Correspondence
Secretin Treatment for Autism
N Engl J Med 2000; 342:1216-1218April 20, 2000
- Article
To the Editor:
My colleagues and I began evaluating children with autism and gastrointestinal problems in 1997 and reported our findings initially in 1998.1 The main secretin-related observation was a significant increase in the volume of pancreaticobiliary fluid after intravenous injection of secretin during endoscopy in 75 percent of the 36 autistic children we studied.2 Most of the children had gastrointestinal changes after a single dose of secretin. The majority of these patients had gradual improvements in social and behavioral skills after repeated injections. Of course, to prove scientifically that secretin is beneficial in the treatment of autism, randomized, double-blind, placebo-controlled studies should be conducted.
The lack of benefit from a single injection of secretin in the study by Sandler et al. (Dec. 9 issue)3 is not surprising. It is unusual for a single dose of a drug to result in full recovery from a chronic disease. In addition, the age range, diagnosis, and absence of serious gastrointestinal problems in the study sample were all predictive factors for a negative result. We studied younger, nonverbal patients with autism who had gastrointestinal symptoms and a low level of functioning. Many of our patients with diarrhea had an immediate improvement in stool consistency after a single injection of secretin; however, marked responses in terms of behavior were rare.
It was striking to read that, despite learning of the negative study results, 63 percent of the parents of children in the secretin group were interested in further treatment with secretin injections. In the early stage of our studies, we became aware of the poor sensitivity of the various diagnostic behavioral tests used in children with autism. These tests were designed to diagnose autism, not to assess drug-induced changes.
In our initial case studies, secretin was administered intravenously during endoscopy. After meals, several gastrointestinal hormones are released. One or more of these hormones may antagonize the effects of secretin. Sandler et al. did not report whether their patients had fasted.
3 ReferencesKaroly Horvath, M.D., Ph.D.
University of Maryland at Baltimore, Baltimore, MD 21201-15951
Horvath K ,Stefanatos G ,Sokolski KN ,Wachtel R ,Nabors L ,Tildon JT . Improved social and language skills after secretin administration in patients with autistic spectrum disorders. J Assoc Acad Minor Phys 1998;9:9-15
Medline2
Horvath K ,Papadimitriou JC ,Rabsztyn A ,Drachenberg C ,Tildon JT . Gastrointestinal abnormalities in children with autistic disorder. J Pediatr 1999;135:559-563
CrossRef | Web of Science | Medline3
Sandler AD ,Sutton KA ,DeWeese J ,Girardi MA ,Sheppard V ,Bodfish JW . Lack of benefit of a single dose of synthetic human secretin in the treatment of autism and pervasive developmental disorder. N Engl J Med 1999;341:1801-1806
Full Text | Web of Science | Medline
To the Editor:
As the father of an adult son with autism, I commend Sandler et al. for their report on the administration of secretin in children with autism and related disorders. I am impressed with their straightforward acknowledgment of several major limitations of their study and their candor in revealing that many parents continued to be interested in secretin after being told of the negative study results. A number of parents whose children participated in the study have informed me that their children have continued to receive secretin, with noteworthy improvement.
In sharp contrast to the appropriately temperate position of Sandler et al., Volkmar,1 in an editorial accompanying the report, displays an unseemly eagerness to discard secretin. He overlooks a major deficiency in the study: the lack of instruments designed to evaluate the effectiveness of treatments for autism. Volkmar is coeditor of the Handbook of Autism and Pervasive Developmental Disorders, 2 which includes the statement that diagnostic checklists, such as those used by Sandler et al., are not sensitive enough “to measure change in response to treatment.” Sandler et al. acknowledge this limitation, but in his editorial, Volkmar does not.
We are in the top of the first inning — it is much too early to reach a negative conclusion. There is a great deal to learn, and a great deal to be gained, from the continued intensive study of the effects of secretin on autism.
2 ReferencesBernard Rimland, Ph.D.
Autism Research Institute, San Diego, CA 921161
Volkmar FR . Lessons from secretin. N Engl J Med 1999;341:1842-1844
Full Text | Web of Science | Medline2
Lord C. Diagnostic instruments in autism spectrum disorders. In: Cohen DJ, Volkmar FR, eds. Handbook of autism and pervasive developmental disorders. 2nd ed. New York: Wiley, 1997:460-83.
