Meaningful differentiation of children who are prenatally

Original Article Neurodevelopmental Functioning in Children With FAS, pfas, and ARND Ira J. Chasnoff, MD, Anne M. Wells, PhD, Erin Telford, PsyD, Christine Schmidt, PsyD, Gwendolyn Messer, MD ABSTRACT: Objective: The purpose of this article is to compare the neurodevelopmental profiles of 78 foster and adopted children with fetal alcohol syndrome (FAS), partial FAS (pfas), or alcohol-related neurodevelopmental disorder (ARND). Method: Seventy-eight foster and adopted children underwent a comprehensive diagnostic evaluation. By using criteria more stringent than those required by current guidelines, the children were placed in 1 of 3 diagnostic categories: FAS, pfas, or ARND. Each child was evaluated across the domains of neuropsychological functioning most frequently affected by prenatal exposure to alcohol. Multivariate analyses of variance were conducted to examine differences in neuropsychological functioning between the 3 diagnostic groups. Descriptive discriminant analyses were performed in follow-up to the multivariate analyses of variance. Results: The children in the 3 diagnostic categories were similar for descriptive and child welfare variables. Children with FAS had significantly decreased mean weight, height, and head circumference. Children with FAS exhibited the most impaired level of general intelligence, significantly worse language-based memory compared with children with ARND, and significantly poorer functional communication skills than children with pfas. On executive functioning, the FAS group of children performed significantly worse on sequencing and shift than either the pfas or ARND groups. Children with pfas and ARND were similar in all neurodevelopmental domains that were tested. Conclusion: The children who met tightly defined physical criteria for a diagnosis of FAS demonstrated significantly poorer neurodevelopmental functioning than children with pfas and ARND. Children in these latter 2 groups were similar in all neurodevelopmental domains that were tested. (J Dev Behav Pediatr 31:192 201, 2010) Index terms: fetal alcohol syndrome, alcohol related neurodevelopmental disorder, neurodevelopment. Meaningful differentiation of children who are prenatally exposed to alcohol has proven to be an ongoing challenge. Although the diagnosis of fetal alcohol syndrome (FAS) is based on 3 criteria: growth retardation, central nervous system impairment, and characteristic facial dysmorphology, 1 most children with prenatal exposure to alcohol do not fully manifest these criteria but demonstrate a wide range of less specific physical changes and neurodevelopmental deficits. To address this issue, diagnostic terminology has been expanded to include partial FAS (pfas), which is applied to those children who have been prenatally exposed to alcohol, meet criteria for facial dysmorphology and neurodevelopmental deficits, but have normal growth (height and weight) patterns, and alcohol-related neurodevelopmen- From the Children s Research Triangle, Chicago, IL. Received November 11, 2009; accepted January 18, 2010. This investigation was supported by Grant U84/CCU520164 from the Centers for Disease Control and Prevention, US Department of Health and Human Services. Registered as a clinical trial at www.clinicaltrials.gov, number NCT00164346. Address for reprints: Ira J. Chasnoff, MD, 180 North Michigan Avenue, Suite 700, Chicago, IL 60601; e-mail: ichasnoff@cr-triangle.org. Copyright 2010 Lippincott Williams & Wilkins tal disorder for those children who have confirmed prenatal alcohol exposure and neurodevelopmental deficits but do not meet physical criteria required for the diagnosis of FAS or pfas. 2 5 Over the years, the relationship between functional impairment and the physical manifestations of FAS, pfas, and alcohol-related neurodevelopmental disorder has remained unclear. Although the presence of facial dysmorphology has always been recognized as an essential component of diagnosis, the degree of dysmorphology necessary to make a diagnosis of FAS varies between the various diagnostic guidelines. 4,5 In 2001, Astley and Clarren 4 found that alcohol-exposed children with more severe facial phenotypes demonstrated more impaired levels of cognitive, neurodevelopmental, and visual motor functioning. Subsequently, Astley 5 emphasized the importance of facial criteria in concluding that a child in fact has FAS. Most recently, magnetic resonance studies of children with fetal alcohol spectrum disorders demonstrated increased brain damage that correlated with more severe facial dysmorphology and more severe neurodevelopmental dysfunction. 6,7 On the other hand, Mattson et al 8 found that the degree of the severity of neurodevelopmental deficits evident in children with prenatal alcohol exposure is independent of the physical features associated with FAS. There is also research that 192 www.jdbp.org Journal of Developmental & Behavioral Pediatrics

suggests that some children with alcohol-related neurodevelopmental disorder may have more pronounced executive functioning deficits than those children with the full expression of FAS. 9,10 Although a conceptual model of common deficits that occur across the fetal alcohol spectrum has been proposed, 11 it is important to clarify the relationship between the presence and severity of physical features of FAS (growth impairment and facial dysmorphology) and degree of neurodevelopmental deficit. This is especially true in light of current work to identify neurodevelopmental profiles for children with complete, partial, or no physical features of prenatal alcohol exposure. 