A gen, and fetal alcohol exposure often leads to evidence

Transcription

1 /97/ $03.00/0 ALCOHOLISM: CLINICAL AND EXPERIMENTAL RESEARCH Vol. 21, No. 1 Fehruary 1997 A Comparison of Children Affected by Prenatal Alcohol Exposure and Attention Deficit, Hyperactivity Disorder Claire 0. Coles, Kathleen A. Platzman, Cheryl L. Raskind-Hood, Ronald T. Brown, Arthur Falek, and Iris E. Smith Behavioral deficits are often noted in children with fetal alcohol syndrome (FAS) and other individuals with prenatal alcohol exposure, including mental retardation, learning problems, social problems, and deficits in attention. Because attention deficit, hyperactivity disorder () has been diagnosed so frequently in children with FAS and other alcohol related birth defects, there has been speculation that alcohol is an etiological factor in. To examine the relationship between behavior characteristics of children with fetal alcohol exposure and those seen in children with a diagnosis of, 149 low socioeconomic status (SES), African-American children (mean age = 7.63 years) were given a battery of neuropsychological and behavioral tests. One hundred and twenty-two were a subsample from a longitudinal study of prenatal alcohol exposure, whereas twenty-seven were identified in an Clinic. Children were given two sets of tests: (1) traditional model of conventional behavioral and psychiatric measures of and externalizing behavior; and (2) measures of neurocognitive functioning reflecting a four-factor model of the neurological basis of the components of attention (Mirsky AF, in Integrated Theory and Practice in Clinical Neuropsychology, Hillsdale, NJ, Lawrence Erlbaum Associates, 1989). Results indicated that children with the physical characteristics associated with prenatal alcohol exposure and those with a diagnosis of had equivalent intellectual abilities with both clinical groups performing more poorly than contrast children from the same SES and ethnic groups. However, there were clear distinctions on behavioral and neurocognitive measures between the two clinical groups with those with performing more poorly on conventional tests sensitive to attentional problems and conduct disorder. When these two groups were compared on measures designed to measure the model of the four factors of attention by Mirsky, they were noted to have distinct patterns of deficits. These results suggested that the alcohol-affected children did not have the same neurocognitive and behavioral characteristics as children with a primary diagnosis of. Key Words: Fetal Alcohol Syndrome, Attention Deficit, Hyperactivity Disorder (), Maternal Alcoholism, Attention. ~~ ~~ From the Deparhnent of Psychiatry and Behavioral Services (C.D.C., U.P., C. L.R-H., R. T. B., A. F.), Emory University School of Medicine, and the Georgia Department of Children and Youth Services (I.E.S.), Atlanta, Georgia. Received for publication July 29, 1996; accepted October 10, 1996 This research was fitnded by a National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health grant 5 ROI AAO and an Alcohol and Drug Abuse Section, Deparhnent of Human Resources, Georgia, grant. Previous draji presented at a Symposium entitled, Is Fetal Alcohol Syndrome a Unique Diagnostic Category? (J. Nanson, Ph.D., Chair) at the Research Society on Alcoholism annual meeting, June 19, 1994, Maui, Hawaii. Reprint requests: Claire D. Coles, Ph.D., Human and Behavior Genetics Laboratory, Georgia Mental Health Institute, 1256 Briarcliff Road, N.E., Atlan fa, Geop.a, Copyright by The Research Society on Alcoholism. 150 LCOHOL IS both a physical and a behavioral terato- A gen, and fetal alcohol exposure often leads to evidence of neurological damage in surviving offspring. The defining characteristics of fetal alcohol syndrome (FAS) are facial dysmorphia, growth retardation, and evidence of neurological damage. Behavioral differences in children with fetal alcohol syndrome (FAS) and what has been called fetal alcohol effects (FAE) have been noted since these disorders were first described by Jones and Smith in In follow-up studies of alcohol-affected and dysmorphic children who were identified in clinical settings, Streissguth and her colleag~es~*~ and Steinhausen, Nestler, and Spoh? have reported a number of negative consequences including cognitive deficits, hyperactivity, relative deficits in adaptive behavior, and psychiatric disorders. A review of these and other papers (e.g., Lemoine and Lemoine6), suggests that negative behavioral outcomes can be predicted by severity of physical effects, cognitive deficits, and instability in rearing environment. One of the characteristics often identified in children with prenatal alcohol exposure is attention deficit, hyperactivity disorder (). In fact, some clinicians regard this as the primary diagnostic sign in preschool and older children and stimulant drugs are often prescribed to children with FAS who show externalizing behaviors. In addition, a relationship between and familial alcoholism is often described retrospectively by alcoholics and in children of alcoholics, leading to suggestions of an association between this disorder and a genetic predisposition to alcoholism. When the clinical literature is examined, the results suggest support for the idea that children with alcohol-related disorders have deficits in attention. For instance, Nanson and Hiscock compared two clinically referred populations, including 29 children with FASFAE and 20 children with a diagnosis of, with 20 normal controls. Three experimental tasks that focused on different components of attention were used: behavior (measured by parent checklists); cognition [e.g., Wechsler Intelligence Scale for Children-Revised (WISC-R): Short Form]; and computerized attention-demanding tasks. Three attention-demanding tasks were examined: tasks demanding maintenance of attention (often called vigilance tasks); tasks demanding inhibition of impulsive responding; and tasks involving reinforcement seeking. This study reported that FASFAE children and those with were very similar Alcohol Clin Erp Res, VDI 21, No 1, 1997: pp

