Fatty Acid Ethyl Esters in Meconium and Substance Use in Adolescence

Abstract

Prenatal alcohol exposure (PAE) continues to be a serious public health problem, yet no reliable clinical tools are available for assessing levels of drinking during pregnancy. Fatty acid ethyl esters (FAEEs), the nonoxidative metabolites of ethanol measured in meconium, are potential biomarkers to quantify the level of PAE. The association between the concentrations of FAEEs from meconium and adolescent substance use and related problems was examined in a prospective birth-cohort of adolescents exposed to alcohol and drugs in utero. FAEEs were quantified with gas chromatography via a flame ionization detector. Meconium was analyzed for FAEEs in 216 newborns; 183 of them (81 boys, 102 girls) were assessed at age 15 for alcohol, tobacco, and marijuana use using biologic assays and self-report. Substance use problems were assessed using the Problem Oriented Screening Instrument for Teenagers. Findings from multivariable logistic regression analyses indicated that, after controlling for other prenatal drug exposure and covariates, higher concentrations of FAEEs (ethyl myristate, ethyl palmitate, ethyl oleate, ethyl linoleate, ethyl linolenate, and ethyl arachidonate) were related to a greater likelihood of marijuana use and experiencing substance use problems, but not tobacco or alcohol use, at age 15. Elevated levels of FAEEs in meconium may be promising markers for PAE, identifying newborns at risk for early substance use and developing substance use problems.

1. Introduction

Prenatal alcohol exposure (PAE) from maternal drinking during pregnancy continues to be a serious public health problem. It can result in lifelong physical and neurological impairments due to its effects on the developing fetal brain, collectively known as Fetal Alcohol Spectrum Disorders (FASD). Despite no known safe amount of alcohol use in pregnancy (Mukherjee et al., 2005), the most recent surveillance survey in 2015–2017 (Denny et al., 2019) found that about 1 in 9 US women (11.5%) reported drinking any alcohol while pregnant, with one third of those drinkers (3.9%) reporting binge drinking (≥ 4 drinks in one occasion). Both measures increased from the previous 2011–2013 survey, from 10.2% and 3.1% respectively.

FASD encompasses a wide range of developmental disabilities. The most severe form, called Fetal Alcohol Syndrome (FAS), is characterized by growth restriction, a distinct pattern of facial features, and central nervous system dysfunction (Hoyme et al., 2016), which are often recognized by teachers. FASD also includes a milder, more subtle expression of PAE, called alcohol-related neurodevelopmental disorder (ARND), which involves various neurologic abnormalities, in the absence of the classic facial dysmorphology and growth deficiency. These abnormalities include problems with memory, learning, attention, executive function, and self-regulation (Mattson et al., 2011), all of which are risk factors commonly associated with adolescent substance use (Huizink & Mulder, 2006).

The estimated prevalence of PAE varies depending on surveillance methods. Although approximately 40,000 (~ 1%) newborns are thought to be affected by FASD each year in the US (May & Gossage, 2001), a recent study indicated that it may be more prevalent than previously estimated, reporting as many as 3–10% of first-graders met criteria for FASD (May et al., 2018). Further, the prevalence of FASD is even greater in certain subpopulations. For instance, FASD is estimated to be 17% among children in the child welfare system (Lange et al., 2013). In the Western Cape Providence of South Africa, the prevalence of FASD is estimated to range from 19.6% to 27.6%, the highest rate globally (May et al., 2017).

Early identification of infants at risk for FASD is critical to reduce secondary disabilities, yet the detection of FASD remains challenging, as no reliable clinical tools for determining PAE have been established. Self-report measures are the most commonly used methods to assess PAE. Although structured in-depth interviews may elicit reliable information (Jacobson et al., 2002), the nature of self-report is likely to underestimate true levels of alcohol use due to the stigma and shame associated with prenatal drinking. Given the difficulty of identifying infants with ARND, a biological marker for PAE would allow for earlier identification and intervention for affected infants and their families, thereby improving prevention efforts to reduce PAE-related secondary disabilities.

