Abstract
Introduction
Several studies have reported increasing prevalence of prescription opioid use among pregnant women. However, little is known regarding the effects of maternal opioid use on neurodevelopmental disorders in early childhood in pregnant women with no evidence of opioid use disorders or drug dependence.
Objective
The aim of this study was to quantify the association between prenatal opioid exposure from maternal prescription use and neurodevelopmental outcomes in early childhood.
Methods
This retrospective study included pregnant women aged 12–55 years and their live-birth infants born from 2010 to 2012 present in Optum’s deidentified Clinformatics® Data Mart database. Eligible infants born to mothers without opioid use disorders or drug dependence were followed till occurrence of neurodevelopmental disorders, loss to follow-up, or study end (December 31, 2017), whichever came first. Propensity score by fine stratification was applied to adjust for confounding by demographic characteristics, obstetric characteristics, maternal comorbid mental and pain conditions, and measures of burden of illnesses and to obtain adjusted hazard ratios (HR) and 95% confidence intervals (CI). Exposed and unexposed infants were compared on the incidence of neurodevelopmental disorders.
Results
Of 24,910 newborns, 7.6% (1899) were prenatally exposed to prescription opioids. Overall, 1562 children were diagnosed with neurodevelopmental disorders, with crude incidence rates of 2.9 per 100 person-years in exposed children versus 2.5 per 100 person-years in unexposed children. After adjustment, we observed no association between fetal opioid exposure and the risk of neurodevelopmental disorders (HR 1.10; 95% CI 0.92–1.32). However, increased risk of neurodevelopmental disorders were observed in children with longer cumulative exposure duration (HR 1.70; 95% CI 1.05–2.96) or high cumulative opioid doses (HR 1.22; 95% CI 1.01–1.54).
Conclusion and Relevance
In pregnant women without opioid use disorders or drug dependence, maternal opioid use was not associated with increased risk of neurodevelopmental disorders in early childhood. However, increased risks of early neurodevelopmental disorders were observed in children born to women receiving prescription opioids for longer duration and at higher doses during pregnancy.
1. Introduction
Maternal use of opioids has been associated with several negative maternal and fetal outcomes including altered fetal growth and congenital malformations [1–6]. Exposure of rodents to opioids during critical developmental periods has also been reported to induce neuronal apoptosis [7–9] and was associated with structural and functional alterations [10]. Despite this safety concern, maternal opioid use has steadily increased in the past decade [11, 12].
Little is known about the effects of maternal prescription opioid use on the neurodevelopmental outcomes of children of mothers without diagnosis of opioid user disorders (OUD) or drug dependence during pregnancy. So far, knowledge of the potential adverse effects of fetal opioid exposure has been primarily aggregated from studies investigating cognitive impairment in relation to birth defects [13, 14] and small sample sizes conducted in mothers with OUD or illicit drug use (eTable 1, see electronic supplementary material [ESM]) [15–19]. These findings are, however, not generalizable to offspring of mothers receiving prescribed opioids since opioid-dependent mothers or those with history of illicit drug use during pregnancy are more likely to have higher incidence of comorbid conditions, risk factors like smoking and alcohol use, or socioeconomic factors like disruptions in maternal care that can markedly increase the risk of neurodevelopmental disorders in the developing fetus [17, 19, 20]. Additionally, although birth defects have been associated with higher likelihood of cognitive impairments and lower academic performance in childhood [13, 21], neurodevelopmental disorders can occur with or without birth defects. Moreover, it is unknown whether the potential effects of opioids on neurodevelopment differs by trimester, cumulative dose, exposure duration or between specific opioids. Therefore, we conducted a large population-based study to quantify the association of gestational prescription opioid exposure on neurodevelopmental outcomes in early childhood of offspring of women without OUD during pregnancy.
2. Methods
2.1. Data Source
This retrospective cohort study was conducted using administrative claims from January 1, 2010 to December 31, 2017 in Optum’s deidentified Clinformatics® Data Mart database [22]. This database contains yearly information on demographics, medical services, laboratory results, and prescription services for more than 20 million unique patients enrolled in commercial and Medicare Advantage health insurance plans in the US. Inpatient or outpatient diagnoses are recorded using International Classification of Disease, Ninth or Tenth Revision, Clinical Modification (ICD-9-CM or ICD-10-CM) codes, while procedures are recorded using ICD-9 procedure codes, Current Procedural Terminology, 4th Edition codes, and the Health Common Procedure Coding System. Information on drugs filled by enrollees, including fill date, medication name, quantity dispensed, days’ supply, National Drug Code (NDC), formulation type, and drug strength are available in the pharmacy claims data. The Clinformatics database contains deidentified individual data; therefore, this study was considered exempt by the University of Rhode Island Institutional Review Board.
2.2. Mother–Infant Linkage and Estimation of Pregnancy Window
Live births were identified from the medical claims using diagnoses (ICD-9-CM) and procedure codes (ICD-9 procedure and CPT-4 codes) indicative of deliveries [23]. Similar to previously reported algorithms in the literature, pregnant women were linked to their live-born infant(s) using a unique family identification number; date of delivery-related service; and the infant’s year of birth and earliest date of insurance enrollment [23–25]. The pregnancy window was defined by the estimated delivery date and last menstrual period (LMP). The LMP was calculated by subtracting 35 weeks for preterm births and 39 weeks for full-term births from estimated delivery date; an algorithm reported to identify 75% and 99% of preterm and term deliveries, respectively, when compared against clinical gestational age [26].
2.3. Eligibility Criteria
The source population comprised mothers aged 12–55 years and their liveborn infants born between study start and December 31, 2012. Mothers were required to have continuous insurance eligibility 3 months prior to the estimated LMP (i.e., baseline period) to 1-month post-partum. We excluded mothers with diagnosis of cancer or OUD during baseline or pregnancy [27, 28]; women exposed to confirmed teratogenic medication during pregnancy; and women or infants with chromosomal abnormality during pregnancy or 1-month post-partum (Fig. 1). Operational definitions and time of assessment for the exclusion criteria are presented in eTable 2 (see ESM).
Figure 1.
Study selection
2.4. Opioid Exposure
Infants were considered exposed if their mother filled one or more prescription of single or combination opioids during pregnancy and unexposed if their mothers did not fill an opioid prescription any time during the baseline and pregnancy period. We required no filling of opioid prescriptions during the baseline period in the unexposed group to prevent exposure misclassification arising from women who may still have had opioids remaining from an earlier prescription, but available at the start of pregnancy. As such, mothers who filled opioids during baseline, but not during pregnancy, were excluded (Fig. 1). To allow meaningful comparisons across different opioid medications, opioid doses were converted to morphine milligram equivalents (MMEs) using opioid-conversion tables [29, 30]. Prescription opioid medications included in the analyses, and method for computing MME for each opioid prescription, are listed in eTable 3 (see ESM).
2.5. Outcome Measure
The outcome of interest was neurodevelopmental disorders comprising a range of disabilities including intellectual disorders, specific delays in development such as mathematics disorders and developmental reading disorders, and pervasive developmental disorders such as autism spectrum disorders (ASD)—conditions that have been associated with disruptions to normal brain development and detectable particularly in early childhood [31]. The term ‘neurodevelopmental disorders’ has been used broadly in literature to refer to overlapping and sometimes distinct neurological conditions [32–35]. In this study, we adopted the diagnostic grouping of the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) for neurodevelopmental disorders, which comprises of any of the following: intellectual disorders, communication disorders, ASD, attention-deficit/ hyperactivity disorder (ADHD), specific learning disorders, motor disorders, and other neurodevelopmental disorders [36]. The outcome measure was defined on the basis of the presence of one or more inpatient or outpatient ICD-9/10 diagnosis codes in medical claims of eligible children, corresponding to the seven categories of neurodevelopmental disorders listed by the DSM-5 definition (eTable 4, see ESM).
