global prevalence of autism and other pervasive developmental disorders

Global Prevalence of Autism and Other PervasiveDevelopmental DisordersMayada Elsabbagh, Gauri Divan, Yun-Joo Koh, Young Shin Kim, Shuaib Kauchali, Carlos Marcín,Cecilia Montiel-Nava, Vikram Patel, Cristiane S. Paula, Chongying Wang, Mohammad Taghi Yasamy,and Eric Fombonne

We provide a systematic review of epidemiological surveys of autistic disorder and pervasive developmental disorders(PDDs) worldwide. A secondary aim was to consider the possible impact of geographic, cultural/ethnic, and socioeco-nomic factors on prevalence estimates and on clinical presentation of PDD. Based on the evidence reviewed, the medianof prevalence estimates of autism spectrum disorders was 62/10 000. While existing estimates are variable, the evidencereviewed does not support differences in PDD prevalence by geographic region nor of a strong impact of ethnic/culturalor socioeconomic factors. However, power to detect such effects is seriously limited in existing data sets, particularly inlow-income countries. While it is clear that prevalence estimates have increased over time and these vary in differentneighboring and distant regions, these findings most likely represent broadening of the diagnostic concets, diagnosticswitching from other developmental disabilities to PDD, service availability, and awareness of autistic spectrum disordersin both the lay and professional public. The lack of evidence from the majority of the world’s population suggests acritical need for further research and capacity building in low- and middle-income countries. Autism Res 2012, 5:160–179. © 2012 International Society for Autism Research, Wiley Periodicals, Inc.

Keywords: epidemiology; prevalence; global health; low- and middle-income countries

Introduction

Autism is a life-long neurodevelopmental condition in-terfering with the person’s ability to communicate andrelate to others. Since the earliest epidemiological sur-veys in the 1960s, a wealth of data has become available,indicating a much higher prevalence of the conditionthan previously thought [Fombonne, 2003a, 2005; Fom-bonne, Quirke, & Hagen, 2011]. It is now recognizedthat some individuals with the condition are able tolead independent and fulfilling lives, whereas for othersthe impact can be severe, interfering significantly withquality of life [Farley et al., 2009]. While the globalburden of autism is currently unknown, in the UnitedStates and in the UK, the annual societal cost of thecondition exceeds several billions [Ganz, 2007; Knapp,Romeo, & Beecham, 2007].

Increased recognition, understanding, and awarenessof autism in the last few decades have been, in part,

driven by the significant growth in research evidence.While many aspects of autism remain poorly under-stood, major advances have been made in terms ofhighlighting the genetic [Abrahams & Geschwind,2008], biological [Belmonte et al., 2004], environmental[Currenti, 2009], and developmental [Elsabbagh &Johnson, 2010] origins of the condition. Large-scaleand well controlled cohort studies (e.g., http://www.earlistudy.org) following-up pregnant mothers are likelyto clarify the effects of some pre- and perinatal risk factorsimplicated in autism. Significant strides have also con-tributed towards developing and validating screeningand diagnostic instruments, helping to reduce heteroge-neity in clinical characterization in research studies.While some of these diagnostic tools remain highlyresource intensive, they are increasingly used in clinicalsettings, as they provide rich and systematic informationto inform service provisions where those are available.However, even in high-income countries, provisions

From the Department of Psychiatry, Montreal Children’s Hospital, Montreal, Canada and Birkbeck, University of London, London, UK (M.E.); Sangath,Goa, India (G.D.); The Korea Institute for Children’s Social Development, Seoul, South Korea (Y.-J.K.); Child Study Center, Yale University School ofMedicine, New Haven, USA (Y. S.K.); Nelson R Mandela School of Medicine, University of KwaZulu-Natal, South Africa and Mailman School of PublicHealth, Columbia University, New York, NY, USA (S.K.); Mexican Autism Clinic, Mexico City Mexico (C.M.); Psychology Department, La Universidad delZulia, Maracaibo, Venezuela (C.M.-N.); Centre for Global Mental Health, London School of Hygiene and Tropical Medicine, London, UK and SangathCentre, Goa, India (V.P.); Developmental Disorders Program, Mackenzie Presbyterian University, São Paulo, Brazil (C.S.P.); Center for Behavioural Scienceand School of Medicine, Nankai University, Tianjin, China (C.W.); World Health Organization (M.T.Y.); Department of Psychiatry, Montreal Children’sHospital, Montreal, Canada (E.F.)

Received June 7, 2011; accepted for publication November 11, 2011Address for correspondence and reprints: Eric Fombonne, Department of Psychiatry, Montreal Children’s Hospital, 4018 Ste-Catherine West, Montreal,

QC, H3Z 1P2, Canada. E-mail: [email protected] online 11 April 2012 in Wiley Online Library (wileyonlinelibrary.com)DOI: 10.1002/aur.239© 2012 International Society for Autism Research, Wiley Periodicals, Inc.

INSAR160 Autism Research 5: 160–179, 2012

REVIEW ARTICLE

for screening, diagnosis, and intervention are highlyvariable and many cases absent in community settings.

Such advances in autism research have contributedtowards bridging the gap between evidence and practicein some countries, but there is little systematic infor-mation available with regards to the impact of the con-dition on most of the world’s population. Frequentlyregarded as essential for advancing basic research andstrategic for informing policy and developing services,epidemiological studies have emerged as a clear prioritywithin several global initiatives. The charity AutismSpeaks in partnership with the US Center for DiseaseControl (CDC) launched the International Autism Epide-miology Network (http://www.autismepidemiology.net),bringing together researchers worldwide and focusingspecifically on service improvements in developingcountries. According to the network, prevalence studiesfor PDD are ongoing in Australia, Mexico, Finland,Portugal, Iceland, India, Vietnam, Taiwan, South Africa,and Uganda. Focusing on a broader context thanautism, the Movement for Global Mental Health hasidentified a clear treatment gap, particularly pronoun-ced in low- and middle-income countries [http://www.globalmentalhealth.org; Patel et al., 2008]. Epide-miological data on the burden of mental and neurolo-gical disorders and systematic mapping of relevantservices in low- and middle-income countries encou-raged World Health Organization (WHO) to launch themental health Gap Action Programme (mhGAP) [WHO,2008]. Moreover, the Global Alliance for Chronic Disease,which groups several agencies including Australia’sNational Health Medical Research Council, the CanadianInstitutes of Health Research, the Chinese Academy ofMedical Science, the UK’s Medical Research Council,and the United States’ National Institutes of Health,announced a program to identify the world’s “GrandChallenges in Global Mental Health” (http://www.grandchallenges.org). The reasons for why epidemio-logical surveys are viewed as a priority extend beyondthe need for objective and robust estimates of pre-valence. These provide additional valuable benefits asthey often result in systematic information regardingexisting services and may help in assessing the needsand priorities for each community. In the long run,the availability of comparable estimates from differentgeographic regions may also enable testing challenginghypotheses regarding the etiology of PDD.

Systematic Review Methodology

The primary aim of this report is to provide an up-to-date and systematic review of the methodological fea-tures and substantive results of epidemiological surveysof the prevalence of pervasive developmental disorders

(PDDs), including autistic disorder (AD) worldwide. Thisis to our knowledge the first attempt to review availableevidence from different world regions, including low-and middle-income countries. While the scope of thecurrent review does not provide extensive coverage ofthe situation in all countries, every effort was made torepresent objectively available evidence from regionswhere research capacity is limited.

Search Strategies

We adopted a systematic review methodology in theidentification of epidemiological reports included inthe current article. First, as a preliminary step, key find-ings from previous reviews of epidemiological surveyswere incorporated into the current report [Fombonne,2003a, 2005; Fombonne et al., 2011; Sun & Allison, 2010;Williams, Higgins, & Brayne, 2006]. Second, severalauthors or group of authors undertook a comprehensivereview of available evidence from regions that wereunderrepresented in previous reviews. Different coun-tries were grouped into subregions according to theWHO classification. The main search strategy was toperform extensive region and/or country-specific searchof Medline publications. When the authors were familiarwith additional local or regional search engines, thesewere also used. Additional search sources includedIndMed search engine, indmed.nic.in national medicaldatabase, VIP.cn Chinese search engine, and arabpsy-net.com regional engine. Search terms included Autis*,Asperger, Autistic Disorder and Child Development andDisorders, Pervasive (mesh term), or their other-languageequivalent. In regions where it was difficult to identifyepidemiological reports, authors also consulted otherresearchers and practitioners who supplied referencesor information. Using these strategies authors identifiedand reviewed over 600 studies typically not included inprevious reviews (133 from South and Central America,and the Caribbean, 319 from South East Asia and theWestern Pacific, and 130 from the Eastern Mediterranean,54 from Africa). Languages of the identified articlesincluded Arabic, Chinese, Dutch, English, French, Portu-guese, and Spanish.

Inclusion and Exclusion Criteria of Prevalence Estimate

From the identified publications, the following crite-ria were used for selecting epidemiological surveysincluded in Tables I–III:

• A full article (published or otherwise) was available ina language that could be read by one of the authors.

• The primary aim of the survey was to estimateprevalence of PDD and/or narrower AD. Studiesexclusively focusing on other categories of PDD,

161Elsabbagh et al./Global epidemiology of autismINSAR

Tabl

eI.

Euro

pe

Regi

onCa

tego

ryYe

arAu

thor

sCo

untr

yAr

eaPo

pula

tion

Age

Num

ber

affe

cted

Diag

nost

iccr

iter

ia%

wit

hav

erag

eIQ

Gend

erra

tio

Prev

alen

ce/

1000

095

%CI

Euro

peAD

1966

Lott

erU

KM

iddl

esex

7800

08–

1032

Rati

ngsc

ale

15.6

2.6

(23/

9)4.

12.

7;5.

5

AD19

72Br

ask

Denm

ark

Aarh

usCo

unty

4650

02–

1420

Clin

ical

—1.

4(1

2/7)

4.3

2.4;

6.2

AD19

76W

ing

etal

.U

KCa

mbe

rwel

l25

000

5–14

17a

24-i

tem

rati

ngsc

ale

ofLo

tter

3016

(16/

1)4.

8b2.

1;7.

5

AD19

83Bo

hman

etal

.Sw

eden

Coun

tyof

Väst

erbo

tten

6900

00–

2039

Rutt

ercr

iter

ia20

.51.

6(2

4/15

)5.

63.

9;7.

4

AD19

84M

cCar

thy

etal

.Ir

elan

dEa

st65

000

8–10

28Ka

nner

—1.

33(1

6/12

)4.

32.

7;5.

9

AD19

86St

einh

ause

net

al.

Germ

any

Wes

tBer

lin27

961

60–

1452

Rutt

er55

.82.

25(3

6/16

)1.

91.

4;2.

4

AD19

89Ci

alde

llaan

dM

amel

leFr

ance

Rhôn

e13

518

03–

961

DSM

-III

like

—2.

34.

53.

4;5.

6AD

1991

Gillb

erg

etal

.Sw

eden

Sout

h-W

estG

othe

nbur

g,Bo

husl

än78

106

4–13

74DS

M-I

II-R

182.

7(5

4/20

)9.

57.

3;11

.6

AD19

92Fo

mbo

nne

and

duM

azau

brun

Fran

ce4

regi

ons

274

816

9&

1315

4Cl

inic

al-I

CD-1

0lik

e13

.32.

1(1

05/4

9)4.

94.

1;5.

7

AD19

97Fo

mbo

nne

etal

.Fr

ance

3re

gion

s32

534

78–

1617

4Cl

inic

alIC

D-10

-lik

e12

.11.

81(1

12/6

2)5.

354.

6;6.

