24th January 03:40
Autism and the Human Gut Flora
Autism and the Human Gut Flora
Dr. Max Bingham
Food Microbial Sciences Unit, Science and Technology Centre, Earley
Gate, University of Reading,
Whiteknights Road, Reading, Berkshire, UK. RG6 6BZ
Previous research into autism has raised the possibility that a link
exists between the disorder and the human gut microflora. Research
has remained limited due to the unwillingness of the orthodox medical
establishment to adopt treatments suggested by this research. This
short review aims to summarise the research completed to date and
evaluate the relevance of a link between autism and a possible
imbalance in the human gut microflora. It appears that yeasts
(Candida species in particular) and clostridia may play an important
role in the development of autistic symptoms. It is suggested that
the control of the growth of these species may reduce the severity of
autistic symptoms but is unlikely to offer a cure.
Previous research into autism has raised the possibility that a link
exists between the disorder and the human gut microflora. However the
relevance of this to sufferers of Autism has remained somewhat un-
researched. As a consequence, it appears that the treatments
suggested previously have remained unused and regarded by the orthodox
medical establishment as irrelevant. This short review will consider
the limited research completed to date.
The term autism is usually associated with the syndrome first
described by Kanner (1943). In more recent years specific criteria
have been set out to aid in the diagnosis of the disorder (Shaw et al,
1999). Autism typically develops early in childhood however
causality, explanation and treatment are often hotly debated.
Symptoms can include (but not necessarily), hyperactivity, loss of eye
contact, decreased vocalisation (i.e. loss of language), stereotypical
behaviours, poor academic performance and other similar social
deficits. Other similar disorders exist. These include Asperger
Syndrome, Attention Deficit Hyperactivity Disorder (ADHD), Pervasive
developmental disorder (PDD) and many others, where symptoms are
similar to Autism but specific differences are demonstrated.
Yeast metabolites in the Urine of Autistic Children
Yeasts constitute only a very small proportion of the population of
the gut (Holzapfel et al, 1998) under normal conditions possibly
kept from growing via competition from bacterial species and certain
immune functions. However, Shaw et al (1999) has proposed that Autism
(or at least many of the symptoms) may be a consequence of an
overgrowth of candida species and a selective IgA deficiency.
Following treatment with antifungal drugs and a gluten and casein free
diet, a child rated as having severe autism improved such an amount as
to be classed as a higher functioning individual with autism.*
It has been shown previously that children exhibiting autistic
features have increased excretion of arabinose and the ****ogs of
Krebs Cycle metabolites (including tartaric acid) (Shaw, Kassen &
Chaves, 1995). Using Gas Chromatography/Mass Spectrometry (GC/MS) to
test for urinary metabolites, it was found that the children had
extremely high values of tartaric acid. The only source of tartaric
acid is yeast (Shaw, 1999). Many reports have suggested that the
onset of autism may be related to the occurrence in children of otitis
media (Kontstantareas & Homatidis, 1987). It is common to treat
otitis media with some sort of broad-spectrum antibiotic. Intestinal
overgrowth of yeast and certain anaerobic bacteria are a well
do***ented outcome of the administration of broad spectrum antibiotics
(Kennedy & Volz, 1983; Danna et al, 1991; Ostfield et al, 1977;
Kinsman et al, 1989; Van der Waaij, 1987; Samsonis et al, 1993,
To evaluate this, it is known that Shaw (1999, 1996) has used GC/MS to
test the urinary content of metabolites following the administration
of Nystatin, an antifungal drug. Accordingly it was found that
urinary tartaric acid declined to zero after about 60 days. When the
Nystatin dose was cut to half, levels appeared to rise. Associated
with this was seen an improvement in eye contact, a reduction in
hyperactivity and an improvement in sleep patterns. It is unclear as
to the method used, however these are interesting results. In other
work, Shaw (1999) has evaluated the progress of the Herxheimer
reaction of yeast die off. Values for the microbial metabolites
described above increased dramatically during the first three days and
began to normalise near day four. It is unclear as to the relevance
of the point that many children take antibiotics for a whole series of
illnesses, but do not develop autism.