12 The purpose of this article is to compare the neurodevelopmental profiles of 78 foster and adopted children with FAS, pfas, or alcohol-related neurodevelopmental disorder, using tightly defined diagnostic criteria based on multidisciplinary assessments by a team of pediatricians, psychologists, and mental health specialists. METHODS Children s Research Triangle focuses on evaluation and treatment of high-risk children, key among them foster and adopted children. Over a 4-year period, 90 foster and adopted children, ages 6 years to 11 years 11 months, who had a confirmed history of prenatal alcohol exposure were recruited to participate in a study of the treatment of fetal alcohol syndrome (FAS)/alcohol-related neurodevelopmental disorder. Twelve declined participation because of geographic distance from the program; these children did not differ in demographic characteristics, child welfare background, reason for referral, or substance exposure histories from the 78 children whose families agreed to participate. Consents for participation were signed by the foster or adoptive parent and, as appropriate, the Office of the Guardian of the Illinois Department of Children and Family Services. In addition, each child signed an assent for participation. All procedures for this study were approved by the Institutional Review Board of Department of Children and Family Services and the Institutional Review Board of the Centers for Disease Control and Prevention, United States Department of Health and Human Services. Child Assessment The initial evaluation for each child consisted of a full pediatric, neurological, and dysmorphology diagnostic examination conducted by a team of board-certified pediatricians. Each child s prenatal alcohol exposure (yes/ no) and exposure to tobacco or illicit drugs was verified through maternal admission or positive toxicology for the mother or newborn as documented in the mother s prenatal or birth records and/or the child s birth, medical, child welfare, or adoption records. Information regarding specific dosage and frequency of maternal alcohol, tobacco, and illicit drug use was not available for most children. The child s general health, neurological functioning, and developmental status were evaluated through a review of the child s health, behavioral health, adoption, child welfare, and school records, and full physical examination. Children with a history of significant head trauma, current or historical evidence of lead poisoning, or evidence of a genetic or dysmorphic syndrome other than FAS were excluded from the study. A digital facial photograph of each child was taken following the guidelines established by Astley and Clarren, 13 and measurements of palpebral fissure length and intercanthal distance were calculated using the recommended formulae. The philtrum and lip ranks (Ranks 1 through 5) were assigned by the pediatrician during the examination based on the established grading system 13 and were confirmed through computer-generated upper lip circularity calculations. After the medical examination, the child and family underwent a clinical interview with a licensed psychologist, and the child was evaluated under the direction of a doctoral level psychologist using instruments that assess child psychologic and neurodevelopmental functioning across several domains. Measures Instruments used for the neurodevelopmental battery were selected to assess key domains of neuropsychologic impairment frequently seen as the result of prenatal exposure to alcohol. 3,7,8 General Intelligence The Wechsler Intelligence Scale for Children-third edition 14 was used to assess the children s global intellectual functioning. The Wechsler Intelligence Scale for Children-third edition consists of 12 subtests that combine to form a Verbal IQ score, a Performance IQ score, and a Full Scale IQ score, and 4 other indices (Verbal Comprehension, Perceptual Organization, Freedom from Distractibility, and Processing Speed). Average reliability coefficients are.95,.91, and.96 for the Verbal IQ, Performance IQ, and Full Scale IQ scores, respectively, across all age groups. For the 4 indices, the average reliability coefficients range from 0.85 to 0.94. Executive Functioning Executive functioning was assessed through administration of 3 instruments, each of which addresses a different aspect of executive functioning skills. The Behavior Rating Inventory of Executive Function 15 is a standardized questionnaire that asks parents to rate the frequency (i.e., never, sometimes, and often) of 86 behaviors related to executive functioning in children. The Behavior Rating Inventory of Executive Function assesses aspects of daily functioning that are accessible to parents and may not be assessed in the structured nature of the clinical evaluation setting, which can hide deficits that are apparent in the child s everyday environment. 16 The Behavior Rating Inventory of Executive Function has shown strong internal consistency (r.80.95) and test/retest reliability (r.76.85 for normative samples and.72.84 for clinical samples). A second assessment of executive functioning, the Children s Color Trails Test, 17 measures cognitive flexi- Vol. 31, No. 3, April 2010 2010 Lippincott Williams & Wilkins 193

bility. To cover all children in our sample, we used the published clinical norms. 17 The Children s Color Trails Test has demonstrated good reliability, with alternate forms showing high correlations (r.85.90), and significant test-retest correlations (r.46.68). The Wisconsin Card Sorting Test-64 18 was originally developed to assess concept formation and the ability to shift cognitive strategies in response to changing environmental contingencies. The Wisconsin Card Sorting Test-64 often is referred to as a measure of frontal or prefrontal functioning. The Wisconsin Card Sorting Test-64 is valid for individuals aged 6 years 6 months through 89 years old with high levels of validity and reliability (r.91.96). Achievement To examine basic achievement in word reading, spelling, and arithmetic, children completed the Wide Range Achievement Test third edition. 