Primary gastrointestinal pathology may play an important role in the inception and clinical expression of some childhood developmental disorders, including autism. During the past 4 years, it has been my privilege to work with one of the finest pediatric gastroenterology teams in the world, headed by John Walker-Smith, on an innovative and challenging investigation of gastrointestinal pathology in children with autism. We believe that this work will provide new and important insights into the pathogenesis of this devastating condition. Although the primary causes of autism may be diverse, clues to the possible origin of the disease may be found in the history and clinical investigation of affected children. This talk focuses on the significance of gastrointestinal symptoms in autistic children, in particular, a subset of children for whom the clinical course is characterized by regression after at least 12 to 15 months of normal development.
In addition to frequent gastrointestinal symptoms, children with autism often manifest complex biochemical, metabolic, and immunologic abnormalities that a primary genetic cause cannot readily account for. The gut–brain axis is central to certain encephalopathies of extracranial origin, hepatic encephalopathy being the best characterized. Certain commonalties between the clinical characteristics of hepatic encephalopathy and an increasingly common autistic phenotype (developmental regression in a previously normal child accompanied by immune-mediated gastrointestinal pathology) have led to the hypothesis that an analogous mechanism of toxic encephalopathy may exist in patients with liver failure and some children with autism. Aberrations in opioid biochemistry are common to these two conditions, and evidence suggests that opioid peptides may be among the central mediators of the respective syndromes. Generating biologically plausible and testable hypotheses in this area may help to identify new treatment options in encephalopathies of extracranial origin.
Interest is growing in the role of the gut in childhood developmental disorders, principally autism. Some of the earliest commentators on pathobiologic mechanisms in autism made the observation that gastrointestinal symptoms are common in children with developmental disorders. Dohan wrote: “K Soddy (University College Hospital, London) wrote me that he noted that recurrent gastrointestinal upsets were a constant feature of autistic children and that, among other symptoms, the deteriorating autistic child often has acute diarrhea” (1). These observations have been iterated in parental accounts (2), but largely ignored by the medical profession (Bauman M, personal communication). More recently, in a systematic analysis of an unselected population of 385 children on the autistic spectrum, clinically significant gastrointestinal symptoms occurred in 46% compared with 10% of 97 developmentally normal pediatric controls (odds ratio, 7.4; confidence intervals, 3.60–15.65;P < 0.0001) (3). We investigated gastrointestinal symptoms in more than 150 autistic children and reported out initial experience (4,5). In this cohort, developmental regression and loss of acquired skills, sometimes occurring precipitously over a period of days to weeks, followed a period of initial normal development; in short, the children became encephalopathic. Longstanding intestinal symptoms, including chronic constipation with overflow, and provocation of both gastrointestinal and behavioral symptoms by certain foods, are typical of this group of children and have been described by others (6). However, even among some pediatric gastroenterologists, the perception remains that such symptoms are to be expected in children with developmental disorders, reflecting effect rather than possible cause. In the absence of thorough investigation, the basis for this presumption is unclear. It is essential, ab initio, to ask whether these symptoms reflect underlying pathology of the gastrointestinal tract.
In 1972, Walker-Smith and Andrews (7) reported low concentrations of α-1-antitrypsin in children with classic autism, a finding indicative of intestinal protein loss. We have subsequently reported increased fecal excretion of calprotectin in children with autistic regression associated with enterocolitis. There was a high degree of correlation between fecal calprotectin and neutrophil infiltration of the colonic mucosa (8). Calprotectin is a stable neutrophil product, and its presence in feces is an established marker of active inflammation in inflammatory bowel disease.
Intestinal permeability, as measured by urinary excretion of metabolically inert sugars after oral dosing, is a surrogate marker of mucosal integrity and is increased in the presence of intestinal inflammation, as in Crohn disease (9) and celiac disease (10). D'Eufemia et al. (11) reported that approximately half of a cohort of autistic children without gastrointestinal symptoms had increased intestinal permeability. Horvath et al. (12) have confirmed the increased intestinal permeability in children with autism. D'Eufemia's detection of aberrant intestinal permeability in asymptomatic autists indicates that reliance on overt symptomatology will substantially underestimate the proportion of autistic children with possible gastrointestinal pathology. The combination of an increased pain threshold, commonly observed in affected children, (13), and restricted ability to communicate symptoms will compound this underestimation.
