI and statin use
had the strongest impact on microbiome composition.
Conclusion: Dementia is associated with changes in gut microbiome composition and increased biomarkers of gut
permeability and inflammation. Lachnospiraceae NK4A136 group as potential butyrate producer was reduced in dementia.
Malnutrition and drug intake were factors, that impact on microbiome composition. Increasing butyrate producing
bacteria and targeting malnutrition may be promising therapeutic targets in dementia.
Trial registration: NCT03167983.
Keywords: Microbiome, Diversity, Gut barrier, Inflammation, Cognitive function, Butyrate producer

Background disease (AD) is the most common form of dementia ac-

Dementia leads to disability and dependency among counting for 60–70% of the cases [1]. In AD, pathologic
older people worldwide and thereby has enormous phys- protein aggregates and neuroinflammation mediated by
ical, psychological, social and economic impact on pa- microglia cells are involved in the pathogenesis, how-
tients, caregivers, families and society [1]. Alzheimer’s ever, the exact mechanism is still unclear [2]. Microglia
maturation and function is critically dependent on
* Correspondence: [email protected] short-chain fatty acids produced by the gut microbiome
Vanessa Stadlbauer and Lara Engertsberger shared first authorship and therefore highlights the microbiome as a potential
1
Department of Internal Medicine, Division of Gastroenterology and
Hepatology, Medical University of Graz, Auenbruggerplatz 15, 8036 Graz,
diagnostic and therapeutic target in dementia [3]. During
Austria ageing, the gut microbiome decreases in diversity, loses
2
Center of Biomarker Research in Medicine (CBmed), Graz, Austria beneficial taxa and facultative pathogens increase [4].
Full list of author information is available at the end of the article

© 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 give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if
changes were made. The images or other third party material in this article are included in the article's Creative Commons
licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons
licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain
permission 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 the
data made available in this article, unless otherwise stated in a credit line to the data.
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 2 of 13

Diet and the place of residence play an important role in forms of dementia, inflammatory bowel diseases, liver
the shaping of the microbiome [5, 6]. Ageing is also as- cirrhosis or recent (< 4 weeks) antibiotic or probiotic
sociated with inflammation – often termed as “inflam- treatment. The study (29–420 ex 16/17) was approved
mageing”. Inflammation is further associated with an by the ethics committee Ethic Committee of the Medical
increase in gut permeability, mucosal inflammation and University of Graz (IRB00002556) and has been regis-
bacterial translocation [2]. tered at clinicaltrials.gov (NCT03167983) before the
Since the main risk factor for developing dementia, es- study started. The study was performed according to the
pecially AD, is ageing, it can be hypothesized that the Declaration of Helsinki and Good Clinical Practise
gut-brain axis is a possible link between age and disease guidelines. Written informed consent was obtained be-
related dysbiosis and inflammation. Animal studies sug- fore any study specific procedure was performed from
gest that AD is associated with changes in the gut participants or their legal representatives (in case pa-
microbiome composition with a decrease in beneficial, tients were not able to give written consent any more
anti-inflammatory genera [7]. Furthermore, genetic alter- due to the severity of cognitive dysfunction). Routine
ations in amyloid genes can influence microbiome com- blood biochemistry analysis including full blood count,
position in mice, pointing towards a vicious cycle in AD electrolytes, renal function, liver function, albumin and
development [8]. Recently studies from the USA have total protein levels and inflammation parameter and a
identified a loss in species diversity and differences in detailed medical history was performed. Stool and serum
bacterial composition in the stool of AD patients com- samples were collected for analysis of gut microbiome
pared to matched controls [9]. A study from Japan has composition and biomarkers of intestinal permeability,
also shown that gut microbiome composition is inde- inflammation and bacterial translocation. Serum samples
pendently and strongly associated with dementia [10]. were collected after overnight fasting. Stool samples
Furthermore, it has been recently published that the were collected by the patients or caregivers in sterile col-
microbiome of dementia patients causes a dysregulation lection tubes either on the same day or the evening be-
of the anti-inflammatory P-glycoprotein pathway [11]. fore the study visit. Samples were kept on 4 °C until
So far factors that may influence the composition of arrival at the hospital and then frozen immediately at −
the gut microbiome in patients with dementia have not 80 °C. Mini Nutritional Assessment Short Form [18] was
been studied in detail. Potentially influencing factors used to assess nutritional status.
could be malnutrition, which is common in dementia
and associated with disease severity, [12–14] or drug in- Cognitive function
take, since polypharmacy is a common problem in eld- The Mini-Mental State Examination [19] and the clock
erly persons and the impact of drugs on the microbiome drawing test [20] were used to quantify cognitive func-
has recently gained attention [15–17]. tion. We classified cognitive dysfunction according to
We hypothesize that dementia is associated with dys- the German S3-guideline on Dementia 2016 as MMSE
biosis, gut barrier dysfunction and inflammation and we 0–9: severe; MMSE 10–19: moderate; MMSE 20–26:
aim to identify external factors influencing microbiome mild; MMSE 27–30: no dementia [21].
composition in dementia, such as nutrition and drug in-
take. To study this, we conducted a prospective con- Microbiome analysis
trolled cohort study in patients with dementia and age Total DNA was isolated from frozen stool samples using
matched controls. MagnaPure LC DNA Isolation Kit III (Bacteria, Fungi)
(Roche, Mannheim, Germany) according to manufacturer’s
Methods instructions including mechanic and enzymatic lysis [22].
Between July 2017 and March 2018 we recruited 25 pa- Hypervariable regions V1-V2 were amplified in a target spe-
tients with diagnosis of Alzheimer type (n = 21) or cific PCR using the primers 27F and R357 (27F-AGAGTT
mixed type (Alzheimer type plus vascular type, n = 4) TGATCCTGGCTCAG; R357-CTGCTGCCTYCCGTA)
dementia with a Mini Mental State Examination and sequenced with the Illumina MiSeq technique (Illumina,
(MMSE) ≤ 26 and 18 age and sex matched controls with- Eindhoven, The Netherlands) [22]. Sequencing was done in
out evidence of dementia and a MMSE > 26 at the cooperation with the Core Facility for Molecular Biology at
Albert-Schweitzer Hospital Graz and at the University the Center for Medical Research in Graz.
Hospital Graz. Diagnosis of dementia was made by a
board-certified neurologist/psychiatrist and according to Gut permeability, inflammation and bacterial
ICD10 criteria including cerebral imaging and exclusion translocation
of differential diagnosis by full laboratory workup. Par- Enzyme linked immunosorbent assays (ELISA) were used
ticipants or their legal representative gave written in- to measure fecal and serum calprotectin, fecal and serum
formed consent. We excluded participants with other zonulin, serum diamine oxidase (Immundiagnostic AG,
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 3 of 13