To the Editor:
The study by Sandler et al. has numerous critical limitations. As the authors note, only a single dose of secretin was evaluated over a period of one month, an unrealistically short assessment period for a chronic disease. A second limitation was the divergent diagnoses and variations in the severity of disease. Fully 33 percent of the children did not meet the diagnostic criteria for autism, and the variance in severity was statistically equivalent to the full range of functioning found in autism.
A third limitation was the use of the Autism Behavior Checklist1 as a primary method of assessing changes in autistic symptoms. This checklist was designed as a screening tool to distinguish children with autism from those with mental retardation or emotional disturbances. The checklist consists of 57 symptoms, which are graded as present or not present. Each symptom is assigned a weight of one to four in the calculation of the overall score. For example, “Does not follow simple commands given once (e.g., sit down)” has a weight of one, whereas “Speech is atonal and arrhythmic” has a weight of four. Atonal speech is more heavily weighted because it better distinguishes autism from other disabilities. Such weighting is not justified when assessing change, particularly in the context of a study in which compliance with commands is an anecdotally reported benefit in some children and a change in speech patterns is not.
In addition, the Autism Behavior Checklist assesses various levels of symptoms, including many that are appropriate to evaluate only in persons with a relatively high level of functioning. “Has pronoun reversals (e.g., you for I)” is an inappropriate assessment for a nonverbal child. Of the 13 items in the language section, only 2 assess the emergence of expressive language, which has been anecdotally reported in some children treated with secretin. Furthermore, several items on the checklist involve historical facts, which by definition are unresponsive to drug treatment.
Given the limitations of the single dose, the heterogeneity of the patients, and the problems with the primary assessment tool, is it any surprise that many parents and Dr. Sandler himself 2 continue to express an interest in the evaluation of multiple doses of secretin in patients with autism?
It should be noted that Repligen, of which I am the president and chief executive officer, is developing secretin as a treatment for autism.
2 ReferencesWalter C. Herlihy, Ph.D.
Repligen, Needham, MA 02494To the Editor:
We find it difficult to ignore the numerous, albeit mostly anecdotal findings of other investigators, who noted dramatic improvement after a single injection of secretin. Although the apparent improvement could indeed have resulted from a placebo effect, we believe the discrepancy between the negative results reported by Sandler et al. and the positive results reported by others may be due to the fact that two different preparations were used. The positive effects were observed with the use of secretin isolated from hog intestines, whereas the negative effects were reported with the use of a synthetic preparation of human secretin.
It is conceivable, though unlikely, that the small difference between the sequences of human and porcine secretin (a difference in 2 of 27 amino acid residues) accounts for the strikingly different results. It is possible that the natural material contained an active “impurity” — for instance, another peptide, most likely vasoactive intestinal peptide. This peptide is closely related to secretin, and the two are present together in the same peptide fractions during the process of isolation from intestinal extracts.1 Vasoactive intestinal peptide and its receptors are widely distributed in many organs, especially the brain, where the actions of the peptide include the protection of neuronal cells against a variety of insults.2
The possibility that the synthetic preparation was not intact secretin should also be considered. The amino acid sequence of secretin is conducive to the conversion of the two aspartyl residues to aminosuccinyl residues, with the possible subsequent formation of inactive β-aspartyl peptides.3 Investigators experienced in the synthesis of secretin have reported the occurrence of this process during its synthesis and purification or storage.4 Sandler et al. did not characterize the synthetic material with respect to its homogeneity, its amino acid composition and sequence, its conformation, or its biologic activity on the basis of the stimulation of pancreatic bicarbonate secretion.
4 ReferencesSami I. Said, M.D.
State University of New York at Stony Brook, Stony Brook, NY 11794-8172Miklos Bodanszky, Ph.D.
Case Western Reserve University, Cleveland, OH 441061
Said SI ,Mutt V . Polypeptide with broad biological activity: isolation from small intestine. Science 1970;169:1217-1218
CrossRef | Web of Science | Medline2
Said SI . Molecules that protect: the defense of neurons and other cells. J Clin Invest 1996;97:2163-2164
CrossRef | Web of Science | Medline3
Bodanszky M ,Kwei JZ . Side reactions in peptide synthesis. VII. Sequence dependence in the formation of aminosuccinyl derivatives from β-benzyl-aspartyl peptides. Int J Pept Protein Res 1978;12:69-74
CrossRef | Medline4
Ondetti MA ,Deer A ,Sheehan JT ,Pluscec J ,Kocy O . Side reactions in the synthesis of peptides containing the aspartylglycyl sequence. Biochemistry 1968;7:4069-4075
CrossRef | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: My colleagues and I have been surprised and gratified by the great interest sparked by our negative study of secretin for the treatment of autism. We realize that ours is only one study, and we eagerly await the results of other trials. Despite the limitations, we believe our methods and findings are robust.
