Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jul 1.
Published in final edited form as: Neurotoxicol Teratol. 2021 Jun 8;86:107000. doi: 10.1016/j.ntt.2021.107000

Infant Neurodevelopmental Outcomes of Prenatal Opioid Exposure and Polysubstance Use

Madelyn H Labella 1,*, Rina D Eiden 2, Alexandra R Tabachnick 1, Tabitha Sellers 1, Mary Dozier 1
PMCID: PMC8277730  NIHMSID: NIHMS1716079  PMID: 34116198

Abstract

Background:

Prenatal opioid exposure has been linked to adverse birth outcomes and delays in infant development. Existing literature is limited by a simple group-differences approach as well as inadequate controls for sociodemographic factors and polysubstance exposure co-occurring with prenatal opioid use.

Method:

The current study assessed cumulative opioid exposure (duration of prescribed and illicit opioid use) as a predictor of infant birth outcomes and mother-reported developmental status at three and six months of age, controlling for polysubstance exposure. Participants were predominantly low-income pregnant and peripartum women, oversampled for mothers receiving medication-assisted treatment (MAT) for opioid use disorder. Prenatal opioid and non-opioid substance use were reported by mothers using a Timeline Follow-Back Interview completed during the third trimester and updated postnatally (infant age six months).

Results:

Developmental scores were in the normal range. However, total opioid exposure was positively related to premature birth and inversely related to mother-reported developmental status in specific domains. Associations with three-month fine motor skills and six-month communication skills were robust to controls for polysubstance exposure and sociodemographic covariates.

Conclusions:

Results suggest unique effects of prenatal opioid exposure on the early development of fine motor and communication skills. Similar findings were obtained for prescribed and illicit opioid use, underscoring developmental risks of both MAT and untreated substance use. Exploratory analyses investigating type and timing of MAT suggest directions for future research.

Keywords: prenatal opioid exposure, medication-assisted treatment, polysubstance use, Timeline Follow-Back Interview, infant development

1. Introduction

Prenatal opioid exposure is a public health problem of increasing urgency. Over the past two decades, diagnoses of opioid use disorder among pregnant women in the United States more than quadrupled (Haight et al., 2018), with an estimated one in five pregnant women consuming opioids during pregnancy (Patrick et al., 2015). Recommended treatment for pregnant women with opioid use disorder involves pharmacotherapy with a full or partial mu-opioid agonist (methadone and buprenorphine, respectively; Jones et al., 2011). Medication-assisted treatment (MAT) with an opioid agonist reduces fetal risks related to cycles of intoxication and withdrawal but is often associated with acute withdrawal symptoms in the neonatal period (Conradt et al., 2019; Konijnenberg & Melinger, 2011; Nelson et al., 2020).

1.1. Prenatal opioid exposure and infant development.

Prenatal opioid exposure has also been associated with poor fetal growth outcomes. Infants prenatally exposed to illicit opioids (Azuine et al., 2019; Nygaard et al., 2017) and/or MAT (Konijnenberg et al., 2001; Nelson et al., 2020) have lower average birth weights than non-exposed infants. Similarly, prenatal opioid exposure has been linked to shorter gestational age and higher risk for preterm birth among infants whose mothers used heroin (Azuine et al., 2019; Fajemirokun-Odudeyi et al., 2006) and/or MAT (Fajemirokun-Odudeyi et al., 2006; Lejeune et al., 2006; Nelson et al., 2020) than among infants without such exposure.

Although neonatal risks are well-established, less is known about neurodevelopmental outcomes of prenatal opioid exposure beyond the first few weeks of life. Whereas most infants exposed to opioids score within the normative range on standardized measures of neurodevelopment (Kaltenbach et al., 2018; Konijnberg et al., 2001), they perform worse than children from non-exposed comparison groups on measures of cognitive, motor, and behavioral functioning across early development (Konijnenberg & Melinder, 2011; Maguire et al., 2016). Three recent meta-analyses concluded that young children exposed to illicit opioids and/or MAT show worse cognitive, language, and motor skills than children with no prenatal substance exposure; pooled effect sizes were in the moderate range (d = 0.49–0.77; Andersen et al., 2020; Lee et al., 2020; Yeoh et al., 2019).

Importantly, the research reflected in these meta-analyses is limited by small sample size and inadequate consideration of confounding variables. Compared to other pregnant women, women using opioids prenatally report lower socioeconomic status and related risk factors, including inadequate nutrition and disruptions in maternal care (Andersen et al., 2020; Azuine et al., 2019; Singer et al., 2020). Differences in rearing environments may partially account for differences in cognitive and motor development, particularly given evidence that opioid exposure may confer increased vulnerability to environmental disadvantage (Konijnenberg & Melinder, 2011). Indeed, a recent review concluded that the majority of studies controlling appropriately for environmental confounders found no differences in infant neurodevelopment as a function of prenatal opioid exposure (Conradt et al., 2019).

In addition to sociodemographic differences, pregnant women using opioids also use more non-opioid substances – including tobacco, alcohol, marijuana, and cocaine – than pregnant women who do not use opioids (Azuine et al., 2019; Nelson et al., 2020; Singer et al., 2020). Heterogeneity in non-opioid substance use (i.e., polysubstance use) is of particular relevance as the use of tobacco, alcohol, and illicit drugs in combination with opioids may alter or enhance teratogenic potential (Boggess & Risher, 2020). Tobacco and alcohol use may be particularly important co-factors influencing neurodevelopment in the context of prenatal opioid use, given their well-established adverse effects on child development (Guille & Aujla, 2019) and relatively high prevalence among pregnant mothers with opioid use disorder: an estimated 85% to 95% of pregnant mothers receiving MAT report regular tobacco use (Azuine et al., 2019; Akerman et al., 2015; Chisholm et al., 2014). One meta-analysis found that MAT was not significantly associated with cognitive development in a subset of studies that matched MAT and comparison groups on prenatal tobacco exposure (Nelson et al., 2020). This null finding highlights the importance of appropriately controlling for non-opioid substance exposure.

Discrepant findings may also reflect differences in timing of assessment, which may obscure dynamic changes in opioid-related risk across early development. For example, one early longitudinal study found that, compared to non-exposed infants, infants exposed to opioids showed worse motor coordination at four months of age, but not eight months or twelve months (Marcus et al., 1989). Another longitudinal study found that neurological differences were evident at four weeks but not one year, with decrements in cognition and language skills emerging at two years (van Baar et al., 1994). Additional longitudinal research is needed to more precisely delineate effects of opioid exposure across developmental time.

1.2. Limitations of the current literature.

Existing literature is thus limited by a preponderance of low sample size studies with inadequate controls for sociodemographic differences and polysubstance use and provides insufficient information regarding longitudinal changes in opioid-related risk. Furthermore, research on prenatal exposure often combines multiple types of opioids, complicating efforts to isolate effects of prescribed versus illicit use (Guille & Aujla, 2019). A recent meta-analysis found that infants exposed to opioids, whether prescribed or illicit, showed similar decrements in cognitive performance compared to non-exposed infants (Lee et al., 2020). Importantly, however, many mothers receiving MAT report illicit opioid use during pregnancy prior to the initiation of treatment and/or in addition to prescribed medications, raising the possibility that illicit opioid use rather than MAT is responsible for any neurodevelopmental effects (Hans & Jeremy, 2001; Jones et al., 2015; Konijnenberg & Melinder, 2011)

In addition to potential differences in outcomes associated with illicit versus prescribed opioids, there may be further differences in outcomes by type of MAT. A growing body of evidence suggests that neonatal abstinence syndrome is less severe among infants exposed to buprenorphine than among infants exposed to methadone (Coyle et al., 2012; Wiegand et al., 2015), Similarly, infants of mothers maintained on buprenorphine may have lower risk for premature birth and higher birth weight compared to those on methadone (Zedler et al., 2016). Although a recent randomized controlled trial found few differences in cognitive or behavioral outcomes associated with prenatal exposure to methadone versus buprenorphine (Kaltenbach et al., 2018), more research is needed to characterize effects of methadone, buprenorphine, and illicit opioids on infant neurodevelopment.

