PEDIATRICS Vol. 106 No. 4 October 2000, pp. 782-791

Teacher-Assessed Behavior of Children Prenatally Exposed to Cocaine

Virginia Delaney-Black, MD, MPH^*, Chandice Covington, RN, PhD, Thomas Templin, PhD, Joel Ager, PhD^§, Beth Nordstrom-Klee, MA^*, Susan Martier, PhD, Linda Leddick, PhD^¶, R. Harvey Czerwinski^¶, and Robert J. Sokol, MD

From the ^* Department of Pediatrics, School of Medicine, Wayne State University;  College of Nursing, Wayne State University; ^§ Center for Healthcare Effectiveness Research, School of Medicine, Wayne State University;  Department of Obstetrics and Gynecology, School of Medicine, Wayne State University; and ^¶ Detroit Public Schools, Detroit, Michigan.

	ABSTRACT

Top Abstract Methods Results Discussion Conclusion References

Objective.  Prenatal cocaine exposure has been associated with alterations in neonatal behavior and more recently a dose-response relationship has been identified. However, few data are available to address the long-term behavioral effects of prenatal exposures in humans. The specific aim of this report is to evaluate the school-age behavior of children prenatally exposed to cocaine.

Methods.  All black non-human immunodeficiency virus-positive participants in a larger pregnancy outcomes study who delivered singleton live born infants between September 1, 1989 and August 31, 1991 were eligible for study participation. Staff members of the larger study extensively screened study participants during pregnancy for cocaine, alcohol, cigarettes, and other illicit drugs. Prenatal drug exposure was defined by maternal history elicited by structured interviews with maternal and infant drug testing as clinically indicated. Cocaine exposure was considered positive if either history or laboratory results were positive. Six years later, 665 families were contacted; 94% agreed to participate. The child, primary caretaker (parent), and, when available, the biologic mothers were tested in our research facilities. Permission was elicited to obtain blinded teacher assessments of child behavior with the Achenbach Teacher's Report Form (TRF). Drug use since the child's birth was assessed by trained researchers using a structured interview.

Results.  Complete laboratory and teacher data were available for 499 parent-child dyads, with a final sample size for all analyses of 471 (201 cocaine-exposed) after the elimination of mentally retarded subjects. A comparison of relative Externalizing (Aggressive, Delinquent) to Internalizing (Anxious/Depressed, Withdrawn, Somatic Complaints) behaviors of the offspring was computed for the TRF by taking the difference between the 2 subscales to create an Externalizing-Internalizing Difference (T. M. Achenbach, personal communication, 1998). Univariate comparisons revealed that boys were significantly more likely to score in the clinically significant range on total TRF, Externalizing-Internalizing, and Aggressive Behaviors than were girls. Children prenatally exposed to cocaine had higher Externalizing-Internalizing Differences compared with controls but did not have significantly higher scores on any of the other TRF variables. Additionally, boys prenatally exposed to cocaine were twice as likely as controls to have clinically significant scores for externalizing (25% vs 13%) and delinquent behavior (22% vs 11%). Gender, prenatal exposures (cocaine and alcohol), and postnatal risk factors (custody changes, current drug use in the home, child's report of violence exposure) were all related to problem behaviors. Even after controlling for gender, other prenatal substance exposures, and home environment variables, cocaine-exposed children had higher Externalizing-Internalizing Difference scores. Prenatal exposure to alcohol was associated with higher total score, increased attention problems, and more delinquent behaviors. Prenatal exposure to cigarettes was not significantly related to the total TRF score or any of the TRF subscales. Postnatal factors associated with problem behaviors included both changes in custody status and current drug use in the home. Change in custody status of the cocaine-exposed children, but not of the controls, was related to higher total scores on the TRF and more externalizing and aggressive behaviors. Current drug use in the home was associated with higher scores on the externalizing and aggressive subscales.

Conclusions.  Results of this study suggest gender-specific behavioral effects related to prenatal cocaine exposure. Prenatal alcohol exposure also had a significant impact on the TRF. Postnatal exposures, including current drug use in the home and the child's report of violence exposure, had an independent effect on teacher-assessed child behavioral problems. Furthermore, among the children prenatally exposed to cocaine, change in the child's custody status was a significant predictor of TRF scores. It remains possible that other unmeasured postnatal characteristics of the cocaine-using household may play important roles in teacher-assessed child behavior.  Key words:  cocaine, alcohol, child behavior, pregnancy, substance abuse.

Few controversies in perinatal medicine have received as much media attention as the outcome of children who have had prenatal exposure to cocaine.^1,2 In the early years of the last decade, the lay press warned that the children prenatally exposed to cocaine would face dire and insurmountable problems as they entered childhood.^1,2 Not surprisingly, these predictions have not come to fruition. Yet an equally dangerous and premature conclusion is to suggest that the cocaine epidemic has not had an adverse impact on children and families. Evidence to suggest that there is indeed a fetal cost has been widespread. Even after control for other variables, children born to cocaine-abusing women are smaller^7,8 and are delivered at an earlier gestational age.⁹ Furthermore, multiple studies from across the United States have also demonstrated an association between prenatal cocaine exposure and alterations of neonatal behavior.^10-14 Evidence of a dose-response relation between prenatal cocaine exposure and behavior is also available. Several centers, using a variety of techniques for assessing the relative amounts of prenatal cocaine exposure, all found evidence of cocaine's effects on neonatal behavior, even after controlling for other substances of abuse.^11-14

