Gastrointestinal abnormalities in children with autistic disorder

Journal of Pediatrics
Volume 135 • Number 5 • November 1999
Copyright 1999 Mosby, Inc.

Karoly Horvath MD, PhD
John C. Papadimitriou MD, PhD
Anna Rabsztyn
Cinthia Drachenberg MD
J. Tyson Tildon PhD

From the Departments of Pediatrics and Pathology, University of Maryland School of Medicine, Baltimore.
Supported by an intramural grant by the University of Maryland School of Medicine.
Submitted for publication Dec 31, 1998.
Revision received May 20, 1999.
Accepted July 21, 1999.
Reprint requests: Karoly Horvath, MD, PhD, Department of Pediatrics, 22 S Greene St, N5W70, Box 140, Baltimore, MD 21201-1595.

Copyright 1999 by Mosby, Inc.

0022-3476/99/$8.00 + 0  9/21/101636

Objectives: Our aim was to evaluate the structure and function of the upper gastrointestinal tract in a group of patients with autism who had gastrointestinal symptoms.
Study design: Thirty-six children (age: 5.7 2 years, mean SD) with autistic disorder underwent upper gastrointestinal endoscopy with biopsies, intestinal and pancreatic enzyme analyses, and bacterial and fungal cultures. The most frequent gastrointestinal complaints were chronic diarrhea, gaseousness, and abdominal discomfort and distension.
Results: Histologic examination in these 36 children revealed grade I or II reflux esophagitis in 25 (69.4%), chronic gastritis in 15, and chronic duodenitis in 24. The number of Paneth's cells in the duodenal crypts was significantly elevated in autistic children compared with non-autistic control subjects. Low intestinal carbohydrate digestive enzyme activity was reported in 21 children (58.3%), although there was no abnormality found in pancreatic function. Seventy-five percent of the autistic children (27/36) had an increased pancreatico-biliary fluid output after intravenous secretin administration. Nineteen of the 21 patients with diarrhea had significantly higher fluid output than those without diarrhea.
Conclusions: Unrecognized gastrointestinal disorders, especially reflux esophagitis and disaccharide malabsorption, may contribute to the behavioral problems of the non-verbal autistic patients. The observed increase in pancreatico-biliary secretion after secretin infusion suggests an upregulation of secretin receptors in the pancreas and liver. Further studies are required to determine the possible association between the brain and gastrointestinal dysfunctions in children with autistic disorder. (J Pediatr 1999;135:559-63)

See editorial, p. 533.

Autistic disorder belongs to the group of pervasive developmental disorders as defined by both the American (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition)[1] and international (International Classification of Diseases, Ninth Revision) diagnostic systems. Autistic disorder implies severity of disturbance in multiple areas of development reflected in a marked lack of development of social interaction and communication; restricted, repetitive, and stereotyped patterns; and typical prelinguistic communicative behaviors. In addition to the abnormalities in communication and language skills, these children frequently have aggressive and self-injurious behaviors. Sudden unexplained irritability or aggressive behavior, nighttime awakening, and pushing on the abdomen are usually considered part of the behavioral problems associated with autism.

Many parents report gastrointestinal symptoms in their autistic child; however, until recently, gastrointestinal symptoms of these children received little attention. In 1996, D'Eufemia et al[2] reported increased intestinal permeability in 9 of 21 (43%) patients with autistic disorder. The report of Wakefield et al[3] represents the first effort to evaluate the gastrointestinal tract in children with autism. In a recent case report we described 3 children with autistic spectrum disorder and chronic diarrhea who had an increased pancreatico-biliary secretory response after secretin injection, suggesting that gastrointestinal dysfunction might be associated with this pervasive developmental disorder.[4]

This report describes several gastrointestinal abnormalities in low-functioning autistic children and underlines the importance of comprehensive gastrointestinal evaluations.


Children diagnosed with autistic disorder or pervasive developmental disorder, not otherwise specified by professionals with expertise in behavioral pediatrics were referred to our gastroenterology clinic for further evaluation.

Thirty-six children with one or more of the following symptoms including abdominal pain (n = 25), chronic diarrhea (n = 21), gaseousness/bloating (n = 21), nighttime awakening (n = 15), and unexplained irritability (n = 18) underwent esophagogastroduodenoscopy. The mean age of these patients was 5.7 years (5.7 2 years, mean SD; range, 2.5-10 years; 33 boys).