2 PRENATAL ALCOHOL EXPOSURE AND AlTENTION DEFICIT, HYPERACTIVITY DISORDER 151 on a number of the parameters tested, including parental ratings. In addition, the alcohol-affected children had more problems on the computerized tasks involving attention than did those with who were, however, more impulsive in their responses. However, there were significant differences between the two clinical groups with the FAS/FAE children having IQ scores in the borderline intellectual range (IQ, 70-85), whereas those with scored in the average range (IQ, ). There were also differences in the speed of responding with those with FAS showing decrements in performance relative to both children with and controls. Because of the behavioral characteristics and social histories of children who are diagnosed with FAS, it is difficult to be sure that externalizing ( acting out ) behavior observed in children who are referred to a clinical setting is truly neurologically based. The majority of referrals of school-aged children to psychiatric and developmental clinics require that be ruled out because this condition is currently the most popular diagnosis for this age group. In fact, this disorder has been so universally diagnosed, with so much attendant misdiagnosis, that concerned clinicians and researchers are beginning to reconsider the basis for the diagnosis. 2 One of the reasons for this confusion is that a number of other conditions are associated both with attentional problems and with behaviors that mimic those seen in children with attentional problems. At the present time, it appears that a number of other conditions can be misdiagnosed as. These other conditions include some kinds of attachment disorder^'^ resulting from the effects of neglect and abuse, anxiety and depression, and expressive language disorders. Given the environments that may be provided by alcoholic parents,16 it is easy to see why there is some potential for confusion when children come into a clinical setting. In prospective studies of alcohol-exposed children whose mothers were social drinkers, some authors have identified statistically significant, although mild, effects on computerized tests of attenti~n ~ ~ and, in clinical studies, deficits in memory and executive functioning. Other researchers, however, have not found the same kinds of outcomes. In a middle class Canadian sample, Fried et a1.20 found that those middle class children exposed to small amounts of alcohol were less likely to show impulsive responding on a CPT task when compared with unexposed children. In contrast, in a low income sample in Cleveland, Boyd et al21 found no relationship between attention and prenatal alcohol exposure. Brown et a1.22 evaluated a subsample of preschool children from an Atlanta cohort, some of whom were heavily exposed, and found that, using a computerized test of attention, there was an increase in errors of omission (omitting response to the appropriate stimulus) in children who had been exposed throughout pregnancy in contrast to those whose mothers had (a) stopped drinking, and (b) those whose mothers had never drunk. However, there was no evidence of impulsivity or hyperactivity in this group of children despite evaluation with a number of other standardized and experimental tests. Observational measures were used as well without identifying increased levels of hyperactivity. Given these findings, the question of the effect of prenatal alcohol exposure on computerized measures of attention and impulsivity seems unresolved. although several studies have found that there is less impulsive behavior in alcohol-exposed children who were not identified in a clinical setting. Definition of Attention Another potentially confusing issue in understanding the relationship among alcohol, attentional problems, and behavior is the nature of the attentional process(es) itself. A great many different behaviors and functions are labeled attention both in everyday language and in the psychological and psychiatric literature. Because, in general usage, there is no well defined meaning to this term in a biological or psychological sense, is usually diagnosed using both clinical judgment and various observational measures and checklists used to describe the child s behavior. Some of the measures conventionally used are the Achenbach Child Behavior Checklist (CBCL),23 the SNAP,24 and the Connors Checklist. s In addition, assumptions about attention are made based on outcomes of educational and neurocognitive tests like measures of intelligence and memory. The difficulty in applying these measures both clinically and for experimental purposes is that there are so many different causes for the same kinds of observable behaviors. The checklists that are usually used simply record adult opinions of children s behaviors, which are then assumed to be related to underlying attentional processes. The ability tests are affected by many different conditions and no research basis has been established for interpreting subtests to reflect underlying attentional deficiencies.26 One model that has been proposed to describe the attentional process and to provide physically and theoretically logical basis for the understanding of deficits in attention is that described by Mirsky and his colleagues In this model, four factors of attention are posited: Focus, involving perceptual/ motor efficiency, which is believed to be supported by the temporal and parietal regions of the brain; Sustain, involving vigilance, which is believed to be supported by the reticular formation, corpus striatum, and medial thalamus; Encode, involving sequential memory and learning, which is usually thought of as a hippocampal function; and Shift, which is related to flexibility and executive functioning, and is the function of the prefrontal cortex. This model is empirically based, derived from studies of the neurocognitive performance of individuals who have lesions in various brain regions as well as individuals with specific clinical problems (e.g., schizophrenia, ). In the development of this model, neuropsychological tests have been used to measure these four aspects of attention. The advantage of examining attention in this way is that it provides a method to discriminate different functions un-