Data from animal studies over the last decade have also suggested strong evidence of behavioral effects related to prenatal exposure to cocaine in mice, rats, rabbits, and nonhuman primates,^15-22 including both maturational and stress-related differences. Furthermore, both animal^23-27 and human²⁸ studies have identified the effects of prenatal cocaine exposure on neurotransmitter systems. In both the human and rat models, these prenatal effects were associated with a significant behavioral alteration. In the human study²⁹ the behavioral effects, while directly related to the elevated concentration of circulating norepinephrine, were transient, whereas in the rat model, adult behavior was affected.³⁰

Although the negative effect of cocaine on neonatal behavior has been well documented, few human data have been available to evaluate the more long-term behavioral effects of prenatal exposure to cocaine. In their 1998 review of the human neurobehavioral studies, Lester et al³¹ identified fewer than 10 peer-reviewed publications with appropriate statistical control that reported child behavior after 24 months of age. Furthermore, what data do exist have often failed to present a clear and definitive answer to the question of whether prenatal cocaine exposure may be associated with childhood behavior problems. The purpose of this study was to evaluate the relationship between teacher-identified behavior problems of first-grade students and prenatal cocaine exposure after appropriate control for other prenatal exposures and postnatal risk factors.

	METHODS

Top Abstract Methods Results Discussion Conclusion References

The study design utilized was historical prospective. Beginning in 1986, women attending our urban university-based maternity clinic were routinely screened at their first prenatal visit for alcohol and drug use by trained research staff. Women were invited to participate in the ongoing prospective Fetal Alcohol Research Center Study based on a rectangular sampling distribution. Random, unselected samples would unlikely identify women who used only drugs and not alcohol because of the significant correlation between drug and alcohol abuse. To reduce colinearity between cocaine and alcohol, a block-sampling design was used, with oversampling of the heavily and moderately exposed pregnancies. This technique was used in an attempt to find sufficient numbers of women at different levels of drug and alcohol use. The Fetal Alcohol Research Center Study screened >2400 women annually during this period and enrolled >600 women each year.

Sample

Because >90% of women seeking prenatal care at this maternity clinic were black, there was inadequate representation of other ethnic/racial groups. Hence, study subjects were the singleton children born to these black women who had been extensively screened during pregnancy by research staff for tobacco, alcohol, cocaine, and other drug use. As clinically indicated, maternal and infant drug testing at delivery was also performed (20% of mothers and their infants). All nonhuman immunodeficiency virus-positive participants in the pregnancy outcome study who delivered singleton live born infants between September 1, 1989 and August 31, 1991 were considered eligible for study entry in the 6-year follow-up. Women with no prenatal care were excluded from study because of the recruitment criteria for prospective prenatal assessment of drug and alcohol exposure. At follow-up families were sought by telephone, mail, or home visit to the last known address. The client files of all 6 Detroit-based university-affiliated hospitals, the pediatric ambulatory service, and the major internal medicine provider for the university were searched for updated contact information. Telephone numbers and state driver's license numbers were also searched. Additionally children were sought through the private and public school system and through an advertisement in a community newspaper. The 665 families who could be contacted represent the potential study sample for this report.

A child participant was considered to be prenatally exposed to cocaine if any of the following were true: maternal history of cocaine use during pregnancy based on data from the prospective structured research interview during pregnancy, the prenatal or neonatal record, maternal urine, infant urine, or infant meconium. Because of institution limitations, meconium analyses were performed only at the end of the subject recruitment period (20% of subjects). An additional 13 families retrospectively admitted to cocaine use during the study pregnancy at the time of the 6-year follow-up interview. These children were also considered to be cocaine-exposed.

Measures

Pregnancy and neonatal data were available from the prospective research database as well as retrospective chart review. Information was prospectively collected during pregnancy on other substances of abuse including alcohol, heroin, marijuana, and cigarettes. At child age 6 years, after receiving informed consent, the child and parent (biologic mother, when available, or the primary caretaker) were tested in our research facility. Testing included the parent's self-reported psychopathology,³² parent-reported social support,³³ a modified Home Observation for Measurement of the Environment³⁴ assessment, family socioeconomic status,³⁵ child IQ,³⁶ and the child's self-report of exposure to violence.³⁷ Drug, alcohol, and cigarette use since the child's birth was assessed by trained researchers using a structured interview.

Permission to obtain blinded teacher assessments of child behavior using the Teacher Report Form (TRF)³⁸ was elicited. This widely used measure is self-administered and can be completed by the child's teacher in <15 minutes. A Total Problem Score and 8 syndrome scores can be derived. The syndrome scores are further grouped into Internalizing Behaviors (Withdrawn, Somatic Complaints, and Anxious/Depressed) and Externalizing Behaviors (Delinquent and Aggressive). Three syndrome scores (Social, Thought, and Attention Problems) fit neither group. This report focuses on Externalizing Behaviors (what Achenbach and Edelbrock³⁹ refer to as undercontrolled), Internalizing Behaviors (overcontrolled),³⁹ and Attention Problems. Clinically significant (or grouped) syndrome scores were also derived based on the cutoff scores recommended by Achenbach.³⁸ In addition, to observe the spectrum of undercontrolled behavior as recommended by the author, a difference score (Externalizing-Internalizing Difference Score) was computed by subtracting the T score for Internalizing Behaviors from the T score for Externalizing Behaviors. Development of the difference score was suggested during consultations with Dr Achenbach as an apriori opportunity to evaluate not only children with acting-out behavior, but specifically children who demonstrated both low anxiety and high externalizing scores (T. M. Achenbach, personal communication, 1998).