Gupta et al (1996) have studied some possible modulating factors in
Autism. These include immuno-deficiencies and probable differences in
biochemical detoxification factors which appear to be very common in
autism. It has been estimated that a high percentage of autistic
children have a significant immune dysfunction that may include
myeloperoxidase deficiency, a genetic deficiency that impairs the
action of white blood cells on yeast cells, IgA deficiency, complement
C4b deficiency, IgG deficiency or IgG subclass deficiency (Gupta et
al, 1996). In the above study, a complete remission of autistic
symptoms was achieved by infusions of gamma globulin in one case.
Shaw (1999) has suggested that it seems increasingly likely that the
immune system takes an inventory of bacteria and yeast cells present
in the gut soon after birth. This inventory is performed by CD5+ B-
cells. These cells may play a role in regulating the secretion of
IgA, the antibody class that is secreted into the intestinal tract and
which may select which micro-organisms are tolerated in the gut.
Continuing, it is suggested that the eradication of the normal gut
flora when antibiotics are administered repeatedly during infancy may
cause the CD5+ cells to reject normal cells and award immune tolerance
to species that potentially could cause harm. Either antibiotic use
in infancy or yeast infection of the mother during pregnancy may
result in later tolerance to yeast. This would possibly extend any
remission from symptoms induced by methods of treatment suggested for
Yeast and Tartaric Acid Toxicity
Shaw, Kassen & Chaves (1995) noted in their initial study of the two
brothers with autism, that in addition to their autistic
characteristics, they exhibited severe muscle weakness. Both of these
brothers excreted large amounts of tartaric acid in their urine. It
is known that tartaric acid is an ****og of malic acid. Malic acid is
a key intermediate in the Krebs cycle which is responsible for the
extraction of most of the energy from food substrates. Tartaric acid
is presumably seen as toxic since it would inhibit this pathway and
limit energy production. This may be the reason for the muscle
weakness seen in the two autistic brothers. It has also been shown
that Candida albicans produce gliotoxins (Shah and Larsen, 1991 and
1992) and immunotoxins (Podzorski et al, 1989; Witken, 1985), which
may further impair the immune system. This would have relevance in
terms of promoting yeast growth and increasing the chances of
additional infections from bacteria leading to antibiotic usage again.
Arabinose, Yeast and Autism
Shaw, Kassen & Chaves (1995), also identified high levels of arabinose
in the urine of the autistic children. The exact biochemical role of
arabinose is unknown, but a closely related yeast alcohol, arabitol
has been used as a biochemical indicator of invasive candidiasis
(Kiehn et al, 1979; Wong et al, 1990, Roboz & Katz, 1992). It is
thought that arabitol produced by the yeast in the gut is absorbed
into the portal circulation and then converted to arabinose by the
liver. Elevated protein-bound arabinose has been found in the serum
glycoproteins of schizophrenics (Varma & Hoshino, 1980) and in
children with conduct disorders (Varma et al, 1983).
Arabinose reacts with the epsilon amino group of lysine in a wide
variety of proteins and may then form cross-links with arginine
residues in an adjoining protein (Sell & Monnier, 1989). Arabinose
has therefore been implicated in protein modification via cross-
linking the proteins and altering both biological structures and
functions of a wide variety of proteins. Shaw et al (1999) has
proposed that this may include proteins involved in the
interconnection of neurons. Decreased clinical symptoms of autism
after antifungal treatment might be due to decreased arabinose and
pentosidine formation (cross-linked proteins described above). This
would result in fewer random neural connections and increased numbers
of neural connections that are oriented to the childs environment.
The influence of a number of vitamins on candida activity and the
operation of many metabolites of candida has been discussed (Mahler &
Cordes, 1966). It is commonly found that children with autism
experience improved symptoms following removal of gluten and casein
from the diet (Shattock et al 1990, 1991, 1996, 1997; Reichelt et al,
1981, 1986; Knivsberg et al, 1990). It is unclear whether there is a
link with arabinose and pentosidine formation, however given the wide
distribution of these proteins in foods, there may be a relevant
process involved with respect to autistic symptoms. Research into
this may be useful.