The literature’s reliance on simple group comparisons also obscures differences in amount of opioid exposure, reducing continuous variation in exposure to a simple binary variable (Conradt et al., 2019). Evidence suggests that milligram dosage of MAT is not a significant predictor of perinatal outcomes (Cleary et al., 2010; Jones et al., 2014); however, little is known about the predictive power of duration of opioid exposure. A group-differences approach also fails to account for variation in timing of opioid exposure during pregnancy. Failure to examine timing is a substantial limitation as vulnerability to teratogens varies across fetal development, peaking during periods of rapid cell differentiation and tissue specialization. For example, Konijnenberg and Melinder (2011) propose that prenatal opioid exposure may have greatest impact during the rapid brain growth that occurs around 24 weeks gestation.

1.3. The current study.

The current study addresses existing gaps in the literature by investigating the effect of continuous variation in total opioid exposure on infant birth outcomes and mother-reported developmental status at two different timepoints in early development, controlling for sociodemographic factors and polysubstance exposure. We hypothesized that more opioid exposure (i.e., longer duration of prescribed and illicit opioid use) would be uniquely associated with more adverse birth outcomes (i.e., shorter gestational age, more premature birth, and lower birth weight) as well as worse cognitive, motor, and behavioral development reported by mothers at three months and six months. Follow-up tests were planned to probe significant effects of total opioid exposure by testing unique contributions of prescribed and illicit opioid use. Additional exploratory analyses examining type and timing of MAT exposure were planned. We hypothesized that both prescribed and non-prescribed prenatal opioid exposure would be associated with adverse infant outcomes. Regarding exploratory analyses, we anticipated that methadone would be associated with more deleterious outcomes than buprenorphine. Given lack of prior research, we did not formulate hypotheses about timing of exposure.

2. Methods

2.1. Participants.

Participants were 82 mothers of 85 infants; one mother had twins and two mothers were pregnant twice during active data collection, resulting in 85 births. Women were enrolled during the third trimester of pregnancy (after 28 weeks’ gestation; n = 70 out of 84 pregnancies) or up to one month postpartum (n = 14). Sixty-two mothers of 65 infants were enrolled in a randomized controlled trial evaluating an attachment-based parenting program adapted for mothers on MAT for opioid use during pregnancy (ClinicalTrials.gov, NCT03891628). Twenty mothers not on MAT during pregnancy were enrolled as part of a no-treatment comparison group and completed the same research assessments as mothers in the randomized trial.

Mothers in treatment for opioid dependence were recruited through local clinics offering MAT and through personal referrals from enrolled participants. Mothers in the MAT group reported prescribed use of methadone (n = 46) and/or buprenorphine (n = 20) during the target pregnancy, with four mothers reporting non-simultaneous use of both opioid treatment medications. Mothers in the non-MAT group were recruited using online advertisements, flyers in an obstetrics-gynecology clinic serving predominantly low-income patients, and personal referrals from enrolled participants. Comparison mothers indicated that they had not received treatment for opioid use during the target pregnancy but were not excluded on the basis of other prenatal substance exposure; 80% reported at least some substance use during pregnancy, including 10% who reported prenatal opioid use. Comparison mothers with more than a baccalaureate education were excluded to minimize sociodemographic differences from mothers with MAT exposure. Given the small number of opioid-naïve mothers (n = 18), the current study does not address simple group differences in infant outcomes. Rather, participants were combined into a single sample, oversampled for opioid exposure, with focal analyses examining infant outcomes as function of continuously measured prenatal opioid and polysubstance exposure.

Demographic characteristics are presented in Table 1. Mothers ranged in age from 18.8 to 40.8 years, with a mean age of 29.2 years (SD = 4.5). Mothers were predominantly White (63.4%) and low socioeconomic status; one-quarter (25.6%) reported less than a high school education and almost half (43.9%) had a household income below $10,000 a year. Infants were identified as White (37.2%), Multiracial (36.4%) or Black (26.9%), and half were identified as female (51.8%).

Table 1.

Mother and infant demographics.

Percentage
Maternal demographics
Race/Ethnicity
 Black/African American 25.6%
 White/Non-Hispanic 63.4%
 Multiracial 9.8%
 Other 1.2%
Education
 Less than High School 25.6%
 High School Degree/ GED 41.5%
 Some College 19.5%
 Certificate or Associate’s Degree 8.5%
 Baccalaureate Degree 4.9%
Household Income
 Less than $10,000 43.9%
 $10,000 – $19,999 13.4%
 $20,000 – $34,999 17.1%
 $35,000 – $49,999 8.5%
 $50,000 – $74,999 6.1%
 $75,000 – $99,999 0.0%
 More than $100,000 2.4%
  Missing 8.5%
Infant demographics
Race/Ethnicity
 Black/African American 26.9%
 White/Non-Hispanic 36.4%
 Multiracial 37.2%
 Other 1.3%
Sex (% Female) 51.8%
Open in a new tab

2.2. Procedures.

The University of Delaware Institutional Review Board approved all study procedures. After providing informed consent for themselves and their infants, mothers completed a baseline research assessment during their third trimester of pregnancy (after 28 weeks) or up to one month postpartum. At this time, mothers reported on demographics and substance use during the pregnancy to date, in addition to completing baseline physiological and parenting assessments that are beyond the scope of the current study. After completing baseline assessments, mothers on MAT were randomized to receive an attachment-based parenting intervention (modified Attachment and Biobehavioral Catch-up) or an active comparison intervention focused on healthy development (modified Developmental Education for Families). Parenting interventions were in progress at the time of the follow-up research assessments. All mother-infant dyads, regardless of MAT status, were invited to complete follow-up assessments when infants were three months and six months old, adjusted for prematurity. At follow-up research assessments, mothers reported on infant demographics, birth outcomes, and developmental status. At the six-month research assessment, mothers reported their substance use since the baseline research assessment.

2.3. Measures.

2.3.1. Maternal Substance Use.

Prenatal substance exposure was assessed at the baseline and six-month research assessments using the Timeline Follow-Back Interview (TLFB), a well-established calendar-based method of eliciting reports of substance use over time (Sobell et al., 1986). The TLFB has demonstrated good test-retest reliability and convergent validity with other intensive self-report and physiological measures of substance use (e.g., salivary cotinine; Brown et al., 1998; Carey et al., 2004; Eiden et al., 2013; Schuetze et al., 2008). The TLFB was used to elicit daily reports of prescribed and illicit opioid use, as well as use of non-opioid substances, during pregnancy. For mothers who completed the baseline assessment prenatally, reports of third trimester substance use were pooled from baseline and six-month interviews in order to capture substance use throughout the entire pregnancy.