2.6. Covariates
We considered known and suspected risk factors and their proxies for adjustment in our analyses. The covariates taken into consideration were grouped as maternal sociodemographic characteristics, obstetric characteristics, maternal comorbidities including pain conditions and mental disorders, concomitant drug use, and measures of burden of illnesses such as the Obstetric Morbidity Index [37]. Obstetric characteristics and maternal comorbid conditions were defined based on the presence of ICD-9 diagnosis and procedure codes in the medical claims of mothers, and concomitant medication use by the presence of relevant drugs in mothers’ prescription claims. A complete list of the covariates and the time of assessment is presented in eTable 5 (see ESM).
2.7. Statistical Analysis
Following delivery, infants born to eligible mothers before December 31, 2012 (inclusive) were followed until first diagnosis of neurodevelopmental disorders, lost to follow-up due to health insurance disenrollment, or December 31, 2017 (the study end date), whichever came first. Cox proportional-hazard models were used to estimate the absolute rates (per 1000 person-years) and hazard ratios (HRs) with their 95% confidence intervals. A time-to-event model was used because our outcome measure comprised different types of neurodevelopmental disorders that may be diagnosed at different times during early childhood. For example, language and intellectual disabilities are more commonly identifiable and diagnosed within the first 4 years of life, but up to age 12 years for ADHD diagnosis. More so, individuals were followed for different lengths of time arising from varying patterns of health insurance disenrollment—a censoring event—between individuals.
Analyses of Schoenfeld residuals showed no violations of the proportional hazards assumption [38]. Adjusted effect estimates—whether for primary, secondary, or sensitivity analyses—were derived using propensity score fine stratification, a confounding adjustment approach shown to be suitable for infrequent exposure [39]. Specifically, we accounted for potential confounding by creating 50 equally sized propensity score strata based on the propensity score distribution among the opioid-exposed women, and then weighted the untreated group in the final outcome model using the distribution of the exposed group [39]. Following propensity score adjustment, standardized differences were used to assess balance in baseline covariates between opioid-exposed and -unexposed individuals [40]. The maximum allowed standardized difference between the opioid-exposed and -unexposed groups was 0.1. We accounted for correlation between siblings by using the Huber sandwich estimator for correction of standard errors [41]. Accounting for clustering between siblings did not meaningfully change the confidence interval. As such, clustering structures were omitted from our analyses. All analyses were conducted using SAS 9.4 (SAS Institute Inc). No adjustment was made for multiple comparisons.
2.8. Secondary Analyses
We assessed the relationship between different cumulative dose categories or duration of fetal opioid exposure and the risk of neurodevelopmental disorders. First, exposed infants were dichotomized by median of the cumulative opioid dose (in MME) during the pregnancy period into low and high cumulative dose (< 37.5 and ≥ 37.5 MME, respectively). Second, we assessed the relationship between cumulative duration of in utero opioid exposure—dichotomized by cumulated days’ supply into ‘short-term’ (< 14 days) and ‘long-term exposure’ (≥ 14 days)—and the incidence of neurodevelopmental disorders. We regarded ≥ 14 cumulative days of opioid supply as ‘long-term exposure’ based on previous research indicating that long-term opioid use is infrequent in pregnancy and the median of the total days of opioid supply during pregnancy is often < 2 weeks [42]. Since individual opioids may be associated with different risk profiles, we further evaluated whether the association between the outcomes of interest differed across individual opioids. Therefore, we compared the incidence of outcomes in unexposed versus newborns exposed to monotherapy of a specific opioid during pregnancy.
2.9. Sensitivity Analyses
To our knowledge, little-to-nothing is known regarding the etiologically relevant gestational window between prenatal opioid exposure and neurodevelopmental disorders. Therefore, we assessed the relationship between opioid exposure during specific trimesters—first trimester (LMP to 90 days of pregnancy), second trimester (91 to 180 days of pregnancy), or third trimester (181 days of pregnancy till date of birth) and the incidence of neurodevelopmental disorders.
To address the issue of exposure misclassification during specific trimesters or during pregnancy, we alternatively considered newborns exposed if the days’ supply of prescribed opioid(s) overlapped with the pregnancy window or any trimester. For example, exposure misclassification can potentially occur with our primary exposure definition if, for example, a mother filled an opioid prescription close to the end of a trimester but had days of supply from a previous prescription extending into the subsequent trimester. Next, we used a more conservative exposure definition. Here, exposed infants comprised individuals whose mothers filled at least two opioid prescriptions during pregnancy. While prescription filling does not necessarily imply actual use, we assumed that mothers who refilled opioid prescriptions during pregnancy are probably consuming them or had underlying indications that may warrant refilling of prescription opioids. To overcome concerns of potential outcome misclassification, we redefined our outcome measure using more conversation operational definitions including (a) two or more hospital contacts for neurodevelopmental disorders; and (b) two or more neurodevelopmental disorder-related claims on two separate days.
Additionally, since neurodevelopmental disorders in early childhood may become more apparent in preschool or pre-kindergarten ages, we repeated our primary analysis in a restricted cohort comprising children with ≥ 3 or ≥ 5 years of follow-up. To mitigate confounding by underlying indications for opioids, we performed two additional analyses. First, we repeated our primary analysis in a restricted cohort of children born to mothers with diagnosis of acute or chronic pain conditions, major or minor surgeries, mental disorders, or receipt of psychotropic medications. Second, we further adjusted the restricted cohort for confounding by all other prespecified covariates. Finally, we calculated the E-value to assess the potential effect of unmeasured confounding on the primary estimates [43]; that is, the minimum magnitude of association an unmeasured confounder would need to have with the treatment–outcome relationship, conditional on the measured covariates to fully explain away the association between prenatal opioid exposure and neurodevelopmental disorders.
3. Results
3.1. Patients
Prior to propensity score adjustment, compared with mothers who were unexposed, mothers who received opioids during pregnancy were older, had higher prevalence of maternal illnesses and comorbidities, and concomitant medications use (Table 1).
Table 1.