Clostridia and Autism
Bolte (1998) outlined the possibility of a sub-acute, chronic tetanus
infection of the gut as the underlying cause for symptoms of autism
observed in some individuals. Here it is postulated that a percentage
of individuals with autism have a history of extensive antibiotic
use. As above it is known that oral antibiotics significantly disrupt
protective gut microflora, creating a favourable environment for
colonisation by opportunistic pathogens. Clostridium tetani is a
ubiquitous anaerobic bacillus that is known to produce a potent
neurotoxin. The normal site of binding for tetanus neurotoxin is the
spinal cord, yet the vagus nerve is capable of transporting tetanus
neurotoxin, providing a route of ascent from the intestinal tract to
the central nervous system and bypassing the spinal cord. The result
would be the typical symptoms of a tetanus infection would not be
evident. Once in the brain the tetanus neurotoxin disrupts the
release of neurotransmitters. This may explain the wide variety of
behavioural deficits apparent in autism. Bolte (1998) presents
evidence of lab animals exhibiting many of these behaviours after
being injected in the brain with tetanus neurotoxin, and also of
children with autism showing a significant reduction in stereotyped
behaviour following treatment with antimicrobials effective against
Shaw (1999) has developed a similar technique to yeast metabolite
****ysis by GC/MS for clostridia metabolites. While it has not been
fully identified it has been shown that tyrosine derivative, a
compound very similar but not identical to 3,4-dihydroyphenylpropionic
acid is raised in the urine of children with various conduct
disorders. It has also been pointed out that tyrosine, a substrate
used by the body for the production of neurotransmitters, might imply
that this unidentified product is important in altering key
biochemical pathways for neurotransmitters in the brain. The
relevance to autism remains unclear. Clearly a treatment to reduce
the numbers of clostridia in the gut might help to reduce the effects
of this inhabitation. However, clostridia are spore forming meaning
re-colonisation is possible even after treatment. The relevance of
probiotic and prebiotic treatments has not been researched, yet this
appears a reasonable approach to aid suppression of the clostridia
Human Gut Flora and Gluten and Casein Intolerance in Autistic Children
Yeasts, including Candida albicans are known to secrete a number of
enzymes. These may include phospholipase, which will break down
phopholipids and proteases such as secretory aspartate protease which
break down proteins. These enzymes may partially digest the gut
membranes and lining itself. Furthermore, it is known that the
mycelium and chlamydospore are capable of tissue invasion (Nolting et
al, 1994). It is likely that such factors could increase the
permeability of the gut. This has important consequences in terms of
food absorption and digestion. In terms of autism symptoms this may
be highly relevant. It has been shown that incompletely broken down
portions of gluten and casein may be crossing the gut into the blood
and having an opioid effect in autistic children. Their symptoms
being a consequence of this opioid action (Reichelt, 1981; Shattock et
al, 1990). While many mechanisms have been suggested for this
incomplete breakdown, it seems key that a yeast and/or clostridia
overgrowth would affect this in some way.
Products of the human gut microflora in relation to autism and its
symptoms appear to have been largely ignored in the past. However
they appear to be relevant certainly in terms of yeast and clostridia
species. Abnormal bacterial metabolites of tyrosine appear to be
elevated in autism and this seems related largely to an overgrowth in
clostridia species. Treatments to control this have resulted in
significant clinical improvement or complete remission of symptoms in
some cases. The issue of whether probiotic and prebiotic treatments
are relevant needs to be researched. Elevation of yeast metabolites
such as tartaric acid and arabinose appears to be even more common in
autism. The arabinose appears to be involved in abnormal protein
binding which may adversely affect neuron connection and this may have
relevance to the appearance of autistic symptoms. Given the
intolerance autistic children appear to have to gluten and casein, an
involvement of arabinose may be important. However, the exact
biochemical role of arabinose remains unclear. Tartaric acid from the
yeast overgrowth may have a direct toxic effect on the muscles and is
a key inhibitor of the Krebs cycle that supplies raw materials for
gluconeogenesis. The implications of this appear to be detrimental to
While it is clear that abnormal metabolites from the human gut flora
may contribute to autistic symptoms, it is certain that many other
systems and pathways are involved. It is possible that many are
functioning simultaneously but at varying levels between individuals.
