Content uploaded by Samantha Coulson
Author content
All content in this area was uploaded by Samantha Coulson
Content may be subject to copyright.
RESEARCH ARTICLE
Green-lipped mussel extract (Perna canaliculus) and glucosamine
sulphate in patients with knee osteoarthritis: therapeutic efficacy
and effects on gastrointestinal microbiota profiles
Samantha Coulson •Henry Butt •Phillip Vecchio •
Helen Gramotnev •Luis Vitetta
Received: 30 May 2012 / Accepted: 21 June 2012
ÓSpringer Basel AG 2012
Abstract
Objective To investigate how changes in the gastroin-
testinal tract (GIT) microbiota profile may influence
nutraceutical efficacy in osteoarthritis (OA) and allow the
formulation of a hypothesis that explains in part the
inconsistent and contentious findings from OA clinical
studies with green-lipped mussel (GLM) and glucosamine.
Methods A non-blinded randomised clinical trial was
conducted with 38 subjects diagnosed with knee OA. Each
participant received either 3,000 mg/day of a whole GLM
extract or 3,000 mg/day of glucosamine sulphate (GS), p.o.
for 12 weeks. Faecal microbial analyses were carried out after
collecting stools at T
0
and T
12
weeks. Additional pharmaco-
metric measures were obtained from changes in arthritic
scores in the Western OntarioMcMaster UniversitiesArthritis
Index (WOMAC) and the Lequesne algofunctional indices
and the Gastrointestinal Symptom Rating Scale (GSRS). An
intention-to-treat analysis was employed and participant data
collected at T
0
,T
6
and T
12
weeks.
Results There were no statistically significant changes in
bacterial growth patterns determined by the Wilcoxon test.
In both groups there was a trend towards a decrease in
Clostridium and Staphylococcus species and increase in
Lactobacillus, Streptococcus and Eubacterium species. In
the GLM group Bifidobacterium tended to increase and
Enterococcus and yeast species to decrease. The GS-trea-
ted group demonstrated a trend towards a decrease in
Bacteroides and an increase in yeasts and Coliforms spe-
cies, most notably Escherichia coli. We further confirm
significant improvement (p\0.05) in all OA outcome
measures from T
0
to T
12
weeks for both the GLM and GS
groups. The GSRS scores indicated that GIT function
significantly improved over the 12 weeks duration with
GLM and GS supplementation.
Conclusion Both GLM and GS reduced OA symptoms
and non-significantly altered the gut microbiota profile
from baseline. Changes in the microbiota profiles occurred
in both treatment groups; the most notable being a reduc-
tion in the Clostridia sp. This study suggests that
nutritional supplements such as GLM and GS may regulate
some of the metabolic and immunological activities of the
GIT microbiota. The decrease in Clostridia, a potent
modulator of colonic Th17 and CD4?regulatory T cells,
was consistent with a decrease in inflammation; improved
GSRS scores and OA symptoms for these OA participants.
The GIT microbiota may be important factor in the first-
pass metabolism of these nutraceuticals.
Keywords Gastrointestinal tract Microbiota
Green-lipped mussel Perna canaliculus Glucosamine
Osteoarthritis Clostridia sp.
Introduction
Natural medicines are an attractive option for patients
suffering from osteoarthritis (OA), (Ehrlich 2003) the most
S. Coulson H. Gramotnev L. Vitetta (&)
The Centre for Integrative Clinical and Molecular Medicine,
The University of Queensland, School of Medicine, Princess
Alexandra Hospital, Lvl 2, R Wing, 199 Ipswich Road,
Woolloongabba, QLD 4102, Australia
e-mail: l.vitetta@uq.edu.au
H. Butt
Bioscreen, Bio21, Molecular Science and Biotechnology
Institute, University of Melbourne, Melbourne, Australia
P. Vecchio
Department of Rheumatology, Princess Alexandra Hospital,
Brisbane, QLD, Australia
Inflammopharmacol
DOI 10.1007/s10787-012-0146-4 Inflammopharmacology
123
common, debilitating musculoskeletal disorder known
(Woolf and Pfleger 2003; Reginster 2002). The high rate of
self-medication with natural products is due to (1) lack of
an available cure (Ernst 1998) and (2) the serious adverse
events associated with chronic use of prescribed drugs,
particularly non-steroidal anti-inflammatory drugs (NSA-
IDs) (Bjordal et al. 2004) and high dose paracetamol
(acetaminophen) (Garcia Rodriguez and Hernandez-Diaz
2001). However, the clinical efficacy of nutraceuticals for
treating OA symptoms, such as joint pain, stiffness and
limited range of motion remains collectively inconsistent,
limiting recommendations for their routine use. This
includes (S-adenosylmethionine (Rutjes et al. 2008), green-
lipped mussel (GLM) extract (Ulbricht et al. 2009) and
D-glucosamine (Wandel et al. 2010).
The GIT and microbiota combined comprise one of the
most metabolically and immunologically active organs
(Egert et al. 2006). Nutraceutical supplements such as
GLM and GS orally administered are metabolised by
gastrointestinal bacteria, potentially modifying and influ-
encing exposure to these compounds (Laparra and Sanz
2010; Egert et al. 2006). The metabolism of GLM extract
by GIT microbial species has not previously been investi-
gated. Early in vitro studies have confirmed that
commensal GIT bacteria can ferment and metabolise glu-
cosamine (Foley et al. 2008; Koser et al. 1961; Wolfe et al.
1956). Furthermore, microbial diversity and colonisation
are restricted in the proximal GIT (*10
2
to 10
4
cfu ml
-1
)
and increases in density and species diversity within the
distal GIT (10
6
to 10
8
cfu ml
-1
) (Holzapfel et al. 1998).
However, the large bowel is the most highly populated and
most metabolically active region of the GIT (10
10
to
10
12
cfu g
-1
) with a truly complex and diverse bacterial
load estimated at more than 1,000 species with 30–40
normally predominating (Kelly et al. 2005; Guarner and
Malagelada 2003).
Anaerobes outnumber aerobic bacteria by a factor of l0
2
to 10
4
within the large bowel. The genera Bacteroides,
Bifidobacterium, Eubacterium, Clostridium, Peptococcus,
Peptostreptococcus, and Ruminococcus are predominant
genera whereas aerobes (facultative anaerobes) such as
Escherichia, Lactobacillus, Enterobacter, Enterococcus,
Klebsiella, and Proteus are among the subdominant genera
(Guarner and Malagelada 2003). Bacterial counts of indi-
vidual species range over several orders of magnitude and
the metabolic products of bacterial groups vary consider-
ably (Egert et al. 2006). The intestinal microbiota can have
marked effects on mucosal defence mechanisms (i.e.
competition for mucosal colonisation and metabolic sub-
strates plus synthesis of regulatory factors such as short-
chain fatty acids and bacteriocins) and the innate and
adaptive immune responses of the host, and are therefore
integral to maintaining immune homeostasis within the
developing and adult gut. Commensal bacteria, however,
differ in their ability both to promote development of the
gut-associated lymphoid tissues and to maintain its func-
tion (Kelly et al. 2005; Umesaki et al. 1999).
