Prenatal prescription opioid analgesic exposure and academic performance in third grade children: An Australian population-based cohort study

Objective To evaluate whether prenatal exposure to opioids analgesics is associated with poor academic performance in third grade children. Design: Population-based cohort study using linked data from New South Wales birth records, medicine dispensing data, and national standardised test results. Setting: New South Wales, Australia (2003-2011). Population: Liveborn children of concessional women, excluding those whose mothers lived interstate, were overseas visitors, or had records of opioid dependence. Methods: Exposure was defined as ≥1 opioid dispensed during pregnancy stratified by pregnancy timing, dose, and type of monotherapy. Reading and numeracy z-scores were compared between children with and without prenatal opioid exposure using linear mixed-effects models incorporating propensity score weights, and accounting for maternal and school clustering. Main Outcome Measures: Standardised national reading and numeracy Z-scores for third-grade children. Results Among 85,478 eligible children, 70,882 (82.9%) had test scores, with 7,664 (10.2%) prenatally exposed to opioids analgesics, mainly codeine. Compared with unexposed children, differences in mean z-scores were small for reading and numeracy among children with any prenatal opioid exposure (adjusted β [aβ] -0.05, 95% CI -0.06 to -0.03 for both). Results for codeine and oxycodone were similar, but a greater difference was observed for tramadol (reading aβ -0.25, 95% CI -0.31 to -0.18; numeracy aβ -0.22, 95% CI -0.29 to -0.16). Conclusions Prenatal exposure to codeine and oxycodone do not have meaningful impacts on third-grade academic performance. It is unclear if the lower academic performance observed with tramadol exposure is attributable to unmeasured confounding or a true effect, warranting further investigation.

Pain is a commonly reported symptom during pregnancy (1). Due to limited alternatives, opioid analgesics are often used to treat pain during pregnancy, with global prevalence ranging from 4 to 191 per 1000 pregnancies (2). However, opioids freely cross the placenta resulting in fetal exposure. Animal studies have demonstrated that opioids inhibit neuronal development in the central nervous system (CNS) by altering growth factors and signalling molecules involved in neurogenesis and apoptosis (3-5). Human studies of infants with neonatal abstinence syndrome (NAS) following prenatal opioid exposure have reported reduced plasma concentrations of proteins critical for neuronal survival, growth, and plasticity, key processes underlying learning and memory (6, 7). Therefore, it is biologically plausible that prenatal opioid exposure could disrupt fetal CNS development, contributing to later difficulties with academic skills such as reading and numeracy. Three studies from the Norwegian Mother, Father, and Child Cohort (MoBa) have examined prenatal opioid analgesic exposure and childhood educational outcomes, finding no clinically meaningful increased risk of impaired language development or academic performance (8-10). However, none of the studies explored the effects of cumulative dose nor individual opioid monotherapies which may have different risks. This highlights the need for further research to replicate these findings in other cohorts and expand our understanding of the long-term safety of opioid analgesics during pregnancy. Our study investigated whether prenatal exposure to opioid analgesics is associated with poor academic performance on reading and numeracy tests in third grade children. We also aimed to examine whether these associations varied by opioid type, timing of exposure, or cumulative dose. Study Setting, Study Design and Data Sources: We conducted a population-based cohort study of all liveborn children (≥400g or ≥20 weeks gestation) in New South Wales (NSW), Australia, using the Early Life Course (ELC) data (11). The ELC dataset comprises linked administrative records including perinatal data, school performance records, social security information, hospital admissions, emergency department (ED) presentations, records of opioid agonist treatment (OAT) for opioid dependence (OD), mortality data, and dispensing records for medicines subsidised through the Pharmaceutical Benefits Scheme (PBS). See eAppendix 1 for a more detailed description of the data sources. We report our study in accordance with Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (eTable 9). Study Population We identified all liveborn children in NSW whose mothers had an estimated last menstrual period (LMP; date of childbirth – gestational age at delivery [weeks] x 7) between January 1, 2003, and March 26, 2011. The study’s end date was selected to ensure that most children reached the required age of (8-9 years) for participation in third grade national standardised testing and to avoid overrepresenting pregnancies shorter than 40 weeks’ gestation. Our cohort was restricted to children born to women with continuous concessional beneficiary status from 365 days before their LMP until childbirth. We restricted to this population because