Proportion of pregnancy days on MAT (defined as the number of days on MAT divided by the number of days in the pregnancy), proportion of pregnancy days on prescription opioid analgesics, and proportion of pregnancy days of illicit opioid use (defined as the sum of the number of days of heroin use, prescription opioid misuse, and opioid treatment misuse, divided by the number of days in the pregnancy) were summed to form a single measure of total opioid exposure. This index reflects cumulative variation in prenatal opioid use but is not interpretable as proportion of pregnancy due to some mothers’ use of prescribed and non-prescribed opioids on the same day. As such, the total opioid exposure score reflects variation in both duration (i.e., proportion of pregnancy days) and type (i.e., prescribed and illicit) of opioid exposure.

To account for prenatal exposure to non-opioid substances, we calculated average number of cigarettes per week, average number of standard drinks per week, and average number of binge-drinking days per week, with binge-drinking defined as consuming four or more standard drinks in one day. A polysubstance risk score was computed by tallying the number of additional non-opioid substances used during pregnancy, with a possible range of 0–9: marijuana, e-cigarettes, prescription stimulant misuse, prescription sedative misuse, cocaine, inhalants, hallucinogens, methamphetamines, and ecstasy.

2.3.2. Infant birth outcomes.

Infant birth outcomes included continuous measures of gestational age and birth weight. In addition, given the clinical significance of prematurity for infant outcomes, a dichotomous variable was used to capture premature birth before 37 weeks of gestation. Gestational age at birth was calculated in weeks using the difference between an infant’s estimated due date and actual date of birth as reported by mothers. Mothers reported infants’ birth weight at their first follow-up research assessment (typically three months).

Of note, two infants in the MAT group died shortly after birth. Both mothers had been prescribed methadone during their pregnancy, and one infant was born extremely preterm at 26 weeks gestation. Given the low base rate of infant death, this was not included as an outcome in our analyses.

2.3.3. Infant developmental status.

Mothers reported their infants’ developmental status at both the three-month and six-month assessment using the Ages and Stages Questionnaire-Third Edition (ASQ-3; Squires & Bricker, 2009). The ASQ-3 is a well-established parent-report measure of infant development in each of five domains: communication, fine motor skills, gross motor skills, problem-solving, and personal-social. Parents indicate infants’ mastery of milestones on a three-point scale (Yes, Sometimes, Not Yet). Items are scored and summed using standard procedures to yield domain scores that range from 0 to 60. ASQ-3 scores have demonstrated strong test-retest and inter-rater reliability, as well as concurrent validity with standardized measures of infant development (Squires et al., 2009). Each domain scale has an empirically derived cut-off indicating areas of clinical concern; however, continuous scale scores were selected as the primary outcome given limited statistical power and the value of understanding predictors of variation within the normal range.

At the three-month assessment, mothers completed ASQ-3: Four Months, which is validated as a screener for infants aged 3 months, 0 days to 4 months, 30 days. One mother completed the ASQ-3 early (at 2 months, 28 days) due to an error in recording the infant’s birth date. At the six-month assessment, mothers completed ASQ-3: Six Months, validated as a screener for infants aged 5 months, 0 days to 6 months, 30 days. Twelve mothers completed the ASQ-3 later than this age range (infant age seven to nine months) due to inability to schedule during the window of eligibility. Significance judgments and approximate effect sizes did not change when these infants were excluded; thus, all infants were included in focal analyses controlling for adjusted age in days to minimize data loss. Adjusted age is equivalent to postconceptional age for infants born before 37 weeks gestation.

2.3.4. Demographic covariates.

Mothers reported on their own age, racial/ethnic identity, educational attainment, and household income at the baseline research assessment. Maternal education was rated on a nine-point scale ranging from 1-Less than high school to 9-Doctorate degree; no mothers in the current sample had higher than a baccalaureate degree. Yearly household income was rated on a seven-point scale ranging from 1-Less than $10,000 to 7-More than $100,000. Education and income were treated as continuous variables for purposes of analyses. Mothers also reported on infant racial/ethnic identity and sex assigned at birth at follow-up research assessments.

2.4. Missing data.

Seventy dyads completed the three-month research assessment and 68 completed the six-month research assessment. Percent missing data ranged from 0% (maternal race, maternal age, maternal education, infant sex, and infant gestational age) to 20% (infant developmental status at six months). Seventy-six infants had ASQ data from at least one follow-up research assessment. For the remaining nine infants, ASQ data were missing due to infant death (n = 2), transfer of custody to another caregiver (n = 5), or difficulty scheduling follow-up assessments (n = 2). Compared to infants with parent reports of developmental status, those without ASQ data were more likely to be exposed to prenatal methadone than not (88.9% versus 52.6%, p = 0.038).

Mothers did not significantly differ on demographic variables, illicit opioid use, or polysubstance use. Because missingness was conditioned on variables included in the analysis, data were assumed to be missing at random and full-information maximum likelihood was used to generate unbiased parameter estimates (Rosseel, 2012).

2.5. Plan for analysis.

Bivariate associations were calculated linking prenatal substance exposure variables with infant birth outcomes and developmental status. To test the robustness of opioid effects on infant outcomes, significant associations with prenatal opioid exposure were followed up with multiple regressions controlling for non-opioid substance exposure (i.e., cigarettes per week, standard drinks per week, and polysubstance risk score), as well as adjusted age in days for ASQ outcomes only. In addition, maternal education, household income, child sex, and child race (White versus non-White) were evaluated as potential covariates for all outcomes, and gestational age and birth weight were evaluated as potential covariates for models predicting ASQ scores. Candidate demographic covariates were included on a second step of the multiple regression if they were significantly associated with the predicted outcome at the bivariate level.

Robust associations between prenatal opioid exposure and infant outcomes were further evaluated to delineate the relative contributions of prescribed versus illicit use. Focal outcomes were simultaneously regressed on proportion of pregnancy days on MAT and proportion of pregnancy days of illicit opioid use, retaining all non-opioid covariates. Wald tests for equality of coefficients were used to assess the statistical significance of differences in parameter estimates.

Finally, exploratory follow-up tests of significant MAT effects were planned to examine differential effects of methadone versus buprenorphine exposure across pregnancy and by trimester. Focal outcomes were regressed on proportion of pregnancy days on methadone and proportion of pregnancy days on buprenorphine, controlling for illicit opioid use and non-opioid covariates. Proportion of pregnancy variables were further decomposed into proportion of first, second, and third trimester days on methadone and buprenorphine, respectively, and tested in three trimester-specific models. Testing effects by trimester was not attempted for illicit opioid use due to relatively low base rate, particularly in the third trimester (n = 19). Due to high correlations among trimester-specific scores for methadone (r’s = 0.72–0.82, p’s <0.001), buprenorphine (r’s = 0.88–0.92, p’s <0.001), and illicit opioid use (r’s = 0.54–0.81, p’s <0.001), scores reflecting multiple trimesters of opioid exposure were not entered as simultaneous predictors. Wald tests for equality of coefficients were used to assess the statistical significance of differences in parameter estimates.

Analyses were completed in R (R Core Team, 2016). Logistic regressions predicting prematurity were conducted using the package ‘logistf,’ which employs Firth’s method to reduce bias associated with predicting rare events (Heinze et al., 2020). Linear regressions predicting continuous outcomes were modeled in a path analytic framework using the package ‘lavaan’ (Rosseel, 2012). Non-dichotomous variables were standardized prior to regression analyses to enhance interpretation. Maximum likelihood estimation with robust (Huber-White) standard errors was used to account for non-normality in prenatal substance exposure variables and full-informed maximum likelihood was using to account for missing data (Rosseel, 2012).

