Original article

Genotype and early development in Rett syndrome:

The value of international data

Helen Leonarda,*, Hannah Moorea, Mary Careya, Susan Fyfeb, Sonj Hallc, Laila Robertsona,

Xi Ru Wud, Xinhua Baod, Hong Pand, John Christodouloue,f,

Sarah Williamsone,f, Nick de Klerka

aTelethon Institute for Child Health Research, Centre for Child Health Research, The University of Western Australia, Perth, Western AustraliabCurtin University of Technology, Perth, Western Australia

cSchool of Population Health, The University of Western Australia, Perth, Western AustraliadPeking University First Hospital, Beijing, China

eDiscipline of Paediatrics and Child Health, University of Sydney, New South Wales, AustraliafWestern Sydney Genetics Program, Children’s Hospital at Westmead, New South Wales, Australia

Received 17 August 2004; received in revised form 1 October 2004; accepted 5 March 2005

Abstract

Background: Rett syndrome is a neurodevelopmental disorder mostly affecting females and caused by mutations in the MECP2 gene.

Originally the syndrome was characterised as having a normal prenatal and perinatal period with later regression. Previous work has speculated

that the girl with Rett syndrome may not be normal at birth. Aims: to examine whether early development between birth and ten months varies

by genotype in Rett syndrome. Methods: cases were sourced from two databases, the Australian Rett Syndrome Database (est. 1993) and the

newly formed InterRett - IRSA Rett Phenotype Database. Data available on 320 cases included information provided by parents on perinatal

problems, early developmental behaviour and mobility. Problem scores, mobility scores and a total composite score for each mutation were

generated and compared. Results: overall, 58% of respondents noted unusual behaviour during the first six months and 70.6% from the period

between 6 and 10 months of life. Statistically significant differences were detected between some of the common mutations. Infants with R294X

(PZ0.05) and R133C (PZ0.03) were less likely than those with R255X to have problems in the perinatal period. The most severe profile

overall for early development was associated with mutations R255X and R270X. Conclusion: This is the largest study to date examining the

effects of individual mutations in Rett syndrome. With the ongoing case ascertainment and expansion of InterRett, sample size will increase

rapidly and provide improved statistical power for future analyses. Results from this study will contribute to understanding the mechanism of

early development in Rett syndrome and determining if and at which time(s) early intervention might be feasible.

q 2005 Elsevier B.V. All rights reserved.

Keywords: Rett syndrome; MECP2; Phenotype; Early development; International; Internet; Database

1. Introduction

The clinical criteria for Rett syndrome identified by

Hagberg et al. [1] and then revised by Trevarthan and Moser

[2] characterised Rett syndrome as a neurodevelopmental

disorder with a normal prenatal and perinatal period and

early development in the normal range with a later

0387-7604/$ - see front matter q 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.braindev.2005.03.023

* Corresponding author. Tel.: C61 8 9489 7790; fax: C61 8 9489 7700.

E-mail address: hleonard@ichr.uwa.edu.au (H. Leonard).

regression period. However, it has become apparent that

girls with Rett syndrome may not be normal at birth [3,4]

and that there are subtle differences which indicate

developmental delay early in life [5,6]. Early hypotonia

[7–9], motor problems [10,11] including early jerky

incoordination [7], placidity and perinatal difficulties

requiring admission to a special care nursery [3] have

been noted in the pre-regression period.

Since the association between mutations in MECP2 and

Rett syndrome was recognized [12], many studies have tried

to characterise the relationship between genotype and

clinical severity. In a sample of 56 girls Amir et al. [13]

found respiratory dysfunction was more frequently

Brain & Development 27 (2005) S59–S68

www.elsevier.com/locate/braindev

H. Leonard et al. / Brain & Development 27 (2005) S59–S68S60

associated with truncating and scoliosis with missense

mutations. Other studies have found a more severe clinical

phenotype in truncating when compared with missense

mutations [14–17]. Where investigations have taken into

account mutation location as well as mutation type a more

refined picture appears. Cheadle et al. [14] showed that early

truncating mutations were likely to be more severe than late

truncating mutations. Similarly, Hoffbuhr et al. [18] found

that missense mutations in the MBD and mutations

truncating the entire TRD had a more severe clinical

phenotype than missense and nonsense mutations in the

TRD and C-terminal. Huppke et al. [16] found that

mutations in the TRD NLS were associated with a more

severe phenotype than those in the MBD and in the rest of

the TRD. Characterisation of phenotypic severity using a

large population by Leonard et al. [19] and Colvin et al. [20]

has shown differential phenotypic severity between the

common mutations including confirmation that, in com-

parison with other MECP2 mutations, the R133C mutation

has a milder clinical presentation. This was the first clear

and clinically meaningful description of the association

between genotype and clinical phenotype. The present study

aimed to investigate the relationship between early

development and individual mutations using both Austra-

lian population based data and international data.

Our previous work showed that, compared with singleton

births of a similar gestation, a greater proportion of Rett

syndrome cases had birthweights !3500 g and lower birth

head circumference; 41% of families reported some

perinatal problem and nearly half (46.5%) of families had

reported their child’s behaviour to be unusual in the first 6

months of life [3]. This led us to speculate that the girl with

Rett syndrome may not be normal at birth. The purpose of

this current study has been to examine whether any of these

early indicators of reduced optimality at birth and in the

early months vary by genotype.

2. Methods

2.1. Case ascertainment

Cases were sourced from two databases, the Australian

Rett Syndrome Database (ARSD) and InterRett, the IRSA

Rett Phenotype Database. Established in 1993, the ARSD is

a population-based database with ongoing ascertainment of

Australian Rett syndrome cases born since 1976 [21]. It

contains demographic, clinical, behavioural, developmen-

tal, functional and genetic information [22]. InterRett is an

international phenotype database, established in 2003 with

funding from the International Rett Syndrome Association

and managed from the Telethon Institute for Child Health

Research in Western Australia [23]. Data contributions to

InterRett cases can be made individually or as multiple

submissions from international collaborating centres.

Both databases collect data from families and clinicians.

