Journal of Autism and Developmental Disorders

, Volume 37, Issue 1, pp 171–183

The Development of Young Siblings of Children with Autism from 4 to 54 Months

Authors

  • Ifat Gamliel
    • Department of Psychology and School of EducationHebrew University of Jerusalem
    • Department of Psychology and School of EducationHebrew University of Jerusalem
  • Marian Sigman
    • Department of PsychiatryUniversity of California
Original paper

DOI: 10.1007/s10803-006-0341-5

Cite this article as:
Gamliel, I., Yirmiya, N. & Sigman, M. J Autism Dev Disord (2007) 37: 171. doi:10.1007/s10803-006-0341-5

Abstract

Cognitive and language skills of 39 siblings of children with autism (SIBS-A) and 39 siblings of typically developing children (SIBS-TD) at ages 4, 14, 24, 36, and 54 months were compared. Twelve of the 39 SIBS-A revealed a delay in cognition and/or language (including one child diagnosed with autism) compared to only two SIBS-TD. Developmental trajectories revealed that the cognitive differences disappeared by age 54 months, but some differences in language ability remained. Thus, most SIBS-A were well-functioning, but some revealed cognitive and/or language difficulties during the preschool years. Even these siblings by and large caught up by the age of 54 months, with receptive and expressive language abilities remaining an area of difficulty for some earlier identified siblings.

Keywords

AutismSiblingsEarly phenotypeCognitive skillsLanguage skills

Introduction

With the growing number of children who receive a diagnosis of autistic disorder, autism spectrum disorders, or related difficulties (Rutter, 2005), and with the better prognosis associated with early detection and intervention, researchers are currently focusing efforts on the identification of early markers for autism and autism related conditions. Various methodologies are being employed to diagnose autism and autism related difficulties at the youngest possible age. Some researchers are concentrating on early screening measures for autism (Baron-Cohen, Allen, & Gillberg, 1992; Robins, Fein, Barton, & Green, 2001; Stone, Coonrod, & Ousley, 2000); other researchers are scrutinizing early signs as manifested in home videos of children who later receive diagnoses within the autistic spectrum (i.e., Baraneck, 1999; Maestro et al., 2005; Werner, Dawson, Osterling, & Dinno, 2000); and other researchers are conducting prospective studies on the development of siblings of children with autism (Landa & Garrett-Mayer, 2006; Yirmiya et al., 2006a; Yirmiya, Gamliel, Shaked, & Sigman, 2006b; Zwaigenbaum et al., 2005).

Several screening studies have been initiated to identify young children at risk for autism. For example, using their Checklist for Autism in Toddlers (CHAT) at age 18 months, Baron-Cohen et al. (1992, 1996) found that only those children who later received a diagnosis of autism by 30 months of age failed to pass two or more of the five key items of lack of pretend play, protodeclarative pointing, joint-attention, social interest, and social play. Robins et al. (2001) validated the Modified Checklist for Autism in Toddlers (M-CHAT), showing that failure to pass items of protodeclarative pointing, following a point, bringing objects to show parents, social interest, imitation, and responding to one’s name discriminated well between toddlers later diagnosed with or without autism or pervasive developmental disorder. Another recent screening measure, the Screening Tool for Autism in Two-year-olds (STAT; Stone et al., 2000), is an interactive, play-based measure that was developed to facilitate identification of children between the ages of 24 and 36 months who are at risk for autism.

Analysis of home movies has revealed evidence of early impairments in social engagement among young infants who were later diagnosed with autism. Osterling and Dawson’s (1994) analysis of home videos of first birthday parties demonstrated that children who later developed autism showed significantly less responsiveness to their name being called and significantly less eye contact, compared to typically developing children. Furthermore, social deficits were already apparent in home videos taken between 8 and 10 months of age for those children whose parents reported noticing early signs (Werner et al., 2000). Adrien et al. (1993) also found that children who were later diagnosed with autism revealed impairments within the first year of life in social interaction, social smiling, and facial expressions, combined with hypoactivity and poor attention. Comparing children later diagnosed with autism to children with developmental delays, Baranek (1999) validated that only the children who were later diagnosed with autism demonstrated impairments in response to name calling and in orientation to visual stimuli, as well as aversion to touch. Maestro et al. (2005) also suggested that early signs of autism may be detected in home videos even before the first birthday.

In addition to screening studies and retrospective studies of home videos, researchers have also begun to examine groups of siblings of children with autism (SIBS-A) in light of the estimated 3–6% reoccurrence rate for autism in these siblings, compared to a less than 1% risk in the general population (Bailey, Palferman, Heavey, & LeCouteur, 1998; Bolton et al., 1994; Pickles et al., 1995). Employing such prospective studies of younger siblings enables the employment of rigorous methodologies in which all siblings are examined at the same ages with the same measures. As development unfolds and diagnoses of autism and autism related conditions are made, it is possible to compare affected versus unaffected siblings on earlier measures that researchers collected when blind to later outcomes.

Zwaigenbaum et al. (2005) reported preliminary findings regarding the development of SIBS-A from 6 to 24 months. These researchers identified several behavioral markers at 12 months (but not at 6 months) that predicted later diagnoses of autism at 24 months. The risk markers included atypical eye contact, visual tracking, disengagement of visual attention, orienting to name, imitation, social smiling, reactivity, social interest, and sensory-oriented behaviors. Furthermore, siblings with a classification of autism on the ADOS (Lord et al., 2000) at 24 months exhibited temperament characteristics of decreased activity at 6 months, intense distress reactions as well as a tendency to fixate on objects at 12 months, and early language delays at 12 months.