19 The Wide Range Achievement Test third edition shows adequate internal consistency, with median test coefficient alphas across age groups ranging from.85 to.95 over subtests. Median correlations across age groups between alternate forms of the Wide Range Achievement Test third edition are.92,.93, and.89 for the word recognition, spelling, and arithmetic subtests, respectively. Test-retest reliability also is strong, as corrected stability coefficients range from.91 to.98 on individual subtests. Memory The Wide Range Assessment of Memory and Learning Screening 20 was administered to the children to assess memory. The Wide Range Assessment of Memory and Learning Screening includes 4 subtests. The Verbal Learning subtest assesses the child s verbal memory by asking the child to actively learn a list of nonrelated words. The Story Memory subtest examines memory for semantically related verbal material. The Picture Memory subtest evaluates visual memory for objects presented in a complex, meaningful picture context, and the Design Memory subtest evaluates the child s ability to remember and reproduce simple visual material after a brief delay. Median coefficient alphas across age groups for the 4 subtests of the Wide Range Assessment of Memory and Learning Screening range from 0.78 to 0.86, indicating strong internal consistency. Adaptive Living Skills The Vineland Adaptive Behavior Scales, 21 an interview conducted with parents, was administered to assess the child s adaptive functioning in the areas of communication, daily living, and socialization. With regard to internal consistency, split half reliability medians range from.83 to.94 for Domain and Composite scores and range from.69 to.84 for subdomains. Test-retest reliability for the 3 domains and the composite score has been strong, with most coefficients falling in the.80 to.90 range. Behavior The Child Behavior Checklist 22 is an empirically based instrument that provides standardized ratings of a child s behavioral and emotional problems and competencies. The Child Behavior Checklist is completed by the parent or caregiver and consists of 118 questions related to the child s behavior. Each of the items is scored on a0to2 scale. The 1991 Child Behavior Checklist profile provides normalized T scores and percentiles for 3 competence scales, total competence, 8 cross-informant syndrome scales, and internalizing, externalizing, and total problems. Syndrome scores on the Child Behavior Checklist are commonly analyzed in terms of discrete levels that delineate between normal and abnormal. Syndrome scales derived from these procedures are labeled withdrawn, somatic complaints, anxious/depressed, social problems, thought problems, attention problems, delinquent behavior, and aggressive behavior. Borderline and clinical ranges for each syndrome were calculated based on normative sample results. Diagnostic Assignment Based on their completed comprehensive evaluation, children were assigned an alcohol exposure-related diagnosis based on the following criteria: Growth retardation: current or past weight and/or height less than 3rd percentile adjusted for age and gender. We tightened the growth criteria from 10th percentile, as recommended in the CDC diagnostic and referral guidelines, 3 to 3rd percentile, as recommended by Astley, 5 because of the high rates of diagnostic misclassification 2,23 and the reduced specificity of using a threshold of 10th percentile for denoting growth retardation. Growth criteria below 3rd percentile also align more closely with the original definition of FAS. 1 Facial dysmorphology: abnormal measurements of the upper lip (Rank 4 or 5) and the philtrum (Rank 4 or 5) and shortened palpebral fissures greater than 2 SD below the mean according to analysis of facial features using the Lip-Philtrum Guide and digital facial photograph based on the criteria of Astley and Clarren. 4,13 Central nervous system abnormalities: demonstration of structural, neurological, or functional central nervous system deficits 12 as documented by the presence of microcephaly (current head circumference below 3rd percentile for age and gender) and/or functional deficits demonstrated as global cognitive delays with performance below 3rd percentile on standardized testing or 3 or more domains of neurodevelopmental functioning more than 2 SD below the normed mean on standardized measures of cognitive, executive, memory, adaptive, motor, attentional, social skills, or sensory functioning. Data collected through evaluation of the children were entered into the 4-digit diagnostic code 13 to reflect the magnitude of expression of the 4 key diagnostic features of FAS: (1) growth deficiency; (2) FAS facial phenotype, (3) central nervous system dysfunction; and (4) documented gestational alcohol exposure. These codes were then converted to diagnostic categories. 194 Neurodevelopmental Functioning: FAS, pfas, and ARND Journal of Developmental & Behavioral Pediatrics

Children who met all physical criteria for growth impairment and facial dysmorphology, and neurodevelopmental deficits were assigned a diagnosis of FAS. Children who demonstrated facial dysmorphology and neurodevelopmental deficits but had normal growth (height and weight) patterns were placed in the pfas category, and children who met criteria for neurodevelopmental deficits but had normal growth and did not meet criteria for facial dysmorphology were classified as alcohol-related neurodevelopmental disorder. 3 5,12,13 All 78 children met requirements for 1 of these 3 diagnostic categories. Data Analytic Approach Before data analysis, the extent of missing data and normality of data were examined. A relatively small amount of data was missing, and there were no systematic patterns associated with why the data were missing. Examination of histograms and skewness and kurtosis statistics on all variables revealed that none of the distributions exceeded the cutoff points of 2* the standard error of skewness or 2* the standard error of kurtosis. 