MPC Member Publications

This database contains a listing of population studies publications written by MPC Members. Anyone can add a publication by an MPC student, faculty, or staff member to this database; new citations will be reviewed and approved by MPC administrators.

Full Citation

Title: The association between intrauterine exposure to opioids, tobacco, alcohol, and cannabis and length of birth hospitalization among neonates without NOWS

Citation Type: Journal Article

Publication Year: 2023

ISSN: 1476-5543

DOI: 10.1038/s41372-023-01694-5

Abstract: INTRODUCTION Birth hospitalization length of stay (LOS) is extended for infants with neonatal opioid withdrawal syndrome (NOWS) [1], raising health care costs and complicating efforts to promote maternal-infant bonding [1, 2]. However, many infants exposed to opioids do not develop NOWS. The overall costs of intrauterine opioid exposure may be significantly greater if infants who do not develop withdrawal symptoms also have an increased LOS due to observation for the emergence of NOWS or other medical or social complications. Using data from the Environmental influences on Child Health Outcomes (ECHO)-wide cohort study funded by the National Institutes of Health, we evaluated whether opioid exposure not resulting in NOWS was associated with a longer LOS. We also performed similar evaluations of tobacco, alcohol, and cannabis, and considered the effect of combined exposures. METHODS Study population The current analysis used previously collected data to evaluate the association between intrauterine substance exposure and postnatal hospital LOS among neonates without NOWS. The analytic sample included infants from eight cohorts with complete data on intrauterine substance exposure and postnatal hospital LOS. Cohorts exclusively enrolling very low birthweight or preterm infants were excluded. Infants diagnosed with and/or receiving treatment for NOWS (parent report of opioid or drug withdrawal in the infant, or one or more NOWS screening tools administered, with either a positive score on that measure or pharmacological treatment for NOWS administered); 14 of the 5989 participants) were removed from analyses. Measures The primary exposure was prenatal maternal opioid use, defined as any use of opioid medications or heroin during pregnancy, other than those administered during labor and delivery, as indicated by self-report, medical record abstraction, or maternal toxicology, all of which were taken from cohort study records. Prenatal maternal use of other substances was defined in the same way. The primary outcome was postnatal LOS, calculated as a continuous variable from medical record abstraction data on the date of discharge from the hospital (or final placement if discharged to a different level of care) and date of birth. Covariates were selected based on known or theoretical associations with either polysubstance exposure or LOS, and included maternal age, race, ethnicity, education, calendar year of pregnancy; and infant sex, race, and ethnicity. Statistical analyses To reduce confounding, we estimated propensity scores for exposure to opioids, alcohol, cannabis, and tobacco using maternal covariates. We used the parameter estimates to calculate propensity scores for exposed and non-exposed infants. The inverse probability weights were stabilized by the marginal probability of the actual infant exposure level. Generalized Estimating Equations (GEE), accounting for clustering of participants in original cohort, models were used to examine associations between each prenatal substance use measure and continuous LOS in days, controlling for infant-level covariates. Sensitivity analyses also controlled for con-comitant substance use. R 4.1.0 was used for all analyses, [3] and Multiple Imputation by Chained Equations (MICE) [4] was used to impute missing data (<10% for race, and 4% for all others). RESULTS The analytic sample included 5975 infants (from 5913 pregnancies) who were not selected from NICU-enriched cohorts and did Journal of Perinatology 1234567890();,: not have a diagnosis of NOWS (n = 14) at birth. For this analysis, among observations with data, maternal race was primarily White (3488; 63%) and other race/multiple races (1730; 31%), with 324 (5.8%) describing themselves as Black, and 1057 (18%) reporting Hispanic ethnicity; 4642 pregnant people (80%) reported at least some college education. Among infants, 5315 (90.9%) were not admitted to the NICU, 4326 (74.9%) were of appropriate weight for gestational age, and 198 (3.3%) tested positive for opioid exposure. No significant differences were detected in postnatal LOS for any of the intrauterine exposure groups, in either unadjusted or adjusted analyses (Table 1). Prenatal substance exposure (other than tobacco) was not associated with NICU admission. Sensitivity analyses controlling for other prenatal substance use or evaluating the effect of LOS outliers were also not significant (results not presented). DISCUSSION Our sample included many infants with intrauterine exposure to opioids who were not diagnosed with or treated for NOWS, a finding that is consistent with differences in the prevalence of opioid use vs. opioid use disorders. Although NOWS is associated with extended LOS, our aggregated data for substance-exposed neonates from a variety of United States (U.S.) hospital settings suggest no differences in duration of birth hospitalization based on exposure history alone when NOWS does not manifest. This could be positive news, and-as with evidence that the negative effects of other substances are restricted to those with higher exposure [5]-suggests that harm cannot be inferred solely on the basis of self-reported use or a positive drug screen. Alternatively, these findings may also mean that some infants with intrauterine opioid exposure are discharged before NOWS manifests (up to 5 days after birth), which would present clear risks. Although current clinical practice guidelines suggest observing infants with intrauterine opioid exposure for 3-7 days after birth, [6] there is significant variability in hospital policy and even in NOWS definitions-particularly during the period in which most original data in this study were collected. Unfortunately the available data do not contain information about readmission for NOWS. Further research should clarify how frequently infants are discharged prior to the emergence of withdrawal symptoms.

Url: https://www.nature.com/articles/s41372-023-01694-5

User Submitted?: No

Authors: Ondersma, Steven J.; Kress, Amii M.; Stroustrup, Annemarie; Annett, Robert D.; Avalos, Lyndsay A.; Talavera-Barber, Maria; Brennan, Patricia A.; Camargo, Carlos A.; Conradt, Elisabeth; Dunlop, Anne L.; Elliott, Amy J.; Hedderson, Monique M.; Li, Ximin; McGrath, Monica; Nguyen, Ruby H. N.; Page, Grier P.; Sathyanarayana, Sheela; Lester, Barry

Periodical (Full): Journal of Perinatology 2023

Issue:

Volume:

Pages: 1-3

Countries:

IPUMS NHGIS NAPP IHIS ATUS Terrapop