1

AD19

97W

ebb

etal

.U

KSo

uth

Glam

orga

n,W

ales

7330

13–

1553

DSM

-III

-R—

6.57

(46/

7)7.

25.

3;9.

3

AD19

97Ar

vids

son

etal

.Sw

eden

Möl

nlyc

ke1

941

3–6

9IC

D-10

22.2

3.5

(7/2

)46

.416

.1;7

6.6

AD19

98Sp

onhe

im&

Skje

ldal

Norw

ayAk

ersh

usCo

unty

6568

83–

1434

ICD-

1047

.1c

2.09

(23/

11)

5.2

3.4;

6.9

AD19

99Ta

ylor

etal

.U

KNo

rth

Tham

es49

000

00–

1642

7IC

D-10

——

8.7

7.9;

9.5

AD19

99Ka

desj

öet

al.

Swed

enKa

rlsta

d82

66.

7–7.

76

DSM

-III

-R/I

CD-1

0Gi

llber

gcr

iter

iafo

rAS

50.0

5.0

(5/1

)72

.614

.7;1

30.6

AD20

00Ba

ird

etal

.U

KSo

uth-

East

Tham

es16

235

750

ICD-

1060

15.7

(47/

3)30

.822

.9;4

0.6

AD20

00Po

wel

let

al.

UK

Wes

tMid

land

s25

377

1–5

62Cl

inic

al/I

CD10

/DSM

-IV

——

7.8

5.8;

10.5

AD20

00Ki

elin

enet

al.

Finl

and

Nort

h(O

ulu

etLa

plan

d)27

572

5–7

57DS

M-I

V49

.8d

4.12

(156

/50)

20.7

15.3

;26.

0

AD20

01Da

vido

vitc

het

al.

Isra

elH

aifa

2616

07–

1126

DSM

-III

-R/D

SM-I

V—

4.2

(21/

5)10

.06.

6;14

.4

AD20

07Ol

ivei

raet

al.

Port

ugal

Mai

nlan

dan

dAz

ores

6779

56–

911

5DS

M-I

V17

2.9

16.7

14.0

;20.

0AD

2001

Fom

bonn

eet

al.

UK

Engl

and

and

Wal

es10

438

5–15

27DS

M-I

V/IC

D-10

55.5

8.0

(24/

3)26

.116

.2;3

6.0

AD20

01M

agnú

sson

and

Saem

unds

enIc

elan

dW

hole

Isla

nd43

153

5–14

57M

ostl

yIC

D-10

15.8

4.2

(46/

11)

13.2

9.8;

16.6

AD20

01Ch

akra

bart

iand

Fom

bonn

eU

KSt

affo

rd15

500

2.5–

6.5

26IC

D10/

DSM

-IV

29.2

3.3

(20/

6)16

.810

.3;2

3.2

AD20

02M

adse

net

al.

Denm

ark

Nati

onal

regi

ster

6385

98

46IC

D-10

——

7.2

5.0–

10.0

AD20

04Te

brue

gge

etal

.U

KKe

nt2

536

8–9

6IC

D-10

—(6

/0)

23.7

9.6;

49.1

AD20

05Ch

akra

bart

i&Fo

mbo

nne

UK

Staf

ford

1090

34–

724

ICD-

10/D

SM-I

V33

.33.

8(1

9/5)

22.0

14.4

;32.

2

INSAR162 Elsabbagh et al./Global epidemiology of autism

AD20

06Gi

llber

get

al.

Swed

enGö

tebo

rg32

568

7–12

115

Gillb

erg’

scri

teri

a—

3.6

(90/

25)

35.3

29.2

;42.

2

AD20

06Ba

ird

etal

.U

KLo

ndon

(Sou

thTh

ames

)56

946

9–10

81IC

D-10

478.

3(ª

72/9

)38

.929

.9;4

7.8

AD20

07El

lefs

enet

al.

Denm

ark

Faro

eIs

land

s7

689

8–17

12IC

D-10

Gillb

erg

crit

eria

for

AS—

3.0

(9/3

)16

.07.

0;25

.0

AD20

07La

tifa

ndW

illia

ms

UK

Wal

es39

220

0–17

50Ka

nner

——

12.7

9.0;

17.0

AD20

08W

illia

mse

tal

.U

KSo

uth

Wes

t(Av

on)

1406

211

30IC

D-10

86.7

5.0

(25/

5)21

.613

.9;2

9.3

PDD

2000

Bair

det

al.

UK

Lond

on(S

outh

Tham

es)

1623

57

94IC

D-10

6015

.7(8

3/11

)57

.946

.8;7

0.9

PDD

2001

Chak

raba

rtia

ndFo

mbo

nne

UK

Staf

ford

1550

04–

796

ICD-

1074

.23.

8(7

7/20

)61

.950

.2;7

5.6

PDD

2002

Mad

sen

etal

.De

nmar

kNa

tion

alre

gist

er—

873

8IC

D-10

——

30.0

—PD

D20

02Sc

otte

tal

.U

KCa

mbr

idge

3359

85–

1119

6IC

D-10

—4.

0(—

)58

.3*

50;6

7*

PDD

2004

Tebr

uegg

eet

al.

UK

Kent

253

68–

921

ICD-

10—

6.0

(18/

3)82

.851

.3;1

26.3

PDD

2005

Chak

raba

rtia

ndFo

mbo

nne

UK

Staf

ford

1090

34–

664

ICD-

1070

.26.

1(5

5/9)

58.7

45.2

;74.

9

PDD

2006

Bair

det

al.

UK

Sout

hTh

ames

5694

69–

1015

8IC

D-10

453.

3(1

21/3

7)11

6.1

90.4

;141

.8

PDD

2006

Har

riso

net

al.

UK

Scot

land

134

661

0–15

443e

ICD-

10,D

SMIV

—7.

0(3

69/5

3)44

.239

.5;4

8.9

PDD

2006

Gillb

erg

etal

.Sw

eden

Göte

borg

3256

871

226

2DS

M-I

V—

3.6

(205

/57)

80.4

71.3

;90.

3

PDD

2007

Lati

fand

Will

iam

sU

KSo

uth

Wal

es39

220

0–17

240

ICD-

10,D

SM-I

V,Ka

nner

’s&

Gillb

erg’

scri

teri

a—

6.8

—61

.254

;69*

PDD

2007

Elle

fsen

etal

.De

nmar

kFa

roe

Isla

nds

768

98–

1741

DSM

-IV,

Gillb

erg’

scri

teri

a68

.35.

8(3

5/6)

53.3

36;7

0

PDD

2008

Will

iam

set

al.

UK

Avon

1406

211

86IC

D-10

85.3

6.8

(75/

11)

61.9

48.8

;74.

9

PDD

2009

Baro

n—Co

hen

etal

.U

KCa

mbr

idge

882

45–

983

ICD-

10—

—94

f75

;116

PDD

2011

Brug

haet

al.

UK

Nati

onal

733

3>1

672

ADOS

-410

0%3.

8—

9830

;165

*Cal

cula

ted

byth

eau

thor

.a T

hisn

umbe

rcor

resp

onds

toth

esa

mpl

ede

scri

bed

inW

ing

and

Goul

d[1

979]

.b T

hisr

ate

corr

espo

ndst

oth

efi

rstp

ublis

hed

pape

ron

this

surv

eyan

dis

base

don

12pa

rtic

ipan

tsam

ongs

tchi

ldre

nag

ed5

to14

year

s.c I

nth

isst

udy,

mild

men

talr

etar

dati

onw

asco

mbi

ned

wit

hav

erag

eIQ

,whe

reas

mod

erat

ean

dse

vere

men

talr

etar

dati

onw

ere

grou

ped

toge

ther

.d T

hese

figu

resa

pply

toth

ew

hole

stud

ysa

mpl

eof

206

part

icip

ants

wit

han

auti

smsp

ectr

umdi

sord

er(A

SD).

e Est

imat

edus

ing

aca

ptur

e-re

capt

ure

anal

ysis

,the

num

bero

fcas

esus

edto

calc

ulat

epr

eval

ence

was

esti

mat

edto

be59

6.f R

ate

base

don

Spec

ialE

duca

tion

Need

sreg

iste

r.A

figu

reof

99/1

000

0is

prov

ided

from

apa

rent

alan

ddi

agno

stic

surv

ey.O

ther

esti

mat

esin

this

stud

yva

ryfr

om47

to16

5/10

000

deri

ving

from

vari

ousa

ssum

ptio

nsm

ade

byth

eau

thor

s.AD

,aut

isti

cdi

sord

er;A

S,As

perg

ersy

ndro

me;

DSM

,Dia

gnos

tic

and

Stat

isti

calM

anua

lofM

enta

lDis

orde

rs;I

CD,I

nter

nati

onal

Clas

sifi

cati

onof

Dise

ases

.PDD

,dev

elop

men

tald

isor

der.


Tabl

eII

.W

este

rnPa

cifi

c,So

uth

East

Asia

,and

the

East

ern

Med

iter

rane

an

Regi

onCa

tego

ryYe

arAu

thor

sCo

untr

yAr

eaPo

pula

tion

Age

Num

ber

affe

cted

Diag

nost

iccr

iter

ia%

wit

hav

erag

eIQ

Gend

erra

tio

Prev

alen

ce/

1000

095

%CI

Wes

tern

Paci

fic

AD19

82H

oshi

noet

al.

Japa

nFu

kush

ima-

Ken

609

848

0–18

142

Kann

er’s

crit

eria

—9.

9(1

29/1

3)2.

331.

9;2.

7

AD19

87M

atsu

ishi

etal

.Ja

pan

Kuru

me

City

3283

44–

1251

DSM

-III

—4.

7(4

2/9)

15.5

11.3

;19.

8

AD19

88Ta

noue

etal

.Ja

pan

Sout

hern

Ibar

aki

9539

47

132

DSM

-III

—4.

07(1

06/2

6)13

.811

.5;1

6.2

AD19

89Su

giya

ma

etAb

eJa

pan

Nago

ya12

263

316

DSM

-III

——

13.0

6.7;

19.4

AD19

96H

onda

etal

.Ja

pan

Yoko

ham

a8

537

518

ICD-

1050

.02.

6(1

3.5)

21.0

811

.4;3

0.8

AD20

00Lu

oCh

ina

Fujia

n10

802

0–14

3AB

C,CC

MD-

2-R

&DS

M-I

II-R

—2:

12.

80.

6;8.

1*AD

2002

Wan

get

al.

Chin

aJi

angS

u3

912

2–6

7CA

BS,C

ARS,

CCM

D2-R

—2.

5:1

17.9

7.2;

36.8

*AD

2003

Wan

get

al.

Chin

aJi

angS

u7

344

2–6

9CA

BS,C

ARS,

CCM

D2-R

,—

2:1

12.3

5.6;

23.3

*AD

2004

Guo

etal

.Ch

ina

Tian

jin3

606

2–6

5CA

RS,C

CMD-

2-R,

CABS

—Al

lboy

s13

.94.

5;32

.3*

AD20

04Zh

ang

etal

.Ch

ina

Tian

jin7

316

2–6

8CA

RS,D

SM–I

V0

(7:1

)10

.94.

7;21

.5*

AD20

05Zh

ang

&Ji

Chin

aTi

anjin

734

52–

68

DSM

-IV,

CARS

—7

10.9

4.7;

21.4

*AD

2005

Hon

daet

al.a

Japa

nYo

koha

ma

3279

15

123

ICD-

1025

.32.

5(7

0/27

)37

.531

.0;4

5.0

AD20

07Ya

nget

al.

Chin

aZu

nyic

ity

1041

23–

126

DSM

-IV,

ABC

—(5

/1)

5.8

2.1;

12.5

*AD

2007

Chen

etal

.Ch

ina

Taiw

an32

953

93–

811

15IC

D-10

—4.

1334

*32

;36*

AD20

08Zh

ang

etal

.Ch

ina

Wux

icit

y25

521

1–6

25CA

RS/D

SM-I

V—

3.2

9.8

6.3;

14.5

*AD

2009

Zhan

get

al.

Chin

aGu

iZho

u5

000

1–6

5CC

MD-

2-R,

CARS

,CAB

S—

4:1

10.0

3.2;

23.3

*AD

2011

Kim

etal

.So

uth

Kore

aGo

yang

5526

67–

1227

DSM

-IV

55.6

4.4

9456

;134

PDD

2004

Icas

iano

etal

.Au

stra

liaBa

rwon

4515

3*2–

1717

7DS

M-I

V53

.48.

3(1

58/1

9)39

.234

;45*

PDD

2008

Won

gan

dH

uiCh

ina

Hon

gKo

ng4

247

206

0–14

682

DSM

-IV

306.

6(5

92/9

0)16

.114

.9;1

7.3*

PDD

2008

Kaw

amur

aet

al.

Japa

nTo

yota

1258

95–

822

8DS

M-I

V66

.42.

8(1

68/6

0)18

1.1

158.

5;20

5.9*

PDD

2011

Kim

etal

.So

uth

Kore

aGo

yang

5526

67–

1210

4DS

M-I

V72

2.4

189

143;

236

Sout

hEa

stAs

iaAD

1992

Wig

nyo-

sum

arto

etal

.In

do-n

esia

Yogy

akar

ita

512

04–

76

CARS

02.

0(4

/2)

11.7

2.3;

21.1

PDD

2009

Pere

raet

al.

SriL

anka

Sem

iurb

anar

ea37

41.

5–2

4DS

M-I

V,Pa

rent

repo

rt—

Allb

oys

107

2.9.

2;27

1.6*

East

ern

Med

i-te

rran

ean

PDD

2007

Eape

net

al.

UAE

Thre

ere

gion

s69

43.

52

DSM

-IV

—5.

14(—

)29

3.5;

103.

7*

PDD

2010

Al-F

arsi

etal

.Om

anCo

untr

y-w

ide

798

913

0–14

113

DSM

-IV

—2.

9(8

4/29

)1.

41.

2;1.

7

PDD

2012

Sam

adie

tal

.Ir

anCo

untr

y-w

ide

132

033

45

826

ADI-

R—

4.3

6.26

5.84

;6.7

0

*Cal

cula

ted

byth

eau

thor

a Thi

sfig

ure

was

calc

ulat

edby

the

auth

oran

dre

fers

topr

eval

ence

data

(not

cum

ulat

ive

inci

denc

e)pr

esen

ted

inth

epa

per(

the

M/F

rati

ois

base

don

asu

bsam

ple)

.AD

,aut

isti

cdis

orde

r;AD

I-R,

Auti

smDi

agno

stic

Inte

rvie

w-R

evis

ed;C

ARS,

Child

hood

Auti

smRa

ting

Scal

e;CA

BS,C

lanc

yAut

ism

Beha

vior

Scal

e;CC

MD,

Chin

eseCl

assi

fica

tion

ofM

enta

lDis

orde

rs;D

SM,D

iagn

osti

can

dSt

atis

tica

lMan

ualo

fMen

talD

isor

ders

;ICD

,Int

erna

tion

alCl

assi

fica

tion

ofDi

seas

es.P

DD,d

evel

opm

enta

ldis

orde

r.


Tabl

eII

I.Am

eric

a

Regi

onCa

tego

ryYe

arAu

thor

sCo

untr

yAr

eaPo

pula

tion

Age

Num

ber

affe

cted

Diag

nost

iccr

iter

ia%

wit

hav

erag

eIQ

Gend

erra

tio

Prev

alen

ce/

1000

095

%CI

Amer

ica

AD19

70Tr

effe

rtU

SAW

isco

nsin

899

750

3–12

69Ka

nner

—3.

06(5

2/17

)0.

70.

6;0.

9

AD19

87Bu

rdet

al.

USA

Nort

hDa

kota

180

986

2–18

59DS

M-I

II—

2.7

(43/

16)

3.26

2.4;

4.1

AD19

88Br

yson

etal

.Ca

nada

Part

ofNo

va-S

coti

a20

800

6–14

21Ne

wRD

C23

.82.

5(1

5/6)

10.1

5.8;

14.4

AD19

89Ri

tvo

etal

.U

SAU

tah

769

620

3–27

241

DSM

-III

343.

73(1

90/5

1)2.

472.

1;2.

8

AD20

01Be

rtra

ndet

al.

USA

Bric

kTo

wns

hip,

New

Jers

ey8

896

3–10

36DS

M-I

V36

.72.

2(2

5/11

)40

.528

.0;5

6.0

AD20

02Cr

oen

etal

.U

SACa

lifor

nia

DDS

495

033

35–

125

038

CDER

«Fu

llsy

ndro

me

»62

.8a

4.47

(411

6/9

21)

11.0

10.7

;11.

3

AD20

05Ba

rbar

esie

tal

USA

Min

neso

ta(O

lmst

ead

Coun

ty)

3772

60–

2111

2DS

M-I

V—

—29

.724

.0;3

6.0

AD20

06Fo

mbo

nne

etal

.Ca

nada

Mon

trea

lisl

and

2774

95–

1760

DSM

-IV

—5.

7(5

1/9)

21.6

16.5

;27.

8

AD20

08M

onti

el-N

ava

etal

.Ve

nez-

uela

Mar

acai

bo25

490

53–

943

0DS

M-I

V-TR

—3:

117

15;1

9*AD

2009

van

Balk

omet

al.

Arub

aNa

tion

al13

109

0–13

25DS

M-I

V36

.07.

3(2

2/3)

19.1

12.3

;28.

1

AD20

10La

zoff

etal

.Ca

nada

Mon

trea

l23

635

5–17

68DS

M-I

V—

5.8

(58/

10)

25.4

19.0

;31.

8

PDD

2001

Bert

rand

etal

.U

SANe

wJe

rsey

889

63–

1060

DSM

-IV

512.

7(4

4/16

)67

.451

.5;8

6.7*

PDD

2003

Year

gin-

Alls

opp

etal

.U

SAAt

lant

a28

945

63–

1098

7DS

M-I

V31

.84.

0(7

87/1

97)

34.0

32;3

6

PDD

2003

Gurn

eyet

al.

USA

Min

neso

ta—

8–10

——

——

52.0

-66

.0b

— —PD

D20

06Fo

mbo

nne

etal

.Ca

nada

Mon

trea

l27

749

5–17

180

DSM

-IV

—4.

8(1

49/3

1)64

.955

.8;7

5.0

PDD

2007

aCD

CU

SA6

stat

es18

776

18

125

2DS

M-I

V-TR

38to

60c

2.8

to5.

567

.0—

d

PDD

2007

bCD

CU

SA14

stat

es40

757

88

268

5DS

M-I

V-TR

55.4

e3.

4to

6.5

66.0

63;6

8PD

D20

08Ni

chol

aset

al.

USA

Sout

hCa

rolin

af47

726

829

5DS

M-I

V-TR

39.6

3.1

(224

/71)

62.0

56;7

0

PDD

2009

Koga

net

al.

USA

Nati

onw

ide

7791

13–

1791

3—

—4.

5(7

46/1

67)

110

94;1

28

PDD

2009

CDC

USA

11st

ates

3307

790

82

757

DSM

-IV

594.

5(—

)89

.686

;93

PDD

2010

Lazo

ffet

al.

Cana

daM

ontr

eal

2363

55–

1718

7DS

M-I

V—

5.4

(158

/29)

79.1

67.8

;90.

4

Sout

hAm

eric

a&

Cari

bbea

nPD

D20

08Le

jarr

aga

etal

.Ar

gent

ina

San

Isid

ro83

90–

511

DSM

-IV

——

13.1

7.3;

23.6

*

PDD

2009

Van

Balk

omet

al.

Arub

aNa

tion

al13

109

0–13

69DS

M-I

V58

.86.

7(6

0/9)

52.6

41.0

;66.

6

PDD

2010

Paul

aet

al.

Braz

ilAt

ibai

a1

470

7–12

4DS

M-I

V—

(4/0

)27

.217

.6;3

6.8

*cal

cula

ted

byth

eau

thor

.a T

hisp

ropo

rtio

nis

likel

yto

beov

eres

tim

ated

and

tore

flect

anun

derr

epor

ting

ofm

enta

lret

arda

tion

inth

eCD

ERev

alua

tion

s.b T

hese

are

the

high

estp

reva

lenc

ees

tim

ated

repo

rted

inth

isst

udy

ofti

me

tren

ds.T

hepr

eval

ence

in10

-yea

rold

sis

fort

he19

91bi

rth

coho

rt,a

ndth

atfo

r8-y

earo

lds

fort

he19

93bi

rth

coho

rt.B

oth

prev

alen

ces

wer

eca

lcul

ated

inth

e20

01–2

002

scho

olye

ar.

c Spe

cifi

cva

lues

for%

wit

hav

erag

eIQ

are

avai

labl

efo

reac

hst

ate

prev

alen

ce.

d Con

fide

nce

inte

rval

sare

avai

labl

efo

reac

hst

ate

prev

alen

ce.

e Ave

rage

acro

ssse

ven

stat

es.

f Chi

ldre

nag

ed8,

born

eith

erin

2000

and

2002

,and

incl

uded

inth

etw

oCD

Cm

ulti

site

repo

rts.

AD,a

utis

tic

diso

rder

;CDE

R,Ca

lifor

nia

Depa

rtm

ento

fDis

abili

ties

;DSM

,Dia

gnos

tic

and

Stat

isti

calM

anua

lofM

enta

lDis

orde

rs;P

DD,d

evel

opm

enta

ldis

orde

r;RD

C,Re

sear

chDi

agno

stic

Crit

eria

.


including childhood disintegrative disorder andAsperger disorder (see details in the following) wereexcluded from the current review as these were veryrare and have been reviewed elsewhere [Fombonneet al., 2011]. Other articles, including a wealth of casereports, surveys of more general intellectual ormental disabilities, and surveys estimating PDDprevalence among comorbid conditions, were alsoexcluded from the current report.

• The survey included a step of diagnostic confirmationregardless of the specific approach taken to verifycaseness (see following discussion for further details).Those studies providing prevalence estimates withoutdiagnostic confirmation, particularly from regionswhere the evidence base is overall limited, werereviewed in the following main text but are not pre-sented in the tables.

• The following information categories were includedor could be ascertained based on information fromthe survey: the country and area where the surveywas conducted, the size of the population based onwhich the prevalence estimate was ascertained,the age range of the participants, the numberaffected by AD or PDD, the diagnostic criteria usedin case ascertainment, gender ratio within theaffected sample, prevalence estimate (number per10 000), and 95% confidence intervals (CI) for theestimate.

Tables I–III summarize the results of epidemiologicalsurveys identified in different world regions. Studiesestimating either AD or PDD as a spectrum of disordersfrom each region were grouped and ordered by publica-tion year. Characteristics of samples surveyed in thesestudies are briefly summarized. A secondary aim ofthis report was to review evidence for the possibleimpact of geographic, economic, social, or cultural/ethnic factors on the clinical presentation and theprevalence of PDD. To that end, selected studies rel-evant for clinical presentation, particularly from under-represented regions were reviewed. This includedinformation on gender differences in the rates of PDD,cognitive characteristics of the identified cases, medicalhistory, or socioeconomic and ethnic characteristics.