Similarly, Landa and Garrett-Mayer (2006) also initiated a prospective study of a sample of 6-month-old infant SIBS-A and infants with no family history of autism. In this study, at 24 months, participants were classified into three groups: ASD group, LD (language delayed) group, and unaffected group. At 6 months, no significant differences emerged among the three groups on the Mullen Scales of Early Learning (MSEL; Mullen, 1995), but at 14 months the ASD group performed significantly poorer than the unaffected group on most of the MSEL domains with the exception of Visual Reception. By 24 months, the ASD group performed significantly poorer than the unaffected group on all the MSEL scales and significantly poorer than the LD group on Gross and Fine Motor and Receptive Language scales. Furthermore, analyses of the developmental trajectories of these three groups revealed that the ASD group had the slowest developmental trajectory, with a significant decrease in development between 14 and 24 months.

Recently, we reported on the development of SIBS-A at 4 and 14 months (Yirmiya et al., 2006a) and at 24 and 36 months (Yirmiya et al., 2006b) as compared to the development of well-matched siblings of children with typical development (SIBS-TD). We examined siblings’ social engagement as well as cognition and language. Altogether, SIBS-A functioned well on most measures at age 4 months. However, a few differences did emerge. SIBS-A displayed significantly lower levels of social engagement in mother–child face-to-face interaction. Also, during the still-face paradigm, SIBS-A displayed significantly more neutral affect and were less upset by the still face, although on a behavior regulation paradigm (name calling), significantly more SIBS-A searched and found their mother compared to SIBS-TD. At age 14 months, some SIBS-A (about 20%) already showed evidence of communication and language difficulties. Furthermore, correlational analyses revealed that SIBS-A who earlier (at age 4 months) showed more neutral affect and who responded less well to their name being called, employed fewer nonverbal joint attention and requesting behaviors at age 14 months (Yirmiya et al., 2006a). At 24 and 36 months we found that significantly more SIBS-A revealed specific language impairments compared to SIBS-TD. Despite differences in language abilities, the two groups of SIBS-A and SIBS-TD did not differ in their cognitive abilities at either age (Yirmiya et al., 2006b).

The aim of the current paper is, first, to continue our report on the development of cognition and language in these same SIBS-A and SIBS-TD, now aged 4.5 years, and to examine the developmental trajectories from 4 months to 4.5 years of those siblings who revealed an earlier cognitive and/or language delay as compared to the remaining SIBS-A and the group of SIBS-TD. We decided to focus on cognition and language because these two domains comprise part of the phenotype as well as the genotype of autism (Dawson et al., 2002; Muhle, Trentascosta, & Rapin, 2004; Rapin & Dunn, 2003; Tager-Flusberg, & Joseph, 2003). Thus, the present study contributes both to the growing literature on the development of SIBS-A and to the concentrated effort of professionals examining early markers for autism spectrum disorders (ASD) and conditions related to the broad phenotype of ASD.

Method

Participants

The SIBS-A group comprised 39 children having an older sibling with autism. SIBS-A have been seen thus far at five ages (see Table 1): 4 months (n = 21), 14 months (n = 30), 24 months (n = 38), 36 months (n = 39), and 54 months (n = 39).
Table 1

Group characteristics for siblings of children with autism (SIBS-A) and siblings of children with typical development (SIBS-TD) at five ages

 

 

4 months

14 months

24 months

36 months

54 months

SIBS-A (n = 21)

SIBS-TD (n = 21)

SIBS-A (n = 30)

SIBS-TD (n = 31)

SIBS-A (n = 38)

SIBS-TD (n = 38)

SIBS-A (n = 39)

SIBS-TD (n = 39)

SIBS-A (n = 39)

SIBS-TD (n = 39)

Female:Male

8:13

9:12

11:19

13:18

15:23

15:23

15:24

15:24

14:25

14:25

Age in months

M (SD)

4.48 (.81)

4.38 (.74)

14.13 (.29)

14.16 (.27)

24.24 (.42)

24.16 (.35)

36.47 (.64)

36.33 (.62)

54.78 (.92)

55.05 (1.32)

Range

4–6

4–6

14–15

14–16

24–25

24–26

35–38

36–39

54–57

53–59

All probands with autism (10 girls/29 boys) had a previous diagnosis of autism made by independent clinicians. However, for the purpose of this study two trained clinicians who were blind to all the other study procedures confirmed diagnoses of autism using the Autism Diagnostic Observation Schedule—Generic (ADOS-G; Lord et al., 2000) and/or the Autism Diagnostic Interview—Revised (ADI-R; Lord, Rutter, & Le Couteur, 1994). (Some of the probands were recruited before the availability of the ADOS-G in Israel and therefore were administered only the ADI-R). Parents were interviewed with the ADI-R and the Vineland Adaptive Behavior Scales—Interview Edition (VABS; Sparrow, Balla, & Cicchetti, 1985) regarding the probands with autism. Proband IQ was assessed with a standard IQ test appropriate for the proband’s age and/or developmental abilities, i.e., Bayley Scales of Infant Development—2nd edition (Bayley, 1993); Kaufman Assessment Battery for Children (Kaufman & Kaufman, 1983); or Wechsler Intelligence Scale for Children—3rd edition (Wechsler, 1991). Nine of the 39 probands with autism were classified as high functioning (IQ and/or daily living skills scores >70), whereas the remaining 30 displayed IQ and daily living skills scores below 70. Inclusion criteria for families in this group were intact families who had a child with autism and a younger sibling. The families were recruited through treatment centers, special schools, the Israeli national organization for children with autism, and families of children with autism.

The SIBS-TD group comprised 39 siblings of typically developing children (see Table 1 for ns at the five ages). Families in this comparison group were recruited from maternity wards at Hadassa Hospital, Mount Scopus, Jerusalem. Inclusion criteria required that the families were intact at the time of the study’s initiation and that the older child exhibited typical development with no history of any learning and/or emotional difficulties and had not received any specialized interventions such as occupational therapy, speech and language therapy, or psychotherapy based on parental report. In addition, for both groups, inclusion criteria included normal pregnancies with the participating sibling as reported by the parents, with no peri-, pre-, or post-natal difficulties as well as normal gestational age (>36 weeks).