3. Results

3.1. Prenatal substance use.

Endorsement rates for prenatal substance exposure are presented in Table 2. Approximately three-quarters of infants were exposed to MAT medications and half were exposed to illicit opioids, including heroin (37.6%), opioid treatment medication misuse (11.8%), and pain prescription misuse (14.1%). Proportion of pregnancy days on MAT ranged from 0 to 100% (any: M = 62.0, SD = 40.8; methadone: M = 47.5, SD = 44.7; buprenorphine: M = 16.2, SD = 34.1). Seven mothers (all of whom were on MAT) took prescription opioid analgesics as prescribed during pregnancy. Proportion of pregnancy days using prescription opioid analgesics was included in the measure of total opioid exposure; however, this variable was highly skewed (M = 0.9, SD = 5.7, skewness = 8.4) and excluded from follow-up analyses. On average, illicit opioid use occurred on 11.8% of pregnancy days (SD = 21.8). Total opioid exposure scores ranged from 0 to 144 (M = 76.4, SD = 44.4); higher scores reflect more days and/or more types of exposure (i.e., both illicit and prescribed).

Table 2.

Percentage of infants exposed to opioid and non-opioid substances.

%
Prenatal opioid exposure
Medication-assisted treatment 76.5
 Methadone 56.5
 Buprenorphine 24.7
Prescription opioid analgesics 8.2
Illicit opioid use 51.8
 Heroin 37.6
 Opioid treatment prescription misuse 11.8
 Pain prescription misuse 14.1
Prenatal non-opioid exposure
Tobacco 87.1
Alcohol 35.3
Marijuana 42.4
Cocaine 24.7
E-cigarettes 7.1
Methamphetamines 4.7
Stimulant misuse 3.5
Sedative misuse 5.9
Hallucinogens 1.2
Other 3.5
Open in a new tab

Note. No mothers endorsed prenatal use of inhalants or ecstasy.

Regarding non-opioid substance exposure, 87.1% of infants were prenatally exposed to tobacco, with a sample mean of 53.3 cigarettes smoked per week (SD = 43.1). Approximately one-third (35.3%) of infants were exposed to alcohol, with a sample mean of 0.7 standard drinks per week (SD = 3.2). A minority of mothers (17.6%) engaged in binge-drinking, defined as consuming four or more standard drinks in a day. However, number of binge-drinking days was highly skewed (M = 2.1, SD = 8.2, skewness = 4.9) and was not used in further analyses. Polysubstance risk scores ranged from 0 to 4, with a median and mode of one (M = 0.9, SD = 0.9).

3.2. Infant birth outcomes.

Infants’ average gestational age and birth weight fell within the normal range (Table 3). Out of 85 infants in the current sample, thirteen were born before 37 weeks gestation. All premature infants had been prenatally exposed to MAT. Continuous birth outcomes were not significantly related to prenatal substance exposure or demographic variables, including infant race and sex. However, an independent-samples t-test adjusted for unequal variances revealed that premature infants had higher total opioid exposure during pregnancy than infants born at term (M = 101.9, SD = 17.0, versus M = 71.8, SD = 46.2, p < 0.001). Groups did not differ significantly with regard to non-opioid substance or demographic variables.

Table 3.

Descriptive statistics and bivariate correlations between prenatal substance exposure and infant developmental outcomes.

Gest. Age Birth Weight 3M Comm 3M Gross 3M Fine 3M Prob 3M Pers 6M Comm 6M Gross 6M Fine 6M Prob 6M Pers
Birth Outcomes
 Gestational age - - −0.01 (0.96) 0.00 (1.00) 0.06 (0.60) 0.08 (0.53) −0.05 (0.67) 0.13 (0.29) 0.35 (0.003) 0.12 (0.31) 0.23 (0.06) 0.13 (0.28)
 Birth weight 0.72 (< .001) - −0.02 (0.89) −0.00 (1.00) 0.18 (0.14) −0.08 (0.49) −0.04 (0.77) −0.01 (0.96) 0.09 (0.47) 0.18 (0.14) 0.14 (0.26) −0.03 (0.79)
Demographics
 Maternal age 0.04 (0.69) −0.06 (0.62) −0.12 (0.31) −0.01 (0.95) −0.01 (0.92) 0.13 (0.29) 0.06 (0.62) 0.05 (0.66) −0.15 (0.21) −0.04 (0.75) 0.07 (0.57) 0.15 (0.23)
 Maternal education 0.09 (0.39) 0.15 (0.20) −0.07 (0.56) 0.08 (0.52) 0.06 (0.65) −0.18 (0.14) 0.11 (0.37) 0.16 (0.20) 0.09 (0.45) 0.00 (0.99) 0.08 (0.50) 0.05 (0.66)
 Household income 0.09 (0.41) 0.21 (0.08) −0.02 (0.86) 0.11 (0.39) 0.22 (0.08) 0.03 (0.80) 0.15 (0.25) 0.36 (0.003) 0.20 (0.13) 0.19 (0.14) 0.21 (0.10) 0.12 (0.35)
Prenatal Substance Use
 Total opioid exposurea −0.13 (0.22) −0.12 (0.31) −0.08 (0.51) −0.16 (0.18) −0.25 (0.04) 0.03 (0.79) −0.11 (0.39) −0.29 (0.02) −0.15 (0.22) 0.03 (0.82) −0.14 (0.26) −0.06 (0.60)
 Cigarettes/week −0.13 (0.23) −0.08 (0.51) 0.10 (0.42) −0.22 (0.07) −0.07 (0.56) 0.07 (0.59) −0.05 (0.69) 0.02 (0.90) −0.17 (0.16) 0.04 (0.77) −0.08 (0.51) −0.14 (0.27)
 Standard drinks/week −0.05 (0.63) 0.09 (0.41) 0.02 (0.87) 0.01 (0.97) −0.07 (0.58) −0.02 (0.89) 0.07 (0.55) −0.13 (0.29) −0.09 (0.46) 0.17 (0.17) 0.05 (0.71) −0.15 (0.21)
 Polysubstance riskb −0.16 (0.15) −0.13 (0.25) −0.01 (0.91) 0.04 (0.75) 0.04 (0.73) 0.23 (0.06) 0.27 (0.02) 0.04 (0.75) 0.01 (0.95) 0.28 (0.02) 0.20 (0.10) 0.10 (0.43)
Mean 38.6 6.6 49.5 54.0 41.6 49.7 49.6 52.9 50.3 50.4 53.5 50.2
Standard deviation 2.1 1.4 10.2 9.0 13.0 11.0 10.6 7.3 9.8 11.3 9.9 11.2
Min-Max 26.3–42.6 1.0c–9.6 10–60 15–60 10–60 20–60 15–60 30–60 20–60 25–60 15–60 20–60
Open in a new tab

Note. p-values presented in parentheses. Significant findings in bold, marginal findings in italics. Gest. Age = gestational age in weeks. Birth weight is given in lbs. Ages and Stages-3rd Edition scales: Comm = Communication; Gross = Gross Motor; Fine = Fine Motor; Prob = Problem-Solving; Pers = Personal-Social.

a

Sum of proportion of pregnancy days on Medication-Assisted Treatment and proportion of pregnancy days of illicit opioid use.

b

Tally of other substances used prenatally (marijuana, e-cigarettes, cocaine, prescription stimulant misuse, prescription sedative misuse, methamphetamines, hallucinogens, other).

c

Baby reported as one pound was born at 26 weeks’ gestation and did not survive.