Bensheim, Germany), soluble (s)CD14 (R&D Systems, more microbial communities. All analyses were per-
Minneapolis, USA), and lipopolysaccharide binding pro- formed on feature, genus, family, class, order and
tein (LBP) (Hycult biotech, Uden, The Netherlands). All phylum level. Sequence data is publicly available at the
assays were performed according to manufacturers’ in- NCBI Sequence Read Archive (SRA accession:
structions. Bacterial products (endotoxin, peptidoglycan PRJNA608281).
and bacterial DNA) were detected in serum using HEK- All other statistical analyses were performed using
Blue hTLR4, HEK-Blue hNOD2 and HEK-Blue hTLR9 re- SPSS version 25.0 (SPSS Inc., Chicago, Illinois, USA)
porter cells (Invivogen, Toulouse, France), respectively as and R [30] version 3.5.2 (packages: “mice”, “ggcorrplot”,
published previously [23]. “psych”, “randomForest”, “fmsb”, “stats”, “robustHD”)
[31–37]. Tests (t-test or Mann-Whitney) were chosen
Statistical analysis depending on the distribution of the data assessed by
For microbiome analysis generated FASTQ files were Shapiro-Wilk normality test. Spearman rank correlation
processed for analysis using Qiime2 [24] tools imple- with Benjamini-Hochberg correction for multiple testing
mented in Galaxy (https://galaxy.medunigraz.at). Denois- was used to assess strength and direction of linear rela-
ing (primers removing, quality filtering, correcting errors tionships between variables. All statistical tests were 2-
in marginal sequences, removing chimeric sequences, re- sided, and p-values < 0.05 were considered statistically
moving singletons, joining paired-end reads, and derepli- significant. Data are presented as median and 95% confi-
cation) was done with DADA2 [25]. Taxonomy was dence interval unless stated otherwise. Missing values
assigned based on Silva 132 database release at 99% were imputed by multivariate imputation by chained
OTU level, trained using a Naïve Bayes classifier. Se- equations (package “mice”) [33] based on random forests
quences were blasted in the NCBI database for further (package “randomForest”) [31]. Univariate and multivari-
classification [26]. Features with a total sequence count ate RDA was performed to find out which variables ex-
of less than 10 and/or present in less than two patient plain the variance in microbiome composition. VIF
samples were excluded from analysis. Chloroplast and values were calculated to account for collinearity be-
cyanobacteria filtering were performed to remove con- tween the explanatory variables (package “fmsb”), [37]
taminants. The resulting mean sequencing depth was 41, explanatory variables were standardized for multivariate
631 (range 21,774–53,719) reads per sample. In QIIME2, RDA (package “robustHD”) [35]. Network analysis was
“feature” is the observational unit and describes a se- based on Spearman’s rho associations between taxa and
quence variant/operational taxonomic unit. Analysis was converting the pairwise correlations into dissimilarities
done using the web-based software Calypso 7.14 (http:// to ordinate nodes in a two dimensional PCoA plot.
cgenome.net/calypso/) [27]. For alpha diversity assess-
ment, data was rarefied with a sampling depth of 24,771
reads and Chao1 index, Simpson reciprocal index and Results
Faith phylogenetic diversity were calculated to quantify Patient characteristics
microbial diversity. We recruited 25 patients with dementia (Alzheimer type
Beta diversity and taxon comparison was done on an and mixed type) and 18 matched controls without cogni-
unrarefied feature table after total sum scaling and tive impairment in this prospective controlled cohort
square root transformation. Redundancy analysis based study. From 2 dementia patients we were not able to
on Bray Curtis dissimilarity was used to compare beta collect enough stool and blood samples to do the
diversity between groups and to identify significant con- intended analyses; therefore, they were excluded from
founders. Differentially abundant taxa were identified the final analysis. (Fig. 1) Dementia patients had a lower
with Analysis of Composition of Microbiomes body mass index (BMI) and erythrocyte count as well as
(ANCOM) [28]. This method accounts for compos- lower serum albumin and total protein levels compared
itional constraints and reduces false discovery rates to controls. Accordingly, nutritional status according to
while maintaining high statistical power during the de- MNA-SF was significantly worse in dementia patients.
tection of differentially abundant taxa. This test utilizes Within the dementia group, erythrocyte count (r =
multiple taxon-to-taxon comparisons and infers differen- 0.669, p = 0.002) and albumin (r = 0.707, p < 0.002)
tial abundance of a taxon based on the number of sig- showed a significant positive correlation with MMSE
nificant group comparisons relative to other taxa (W- and clock drawing test showed a weak positive correl-
value). Feature selection was performed using the super- ation with albumin (r = 0.485, p = 0.019). No significant
vised machine learning tool Linear Discriminant analysis differences were found regarding age, gender, and other
Effect Size (LEfSe) [29]. LEfSe is a tool to discover fea- routine biochemistry parameters. BMI did not correlate
tures by way of class comparison, tests of biological with MMSE or clock drawing test results. Collinearity
consistency and effect size estimation between two or analysis showed variance inflation factors (VIF) below 2
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 4 of 13

Fig. 1 Patient flow chart

for MMSE, clock-drawing test, BMI, albumin and MNA- sedatives, metamizole and paracetamol were usually pre-
SF. scribed as needed, whereas the other drugs were pre-
Prescription drug intake was significantly different be- scribed as fixed dose medication. Patient characteristics
tween dementia patients and controls. Dementia patients are shown in Table 1.
took three times more prescription drugs compared to
controls. Antidepressants, laxatives, opioids, anti- Gut microbiome composition
dementia drugs, sedatives, vitamin D and metamizole Alpha diversity using Chao 1 index (Fig. 2a), Simpson
were prescribed nearly exclusively in the dementia reciprocal index (Supplementary figure S1A) or Faith
group, whereas proton pump inhibitors (PPI), antihyper- phylogenetic diversity (Supplementary figure S1C) was
tensive drugs, statins, nonsteroidal anti-inflammatory not significantly different in dementia patients compared
drugs (NSAIDS), paracetamol, antidiabetics, thyroid hor- to age matched controls. Redundancy Analysis (RDA)
mones, calcium and magnesium supplements, anticoagu- showed clear clustering of dementia patients compared
lation and phytotherapeutics were equally prescribed for to controls (Variance 34.3, F = 1.31 p = 0.003). (Fig. 2b)
dementia patients and age matched controls. Laxatives, Alpha diversity also did not change significantly with