Procedure

Children and parents were evaluated after the date of the child's expected entry into first grade. Two female research assistants trained in the measures and blinded to the child's exposure status interviewed each child and mother independently. Another research assistant, also blinded to drug exposure status, collected teacher data. Mailings were sent to each teacher and, when forms were not returned by mail, an appointment was made to meet with the teacher. A compliance rate of 89% was obtained for teacher-generated data.

Statistical Analysis

A power analysis was performed using the Power Analysis Program by Borenstein and Cohen.⁴⁰ For the regression equations reported below, sufficient power existed (.8) to detect a variable that accounts for 1.5% of the unique variance in the outcomes. Data analysis was conducted using SPSS for Windows (SPSS, Chicago, IL). All variables were checked for possible violations of normality. Bivariate relationships were tested using 2-tailed t tests, ² analyses, and Pearson correlations. Stepwise multiple regression was used to examine relationships between exposure and behavioral outcomes while controlling for potential confounders.

	RESULTS

Top Abstract Methods Results Discussion Conclusion References

Sample

The primary research hypothesis to be tested was that prenatal cocaine exposure would be associated with an increase in problem behaviors reported by first-grade teachers after controlling for important covariates. Of the 665 child-parent dyads located, 94% (626) agreed to participate. However, 40 subjects (6%) missed multiple testing appointments, and these subjects as well as those with limited child testing (n = 28; 4%) were eliminated from further analysis. The remaining study sample consisted of 556 subjects. These 556 children generally did not differ significantly from the 109 who did not participate. The mothers of the participants were significantly older and had had more children than those who did not participate. However, the 2 groups of children did not differ significantly on any newborn characteristics, and the mothers did not differ on prenatal use of cigarettes, alcohol, or cocaine. Before data analysis, 29 children were identified as mentally retarded (performance IQ  65 or 2 standard deviations below the mean of this sample). Data for these children were excluded from this report because mental retardation may be associated with behavioral abnormalities. Additionally, although teacher compliance was 89%, no reports were available for 56 subjects. The remaining 471 subjects, 270 controls, and 201 children exposed to cocaine prenatally were the basis for this report.

As in previous studies, women who used cocaine during pregnancy had multiple risk factors (Table 1). Cocaine users were older and used more alcohol, cigarettes, and marijuana during the index pregnancy than did control mothers. Additionally, although the cocaine-exposed children were smaller at birth and had shorter gestational duration, exposed infants did not differ from controls on 1- or 5-minute Apgar scores. At the 6-year follow-up, exposed children were less likely to be in the care of their mother; however, the marital status of the primary caretaker at follow-up did not differ between groups and few mothers in either group were married. There was no difference between cocaine-exposed and control mothers on education status or income, with both groups having a mean education of less than a high school diploma and a mean yearly income of <$20 000.

Univariate Analyses

Comparison of teacher-assessed child behavior is provided in Fig 1. No differences were observed between control and cocaine-exposed children for the Total Score, Internalizing, Externalizing, or Attention Problems. However, children prenatally exposed to cocaine had a significantly higher mean Externalizing-Internalizing Difference Score than did the control subjects (P = .018).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   TRF scores by cocaine exposure status (boys and girls).

Preliminary analyses, as well as previous research, indicated different levels of problem behavior, as well as different patterns of relationships between behavior and other variables for boys and girls. In the current study, girls had significantly lower scores than boys on all TRF measures except Internalizing Behaviors and the Externalizing-Internalizing Difference Score. However, using criteria established by Achenbach,³⁸ the frequency of clinically significant scores for each of the syndrome scales and Internalizing and Externalizing Behaviors were computed for boys and girls in our sample (Fig 2). Boys were more likely to score in the clinically significant range on Total, Externalizing, Internalizing, and Aggression. A similar trend existed for Attention.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Percentage of clinically significant scores by gender.

No significant difference was found in the frequency of clinically significant scores by cocaine exposure status for girls. However, cocaine-exposed boys were more likely to have clinically significant scores for Externalizing and Delinquent behaviors. Comparison of the frequency of clinically significant scores by cocaine exposure status for boys is provided in Fig 3.

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Percentage of boys with clinically significant behavior (TRF) scores by cocaine exposure status.

Several postnatal risk factors were considered as potential control variables. One such factor, custody status, had changed significantly over the first 6 years of life, particularly among the cocaine-exposed children. Compared with the exposed group, control children were less likely to have been removed from the home at any time (10.5% vs 36.9%; P < .001). Among the cocaine-exposed children, 114 (59%) had always lived with their biologic mother, 16 (8%) had never lived with her, and the remaining children had experienced custody changes. In the latter group, 27 (14%) were currently living with the mother and 35 (18%) were not. There were no TRF differences by custody status for the control group. However, for the cocaine-exposed children, custody status was related to TRF scale scores (Fig 4). Mean scores for the cocaine-exposed children who had experienced a custody change and were currently living with their mother were significantly higher compared with controls for the Total Score, Externalizing, and Externalizing-Internalizing Difference.

View larger version (51K): [in this window] [in a new window]   Fig. 4.   TRF scores by cocaine exposure and custody status (boys and girls).