24 Thus, all data were included in the analysis. To examine for significant differences between the 3 groups on the descriptive variables, 2 analyses were completed on nominal level variables and one-way ANOVAs, followed by post hoc contrasts were used for interval and ratio level variables. Multivariate analyses of variance (MANOVA) were conducted to examine the differences between the 3 diagnostic groups on measures of intellectual, executive, memory, academic, adaptive, and behavioral functioning. In choosing among the 4 multivariate test statistics available for MANOVA, Wilks lambda ( ), Pillai s trace, Hotelling s trace, and Roy s largest root, the assumptions of the MANOVA were considered, and the statistical approach was chosen based on the characteristics of the sample. 25 Descriptive discriminant analyses 26 were performed in follow-up to the MANOVA, labeling the significant underlying constructs based on the level of loading of the indices on the linear discriminant functions. The equality of the 3 group covariance matrices on each discriminant analysis was examined using Box s Test of Equality of Covariance Matrices. 27 This assumption was met in all cases by nonsignificant F tests. As none of the MANOVA assumptions (normality and homogeneity of variance) were violated, Wilk s lambda was used as the MANOVA test statistic unless otherwise noted. All statistical computation was performed using SPSS 13.0: the General Linear Model Multivariate program for the MANOVAs and the Discriminant Analysis program for the descriptive discriminant analyses. RESULTS Descriptive Data By using the diagnostic criteria presented in the Methods section, 21 children met criteria for fetal alcohol syndrome (FAS), 10 children had partial FAS (pfas), and 47 met criteria for alcohol-related neurodevelopmental disorder (ARND). The children in the 3 diagnostic categories were similar for gender, racial/ethnic distribution, country of birth, adoption status, child s age, age of the primary caregiver, and rates of polydrug exposure to tobacco, cocaine, marijuana, and opiates (Table 1). Child welfare histories were similar in that there were no differences across the 3 groups of children as to number of placements or age at adoption. The 3 groups demonstrated similar academic background and experiences with no differences in rates of participation in special education services, suspension and expulsion from school, or having been held back at least 1 grade (Table 1). As would be expected based on the criteria by which the children were placed in the 3 groups, there were significant differences in current mean growth parameters (Table 1), with the children with FAS having significantly decreased mean weight, height, and head circumference measurements and significantly higher percentage of children with a head circumference below 3rd percentile. There was no difference in growth patterns between the children with pfas versus children with ARND. Neurodevelopmental Functioning General Intelligence Although the Wechsler Intelligence Scale for Childrenthird edition provides a Performance, Verbal, and Full Scale IQ, we chose to use the 4 index scores that emerged from factor analysis during the development of the instrument 14 (Verbal Comprehension, Perceptual Organization, Freedom from Distractibility, and Processing Speed) rather than the global IQ scores because each subtest of the Wechsler Intelligence Scale for Children-third edition is represented only once on the 4 indices. We used Pillai s criterion as our multivariate statistic because we obtained statistical significance on Box s Test of Equality of Covariance Matrices in the means on these 4 indices across the 3 groups (Table 2). The multivariate analyses of variance (MANOVA) for the 4 index scores was significant (F (8,138) 3.62, p.001). The observed power of this analysis was 0.981. On follow-up descriptive discriminant analyses (DDA), 2 linear discriminant functions were generated by the DDA, but only 1 reached statistical significance (Table 3). Because of the heavy loading of all 4 indices on the first linear discriminant function, we labeled the construct general intelligence. To understand how the 3 groups differed from one another on this construct, we conducted a multivariate contrast (Fig. 1), which revealed that the FAS group scores were significantly lower than the scores for the pfas group and the ARND group (F (4,68) 3.18, p.019; F (4,68) 6.60, p.001, respectively), which did not differ significantly from one another (F (4,68) 1.16, p.34). Memory The overall MANOVA for the Wide Range Assessment of Memory and Learning Screening (Table 2) was significant (F (8,146) 2.38, p.019). The observed power of this analysis was 0.880. On follow-up DDA, 2 linear discrimi- Vol. 31, No. 3, April 2010 2010 Lippincott Williams & Wilkins 195

Table 1. Descriptive Characteristics and Current Growth Parameters Variable FAS (N 21) pfas (N 10) ARND (N 47) F/ 2 p Gender female, n (%) 8 (38.1) 4 (40.0) 13 (27.7) 1.06.59 Race/ethnicity, n (%) 2.42.96 Caucasian 9 (42.9) 4 (40.0) 16 (34.0) African American 7 (33.3) 5 (50.0) 21 (44.7) Hispanic 1 (4.8) 0 (0) 2 (4.3) Native American 0 (0) 0 (0) 1 (2.1) Mixed race 4 (19.0) 1 (10.0) 7 (14.9) Birthplace, n (%) 1.88.39 Overseas 6 (28.6) 1 (10.0) 8 (17.0) United States 15 (71.4) 9 (90.0) 39 (83.0) Adoption status, n (%) 3.76.15 Adopted 19 (90.5) 10 (100.0) 36 (78.3) Not adopted 2 (9.5) 0 (0) 10 (21.7) Polydrug exposure pattern, n (%) Amphetamines 0 (0) 0 (0) 1 (2.1) 0.98.61 Cocaine 11 (52.4) 4 (40.0) 21 (44.7) 0.65.72 Heroin 1 (4.8) 1 (10.0) 6 (12.8) 1.01.60 Marijuana 5 (23.8) 0 (0) 5 (10.6) 1.46.48 Tobacco 6 (28.6) 3 (30.0) 15 (31.9) 0.45.80 Receiving special education services, n (%) 3 (14.3) 2 (20.0) 8 (19.0) 7.75.46 Held back at least one grade, n (%) 7 (33.3) 0 (0) 7 (14.9) 5.54.06 Age (yr), mean (SD) 9.66 (2.24) 8.99 (0.98) 9.57 (1.65) 