Disclaimer: The author of this article is not a medical professional, and none of what follows is intended to be authoritative information. The intent of this article is to provide a general approximation of modern medical knowledge on the subject of the causes of autism, simplified for ease of understanding. Since much of modern medical knowledge on this subject is theory and speculation, nothing that follows can be considered absolute; since much of the following is simplified, this is doubly true.
The following is a potentially dangerous subject, one where the majority of my readers may have already formed their opinions before coming here. It’s a subject that naturally generates strong opinions, and this is entirely understandable: Childhood autism diagnoses (and misdiagnoses) have risen drastically in the industrialized world over the past few decades, and yet as a general rule it’s difficult in any one case to point to any definitive cause.
One of the chief obstacles to the correct diagnosis, and indeed to common understanding, of autism spectrum disorders is the broadness of that very spectrum. While symptomatically there are vast differences between Asperger’s Syndrome and Heller’s Syndrome (to name but two varieties), and although prognosis and treatment vary hugely, both are classed the same in the current DSM, in clinical diagnoses, and in many studies. Likewise, due to a knowledge base that is expanding faster than continuing professional education can keep pace with, some individuals have been diagnosed with autism instead of a more precise (and distinct) disorder such as Rett Syndrome or cerebral palsy.
Given these as starting conditions, one might wonder what if anything can be absolutely determined about autism, especially with regard to such questions as whether childhood vaccines can be said to be a contributory or even a primary cause. This article is intended as a brief summary, written (insofar as is practicable) in layman’s terms by a layman, of the state of current knowledge on the subject.
NOTE: Multiple studies have been referred to in the composition of this, but due to their technical nature it is presumed that the average reader will have no desire to peruse them. Moreover, the majority of them are not available except by subscription to the journals in which they were published. Inline links to articles in the public domain have been provided, and a truncated informal bibliography has been appended to the end of this document. Anyone wishing to read the full articles can access them easily, using these citations, either at a public university or in the medical libraries of most hospitals; both are usually open to the public. Standard medical references such as the DSM are commonly available; references in this article are to versions IV and V, the two most recent. Descriptions are drawn from the DSM and from ICD-9 as appropriate.
An Autism Spectrum Disorder is defined by the DSM as causing “persistent deficits in social communication and social interaction.” It’s a broad catch-all diagnosis for otherwise unexplained symptoms; notably, once a cause and etiology have been differentiated for a specific syndrome, it is removed from the DSM forthwith as a medical rather than psychiatric disorder (for example, Rett Syndrome et cetera). Presently, the DSM includes the following broad categories:
Autistic Disorder: This most common variant (up to 0.6% of the general population) is characterized by a triad of symptoms, to wit: Impaired communication, impaired social interaction (intuition), and repetitive behavior / restricted interests. Historically, it has been described since the middle ages; modern systematic research began in the early 1900s.
Asperger Syndrome: Often considered a subset of general autistic disorder, Asperger’s is commonly differentiated by an absence of communication disability; nevertheless, atypical development is common. Historically, it has been described as distinct since the earliest systematic research, but the validity of this distinction is frequently questioned, as under the current DSM standard.
Childhood Disintegrative Disorder: This is an extremely rare presentation of autism; it affects fewer than two per 100,000 (also known as Heller’s Syndrome). It is characterized by a spontaneous regression at a later age after several skills have been achieved, often with full awareness of the child. Notably, while this has no single known (determined) cause, it is frequently associated with other conditions and causal relationships are being studied.
Rett Syndrome: This is similar to autism in symptoms and diagnosis but has a recently-discovered distinct demonstrated cause and has been specifically separated in diagnostic manuals — simply put, it’s no longer considered autism because the cause is known (1). Some schools of thought suggest that all autism-spectrum syndromes will one day have similar descriptions and will at that time be described similarly; hence its inclusion in this list. In terms of projected causality, it is useful for comparison.
Pervasive Developmental Disorder: In short, this is another catch-all, and it refers to atypical presentations of autistic disorder and Asperger syndrome. It is no longer differentiated in the DSM, but in other diagnostic guides and systems it exists as a label. Due to the atypical presentation and low incidence, even less is known about this.
While there’s a lot that is not known, a great deal has been determined with a very high degree of certainty. Generally speaking, theoretical causes and observed direct relationships can be divided between Autistic Disorder (including Asperger’s and PDD) and Childhood Disintegrative Disorder. They will be listed separately where practicable as the two disorders are distinct in presentation and therefore (likely) cause.
While specific causality is difficult to link, all examined confirmed cases result (through multiple pathophysiologies) from maturation-related changes in various brain systems (Singh (3)). No single cause has been implicated in all cases; however, given the variety in presentation, that’s hardly surprising. The most common correlations are described below.
Fragile X syndrome: Caused by a genetic mutation which apparently restricts production of something called FMRP in cell replication, fully half of Fragile X diagnoses are comorbid with autistic disorder and a causal relationship is presumed but impracticable to determine. Fragile X is a nonstandard inheritable condition. It is currently presumed by some researchers that anywhere from one in eight to one in two cases of autistic disorder are caused by this, but a great deal of further testing is required, especially as only half of Fragile X patients are diagnosably autistic. (In lay terms, it can safely be said that there’s a correlation in at least one in eight cases, though that is an oversimplification.)
