Phytoscreening the homes of pregnant residents for volatile organic compounds: a pilot study in metro Detroit.

Exposure to anthropogenic volatile organic compounds (VOCs) has been associated with preterm birth, yet the spatial distribution of VOC exposures within urban environments remains poorly characterized, particularly for vapor intrusion pathways. Plants have recently been used as a cost-effective, non-invasive tool to "phytoscreen" for belowground contaminants. Here, we evaluated the use of phytoscreening within a birth cohort in metro Detroit, Michigan, USA by (a) assessing participant willingness to allow sampling in residential front yards, (b) quantifying VOC concentrations across plant tissue types (leaves and twigs), and (c) testing whether VOC detections in plants are associated with proximity to brownfield sites. Of 20 pregnant participants, 18 agreed to phytoscreening and sampling was completed at 13 residences, indicating high feasibility. Across 15 plant species, five of six target VOCs (benzene, toluene, ethylbenzene, xylene, trichloroethylene, tetrachloroethylene) were detected. Specifically, toluene was detected at all 13 homes and in 68% of samples. While less common, ethylbenzene was detected in 12 samples across 3 homes and trichloroethylene was detected in one sample that exceeded Michigan's risk-based screening level for soil vapor intrusion. Additionally, VOCs were detected more frequently and at higher concentrations in leaves compared to twigs. However, VOC concentrations in plants were not associated with proximity to known or suspected brownfield sites in this small sample size. Overall, these findings demonstrate the feasibility of phytoscreening in epidemiological studies and highlight its utility for characterizing spatial patterns of VOC exposure. Incorporating plant-based sampling approaches may help identify localized hotspots of human exposure and inform targeted environmental remediation efforts.