0.49.62 Primary caregiver age (yr), mean (SD) 46.7 (4.65) 46.1 (4.68) 47.5 (8.14) 0.23.79 No. prior placements, mean (SD) 1.7 (1.4) 1.2 (1.3) 2.0 (1.7) 1.04.36 Age (mo) at adoption, mean (SD) 53.9 (30.2) 30.3 (18.8) 35.5 (27.9) 3.17.05 Current growth parameters Weight (g), mean (SD) 22,602 (5,640.5) 27,114 (6,240.8) 26,927 (5,827.7) 4.29.02 Height (cm), mean (SD) 120.3 (9.7) 127.8 (10.9) 127.6 (9.5) 4.32.02 Head circumference, mean (SD) 49.9 (1.6) 51.6 (2.2) 51.6 (1.7) 7.46.00 Head circumference 3rd percentile, n (%) 16 (80) 2 (20.0) 17 (36.1) 10.63.005 FAS, fetal alcohol syndrome; pfas, partial FAS; ARND, alcohol-related neurodevelopmental disorder. nant functions were generated, only the first of which reached statistical significance (Table 3). This construct was labeled language-based memory because the Story Memory and Verbal Learning subtests were weighted so heavily on the construct. A multivariate contrast on the 3 groups (Fig. 2) demonstrated that although the pfas group had the lowest scores on the language-based memory construct, the only differences that reached statistical significance were the differences between the FAS group and the ARND group (F (4,72) 2.61, p.042). The pfas and the ARND groups did not differ significantly from one another (F (4,72) 2.37, p.06), nor did the FAS and the pfas groups (F (4,72) 1.80, p.14). Executive Functioning Executive functioning as measured on the Children s Color Trails Test differed across the groups, with the MANOVA on the Children s Color Trails Test reaching statistical significance (F (4,140) 3.28, p.013; Table 2). The observed power of this analysis was.826. The follow-up DDA generated 2 linear discriminant functions, but only the first reached statistical significance (Table 3). This function loaded heavily on the second color trail, so, we named the factor sequencing and shifting. Multivariate contrasts demonstrated that the FAS group s times were significantly longer on this sequencing and shifting task than the pfas and ARND groups times (F (2,69) 3.42, p.038; F (2,69) 6.35, p.003), which did not significantly differ from each other (F (2,69) 0.04, p.96; Fig. 3). Evaluation with the Behavior Rating Inventory of Executive Function and the Wisconsin Card Sorting Test did not demonstrate any statistically significant differences between the 3 groups. The observed power associated with the Behavior Rating Inventory of Executive Function was 0.847, and the 196 Neurodevelopmental Functioning: FAS, pfas, and ARND Journal of Developmental & Behavioral Pediatrics

Table 2. Variables Means of the Dependent Measures FAS, Mean (SD) pfas, Mean (SD) ARND, Mean (SD) F p WISC-III 3.62.001 VCI 80.2 (15.2) 93.3 (7.0) 98.8 (15.0) POI 78.5 (15.2) 92.7 (13.4) 96.3 (16.8) FDI 80.9 (17.4) 97.7 (12.0) 93.9 (14.4) PSI 84.0 (17.3) 104.4 (16.6) 97.6 (17.0) WRAML-S 2.38.019 Picture memory 7.7 (2.8) 10.3 (3.2) 8.7 (2.6) Design memory 6.9 (3.4) 9.0 (4.0) 8.1 (3.5) Verbal learning 7.2 (2.8) 7.1 (3.2) 9.3 (3.6) Story memory 6.1 (3.3) 6.7 (3.5) 8.5 (2.9) CCTT a 3.28.013 Trail 1 1.2 (1.8) 0.3 (1.1) 0.5 (1.6) Trail 2 1.0 (1.2) 0.1 (1.0) 0.02 (0.9) WRAT-3 1.93.08 Reading 85.5 (18.6) 94.4 (18.1) 97.0 (15.4) Spelling 87.2 (16.9) 91.6 (19.2) 97.8 (15.1) Arithmetic 77.2 (15.1) 86.8 (12.5) 90.7 (17.5) Vineland 2.85.012 Communication 75.3 (18.4) 99.1 (20.6) 89.0 (17.9) Daily living skills 74.1 (15.8) 80.0 (22.3) 85.2 (20.7) Social 76.6 (16.6) 75.0 (15.0) 82.4 (15.8) Child behavior checklist 1.41.14 Anxious/depressed 58.4 (8.8) 62.3 (9.0) 63.7 (7.7) Withdrawn/depressed 57.1 (7.8) 61.8 (6.9) 62.9 (8.9) Somatic complaints 57.0 (6.5) 59.6 (8.8) 60.8 (7.6) Social problems 61.4 (8.7) 66.2 (8.6) 63.8 (7.8) Thought problems 66.5 (8.3) 64.2 (8.4) 66.6 (8.5) Attention problems 70.7 (11.8) 71.5 (9.4) 69.6 (10.0) Rule-breaking behavior 60.7 (8.4) 68.3 (9.3) 64.8 (8.7) Aggressive behavior 62.0 (9.9) 71.8 (12.5) 68.6 (10.2) a Results are expressed in z-scores calculated using clinical norms. FAS, fetal alcohol syndrome; pfas, partial FAS; ARND, alcohol-related neurodevelopmental disorder; WISC-III, Wechsler Intelligence Scale for Children-third edition; VCI, Verbal Comprehension Index; POI, Perceptual Organization Index; FDI, Freedom from Distractibility Index; PSI, Processing Speed Index; WRAML-S, Wide Range Assessment of Memory and Learning Screening; CCTT, Children s Color Trails Test; WRAT-3, Wide-3, Range Achievement Test third edition. power associated with the Wisconsin Card Sorting Test-64 was 0.283. Academic Functioning Evaluation of academic functioning through MANOVA testing of the Wide Range Achievement Test third edition data revealed no statistically significant differences across the 3 groups (F (6,148) 1.93, p.08). The observed power of this analysis was 0.697. Adaptive Functioning There was a difference in adaptive functioning (Table 2) across the 3 groups as demonstrated on MANOVA of the Vineland (F (6,130) 2.85, p.012). The observed power of this analysis was 0.876. On follow-up analysis with DDA, the first linear discriminant function attained statistical significance (Table 3). The first linear discriminant function loaded most heavily on the Communication Domain with positive, but notably lower loadings on the remaining 2 subscales. As a result, we labeled this construct Functional Communication. The F tests for the multivariate contrasts (Fig. 4) demonstrated that the FAS group had a significantly lower mean Functional Communication score than the pfas group (F (3,64) 4.48, p.006), which had the highest mean score. The ARND group, which had a mean score in between the other 2 groups was not significantly different from either the FAS (F (3,64) 2.48, p.069) or the pfas group (F (3,64) 2.54, p.064). Behavior A MANOVA on the problem scales of the Child Behavior Checklist revealed that there were no statistically Vol. 31, No. 3, April 2010 2010 Lippincott Williams & Wilkins 197