In the focal logistic regression controlling for non-opioid exposure variables (cigarettes per week, standard drinks per week, and polysubstance risk; Table 4), prematurity was no longer significantly associated with total opioid exposure (b = 0.83, OR = 2.28, p = 0.06). The model yielded no significant predictors.

Table 4.

Predicting infant outcomes from prenatal opioid and non-opioid substance exposure and covariates.

Prematurity
 3M ASQ Fine Motor
 6M ASQ Communication
b OR p b p R2 b p R2
1. Total opioid exposure 0.83 2.30 0.06 −0.32 0.01 0.12 −0.32 0.02 0.21
 Cigarettes/week 0.14 1.14 0.68 0.10 0.50 0.19 0.17
 Standard drinks/week 0.07 1.07 0.81 −0.14 0.17 −0.16 0.07
 Polysubstance risk 0.30 1.34 0.28 0.07 0.52 0.07 0.49
 Adjusted age (days) - - - 0.19 0.12 0.10 0.29
 Income - - - - - 0.27 0.01
1. MAT (% pregnancy) - - - −0.23 0.07 0.15 −0.25 0.02 0.24
 Illicit opioid use (% pregnancy) - - - −0.34 0.03 −0.24 0.11
 Cigarettes/week - - - 0.12 0.43 0.24 0.06
 Standard drinks/week - - - −0.15 0.09 −0.15 0.21
 Polysubstance risk - - - 0.13 0.16 0.12 0.31
 Adjusted age (days) - - - 0.20 0.08 0.12 0.20
 Adjusted age (days) - - - - - - 0.30 0.004
Open in a new tab

Note. Non-dichotomous variables were standardized prior to analyses. Significant findings in bold, marginal findings in italics. OR= Odds Ratio. ASQ = Ages and Stages Questionnaire-3rd Edition. MAT = medication-assisted treatment.

3.3. Developmental status at three months.

At three months, all mean ASQ scores were in the normal range and the majority of infants scored above the clinical cutoff in each domain, indicating normative development: Communication (91.3%), Gross Motor (94.2%), Fine Motor (70.4%), Problem-Solving (91.4%), and Personal-Social (91.4%). Continuous communication, gross motor, and problem-solving scores were not significantly related to prenatal substance exposure or candidate covariates, including infant race and sex. Unexpectedly, higher polysubstance risk was associated with better personal-social skills (r = 0.27, p = 0.02). Opioid exposure was inversely associated with scores in the fine motor domain (r = −0.25, p = 0.03).

In the focal linear regression controlling for non-opioid exposure and adjusted age at three months (Table 4), total opioid exposure continued to predict three-month fine motor skills (b = −0.32, p = 0.01). No other predictors were significant, and the model accounted for an estimated 11.5% of the variance in three-month ASQ Fine Motor scores.

In a planned follow-up analysis assessing the relative contributions of prescribed versus illicit opioid exposure, only illicit opioid use significantly predicted fine motor skills (b = −0.34, p = 0.03). The association between three-month fine motor skills and proportion of pregnancy on MAT dwindled to marginal significance (b = −0.23, p = 0.07). A Wald test for equality of coefficients indicated that the difference between parameter estimates for illicit opioid use and MAT was not statistically significant (p = 0.47).

3.4. Developmental status at six months.

At six months, all mean ASQ scores were in the normal range and the majority of infants scored above the clinical cutoff in each domain, indicating normative development: Communication (100.0%), Gross Motor (98.5%), Fine Motor (97.1%), Problem-Solving (97.1%), and Personal-Social (98.5%). Continuous gross motor, problem-solving, and personal-social scores were not significantly related to prenatal substance exposure, although gross motor development was significantly associated with non-White race (non-White M = 52.1, SD = 8.5 versus White M = 46.7, SD = 11.7, p = 0.04) and older gestational age (r = 0.37, p = 0.003). Unexpectedly, more polysubstance risk was associated with better parent-reported fine motor skills (r = 0.28, p = 0.02). Six-month communication scores were positively associated with household income (r = 0.36, p < 0.01) and negatively related to total opioid exposure (r = −0.30, p = 0.01).

In the focal linear regression controlling for non-opioid exposure and adjusted age at six months, six-month communication scores were associated with total opioid exposure (𝑏 = −0.43, p = 0.001) and number of standard drinks of alcohol per week (𝑏 = −0.26, p = 0.004). When household income was entered on a subsequent step (Table 4), total opioid exposure (but not prenatal alcohol exposure) continued to inversely predict communication skills (𝑏 = −0.32, p = 0.02). Overall, the final model accounted for 20.8% of the variance in six-month communication scores.

In a planned follow-up analysis assessing the predictive significance of prescribed versus illicit opioid exposure, six-month communication scores were robustly related to household income and proportion of days on MAT (𝑏 = −0.25, p = 0.02). Controlling for these variables, communication scores were not associated significantly with illicit opioid use (𝑏 = −0.24, p = 0.11) or non-opioid substance exposure. Parameter estimates for MAT and illicit opioid exposure did not significantly differ from each other (p = 0.95).

3.4.a. Exploratory analyses by type and timing of MAT.

Given persistent links between MAT exposure and six-month infant communication, exploratory follow-up tests were conducted to investigate differential effects by type and timing of MAT. Accounting for illicit opioid use, non-opioid exposure, adjusted age, and household income, six-month communication skills were significantly associated with proportion of days on methadone (𝑏 = −0.29, p = 0.03) but not buprenorphine (𝑏 = −0.17, p = 0.17). A Wald test indicated that regression coefficients for methadone and buprenorphine did not significantly differ from each other (p = 0.59).

Similar findings were obtained when testing proportion of days on each medication in the first trimester (methadone 𝑏 = −0.29, p = 0.02, buprenorphine 𝑏 = −0.16, p = 0.19) and second trimester (methadone 𝑏 = −0.28, p = 0.03, buprenorphine 𝑏 = −0.17, p = 0.16), controlling for the same covariates. The size of effects dwindled when testing proportion of days on each medication in the third trimester, with neither medication significantly predicting lower communication skills (methadone 𝑏 = −0.22, p = 0.11, buprenorphine 𝑏 = −0.13, p = 0.35).

4. Discussion

The purpose of this study was to evaluate effects of continuous variation in prenatal opioid exposure on infant development, controlling for sociodemographic variables and polysubstance exposure. We hypothesized that more total opioid exposure (defined as the sum of durations of prescribed and illicit opioid use) would be associated with more adverse birth outcomes and worse mother-reported cognitive, motor, and behavioral development. We further anticipated that, among infants exposed to MAT, methadone would be associated with more adverse outcomes than buprenorphine.

Hypotheses were partially supported. Regarding birth outcomes, more total opioid exposure was significantly associated with greater risk for premature birth, but not shorter gestational age or lower birth weight. Furthermore, when entered simultaneously into a logistic regression, neither opioid nor non-opioid substance exposure (as captured by cigarettes per week, standard drinks per week, and polysubstance risk) significantly predicted prematurity. Non-significant results in this more complex model may reflect low statistical power to predict a dichotomous outcome. Alternately, given prior research documenting elevated rates of prematurity among infants whose mothers received methadone versus buprenorphine (Zedler et al., 2016), prematurity may be better predicted by methadone-specific exposure rather than total opioid exposure. To minimize multiple comparisons, we did not test specific effects of methadone in the absence of significant findings for total opioid exposure. However, this should be investigated in future studies with larger samples.

