Cognitive and Functional Outcomes at Age 21 After Prenatal Cocaine/Polydrug Exposure and Foster/Adoptive Care

Abstract

Objective:

Prenatal cocaine exposure (PCE) has been linked to specific cognitive deficits and behavioral outcomes through early adolescence but there is little information on adult outcomes nor on the relationship of environmental interventions, such as foster/adoptive care, to outcomes.

Methods:

At 21 years, data were available on 325 young adults, [163 PCE and 162 nonexposed (NCE)], primarily African-American, with low SES, who were followed from birth in a prospective longitudinal cohort study. Participants were administered the Wechsler Abbreviated Scale of Intelligence (WASI-II) and surveyed regarding high school completion, problematic substance use, and incarceration/ probation history. In the PCE group, 32 remained in nonkinship foster/adoptive care (PCE/FA) from early in life (< 4 years) to 17 years. Group differences were examined through t-tests, MANOVA/ MANCOVA with post-hoc analyses, comparing outcomes and environmental correlates of young adults with PCE vs. NCE, as well as outcomes of PCE young adults in non-kinship foster/adoptive care (PCE/ FA) vs. PCE in birth/kinship care and NCE young adults.

Results:

At 21 years, young adults with PCE had lower mean Full Scale (83.7 ± 10.4 vs. 87.3 ± 12.5, p < .01) and Perceptual Reasoning IQs (87.3 ± 11.5 vs. 91.4 ± 13.9, (p < .02), lower high school completion rates (75% vs. 86%, p < .02), and were marginally more likely to have been on probation than NCE young adults, but did not differ in Verbal IQ, self-report of problematic substance use or incarceration. Young adults with PCE in F/A had similar lower IQ scores but had better verbal skills and high school graduation rates that did not differ from NCE young adults (80.6 vs 86.2%, p > .05). They had higher drug exposure at birth and more experiences of maltreatment (p’s <.05) but their home environment quality was better and lead levels lower (p’s <.05) than those of young adults with PCE in birth/kinship care.

Conclusions:

Young adults with PCE had lower Perceptual Reasoning and Full-Scale IQ scores, independent of caregiving placement, compared to non-exposed young adults. Young adults with PCE placed in non-kinship foster/adoptive care had lower lead levels, more stimulating home environments, better vocabulary skills and were more likely to graduate from high school than those in birth/kinship care,but were not different in their self-report of problematic substance use, or experiences of incarceration or probation. Our data suggest that some cognitive deficits observed in young adults with PCE may be biologically based, but that some functional outcomes can be modified through environmental interventions. Our data also reflect the complexity of disentangling the effects of teratologic exposures on long term outcomes across a variety of domains and the need for studies of children in the foster care system.

1. Introduction

Prenatal cocaine/polydrug exposure (PCE) affected over a million births during the 1980s-1990s, primarily in poor urban areas of the United States, raising concerns about long-term developmental effects on offspring. Prenatal cocaine exposure is considered a risk factor for later developmental problems due to its biologic effects on multiple neurotransmitter systems affecting fetal development (Martin et al., 2016) but also due to a range of caregiving risk factors frequently associated with maternal drug use known to have negative effects on child development, including poverty, maternal low education, mental health problems, and violence exposure (Singer, Salvator, Arendt et al., 2002). A large number of well-designed studies have linked PCE to a range of developmental problems, including: fetal growth restriction (Eyler et al., 1998; Frank et al., 1998; Singer et al., 2002), cardiac alterations (Mehta et al., 2001, 2002a, 2002b), behavioral problems extending from infancy (Frank et al., 1998; Singer et al., 2000; Richardson et al., 2008, Eiden et al., 2009) to aggressive behavior and conduct difficulties through adolescence (Accornero et al., 2006; Minnes et al., 2010; Delaney-Black et al., 2011; Min et al., 2014, 2015, 2018). Specific cognitive and language deficits have been described from the preschool years through adolescence (Behnke et al., 2006; Lewis et al., 2004, 2011; Singer et al., 2004; Bennett et al., 2008; Singer et al., 2008; Mayes et al., 2007). However, there is little information on outcomes of individuals with PCE during emerging or young adulthood, a critical developmental transition period during which these earlier identified problems may be exacerbated or resolved.

Only two studies to date, with somewhat conflicting results, have addressed the longitudinal outcomes of offspring with PCE in emerging adulthood. With excellent (76%) retention of a birth sample, Richardson and colleagues (Richardson et al., 2019) found direct associations of PCE with emotional regulation problems, arrest history and a diagnosis of Conduct Disorder at age 21, as well as with early adolescent initiation of marijuana use that predicted later marijuana use at age 21. Another prospective longitudinal cohort study (Forman et al., 2017) with lower (54%) retention, found no association of PCE with a composite measure of adaptive functioning at ages 18–21. Neither of these studies addressed whether non-parental caregiving was a potential factor related to offspring outcomes, although Richardson’s study (Richardson et al., 2019) noted that 13% of their sample was living with a non-maternal caregiver at follow-up. Forman et al, (Forman et al., 2017) do not mention placement history at follow-up, but (Frank et al., 2002) reported in earlier studies of their cohort that cocaine-exposed children in kinship or unrelated foster care performed more poorly on cognitive assessments at 2 years of age than those in the care of their birth mother.