OA and rheumatoid arthritis (RA) can predispose
patients to GIT symptoms that can be attributed to orally
administered analgesics (i.e. paracetamol/acetaminophen
and NSAIDs) (Chong and Wang 2008; Wolfe et al. 2000).
GIT symptoms can include dyspepsia, epigastric pain and
burning, bloating, early satiety, nausea and belching,
mucosal ulceration and altered bowel habits (constipation/
diarrhoea). In addition to analgesic and anti-inflammatory
medicines inducing such GIT symptoms, altered GIT
microbial growth trends may also contribute (Tana et al.
2010). We have recently reported that administration of
GLM extract (3,000 mg/day) to OA patients significantly
improved GIT symptoms, concurrent with significant
improvement in knee pain, stiffness and function (Coulson
et al. 2012). Profiling the GIT microbiota has not been
previously studied in OA patients. Altered GIT microbiota
profiles may be a significant factor that largely determines
the therapeutic efficacy of nutraceutical supplements such
as GLM and GS, in OA patients.
The GIT microbiota of OA patients may therefore be a
target of nutritional intervention that influences bacterial
viability, growth, and metabolic activity. Therefore, the
aim of this clinical study was to assess faecal microbiota
profiles in patients diagnosed with OA of the knee and then
to report whether supplementation with GLM extract and
GS affected the growth patterns of the faecal microbiota.
Furthermore, we investigated whether bacterial growth
patterns could be correlated to the therapeutic efficacy
outcomes with GLM and GS.
Patients and methods
Study design
The trial was conducted according to a non-blinded, ran-
domised, comparator-controlled, parallel-group design in
one outpatient rheumatology clinic in Brisbane, QLD. The
study protocol complied with the Helsinki Declaration and
was approved by the Human Research Ethics Committees
of the University of Queensland and The Princess Alex-
andra Hospital. Participants who met the inclusion criteria
were then randomly assigned to one of the two treatment
arms, i.e. GLM or GS. Each patient received written and
verbal explanations regarding their involvement in the
study before signing informed consent. They were exam-
ined before receiving their first treatment at baseline (T
0
)
and then re-examined at T
6
and T
12
weeks of treatment at
which time their participation was concluded.
S. Coulson et al.
123
Participant selection
The study group comprised of 40 participants diagnosed
with OA in one or both knees (11 male, 29 female) who
satisfied the inclusion and exclusion criteria. Patients were
recruited if their knee OA was reviewed and confirmed by
a Rheumatologist (PV) and they were not taking antibiotics
or any form of herbal/multivitamin/nutritional supplements
4 weeks before entering this study. Participants were
excluded if they had uncontrolled systemic disease, were
pregnant or breast feeding or had allergies/intolerances to
shellfish. They were also instructed to avoid antibiotics
while enrolled in the study. All participants satisfied the
decision tree format of the American College of Rheuma-
tology (ACR) classification criteria for idiopathic clinical
OA of the knee (Altman et al. 1986). The ACR clinical
criteria for knee OA were fulfilled by the uniform presence
of knee pain and at least three of the following six criteria:
C50 years of age, morning stiffness lasting B30 min,
crepitus on movement, bony tenderness, bony enlargement
and no palpable warmth of synovium, or at least one of the
following three criteria plus radiographic evidence of
osteophytes: C50 years of age, morning stiffness lasting
B30 min and crepitus on movement.
The subject’s demographic data and medical history
were obtained at baseline. Anthropometric measurements,
blood pressure and outcome measures were performed at
T
0
,T
6
and T
12
weeks. Height was measured with the sub-
ject standing barefoot using a body meter measuring tape
with wall stop. Body mass was measured using calibrated
scales and body mass index (BMI) was calculated using the
formula: mass divided by height squared (kg/m
2
); waist hip
ratios were calculated by dividing the waist circumference
(cm) by the hip circumference (cm). Participants were
recruited from the local Brisbane area from June to
November 2011 and were seen for each visit at the Princess
Alexandra Hospital, Brisbane QLD. Patients were ran-
domised for treatment with either the GLM (5 male, 16
female) or GS (6 male, 13 female) groups.
Intervention
The GLM extract sourced for this trial was GlycOmega
TM
PLUS from Aroma NZ Ltd, Christchurch NZ, a propriety
blend of freeze-dried mussel meat (no shell), stabilised
with Rosemary oil extract. The glucosamine sulphate-
potassium chloride was sourced from BJP Laboratories
(Brisbane, QLD). Patients were randomly allocated (non-
blinded) to 3,000 mg/day (3 9500 mg capsules b.i.d) of
GlycOmega
TM
PLUS or 3,000 mg/day (3 9500 mg cap-
sules b.i.d) of GS. Capsules were dispensed in opaque
white bottles for the 12 weeks duration. Compliance was
checked by examining the patients’ records of their daily
intake of GLM or GS in provided diaries. Participants in
this trial were allowed to continue taking analgesic medi-
cations (paracetamol and/or NSAIDs) as required.
Participants were asked to complete two diaries that
recorded daily administration of (1) GLM or GS; (2)
analgesic medication and (3) knee pain experienced on a
5-point Likert scale (T
0
–T
6
and T
6
–T
12
).
Randomisation
A randomisation list was generated by computer program
and maintained by the investigators. Participants were
sequentially assigned to a randomisation schedule number
(1 =GLM or 2 =GS) at their first visit to the trial clinic
post-adherence to all inclusion and exclusion criteria.
Objectives and outcomes
The primary outcome measure was to assess at T
0
and T
12
weeks, the common GIT microbial genera.
Faecal sample collection
A morning bowel motion was collected in a faecal con-
tainer by patients in their own home in a modified faecal
container with the lid of the container perforated to assist
sample anaerobiasis. The sample was immediately trans-
ferred into a sealed anaerobic pouch system (Oxoid,
Thermo Fisher Scientific, Australia). Anaerobiasis was
achieved by activating the Anaero Gen Compact (Oxoid,
Thermo Fisher Scientific, Australia) before sealing the
pouch. Samples were transported cold (\12 °C) to the
laboratory and analysed within 48 h after collection.
Samples were rejected for analysis if they were delayed
during transit, had inadequate refrigeration or if the
anaerobic storage during transit did not meet the criteria
and remained aerobic. Patients with rejected samples were
requested to resubmit another sample for analysis.
Quantitation of faecal microorganisms
All faecal samples, once removed from the anaerobic
pouch system were processed within 10–15 min. A deter-
mined quantity of faecal sample (range 0.5–1.0 g) was
transferred to 10 mL of 1 % glucose-saline buffer (Willis
2007). Dilution factor was determined by the difference in
the weight of the glucose-saline buffer with and without the
sample. One hundred and one thousand fold dilutions
(beginning from 10
-1
to 10
-7
) of homogenised faecal
samples were prepared (Thrupp 1980). A 10 and/or 1 lL
amount of the appropriate dilutions were transferred each
onto previously dried Columbia horse blood agar (Oxoid),
chromogenic medium (Oxoid), colistin and nalidixic acid
Gastrointestinal microbiota profiles in patients with OA
123
blood selective agar (Oxoid), and chloramphenicol-genta-
micin selective Sabouraud agar for aerobic incubation. Pre-
reduced Columbia horse blood haemin agar and Raka Ray
medium were used for anaerobic incubation. All aerobic
media were incubated at 35 °C for 48 h and anaerobic media
in anaerobic jars (Oxoid) for 4 days. All aerobicand anaerobic
culture plates were examined under a stereomicroscope for a
minimum of 20 min/plate before identification of the bacteria.