Combining and Appraising Results

Each group of authors focusing on specific WHO regionsentered their findings into the same template thatincluded the following: (a) general context and back-ground for the region covered; (b) key findings on preva-lence estimates entered into the same format presentedin Tables I–III; (c) key findings clinical characterizationsummarized qualitatively; and (d) commentary on oppor-tunities and barriers in conducting research in the regioncovered. Studies included were appraised by each group

of authors who decided on inclusion and exclusionbased on the criteria outlined earlier. M.E. reviewed allentries and combined the results, and together with E.F.performed overall appraisal of the data included. Allauthors subsequently verified the information in the finalarticle. The diversity in the methodological approachesprecluded the possibility of any further analysis of theseestimates included meta-analysis but average figuresand basic analyses were included in later sections.

Methodological Issues and Interpretation ofPrevalence Estimates

Before summarizing the key findings related to preva-lence and clinical characterization, we will discuss somemethodological issue that need to be accounted for wheninterpreting epidemiological data. These include charac-terization of PDD, changes in concepts over time, andcase identification and evaluation.

Characterization of PDD

Over time, the definitions of autism have changed asillustrated by the numerous diagnostic criteria that wereused in both epidemiological and clinical settings. Start-ing with the narrowly defined Kanner’s autism, defini-tions progressively broadened in the criteria proposed byRutter [Rutter, 1970], and then International Classifica-tion of Diseases (ICD)-9 [WHO, 1975], Diagnostic andStatistical Manual of Mental Disorders (DSM)-III [Ameri-can Psychiatric Association, 1980] and DSM-III-R [Ameri-can Psychiatric Association, 1987], and more recently inthe two major nosographies used worldwide: ICD-10[WHO, 1992] and DSM-IV [American Psychiatric Associa-tion, 1994]/DSM-IV-TR [American Psychiatric Associa-tion, 2000]. The first diagnostic criteria reflected the morequalitatively severe forms of the phenotype of autism,usually associated with severe delays in language andcognitive skills. It is only in the 1980s that less severeforms of autism were recognized, either as a qualifier forautism occurring without intellectual disability (so-called“high-functioning” autism), or as separate diagnostic cat-egories (PDDs not otherwise specified [PDDNOS]) withina broader class of autism spectrum disorders (ASD)denominated “PDDs” (an equivalent to ASD) in currentnosographies. Whilst it had been described in the litera-ture as early as 1944, one type of PDD, Asperger disorder,appeared in official nosographies only in the 1990s and,then, with unclear validity, especially with respect to itsdifferentiation from “high-functioning” autism. Subtypesof PDD that existed in DSM-III subsequently disappeared(i.e., autism—residual state). While there is generally highinter-rater reliability on the diagnosis of PDDs and com-monality of concepts across experts, some differences


persist between nomenclatures about the terminologyand precise operationalized criteria of PDDs. For example,DSM-IV [American Psychiatric Association, 1994] has abroad category of PDDNOS, sometimes referred to loose-ly as “atypical autism,” whereas ICD-10 [WHO, 1992]has several corresponding diagnoses for clinical presenta-tions that fall short of AD, which include atypical autism(F84.1, a diagnostic category that existed already inICD-9), Other PDD (F84.8), and PDD, unspecified (F84.9).

In the recent years, the research definitions of syn-dromes falling on the autism spectrum have beenexpanded further with an increasing reliance on a dimen-sionalization of the autism phenotype. As no impair-ment or diagnostic criteria are available for these milderforms, the resulting boundaries with the spectrum ofPDDs are left uncertain. Whether or not this plays a rolein more recent epidemiological studies is difficult toknow, but the possibility should be considered in assess-ing results for the new generation of surveys. While thevast majority of studies have relied on these commonnosographies, culturally specific instruments and systemshave been used. For example, the Chinese Classificationof Mental Disorders (CCMD) has been used in some ofthe studies appearing in Chinese journals and often com-bined with other classification systems.

Case Identification

When an area or population has been identified for asurvey, different strategies have been used to find partici-pants matching the case definition retained for the study.Some studies have solely relied on existing service pro-viders databases [Croen, Grether, Hoogstrate, & Selvin,2002]; on special educational databases [Fombonne,Zakarian, Bennett, Meng, & McLean-Heywood, 2006;Gurney et al., 2003; Lazoff, L., Piperni, & Fombonne,2010]; or on national registers [Madsen et al., 2002] forcase identification. These studies have the common limi-tation of relying on a population group who happenedto access the service provider or agencies rather thansampling from the population at large. As a result, par-ticipants with the disorder who are not in contact withthese services are yet unidentified and not included ascases, leading to underestimation of prevalence propor-tion. This is a particularly significant issue when estimat-ing prevalence using such methods in communities withrecognized limitations in service provisions.

Other investigations have relied on a multistageapproach to identify cases in underlying populations [e.g.,Baird et al. 2006; Kim et al., 2011]. The aim of the firstscreening stage of these studies is to cast a wide net in orderto identify participants possibly affected with a PDD, withthe final diagnostic status being determined at a subse-quent phase. The process used by researchers often con-sisted of sending letters or brief screening scales requesting

school and health professionals and/or other data sourcesto identify possible cases of autism. Few investigationshave relied on systematic sampling techniques that wouldensure a near complete coverage of the target population.Moreover, each investigation differed in several keyaspects of this screening stage. First, the thoroughness ofthe coverage of all relevant data sources varied enormouslyfrom one study to another. In addition, the surveyed areaswere not comparable in terms of service development,reflecting the specific educational or health care systems ofeach country and of the period of investigation. Second,the type of information sent out to professionals invited toidentify children varied from simple letters, including afew clinical descriptors of autism-related symptoms ordiagnostic checklists rephrased in nontechnical terms, tomore systematic screening strategy based on question-naires or rating scales of known reliability and validity.Third, variable participation rates in the first screeningstages provide another source of variation in the screeningefficiency of surveys, although refusal rates tended, onaverage, to be very low.

Few studies provided an estimate of the reliabilityof the screening procedure. The sensitivity of the screen-ing methodology is also difficult to gauge in autismsurveys, and the proportion of children truly affectedwith the disorder but not identified in the screeningstage (the “false negatives”) remains generally unmea-sured. The usual approach, which consists of sampling atrandom screened negative participants in order to esti-mate the proportion of false negatives, has not beenused in these surveys for the obvious reason that, becauseof the relatively low frequency of the disorder, it wouldbe both imprecise and very costly to undertake suchestimations. This approach may gradually change inview of recent prevalence studies suggesting that autismcan no longer be regarded as a rare condition. In anycase, the magnitude of this of under- or over-estimationof prevalence studies relative to “true” prevalence is cur-rently unknown.

Case Evaluation

When the screening phase is completed, participantsidentified as positive screens go through the next stepinvolving a more in-depth diagnostic evaluation toconfirm their case status. Similar considerations aboutthe methodological variability across studies apply tothese more intensive assessment phases. Participationrates in second stage assessments within the studiesreviewed were generally high. The source of informationused to determine caseness usually involved a combina-tion of data coming from different informants and datasources (medical records, educational sources, parents,teachers, health professionals, etc.), with a direct assess-ment of the person with autism being offered in some but


not all studies. Obviously, surveys of very large popula-tions as those conducted in the United States by the CDC[CDC, 2007a, 2007b, 2009] or in national registers[Madsen et al., 2002] did not include a direct diagnosticassessment of all participants by the research team.However, these investigators could generally confirm theaccuracy of their final caseness determination by under-taking, on a randomly selected subsample, a more com-plete diagnostic work-up. The CDC surveys haveestablished a methodology for surveys of large popula-tions that relies on screening of population using mul-tiple data sources, a systematic review and scoring systemfor the data gathered in the screening phase combinedwith, in the less obvious cases, input from experiencedclinicians. This methodology is adequate for largesamples, and is likely to be used in the future for surveil-lance efforts. When participants are directly examined,the assessments were conducted with various diagnosticinstruments, ranging from an unstructured examinationby a clinical expert to the use of batteries of standardizedmeasures by trained research staff. The Autism DiagnosticInterview [ADI; Le Couteur et al., 1989] and/or theAutism Diagnostic Observational Schedule [ADOS; Lordet al., 2000] have been increasingly used in the mostrecent surveys in North America and some Europeancountries. Against this background and in view of thesediverse methodological features in the studies we report,we now turn to the available evidence from epidemio-logical surveys and studies on clinical characterization.

EuropeOlder Prevalence Estimates

Prior to the modern conceptualization of PDD, epidemio-logical surveys presented in Table I conducted as early asthe 1960s in Europe, mainly in Northern countries, pro-vided useful information on rates of syndromes similar toautism but not meeting strict diagnostic criteria for ADthen in use [Fombonne, 2003b, 2005]. At the time, dif-ferent labels were used by authors to characterize theseclinical pictures such as the triad of impairments involv-ing deficits in reciprocal social interaction, communica-tion, and imagination [Wing & Gould, 1979], or “autistic-like” syndromes [Burd, Fisher, & Kerbeshan, 1987]. Thesesyndromes would be falling within our currently definedautistic spectrum, probably with diagnostic labels such asatypical autism and/or PDD-NOS. In surveys providingseparate estimates of the prevalence of these develop-mental disorders, higher rates for the atypical forms werefound compared with those for more narrowly definedAD [see Fombonne, 2003a, p.172]. This group receivedlittle attention in previous epidemiological studies andthese participants were not counted in the numerators ofprevalence calculations, thereby underestimating system-atically the prevalence of what would be defined today as

the spectrum of ADs. For example, in the surveys con-ducted by Lotter [Lotter, 1966] and Wing et al. [Wing,Yeates, Brierly, & Gould, 1976], prevalence figures wouldrise substantially if atypical forms had been included inthe case definition. For purpose of historical compari-sons, it is important to be attentive to this earlier figure,bearing in mind that the study was conducted in theearly 1970s for the field work and that autism occurringin participants with an IQ within the normal range wasnot yet being investigated. Progressive recognition of theimportance and relevance to autism of these less typicalclinical presentations has led to changes in the design ofmore recent epidemiological surveys (see following dis-cussion) that are now using case definitions that incor-porate upfront these milder phenotypes.

Recent Prevalence Estimates

Available studies from Northern European countries (UK,Iceland, Denmark, Sweden) provide estimates for com-bined PDD as well as AD. Much less data are availablefrom other European countries, namely from France,Germany, Portugal, and Israel. Sample sizes of multiplesurveys estimating AD varied from 826 to 490 000 par-ticipants, with an age range of birth to adulthood. Preva-lence rates varied from 1.9/10 000 to 72.6/10 000 witha median value of 10.0/10 000. In studies published since1999, the median rate was 18.75/10 000. Studies provid-ing estimate for combined PDD ranged in sample sizefrom 2536 to 134 661 participants. All these surveyswere published since 2000, and the majority since 2006.The diagnostic criteria reflect reliance on modern diag-nostic schemes. There was high variation in prevalenceproportions that ranged from a low 30.0/10 000 to ahigh of 116.1/10 000, with a median rate of 61.9/10 000.

The wide range of prevalence estimates reportedin these studies may be attributed, at least in part, tomethodological issues outlined earlier. Furthermore, suchestimates should always be regarded in the context of theimperfect sensitivity of case ascertainment that resultsin downward biases in prevalence proportions. Forexample, in the Danish investigation [Madsen et al.,2002], case finding depended upon notification to aNational Registry, a method that is usually associatedwith lower sensitivity for case finding. By contrast, case-finding techniques used in other surveys relied on mul-tiple and repeated screening phases, involving bothdifferent informants at each phase and surveying thesame cohorts at different ages, which certainly maxi-mized the sensitivity of case identification [Baird et al.,2006; Chakrabarti & Fombonne, 2005].

Clinical Presentation

Clinical characterization in Europe has increasinglyrelied on standardized measures, serving to reduce


heterogeneity in clinical judgment. Assessments in seve-ral of the studies summarized earlier were often perfor-med with these diagnostic measures that match the moredimensional approach for case definition. Extensive dis-cussion of clinical presentation has been previously pre-sented based on these studies [Fombonne, 2003a, 2005].To summarize, males consistently outnumbered femalesin the vast majority of studies, with a ratio ranging from1.33 to 16.0 for AD and 3.3 to 15.7 for PDD. In studieswhere IQ scores were available, the proportion of partici-pants with normal IQ was 15.6–86.7% for AD and45–85.3% for PDD. A few studies provided informationon the social class of the families of autistic children. Ofthese, two older studies [Brask, 1972; Lotter, 1966] sugges-ted an association between autism and social class orparental education. Studies conducted thereafter provi-ded no evidence for the association. Thus, the epidemio-logical results suggest that the earlier findings wereprobably due to artifacts in the availability of services andin the case-finding methods, as already shown in othersamples [Schopler, Andrews, & Strupp, 1979; Wing, 1980].

Some investigators have mentioned the possibility thatrates of autism might be higher among immigrants inNorthern Europe [Gillberg, 1987; Gillberg, Schaumann, &Gillberg, 1995; Gillberg, Steffenburg, & Schaumann,1991; Wing, 1980]. A meta analysis including five studiesfound weak (nonsignificant) association between autismand mother’s birthplace [in the country where the studywas conducted vs. abroad; Gardener, Spiegelman, & Buka,2009]. Serious methodological caveats related to thesecomparisons have been extensively discussed, includingsmall samples, variation in rate of immigration in theareas samples, and the lack of a reasoned biologicallyplausible hypothesis linking immigration to autism[Fombonne, 2005].

Complex sociobiological [Gillberg et al., 1995; Gillberget al., 1991] and environmental routes [Dealberto, 2011]have been suggested to explain the differences reportedacross a range of studies but none have been formallytested. In trying to ascertain factors that might explaindifferences in rates among immigrants and nonimmi-grants, a recent study in the Netherlands [Begeer, Bouk,Boussaid, Terwogt, & Koot, 2009] showed that ethnicminorities were underrepresented among cases referredto autism institutions. Yet, the study also found that biasin clinical judgment among pediatricians may lead tounder-referral. Moreover, the same bias was no longerapparent when pediatricians were asked to provideexplicit structured ratings of the probability of autism.Taken together, in view of methodological limitationsrelated to low numbers of identified cases, appropriate-ness of the comparison data that were used and largeheterogeneity in the definition of immigrant status thesefindings cannot support a difference in PDD rates amongimmigrants and nonimmigrants.

Western Pacific, South East Asia, and theEastern MediterraneanPrevalence Estimates

The oldest and most comprehensive prevalence surveysof AD in the Western Pacific Region come from Japanand China. While studies from Japan have tended to bepublished in English and have been reported in previousreviews, Chinese search engines provided a substantialbody of studies published in local journals. Studies inthis region had population size ranging from 3606 to609 848. Since 2000, prevalence rates varied from 2.8/10 000 to 94/10 000 with a median value of 11.6/10 000.The median is somewhat similar (13/10 000) when olderstudies are included. The median value does not differwhen older studies are excluded. This rate of AD is lowerthan current estimates from a comparable number ofstudies in Northern Europe but resembles estimates fromolder studies in that region. Only one estimate of AD wasavailable from South East Asia. The study conducted inIndonesia estimated the rate of AD to be 11.7/10 000[Wignyosumarto, Mukhlas, & Shirataki, 1992]. No esti-mates of AD were found in the Eastern Mediterranean.

In relation to combined PDD, three were available fromthe Western Pacific Region. One study in Australia esti-mated prevalence of PDD to be 39.2/10000 [Icasiano,Hewson, Machet, Cooper, & Marshall, 2004]. The highestrate of PDD internationally comes from a recent Koreanstudy estimating the prevalence to be 189/10 000 [Kimet al., 2011] followed by a Japanese study estimating theprevalence to be 181.1/10 000 [Kawamura, Takahashi, &Ishii, 2008]. In fact, the former study [Kim et al., 2011]also reports a high probability estimate of 264/10 000in addition to the general population estimate. By con-trast, a study from China [Wong & Hui, 2008] reported amuch lower estimate of 16.1/10 000. The Chinese studyalso reported that ASD incidence based on registry datain Hong Kong has been rather stable during 20 years.Estimated rates of 5-year incidence increased slightlyfrom 5.19 (CI: 4.43–5.94) to 7.86 (CI: 6.68–9.05) per10 000 over the study period.

Four estimate of PDD was available from South EastAsia and the Eastern Mediterranean. These were 100/10 000 in Sri Lanka [Perera, Wijewardena, & Aluth-welage, 2009], 29/10 000 in the United Arab Emirates[Eapen, Mabrouk, Zoubeidi, & Yunis, 2007], 1.4/10 000in Oman [Al-Farsi et al., 2010], and 6.3/10 000 inIran [Samadi, Mahmoodizadeh, & McConkey, 2012].Another estimate (not included in Table II) from Iran[1.9%, Ghanizadeh, 2008] was available but the studydid not confirm diagnosis in the identified group. Inaddition to these estimates, ASD incidence based onreferral statistics in Bangkok were reported to haveincreased from 1.43 per 10 000 in 1998 to 6.94 per10 000 in 2002 [Plubrukarn, Piyasil, Moungnoi, Tanpra-


sert, & Chutchawalitsakul, 2005]. Factors related tobroadening of the diagnostic concepts and increasingservice seeking discussed earlier are likely to account forthe variability in estimates as well as the observedincrease. In this study, most of incidence changesstemmed from the increase of the AD cases, not froman increase of PDD-NOS. Additionally, incidencein children older than five did not change (1.52 to 1.66)during the same period [Plubrukarn et al. 2005]. Thesefindings suggest that younger and more severe cases ofPDD rather than older and milder cases of children overfive years account for the increase. By contrast, therelatively low prevalence in the Omani and Iranianstudies was attributed to prevalence caused by underdi-agnosis and underreporting of cases resulting fromlimited access to service centers [Al-Farsi et al., 2010].

Clinical Presentation—Western Pacific

While detailed medical conditions and ASD phenotypedata were not available, several studies reported simila-rities in the PDD phenotypes within this region. Exclud-ing older studies from Japan, AD and PDD were moreprevalent in boys than girls in the epidemiological studiesin Table II. In Japan, 66.4% of participants in the studywith the highest prevalence estimate had IQ scoreswithin the normal range. By contrast, all cases with ADidentified in the population-based prevalence study inChina had intellectual disabilities. Zhang and Ji, [2005]speculated that autistic children with average cognitivefunction or mild intellectual disabilities might be neg-lected during the screening and case-ascertainmentprocess. The Tanjin study [Zhang & Ji, 2005] reported thatthe AD prevalence was higher among urban childrenthan children in suburban area (1.37% vs. 0.80%, respec-tively). Authors also noted that parents in urban areashad higher education and higher family income, leadingto better accessibility for clinical care and rehabilitationservices in these families.

Clinical Presentation—South East Asia

In addition to the two prevalence studies in Table II,a few studies characterized relatively small samplesof children in tertiary care settings in India among alarger literature on intellectual disabilities from theregion. The studies had a high but variable male-to-female ratios [Juneja, Mukherjee, & Sharma, 2004;Kalra, Seth, & Sapra, 2005; Malhi & Singhi, 2005; Singhi& Malhi, 2001]. Across studies, the concern that mostcommonly led to referral from medical professionalswas language delay or regression in language skills, fol-lowed by social difficulties and hyperactivity. Mostchildren received the diagnosis of ASD between 3 and 6years [Juneja et al., 2004; Kalra et al., 2005]. The time

between recognition of symptoms by caregivers anddiagnosis averaged about 2 years. Three Indian studiesnoted that the majority of families were from middle-class backgrounds [Daley, 2004; Juneja et al., 2004;Singhi & Malhi, 2001] and postulated that the highersocioeconomic families do not attend state-run faci-lities while the lower socioeconomic groups may notaccess care unless the child is acutely ill. Regression ofskills was found in 25% of the children in one study[Singhi & Malhi, 2001]. Seizures were associated withASD in 6.8–31% of children [Daley, 2004; Juneja et al.,2004; Kalra et al., 2005]. Intellectual disability was themost common co-morbid condition ranging from 24 to95% of children. Perinatal events were examined inthree studies [Juneja et al., 2004; Kalra et al., 2005; Malhi& Singhi, 2005]; two reported such events in up to 25%of the children whereas the third found no significantperinatal events.

Clinical Presentation—Eastern Mediterranean

Similar to other regions, a number of studies report ahigh but variable male-to-female ratios [Al-Farsi et al.,2010; Al-Salehi, Al-Hifthy, & Ghaziuddin, 2009; Eapenet al., 2007; Seif Eldin et al., 2008]. The gender ratio inthe Iranian study was more equally distributed betweenmales and females [Ghanizadeh, 2008], but the ratereported in this study most likely reflects that of PDDsymptoms among school-aged children rather than esti-mates of the disorder because a stage of diagnostic con-firmation was not conducted [Ghanizadeh, 2008]. Onestudy from Saudi Arabia reported that girls were olderthan boys in a tertiary referral center [Al-Salehi et al.,2009].

Parallel findings were observed across a number ofstudies from neighboring Arabic-speaking countries[Al-Salehi et al., 2009; Eapen et al., 2007], including onestudy that combined groups of children from nine coun-tries [Seif Eldin et al., 2008]. These studies suggest highrates of behavioral, cognitive, and/or medical problemsamong children with PDD. One study highlighted thatcommunication difficulties were a main reason for refer-ral and reported a regressive profile in 10% of cases[Al-Farsi et al., 2010; Al-Salehi et al., 2009]. Two studiesexamining family characteristics of groups of childrenwith PDD reported that the rate of consanguinity didnot differ from a control group [Seif Eldin et al., 2008] orfrom population estimates [Al-Salehi et al., 2009]. Yet, theaffected groups typically had higher parental age, inci-dence of pregnancy and labor complications [Seif Eldinet al., 2008], and incidence of familial mental healthproblems [Eapen et al., 2007; Seif Eldin et al., 2008]. Inthe Omani study, most cases identified were in the capitalMuscat and the rates were variable across the otherregions examined [Al-Farsi et al., 2010].


AmericaPrevalence Estimates

The vast majority of data from this region (Table III)comes from the United States and Canada from as earlyas the 1970s. The sample sizes of studies estimating ADwere fairly large ranging from 8896 to 4 950 333 indi-viduals. Estimates of AD varied from 0.7–40.5/10 000,with a median of 11/10 000. Studies done since 2000,however, yield estimates ranging from 11 to 50.5/10 000,with a median of 21.6/10 000. Studies estimating com-bined PDD, all of which were conducted since 2000ranged from 8896 to 407 578 individuals in the targetpopulation. PDD estimates varied from 34 to 90/10 000,with a median of 65.5/10 000. Despite huge variation inranges, both AD and PDD median estimates correspondclosely to those derived from Northern Europe. Similar tothe latter studies, a likely underestimation of the truepopulation rates holds in this region as well. For example,the Atlanta survey by the CDC [Yeargin-Allsopp et al.,2003] was based on a very large population and includedyounger age groups than subsequent CDC surveys, andage specific rates were in fact in the 40–45/10 000 rangein some birth cohorts [Fombonne, 2003b].

Less data were available from other countries inAmerica. Two published studies from Argentina [Lejar-raga et al., 2008] and Venezuela [Montiel-Nava & Peña,2008] along with unpublished reports from Brazil [Paula,Ribeiro, Fombonne, & Mercadante, 2011] and Mexico[Marcín, personal communication] provided prevalencedata (Table III). There was also one study from the Carib-bean (Aruba) [van Balkom et al., 2009]. All of the studieshad relatively small sample sizes. The two studies fromVenezuela and Aruba provided similar AD estimates of13 and 19/10 000. With the exception of one study con-ducted in 1982, the remaining estimates of PDD rangedfrom 27 to 39.2/10 000. Administrative prevalence wasavailable from Mexico through a registry of minors withdisabilities reporting the number of children identifiedwith autism, combined with Mexican census figures.Based on these figures the rate for childhood autism inMexico was estimated to be 14.3 per 10 000 [Marcín,personal communication]. The estimates in those studiesare smaller than those reviewed earlier but it is difficultto compare the findings given the limited data andmethodological differences among studies.


Similar to Northern Europe, clinical characterization ofPDD in the United States and Canada has increasinglyrelied on standardized measures and have been previous-ly reviewed [Fombonne, 2003a, 2005]. Across all studiesreported in Table III, the male-to-female ratio ranged from2.2 to 5.8 for AD and from 2.7 to 6.7 for PDD. In studiesreporting IQ scores, 23.8–62.5% of participants with AD

and 31.8–60.0% of participants with PDD were reportedto have normal IQ, with the higher estimate correspond-ing to the United States (South Carolina) and Aruba.

The largest studies internationally to utilize comparablemethodology for examining variation of prevalenceamong neighboring geographical regions were conductedby the CDC in the United States. Findings from geo-graphic regions falling within 10 American states founda threefold variation of rate by state, where Alabama hadthe lowest rate (3.3/1000) whereas New Jersey had thehighest (10.6/1000) [CDC, 2007b]. Subsequent CDCsurveys found a similar variation of rates but the lowestwas in Florida (4.2/1000) and highest in Missouri andArizona (12.1/1000) [CDC, 2009]. The CDC surveys foundthat prevalence estimates were lower in sites that reliedsolely on health sources to identify cases compared withsites that also had access to education records, and hencea less rich database for those areas where ASD indicatorscould be found. As surveillance efforts continue, it is likelythat awareness and services will develop in states thatwere lagging behind, resulting in a predictable increase inthe average rate for the United States as time elapses.

CDC surveys also monitor variation in prevalence esti-mates among different ethnic communities. In the lastsurvey [CDC, 2009] prevalence among white childrenwas greater than that among black and Hispanic children.Similar to other world regions, service availability mayaccount for the observed differences. In support of this,prevalence has increased over time across all sex, racial/ethnic, and cognitive functioning subgroups, making itunlikely that one group is differentially affected. Variabil-ity of estimates as a function of income level has alsobeen noted. As an example, a study conducted in Texasshowed that the top decile of income had a 6 timesgreater administrative prevalence than the bottom decile[Palmer, Blanchard, Jean, & Mandell, 2005].

AfricaPrevalence Estimates

As early as 1970s, a report describing autism amongchildren with mental handicaps who were known toauthorities in six central and south African countriesreported a preliminary observation that the number ofcases of autism maybe smaller than those observed inthe UK [Lotter, 1978]. Findings from this preliminaryreport, which was not designed as an epidemiologicalsurvey, were never corroborated. At the time, however,this served as crucial evidence that autism can be identi-fied across varied geographic region. Similar to the con-clusion of a recent review on autism in Africa [Bakare &Munir, 2011], we could not identify published data onpopulation-based estimates of prevalence of PDD fromAfrican region.



A few studies [reviewed in Ametepee & Chitiyo, 2009]including Lotter’s [Lotter, 1978], documented unambigu-ous recognition of autism among African children, albeitraising some questions about the relative frequency ofsome of the manifestations [Lotter, 1978; Mankoski et al.,2006]. A higher male-to-female ratio was consistent[Khan & Hombarume, 1996; Lotter, 1978; Mankoskiet al., 2006]. Similar to observations from other regions,those studies suggested an over representation of highersocioeconomic background and higher frequency of caseswith intellectual difficulties, i.e., the children identifiedwere mainly nonverbal. Yet, none of the studies usedappropriate tools for the evaluation of cognitive abilitiesand similar arguments related to service accessibilityapply here as well. In one study, 3/14 cases described weredescribed as having developed autism after the secondyear subsequent to a severe malaria infection [Mankoskiet al., 2006]. The findings, however, were never repli-cated with appropriate sample sizes or with a prospectiverather than retrospective methodology to account forthe possibility of biased recall. In contrast to the dearth ofevidence, there has been an abundance of claims regard-ing the prevalence and etiology of the autism in Africa[reviewed in Ametepee & Chitiyo, 2009; Bakare & Munir,2011]. The clinical presentation of the condition in thisregion remains elusive.

PDD in the Global Context: Prevalence,Determinants, and BurdenGlobal Variation in Prevalence Estimates

Epidemiological surveys of autism and AD and PDD(summarized inFigs 1 and 2) have now been conducted inseveral countries. Methodological differences in case defi-nition and case-finding procedures make between surveycomparisons difficult to perform. However, most studiesconducted since the year 2000 in different geographicalregions and by different teams converge to estimates to amedian of 17/10 000 for AD and 62/10 000 for all PDDscombined (Table IV). This is currently the best estimatefor the prevalence of PDDs available. The estimate repre-sents an average figure and there is substantial variabilityacross studies.

Notwithstanding considerable variability across and wi-thin geographical regions, estimates of AD collected sincethe year 2000 in America, Western Pacific, and Europe donot differ statistically (P = 0.3). While considerably lessstudies have been conducted on PDD, those estimates donot differ between America and Europe. The global meanprevalence of 62/10 000 translates into one child out of160 with a PDD. Some well-controlled studies have,however, reported rates that are two to three times higher[Baird et al., 2006; Kawamura et al., 2008; Kim et al.,

2011]. By contrast, in many regions of the world includ-ing Africa, prevalence estimates are either unavailable orpreliminary. Moreover, within each of the regions cove-red, the majority of studies were conducted in circum-scribed regions, namely, in northern Europe, Japan,China, and the United States. With the exception ofChina, countries where a relatively large evidence baseexists are high-income countries. A few studies have beenconducted in middle-income countries and no prevalenceestimates were available from any low-income country.

Our minimal criteria for inclusion of prevalence esti-mates described in the previous section were motivatedby the need for a systematic and comprehensive accountof available estimates. However, given the significantvariability in resulting prevalence estimates, and the vari-ability in methodological features outlined earlier, theestimates should be considered with caution. Specifically,these should not be considered as reflecting “world-wide”estimates, but rather, a general reflection of the currentstate of evidence from various regions. As we note in thefinal section, our findings offer critical public healthmessages regarding the need for scaling of services as wellas research in most of the world’s regions.

Phenotypic presentation of PDD

As early as the 1970s, the phenotypic characteristics ofautism have been observed and described across a widerange of geographical, economic, social and cultural set-tings. Based on available evidence world-wide, the higherratio in males relative to females is consistently observed.It is also clear that PDD is associated with a degree ofintellectual impairment among a significant portion ofthose affected. Yet, the range of variation in gender ratioand intellectual functioning is vast. By contrast, very fewstudies examined the impact geographical, economic,social and cultural factors directly and, on the whole, thefindings indicate little impact. Differences in such fac-tors reported retrospectively in large-scale surveys suchas those conducted by the CDC [CDC, 2007a, 2007b] maybe accounted for entirely by differences in service provi-sions, richness of registered data, and community aware-ness. Moreover, there is a striking absence of controlledstudies examining the potential impact of such factorson the understanding, experience, expression, and help-seeking for autism, masking the true global burden of thecondition. Some positive indicators suggest improvedidentification of autism in low-resource countries, asinferred from the increase in referrals and identified cases[e.g., Plubrukarn et al., 2005]. Yet, it has also been sug-gested that regional variation in parental educationmay still determine their help seeking for autism [SeifEldin et al., 2008] and that it is predominantly severecases that are currently being identified in poor commu-nities [Daley, 2004; Juneja et al., 2004; Singhi & Malhi,


Figure 1. Prevalence of autistic disorder from Tables I and II (rate/10 000 and 95% confidence interval).


Figure 2. Prevalence of PDD from Tables I and II (rate/10 000 and 95% confidence interval).


2001]. Such possibilities require confirmation in futurelarge-scale controlled studies.

The data we reviewed is also relevant for the ongoingdebate regarding time trends in autism. Consideringall studies conducted since the 1960s, prevalence datasuggest that there is a correlation between AD estimatesand year of publication (r = 0.4, P < 0.01, R2 = 18.5%).While this simple test is limited by the considerablevariation, it does confirm that when data is consideredfrom several world regions there is indeed a rise inprevalence over time, confirming previous conclusions.Such time trends have given rise to the hypothesis thatthere is a secular increase in rates of autism. However inview of a number of conceptual and methodologicalreasons, no inference related to trends in the incidenceof PDDs can be derived from a simple comparison ofprevalence rates over time. In particular, it is crucial todifferentiate prevalence (the proportion of individuals ina population who suffer from a defined disorder at anyone point in time) from incidence (the number of newcases occurring in a population over a period of time).Prevalence is useful to estimate needs and plan services;only incidence rates can be used for causal research.Both prevalence and incidence estimates will increasewhen case definition is broadened and case ascertain-ment is improved. Time trends in rates can thereforeonly be gauged in investigations that hold these param-eters under strict control over time. Several factorsaccounting for this rise have been proposed and dis-cussed extensively in previous studies. These factorsinclude: (a) the broadening of diagnostic criteria; (b) theincreased efficiency over time in case identificationmethods used in surveys as well as changes in diagnosticpractices; and (c) the diagnostic substitution (or switch-ing) when some diagnostic categories become increas-ingly familiar to health professionals and/or when accessto better services is insured by using a new diagnosticcategory.

The Way ForwardProspects and Challenges in Global EpidemiologicalResearch

Because significant costs are associated with prevalencestudies, some have questioned whether country-specificestimates should be a priority, especially where resourcesare limited and priorities include preventable life-threatening conditions. Such epidemiological studies areuseful, however, for assessing needs and priorities withineach community. Epidemiological studies often providesystematic information regarding the availability, quality,and accessibility of existing services. These studies alsorequire development of valid tools for systematic clinicalcharacterization, including screening and/or diagnosticinstruments that can be useful for immediate improve-ments in training, services, and awareness, as well asfacilitating future research. As such, epidemiologicalresearch may be viewed as a systematic framework forunveiling local burden and informing policy and research.

The power of epidemiological research extends wellbeyond the communities in which it is conducted, par-ticularly in view of mounting speculation regarding timetrends and the potential impact of geographical andsocioeconomic factors on the prevalence and possibly onthe incidence of autism. While research within a globalcontext is likely to help address such perplexing puzzles,currently, available evidence is extremely limited. Thereis also a clear need for better theoretical and experimentalmodelling to test the possible mechanisms throughwhich various factors may have an impact on incidence,clinical presentation, or both. Studies using comparablemethodology to estimate prevalence across different geo-graphical regions and to monitor changes to it over timeprovide a powerful approach. Such findings may addresswhether certain factors, environmental or otherwise,operating in specific regions, have a disproportionateimpact on prevalence over time as well as ascertain the

Table IV. Summary of Prevalence Estimates of AD and PDD Across World Regions Since the Year 2000. Medians are not PresentedWhen there were Too Few Estimates Available Within a Given Region

Region

AD estimates PDD estimates

Median Range Number of estimates Median Range Number of estimates

Europe 19 7–39 16 62 30–116 14America 22 11–40 7 65 13–110 12Western Pacific 12 2.8–94 12 — — 3South East Asia — — 1 — — 1Eastern Mediterranean — — 0 — — 3Africa — — 0 — — 0All 17 2.8–94 36 62 1–189 33

AD, autistic disorder; PDD, developmental disorder.


true global burden of the condition. Moreover, epidemio-logical studies can be potentially translated into publichealth practices through developing or validating screen-ing and diagnostic methods and though strengtheningcapacity for services.