Dr. Horvath dismisses our finding that treatment with secretin has no benefit because we used a single dose. There is much anecdotal evidence that children have had responses to a single dose. In fact, Parker Beck, Dr. Horvath's patient who has received such wide publicity, had a dramatic response to the first dose. We agree that it would be unrealistic to expect a “full recovery from a chronic disease” after treatment with a single dose; we evaluated our patients for any improvement in symptoms. One does not have to look beyond the use of a single dose of insulin in diabetes or a bronchodilator in asthma to find examples of improvements in the symptoms of a chronic disease. Other controlled studies currently under way are evaluating the use of multiple doses.
Dr. Horvath also points to the age range, diagnosis, and absence of gastrointestinal problems in our sample. Our patients ranged in age from 3 to 14 years, and all of them had autism or pervasive developmental disorder, not otherwise specified. There was a wide range in level of functioning and severity, and some children had a low level of functioning and were nonverbal. Among the 30 children who received secretin, 7 had a history of gastrointestinal symptoms. We found that the presence of such symptoms was not related to the patients' response to secretin. Unlike Dr. Horvath, we did not select children with specific symptoms for our study, and our results are therefore not subject to a referral bias, which would limit the generalizability of the findings.
Drs. Said and Bodanszky raise questions about the synthetic preparation of human secretin. We did not include independent assays of the secretin we used. Previous studies of this product have shown that its biologic activity is identical to that of porcine secretin. Other investigators are using both human and porcine secretin.
Drs. Horvath, Rimland, and Herlihy all point out problems with the measures we used to monitor treatment effects. Although we acknowledged these measurement issues in our article, we maintain that the Autism Behavior Checklist and the Clinical Global Impression Scale are sensitive to changes in a broad array of core autistic symptoms and associated behavior, and both have been validated for this purpose. In fact, our results confirm that these measures were sensitive to changes in our sample. There was significant improvement in these measures over time (although there were no significant differences between the secretin group and the placebo group).
Adrian Sandler, M.D.
Thoms Rehabilitation Hospital, Asheville, NC 28803Author/Editor Response
Horvath notes several potential limitations of the study by Sandler et al. It is of course possible that repeated infusions are beneficial or that the presence of gastrointestinal difficulties, fasting status, or both affect the response to the drug; these possibilities await further research. We all agree that more sophisticated instruments for measuring changes in response to treatment for autism are needed; however, the instruments currently available have been used to demonstrate significant improvement in studies of other agents.1
The pessimism that Rimland senses in my editorial reflects my awareness that this is the second negative study of secretin.2 Both negative studies were conducted after the initial claims of marked improvement and the subsequent frenzy surrounding this agent. Horvath's account of the improvement he has observed is welcome; the importance of his observation of improved stool consistency with regard to the child's development and behavior remains to be established.
Both Horvath and Rimland comment on the interest of parents in continuing a treatment that has not been demonstrated to be effective. In the best of all possible worlds, we would indeed be able to judge the scientific merit of a treatment by its popularity; unfortunately, this is not the case. As I noted, there may be important but nonspecific gains as a result of the participation in research and the heightened expectations of parents, and such gains may encourage parents to seek continued treatment. I also noted the limitations of the study by Sandler et al. and the need, as with all negative studies, for additional research. One of the most important lessons to be learned from secretin may indeed be the importance of encouraging substantive research on treatments for autism.
2 ReferencesFred Volkmar, M.D.
Yale University School of Medicine, New Haven, CT 065201
McDougle CJ. Psychopharmacology. In: Volkmar FR, ed. Autism and pervasive developmental disorders. Cambridge, England: Cambridge University Press, 1998:169-94.
2
Owley T, Steele E, Corsello C, et al. A double-blind, placebo-controlled trial of secretin for the treatment of autistic disorder. Medscape General Medicine 1999. (See: http://www.medscape.com/medscape/generalmedicine/journal/1999/v01.n10/mgm1006.owle-01.html.)
- Citing Articles (1)
Citing Articles
1
Barbara E. Esch, James E. Carr. (2004) Secretin as a Treatment for Autism: A Review of the Evidence. Journal of Autism and Developmental Disorders 34:5, 543-556
CrossRef