Fatty acid ethyl esters (FAEEs) are the non-oxidative metabolites of ethanol analyzed in meconium. Meconium, an infant’s first stool, can detect PAE incurred as early as 12 weeks of gestation (Burd & Hofer, 2008). This non-invasive detection method allows for the identification of moderate and episodic PAE. FAEEs accumulate in meconium as a result of ethanol metabolism by the fetus and have been investigated as biomarkers for identifying alcohol-exposed neonates (Burd & Hofer, 2008). We previously established that increased concentrations of FAEEs in meconium were correlated with maternal alcohol consumption during pregnancy in a high-risk urban population (Bearer et al., 1999, 2003, 2005). Further, high concentrations of FAEEs in meconium were associated with poorer mental and psychomotor developmental scores on the Bayley Scales of Infant Development at 6.5 months, 1 year, and 2 years of age (Peterson et al., 2008) and with poorer verbal comprehension, working memory, and Full-Scale IQs on the Wechsler Intelligence Scales for Children at ages 9, 11, and 15 years (Min et al., 2015). The purpose of the present study is to extend our previous findings to examine whether the concentration of FAEEs in infant meconium is related to early adolescent substance use occasionally reported in studies of children with PAE (Dodge et al., 2019). To our knowledge, no studies have examined the association between FAEEs in meconium and adolescent substance use. Variables that may confound the association between PAE and adolescent substance use, such as prenatal exposure to tobacco, marijuana, and cocaine, ongoing caregiver substance use and psychological distress (Minnes et al., 2014), exposure to violence (Frank et al., 2011; Kobulsky et al., 2016), inadequate parental monitoring (Min et al., 2014), and out-of-home placement (Linares et al., 2006) were considered in the analyses. Given our prior findings of a significant negative association between FAEEs and cognitive development, we also assessed the association of adolescent cognitive development with substance use. We hypothesized that higher concentrations of FAEEs in meconium would be related to adolescent substance use at age 15.

2. Methods

2.1. Sample and procedure

The present study included 183 adolescents (81 male, 102 female) recruited at birth from a metropolitan teaching hospital for a longitudinal study on the developmental effects of prenatal cocaine exposure (Singer et al., 2015). Pregnant women at high risk for drug use, determined by behavior suggesting intoxication, self-admitted substance use, a lack of prenatal care, or a history of involvement with the Department of Human Services, underwent drug toxicology screenings at delivery. Women with a HIV-positive status, chronic medical illness, a psychiatric history, or a diagnosis of intellectual disability, were excluded, as were infants with Down syndrome, FAS, or congenital heart defects. After informed consent, random samples of meconium were obtained from 248 newborns, and 216 had adequate analysis of meconium (≥ 0.5 g meconium available and ≥ 50% recovery of internal standard; Bearer et al., 2005). Of the 216 children, 14 had missing prenatal substance use interview data, 2 children died, and 17 dropped out or were lost to contact. The current study utilizes data from 183 children who completed a substance use assessment at age 15, which represents an 85% retention rate. Children and their caregivers were seen at the developmental research laboratory by separate examiners who were blinded to both the mother and child’s alcohol and drug exposure status. Children were assessed by a clinical psychologist or master’s level research assistant, and caregivers were assessed by a social worker or trained research assistant. Parental written informed consent and child assent were obtained prior to data collection. All participants were compensated with a monetary stipend, lunch, and/or transportation costs for their time. A Certificate of Confidentiality (DA-98–91) was obtained from the Department of Health and Human Services to protect against release of confidential health information from women participating in the study. The Institutional Review Board of the participating hospital approved the study.

2.2. Measures

Meconium was collected within 24 hours after birth and frozen at −70°C until analysis. FAEEs were extracted with acetone/hexane and isolated using silica gel chromatography. The isolated FAEEs were identified and quantitated by gas chromatography using a flame ionization detector (GC/FID) (See Bearer et al., 1999, 2005 for further details). Six FAEE analytes were examined: ethyl myristate, ethyl palmitate, ethyl oleate, ethyl linoleate, ethyl linolenate, and ethyl arachidonate.