Table 1 Patient characteristics. Data are given as median and 95% confidence interval unless stated otherwise
Dementia patients (n = 23) Controls (n = 18) p-value
Age (years) 88 (73;85) 75 (74;76) n.s.
Gender (f/m) (n) 15/8 11/7 n.s.
BMI (kg/m2) 24.8 (22.6; 25.9) 28.1 (25.2; 31.0) p = 0.028
MMSE 16 (13;21) 29 (30;30) p < 0.0001
Clock drawing test 3 (0;5) 7 (7;9) p < 0.0001
Number of prescription drugs 9 (6;11) 3 (1;4) p < 0.0001
MNA-SF 10 (9;12) 14 (14;14) p < 0.0001
Leukocytes (109/L) 6.6 (6.2;8.3) 6.1 (5.4;7.5) n.s
Erythrocytes (1012/L) 4.5 (4.0;4.7) 4.7 (4.4; 5.1) p = 0.028
Thrombocytes (109/L) 220 (216;248) 216 (205;222) n.s
Hemoglobin g/dL 13.2 (12.7;14.4) 14.1 (13.3;14.7) n.s
Creatinine (mg/dL) 0.9 (0.8;1.0) 1.0 (0.9;1.1) n.s
Bilirubin (mg/dL) 0.6 (0.5;0.9) 0.6 (0.5;0.6) n.s
Albumin (g/dL) 3.9 (3.7; 4.1) 4.2 (4.1;4.4) p = 0.006
total protein (g/dL) 7.0 (6.8;7.3) 7.5 (7.3;7.6) p = 0.014
CRP (mg/l) 5 (3;11) 2 (1.2; 3.4) n.s.
BMI body mass index, MMSE Mini mental state examination, MNA-SF Mini Nutritional Assessment Short Form, CRP C reactive protein
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 5 of 13

Fig. 2 Differences in alpha and beta diversity in stool microbiome between dementia patients and controls a Alpha diversity (Chao1) in dementia
patients and controls b Beta diversity (RDA) in dementia patients and controls c Alpha diversity between controls and different stages of
cognitive dysfunction d Beta diversity (RDA) in controls and different stages of dementia

increasing degree of dementia. (Fig. 2c and supplemen- When looking at different stages of dementia, LEfSe
tary figure S1B and D) RDA showed clear clustering of identified one class, 3 orders, 3 families, 18 genera and
different stages of dementia (Variance 94.7, F = 1.2 p = 20 features, being associated with severity of cognitive
0.001). (Fig. 2d) Linear discriminant analysis of effect impairment. Most notably, three Lachnospiraceae spe-
size (LEfSe) identified one family, 5 genera and 7 fea- cies with the corresponding genus Lachnospiraceae
tures to differ between patients with dementia and con- NK4A136 group and the genus Lachnospira were associ-
trols. For example, the features Clostridium ated with health; Faecalibacterium prausnitzii was asso-
clostridioforme, Anaerostipes hadrus and Bacteroides ciated with mild dementia; moderate dementia was
dorei were associated with dementia; Lachnospiraceae associated with Lactobacillus amylovorus and the corre-
bacterium MC-35, another Lachnospiraceae sp., and the sponding higher taxonomic levels (the genus Lactobacil-
genus Lachnospiraceae NK4A136 group were associated lus, the family Lactobacillaceae and the order
with health. (Fig. 3a) Analysis of Composition of Micro- Lactobacillales). Severe dementia was associated with
biomes (ANCOM) confirmed that from the taxa identi- several potential pathogens (e.g. Clostridium clostridi-
fied by LEfSe to discriminate between dementia and forme, Streptococcus salivarius) (Fig. 4a) From these dis-
control, one uncultured Lachnospiraceae feature as well criminating taxa, ANCOM analysis identified the feature
as the genus Lachnospiraceae NK4A136 group were sig- C. clostridioforme and the genus Eisenbergiella to in-
nificantly less abundant in stool of dementia patients. crease with severity of cognitive impairment and the
Additionally, the feature Eubacterium rectale was also family Lactobacillaceae to be highest in patients with
less abundant in stool of dementia patients (Fig. 3b). moderate cognitive impairment. (Fig. 4b).
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 6 of 13

Fig. 3 a Features selected by Linear discriminant analysis Effect Size (LEfSe) to discriminate between dementia patients and controls. b
Differentially abundant taxa between dementia and controls

Association of drug intake and nutrition on microbiome Statins, but none of the other drugs had a significant im-
composition pact on beta diversity (RDA, Variance 35.58, F 1.36, p =
While some drugs were nearly exclusively prescribed in 0.003). LEfSe identified several features, genera, families,
dementia patients (for details see supplementary table S1), orders and classes being associated with use or non-use of
other drugs were equally prescribed between dementia each drug. ANCOM identified several of these taxa as well
and control subjects. To understand how drug use may in- as some taxa that were not discriminative on LEfSe to be
fluence microbiome composition irrespective of the dis- differentially abundant taxa between drug user and non-
ease, we studied the effect of drugs that were equally user. For details see supplementary tables S2–6. Interest-
prescribed in dementia patients and healthy controls on ingly, PPI use was associated with increased abundance of
diversity measures and taxonomic composition: namely oral bacteria (e.g. Streptococcus salivarius) whereas statin
PPI, antihypertensive drugs, statins, thyroid hormones and and antihypertensive drug use was associated with in-
NSAIDS. Paracetamol, antidiabetics and calcium and creased abundance of bacteria known to produce butyrate
magnesium supplements were taken by less than 15% of (e.g. Faecalibacterium sp.).
the cohort and therefore these drugs were not included Since malnutrition was present in 74% of dementia pa-
into the analysis. None of the drugs influenced alpha di- tients but in none of the control persons, microbiome
versity (Chao1, Simpson, Faith phylogenetic diversity). composition in malnourished versus non-malnourished
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 7 of 13

Fig. 4 a Features selected by Linear discriminant analysis Effect Size (LEfSe) to discriminate between dementia different stages of cognitive
dysfunction and controls. b Differentially abundant taxa between stages of cognitive dysfunction

patients was very similar to the results obtained when as well as bacterial translocation by endotoxin, peptido-
comparing dementia versus controls. LEfSe identified glycanes and bacterial DNA in serum. Patients with de-
the feature Ruminococcaceae UCG-014 sp with the cor- mentia had higher DAO levels and sCD14 levels,
responding genus Ruminococcaceae UCG014 and the indicative for an association with increased gut perme-
genus Lachnospiraceae NK4A136 group to be associated ability and increased endotoxin load. (Table 2).
with normal nutritional state. These taxa were also PPI use was associated with significantly increased fae-
found to be associated with healthy controls. The genus cal calprotectin levels (PPI use: 92.5 ng/ml (50.2; 120.5);
Eubacterium hallii group was associated with dementia. PPI non-use: 28.1 ng/ml (20.8; 47.9); p = 0.008). Antihy-
(supplementary table S7) ANCOM confirmed the feature pertensive use was associated with significantly increased
Ruminococcaceae UCG-014 sp. and the genus Lachnos- CRP levels (antihypertensive use: 6 mg/dl (3; 11); antihy-
piraceae NK4A136 group to be differentially abundant pertensive non-use 1.3 mg/dl (1;4); p = 0.016), suggesting
between malnutrition and normal nutrition. complex relations between disease, drug use and
inflammation.
Gut barrier dysfunction, inflammation and bacterial
translocation Multivariate and network analysis of potential factors
We assessed intestinal permeability by serum diami- influencing microbiome composition in dementia
nooxidase (DAO) and fecal zonulin; inflammation by C- To understand the main drivers of dysbiosis in dementia
reactive protein, serum lipopolysaccharide binding pro- we further performed univariate and multivariate RDA
tein (LBP), soluble CD 14 (sCD14) and fecal calprotectin to assess the association of clinical variables and
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 8 of 13