Every colony from each medium was examined microscopi-
cally and quantitatively recorded. Microscopic colonies of
similar morphotypes were sub-cultured onto horse blood agar
to check for purity prior to identification. To determine the
validity of the anaerobic transport and culture methods, two
faecal samples from one of the investigators (HB), collected at
different days, were processed each within 2 h after collec-
tion, and again on two different occasions, at 24 and 48 h. On
each occasion the specimen was re-sealed and stored refrig-
erated anaerobically immediately after being processed.
Results from this internal quality assurance investigation
showed that there were no significant quantitative changes in
either the aerobes or anaerobes processed atthe three different
periods. The incidence of the predominant aerobes and
anaerobes remained unchanged.
Identification of faecal bacteria: MALDI-TOF MS
analysis
Index bacterial colonies, from overnight purity checks,
were transferred to a target polished steel plate (MSP 96
target polished steel, Bruker Daltonics Inc.) for drying
under exhaust ventilation in a Class II Biohazard Hood
(Gelman Sciences Australia) at room temperature. Air-
dried samples were subjected to protein extraction with
1lL 70 % formic acid (Sigma). Samples were again
allowed to air dry, under exhaust ventilation, before being
overlaid with 1 lL of matrix solution (saturated solution of
a-cyano-4-hydroxycinnamic acid (HCCA) in a mixture of
47.5 % ultra-pure water, 2.5 % trifluoroacetic acid, and
50 % acetonitrile). Once dried, samples were subjected to
analysis using a Microflex MALDI-TOF mass spectrometer
(Bruker Daltonik GmbH, Leipzig, Germany) equipped with a
60 Hz nitrogen laser. Spectra were recorded in the positive
linear mode for the mass range of 2,000–20,000 Da at maxi-
mum laser frequency. Raw spectra were analysed, using the
MALDI Biotyper 3.0 software package (Bruker Daltonik
GmbH, Bremen, Germany) with default settings. Measure-
ments were performed automatically without any user
intervention.
Secondary outcome measures
Secondary outcome measures were recorded at T
0
,T
6
and
T
12
weeks and included the Western Ontario McMaster
Universities Arthritis Index (WOMAC) (Bellamy 2002) the
Lequesne algofunctional index (Lequesne et al. 1987) the
Gastrointestinal Symptom Rating Scale (GSRS) question-
naire (Svedlund et al. 1988) and the SF-12v2
TM
health
survey (Tucker et al. 2010). The WOMAC is a validated
questionnaire containing 24 questions, designed to assess
lower extremity pain and function in OA of the knee or hip
by assessing severity of pain, stiffness and limitation of
physical function with a maximum score of 96 (Bellamy
et al. 1988). The Lequesne algofunctional index includes
10 questions for the measurement of pain, walking distance
and activities of daily living (Lequesne et al. 1987). Scores
for each question are added together to provide a combined
disease severity score with a maximum score of 24. Scores
of 1–4 are classified as mild OA, 5–7 moderate, 8–10
severe, 11–13 very severe, and greater than 14 as extremely
severe OA. GIT function was assessed using the validated
GSRS questionnaire which contains 15 items to evaluate
abdominal pain, gastro-oesophageal reflux, indigestion,
diarrhoea and constipation scored on a 7-point Likert scale.
The GSRS has a maximum score of 105. General quality of
life was assessed using the SF-12 questionnaire that is
expressed in terms of two meta-scores: the physical com-
ponent summary (PCS) and the mental component
summary (MCS) (Tucker et al. 2010; Ware et al. 1996).
Assessment of safety
Any adverse events were documented by patients in their
diaries kept over the 12 weeks duration of the trial. These
events were recorded by giving the start and finish dates of
symptoms and their severity (from 1, mild; 2, moderate and
3, severe). Medications required to treat the adverse events
were also recorded in diaries provided, listing the medi-
cation’s name, dose taken per day and the start and finish
dates of taking the medications. Routine laboratory mea-
surements including full blood count, serum electrolytes,
liver function tests, blood glucose (non-fasting) and
inflammatory markers, C-reactive protein (CRP) and
erythrocyte sedimentation rate (ESR), were performed at
time of enrolment and after 12 weeks of treatment.
Statistical analysis
All microbiology data were log10 transformed prior to
statistical analysis. Normality tests were performed on
bacterial data. Wilcoxon test for paired data was used to
analyse the changes in bacteria within treatment groups.
Kolmogorov–Smirnov tests were performed for all the
scored variables to confirm the assumption of normality,
justifying the use of ttests. Paired ttests were then used to
compare the changes in the scores over the study period
(T
0
–T
12
weeks) within the GLM and GS groups.
S. Coulson et al.
123
Independent group ttests were used to test for differences
in changes in the scores between the GLM and GS groups.
The scoring analysis was by assessment of the difference
between groups and within groups in the change from
baseline to 12 weeks, in the intention to treat (ITT) pop-
ulation (i.e. including all randomised patients with at least
1 efficacy assessment after randomisation). The signifi-
cance level was set at p\0.05. Analgesic usage was
assessed in patients who completed, or partially completed,
the 12 weeks treatment course. The mean number of medi-
cations per 7-day period was tabulated for descriptive
purposes.A similar graphwas produced for the amount of pain
reported, also as a 7-day average over 12 weeks of the study.
Results
The demographic and clinical characteristics of the ran-
domised patients were comparable at baseline (Table 1).
The majority of patients were female, being 74 % of the
cohort. The mean age was 58.6 ±8.9 years and 89 % of
patients were overweight (BMI[25 kg/m
2
) and 58 % were
obese (BMI [30 kg/m
2
). The mean duration of knee OA
was 9.4 ±8.3 years. For the overall cohort mean WO-
MAC score at baseline was 36.3 and the Lequesne index
score 11.7 indicating chronic disease with moderate
symptom severity (Table 1). Four patients withdrew from
the trial with 2 patients providing their first diary and
completing the questionnaires at T
6
weeks yielding a total
of 38 participants (10 male, 28 female). Twenty-one
patients in the GLM group and 17 in the GS group pro-
vided data for an analysis of Lequesne, WOMAC, GSRS
and SF-12 scores (Fig. 1). Patient numbers for bacterial
analysis differed slightly, the GLM group providing 21
faecal samples for analysis at baseline but only 18 samples
at week 12 (due to samples being lost in transit to the
analytical laboratory). The GS group provided 17 faecal
samples at both baseline and week 12.
Weight, BMI and waist-to-hip ratios did not change sig-
nificantly within each group between time points. Current and
stable medications were continued throughout the trial. The
type of medications and number of patients taking them
included cholesterol lowering agents (n=10), anti-diabetic
agents (n=4), diuretic (n=1), anti-hypertensive medica-
tions (n=8), anti-inflammatory medications (n=3), proton
pump inhibitors (n=11), anti-depressant medications
(n=7), thyroid medications (n=3), anti-histamine medi-
cation (n=5), anti-coagulant medications (n=9),
benzodiazepines (n=2), medications for migraine (n=2),
gout medication (n=1). We also importantly note that these
medications prescribed to participants by their clinicians for
management of chronic diseases had been administered at
length prior to participating in the trial and remained as such
during the trial. Therefore, we can safely deduce that baseline
microbiota readings should not have changed over the
12 weeks due to these prescribed medications. This further
substantiating that any changes that occurred at 12 weeks
were not influenced by these pharmaceuticals.