The InterRett family questionnaire administered on case

enrolment is similar to and has been developed from the

original Australian version providing comparability

between the two datasets.

As at 31st October 2003, 251 cases had been reported to

ARSD and verified according to the clinical criteria for

classical or atypical Rett syndrome [3]. Of the 251 verified

cases, data from families have been provided in 235 and in

16, information was currently only available from the

clinician. Data provision to InterRett commenced in January

2003. Cases were included in this analysis either if they had

a valid pathogenic mutation or if the diagnosis of Rett

syndrome was considered by the child’s doctor to be

‘definite’. As at June 2003, 88 InterRett cases meeting this

study definition had been reported. Three of these cases

were duplicates of ARSD cases and so were excluded from

this analysis. Thus a total of 235 ARSD (including 130 cases

from the original cohort [3]) and 85 InterRett cases with

family information were available for this study. Ethical

approval for the study was granted by the local institutional

ethics committee.

2.2. Clinical data collection

This report focuses on information relating to early

development provided by the family. As previously

described, data were collected by paper-based question-

naires for ARSD cases [21] and by paper-based or Internet

questionnaires for InterRett cases [23]. Questions were

asked about problems in the first week of life, abnormal

behaviours and development in the first 6 months and from

after 6–10 months. Open responses were coded according to

type of problem as identified by the family e.g. placid,

sleeping problems, gastro-intestinal, developmental

problems.

For each question, cases were generally allocated a score

of 0 if there were no problems identified and a score of 1 for

each problem type identified. However for the questions on

admission to special nursery and for help with breathing at

birth, responses were simply scored as 0 or 1. A composite

perinatal score was derived from the three questions

involving the perinatal period. The maximum number of

problems coded for any of the three problem questions was

3. Thus each case could receive a score of between 0 and 5

for the perinatal period and a score of between 0 and 3 for

the other two time periods and thus a total problem score

ranging from 0 to 11. Additionally, cases were scored

according to their reported mobility at 10 months with the

lowest score of 0 allocated to those performing best (i.e.

walking independently) and the highest score of 12 to those

who could not move around. A score of 10 was allocated for

rolling, 8 for commando crawling, 6 for bottom shuffling, 4

for crawling, and 2 for moving around by furniture walking

or knee walking. A total composite (problem and mobility)

score was computed with a possible range from 0 to 23.

Cases with missing data were not allocated scores.

Table 1

Distribution of cases by year of birth, country of birth and age at diagnosis

and data source

Category Data source

ARSD

(NZ235) n(%)

InterRett

(NZ85) n(%)

All cases*

(NZ320) n(%)

Year of birth

1962–1970 – 6(7.1) 6(1.9)

1971–1975 – 2(2.4) 2(0.6)

1976–1980 33(14.0) 9(10.6) 42(13.1)

1981–1985 49(20.9) 10(11.8) 59(18.4)

1986–1990 59(25.1) 13(15.3) 72(22.5)

1991–1995 58(24.7) 17(20.0) 75(23.4)

1996–2000 35(14.9) 28(32.9) 63(19.7)

2001-present 1(0.4) – 1(0.3)

Country of birth

Argentina – 1(1.2) 1(0.3)

Australia 222(94.5) 2(2.4) 224(70.0)

Canada – 2(2.4) 2(0.6)

China – 15(17.6) 15(4.7)

Hong Kong 1(0.4) – 1(0.3)

India – 1(1.2) 1(0.3)

Jordan 1(0.4) – 1(0.3)

Mexico – 1(1.2) 1(0.3)

New Zealand 3(1.3) 1(1.2) 4(1.3)

Philippines 1(0.4) – 1(0.3)

Romania 1(0.4) – 1(0.3)

United Kingdom 3(1.3) 1(1.2) 4(1.3)

United States 1(0.4) 61(71.8) 62(19.4)

Age at diagnosis (yrs)

0!6 132(76.3) 57(68.7) 189(73.8)

6!11 28(16.2) 14(16.9) 42(16.4)

11!16 11(6.4) 6(7.2) 17(6.6)

16!21 2(1.2) 4(4.8) 6(2.3)

21 and above – 2(2.40) 2(0.8)

*Denominator used taking into account missing values for individual

questions.

H. Leonard et al. / Brain & Development 27 (2005) S59–S68 S61

2.3. Mutational analyses

Mutation testing has been carried out on 197/235

(83.8%) ARSD and on 66/85 (77.6%) InterRett cases with

pathogenic mutations identified in 136/197 (69.0%) and

45/66 (68.2%), respectively. This analysis concentrates

particularly on those cases where a pathogenic MECP2

mutation was identified. Pathogenicity was verified as

previously [19,20] by laboratory report for ARSD cases.

For InterRett cases, multiple sources (clinician and family

questionnaires) were used to confirm mutation status and

pathogenicity was validated using the mutation database,

RettBASE [24]. For the purposes of this study any

ambiguous mutation data were not treated as pathogenic.

Cases with any pathogenic mutation were identified and

then cases with any of the seven most common pathogenic

mutations (R133C, T158M, R168X, R255X, R270X,

R294X and R306C) were grouped together and according

to their specific mutation. Other groupings were cases with

other pathogenic mutations and cases who had either not

been tested or in whom a mutation had not been found. The

ARSD and InterRett data were stored in Filemaker Pro 6

databases and transferred into SPSS Version 10.0 for

analysis.

2.4. Data analysis

For each group the proportions reporting any or a specific

problem for the three time periods, perinatal, 0–6, O6–10

months were presented. In all cases the denominators were

adjusted to account for missing values on individual

questions. The chi-square test was used to determine the

presence of associations between categorical variables [25].

Means were calculated for individual and total problem

scores, the mobility score and the composite score

[including problems and mobility]. One-way analysis of

variance [Kruskal Wallis for non-normally distributed

variables] was used to compare differences in means

according to mutation categories. P values of less than 0.1

were considered indicative of important differences and no

adjustments were made for multiple testing [26].