Table 2 Biomarker for gut barrier dysfunction, inflammation and bacterial translocation, Data are shown as median and 95%
confidence interval
Dementia patients (n = 23) Controls (n = 18) p-value
Serum diaminooxidase (U/ml) 20.8 (9.7;29) 11.2 (8.4; 13.8) 0.025
Fecal zonulin (ng/ml) 33.8 (31.2; 57) 55.1 (40.8; 76.7) n.s
C-reactive protein (mg/L) 5 (4; 11) 2 (1.2;3.4) n.s
Serum lipopolysaccharide binding protein (μg/ml) 17.9 (16.1; 18.6) 20.0 (14.6; 21.3) n.s
Soluble CD 14 (μg/ml) 2.4 (1.9; 3.1) 1.8 (1.7; 2.1) 0.022
Fecal calprotectin (ng/ml) 31.5 (26.6; 85.8) 49.0 (18.2; 66.3) n.s
Endotoxin (EU/ml) 0.26 (0.0; 0.33) 0.25 (0.09; 0.53) n.s
Peptidoglycana (ng/mL) 0.96 (0.26; 1.66) 0.42 (0.30;1.05) n.s.
Bacterial DNA (μM) 0.06 (0.00;1.46) 0.7 (0.0; 1.29) n.s
a
peptidoglycan was only measurable in 12% of the samples, therefore median and confidence interval only for the positive samples are shown. CD cluster of
differentiation, EU endotoxin units

biomarkers with microbiome composition. RDA showed composition in dementia. This study therefore supports
that BMI, albumin, total protein, sCD14, statins, NSAI the concept of a disrupted gut-brain axis in dementia.
Ds, number of drugs, MNA-SF, MMSE, clock-drawing The concept of a disrupted gut-brain axis in dementia
test, sex, number of drugs were explanatory variables for has recently emerged. Several animal studies show that
microbiome composition in controls compared to de- induction of dysbiosis by antibiotics, irradiation or
mentia and between different stages of cognitive dys-
function (p < 0.1) (supplementary table S8). To the final Table 3 Multivariate RDA to identify the most important
multivariate RDA model explanatory variables with p < explanatory variables for microbiome composition changes
0.1 in the univariate analysis were included and variables Variable Control versus Dementia Severity of dementia
with VIF > 2 in multicollinearity analysis were excluded. BMI Variance = 33.18 Variance = 33.18
(Table 3) In the multivariate model BMI and statin use F = 1.29 F = 1.29
were the remaining significant explanatory variables for P = 0.006 P = 0.008
differences in microbiome composition between demen-
Total protein Variance = 29.19 Variance = 29.19
tia and control groups and between the groups of de-
mentia severity in the dementia group only. (Table 3) F = 1.14 F = 1.14
Network analysis also illustrates the overlap between fac- P = 0.067 P = 0.070
tors influencing microbiome composition: Genera asso- soluble CD14 Variance = 29.06 Variance = 29.06
ciated with dementia (red) overlap with genera F = 1.13 F = 1.13
associated with no statin intake (yellow) and BMI P = 0.072 P = 0.079
(green), whereas genera associated with health (blue)
Statins Variance = 32.06 Variance = 32.06
overlap with genera associated with statin intake (pur-
ple). (Fig. 5a) When performing network analysis in the F = 1.25 F = 1.25
subgroup of dementia patients only, the overlaps are less P = 0.009 P = 0.014
clear, but again genera associated with severe dementia Clock-drawing test Variance = 25.89 Variance = 25.89
(red) overlap with genera associated with no statin in- F = 1.01 F = 1.01
take (yellow) and genera associated with mild dementia P = 0.376 P = 0.374
(blue) overlap with genera associated with statin intake
Age Variance = 25.88 Variance = 25.88
(purple). The association with BMI is less pronounced in
the dementia subgroup. (Fig. 5b). F = 1.01 F = 1.01
P = 0.427 P = 0.409
Discussion Sex Variance = 27.63 Variance = 27.63
Our cross-sectional controlled pilot cohort study shows F = 1.08 F = 1.08
that dementia is associated with changes in microbiome P = 0.137 P = 0.154
composition including a reduction in bacteria known to
NSAIDS Variance = 28.17 Variance = 28.17
produce short chain fatty acids (SCFA) and increased
biomarkers of gut permeability and inflammation. Fur- F = 1.10 F = 1.10
thermore, we could show that both malnutrition and P = 0.098 P = 0.107
drug intake are factors associated with microbiome BMI body mass index, NSAIDS non-steroidal anti-inflammatory drugs
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 9 of 13

Fig. 5 Network analysis to identify associations between bacteria and selected host variables. Taxa and explanatory variables are represented as
nodes, taxa abundance as node size, and edges represent positive associations. Nodes (genera) are coloured based on their association with
selected host variables (dementia/health, dementia stages, statin use or non-use, BMI). a whole cohort (n = 41), b dementia patients (n = 21)

germ-free conditions negatively impact on cognitive and Lachnospiraceae NK4A136 group was lower in de-
function and plaque deposition as recently reviewed by mentia patients compared to controls. LEfSe also identi-
Ticinesi et al. [38] Recently, in patients with dementia a fied the family Lachospiraceae with its genus
reduction in diversity of the microbiome has been de- Lachnospiraceae NK4A136 and several Lachnospiraceae
scribed, however data on taxonomic microbiome com- species to be associated with health. Eubacterium rectale
position are varying between different studies from is a well-known butyrate producing bacterium [43] and
different geographical locations [9, 10, 39]. We describe has already previously been associated with cognitive de-
altered beta diversity and distinct taxonomic changes in cline [44]. Members of the Lachnospiraceae family have
dementia in a European cohort. Alpha diversity data in been linked to obesity on the one hand and protection
dementia so far are conflicting, since a lower alpha di- from colon cancer in humans on the other hand, likely
versity has been observed in the study from the USA [9], due to the association of many species within the group
whereas in the Japanese study a lower alpha diversity with the production of butyric acid, a SCFA that is im-
was observed in the control group [10] but we found un- portant for host epithelial cell growth and integrity [45].
changed alpha diversity of the gut microbiome between Mild dementia was also associated with another butyate
European dementia patients and healthy age matched producer – F. prausnitzii [46]. Increasing the number of
controls. Microbial diversity depends on many factors, butyrate producing bacteria in dementia may therefore
especially in elderly, where the microbiome is likely to be a promising therapeutic approach, since SCFA such
be less stable [40]. Elderly people are more often ex- as butyrate are critically involved in microglia matur-
posed to microbial community-altering events, such as ation and function [2, 3]. Data from animal and human
infections and concomitant antibiotic use, polypharmacy pilot studies support this concept. A dietary intervention
or hospital stays. Therefore, elderly controls may not be with bilberry anthocyanin extract was able to increase
healthy in a strict sense and selection of the control co- Lachnospiraceae NK4A136 group abundance and im-
horts may account for the different findings in different prove gut barrier function in ageing rats [47]. An ex-
studies. Differences in analysis techniques [41] and also ploratory pilot study in patients with dementia showed
in geographic location [42] may be further factors that that a multispecies probiotic can increase the abundance
impact on diversity and composition of the gut of butyrate producing bacterial strains [48]. We also
microbiome. found that the abundance of C. clostridioforme and the
When looking at taxonomic differences, abundance of genus Eisenbergiella increased with increasing cognitive
Eubacterium rectale, an uncultured Lachnospiraceae sp. impairment. C. clostridioforme has mainly been
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 10 of 13