Maternal reports of infant development fell predominantly in the normal range, corroborating prior evidence that most children with prenatal opioid exposure show age-appropriate developmental progress (Kaltenbach et al., 2018; Konijnberg et al., 2001). However, opioid exposure over a greater proportion of the pregnancy was linked to lower performance in specific domains. At the three-month assessment, more total opioid exposure was associated with worse fine motor skills, but not communication, gross motor skills, problem-solving, or personal-social development, as reported by mothers on the ASQ. At the six-month assessment, more total opioid exposure was associated with worse parent-reported communication, but not gross or fine motor skills, problem-solving, or personal-social development. Transient decrements in fine motor skills are consistent with prior longitudinal evidence that opioid-exposed infants show worse motor coordination than non-exposed infants at four months, but not eight or twelve months (Marcus et al., 1984). No sociodemographic variables were significantly related to fine motor skills at the bivariate level, increasing confidence that results reflect substance exposure rather than other environmental risk. Importantly, amount of opioid exposure continued to predict fine motor skills after controlling for polysubstance use and adjusted age, suggesting opioid-specific effects.

Planned follow-up analyses differentiating by type of opioid found that three-month fine motor skills were uniquely associated with proportion of days of illicit opioid exposure, an effect that was moderate in size (b = −0.34, p = 0.03). Controlling for other opioid and polysubstance exposure, proportion of days on MAT showed a modest and marginally significant association with early fine motor skills (b = −0.23, p = 0.07). Parameter estimates for MAT and illicit opioid use did not significantly differ from each other, suggesting that prenatal opioid exposure shows similar associations with three-month fine motor skills whether opioids are prescribed or not. Overall, results suggest that in utero exposure to opioids, whether prescribed or not, may be damaging for early fine motor development.

Associations between total opioid exposure and parent-reported communication skills emerged at the six-month assessment. Lower communication skills during infancy may foreshadow difficulties with language development, consistent with prior research showing weaker verbal skills among young children exposed to opioids compared to those without prenatal exposure (Salo et al., 2009; van Baar et al., 1994), as well as elevated rates of language problems in both treatment groups of a randomized controlled trial comparing prenatal exposure to methadone versus buprenorphine (Czynski et al., 2020; deRegnier, 2020). Six-month communication skills were also positively associated with income, underscoring the importance of controlling for sociodemographic factors. In multiple regression models, total opioid exposure and household income each independently predicted six-month communication skills with moderate effect sizes (b = −0.32, p = 0.02, and b = 0.27, p = 0.01, respectively; comparable to betas as variables were standardized), suggesting that opioid effects are not reducible to sociodemographic risk. Associations with prenatal alcohol exposure were nonsignificant controlling for household income; non-opioid substance exposure did not uniquely predict six-month communication skills.

Follow-up analyses differentiating by type of opioid found that six-month communication skills were uniquely associated with proportion of days on MAT, with a small-to-medium effect size (b = −0.25, p = 0.02). The parameter associated with illicit opioid use was similar in magnitude, but a large standard error rendered it nonsignificant (b = −0.24, p = 0.11). This finding indicates that opioids may have downstream effect on infant neurodevelopment even when taken as prescribed.

Exploratory analyses differentiating proportion of days on a full versus partial opioid agonist found significant effects for methadone but not buprenorphine. A Wald test indicated that parameter estimates for the two medications were not significantly different from each other, consistent with results documenting comparable cognitive and behavioral development among infants of mothers randomized to receive methadone or buprenorphine during pregnancy (Kaltenbach et al., 2018). However, the non-significant advantage of buprenorphine over methadone is consistent with a growing body of evidence that MAT with buprenorphine is associated with lower neonatal risks than methadone (Guille & Aujla, 2019). Further research is needed to elucidate the relative risks of methadone and buprenorphine exposure measured continuously in samples that are powered to detect small differences in parameter estimates.

A final set of exploratory analyses sought to elucidate the role of timing of substance exposure during gestation. Trimester-specific models revealed robust associations between six-month communication skills and proportion of days exposed to methadone versus buprenorphine during the first and second trimester. In contrast, proportion of days on methadone during the third trimester did not significantly predict six-month communication skills controlling for third trimester buprenorphine exposure and standard covariates (i.e., adjusted age, income, illicit opioid use, and polysubstance variables). This may indicate stronger effects of opioid exposure earlier in prenatal development; however, results should be interpreted with caution given the exploratory nature of these analyses.

Overall, results reveal unique associations linking prenatal opioid exposure, measured continuously, with specific facets of infant development: variation in total opioid exposure and illicit opioid use each significantly predicted three-month fine motor skills, whereas total opioid exposure and prescribed exposure to MAT (particularly methadone during the first and second trimester) predicted six-month communication skills. In contrast, prenatal opioid exposure was not significantly associated with birth outcomes controlling for key covariates capturing polysubstance use. This study has several notable strengths, including detailed assessment of prenatal substance use with the Timeline Follow-Back Interview, permitting quantification of continuous variation in opioid and polysubstance exposure. Analyses are substantially strengthened by appropriate controls for sociodemographic factors and non-opioid exposure, including the widely used teratogens tobacco and alcohol. Exploratory follow-up analyses offer novel insight into timing-specific effects of prenatal MAT exposure, which, if replicated, may help to inform treatment decisions early in pregnancy for women with opioid use disorder.

Limitations of the current study include its small sample size, which precluded rigorous correction for multiple statistical tests as well as tests of moderation by key demographic factors (e.g., infant race or sex). This is an important direction for future research particularly given racial disparities in birth outcomes (Martin et al., 2021). Additionally, measurement of prenatal substance use relied on maternal report and did not include biomarkers. Importantly, however, for a majority of participants, reports of opioid exposure were supported by enrollment in a MAT program, and prior research has validated TLFB data using biomarkers of substance exposure (Eiden et al., 2013; Schuetze et al., 2008). Additional limitations involve reliance on maternal report of gestational age and birth weight, lack of data regarding neonatal opioid withdrawal symptoms, and limited information on maternal history (e.g., mothers’ own prenatal exposure to substances). Future research should seek to replicate these findings with a larger sample, multi-method assessment of substance exposure, and objective measurement of birth outcomes, including infant head circumference and neonatal opioid withdrawal symptoms by trained clinicians. Larger studies should investigate methadone-specific associations with prematurity, trimester-specific effects of illicit opioid use, and links between binge-drinking and infant neurodevelopment in the context of prenatal opioid use. Finally, longitudinal follow-up is needed to elucidate effects of opioid exposure later in neurodevelopment (e.g., Benninger et al., 2020; Byrnes & Vassoler, 2018; van Baar et al., 1984) and/or across generations (Manikkahm et al., 2012).

Despite these limitations, this study represents an important contribution to the growing literature on prenatal opioid exposure. Results confirm that most infants with prenatal opioid exposure show neurodevelopment in the normal range, but also document unique effects of both prescribed and illicit opioids on early development. Effective substance use treatment for pregnant women and careful monitoring of opioid-exposed infants are urgently needed to mitigate these developmental consequences of the opioid crisis.

