Aetiology of Autism: Findings and Questions*

Journal of Intellectual Disability Research 231     pp –  

Blackwell Science, LtdOxford, UKJIRJournal of Intellectual Disability Research-Blackwell Publishing Ltd, 4231238Original ArticleAetiology of autismM. Rutter

invited review

Aetiology of autism: ﬁndings and questions*

M. Rutter

SGDP Centre, Institute of Psychiatry, Denmark Hill, London, UK

Abstract Keywords autism spectrum disorders, epidemiology, genetic influences, incidence, measles- Background Although there is good evidence that mumps-rubella vaccine autism is a multifactorial disorder, an adequate understanding of the genetic and non-genetic causes has yet to be achieved. Introduction Methods Empirical research findings and concep- tual reviews are reviewed with respect to evidence on Before turning to what is known on the aetiology of possible causal influences. autism, it is necessary to note that, over the last few Results Much the strongest evidence concerns the decades, there have been important changes in the importance of susceptibility genes, but such genes diagnostic concept. During the s, it was gener- have yet to be identified. Specific somatic conditions ally assumed that autism was a rare, seriously hand- (such as tuberous sclerosis and the fragile X anomaly) icapping disorder, usually associated with intellectual account for a small proportion of cases. Over recent disability (ID), constituting a condition that was decades there has been a major rise in the rate of qualitatively distinct from variations in social and diagnosed autism. The main explanation for this rise communicative competence within the normal range. is to be found in better ascertainment and a broad- Both epidemiological and genetic research findings ening of the diagnostic concept. Nevertheless, some have forced a change in concept as a result of the degree of true rise cannot be firmly excluded. How- evidence that autistic-like abnormalities can and do ever, the epidemiological evidence on the main occur in individuals of normal intelligence (see Rut- hypothesized environmental explanation, namely the ter ). There appears to be a broad spectrum of measles-mumps-rubella vaccine, is consistently disorders that are closely similar in quality but milder negative. in some respects and often occurring in individuals Conclusion Progress on the elucidation of the causes of normal intelligence. Even further outside the core of autism will be crucially dependent on the combi- is a group of much milder, but apparently similar, nation of epidemiology with more basic science lab- conditions that have come to be termed the ‘broader oratory studies. phenotype’. They are found in some one in five first- degree relatives of individuals with autism. But outside even this broad spectrum there are  Correspondence: Michael Rutter, PO , SGDP Centre, Institute other disorders that may sometimes be confused with of Psychiatry, De Crespigny Park, Denmark Hill, London SE AF, autism. First, there is the well-defined condition UK. Tel.    (e-mail: [email protected]). *Based on a paper given at the IASSID Conference in Montpellier, called Rett syndrome after the physician who first France, June . described it (Rett ; Hagberg et al. ); it has

Journal of Intellectual Disability Research       232 M. Rutter • Aetiology of autism