Complex Chromosome Abnormality: In many cases, a variety of genetic abnormalities have been detected in autistic patients. There is no single “autism gene”; rather, a number of coextant heritable — but not necessarily inherited — mutations are often present, far more commonly with autism or similar syndromes than with the unaffected. Teratogen exposure, including active viral infection, (see Rubella below), pollutants, toxins (such as alcohol), and even stress, specifically during the first eight weeks of development, have some observed links, but studies are ongoing.
Rubella: A link between Rubella infection and autistic disorder in pregnant women during early development has recently been established (6). Studies are continuing, but it has been concluded that around 1% of cases are directly linked, and further that vaccination is preventative. Further viral relationships are also being studied, particularly measles and Zika; generally, correlations have been observed but that alone is insufficient to demonstrate causality.
Mitochondrial Dysfunction: There is a moderate (5%) observed correlation between autism and mitochondrial disease; however, since the mechanisms of mitochondrial dysfunction are as yet unknown, the relationship is difficult to qualify.
Maternal immunity: There is an observed relationship (Singh (2)) between certain maternal autoimmune markers and autism incidence; however, causality in the absence of complete developmental knowledge is impossible to determine. Moreover, the mechanism involved, if any, is unknown. This is most commonly observed in cases where the mother has suffered miscarriages and is coextant with other developmental disorders in siblings.
Diabetes: There has been a direct correlation observed between diabetes in a pregnant mother and a future autism diagnosis of the child. It is not known whether this is a causal or incident relationship, especially as diabetes is being researched now as an autoimmune disorder.
Psychiatric Drugs: There has been a direct correlation observed between the use of psychiatric drugs in a pregnant mother and a future autism diagnosis of the child. Studies are difficult due to the possible relationship between undiagnosed autism in the mother combined with comorbid psychiatric disorders; alternately, fetal toxicity with the drugs as teratogens is possible, even likely — but difficult to demonstrate.
Parental Age: Advanced age in either parent (late 30s and up) has been directly linked to an increased incidence of autism. It is presumed that this is due to a known increased mutation burden over time; anything from viral infection through environmental radiation, teratogen exposure, and cosmic ray events may be the cause, but there is certainly a demonstrable relationship. This would be causally linked to complex chromosome abnormalities (above) and maternal immunity (above).
It is worthy of mention that there is a common mechanism with several of the above causes (and some below) which seems to be autoimmune in nature (3). It is not exclusive to any one cause and is not easily consistently observed. Nevertheless, it seems likely that any medical preventative treatment will be in this area.
Childhood Disintegrative Disorder
Due to its rarity, studies of this condition are difficult. General medical thought is that no single underlying medical cause is responsible; however, CDD often occurs alongside these other conditions:
Epilepsy: There’s no way to tell which if either is causal, but CDD sufferers are likely to be epileptic. Observation of the degenerative impact of certain forms of epilepsy shows a logical chain indicating a causal relationship; however, analysis is problematic.
Lipid Storage Diseases: There are several of these; generally, a toxic buildup of fats occurs in the brain and general nervous system. There may well be a causal relationship.
Measles: A chronic measles infection can cause a form of encephalitis (6) leading to brain inflammation and the subsequent death of brain cells. While a correlation seems likely, a causal relationship is difficult to determine. (The difference is important.)
Tuberous sclerosis (TSC): Certain genetic abnormalities can cause benign (but potentially damaging) tumors to grow in the brain. It’s unknown whether the tumors cause the autism symptoms or, instead, autism is caused by different genetic abnormalities that occur in parallel.
Leukodystrophy: There are more than forty known types of leukodystrophy, any of which can cause the protective myelin sheath of the brain to disintegrate (or never properly form), leading to the dissolution of white matter. As with TSC (above), it’s difficult to determine whether the relationship is causal or merely coincident.
Speculation And Conclusions
Due to the broad spectrum of autism in presentation, symptomology, and above all in diagnosability, it is useless to search for any single cause. Indeed, in a strictly medical sense, it can safely be theorized that there are multiple varieties of autism-spectrum disorders or syndromes, each with a specific (yet thus far unknown) cause or set of causes; Rett Syndrome (above) is an example demonstrating the validity this school of thought. As a result, most modern causality and medical treatment research proceeds along these lines.
However, it must always be borne in mind that, absent a demonstrated cause and mechanism, any definitions, classification, and treatment will be not strictly medical but rather psychiatric in nature. This is a result of basic classification, and a vast number of misunderstandings can easily be generated by this difference.
While an oversimplification, for practical purposes Autistic Disorder can, generally, be said to have a definite link to genetic abnormalities; some of these are inherited, while others are of indeterminate origin. Teratogen and viral exposure during the first eight weeks of pregnancy are likewise linked. Presumed causes therefore include Fragile X, specific viral infections, and teratogen exposure at a specific time in pregnancy. Logically, therefore, since one’s genetic code cannot be completely changed after extremely early development, it can safely be stated that autism cannot be caused by any post-birth factor.