Regression Analyses

Potential confounding variables were identified using zero-order correlations. All variables that were related to either the predictor (cocaine exposure) or outcome measures (TRF scores) with a P value .2 were selected for inclusion in the stepwise regression models. Additionally, all regression analyses reported were controlled for gender. While previous studies have identified a relation between prenatal smoking and child behavior,⁴¹ prospectively collected pregnancy data from this study failed to confirm this relationship. Similarly, prenatal heroin and marijuana exposure did not enter the final regression model. Prenatal alcohol exposure, expressed as the log of the ounces of absolute alcohol consumed per day across pregnancy, was a significant predictor of the Total Score, Delinquent Behaviors, and Attention Problems. Postnatal environmental factors including the child's lead level at 6 years old, and exposure to both violence and current drug use had a significant role in predicting the Total Score, Externalizing Behaviors (both Aggressive and Delinquent syndromes), and Attention Problems. An effect of prenatal cocaine exposure was observed for the Externalizing-Internalizing Difference, after controlling for all other covariates (Table 2).

Custody status was further investigated by omitting from the regression 3 postnatal variables that could be influenced by parental care-taking. Exposure to family violence or current drug abuse may be related to child custody (furthermore, it is likely that current drug abuse may be underreported and, hence, subject to significant misclassification). Even whole blood lead levels may be influenced by parental inattention during the toddler years or residence in unsafe housing. In the series of simplified regressions that follow, prenatal risk factors, socioeconomic status, and child age and gender were included in the models predicting each of the TRF outcomes. Additionally, dummy variables were created to contrast the 4 custody status groups: always with the biologic mother, currently but not always with the biologic mother, previously but not currently with the biologic mother, and never with the biologic mother. In the stepwise regressions presented in Table 3, custody change (both sometimes with mother groups vs never a custody change) was a significant predictor of TRF Total Score, Externalizing, and Aggression. Children with at least 1 custody change had higher behavior scores. Prenatal cocaine exposure continued to predict the TRF Externalizing-Internalizing Difference, whereas prenatal alcohol exposure remained significantly related to both Delinquent Behaviors and Attention Problems. Gender was also a significant predictor for the Total Score, Externalizing, Aggressive, Delinquent, and Attention, with boys having significantly higher scores.

	DISCUSSION

Top Abstract Methods Results Discussion Conclusion References

In this prospective, blinded study of nearly 500 children at early school age, prenatal exposure to both alcohol and cocaine were associated with more teacher-reported problem behaviors. Specifically, cocaine-exposed children had a significantly greater difference between externalizing and internalizing behavior scores. These findings persisted after control for other prenatal exposures and postnatal risk factors. Although cocaine-exposed and control children did not differ significantly on either Externalizing or Internalizing Behavior scores alone or in Attention Problem scores, cocaine-exposed boys were significantly more likely to have scores in the clinically significant range for Total Externalizing Behaviors and Delinquent Behaviors. The effects of prenatal alcohol observed in this study confirm findings of previous investigators: exposure to increasing amounts of alcohol across pregnancy has a detrimental effect on both child attention and other problem behaviors.⁴¹ However, despite repeated reports of significant effects of prenatal cocaine exposure on neonatal behavior (see the recent review by Eyler and Behnke⁴²), previous investigators have failed to consistently identify more long-term effects of cocaine on child behavior. Tronick and Beeghly⁴³ suggest that several factors may account for these previously reported negative results. In addition to previous problems of study design, including small sample sizes and insensitive measures, many investigators have failed to account for other postnatal risk factors. These frequently unmeasured variables include poor parent-child interaction as well as a model of cumulative risk. Parent-infant interaction of the drug-exposed dyad has also been reported to be negatively impacted.^44,45 Furthermore, both human and animal studies provide evidence of the critical nature of the parent-child dyad in the cocaine-exposed pregnancy.⁴⁶ Children with prenatal cocaine exposure have been reported to have more attachment problems. Goodwin et al¹⁷ reported that Sprague Dawley rat pups raised by a cocaine-exposed dam had a shorter latency to induced aggression, independent of the pup's prenatal exposure status. Furthermore, in 1 study, dams exposed to cocaine during pregnancy seemed to have differential aggressive responses, exhibiting more aggression when they fostered their own biologic pups rather than unexposed foster pups.⁴⁷ This later finding was not confirmed by Johns and colleagues⁴⁸ who reported that maternal behaviors toward surrogate pups and aggression toward an intruder were affected by exposure to cocaine during pregnancy.

To address these issues in the study reported here, careful attention was paid to methodological decisions in an attempt to improve the likelihood of identifying a cocaine effect, should one exist. Specifically, an adequate sample size was calculated, prospective pregnancy data were collected, and blinded assessments were made of both the outcome and control variables. Even with these precautions, regular recurrent assessment of maternal urine across pregnancy was not performed. Hence, misclassification of women is still possible in this study. Every attempt was made to determine the impact of prenatal cocaine exposure after controlling for important covariates, including characteristics of the primary caretaker, other prenatal exposures, and postnatal risk factors. Additionally, children with severe reductions in IQ were eliminated from the analyses because of a possible association between mental retardation and behavioral problems. Tronick and Beeghly⁴³ also caution that previous studies may have failed to identify a cocaine effect because of the selection of global or insensitive measures. To minimize this likelihood, consultation with Dr Achenbach lead to the decision to evaluate the full spectrum of undercontrolled behavior by computing the Externalizing-Internalizing Difference from the TRF of the Child Behavior Checklist. In both univariate and multiple regression analyses, the Externalizing-Internalizing Difference was adversely affected by prenatal exposure to cocaine.