Selected baseline characteristics of included women by maternal prescription opioid use
| Characteristics | Unmatched cohort |
Propensity score adjusted cohort |
||||
|---|---|---|---|---|---|---|
| Exposed (n = 1899; 7.6%) | Unexposed (n = 23,011; 92.4%) | Standardized difference | Exposed (n = 1899; 7.6%) | Unexposed (n = 22,982; 92.4%) | Standardized difference | |
|
| ||||||
| Maternal age at delivery | ||||||
| Mean age (y), SD | 30.7 (4.3) | 30.5 (4.1) | 0.040 | 30.7 (4.3) | 30.7 (4.2) | −0.003 |
| <18 | 3 (0.2) | 51 (0.2) | −0.015 | 3 (0.2) | 34 (0.2) | 0.002 |
| 18–24 | 124 (6.5) | 1421 (6.2) | 0.015 | 124 (6.5) | 1502 (6.5) | 0 |
| 25–34 | 1464 (77.1) | 18291 (79.5) | −0.058 | 1464 (77.1) | 17709 (77.1) | 0.001 |
| ≥35 | 308 (16.2) | 3248 (14.1) | 0.059 | 308 (16.2) | 3735 (16.3) | −0.001 |
| Year of birth | ||||||
| 2010 | 3 (0.2) | 49 (0.2) | −0.013 | 3 (0.2) | 37 (0.2) | −0.001 |
| 2011 | 1057 (55.7) | 12169 (52.9) | 0.056 | 1057 (55.7) | 12898 (56.1) | −0.009 |
| 2012 | 839 (44.2) | 10793 (46.9) | −0.055 | 839 (44.2) | 10047 (43.7) | 0.009 |
| Geographic region | ||||||
| Northeast | 71 (3.7) | 1495 (6.5) | −0.125 | 71 (3.7) | 906 (3.9) | −0.011 |
| Midwest | 483 (25.4) | 6587 (28.6) | −0.072 | 483 (25.4) | 5662 (24.6) | 0.018 |
| South | 782 (41.2) | 8675 (37.7) | 0.071 | 782 (41.2) | 9550 (41.6) | −0.008 |
| West | 553 (29.1) | 6156 (26.8) | 0.053 | 553 (29.1) | 6739 (29.3) | −0.004 |
| Unknown | 10 (0.5) | 98 (0.4) | 0.015 | 10 (0.5) | 124 (0.5) | −0.002 |
| Obstetric characteristics | ||||||
| Tobacco use | 59 (3.1) | 349 (1.5) | 0.106 | 59 (3.1) | 701 (3) | 0.003 |
| Alcohol | 5 (0.3) | 17 (0.1) | 0.046 | 5 (0.3) | 47 (0.2) | 0.012 |
| Obesity | 76 (4.0) | 689 (3.0) | 0.055 | 76 (4.0) | 919 (4.0) | 0 |
| Multiple gestation | 13 (0.7) | 111 (0.5) | 0.027 | 13 (0.7) | 152 (0.7) | 0.003 |
| Substance use/abuse | 2 (0.1) | 11 (0.0) | 0.021 | 2 (0.1) | 21 (0.1) | 0.004 |
| Comorbid conditions | ||||||
| Psychiatric conditions | ||||||
| Depression | 128 (6.7) | 826 (3.6) | 0.143 | 128 (6.7) | 1524 (6.6) | 0.004 |
| Anxiety | 113 (6.0) | 886 (3.9) | 0.097 | 113 (6.0) | 1364 (5.9) | 0.001 |
| ADHD | 26 (1.4) | 168 (0.7) | 0.063 | 26 (1.4) | 321 (1.4) | −0.002 |
| Bipolar disorder | 10 (0.5) | 42 (0.2) | 0.058 | 10 (0.5) | 111 (0.5) | 0.006 |
| Epilepsy | 4 (0.2) | 36 (0.2) | 0.013 | 4 (0.2) | 42 (0.2) | 0.006 |
| Pain conditions | ||||||
| Migraine | 217 (11.4) | 1030 (4.5) | 0.259 | 217 (11.4) | 2599 (11.3) | 0.004 |
| Chronic pain | 553 (29.1) | 3955 (17.2) | 0.286 | 553 (29.1) | 6823 (29.7) | −0.013 |
| Major or minor surgery | 56 (2.9) | 278 (1.2) | 0.12 | 55 (2.9) | 651 (2.8) | 0.001 |
| Other maternal illnesses | ||||||
| Diabetes mellitus | 29 (1.5) | 302 (1.3) | 0.029 | 156 (8.2) | 1885 (8.2) | 0 |
| Hypertension | 44 (2.3) | 321 (1.4) | 0.059 | 65 (3.4) | 774 (3.4) | 0.003 |
| Concomitant medication use | ||||||
| Psychotropic medications | ||||||
| Benzodiazepines | 121 (6.4) | 504 (2.2) | 0.208 | 121 (6.4) | 1433 (6.2) | 0.006 |
| Antidepressants | 240 (12.6) | 1322 (5.7) | 0.24 | 240 (12.6) | 2910 (12.7) | −0.001 |
| Psychostimulants | 44 (2.3) | 248 (1.1) | 0.096 | 44 (2.3) | 536 (2.3) | −0.001 |
| Antipsychotics | 5 (0.3) | 36 (0.2) | 0.023 | 5 (0.3) | 55 (0.2) | 0.004 |
| Other medications | ||||||
| Antidiabetics | 57 (3) | 430 (1.9) | 0.074 | 57 (3.0) | 706 (3.1) | −0.004 |
| Antihypertensives | 79 (4.2) | 465 (2.0) | 0.124 | 79 (4.2) | 910 (4.0) | 0.01 |
| Antiepileptics | 38 (2.0) | 141 (0.6) | 0.122 | 38 (2.0) | 419 (1.8) | 0.013 |
| Maternal comorbidity markers | ||||||
| Maternal comorbidity index, mean (SD) | 0.5 (0.9) | 0.4 (0.8) | 0.093 | 0.4 (0.8) | 0.4 (0.8) | 0.004 |
| Number of hospitalizations | 1.2 (2.7) | 1.1 (1.9) | 0.039 | 1.2 (2.7) | 1.2 (3.0) | 0.003 |
| Length of hospitalization, mean (SD) | 0.4 (0.5) | 0.5 (0.5) | −0.043 | 0.4 (0.5) | 0.4 (0.5) | 0.009 |
| Number of outpatient visits, mean (SD) | 36.7 (25.2) | 33.1 (22.4) | 0.152 | 36.7 (25.2) | 36.9 (26.7) | −0.008 |
Unless otherwise stated, number (%) are presented
ADHD attention-deficit hyperactivity disorder
Overall, 24,729 mothers with 24,910 unique newborns (24,550 singletons and 360 nested siblings) were included in our analyses (Fig. 1). Of these infants, 7.6% (n = 1899) were exposed to prescription opioids during pregnancy with maternal opioid use more prevalent during the third trimester alone (34.4%, n = 653) (Table 2). Additional results on opioid utilization including cumulative dose and frequently prescribed opioids are presented in Table 2 and eTable 6 (see ESM). Included children were followed for a maximum of 2580 days; median (IQR) and average (standard deviation [SD]) follow-up times were 725 (271–1665) and 964 (784) days, respectively. No differences in follow-up time were observed by maternal opioid use (exposed vs unexposed: 941.3 [776.3] vs 966.2 [784.6]).
Table 2.