  • More prenatal opioid exposure is related to worse fine motor skills at 3 months.

  • More prenatal opioid exposure is related to worse communication skills at 6 months.

  • Prenatal opioid effects persist controlling for demographics and polysubstance use.

  • Effect sizes are similar whether prenatal opioid use is illicit or prescribed.

Acknowledgements:

This research was supported by funding awarded to the last author by the National Institutes of Health (R01 HD098525) and the Mortimer Sackler Foundation, and by an NRSA predoctoral fellowship from NIH/NIDA (F31DA050426) awarded to the third author. We would like to thank our participants, referral partners in the community, and study staff at the University of Delaware.

Footnotes

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Akerman SC, Brunette MF, Green AI, Goodman DJ, Blunt HB, & Heil SH (2015). Treating tobacco use disorder in pregnant women in medication-assisted treatment for an opioid use disorder: a systematic review. Journal of Substance Abuse Treatment, 52, 40–47. 10.1016/j.jsat.2014.12.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Andersen JM, Høiseth G, & Nygaard E (2020). Prenatal exposure to methadone or buprenorphine and long-term outcomes: A meta-analysis. Early Human Development, 143, 104997. 10.1016/j.earlhumdev.2020.104997 [DOI] [PubMed] [Google Scholar]
  3. Azuine RE, Ji Y, Chang HY, Kim Y, Ji H, Dibari J, Hong X, Wang G, Singh GK, Pearson C, Zuckerman B, Surkan PJ, & Wang X (2019). Prenatal risk factors and perinatal and postnatal outcomes associated with maternal opioid exposure in an urban, low-income, multiethnic US population. JAMA Network Open, 2, 1–14. 10.1001/jamanetworkopen.2019.6405 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Beckwith AM, & Burke SA (2015). Identification of early developmental deficits in infants with prenatal heroin, methadone, and other opioid exposure. Clinical Pediatrics, 54, 328–335. 10.1177/0009922814549545 [DOI] [PubMed] [Google Scholar]
  5. Benninger KL, Borghese T, Kovalcik JT, Moore-Clingenpeen M, Isler C, Bonachea EM, Stark AR, Patrick SW, & Maitre NL (2020). Prenatal exposures are associated with worse neurodevelopmental outcomes in infants with neonatal opioid withdrawal syndrome. Frontiers in Pediatrics, 8, 482. 10.3389/fped.2020.00462. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Boggess T, & Risher WC (2020). Clinical and basic research investigations into the long-term effects of prenatal opioid exposure on brain development. Journal of Neuroscience Research, 1–14. 10.1002/jnr.24642 [DOI] [PubMed] [Google Scholar]
  7. Brown RA, Burgess ES, Sales SD, Whiteley JA, Evans DM, & Miller IW (1998). Reliability and validity of a smoking timeline follow-back interview. Psychology of Addictive Behaviors, 12, 101–112. 10.1037/0893-164X.12.2.101 [DOI] [Google Scholar]
  8. Byrne EM, & Vassoler FM (2018). Modeling prenatal opioid exposure in animals: Current findings and future directions. Frontiers in Neuroendocrinology, 51, 1–13. 10.1016/j.yfrne.2017.09.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Carey KB, Carey MP, Maisto SA, & Henson JM (2004). Temporal stability of the timeline follow-back interview for alcohol and drug use with psychiatric outpatients. Journal of Studies on Alcohol, 65, 774–781. 10.15288/jsa.2004.65.774 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chisolm MS, Fitzsimons H, Leoutsakos JMS, Acquavita SP, Heil SH, Wilson-Murphy M, Tuten M, Kaltenbach K, Martin PR, Winklbaur B, Jansson LM, & Jones HE (2013). A comparison of cigarette smoking profiles in opioid-dependent pregnant patients receiving methadone or buprenorphine. Nicotine and Tobacco Research, 15, 1297–1304. 10.1093/ntr/nts274 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cleary BJ, Donnelly J, Strawbridge J, Gallagher PJ, Fahey T, Clarke M, & Murphy DJ (2010). Methadone dose and neonatal abstinence syndrome — systematic review and meta-analysis. 2071–2084. 10.1111/j.1360-0443.2010.03120.x [DOI] [PubMed] [Google Scholar]
  12. Coyle MG, Salisbury AL, Lester BM, Jones HE, Lin H, Graf-Rohrmeister K, & Fischer G (2012). Neonatal neurobehavior effects following buprenorphine versus methadone exposure. 107, 63–73. 10.1111/j.1360-0443.2012.04040.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Conradt E, Flannery T, Aschner JL, Annett RD, & Croen LA (2019). Prenatal opioid exposure: Neurodevelopmental consequences and future research priorities. Pediatrics. 144(3): e20190128. 10.1542/peds.2019-0128 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Czynski AJ, Davis JM, Dansereau LM, Engelhardt B, Marro P, Bogen DL, Hudak ML, Shenberger J, Wachman EM, Oliveira EL, & Lester BM (2020). Neurodevelopmental outcomes of neonates randomized to morphine or methadone for treatment of Neonatal Abstinence Syndrome. Journal of Pediatrics, 219, 146–151.e1. 10.1016/j.jpeds.2019.12.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. deRegnier RA (2020). Developmental and behavioral outcomes of children with neonatal abstinence syndrome. Journal of Pediatrics, 219, 1–3. 10.1016/j.jpeds.2020.02.027 [DOI] [PubMed] [Google Scholar]
  16. Eiden RD, Homish GG, Colder CR, Schuetze P, Gray TR, & Huestis MA (2013). Changes in smoking patterns during pregnancy. Substance Use & Misuse, 48, 513–522. 10.3109/10826084.2013.787091 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fajemirokun-Odudeyi O, Sinha C, Tutty S, Pairaudeau P, Armstrong D, Phillips T, & Lindow SW (2006). Pregnancy outcome in women who use opiates. European Journal of Obstetrics & Gynecology and Reproductive Biology, 126, 170–175. 10.1016/j.ejogrb.2005.08.010 [DOI] [PubMed] [Google Scholar]
  18. Guille C, & Aujla R (2019). Developmental consequences of prenatal substance use in children and adolescents. Journal of Child and Adolescent Psychopharmacology, 29, 479–486. 10.1089/cap.2018.0177 [DOI] [PubMed] [Google Scholar]
  19. Haight SC, Ko JY, Tong VT, Bohm MK, & Callaghan WM (2018). Opioid use disorder documented at delivery hospitalization — United States, 1999–2014. MMWR. Morbidity and Mortality Weekly Report, 67, 845–849. 10.15585/mmwr.mm6731a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hans SL, & Jeremy RJ (2001). Postneonatal mental and motor development of infants exposed in utero to opioid drugs. Infant Mental Health Journal, 22(3), 300–315. 10.1002/imhj.1003 [DOI] [Google Scholar]
  21. Heinze G, Ploner M, & Jiricka L (2020). logistf: Firth’s Bias-Reduced Logistic Regression. R package version 1.24. https://CRAN.R-project.org/package=logistf. [Google Scholar]