3. Results

The distribution of cases by year, country of birth and age

at diagnosis by data source is shown in Table 1. Of those

262 (81.9%) had results from molecular testing. In 181 of

those tested, a valid pathogenic mutation as defined by study

criteria was present. For each of the mutations R133C,

T158M, R168X, R255X, R270X, R294X and R306C

there were eleven or more cases. This group accounted for

122/181 (67.4%) of those cases with pathogenic mutations.

Over a third of respondents reported that their child had

experienced a problem in the first week of life, with almost

the same proportion (39.3%) in the group with pathogenic

mutations. Within the seven common mutations the

proportions ranged from 3/11 (27.3%) in those with

R133C to 9/16 (56.3%) in those with R255X (PZ0.14,

Table 2). Feeding problems were the most common

individual problem reported in 24 (13.5%) of those with

valid mutations and in 18 (15%) of those with the seven

common mutations. Parents were also asked whether the

infant needed help breathing. Again there was variation with

no such reports in the 11 with R133C compared with 4/15

(26.7%) in those with R255X (PZ0.06). Only 1/16 (6.3%)

of R294X cases had been admitted to a special nursery

compared with 5/16 (31.3%) of those with R255X (PZ0.07). Overall 81/180 (45.0%) of those with pathogenic

mutations were reported to have some perinatal difficulty:

either help to start breathing, admission to a special nursery

or a problem in the first week of life.

Overall 184/317 (58%) of respondents reported that they

had noted unusual behaviour or development during the first

6 months of their daughter’s life (Table 3). Nineteen reports

included in the analysis were ambiguous in that respondents

provided a response to the open but not the closed question.

Table 2

Presence and nature of perinatal problems by mutation type

Observation/problem R133C

(NZ11)

n (%)

R294X

(NZ17)

n (%)

R168X

(NZ22)

n (%)

T158M

(NZ28)

n (%)

R255X

(NZ16)

n (%)

R270X

(NZ16)

n (%)

R306C

(NZ12)

n (%)

Seven common

mutations (NZ122)

n (%)

Other patho-

genic mutations

(NZ59) n (%)

All pathogenic

mutations

(NZ181) n (%)

No Mutation or

Not tested

(NZ139) n (%)

All cases*

(NZ320)

n (%)

Problems in the 1st week

of life

3(27.3) 6(35.3) 9(45.0) 13(46.4) 9(56.3) 6(37.5) 6(50.0) 52(43.3) 18(31.0) 70(39.3) 63(46.3) 123(38.8)

Feeding – 2(11.1) 2(10.0) 7(25.0) 3(18.8) 1(6.3) 3(25.0) 18(15.0) 6(10.3) 24(13.5) 27(19.8) 51(16.1)

Jaundice – 1(5.6) 4(20.0) 2(7.1) – 1(6.3) 2(16.7) 10(8.3) 7(12.1) 17(9.6) 5(3.7) 22(7.0)

Gastrointestinal 1(9.1) 1(5.6) – 2(7.1) 1(6.3) 2(12.6) 1(8.3) 8(6.7) 2(3.4) 10(5.6) 6(4.4) 16(5.1)

Breathing problems – – 1(5.0) 3(10.7) 2(12.5) 1(6.3) 2(16.7) 9(7.5) 1(1.7) 10(5.6) 4(2.9) 14(4.4)

Sleep – – – 2(7.1) 1(6.3) 3(18.8) 6(5.0) 2(3.4) 8(4.5) 4(2.9) 12(3.8)

Hypoglycaemia – – 2(10.0) – – – 2(1.7) 2(3.4) 4(2.2) – 4(1.3)

Hypothermia – – – 1(3.6) – – 1(1.0) 1(1.7) 2(1.1) – 2(0.6)

Hypotonia – – – 1(3.6) 1(6.3) – 2(1.7) – 2(1.1) – 2(0.6)

Birth defect – – – – 1(6.3) – 1(1.0) – – 1(0.3)

Other 2(18.2) 2(11.1) – 2(7.1) 2(12.5) – 2(16.7) 10(8.3) 2(3.4) 12(6.7) 11(8.1) 23(7.3)

Special nursery 2 (8.2) 1(6.3) 5(22.7) 4(15.4) 5(31.3) – 2(16.7) 19(16.0) 8(13.8) 27(15.3) 24(17.8) 51(16.8)

Help breathing – 1(6.7) – 6(23.1) 4(26.7) 1(6.7) 2(16.7) 14(8.7) 5(8.6) 19(11.0) 22(16.4) 41(13.4)

Any perinatal difficulty 4(36.4) 6(35.3) 10(45.5) 13(46.4) 10(62.5) 7(43.8) 7(58.3) 57(46.7) 24(41.4) 81(45.0) 64(46.4) 145(45.6)

*Denominator used taking into account missing values for individual questions.

Table 3

Presence and nature of abnormal development or behaviour in first 6 months of life by mutation type

Observation/problem R133C

(NZ11)

n (%)

R294X

(NZ17)

n (%)

R168X

(NZ22)

n (%)

T158M

(NZ28)

n (%)

R255X

(NZ16)

n (%)

R270X

(NZ16)

n (%)

R306C

(NZ12)

n (%)

Seven com-

mon

mutations

(NZ122)

n (%)

Other

pathogenic

mutations

(NZ59)

n (%)

All patho-

genic

mutations

(NZ181)

n (%)

No

mutation or

not tested

(NZ139)

n (%)

All cases*

(NZ320)

n (%)

Unusual behaviour in

the 1st 6 months

4(36.4) 9(52.9) 15(68.2) 19(67.9) 11(68.8) 9(56.3) 6(50.0) 73(59.8) 33(56.9) 106(58.9) 78(56.9) 184(58.0)

Placid – 3(17.6) 3(13.6) 4(14.3) 4(25.0) 2(12.5) 4(33.3) 20(16.4) 12(20.7) 32(17.8) 26(19.0) 58(18.3)

Feeding – – 3(13.6) 5(17.9) 1(6.3) 4(25.0) 1(8.3) 14(11.5) 8(13.8) 22(12.2) 17(12.4) 39(12.3)