This may give rise to the differences in symptoms exhibited by
sufferers. This may also explain the similarity of autism and other
chronic behavioural disorders such as Asperger Syndrome, PDD, ADHD,
ADD and Rett syndrome. While other pathways may exist, research into
the role of human gut flora and the full identification of species
involved may allow for more directed treatments. While it will not
cure the disorder, modifications of gut flora function might improve
Bolte ER, (1998). Autism and Clostridium tetani. Medical Hypotheses
51(2): 133-144 [View abstract]
Danna P, Urban C, Bellin E and Rahal J, (1991). Role of Candida in
pathogenesis of antibiotic associated diarrhoea in elderly patients.
Lancet 337: 511-14.
Gupta S, Aggarwal S and Heads C (1996). Dysregulated immune system in
children with autism: beneficial effects of intravenous immune
globulin on autistic characteristics. Journal of Autism and
Developmental Disorders 26(4): 439-52.
Holzapfel WH, Haberer P, Snel J, Scillinger U, Jos HJ, Velt Hi,
(1998). Overview of gut flora and probiotics. International Journal
of Food Microbiology 41: 85-101.
Kanner L, (1943). Autistic disturbances of affective contact. The
Nervous Child 2: 217-250.
Kiehn T, Bernard E, Gold J & Armstrong D, (1979). Candidiasis:
detection by gas-liquid chromatography of D-arabitol, a fungal
metabolite, in human serum. Science 206: 577-580.
Kinsman OS and Pitblado K, (1989). Candida albicans gastrointestinal
colonization and invasion in the mouse: effect of antibacterial
dosing, antifungal therapy, and immunosuppression. Mycoses 32:
Knivsberg A-M, Wiig K, Lind G, Nødland M & Reichelt K-L, (1990).
Dietary Intervention in Autistic Syndromes. Brain Dysfunction 3:
Kontstanareas M and Homatidis S, (1987). Ear infections in autistic
and normal children. Journal of Autism and Developmental Disorders
Kennedy M and Volz P (1983). Dissemination of yeasts after
gastrointestinal inoculation in antibiotic-treated mice. Sabouradia
Mahler H & Cordes E (1966). Biological Chemistry. Harpur and Row, New
York pp 322-375.
Nolting S, Guzek B, Hauss R, (1995). Mykosen des Verdauungstraktes,
Med. Wiss. Mediendienst Hamburg.
Ostfeld E, Rubinstein E, Gazit E and Smetana Z, (1977). Effect of
systemic antibiotics on the microbial flora of the external ear c****
in hospitalized children. Pediatrics 60: 364-66.
Podzorski R, Herron M, Fast D & Nelson R (1989). Pathogenesis of
candidiasis. Immunosuppression by cell wall mannan catabolites.
Archives of Surgery 124: 1290-1294.
Reichelt K-L, Saelid G, Lindback T & Boler JB, (1986). Child Autism:
a complex disorder. Biological Psychiatry 21: 1279-1290.
Reichelt K-L, Hole K, Hamberger A, Saelid G, Edminson PD, Braestrup
CB, Lingjaerde P & Orbeck H, (1981). Biologically active peptide
containing fractions in schizophrenia and childhood autism. Advances
in Biochemical Psychopharmacology 28: 627-643.
Roboz J & Katz R, (1992). Diagnosis of disseminated candidiasis based
on serum D/L arabinitol ratios using negative chemical ionisation mass
spectrometry. Journal of Chromatography 575: 281-286.