This is not the first study to report that children prenatally exposed to cocaine differed from control children in some aspect of behavior. Blinded reports of spontaneous play behavior have suggested that children prenatally exposed to drugs, cocaine, and polychlorinated biphenyls had more immature play strategies, reduced attention, and more abrupt transitions.⁴⁹ Boys in the drug-exposed group were particularly affected. Allesandri and colleagues⁵⁰ evaluating early infancy behavior found that the cocaine-exposed children were less aroused, less engaged in contingency testing, and demonstrated less evidence of interest, joy, surprise, sadness, and anger. Additionally, Karmel and Gardner⁵¹ reported that unlike control infants, the cocaine-exposed 1-month-old had no state modification of his preference for rapid frequency of visual stimuli. Similarly, Jacobson et al⁵² reported faster responsiveness on a visual expectancy paradigm in 6-month-old infants after prenatal cocaine exposure. Both of these studies controlled for prenatal alcohol exposure. In a controlled, blinded study, Mayes and her colleagues⁵³ also reported impaired regulation of arousal in 3-month-old infants after exposure to cocaine and other drugs during pregnancy. The exposed infants responded to novel stimuli with more and longer periods of crying and had more negative facial affect even when the analyses were controlled for perinatal and maternal demographic variables.

Although limited to infancy, the previously described studies do extend the observation period beyond the expected acute effects of residual cocaine metabolites. However, few previous studies have attempted to evaluate children beyond infancy. One small, inconsistently blinded study of children between 2 and 5 years of age failed to identify an effect of cocaine exposure.⁵⁴ A larger Dutch study of 29 polydrug-exposed and 35 comparison children tested through 30 months⁵⁵ failed to identify any differences between groups in univariate analyses. Similarly, parental reports did not demonstrate differences in child behavior.

Other studies have investigated the relationship between prenatal exposure to cocaine and parent reports of child behavior to 3 years of age. At age 3 years, prenatal drug abuse was found to have both a direct (higher externalizing scores) and an indirect effect on the parents' reports of child behavior.⁵⁶ In addition to higher externalizing scores, prenatal exposure negatively impacted characteristics of the home environment, which in turn had a negative effect on the parental report of child behavior. In a more detailed description, Griffith et al⁵⁷ reported that compared with controls, cocaine-exposed children were more likely to be rated as aggressive and destructive (Externalizing Behaviors). After control for other variables, the cocaine effect on Externalizing Behaviors persisted, while prenatal alcohol was found to be related to higher aggressive scores. Field⁵⁸ has also reported that grade school children prenatally exposed to cocaine were more likely to have clinically significant Externalizing scores on the Achenbach measure. In an uncontrolled study of women receiving services for drug-dependence, Hawley et al⁵⁹ reported higher Internalizing scores and more children with clinically significant Child Behavior Checklist scores among 20 cocaine-exposed subjects, compared with a control group recruited from Head Start. The Externalizing scores also approached significance. Kelley⁶⁰ in a small, nonblinded, and uncontrolled study reported that both biologic mothers and foster mothers perceived more behavioral abnormalities on standardized measures of child behavior among cocaine-exposed youngsters. These abnormalities included both hyperactivity and short attention span.

In contrast to these studies, a well-designed investigation reported by Richardson and colleagues⁶¹ failed to identify any TRF behavioral effects related to light to moderate prenatal cocaine exposure among 28 school-aged children. Study participants received prenatal care and most (64%) limited cocaine use to the first trimester of pregnancy. Teachers rated the cocaine-exposed children as having fewer social or attention problems than the unexposed children, although these differences did not persist after control for other important covariates, including first trimester use of alcohol, marijuana, and other illicit substances. A later report by the same research group identified more errors of omission on a continuous performance test related to prenatal cocaine exposure during the first trimester of pregnancy.⁶²

Despite previous reports of the detrimental effects of prenatal smoking on child behavior,⁶³ no such effects were identified in our study. Reports identifying a relationship between maternal smoking and child behavior have suggested that these findings were not consistently present at all ages.^64-66 Teachers reported no significant behavior problems (conduct, attention, or disruptive behaviors) related to maternal cigarette smoking at the 8-year assessment, although differences began by age 10 years.⁶⁷ Wakschlag and colleagues⁶⁵ began their investigation when the boys were 7 to 10 years of age. Although the boys were older than our sample at study entry, one third were not identified as having conduct disorder until evaluated at a later age. It is possible that continued follow-up of children in our study may also show age-related differences.

Finally, it is important to recognize that no one study is adequate to describe the effects of prenatal exposure of cocaine on child behavior. Cocaine is seldom used alone, women often underreport use, and no biologic measure currently available is without technical difficulties. Furthermore, local differences may occur in the purity and contaminant of illicit drugs. Additionally, animal studies have suggested that the effect of prenatal cocaine is related not only to the quantity, but also to the timing, of exposure. Only a few human studies have even begun to address this issue.⁶² Finally, because the subjects in this report all received prenatal care, the results of this study may not be generalizable to women with more chaotic lifestyles that preclude medical attention. Furthermore, as Tronick and Beeghly⁴³ so aptly reported, most studies of cocaine abuse in the available literature are associated with significant cumulative risk. Despite these cautions, in the study reported here, when prenatal cocaine abuse and current drug use were evaluated together, the overwhelming cost of illicit drug abuse on child outcome begins to be apparent. The studies of aggressive behavior in animals and the emerging data from human studies at least require our continued concern and future study.