before July 2012, the PBS database only recorded medicines dispensed that cost more than the co-payment threshold. The Australian government subsidises PBS-listed medicines when their prices are above these thresholds. Concessional beneficiaries, including social security recipients, low-income earners, individuals with disabilities, and caregivers, pay a reduced co-payment ($7.70 in 2024) compared to general beneficiaries ($31.60). Since all medicines cost more than the concessional co-payment threshold, complete dispensing data were only available for concessional beneficiaries. We excluded children with missing or implausible gestational age, those whose mothers were interstate residents, overseas visitors, or had an OD diagnosis or received OAT from 12 months before LMP until childbirth (eTable 1 – exclusion criteria codes). eFigure 1 illustrates the study design. Exposure We included all PBS-subsidised opioid analgesics available during the study period (eTable 2). During the study period, methadone and buprenorphine were PBS-subsidised only for the indication of pain. Children were classified as prenatally exposed if their mother was dispensed ≥1 opioid analgesic during pregnancy (LMP until childbirth), and unexposed if no opioid was dispensed from 90 days before LMP through childbirth. As synaptogenesis is known to peak during late pregnancy (12), we examined exposure timing across distinct gestational periods: early pregnancy only (LMP through until gestational week 20 with no late pregnancy exposure), late pregnancy only (gestational week 20 until childbirth with no early pregnancy exposure), and both early and late pregnancy exposure. To examine effects of the most common individual opioids, we stratified exposure into monotherapy groups for codeine (including combinations with paracetamol or aspirin), oxycodone (including oxycodone-naloxone), and tramadol, where women received only that specific opioid type during pregnancy. We also evaluated cumulative opioid dose by calculating total oral morphine equivalents (OME), dichotomised into low (<60 mg) OME and higher (≥60 mg) OME. We calculated total OME by multiplying each dispensed opioid’s strength (mg), quantity, and OME conversion factor (13), then summing the OME for all opioids dispensed during pregnancy. Academic Performance Academic performance was assessed through third grade reading and numeracy scores from the National Assessment Program - Literacy and Numeracy (NAPLAN), an annual standardised assessment administered nationwide by each state and territory (in our case NSW) since 2008 that evaluates students’ proficiency in reading, writing, language conventions, and numeracy. We focused on reading and numeracy scores as these domains have demonstrated the highest stability across multiple testing periods (14). We calculated z-scores using the mean and standard deviation (SD) for the first available third-grade test results for all children in our cohort who completed each subject in a given assessment year. As a secondary outcome, we examined the performance of third-grade students who scored below the National Minimum Standard (NMS). These students lack the fundamental skills needed for academic progression and are typically targeted for educational interventions (15). Covariates Covariates included maternal, parental, and child-related factors, with the corresponding codes, data sources, and ascertainment windows detailed in eTable 3. Maternal factors included sociodemographics, comorbidities, medicine use, and pre-pregnancy healthcare utilisation as a marker of a woman’s overall comorbidity burden. Parental characteristics included highest education level and occupation of first and second parent (if available). Child-related factors comprised birth year, sex, year of testing, and primary language spoken at home. A directed acyclic graph informed the identification of confounders (eFigure 2). Missing data were addressed by either assigning a ’missing’ category or imputing values using the mode or median (eTable 3). Follow-up and Missing Test Scores Follow-up commenced at childbirth and continued until the earliest occurrence of an outcome, death, or the administrative end of the study (31 st December 2019). Children who died before age nine were classified as deceased; those younger than nine years old at the last available test date (31st May 2019) were classified as too young for testing. Missing test scores occurred in two scenarios: non-participation in NAPLAN tests (absences, withdrawals, or exemptions for intellectual disability, significant co-existing conditions, or limited English proficiency), and no NAPLAN records (classified as lost to follow-up). Statistical Analysis Among children with prenatal exposure to opioid analgesics, we characterised exposure profiles by examining the proportion with 1, 2, or ≥3 dispensings and reporting median and interquartile (IQR) range of total OME, stratified by opioid type. We compared covariate characteristics between: 1) children with prenatal opioid exposure (any or specific types) and unexposed children, 2) children with test scores and those lost to follow-up to evaluate attrition bias, and 3) children included