367 - Agreement between maternal urine substance screenings and neonatal urine and meconium substance screenings

Publication Number: 367.3984

Courtney W. Tam, Children's Hospital Los Angeles, Los Angeles, CA, United States; Amy Yeh, Los Angeles General Medical Center/USC, Los Angeles, CA, United States; Hemananda Muniraman, Keck School of Medicine of the University of Southern California, Los Angeles, CA, United States

Background: American College of Obstetricians and Gynecologists recommends routine screening for substance use disorder in pregnant women should be applied equally to all and can be accomplished via validated questionnaires or patient conversations. Routine laboratory testing of biologic samples is not required and is associated with bias and disparity. Similarly, neonatal urine and meconium substance screening tests used to detect intrauterine substance exposure (IUSE) have been associated with bias and disparity and used as a factor in determining family separation. Neonatal laboratory screenings for IUSE vary by institution.

Objective: To evaluate the agreement between maternal urine substance screening with neonatal urine and meconium screening results for IUSE.

Design/Methods: A single-center retrospective study included pregnant women admitted with a positive history and urine screen for substance use during pregnancy over a 4-year period (February 2018-December 2022). Agreement of maternal urine screening with neonatal urine and meconium screening results was analyzed using kappa statistics and accuracy measures were evaluated with sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under receiver operating characteristic curve (AUC).

Results: 931 paired maternal and neonatal urine screening results and 918 paired maternal urine and neonatal meconium screening results were included. Marijuana, methamphetamines, methadone and fentanyl were the most common substances found positive in maternal and neonatal screens (Table 1,2). Overall, neonatal urine and meconium screening had high specificity and NPV; however, sensitivity and PPV varied between substances and source of neonatal screening. Methadone had highest accuracy measures and agreement for both neonatal urine (AUC 0.91, k 0.86) and meconium screening (AUC 0.87, k 0.78). Neonatal urine screen was found to have very good accuracy measures for marijuana (AUC 0.80) and methamphetamines (AUC 0.88) but moderate accuracy measures and agreement for meconium screens (AUC 0.75, 0.77). In contrast to the high accuracy of neonatal urine fentanyl (AUC 0.84), meconium was found to be non-discriminating for fentanyl (Table 1, 2).

Conclusion(s): Neonatal urine and meconium screenings had high specificity and NPV, but varying sensitivity and PPV. Methadone, marijuana and methamphetamines had the highest accuracy measures. Meconium was found to be non-discriminating when compared to urine screening for fentanyl. The utility of neonatal substance screening when maternal substance screening is negative needs further evaluation.

Table 1. Comparison of maternal urine substance screening results to neonatal urine substance screening results
Table delineating sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), area under receiver operating characteristic curve (AUC) and agreement using kappa statistics when comparing maternal urine screening results with neonatal urine screening results.

Table 2. Comparison of maternal urine substance screening results to neonatal meconium substance screening results
Table delineating sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), area under receiver operating characteristic curve (AUC) and agreement using kappa statistics when comparing maternal urine screening results with neonatal meconium screening results.