	CONCLUSIONS

Top Abstract Methods Results Discussion Conclusion References

In the study reported here, prenatal cocaine exposure adversely affected the Externalizing-Internalizing Difference even after controlling for other prenatal exposures and postnatal risk factors, suggesting persistence of a biologic effect of prenatal cocaine exposure. A similar effect of prenatal exposure to cocaine on behavior has previously been identified in the neonatal period and early childhood. It is likely that misclassification of exposure status (failing to identify women with prenatal cocaine exposure) as well as the early age of the children in our report underestimates the true size of this effect. Prenatal alcohol exposure was associated with both Attention Problems and Delinquent Behaviors. Continued evaluation of the interaction between prenatal exposures and postnatal parental status is essential to best serve high-risk populations.

In this study, teacher-reported behavioral problems were also significantly related to gender and to several postnatal risk factors including the child's self-report of exposure to violence, current drug use around the child, and the child's whole blood lead level at 6 years old. Additionally, among the cocaine-exposed children (75% of whom remained in the custody of their biologic mother), changes in the child custody status predicted behavioral problems. Further analyses to identify the maternal characteristics that predict custody change in this cohort are underway.

The very significant effects of postnatal risk factors on child outcome is both distressing and, at the same time, cause for optimism. Irrespective of prenatal drug exposure, postnatal exposure to violence and drugs in the home were associated with teacher-reported behavioral problems. Although several authors have suggested mother-infant interaction problems in the cocaine-exposed dyad, at least 2 authors have been able to reduce these effects by well-designed intervention studies.^44,45 Further investigation is essential to identify the potential for amelioration of child outcome through intensive family intervention.⁶⁸ As is obvious from our report, simply removing the child from the home may not prevent poor outcome.

	ACKNOWLEDGMENTS

This research was funded by Grant DA08524 from the National Institute on Drug Abuse.

We thank the children, families, research assistants, and teachers who made this research possible, and Donna Dokho for secretarial assistance. We also thank Dr Thomas Achenbach for his invaluable suggestions and comments.

	FOOTNOTES

Received for publication Nov 9, 1999; accepted Apr 17, 2000.

Reprint requests to (V.D.-B.) Children's Hospital of Michigan, 3901 Beaubien, Detroit, MI 48201. E-mail: vdelaney{at}med.wayne.edu

	ABBREVIATIONS

TRF, Teacher's Report Form.