Their medical history revealed that 64% of these children were breast-fed for an average of 6.7 months, 36% had a history of cow's milk and/or soy protein intolerance, and 44.4% had allergy to foods according to the parents and care givers interviewed at the outpatient clinic. Seventeen (47.2%) children were on a casein-free and/or gluten-free diet at the time of evaluation.

Clinical Investigations

Patients fasted after midnight, and the procedures were performed the next morning with patients under general anesthesia. The full upper gastrointestinal workup included esophageal, gastric, and duodenal biopsies for histology, measurement of the digestive enzymes of the small intestine (lactase, maltase, sucrase, palatinase, glucoamylase) and pancreas (lipase, amylase, trypsin, chymotrypsin, and carboxypeptidases A and B), and bacterial and fungal cultures. The normal values for intestinal enzymes were based on the measurements on 104 histologically normal intestinal biopsy tissues. The normal values for pancreatic enzymes were established by measurement on 215 specimens collected from children. Antibiotic and antifungal therapies were discontinued a week before the endoscopy to allow culture of the duodenal fluid. All fluids for cultures were obtained before pancreatic stimulation to avoid the dilution of juice resulting in falsely low quantitative counts. The pancreatic enzyme activities were measured from specimens collected before and after secretin stimulation.

For the collection of pancreatic juice, we stimulated the fluid secretion with secretin (Ferring Laboratories, Inc, Suffern, NY), 2 cat units (CU)/kg body weight, given intravenously within 1 minute. The pancreatico-biliary juice was collected after positioning the endoscope distal to the ampulla of Vater, and the fluid was collected by moving the tip into the outcoming fluid and suctioning it into a collector trap. In a few cases an endoscopic retrograde cholangiopancreatography catheter was placed into the channel of the endoscope, and fluid was suctioned into a syringe.

A basal sample was collected in the duodenum around the Vater papilla before the secretin injection and was sent for enzyme analysis and bacterial and fungal cultures. Three additional specimens were collected after the secretin injection within a 5- to 10-minute period. The pancreatic enzyme activities were measured in all collected aliquots. The volume of secreted fluid after secretin administration was measured and recorded in milliliters per minute. The normal fluid output was based on the data from 26 non-autistic patients who underwent the same procedure. The duodenal biopsy specimens were obtained after the fluid collections to avoid blood contamination. Endoscopic grading of esophagitis was based on the description of Leape at al[5] for pediatric patients.


The histologic grading of esophagitis followed the scores described by the Working Group on Gastro-Oesophageal Reflux Disease of the European Society of Paediatric Gastroenterology and Nutrition.[6] The histologic criteria for reflux esophagitis included eosinophilic and/or polymorphonuclear infiltrate, basal layer thickening, and papillary hypertrophy.

Gastric biopsy specimens were stained with Giemsa stain to determine the presence of Helicobacter pylori infection.

The histology slides were examined by surgical pathologists and were further reviewed by 3 authors (J.P., C.D., and K.H.) in an observer-blinded fashion. In the intestinal specimens, the number of intraepithelial lymphocytes, lamina propria cell density, villus/crypt ratio, and mitoses in crypts were assessed. The number of Paneth's cells per crypt was counted in all 36 patients with autism and compared with 22 biopsy specimens from non-autistic pediatric patients (12 immunocompetent and 10 immunodeficient) who underwent endoscopy because of failure to thrive, chronic diarrhea, or suspected celiac disease.

Ethical Approval and Consent

The project for the examination of the effect of secretin injection was approved by the Internal Review Board of University of Maryland School of Medicine. Informed written consent was obtained from each parent before endoscopy and secretin infusion.

Statistical Analysis

Data were expressed as mean SD. In all comparisons between groups the Student t test was used, and a P value of <.05 was considered significant.


Histologic Findings

The most frequent histologic finding was the presence of reflux esophagitis in 25 of 36 children (69.4%). Twenty-two of these 25 children (88%) had symptoms such as nighttime awakening with irritability, signs of abdominal discomfort, or pushing on the abdomen, which are typically reported by non-autistic children with esophagitis. Chronic inflammation of the gastric mucosa was present in 15 children. None of the patients had H pylori infection. Chronic nonspecific duodenal inflammation was found in 24 children (66.6%). Two children had grade II partial villus atrophy, but they did not have serologic or histologic evidence of celiac disease (on gamma/delta lymphocyte staining).