Developmental – 1(5.9) 1(4.5) 5(17.9) 3(18.8) 3(18.8) – 12(9.8) 5(8.6) 17(9.4) 18(13.1) 35(11.0)

Floppy – – 5(22.7) 5(17.9) 1(6.3) 2(12.5) – 13(10.7) 5(8.6) 18(10.0) 14(10.2) 32(10.1)

Gastrointestinal 1(9.1) 3(17.6) 1(4.5) 2(7.1) 2(12.5) 2(12.5) – 11(9.0) 1(1.7) 12(6.7) 9(6.6) 21(6.6)

Eyes – – 1(4.5) 1(3.6) 1(6.3) – 1(8.3) 4(3.3) – 4(2.2) 11(8.0) 15(4.7)

Sleep 1(9.1) 1(5.9) – – – – – 2(1.6) 4(6.9) 6(3.3) 3(2.2) 9(2.8)

Strange behaviour – 1(5.9) – 2(7.1) – – – 3(2.5) 1(1.7) 4(2.2) 4(2.9) 8(2.5)

Hand behaviour – – – – – – – – – – 5(3.6) 5(1.6)

Other 3(27.3) 2(11.8) 1(4.5) 1(3.6) 1(6.3) – 1(8.3) 9(7.4) 2(3.4) 11(6.1) 1(0.7) 12(3.8)

*Denominator used taking into account missing values for individual questions.

H.

Leo

na

rdet

al.

/B

rain

&D

evelopm

ent

27

(20

05

)S

59

–S

68

S6

2

Box 1. Examples of family expression of unusual behaviour from the first 6 months of life

Code: Placid

‘Unusually passive—happy always! Never crying not even when hungry—or when had chicken pox at 6 months’

‘Was very placid and slept a lot compared to friends babies’

‘She was always ‘too good’ a baby. Very placid, fed well (breast fed)’

‘Over placid. Did not cry much’

‘She was ‘too good’. Extremely placid with no acknowledgement of the world around her’

‘Extremely placid, very good sleeper’

‘Very placid, easily satisfied’

‘Low tone, slow to smile, very placid, sleepy’

‘She was a placid baby with a voracious appetite and a good feeder’

Code: Floppy

‘Poor feeder. Very floppy. Cried a lot’

‘She was floppy compared to other children’

‘Until 6 months was unable to support her own head—hypotonic—floppy’

‘She always seemed to have a very floppy tone compared with my first child and did not seem to be making her

milestones’

‘Her doctor as the time classed her as a ‘floppy baby’’

‘**** was a very ‘floppy’ baby—however in the first 6 months we just saw this as her being a very cuddly baby’

‘Kind of ‘floppy’ tone’

‘Lots of reflux, otherwise a very content ‘floppy’ baby’

‘She was a little floppy/soft’

H. Leonard et al. / Brain & Development 27 (2005) S59–S68 S63

A very similar proportion of unusual behaviour was

reported in those cases with pathogenic mutations. For the

seven specific mutations examined, the proportion varied

from 4/11 (36.4%) in cases with R133C to 11/16 (68.8%) in

cases with R255X mutations (PZ0.10, Table 3). When the

open responses were coded into specific categories the most

commonly reported category was placid reported in 18.3%

of all cases and in 16.4% of those with the common

mutations. Some examples of how the families expressed

this are shown in Box 1. Other high rankings were reports of

a floppy baby in 10.1% of all cases and 10.7% of those with

common mutations and of developmental problems in

11.0% of all cases and 9.8% of those with common

mutations.

Participants were then asked whether they had noticed

anything unusual about their child’s development or

behaviour in the period from after six to ten months of

age. Overall 221 of the 313 (70.6%) reported in the

affirmative (7 cases with ambiguous responses in that they

provided a response to the open but not the closed question

were also included) and 122/180 (67.8%) of those cases

with a valid mutation (Table 4). Thus the proportion

reporting in the affirmative among those who did not have a

mutation was marginally higher than among those who did

(74.4 vs 67.8%, PZ0.20). For the seven specific mutations

the proportions varied from 9/17 (52.9%) in those with

R294X to 13/16 (81.3%) in those with R270X. The

comparison between these two approached significance at

PZ0.08. Respondents’ most predominant recollection

within this time period, was concern specifically about

their child’s development (Box 2). This was mentioned in

nearly half (45.7%) of all cases and in 42.6% of those with

common mutations. Other categories were much less

frequently reported and generally in less than 10% of all

cases (Table 4).

For 311 cases a description of the child’s mobility at 10

months was available. In 62/311 (19.9%) of all cases and in

15.0% of those with the common mutations the child was

reported not to be moving around at 10 months. In 88/311

(28.3%) of all cases and a quarter (25.8%) of those with the

common mutations the child was reported to be moving

around by rolling. The child was commando crawling in

29/311 (9.3%) of all cases and in 7.5% of those with the

common mutations and bottom shuffling in 40/311 (12.9%)

and 15.8%, respectively. Sixty one (19.6%) of all cases and

nearly a quarter (24.2%) of those with the common

mutations were crawling at this time. Whilst 2.3 and

3.3%, were said to be able to walk independently, a further

8.0% of all cases and 8.3% of those with the common

mutations were using other forms of mobility such as

furniture or knee walking.

The mobility and other scores for each of the seven

common mutations are shown in Fig. 1. For mobility the

highest score indicating the least mobility at 10 months was

for R270X (meanZ8.25) and the lowest mean scores

indicating the best mobility were for R306C and R294X

(5.17, 5.76). After R294X and R306C the next most mobile

were the R133C and T158M cases, followed by R168X, and

R255X. However the only statistically significant differ-

ences were between R306C and both R255X and R270X

Tab

le4

Pre

sen

cean

dn

ature

of

abn

orm

ald

evel

op

men

to

rb

ehav

iou

rfr

om

6to

10

mon

ths

of

life

by

mu

tati

on

typ

e

Ob

serv

atio

n/p

rob

lem

R1

33

C

(NZ

11

)n

(%)

R2

94

X

(NZ

17

)n

(%)

R1

68

X

(NZ

22

)n

(%)

T1

58

M

(NZ

28

)n

(%)