3 ~ ~~ ~ 152 COLES ET AL. Variable Neonatal % Male Birthweight (9) Head circumference (cm) Length (cm) Dysmorphia score Table I. Children s Physical Characteristics: Birth and Follow-Up ETOH ni- Controls FAS/FAE Dysmorphic (n = 35) (n = 25) (n = 62) (n = 27) (448.29) (1.12) (2.38) 2.51 (2.89) (716.32) (2.27) (3.92) (9.26) (545.69) (1.70) (3.22) 2.23 (2.47) 81 S O (513.22) Not available Not available Not available Significance x2 = p < F = p < 0,000 All > 2 F = p < > 2 F = 11.9op < &3 > 2 F = p < > la3 Follow-Up Age (Yr) Weight (lb) Head circumference (cm) Height (in) Dysmorphia score 7.73 (0.54) (13.95) (1.69) (2.45) 2.09 (2.08) 7.46 (0.41) (12.94) (2.31) (2.23) 7.24 (4.64) 7.64 (0.49) (13.98) (1.55) (2.59) 1.32 (126) 7.68 (0.54) (15.12) (1.49) (2.52) 1.58 (1.71) F = 5.91 p < All > 2 F= 11.1Op<0.000^ All > 2 F = 9.74 p < All > 2 F = p < > All Controlling for gender. derlying the child s behavior and performance and, therefore, may help in refining treatment and educational interventions. If children with a diagnosis of FAS or evidence of other alcohol effects are more impacted in specific brain areas, certain processes should be differentially impacted and certain neuropsychological measures performed less well. If these functions can be identified, deficits characteristic of alcohol exposure may be specified. In the current study, to measure the effects of prenatal alcohol exposure on attention and neurocognitive functioning, comparisons were made among several groups of children using these two methods of measuring attentional functioning: (1) traditional model of conventional observation, checklists, and cognitive tasks commonly used in the diagnosis of ; and (2) tasks associated with the four factors of attention suggested by Mirsky et al It was hypothesized that, if alcohol-affected children have, they should perform similarly on both kinds of tests to those children who had received a clinical diagnosis of. Furthermore, it was hypothesized that children who were exposed to alcohol but who did not have physical effects (i.e., dysmorphia and growth retardation) might show similar but attenuated effects similar to those of the more severely affected children, whereas the performance of nonexposed children would not show similar defects. METHODS Subjects One hundred forty-nine low-income children and their mothers/caregivers were included in the current study. All children were selected from the same inner city teaching hospital, either prenatally (n = 122), as part of a longitudinal study of the effects of prenatal exposure, or when they were diagnosed as having (n = 27) in the hospital s child psychiatry clinic or in a county Community Mental Health Center. To control for the known effects of low socioeconomic status (SES) and/or race, all children were drawn from the same low-income, African-American population. Children s demographic characteristics at birth and at follow-up are shown in Table 1. Based on their scores on an empirically based dysmorphia checklist that was given at birth and at follow-up,29 children with prenatal alcohol exposure were classified as either Dysmorphic (FASI FAE) or Nondysmorphic. Children whose mothers did not use alcohol in pregnancy were classified as Controls. The fourth group () was selected from the Department of Psychiatry s Attention Deficit Disorder Clinic. All were clinically referred and had received a diagnostic work-up. All met the American Psychiatric Association s Diagnostic and Statistical Manual s (DSM-III-R)3 criteria for based on a structured interview, the Diagnostic Interview Schedule for Children3 (DISC), as well as a psychiatric evaluation. All children in this group were taking medication for, the majority, methylphenidate or other stimulants. Mothers or caregivers of all children were interviewed to obtain demographic information and information about prior and current drug and alcohol use. Women who used drugs other than alcohol, cigarettes, and/or marijuana in pregnancy were not included in the study. The exception was occasionai cocaine (not crack) use by a few of the mothers first identified in pregnancy between 1980 and Mothers of children with a diagnosis of who reported alcohol use during pregnancy were not included in the study. Table 2 shows the demographic characteristics and substance use of these women, as well as results of blood and urine screens, which were done at follow-up. Procedure There were two methods of subject selection. Children in the longitudinal cohort who were years old at the time of testing were identified and their mothers were recontacted, if possible. The study was explained and those families who agreed to participate and signed a consent form approved by the University Human Investigations Committee were brought to the Human Genetics Laboratory for the assessment. Children with a diagnosis of were identified by a co-investigator in the course of his work at the Clinic. Several others were identified by staff at a county Community Mental Health Center. When an appropri-

4 PRENATAL ALCOHOL EXPOSURE AND ATENTION DEFICIT, HYPERACTIVITY DISORDER 153 Table 2. Caregiver Demographics at Follow-up Controls FAS/FAE Dysmorphic Variable (n = 35) (n = 25) (n = 62) (n = 27) Significance Characteristics Maternal age (yr) Education (yr) % Married % Black Monthly income ($) Dependents (n) Cognitive WAlS Full Scale IQ WAlS Verbal IQ WAlS Performance 10 Drug Use AAoz./week during pregnancy Current AAoz./week %Cocaine during pregnancy Current % cocaine (5.49) (1.44) (768.87) 2.88 (1.75) (9.1 1) (8.25) (1 1.I 4) (3.57) (10.41) 11.oo (1.72) (794.41) 2.25 (1.59) (6.67) (7.05) (8.23) (1 1.52) (1 1.88) 6.70 % Marijuana during pregnancy Current % marijuana GGTP (10.87) (205.66) (84.57) (11.32) (7.26) (154) (423.94) 2.81 (1.38) (9.92) (8.57) (1 1.89) 6.67 (6.48) 7.41 (12.15) (6.50) (1.74) (404.45) 2.08 (0.96) (11.45) (13.56) (10.14) (0.45) F = 6.29 p < >2&34>3 x2 = p < > All x2 = p < 0.04 All > 4 F = 3.29 p < > 2&3 F = 5.55 p < > 2&3 F = p < >all 3 > 1&4 F = 7.88 p < > 184 x2 = 11.29~ < &3 > 1&4 x2 = 21.I5 p < &3 > l&4 x2 = 21.15~ < > 1&4 F = 2.99 p < > 1&4 ately aged child was identified who had the equivalent socioeconomic status (SES) characteristics, the study was explained to the parent and participation requested. If both mother and child met the study requirements, the study proceeded in the same way as for the other groups. Each mothedchild pair was seen during a single day. Testing took about 5 hr with a break for lunch. All testing was done by master s level psychometricists or psychologists in a fashion blind to the mother s drinking history and the child s diagnostic category. Children who were receiving Ritalin or other stimulant medications were not given these drugs on the day of the test. Children who were receiving other medications (e.g., antidepressants) continued to receive these medications. Medication regimen was monitored by a psychiatrist associated with the study. Measures Conventional Measures of Neurocognitive Functioning and Attention. The measures used in the conventional assessment of included the following: Kaufman-Assessment Battery for Children3 (K-ABC), a measure of intellectual and scholastic aptitude, which includes global summary measures as well as subtests making