Prescribed antibiotics, but not within 4 weeks preceding
trial commencement, was reported by 9 patients while 26
patients reported recent influenza vaccinations. The rate of
compliance with GLM or GS intake as per the protocol (3
caps b.i.d for 12 weeks) was 34 % of the participants not
missing any doses, while the remaining 66 % missed 1 or
more doses over the 12-week study period.
Four patients from the original 40 did not complete the
study namely, 1 male from the GLM group and 2 males
and 1 female from the GS group. Drop-outs were due to
either lack of interest or an adverse event unrelated to the
GLM or GS. Two of the male dropouts (1 GLM, 1 GS) did
return their first diaries and completed week 6 question-
naires. The data was included for analysis as intention-to-
treat with one also providing a faecal sample for bacterial
analysis at week 12.
Faecal analysis
Of the 73 faecal samples analysed, 12 genera and 137
species were recovered (data not shown). The mean viable
count [colony forming units (cfu)/g] of the total aerobic
and anaerobic microbial flora at baseline was not signifi-
cantly different between the two treatment groups. The
Table 1 Demographic and baseline clinical characteristics of
patients in the intention-to-treat population
Characteristics Intervention
Green-lipped
mussel group
(n=21)
Glucosamine
sulphate group
(n=17)
Women, no (%) 16 (76) 12 (71)
Age (years) 56.7 ±8.9 60 ±8.6
BMI (kg/m
2
) 31.3 ±6.1 30.2 ±4.8
Weight (kg) 90.1 ±17.1 87.1 ±15.4
Waist:hip ratio 0.9 ±0.1 0.9 ±0.1
Duration of knee OA (years) 7.5 ±5.9 11.4 ±9.5
Lequesne index 12.2 ±3.4 11.1 ±2.6
WOMAC
Total 37.4 ±13.8 34.9 ±10.8
Pain 8.0 ±2.8 7.5 ±2.7
Stiffness 4.2 ±1.5 3.6 ±1.3
Function 25.2 ±10.2 23.8 ±7.8
GSRS 34.4 ±19.5 24.3 ±9.9
Except where indicated otherwise, values are represented as
mean ±SD
Gastrointestinal microbiota profiles in patients with OA
123
mean viable counts and pvalues (determined by the Wil-
coxon test) of aerobic and anaerobic species are presented
in Table 2. A number of GIT bacterial species in these OA
participants at baseline demonstrated variation from the
GIT species in the control data (n=177) with no gastro-
intestinal symptoms (data on file, Bioscreen). The viable
count of microorganisms that were increased from the
control data (at baseline) included Enterococcus, Strepto-
coccus, Staphylococcus, Eubacterium, Lactobacillus,
Bifidobacterium and Clostridium. The bacterial counts of
these microorganisms remained high, at week 12 when
compared to the control data. However, Clostridial and
Staphylococcus species were observed to decrease from
baseline to 12 weeks in both treatment groups.
Within-group faecal microbial analysis did not demon-
strate statistically significant changes within any of the
bacterial species from T
0
to T
12
weeks after supplementa-
tion with GLM or GS. Further analyses of the data did,
however, demonstrate an effect on the growth trends (i.e.
decreased or increased) among the microbiota profiles that
differed between the two treatment groups. Overall, the
GLM-treated group demonstrated a trend towards
decreasing Enterococcus sp., Staphylococcus sp. Clostrid-
ium sp. and yeasts and an increased Lactobacillus sp.,
Bifidobacterium sp., Eubacterium sp., Bacteroides sp. and
Streptococcus sp. By comparison the GS-treated group
demonstrated a trend towards decreasing Staphylococcus
sp., Bacteroides sp. and Clostridium sp. and increased
Coliforms, Streptococcus sp., Eubacterium sp. and Lacto-
bacillus sp. The most prevalent species detected within
each genera were as follows: Coliforms (Escherichia coli,
Klebsiella pneumoniae); Enterococcus (E. faecalis, E. fae-
cium); Streptococcus (S. mutans, S. parasanguinis,
S. salivarius); Staphylococcus (S. aureus, S. epidermidis,
S. capitis); Yeasts (Candida albicans); Bacteroides
(B. vulgates, B. ovatus, B. fragilis, B. uniformis, B. theta-
iotaomicron, B. stercoris); Eubacterium (Collinsella
aerofaciens); Lactobacillus (L. acidophilus, L. gasseri,
L. paracasei, L. fermentum); Bifidobacterium (B. longum,
B. animalis, B. adolescentis, B. bifidum); Clostridium
(C. innocuum, C. tertium) with C. perfringens detected in
one patient at baseline.
Paired sample ttests were performed on the Lequesne,
WOMAC (total and subscores), GSRS, and two component
SF-12 scores (PCS and MCS) between T
0
and T
6
weeks and
T
0
and T
12
weeks for each treatment group. The results are
reported in Table 3. Both the GLM and GS-treatment
groups demonstrated statistical significance for the
Lequesne and WOMAC total and subscores between
all intervals, p\0.001 and p=0.001, respectively. The
GLM group also demonstrated statistical significance for
GSRS scores (p=0.02) and borderline significance for the
GS group (p=0.044). The incidence of GSRS symptoms
is reported in Table 4. The PCS scores for both the GLM
(p=0.004) and GS (p=0.001) groups improved signifi-
cantly from baseline to week 12, whereas the MCS scores
showed no statistical improvement for both the GLM
(p=0.20) or GS (p=0.70) groups.
