A probabilistic toxicokinetic modeling approach to the assessment of the impact of daily variations of lead concentrations in tap water from schools and daycares on blood lead levels in children

Highlights

• A simple stochastic kinetic BLL model addressing time-varying Pb exposure was built.

• The model was applied to children exposed to Pb in tap water at school or daycare.

• In problematic institutions, resulting BLL indicators can increase significantly.

• Further, BLL may exceed CDC's threshold in up to 12% of infants and 3% of toddlers.

• This impact may remain unnoticed at large population level.

Abstract

The aim of this study was to assess the impact of exposure to tap water lead concentration ([Pb]_TW) occurring in schools or daycares on blood lead level (BLL) of attending children. Given the potentially wide variations in space and time of ([Pb]_TW) documented in the literature, a simple probabilistic toxicokinetic (STK) model that allows the simulation of the time-varying evolution of BLL in response to these variations was developed. Thus, basic toxicokinetic equations were assembled to simulate BLL in a typical infant, toddler and pupil. The STK model's steady-state BLL predictions showed good correspondence when validated against Integrated Exposure and Uptake BioKinetic model predictions for comparable [Pb]_TW values. Exposures to three distributions of [Pb]_TW in specific sets of Canadian schools and daycares documented in the scientific literature were simulated probabilistically with Monte Carlo simulations. For the highest distribution of [Pb]_TW simulated (median, 90th percentile = 24, 412 μg/L), average annual BLL (median, 97.5th percentile) varies between 1.5 and 6.4 μg/dL in infant and 1.1 and 3 μg/dL in pupils. Toddler's results were midway between those from the infants and pupils. Under this exposure scenario, the infant may present BLL > 5 μg/dL for a significant number of days over the course of the academic year (median; 97.5th: 17; 227 days). However, peak exposure may remain unnoticed if rare and drowned out by the background BLL. In conclusion, even if they may be sparse, peak exposure episodes to [Pb]_TW in schools and daycares may suffice to increased BLL in attending individuals. This finding emphasizes the need for further characterization of [Pb]_TW in schools and daycares in order to identify potentially problematic institutions and therefore avoid undesirable exposures for the children attending them.

Introduction

Lead (Pb) is a systemic human toxicant that notably affects neurological function in exposed children. This translates into adverse effects such as altered behavior and mood, deficit in neuromotor and neurosensory function, and decrement in cognitive function and learning capacities (ATSDR, 2020). Intellectual disability caused by Pb exposure in children is a concern worldwide (Carrington et al., 2019). In this regard, it is now recognized that loss in intellectual quotient (IQ) may result from exposures leading to blood Pb levels (BLL) below 10 μg/dL (Jusko et al., 2008; Lanphear et al., 2005; Surkan et al., 2007; Téllez-Rojo et al., 2006) or even 5 μg/dL (Desrochers-Couture et al., 2018; Santa Maria et al., 2019). In the seminal pooling study from Lanphear et al. (2005), blood samples were taken at least four times in more than 1300 children between birth and the end of the follow-up period, when the IQ test was administered (between 5 and 10 years old). The authors observed that greater IQ loss was observed, relatively per blood Pb concentration unit, below 10 μg/dL as compared to above this level, the reference value at the time. This relation, further validated by Crump et al. (2013), has become the basis for Pb's risk assessment by the public health authorities, and the scientific community now considers that there is no safe threshold identified for Pb exposure (AAP, 2017; EFSA, 2010; Joint FAO/WHO Expert Committee on Food Additives, 2011; WHO, 2010a). However, given the above-mentioned studies, the CDC sets the threshold for intervention at 5 μg/dL (CDC, 2012).

In view of these evidences, the International Society for Environmental Epidemiology has called for the reduction as much as possible of human Pb exposure, particularly in children (ISEE, 2015). Tap water, distributed either in residential settings as well as in schools and daycares (SDC), is a source of exposure to target in this regard. Indeed, among the various sources of lead exposure, it appears as the one on which actions can more readily be taken (Lanphear et al., 1998; Levallois et al., 2014; Oulhote et al., 2013; Stanek et al., 2020; Triantafyllidou and Edwards, 2012). As a matter of facts, regulation of lead in drinking water has been constantly improved and various protocols and standards have been proposed to guide drinking water providers on this issue (Levallois et al., 2018). Currently Health Canada's maximum acceptable concentration (MAC) is 5 μg/L (Health Canada, 2019).

The quantitative evaluation of Pb exposure through tap water is a complex issue. In particular, some Pb concentrations measured in tap water of SDC may, depending of the sampling protocol and existence of corrosion control, exhibit spatio-temporal variations that exceed three orders of magnitude within a given SDC, sometimes superating the mg/L value (Barn et al., 2014; Deshommes et al., 2016; Doré et al., 2018; Triantafyllidou et al., 2014b; Triantafyllidou and Edwards, 2009). A single fountain within a given SDC can present close to 100-fold variations in Pb concentrations (Deshommes et al., 2016; Doré et al., 2018). Yet, some authors suggested that even as sparse as they might be, peak concentrations within this variation pattern could contribute to a worrying increase BLL of attending children (Lakind, 1998; Lambrinidou et al., 2010; Triantafyllidou et al., 2014a). Besides, Pb in drinking water from schools have recently been associated with significant, though mild, deprivation of educational outcomes in Ontario, Canada (Buajitti et al., 2021).

The unpredictable nature of such peak exposures is quite specific to the case of Pb levels in tap water distributed in large buildings such as SDC (Deshommes et al., 2016), and it precludes the design of experimental settings using blood measurements to verify their impact on exposure. Else, toxicokinetic models can be used, but available ones present important limitations. Indeed, US EPA's Integrated Exposure and Uptake Biokinetic (IEUBK) model, the most commonly used toxicokinetic model to calculate BLL in children as a result of environmental exposures to Pb (Laidlaw et al., 2017; Li et al., 2016; Zartarian et al., 2017; Zhong et al., 2017), relies on a premise of continuous exposure to constant concentrations of Pb in the various environmental media (Hogan et al., 1998). At best, it can reflect exposure conditions that are averaged over a 30-days period (Deshommes et al., 2013; Ngueta et al., 2014; Zartarian et al., 2017). Thus, its design and settings are not built to reflect the impact of short term (daily) variations in exposure concentrations such as those describe above for tap water concentrations in SDC. Other pharmacokinetic models exist for Pb, such as the International Commission on Radioprotection (ICRP) model (Leggett, 1993) and the O'Flaherty, 1993, O'Flaherty, 1995 model, but their parametrization and use is more complex. Noteworthingly, the probability of occurrence of high BLL levels in given exposure conditions through tap water can also not be determined with these models, as they are fundamentally deterministic. This lack of probabilistic dimension to the available models limits the possibilities in terms of interpretation of BLL modeling results by risk managers, and the present work also aims to address this caveat.

Therefore, the objective of the present work was to probabilistically assess the relative impact of daily variations of exposure to tap water Pb concentration in SDC on BLL of attending children. Conceptually, this requires the development of a simple stochastic toxicokinetic modeling tool that allows the estimation of time-varying BLL in response to the specific rapid, and potentially wide, time-varying nature of Pb levels mentioned above.