The two groups were matched at 4 months on a one-to-one basis according to chronological age (CA), sex, birth order, number of children in the family, sex of the older proband, temperament profile (ICQ: Infant Characteristics Questionnaire; Bates, Freeland, & Lounsbury, 1979), and Bayley (1993) mental and motor scores. In addition, parents’ age, ethnicity, income, and education level did not significantly differ between the two groups. SIBS-A who joined the study after the age of 4 months were matched to SIBS-TD according to the same variables, as best as possible. Group characteristics at age 4, 14, 24, 36 and 54 months are presented in Table 1.

Measures

The various measures administered to the two groups of siblings at each age are presented in Table 2.
Table 2

Procedures employed for siblings by age

Procedure

Age in months

4

14

24

36

54

Bayley Scales of Infant Development (BSID-II)

  

Reynell Developmental Language Scales (RDLS)

  

  

Kaufman Assessment Battery for Children (K-ABC)

   

Clinical Evaluation of Language Fundamentals-Preschool (CELF-P)

   

Services survey

  

Bayley Scales of Infant Development—2nd Edition (BSID-II)

The BSID-II (Bayley, 1993) is a standardized measure designed to assess the developmental level of infants and toddlers between the ages of 1–42 months. The Mental Developmental Index (MDI) and the Psychomotor Developmental Index (PDI) were calculated at 4, 14, and 24 months. An additional Language Developmental Age score was calculated at 14 months and was used as an index for both receptive and expressive language in all pertaining analyses.

Reynell Developmental Language Scales (RDLS)

The RDLS (Reynell & Grubber, 1990) comprises a standardized measure providing both quantitative and qualitative assessments of expressive language and verbal comprehension. Each of these scales has a mean score of 100 with a standard deviation of 15. An additional average score of the RDLS receptive and expressive scores was calculated.

Kaufman Assessment Battery for Children (K-ABC)

The K-ABC (Kaufman & Kaufman, 1983) consists of a standardized intelligence test designed for children. Its subtests are combined to provide a scaled score for a mental processing scale, with a mean standard score of 100 and a standard deviation of 15.

Clinical Evaluation of Language Fundamentals-Preschool (CELF-Preschool)

The CELF-Preschool (Wiig, Secord, & Semel, 1992) is designed for identification, diagnosis, and follow-up evaluations of language abilities in preschool children. This individually administered test is standardized for children between the ages of 3–6 years. The test provides scaled scores for receptive and expressive verbal abilities as well as a total scale score for all subtests combined.

Survey of Clinical and/or Educational Services

Parents completed a questionnaire (available per request from the authors) regarding possible interventions conducted with the sibling up to the time of the questionnaires’ completion. This questionnaire included a list of ten intervention settings as well as an open question regarding other possible interventions that the parents may have employed for the sibling.

Results

Siblings Identified with Delays

At each age, we identified children who exhibited cognitive and/or language difficulties, i.e., siblings who received scores of at least two standard deviations below average on the cognitive and/or language measures. At 14 months, we identified six SIBS-A who manifested a language delay of 5 months (Verbal IQ of 64), (Yirmiya et al., 2006a). One of these six siblings was suspected at 14 months as possibly having an ASD, and a diagnosis of autistic disorder was confirmed at ages 24 and 36 months, using the ADI-R (Lord et al., 1994) and the ADOS–G (Lord et al., 2000) conducted by two independent trained professionals. This boy’s data (Child E in Table 3) was excluded from all further analyses. No other siblings showed any difficulties that necessitated an evaluation for ASD or for any other diagnostic entity. In this report, we continue to follow the developmental trajectories of the five SIBS-A who were identified as delayed at 14 months, known henceforth as siblings of children with autism who show signs of the broad phenotype at 14 months or “SIBS-A-BP14” (Children A, B, C, D, and F in Table 3).
Table 3

Siblings identified with delays

SIBS-A-BP14 (n = 6)

Sex

4 months

14 months

24 months

36 months

54 months

Cognitiona

Cognitiona

Languageb

Cognitiona

Languagec

Cognitiond

Languagee

Cognitiond

Languagee

A

F

113/109

50/50

9

68/92

63/63/63

91

65/50/61

104

104/96/100

B

F

No data

95/107

9

96/96

76/76/76

94

87/84/84

87

110/98/104

C

M

105/98

97/120

9

98/100

71/76/74

99

100/81/90

106

104/88/97

D

F

91/102

101/100

9

86/84

74/70/72

85

71/90/80

No data

E*

M

99/92

91/92

9

50/53

63/63/63

50

50/50/50

77

50/62/58

F

M

107/104

112/112

9

98/92

76/63/70

107

112/92/102

115

110/99/104

SIBS-A-BP24 (n = 6)

G

M

No data

No data

82/110

71/63/67

114

89/70/80

128

97/85/91

H

M

102/88

101/84

14

76/100

63/64/64

91

89/69/77

91

91/85/87

I

F

105/116

110/116

13

88/100

63/63/63

75

61/50/58

111

98/83/90

J

M

No data

No data

84/107

63/63/63

101

104/86/96

111

106/96/101

K

M

No data

95/120

14

66/92

63/64/64

73

69/50/63

77

91/75/81

L

M

97/101

110/88

14

78/92

63/63/63

77

50/50/50

96

108/85/97

aBSID: Mental score/Motor score

bBSID: Developmental language age (in months)

cRDLS: Receptive score/Expressive score/Average of Receptive and Expressive scores

dK-ABC: Mental score

eCELF-P: Receptive score/Expressive score/Total Language score

* Data of child diagnosed with autism

Note: All scores (except the BSID: Developmental language age) have a mean of 100 and a standard deviation of 15

In our second report (Yirmiya et al., 2006b), we retested participants at ages 24 and 36 months and had six SIBS-A who exhibited cognitive and/or language difficulties according to the same criterion of scoring at least two standard deviations below average on the cognitive (i.e., the child’s BSID-II mental score) and/or language test (i.e., the average of the child’s RDLS receptive and expressive scores, not counting the child with autism). These included four siblings who did not manifest delays at 14 months—Children H, I, K, and L—and two children who were already identified at age 14 months, A and F and who continued to show difficulties at 24 months.