Identifying decreased serum α-1-antitrypsin, increased fecal calprotectin, and increased intestinal permeability in a patient is not a diagnostic endpoint, but indicates the need for further detailed investigation. As a corollary to these changes, does evidence suggests that children with autism and gut symptoms have demonstrable organic pathology of the gastrointestinal tract? The answer seems to be yes. We recently described a characteristic pattern of intestinal pathology—ileocolonic lymphoid nodular hyperplasia and enterocolitis—in a large cohort of autistic children. The endoscopic and histopathologic characteristics of this condition have been reported in detail elsewhere (4,5). A comparison of the mucosal lesion in the colon and small intestine, with appropriate controls, shows a subtle but characteristic disease process. Briefly, the colonic lesion consists of a mucosal infiltrate of γδ T cells and CD8+ T cells, significantly in excess of that seen in either healthy or disease control groups (14). Crypt cell proliferation is substantially enhanced, and the epithelial basement membrane is thicker than in either healthy or disease control groups. Neutrophil and eosinophil infiltration of the mucosa is evident. The absence of colonic epithelial HLA-DR in autistic children suggests a Th2-dominated immune response (14,15). Studies of the corresponding small intestinal lesion also indicate a distinct cell-mediated immunopathology in which immune-mediated epithelial damage is predominant, serum immunoglobulin G colocalizes with complement (C1q) at the epithelial basolateral membrane, and epithelial proliferation is grossly increased (16). This is not seen in either healthy children or those with cerebral palsy. The intestinal changes are consistent with an autoimmune pathology and, in view of the increasing evidence for gut epithelial dysfunction in autism, are indicative of a specific and possibly important lesion. Sabra and Bellanti (17) presented preliminary evidence of similar findings in children with attention deficit hyperactivity disorder, suggesting that gastrointestinal pathology may be relevant to a broader spectrum of childhood developmental/behavioral disorders.
Horvath et al. (6) reported their findings in the upper gastrointestinal tract in 36 autistic children whose symptoms included chronic diarrhea, gaseousness, abdominal discomfort, and distension. They detected grade I to II reflux esophagitis in 25 (69%), chronic gastritis in 15 (42%), and chronic duodenitis with associated Paneth cell hyperplasia in 24 (67%). Digestive enzyme activity was decreased in 21 (58%) autistic children, and pancreaticobiliary fluid output in response to intravenous secretin was increased in 27 (75%). Subsequent to Horvath's early report (6), we have included upper gastrointestinal endoscopy in our routine assessment of affected children and our findings support his.
THE GUT–BRAIN AXIS
Neuroactive compounds derived from the intestinal lumen can permeate the mucosa; cross the blood–brain barrier; and cause psychiatric, cognitive, and behavioral disturbances. Indeed, this axis is critical in, for example, oral medication of psychopathology. Awareness is growing, particularly within the field of childhood developmental disorders, that in a substantial proportion of affected children, gut–brain interactions may be central to abnormal neural development and the subsequent expression of aberrant behaviors. Difficulties in accepting the biologic plausibility of such a model, particularly among those whose interests have focused on primary pathogenetic mechanisms operating within the central nervous system, may reflect, in part, a perceived lack of an analogous gut–brain interaction in either human or experimental models of encephalopathy. Among gastroenterologists and hepatologists, however, the evidence for such a mechanism is readily apparent. Seeking analogy with circumstances in which clear evidence shows an influence of the gut on the normal brain may help advance the argument.
Untreated celiac disease—an aberrant immune response to dietary gliadin—is associated with intestinal mucosal inflammation, increased intestinal permeability (10), increased absorption and urinary excretion of neuroactive dietary peptides (18), autistic and psychotic behaviors (19,20), and neurologic complications(21,22). The precise mechanism(s) of central nervous system sequelae has not been established, although toxicity from the gut (18) and autoimmunity (23) are pathogenetic forerunners. D-lactic acidosis, a complication of acid-tolerant bacterial overgrowth in patients with short bowel syndrome and those undergoing intestinal bypass surgery for obesity, is associated with a range of psychiatric and neurologic sequelae (24). Patients may experience altered mental state, aggression, stupor, ataxia, and asterixis; these symptoms respond rapidly to oral antibiotic treatment. Encephalopathy is a recognized presenting feature of intestinal intussusception in infants (25–27) and, intriguingly, may be reversible with naloxone (25).