in our cohort and those born to women who did not meet the continuous concessional beneficiary definition (i.e, representing the remainder of the NSW population) to evaluate generalisability. For these comparisons, an absolute standardized mean difference >0.1 indicated meaningful group differences. Primary Analysis We assessed the total effect of prenatal opioid exposure on either reading or numeracy z-scores using mixed-effects linear regression models with random intercepts for school and maternal clustering. Models yielded beta (β) coefficients with 95% confidence intervals calculated using Satterthwaite’s approximation (16). To control for confounding and ensure balance across maternal, paternal and child-related characteristics, we applied inverse probability weighting incorporating all covariates listed in the covariate section, including those related to the outcome regardless of their association with the exposure (17). Propensity scores were calculated using logistic regression, with maternal age modelled as a natural cubic spline with 4 degrees of freedom, selected based on the Akaike Information Criterion. Beta coefficients >0.2 were considered indicative of meaningful academic differences (18). For our secondary outcome, we calculated the relative risk (RR) of performing below the NMS using Poisson regression with log-link function. Models used generalized estimating equations with an exchangeable correlation structure to account for school clustering and adjusted analyses incorporated propensity score weights. All analyses were restricted to children with available test scores. Sensitivity Analysis We conducted several sensitivity analyses to explore the robustness of our findings. To evaluate the impact of exposure misclassification, we redefined exposure as ≥2 dispensing records during pregnancy. To examine potential unmeasured confounding, we redefined the unexposed group as children whose mothers discontinued opioid use before pregnancy (i.e., ≥ 1 opioid dispensing between 365 and 90 days prior to the LMP, with no dispensing from 90 days before LMP through the end of pregnancy) and conducted a sibling analysis restricted to exposure-discordant siblings born to the same mother. To assess generalisability, we recalculated z-scores for an expanded cohort including children born to both concessional and children born to women not meeting the continuous concessional definition. Finally, we assessed potential selection bias from restricting analyses to children with test scores using two approaches: first, by including exempt students (assigning them z-scores 2 SD below the mean for their respective exposure group or classifying them as below NMS), and second, by applying inverse probability of censoring weights (IPCW) to account for all missing test data, including non-participants (exempt, absent, or withdrawn) and those lost to follow-up (methods in eAppendix 2). Post-hoc analyses We compared academic outcomes in children exposed to low versus higher OME during pregnancy to assess the dose-response relationship. For the potential safety signal associated with tramadol, we calculated the E-value to estimate the strength needed for an unmeasured confounder to explain observed associations (19). Finally, recognising that prenatal opioid exposure may increase the risk of adverse birth outcomes (20-23), which are risk factors for poor childhood development and academic performance (24, 25), we examined the potential role of these outcomes in the association between prenatal opioid exposure and academic achievement. We compared both crude and standardised proportions of adverse birth outcomes across exposure groups, with all estimates accounting for maternal clustering and standardized proportions incorporating propensity score weights based on all covariates except child sex and test year, as these are not confounders in the exposure-outcome pathway (eTable 4). Ethics Approvals The study received approval from the Australian Institute of Health and Welfare Human Research Ethics Committee (EO2020-2-1130) and the New South Wales Population and Health Services Research Ethics Committee (2019/ETH11830).

There were 85,478 children born to concessional beneficiaries (64,812 unique mothers; 84,242 pregnancies) who were eligible to be included in the study cohort (Figure 1), of whom 70,882 (82.9%) children had test scores. There was no differential pattern of missing data between exposure groups among NAPLAN non-participants and those lost to follow-up (Figure 1). Furthermore, children lost to follow-up had similar characteristics to those with test scores (eTable 5). Among children who had test scores, 7,116 (10.0%) were prenatally exposed to analgesic opioids; 3,084 (4.4%) in early pregnancy, 3,083 (4.3%) in late pregnancy, and 949 (1.3%) in both early and late pregnancy. Codeine was the most common monotherapy (89.6% of exposures), followed by tramadol (3.5%) and oxycodone (2.3%). Most children were exposed to a single opioid dispensing during pregnancy (72.0%), with 14.8% exposed to two and 13.2% to three or more (eFigure 3). Children prenatally exposed to tramadol had the highest median total OME (400 mg, IQR: 200 - 600), followed