up the cognitive and academic scales; subtests from the Wechsler Intelligence Scale for Children- Revised (WISC-R), including Coding and VocabulaIy, the Developmental Test of VisuaVMotor Integration (VMI),S3 a sensitive measure of perceptual/motor performance; two parent checklists designed to identify children with ; the SNAPz4 and the AchenbachZ3 Child Behavior Checklist (both parent and teacher versions); and the DISC,31 a structured interview used to establish psychiatric diagnosis in childhood. In addition, an automated Continuous Performance Task (CPT), lasting about 90 min, was administered to each child, although some were not able to complete this measure. This computer-driven application consisted of a 40-trial baseline Simple Reaction Time (SRT) measure, a 1800 trial Vigilance measure, and a 1800 trial CPT measure. Real-time reaction times and performance accuracy were recorded and summarized automatically by the application. Each subject was provided three practice trials to meet criterion (set at 0.80) before the actual experimental portion of each task. If criterion was not met, the child was pardoned from this phase of the testing. Four Factors ofartention. In addition to the measures described, several tests were added to allow the Four-Factor Model of Attention assessment of Mirsky et al.27.2x For some factors, due to the child s age, tests with similar neuropsychological properties have been substituted for those used in the original studies. The tests used in this study include the following: Focus: WISC-R Coding (equivalent to the Digit Symbol); Encode: K-ABC Number Recall (equivalent to Digit Span), K-ABC Arithmetic, and Kinsborne s3s Paired Associate Memory Task, Encoding (List 1 and List 2) as well as Delayed Recall of each list; Sustain: SRT, Vigilance Tasks, and CPT Tasks; and Shift: Wisconsin Card Sorting test, 6 Number of Categories Completed, Perseverative Responses, and Errors. RESULTS Demographic and outcome variables from the four groups of children and their mothers/caretakers were compared using ANOVA techniques, including post hoc procedures, as well as x2 analysis procedures for ordinal and categorical variables. Not all children were able to complete every task. Reasons for subject loss were examined and analyzed for systematic biases. When appropriate,

5 154 COLES ET AL. Table 3. K-ABC Standardized Composite Scores at Follow-up Controls FAS/FAE Dysmorphic Variable (n = 35) (fl = 25) (n = 62) (fl = 27) Significance Mental processing Composite Sequential processing Composite Simultaneous processing Composite Nonverbal processing Composite Achievement composite (10.49) (13.18) (9.54) (9.67) (9.63) (11.34) (11.09) (12.89) (11.74) (12.91) (12.79) (14.07) (12.16) (1 1.59) (1 1.47) (11.85) (13.66) (12.65) (11.72) (8.92) F = 5.32 p < la3 > 2a4 F = 3.68 p < 0.01 la3 B 4 F = 4.73 p < ia3>2i >4 F = 6.28 p < ia3 B 284 F = 3.27 p < > 2 these biases are discussed below. When only a few cases were omitted, mean substitution was used. Maternal Characteristics Maternalharegiver characteristics are shown in Table 2. Note that the mothers, like the control mothers, reported that they did not use alcohol during pregnancy and, at follow-up, these two groups reported significantly less current alcohol consumption than both groups of women who reported alcohol use in pregnancy. Their selfreport is confirmed by the results of the y-glutamyl traspepsidate (GGTP) blood test, which measures liver function and is a sensitive measure of the current and cumulative effects of alcohol use? Among the mothers of children who showed alcohol-related dysmorphia, GGTP levels were greatly elevated (M = units) when compared with the other groups (see Table 2); however, even the mothers of the nondysmorphic, alcohol-exposed children had much higher GGTP levels (M = units) than either of the abstemious groups whose levels were both in the range typically seen for adult women (that is, 3 to 33 units). Although demographic factors (age, parity, and income) were not significantly different among these groups of women, overall IQ scores on the Wechsler Adult Intelligence Scale-Revised3* did show differences with mothers of the alcohol-affected children being different from the other groups. As shown in Table 2, Verbal IQ was more discrepant than Performance IQ. Children s Physical Characteristics Analysis of the children s physical characteristics at the neonatal period and at follow-up revealed a number of significant group differences (see Table 1). On average, the children in the FASFAE group were significantly lighter in weight, had smaller head circumferences, and were shorter in length when compared with the other groups of children at birth and at follow-up. Dysmorphia scores were also significantly higher among the FASFAE children at the neonatal period (M = 10.73) and at follow-up (M = 7.24) compared with scores in the other groups, indicating a higher incidence of minor physical anomalies. Because the groups were selected on this variable, this outcome is to be expected. Children s Neurocognitive Characteristics Tables 3,4 and 5 contain the MANOVA results from the neurocognitive functioning measures. General Cognitive Characteristics. Analysis of the summary standard scores from the Kaufman Assessment Battery for Children (K-ABC)32 revealed that the children in the FASFAE and the groups had similar global intellectual deficits on all composites compared with the controls and the non-dysmorphic, alcohol-exposed children (see Table 3). Specific Cognitive Tests. On some of the specific subtests of the K-ABC, particularly those associated with sequential functioning (see Table 4), both clinical groups (FASFAE and ) showed deficits (i.e., Word Order, Spatial Memory, Photo Series). On other subtests, particularly those associated with visual/spatial reasoning, the alcoholaffected children s scores were lower than the other groups (i.e., Triangles, Matrix Analogies). On the Achievement Battery, the alcohol-affected dysmorphic children were significantly poorer on Arithmetic, whereas the group was significantly poorer on ReadingDecoding; note that the scores of the FASFAE group were similarly depressed on this subtest but did not reach the conventional level of significance. No significant differences among the groups of children were revealed in verbal skills or vocabulary, but the patterns of deficits differed between the FASFAE children and the groups when they were presented with measures of memory and visualhpatial skills (see Table 5). The group, as would be predicted, had difficulty with WISC-R Coding (Digit Symbol) and so did both groups of alcohol-exposed children (see Table 5). The FASFAE children showed problems with encoding of the second list on the Paired Associate Task and the number of categories completed on the Wisconsin Card Sorting Test36 (see Table 5).