Analgesic medication was allowed and recorded daily
by the participants over the 12-week trial period. The need
for analgesic medications decreased in the GLM group
with a mean consumption at baseline of 1.07 ±1.62
Fig. 1 CONSORT diagram of participant inductions
S. Coulson et al.
123
Table 2 Mean viable counts (cfu/g) and Wilcoxon signed ranks test (pvalue) of faecal aerobes and anaerobes in the green-lipped mussel and
glucosamine sulphate groups at baseline and week 12
Organism Normal range
a
Green-lipped mussel group Glucosamine sulphate group
Baseline Week 12 pvalue Baseline Week 12 pvalue
(n=21) (n=18) (n=17) (n=17)
Total bacterial count 1 910
9
to 1 910
12
2.31 910
10
3.29 910
10
0.36 4.04 910
10
3.10 910
10
0.72
Total aerobes 1 910
7
to 1 910
8
7.17 910
7
8.83 910
7
0.71 6.98 910
7
1.02 910
8
0.35
Coliforms 7910
6
to 9 910
7
4.38 910
7
5.82 910
7
0.68 6.01 910
7
7.52 910
7
0.77
Enterococcus \5910
5
9.32 910
6
5.75 910
6
0.48 4.86 910
6
8.41 910
6
0.64
Streptococcus \3910
5
1.12 910
7
2.84 910
7
0.35 1.76 910
7
2.98 910
7
0.83
Staphylococcus \2910
5
4.51 910
6
1.97 910
4
1.00 7.03 910
5
4.78 910
4
0.36
Yeast \1910
4
9.49 910
3
3.46 910
3
0.13 3.85 910
3
3.92 910
5
0.75
Total anaerobes 1 910
8
to 1 910
12
2.30 910
10
3.28 910
10
0.35 4.03 910
10
3.09 910
10
0.74
Bacteroides 9910
7
to 9.5 910
11
1.31 910
10
1.82 910
10
0.23 3.05 910
10
1.98 910
10
0.27
Prevotella \5910
8
ND ND 1.90 910
8
ND
Eubacterium \1910
9
6.36 910
9
1.24 910
10
0.39 5.02 910
9
7.60 910
9
0.25
Lactobacillus 5910
5
to 1 910
7
8.06 910
8
8.44 910
8
0.75 1.87 910
8
8.27 910
8
0.76
Bifidobacterium 5910
5
to 5 910
8
3.83 910
9
4.19 910
9
0.51 7.75 910
9
5.58 910
9
0.24
Clostridium \5910
8
1.91 910
9
8.11 910
8
0.70 1.84 910
9
9.95 910
8
0.94
ND not detected
a
Normal range determined by normal healthy population (Bioscreen data)
Table 3 Mean (95 %
confidence interval) change
from baseline in secondary
outcome measures. Mean (95%
confidence interval) change
from baseline in secondary
outcome measures and
differences between groups
pvalues and 95 % confidence
intervals of the differences
between groups were derived by
independent samples ttests
GSRS gastrointestinal symptom
rating score
*pvalues represent change in
score from T
0
–T
12
Outcomes Green-lipped
mussel (n=21)
Glucosamine
sulphate (n=17)
Difference
between groups
GSRS
Change T
0
–T
6
-10.2 (2.6, 17.9) -3.8 (-0.7, 8.3)
Change T
0
–T
12
-10.4 (1.8, 18.9) -3.9 (0.1, 7.8) -6.4 (-15.6, 2.7)
p* 0.02 0.044 0.162
Lequesne index
Change T
0
–T
6
-3.7 (2.1, 5.3) -2.9 (1.3, 4.6)
Change T
0
–T
12
-4.2 (2.2, 6.1) -3.3 (1.6, 4.9) -0.9 (-3.4, 1.7)
p\0.001 0.001 0.5
WOMAC (total)
Change T
0
–T
6
-16.4 (10.6, 22.3) -10.4 (3.1, 17.6)
Change T
0
–T
12
-18.9 (12.7, 25.1) -14.8 (7.3, 22.2) -4.1 (-13.4, 5.1)
p\0.001 0.001 0.372
WOMAC (pain)
Change T
0
–T
6
-4.0 (2.5, 5.4) -2.4 (0.2, 4.5)
Change T
0
–T
12
-4.9 (3.4, 6.3) -3.3 (1.5, 5.1) -1.6 (-3.7, 0.6)
p\0.001 0.001 0.157
WOMAC (stiffness)
Change T
0
–T
6
-1.5 (0.8, 2.3) -1.1 (0.5, 1.7)
Change T
0
–T
12
-2.4 (1.6, 3.1) -1.2 (0.7, 1.8) -1.1 (-2.0, 0.2)
p\0.001 \0.001 0.02
WOMAC (physical)
Change T
0
–T
6
-11.0 (6.6, 15.3) -6.9 (1.8, 12.0)
Change T
0
–T
12
-11.8 (7.2, 16.3) -10.2 (4.7, 15.6) -1.6 (-8.3, 5.2)
p\0.001 0.001 0.639
Gastrointestinal microbiota profiles in patients with OA
123
tablets/capsules per day to 0.66 ±1.10 tablets/capsules at
week 12. The intake of analgesic medication in the GS
group, however, did not change with a mean consumption
at baseline of 0.48 ±0.80 tablets/capsules per day to
0.47 ±1.02 tablets/capsules per day at week 12 (Fig. 2).
Only 18 % of patients in the GS group did not use anal-
gesic medication at all, compared with 29 % in the GLM
group. Analgesic medication and doses per capsule inclu-
ded panadol (paracetamol 500 mg), panadol osteo
(paracetamol 665 mg), panadeine forte (paracetamol
500 mg ?codeine phosphate 30 mg), and various NSA-
IDs (aspirin 100 mg; indomethacin 25 mg; diclofenac
sodium 50 mg; ibuprofen 200 mg).
Adverse events
The number of adverse events reported during the trial
period was similar in both groups and is presented in
Table 5. The most frequent adverse events were upper
respiratory tract infections (URTI) for which seven patients
(6 GLM, 1 GS) took antibiotics over a period of 4–26 days
while enrolled within the study. Other events included
gastrointestinal disturbances for which 1 patient in the GS
group took probiotics for 21 days (pharmacist recommen-
dation), headaches, migraines, injuries, angina and hip
surgery. None of the adverse events were deemed related to
the GLM or GS preparations.
A subsequent sub-analysis of faecal microbiota profiles
at 12 weeks after correcting for the antibiotic (n=7) and
probiotic (n=1) use, showed that the remaining partici-
pants’ microbiota profile trends did not deviate from the
uncorrected trend in both groups (compare Tables 2and 6).
Specifically participants who were prescribed an antibiotic
(6 in GLM and 1 GS group) demonstrated at 12 weeks a
trend towards an increase in Enterococcus, Streptococcus,
Eubacterium and Lactobacillus species. Decreased growth
was noted for Coliforms,Yeasts,Bifidobacterium and Clos-
tridia species. For the sole participant who was prescribed
a single-strain probiotic (containing 3 Lactobacillus sp.)
Table 4 Gastrointestinal complaints: number of patients in both treatment groups at baseline and week 12 and range of severity according to the
GSRS 7-point Likert scale
Gastrointestinal symptom
a
Green-lipped mussel group Glucosamine sulphate group
Baseline Week 12 Baseline Week 12
Reflux 5 (moderate–severe) 2 (mild–moderate) 4 (mild–severe) 1 (severe)
Constipation 6 (mild–severe) 3 (mild–moderate) 2 (mild–moderate) 1 (mild)
Diarrhoea 5 (mild–very severe) 3 (mild–very severe) 2 (mild–moderate) 0
Heart burn 5 (moderate–severe) 2 (mild–moderate) 5 (mild–severe) 2 (moderate–severe)
Nausea 6 (mild–mod/sev) 1 (moderate) 2 (mild–mod/sev) 1 (mod/sev)
Bloating 8 (mild–very severe) 5 (mild–moderate) 7 (mild–severe) 3 (mild)
Abdominal pain 6 (mild–severe) 4 (mild–severe) 4 (mild–severe) 2 (mild–moderate)
Flatulence 9 (mild–very severe) 6 (mild–severe) 5 (mild–severe) 5 (mild–severe)
Burping 7 (mild–very severe) 4 (mild–moderate) 3 (mild–severe) 1 (mild)
Rumbling 6 (mild–mod/sev) 5 (mild–mod/sev) 5 (mild–mod/sev) 4 (mild–moderate)
mod/sev: moderate/severe symptoms, rumbling refers to stomach rumbling
a
Complaints reported in the Gastrointestinal Symptom Rating Scale (GSRS)
Fig. 2 a Average daily pain scores and baverage analgesic
medication administered for GLM and GS groups
S. Coulson et al.
123
showed at 12 weeks an increased trend in Coliforms, Eubac-
terium, Lactobacillus sp. and decreased growth of
Streptococcus and Clostridia sp. (Table 6).