Table 3. Structure Coefficients for Dependent Measures Indices LDF p WISC-III 0.67.001 VCI 0.90 FDI 0.78 POI 0.65 PSI 0.61 WRAML-S 0.78.021 Story memory 0.88 Verbal learning 0.79 Design memory 0.17 Picture memory 0.04 CCTT 0.83.013 Trail 2 1.00 Trail 1 0.48 Vineland 0.78.012 Communication 0.900 Daily living skills 0.280 Social 0.005 LDF, linear discriminant function; WISC-III, Wechsler Intelligence Scale for Children-third edition; VCI, Verbal Comprehension Index; POI, Perceptual Organization Index; PSI, Processing Speed Index; WRAML-S, Wide Range Assessment of Memory and Learning Screening; CCTT, Children s Color Trails Test; significant differences between the 3 groups (F (16,134) 1.41, p.14). The observed power of this analysis was 0.831. The 2 analyses of differences across the 3 groups with respect to the percentages of children who scored in the clinical range on each of the domain scale scores revealed that none of these differences reached statistical significance. DISCUSSION The purpose of this study was to compare neurodevelopmental functioning among children with fetal alcohol syndrome (FAS), partial FAS (pfas), and alcoholrelated neurodevelopmental disorder (ARND). Seventyeight foster and adopted children with confirmed prenatal exposure to alcohol were distributed across 3 diagnostic categories based on the presence of physical Figure 2. Group means on language-based memory. and central nervous system (CNS) characteristics that are fundamental to the diagnoses of FAS, pfas, and ARND. 3,5,12,13 Clinical criteria for placement in diagnostic groups were based on stringent guidelines, including documentation of all 3 criteria for facial dysmorphology and past or current growth impairment (height or weight) below 3rd percentile. In addition, to meet criteria for abnormal CNS functioning, children had to have a head circumference below 3rd percentile, evidence of global cognitive functioning below 3rd percentile, and/or 3 domains of neurodevelopmental functioning greater than 2 SDs below the normed mean, to ensure that children were not misclassified. The resulting 3 groups were similar in demographic and child welfare background information. The children who met tightly defined physical criteria for a diagnosis of FAS were significantly different from alcohol-exposed children who did not meet all criteria, whereas children with pfas and ARND were similar in all neurodevelopmental domains that were tested. Children with FAS exhibited the most impaired level of general intelligence as documented on the Wechsler Intelligence Scale for Children-third edition, significantly worse language-based memory when compared with children with ARND and significantly poorer functional Figure 1. Three group means on general intelligence. Figure 3. Group means on sequencing and shifting. 198 Neurodevelopmental Functioning: FAS, pfas, and ARND Journal of Developmental & Behavioral Pediatrics

Figure 4. Group means on functional communication. communication skills than children with pfas. Although one might question if the differences in the cognitive measures are simply the reflection of the categorization process, the classification scheme for diagnosis of FAS, pfas, and ARND was driven by physical criteria of growth impairment and facial dysmorphology; neurodevelopmental functioning was not used as a classification element as all groups demonstrated CNS impairments. Neurodevelopmental functioning and global cognitive functioning were significantly worse in those children with physical manifestations of prenatal alcohol exposure. The Children s Color Trails Test differentiated degree of executive functioning difficulties among the 3 groups of alcohol-exposed children, with the FAS group of children performing significantly worse on sequencing and shift tasks than either the pfas or ARND groups. On the other hand, all 3 groups performed similarly on the Wisconsin Card Sorting Test-64 and the Behavior Rating Inventory of Executive Function. Executive functioning refers to multiple complex cognitive skills that relate to higher order thinking, including behavioral inhibition, working memory skills, planning, and shifting between different activities and mental states. 28 These skills derive from integration of brain functioning across all levels, including higher cortical frontal systems, subcortical regions, and their neuronal connections. Exposure to alcohol at any point in gestation could affect the system most vulnerable at that specific time in fetal development, disrupting communication at any point in the interconnective circuitry. Thus, it is not surprising that executive functioning difficulties are among the most common problems found in children across the fetal alcohol spectrum. 8,9 In analyzing these results, one must consider that there is covariance in the cognitive outcomes; i.e., a lower IQ could account for poorer memory and executive functioning. Reanalysis of the memory and executive functioning findings with the Wechsler Intelligence Scale for Children-third edition as a covariate revealed that the Freedom from Distractibility index shared notable covariance with the Wide Range Assessment of Memory and Learning Screening and rendered the Wide Range Assessment of Memory and Learning Screening findings nonsignificant. Similarly, the Processing Speed index shared notable covariance with the Children s Color Trails Test and rendered those findings nonsignificant. These results suggest that the Wide Range Assessment of Memory and Learning Screening and Children s Color Trails Test tap into attention and concentration, areas that the Freedom from Distractibility and Processing Speed indices access. In reporting this, however, we must point out that using IQ as a covariate when assessing neurodevelopmental disorders can be misleading. The IQ differences seen in our sample are an inherent group characteristic; removing the variability because of IQ will result in a situation that is not descriptive of the actual groups. 