In the current report, with a larger sample than that included in Yirmiya et al. (2006b), we identified 2 additional SIBS-A who manifested delays in cognition and/or language at age 24 months (children G and J in Table 3, who entered the study prior to the 24 months visit). Thus, this subgroup of “SIBS-A-BP24” i.e., SIBS-A, who did not reveal any difficulties at 14 months of age but at age 24 months did reveal difficulties considered to be signs of the broad phenotype of autism, included six siblings. The SIBS-A-BP14 and SIBS-A-BP24 groups do not overlap (that is, the two children who manifested delays at both 14 and 24 months are included only in the SIBS-A-BP14 group).

In comparison to the total of 11 siblings identified with signs of the BAP in the SIBS-A group at 14 and 24 months, only 2 siblings in the SIBS-TD group manifested delays in cognition and/or language, (according to the same criteria, i.e., siblings who received scores of at least 2 standard deviations below average on the cognitive and/or language measures), both identified only at the 24 month time point. No other siblings in the SIBS-TD group were identified at any of the time points.

Developmental Trajectories of SIBS-A and SIBS-TD

To compare developmental trajectories of the sibling groups, analyses of variance were employed for four groups: SIBS-A-BP14 (n = 5, excluding the child with autism), SIBS-A-BP24 (n = 6), SIBS-A-nonBP (n = 27, the remaining SIBS-A who did not reveal cognitive and/or language difficulties), and SIBS-TD (n = 37, excluding the two children identified at 24 months with cognitive and/or language difficulties). Due to the small groups, we also examined effect sizes. Unlike significance tests, effect sizes are independent of sample size. Effect sizes can be interpreted in terms of the percent of nonoverlap of one group’s scores with those of the other group. An ES of 0.0 indicates that the distribution of scores for the one group overlaps completely with the distribution of scores for the other group, thus there is 0% of nonoverlap. An effect size of .80, which is considered as large (Cohen, 1988) indicates a nonoverlap of almost 50% in the two distributions, whereas an effect size of 1.7 indicates a nonoverlap of about 75%. Since our SIBS-A-BP groups were quite small (n = 5 and n = 6 at 14 and 24 months respectively), in addition to significant findings based on group comparisons, we interpreted effect sizes of .80 and larger as indicative of meaningful group differences. All effect sizes are included in Tables 4–6.
Table 4

Cognition: group differences between siblings of children with autism (SIBS-A) and siblings of children with typical development (SIBS-TD)

Age

Scale

 

SIBS-A-BP14

SIBS-A-BP24

SIBS-A-nonBP

SIBS-TD

Post hoc LSD

Effect size

4 months

BSID-II mental

M

104.00

101.33

102.69

108.60

SIBS-A-BP14 = SIBS-A-BP24

d = .37

SD

9.31

4.04

8.72

7.29

SIBS-A-BP14 = SIBS-A-nonBP

d = .15

     

SIBS-A-BP14 = SIBS-TD

d = .55

     

SIBS-A-BP24 = SIBS-A-nonBP

d = .02

     

SIBS-A-BP24 = SIBS-TD

d = 1.23

     

SIBS-A-nonBP < SIBS-TD*

d = .74

BSID-II motor

M

103.25

101.67

98.23

101.98

SIBS-A-BP14 = SIBS-A-BP24

d = .15

SD

4.57

14.01

9.94

9.54

SIBS-A-BP14 = SIBS-A-nonBP

d = .65

     

SIBS-A-BP14 = SIBS-TD

d = .17

     

SIBS-A-BP24 = SIBS-A-nonBP

d = .28

     

SIBS-A-BP24 = SIBS-TD

d = .03

     

SIBS-A-nonBP = SIBS-TD

d = .39

14 months

BSID-II ment al

M

91.00

104.00

114.15

111.75

SIBS-A-BP14 = SIBS-A-BP24

d = .74

SD

23.84

7.35

9.92

8.21

SIBS-A-BP14 < SIBS-A-nonBP****

d = 1.27

     

SIBS-A-BP14 < SIBS-TD****

d = 1.16

     

SIBS-A-BP24 = SIBS-A-nonBP

d = 1.16

     

SIBS-A-BP24 = SIBS-TD

d = .99

     

SIBS-A-nonBP = SIBS-TD

d = .26

BSID -II moto r

M

97.80

102.00

109.25

106.88

SIBS-A-BP14 = SIBS-A-BP24

d = .18

SD

27.70

18.62

9.23

11.66

SIBS-A-BP14 = SIBS-A-nonBP

d = .55

     

SIBS-A-BP14 = SIBS-TD

d = .43

     

SIBS-A-BP24 = SIBS-A-nonBP

d = .49

     

SIBS-A-BP24 = SIBS-TD

d = .31

     

SIBS-A-nonBP = SIBS-TD

d = .23

24 months

BSID-II mental

M

89.20

79.00

107.23

102.83

SIBS-A-BP14 = SIBS-A-BP24

d = .96

SD

12.85

7.67

12.29

11.40

SIBS-A-BP14 < SIBS-A-nonBP***

d = 1.43

     

SIBS-A-BP14 < SIBS-TD*

d = 1.12

     

SIBS-A-BP24 < SIBS-A-nonBP****

d = 2.71

     

SIBS-A-BP24 < SIBS-TD****

d = 2.45

     