been shown to be a result, in almost all cases, of a population, it is very high. Putting together the twin mutant gene on the X chromosome. Its downward and family findings, the figures indicate that it is likely clinical course and neurological features mark it out that there are somewhere between  and  suscep- as different from autism but many of those who suffer tibility genes for autism that act synergistically. The from this condition go through a phase that involves reason why most siblings do not have ASD, despite social impairments superficially similar to autism. In the high heritability, is that they have only some of addition, there are several less well-defined quasi- the relevant genes. autistic patterns, similar to autism in many respects, but different in some features. These include the patterns seen in some congenitally blind children (Hob- Susceptibility genes for autism son et al. ) and some children reared in seriously A major growth area in psychiatric genetics as applied depriving institutions (Rutter et al. ); in some to ASD concerns the search, using linkage and young adults who had a serious developmental disor- association strategies, to identify susceptibility der of receptive language when young (Clegg et al. genes (Rutter ). There are very promising, par- ; Howlin et al. ), and in some children with tially replicated, findings with respect to loci on a semantic pragmatic language disorder (Bishop et al. chromosomes  and , as well as leads on other ; Bishop ). In the remainder of this paper I chromosome locations. It may be anticipated, with will confine attention to what is known about the some confidence, that the actual susceptibility genes aetiology of autism and the more closely associated will be determined during the next decade, if not autism spectrum disorders (ASD), and for the most rather earlier than that. Linkage strategies study the part will not deal with either the broader phenotype degree to which affected members in the same family or these varied more atypical patterns. show coinheritance of the same genetic loci on particular chromosomes. Association strategies, by con- trast, determine whether individuals with ASD differ Susceptibility genes for autism from controls in their pattern of allelic variations of spectrum disorders specific genes. That is, each person inherits one out of several possible allelic copies of each gene. Some Much the best-established risk factor for ASD is copies will carry risk whereas others will not. Note genetic liability (Rutter ). In idiopathic cases that the susceptibility genes may well turn out to be without a known medical cause, twin studies have common genes that, on their own, do not directly shown a concordance rate in monozygotic (identical) cause disease – rather than rare pathogenic muta- twin pairs of about % as compared with a rate of tions. Although it may be expected that the identifi- % in dizygotic (fraternal) pairs. Taken together with cation of susceptibility genes will be enormously the population base rate for autism, this implies that helpful in shaping the biological research that will the heritability or underlying genetic liability is about determine the neural basis of autism, it is much less % – the highest figure among all multifactorial likely that the genes themselves will be of much prac- child psychiatric disorders. However, the twin data tical utility in terms of either screening or diagnosis. also show that the genetic liability extends well beyond the traditional core diagnosis of autism to include a wider spectrum of autistic-like disorders Single gene conditions including the broader phenotype. In other words, in twin pairs discordant for core autism, the monozy- The only single gene condition with an established gotic concordance for the broader phenotype far association with ASD is tuberous sclerosis (Smalley exceeds the dizygotic concordance. ). The best estimates suggest that this is found Family studies tell a similar story. The rate of ASD in about % to % of cases of ASD (Harrison & in the siblings of individuals with autism is about % Bolton ). Although the association with ASD is – a rate many times higher than the rate of about well established, its meaning in terms of the causal .% in the general population. The rate of % seems mechanisms that are involved is less well understood. low in absolute terms but, relative to the general It is probably important that the association with

Journal of Intellectual Disability Research       233 M. Rutter • Aetiology of autism