Prenatal drug exposure is one of the most common precipitants of non-parental caregiving placements for infants (Nair et al., 1997; Tung et al., 2020). Despite its significance as an intervention for prenatal substance exposure, there is little information on the contextual characteristics of non-kinship placement nor its relationship to health and developmental outcomes in drug exposed cohorts. A number of maternal factors other than substance use are related to placement of infants in foster, kinship, or adoptive care, including younger maternal age, having two or more children, and depressive symptoms (Nair et al., 1997). In a group of cocaine using women, those whose infants were placed in non-maternal care at birth had evidence of heavier drug use during pregnancy than those not placed, reported more psychological distress and negative coping strategies and were more likely to have a history of childhood physical abuse and neglect (Minnes et al., 2008). Infant vulnerability as manifest through prematurity or low birthweight also increases the likelihood of placement out of maternal care (Tung et al., 2020).

A small number of available research studies are inconclusive about the effects of placement in non-maternal care for infants with prenatal drug exposure. In a sample of 1,092 three-year-old children, Bada (Bada et al., 2008) found that caregivers of children in kinship care reported fewer externalizing problems in their children than did caregivers of children in non-kinship care, and children in kinship care also scored better in daily living and communication skills. In contrast, Brown et al (Brown et al., 2004) found a protective effect of non-kinship care for drug exposed infants, with cocaine exposed children in non-parental care experiencing a measurably better home environment and performing better on assessments of development at 2 years of age. Similarly, infants adopted before 20 months of age showed language and motor improvements at follow-up one year after placement (Tung et al., 2020). Notably, none of these studies followed children beyond the preschool years.

In our longitudinal study (Singer, Arendt, Minnes et al., 2002), protective effects of nonkinship foster or adoptive placement were found for children with PCE on verbal and language skills and on overall IQ during the preschool years through early adolescence (Lewis et al., 2004; Singer et al., 2004; Singer et al., 2008; Lewis et al., 2011), but these positive associations were not apparent on measures of child or caregiver-reported behavioral problems at 4–12 years (Linares et al., 2006; Minnes et al., 2010; McLaughlin et al., 2011), or on adolescent substance use (Min et al., 2014; Minnes et al., 2014), or on specific cognitive functions of perceptual reasoning skills at 4, 9 years and 15 years (Singer et al., 2004, 2008, 2018).

In the present study, we assessed the association of prenatal cocaine exposure with measures of cognitive ability as well as functional outcomes of high school completion, engagement with the legal system through incarceration or probation, and self-reported problematic substance use at age 21 in a prospective cohort followed since birth. We also assessed whether there were differences within the cocaine-exposed group on cognitive and functional outcomes based on long term (<4 to 17 years) of non-kinship foster or adoptive care, and whether any outcomes differed by sex, as teratologic sequelae may vary by sex.

2. Methods

2.1. Sample and procedures

This study included 325 (PCE, NCE) young adults and their birth mothers or caregivers recruited at birth (September 1994 - June 1996) from an urban county hospital for a longitudinal Investigation of the effects of PCE. All recruited mothers were identified from a high-risk population screened for drug use. Urine drug toxicology screens were performed by the hospital staff on women who received no prenatal care, seemed to be intoxicated or taking drugs, had a history of involvement with the Department of Human Services in previous pregnancies due to drug use, self-admitted drug use, or appeared to be at high risk for drug use after an interview with hospital staff. Women with a psychiatric history, low intellectual functioning indicated in medical chart review, HIV-positive status, or chronic medical illness were excluded, as were infants with Down syndrome, fetal alcohol syndrome, or congenital heart defects. A nurse recruiter approached 647 screened women immediately before or after infant birth; of these 647 women, 54 were excluded, 155 refused to participate, and 23 did not come to the enrollment visit.

Maternal and infant urine and infant meconium samples were obtained shortly before or after infant birth and analyzed for cocaine and other drug metabolites, including benzoylecgonine, metahydroxybenzoylecgonine, cannabinoids, opiates, phencyclidine, amphetamines, cocaethylene, and benzodiazepines. A total of 415 newborns and their birth mothers were enrolled, of which 218 infants were identified as PCE based on positive screens of maternal or infant urine, infant meconium, or maternal self-report. Infants negative on all indicators of PCE were identified as NCE,but may have been exposed to other substances (i.e., alcohol, tobacco, marijuana), forming a comparison group. Subjects and caregivers were assessed by separate examiners masked to exposure status at follow-up at 1 week, 6, 12, and 24 months and 4, 6, 9–12, 15, 17 and 21 years postpartum.

Since birth, 14(10 PCE, 4 NCE) enrolled children died from sudden infant death syndrome (4 PCE, 2 NCE), cardiopulmonary arrest (1 PCE), pneumonia (1 PCE), accidental asphyxia (1 PCE), respiratory distress syndrome (1 PCE, 1 NCE), gunshot (1 PCE, 1 NCE) and unknown illness (1 PCE). At 21 years, data were available for 325 (81% of the 401 living participants). Compared to participants, attrition in the PCE group was associated with male sex, white race, prenatal alcohol exposure and younger maternal age. The NCE adults lost to follow up were more likely to be white, of higher birthweight compared to those remaining in the study. Attrition in the PCE group in birth or kinship care was also associated with male sex and prenatal alcohol exposure and low birthweight, while those in the PCE group in nonkinship foster or adoptive care lost to follow up did not differ from those retained at 21.

The Institutional Review Board of the participating hospital approved this study. All participants were given a monetary stipend, lunch, and transportation costs. Written informed consent was obtained. A Certificate of Confidentiality (DA-09-146) was obtained from the U.S. Department of Health and Human Services to protect the release of drug-related information from forced disclosure. For measures taken before 18 years of age, consent was obtained from the caregiver and assent from the child was obtained beginning at 9 years of age.