By contrast, however, Childhood Disintegrative Disorder may well proceed from a variety of causes, some few of which could be due to post-birth factors. It is important to note that CDD has been linked to several delayed-onset genetic disorders which are, by definition, certainly due to either a combination of inherited traits or mutation during early development. Likewise, as it is an extremely rare condition to begin with, any post-birth cause would be highly unlikely in general as well as nearly impossible to demonstrate.
In conclusion, it can safely (albeit oversimply) be stated that current medical wisdom with regard to the likely causes of Autism Disorder are:
- Inherited genetic disorder (such as Fragile X)
- Teratogen exposure
- Viral exposure
- High parental age
CDD, viewed as a distinct entity, may possibly have other causes. These are widely varied, but are almost all, as with AD, quite probably genetic in origin.
Addendum: Vaccine Speculation
Due to the recent controversy, it is necessary to state explicitly the following conclusion: No cases of Autism Disorder can possibly have been caused by post-birth or even late-pregnancy vaccination. Moreover, while it cannot be definitively stated that no case of Childhood Disintegrative Disorder is due to vaccination, several studies have been performed which demonstrate zero correlation of autism with thiomersal use in particular and vaccines in general.
Again, the following is an oversimplification, but it should serve as an understandable description: Due to the nature of the MMR vaccine (which contains live attenuated viruses), it is theoretically remotely possible that a measles infection could be caused (7) that, in turn, causes encephalitis; while no causal relationship has been demonstrated, it is arguable that this could in turn cause CDD. Several large studies have been undertaken to examine this relationship, and no correlation has been found (5), but since CDD is extremely rare, it is possible that a causal relationship on the order of approximately one in ten million could exist and still fall within the margin of error of some of the studies.
With regard to the above, it must be observed that a natural measles infection is far more likely for a nonimmunized child. Since the resulting likelihood of encephalitis is several orders of magnitude more likely (without considering potential mortality or other side effects), it seems evident that the MMR vaccine is far safer administered than not with specific regard to CDD (Schifle et al). (Specific odds are cited at 1:10,000 or less for panencephalitis with a measles infection (8). It’s usually fatal but even if not, compared to the 1:10,000,000 approximation quoted above for the vaccine, the vaccine is at least 300 times and as much as 1000 times less likely to cause the indicated condition even if there exists a causal link with vaccination.)
NOTE: It should also be noted that any anecdotal observation of CDD/vaccination relationships may easily fall prey to the logical fallacy known as “Post hoc, ergo propter hoc“; after, therefore because of. Simply, that two events occur in sequence in no way indicates that they must be related, especially causally.
With specific regard to Rubella: The long-term impact of this virus is primarily observed on pregnant women. It has no direct link to autism-spectrum disorders in infant or childhood cases, and in the context of this discussion its primary danger is as a teratogenic agent in early pregnancy — a correlation long since confirmed. Herd immunity is therefore of vital importance in the prevention of autism and other, often more serious, disorders.
It should be observed that immunization with MMR during a pregnancy is usually contraindicated in the extreme due to the above known relationships.
Finally, it should be mentioned that this article does not examine any relationship between vaccination and non-autism disorders. That there can be side effects is certain; it is my hope that these can be more thoroughly examined at a later time in this series of articles. Likewise, an examination of the philosophy of pervasive or required immunization should follow; discussion of such points should therefore properly occur there.
- “Rett syndrome is caused by mutations in…” Van den Veyver et al; Nature Genetics, 1999
- “Serological association of measles virus and human herpesvirus-6 with brain autoantibodies in autism.” Singh, VK et al; Journal of Clinical Immunology and Immunopathology, October 1998 — Public-access precis is available here courtesy NIH.
- “Phenotypic expression of autoimmune autistic disorder (AAD): a major subset of autism.” Singh, VK; Annals of Clinical Psychiatry, Sept. 2009
- “Measles-mumps-rubella vaccine and autistic spectrum disorder…”, Halsey, Hyman et al; Pediatrics May 2001
- “MMR vaccine and autism…”, Poland, Mayo Clinic Proceedings, Sept 2011
- “Epidemiology of Autism Spectrum disorders”, Patel et al; Pediatric Clinics of North America 2013 (59/I)
- “Do childhood vaccines cause thrombocytopenia?”, Schifle et al; Paediatric and Child Health, January 2009
- “Nectin 4 is the epithelial cell receptor for measles virus”, Noyce, Richardson; Trends in Microbiology, September 2012
Other studies were referred to in the creation of this article. Those listed are either the earliest found with the information used or, in case of reviews, contain references to both the earliest studies and later examination. There exists a vast amount of literature available on the subject; specifically, I would recommend VK Singh’s other work as comprehensive, reliable, and informative (although not particularly easy to read).
NOTE: An informative opinion article on the propriety of a distinct diagnosis and research category for CDD can be found here.