At the newborn visit, birth mothers were asked to recall the frequency and amount of alcohol and drug use for the month prior to and for each trimester of pregnancy. The number of standard drinks (0.5 oz. of absolute alcohol) of beer, wine, or hard liquor per drinking day was computed. The number of drinking days per week was recorded using a Likert-type scale ranging from 1 (less than once a month) to 7 (daily use). Number of drinks per week was calculated by multiplying the number of standard drinks per drinking day with the number of drinking days per week. Risk drinking was assessed via the TWEAK (Russell, 1994). A total score of TWEAK ≥ 2 indicates risk drinking (Chan et al., 1993). Other substance use during pregnancy, such as the number of tobacco cigarettes smoked, marijuana joints smoked, and crack cocaine “rocks” consumed and the amount of money spent per day for crack cocaine, was also collected along with the frequency of use. The alcohol and drug assessment was updated with the adolescent’s current caregiver at each follow-up visit to obtain an assessment of recent (prior 30 day period) postnatal, caregiver alcohol and drug use.

Demographic and medical characteristics, including maternal age at birth, gestational age, birth weight and length, head circumference, and APGAR scores were extracted from hospital medical records. At the newborn visit, socioeconomic status (SES) was assessed via the Hollingshead Index (Hollingshead, 1957). Maternal psychological distress was assessed using the Global Severity Index (α = .95) from the Brief Symptom Inventory (Derogatis, 1992) at each visit. Maternal receptive vocabulary was assessed using the Peabody Picture Vocabulary Test-Revised (PPVT-R) (Dunn & Dunn, 1981) at the newborn visit and updated using its third edition (PPVT-III) (Dunn et al., 1997) at later assessments. The child’s placement (with either biological mother/relative or non-kinship adoptive/foster caregiver) was also recorded at each visit, and data on the current caregiver were updated to provide concurrent assessment of caregiver intelligence, psychological distress and substance use. At 12-year visit, parental monitoring (α =.74) and lifetime frequency of exposure to violence, either as a direct victim or witness (α=.76), were assessed using the Assessment of Liability and Exposure to Substance Use and Antisocial Behavior, an illustration-based, computer-assisted self-interview for children ages 9–12 (Ridenour et al., 2009).

At 15 years, adolescent substance use was assessed using self-report and biologic assays for drug metabolites. The Youth Risk Behavior Surveillance System (Centers for Disease Control and Prevention, 2009) was used to collect self-reported alcohol, tobacco, and marijuana use. Samples of participants’ hair, urine, and/or bloodspots were collected by research nurses from the university’s NIH-funded Clinical Research Unit and sent to the United States Drug Testing Laboratory for analysis (See Minnes et al., 2014 for a complete description of adolescent substance use assessment). Respondents’ positive result on either self-report or biologic assays for a particular drug were coded 1 (yes) for that drug. Substance use related problems (e.g., getting into trouble, drunk driving, mood swings) were assessed with the 17-item Substance Use and Abuse Scale from the Problem Oriented Screening Instrument for Teenagers (Radhert, 1991). Due to skewed distribution, respondents endorsing ≥ 1 problem(s) were coded 1 (yes). The Wechsler Intelligence Scales for Children-Fourth Edition (WISC-IV) was used to assess adolescent’s intelligence (Wechsler, 2003) at 15 years. The Home Observation for Measurement of the Environment (HOME; Caldwell & Bradley, 2003) was used to assess the quality of the caregiving environment in an interview format with caregivers at age 15.

2.3. Statistical analyses

Study variables positively skewed were normalized using a log transformation prior to analyses. Means and standard deviations (SD) were presented by the variables’ original distribution, although transformed data were used in multivariable analyses. Concentrations of each FAEE (ng/g) were transformed by log₁₀ (FAEE+100) to correct skewed distribution. A constant value of 100 was added, so that the value of the cases below the limit of detection at 100 could be log transformed. Multivariable logistic regressions were conducted to evaluate the relationship of each FAEE with adolescent substance use and substance use related problems. In order to avoid multicollinearity and saturation of the model, confounding variables correlated with FAEE analytes or with the outcomes at p ≤ .20 were entered into the model using backward elimination approach using the criterion p < .10. In the backward elimination approach, the analysis begins with all covariates; each variable is treated as if it was entered last in the equation; in each step, a covariate that fails to meet the selection criterion (selected p < .10 for our study) is eliminated until no further variables can be deleted without significant loss of model fit (Warner 2013). Separate logistic regressions were conducted for each FAEE analyte, as our previous study demonstrated no better sensitivity/specificity for a linear combination of all FAEEs than individual FAEE analyte (Bearer et al., 2005).