by oxycodone (150 mg, IQR: 150 - 300) and codeine (60 mg, IQR: 60 - 60) (eFigure 4). Mothers of children who were prenatally exposed to opioids were more likely to smoke during pregnancy, have mental health conditions, use both opioid (pre-pregnancy) and non-opioid analgesics, and have higher pre-pregnancy healthcare utilisation compared to mothers of children not exposed (Table 1, eTable 6). After propensity score weighting, all covariates were balanced (eFigure 5). Differences in Academic Performance Compared with unexposed children, the adjusted differences in mean z-score for those prenatally exposed to any opioid were −0.05 (95% CI, −0.06 to −0.03) for both reading and numeracy; for codeine, −0.04 (−0.06 to −0.03) for both reading and numeracy; and for oxycodone, −0.01 (−0.07 to 0.09) for reading and −0.02 (−0.06 to 0.09) for numeracy (Figures 2–3). Differences in mean z-scores were larger with exposure to tramadol, (reading: adjusted β [aβ] -0.25, 95% CI -0.31 to -0.18; numeracy: aβ -0.22, 95% CI -0.29 to -0.16), during both early and late pregnancy (reading: aβ -0.11, 95% CI -0.14 to -0.07; numeracy: aβ -0.13, 95% CI -0.16 to -0.10) and with higher OME doses (reading: aβ -0.14, 95% CI -0.16 to -0.12; numeracy: aβ -0.11, 95% CI -0.13 to -0.09). Secondary Outcome Similar patterns to the primary outcome were observed for secondary outcomes based on scores below the NMS threshold, with the highest relative risks (RR) observed among children exposed in both early and late pregnancy (reading 1.20, 95% CI 0.98–1.46; numeracy 1.16, 95% CI 0.92–1.44), and to tramadol (reading 1.20, 95% CI 0.85–1.70; numeracy 1.08, 95% CI 0.74–1.58) (eFigure 6-7). Sensitivity analysis Compared to the primary analysis, there was a slightly larger difference between mean z-scores when prenatal exposure was redefined as at least two dispensings (reading aβ -0.13, 95% -0.17 to -0.10, numeracy aβ -0.10, 95% -0.13 to -0.07). For discontinuer and sibling analyses results were close to the null for any opioid, codeine or oxycodone exposure, though confidence intervals were wider for individual opioids. For tramadol, estimates were only slightly attenuated although confidence intervals were wide (reading: aβ -0.17, 95% CI -0.26 to -0.08; numeracy: aβ -0.18, 95% CI -0.27 to -0.09) (eFigure 8). Results were consistent in the analysis based on z-scores from test results of all students (general and concessional), as well as in our analyses accounting for potential selection bias (Figure 2-3; eFigure 6-7). Post Hoc Analyses The association between higher OME exposure and poor academic performance was attenuated when compared to low OME exposure (Figures 2, 3; eFigures 7, 8). The E-value indicates an unmeasured confounder would need a risk ratio of 1.82 (95% CI: 1.36–2.58) with both tramadol exposure and academic performance to fully explain away the observed association. We found similar standardized proportions of adverse birth outcomes between the exposed and unexposed groups, suggesting these outcomes likely play a minimal role in the relationship between opioid exposure and academic performance (eTable 7).

Main Findings In this population-based cohort study of children born to concessional beneficiaries in NSW, Australia, prenatal exposure to opioid analgesics across various pregnancy periods, doses, and monotherapies, with the potential exception of tramadol, had negligible impacts on third grade reading and numeracy skills. The consistency of these small effects across primary and secondary outcomes, and the near-null values in sensitivity analyses with improved confounding control suggests that prenatal opioid exposure is unlikely to have clinically meaningful impacts on academic performance. Interpretation While tramadol was the only exposure to exceed our predefined threshold for a meaningful academic difference, this finding should be interpreted in the context of additional evidence. First, our discontinuer and sibling analyses suggest that unmeasured confounding factors may partially account for the association with academic performance. Second, post-hoc sensitivity analyses indicate that even a modest confounder could account for the observed effect. Third, the higher median tramadol dose compared to other opioids raises the possibility of a dose–response relationship or confounding by indication. Fourth, the effect sizes for the secondary outcome were small and inconsistent, although confidence intervals were wider due to smaller group sizes. Fifth, although tramadol is unique in acting as both a μ-opioid receptor agonist and a serotonin/norepinephrine reuptake inhibitor, previous research has not demonstrated any association between academic performance and prenatal exposure to antidepressants with similar reuptake inhibition properties (26, 27). Taken together, these results raise the possibility that unmeasured confounding factors specifically associated with prenatal tramadol exposure may have influenced our findings. In our data, women prescribed tramadol exhibited higher rates of mental health conditions, analgesic use, and exposure to teratogenic medicines compared to unexposed women and codeine users, suggesting greater pain burden and more