Hepatic encephalopathy—a variable impairment of cerebral functioning in patients with acute or chronic liver disease—is the result of multiple biochemical influences on central neurotransmitter systems. In addition to the neurotoxic effects of ammonia (28), derangements in the gamma-aminobutyric acidergic (28) and serotoninergic (29) systems are evident (reviewed by Butterworth (30), and Albrecht and Jones (31)). It may be more than just coincidence that changes in these neurotransmitter systems also have been described in autism (32,33). Central to changes in the activity of these various systems in hepatic encephalopathy is a failure of the diseased liver to metabolize or eliminate toxic compounds derived from the gut, which may itself be entirely healthy. Accordingly, porta-systemic shunting (rerouting the venous drainage of the gut to bypass the liver and directly enter the systemic circulation) may produce encephalopathy for the amelioration of portal hypertension. Precipitants of hepatic encephalopathy in patients with liver disease include a high enteroluminal protein load, constipation, and sepsis (30,31). The clinical management of hepatic encephalopathy includes restricting dietary protein, promptly treating constipation, and antibiotic therapy directed against luminal colonic bacteria. As such, hepatic encephalopathy represents a prototypic afferent gut–brain interaction that may provide insights into other encephalopathic states that have been linked to extracranial disease.
During the course of clinical assessment and management of these children, we have been impressed by the symptomatic improvement in their behavior and general well-being after bowel clearance before colonoscopy; treatment of intestinal inflammation with 5-amino salicylate-based compounds or a polymeric diet (34); relief of chronic constipation; and, in particular, the elimination of certain proteins (casein or gluten) from the diet (34). Bolte (35) proposed a role for intestinal clostridial dysbiosis in autism, specifically through neurotoxic encephalopathy. In seeking to test this hypothesis Sandler et al. (36) noted objective cognitive improvement in autistic children in an open label study of oral vancomycin, an antibiotic that exhibits minimal systemic absorption. Children regressed after cessation of therapy, suggesting that any colonic dysbiosis and associated toxic sequelae probably were secondary to underlying intestinal disease rather than the primary problem. Failure of vancomycin to eliminate clostridial spores is also a possibility, although less likely, given the efficacy of this drug in treating Clostridium difficile. These observations, although empirical or observed in open label studies, were largely unexpected. In addition, they are reminiscent of certain aspects of the clinical course of hepatic encephalopathy, in which gut–brain interactions are paramount. An analogous interaction may be operating in a subset of autistic children; any biochemical basis for such an interaction in autism might be identifiable in syndromes such as hepatic encephalopathy.
In summary, within the autistic spectrum, a substantial group of children have what may be primary intestinal pathology. The constellation of developmental disorder and gastrointestinal pathology (provisionally termed “autistic enterocolitis”) combines the paradoxic elements of a motility disorder—esophageal reflux and constipation with spurious diarrhea—and enterocolonic mucosal inflammation, a feature more commonly associated with frank diarrhea. Understanding the neurochemical basis of any gut–brain interaction in autistic enterocolitis may help to resolve this paradox and help to develop rational therapeutic approaches.
1. Dohan FC. Schizophrenia: possible relationship to cereal grains and celiac disease. In: S. Sankar, ed. Schizophrenia: Current Concepts and Research. Hicksville, NY: PJD Publications; 1968.
2. Sullivan RC. Hunches on some biological factors in autism. J Autism Child Schizophr 1975; 5:177–84.
3. Melmed R, Schneider CK, Fabes RA, et al. Metabolic markers and gastrointestinal symptoms in children with autism and related disorders. J Pediatr Gastroenterol Nutr 2000; 31(suppl 2):A116.
4. Wakefield AJ, Murch SH, Anthony A, et al. Ileal-lymphoid nodular hyperplasian non-specific colitis, and pervasive developmental disorder in children. Lancet 1998; 351:637–41.
5. Wakefield AJ, Anthony A, Murch SH, et al. Enterocolitis in children with developmental disorders. Am J Gastroenterol 2000; 95:2285–95.
6. Horvath K, Papdimitriou JC, Rabsztyn A, et al. Gastrointestinal abnormalities in children with autistic disorder. J Pediatr 1999; 135:559–63.
7. Walker-Smith JA, Andrews J. Alpha-antitrypsin, autism and celiac disease. Lancet 1972; 2:883–4.
8. Anthony A, Bjnarson I, Sigthorsson G, et al. Fetal calprotectin levels correlate with acute inflammation in autistic enterocolitis. Gut 2000; 46(Suppl 2):A3.
9. Tibble JA, Sigthorsson G, Bridger S, et al. Surrogate markers of intestinal inflammation are predictive of relapse in patients with inflammatory bowel disease. Gastroenterology 2000; 119:15–22.
10. Cummins AG, Celli M, Finocchario R, et al. Recovery of the small intestine in celiac disease on a gluten-free diet: changes in intestinal permeability, small bowel morphology and T-cell activity. J Gastroenterol Hepatol 1991; 6:53–7.
11. D'Eufemia P, Celli M, Finocchinario R, et al. Abnormal intestinal permeability in children with autism. Acta Paediatr 1996; 85:1076–9.