SIBS-A-nonBP = SIBS-TD

d = .37

BSID-II motor

M

92.80

100.17

103.84

103.23

SIBS-A-BP14 = SIBS-A-BP24

d = 1.1

SD

5.93

7.44

8.32

8.40

SIBS-A-BP14 < SIBS-A-nonBP**

d = 1.53

     

SIBS-A-BP14 < SIBS-TD**

d = 1.44

     

SIBS-A-BP24 = SIBS-A-nonBP

d = .47

     

SIBS-A-BP24 = SIBS-TD

d = .39

     

SIBS-A-nonBP = SIBS-TD

d = .07

36 mo nths

K- ABC mental

M

95.20

88.50

105.19

103.93

SIBS-A-BP14 = SIBS-A-BP24

d = .51

SD

8.32

16.54

14.69

16.12

SIBS-A-BP14 = SIBS-A-nonBP

d = .84

     

SIBS-A-BP14 = SIBS-TD

d = .68

     

SIBS-A-BP24 = SIBS-A-nonBP

d = 1.07

     

SIBS-A-BP24 = SIBS-TD

d = .94

     

SIBS-A-nonBP = SIBS-TD

d = .08

54 mont hs

K-ABC mental

M

103.00

102.33

107.71

111.59

SIBS-A-BP14 = SIBS-A-BP24

d = .05

SD

11.69

17.10

16.57

13.94

SIBS-A-BP14 = SIBS-A-nonBP

d = .33

     

SIBS-A-BP14 = SIBS-TD

d = .67

     

SIBS-A-BP24 = SIBS-A-nonBP

d = .32

     

SIBS-A-BP24 = SIBS-TD

d = .59

     

SIBS-A-nonBP = SIBS-TD

d = .25

Note: SIBS-A-BP14 = siblings of children with autism who revealed difficulties within the broad phenotype of autism at age 14 months; SIBS-A-BP24 = siblings of children with autism who revealed difficulties within the broad phenotype of autism at age 24 months; SIBS-A-nonBP = siblings who did not reveal such difficulties; SIBS-TD = siblings of children with typical development

* P < .05, ** P < .01, *** P < .005, **** P < .001

Cognition

Descriptive data and analyses for the four groups’ cognitive scores at each age are presented in Table 4 and in Fig. 1. At 4 months, these analyses revealed a near significant group effect for the BSID-II mental score (F(3, 56) = 2.67, P = .056, ηp2 = .13). As seen in Table 4 and Fig. 1, LSD post-hoc tests revealed that, as a group, SIBS-A-nonBP had a significantly lower BSID-II mental score than SIBS-TD. In addition, examination of effect sizes indicated a large effect size for the comparison between SIBS-A-BP24 and SIBS-TD. Thus, examination of significant group differences and effect sizes revealed that those siblings who were later identified as showing some delays at age 24 months (SIBS-A-BP24) had lower mental scores compared to SIBS-TD at age 4 months, and those siblings who developed well throughout the study (SIBS-A-nonBP) still scored lower compared to SIBS-TD at age 4 months. There were neither significant group differences nor any large size effects on the BSID-II motor score.
https://static-content.springer.com/image/art%3A10.1007%2Fs10803-006-0341-5/MediaObjects/10803_2006_341_Fig1_HTML.gif
Fig. 1

Cognition: developmental trajectories of SIBS-A and SIBS-TD

At 14 months, as expected based on group membership, a significant overall group effect emerged for the BSID-II mental score (F(3, 65) = 7.50, P = .00, ηp2 = .26). As seen in Table 4 and Fig. 1, LSD post-hoc tests revealed that, as a group, SIBS-A-BP14 had a significantly lower BSID-II mental score than SIBS-A-nonBP and SIBS-TD. No significant differences emerged on the BSID-II motor score among the four groups. Examination of effect sizes revealed that in addition to the aforementioned group differences, large effect sizes emerged for the comparisons between SIBS-A-BP24 and SIBS-nonBP and between SIBS-A-BP24 and SIBS-TD. Thus, once more, there is indication that siblings who were later identified as showing some delays (SIBS-A-BP14 and SIBS-A-BP24) were already showing some cognitive delays compared to SIBS-A-nonBP and SIBS-TD at age 14 months.

At age 24 months, a significant overall group effect emerged for the BSID-II mental score (F(3, 73) = 11.72, P = .00, ηp2 = .33) and for the BSID-II motor score (F(3, 72) = 2.83, P = .04, ηp2 = .11). Post-hoc LSD tests revealed that as a group, SIBS-A-BP14 had a significantly lower mental and motor score than both SIBS-A-nonBP and SIBS-TD. SIBS-A-BP24 had a significantly lower mental score (but not motor score) than both SIBS-A-nonBP and SIBS-TD. Additionally, examination of effect sizes revealed that the difference in performance between SIBS-A-BP14 and SIBS-A-BP24 was large, with SIBS-A-BP14 performing better on the mental scale, and SIBS-A-BP24 performing better on the motor scale.

At age 36 months, there were no significant group differences on the K-ABC mental score. Yet examination of effect sizes revealed large size group differences for the comparisons between SIBS-A-BP14 and SIBS-A-nonBP, SIBS-A-BP24 and SIBS-A-nonBP, and SIBS-A-BP24 and SIBS-TD, indicating lower cognitive scores for those SIBS-A who were identified with delays at 14 and 24 months.

At age 54 months there were no significant differences among the four groups. Effect sizes for all post hoc comparisons were smaller than .80.