complex health profiles. However, similar patterns were observed among those who used oxycodone, meaning women that used tramadol would need to differ in some unmeasured way to explain the tramadol-specific effects. Since we cannot definitively rule out unmeasured confounding, we view these findings as a potential signal warranting further investigation in future studies. Our findings regarding prenatal exposure to any opioid analgesic are consistent with those of a previous Norwegian cohort study of fifth-grade children, although it should be noted that both sets of results likely reflect the effect of codeine exposure given its predominance in both study populations (8). Our research expands the evidence on long-term safety of analgesic opioids used during pregnancy by evaluating academic performance related to commonly used opioids worldwide, including tramadol (2). Our findings primarily provide evidence of safety for acute and low-dose opioid exposure during pregnancy, and they may not be generalisable to contexts involving higher dosages or repeated opioid dispensings, particularly given some evidence of potential dose-response and duration effects. Strengths and Limitations: The primary strength of this study is the use of real-world, population-level data to understand the impact of prenatal opioid exposure on academic performance among children. Our study has several limitations. First, dispensing records do not reflect actual consumption. Second, limited exposure data prevented sensitivity analyses requiring ≥2 dispensings. This constraint may have contributed to some of the observed null associations through exposure misclassification. Third, our data do not capture medicines supplied in public hospitals, privately funded prescriptions, or over the counter (OTC) low-dose codeine (≤15mg), which was available until February 2018 (28). Privately dispensed opioids likely had minimal impact on our findings given that our concessional beneficiary population would have preferentially accessed lower-cost, subsidized medicines. However, incomplete capture of codeine use may have resulted in exposure misclassification, potentially biasing estimates toward the null. In 2014, privately funded and OTC codeine comprised approximately one-fifth of total opioid utilisation in Australia (29). Fourth, we restricted our study population to children of concessional beneficiaries to ensure complete capture of medicine dispensings. Consequently, compared to the general population, mothers in our sample were disproportionally younger, had more co-morbidities and experienced greater socio-economic disadvantage (eTable 8). Fifth, selection bias may have influenced our findings, as approximately 17% of eligible children lacked test scores due to non-participation in tests or loss to follow-up. However, we believe our findings are robust because: 1) proportions of exposed and unexposed children were similar across missing data categories, 2) children lost to follow-up were largely comparable to those with test scores, and 3) sensitivity analyses incorporating exempt children and applying inverse probability of censoring weights yielded results consistent with our primary analysis. Finally, we lacked information on clinical indications for opioid use, potentially introducing confounding by indication which could bias results away from the null if the underlying conditions are independently associated with poorer academic outcomes.

Overall, our findings suggest that prenatal exposure to opioid analgesics, with the possible exception of tramadol, does not have an academically meaningful impact on third-grade academic performance. However, whether the observed tramadol results indicate a causal relationship or reflect unmeasured confounding remains unclear, warranting further investigation. These long-term safety findings should be considered alongside evidence of short-term risks when assessing the benefits and risks of prescribing opioids during pregnancy.

This research was completed using the Early Life Course (ELC) data platform. The authors would like to thank data providers: the Australian Government Department of Health and Aged Care, NSW Ministry of Health, the Australian Government Department of Social Services, NSW Department of Education, as the data custodian for the NSW School Enrolments data. The authors gratefully acknowledge the assistance of the Centre for Health Record Linkage (CHeReL) and the Australian Institute of Health and Welfare (AIHW) Data Integration Unit in linking the numerous data collections. They also wish to thank the investigators named on the UNSW Research Infrastructure Scheme grant for their assistance with securing funds to build the ELC platform, and investigators at the Faculty of Medicine and Health, UNSW, for their ongoing management and governance of the ELC platform. The linkage was funded by a UNSW Research Infrastructure Scheme grant and a UNSW Scientia Program Award. UNSW’s National Perinatal Epidemiology and Statistics Unit contributed research support funds for project management. Initial data checking and cleaning was resourced by a National Health and Medical Research Council Ideas grant (#2010778).