6 PRENATAL ALCOHOL EXPOSURE AND ATTENTION DEFICIT, HYPERACTIVITY DISORDER 155 Variable Mental subscales Hand Movements Gestalt Closure Number Recall Triangles Word Order Matrix Analogies Spatial Memory Photo Series Achievement subscales Faces & Places Arithmetic Riddles Reading/Decoding ReadingNnderstanding VMI Standard Score Table 4. K-ABC Standardized Subtest Scores and VMI Scores at Follow-up Controls FAS/FAE Dysmorphic (n = 35) (n = 25) (n = 62) (n = 27) Significance 8.56 (2.23) 9.06 (2.70) 8.29 (2.75) 8.71 (2.43) 8.79 (2.54) 9.65 (1.86) 9.29 (2.39) 9.00 (1.79) (10.04) (11.06) (11.33) (12.55) (10.96) 7.56 (1.87) 8.56 (3.19) 7.28 (2.97) 7.08 (2.48) 7.20 (2.24) 8.16 (2.21) 6.96 (2.68) 6.64 (2.63) (14.46) (14.32) (10.36) (15.69) (7.14) 8.21 (2.59) 8.79 (3.27) 8.65 (2.99) 7.77 (2.27) 8.34 (2.70) 9.40 (2.19) 8.45 (2.54) 8.58 (2.15) (1 1.26) (1 1.59) (10.96) (13.59) (14.13) 7.41 (2.01) 9.07 (3.41) 7.00 (4.01) 7.59 (2.32) 6.70 (2.52) 8.56 (2.14) 7.44 (2.68) 7.26 (2.49) (20.42) (10.46) 1&3 > (10.79) (11.51) 1> (8.51) F = 2.55 p < >2 F = 4.59 p < > All 3&1 > 4 F = 3.45 p < 0.02 l&3 > 2 F = 5.08 p < &3>21 >4 F = 7.74 p < &3 > 2&4 F = 4.10~ < F = 3.28 p < F = 4.41 p < (10.31) (13.90) (11.37) (10.95) 1 > All Behavioral Functioning On those measures designed to assess the behaviors associated with, children with that diagnosis were significantly higher than all the other groups on the CBCL- T,23 the SNAPz4 (particularly the barred items ), and on the DISC3 interview items (see Table 6). In contrast, those who were alcohol exposed and affected did not differ from the control group with the exception of the Attention T- score on the CBCL for the FAS group. With the exception of some of the scores of those in the group, none of the T-scores on the CBCL are in the clinical range (see Table 6). Computerized Attentional Measures Because many of the clinically diagnosed children found this task difficult to complete, we analyzed the subject loss for systematic bias. Of the children with a diagnosis of, 60% dropped out of the CPT task, as did 52% of FASFAE group, 43% of the controls, and 25% of the non-dysmorphic, alcohol-exposed group (x2 = 10.97, p < 0.01). Thus, there was differential subject loss by group with those in the clinical groups being less likely to complete the task. Overall Vigilance and CPT task performance was analyzed using univariate ANOVA procedures as well as 3(Group) X 4(Block) MANOVAs. Findings from these data revealed that both speed and accuracy were particularly impaired for the children compared with the other groups of children (see Tables 7 and 8). On the Vigilance task, the group had significantly slower reaction times and more false alarm responding, associated with impulsivity. On the CPT, the children in the group had fewer correct responses (hits) and more response omissions (misses) than the other groups on either task. The scores of children with FASFAE, who were able to complete the task, did not differ significantly from those of controls or the nondysmorphic alcohol-exposed children. The children with FASFAE who were able to complete the task generally had better scores on these measures than the other groups. Discriminant Functions To assess how well these groups of children could be discriminated from each other based on tasks commonly used to diagnosis, a discriminate function of the traditional model was performed. In addition to the Arithmetic and Number Recall subtests from the K-ABC?2 this model used measures commonly used in clinical diagnosis of this disorder, including the average number of symptoms of conduct disorder and from the DISC3 interview, the aggression, attention problems and delinquency T-scores from the Achenbach CBCL-T,23 and the barred

7 156 COLES ET AL. WISC-R Subscales Vocabulary Coding Table 5. Measures of Neurocognitive Functioning at Follow-up Variable = = = = Controls (n 35) FAS/FAE (n 25) Dysmorphic (n 62) (n 27) Sianificance Kinsborne Paired Associates List 1 (Trials to Criterion) List 2 (Trials to Criterion) Trial 1 (Delayed Recall) Trial 2 (Delayed Recall) Wisconsin Card Sort' # Categories Completed # correct # Errors # Perseverative Responses # Perseverative Errors # Non-Perseverative Errors 9.29 (2.75) (3.55) 5.28 (3.04) 5.84 (2.49) 6.97 (1.45) 6.89 (1.36) n = (1.33) (15.61) (15.61) 51.OO (32.97) (23.26) (16.53) * Note changes in sample size (2.89) 8.29 (3.31) 6.13 (3.53) 8.26 (3.11) 6.77 (1.15) 6.09 (1.71) n = (1.12) (18.29) (18.29) (38.94) (27.05) (19.72) 9.27 (3.03) 9.00 (2.99) 5.58 (3.18) 6.89 (2.87) 6.83 (152) 6.53 (1.87) n = (1.74) (17.19) (20.01) (32.86) (23.55) (3.06) 8.29 (2.66) 7.20 (2.87) 4.78 (2.46) 6.87 (3.13) 6.78 (0.99) 6.69 (1.19) n = (1.64) (15.91) (17.86) (25.90) (18.29) (12.26) F = 5.21 p < >All F = 3.44 p < > 1&3 F = 2.86 p < >2 SNAP Total score Barred items CBCL-T Externalizing 7" Variable Attention Problems Delinquency Aggressive Internalizing '7" DISC-T Scores' Mean # Conduct Disorder Symptoms Mean # Symptoms Table 6. Measures of -Related Behaviors at Follow-up Controls FAS/FAE Dysmorphic (n = 35) (n = 25) (n = 62) (n = 27) Significance (7.74) 1.24 (1.47) (1 0.52) (7.21) (8.27) (8.44) (10.99) n = (0.70) 3.61 (4.1 8) Note changes in sample size (8.89) 3.33 (3.31) (11.32) (9.96) (8.03) (9.98) (6.15) n = (0.74) 4.95 (3.72) (8.29) 2.16 (2.46) (12.79) (7.82) (8.47) (12.41) (8.49) n = (0.76) 4.58 (4.03) (6.42) 9.38 (3.39) (6.96) (8.74) ( (7.39) (8.39) n = (1.06) 8.50 (2.55) F = p < All > 1 F = p < All > 1 F = 7.16 p < F = p < >All 2 > 1 F = 6.80 p < > 1&3 3 > 1 F = 5.43 p < F = 3.72 p < 0.01 F = 9.42 p < items from the SNAP24 inventory, which are considered those that are most diagnostic of. If the behavioral characteristics of alcohol-affected children are similar to those with, relatively poor discrimination of these groups should result. For this discriminant function, controls and nondysmorphic groups were combined into a "nonclinical" group, which was compared with the alcohol dysmorphic group and the group. Overall, 71% of the subjects were correctly classified with the highest discrimination being shown by the group (85%). In contrast, only 44% of ETOH dysmorphic subjects were correctly classified with these traditional measures of and only 12% were misclassified as (see Table 9).