3. Results

The majority of the 183 mothers and adolescent were African American and of low SES (Table 1). Only 12% of mothers were married at the child’s birth. About one-third (38%, n=69) had not finished high school, with a mean number of education years of 11.8 (SD=1.5). Of the 183 birth mothers, 108 (59%) reported alcohol use during pregnancy, with 6.6 (SD=12.2) alcohol drinks per week, and 67 (37%) engaged in risk drinking (TWEAK ≥ 2). Twenty-five (14%) mothers reported ≥ 7 drinks per week and 14 mothers reported ≥ 14 drinks per week, indicating that most of the women in this study were low-to-moderate drinkers. More than half of the mothers (n=111, 61%) smoked cigarettes, 44 (24%) used marijuana, and 87 (48%) used crack cocaine during pregnancy. In terms of birth outcomes (Table 2), the mean gestational age was 38 (SD=3) weeks with a mean birth weight of 3002 (SD=688) grams. About 49% of the children were placed with non-kinship foster or adoptive parents some time in their lifetime, and 86% (n=147) reported receiving free lunch at school. At age 15, the mean Full Scale IQ was almost 1 SD below the normative mean (mean = 87, SD = 13). Approximately one-third of the offspring used tobacco, alcohol, or marijuana by age 15, and 22% reported experiencing substance use related problems.

Table 1.

Study Population Demographics: Biologic Mother and Caregiver at Subject Age of 15 (N = 183)

	n (%)/ Mean ± SD	Median (10% – 90%)
Characteristics of the biological mother
African American	148 (80.9)
Low socioeconomic status	183 (100)
Married	22 (12.0)
Age at delivery	27.30 ± 5.32	26 (21 – 35)
Years of education	11.76 ± 1.48	12 (10 – 14)
Number of prenatal visits	7.49 ± 4.84	8 (1 – 14)
Global Severity Index	0.68 ± 0.70	0.45 (0.08 – 1.55)
PPVT Standard Score	76.96 ± 13.37	77 (60 – 93)
Measures of drinking during pregnancy (n=108)a
Alcohol drinks on drinking day	2.15 ± 2.85	1.5 (0.25 – 5.2)
Alcohol drinking days per week	1.37 ± 1.50	0.75 (0.06 – 3.5)
Alcohol drinks per week	6.55 ± 12.17	1.98 (0.06 – 18)
Risk drinking (TWEAK ≥ 2)	67 (62.0)
Other substance use during pregnancya
Cigarettes per day (n = 111)	11.59 ± 10.80	10.5 (1.2 – 20)
Marijuana joints per week (n = 44)	2.96 ± 4.36	1.1 (0.1 – 7)
Cocaine units per week (n = 87)	24.71 ± 51.83	5 (0.4 – 54)
Characteristics of the caregiver at subject age 15
HOME	43.57 ± 6.60	44 (35 – 51)
Global Severity Index	0.38 ± 0.45	0.21 (0.02 – 0.85)
PPVT Standard Score	79.93 ± 14.26	79 (63 – 97.25)
Substance use past 30 daysab
Alcohol drinks per week (n = 66)	3.26 ± 3.71	2.25 (0.5–9)
Cigarettes per day (n = 85)	10.93 ± 8.31	10 (2–20)
Marijuana joints per week (n = 6)	8.58 ± 10.73	4.25 (0.25–28)

PPVT, Peabody Picture Vocabulary Test; TWEAK, Tolerance, Worried, Eye-openers, Amnesia, and K/Cut down; HOME, Home Observation for Measurement of the Environment

^aBased on users (n) only

^bOnly one caregiver reported cocaine use in the past 30 days.

Table 2.

Study Population Demographics: Adolescent Characteristics (N = 183)