Paneth's cell hyperplasia was evident in the duodenal biopsy specimens. We performed a morphometric analysis and compared the number of Paneth's cells seen in the crypts with those of 22 non-autistic control subjects and found an elevated number of Paneth's cells per crypt (3.09 0.46 vs 2.07 0.32; P < .05). Furthermore, the Paneth's cells were frequently enlarged, and discharge of granules into the crypt lumen was a typical finding. There was no difference in the number of cells between patients with and those without diarrhea (3.04 0.53 vs 3.16 0.34 cells/crypt). Fig 1 shows the number of Paneth's cells counted in the studied patients and the 2 groups of control subjects.

Fig. 1. Paneth's cell counts in children with autistic disorder, normal control subjects, and immunodeficient patients (HIV). PDD, Pervasive developmental disorder; HIV, human immunodeficiency virus.

Enzyme Assays

Decreased activity (<1 SD below normal values) of one or more disaccharidases or glucoamylase was found in 21 children (58.3%); 10 children had decreased activity in 2 or more enzymes. The most frequent finding was a low lactase level, which was present in 14 patients (<9.4 IU/g/min). All of the 21 children with low enzyme activities had loose stools and/or gaseousness. None of these 36 children had pathologic pancreatic enzyme activities after secretin stimulation.

Pancreatico-Biliary Fluid Output for Secretin

Fig 2 shows the secretory responses in the 2 groups of children with autistic disorder and control subjects.

Fig. 2. Pancreatico-biliary fluid output after administration of secretin, 2 CU/kg body weight, in children with autistic disorder with and without diarrhea and in control subjects.
The average pancreatico-biliary fluid output was significantly higher (3.8 2.2 mL/min) for the autistic group compared with the control group (1.46 0.57 mL/min; P < .05). In 27 children (75%) the volume of pancreatico-biliary fluid output after secretin stimulation was 1 SD above the values of non-autistic patients. Nineteen of the 21 patients (90.47%) with the main symptom of chronic diarrhea had significantly higher fluid output compared with that of control subjects. There was a statistically significant difference between patients with diarrhea (n = 21) and those without diarrhea (n = 15) (4.8 2.3 mL/min vs 2.4 1.3 mL/min; P < .05). An important clinical observation was that autistic children with chronic diarrhea showing the high fluid response with secretin had an improved stool consistency after the procedure, and it lasted for a few weeks or was sustained.

Duodenal Fluid Culture

There was no evidence of either fungal or bacterial overgrowth in the duodenum, even in those 12 patients who, on the basis of urine organic acid test results, were suspected of having such overgrowth.


Few studies have addressed gastrointestinal problems in children with autistic disorder. Goodwin et al[7] studied 15 randomly selected children with autism and found that 6 had either bulky, odorous, or loose stools or intermittent diarrhea; one had celiac disease. In a recentstudy, 43% of the autistic patients without symptoms or evidence of any gastrointestinal disease had altered intestinal permeability.[2] Low concentrations of serum alpha1 -antitrypsin were reported in children with typical autism,[8] a finding that is indicative of intestinal protein loss. In a recent case report we presented gastrointestinal and behavioral observations on 3 children with autistic spectrum disorder.[4]

Although gastrointestinal symptoms frequently accompany the manifestations of autism, little attention has been paid to this aspect of this developmental behavioral disorder, and a gastrointestinal workup has not been part of the regular medical evaluations. Sudden unexplained irritability or aggressive behavior, mood change, discomfort, and nighttime awakenings in these children were considered to be part of the brain dysfunction and not manifestations of organic problems. A significant percentage of children with autistic disorder are reported to be low functioning and have only prelinquistic communicative behavior. A plausible reason for the paucity of gastrointestinal evaluation of these children may be their inability to verbalize and describe their abdominal pain or discomfort and a lack of cooperation in non-invasive studies, such as breath tests.

The upper gastrointestinal evaluations of children with autistic disorder support the presence of a chronic inflammatory process in the gut, as reported by Wakefield et al.[3] They performedcolonoscopy with histologic examinations in 12 children and reported that all had intestinal abnormalities, ranging from lymphoid nodular hyperplasia to aphthoid ulceration. [3] In our study chronic inflammation of the esophagus, stomach, and duodenum was the major and most consistent finding.