2.2. Measures

2.2.1. Prenatal cocaine and other substance exposures

After birth, a research assistant interviewed mothers about prenatal drug use, using a Timeline Follow Back method (Singer, Salvator, Arendt et al., 2002). To quantify use, women were asked to recall frequency and amount for the month prior to and for each trimester of pregnancy. The number of tobacco cigarettes and marijuana “joints” smoked, and the number of drinks of beer, wine, or hard liquor per day were computed, with each drink equivalent to 0.5 oz of absolute alcohol. For cocaine, the number of “rocks” consumed and the amount of money spent per day were noted. For each drug, frequency of use was recorded on a Likert-type scale ranging from 0 (not at all) to 7 (daily use), converted to reflect the average number of days per week a drug was used, except for cigarettes, collected as the number smoked per day. Frequency was multiplied by the amount used per day to compute an average use score for the month prior to pregnancy and for each trimester. This score was then averaged to obtain a total average score. Measures were updated at each follow-up. The current caregiver was interviewed at follow-up if the birth mother was not the caregiver.

Maternal education level and socioeconomic status (on public assistance vs. none) were determined. The Peabody Picture Vocabulary Scale-Revised (PPVT-R)(Dunn & Dunn, 1981) and the block design and picture completion subscales of the WAIS-R (Wechsler, 1981) were used to measure maternal vocabulary and nonverbal intelligence. The Global Severity Index (GSI), a summary scale of the Brief Symptom Inventory (Derogatis, 2001), measured psychological distress. Maternal race, age, parity, number of prenatal visits, infant gestational age, birth weight, length, head circumference, were taken from hospital records. At each follow-up, caregiver information was updated. For the present analyses, the quality of the caregiving environment was assessed at the 15-year visit with the Home Observation for Measurement of the Environment: Adolescent version (EA-HOME; α = .83)(Caldwell & Bradley, 1984).

At 2 and 4 years, children participated in a separate study of blood lead levels. Blood samples could not be obtained from some children due to parental refusal, inability to draw blood without undue stress, child sickness or logistical difficulties. The numbers of subjects with valid blood measures at two and four years were 143 and 274, respectively. For 122 children seen at both times, measures were averaged. A higher percentage of African-American and married women, and a lower percentage of foster parents, consented for blood collection.

At 4 years, children were classified as having been placed in birth/kinship care, or nonkinship foster/adoptive care. To be considered in the latter category, placement needed to be with a non-relative caregiver. At birth, 63 infants who were PCE were placed out of maternal care, (37 in non-kinship foster/adoptive care, 26 in kinship care) ( Minnes et al., 2008). Forty-four children with PCE placed in non-kinship foster/adoptive care were identified at 4 years, and 32 remained in non-kinship care and were seen at 17 years for follow up. Examining demographic, lead levels and outcome variables of this group at 4 years indicated that they differed significantly from both the remaining PCE children and the NCE group, having higher HOME scores, lower lead levels, higher caregiver vocabulary and lower caregiver stress. Children with PCE placed in kinship care did not differ on any demographic variable or outcome measure from those remaining with their birth family at 4 years; therefore they remained with the birth family group for analyses. In the NCE group at age 4, 7 children were identified to be in nonkinship foster care, with 4 remaining at 17 years. Four nonexposed children were in kinship care at age 4 (3 with grandmother, 1 with other relative), with 3 remaining at age 17. Because of the small numbers, they were retained in the NCE group for analyses.

At age 17, adolescents’ lifetime exposure to violence, crime, and abuse was assessed retrospectively using the Juvenile Victimization Questionnaire (JVQ), composed of 33 items (Hamby et al., 2004; Finkelhor et al., 2005; 2009) covering five areas of victimization: conventional crime, child maltreatment, peer and sibling victimization, sexual victimization, and witnessing/indirect exposure to violence. Lifetime poly-victimization (ranged 0 – 5) was also calculated, indicating the total number of areas of violence to which adolescents were exposed. The Cronbach’s α of the JVQ total scores (i.e., lifetime poly-victimization) in our sample was .91. Criterion-related validity has been established with the measures of trauma symptomatology in children and adolescents (Finkelhor et al., 2005). The JVQ also shows acceptable test-rest reliability with mean κ = .59 and acceptable internal consistency with α = .80.

At 21 years, examiners individually administered the Wechsler Abbreviated Scale of Intelligence-Second Edition (WASI-II) (Wechsler, 2011), a brief, reliable estimate of general cognitive ability (FSIQ), perceptual reasoning (PRI) and verbal comprehension (VIQ) for ages 6–90. The 4-subscale form was administered (Vocabulary, Similarities, Block Design and Matrix Reasoning). Average reliability coefficients (α) range from .91 to .97, with concurrent validity with comprehensive measures of intelligence. For 14 participants who were incarcerated during the 21-year visit, WASI-II scores were imputed from their 15-year WISC-IV data, as correlations between 15 and 21 years for the remaining sample were above .80 for summary scales.

At 21 years, the Substance Abuse Module (SAM) of the Composite International Diagnostic Interview (Robins et al., 1988) was administered that yields an assessment of self reported substance use disorders (SUDs) (yes/no)) and severity (none, mild, moderate, severe) for individuals ≥ 15 years. The interview consists of questions about onset, recency of symptoms, withdrawal symptoms, consequences for each category of substances used, quantity/frequency of heaviest use, use in the past 12 months, impairment and treatment seeking. Reliability is high for both Black and white respondents. The dichotomized outcome (yes/no to mild, moderate, or severe problematic use for each substance) was used in analyses, requiring 2 or more symptoms of problematic use for inclusion.

Achievement of high school graduation (yes/no) was self-reported. Participants were also asked if they had been incarcerated and/or ever on probation since age 18.