https://www.tga.gov.au/codeine-info-hub 1. Pennick V, Liddle SD. Interventions for preventing and treating pelvic and back pain in pregnancy. Cochrane Database Syst Rev. 2013(8):CD001139.2. Brett J, Cesta CE, Gillies MB, Bateman BT, Chan A, Cheng M, et al. Global trends in analgesic opioid use in pregnancy: A retrospective cohort study. Anesthesiology. 2025.3. Boggess T, Risher WC. Clinical and basic research investigations into the long-term effects of prenatal opioid exposure on brain development. J Neurosci Res. 2022;100(1):396-409.4. Goldfarb SS, Stanwood GD, Flynn HA, Graham DL. Developmental opioid exposures: Neurobiological underpinnings, behavioral impacts, and policy implications. Exp Biol Med (Maywood). 2020;245(2):131-7.5. 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(1):61-87.6. Subedi L, Huang H, Pant A, Westgate PM, Bada HS, Bauer JA, et al. Plasma Brain-Derived Neurotrophic Factor Levels in Newborn Infants with Neonatal Abstinence Syndrome. Front Pediatr. 2017;5:238.7. Bathina S, Das UN. Brain-derived neurotrophic factor and its clinical implications. Arch Med Sci. 2015;11(6):1164-78.8. Tronnes JN, Lupattelli A, Ystrom E, Nordeng H. Analysis of Prenatal Exposure to Opioid Analgesics and Scholastic Skills in Children in Fifth Grade in Norway. JAMA Netw Open. 2022;5(7):e2222425.9. 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;29(6):736-44.10. 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(6):625-34.11. Tran DT, Robijn AL, Varney B, Zoega H, Brew BK, Chambers GM, et al. Data Resource Profile: The Early Life Course data platform for research on perinatal and early childhood exposures and outcomes in Australia. Int J Epidemiol. 2024;53(2).12. Tau GZ, Peterson BS. Normal development of brain circuits. Neuropsychopharmacology. 2010;35(1):147-68.13. Nielsen S, Degenhardt L, Hoban B, Gisev N. A synthesis of oral morphine equivalents (OME) for opioid utilisation studies. Pharmacoepidemiol Drug Saf. 2016;25(6):733-7.14. He WQ, Moore HC, Miller JE, Burgner DP, Swann O, Lain SJ, et al. Impact of early childhood infection on child development and school performance: a population-based study. J Epidemiol Community Health. 2024;79(1):27-35.15. Australian Institute of Health and Welfare. National Opioid Pharmacotherapy Statistics Annual Data Collection (NOPSAD) 2025. Australian Institute of Health and Welfare, Canberra 2025.16. Satterthwaite FE. An Approximate Distribution of Estimates of Variance Components. Biometrics Bulletin. 1946;2(6):110-4.17. Brookhart MA, Schneeweiss S, Rothman KJ, Glynn RJ, Avorn J, Sturmer T. Variable selection for propensity score models. Am J Epidemiol. 2006;163(12):1149-56.18. Australian Curriculum, Assessment and Reporting Authority 2020, National Assessment Program – Literacy and Numeracy 2019: Technical Report, ACARA, Sydney.19. Mathur MB, Ding P, Riddell CA, VanderWeele TJ. Web Site and R Package for Computing E-values. Epidemiology. 2018;29(5):e45-e7.20. Esposito DB, Bateman B, Werler M, Straub L, Mogun H, Hernandez-Diaz S, et al. Ischemic Placental Disease, Preterm Delivery, and Their Association With Opioid Use During Pregnancy. Am J Epidemiol. 2022;191(5):759-68.21. Brogly SB, Velez MP, Werler MM, Li W, Camden A, Guttmann A. Prenatal Opioid Analgesics and the Risk of Adverse Birth Outcomes. Epidemiology. 2021;32(3):448-56.22. Sujan AC, Quinn PD, Rickert ME, Wiggs KK, Lichtenstein P, Larsson H, et al. Maternal prescribed opioid analgesic use during pregnancy and associations with adverse birth outcomes: A population-based study. PLoS Med. 2019;16(12):e1002980.23. Kallen B, Borg N, Reis M. The use of central nervous system active drugs during pregnancy. Pharmaceuticals (Basel). 2013;6(10):1221-86.24. Rees P, Stilwell PA, Bolton C, Akillioglu M, Carter B, Gale C, et al. Childhood Health and Educational Outcomes After Neonatal Abstinence Syndrome: A Systematic Review and Meta-analysis. J Pediatr. 2020;226:149-56 e16.25. Twilhaar ES, de Kieviet JF, Aarnoudse-Moens CS, van Elburg RM, Oosterlaan J. Academic performance of children born preterm: a meta-analysis and meta-regression. Arch Dis Child Fetal Neonatal Ed. 2018;103(4):F322-F30.26. Besag FMC, Vasey MJ. Should Antidepressants be Avoided in Pregnancy? Drug Saf. 2023;46(1):1-17.27. Leshem R, Bar-Oz B, Diav-Citrin O, Gbaly S, Soliman J, Renoux C, et al. Selective Serotonin Reuptake Inhibitors (SSRIs) and Serotonin Norepinephrine Reuptake Inhibitors (SNRIs) During Pregnancy and the Risk for Autism spectrum disorder (ASD) and Attention deficit hyperactivity disorder (ADHD) in the Offspring: A True Effect or a Bias? A Systematic Review & Meta-Analysis. Curr Neuropharmacol. 