Shah D & Larsen B, (1992). Identity of a Candida albicans toxin and
its production in ******l secretions. Medical Science Research 20:
Shah D & Larsen B, (1991). Clinical isolates of a yeast produce a
gliotoxin-like substance. Mycopathologica 116: 203-208.
Samonis G and Dassiou M, (1994a). Antibiotics affecting
gastrointestinal colonization of mice by yeasts. Chemotherapy 6: 50-2.
Samonis G, Gikas A and Toloudis P, (1994b). Prospective evaluation of
the impact of broad-spectrun antibiotics on the yeast flora of the
human gut. European Journal of Clinical Microbiology & Infectious
Diseases 13: 665-7.
Samonis G, Gikas A and Anaissie E, (1993). Prospective evaluation of
the impact of broad-spectrun antibiotics on gastrointestinal yeast
colonization of humans. Antimicrobial Agents and Chemotherapy 37:
Sell D & Monnier V, (1989). Structure elucidation of a senescence
cross-link from human extracellular matrix. Implications of pentoses
in the aging process. Journal of Biological Chemistry 264:
Shattock P & Savery D, (1997). Autism as a Metabolic Disorder. Autism
Research Unit, University of Sunderland, Sunderland, UK. [Link]
Shattock P & Savery D, (1996). Urinary profiles of people with
autism: Possible implications and relevance to other research. In:
Proceedings of Conference on The****utic Interventions in Autism:
Perspectives from Research and Practice. 309-325.
Shattock P & Lowdon G, (1991). Proteins, Peptides and Autism. Part 2:
Implications for the education and care of people with autism. Brain
Dysfunction 4(6): 323-334.
Shattock P, Kennedy A, Rowell F & Berney T, (1990). Role of
neuropeptides in autism and their relationships with classical
neurotransmitters. Brain Dysfunction 3: 328-345.
Shaw W, Baptist J, Geenens D, (1999). Immunodeficiency,
gastrointestinal Candidiasis, wheat and dairy sensitivity, abnormal
urinary arabinose and autism: a case study. Unpublished: Great Plains
Laboratory, Overland Park, Kansas 66204.
Shaw W, (1999) Role for certain yeast and bacteria byproducts
discovered by organic acid testing in the etiology of a wide variety
of human diseases. Bulletin of The Great Plains Laboratory. Overland
park, KS 66204 (913) 341-8949.
Shaw, W and Chaves E, (1996). Experience with organic acid testing to
evaluate abnormal microbial metabolites in the urine of children with
autism. Published in: The Proceedings of the Autism Society of America
National Conference on Autism. Milwaukee: Wisconsin.
Shaw W, Kassen E and Chaves E, (1995). Increased excretion of ****ogs
of Krebs cycle metabolites and arabinose in two brothers with autistic
features. Clinical Chemistry 41: 1094-1104.
Van der Waaij D, (1987). Colonization resistance of the digestive
tract--mechanism and clinical consequences. Nahrung 31: 507-17.
Varma R & Hoshino AY, (1980). Serum glycoproteins in schizophrenia.
Carbohydrate Research. 82(2): 343-51.
Varma R, Michos GA, Gordon BJ, Varma RS, Shirey RE (1983). Serum
glycoconjugates in children with schizophrenia and conduct and
adjustment disorders. Biochemical Medicine. 30(2): 206-14.
Witken S (1985). Defective immune responses in patients with
recurrent candidiasis. Infections in Medicine May/ June pp 129-132.
Wong B, Brauer K, Clemens J & Beggs S, (1990). Effects of
gastrointestinal candidiasis, antibiotics, dietary arabinitol, and
cortisone acetate on levels of Candida metabolite D-arabitol in rat
serum and urine. Infection and Immunity 58: 283-288.
* The Childhood Autism Rating Scale, an observational measure of
various aspects of autism, for the child in question decreased from 43
(severely autistic) prior to introduction of these therapies to a
value of 29 (non-autistic) after therapy.