ASD is particularly marked only when tuberous scle- Medical conditions and autism rosis is associated with severe ID, severe epilepsy and There have been various debates in the literature the location of tubers in the temporal lobe (Bolton concerning the frequency with which ASD are asso- et al. ). In other words, the implication is that ciated with definite diagnosable medical conditions the causal pathway may involve the type and location that are likely to have been implicated in the causal of the pathophysiology of brain disturbance deriving processes (Rutter et al. ; Gillberg & Coleman from tuberous sclerosis, rather than from anything ). One of the major difficulties in coming to a directly associated with the genes on chromosome  specific figure concerns the major extent to which and chromosome  that give rise to tuberous findings are likely to be influenced by both the nature sclerosis. of the samples investigated and the thoroughness of the medical investigations undertaken. However, a reasonable estimate would be that something in the Chromosome anomalies order of % of individuals with ASD has some potentially relevant identifiable somatic disease or Most of the evidence on connections between chro- disorder. This means that an appropriately thorough mosome anomalies and ASD come from isolated medical assessment is essential in all cases. The gen- case reports (Gillberg ). These are of little use eral consensus would be that this should include care- in testing causal hypotheses and greater reliance ful medical examination, including the use of Wood’s needs to be placed on systematic studies of either light, in order to detect tuberous sclerosis, that kary- general population or clinic samples. Initially, the otyping should be routinely undertaken, but also that strongest claims concerned the supposed associa- this should include the use of DNA methods to diag- tion between the Fragile X anomaly and autism nose the Fragile X anomaly. Although some Scandi- (Gillberg & Wahlstrom ). The initial claims of a navian researchers have advocated intrusive further strong association were based on unsatisfactory cell investigations as a routine – including lumbar punc- culture methods and, once DNA methods became ture, brain scanning using a general anaesthetic, etc. available, it was evident that the cytological identifi- (Gillberg ), the evidence would seem to suggest cation of fragile sites led to many false positives a more conservative approach in which the extent, (Gurling et al. ). Systematic surveys of large and type, of medical investigations are determined on samples of individuals with an ASD have shown the basis of the clinical history and clinical examina- that only about % to % show the Fragile X tion findings in the individual patient. anomaly (Bailey et al. ; Chakrabarti & Fom- bonne ). This is still a meaningful and significant association but it is evident that it accounts for Prenatal influences a very small proportion of cases of ASD. On the Intra-uterine infections and toxins other hand, surveys of individuals known to have the Fragile X anomaly have shown that quite a high Although the main research attention has focused on proportion show social and communicative abnor- genetic influences in autism and on associated med- malities of a kind that could be confused with ical conditions, the evidence is clear cut that most autism, even though they do not meet the usual cri- ASD constitute multifactorial disorders. That means teria for an ASD (Reiss & Dant ). The only that some kinds of non-genetic factors are also likely other chromosome anomaly at all commonly associ- to play a part in aetiology, even if we know little about ated with ASD concerns the maternally transmitted them. interstitial duplications of chromosome  (Folstein Isolated case reports have suggested that a range of & Rosen-Sheidley ). Systematic surveys of possible intrauterine infections and toxins could play chromosome anomalies in a series of individuals a contributory causal role in the development of ASD with an ASD have shown that approximately % in individual cases (Nelson ; Rodier & Hyman show anomalies of one kind or another. These are ; Folstein & Rosen-Sheidley ; Medical quite varied and, in most cases, their clinical signifi- Research Council ). These include various pos- cance remains uncertain. sible maternal circumstances that could affect the

Journal of Intellectual Disability Research       234 M. Rutter • Aetiology of autism