	REFERENCES

Top Abstract Methods Results Discussion Conclusion References

1. Newborns and addiction: a surprising study on the effects of cocaine. Newsweek. 1992
2. Toufexis A Innocent victims: damaged by the drugs their mothers took, crack kids will face social and educational hurdles and must count on society's compassion. Time 1991; 137:56
3. Frank DA, Bauchner H, Parker S, Neonatal body proportionality and body composition after in utero exposure to cocaine and marijuana. J Pediatr 1990; 117:622-626 [CrossRef][Medline]
4. Microchnick M, Frank DA, Cabral H, Turner A, Zuckerman B Relation between meconium concentration of the cocaine metabolites: benzoylecgonine and fetal growth. J Pediatr 1995; 126:636-638 [CrossRef][Medline]
5. Singer L, Arendt R, Song LY, Warshawsky E, Kliegman R Direct and indirect interactions of cocaine with childbirth outcomes. Arch Pediatr Adolesc Med 1994; 148:959-964 [Abstract/Free Full Text]
6. Coles CD, Platzman KA, Smith I, James ME, Falek A Effects of cocaine and alcohol use in pregnancy on neonatal growth and neurobehavioral status Neurolotoxicol Teratol 1992; 14:3-33
7. Eisen LN, Field T, Bandstra ES, Roberts JP, Morrow C, Laron SK, Steele BM Perinatal cocaine effects on neonatal stress behavior and performance on the Brazelton scale. Pediatrics 1991; 88:477-480 [Abstract/Free Full Text]
8. Mayes LC, Granger RH, Frank MA, Schottenfeld R, Bornstein MH Neurobehavioral profiles of neonates exposed to cocaine prenatally. Pediatrics 1993; 91:778-783 [Abstract/Free Full Text]
9. Richardson GA, Hamel SC, Goldschmidt L, Day NL The effects of prenatal cocaine use on neonatal neurobehavioral status. Neurotoxicol Teratol 1996; 18:519-528 [CrossRef][Medline]
10. Legido A, Clancy RR, Spitzer AR, Finnegan LP Electroencephalographic and behavioral-state studies in infants of cocaine-addicted mothers. Am J Dis Child 1992; 156:748-752
11. Delaney-Black V, Covington C, Ostrea E, Prenatal cocaine and neonatal outcome: evaluation of a dose-response relationship. Pediatrics 1996; 98:735-740 [Abstract/Free Full Text]
12. Eyler FD, Behnke M, Conlon M, Woods NW, Wobie K Birth outcomes from a prospective matched study of prenatal crack cocaine use. II. Interactive and dose effects on neurobehavioral assessment. Pediatrics 1998; 101:237-241 [Abstract/Free Full Text]
13. Tronick EZ, Frank DA, Cabral H, Mirochnick M, Zuckerman B Late dose-response effects of prenatal cocaine-exposure on newborn neurobehavioral performance. Pediatrics 1996; 98:76-83 [Abstract/Free Full Text]
14. Martin JC, Barr HM, Martin DC, Streissguth AP Neonatal neurobehavioral outcome following prenatal exposure to cocaine. Neurotoxicol Teratol 1996; 18:617-625 [CrossRef][Medline]
15. Smith RF, Mattran KM, Kurkjian MF, Kurtz SL Alterations in offspring behavior induced by chronic prenatal cocaine dosing. Neurotoxicol Teratol 1989; 11:35-38 [CrossRef][Medline]
16. Spear LP, Campbell J, Snyder K, Silveri M, Katovic N Animal behavior models. Increased sensitivity to stressors and other environmental experiences after prenatal cocaine exposure. Ann N Y Acad Sci 1998; 846:76-88 [CrossRef][Medline]
17. Goodwin GA, Heyser CJ, Moody CA, A fostering study of the effects of prenatal cocaine exposure: offspring behavior measures. Neurotoxicol Teratol 1992; 14:423-432 [CrossRef][Medline]
18. Johns JM, Noonan LR Prenatal cocaine exposure affects social behavior in Sprague Dawley rats. Neurotoxicol Teratol 1995; 17:569-576 [CrossRef][Medline]
19. Riley EP, Lafiette MH The effects of prenatal cocaine exposure on subsequent learning in the rat. NIDA Res Monogr 1996; 164:53-77 [Medline]
20. Molina VA, Wagner JM, Spear LP The behavioral response to stress is altered in adult rats exposed prenatally to cocaine. Physiol Behav 1994; 55:941-945 [CrossRef][Medline]
21. Heyser CJ, Spear NE, Spear LP Effects of prenatal exposure to cocaine on conditional discrimination learning in adult rats. Behav Neurosci 1992; 106:837-845 [CrossRef][Medline]
22. Church MW, Crossland WJ, Holmes PA, Overbeck GW, Tilak JP Effects of prenatal cocaine on hearing, vision, growth and behavior. Ann NY Acad Sci 1998; 846:12-28 [CrossRef][Medline]
23. Dow-Edwards DL, Freed LA, Fico T Structural and functional effects of prenatal cocaine exposure in adult rats. Pharmacol Biochem Behav 1990; 42:509-515
24. Akbari HM, Kramer HK, Whitaker-Azmitia PM, Spear LP, Azmiia EC Prenatal cocaine disrupts the development of the serotonergic system. Brain Res 1992; 572:57-63 [CrossRef][Medline]
25. Wang L, Pitts DK Perinatal cocaine exposure decreases the number of spontaneously active midbrain dopamine neurons in neonatal rats. Synapse 1994; 17:210-215 [CrossRef][Medline]
26. Leslie CA, Robertson MW, Jung AB, Liebermann J, Bennett JP Effects of prenatal cocaine exposure on postnatal development of neostriatal dopaminergic function. Synapse 1994; 17:210-215
27. Fang Y, Janowsky A, Ronnekliv OK Cocaine exposure in fetal rheusus monkey: consequences for dopamine D1- and D2-like receptor binding densities. Dev Brain Res 1997; 104:163-174 [CrossRef][Medline]
28. Mirochnick M, Meyer J, Cole J, Herren T, Zuckerman B Circulating cetecholamine concentrations in cocaine-exposed neonates: a pilot study. Pediatrics 1991; 88:481-485 [Abstract/Free Full Text]
29. Mirochnick M, Meyer J, Frank DA, Cabral H, Tronick EZ, Zuckerman B Elevated plasma norepinephrine after in utero exposure to cocaine and marijuana. Pediatrics 1997; 99:555-559 [Abstract/Free Full Text]
30. Whitaker-Azmitia PM Role of the neurotrophic properties of seratonin in the delay of brain maturation induced by cocaine. Ann N Y Acad Sci 1998; 846:12-28
31. Lester BM, LaGasse LL, Bigsby R Prenatal cocaine exposure and child development: what do we know and what do we do? Semin Speech Lang 1998; 19:123-146 [Medline]
32. Derogatis LR, Lipman RS, Covi L SCL-90: an outpatient psychiatric rating scale, preliminary report. Psychopharmacol Bull 1973; 1:13-28