2.3. Statistical Analyses

We conducted statistical analyses in several steps. Two-way multivariate or univariate analyses of variance (MANOVA/ANOVA), or covariance (MANCOVAs) to assess sex and its interaction effects with PCE, were conducted on unadjusted values of outcomes to assess 1) group differences by cocaine exposure and 2) group differences by cocaine exposure with and without foster/adoptive care in comparison to non-exposure. Significant (p < .05) or marginally significant (p <.10) overall main effects were followed by Tukey-Kramer post-hoc tests to adjust for multiple comparisons. Groups were also compared for descriptive purposes on incidence of standard scores reflecting intellectual disability (< 70) or above the normative mean (> 100).

3. Results

3.1. Caregiver characteristics by cocaine exposure

Table 1 presents birth mother and current caregiver characteristics by prenatal cocaine exposure for the 325 young adults at 21-year follow-up. As per our prior studies, birth mothers in the cocaine group had fewer years of education, were less likely to be married, and marginally more likely to be psychologically distressed. They also used more alcohol and marijuana throughout pregnancy and smoked more cigarettes the month prior to pregnancy than mothers of non-exposed infants. Current caregiver characteristics when offspring were 17 years of age did not differ between PCE and NCE caregivers in years of education, vocabulary, HOME environment, psychological distress or current cigarette, alcohol, and other drug use, nor on maternal age, parity, number of prenatal visits, or vocabulary score.

Table 1.

Maternal and Current Caregiver Characteristics by Prenatal Cocaine Exposure (PCE) and Non-Exposure (NCE) (N = 325)

	Total	PCE Group 1, Birth Maternal/Relative	Group II, Foster/Adoptive	NCE Group III Total	PCE/NCE	Groups I, II, III
	(n = 163)	(n = 131)	(n = 32)	(n = 162)
	M (SD)	M (SD)	M (SD)	M (SD)	P	p
Birth mother
Mother’s age at birth	29.96 (4.86)	29.79a (4.86)	30.63b (4.88)	25.59ab (4.72)	0.56	< .001
Parity	3.51 (1.87)	3.25ab (1.57)	4.56ac (2.56)	2.77bc (1.88)	0.64	< .001
Number of prenatal visits	5.39 (4.67)	5.83ab (4.68)	3.56ac (4.22)	8.85bc (4.84)	0.64	< .001
African American, n (%)	135 (82.8)	110 (83.97)	25 (78.12)	134 (84.72)	0.98	0.74
Maternal years of education	11.61 (1.69)	11.7 (1.72)	11.25a (1.54)	11.96a (1.4)	0.02	0.04
Married n (%)	14 (8.6)	13 (9.92)	1 (3.12)	27 (16.67)	0.03	0.05
Low SES, n (%)	159 (98.1)	129 (99.23)	30 (93.75)	158 (97.53)	0.7	0.15
PPVT Standard score	73.68 (13.97)	74.77 (13.82)	68.97a (13.83)	77.89a (15.11)	0.29	0.01
WAIS-R Block Design	6.96 (2.08)	7.06 (1.95)	6.5 (2.54)	7.14 (2.05)	0.94	0.3
WAIS-R Picture Completion	6.68 (2.13)	6.78 (2.13)	6.27 (2.13)	6.87 (2.34)	0.71	0.4
BSI Global Severity Index	0.85 (0.78)	0.80a (0.74)	1.05b (0.91)	0.52ab (0.55)	0.06	< .001
Prenatal drug use
Cigarette per day	11.56 (11.05)	10.25ab (8.06)	17.17ac (18.35)	3.81abc (7.28)	0.06	< .001
Month prior cigarettes	13.29 (12.34)	12.38ab (10.41)	17.20ac (18.19)	5.18abc (9.74)	0.001	< .001
1st trimester cigarettes	12.67 (12.4)	11.21ab (9.72)	18.93ac (19.22)	3.94abc (8.03)	0.19	< .001
2nd trimester cigarettes	10.58 (11.56)	9.05ab (8.27)	17.17ac (19.28)	3.12abc (7.09)	0.16	< .001
3rd trimester cigarettes	9.43 (11.16)	8.05ab (7.71)	15.37ac (19.25)	3.00abc (6.06)	0.21	< .001
Alcohol dose per week	10.64 (18.54)	9.25ab (15.52)	16.59ac (27.57)	1.50abc (4.82)	< .001	< .001
Month prior drinks	14.06 (23.9)	13.3 (21.96)	17.3 (31.08)	2.72 (8.19)	< .001	< .001
1st trimester drinks	13.14 (25.07)	11.22ab (20.54)	21.35ac (38.32)	1.41abc (4.05)	< .001	< .001
2nd trimester drinks	8.61 (20.81)	7.05ab (17.42)	15.28ac (30.94)	0.63abc (3.18)	< .001	< .001
3rd trimester drinks	6.76 (18.19)	5.43ab (13.48)	12.43ac (30.89)	1.23abc (8.28)	< .001	< .001
Marijuana dose per week	1.47 (3.68)	1.22 (2.96)	2.54a (5.79)	0.52a (3.25)	< .001	< .001
Month prior marijuana	1.7 (3.98)	1.62a (3.66)	2.06b (5.19)	1.44ab (10.12)	< .001	< .01
1st trimester marijuana	1.63 (4.32)	1.45a (3.79)	2.38b (6.08)	0.37ab (2.18)	< .001	< .001
2nd trimester marijuana	1.48 (4.56)	1.09ab (3.55)	3.10ac (7.32)	0.18abc (1.76)	< .001	< .001
3rd trimester marijuana	1.1 (4.09)	0.74a (2.94)	2.63ab (7.04)	0.09b (0.79)	< .001	< .001
Cocaine units per week	23.38 (38.37)	21.87 (37.42)	29.55 (42.1)	0 (0)	n/a	n/a
Month prior cocaine	30.27 (54.4)	29.72 (51.77)	32.52 (64.94)	0 (0)	n/a	n/a
1st trimester cocaine	31.16 (59.84)	28.11 (53.41)	43.63 (80.92)	0 (0)	n/a	n/a
2nd trimester cocaine	22.59 (53.18)	19.56 (48.7)	34.99 (68.07)	0 (0)	n/a	n/a
3rd trimester cocaine*	12.89 (28.16)	10.19 (25.25)	23.52 (36.02)	0 (0)	n/a	n/a
Current caregiver at 17 years
Years of education	12 . 4 9 (2.54)	12.17ab (2.29)	14.04ac (3.15)	12.99bc (1.91)	0.11	<. 001
PPVT Standard score.	79.05 (15.17)	76.48a (13.74)	91.32ab (15.99)	78.92b (16.11)	0.31	< .001
WAIS-R Block Design Scale	7.19 (2.06)	7.12 (1.97)	7.5 (2.5)	7.23 (1.92)	0.58	0.71
WAIS-R Picture Completion	7.55 (2.54)	7.4 (2.39)	8.22 (3.12)	7.14 (2.38)	0.17	0.14
BSI Global Severity Index	0.39 (0.43)	0.39 (0.44)	0.42 (0.39)	0.36 (0.45)	0.87	0.49
HOME score at 15 years	47.63 (6.67)	46.88ab (6.73)	50.74a (5.5)	48.72b (6.15)	0.49	<.01
Postnatal drug use
Cigarette per day	4 .59 (8.53)	5.43ab (9.12)	0.62a (2.12)	3.38b (6.11)	0.45	< .001
Alcohol dose per	2.1 (4.91)	2.44 (5.33)	0.47 (0.95)	2.7 (8.47)	0.25	0.07
week Marijuana dose per week	0.56 (3.35)	0.68 (3.68)	0 (0)	0.44 (3.62)	0.6	0.46
Cocaine units per week	0 (0)	0 (0)	0 (0)	0 (0)	n/a	n/a