Receptive Language

Descriptive data and analyses for receptive language at 14, 24, 36, and 54 months are presented in Table 5 and in Fig. 2. As expected, a significant overall group effect emerged for the language score at age 14 months (F(3, 65) = 79.98, P = .00, ηp2 = .78), as well as significant differences between subgroups employing LSD post-hoc tests. As a group, SIBS-A-BP14 revealed a significantly lower language score than did the SIBS-A-BP24, SIBS-A-nonBP, and SIBS-TD. In addition, examination of effect sizes revealed large group differences between SIBS-A-BP24 and SIBS-A-nonBP and between SIBS-A-BP24 and SIBS-TD, indicating lower language scores for SIBS-A-BP24. Thus, similar to cognition, siblings who were later identified at 24 months as showing delays were already showing some delays in their language development compared to SIBS-A-nonBP and SIBS-TD at age 14 months.
Table 5

Receptive language scores: differences between siblings of children with autism (SIBS-A) and siblings of children with typical development (SIBS-TD)

Age

Scale

 

SIBS-A-BP14

SIBS-A-BP24

SIBS-A-nonBP

SIBS-TD

Post hoc LSD

Effect size

14 months

BSID-II language

M

64.29

98.21

103.93

102.86

SIBS-A-BP14 < SIBS-A-BP24****

d = 13.41

SD

.00

3.57

7.50

4.52

SIBS-A-BP14 < SIBS-A-nonBP****

d = 7.48

     

SIBS-A-BP14 < SIBS-TD****

d = 12.05

     

SIBS-A-BP24 = SIBS-A-nonBP

d = .97

     

SIBS-A-BP24 = SIBS-TD

d = 1.14

     

SIBS-A-nonBP = SIBS-TD

d = .17

24 months

RDLS receptive

M

72.00

64.33

106.15

98.97

SIBS-A-BP14 = SIBS-A-BP24

d = 1.71

SD

5.43

3.27

14.27

13.20

SIBS-A-BP14 < SIBS-A-nonBP****

d = 3.16

     

SIBS-A-BP14 < SIBS-TD****

d = 2.67

     

SIBS-A-BP24 < SIBS-A-nonBP****

d = 4.04

     

SIBS-A-BP24 < SIBS-TD****

d = 3.60

     

SIBS-A-nonBP > SIBS-TD*

d = .52

36 months

CELF-P receptive

M

87.00

77.00

113.65

113.55

SIBS-A-BP14 = SIBS-A-BP24

d = .50

SD

19.58

20.33

11.72

12.66

SIBS-A-BP14 < SIBS-A-nonBP****

d = 1.65

     

SIBS-A-BP14 < SIBS-TD****

d = 1.62

     

SIBS-A-BP24 < SIBS-A-nonBP****

d = 2.16

     

SIBS-A-BP24 < SIBS-TD****

d = 2.49

     

SIBS-A-nonBP = SIBS-TD

d = .008

54 mont hs

CELF-P receptive

M

107.00

98.50

107.46

105.61

SIBS-A-BP14 = SIBS-A-BP24

d = 1.50

SD

3.46

7.23

14.19

9.80

SIBS-A-BP14 = SIBS-A-nonBP

d = .04

     

SIBS-A-BP14 = SIBS-TD

d = .19

     

SIBS-A-BP24 = SIBS-A-nonBP

d = .80

     

SIBS-A-BP24 = SIBS-TD

d = .82

     

SIBS-A-nonBP = SIBS-TD

d = .15

Note: SIBS-A-BP14 = siblings of children with autism who revealed difficulties within the broad phenotype of autism at age 14 months; SIBS-A-BP24 = siblings of children with autism who revealed difficulties within the broad phenotype of autism at age 24 months; SIBS-A-nonBP = siblings who did not reveal such difficulties; SIBS-TD = siblings of children with typical development

* P < .05, **** P < .001

https://static-content.springer.com/image/art%3A10.1007%2Fs10803-006-0341-5/MediaObjects/10803_2006_341_Fig2_HTML.gif
Fig. 2

Receptive language score: developmental trajectories of SIBS-A and SIBS-TD

At age 24 months, a significant difference emerged among the four groups on the RDLS receptive score (F(3, 72) = 23.95, P = .00, ηp2 = .50). According to the post-hoc LSD analyses, as a group, SIBS-A-BP14 revealed a significantly lower receptive score than both SIBS-A-nonBP and SIBS-TD. As a group, SIBS-A-BP24 revealed a significantly lower receptive score than both SIBS-A-nonBP and SIBS-TD. Furthermore, SIBS-TD revealed a significantly lower receptive score than SIBS-A-nonBP. Examination of the effect sizes revealed a large effect size for the comparison between SIBS-A-BP14 and SIBS-A-BP24. Thus at age 24 months, SIBS-A-BP24 had a lower receptive language score compared to SIBS-BP14.

At age 36 months, significant differences also emerged among the four groups for the CELF-P receptive score (F(3, 73) = 18.24, P = .00, ηp2 = .43). According to the post-hoc LSD test, as a group, SIBS-A-BP14 revealed a significantly lower receptive score than both SIBS-A-nonBP and SIBS-TD. As a group, SIBS-A-BP24 revealed a significantly lower receptive score than both SIBS-A-nonBP and SIBS-TD. Examination of effect sizes did not reveal any additional differences.

At 54 months, no significant differences emerged among the four groups on the CELF-P receptive score, yet examination of effect sizes revealed that similar to age 24 months, SIBS-A-BP14 had a higher language score compared to SIBS-A-BP24 now at age 54 months, and that SIBS-A-BP24 had a lower language score compared to SIBS-A-nonBP and SIBS-TD.

Expressive Language

Descriptive data and analyses of expressive language at 14, 24, 36, and 54 months are presented in Table 6 and in Fig. 3. Data regarding the BSID-II developmental language scores at 14 months of age are the same as those presented above for receptive language at age 14 months, because we assessed global language, undifferentiated into receptive versus expressive abilities (see Method section).
Table 6

Expressive language scores: group differences between siblings of children with autism (SIBS-A) and siblings of children with typical development (SIBS-TD)

Age

Scale

 

SIBS-A-BP14

SIBS-A-BP24

SIBS-A-nonBP

SIBS-TD

Post hoc LSD

Effect size

14 months

BSID-II language

M

64.29

98.21

103.93

102.86

SIBS-A-BP14 < SIBS-A-BP24****

d = 13.41

SD

.00

3.57

7.50

4.52

SIBS-A-BP14 < SIBS-A-nonBP****

d = 7.48

     

SIBS-A-BP14 < SIBS-TD****

d = 12.05

     

SIBS-A-BP24 = SIBS-A-nonBP

d = .97

     

SIBS-A-BP24 = SIBS-TD

d = 1.14

     

SIBS-A-nonBP = SIBS-TD

d = .17

24 months

RDLS expressive

M

69.60

63.33

90.54

84.26

SIBS-A-BP14 = SIBS-A-BP24

d = 1.36

SD

6.50

.52

16.73

12.87

SIBS-A-BP14 < SIBS-A-nonBP***

d = 1.65

     

SIBS-A-BP14 < SIBS-TD*

d = 1.44

     

SIBS-A-BP24 < SIBS-A-nonBP****

d = 2.30

     

SIBS-A-BP24 < SIBS-TD***

d = 2.30

     

SIBS-A-nonBP = SIBS-TD

d = .42

36 months

CELF-P expressive

M

79.40

62.50

92.35

91.63

SIBS-A-BP14 > SIBS-A-BP24*

d = 1.05

SD

17.02

14.96

11.52

12.42

SIBS-A-BP14 < SIBS-A-nonBP*

d = .89

     

SIBS-A-BP14 < SIBS-TD*

d = .82

     

SIBS-A-BP24 < SIBS-A- nonBP****

d = 2.24

     

SIBS-A-BP24 < SIBS-TD****

d = 2.12

     

SIBS-A-nonBP = SIBS-TD

d = .06

54 months

CELF-P expre ssive

M

95.25

84.83

99.07

96.79

SIBS-A-BP14 = SIBS-A-BP24

d = 1.76

SD

4.99

6.71

11.85

9.78

SIBS-A-BP14 = SIBS-A- nonBP

d = .42

     

SIBS-A-BP14 = SIBS-TD

d = .20

     

SIBS-A-BP24 < SIBS-A- nonBP***

d = 1.48

     

SIBS-A-BP24 < SIBS-TD***

d = 1.43

     

SIBS-A-nonBP = SIBS-TD

d = .20

Note: SIBS-A-BP14 = siblings of children with autism who revealed difficulties within the broad phenotype of autism at age 14 months; SIBS-A-BP24 = siblings of children with autism who revealed difficulties within the broad phenotype of autism at age 24 months; SIBS-A-nonBP = siblings who did not reveal such difficulties; SIBS-TD = siblings of children with typical development

* P < .05, *** P < .005, **** P < .001

https://static-content.springer.com/image/art%3A10.1007%2Fs10803-006-0341-5/MediaObjects/10803_2006_341_Fig3_HTML.gif
Fig. 3

Expressive language score: developmental trajectories of SIBS-A and SIBS-TD

At age 24 months, an overall group effect emerged among the four groups on the RDLS expressive score (F(3, 72) = 8.40, P = .00, ηp2 = .26). According to the post-hoc LSD tests, as a group, SIBS-A-BP14 revealed a significantly lower expressive score than both SIBS-A-nonBP and SIBS-TD. As a group, SIBS-A-BP24 also revealed a significantly lower expressive score than both SIBS-A-nonBP and SIBS-TD. Examination of the effect sizes revealed a large effect for the comparison between SIBS-A-BP14 and SIBS-A-BP24. Thus at age 24 months, similar to the findings regarding receptive language ability, SIBS-A-BP24 had a lower expressive language score compared to SIBS-A-BP14.

At age 36 months, significant differences also emerged among the four groups for the CELF-P expressive score (F(3, 73) = 10.98, P = .00, ηp2 = .31). According to the post-hoc LSD tests, as a group, SIBS-A-BP14 revealed a significantly higher expressive score than SIBS-A-BP24 and a significantly lower score than both SIBS-A-nonBP and SIBS-TD. As a group, SIBS-A-BP24 revealed a significantly lower expressive score than both SIBS-A-nonBP and SIBS-TD. Examination of effect sizes did not reveal any additional differences.

At 54 months of age, a significant overall group effect emerged among the four groups on the CELF-P expressive score (F(3, 72) = 3.18, P = .03, ηp2 = .12). According to the post-hoc LSD tests, only the SIBS-A-BP24 subgroup revealed a significantly lower expressive score than both SIBS-A-nonBP and SIBS-TD. Yet examination of effect sizes revealed that similar to age 24 months and to findings regarding receptive language, SIBS-A-BP14 had a higher expressive language score compared to SIBS-A-BP24 now at age 54 months.

Clinical and/or Educational Interventions

Only nine of the SIBS-A (including the child diagnosed with autism) were receiving, or had received, some professional help when we saw them at 54 months. These included six who were part of a BP group and three who were part of the SIBS-A-nonBP group. The services reported were mostly occupational therapy and speech and language therapy, as well as remedial horseback riding. Of the six SIBS-A who manifested delays in language at age 14 months (i.e., the SIBS-A-BP14 group), only two received services. The child diagnosed with autism (Child E on Table 3) received speech and language therapy, occupational therapy, remedial horseback riding, music therapy, and psychological counseling from age 1 year to the present data collection at 54 months. The other child who received services in the SIBS-A-BP14 group (Child D in Table 3) began occupational therapy after receiving the testing results at 36 months.

Of the six children who were identified with delays in cognition and/or language at age 24 months (i.e., the SIBS-A-BP24 group), four children (Children I, J, K, and L) received speech and language therapy that began around age 2 years and was continuing at the 54-month testing period. These data suggest that the four children who were identified as showing a language delay at age 14 months (Children A, B, C, and F) actually caught up spontaneously without any intervention.

Discussion

Our data suggests that, altogether, most SIBS-A who participated in this study developed well in terms of cognition and language. However, although matched to typical peers on a one-on-one basis at age 4 months, by 14 months, a sub-sample of SIBS-A emerged who exhibited significantly lower cognitive and/or language abilities compared to the remaining SIBS-A and to the SIBS-TD. Similarly, at age 24 months, another sub-sample of SIBS-A was identified, and this group, too, showed significantly lower scores on the cognitive and/or language measures employed in the current study. Whereas differences in cognitive ability disappeared by age 54 months, some language difficulties were still evident for some siblings at 54 months. These data are intriguing in that this is the first time in which cognition and language are studied longitudinally in a sample of young SIBS-A. Our data suggests that at 54 months of age, receptive and expressive language abilities remain areas of difficulty for some SIBS-A. It is important to note that although language difficulties were still experienced at age 54 months by SIBS-A who were identified at age 14 and/or 24 months as showing some delays, most SIBS-A (i.e., SIBS-A-nonBP) did not differ from SIBS-TD in their language abilities as indicated by nonsignificant group differences and very small effect sizes.

Our findings of both similarities and differences between SIBS-A and SIBS-TD in language are similar to the findings reported for older siblings. Whereas some researchers report lower verbal abilities in SIBS-A (August, Stewart & Tsai, 1981; Bolton et al., 1994; Leboyer, Plumet, Goldblum, Perez-Diaz, & Marchaland, 1995; Piven et al., 1990; Plumet, Goldblum, & Leboyer, 1995), others report no significant differences (Folstein et al., 1999; Gillberg, Gillberg, & Stuffenburg, 1992; Happè, Briksman, & Frith, 2001; Smalley & Asarnow, 1990) and still others actually report that SIBS-A have higher verbal abilities (i.e., Pilowsky, Yirmiya, Shalev, & Gross-Tsur, 2003). These contradictory findings may be due to differences in the tests that were employed, and in the comparison groups tested (Shaked & Yirmiya, 2004; Yirmiya, Shaked, & Erel, 2001). Moreover, as language abilities develop throughout childhood and adolescence, it may also be that some children are more vulnerable to, and experience more difficulties in specific developmental stages. As additional data from other longitudinal studies involving SIBS-A become available, the developmental trajectory of language development and abilities will become clearer, as well as the association among cognition, language and the broad phenotype of autism in siblings.

Although the autism and typical development groups were matched on a one-to-one basis, somewhat different developmental trajectories emerged. More specifically, within the group of SIBS-A, at 14 months of age, five children were identified as having delays in language development and one child was diagnosed with autism. At 24 months of age, eight SIBS-A were identified as having cognitive and/or language delay/s (comprising two who were already identified at age 14 months, and six additional siblings who were not identified at age 14 months). Thus, altogether between the ages of 14 and 36 months, 11 of the 39 SIBS-A revealed delays in cognitive or language development, and 1 child was diagnosed with autism.

As can be seen, for three children delays were transient—found only at 14 months and not later on again (although still they had some minor-mild difficulties as revealed by language scores lower than average at 24 months but not low enough to be included in the BP group) and for six others, the delays were not found early (either because they were functioning well n = 4 or because they entered the study prior to testing at 24 months, n = 2). These findings contrasted with the SIBS-TD group, in which only 2 of the 39 siblings manifested delays in cognition and/or language. Thus, 11 of the 39 SIBS-A revealed difficulties considered to be signs of the broad phenotype of autism at a very young age. These data indicating lower language abilities replicate those reported by Landa and Garrett-Mayer (2006), Stone and Henderson (2005) and Zwaigenbaum et al. (2005) regarding young SIBS-A and those previously reported for older siblings and for parents of children with autism (Bolton et al., 1994; Leboyer et al., 1995; Le Couteur et al., 1996; Piven et al., 1990; Plumet et al., 1995). Similar data regarding early language delays that diminished during the first few years of life were recently reported by Stone and Henderson (2005).

Our sample is relatively small and some of the identified SIBS-A with delays joined the study close to age 24 months, possibly due to parental concerns (n = 2). Therefore, it remains to be seen whether other groups studying larger samples of young SIBS-A will replicate these data. Furthermore, we investigated only two domains—cognition and language—and even within these domains, our assessment did not include important aspects of language such as pragmatics, which is known to be impaired in individuals with autism and members of their families (Landa et al., 1992; Ozonoff & Miller, 1996). Thus, future studies investigating other domains of development will be crucial to elaborate our understanding of the development, abilities, and difficulties of SIBS-A. In addition, perhaps some of the effects observed at younger ages will reappear now during elementary school. Developmentally, children may possibly catch up with language at the age of 54 months, but some of their earlier difficulties may reappear later as learning difficulties in school. We are currently continuing our study with the siblings and are retesting them at age 7 years, with a focus on cognition, language, and learning difficulties.

Finally, the ethical considerations involved in studies such as ours must be noted. Although well known that SIBS-A are at greater risk for developmental delays, research is at a very early stage regarding knowledge about the development of these siblings during the first few years of life. Our data suggests that most siblings who manifested lower language abilities at age 14 months were functioning well at 54 months without receiving any intervention. Although one may argue that intervention or prevention services may enhance the development of these siblings, one needs to consider the enormous stress that these families are already experiencing regarding their child who is diagnosed with autism. Until we know more about the developmental sequels of these siblings, we should be careful not to set a false alarm for parents while alerting them about their children’s possible difficulties.

Acknowledgments

This study was supported by the United States-Israel Binational Science Foundation (BSF), grants number 94–66/3 and 97-00073, awarded to NY and MS. We are grateful to the families who took part, for their cooperation. We also thank Tammy Pilowsky, Simon Baron-Cohen, and Ruth Feldman for their contributions.

Copyright information

© Springer Science+Business Media, LLC 2006