8 PRENATAL ALCOHOL EXPOSURE AND A ITENTION DEFICIT, HYPERACTIVITY DISORDER 157 # Hits Table 7. Vigilance Task Performance at Follow-up Variable Controls (n = 26) FAWFAE (n = 15) Dysmorphic (n = 53) (n = 17) # False alarms Mean reaction time hits (msec) Mean reaction time false alarms (msec) Standard deviation reaction time hits Standard deviation reaction time false alarms (41.86) (32.87) (41.42) (86.79) (23.06) (55.27) (54.36) (25.38) (70.06) (72.30) (36.76) (61.47) (4127) (36.79) (65.54) (80.55) (23.57) (62.28) (38.27) (37.47) (44.64) (53.83) (30.00) (42.78) Significance F = 4.56 p < F = 9.74 p < Table 8. CPT Performance at Follow-up Variable Controls (n = 26) FAS/FAE (n = 15) Dysmorphic (n = 53) (n = 17) Sianificance # Hits # False alarms Mean reaction time hits (msec) Mean reaction time false alarms (rnsec) Standard deviation reaction time hits Standard deviation reaction time false alarms (64.99) (58.92) (60.79) (67.00) (52.04) (68.30) (83.47) (66.43) (74.73) (70.82) (55.44) (99.47) (70.69) (73.33) (43.19) (44.90) (114.51) (93.38) (55.12) (52.23) ~~ ~ F = 2.92 p < >4 (34.71) (49.10) (41.22) (28.67) Table 9. Discriminant Function Using a Traditional Model Cases FAS/FAE CTLS + ETOH ND FAS/FAE (44) 3 (12) 11 (44) 27 3 (11) 23 (85) 1 (4) CTLS + ETOH ND 97 22(23) 3(3) 72 (74) Variables included the following: Aggressive, Attention Problems, and Delinquency (CBCL-T), Barred Items (SNAP), Arithmetic and Number Recall (K-ABC). Mean Number of Conduct Disorder Symptoms & Symptoms (DISC-T). Cases correctly classified, 71 %. Percentages are in parentheses. Patterns of Outcome On Experimental Model of Attention Figure 1 displays the performance of all groups on attentional processing tasks based on the Four-Factor Attentional Model of Mirsky et al To allow a graphic comparison across these different tests, results of each test have been converted to z-scores to compare performance among all tasks and groups simultaneously. In Fig. 1, differences in performance patterns between the FASFAE and the groups can be observed. For example, the FAS/ FAE group performed significantly less well on the Encoding dimension [K-ABC Arithmetic, Kinsborne s Paired Associates Task List 2 (which is affected by Proactive Inhibition)] and Shift (Wisconsin Card Sorting Test) variables, although not differing from the control group on the Sustain dimension. In contrast, the children encountered more problems with Focus (Digit Symbol) and Sustain (Vigilance and CPT) dimensions (see Fig. 1). To evaluate the extent to which children affected by alcohol resembled those with on these factors, additional discriminant functions were done for each of these four factors using these two clinical groups and the control and non-dysmorphic groups combined into a nonclinical group. The factors were considered separately both for clarity and because the number of subjects varied slightly for each set of tests. The results of these analyses are shown in Fig. 2. The Focus dimension was represented by the Coding subtest of the WISC-R. When this task was used, 59% of the cases were classified appropriately. However, it was the (56%) and the nonclinical (69%) groups that were more likely to be correctly classified while the alcohol affected group was not (19%) and was most likely to be classified as like the nonclinical group (43%). The Encoding dimension was represented by the Kin% borne paired associate particularly by learning on List 2, a task involving both learning and resistance to

9 158 COLES ET AL WlSC Dig11 Symbol PA Trial 2 WCST WCST 1 Cats # Persev Responses VIG VIG # False Alarms Mean RT False Alarms c s o (? -0.2 N -0.4 n 6 Fig. 1. Attention tasks by four groups. ENCODING SHIFT SUSTAIN Task 0 CONTROLS FASIFAE 0 ETOH NON-DYShlORF HIC FASFAE Focus: WlSC Coding Shift: WCST Y Categorlea. L Peraeveratlve Responses l FASFAE 50 %..., ,.,,,.,,,,,.,,,.,, CTLS+ETOH ND P f 0% 25% 50% 75 % 100% % Predicted Group Membership 9% cases correctly classified overall CTLScETOH ND 0% 25% 50% 75% T( % Predicted Group Membership $% cases correctly classified overall Encoding: Paired Associates -Trial 2 Sustain: Vigilance I Fdas Alarms FASFAE 59% I FASFAE 46% 89% CTLS+ETOH ND 0% 25% 50% 75% 100% % Predicted Group Membership 1% cases correctly classified overall CTLS+ETOH ND I 0% 25% 50% 75% 100% % Predicted Group Membership I 49% cases correctly classified overall Fig. 2. Comparison of discrminant functions using a four-factor model. proactive inhibition. On this task, 50% of the children are correctly classified, with 59% of both the alcohol group and the nonclinical group, but only 9% of the group. The rest of the group were distributed evenly between the other two groups. The Shift dimension was represented by the number of categories identified correctly and the number of perseverative responses on the Wisconsin Card Sorting Test.36 Using these outcomes, 53% of the cases were classified correctly with 50% of the alcohol-affected children, 67% of the controls, but only 20% of the group. When the children were incorrectly classified, they were usually grouped with the nonclinical cases. The Sustain dimension was represented by the number of False Alarms (impulsive responding) on the Vigilance procedure with 49% of cases correctly classified. However, once again, there was a difference in the effectiveness of the discrimination for the different groups, with 89% of the

10 PRENATAL ALCOHOL EXPOSURE AND ATTENTION DEFICIT, HYPERACTIVITY DISORDER 159 group being correctly classified, 46% of the alcohol affected group and only 40% of the nonclinical group. Only 7.7% of the alcohol affected group were misclassified as. DISCUSSION Children with FAS are often diagnosed with and prescribed stimulant medication as a result of their behavior and because in clinical practice there is a strong presumption that is a defining characteristic of the disorder. Because there are many alternative hypotheses that might account for observed behavior, the current study investigated the extent to which children with documented prenatal alcohol exposure and the physical features associated with FAS and fetal alcohol effects show the same neurocognitive and behavioral characteristics as children, from the same SES and ethnic group, who have a diagnosis of. The hypothesis that children with FAS have was tested in two ways. First, by comparing behavior on the cognitive and behavioral measures typically used in clinical settings to make this diagnosis and second, by comparing some members of these groups using the factors of attention posited by Mirsky and his colleague^.