	n (%)/ Mean ± SD	Median (10% – 90%)
Characteristics assessed at birth
Male	81 (44.3)
Gestational age, weeks	38.3 ± 3.0	39 (35 – 41)
Prematurity (< 37 weeks gestational age)	39 (21.3)
Birth weight, grams	3,002 ± 688	3,135 (2,000 – 3,735)
Boysa	3,044 (46.3)
Girlsa	2,970 (41.2)
Birth length, cm	48.8 ± 4.0	49.0 (44.4 – 53.0)
Boysa	49.2 (0.30)
Girlsa	48.5 (0.27)
Head circumference, cm	33.1 ± 2.4	33.0 (30.2 – 35.5)
Boysa	33.3 (0.19)
Girlsa	33.0 (0.16)
APGAR score – 1 min < 7	23 (12.6)
APGAR score – 5 min < 7	3 (1.7)
FAEE levels
Ethyl myristate	582 ± 1,746	48 (0 – 1,535)
Ethyl palmitate	993 ± 2,390	134 (32 – 2,687)
Ethyl oleate	12,412 ± 38,026	270 (55 – 34,074)
Ethyl linoleate	19,036 ± 65,369	241 (0 – 42,642)
Ethyl linolenate	5,412 ± 17,649	129 (0 – 18,150)
Ethyl archidonate	754.4 ± 1,516	179 (0 – 2,246)
Characteristics assessed at adolescence
Age at assessment	15.67 (0.28)
Parental monitoringb	2.45 ± 0.57	2.6 (1.75 – 3.0)
Violence exposurec	0.63 ± 0.79	0.25 (0 – 1.88)
WISC-IV Full Scale IQ	86.7 ± 13.12	86 (71 – 106)
Receiving free lunch at school	147 (86.0)
Always in birth parents’ care	91 (51.1)
Substance use
Alcohol	69 (38.8)
Using alcohol more than 2 days in lifetime	8 (4.2)
Using alcohol more than 2 days in last 30 days	0 (0)
Tobacco	60 (33.7)
Using cigarette more than 1 per day in last 30 daysd	2 (1.1)
Marijuana	56 (31.5)
Using marijuana more than 2 times in lifetime	19 (10.3)
Using marijuana more than 2 times in last 30 days	9 (5.0)
Substance related problem	39 (22.3)

WISC-IV, Wechsler Intelligence Scales for Children-Fourth Edition

^aGestational age adjusted mean (standard error)

^bYouth-perceived parents’ awareness of the youth’s activities and whereabouts, 0 = none of the time to 3 = all of the times

^cLifetime frequency, 1 = none to 5 = 5 times or more

^dYRBSS does not ask life time use of tobacco

Table 3 summarizes the relationships of the six FAEE analytes (ethyl myristate, ethyl palmitate, ethyl oleate, ethyl linoleate, ethyl linolenate, and ethyl arachidonate) with adolescent substance use outcomes after adjusting for other prenatal drug exposure and covariates. Higher levels of FAEEs were related to a greater likelihood of marijuana use and experiencing substance use problems. For example, after adjusting for other prenatal drug exposure and covariates, each 10-fold increase of ethyl myristate was related to a 120% increase in marijuana use by age 15 (OR = 2.20, 95% CI = 1.04 – 4.67) and to a 243% increase in experiencing substance use problems (OR = 3.43, 95% CI = 1.42 – 8.24). No significant relationship (p > .10) was observed between FAEEs and adolescent tobacco or alcohol use (Data not shown). Boys were more likely to report substance use related problems (OR = 2.64, 95% CI = 1.08 – 6.48, in the model with ethyl myristate as a predictor, for example), despite no gender differences found in alcohol, tobacco, or marijuana use. Parental monitoring was inversely related to a likelihood of substance use related problems (OR = 0.40, 95% CI = 0.19 – 0.83, in the model with ethyl myristate as a predictor). Prenatal cocaine exposure was independently related to a higher likelihood of marijuana use (OR = 2.51, 95% CI = 1.21 – 5.20) and substance use problems (OR = 3.48, 95% CI = 1.40 – 8.68), consistent with our previous studies with a larger sample size, which included children who did not have measures of FAEEs (Min et al., 2014; Minnes et al., 2014). Adolescent IQ was not related to adolescent substance use or related problems.

Table 3.

Adjusted Association of Fatty Acid Ethyl Esters with Substance Use in Adolescent Offspring

	Marijuana Usea		Substance Use Problemsb
	OR (95% CI)	P	OR (95% CI)	P
Ethyl myristate	2.20 (1.04 – 4.67)	.039	3.43 (1.42 – 8.24)	.006
Ethyl palmitate	1.97 (0.98 – 3.96)	.057	3.28 (1.43 – 7.51)	.005
Ethyl oleate	1.67 (1.02 – 2.52)	.015	1.91 (1.17 – 3.11)	.009
Ethyl linoleate	1.62 (1.11 – 2.36)	.011	1.85 (1.19 – 2.88)	.006
Ethyl linolenate	1.71 (1.07 – 2.72)	.025	1.95 (1.14 – 3.33)	.015
Ethyl archidonate	2.16 (1.10 – 4.27)	.026	2.47 (1.10 – 5.55)	.028

OR = odds ratio; Variables examined for covariates but failed to meet the selection criteria in logistic regression modeling include maternal education, prenatal tobacco exposure, prenatal marijuana exposure, offspring sex and race, receiving free lunch, and non-kinship foster/adoptive care.