12. Horvath K, Zielke RH, Collins RM, et al. Secretin improves intestinal permeability in autistic children. J Pediatr Gastroenterol Nutr 2000; 31(suppl 2):A112.
13. Gillberg C. The role of endogenous opioids in autism and possible relationships to clinical features. In: Wing L, ed. Aspects of Autism: Biological Research. London: Gaskell/NAS; 1988:31–7.
14. Furlano R, Anthony A, Day R, et al. Quantitative immunohistochemistry shows colonic epithelial pathology and γδ-T cell infiltration in autistic enterocolitis. Gastroenterology
1999;116(suppl): (in press).
15. Anthony A, Sim R, Murch SM, et al. Lymphonodular hyperplasia of the ileum with increased MHC class II antigen expression and macrophage infiltration of the colon in children with regressive developmental disorder. Gut 1998; 42(suppl 1):A24.
16. Torrente F, Machado N, Perez-Machado M, et al. Enteropathy with T cell infiltration and epithelial IgG deposition in autism. J Pediatr Gastroenterol Nutr 2000; 31(suppl 2):A546.
17. Sabra S, Bellanti JA, Colon AR. Ileal lymphoid nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children. Lancet 1998; 352:234–5.
18. Reichelt WH, Stensrud J-EM, Reichelt KL. Peptide excretion in celiac disease. J Pediatr Gastroenterol Nutr 1998; 26:305–9.
19. Hallert C, Derefeldt T, Psychic disturbances in adult coeliac disease. Scand J Gastroenterol 1982; 17:17–9.
20. Asperger H. Der psychopathologie des coeliaki kranken kindes. Ann Paediatr 1961; 187:346–51.
21. Cooke WT, Smith WT. Neurological disorders associated with adult celiac disease. Brain 1966; 89:683–722.
22. Gobbi G, Bouquet F, Greco L, et al. Coelaic disease, epilepsy and cerebral calcifications. Lancet 1992; 340:439–43.
23. Hadjivassiliou M, Gibson A, Davies-Jones GAB, et al. Does cryptic gluten sensitivity play a part in neurological illness? Lancet 1996; 347:369–71.
24. Stolberg L, Rolfe R, Gitlin N, et al. D-Lactic acidosis due to abnormal gut flora: diagnosis and treatment of two cases. N Engl J Med 1982; 306:1344–8.
25. Tenenbein M, Wiseman NE. Early coma in intussusception: endogenous opioid induced? Pediatr Emerg Care 1987; 3:22–3.
26. Goetting MG, Tiznado-Garcia E, Bakdash TF. Intussusception encephalopathy: an underrecognised cause of coma in children. Pediatr Neurol 1990; 6:419–21.
27. Branski D, Shatsberg G, Gross-Kieselstein E, et al. Neurological dysfunction as a presentation of intussusception in an infant. J Clin Gastroenterol 1986; 8:604–5.
28. Basile AS, Jones EA. Ammonia and GABA-ergic neurotransmission: interrelated factors in the pathogenesis of hepatic encephalopathy. Hepatology 1997; 25:1303–5.
29. Yurdaydin C, Hortnagl H, Steindl P, et al. Increased serotoninergic and noradrenergic activity in hepatic encephalopathy in rats with thioacetamide-induced acute liver failure. Hepatology 1990; 11:371–8.
30. Butterworth RF. Complications of cirrhosis III hepatic encephalopathy. J Hepatol 2000; 32(suppl 1):171–80.
31. Albrecht J, Jones EA. Hepatic encephalopathy: molecular mechanisms underlying the clinical syndrome. J Neurol Sci 1999; 30:138–46.
32. Shattock P, Kennedy A, Rowell F, et al. Role of neuropeptides in autism and their relationships with classical neurotransmitters. Brain Dysfunction 1991; 3:328–325.
33. Du Verglas G, Banks SR, Guyer KE. Clinical effects of fenfluramine on children with autism. Review of the research. J Autism Dev Disord 1988; 18:297–308.
34. Walker-Smith JA, Davies SE, Murch SH, et al. Ileo-caecal lymphoid nodular hyperplasia, non-specific ileo-colitis with regressive behavioural disorder and food intolerance: a case study. J Pediatr Gastroenterol Nutr 1998; 25(suppl 1):548.
35. Bolte ER. Autism and Clostridium tetani.
Med Hypotheses 1998; 51:133–44.
36. Sandler RH, Finegold SM, Bolte ER, et al. Short-term benefit from oral vancomycin treatment of regressive-onset autism. J Child Neurol 2000; 15:429–35.