B-283 Concordance of Drug Detection Between Neonatal Urine, Maternal Urine, and Umbilical Cord Tissue: A Multi-Matrix Investigation of Prenatal Substance Exposure

Background Identification of neonates exposed to substances in-utero is crucial for appropriate intervention and follow-up care. Several biological matrices can be tested to identify drug exposure, including neonatal urine (NU), umbilical cord tissue (UCT), meconium, and maternal urine (MU). However, there are key differences between these matrices that influence drug detection, potentially resulting in discordant results between sample types. Such discrepancies may pose challenges with result interpretation, especially for the clinical team. A clear understanding of these matrices and their differences can help with interpretation and ordering. A number of studies have already assessed agreement between meconium and UCT, however, there are limited data comparing NU or MU with UCT. The primary focus of this study was to evaluate the concordance between 1. NU with UCT, 2. MU with UCT, and 3. NU and MU in paired samples. Frequency of drug positivity across all three matrices was also assessed. Methods Toxicology results for 536 NU samples that were ordered as part of routine clinical testing over a 12-month period were reviewed retrospectively. Of those infants with NU samples, there were 497 corresponding MU samples and 179 UCT samples. All testing was performed by liquid chromatography-tandem mass spectrometry; NU and MU testing was conducted at the same laboratory, while UCT analysis was performed at a send-out laboratory. Positivity rates were assessed for each of the three sample types. Agreement between NU and UCT, MU and UCT, and NU and MU was evaluated using Cohen’s kappa (k) for various drugs or drug classes, including barbiturates, amphetamines, cocaine, opioids, and THC. Results One or more drugs was detected in 40% of NU samples, 85% of UCT samples, and 69% of MU samples. Across all matrices, 40 unique analytes were detected in NU, 27 in UCT, and 33 in MU samples. Carboxy-THC was the most frequently detected substance, found in 10% of NU,15% of UCT, and 57% of MU samples. Other substances that were observed most frequently included fentanyl/norfentanyl, cocaine (as benzoylecgonine), amphetamine/methamphetamine, buprenorphine/norbuprenorphine, and butalbital. Depending on the matrix, positivity rates for these compounds ranged from 2-9%. Overall agreement (for NU vs. UCT, NU vs. MU, and UCT vs. MU, respectively) was strongest for amphetamines (k=0.81, 0.89, 0.84), and benzoylecgonine (k=0.88, 0.81, 0.81), whereas it was generally moderate for barbiturates (k=0.92, 0.78, 0.72) and opioids (k=0.75, 0.64, 0.80), and poor for carboxy-THC (k=0.17, 0.11, 0.83) when comparing NU with MU or UCT. A total of 168 cases had carboxy-THC results for all three matrices, of which 71 were negative across all sample types. Only 15 were positive in all three matrices, while 11 were positive in MU alone, 4 positive in UCT alone, and 67 were positive in MU and UCT but negative in NU. Conclusion Significant discrepancies in drug detection across NU, UCT, and MU highlight the complexity of neonatal drug testing. These findings highlight that reliance on a single specimen type may miss fetal drug exposure, supporting the need for comprehensive testing strategies and careful evaluation of discordant or unexpected results.