^^^^^ The results of both of these comparisons suggest that these two clinical groups have unique attentional profiles. In addition, these results call into question the assumption that behaviors seen in children with FAS result from the same neurocognitive deficits as those seen in individuals with. In this study, alcohol-affected children do show some differences in several neurocognitive dimensions and behavior on many of the variables tested when they are compared with nonexposed children and exposed children who do not have dysmorphia and growth retardation; however, when dysmorphic children are compared with children diagnosed with, even though both groups are equally impaired intellectually, there is little similarity in their pattern of responses. In fact, discriminant functions separate the two groups reliably. To generalize, children with are best identified by behavior checklists and measures of the ability to focus and sustain attention, whereas those with FAS appear to have deficits in visual/ spatial skills, encoding of information, and flexibility in problem solving. Why Are These Data Inconsistent with Previous Reports? The assumption that prenatal alcohol exposure is associated with is based on case study reports, descriptions of patients identified in clinical settings, and results of prospective studies of children who, although exposed to alcohol prenatally, do not have FAS. The current study involves a different sample, one that includes young children with this diagnosis who were not identified through referral to medical and psychiatric settings. The children in the longitudinal sample were living with mothers or relatives or had been adopted early in life. For the most part, they had not experienced highly disorganized environments or multiple custody arrangements. Although many are in special education programs, their behavior has not been sufficiently difficult to lead to psychiatric referrals. At early school-age, scores on measures of adaptive functioning3 were not significantly different from those of peers. Therefore, the group of children in this study may differ from others reported in the clinical literature. This explanation would imply either that the behaviors attributed to in children with FAS result from environmental effects (e.g., attachment disorders, anxiety, posttraumatic stress disorder, socialization deficits associated with neglect), or that there is some difference in degree of FAS with only the more severe FAS cases exhibiting externalizing behaviors associated with. Both of these possibilities should be explored further. There are other possible explanations of apparent differences in outcomes between this study and others in the literature. It may be that there are mild effects on attention in exposed children, without physical features, which become evident when they are compared with those without alcohol exposure. In neuropsychology, it is well known that memory and attention are the cognitive processes most sensitive to any kind of stressors with similar decrements noticed across a number of conditions. In the more seriously affected dysmorphic group who may have frank brain damage, other patterns of effects may emerge. Finally, when previous research is examined carefully, it may be that the outcomes are not as discrepant as these data at first appear. Most other prospective studies that have examined attention in alcohol-exposed children have not found the impulsivity that is generally characteristic of, although children have been found to have deficits in their performance on many of the tests used. In most of the clinical studies that have been reported, cognitive status and environmental factors have not been controlled so that some of the observed effects might be attributable to confounding factors. In Nanson and Hiscock s study, behavioral measures (i.e., checklists) were similar between the group, who had average IQ scores, and the FAS group, who were in the borderline IQ range. However, these two groups of children showed significant differences in performance on measures of attention, with those in the group being more similar to those in the control group. There are clear limitations in making interpretations from the current study. First, this is a single study and the effects must be replicated to insure that they are not characteristic only of this particular sample. Second, clinically referred children with a diagnosis of FAS should be examined in a similar way to evaluate the extent to which these patterns of attentional factors can be observed in a different situation. In addition, children with carefully diagnosed whose mothers drank during pregnancy and afterward should be examined as well.

11 I60 COLES ET AL. To assure that the behavioral differences seen in this sample are really associated with FAS and not with some other attributes of this sample, comparisons should be made also with individuals affected by other teratogens (e.g., maternal phenylketonuria), other children with cognitive deficits associated with mental retardation and specific learning disabilities, and those who have a history of neglect and abuse. What Are the Implications of the Current Study? From the data collected in the current study, it is possible to suggest that there may be specific brain regions affected by prenatal alcohol exposure, that these can be measured through certain neurocognitive tests, and that these regions are distinct from those that are affected in children with a diagnosis of. For instance, these data, indicating that there are difficulties in the encoding and retrieval of information, support findings in earlier animal research4" that reported that the hippocampus was more vulnerable to alcohol exposure. Other research with human^'^.'^ also supports the suggestion that short-term memory and executive functioning are affected. Similarly, results suggest that there are specific deficits in visual/spatial skills, in mathematical skills and in executive functioning (e.g., Wisconsin Card Sort) in alcohol-exposed children. If further study supports these specific findings, there may be implications for the education and treatment of alcohol-affected children. For instance, many children with FAS are given stimulant medication. However, this treatment does not seem to have prevented the negative consequences often noted in adolescence.' Examination of the data from this study suggests that children with a diagnosis of have specific problems with focusing and maintaining attention on stimuli. The -diagnosed children in this study were responsive to Ritalin, which may have helped them to maintain attention appropriately. It appears that, when their attention is focused, such children perform relatively well on the encoding dimension-that is, on short-term memory and learning. In contrast, the children with FAS/FAEi were able to focus and maintain their attention very well, but were not able to encode the information they attended to or to use new information meaningfully in problem solving. This pattern of deficits may have implication for medication regimen and for educational interventions with these children. For instance, stimulant medications, while helping to focus attention, may not improve learning or problem solving. Similarly, educational interventions might be modified to provide children with strategies for encoding and problems solving. Finally, it will be interesting to examine the relationship between specific patterns of cognitive dysfunction and the social outcomes, which are reported in older children with FAS. Among individuals with PKU who are reported to have specific deficits in tasks affected by the prefrontal ~ortex,~" social competence is reported to be impaired. It is possible that similar patterns of dysfunction may be found among individuals with FAS if confounding environmental and emotional factors can be controlled. In