R2

55

X

(NZ

16

)n

(%)

R2

70

X

(NZ

16

)n

(%)

R3

06

C

(NZ

12

)n

(%)

Sev

enco

mm

on

mu

tati

on

s(N

Z1

22

)n

(%)

Oth

er

pat

ho

gen

ic

mu

tati

on

s

(NZ

59

)

All

pat

ho

-

gen

ic

mu

tati

on

s

(NZ

18

1)

No

mu

tati

on

or

no

tte

sted

(NZ

13

9)

All

case

s*

(NZ

32

0)

Un

usu

al

beh

av

iou

rO

6–

10

mo

nth

s

6(5

4.5

)9

(52

.9)

15

(68.2

)1

7(6

0.7

)1

2(7

5.0

)1

3(8

1.3

)9

(75

.0)

81

(66.4

)4

1(7

0.7

)1

22

(67

.8)

99

(74.4

)2

21

(70

.6)

Dev

elop

men

t3

(27

.3)

5(2

9.4

)1

0(4

5.5

)1

3(4

6.4

)7

(43

.8)

8(5

0.0

)6

(50

.0)

52

(42.6

)2

8(4

8.3

)8

0(4

4.4

)6

3(4

7.4

)1

43

(45

.7)

Flo

pp

y–

–2

(9.1

)3

(10

.7)

–2

(12

.5)

1(8

.3)

8(6

.6)

10

(17.2

)1

8(1

0.0

)1

2(9

.0)

30

(9.6

)

Pla

cid

–1

(5.9

)1

(4.5

)1

(3.6

)2

(12

.5)

1(6

.3)

–6

(4.9

)6

(10

.3)

12

(6.7

)1

3(9

.8)

25

(8.0

)

Ey

es1

(9.1

)1

(5.9

)3

(13

.6)

2(7

.1)

2(1

2.5

)2

(12

.5)

2(1

6.7

)1

3(1

0.7

)1

(1.7

)1

4(7

.8)

9(6

.8)

23

(7.3

)

Oth

erin

cl.lo

sso

fsk

ills

1(9

.1)

1(5

.9)

3(1

3.6

)3

(10

.7)

2(1

2.5

)2

(12

.5)

2(1

6.7

)1

4(1

1.5

)4

(6.9

)1

8(1

0.0

)5

(3.8

)1

9(6

.1)

Str

ang

eb

ehav

iou

r2

(18

.2)

2(1

1.8

)–

–2

(12

.5)

1(6

.3)

–7

(5.7

)2

(3.4

)9

(5.0

)8

(6.0

)1

7(5

.4)

Han

db

ehav

iou

r1

(9.1

)1

(5.9

)–

––

–1

(8.3

)3

(2.5

)1

(1.7

)4

(2.2

)9

(6.8

)1

3(4

.2)

Sle

ep–

1(5

.9)

––

––

1(8

.3)

2(1

.6)

–2

(1.1

)5

(3.8

)1

1(3

.5)

Gas

tro

inte

stin

al–

3(1

6.7

)–

––

––

3(2

.5)

–3

(1.7

)4

(3.0

)7

(2.2

)

Fee

din

g–

––

1(3

.6)

––

–1

(0.1

)1

(1.7

)2

(1.1

)4

(3.0

)6

(1.9

)

*D

eno

min

ato

ru

sed

tak

ing

into

acco

un

tm

issi

ng

val

ues

for

indiv

idu

alq

ues

tio

ns.

H. Leonard et al. / Brain & Development 27 (2005) S59–S68S64

(PZ0.04, 0.02) and between R294X and R270X (PZ0.04).

For the combined perinatal score, which included data from

three questions, R255X had the highest mean score of 1.27

followed by R306C (1.17) and T158M (1.08) and the lowest

scores were for R294X (0.43) and R133C (0.45). The

differences between R255X and R133C and between

R255X and R294X were significant (PZ0.05, 0.03).

Additionally the mean of the total problem scores for each

mutation are also shown (Fig. 1). R255X, T158M and

R306C had the highest mean scores (3.00, 2.84 and 2.75)

and R294X and R133C had the lowest (1.50, 1.64). The

differences between the lowest scoring mutation R294X and

the highest two R255X and T158M were just significant

(PZ0.05). However when the mobility and problem scores

were combined the most severely affected was R255X

(meanZ10.73) closely followed by R270X (meanZ10.60)

with R294X and R306C the least severe (meanZ7.50,7.91).

The differences between R294X and R255X (PZ0.06) and

R270X (PZ0.07) approached significance.

4. Discussion

This study has found that, in general, cases with R255X

and R306C mutations were slightly more likely to have

problems at birth with appreciable differences between

R255X and both R294X and R133C. For the period after

birth up to 10 months cases with R255X and R270X

mutations were more likely to have problems. Cases with

R270X were likely to be least and cases with R306C most

mobile at 10 months. When all the scores were combined

R255X and R270X were the most severe and R294X the

least severe. R306C had the least consistent pattern overall,

with relatively high combined perinatal problems but

particularly good mobility and thus scoring overall on the

mild side.

This is the largest study to date examining the effects of

individual mutations in numbers of this magnitude. This has

only been possible because we have been able to combine

cases from our Australian population-based registry of

juvenile Rett syndrome ascertained over a 10 year period

[21,22] with cases from the international phenotype

database InterRett ascertained so far over only a 6 month

period [23]. Rett syndrome is fortunately a rare disorder

with an incidence of 1:10,000 females such that in Australia

we only expect 12 cases per birth year [21]. Given that

T158M, the commonest mutation, only represents less than

10% of Australian Rett Syndrome cases we would only

expect to ascertain approximately one case per birth year

with the frequency for the other common mutations being

even less. With the introduction of online-based ques-

tionnaires [23,27] we are able to reach a wider international

audience with minimal costs. Therefore, this form of data

collection utilised by InterRett provides an efficient

mechanism to maximise case numbers for genotype-

phenotype studies in a rare condition, which has multiple

Box 2. Examples of family expression of unusual behaviour from 6 to 10 months of life

Code: Development

‘Her development was slower than my other two children’

‘Did not sit up and very slow to roll over’

‘No crawling or weight bearing/standing’

‘No attempts to crawl despite being able to sit well from 5.5 months’

‘In between those months rolling around on the floor became less frequent, and moving around in a baby walker began to

stop. Also, no crawling or sitting independently eventuated’

H. Leonard et al. / Brain & Development 27 (2005) S59–S68 S65

different mutations. An additional strength of this and

further studies that will follow is that it provides the

opportunity for collaboration with other centres such as has

already occurred with the Chinese group (XW, XB, HP).