Patterns of prenatal opioid exposure from maternal opioid use
| (A) | First trimester alone (n = 653; 34.4%) | Second trimester alone (n = 395; 20.8%) | Third trimester alone (n = 672; 35.4%) | More than two trimesters (n = 179; 9.4%) | Anytime during pregnancy (n = 1899; 100%) |
|
| |||||
| Number of prescriptions | |||||
| 1 | 580 (88.8) | 359 (90.9) | 599 (89.1) | 0 (0) | 1538 (81.0) |
| 2 | 57 (8.7) | 32 (8.1) | 55 (8.2) | 117 (65.4) | 261 (13.7) |
| 3 | 10 (1.5) | 3 (0.8) | 11 (1.6) | 32 (17.9) | 56 (3.0) |
| 4 | 6 (1.0) | 1 (0.3) | 6 (0.9) | 16 (8.9) | 29 (1.5) |
| 5 | 0 | 0 | 0 | 4 (2.2) | 4 (0.2) |
| ≥6 | 0 | 0 | 1 (0.2) | 10 (5.6) | 11 (0.6) |
| Cumulative days of opioid supply | |||||
| Mean (SD) | 6.5 (6.3) | 6.3 (5.9) | 5.6 (5.7) | 17.2 (32.2) | 7.2 (11.8) |
| Median | 5.0 | 5.0 | 4.0 | 10.0 | 5.0 |
| Min, max | 1, 65 | 1, 45 | 1, 95 | 3, 392 | 1, 392 |
| Q1, Q3 | 3, 8 | 3, 7 | 3, 6 | 7, 16 | 3, 8 |
| Cumulative opioid dose (in MME) | |||||
| Mean (SD) | 51.0 (56.7) | 44.1 (38.8) | 51.2 (42.1) | 125.6 (182.4) | 56.7 (75.3) |
| Median | 37.5 | 30.0 | 37.5 | 83.3 | 37.5 |
| Min, max | 2.7, 709 | 3.8, 450 | 2.25, 450 | 10, 2205 | 2.25, 2205 |
| Q1, Q3 | 25, 56.3 | 22.5, 50 | 30, 58.7 | 60, 135 | 28.3, 64.5 |
| Prescribed opioids | |||||
| Exposed to a single opioid | |||||
| Hydrocodone | 351 (53.8) | 151 (38.2) | 289 (43.0) | 45 (25.1) | 836 (44.0) |
| Codeine | 194 (29.7) | 209 (52.91) | 261 (38.8) | 44 (24.6) | 708 (37.3) |
| Oxycodone | 54 (8.3) | 11 (2.8) | 102 (15.2) | 4 (2.2) | 171 (9.0) |
| Tramadol | 28 (4.3) | 11 (2.8) | 3 (0.5) | 1 (0.6) | 43 (2.3) |
| Meperidine | 1 (0.2) | 1 (0.3) | 0 | 0 | 2 (0.1) |
| Hydromorphone | 1 (0.2) | 0 | 0 | 0 | 1 (0.1) |
| Tapentadol | 1 (0.2) | 0 | 0 | 0 | 1 (0.1) |
| Exposed ≥2 opioids | 23 (3.5) | 12 (3.0) | 17 (2.5) | 85 (62.0) | 137 (9.0) |
|
| |||||
| (B) | N (%) | Mean (SD) | Median | Min, max | Q1, Q3 |
|
| |||||
| Cumulative opioid dose (in MME) by opioid type | |||||
| Exposed to a single opioid | |||||
| Hydrocodone | 836 (44.0) | 53.1 (57.6) | 36.5 | 5, 709.5 | 28.6, 59.2 |
| Codeine | 708 (37.3) | 45.5 (37.5) | 35.9 | 2.3, 360 | 22.5, 54 |
| Oxycodone | 171 (9.0) | 87.2 (182.3) | 56.3 | 7.5, 2205 | 42.9, 75 |
| Tramadol | 43 (2.3) | 24.3 (9.3) | 22.5 | 5, 50 | 20. 30 |
| Meperidine | 2 (0.1) | 12.5 (10.6) | 12.5 | 5, 20 | 5, 20 |
| Hydromorphone | 1 (0.1) | 34.3 | 34.3 | 34.3 | 34.3, 34.3 |
| Tapentadol | 1 (0.1) | 150 | 150 | 150, 150 | 150, 150 |
| Exposed ≥2 opioids | 137 (9.0) | 108.4 | 90.0 | 23.5, 484.6 | 64.5, 130.6 |
| Cumulative days of supply by opioid type | |||||
| Exposed to a single opioid | |||||
| Hydrocodone | 836 (44.0) | 6.2 (7.2) | 4.0 | 1, 95 | 3, 6 |
| Codeine | 708 (37.3) | 6.8 (6.2) | 5.0 | 1, 75 | 3, 8 |
| Oxycodone | 171 (9.0) | 7.7 (30.4) | 4.0 | 1, 392 | 3, 6 |
| Tramadol | 43 (2.3) | 9.0 (10.0) | 6.0 | 1, 60 | 3, 10 |
| Meperidine | 2 (0.1) | 5 (0.7) | 5.5 | 5, 6 | 5, 6 |
| Hydromorphone | 1 (0.1) | 7.0 | 7.0 | 7, 7 | 7, 7 |
| Tapentadol | 1 (0.1) | 2.0 | 2.0 | 2, 2 | 2, 2 |
| Exposed ≥2 opioids | 137 (9.0) | 14.0 (13.9) | 10.0 | 3, 120 | 7, 15 |
max maximum, min minimum, MME morphine milligram equivalent, n number, Q1 25th percentile, Q3 75th percentile, SD standard deviation
3.2. Outcomes
We identified 1562 infants with neurodevelopmental disorders with 134 (7.1%) in prenatally opioid-exposed infants versus 1951 (6.2%) in unexposed infants. The mean and median age at diagnosis of neurodevelopmental disorders was 857.6 (526) days and 736 (523–1130) days. The crude incidence rate among exposed and unexposed infants was 2.90 (2.43–3.43) and 2.46 (2.34–2.59) per 100 person-years (PY), respectively, equivalent to an unadjusted HR (95% CI) of 1.18 (0.99–1.41) (Table 3). Following propensity-score adjustment, we found no association between prenatal opioid exposure and the risk of neurodevelopmental disorders in early childhood (adjusted HR 1.10 [0.92–1.32]).
Table 3.
Hazard ratios of the risk of neurodevelopmental disorders in children by maternal exposure to opioids (primary and secondary analyses)
| Outcome | Total number | Cases, n (%) | Crude IR (95% CI) per 100 PY | Crude HR (95% CI) | Adjusted HR (95% CI) |
|---|---|---|---|---|---|
|
| |||||
| Primary analysis | |||||
| Unexposed | 23011 | 1428 (6.2) | 2.46 (2.34–2.59) | 1 | 1 |
| Exposed | 1899 | 134 (7.1) | 2.90 (2.43–3.43) | 1.18 (0.99–1.41) | 1.10 (0.92–1.32) |
| Secondary analyses | |||||
| Timing of exposure | |||||
| Unexposed | 23011 | 1428 (6.2) | 2.46 (2.34–2.59) | 1 | 1 |
| First trimester alone | 653 | 47 (7.2) | 2.92 (2.14–3.88) | 1.18 (0.89–1.58) | 1.08 (0.81–1.46) |
| Second trimester alone | 395 | 30 (7.6) | 3.22 (2.17–4.60) | 1.32 (0.92–1.77) | 1.23 (0.85–1.77) |
| Third trimester alone | 672 | 42 (6.3) | 2.53 (1.82–3.42) | 1.02 (0.76–1.42) | 1.05 (0.77–1.42) |
| Cumulative duration of use | |||||
| Unexposed | 23011 | 1428 (6.2) | 2.46 (2.34–2.59) | 1 | 1 |
| Less than 14 days | 1717 | 116 (6.8) | 2.75 (2.28–3.30) | 1.12 (0.93–1.35) | 1.05 (0.87–1.28) |
| At least 14 days | 182 | 18 (9.9) | 4.35 (2.58–6.87) | 1.80 (1.13–2.87) | 1.70 (1.05–2.76) |
| Cumulative opioid dosea | |||||
| Unexposed | 23011 | 1428 (6.1) | 2.46 (2.34–2.59) | 1 | 1 |
| Low dose (<37.5 MME) | 986 | 60 (6.1) | 2.73 (2.08–3.51) | 1.11 (0.86–1.44) | 1.05 (0.81–1.37) |
| High dose (≥37.5 MME) | 913 | 74 (8.1) | 3.05 (2.39–3.83) | 1.24 (0.98–1.57) | 1.22 (1.01–1.54) |
| Individual opioids | |||||
| Unexposed | 23011 | 1428 (6.2) | 2.46 (2.34–2.59) | 1 | 1 |
| Hydrocodone monotherapy | 836 | 70 (8.4) | 3.42 (2.66–4.32) | 1.39 (1.09–1.77) | 1.33 (1.04–1.70) |
| Codeine monotherapy | 708 | 41 (5.8) | 2.35 (1.68–3.18) | 0.96 (0.70–1.30) | 0.93 (0.68–1.28) |
Fetal cumulative opioid exposure was categorized by median split of the total opioid doses during pregnancy. The low and high categories correspond to 2.25–37.49 MME and 37.5–2250 MME, respectively
HR hazard ratio, PY person-years, n number, IR incidence rate, MME morphine milligram equivalent
3.3. Secondary Analyses
The association between maternal opioid use and neurodevelopmental disorders was significantly increased in children exposed to high cumulative opioid doses or longer duration of opioid exposure during pregnancy (adjusted HRs 1.22 [1.01–1.54] and 1.70 [1.07–2.70], respectively) (Table 3). Although we intended to assess the relationship between each opioid molecule included in our analyses and the incidence of outcomes, due to limited sample size this specific analysis was limited to hydrocodone and codeine. When we assessed the effects of individual opioids, children born to mothers on hydrocodone monotherapy during pregnancy had 1.3 times the risk of neurodevelopmental disorder in early childhood (adjusted HR 1.33 [1.04–1.70]) (Table 3).