Rapid developments in the public policy context forautism in various regions have given rise to promisingopportunities for furthering research in this area. A num-ber of parent and professional led groups as well as gov-ernmental agencies have been established around theworld, who are actively working on creating awareness,services, and advocacy for people with PDD. While thecritical need for population-based studies of the epidemi-ology of PDD may be evident, it is essential to also con-sider the likely reasons for the current lack of such data.

Funding for research in some countries is dispropor-tionate to the burden of disease they face (referred to asthe 10/90 gap, that is, only 10% of global spending onhealth research is directed towards the problems thatprimarily affect the poorest 90% of the world’s popula-tion). There is limited local capacity to conduct researchin these settings, let alone research in field of childhoodautism. Despite this situation being common to severallow- and middle-income countries, our review high-lighted that regional or country-specific economic anddevelopment capacity only one factor contributing tothe scarcity of data. More general contextual difficultiesmay be encountered in conducting epidemiologicalsurveys worldwide. First, the epidemiological researchconcept of PDD is relatively new, thus little cumulativeinformation on epidemiology of PDD are available.Second, service availability/accessibility in many regionsremains low, which creates problems for case finding.Third, awareness of PDD among the public as well asclinical experts remains limited. Such contextual factorsmay help explain why epidemiological data is limitedin some countries within Western Europe, where psycho-analytic frameworks for management of PDD may still beoperative, as they were in the 1960s and 1970s in otherregions of Europe and North America (and remain opera-tive is some parts of Europe). Hence, there is a need forcontinued educational, training, and awareness programswith specific targets tailored to the needs and resourcesavailable for each community.

Further challenges face the research community inparticular. Diagnosis of PDD still relies on behavioralobservation, leaving room for wide variation in clinicaljudgment, and in cultural and social impact in theidentification of biological conditions. While thesechallenges have been addressed in some countries withstrong research capacity through the development andstandardization of diagnostic tools, progress has beenslow in using such instruments widely within thosecountries themselves, let alone, within a global context,where significant cultural and linguistic variation across

and within communities needs to be accounted for. Anumber of instruments have been or are in the processof being translated and/or validated. Translation ofinstruments has also been identified as a priority withinthe International Epidemiology Network (http://www.autismepidemiology.net). Yet the biggest challenge todescribing the true burden of PDD is the availability oflow cost tools for epidemiological surveys, which wouldallow those to be conducted with adequate samples indifferent world regions.

Further questions are often raised as to whetherresearch efforts need to focus on PDD in particular orincorporate a wider range of disabilities. For example,a group of scientists and practitioners, led by INCLEN-India (http://www.inclentrust.org), have been developingand validating a neurodevelopmental screening testfor use by community health workers for the detection of10 neurodevelopmental disabilities (including ASD). Thiswork is leading to the emergence of a collaborative frame-work for research on neurodevelopmental disorders inthe region to carry out multicenter epidemiologicalstudies with a common protocol. The availability of thesetools will provide a new opportunity for studies of theburden, determinants and needs of families affected byPDD in the region. Epidemiological studies themselvesmay also eventually provide revealing findings regardingthe manifestation of PDD across cultures. In view of theincreased awareness of autism worldwide and the interestfrom a wide range of stakeholders, research in this rela-tively narrow area may become a vehicle for broaderimprovements in evidence, and consequently in practicestandards, in child mental health in general.

Public Health Implications

Based on already available knowledge indicating thatPDDs are not limited to high-income countries; the needfor services especially in low and middle-income coun-tries is felt more than before. It is imperative to engagecommunity resources and more peripheral extensions ofhealth systems as well. The situation in low- and middle-income countries appears to be that child health pro-grams focus mainly on child survival issues. Very littleattention is paid to developmental disabilities at policyand implementation level and as a result budget alloca-tions and human resource deployment is directed awayfrom these programs. For example, some regions in Africaface a dual burden of communicable and noncom-municable conditions, including childhood disabilities[Murray, Lopez, Mathers, & Stein, 2001]. Dysfunctionalhealth systems contribute further to lack of service deliv-ery for children with developmental disabilities. Wherethey exist, access to these facilities is also hindered bylack of effective identification and referral programs. Aneven bigger barrier has been the lack of an evidence-


based, affordable, package of care for people with PDD,which could be delivered if such facilities were available.Only recently WHO’s mhGAP has included evidence-based and simple interventions for PPD and intellectualdisability in a package that needs to be pilot tested andevaluated in low- and middle-income countries. Thisreview indicates that increase in prevalence estimates overtime most likely represent broadening of the diagnosticconcepts, diagnostic switching from other developmentaldisabilities to PDD, service availability, and awareness ofautistic spectrum disorders in both the lay and profes-sional public. This has clear public health implicationsthat services for different categories of developmentaldisorders should not be segregated prematurely andhealth systems need to consider overarching programmesto fill the gap for all developmental disorders.

Acknowledgments

WHO commissioned and supported this review. How-ever, the views expressed in this article do not necessarilyrepresent the decisions, policy, or views of WHO. Weappreciate the Shirley Foundation, Autism Speaks, andAutistica for supporting this work. We would like tothank Eileen Hopkins from Autistica, Andy Shih, MichaelRosanoff, and Jessica Napolitano from Autism Speaksand all the professionals and researchers who provideda wealth of information included in the current review.We are also grateful to Rachel Wu for valuable assistancein preparing this article.

References

Abrahams, B.S., & Geschwind, D.H. (2008). Advances in ASDgenetics: on the threshold of a new neurobiology. NatureReviews Genetics, 9, 341–355.

Al-Farsi, Y., Al-Sharbati, M., Al-Farsi, O., Al- Shafaee, M., Brooks,D., & Waly, M. (2010). Brief report: prevalence of autisticspectrum disorders in the Sultanate of Oman.

Al-Salehi, S.M., Al-Hifthy, E.H., & Ghaziuddin, M. (2009).Autism in Saudi Arabia: presentation, clinical correlates andcomorbidity. Transcultural Psychiatry, 46, 340–347.

American Psychiatric Association. (1980). Diagnostic and statis-tical manual of mental disorders, 3rd ed. Arlington, VA:American Psychiatric Association.

American Psychiatric Association. (1987). Diagnostic and statis-tical manual of mental disorders, 3rd ed. Revised. Arlington,VA: American Psychiatric Association.

American Psychiatric Association. (1994). Diagnostic and statis-tical manual of mental disorders, 4th ed. Arlington, VA:American Psychiatric Association.

American Psychiatric Association. (2000). Diagnostic and statis-tical manual of mental disorders, 4th ed. Arlington, VA:American Psychiatric Association.

Ametepee, L.K., & Chitiyo, M. (2009). What we know aboutautism in Africa: a brief research synthesis. The Journal of theInternational Association of Special Education, 10, 11–13.

Arvidsson, T., Danielsson, B., Forsberg, P., Gillberg, C., Johans-son, M., & Kjellgren, G. (1997). Autism in 3–6 year-old in asuburb of Goteborg, Sweden. Autism, 2, 163–173.

Baird, G., Simonoff, E., Pickles, A., Chandler, S., Loucas, T., et al.(2006). Prevalence of disorders of the autism spectrum in apopulation cohort of children in South Thames: the specialneeds and autism project (SNAP). Lancet, 368, 210–215.

Bakare, M.O., & Munir, K.M. (2011). Autism spectrum disorders(ASD) in Africa: a perspective. African Journal of Psychiatry(Johannesbg), 14, 208–210.

van Balkom, I.D.C., Bresnahan, M., Vogtländer, M.F., vanHoeken, D., Minderaa, R., et al. (2009). Prevalence of treatedautism spectrum disorders in Aruba. Journal of Neurodevel-opmental Disorders, 1, 197–204.

Begeer, S., Bouk, S.E., Boussaid, W., Terwogt, M.M., & Koot, H.M.(2009). Underdiagnosis and referral bias of autism in ethnicminorities. Journal of Autism and Developmental Disorders,39, 142–148.

Belmonte, M.K., Cook, E.H.J., Anderson, G.M., Rubenstein, J.L.,Greenough, W.T., et al. (2004). Autism as a disorder of neuralinformation processing: directions for research and targets fortherapy. Molecular Psychiatry, 9, 646–663.

Brask, B. (1972). A prevalence investigation of childhood psychoses.Paper presented at the Nordic Symposium on the Care ofPsychotic Children.

Burd, L., Fisher, W., & Kerbeshan, J. (1987). A prevalence study ofpervasive developmental disorders in North Dakota. Journalof the American Academy of Child and Adolescent Psychia-try, 26, 700–703.

CDC. (2007a). Center for Disease Control: Prevalence of autismspectrum disorder—autism and developmental disabilitiesmonitoring network, six sites, United States (2000). MMWRSurveillance Summary, 56, 1–11.

Center for Disease Control. (2007b). Prevalence of autism spec-trum disorders—autism and developmental disabilities moni-toring network, 14 sites, United States (2002). MMWRSurveillance Summary, 56, 12–28.

Center for Disease Control. (2009). Prevalence of autismspectrum disorders—Autism and Developmental Disabi-lities Monitoring network, United States, 2006. Morbidityand Mortality Weekly Report Surveillance Summary, 58,1–14.

Chakrabarti, S., & Fombonne, E. (2001). Pervasive developmen-tal disorders in preschool children. Journal of the AmericanMedical Association, 285, 3093–3099.

Chakrabarti, S., & Fombonne, E. (2005). Pervasive developmen-tal disorders in preschool children: confirmation of highprevalence. American Journal, 162, 1133–1141.

Croen, L.A., Grether, J.K., Hoogstrate, J., & Selvin, S. (2002). Thechanging prevalence of autism in California. Journal ofAutism and Developmental Disorders, 32, 207–215.

Currenti, S.A. (2009). Understanding and determining the etiol-ogy of autism. Cellular & Molecular Neurobiology, 30, 161–171.

Daley, T.C. (2004). From symptom recognition to diagnosis:children with autism in urban India. Social Science and Medi-cine, 58, 1323–1335.

Dealberto, M.J. (2011). Prevalence of autism according to mater-nal immigrant status and ethnic origin. Acta PsychiatricaScandinavica, 123, 339–348.


Eapen, V., Mabrouk, A.A., Zoubeidi, T., & Yunis, F. (2007). Preva-lence of pervasive developmental disorders in preschoolchildren in the UAE. Journal of Tropical Pediatrics, 53, 202–205.

Elsabbagh, M., & Johnson, M.H. (2010). Getting answersfrom babies about autism. Trends in Cognitive Science, 14,81–87.

Farley, M., McMahon, W., Fombonne, E., Jenson, W.R., Miller, J.,et al. (2009). Twenty-year outcome for individuals withautism and average or near-average cognitive abilities. AutismResearch, 2, 109–118.

Fombonne, E. (2003a). Epidemiological surveys of autism andother pervasive developmental disorders: an update. Journalof Autism and Developmental Disorders, 33, 365–382.