The advantage of studies using the ARSD over other

studies has always been the fact that it is population-based

and has shown to be effective in representing juvenile Rett

syndrome in Australia (population approximately

20 million) [21]. InterRett has the capacity to provide

much larger numbers and although we hope that in the

future it will contain representative data for many countries

it will never fully represent its target population—the world.

Similarly, although for our Australian study [3] we were

able to use state and national population norms for

comparing birthweight and head circumference, this

becomes much more difficult in an international multi-

ethnic sample and for this reason we did not include such an

analysis here.

We acknowledge that our research relies on retrospective

parental recall. Nevertheless with a rare condition such as

this a prospective epidemiological study involving the

period prior to regression is not logistically feasible and

hence for information on early development one is

dependent on a study design, which uses retrospective

0 2 4

Problem Perinatal

Problem 0-6mth

Problem >6-10mth

Total Problem

Mobility

Total Composite

Me

Fig. 1. Mean score for problem and mobility and problem/m

report. In studies of this nature there is always concern about

whether the information is recalled accurately but more

importantly about whether differential recall between

groups is biasing the results [28]. Our experience of

collecting data from both clinicians and families is that

the information provided by families is often more detailed

and comprehensive than that provided by clinicians. Also

when information is requested of them we note that

clinicians will often return to the families to supplement

or validate their own data. With respect to differential recall

we do not have any evidence to suggest that this is likely to

be systematically different between families of children

with Rett syndrome whose subsequent outcome is more or

less severe because of their mutation type. Whilst the

primary purpose of the present study was to compare early

development in different categories of children with Rett

syndrome, Majnemer and Rosenblatt [29] found little

evidence of bias in a case control study of parental recall

of developmental milestones in high risk and healthy

newborns. In our study we encouraged parents to use the

child’s baby book to help answer some of the questions. As

this record was made at the time there should be little

chance that later outcome would affect their recording of

this event and thus their reporting.

6 8 10 12an Score

R306CR270XR255XT158MR168XR294XR133C

obility composite scores for seven common mutations.

H. Leonard et al. / Brain & Development 27 (2005) S59–S68S66

The data collection methods have not been entirely

uniform for the whole dataset in that all the Australian data

were provided by paper-based questionnaires while for the

InterRett cohort, the emphasis was on online data collection

using the Internet. In future research we plan to investigate

differences in data produced by the two approaches but this

is beyond the scope of this paper. A proportion (40%) of the

records used were previously presented [3] but the early

analysis did not examine the effect of genotype. In other

reports [22] we have used four scales – Kerr, Percy, Pineda

and WeeFIM to help characterise phenotype. One of the

twenty items contributing to the Kerr scale did relate to

early development, whilst the Percy scale did seek

information about crawling and the Pineda scale included

an item on age at sitting. Age at sitting was not used in this

analysis. The nature of our data makes our information on

crawling incomplete and so in this previous analysis we had

to use surrogate measures [22]. However, in both cases (for

the Kerr and Percy scores) these developmental items only

contributed w5–6% of the total score.

In our previous reports [19,20] we have included X

inactivation data and have found no protective effect on the

phenotype by the presence of skewed X inactivation when

specific mutations have been examined. However more

recently, we demonstrated that skewed X inactivation was

associated with an overall protective effect (including in

cases where no mutation was found) in particular reference

to health service utilisation and hospital admissions [30]. In

this present study X inactivation data were only available

for the Australian cohort and therefore were not included.

Apparently normal pre- and perinatal period and

apparently normal psychomotor development through the

first six months of life are two of the diagnostic criteria for

Rett syndrome set out by Trevathan et al. in 1988 [2].

Despite this the issue of early normality in Rett syndrome

has been questioned for many years in case reports [5,7,8,

11,31]. These clinical observations were corroborated by

our epidemiological studies where we compared birth

weight and head circumference in Rett syndrome with

population norms and reported on parental perceptions of

the perinatal period and early months of life [3]. Much more

recently and independently Huppke et al. [4] also reported

that birth weight, length and head circumference in a group

of 120 mutation positive Rett syndrome cases were

significantly lower than in the general population. In 2002

it was suggested that the clinical criteria be revised such that

early normality was no longer mandatory [32].

The results of the present analysis with respect to

parental recall of the perinatal period and first 6 months of

life confirm our previous findings [3]. The proportion

reporting any kind of perinatal abnormality (46.4%) in this

study was very similar to the 41% reported in the previous

study. However a higher proportion, nearly 60% of

respondents indicated that they had concerns about their

daughter’s behaviour or development in the first 6 months

compared with just under half in the original study.

Considering that the 108 Australian cases ascertained

since 1995 had a younger average age at ascertainment

than the original cohort, this could be due to better recall

when parents did not have access to baby books.. This age

differentiation was a phenomenon noted by Charman et al.

[6] who also found that normal pre-regression development

was the exception and occurred in only 30% cases. Our

findings are also corroborated by the recent work of Burford

et al. [33] and Einspieler et al. [34]. In Burford’s study

health visitors were asked to review home videos taken

during the first year of life in girls who were later diagnosed

with Rett syndrome and in a comparison group without

developmental problems at 5 years. The nurses were

significantly more likely to be alerted to signs of

developmental deviation in the Rett syndrome cases again

suggesting that the neurodevelopmental aberration in Rett

syndrome is present at a much earlier age than has typically

been recognised. Einspeler et al. on the other hand

conducted a detailed video analysis of movements, posture

and behaviour of 22 cases of Rett syndrome during the first

6 months of life. They found that there was an abnormal

quality of general movements in all cases and in over half

the quality of tongue protrusion, postural stiffness, eye and

finger movements were abnormal. Thus our research adds to

the accumulating and complementary body of literature

which is scrutinising early development in Rett syndrome

through a range of methodologies.