33. Norbeck J, Lindsey AM, Carrieri VL, The development of an instrument to measure social support. Nurs Res 1981; 30:264-269 [Medline]
34. Caldwell BM, Bradley RH. Home Observation for Measurement of the Environment. Little Rock, AR: University of Arkansas; 1979
35. Hollingshead AB. Four Factor Index of Social Status. New Haven, CT: Yale University, Department of Social Work; 1975
36. Wechsler D. Weschler Preschool and Primary Scale of Intelligence-Revised Manual. San Antonio, TX: Psychological Corporation; 1989
37. Richters JE, Martinez P The NIMH Community Violence Project: children as victims of and witnesses to violence. Psychiatry 1993; 56:7-20 [Medline]
38. Achenbach TM. Manual for the Teacher's Report Form and 1991 Profile. Burlington, VT: University of Vermont Department of Psychiatry; 1991
39. Achenbach TM, Edelbrock C The classification of child psychopathology: a review and analysis of empirical efforts. Psychol Bull 1978; 85:1275-1301 [CrossRef][Medline]
40. Borenstein M, Cohen J. Power Analysis: A Computer Program. Hillsdale, NJ: Lawrence Erlbaum; 1988
41. Brown RT, Coles CD, Smith IE, Effects of prenatal alcohol exposure at school age: attention and behavior. Neurotoxicol Teratol 1991; 13:369-376 [CrossRef][Medline]
42. Eyler FD, Behnke M Early development of infants exposed to drugs prenatally. Clin Perinatol 1999; 26:107-150 [Medline]
43. Tronick EZ, Beeghly M Prenatal cocaine exposure, child development, and the compromising effects of cumulative risk. Clin Perinatol 1999; 26:151-171 [Medline]
44. Field T, Scafaldi F, Pickens J, Polydrug-using adolescent mothers and their infants receiving early intervention. Adolescence 1998; 33:117-143 [Medline]
45. Black M, Nair P, Kigh C, Wachtell R, Roby P, Schuler M Parenting and early development among children of drug-abusing women: effects home intervention. Pediatrics 1994; 94:13 [Abstract/Free Full Text]
46. Rodning C, Beckwith L, Howard J Characteristics of attachment organization and play organization in prenatally drug-exposed toddlers. Dev Psychopathol 1989; 1:277-289 [CrossRef]
47. Heyser CJ, Molina VA, Spear LP A fostering study of the effects of prenatal cocaine exposure: maternal behaviors. Neurotoxicol Teratol 1992; 14:415-421 [CrossRef][Medline]
48. Johns JM, Noonan, LR, Zimmerman LI, Li L, Pedersen CA Effects of chronic and acute cocaine treatment on the onset of maternal behavior and aggression in Sprague Dawley rats. Behav Neurosci 1994; 108:107-112 [CrossRef][Medline]
49. Beckwith L, Rodning, C, Norris D, Phillipsen L, Khandabi P, Howard J Spontaneous play in two-year-olds born to substance abusing mothers. Infant Ment Health J 1994; 15:189-201
50. Allesandri SM, Sullivan MW, Imaizumi S, Lewis M Learning and emotional responsivity in cocaine-exposed infants. Dev Psychol 1993; 29:989-997 [CrossRef]
51. Karmel BZ, Gardner JM Prenatal cocaine exposure effects on arousal-modulated attention during the neonatal period. Dev Psychobiol 1996; 29:463-480 [CrossRef][Medline]
52. Jacobson SW, Jacobson JL, Sokol RJ, Martier SS, Chiodo LM New evidence for neurobehavioral effects of in utero cocaine exposure. Pediatrics 1996; 129:581-590
53. Mayes LC, Bornstein MH, Chawarska K, Haynes M, Granger RH Impaired regulation of arousal in 3-month-old infants exposed prenatally to cocaine and other drugs. Dev Psychopathol 1996; 8:29-41
54. Rotholz DA, Snyder P, Peter G A behavioral comparison of preschool children at high and low risk from prenatal cocaine exposure. Educ Treat Child 1995; 18:1-18
55. vanBaar A Development of infants of drug dependent mothers. J Child Psychol Psychiatry 1990; 31:911-920 [CrossRef][Medline]
56. Azuma S, Chasnoff I Outcome of children prenatally exposed to cocaine and other drugs: a path analysis of three-year data. Pediatrics 1993; 92:396-402 [Abstract/Free Full Text]
57. Griffith DR, Azuma S, Casnoff I Three year outcome of children exposed prenatally to drugs. J Am Acad Adolesc Psychiatry 1994; 33:20-27 [CrossRef]
58. Field T. Cocaine exposure and intervention in early development: mothers, babies and cocaine. Mahwah, NJ: Lawrence Erlbaum Associates; 1994:355-368
59. Hawley LT, Halle TG, Drasin RE, Thamas NG Children of addicted mothers: effects of the "crack epidemic" on the caregiving environment and the development of preschoolers. Am J Orthopsychiatry 1995; 65:364-379 [Medline]
60. Kelley SJ Parenting stress and child maltreatment in drug-exposed children. Child Abuse Negl 1992; 16:317-328 [CrossRef][Medline]
61. Richardson GA, Conroy ML, Day NL Prenatal cocaine exposure: effects on the development of school-age children. Neurotoxicol Teratol 1996; 18:627-633 [CrossRef][Medline]
62. Leech SL, Richardson GA, Goldschmidt L, Day NL Prenatal substance exposure: effects on attention and impulsivity of 6-year-olds. Neurotoxicol Teratol 1999; 21:109-118 [CrossRef][Medline]
63. Fergusson DM, Woodward LJ, Horwood LJ Maternal smoking during pregnancy and psychiatric adjustment in late adolescence. Arch Gen Psychiatry 1998; 55:721-727 [Abstract/Free Full Text]
64. Orlebeke JF, Knol DL, Verhurst FC Increase in child behavior problems resulting from maternal smoking during pregnancy. Arch Environ Health 1997; 52:317-321 [Medline]
65. Wakschlag LS, Lahey BB, Loeber R, Green SM, Gordon RA, Leventha BL Maternal smoking during pregnancy and the risk of conduct disorder in boys. Arch Gen Psychiatry 1997; 54:670-676 [Abstract/Free Full Text]
66. Millberger S, Biederman J, Farone SV, Chen L, Jones J Is maternal smoking during pregnancy a risk factor for attention deficit hyperactivity disorder in children? Am J Psychiatry 1966; 153:1138-1142 [Abstract/Free Full Text]
67. Fergusson DM, Horwood J, Lynskey MT Maternal smoking before and after pregnancy: effects on behavioral outcomes in middle childhood. Pediatrics 1993; 92:815-822 [Abstract/Free Full Text]
68. Kilbride H, Castor C, Hoffman E, Fuger KL Thirty-six-month outcome of prenatal cocaine exposure for term or near-term infants: impact of early case management. J Dev Behav Pediatr 2000; 21:19-26 [Medline]