Note. BSI = Brief Symptom Inventory; HOME = Home Observation for Measurement of the Environment-Early Adolescent version; NCE = Non-prenatally Cocaine Exposed; SES = socioeconomic status; PCE = Prenatally Cocaine Exposed; PPVT = Peabody Picture Vocabulary Test; WAIS-R = Wechsler Adult Intelligence Scale-Revised; X = average. Within the same row, means identified with the same superscripts [a or b] are significantly different from each other (p < .05).

^*Group 1 differs from Group 2.

3.2. Caregiver characteristics by cocaine exposure and placement at age 17

Table 1 also presents the same characteristics by 3 groups: cocaine exposed young adults who remained in birth maternal or kinship care (PCE); cocaine exposed placed in non-kinship foster or adoptive care (PCE/F/A), and non-cocaine exposed (NCE). Since fewer than 5 nonexposed young adults had been in early nonmaternal care, they were not considered in analyses.

Infants with PCE placed in non-kinship foster/adoptive care before 4 years of age had birth mothers who were older, had less education, lower vocabulary scores, and more psychological distress than non-exposed infants and had more children and fewer prenatal care visits than both NCE infants and PCE infants in birth maternal care. Their mothers also used more cigarettes the month prior to and through pregnancy, more alcohol during pregnancy, and more marijuana in trimesters 2 and 3, than mothers of infants who remained in birth/kinship care and those of NCE infants. Within the cocaine exposure group, mothers of infants placed in nonkinship foster or adoptive care also used more cocaine in trimester 3. Current caregivers of the PCE/F/A group at 17 years differed by having more years of education and higher vocabulary scores than both NCE and birth/kinship caregivers. They also had higher current HOME environment scores and smoked fewer cigarettes a day than the caregivers in the birth/kinship group. Concurrent alcohol and marijuana use were not different across all groups, and no caregiver endorsed any cocaine use at 17 years.

3.3. Young adult characteristics by cocaine exposure

Table 2 shows differences in the young adult outcomes at follow-up. Based on prenatal cocaine exposure alone, there were no differences between groups in sex, race, birthweight, gestational age, preschool lead levels, alcohol or marijuana use by 17 years, or self-reported problematic cocaine, tobacco, alcohol, or marijuana use at 21 years. Five young adults with PCE reported using cocaine in the last 12 months vs. 1 in the non-exposed group, (p > .10). A higher percentage of the PCE group had reported experiencing maltreatment, (29.9% vs. 18.6%, p <.02) compared to the NCE group, and there was a nonsignificant trend for the PCE group to report higher peer/sibling victimization. High school completion was lower in the PCE group (74.5% vs. 86.2%, p < .01) and a probation history marginally more likely (14.5% vs. 8.0%, p < .07).

Table 2.