2021;19(6):896-906.28. Therapeutic Goods Administration. Codeine information hub: how and where to get advice. Canberra: TGA; 2018. Available from: . (Accessed 04 February 2024).29. Gisev N, Pearson SA, Karanges EA, Larance B, Buckley NA, Larney S, et al. To what extent do data from pharmaceutical claims under-estimate opioid analgesic utilisation in Australia? Pharmacoepidemiol Drug Saf. 2018;27(5):550-5. Figure 1: Study cohort flowchart Abbreviations: LMP, last menstrual period; NSW, New South Wales; NAPLAN, National Assessment Program – Literacy and Numeracy; OUD, opioid use disorder; SD, standard deviation. Table 1: Maternal and child characteristics by exposure to opioid analgesics during pregnancy, for children of concessional beneficiaries between January 01, 2003, to March 26, 2011. Figure 2: Differences in mean z scores in reading between third-grade children with and without prenatal opioid analgesic exposure: Crude and adjusted beta coefficients (β) with 95% confidence intervals (CI). Covariates included in the adjusted analyses included child’s sex, child’s year of birth, test year, language spoken at home, maternal age, born in western country, does not have a partner, social security benefits received, previous caesarean delivery, smoking during pregnancy, area-based socioeconomic disadvantage, area-based remoteness, multifetal gestation, parental highest educational attainment, parent occupation, maternal conditions, pre-pregnancy maternal medicine use, and pre-pregnancy healthcare utilisation. *Comparator reference groups consisted of children that were not prenatally exposed to opioid analgesics, with three exceptions: (1) For discontinuers, the reference group included children prenatally exposed in the 90 days before the last menstrual period (LMP) but not during pregnancy; (2) for the sibling analysis, the reference group comprised of siblings with no prenatal opioid exposure; and (3) for the OME comparison, low OME exposure was used as the reference group. † z scores based on both concessional and general beneficiaries’ test scores. Figure 3: Differences in mean z scores in numeracy between third-grade children with and without prenatal opioid analgesic exposure: Crude and adjusted beta coefficients (β) with 95% confidence intervals (CI). Covariates included in the adjusted analyses included child’s sex, child’s year of birth, test year, language spoken at home, maternal age, born in western country, does not have a partner, social security benefits received, previous caesarean delivery, smoking during pregnancy, area-based socioeconomic disadvantage, area-based remoteness, multifetal gestation, parental highest educational attainment, parent occupation, maternal conditions, pre-pregnancy maternal medicine use, and pre-pregnancy healthcare utilisation. *Comparator reference groups consisted of children that were not prenatally exposed to opioid analgesics, with three exceptions: (1) For discontinuers, the reference group included children prenatally exposed in the 90 days before the last menstrual period (LMP) but not during pregnancy; (2) for the sibling analysis, the reference group comprised of siblings with no prenatal opioid exposure; and (3) for the OME comparison, low OME exposure was used as the reference group. † z scores based on both concessional and general beneficiaries’ test scores. Supplementary Material File (table 1.docx) - Download - 26.65 KB Information & Authors Information Version history Peer review timeline Published BJOG: An International Journal of Obstetrics & Gynaecology Version of Record15 Mar 2026Published Copyright This work is licensed under a Non Exclusive No Reuse License. Collection

Authors Metrics & Citations Metrics Article Usage 209views 130downloads Citations Download citation Bianca Varney, Helga Zoega, Malcolm Bjørn Gillies, et al. Prenatal prescription opioid analgesic exposure and academic performance in third grade children: An Australian population-based cohort study. Authorea. 16 August 2025. DOI: https://doi.org/10.22541/au.175533129.98373096/v1 DOI: https://doi.org/10.22541/au.175533129.98373096/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.