An apple a day? Assessing gardeners' lead exposure in urban agriculture sites to improve the derivation of soil assessment criteria.

Globally, many of our urban agriculture sites (UAS) contain high levels of lead (Pb), a contaminant of toxicological concern to humans. To improve the derivation of soil assessment criteria at UAS, and avoid inappropriate closure of these valuable community spaces, we sampled nearly 280 paired soil and crop samples across 31 UAS gardens. This sampling was coupled with an exposure and food frequency questionnaire and participants blood Pb levels (BLL), (43 gardeners and 29 non-gardening neighbours). In 98% of the sampled soils, Pb concentrations were above the current UK soil guideline for UAS (80 mg/kg), however despite the high soil Pb (geometric mean: 324 mg/kg), and high soil bioaccessible Pb (geometric mean: 58.7%), all participants BLL were <4.1 μg/dL (range: 0.6-4.1 μg/dL). Indeed, there was no statistically significant difference between the BLL of the UAS gardeners and those of their non-gardening neighbours (p = 0.569). Pb uptake, however, varied with crop type and our study highlights the suitability of certain crops for growing at UAS with elevated Pb (e.g. tubers, shrub and tree fruit), whilst limiting the consumption of others (selected root vegetables, such as rhubarb, beetroot, parsnips and carrots, with observed Pb concentrations > 0.1 mg/kg FW). The importance of defining the exposure scenario of a specific sub-population (i.e. UAS gardeners) is highlighted. Our preferred models predict site specific assessment criteria (SSAC) of 722-1634 mg/kg. We found fruit and vegetable consumption rates by all participants, and not just the UAS gardeners, to be considerably higher than those currently used to derive the UK's category 4 screening levels (C4SLs). Furthermore, the soil to plant concentration factors (SPCFs) used to derive the UAS C4SL significantly over predict Pb uptake. Our study indicates it may be appropriate to develop a distinct exposure dataset for UAS. In particular we recommend the derivation of SPCFs that are reflective of urban soils, both in terms of the range of soil Pb concentrations typically observed, but also the sources (and hence human oral bioaccessibility and plant-availability) of this Pb.