Young Adult Characteristics and Outcomes by Prenatal Cocaine Exposure (PCE) and Non-Exposed (NCE) (N = 325)

	Total	PCE Group I, Birth Maternal/Relative	Group II Foster/Adoptive	NCE Group III Total	PCE/NCE	Groups I, II, III
	(n = 163)	(n = 131)	(n = 32)	(n = 162)
	n (%)	n (%)	n (%)	n (%)	P	P
Male	72 (44.17)	57 (43.5)	15 (46.9)	78 (48.1)	0.47	0.73
African American,	135 (82.82)	110 (83.97)	25 (78.12)	133 (82.1)	0.86	0.73
Gestational Age, M, SD	37.87 (2.86)	38.08 (2.34)	37.00a (4.35)	38.36a (2.95)	0.13	0.05
Birth length (cm)x, M, SD	47.54 (0.2)	47.52a (3.57)	46.33b (5.37)	49.02ab (3.9)	< .00 1	< .001
Birth weight (grams)x, M, SD	2754.7 8 (32.8)	2751.40a (588.2)	2533.78b (850.59)	3082.85ab (721.81)	< .00 1	< .001
Head circumference (cm)x, M, SD	32.44 (0.12)	32.44a (1.88)	31.68b (3.11)	33.39ab (2.44)	< .00 1	< .001
2 and 4 years lead level (g/dl)y, M, SD	7.27 (4.13)	7.67a (4.26)	4.93ab (2.22)	8.13b (4.75)	0.17	0.01
JVQ total	2.56 (1.47)	2.51 (1.38)	2.79 (1.82)	2.38 (1.5)	0.29	0.37
Conventional crime	108 (69.2)	81 (64.8)	20 (69)	108 (69.2)	0.49	0.72
Childhood maltreatment	46 (29.87)	34 (27.2)	12a (41.38)	29a (18.59)	0.02	0.02
Peer and sibling victimization	104 (57.7)	83 (66.4)	21 (72.4)	90 (57.7)	0.07	0.17
Sexual victimization	31 (19.9)	20a (16)	11ab (37.93)	31b (19.87)	0.96	0.03
Witnessing and indirect exposure to violence	113 (73.4)	96 (76.8)	17 (58.6)	114 (73.1)	0.95	0.14
Tobacco use by 17 years	77 (50.33)	63 (51.22)	14 (46.67)	60 (38.46)	0.04	0.1
Alcohol by 17 years	87 (55.77)	71 (56.35)	16 (53.33)	79 (49.68)	0.28	0.53
Marijuana use by 17 years	81 (51.92)	69 (54.76)	12 (40)	69 (43.4)	0.13	0.11
Outcomes at 21 years
Problematic substance use at 21yrs
Cocaine	2 (1.3)	2 (1.6)	0 (0)	1 (0.7)	0.62	0.62
Tobacco	48 (32)	41 (33.33)	7 (25.92)	45 (29.22)	0.6	0.65
Alcohol	26 (17.33)	21 (17.07)	5 (18.52)	22 (14.28)	0.47	.71z
Marijuana	50 (33.56)	41 (33.61)	9 (33.33)	49 (32.24)	0.81	0.97
High school completion	114 (74.51)	89a (72.95)	25 (80.64)	131a (86.18)	0.01	0.02
Incarceration/probation	30 (19.35)	24 (19.35)	6 (19.35)	20 (13.24)	0.15	.35z
Incarceration	22 (14.28)	17 (13.82)	5 (16.13)	16 (10.6)	0.33	.53z
Probation	22 (14.47)	19 (15.7)	3 (9.68)	12 (7.95)	0.07	.14z
WASI-II at 21years
Perceptual Reasoning Index, M, SD	87.25 (11.46)	87.61a (11.72)	86.19b (10.27)	91.44ab (13.9)	0.01	0.01
Perceptual Reasoning > 100	18 (11)	16a (12.2)	2b (6.5)	43ab (26.5)	<.001	< .01
Perceptual Reasoning < 70	11 (6.7)	10 (7.63)	1 (3.13)	9 (5.56)	0.66	.70z
Block Design	41.28 (8.46)	41.51a (8.75)	40.58b (7.18)	44.10ab (9.51)	0.01	0.02
Matrix Reasoning	43.39 (7.78)	43.63a (7.86)	42.71 (7.42)	45.62a (8.98)	0.02	0.02
Verbal Comprehension Index, M, SD	84.3 (10.04)	83.98 (10.06)	86.52 (8.72)	86.07 (11.67)	0.17	0.21
Verbal Comprehension > 100	3 (1.8)	3 (2.29)	0 (0)	15 (9.26)	<.01	.01z
Verbal Comprehension < 70	12 (7.4)	10 (7.63)	2 (6.25)	15 (9.26)	0.54	.84z
Vocabulary, M, SD	40.1 (7.53)	39.63a (7.52)	42.61a (6.64)	41.25 (8.23)	0.3	0.08
Similarities, M, SD	40.64 (5.65)	40.66 (5.85)	40.84 (4.7)	41.7 (6.83)	0.17	0.35
Full Scale IQ, M, SD	83.69 (10.36)	83.80a (10.52)	84.22 (8.24)	87.25a (12.48)	0.01	0.03
Full Scale IQ > 100	7 (4.3)	7a (5.34)	0b (0)	21ab (12.96)	<.001	.01z
Full Scale IQ < 70	8 (4.9)	4 (3.28)	0 (0)	7 (4.76)	0.79	.82z

Note. JVQ = Juvenile Victimization Questionnaire- Adult Retrospective Version; NCE = Non-prenatally cocaine exposed; PCE = Prenatally cocaine exposed; WASI-II = Wechsler Abbreviated Scale of Intelligence-2nd Edition. Within the same row, means identified with the same superscripts [a or b] are significantly different from each other (p < .05).

^xp-value is adjusted for prematurity.

^ySub-sample size of lead is 255, included 109 Group I, 19 Group II, and 127 Group III.