Safety measures
Baseline blood analyses indicated that a number of liver
function enzymes were elevated. Gamma-glutamyl transfer-
ase (GGT) was increased in two patients (1 GLM, 1 GS),
mean ±SD, 54.3 ±9.8 U/L (reference range\38 U/L) and
alanine transaminase (ALT) was increased in three patients (2
GLM, 1 GS), mean ±SD, 54.3 ±13.7 U/L (reference range
\34 U/L) and aspartate transaminase (AST) was increased in
three patients (3 GLM), mean ±SD, 41.7 ±3.1 U/L (ref-
erence range\31 U/L). These patients, however, were taking
a number of medications for hyperlipidaemia, hypertension,
diabetes, gastric reflux and depression/anxiety that are known
to elevate liver enzymes (Gillett and Norrell 2011;Chitturi
and George 2002). At 12 weeks, GGT was elevated in four
participants (2 GLM, 2 GS); mean ±SD, 49.8 ±5.9 and
ALT was elevated in two participants (2GLM), mean ±SD,
40.5 ±0.7 U/L. AST levels returned to normal range in the
three patients by 12 weeks. Non-fasting blood glucose levels
were elevated in three patients (2 GLM, 1 GS) at the end of
treatment (mean ±SD, 9.3 ±1.9 mmol/L), however; all
three patients had type II diabetes mellitus prior to enrolment
and were receiving medication for their diabetes. Mean CRP
values differed slightly between the groups at baseline and
12 weeks with the GS group demonstrating a decreasein CRP
(mean ±SD, T
0
=4.69 ±4.76 mg/L; T
12
=2.78 ±1.28
mg/L) compared to GLM (mean ±SD, T
0
=4.62 ±
4.69 mg/L; T
12
=5.22 ±4.75 mg/L). The GS group also
demonstrated a decrease in ESR (mean ±SD, T
0
=19.06 ±
14.53 mm/h; T
12
=16.92 ±12.51 mm/h) compared to the
GLM group (mean ±SD, T
0
=14.65 ±7.15 mm/h;
T
12
=16.89 ±8.95 mm/h). There were no significant
changes in the other routine laboratory parameters. Blood
pressure remained stable throughout the treatment period in
both the GLM (mean ±SD, T
0
=systolic 131.0 ±13.51
mmHg and diastolic 79.06 ±9.97 mmHg; T
12
=systolic
132.69 ±16.1 mmHg and diastolic 79.46 ±11.77 mmHg)
and GS groups (mean ±SD, T
0
=systolic 129.0 ±11.69
mmHg and diastolic 75.38 ±11.75 mmHg; T
12
=systolic
128.74 ±10.57 mmHg and diastolic 76.0 ±8.05 mmHg).
Discussion
This study evaluated some of the commonly encountered
GIT microbial genera in patients with OA before and after
treatments with GLM or GS. Specifically, we have assessed
how GLM and GS, over a 12-week period, might alter
common bacterial growth patterns in the GIT. This clinical
study has provided an alternate hypothesis that could
explain the variable and contentious therapeutic effects
GLM and GS have elicited in clinical trials (Brien et al.
2008; Towheed et al. 2005). To this end, GIT dysbiosis
may have a significant role to play in nutraceutical efficacy
for OA. We observed a significant improvement in GSRS
scores, primarily, in the GLM-treated group, corroborating
previous observations (Coulson et al. 2012). This study
also confirmed that both GLM and GS each significantly
attenuated knee joint pain, stiffness and improved joint
flexibility and function.
Several methodological issues and industry bias have
been suggested as contributing factors that have led to
inconsistent clinical findings when evaluating GLM and
GS efficacy (Brien et al. 2008; Towheed et al. 2005). The
administration of analgesic medications to rescue pain, or
the exclusion of their use, may be an important factor that
Table 5 Summary of adverse events experienced by patients at any
time during the treatment period
Adverse event No. of patients experiencing
an adverse event
Green-lipped
mussel group
Glucosamine
sulphate group
Gastrointestinal symptoms
Reflux 1 1
Constipation 2 2
Diarrhoea 0 1
Heart burn 1 1
Nausea 3 0
Bloating 1 0
Abdominal pain 0 1
Infections
Respiratory tract infections 8 4
Gastroenteritis 2 1
Cutaneous infections 1 0
Respiratory Symptoms
Coughing and associated symptoms 3 2
Allergies 1 2
Neurological symptoms
Headache 4 5
Migraine 1 2
Insomnia 1 0
Musculoskeletal symptoms
Back/neck pain 0 4
Shoulder pain 3 1
Injuries
Fall 1 0
Other 2 2
Cardiovascular symptoms
Angina 1 0
Gastrointestinal microbiota profiles in patients with OA
123
Table 6 Subanalysis of mean viable counts (cfu/g) of faecal aerobes and anaerobes in the green-lipped mussel, glucosamine sulphate, antibiotic and probiotic groups at baseline and week 12
Organism Normal range
a
Green-lipped mussel group Glucosamine sulphate group Administered
Antibiotics Probiotics
Baseline
(n=15)
Week 12
(n=12)
Baseline
(n=15)
Week 12
(n=15)
Baseline
(n=7)
Week 12
(n=7)
Baseline
(n=1)
Week 12
(n=1)
Total bacterial count 1 910
9
to 1 910
12
2.54 910
10
3.93 910
10
4.34 910
10
3.27 910
10
1.77 910
10
1.98 910
10
1.52 910
10
1.75 910
10
Total aerobes 1 910
7
to 1 910
8
5.60 910
7
1.07 910
8
7.59 910
7
8.57 910
7
9.96 910
7
9.06 910
7
2.27 910
7
1.17 910
8
Coliforms 7910
6
to 9 910
7
1.93 910
7
6.70 910
7
6.89 910
7
8.03 910
7
9.02 910
7
3.47 910
7
1.32 910
7
7.80 910
7
Enterococcus \5910
5
1.80 910
7
6.58 910
6
6.18 910
6
6.27 910
5
2.25 910
6
3.96 910
6
4.04 910
5
3.91 910
7
Streptococcus \3910
5
1.09 910
7
3.65 910
7
1.73 910
7
1.04 910
7
1.84 910
7
6.25 910
7
9.08 910
6
ND
Staphylococcus \2910
5
4.51 910
6
1.97 910
4
7.03 910
5
4.78 910
4
ND ND ND ND
Yeast \1910
4
8.52 910
3
4.19 910
3
4.94 910
3
5.22 910
5
1.00 910
4
2.00 910
3
1.01 910
2
1.30 910
2
Total anaerobes 1 910
8
to 1 910
12
2.53 910
10
3.92 910
10
4.34 910
10
3.26 910
10
1.76 910
10
1.97 910
10
1.52 910
10
1.74 910
10
Bacteroides 9910
7
to
9.5 910
11
1.26 910
10
2.05 910
10
3.33 910
10
2.09 910
10
1.46 910
10
1.43 910
10
4.04 910
9
3.90 910
9
Prevotella \5910
8
ND ND 1.62 910
9
ND ND ND ND ND
Porphyromonas \5910
8
ND ND 1.80 910
9
ND ND ND ND ND
Eubacterium \1910
9
8.14 910
9
1.61 910
10
5.40 910
9
7.80 910
9
2.21 910
9
4.24 910
9
ND 5.20 910
9
Lactobacillus 5910
5
to 1 910
7
1.08 910
9
4.50 910
8
2.02 910
8
2.64 910
8
9.24 910
7
1.42 910
9
1.82 910
7
6.72 910
9
Bifidobacterium 5910
5
to 5 910
8
4.85 910
9
5.65 910
9
8.22 910
9
6.13 910
9
1.07 910
9
1.73 910
8
ND ND
Clostridium \5910
8
2.04 910
9
9.48 910
8
1.98 910
9
9.95 910
8
1.01 910
9
5.38 910
8
1.01 910
9
ND
ND not detected
a
Normal range determined by normal healthy population (Bioscreen data)