summary, the results of this study, while preliminary, suggest that it may be possible to examine the cognitive and behavioral functioning of alcohol-affected children and to compare their abilities with typical children and with those with other clinically defined disorders. By doing so, it may be possible to identify specific effects of alcohol exposure and to provide more effective means for treatment and remediation of identified deficits. ACKNOWLEDGMENTS The authors thank the following people for their participation in this research: Mina Dulcan, M.D., for consultation about medication effects in children with, Paul Fernhoff, M.D., for continuing medical support, as well as the staff of the Georgia Addiction and Pregnancy Project at the time this study was done (Krystal Ammonds, M.P.H., Steve Erickson, M.A., Kathy Sloan, R.N., Jeff Silverstein, M.Ed., Helen Mann, Ramelle Edwards, and Beth Jones). Finally, we wish to acknowledge with gratitude the mothers and children who have contributed their time and effort to the continuation of this project. REFERENCES 1. Institute of Medicine, Committee to Study Fetal Alcohol Syndrome: Fetal Alcohol Syndrome: Diagnosis, Epidemiology, Prevention, and Treatment. Washington, D.C., National Academy Press, Jones KL, Smith DW: Recognition of the fetal alcohol syndrome in early infancy. Lancet , Streissguth AP, Clarren SK, Jones KL: Natural history of the fetal alcohol syndrome: A 10-year follow-up of eleven patients. Lancet 10:85-92, Streissguth AP, Aase JM, Clarren SK, Randels SP, LaDue RA, Smith DF: Fetal alcohol syndrome in adolescents and adults. JAMA 265: , Steinhausen HC, Nestler V, Spohr HL Development and psychopathology of children with the fetal alcohol syndrome. Dev Behav Pediatr 3:49-54, Lemoine P, Lemoine Ph: Avenir des enfants de meres alcooliques (etude de 105 das retrouves a Page adulte) et qelques constatations d'interet prophalactique. Ann Pediatr (Paris) 39: , 1992 I. Astley S: Diagnostic Criteria for FAS. Paper presented at Centers for Disease Control and Prevention, Data Collaborative Meeting, 1994, Atlanta, GA, March, Schuckit MA, Sweeney S, Huey L Hyperactivity and the risk for alcoholism. J Clin Psychiatry 48: , Nanson JL, Hiscock M: Attention deficits in children exposed to alcohol prenatally. Alcohol Clin Exp Res 14: , Wechsler D: Wechsler Intelligence Scale for Children-Revised. New York, Psychological Corporation, Rapport MD: Attention-deficit hyperactivity disorder, in Hersen MH, Ammerman RT (eds): Advanced Abnormal Child Psychology. Hillsdale, NJ, Earlbaum, 1995, pp Shaywitz SE, Fletcher JM, Shaywitz BA: Issues in the definition and classification of attention deficit disorder. Topics Language Dis 14:l-25, Zeanah CH, Mammen OK, Lieberman AF: Disorders of attachment, in Zeanah CH (ed): Handbook of Infant Mental Health. New YorWndon, Guilford Press, 1994, pp Jensen JB, Burke M, Garfinkle BD: Depression and symptoms of attention deficit disorder with hyperactivity. J Am Acad Child Adolesc Psychiatry 27: , 1988

12 PRENATAL ALCOHOL EXPOSURE AND ATENTION DEFICIT, HYPERACTIVITY DISORDER Baker L, Cantwell DP: A prospective follow-up of children with speech/language disorders. J Am Acad Child Adolesc Psychiatry , Sher KJ: Psychological characteristics of children of alcoholics: Overview of research methods and findings, in Recent Developments in Alcoholism: Children of Alcoholics, New York, Plenum Press, 1991, pp Streissguth AP, Barr HM, Sampson PD, Parrish-Johnson JC, Kirchner GL, Martin DC: Attention, distraction, reaction time, at age 7-years and prenatal alcohol exposure. Neurobehav Toxic01 Teratol8: , Streissguth AP, Sampson PD, Olson HC, Bookstein FL, Barr HM, Scott M, Feldman J, Mrisky AF Maternal drinking during pregnancy: Attention and short-term memory in 14-year old offspring: A longitudinal prospective study. Alcohol Clin Exp Res 18: Kodituwakku PW, Handmaker, Cutler EK, Weathersby EK, Handmaker SD: Specific impairments in self-regulation in children exposed to alcohol prenatally. Alcohol Clin Exp Res 19: , Fried PA, Watkinson B, Gray R: A follow-up study of attentional behavior in 6-year old children exposed prenatally to marijuana, cigarettes, and alcohol. Neurotoxicol Teratol 14: , Boyd TA, Ernhart CB, Greene TH, Sokol RJ, Martier S: Prenatal alcohol exposure and sustained attention in the preschool period. Neurotoxicol Teratol 13:49-55, Brown RT, Coles CD, Smith IE, Platzman KA, Silverstein J, Erickson S, Falek A Effects of prenatal alcohol exposure at school age. 11. Attention and behavior. Neurotoxicol Teratol 13: , Achenbach T: Manual for the Child Behavior Checklist: 4-18 and 1991 Profile. Burlington, VT, University of Vermont, Department of Psychiatry, Swanson JM, Nolan W, Pelham WE: The SNAP Rating Scale. Resources in Education, Connors CK Rating scales for use in drug studies with children. Pharmacotherapy with children (special issue). Psychopharmacol Bull: 24-28, Massman PJ, Nussbaum NL, Bigter ED: The mediating effect of age on the relationship between child behavior checklist scores and psychological test performance. J Abnorm Child Psychol 16:89-95, Mirsky AF: The neuropsychology of attention: Elements of a com- plex behavior, in Peregman E (ed): Integrated Theory and Practice in Clinical Neuropsychology. Hillsdale, NJ, Lawrence Erlbaum Associates, Mirsky AF, Anthony BJ, Duncan CC, Ahern MB, Kellam SG: Analysis of the elements of attention: A neuropsychological approach. Neuropsychol Rev 2:75-88, Coles CD, Smith IE, Fernhoff PM, Platzman KA, Raskind-Hood C, Brown RT, Falek A Consistency of the effects of prenatal alcohol exposure: Prediction from the prenatal period. Submitted for publication, American Psychiatric Association Diagnostic and Statistical Manual-Revised, ed 3. Washington, DC, APA, Costello AJ, Edelbrok CS, Katas R, Kenter MK, Kbric JA: National Institute of Mental Health Diagnostic Interview Schedule for Children. Bethesda, MD, National Institute of Mental Health, Kaufman AS, Kaufman NL Kaufman assessment battery for children. Circle Pines, MI, American Guidance Services, Beery KE, Buktenica NA Developmental Test of VisuaVMotor Integration. Cleveland, Modern Curriculum Press, Rosvold HE, Mirsky AF, Sarason I, Bransome ED, Beck LH: A continuous performance test of brain damage. J Consult Psychol 20: , Conte R, Kinsbourne M, Swanson J, Zirk H, Samuels M: Presentation rate effects on paired associate learning by attention deficit disordered children. Child Dev 57: , Grant DA, Berg EA: The Wisconsin Card Sorting Test. San Antonio, TX, Psychology Corporation, Wallach J: Interpretation of Diagnostic Tests: A Synopsis of Laboratory Medicine, ed 4. BostowToronto, Little Brown, Wechsler D: Wechsler Adult Intelligence Scale-Revised. San Antonio, TX, Psychological Corporation, Coles CD, Brown RT, Smith IE, Platzman KA, Erickson S, Falek A Effects of prenatal alcohol exposure at school age. I: Physical and cognitive development. Neurotoxicol Teratol 13: , West JR, Hamre Kh4: Effects of alcohol exposure during different periods of development: Changes in hippocampal mossy fibers. Dev Brain Res 17: , Welsh MC, Pennington BF, Ozonoff S, Rouse B, McCabe ERB: Neuropsychology of early-treated phenylketonuria: Specific executive function deficits. Child Dev 61: , 1990