^zFisher’s exact test

The MANCOVA on the Domain IQ scores of the WASI II, controlling for sex, was significant (Wilks’ λ = 0.98 [F = 3.97, df = 2, 320, p < .02]). There were no sex effects nor interaction effects of cocaine exposure with sex. Young adults with PCE had lower Perceptual Reasoning scores on the WASI-II (87.3 ± 12 vs. 91.4 ± 14, p < .02) and were less likely to score above the normative mean (11% vs. 26.5%, p < .001). While Verbal Comprehension scores were marginally lower (84.3 ± 10 vs. 86.1 ± 12, p < .08), Full Scale IQ was reliably lower (M = 83.7 ± 10 vs. 87.3 ± 13, p < .01), with only 4% of young adults with PCE vs. 13% of non-exposed young adults scoring above the normative mean of 100 (p < .01).

The MANCOVA for the subscale scores yielded an overall non-significant trend ((Wilks’ λ = 0.98 [F = 2.01, df = 4, 318, p <.09]), with effects for sex (Wilks’ λ = 0.97 [F = 2.86, df = 4, 318, p < .02]), but no interaction effects, p < 0.44). The ANOVA for the block design subscale was significant (F = 8.04, df = 3,319, p < .005), indicating that young adult women scored higher than men on the block design subscale, (43.8 vs. 41.8, p <.05) and that young adults in the NCE group scored higher than those in the PCE group (see Table 3). The ANCOVA for the matrix reasoning subscale was also significant (F =2.88, df = 3, 319 p < .05), with follow up tests indicating that young adults without PCE scored higher than those with PCE. There were no sex or interaction effects. There were no significant overall effects for the similarities or vocabulary indices.

3.4. Young adult characteristics by cocaine exposure and placement

Table 2 also shows the same birth characteristics and outcomes differentiating the PCE group by caregiver status in comparison to the NCE young adults. In contrast to classification by PCE alone, PCE young adults placed in foster/adoptive care had lower gestational ages at birth than NCE young adults, and lower preschool lead levels than both young adults with PCE who were in birth/kinship care and NCE groups. At age 17, they reported higher lifetime history of maltreatment than non-exposed adolescents (41.4% vs. 18.6%, p<.05), and higher history of sexual victimization, than both cocaine-exposed adolescents in birth or kinship care, and nonexposed adolescents (37.9% vs. 16% and 19.9%, p’s <.05). There were no differences in self report of problematic cocaine, alcohol, tobacco or marijuana use. Self-report of any cocaine use during the prior 12 months was positive for 5 in the birth/ PCE, 0 in the PCE/FA group and 1 in the non- exposed group (p<. 09).

The MANCOVA on the WASI ll Domain IQ scores controlling for sex was significant [Wilks’ λ = 0.97 (F = 2.58, df =4, 636, p < .05)], with a significant main effect for care status (F = 4.23, df = 2, 319, p < .01), but no sex or interaction effects. The MANCOVA on the subscale scores adjusting for sex also yielded a significant overall effect (F = 2.62, df = 4, 317, p <.03)], reflecting main effects of care status (F = 4.23, p < .01) and sex (F = 3.95, p <.05) on the block design subscale, indicating better performance by young adult women and non-exposed young adults. Young adults with PCE whether in birth/kinship or foster or adoptive care performed more poorly then NCE young adults in perceptual reasoning IQ, p’s < .05, < .07, respectively, reflecting poorer block design skills for cocaine exposed young adults regardless of care status and better performance overall for young adult women. The omnibus F test for the Verbal Comprehension IQ or for the similarities subscale was not significant. However, the overall test for the vocabulary scale yielded a marginal trend (p<.08), with post-hoc tests indicating that young adults with PCE in F/A care had better vocabulary skills than those in birth/kinship care (43.6 ± 7 vs. 39.6 ± 8, p < .05) and not different from non-exposed young adults (43.6 ± 7 vs. 41.3 ± 8, p>.05).

Young adults with PCE in F/A care also had high school graduation rates equivalent to NCE young adults (80.6% vs 86.2%, p > .05). The pattern of lower Perceptual Reasoning IQ with prenatal cocaine exposure previously seen was sustained even when non-kinship foster/adoptive care was considered, as both young adults with PCE in birth/kinship care and those in F/A care performed more poorly than young adults who were not exposed.

When groups were also compared (see Tables 3 and 4) for descriptive purposes on incidence of standard scores reflecting intellectual disability (< 70) or above the normative mean (>100), for all cognitive outcomes, there were no differences in intellectual disability across groups. However, on all dimensions, young adults with PCE were less likely to score above the normative mean, independent of care status.

4. Discussion

This study sought to assess the cognitive outcomes, as well as several components of adaptive functioning, (educational attainment, incarceration or probation history, and self-reported problematic substance use) at 21 years of age in a cohort of young adults with history of prenatal cocaine exposure and a comparison group prospectively followed since birth. A secondary aim sought to describe whether young adults with PCE who had been placed in nonkinship/foster or adoptive care early in life differed in outcomes from young adults with PCE who remained in birth/kinship care or non-exposed young adults.

Corroborating our previous findings at 4, 9, and 15 years (Singer et al., 2004, 2008, 2018), young adults with PCE had poorer perceptual reasoning skills than non-exposed young adults. At 21 years, PCE was also associated with lower full-scale IQ, poorer performance in verbal abstraction, lower high school completion rates, and a higher likelihood of engagement with the justice system, but not with self-report of problematic substance use. Early placement in non-kinship foster or adoptive care was associated with better outcomes only in vocabulary skills and high school graduation rate, despite the measurably better home environments and lower lead exposure of foster or adoptive placement. Although incidence was low, placement was also associated with a lower likelihood of self-reported cocaine use at 21 years compared to those in birth or kinship care.