One of the four case reports cited by Huppke et al. [4] to

illustrate the phenotypic spectrum in Rett syndrome (a child

with a C terminal deletion who achieved few milestones)

exemplifies well how abnormal early development may be.

However, none of these or any studies, of which we are

aware, has attempted to examine early development in

the context of specific mutations. In fact other than our own

[19,20], there has been little analysis of the effects of

specific mutations, although Huppke et al. also demon-

strated the severity of truncating mutations involving the

TRD-NLS [16]. We acknowledge that even with a sample

size of 320 cases the statistical power to discriminate

between individual mutations, when even the commonest of

these will not be responsible for more than one sixth of

mutation positive cases (and !9% of all cases in this study),

needs to be improved. Yet the present analysis does show

that two of the three mutations with the lowest problem and

mobility scores, namely R133C and R294X, were those

we had previously identified as having a mild phenotype

[19,20]. Conversely the mutations with the highest scores

were the R255X and R270X. Thus the pattern we [20] and

Huppke et al. [16] had seen in regard to general functioning

of individuals with mutations truncating the TRD-NLS was

also seen in terms of early development. Also particularly

striking was the better mobility with the R294X mutation

which was the aspect of the associated phenotype which

most contributed to its low composite score. As mentioned

the pattern with R306C, a missense mutation located in the

TRD after the NLS was more variable with apparently more

H. Leonard et al. / Brain & Development 27 (2005) S59–S68 S67

problems at birth but good mobility at 10 months.

Previously we would have graded R306C as being neither

particularly severe or mild although we did find that the

onset of regression tended to be a little later on average [20].

The results of this study have important implications

from a variety of perspectives. Our findings support the

increasing body of literature showing that in general the

onset of this disorder is much earlier than originally

appreciated. It now becomes evident that those children

previously coined congenital Rett syndrome [35], and those

subsequently described by Huppke et al.[4] and by ourselves

[36] and those diagnosed as newborn encephalopathy

[37–39] are a clearly defined part of the phenotypic

spectrum. We have now been able to show that the

manifestation of this early developmental aberration does

seem to be associated with the underlying genotype. These

findings also have major implications for the diagnosis and

recognition of this disorder. It is clear we have to reverse the

entrenched clinical dogma that a diagnosis of Rett syndrome

should not be considered unless the child’s early develop-

ment is normal. Previously and especially before the

identification of the association with the MECP2 gene,

such children may have remained underdiagnosed. More-

over the information is particularly relevant to under-

standing the mechanism of early development in Rett

syndrome and determining if and at which time(s) early

intervention might be feasible.

This is the first report to have made use of this new

source of international Rett syndrome data. However,

InterRett is still in its infancy with data collection only

commencing in 2003. Despite having comparatively large

numbers of subjects, statistical power will be further

improved as InterRett expands and population data from

countries such as Israel, Spain, Italy, Netherlands, Belgium,

Turkey and France are included. Current negotiations with

these countries for data contributions will yield in excess of

1000 cases.

Acknowledgements

The authors would like to acknowledge the International

Rett Syndrome Association for their funding of InterRett

and also for their ongoing support and encouragement. The

National Institute of Child Health and Human Development

is also acknowledged for its current funding of the

Australian Rett Syndrome Database under NIH grant

number 1 R01 HD43100-01A1 PI and the Financial Markets

Foundation for Children and the Rett Syndrome Australian

Research Fund for previous funding. HL is part funded by

NHMRC program grant 003209. The Sydney research

group was also in part funded by NHMRC project grant

185202. We also acknowledge Neil Leonard and the

information technology team at the Telethon Institute for

Child Health Research for their information technology

expertise and assistance. Finally we would also like to

express our gratitude to all the families who have

participated in the study; the Australian Paediatric

Surveillance Unit and the Rett Syndrome Association of

Australia who facilitated case ascertainment in Australia;

and those members of the InterRett International Reference

panel who helped with the piloting of InterRett.

References

[1] Hagberg B, Goutieres F, Hanefield F, Rett A, Wilson J. Rett

syndrome: criteria for inclusion and exclusion. Brain Dev 1985;7:

372–3.

[2] Rett Syndrome Diagnostic Criteria Working Group. Diagnostic

Criteria for Rett Syndrome. Ann Neurol 1988;23:425–8.

[3] Leonard H, Bower C. Is the girl with Rett syndrome normal at birth?

Dev Med Child Neurol 1998;40:115–21.

[4] Huppke P, Held M, Laccone F, Hanefeld F. The spectrum of

phenotypes in females with Rett syndrome. Brain Dev 2003;25:

346–51.

[5] Naidu S, Hyman S, Harris EL, Narayanan V, Johns D, Castora F. Rett

syndrome studies of natural history and search for a genetic marker.

Neuropediatrics 1995;26:63–6.

[6] Charman T, Cass H, Owen L, Wigram T, Slonims V, Weeks L, et al.

Regression in individuals with Rett syndrome. Brain Dev 2002;24:

281–3.

[7] Kerr AM, Montague J, Stephenson JB. The hands, and the mind, pre-

and post-regression, in Rett syndrome. Brain Dev 1987;9:487–90.

[8] Nomura Y, Segawa M. Clinical features of the early stage of the Rett

syndrome. Brain Dev 1990;12:16–19.

[9] Heilstedt HA, Shahbazian MD, Lee B. Infantile hypotonia as a

presentation of Rett syndrome. Am J Med Genet 2002;111:238–42.

[10] Witt Engerstrom I. Age-related occurrence of signs and symptoms in

the Rett syndrome. Brain Dev 1992;14(Suppl).

[11] Kerr AM. Early clinical signs in the Rett disorder. Neuropediatrics

1995;26:67–71.