S. Coulson et al.
123
explains the null results reported for OA. Analgesic med-
ications administered by the cohort in this clinical study
showed a decreased trend with study progress. This was
observed to be greater for the GLM group than for the GS
group. We hence hypothesise that bacterial metabolism of
nutraceutical compounds in the GIT may generate metab-
olites that further confound an equivocal understanding of
the mechanism of action of nutritional supplement thera-
pies. This may be especially relevant for analgesic
medications used and administered by patients diagnosed
with OA, given that GIT dysbiosis may play a significant
role in sustaining inflammatory sequelae both locally and
systemically (Vitetta et al. 2012).
GIT microbial profiles have not yet been hitherto con-
sidered relevant in assessing symptom attenuation in OA
patients. In this study we partly demonstrate that the GIT
microbiome may provide new insights into the control of
gastrointestinal pro-inflammatory and/or adverse immune
responses in OA. Therapeutic interventions modulating a
poorly regulated GIT inflammatory response might be of
benefit in musculoskeletal diseases. We observed that an
improvement in GSRS scores correlated with improvement
in OA joint pain and flexibility in this study and this would
tend to support this nexus. Hence a consequent biologically
plausible postulate for GLM and GS efficacy in this study
may intimately involve the rescue of GIT analgesic pro-
moted dysbiosis by high doses of GLM and GS with
concurrently less use of analgesic medications over the
study period.
Studies with germ-free mice have demonstrated a link
between segmented filamentous bacteria—Clostridium-
related, Gram-positive bacteria and adherence to GIT
epithelial cells and generation of TH17 cells (Smith
et al. 2007; Gaboriau-Routhiau et al. 2009). The sig-
nificance of these studies is that segmented filamentous
bacteria (as a member of the Clostridia sp.) may trigger
T-cell driven GIT inflammation (Stepankova et al. 2007)
and arthritis (Wu et al. 2010) in mouse models. Clos-
tridia sp. also induce T regulatory lymphocytes that play
a critical role in the maintenance of immune homeostasis
(Atarashi et al. 2011). In this clinical trial, the Clos-
tridium sp. was observed to be reduced in both treatment
groups over the 12-week period. Concurrently the
decrease in the Clostridium sp. count in both groups was
paralleled to reductions in knee OA symptoms. Thus,
the results of this clinical study suggest that re-regula-
tion of the GIT pro-inflammatory response may have
occurred.
Investigating nutraceuticals such as GLM and GS can
provide further clues as to how the GIT microbiota may
normalise dysregulated inflammatory responses within the
GIT that produce systemic responses elsewhere in the host.
There is no rigidly defined ‘normal’ reference range for
mucosal bacterial species within the human GIT. Rather,
the interpretation is by common growth trends with a large
number of microbiota species shared between individuals.
Hence, understanding the precise molecular interactions
that encourage bacterial species to induce immune-stimu-
lation by promoting either pro- or anti-inflammatory
activity will be an important step towards manipulating the
GIT microbiota for achieving therapeutic responses both
within and beyond the gut (Barnes and Powrie 2011).
We also note that glucosamine has been reported to have
poor bioavailability after oral administration (Aghazadeh-
Habashi et al. 2002), and as such we postulate that the GIT,
rather than the liver, would appear to be the most likely site
for the first-pass effect. A large number of GIT microbial
species are known to ferment glucosamine, including
Lactobacilli (L. casei, L. plantarum, L. acidophilus, and
L. leichmanii), Streptococcus, Staphylococcus and Leuco-
nostoc (Koser et al. 1961) and Escherichia coli,
Enterococcus faecalis, Proteus vulgaris and Bacillus coli
(Lutwak-Mann 1941; Wolfe and Nakada 1956).
We surmise then that the GIT is often a forgotten organ
when considering the pathogenesis of musculoskeletal
diseases. Future clinical studies that include a placebo and
GIT microbiota profiling may further demonstrate and
elucidate clear differences in GIT microbiota shifts that can
be significantly correlated to the rescue of GIT dysbiosis.
Such studies may thus further strengthen the evidence-base
of GLM and GS efficacy in OA symptom improvement and
management.
Acknowledgments We thank Prof Michael Whitehouse for
reviewing the manuscript and Christine Downey [from the PA Hos-
pital Rheumatology Clinic] for assisting with patient recruitment and
follow-ups.
Conflict of interest GLM extract capsules were donated by Aroma
New Zealand Ltd who had no involvement in the collection, analysis
or interpretation of the data, writing the report or the decision to
submit the paper for publication.
Ethical approval and clinical trial registration Approval for this
prospective study was obtained from the Ethics Committee of The
University of Queensland and Princess Alexandra Hospital Human
Research Ethics Committees. The clinical trial was registered with the
Australian and New Zealand Clinical Trail Registry [ANZCTRN:
12611000517976].
References
Aghazadeh-Habashi A, Sattari S, Pasutto F, Jamali F (2002) Single
dose pharmacokinetics and bioavailability of glucosamine in the
rat. J Pharm Pharm Sci 5(2):181–184
Altman R, Ach E, Bloch D et al (1986) Development of criteria for
the classification and reporting of osteoarthritis: classification of
osteoarthritis of the knee. Arthr Rheum 29(8):1039–1049
Gastrointestinal microbiota profiles in patients with OA
123
Atarashi K, Tanoue T, Shima T et al (2011) Induction of colonic
regulatory T cells by indigenous Clostridium species. Science
331(6015):337–341
Barnes MJ, Powrie F (2011) Immunology. The gut’s Clostridium
cocktail. Science 331(6015):289–290
Bellamy N (2002) WOMAC: a 20-year experiential review of a
patient-centred self-reported health status questionnaire. J Rheu-
matol 29(12):2473–2476
Bellamy N, Buchanan W, Goldsmith CH et al (1988) Validation study
of WOMAC: a health status instrument for measuring clinically
important patient relevant outcomes to anti-rheumatic drug
therapy in patients with osteoarthritis of the hip or knee.