foetus including hypothyroidism, thalidomide use, autism. However, both Wing () and Schopler valproic acid use, cocaine or alcohol use, and con- et al. () did not find this. Their studies were both genital cytomegalovirus infection. None of these have flawed by a failure to take ethnicity into account. been prominent in any of the epidemiological studies Most recent surveys have not examined social back- of ASD and it seems unlikely that they constitute ground systematically but, with one possible excep- commonly operating risk factors for ASD. tion (Fombonne et al. ), the few that have done The only other established link is that between so have not found any association. Most reviewers congenital rubella and autism (Chess et al. ; have therefore concluded that probably there is no Chess ). Findings from a systematically studied association – other than that associated with referral large sample of children with congenital rubella bias (MRC ; Fombonne ). That is likely to showed that a substantial minority developed some be true but it has to be said that the evidential base form of ASD. The rate was substantially higher in the for the conclusion is weak. In addition, there has been children whose handicaps included marked ID as some indication that autism may be more common well as visual and hearing defects (but ASD was not in the UK in children born to parents of Afro- confined to that group). It is noteworthy, however, Caribbean background (Wing ; Goodman & that the follow-up showed that the course of ASD in Richards ) and Gillberg and Gillberg reported these children tended, on the whole, to be rather an increase in Sweden for children born to immigrant different from that associated with idiopathic autism parents (Gillberg & Gillberg ). As with social in that, although the children remained markedly class, the findings are contradictory, inconclusive and handicapped at they grew older, the autistic features based on small numbers (MRC ; Fombonne tended to diminish. The findings, of course, are of ). The conclusion has to be ‘not proven’. very limited contemporary relevance in view of the rarity of congenital rubella following the establish- Monozygotic twinning as a risk factor ment of population-wide vaccination programmes. Non-geneticists tend to assume that the non-genetic factors involved in the aetiology of ASD must neces- Obstetric complications and birth order sarily involve some form of specific environmental Since the first pulling together of the evidence risk. It is important to appreciate that that is not (Deykin & MacMahon ), it has generally been necessarily the case (Molenaar et al. ; Jensen found that autism tends to be more common in the ). For example, Greenberg et al. (), and also firstborn in sibships of two but more common in the Betancur et al. (), reported an apparent excess last born in larger sibships. Also it has usually been of twins among affected sibling pairs with autism (but found that obstetric complications (generally of a see Hodge et al. ; Visscher ). If this finding mild variety) are more common in individuals with were to prove valid, it would suggest that being a twin autism than in their unaffected siblings or in controls constituted a risk factor for autism. That could come (Bolton et al. ). The balance of the evidence sug- about either because twinning is associated with an gests that the obstetric complications do not consti- increased risk of obstetric complications or because tute an environmentally mediated risk; rather, they monozygotic twinning itself constitutes a form of may reflect a response to a genetically abnormal foe- congenital anomaly (Hall ). Congenital anoma- tus (Bolton et al. ). lies have been found to be more common in individuals with autism and these probably index the ways in which development, which is probabilistic rather Parental social class and country of origin than deterministically programmed, may go awry Following Kanner’s initial report that children with (Vogel & Motulsky ). Thus, congenital anomalies autism were disproportionately likely to have parents are more common in twins than in singletons and are of a high socio-economic background (Kanner ), more common in children born to older mothers than and Lotter’s finding that there was a slight tendency in those born to younger ones ( Myrianthopoulos & of this kind (Lotter ), there was an interest in the Melnick ; Rutter et al. ). Accordingly, it possibility that high social class was associated with could be that these semi-random developmental per-

turbations could enhance the adverse effects of a MMR hypothesis. If MMR had been responsible for genetic liability to ASD. It should be noted, never- the rise in ASD it would be expected that the intro- theless, that congenital anomalies show an increased duction of the vaccine, in countries in which the take- rate in a wide range of psychiatric disorders, so that up was rapid and very high (as was the case in the the risk is by no means specific to ASD. UK), should be followed by a large step-wise increase More importantly, some scepticism is necessary in ASD, that this should be followed by a plateau in with respect to the supposed finding that the rate of rate, and that when MMR was stopped (as it was in twinning is actually increased in ASD. Ascertainment Japan) this should be followed by a fall in rate. In all biases are likely to have played a major role and it is these phases, the main changes should apply to noteworthy that the most systematic twin sample of regressive autism. The evidence shows that none of Bailey et al. () did not include a significant these expectations were borne out (Rutter ). excess of monozygotic twins; nor did Hallmayer The hypothesis regarding Thimerosal (a preserva- et al.’s  Australian twin sample. It may be con- tive that was, until recently, used in many vaccines) cluded that the postulated increased risk for autism is somewhat different in detail, in that mercury is associated with being a twin remains a speculative known to be a neurotoxin; accordingly, a direct suggestion and, on balance, the evidence indicates adverse effect on the brain was expected. However, that it is not likely that being a monozygotic twin it remains uncertain whether a ‘bolus’ effect causes constitutes a major risk factor. damage (i.e. the immediate, large, but very transient rise in mercury level following vaccination) or whether the damage derives from the cumulative Postnatal risk influences mercury build-up resulting from multiple vaccina- All the evidence suggests that it is rare for postnatal tions. The epidemiological evidence on Thimerosal is somatic disease to give rise to an ASD. There are much less than that on MMR but again the findings isolated case reports of herpes encephalitis causing are negative. autism (Gillberg ; Ghaziuddin et al. ) but These negative conclusions give rise to two main this is decidedly unusual and ASD has not been queries. First, if neither MMR nor Thimerosal is reported as a common consequence of encephalitis responsible for the rise in autism, what has caused in childhood (Rantala et al. ). the increase? It is clear that the main explanation is that it derives from a combination of better ascertainment and a broadening of the diagnostic con- Measles-mumps-rubella and thimerosal cept. However, the possibility that, in addition, there During the last decade, the main focus within the has been a true rise in incidence because of some, as realm of possible postnatal risk factors for ASD has yet unidentified, environmental risk factor cannot be been on the possibility that immunization constitutes ruled out (Rutter ). Second, although it is a contributory factor for ASD. First, there was the no longer plausible that MMR or Thimerosal have suggestion that the measles-mumps-rubella (MMR) led to an overall increase in ASD, the epidemiologi- vaccine was responsible for the huge recent rise in the cal data cannot exclude the possibility that either rate of diagnosed ASD (Rutter ). It was argued might have a risk effect in a small proportion of that through the route of a vaccine-caused gut disor- unusually susceptible children. There is no evidence der, there was leakage of protein products into the supporting this suggestion but it cannot be firmly blood stream and that these then caused a special excluded. regressive form of autism (in which there was a loss of previously acquired social and communicative Other possibilities skills). A range of epidemiological studies was undertaken to determine whether the use of the MMR It needs to be added that there are other possible vaccine might be responsible for the worldwide rise causes of ASD. For example, there are uncertain in the rate of autism as diagnosed, and in particular pointers to the role of immunological abnormalities. whether it led to this postulated regressive form of The supporting evidence is weak but it is a group of autism. The evidence is consistently against the risk factors that warrant further exploration.

Conclusions progress is going to be crucially dependent on the combination and integration of epidemiology with In summary, there are good epidemiological data more basic science laboratory studies. indicating that the true incidence of ASD now is likely to be of the order of – cases per  , as compared with the original estimate of four per References   made some four decades ago (Rutter ). Bailey A., Le Couteur A., Gottesman I., Bolton P., Simo- Administrative data show massive increases over noff E., Yuzda E. & Rutter M. () Autism as a strongly time in the rate of diagnosed ASD and it is clear that, genetic disorder: evidence from a British twin study. Psy-    in large part, this is because of the combination of chological Medicine , – .  better ascertainment and a broadening of the diag- Betancur C., Leboyer M. & Gillberg C. ( ) Increased rate of twins among affected sibling pairs with autism. nostic concept, but a true rise over time in the inci- American Journal of Human Genetics , –. dence of ASD cannot be entirely ruled out. Despite Bishop D. V. M. () The role of genes in the etiology strong claims made about the possible role of MMR of specific language impairment. Journal of Communica- in relation to the causation of autism, there is no tion Disorders , –. convincing evidence in support of this hypothesis. In Bishop D. V. M., North T. & Donlan C. () Genetic particular, the rate of ASD shows no particular asso- basis of specific language impairment: evidence from a ciation with either the stopping or starting of MMR twin study. Developmental Medicine and Child Neurology , –. and there has been no change over time in the pattern of association between ASD and either bowel distur- Bolton P., Murphy M., Macdonald H., Whitlock B., Pickles A. & Rutter M. () Obstetric complications in autism: bance or developmental regression. The evidence consequences or causes of the condition? Journal of the with respect to a possible association with Thimero- American Academy of Child and Adolescent Psychiatry , sal, a preservative in some vaccines, is much more –. limited but, again, there is no supporting epidemio- Bolton P. F., Park R., Higgins N., Griffiths P. D. & Pickles logical evidence of a causal association. It remains A. () Neuro-epileptic determinants of autism spec-  possible that there has been a true rise in incidence trum disorders in tuberous sclerosis complex. Brain , –. because of some environmental risk factor but, if so, Chakrabarti S. & Fombonne E. () Pervasive develop- it remains quite obscure as to what that factor might mental disorders in preschool children. Journal of the be. American Medican Association , –. The genetic evidence is clear cut that ASD are Chess S. () Follow-up report on autism in congenital multifactorial conditions caused by multiple genes rubella. Journal of Autism and Childhood Schizophrenia , and some, as yet to be identified, non-genetic factors. –. The genetic factors that underlie ASD are likely to Chess S., Kern S. J. & Fernandez P. B. () Psychiatric be heterogeneous but it remains unclear whether that Disorders of Children with Congenital Rubella. Brunner/ heterogeneity is indexed by clinical features and, if it Mazel, New York.  is, which they are. The evidence is also clear-cut that Clegg J., Hollis C., Mawhood L. & Rutter M. ( ) Devel- opmental language disorders: a follow-up in later adult the genetic liability to ASD involves a broader phe- life. Journal of Child Psychology and Psychiatry (in press). notype that extends well beyond the traditional Deykin E. Y. & MacMahon B. () Pregnancy, delivery, diagnosis of a handicapping condition of autism. and neonatal complications among autistic children. However, it is significant that the broader phenotype American Journal of Diseases of Children , –. does not seem to be associated with either epilepsy Folstein S. & Rosen-Sheidley B. () Genetics of autism: or ID and very little is known on the factors, genetic complex aetiology for a heterogeneous disorder. Nature or non-genetic, that are implicated in the transition Reviews: Genetics , –. from the milder broader phenotype to a seriously Fombonne E. () Epidemiological surveys of autism handicapping disorder. and other pervasive developmental disorders: an update. Journal of Autism and Developmental Disorders , – Epidemiological findings have been helpful in both . ruling in and ruling out various postulated causal Fombonne E., Simmons H., Ford T., Meltzer H. & Good- influences and they will continue to be formative in man R. () Prevalence of pervasive developmental that connection. Nevertheless, it is evident that disorders in the British nationwide survey of child mental