The most frequently detected abnormalities in children with autistic disorder included a high prevalence of reflux esophagitis, hyperplasia of duodenal Paneth's cells, intestinal carbohydrate digestive enzyme deficiencies, and an unusual hypersecretory response to intravenous secretin administration.

A significant portion (25/36) of autistic children had gastroesophageal reflux and reflux esophagitis. There are no age-related data on the prevalence of gastroesophageal reflux disease in the 2.5- to 10-year-old group. The prevalence of reflux esophagitis is low (estimated 2%) in Western countries.[9] It is known that both neural and humoral factors can have an effect on the lower esophageal sphincter. People under stress are more likely to have dysmotility and reflux. It is known that secretin has a suppressive effect on gastric secretion. [10] Whether a low secretin level may contribute to the high prevalence of acidic reflux in these children warrants further investigation.

An elevation in the number of Paneth's cells was found in most of the studied patients. Reportedly, the average count of Paneth's cells in the crypts is about 5%, and there is no significant difference in the number of these cells between the healthy and inflamed duodenum[11] or in celiac disease.[12] Scott and Brandtzaeg[12] reported that the average number of Paneth's cells in healthy control subjects was 2 per crypt, which is similar to that of our control subjects. We observed an ~50% increase in the number of these cells in patients with autistic disorder. In normal crypt base cross-sections, the size of Paneth's cells is similar to that of the surrounding cells, and they are smaller than the goblet cells. Many patients with autistic disorder had enlarged Paneth's cells filled with huge granules. Hypertrophy and hyperplasia of the Paneth's cells have been reported in hamsters after ligation of pancreatic ducts,[13] and a varying degree of increase in the number of Paneth's cells was described in patients with chronic pancreatitis.[14] It is thought that absence of pancreatic fluid favors the multiplication of Paneth's cells. We did not find evidence of pancreatic insufficiency in our patients. However, the high secretory response to secretin might suggest the absence of regular secretin stimulation of the pancreas and biliary tract. There are no data available regarding the effect of secretin or cholecystokinin on Paneth's cells or local immune defense of the intestine. Studies indicate that Paneth's cells produce and release substances such as lysozyme, defensins,[15] and alpha1 -antitrypsin.[16] The human intestinal defensin 5 (HD-5) has antimicrobial activities against bacteria and Candida albicans.[17] Paneth's cell metaplasia is a typical finding in inflammatory bowel diseases involving the colon.[18] Transmission electron microscopic studies in tissues from patients with Crohn's disease showed that Paneth's cells were increased in number and showed both focal granule extrusion and cytoplasmic lysosomal inclusions.[19] Our studies revealed similar findings, including a greater number of Paneth's cells and increased granule discharge into the lumen of crypts. However, although the presence of Paneth's cells represents a metaplastic change in Crohn's disease, these cells are normally present in the duodenum. The importance of this Paneth's cell activation in patients with autistic disorder is not known and warrants further investigation.

Children with autistic disorder frequently have loose, extremely foul-smelling stools and gaseousness. These symptoms are not associated with growth failure and cannot be explained by the limited diet preference of these children. In most of the cases, results of the routine stool tests are negative. Typically, parents claim that the gastrointestinal and behavioral symptoms are manifested in parallel. Our study showed that 58% of the examined children had disaccharidase/glucoamylase enzyme activities below the normal range. Carbohydrate malabsorption may result in gaseousness with crampy abdominal pain and may be the cause of chronic loose stools. The most frequent finding was a low lactase activity in 14 of the 21 children with pathologic disaccharidase results.

Children with autistic disorder and chronic diarrhea had a higher pancreatico-biliary fluid output after secretin stimulation. This high secretory response to secretin administration may indicate an upregulation of the secretin receptors in the ductal cells of the pancreas or in the bile-duct epithelium. In turn, this may occur in the absence of normal secretin stimulation, which can be the consequence of either a defect in secretin production or a problem of release from the intestinal S cells.

In summary, our findings suggest that gastrointestinal abnormalities may contribute to some of the behavioral problems frequently described in these children. The presence of esophagitis correlated well with the reported symptoms and may in part explain the sudden irritability, aggressive behavior, or nighttime awakenings in many of these children. The diarrhea and gaseousness may be the consequence of decreased disaccharidase activity and may also contribute to the behavioral problems.