  22. Jones HE, Dengler E, Garrison A, O’Grady KE, Seashore C, Horton E, Andringa K, Jansson LM, & Thorp J (2014). Neonatal outcomes and their relationship to maternal buprenorphine dose during pregnancy. Drug and Alcohol Dependence, 134, 414–417. 10.1016/j.drugalcdep.2013.11.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Jones HE, Johnson RE, Jasinski DR, O’Grady KE, Chisholm CA, Choo RE, Crocetti M, Dudas R, Harrow C, Huestis MA, Jansson LM, Lantz M, Lester BM, & Milio L (2005). Buprenorphine versus methadone in the treatment of pregnant opioid-dependent patients: Effects on the neonatal abstinence syndrome. Drug and Alcohol Dependence, 79, 1–10. 10.1016/j.drugalcdep.2004.11.013 [DOI] [PubMed] [Google Scholar]
  24. Jones HE, Kaltenbach K, Heil SH, Stine SM, Coyle MG, Arria AM, Ogrady KE, Selby P, Martin PR, & Fischer G (2011). Neonatal abstinence syndrome after methadone or buprenorphine exposure. Obstetrical and Gynecological Survey, 66, 191–193. 10.1097/OGX.0b013e318225c419 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kaltenbach K, O’Grady KE, Heil SH, Salisbury AL, Coyle MG, Fischer G, ... & Jones HE (2018). Prenatal exposure to methadone or buprenorphine: early childhood developmental outcomes. Drug and Alcohol Dependence, 185, 40–49. 10.1016/j.drugalcdep.2017.11.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Konijnenberg C, & Melinder A (2011). Prenatal exposure to methadone and buprenorphine: A review of the potential effects on cognitive development. Child Neuropsychology, 17, 495–519. 10.1080/09297049.2011.553591 [DOI] [PubMed] [Google Scholar]
  27. Konijnenberg C, Sarfi M, & Melinder A (2016). Mother-child interaction and cognitive development in children prenatally exposed to methadone or buprenorphine. Early Human Development, 101, 91–97. 10.1016/j.earlhumdev.2016.08.013 [DOI] [PubMed] [Google Scholar]
  28. Lee SJ, Bora S, Austin NC, Westerman A, & Henderson JMT (2020). Neurodevelopmental outcomes of children born to opioid-dependent mothers: A systematic review and meta-analysis. Academic Pediatrics, 20, 308–318. 10.1016/j.acap.2019.11.005 [DOI] [PubMed] [Google Scholar]
  29. Lejeune C, Simmat-Durand L, Gourarier L, & Aubisson S (2006). Prospective multicenter observational study of 260 infants born to 259 opiate-dependent mothers on methadone or high-dose buprenophine substitution. Drug and Alcohol Dependence, 82, 250–257. 10.1016/j.drugalcdep.2005.10.001 [DOI] [PubMed] [Google Scholar]
  30. Maguire DJ, Taylor S, Armstrong K, Shaffer-Hudkins E, Germain AM, Brooks SS, Cline GJ, & Clark L (2016). Long-term outcomes of infants with neonatal abstinence syndrome. Neonatal Network, 35(5), 277–286. 10.1891/0730-0832.35.5.277 [DOI] [PubMed] [Google Scholar]
  31. Manikkam M, Guerrero-Bosagna C, Tracey R, Haque MM, & Skinner MK (2012). Transgenerational actions of environmental compounds on reproductive disease and identification of epigenetic biomarkers of ancestral exposures. PLoS One, 7, e319101. 10.1371/journal.pone.0031901 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Marcus J, Hans SL, & Jeremy RJ (1984). A longitudinal study of offspring born to methadone-maintained women. III. Effects of multiple risk factors on development at 4, 8, and 12 months. The American Journal of Drug and Alcohol Abuse, 10, 195–207. 10.3109/00952998409002780 [DOI] [PubMed] [Google Scholar]
  33. Martin JA, Hamilton BE, Osterman MJK, & Driscoll AK (2021). Births: Final data for 2019. National Vital Statistics Reports; vol 70 no 2. National Center for Health Statistics. [PubMed] [Google Scholar]
  34. Nelson LF, Yocum VK, Patel KD, Qeadan F, Hsi A, & Weitzen S (2020). Cognitive outcomes of young children after prenatal exposure to medications for opioid use disorder: A systematic review and meta-analysis. JAMA Network Open, 3, e201195. 10.1001/jamanetworkopen.2020.1195 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Nygaard E, Slinning K, Moe V, & Walhovd KB (2017). Cognitive function of youths born to mothers with opioid and poly-substance abuse problems during pregnancy. Child Neuropsychology, 23(2), 159–187. 10.1080/09297049.2015.1092509 [DOI] [PubMed] [Google Scholar]
  36. R Core Team. (2016). R: A Language and Environment for Statistical Computing. Vienna, Austria. Retrieved from https://www.R-project.org/ [Google Scholar]
  37. Rosseel Y (2012). lavaan: An R Package for Structural Equation Modeling. Journal of Statistical Software, 48, 1–36. Retrieved from http://www.jstatsoft.org/v48/i02/ [Google Scholar]
  38. Salo S, Kivistö K, Korja R, Biringen Z, Tupola S, Kahila H, & Kivitie-Kallio S (2009). Emotional availability, parental self-efficacy beliefs, and child development in caregiver-child relationships with buprenorphine-exposed 3-year-olds. Parenting, 9, 244–259. 10.1080/15295190902844563 [DOI] [Google Scholar]
  39. Schuetze P, Lopez FA, Granger DA, & Eiden RD (2008). The association between prenatal exposure to cigarettes and cortisol reactivity and regulation in 7‐month‐old infants. Developmental Psychobiology, 50, 819–834. 10.1002/dev.20334 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sobell MB, Sobell LC, Klajner F, Pavan D, & Basian E (1986). The reliability of a timeline method for assessing normal drinker college students’ recent drinking history: Utility for alcohol research. Addictive Behaviors, 11, 149–161. 10.1016/0306-4603(86)90040-7 [DOI] [PubMed] [Google Scholar]
  41. Singer LT, Chambers C, Coles C, & Kable J (2020). Fifty years of research on prenatal substances: Lessons learned for the opioid epidemic. Adversity and Resilience Science, 223–234. 10.1007/s42844-020-00021-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. van Baar A, Soepatmi S, Gunning W, & Akkerhuis G (1994). Development after prenatal exposure to cocaine, heroin and methadone. Acta Pædiatrica, 83, 40–46. 10.1111/j.1651-2227.1994.tb13383.x [DOI] [PubMed] [Google Scholar]
  43. Wiegand S, Stringer E, Seashore C, Garcia K, Jones H, Stuebe A, & Thorp J (2014). 750: Buprenorphine/naloxone (B/N) and methadone (M) maintenance during pregnancy: a chart review and comparison of maternal and neonatal outcomes. American Journal of Obstetrics and Gynecology, 210, S368–S369. 10.1016/j.ajog.2013.10.783 [DOI] [Google Scholar]
  44. Yeoh SL, Eastwood J, Wright IM, Morton R, Melhuish E, Ward M, & Oei JL (2019). Cognitive and motor outcomes of children with prenatal opioid exposure: A systematic review and meta-analysis. JAMA Network Open, 2, 1–14. 10.1001/jamanetworkopen.2019.7025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Zedler BK, Mann AL, Kim MM, Amick HR, Joyce AR, Murrelle EL, & Jones HE (2016). Buprenorphine compared with methadone to treat pregnant women with opioid use disorder: a systematic review and meta‐analysis of safety in the mother, fetus and child. Addiction, 111, 2115–2128. 10.1111/add.13462 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES