1. 1st Brazilian Meeting for Autism Research EBPA 10 Porto Alegre April 22-24 2010 Pr. Eric Fombonne McGill University The epidemiology of Autism and PDDs

2. Outline Recent prevalence surveys Best current estimate for ASDs World wide efforts on ASD epidemiology Time trends: is there an epidemic?

4. Great Britain Stafford London

5. Staffordshire surveys 92-95 cohort 96-98 cohort Combined samples N=15,500N=10,903 N=26,403 NPNP NP95% CI Autistic disorder 2616.82422.0 5018.914.1-25.0 PDD NOS 5636.12724.8 8331.425.0-39.0 Asperger 13 8.41211.0 25 9.5 6.1-14.0 CDD 1 0.7 1 0.9 2 0.8 0.1-2.7 All PDDs 96 1 61.96458.7 16060.651.6-70.7 1: One girl with Rett syndrome has been excluded Chakrabarti & Fombonne (2005)

6. New Montreal survey Sample –English Montreal School Board (EMSB) for Anglophone children –71 schools; 23,635 pupils from K to grade 11 –Children with PDD identified through special education code (ASD) that provides extra funding to the school –187 children with PDD identified on April 1 st 2008 Lazoff, Fombonne et al., 2010 (submitted)

7. Male N Female N Total N Prevalence /10,000 (95% confidence interval) PDD -NOS881510343.6 (35.2 – 52.0) Autistic Disorder 50106025.4 (19.0 - 31.8) Asperger194239.7 (5.8 – 13.7) CDD1010.4 (0.0 – 1.3) Total1582918779.1 (67.8 - 90.4) New Montreal survey Lazoff, Fombonne et al. 2010 (submitted)

8. New Montreal survey: trends in birth cohorts 1991-2002

11. Outline Recent prevalence surveys Best current estimate for ASDs World wide efforts on ASD epidemiology Time trends: is there an epidemic?

12. Recent review of surveys 63 published surveys 50 on AD, 14 on ASP, 12 on CDD, 26 on PDD half published since 2001 AD: 2.2/1,000 ASP: 1/1,000 CDD: 2/100,000 PDD: ~ 7/1,000 Some studies have PDD rates over 1% 1.1% in NJ- USA (CDC, 2007) 1.3% in Arizona (CDC, 2009) 1.2% in the UK (Baird et al. 2006) 1.3% in some birth cohorts (Fombonne et al., 2010) 1.8% in Japan (Kawamura et al. 2008) Fombonne et al 2010 (in press)

13. Equivalences 70/10,000 = 7/1,000 = 0.7% = 1 child in 140

14. Brazil estimates for PDDs (2000) Age groups 0- 45 - 910 - 1415 - 19Under 20 Urbana Population 12,749,346 12,787,933 13,509,94314,401,006 53,448,228 N expected 89,24589,51694,570100,807374,138 Rural Population 3,636,893 3,788,326 3,843,740 3,548,28214,817,241 N expected 25,458 26,518 26,906 24,838 103,720 Brasil Population16,386,23916,576,27917,353,68317,949,28968,265,490 N expected 114,704116,034121,476125,645 477,858 Based on Tabela 1.1.1 - População residente, por sexo e situação do domicílio,segundo os grupos de idade – Brasil and a prevalence estimate of 0.7%

15. Outline Recent prevalence surveys Best current estimate for ASDs World wide efforts on ASD epidemiology Time trends: is there an epidemic?

17. Korean Autism Study (KAS) Research Design Total Population Study No sampling Mechanisms to Identify Children 1.Mandatory Educational System 2.Home Schooling 3.Disability Registry Two Stage Design 1.Multi-Informant Screening 2.Confirmative Dx with ADOS & ADI-R

18. School Visit and Screening: East District

19. Cultural differences: Autism versus Reactive attachment disorder ?

20. Value of an epidemiological survey Get an estimate of the magnitude of the health problem Get a baseline useful to monitor time trends Provide information to decision-makers re. service needs and planning Provide data on risk factors/correlates Evaluate access to services and factors that influence it Generate a representative case series from which to sample for further studies (ie case-control, outcome studies, etc…) Generate data on trajectories of ASD subjects in the local health/welfare system Develop up-to-date screening and evaluation techniques Increase awareness, involve professionals, develop expertise

21. Outline Recent prevalence surveys Best current estimate for ASDs World wide efforts on ASD epidemiology Time trends: is there an epidemic?

22. Time trends in autism Problems : –prevalence versus incidence rates –changes in case definition / case finding –secular changes in age at diagnosis –statistical power issues

23. Japan – Honda et al, 2005United Kingdom – Taylor et al, 1999 Minnesota, USA – Gurney et al., 2003Denmark – Madsen et al., 2003 It happened in the 1990s...

24. Approaches used to evaluate time trends in autism – referral statistics – comparison of prevalence surveys over time – repeat surveys in defined areas – trends in rates in consecutive birth cohorts – incidence studies

25. Prevalence and access to services Population Services Population Services Low access to services Same prevalence High access to services

26. Trends in Minnesota Gurney et al., 2003 DSM-III-RICD-10DSM-IV Individual with Disabilities Educational Act (IDEA)

27. Gurney et al., 2003 1991-92 birth cohort as it ages age effect

29. Prevalence of autism and MR of unknown cause in California Croen et al., 2001 King & Bermans replication (Int J Epi, in press): 25% of DDS caseload of Autism is due to shift from MR to ASD

30. Diagnostic substitution: from Language disorders to Autism 38 subject (31 M, 7F), diagnosed with developmental language disorder (20 PLI, 18 SLI) re-evaluated as adults (age 15 to 31) with autism specific instruments (ADI-R and ADOS-G) SLI N=18 PLI N=20 ALL N=38 Autistic disorder on both ADI and ADOS 0-0- 8 (40%) 8 (21%) PDD on both ADI and ADOS 2 (11%) 11 (55%) 13 (34%) PDD on either ADI or ADOS 6 (33%) 19 (95%) 25 (66%) Bishop et al. 2008

31. Population estimates from recent CDC studies 6/10,000 CDC Atlanta survey

32. 1.Odds of being classified in autism category increased by 1.21 during 1994-2003. In the meantime, the odds decreased significantly of being classified in the LD (learning disability; OR=0.98) and the MR (mental retardation; OR=0.97) 2.The growing prevalence of autism was directly associated decreasing prevalence of LD and MR within states 3.A significant downward deflection in the historical trajectories of LD and MR occurred when autism became reported as an independent category in 1993-94. 4.From 1994 to 2003, the mean increase for the combined category of Autism + Other Health Impairments +Trauma Brain Injury + Developmental Delay was 12/1000; the mean decrease for MR and LD was 11/1000 during the same period.

33. Approaches used to evaluate time trends in autism – referral statistics – comparison of prevalence surveys over time – repeat surveys in defined areas – trends in rates in consecutive birth cohorts – incidence studies

34. Relative rates of AD and PDD NOS Study Definition for other PDD ADPDD NOS Ratio Lotter (1966) behaviour similar to 4.1 3.30.8 autistic children Brask (1970) other psychoses or 4.3 1.90.4 borderline psychotic Wing et al (1976) socially impaired 4.9 16.33.3 (triad of impairments) Hoshino et al autistic mental 2.3 2.91.3 (1982) retardation Burd et al (1987) autistic-like 3.3 > 7.82.4 Cialdella & other forms of 4.5 4.71.0 Marmelle (1989) infantile psychosis

35. Impact of diagnostic criteria on rate estimation: example of the Northern Finland survey Kielinen et al., 2000 7.6 Autism Spectrum ICD-10 39,2163015 – 18 6.1 Autism ICD-10/DSM-IV 39,2162815 – 18 2.3Kanner39,216 915 – 18 Rate /10,000 CriteriaPopulationNAge

36. Study design impact on prevalence Example of 4 recent UK surveys PDD rate location sizeage group method /10,000 Baird et al. South.East 16,235 7 Early screening 57.9 2000 Thames + FU identification Chakrabarti & Stafford 15,500 2½ - 6½ intense screening 62.6 Fombonne -shire + assessment 2001 Fombonne England 10,438 5 - 15 household survey 26.1 et al. 2001 & Wales Taylor et al. North 490,000 0 - 16 administrative 10.1 1999 Thames records Six-fold variation in estimates

38. Stafford Surveys Chakrabarti and Fombonne, 2005 Rate /10,000

39. Prevalence rates by birth cohorts (1972-1985) in two surveys Fombonne et al. 1997 Prevalence rate / 10,000 Birth cohort

40. Smeeth, Fombonne et al., 2004 Incidence of PDD in GPRD – UK 1988 – 2001 Incidence Autism 1988199219962001 Incidence other PDDs 0 0.2 0.4 0.6 0.8 1.0 1.2 0 0 0.06 1.06 0 0.4 0.8 1.2 1.6 2.0 0.11 0.340.68 1.92

41. Time trends : conclusions Most epidemiological studies are not informative to gauge trends over time There is evidence that methodological factors account for a substantial part of the observed increase in prevalence Prevalence rates have gone up but this trend cannot be interpreted as evidence of a secular increase in the incidence The hypothesis of an increased incidence is not ruled out, but it remains to be tested with adequate epidemiological data

42. Take home messages... ASDs are amongst the most frequent child neurodevelopmental disorders Increasing numbers/prevalence occurred with changes in diagnostic criteria and awareness Surveillance should be developed in order to detect future changes in the incidence