[12] Amir RE, Van den Veyver IB, Wan M, Tran CQ, Francke U,

Zoghbi HY. Rett syndrome is caused by mutations in X-linked

MECP2, encoding methyl- CpG-binding protein 2. Nat Genet 1999;

23:185–8.

[13] Amir RE, Van den Veyver IB, Schultz R, Malicki DM, Tran CQ,

Dahle EJ, et al. Influence of mutation type and X chromosome

inactivation on Rett syndrome phenotypes. Ann Neurol 2000;47:

670–9.

[14] Cheadle JP, Gill H, Fleming N, Maynard J, Kerr A, Leonard H, et al.

Long-read sequence analysis of the MECP2 gene in Rett syndrome

patients: correlation of disease severity with mutation type and

location. Hum Mol Genet 2000;9:1119–29.

[15] Monros E, Armstrong J, Aibar E, Poo P, Canos I, Pineda M. Rett

syndrome in Spain: mutation analysis and clinical correlations. Brain

Dev 2001;23(suppl 1):S251–S3.

[16] Huppke P, Held M, Handefeld F, Engel W, Laccone F. Influence of

mutation type and location on phenotype in 123 patients with Rett

syndrome. Neuropediatrics 2002;33:63–8.

[17] Weaving LS, Williamson SL, Bennetts B, Davis M, Ellaway CJ,

Leonard H, et al. Effects of MECP2 mutation type, location and

X-inactivation in modulating Rett syndrome phenotype. Am J Med

Genet 2003;118A:103–14.

[18] Hoffbuhr K, Devaney JM, LaFleur B, Sirianni N, Scacheri C, Giron J,

et al. MeCP2 mutations in children with and without the phenotype of

Rett syndrome. Neurology 2001;56:1486–95.

[19] Leonard H, Colvin L, Christodoulou J, Schiavello T, Weaving L,

Williamson S, et al. Patients with the R133C mutation: is their

phenotype different from Rett syndrome patients with other

mutations? J Med Genet 2003;40:e52.

H. Leonard et al. / Brain & Development 27 (2005) S59–S68S68

[20] Colvin L, Leonard H, de Klerk N, Davis M, Weaving L,

Williamson S, et al. Refining the phenotype of common mutations

in Rett syndrome. J Med Genet 2004;41:25–30.

[21] Leonard H, Bower C, English D. The prevalence and incidence of Rett

syndrome in Australia. Eur Child Adolesc Psychiatry 1997;6(suppl 1):

8–10.

[22] Colvin L, Fyfe S, Leonard S, Schiavello T, Ellaway C, De Klerk N,

et al. Describing the phenotype in Rett syndrome using a population

database. Arch Dis Child 2003;88:38–43.

[23] Fyfe S, de Klerk N, Cream A, Christodoulou J, Leonard H. MECP2

and InterRett: international IRSA databases for Rett syndrome.

J Child Neurol 2003;18:709–13.

[24] Christodoulou J, Grimm A, Maher T, Bennetts B. RettBASE: the

IRSA MECP2 variation database-a new mutation database in

evolution. Hum Mutat 2003;21:466–72.

[25] Bland M. An introduction to medical statistics. Oxford: Oxford

Medical Publications; 1987.

[26] Rothman K. In: Rothman K, editor. Modern epidemiology. 1st ed.

Boston/Toronto: Little Brown and Company; 1986. p. 147–50.

[27] Fyfe S, Leonard H, Gelmi R, Tassell A, Strack R. Using the Internet to

pilot a questionnaire on childhood disability in Rett syndrome. Child

Care Health Dev 2001;27:535–43.

[28] Chouinard E, Walter S. Recall bias in case control studies: an

empirical analysis and theoretical framework. J Clin Epidemiol 1995;

48:245–54.

[29] Majnemer A, Rosenblatt B. Reliability of parental recall of

developmental milestones. Pediatr Neurol 1994;10:304–8.

[30] Moore H, Leonard H, de Klerk N, Robertson I, Fyfe S,

Christodoulou J, Weaving L, Davis M, Mulroy S, Colvin L. Health

Service use in Rett syndrome. J Child Neurol 2004;19:42–50.

[31] Nomura Y, Segawa M, Higurashi M. Rett syndrome—an early

catecholamine and indolamine deficient disorder? Brain Dev 1985;7:

334–41.

[32] Hagberg B, Hanefeld F, Percy A, Skjeldal O. An update on

clinically applicable diagnostic criteria in Rett syndrome.

Comments to Rett syndrome clinical criteria consensus panel

satellite to European paediatric neurology society meeting, Baden

Baden, Germany, 11 September 2001. Eur J Paed Neurol 2002;6:

293–7.

[33] Burford B, Kerr AM, Macleod HA. Nurse recognition of early

deviation in development in home videos of infants with Rett disorder.

J Intellect Disabil Res 2003;47(Pt 8):588–96.

[34] Einspieler C, Kerr AM, Prechtl HF. Is the early development of girls

with Rett disorder really normal? Pedratr Res 2005;57:696–700.

[35] Goutieres F, Aicardi J. Atypical forms of Rett syndrome. Am J Med

Genet 1986;24:184–94.

[36] Ellaway CJ, Badawi N, Raffaele L, Christodoulou J, Leonard H. A

case of multiple congenital anomalies in association with Rett

syndrome confirmed by MeCP2 mutation screening. Clin Dysmorphol

2001;10:185–8.

[37] Wan M, Lee SS, Zhang X, Houwink-Manville I, Song HR, Amir RE,

et al. Rett syndrome and beyond: recurrent spontaneous and familial

MECP2 mutations at CpG hotspots. Am J Hum Genet 1999;65:1520–9.

[38] Zeev BB, Yaron Y, Schanen NC, Wolf H, Brandt N, Ginot N, et al.

Rett syndrome: clinical manifestations in males with MECP2

mutations. J Child Neurol 2002;17:20–4.

[39] Lynch SA, Whatley SD, Ramesh V, Sinha S, Ravine D. Sporadic case

of fatal encephalopathy with neonatal onset associated with a T158M

missense mutation in MECP2. Arch Dis Child Fetal Neonatal Ed

2003;88:F250–F2.