Fombonne, E. (2003b). The prevalence of autism. Journal of theAmerican Medical Association, 289, 1–3.

Fombonne, E. (2005). Epidemiological studies of pervasivedevelopmental disorders. In F.R. Volkmar, R. Paul, A. Klin, &D. Cohen (Eds.), Handbook of autism and pervasive deve-lopmental disorders (pp. 42–69). Hoboken, NJ: Wiley.

Fombonne, E., Quirke, S., & Hagen, A. (2011). Epidemiology ofpervasive developmental disorders. In D.G. Amaral, G.Dawson, & D.H. Geschwind (Eds.), Autism spectrum disor-ders Oxford: Oxford University Press.

Fombonne, E., Zakarian, R., Bennett, A., Meng, L., & McLean-Heywood, D. (2006). Pervasive developmental disorders inMontreal, Quebec, Canada: prevalence and links with immu-nizations. Pediatrics, 118, e139–e150.

Ganz, M.L. (2007). The lifetime distribution of the incrementalsocietal costs of autism. Archives of Pediatric and AdolescentMedicine, 161, 343–349.

Gardener, H., Spiegelman, D., & Buka, S.L. (2009). Prenatal riskfactors for autism: comprehensive meta-analysis. BritishJournal of Psychiatry, 195, 7–14.

Ghanizadeh, A. (2008). A preliminary study on screening preva-lence of pervasive developmental disorder in school childrenin Iran. Journal of Autism and Developmental Disorders, 38,759–763.

Gillberg, C. (1987). Infantile autism in children of immigrantparents: a population-based study from Göteborg, Sweden.British Journal of Psychiatry, 150, 856–858.

Gillberg, C., Schaumann, H., & Gillberg, I.C. (1995). Autism inimmigrants: children born in Sweden to mothers born inUganda. Journal of Intellectual Disability Research, 39, 141–144.

Gillberg, C., Steffenburg, S., & Schaumann, H. (1991). Is autismmore common now than ten years ago? British Journal ofPsychiatry, 158, 403–409.

Gurney, J.G., Fritz, M.S., Ness, K.K., Sievers, P., Newschaffer, C.J.,& Shapiro, E.G. (2003). Analysis of prevalence trends ofautism spectrum disorder in Minnesota [comment]. Archivesof Pediatrics and Adolescent Medicine, 157, 622–627.

Honda, H., Shimizu, Y., Misumi, K., Niimi, M., & Ohashi, Y.(1996). Cumulative incidence and prevalence of childhoodautism in children in Japan. British Journal of Psychiatry,169, 228–235.

Honda, H., Shimizu, Y., & Rutter, M. (2005). No effect of MMRwithdrawal on the incidence of autism: a total populationstudy. Journal of Child Psychology and Psychiatry, 46, 572–579.

Icasiano, F., Hewson, P., Machet, P., Cooper, C., & Marshall, A.(2004). Childhood autism spectrum disorder in the Barwonregion: a community based study. Journal of Paediatric ChildHealth, 40, 696–701.

Juneja, M., Mukherjee, S.B., & Sharma, S. (2004). A descriptivehospital based study of children with autism. Indian Pediat-rics, 42, 453–458.

Kalra, V., Seth, R., & Sapra, S. (2005). Autism—experiences in atertiary care hospital. Indian Journal of Pediatrics, 72, 227–230.

Kawamura, Y., Takahashi, O., & Ishii, T. (2008). Reevaluating theincidence of pervasive developmental disorders: impact ofelevated rates of detection through implementation of anintegrated system of screening in Toyota, Japan. Psychiatryand Clinical Neuroscience, 62, 152–159.

Khan, N., & Hombarume, J. (1996). Levels of autistic behavioramong the mentally handicapped children in Zimbabwe.Central African Journal of Medicine, 42, 36–39.

Kielinen, M., Linna, S.L., & Moilanen, I. (2000). Autism in north-ern Finland. European Child and Adolescent Psychiatry, 9,162–167.

Kim, Y.S., Leventhal, B.L., Koh, Y.J., Fombonne, E., Laska,E., et al. (2011). Prevalence of autism spectrum disorders in atotal population sample. American Journal of Psychiatry, 168,904–912.

Knapp, M., Romeo, R., & Beecham, J. (2007). The economicconsequences of autism in the UK. London: Foundation forPeople with Learning Disabilities.

Lazoff, T.L., Piperni, T., & Fombonne, E. (2010). Prevalence ofpervasive developmental disorders among children at theEnglish Montreal School Board. Canadian Journal of Psychia-try, 55, 715–720.

Le Couteur, A., Rutter, M., Lord, C., Rios, P., Robertson, S., et al.(1989). Autism Diagnostic Interview: a standardizedinvestigator-based instrument. Journal of Autism and Devel-opmental Disorders, 19, 363–387.

Lejarraga, H., Menendez, A.M., Menzano, E., Guerra, L., SilviaBiancato, S., et al. (2008). Screening for developmentalproblems at primary care level: a field programme in SanI-sidro, Argentina. Pediatric and Perinatal Epidemiology, 22,180–187.

Lord, C., Risi, S., Lambrecht, L., Cook, E.H., Jr, Leventhal, B.L.,et al. (2000). The autism diagnostic observation schedule-generic: a standard measure of social and communicationdeficits associated with the spectrum of autism. Journal ofAutism and Developmental Disorders, 30, 205–223.

Lotter, V. (1966). Epidemiology of autistic conditions in youngchildren: I. Prevalence. Social Psychiatry, 1, 124–137.

Lotter, V. (1978). Childhood autism in Africa. Journal of ChildPsychology and Psychiatry, 19, 231–244.

Madsen, K.M., Hviid, A., Vestergaard, M., Schendel, D., Wohl-fahrt, J., et al. (2002). A population-based study of measles,mumps, and rubella vaccination and autism. New EnglandJournal of Medicine, 347, 1477–1482.

Magnússon, P., & Saemundsen, E. (2001). Prevalence of autism inIceland. Journal of Autism and Developmental Disorders, 31,153–163.

Malhi, P., & Singhi, S. (2005). Patterns of development in youngchildren with autism. Indian Journal of Pediatrics, 72, 553–556.


Mankoski, R.E., Collins, M., Ndosi, N.K., Mgalla, E.H., Sarwatt,V.V., & Folstein, S.E. (2006). Etiologies of autism in a caseseries from Tanzania. Journal of Autism Developmental Dis-orders, 36, 1039–1051.

Montiel-Nava, C., & Peña, J. (2008). Epidemiological findings ofpervasive developmental disorders in a Venezuelan Study.Autism, 12, 191–202.

Murray, C.J.L., Lopez, A.D., Mathers, C.D., & Stein, C. (2001).The Global Burden of Disease 2000 project: aims, methodsand data sources. Geneva: World Health Organization.

Palmer, F., Blanchard, S., Jean, C.R., & Mandell, D.S. (2005).School district resources and identification of children withautistic disorder. American Journal of Public Health, 95, 125–130.

Patel, V., Garrison, P., de Jesus Mari, J., Minas, H., Prince, M., &Saxena, S. (2008). Advisory group of the Movement forGlobal Mental Health. The Lancet’s series on global mentalhealth: 1 year on. Lancet, 372, 1354–1357.

Paula, C.S., Ribeiro, S., Fombonne, E., & Mercadante, M.T.(2011). Brief report: Prevalence of pervasive developmentaldisorder in Brazil: a pilot study. Journal of Autism and Devel-opmental Disorders., 41, 1738–1741.

Perera, H., Wijewardena, K., & Aluthwelage, R. (2009). Screeningof 18-24-month-old children for autism in a semi-urban com-munity in Sri Lanka. Journal of Tropical Pediatrics, 55, 402–405.

Plubrukarn, R., Piyasil, V., Moungnoi, P., Tanprasert, S., &Chutchawalitsakul, V. (2005). Trend study of autisticspectrum disorders at Queen Sirikit National Institute of ChildHealth. Journal of Medical Association of Thailand, 88, 891–897.

Powell, J., Edwards, A., Edwards, M., Pandit, B., Sungum-Paliwal,S., & Whitehouse, W. (2000). Changes in the incidence ofchildhood autism and other autistic spectrum disorders inpreschool children from two areas of the West Midlands, UK.Developmental Medicine and Child Neurology, 42, 624–628.

Rutter, M. (1970). Autistic children: infancy to adulthood. Semi-nars in Psychiatry, 2, 435–450.

Samadi, S.A., Mahmoodizadeh, A., & McConkey, R. (2012). Anational study of the prevalence of autism among five-year-old children in Iran. Autism, 16, 5–14.

Schopler, E., Andrews, C.E., & Strupp, K. (1979). Do autisticchildren come from upper-middle-class parents. Journal ofAutism and Developmental Disorders, 9, 139–151.

Seif Eldin, A., Habib, D., Noufal, A., Farrag, S., Bazaid, K., et al.(2008). Use of M-CHAT for a multinational screening of

young children with autism in the Arab countries. Interna-tional Review of Psychiatry, 20, 281–289.

Singhi, P., & Malhi, P. (2001). Clinical and neurodevelopmentalprofile of young children with autism. Indian Pediatrics, 38,384–390.

Sun, X., & Allison, C. (2010). A review of the prevalence ofAutism Spectrum Disorder in Asia. Research in Autism Spec-trum Disorders, 4, 156–167.

Webb, E., Lobo, S., Hervas, A., Scourfield, J., & Fraser, W. (1997).The changing prevalence of autistic disorder in a Welshhealth district. Developmental Medicine and Child Neurol-ogy, 39, 150–152.

Wignyosumarto, S., Mukhlas, M., & Shirataki, S. (1992). Epide-miological and clinical study of autistic children in Yogy-akarta, Indonesia. Kobe Journal of Medical Sciences, 38, 1–19.

Williams, J.G., Higgins, J.P., & Brayne, C.E. (2006). Systematicreview of prevalence studies of autism spectrum disorders.Archives of Disease in Childhood, 91, 8–15.

Wing, L. (1980). Childhood autism and social class: a question ofselection? British Journal of Psychiatry, 137, 410–417.

Wing, L., & Gould, J. (1979). Severe impairments of social inter-action and associated abnormalities in children: epidemiol-ogy and classification. Journal of Autism and DevelopmentalDisorders, 9, 11–29.

Wing, L., Yeates, S., Brierly, L., & Gould, J. (1976). The preva-lence of early childhood autism: comparison of administra-tive and epidemiological studies. Psychological Medicine, 6,89–100.

Wong, V.C., & Hui, S.L. (2008). Epidemiological study of autismspectrum disorder in China. Journal of Child Neurology, 23,67–72.

World Health Organization (WHO). (1975). The ICD-9 Classifi-cation of Mental and Behavioural Disorders. Clinical descrip-tions and diagnostic guidelines. Geneva.

World Health Organization (WHO). (1992). The ICD-10 Classi-fication of Mental and Behavioural Disorders. Clinicaldescriptions and diagnostic guidelines. Geneva.

World Health Organization (WHO). (2008). mhGAP MentalHealth Gap Action Programme Scaling up care for mental,neurological, and substance use disorders. Geneva.

Yeargin-Allsopp, M., Rice, C., Karapurkar, T., Doernberg, N.,Boyle, C., & Murphy, C. (2003). Prevalence of autism in a USmetropolitan area [comment. Journal of the AmericanMedical Association, 289, 49–55.

Zhang, X., & Ji, C.Y. (2005). Autism and mental retardationof young children in China. Biomedical and EnvironmentalSciences, 18, 334–340.


global prevalence of autism and other pervasive developmental disorders

Health & Medicine