29 Consistent with findings with the Behavior Rating Inventory of Executive Function and the Wisconsin Card Sorting Test-64 that executive functioning difficulties occur in all 3 groups of children, there was no difference between any of the groups on behavior; in fact, there was a high rate of diagnosis of attention deficit hyperactivity disorder in each of the 3 groups: 60% of children with FAS, 88.9% of children with pfas, and 76.9% of children with ARND met DSM-IV criteria for a diagnosis of attention deficit hyperactivity disorder. The combination of neurodevelopmental deficits across multiple domains of functioning, as required in this study for diagnosis of an alcohol exposure-related disorder, can impede the child s ability to think ahead to self-direct behavior, to maintain and integrate multiple bits of information, to stay on task, to problem solve in an organized manner, or to place information into memory for later recall. These deficits have been noted in children across the fetal alcohol spectrum 8,9 and are the characteristics that most likely lead to the high rate of diagnosis of attention deficit hyperactivity disorder in children with prenatal alcohol exposure. 7,10,11 This also helps explain the finding that all 3 groups of children in this study had significant behavioral problems at school: 5% of the FAS group, 10% of the pfas group, and 11% of the ARND group had been suspended or expelled from school as compared with 1% of children in the Chicago Public Schools system. 30 Children in this study with FAS had a significantly reduced mean head circumference and an increased rate of head circumferences below 3rd percentile compared with the pfas and ARND children, who were similar in measures of head circumference. Since the midface forms early in gestation and growth is primarily a third trimester occurrence, it may be that a diagnosis of FAS implies exposure across the full gestational period, whereas pfas implies exposure only in the first trimester with sparing of brain growth in the third trimester. This is consistent with recent studies by Astley et al, 6 Vol. 31, No. 3, April 2010 2010 Lippincott Williams & Wilkins 199

which demonstrated such a relationship between trimester of exposure and mean volume of the frontal lobes: the mean volume of the frontal lobe was significantly smaller in children with 3 trimesters of exposure compared with those children with exposure through only the first trimester. Children with exposure in the first and second trimesters but no exposure in the third trimester had frontal lobe volumes comparable with nonexposed controls. Astley et al 4,7 documented a correlation between facial dysmorphology and brain dysfunction, validating the consistency between midline defects and underlying brain abnormalities. On the other hand, Mattson et al 8 found similar neurodevelopmental profiles for prenatally alcohol-exposed children with and without facial dysmorphology. In our study population of children in which facial phenotype was measured using Astley and Clarren s methodology, 4,13 growth patterns (height or weight below 3rd percentile for age and gender) provided the strongest differentiation of risk for neurodevelopmental functioning: the FAS group, who, by definition, had facial dysmorphology and growth impairment, performed significantly worse on most measures of neurocognitive functioning, including global intellectual functioning, than the pfas group, which had facial dysmorphology but were less growth restricted. In fact, neurodevelopmental functioning of children with pfas and ARND demonstrated a similar profile, and these 2 groups did not differ in any areas evaluated. The recent study by Astley et al 7 combined children with FAS and pfas into a single group. However, growth impairment as a marker of increased risk among children with prenatal alcohol exposure, as found in this study, is consistent with the early work by Jones and Smith 1 who noted in their original description of FAS that ongoing growth impairment is a key factor demonstrating the adverse effects of prenatal alcohol exposure. A limitation of this study is that the sample size is relatively small. This factor is a potential cause for a lack of significant findings for some areas of neurodevelopment, especially between the pfas and ARND groups. However, although an examination of the means in Table 2 reveals differences of one half to three quarters of a SD between the pfas and ARND groups on some subscales, the powers reported suggest that the lack of significant differences is not because of insufficient power. Power for the WRAT-III was moderate; if there were larger numbers of children, the results might demonstrate significant differences. The lack of findings for the Wisconsin Card Sorting Test-64 could be because of insufficient sample size or heterogeneity in the sample. Future studies with larger sample sizes may help clarify potential differences that appear not to be significant in this study. It also should be emphasized that this is a clinical sample, with its associated biases, and consists entirely of foster and adopted children. Thus, the conclusions cannot be generalized to the general population of children with prenatal alcohol exposure or children who are living with their biological parents. In addition, although we were able to document maternal alcohol use in all cases, we could not obtain information on dosage or drinking patterns during gestation. Such a lack of information regarding prenatal alcohol exposure is not an uncommon problem, especially in working with children in out-of-home placement. It also should be noted that all children in this study, to qualify for a diagnosis of FAS, pfas, or ARND, had to meet criteria for severe CNS dysfunction (Rank 3 on the 4-Digit Code 13 ). Children with a diagnosis of mild to moderate CNS dysfunction (Rank 2) were not