J Rheumatol 15:1833–1840
Bjordal J, Ljunggren AE, Klovning A, Slordal L (2004) Non-steroidal
anti-inflammatory drugs, including cyclo-oxygenase-2 inhibi-
tors, in osteoarthritis knee: meta-analysis of randomised placebo
controlled trials. Br Med J 329(7478):1317–1322
Brien S, Prescott P, Coghlan B et al (2008) Systematic review of the
nutritional supplement Perna Canaliculus (green-lipped mussel)
in the treatment of osteoarthritis. Q J Med 101(3):167–179
Chitturi S, George J (2002) Hepatotoxicity of commonly used drugs:
nonsteroidal anti-inflammatory drugs, antihypertensives, antidia-
betic agents, anticonvulsants, lipid-lowering agents, psychotropic
drugs. Semin Liver Dis 22(2):169–183
Chong V, Wang C (2008) Higher prevalence of gastrointestinal
symptoms among patients with rheumatic disorders. Singap Med
J 49(5):419–424
Coulson S, Vecchio P, Gramotnev H, Vitetta L (2012) Green-lipped
mussel (Perna canaliculus) extract efficacy in knee osteoarthritis
and improvement in gastrointestinal dysfunction: a pilot study.
Inflammopharmacology 20(2):71–76
Egert M, de Graaf AA, Smidt H et al (2006) Beyond diversity:
functional microbiomics of the human colon. Trends Microbiol
14:86–91
Ehrlich G (2003) Global treatment of osteoarthritis. Inflammophar-
macology 11(4–6):333–336
Ernst E (1998) Usage of complementary therapies in rheumatology: a
systematic review. Clin Rheumatol 17:301–305
Foley S, Stolarczyk E, Mouni F et al (2008) Characterisation of
glutamine fructose-6-phosphate amidotransferase (EC 2.6.1.16)
and N-acetylglucosamine metabolism in Bifidobacterium. Arch
Microbiol 189(2):157–167
Gaboriau-Routhiau V, Rakotobe S, Le
´cuyer E et al (2009) The key role
of segmented filamentous bacteria in the coordinated maturation
of gut helper T cell responses. Immunity 31(4):677–689
Garcia Rodriguez LA, Hernandez-Diaz S (2001) Relative risk of
upper gastrointestinal complications among users of acetamino-
phen and nonsteroidal anti-inflammatory drugs. Epidemiology
12(5):570–576
Gillett R, Norrell A (2011) Considerations for safe use of statins: liver
enzyme abnormalities and muscle toxicity. Am Fam Physician
83(6):711–716
Guarner F, Malagelada JR (2003) Gut flora in health and disease.
Lancet 361(9356):512–519
Holzapfel W, Haberer P, Snel J et al (1998) Overview of gut flora and
probiotics. Int J Food Microbiol 41:85–101
Kelly D, Conway S, Aminov R (2005) Commensal gut bacteria:
mechanisms of immune modulation. Trends Immunol 26(6):
326–333
Koser S, Tribby I, Stuedell J (1961) Glucosamine utilization by some
lactic acid bacteria. J Infect Dis 108:324–332
Laparra JM, Sanz Y (2010) Interactions of gut microbiota with
functional food components and nutraceuticals. Pharmacol Res
61(3):219–225
Lequesne MG, Mery C, Samson M, Gerard P (1987) Indexes of
severity for osteoarthritis of the hip and knee validation-value in
comparison with other assessment tests. Scand J Rheumatol
65:85–89
Lutwak-Mann C (1941) Enzymatic decomposition of amino-sugars.
Biochem J 35(5):610–626
Reginster JY (2002) The prevalence and burden of arthritis.
Rheumatology (Oxford) 41(Supp 1):3–6
Rutjes A, Nu
¨esch E, Reichenbach S, Ju
¨ni P (2008) S-Adenosylme-
thionine for osteoarthritis of the knee or hip. Cochrane Database
Syst Rev 4:CD007321
Smith K, McCoy KD, Macpherson AJ (2007) Use of axenic animals
in studying the adaptation of mammals to their commensal
intestinal microbiota. Semin Immunol 19(2):59–69
Stepankova R, Powrie F, Kofronova O et al (2007) Segmented
filamentous bacteria in a defined bacterial cocktail induce intesti-
nal inflammation in SCID mice reconstituted with CD45RBhigh
CD4?T cells. Inflamm Bowel Dis 13(10):1202–1211
Svedlund J, Sjodin I, Dotevall G (1988) GSRS: a clinical rating scale
for gastrointestinal symptoms in patients with irritable bowel
syndrome and peptic ulcer disease. Dig Dis Sci 33:129–134
Tana C, Umesaki Y, Imaoka A et al (2010) Altered profiles of
intestinal microbiota and organic acids may be the origin of
symptoms in irritable bowel syndrome. Neurogastroenterol Motil
22(5):512–519, e114–e115
Thrupp L (1980) Susceptibility testing of antibiotics in liquid media.
In: Lorian V (ed) Antibiotics in laboratory medicine, chap 4.
Williams & Wilkins pub, Baltimore, pp 73–113
Towheed T, Maxwell L, Anastassiades TP et al (2005) Glucosamine
therapy for treating osteoarthritis. Cochrane Database Syst Rev
2:CD002946
Tucker G, Adams R, Wilson D (2010) New Australian population
scoring coefficients for the old version of the SF-36 and SF-12
health status questionnaires. Qual Life Res 19:1069–1076
Ulbricht C, Chao W, Costa D et al (2009) An evidence-based
systematic review of green-lipped mussel (Perna canaliculus)by
the Natural Standard Research Collaboration. J Diet Suppl
6(1):54–90
Umesaki Y, Setoyama H, Matsumoto S et al (1999) Differential roles
of segmented filamentous bacteria and clostridia in development
of the intestinal immune system. Infect Immun 67:3504–3511
Vitetta L, Briskey D, Hayes E et al (2012) A review of the
pharmacobiotic regulation of gastrointestinal inflammation by
probiotics, commensal bacteria and prebiotics. Inflammophar-
macology (Epub ahead of print)
Wandel S, Juni P, Tendal B et al (2010) Effects of glucosamine,
chondroitin or placebo in patients with osteoarthritis of hip or
knee: network meta-analysis. BMJ 341:c4675
Ware J, Kosinski M, Keller S (1996) A 12-item short-form health
survey: construction of scales and preliminary tests of reliability
and validity. Med Care 34(3):220–233
Willis A (2007)Anaerobic culture methods, in anaerobicmicrobiology: a
practical approach. IRL Press at Oxford University Press, Oxford
Wolfe J, Nakada H (1956) Glucosamine degradation by Escherichia
coli. II. The isomeric conversion of glucosamine 6-PO4 to
fructose 6-PO4 and ammonia. Arch Biochem Biophys 64(2):
489–497
Wolfe J, Morita R, Nakada H (1956) Glucosamine degradation by
Escherichia coli, I. Observations with resting cells and dry-cell
preparations. Arch Biochem Biophys 64:480–488
Wolfe F, Kong S, Watson D (2000) Gastrointestinal symptoms and
health related quality of life in patients with arthritis. J Rheuma-
tol 27(6):1373–1378
Woolf AD, Pfleger B (2003) Burden of major musculoskeletal
conditions. Bull World Health Organ 81(9):646–656
Wu H, Ivanov I, Darce J et al (2010) Gut-residing segmented
filamentous bacteria drive autoimmune arthritis via T helper 17
cells. Immunity 32(6):815–827
S. Coulson et al.
123