described as a human pathogen [49] but has also been striking. The known consequences of polypharmacy are
described to be associated with vegetarian diet [50]. The among others, cognitive impairment, a higher risk of falls
genus Eisenbergiella was recently found to be increased and non-compliance, but interventions to reduce polyphar-
in long lived adults [51]. Therefore these findings are macy are difficult [61, 62]. Drug-microbiome interactions
difficult to interpret in the context of cognitive dysfunc- are increasingly recognized. A population based deep se-
tion. The family Lactobacillaceae was differentially quencing study revealed, that proton pump inhibitors (PPI)
abundant in different stages of dementia, with the high- were associated with the most profound microbiome
est abundance in moderate dementia and a lower abun- changes, followed by statins, antibiotics, laxatives and beta
dance in mild dementia and severe dementia. LEfSe also blockers [17]. It has been shown experimentally that not
revealed the bacterium L. amylovorus and the corre- only classic antimicrobials but also many other human-
sponding genus Lactobacillus, the family Lactobacilla- targeted drugs have an extensive impact on human gut bac-
ceae and the order Lactobacillales to be associated with teria [15]. We recently showed that PPI are one of the main
moderate dementia. This finding is also difficult to inter- drivers of dysbiosis in liver cirrhosis [63, 64]. We therefore
pret, since Lactobacillus sp. in general were already more assessed the association of prescription drugs with gut
than 100 years ago associated with longevity by the No- microbiome composition. As expected, effects on overall
bel prize winner Elie Metchnikow in 1907 [52] and sev- community structure (alpha and beta diversity) were small.
eral studies using different Lactobacillus sp. have been Each drug class was associated with distinct associations
conducted with varying success in neurodegenerative throughout different taxonomic levels between users and
diseases [53]. Lactobacillus amylovorus has been de- non-users. PPI use was associated with higher abundance of
scribed as a novel probiotic strain that is able to reduce oral bacteria in the stool and statins and antihypertensive
ammonia levels and may therefore be associated with use was associated with an increase in SCFA producing bac-
cognitive function [54]. teria. Due to the small sample size, the results have to be
Our study also shows that dementia is not only associ- interpreted with caution and can only serve as pilot data that
ated with dysbiosis but also associated with markers of need to be explored in larger cohorts. Additionally, we
increased gut permeability (DAO) and markers of in- assessed the impact of drug intake on markers of gut perme-
flammation (sCD14). Ageing itself has been associated ability, bacterial translocation and inflammation. We found
with an increase in gut permeability, mucosal inflamma- that PPI use was associated with increased intestinal inflam-
tion and bacterial translocation – often termed as mation. This has been previously described in the context of
“inflammageing” [2]. Increased calprotectin levels in other diseases [65–67] and we have recently linked dysbiosis,
stool as a sign of intestinal inflammation have been ob- gut permeability and intestinal inflammation to adverse out-
served in a pilot study [55]. Another study in dementia come in patients with liver cirrhosis who use PPI [63]. Anti-
showed a decrease in previously elevated zonulin levels hypertensive use was associated with slightly, but
after probiotic treatment as a possible hint towards a significantly elevated CRP levels, which is most likely due to
causal link between dysbiosis and gut permeability in de- the underlying disease and not to the drug itself, since arter-
mentia [48]. Although we did not find any differences in ial hypertension is associated with elevated CRP levels [68]
stool zonulin and calprotectin levels, we found an in- and therefore validates the relevance of our findings al-
crease in DAO levels which has been proven to be a though sample size is small.
valuable serum biomarker of gut barrier dysfunction Malnutrition is common in dementia and nutrition
[56–59]. Also, Ginko biloba, a commonly used phy- care is an integral part of dementia care [69]. Although
totherapeutic drug in dementia, was able to reduce DAO all patients in our dementia cohort were treated accord-
levels in an animal model of alcoholic liver disease, indi- ing to nutritional support standards that include oral nu-
cating both the validity of DAO as a permeability bio- tritional supplements in patients with MNA-SF < 9,
marker and that gut hyperpermeability may be a MNA-SF and laboratory parameters showed that more
modifiable and relevant therapeutic target [60]. We fur- than three quarter of the dementia patients in our study
thermore found elevated sCD14 levels in dementia as a were malnourished. It is therefore impossible from this
marker of endotoxemia and inflammation. Recent cross-sectional pilot study to distinguish if malnutrition
in vitro data suggest that the altered stool microbiome or dementia are the starting point of dysbiosis. This
composition in dementia directly modulates intestinal could only be answered by longitudinal studies. In gen-
epithelial homeostasis via the anti-inflammatory P- eral, malnutrition has been associated with differences in
glycoprotein pathway [11]. microbiome composition, such as loss of bifidobacteria,
In our cohort, dementia patients, although not different however, most studies were performed in malnourished
regarding age and gender from our controls, received 3 children and not in elderly people [70].
times more prescription drugs. Although some of these In order to identify the drivers of dysbiosis we per-
drugs were only prescribed on demand, this finding is still formed multivariate analyses with all variables and
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 11 of 13

excluded variables that showed high collinearity to antihypertensive users and non-users. Table S5. Features selected by
understand the driver of dysbiosis. We found that BMI LEfSe to discriminate between NSAID users and non-users. Table S6.
and statin use were the strongest influencing factors, Features selected by LEfSe to discriminate between thyroid hormone
users and non-users. Table S7. Features selected by LEfSe to discriminate
underpinning the notion that malnutrition and prescrip- between malnourished and non-malnourished participants. Table S8.
tion drug use drive microbiome composition in demen- Redundancy analysis with explanatory variables of microbiome
tia. Also network analysis supports the close association composition changes.
of these factors. However, due to the small sample size,
the results of our multivariate analysis have to be inter- Abbreviations
AD: Alzheimer’s disease; ANCOM: Analysis of Composition of Microbiomes;
preted with caution and can be seen as hypothesis gen- BMI: Body mass index; CRP: C reactive protein; DAO: Diamino oxidase;
eration only. The results need to be confirmed in larger LEfSe: Linear Discriminant analysis Effect Size; LBP: Lipopolysaccharide
studies. binding protein; MMSE: Mini Mental State Examination; MNA-SF: Mini
Nutritional Assessment Short Form; NSAIDS: Nonsteroidal anti-inflammatory
Our study has some limitations: First, the single center drugs; PPI: Proton pump inhibitors; RDA: Redundancy Analysis; SCFA: Short
approach and the sample size limit the generalizability of chain fatty acids; sCD14: Soluble CD14 levels; VIF: Variance inflation factor
the data. A combination of all studies on gut micro-
biome in dementia would be desirable, however, due to Acknowledgements
We thank Ms. Bianca Schmerböck and Ms. Birgit Reinhart for their excellent
the lack of standards in sequencing techniques this work as study coordinators and we thank Dr. Ingeborg Klymiuk and her
would not be technically feasible. Second, we could not team from the Core Facility Molecular Biology for performing the 16S
perform the gold standard analysis of gut permeability – sequencing.
the differential sugar absorption test – because of the
Authors’ contributions
cognitive impairment of our patients, who were not able VS planned the study, acquired funding, performed analyses, analyzed data
to follow the instructions of the test. We overcame this and wrote the manuscript; LE performed analyses, analyzed data and wrote
by using a panel of serum and stool markers that do not the manuscript; IK analyzed data and wrote manuscript, NF collected data,
performed analyses, and critically reviewed the manuscript; BL performed
require compliance with test instructions for the partici- analyses and critically reviewed the manuscript, GP, NF and MS recruited
pants. And third, this study only provides cross-sectional patients, collected data and critically reviewed the manuscript, RS and WS
data and can therefore not answer any questions regard- critically reviewed the manuscript, AH analyzed data and wrote the
manuscript. All authors have read and approved the manuscript
ing causality or cause-effect relationship between
cognitive dysfunction dysbiosis and malnutrition. A lon- Funding
gitudinal study is in planning. The project was funded by the Center for Biomarker Research in Medicine, a
COMET K1 center funded by the Austrian Research Promotion Agency
(Project 3.9 Biomarkers of the gut-brain-axis). The funding body was not
Conclusion directly involved in the design, data collection, analysis, interpretation or
writing of the manuscript.
In summary this study provides evidence that structural
changes in microbiome composition in dementia are as- Availability of data and materials
sociated with malnutrition and prescription drug use Sequence data is publicly available at the NCBI Sequence Read Archive (SRA
and that biomarkers of gut permeability are increased in accession: PRJNA608281). The remaining datasets used and/or analyzed
during the current study are available from the corresponding author on
dementia. Further studies to move from associations to reasonable request.
causality in understanding the gut-brain axis in dementia
are necessary. Increasing butyrate producing bacteria Ethics approval and consent to participate
and targeting malnutrition seems to be promising thera- The study (29–420 ex 16/17) was approved by the ethics committee Ethic
Committee of the Medical University of Graz (IRB00002556). The study was
peutic approaches to treat dementia related dysbiosis. performed according to the Declaration of Helsinki and Good Clinical
The effect of microbiome modulating strategies on cog- Practise guidelines. Written informed consent was obtained before any study
nitive function needs to be addressed in future studies. specific procedure was performed from participants or their legal
representatives (in case patients were not able to give written consent any
more due to the severity of cognitive dysfunction). This procedure was
Supplementary information approved by the ethics committee.
Supplementary information accompanies this paper at https://doi.org/10.
1186/s12877-020-01644-2. Consent for publication
Not applicable.
Additional file 1: Figure S1. Alpha diversity in stool microbiome
between dementia patients and controls A: Alpha diversity (Faith PD) in Competing interests
dementia patients and controls B: Alpha diversity (Faith PD) in controls The authors declare no conflict of interest.
and different stages of cognitive dysfunction C: Alpha diversity (Simpson)
Author details
between dementia patients a D: Alpha diversity (Simpson) in controls 1
and different stages of cognitive dysfunction. Table S1. Drug intake in Department of Internal Medicine, Division of Gastroenterology and
dementia patients and controls. Table S2. Features selected by LEfSe to Hepatology, Medical University of Graz, Auenbruggerplatz 15, 8036 Graz,
discriminate between PPI users and non-users. Table S3. Features Austria. 2Center of Biomarker Research in Medicine (CBmed), Graz, Austria.
3
Department of Surgery, Division of Transplantation Surgery, Medical
selected by LEfSe to discriminate between statin users and non-users.
Table S4. Features selected by LEfSe to discriminate between University of Graz, Graz, Austria. 4Department of Neurology, Geriatric Health
Centers Graz, Albert Schweitzer Hospital, Graz, Austria. 5Clinical Division of
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 12 of 13

Neurogeriatrics, Department of Neurology, Medical University of Graz, Graz, 20. Shulman KI. Clock-drawing: is it the ideal cognitive screening test? Int J
Austria. Geriatr Psychiatry. 2000;15(6):548–61.
21. S3-Leitlinie Demenzen. 2016. Deutsche Gesellschaft für Neurologie, Hrsg.
Received: 9 September 2019 Accepted: 9 July 2020 Leitlinien für Diagnostik und Therapie in der Neurologie.
22. Klymiuk I, Bilgilier C, Stadlmann A, Thannesberger J, Kastner MT, Hogenauer
C, Puspok A, Biowski-Frotz S, Schrutka-Kolbl C, Thallinger GG, et al. The
human gastric microbiome is predicated upon infection with helicobacter
References pylori. Front Microbiol. 2017;8:2508.
1. Dementia. WHO fact sheet 2016. 23. Horvath A, Leber B, Schmerboeck B, Tawdrous M, Zettel G, Hartl A, Madl T,
2. Kohler CA, Maes M, Slyepchenko A, Berk M, Solmi M, Lanctot KL, Carvalho Stryeck S, Fuchs D, Lemesch S, et al. Randomised clinical trial: the effects of
AF. The gut-brain Axis, including the microbiome, leaky gut and bacterial a multispecies probiotic vs. placebo on innate immune function, bacterial
translocation: mechanisms and pathophysiological role in Alzheimer's translocation and gut permeability in patients with cirrhosis. Aliment
disease. Curr Pharm Des. 2016;22(40):6152–66. Pharmacol Ther. 2016;44(9):926–35.
3. Erny D, Hrabe de Angelis AL, Jaitin D, Wieghofer P, Staszewski O, David E, 24. Hall M, Beiko RG. 16S rRNA Gene Analysis with QIIME2. Methods Mol Biol
Keren-Shaul H, Mahlakoiv T, Jakobshagen K, Buch T, et al. Host microbiota (Clifton, NJ). 2018;1849:113–29.
constantly control maturation and function of microglia in the CNS. Nat 25. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJ, Holmes SP.
Neurosci. 2015;18(7):965–77. DADA2: high-resolution sample inference from Illumina amplicon data. Nat
4. O'Toole PW, Jeffery IB. Gut microbiota and aging. Science. 2015;350(6265): Methods. 2016;13(7):581–3.
1214–5. 26. Coordinators NR. Database resources of the National Center for
5. Xu Z, Knight R. Dietary effects on human gut microbiome diversity. Br J biotechnology information. Nucleic Acids Res. 2016;44(D1):D7–19.
Nutr. 2015;113(Suppl):S1–5. 27. Zakrzewski M, Proietti C, Ellis JJ, Hasan S, Brion MJ, Berger B, Krause L.
6. Haran JP, Bucci V, Dutta P, Ward D, McCormick B. The nursing home elder Calypso: a user-friendly web-server for mining and visualizing microbiome-
microbiome stability and associations with age, frailty, nutrition and physical environment interactions. Bioinformatics. 2017;33(5):782–3.
location. J Med Microbiol. 2018;67(1):40–51. 28. Mandal S, Van Treuren W, White RA, Eggesbo M, Knight R, Peddada
7. Zhang L, Wang Y, Xiayu X, Shi C, Chen W, Song N, Fu X, Zhou R, Xu YF, SD. Analysis of composition of microbiomes: a novel method for
Huang L, et al. Altered gut microbiota in a mouse model of Alzheimer's studying microbial composition. Microb Ecol Health Dis. 2015;26:
disease. J Alzheimers Dis. 2017;60(4):1241–57. 27663.
8. Harach T, Marungruang N, Duthilleul N, Cheatham V, Mc Coy KD, Frisoni G, 29. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS,
Neher JJ, Fak F, Jucker M, Lasser T, et al. Reduction of Abeta amyloid Huttenhower C. Metagenomic biomarker discovery and explanation.
pathology in APPPS1 transgenic mice in the absence of gut microbiota. Sci Genome Biol. 2011;12(6):R60.
Rep. 2017;7:41802. 30. Team RC. R: a language and environment for statistical computing. Vienna:
9. Vogt NM, Kerby RL, Dill-McFarland KA, Harding SJ, Merluzzi AP, Johnson SC, R Foundation for Statistical Computing; 2019.
Carlsson CM, Asthana S, Zetterberg H, Blennow K, et al. Gut microbiome 31. Liam A, Wiener M. Classification and regression by randomForest. R News.
alterations in Alzheimer's disease. Sci Rep. 2017;7(1):13537. 2002;2(3):18–22.
10. Saji N, Niida S, Murotani K, Hisada T, Tsuduki T, Sugimoto T, Kimura A, Toba 32. Kassambara A. ggcorrplot: Visualization of a Correlation Matrix using
K, Sakurai T. Analysis of the relationship between the gut microbiome and 'ggplot2; 2019.
dementia: a cross-sectional study conducted in Japan. Sci Rep. 2019;9(1):1008. 33. van Buren S, Groothuis-Oudshoorn K. mice: Multivariate Imputation by
11. Haran JP, Bhattarai SK, Foley SE, Dutta P, Ward DV, Bucci V, McCormick BA. Chained Equations in R. J Stat Softw. 2011;45(3):1–67.
Alzheimer's Disease Microbiome Is Associated with Dysregulation of the 34. Revelle W. psych: Procedures for Personality and Psychological Research.
Anti-Inflammatory P-Glycoprotein Pathway. mBio. 2019;10(3):e00632–19. Evaston: Northwestern University; 2018.
12. Auer SR, Hofler M, Linsmayer E, Berankova A, Prieschl D, Ratajczak P, Steffl 35. Alfons A. robustHD: Robust Methods for High-Dimensional Data; 2019.
M, Holmerova I. Cross-sectional study of prevalence of dementia, 36. Rundel C, Cetinkaya-Rundel M, Clyde M, Banks D. statsr: Companion
behavioural symptoms, mobility, pain and other health parameters in package for statistics with; 2018.
nursing homes in Austria and the Czech Republic: results from the DEMD 37. Nakazawa M. fmsb: Functions for Medical Statistics Book with some
ATA project. BMC Geriatr. 2018;18(1):178. Demographic Data; 2018.
13. Sanders CL, Wengreen HJ, Schwartz S, Behrens SJ, Corcoran C, Lyketsos CG, 38. Ticinesi A, Tana C, Nouvenne A, Prati B, Lauretani F, Meschi T. Gut
Tschanz JT, Cache County I. Nutritional status is associated with severe microbiota, cognitive frailty and dementia in older individuals: a systematic
dementia and mortality: the Cache County dementia progression study. review. Clin Interv Aging. 2018;13:1497–511.
Alzheimer Dis Assoc Disord. 2018;32(4):298-304. https://doi.org/10.1097/ 39. Araos R, Andreatos N, Ugalde J, Mitchell S, Mylonakis E, D'Agata EMC. Fecal
WAD.0000000000000274. PMID: 30188355; PMCID: PMC6345543. microbiome among nursing home residents with advanced dementia and
14. Vandewoude MFJ, van Wijngaarden JP, De Maesschalck L, Luiking YC, Van Clostridium difficile. Dig Dis Sci. 2018;63(6):1525–31.
Gossum A. The prevalence and health burden of malnutrition in Belgian 40. Claesson MJ, Cusack S, O'Sullivan O, Greene-Diniz R, de Weerd H, Flannery E,
older people in the community or residing in nursing homes: results of the Marchesi JR, Falush D, Dinan T, Fitzgerald G, et al. Composition, variability,
NutriAction II study. Aging Clin Exp Res. 2019;31(2):175-83. https://doi.org/ and temporal stability of the intestinal microbiota of the elderly. Proc Natl
10.1007/s40520-018-0957-2. Acad Sci U S A. 2011;108(Suppl 1):4586–91.
15. Maier L, Pruteanu M, Kuhn M, Zeller G, Telzerow A, Anderson EE, Brochado 41. Allali I, Arnold JW, Roach J, Cadenas MB, Butz N, Hassan HM, Koci M, Ballou
AR, Fernandez KC, Dose H, Mori H, et al. Extensive impact of non-antibiotic A, Mendoza M, Ali R, et al. A comparison of sequencing platforms and
drugs on human gut bacteria. Nature. 2018;555(7698):623–8. bioinformatics pipelines for compositional analysis of the gut microbiome.
16. Wastesson JW, Cedazo Minguez A, Fastbom J, Maioli S, Johnell K. The BMC Microbiol. 2017;17(1):194.
composition of polypharmacy: a register-based study of swedes aged 75 42. Gupta VK, Paul S, Dutta C. Geography, Ethnicity or Subsistence-Specific Variations in
years and older. PLoS One. 2018;13(3):e0194892. Human Microbiome Composition and Diversity. Front Microbiol. 2017;8:1162.
17. Zhernakova A, Kurilshikov A, Bonder MJ, Tigchelaar EF, Schirmer M, Vatanen 43. Barcenilla A, Pryde SE, Martin JC, Duncan SH, Stewart CS, Henderson C, Flint
T, Mujagic Z, Vila AV, Falony G, Vieira-Silva S, et al. Population-based HJ. Phylogenetic relationships of butyrate-producing bacteria from the
metagenomics analysis reveals markers for gut microbiome composition human gut. Appl Environ Microbiol. 2000;66(4):1654–61.
and diversity. Science. 2016;352(6285):565–9. 44. Cattaneo A, Cattane N, Galluzzi S, Provasi S, Lopizzo N, Festari C, Ferrari C,
18. Rubenstein LZ, Harker JO, Salva A, Guigoz Y, Vellas B. Screening for Guerra UP, Paghera B, Muscio C, et al. Association of brain amyloidosis with
undernutrition in geriatric practice: developing the short-form mini-nutritional pro-inflammatory gut bacterial taxa and peripheral inflammation markers in
assessment (MNA-SF). J Gerontol A Biol Sci Med Sci. 2001;56(6):M366–72. cognitively impaired elderly. Neurobiol Aging. 2017;49:60–8.
19. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state". A practical 45. Meehan CJ, Beiko RG. A phylogenomic view of ecological specialization in
method for grading the cognitive state of patients for the clinician. J the Lachnospiraceae, a family of digestive tract-associated bacteria. Genome
Psychiatr Res. 1975;12(3):189–98. Biol Evol. 2014;6(3):703–13.
Stadlbauer et al. BMC Geriatrics (2020) 20:248 Page 13 of 13

46. Ferreira-Halder CV, Faria AVS, Andrade SS. Action and function of 67. Poullis A, Foster R, Mendall MA, Shreeve D, Wiener K. Proton pump
Faecalibacterium prausnitzii in health and disease. Best Pract Res Clin inhibitors are associated with elevation of faecal calprotectin and may affect
Gastroenterol. 2017;31(6):643–8. specificity. Eur J Gastroenterol Hepatol. 2003;15(5):573–4 author reply 574.
47. Li J, Wu T, Li N, Wang X, Chen G, Lyu X. Bilberry anthocyanin extract 68. Hage FG. C-reactive protein and hypertension. J Hum Hypertens. 2014;28(7):
promotes intestinal barrier function and inhibits digestive enzyme activity 410–5.
by regulating the gut microbiota in aging rats. Food Funct. 2019;10(1):333–43. 69. Volkert D, Chourdakis M, Faxen-Irving G, Fruhwald T, Landi F, Suominen MH,
48. Leblhuber F, Steiner K, Schuetz B, Fuchs D, Gostner JM. Probiotic Vandewoude M, Wirth R, Schneider SM. ESPEN guidelines on nutrition in
supplementation in patients with Alzheimer's dementia - an explorative dementia. Clin Nutr. 2015;34(6):1052–73.
intervention study. Curr Alzheimer Res. 2018;15(12):1106–13. 70. Million M, Diallo A, Raoult D. Gut microbiota and malnutrition. Microb
49. Finegold SM, Song Y, Liu C, Hecht DW, Summanen P, Kononen E, Allen SD. Pathog. 2017;106:127–38.
Clostridium clostridioforme: a mixture of three clinically important species.
Eur J Clin Microbiol Infect Dis. 2005;24(5):319–24.
50. Matijasic BB, Obermajer T, Lipoglavsek L, Grabnar I, Avgustin G, Rogelj I.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
Association of dietary type with fecal microbiota in vegetarians and
published maps and institutional affiliations.
omnivores in Slovenia. Eur J Nutr. 2014;53(4):1051–64.
51. Zhu Q, Huang L, Zhu J, Lu J, Qin H. Analysis of gut microbiota in long-lived
older adults and their relatives: a gradual change with ageing. Mech Ageing
Dev. 2018;178:1–8.
52. Metchnikoff E. The prolongation of life; optimistic studies. New York: G. P.
Putnam´s Son; 1908. https://openlibrary.org/books/OL14036242M/The_
prolongation_of_life.
53. Cheng LH, Liu YW, Wu CC, Wang S, Tsai YC. Psychobiotics in mental health,
neurodegenerative and neurodevelopmental disorders. J Food Drug Anal.
2019;27(3):632–48.
54. Singh P, Chung HJ, Lee IA, D'Souza R, Kim HJ, Hong ST. Elucidation of the
anti-hyperammonemic mechanism of Lactobacillus amylovorus JBD401 by
comparative genomic analysis. BMC Genomics. 2018;19(1):292.
55. Leblhuber F, Geisler S, Steiner K, Fuchs D, Schutz B. Elevated fecal
calprotectin in patients with Alzheimer's dementia indicates leaky gut. J
Neural Transm. 2015;122(9):1319–22.
56. Leber B, Tripolt NJ, Blattl D, Eder M, Wascher TC, Pieber TR, Stauber R, Sourij
H, Oettl K, Stadlbauer V. The influence of probiotic supplementation on gut
permeability in patients with metabolic syndrome: an open label,
randomized pilot study. Eur J Clin Nutr. 2012;66(10):1110–5.
57. Linsalata M, Riezzo G, D'Attoma B, Clemente C, Orlando A, Russo F.
Noninvasive biomarkers of gut barrier function identify two subtypes of
patients suffering from diarrhoea predominant-IBS: a case-control study.
BMC Gastroenterol. 2018;18(1):167.
58. Rainer F, Horvath A, Sandahl TD, Leber B, Schmerboeck B, Blesl A, Groselj-
Strele A, Stauber RE, Fickert P, Stiegler P, et al. Soluble CD163 and soluble
mannose receptor predict survival and decompensation in patients with
liver cirrhosis, and correlate with gut permeability and bacterial
translocation. Aliment Pharmacol Ther. 2018;47(5):657–64.
59. Ruan P, Gong ZJ, Zhang QR. Changes of plasma D(−)-lactate, diamine
oxidase and endotoxin in patients with liver cirrhosis. Hepatobiliary Pancreat
Dis Int. 2004;3(1):58–61.
60. Li H, Qiu P, Wang J, Niu C, Pan S. Effects of compound Ginkgo biloba on
intestinal permeability in rats with alcohol-induced liver injury. Food Funct.
2015;6(2):470–8.
61. Maher RL, Hanlon J, Hajjar ER. Clinical consequences of polypharmacy in
elderly. Expert Opin Drug Saf. 2014;13(1):57–65.
62. Patterson SM, Hughes C, Kerse N, Cardwell CR, Bradley MC. Interventions to
improve the appropriate use of polypharmacy for older people. Cochrane
Database Syst Rev. 2012;5:CD008165.
63. Horvath A, Rainer F, Bashir M, Leber B, Schmerboeck B, Klymiuk I, Groselj-
Strele A, Durdevic M, Freedberg DE, Abrams JA, et al. Biomarkers for
oralization during long-term proton pump inhibitor therapy predict survival
in cirrhosis. Sci Rep. 2019;9(1):12000.
64. Stadlbauer V, Komarova I, Klymiuk I, Durdevic M, Reisinger A, Blesl A, Rainer
F, Horvath A. Disease severity and proton pump inhibitor use impact
strongest on faecal microbiome composition in liver cirrhosis. Liver Int.
2020;40(4):866–77. https://doi.org/10.1111/liv.14382. Epub 2020 Jan 24.
PMID: 31943691; PMCID: PMC7187411.
65. Cohen M. Proton pump inhibitors may cause elevation in faecal calprotectin
levels. Br J Gen Pract. 2016;66(648):350.
66. Lundgren D, Eklof V, Palmqvist R, Hultdin J, Karling P. Proton pump inhibitor
use is associated with elevated faecal calprotectin levels. A cross-sectional
study on subjects referred for colonoscopy. Scand J Gastroenterol. 2019;
54(2):152-7. https://doi.org/10.1080/00365521.2019.1566493. Epub 2019 Jan
24. PMID: 30676120.