included; thus, this study does not address the neurodevelopmental profile of children on the milder end of the fetal alcohol spectrum. The data from this study are consistent with previous studies in that children with prenatal alcohol exposure but without the full manifestations of growth impairment and facial dysmorphology have significant neurodevelopmental deficits. However, the children with the full manifestation of FAS demonstrated a global pattern of impairment and functioned significantly worse in most arenas than children with pfas or ARND. It appears that use of the most stringent criteria for the diagnosis of FAS, especially requiring growth deficits greater than 2 SD below the normed mean, serves to differentiate children with FAS from children with pfas and ARND. This clarification of the diagnosis of FAS is important, especially as ongoing work to identify specific neurodevelopmental profiles across the fetal alcohol spectrum continues. REFERENCES 1. Jones K, Smith DW. Recognition of the fetal alcohol syndrome in early infancy. Lancet. 1973;ii:999 1201. 2. Stratton K, Howe C, Battaglia F. Fetal Alcohol Syndrome: Diagnosis, Epidemiology, Prevention, and Treatment. Washington, DC: National Academy Press, Institute of Medicine; 1996. 3. Bertrand J, Floyd RL, Weber MK. Guidelines for identifying and referring persons with fetal alcohol syndrome. MMWR Recomm Rep. 2005;54:1 15. 4. Astley SJ, Clarren SK. Measuring the facial phenotype of individuals with prenatal alcohol exposure: correlations with brain dysfunction. Alcohol Alcohol. 2001;36:147 159. 5. Astley SJ. Comparison of the 4-digit diagnostic code and the Hoyme diagnostic guidelines for Fetal Alcohol Spectrum Disorders. Pediatrics. 2006;118:1532 1545. 6. Astley SJ, Aylward EH, Olson HC, et al. Magnetic resonance imaging outcomes from a comprehensive magnetic resonance study of children with fetal alcohol spectrum disorders. Alcohol Clin Exp Res. 2009;33:1671 1689. 7. Astley SJ, Olson HC, Kerns K, et al. Neuropsychological and behavioral outcomes from a comprehensive magnetic resonance study of children with fetal alcohol spectrum disorders. Can J Clin Pharmacol. 2009;16:e178 e201. 8. Mattson SN, Riley EP, Graming L, Delis DC, Jones KL. Neuropsychological comparison of alcohol-exposed children with or without physical features of fetal alcohol syndrome. Neuropsychology. 1998;12:146 153. 9. Connor PD, Sampson PD, Bookstein FL, Barr HM, Streissguth AP. Direct and indirect effects of prenatal alcohol damage on executive function. Dev Neuropsychol. 2000;18:331 354. 200 Neurodevelopmental Functioning: FAS, pfas, and ARND Journal of Developmental & Behavioral Pediatrics

10. Kodituwakku PW, May PA, Clericuzio CL, Weers D. Emotionrelated learning in individuals prenatally exposed to alcohol: an investigation of the relation between set shifting, extinction of responses, and behavior. Neuropsychologia. 2001;39:699 708. 11. Kodituwakku PW. Defining the behavioral phenotype in children with fetal alcohol spectrum disorders: a review. Neurosci Biobehav Rev. 2007;31:192 201. 12. National Task Force on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention. Fetal Alcohol Syndrome: Guidelines for Referral and Diagnosis. Washington, DC: US Department of Health and Human Services; 2004. 13. Astley SJ, Clarren SK. Diagnosing the full spectrum of fetal alcohol-exposed individuals: introducing the 4-digit diagnostic code. Alcohol Alcohol. 2000;35:400 412. 14. Wechsler D. WISC-III Manual. San Antonio: The Psychological Corporation; 1991. 15. Gioia G, Isquith P, Guy S, Kenworthy L. Behavior Rating Inventory of Executive Function. Odessa, FL: Psychological Assessment Resources, Inc.; 2000. 16. Vriezen ER, Pigott SE. The relationship between parental report on the BRIEF and performance-based measures of executive function in children with moderate to severe traumatic brain injury. Child Neuropsychol. 2002;8:296 303. 17. Elia LFD. Color Trails Test. Odessa, FL: Psychological Assessment Resources; 1996. 18. Heaton RK, Chelune GJ, Talley JL, Kay GG, Curtiss G. Wisconsin Card Sorting Test Manual: Revised and Expanded. Odessa, FL: Psychological Assessment Resources, Inc.; 1993. 19. Wilkinson G. Wide Range Achievement Test Revision 3. Wilmington, DE: Jastak Associates; 1993. 20. Sheslow D, Adams W. Wide Range Assessment of Memory and Learning. Odessa, FL: Psychological Assessment Resources, Inc.; 1990. 21. Sparrow S, Balla A, Cicchetti D. Vineland Adaptive Behavior Scales, Interview Edition, Survey Form Manual. Circle Pines, MN: American Guidance Service; 1984. 22. Achenbach TM. Manual for the Child Behavior Checklist and 1991 Profile. Burlington, VT: University of Vermont; 1991. 23. Aase JM. Clinical recognition of FAS: difficulties of detection and diagnosis. Alcohol Health Res World. 1994;18:5 9. 24. Tabachnick BG, Fidell LS. Using Multivariate Statistics. 3rd ed. New York, NY: Harper Collins; 1996. 25. Olson CL. Practical considerations in choosing a MANOVA test statistic: a rejoinder to Stevens. Psychol Bull. 1979;86:1350 1352. 26. Huberty CJ, Smith JD. The study of effects in MANOVA. Multivariate Behav Res. 1982;17:417 432. 27. Box GEP. A general distribution theory for a class of likelihood criteria. Biometrika. 1949;36:317 346. 28. Fryer SL, Tapert SF, Mattson SN, Paulus MP, Spadoni AD, Riley EP. Prenatal alcohol exposure affects frontal-striatal BOLD response during inhibitory control. Alcohol Clin Exp Res. 2007; 31:1415 1424. 29. Dennis M, Francis DJ, Cirino PT, Schachar R, Barnes MA, Fletcher JM. Why IQ is not a covariate in cognitive studies of neurodevelopmental disorders. J Int Neuropsychol Soc. 2009;15: 331 343. 30. The Board of Education of the City of Chicago. Chicago Public School Website. Available at: http://www.cps.edu/about_cps/ At-a-glance/Pages/Stats%20and%20facts.aspx. Updated October 2009. Accessed November 3, 2009. Vol. 31, No. 3, April 2010 2010 Lippincott Williams & Wilkins 201