Significant (p < .05) covariates are listed in italics.

^aAdjusted for prenatal cocaine exposure, birth mother’s psychological distress at birth, parental monitoring, and violence exposure

^bAdjusted for adolescent sex, prenatal cocaine exposure, birth mother’s psychological distress at birth, HOME score, adolescent full scale IQ, parental monitoring, and violence exposure

4. Discussion

The present study demonstrated that higher concentrations of FAEEs in meconium were associated with increased likelihood of marijuana use and substance use related problems at age 15 in low SES, primarily African-American, urban adolescents with prenatal poly-drug exposure. This is the first study to document the relationship between FAEEs in the newborn period and substance use in adolescence. Our findings are consistent with previous studies that relied on maternal report of prenatal alcohol use and found relationships with later adolescent substance use (Dodge et al., 2019; O’Brien & Hill, 2014). In a birth-cohort study of 917 mother-offspring dyads, maternal drinking of ≥ 1 drink per day in the first trimester was associated with higher levels of drinking (> 1 drink per week on average for the past year) among their children at age 16 years (Cornelius et al., 2016). Similarly, in a German study using a representative sample of adolescents ages 11 to 17 (Pfinder et al., 2014), retrospective maternal self-report of alcohol intake during pregnancy was related to increased risk in offspring of using illicit drugs, including marijuana, and regular drinking (defined as at least once a week). Our study did not observe a specific relationship between FAEEs and drinking in offspring, probably due to the relatively lower alcohol consumption in African American adolescents than in White counterparts (Duncan et al., 2012). No adolescent reported more than 2 days of alcohol use in the past 30 days, although more adolescents reported lifetime alcohol use than tobacco or marijuana (Table 2). Constricted ranges of PAE alcohol exposure, due to the selection criteria of the parent study, might also limit the observation of the effects of FAEEs to other outcomes. Replication studies that include children diagnosed with FAS in a more demographically diverse sample may ascertain the relationship between FAEEs and substance use in offspring. Nevertheless, our study provides additional evidence of a significant prospective association between FAEEs and substance use in adolescence.

Limitations of our study should be considered. For FAEE determination, the use of the GC/FID is less sensitive than tandem mass spectroscopy and may produce false negative detection, or the concentrations measured could be inflated (Burd & Hofer, 2008). Maternal prenatal tobacco and marijuana use relied on retrospective self-report after delivery, subject to recall error and social desirability bias. The lack of data on behavioral interventions that these children may have received might also obscure the effects of FAEE on substance use outcomes. Since meconium tends to be formed late in the second and third trimesters of pregnancy, FAEEs in meconium do not capture PAE spanning the full gestational period, failing to identify infants of mothers who were able to quit alcohol consumption early in pregnancy. By analyzing each FAEE analyte separately, we might increase the risk of Type I error. Still, our study also has multiple strengths including the prospective design, following children since birth with 85% of the original surviving cohort assessed at age 15, use of biological measures to quantify PAE as well as adolescent substance use, and a large number of potential confounders evaluated including other prenatal substance exposures.

In conclusion, elevated levels of FAEEs in meconium can be promising markers for identifying newborns at risk for early substance use and developing substance use problems. Our study supports the validity of FAEEs as markers of PAE, adding to a convergent body of evidence to increase confidence in the validity of FAEEs. Continued follow-up with this sample will elucidate whether and how PAE, operationalized as FAEEs from meconium, may affect the chronicity or severity of substance use problems and the development of substance use disorder in adulthood.

Highlights.

• Prenatal alcohol exposure (PAE) continues to be a serious public health problem

• Fatty acid ethyl esters (FAEEs) are potential biomarkers to quantify PAE

• FAEEs in the newborn period may be related to marijuana use in adolescence at age 15

• FAEEs in the newborn period may be related to substance use problems at age 15