3.4. Sensitivity Analyses
The results from sensitivity analyses including maternal opioid use by specific trimester, alternative exposure and more stringent outcome definitions, or restriction to children with ≥3 years or at least 5 years of follow-up remained largely consistent with primary findings; that is, no association between maternal opioid use and increased risk of neurodevelopmental disorders in early childhood (Table 4). Our primary conclusion also remained unchanged when we accounted for confounding by underlying indication; that is, restricted analyses in children whose mothers had pain conditions, mental disorders, or received psychotropic medications during pregnancy (adjusted HR 1.16 [0.89–1.50]).
Table 4.
Sensitivity analyses of the risk of neurodevelopmental disorders in children by maternal opioid use
| Analyses | Total number | Case, n (%) | Crude IR (95% CI) per 100 PY | Crude HR (95% CI) | Adjusted HR (95% CI) |
|---|---|---|---|---|---|
|
| |||||
| Alternative exposure definition | |||||
| Exposure defined as ≥2 fills of opioid prescription | |||||
| Unexposed | 23011 | 1428 (6.2) | 2.46 (2.34–2.59) | 1 | 1 |
| Exposed | 361 | 31 (8.6) | 3.60 (2.45–5.12) | 1.48 (1.04–2.11) | 1.25 (0.86–1.81) |
| Exposure defined based on overlapping days of supply | |||||
| Unexposed | 22933 | 1426 (6.2) | 2.47 (2.34–2.560) | 1 | 1 |
| Exposeda | 1957 | 136 (7.0) | 2.84 (2.38–3.36) | 1.15 (0.97–1.38) | 1.10 (0.92–1.32) |
| Alternative outcome definitions | |||||
| ≥2 claims for neurodevelopmental disorders | |||||
| Unexposed | 23011 | 1137 (4.9) | 1.96 (1.85–2.08) | 1 | 1 |
| Exposed | 1899 | 103 (5.4) | 2.23 (1.82–2.70) | 1.14 (0.93–1.39) | 1.07 (0.87–1.31) |
| ≥2 neurodevelopmental disorder claims on ≥2 separate days | |||||
| Unexposed | 23011 | 815 (3.5) | 1.41 (1.31–1.51) | 1 | 1 |
| Exposed | 1899 | 75 (3.9) | 1.62 (1.28–2.03) | 1.16 (0.91–1.46) | 1.09 (0.85–1.39) |
| Minimum follow-up time for outcome measure assessment | |||||
| Individuals with ≥3 years of follow-up | |||||
| Unexposed | 7490 | 570 (7.6) | 1.49 (1.37–1.62) | 1 | 1 |
| Exposed | 584 | 58 (9.9) | 1.97 (1.50–2.54) | 1.34 (1.03–1.76) | 1.32 (1.0–1.74) |
| Individuals with ≥5 years of follow-up | |||||
| Unexposed | 4422 | 174 (174) | 0.66 (0.57–0.77) | 1 | 1 |
| Exposed | 320 | 19 (5.9) | 1.0 (0.60–1.56) | 1.51 (0.94–2.42) | 1.42 (0.87–2.33) |
| Analysis in restricted cohort with diagnosis of pain or mental disorders, or receipt of psychostimulants | |||||
| Unexposed | 6270 | 422 (6.7) | 2.70 (2.45–2.97) | 1 | 1 |
| Exposed | 874 | 70 (8.0) | 3.33 (2.59–4.20) | 1.24 (0.96–1.59) | 1.16 (0.89–1.50) |
Infants were considered exposed if their mothers filled at least one prescription of opioids anytime that overlapped with the pregnancy window
HR hazards ratio, IR incidence rate, PY person-years
The 95% CI of the primary endpoint includes the null (i.e., HR 1.10 [0.92–1.32]), therefore no further unmeasured confounding is needed to move the CI to include the null value of 1. Next, we calculated the non-null E-value needed to shift the lower limit of the 95% CI from 0.92 to 1.01 [43]. We report that an unmeasured confounder that is associated with both maternal opioid use and neurodevelopmental disorders by a risk ratio of 1.40 each could shift the lower 95% CI to exclude the null, however weaker confounding could not.
4. Discussion
The primary findings from this study involving pregnant women with no diagnosis of OUDs or drug dependence showed that maternal opioid use anytime in pregnancy or during specific pregnancy trimesters is not associated with an increase in risk of neurodevelopmental disorders in early childhood. However, children whose mothers received ≥14 days’ cumulative supply or higher doses of opioids during pregnancy had increased risk of neurodevelopmental disorders. Overall, the findings from our secondary analyses suggest the potential dangers of longer duration and higher doses of prenatal opioid exposure on children’s neurodevelopment and are consistent with the known dose/exposure effects of teratogens on risks for malformations and neurodevelopmental outcomes. Since the majority of prescription opioid use in pregnancy is reportedly to manage acute pain [2], and is often prescribed at lower doses or for acute use (which we also observed in the current study) [42], the findings of significantly increased risks of neurodevelopmental disorders at high doses or longer duration of cumulative exposure have particularly important clinical implications in women with chronic pain conditions or at risk of opioid use disorders. Further research is needed, however, to confirm these findings from our secondary analyses.
The exact mechanism linking prenatal opioid exposure to neurodevelopmental disorders is not fully delineated. However, opioids are known to readily cross the placenta, are present in breast milk of opioid-exposed mothers, and have been reported to be capable of altering fetal brain development, particularly in women with a history of substance abuse [10, 44]. Volumetric measurements of the cerebral characteristics of children exposed to opiates during pregnancy indicated smaller intracranial and brain volumes including the cerebral cortex, amygdala, brainstem, cerebellar cortex, and thinner cortex of the right lateral orbitofrontal cortex compared with unexposed infants [45, 46]. Alterations of the hippocampal cholinergic system and opiate receptor system have also been hypothesized in literature [44, 47–49]. More so, morphine induces apoptosis of the human microglia and neurons [44]. Since several regions of the brain such as the limbic system, midbrain, and thalamus contain the highest concentration of opioid receptors, it has been postulated that inappropriate activation of these receptors can alter the normal fetal brain development [50].
So far, evidence on this potential safety issue in human clinical studies has largely accrued from studies evaluating illicit heroin or methadone, mothers with diagnosis of OUDs, and infants with neonatal abstinence syndrome (eTable 1, see ESM). The overall findings from these individual studies (eTable 1) and their resulting meta-analyses [51–53] indicate that in pregnant women with OUDs or drug dependence, maternal opioid use was associated with significantly increased risk of early childhood neurodevelopmental disorders. Since we excluded mothers with history of OUDs, it is difficult to index our primary findings to the studies previously reported in literature. More so, ample evidence from literature indicate risk factors like smoking, alcohol use, or socioeconomic factors like disruptions in maternal care that can markedly increase the risk of neurodevelopmental disorders in the developing fetus are more prevalent in mothers with OUD or opioid dependence compared with women with no diagnosis of OUD or opioid dependence during pregnancy [17, 19, 20]. It may also be the case that compared with offspring of mothers with OUD, children of non-OUD or drug-dependent mothers have higher parental education, better home environment and caregiving or parenting abilities that may help exposed infants reduce or overcome poor neurodevelopmental outcomes in early childhood [54, 55].
The association between maternal opioid exposure and the risk of neurodevelopmental disorders in the general population is largely unknown. To our knowledge, results at 3 and 5 years old from the Norwegian Mother and Child Cohort (MoBa) study by Skovlund et al. represent the only other large population-based reports evaluating this safety concern [56, 57]. The authors reported no association between prenatal opioid exposure and impairments in language competencies and communication skills in children at 3 or 5 years of age but noted that their study could not address opioid use with large doses or long duration during pregnancy. While the results of our overall analysis are similar to the two reports by Skovlund et al., the current study is different in a few ways. Unlike in the Skovlund et al. studies, we did not rely on self-reports to assess exposure status or to identify neurodevelopmental outcomes in children. Rather, we used the presence of diagnosis codes available in medical records of children to identify neurodevelopmental disorders and assessed potential exposure using maternal prescription data. While parental self-reporting is generally regarded as a valid tool for measuring neurocognitive disorders in children [58, 59], there is increased propensity for recall bias with self-reported exposure. Moreover, differences between self-reported use of opioids during pregnancy compared with dispensing information available in prescription records has been reported in the MoBa cohort [60]. Additionally, neurodevelopmental disorders in early childhood can occur as a constellation of overlapping symptoms such that restricting to language disorders, while clinically relevant, may be too restrictive [32, 36, 61].
The current study has several strengths. First, it provides updated information on opioid utilization during pregnancy and extends the limited knowledge in literature on the risks associated with in-utero opioid exposure and early childhood neurodevelopment in offspring of non-OUD mothers. Second, we utilized a relatively large population size, utilized a specific exposure definition and diagnoses for neurodevelopmental disorders as well as performed several additional analyses to mitigate confounding and to assess the robustness of our findings. Most importantly, our study provides much-needed data on the relationship between cumulative dose, timing of exposure, individual opioids, and duration of use of opioids during pregnancy and the risk of neurodevelopmental disorders in children; information otherwise missing in the literature.
There are several limitations in the current study. Considering that administrative claims are not always reported with fidelity, various forms of measurement errors including exposure misclassification may have occurred. For example, we cannot completely guarantee that dispensed opioids were taken, or even how they were taken. This issue of measurement error is particularly pertinent considering opioids were infrequently prescribed in our study. Nevertheless, the observed strength of association between our primary analysis and overall conclusion did not appreciably change with alternative or more conservative exposure and outcome definitions.
Second, we assessed the presence of neurodevelopmental disorders in early childhood as a dichotomous measure. This approach, while similar to other cohort studies utilizing health insurance data to assess neurological functioning or mental disorders, may miss cognitive declines that do not rise to the level of a diagnosis and may have less power compared with objective continuous measures of cognitive abilities such as intelligence quotient (IQ). For example, in the NEAD study that demonstrated robust effects of valproate on IQ, an analysis for the dichotomous measure of cognitive performance (i.e., <70 IQ) was not significant [62]. As such, future studies involving objective measures of cognitive performance are needed to further evaluate this potential safety concern.
Despite our best efforts to address confounding by known and suspected risk factors and their proxies in our analyses, our study results may still be impacted by residual or unmeasured confounding arising from inadequate adjustments or incomplete information on important confounders such as maternal use of folic acid, socioeconomic factors, lifestyle factors such as alcohol drinking during pregnancy, severity of pain, and heritability. This is perhaps evident in our E-value. Although we also planned to perform a nested sibling comparison to further adjust for unmeasured genetic and environmental time-invariant confounders, the limited number of siblings in our study (n = 360) precluded this analysis.
Additionally, our findings may have been impacted by the relatively short duration of follow-up; the median (IQR) and mean (SD) follow-up times were 725 (271–1665) and 964 (784) days, respectively). Therefore, some relevant neurodevelopmental outcomes may go undetected in early infancy until later in childhood. Conversely, we may have inaccurately included false positives; that is, individuals who experienced improvements in intellectual and cognitive function later in childhood and later in life. As such, interpretation of the current study is limited to neurodevelopmental disorders diagnosed in infancy or in early childhood, not during adolescence or adulthood.
Finally, our study population is composed predominantly of mothers and children potentially of higher socioeconomic status or those eligible for enrollment in private health insurance plans in the US. Therefore, our findings may not be generalizable to uninsured or Medicaid-enrolled mothers who are typically of lower socioeconomic status and thus may have a higher prevalence of risk factors predisposing children to higher risks of neurodevelopmental disorders. This specific limitation is, however, unrelated to the internal validity of our study.
5. Conclusions
Overall, in mothers with no diagnosis of opioid use disorders or drug dependence, maternal opioid use was not associated with increased risk of neurodevelopmental disorders in early childhood. However, significantly increased risk of neurodevelopmental disorders were observed in offspring of mothers prescribed opioids for a longer period or at a higher cumulative dose during pregnancy, thus emphasizing the need for careful evaluation of risks versus potential benefits, particularly in pregnant women requiring long-term or high-dose opioid therapy. Further research is needed to confirm these findings.
Supplementary Material
Key Points.
In this retrospective cohort study comprising mothers with no diagnosis of opioid use disorders or drug dependence, maternal opioid use was not associated with an increase in the risk of neurodevelopmental disorders in early childhood.
However, compared with no exposure, children born to mothers receiving high doses or long-term opioid use were associated with an increased risk of neurodevelopmental disorders.
Further research is needed to assess the risks versus potential benefits of prescription opioids in pregnant women without opioid use disorders or drug dependence requiring long-term or high doses of opioid therapy.
Funding/support
Research reported in this publication was supported by the Eunice Kennedy Shriver National Institute Of Child Health & Human Development of the National Institutes of Health under Award Number R15HD097588. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Principal Investigator: Wen X. The funding source had no role in the design or conduct of the study; collection, management, analysis, or the interpretation of the data preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.
Conflicts of interest and financial disclosures
Dr Meador has received research support from the National Institutes of Health and Sunovion Pharmaceuticals, and travel support from Eisai. The Epilepsy Study Consortium pays Dr Meador’s university for his research consultant time related to Eisai, GW Pharmaceuticals, NeuroPace, Novartis, Supernus, Upsher-Smith Laboratories, and UCB Pharma. The other authors have no conflicts of interests to declare.
Footnotes
Ethics approval This study uses deidentified data, hence was considered exempt by the University of Rhode Island Institutional Review Board.
Consent to participate Not applicable.
Consent for publication Not applicable.
Availability of data and material A data use agreement exists between The University of Rhode Island College of Pharmacy and Optum. The Optum Clinformatics Data Mart database cannot be shared by authors with parties external to this agreement.
Code availability Not applicable.
Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s40264-021-01080-0.
References
- 1.Desai RJ, Huybrechts KF, Hernandez-Diaz S, Mogun H, Patorno E, Kaltenbach K, et al. Exposure to prescription opioid analgesics in utero and risk of neonatal abstinence syndrome: population based cohort study. BMJ. 2015;350:h2102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Yazdy MM, Mitchell AA, Tinker SC, Parker SE, Werler MM. Periconceptional use of opioids and the risk of neural tube defects. Obstet Gynecol. 2013;122:838–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Broussard CS, Rasmussen SA, Reefhuis J, Friedman JM, Jann MW, Riehle-Colarusso T, et al. Maternal treatment with opioid analgesics and risk for birth defects. Am J Obstet Gynecol. 2011;204(314):e1–11. [DOI] [PubMed] [Google Scholar]
- 4.Nørgaard M, Nielsson MS, Heide-Jørgensen U. Birth and neonatal outcomes following opioid use in pregnancy: a Danish population-based study. Subst Abuse. 2015;9:5–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Patrick SW, Dudley J, Martin PR, Harrell FE, Warren MD, Hartmann KE, et al. Prescription opioid epidemic and infant outcomes. Pediatrics. 2015;135:842–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Salihu HM, Salemi JL, Aggarwal A, Steele BF, Pepper RC, Mogos MF, et al. Opioid drug use and acute cardiac events among pregnant women in the United States. Am J Med. 2018;131:64–71.e1. [DOI] [PubMed] [Google Scholar]
- 7.Bajic D, Commons KG, Soriano SG. Morphine-enhanced apoptosis in selective brain regions of neonatal rats. Int J Dev Neurosci. 2013;31:258–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Golalipour MJ, Ghafari S. Purkinje cells loss in off spring due to maternal morphine sulfate exposure: a morphometric study. Anat Cell Biol. 2012;45:121–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang Y, Han T-Z. Prenatal exposure to heroin in mice elicits memory deficits that can be attributed to neuronal apoptosis. Neuroscience. 2009;160:330–8. [DOI] [PubMed] [Google Scholar]
- 10.Ross EJ, Graham DL, Money KM, Stanwood GD. Developmental consequences of fetal exposure to drugs: what we know and what we still must learn. Neuropsychopharmacology. 2015;40:61–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Epstein RA, Bobo WV, Martin PR, Morrow JA, Wang W, Chandrasekhar R, et al. Increasing pregnancy-related use of prescribed opioid analgesics. Ann Epidemiol. 2013;23:498–503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Desai RJ, Hernandez-Diaz S, Bateman BT, Huybrechts KF. Increase in prescription opioid use during pregnancy among Medicaid-enrolled women. Obstet Gynecol. 2014;123:997–1002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Knight J, Cassell CH, Meyer RE, Strauss RP. Academic outcomes of children with isolated orofacial clefts compared with children without a major birth defect. Cleft Palate Craniofac J. 2015;52:259–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Steinlin M. Cerebellar disorders in childhood: cognitive problems. Cerebellum. 2008;7:607–10. [DOI] [PubMed] [Google Scholar]
- 15.Hunt RW, Tzioumi D, Collins E, Jeffery HE. Adverse neurodevelopmental outcome of infants exposed to opiate in-utero. Early Hum Dev. 2008;84:29–35. [DOI] [PubMed] [Google Scholar]
- 16.Moe V. Foster-placed and adopted children exposed in utero to opiates and other substances: prediction and outcome at four and a half years. J Dev Behav Pediatr. 2002;23:330–9. [DOI] [PubMed] [Google Scholar]
- 17.Ornoy A. The impact of intrauterine exposure versus postnatal environment in neurodevelopmental toxicity: long-term neurobehavioral studies in children at risk for developmental disorders. Toxicol Lett. 2003;140–141:171–81. [DOI] [PubMed] [Google Scholar]
- 18.Bunikowski R, Grimmer I, Heiser A, Metze B, Schäfer A, Obladen M. Neurodevelopmental outcome after prenatal exposure to opiates. Eur J Pediatr. 1998;157:724–30. [DOI] [PubMed] [Google Scholar]
- 19.Hans SL, Jeremy RJ. Postneonatal mental and motor development of infants exposed in utero to opioid drugs. Infant Ment Health J. 2001;22:300–15. [Google Scholar]
- 20.Ornoy A, Michailevskaya V, Lukashov I, Bar-Hamburger R, Harel S. The developmental outcome of children born to heroin-dependent mothers, raised at home or adopted. Child Abuse Negl. 1996;20:385–96. [DOI] [PubMed] [Google Scholar]
- 21.Speltz ML, Endriga MC, Hill S, Maris CL, Jones K, Omnell ML. Brief report: Cognitive and psychomotor development of infants with orofacial clefts. J Pediatr Psychol Oxf Acad. 2000;25:185–90. [DOI] [PubMed] [Google Scholar]
- 22.Optum Inc. Retrospective Database Analysis [Internet]. Retrospective database analysis. 2013. Available from: https://www.optum.com/content/dam/optum/resources/productSheets/Retrospective-Database-Analysis.pdf
- 23.Palmsten K, Huybrechts KF, Mogun H, Kowal MK, Williams PL, Michels KB, et al. Harnessing the Medicaid Analytic eXtract (MAX) to Evaluate Medications in Pregnancy: Design Considerations. PLoS ONE. 2013;8:e67405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Bandoli G, Chambers CD, Wells A, Palmsten K. Prenatal Antidepressant Use and Risk of Adverse Neonatal Outcomes. Pediatrics [Internet]. American Academy of Pediatrics; 2020. [cited 2020 Jul 6];146. Available from: https://pediatrics.aappublications.org/content/146/1/e20192493 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Veeravalli N, Thayer S, Bandoli G. An Unbreakable Bond: Linking Mothers and Newborns in Large, De-Identified Claims Database. AcademyHealth; 2019. [cited 2021 Feb 21]. Available from: https://academyhealth.confex.com/academyhealth/2019arm/meetingapp.cgi/Paper/31468 [Google Scholar]
- 26.Margulis AV, Setoguchi S, Mittleman MA, Glynn RJ, Dormuth CR, Hernández-Díaz S. Algorithms to estimate the beginning of pregnancy in administrative databases. Pharmacoepidemiol Drug Saf. 2013;22:16–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Chastek B, Harley C, Kallich J, Newcomer L, Paoli CJ, Teitelbaum AH. Health care costs for patients with cancer at the end of life. J Oncol Pract. 2012;8:75s–80s. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Haight SC, Ko JY, Tong VT, Bohm MK, Callaghan WM. Opioid use disorder documented at delivery hospitalization United States, 1999–2014. MMWR Morb Mortal Wkly Rep. 2018;67:845–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Dowell D, Haegerich TM, Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain—United States, 2016. JAMA. 2016;315:1624–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Opioid morphine equivalent conversion factors [Internet]. [cited 2020 Jul 6]. Available from: https://www.cms.gov/Medicare/Prescription-Drug-Coverage/PrescriptionDrugCovContra/Downloads/Opioid-Morphine-EQ-Conversion-Factors-March-2015.pdf
- 31.Halloran DR, Swindle J, Takemoto SK, Schnitzler MA. Multiple psychiatric diagnoses common in privately insured children on atypical antipsychotics. Clin Pediatr (Phila). 2010;49:485–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Thapar A, Cooper M, Rutter M. Neurodevelopmental disorders. Lancet Psychiatry. 2017;4:339–46. [DOI] [PubMed] [Google Scholar]
- 33.Doernberg E, Hollander E. Neurodevelopmental Disorders (ASD and ADHD): DSM-5, ICD-10, and ICD-11. CNS Spectr. 2016;21:295–9. [DOI] [PubMed] [Google Scholar]
- 34.Chakrabarti S, Fombonne E. Pervasive developmental disorders in preschool children. JAMA. 2001;285:3093–9. [DOI] [PubMed] [Google Scholar]
- 35.DiMaggio C, Sun LS, Kakavouli A, Byrne MW, Li G. A retrospective cohort study of the association of anesthesia and hernia repair surgery with behavioral and developmental disorders in young children. J Neurosurg Anesthesiol. 2009;21:286–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.American Psychiatric Association. Neurodevelopmental Disorders. Diagnostic and Statistical Manual of Mental Disorders [Internet]. 5th ed. American Psychiatric Association; 2013. [cited 2021 Feb 21]. Available from: https://dsm.psychiatryonline.org/ doi/abs/https://doi.org/10.1176/appi.books.9780890425596.dsm01 [Google Scholar]
- 37.Bateman BT, Mhyre JM, Hernandez-Diaz S, Huybrechts KF, Fischer MA, Creanga AA, et al. Development of a comorbidity index for use in obstetric patients. Obstet Gynecol. 2013;122:957–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Kleinbaum DG, Klein M. Survival Analysis: A Self-Learning Text, Third Edition [Internet]. 3rd ed. New York: Springer-Verlag; 2012. [cited 2019 Jun 15]. Available from: https://www.springer.com/us/book/9781441966452 [Google Scholar]
- 39.Desai RJ, Rothman KJ, Bateman BT, Hernandez-Diaz S, Huybrechts KF. A propensity-score-based fine stratification approach for confounding adjustment when exposure is infrequent. Epidemiology. 2017;28:249–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28:3083–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Louis GB, Dukic V, Heagerty PJ, Louis TA, Lynch CD, Ryan LM, et al. Analysis of repeated pregnancy outcomes. Stat Methods Med Res. 2006;15:103–26. [DOI] [PubMed] [Google Scholar]
- 42.Bateman BT, Hernandez-Diaz S, Rathmell JP, Seeger JD, Doherty M, Fischer MA, et al. Patterns of opioid utilization in pregnancy in a large cohort of commercial insurance beneficiaries in the United States. Anesthesiology. 2014;120:1216–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med. 2017;167:268–74. [DOI] [PubMed] [Google Scholar]
- 44.Šlamberová R. Drugs in pregnancy: the effects on mother and her progeny. Physiol Res. 2012;61(Suppl 1):S123–135. [DOI] [PubMed] [Google Scholar]
- 45.Walhovd KB, Moe V, Slinning K, Due-Tønnessen P, Bjørnerud A, Dale AM, et al. Volumetric cerebral characteristics of children exposed to opiates and other substances in utero. Neuroimage. 2007;36:1331–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Sirnes E, Oltedal L, Bartsch H, Eide GE, Elgen IB, Aukland SM. Brain morphology in school-aged children with prenatal opioid exposure: a structural MRI study. Early Hum Dev. 2017;106–107:33–9. [DOI] [PubMed] [Google Scholar]
- 47.Yaniv SP, Naor Z, Yanai J. Prenatal heroin exposure alters cholinergic receptor stimulated activation of the PKCbetaII and PKCgamma isoforms. Brain Res Bull. 2004;63:339–49. [DOI] [PubMed] [Google Scholar]
- 48.Steingart RA, Abu-Roumi M, Newman ME, Silverman WF, Slotkin TA, Yanai J. Neurobehavioral damage to cholinergic systems caused by prenatal exposure to heroin or phenobarbital: cellular mechanisms and the reversal of deficits by neural grafts. Brain Res Dev Brain Res. 2000;122:125–33. [DOI] [PubMed] [Google Scholar]
- 49.Steingart RA, Silverman WF, Barron S, Slotkin TA, Awad Y, Yanai J. Neural grafting reverses prenatal drug-induced alterations in hippocampal PKC and related behavioral deficits. Brain Res Dev Brain Res. 2000;125:9–19. [DOI] [PubMed] [Google Scholar]
- 50.Vathy I. Prenatal opiate exposure: long-term CNS consequences in the stress system of the offspring. Psychoneuroendocrinology. 2002;27:273–83. [DOI] [PubMed] [Google Scholar]
- 51.Nelson LF, Yocum VK, Patel KD, Qeadan F, Hsi A, Weitzen S. Cognitive outcomes of young children after prenatal exposure to medications for opioid use disorder: a systematic review and meta-analysis. JAMA Netw Open. 2020;3:e201195. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Lee SJ, Bora S, Austin NC, Westerman A, Henderson JMT. Neurodevelopmental outcomes of children born to opioid-dependent mothers: a systematic review and meta-analysis. Acad Pediatr. 2020;20:308–18. [DOI] [PubMed] [Google Scholar]
- 53.Yeoh SL, Eastwood J, Wright IM, Morton R, Melhuish E, Ward M, et al. Cognitive and motor outcomes of children with prenatal opioid exposure: a systematic review and meta-analysis. JAMA Netw Open. 2019;2:e197025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Nair P, Black MM, Ackerman JP, Schuler ME, Keane VA. Children’s cognitive-behavioral functioning at age 6 and 7: prenatal drug exposure and caregiving environment. Ambul Pediatr. 2008;8:154–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Suchman NE, McMahon TJ, Slade A, Luthar SS. How early bonding, depression, illicit drug use, and perceived support work together to influence drug-dependent mothers’ caregiving. Am J Orthopsychiatry. 2005;75:431–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56.Skovlund E, Handal M, Selmer R, Brandlistuen RE, Skurtveit S. Language competence and communication skills in 3-year-old children after prenatal exposure to analgesic opioids. Pharmacoepidemiol Drug Saf. 2017;26:625–34. [DOI] [PubMed] [Google Scholar]
- 57.Skovlund E, Selmer R, Skurtveit S, Brandlistuen RE, Handal M. In utero exposure to analgesic opioids and language development in 5-year old children. Pharmacoepidemiol Drug Saf. 2020. [DOI] [PubMed] [Google Scholar]
- 58.Feldman HM, Campbell TF, Kurs-Lasky M, Rockette HE, Dale PS, Colborn DK, et al. Concurrent and predictive validity of parent reports of child language at ages 2 and 3 years. Child Dev. 2005;76:856–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Dale PS, Price TS, Bishop DVM, Plomin R. Outcomes of early language delay: I. Predicting persistent and transient language difficulties at 3 and 4 years. J Speech Lang Hear Res. 2003;46:544–60. [DOI] [PubMed] [Google Scholar]
- 60.Skurtveit S, Selmer R, Odsbu I, Handal M. Self-reported data on medicine use in the Norwegian Mother and Child cohort study compared to data from the Norwegian Prescription Database. 1 [Internet]. 2014. [cited 2019 Jun 15];24. Available from: https://www.ntnu.no/ojs/index.php/norepid/article/view/1824
- 61.Thapar A, Rutter M. Neurodevelopmental disorders. Rutter’s Child and Adolescent Psychiatry [Internet]. John Wiley & Sons, Ltd; 2015. [cited 2021 Feb 21]. p. 31–40. Available from: 10.1002/9781118381953.ch3 [DOI] [Google Scholar]
- 62.Meador KJ, Baker GA, Browning N, Cohen MJ, Bromley RL, Clayton-Smith J, et al. Fetal antiepileptic drug exposure and cognitive outcomes at age 6 years (NEAD study): a prospective observational study. Lancet Neurol. 2013;12:244–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.