Our results suggest a possible upregulation of secretin receptors in the pancreas and the biliary tract and a high prevalence of reflux esophagitis and chronic duodenitis associated with Paneth's cell hyperplasia in a group of children with autistic disorder who have various gastrointestinal symptoms. Further gastrointestinal studies of children with autistic disorder may contribute to a better understanding of the etiology of this disorder.

We thank Lisa Medeiros, BSN, and Tina Sewell, BSN, for their help with patient recruitment and during the procedures; their contribution was important to accomplish this work. We are very grateful to Lois Roeder, PhD, for her contribution in the preparation of the manuscript.

1. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington (DC): American Psychiatric Association; 1994.  
2. D'Eufemia P, Celli M, Finocchiaro R, Pacifico L, Viozzi L, Zaccagnini M, et al. Abnormal intestinal permeability in children with autism. Acta Paediatr 1996;85:1076-9.  abstract
3. Wakefield AJ, Murch SH, Anthony A, Linnell J, Casson DM, Malik M, et al. Ileal-lymphoid-nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children. Lancet 1998;351:637-41. abstract
4. Horvath K, Stefanatos G, Sokolski K, Wachtel R, Nabors L, Tildon JT. Improved social and language skills after secretin administration in patients with autistic spectrum disorders. J Assoc Acad Minority Physicians 1998;9:9-15.  
5. Leape LL, Bhan I, Ramenofsky ML. Esophageal biopsy in the diagnosis of reflux esophagitis. J Pediatr Surg 1981;16:379-84.  abstract
6. Vandenplas Y. Reflux esophagitis in infants and children: a report from the Working Group on Gastro-Oesophageal Reflux Disease of the European Society of Paediatric Gastroenterology and Nutrition. J Pediatr Gastroenterol Nutr 1994;18:413-22.  abstract
7. Goodwin MS, Cowen MA, Goodwin TC. Malabsorption and cerebral dysfunction: a multivariate and comparative study of autistic children. J Autism Child Schizophr 1971;1:48-62.  citation
8. Walker-Smith J, Andrews J. Alpha-1-antitrypsin, autism, and coeliac disease. Lancet 1972;2:883-4. citation
9. Wienbeck M, Barnert J. Epidemiology of reflux disease and reflux esophagitis. Scand J Gastroenterol Suppl 1989;156:7-13.  abstract
10. Chung I, Li P, Lee K, Chang T, Chey WY. Dual inhibitory mechanism of secretin action on acid secretion in totally isolated, vascularly perfused rat stomach. Gastroenterology 1994;107: 1751-8.  abstract
11. Schmitz-Moormann P, Schmidt-Slordahl R, Peter JH, Massarrat S. Morphometric studies of normal and inflamed duodenal mucosa. Pathol Res Pract 1980;167:313-21.  abstract
12. Scott H, Brandtzaeg P. Enumeration of Paneth cells in coeliac disease: comparison of conventional light microscopy and immunofluorescence staining for lysozyme. Gut 1981;22: 812-6.  abstract
13. Balas D, Senegas-Balas F, Bertrand C, Frexinos J, Ribet A. Effects of pancreatic duct ligation on the hamster intestinal mucosa. Histological findings. Digestion 1980;20:157-67.  abstract
14. Senegas-Balas F, Bastie MJ, Balas D, Escourrou J, Bommelaer G, Bertrand C, et al. Histological variations of the duodenal mucosa in chronic human pancreatitis. Dig Dis Sci 1982;27:917-22.  abstract
15. Ouellette AJ. Paneth cells and innate immunity in the crypt microenvironment. Gastroenterology 1997;113: 1779-84.  abstract
16. Molmenti EP, Perlmutter DH, Rubin DC. Cell-specific expression of alpha 1-antitrypsin in human intestinal epithelium. J Clin Invest 1993;92:2022-34.  abstract
17. Porter EM, van Dam E, Valore EV, Ganz T. Broad-spectrum antimicrobial activity of human intestinal defensin 5. Infect Immun 1997;65:2396-401.  abstract
18. Stamp GW, Poulsom R, Chung LP, Keshav S, Jeffery RE, Longcroft JA, et al. Lysozyme gene expression in inflammatory bowel disease. Gastroenterology 1992;103:532-8.  abstract
19. Dvorak AM, Dickersin GR. Crohn's disease: transmission electron microscopic studies. I. Barrier function. Possible changes related to alterations of cell coat, mucous coat, epithelial cells, and Paneth cells. Hum Pathol 1980; 11(suppl 5):561-71.  abstract