The current data support our prior findings of persistent specific cognitive deficits in perceptual reasoning related to PCE independent of the enrichment of the placement environment. Perceptual reasoning is a nonverbal executive function reflecting abstract reasoning that has been negatively related to PCE in our and others’ prior studies ((Singer, Arendt, Minnes et al., 2002, Singer et al., 2004, 2008, 2018; Mayes et al., 2007; Arendt et al., 2004). MRI studies have also implicated PCE in deficits of visual- perceptual and abstract reasoning skills (Warner et al., 2006; Dow-Edwards et al., 2006) as have preclinical studies (Gendle et al., 2004; Stanwood et al., 2001). Previously we demonstrated that the perceptual reasoning differences were unrelated to alcohol, marijuana, tobacco or lead exposure in this cohort, but were partially mediated by birth head circumference which in turn was directly related to prenatal cocaine exposure (Singer, Arendt, Salvator et al., 2002).

There were few differences in outcome between prenatally cocaine-exposed young adults placed in non-kinship foster care and those in birth or kinship care, with the exception of better vocabulary and higher rate of high school graduation. Attainment of high school graduation has been shown to be related to lifelong positive outcomes in income, health and life satisfaction (Oreopoulos, 2007; Ross & Wu, 1995). Better vocabulary skills are an important component of communication that underlie academic achievement as well. Studies of children in foster care indicate that academic failure and poor school performance are common, so that the better high school graduation rates and better verbal skills achieved in this group are notable (Luke & O’Higgins, 2018), particularly given their higher birth risk overall, greater prenatal substance exposure and higher victimization levels compared to cocaine exposed infants remaining in birth/kinship care.

Surprisingly there were no differences across all groups in self-reported problematic substance use or incarceration and only marginal differences in probation experiences. Previously we have identified differences in adolescent drug use in this cohort (Min et al., 2015) similar to Richardson’s sample (Richardson et al., 2019). When biomarkers were used, the identification of cocaine use was much higher than in this study, and much higher than self report measures taken (Min et al., 2014; Minnes et al, 2014). While the criteria for problematic substance use in the assessment used in this study were specific for adult diagnosis of Substance Use Disorder and may be too stringent to detect differences during early adulthood that do not meet criteria for a medical diagnosis, rates of problematic tobacco, alcohol and marijuana use were relatively high and suggest this young adult group overall is at risk for later addictive disorders. Caution is warranted related to the low numbers of problematic or any cocaine use in the sample given that the measure used relied on self-report. Self report of any cocaine use was marginally higher for those with PCE in birth/kinship care compared to those in F/A care and non-exposed young adults. While the absence of self-reported use in the nonkinship care group is intriguing, numbers are too small to draw conclusions. Additionally, the non-exposed comparison group was formed at birth from a similar high-risk group of pregnant women, most of whom were positive for alcohol use and drugs other than cocaine and who were of similar low education and socioeconomic status, factors which may have served as more powerful determinants of behavioral outcomes than prenatal exposure to cocaine, alcohol or other drugs. In this cohort at earlier ages, prenatal cocaine exposure was related to externalizing behavior problems, particularly among girls, and cocaine exposed children in nonkinship foster or adoptive care were rated by their caregivers as having more externalizing behaviors than those in birth family/kinship care (Linares et al., 2006; McLaughlin et al., 2011), suggesting that the positive relationship of non kinship placement in a more enriched environment to higher educational attainment did not transfer to behavioral outcomes as measured by engagement with the justice system or problematic substance use.

A recent meta-analysis concluded that children in kinship care may perform better than children in traditional foster care in domains of behavioral development, mental health, and placement stability while those in non-kinship foster care as in this study may be more likely to achieve adoption and access services (Winokur et al., 2014). Placement instability is a common occurrence for children in foster care that has been related to more negative outcomes (Fisher et al., 2012, 2013; Bada et al., 2008). We were unable to accurately determine the extent of placement instability for this cohort at age 21. In an earlier study at 15 years, more than 65% of adolescents with PCE placed in non-kinship care in this sample were found to have 1–2 placement disruptions, and the remainder more than 2, but the reasons for and the nature of the changes are unclear (Min et al, 2014). Placement disruption may have been influenced by the higher lifetime rate of sexual victimization noted for this group at age 17.

Over 400,000 children annually in the U. S. are placed in foster care, with about a third in kinship care (Child Welfare Information Gateway, 2021). Our study highlights the absence of information on long term outcomes for prenatally drug exposed children placed in foster or adoptive care despite the wide use of such placement as an intervention for drug or alcohol exposed infants. There are multiple reasons for the lack of information. State or county agencies may not be cooperative with research and in some cases may even forbid research engagement of children placed in state custody. Foster families planning to adopt infants may want privacy regarding the infant’s birth history or have concerns about birth family engagement. To our knowledge, there are no other studies addressing the relationship of nonkinship foster care to adult outcomes in a drug exposed cohort.

Strengths of the study include its longitudinal prospective design, accuracy of prenatal drug exposure, the numerous assessments of family and environmental variables, the use of cognitive assessments that were not based on family or self-report, and acceptable retention rates despite the attrition and difficulty of following children in foster or adoptive care. Despite these problems, our sample size met a threshold acceptable statistically (n> 30) and sufficient to detect differences. Outcomes related to substance use behaviors, educational achievement, and incarceration were based on self-report and may be less reliable than the cognitive measures used. Our data reflect the complexity of disentangling the effects of teratologic exposures on long term outcomes across a variety of domains and the need for studies of children in the foster care system.

Highlights.

At 21 years, prenatal cocaine exposure was associated with poorer perceptual reasoning, independent of environmental factors.

Young adults with prenatal cocaine exposure in nonkinship care from early childhood were more likely to graduate from high school, had better verbal skills, lower lead levels and more stimulating home environments than those in birth or kinship care.