health. Journal of the American Academy of Child and Ado- Howlin P., Mawhood L. & Rutter M. () Autism and lescent Psychiatry , –. developmental receptive language disorder – a compara- Ghaziuddin M., Tsai L. Y., Eilers L. & Ghaziuddin N. tive follow-up in early adult life. II: Social behavioural, () Brief report: autism and herpes simplex encepha- and psychiatric outcomes. Journal of Child Psychology and litis. Journal of Autism and Developmental Disorders , Psychiatry , –. –. Jensen A. R. () The puzzle of nongenetic variance. In: Gillberg C. () Brief report: onset at age  of a typical Intelligence, Heredity, and Environment (eds R. Sternberg autistic syndrome. A case report of a girl with herpes & E. L. Grigorenko), pp. –. Cambridge University simplex encephalitis. Journal of Autism and Developmental Press, Cambridge. Disorders , –. Kanner L. () Autistic disturbances of affective contact. Gillberg C. () Medical work-up in children with autism Nervous Child , –. and Asperger syndrome. Brain Dysfunction , –. Lotter V. () Epidemiology of autistic conditions in Gillberg C. () Chromosomal disorders and autism. young children: . Prevalence. Social Psychiatry , –. Journal of Autism and Developmental Disorders , –. Medical Research Council () MRC Review of Autism Gillberg C. & Coleman M. () Autism and medical Research: Epidemiology and Causes. MRC, London. disorders: a review of the literature. Developmental Medi- Molenaar P. C. M., Boomsma D. I. & Dolan C. V. () cine and Child Neurology , –. A third source of developmental differences. Behavior Gillberg I. C. & Gillberg C. () Autism in immigrants: Genetics , –. a population-based study from Swedish rural and urban Myrianthopoulos N. C. & Melnick M. () Malforma- areas. Journal of Intellectual Disability Research , –. tions in monozygotic twins: a possible source of environ- Gillberg C. & Wahlstrom J. () Chromosome abnormal- mental inﬂuence on the developmental genetic clock. In: ities in infantile autism and other childhood psychoses: a Gene–Environment Interactions in Common Diseases (eds E. population study of  cases. Developmental Medicine and Inouye & H. Nishimusa), pp. –. University Park Child Neurology , –. Press, Baltimore. Goodman R. & Richards H. () Child and adolescent Nelson K. B. () Prenatal and perinatal factors in the psychiatric presentations of second-generation Afro- etiology of autism. Pediatrics , –. Caribbeans in Britain. British Journal of Psychiatry , Rantala H., Uhari M., Uhari M., Saukkonen A.-L. & Sorri –. M. () Outcome after childhood encephalitis. Devel- Greenberg D. A., Hodge S. E., Sowinski J. & Nicoll D. opmental Medicine and Child Neurology , –. () Excess of twins among affected sibling pairs with Reiss A. L. & Dant C. C. () The behavioral neuroge- autism: implications for the etiology of autism. American netics of Fragile X syndrome: analyzing gene-brain- Journal of Human Genetics , –. behavior relationships in child developmental psycho- Gurling H. M. D., Bolton P. F., Vincent J., Melmer G. & pathologies. Development and Psychopathology , –. Rutter M. () Molecular and cytogenetic investiga- Rett A. () Uber ein eigenartiges himatrophisches Syn- tions of the Fragile X region including the FRAX and drom bei Hyperammonie in Kindesalter. Wiener Mediz- FRAX E CGG trinucleotide repeat sequences in families inische Wochenschrift , –. multiplex for autism and related phenotypes. Human Rodier P. & Hyman S. () Early environmental factors    Heredity , – . in autism. Mental Retardation and Developmental Disabili- Hagberg B., Aicardi J., Dias K. & Ramos O. () A ties Research Reviews , –. progressive syndrome of autism, dementia, ataxia and loss Rutter M. () Incidence of autism spectrum of purposeful hand use in girls: Rett syndrome – report disorders: changes over time and their meaning. Acta of  cases. Annals of Neurology , –. Paediatrica. doi: ./. Hall J. G. () Twinning. Lancet , –. Rutter M. () Genetic inﬂuences and autism. In: Hand- Hallmayer J., Glasson E. J., Bower C., Petterson B., Croen book of Autism, rd edn. (ed. F. Volkmar). John Wiley, L., Grether J. & Risch N. () On the twin risk in New York (in press). autism. American Journal of Human Genetics , –. Rutter M., Andersen-Wood L., Beckett C., Bredenkamp Harrison J. E. & Bolton P. F. () Annotation: tuberous D., Castle J., Groothues C., Kreppner J., Keaveney L., sclerosis. Journal of Child Psychology and Psychiatry , Lord C., O’Connor T. G. and the ERA Study Team –. () Quasi-autistic patterns following severe early glo- Hobson R. P., Lee A. & Brown R. () Autism and bal privation. Journal of Child Psychology and Psychiatry congenital blindness. Journal of Autism and Developmental , –. Disorders , –. Rutter M., Bailey A., Bolton P. & Le Couteur A. () Hodge S. E., Greenberg D. A., Betancur C. & Gillberg C. Autism and known medical conditions: myth and sub- () Response to Visscher. American Journal of Human stance. Journal of Child Psychology and Psychiatry , – Genetics , –. .

Rutter M., Bolton P., Harrington R., Le Couteur A., Visscher P. M. () Increased rate of twins among Macdonald H. & Simonoff E. () Genetic factors in affected sib pairs. American Journal of Human Genetics , child psychiatric disorders: . A review of research strat- –.   egies. Journal of Child Psychology and Psychiatry , – Vogel F. & Motulsky A. G. () Human Genetics: Problems  . and Approaches, rd edn. Springer-Verlag, Berlin.  Schopler E., Andrews C. E. & Strupp K. ( ) Do autistic Wing L. () Childhood autism and social class: a children come from upper middle class parents? Journal question of selection? British Journal of Psychiatry ,    of Autism and Developmental Disorders , – . –. Smalley S. () Autism and tuberous sclerosis. Journal of Autism and Developmental Disorders , –. Accepted  July 