Introduction

This health impact assessment relates to human populations exposed indirectly to lead (Pb) via the environment—abbreviated as MvE (Man via Environment) exposure. In some cases, MvE exposure can be an important route of general population exposure to chemicals and has to be considered in Chemical Safety Assessment to comply with the Regulation (EC) No 1907/2006 on the Regulation, Evaluation, Authorization and Restriction of Chemicals (REACH).

First tier screening tools such as the European Union System for the Evaluation of Substances (EUSES) [1] developed for and used for MvE assessment in the context of risk assessment in REACH Registration dossiers, are designed to be conservative and are appropriate for benchmarking against a threshold value (e.g. Derived No Effect Level, DNEL) since the level of conservatism does not play a role as long as the estimated exposure is below the DNEL. However, for human health impact assessments required for REACH authorisation, a realistic instead of an overly conservative approach for exposure modelling is warranted in order to ensure the health impact assessment and the overall socio-economic analysis are appropriate.

The aims of this study are (i) to present a tiered approach for assessment of MvE exposure for impact assessment (and risk assessment) purposes; (ii) to demonstrate the tiered approach in a case study for the lead batteries manufacturing and recycling industry in Europe in the context of an impact assessment; (iii) to compare the outcome with available human biomonitoring (HBM) data.

Materials and methods

Development of a tiered framework

The conceptual framework for MvE exposure from ECHA [2] has been used as starting point, including: (1) direct exposure via inhalation of air and drinking of water, and (2) indirect exposure following transfer of substances present in the environment (soil, air, groundwater, surface water) to biota (crops, cattle, dairy products, fish) for human nutrition. Additionally, exposure to Pb via dust and soil ingestion has been modelled since these exposure routes have previously been identified as a potentially important exposure source for Pb, especially for children [3].

In this study, the conceptual framework is implemented in a tiered approach. This approach is to start in Tier 1 with EUSES modelling [1], modified with metal specific transfer factors. Next, the Tier 2 approach is to use the higher tier Gaussian Plume Model (GPM) for high-resolution air quality modelling, plus dietary exposure modelling based on intake rates derived from the Comprehensive Food Database of the European Food Safety Authority (EFSA), in combination with crop specific transfer factors. Tier 3 is to consider additional site-specific information such as proximity of inhabitants and agricultural activities in relation to industrial sites. This should provide flexibility to include additional data gathering and modelling efforts for dominant routes of exposures, and to keep efforts limited for cases where conservative models are predicted to result in low exposure, or where input data for higher tier modelling are lacking.

The different tools used as building blocks for the tiered approach are described in Section “Existing tools and databases“.

Regional and local assessment

In a REACH framework, the release of a substance and subsequent exposure of the environment are in principle assessed on two spatial scales: locally in the vicinity of a representative point source of release to the environment, and regionally for a larger area which includes all release sources in that area (ECHA Guidance R.16). Only the release sources relevant to the regulatory scope of REACH authorization need to be assessed at the regional scale. When relying on modelling tools to assess environmental concentrations, the local concentration (Predicted Environmental Concentration, PEClocal) close to a point source emission is calculated as the sum of the predicted concentration caused by the point source (Clocal) and the background concentration (the regional concentration (PECregional)). When using monitoring data to assess environmental concentrations at the local scale, no summation with PEC regional should be made since monitoring data inherently reflect the sum of a local and a regional source.

Existing tools and databases

The subsequent subsections provide an overview of the existing tools and databases that will be used as building blocks in the tiered approach.

Modelling of Pb in the environment

Pb levels in ambient air

The EUSES model [1] is a simplified operational priority substances model that estimates the Pb air concentration at 100 m distance from a central stack (default stack height: 10 m; exhaust temperature identical with temperature in the environment) at standard Dutch weather conditions (default wind speed: 3 m/s, precipitation 700 mm/year). GPM can be used to generate more realistic predictions of local air emissions (Clocal air) than the EUSES model, by accounting for site-specific stack conditions and these models can estimate Pb air concentrations at several distances from the central stack. The input parameters generally required in GPM models are emission rate, wind speed, source height, gas temperature and gas velocity. GPMs assume ideal plume geometry, ideal steady state of air pollutant emissions and meteorological conditions, a uniform flat terrain and a complete conservation of mass. As the gases are heated in the industrial facility, the hot plume will be thrust upward some distance above the top of the stack, i.e., the effective stack height. Once the plume has reached its effective stack height, dispersion will begin in three dimensions depending on Gaussian plume equations, wind speed and atmospheric conditions. The model assumes that dispersion will take the form of a normal Gaussian curve, with the maximum concentration in the centre of the plume. The GPM model used in this project is freely available in Microsoft Excel format at https://www.arche-consulting.be/tools/local-air-modeling-gpm-tool/. and described in more detail in a previous paper [4].

Transfer from environment to crops

Soil-crop transfer

The Transfer Factor (TF) from soil to crops can be defined by the following formula [5]:

$$TF_i\frac{{Pb\;concentration_i}}{{Pb\;concentration\;in\;soil}}$$

Where i ϵ {leafy crops, root crops, grass, fruit, grain, tuber, bulb, shoot}

Pb concentration in plant parts is expressed in mg lead / kg dry weight (dwt).

Pb concentration in soil is expressed in mg lead/kg soil dwt.

TFi will be expressed in kg soil dwt/kgi,dwt.

A literature search for transfer factors was conducted and quality criteria were used (see ref. [11] and summarised in Table 1). The following simplified model equations can be used for the generic categories roots, leaves and grass:

$$C_{roots} = TF_{roots,soil} \cdot C_{soil}$$
$$C_{leaves} = TF_{leaves,soil} \cdot C_{soil}$$
$$C_{grass} = TF_{grass,soil} \cdot C_{soil}$$

with

Table 1 Soil-crop transfer factors (TF) for leaves and roots (from ref. [11]).
Full size table

Csoil is the dwt concentration in soil, expressed in kg/kgsoil

Ccrop is the dwt concentration in plant (roots, leaves or grass), expressed in kg/kgsoil

The same equations are applicable for the subcategories fruit, grain, tuber, shoot, bulb etc… Since Ccrop is obtained on the basis of dry weight data, it is necessary to convert into wet weight (wwt) by using the data of dry matter fraction in a crop of interest using the following formula:

$${{{{{\rm{C}}}}}}_{{{{{{\rm{crop}}}}}}}\left( {{{{{\rm{wwt}}}}}} \right) = {{{{{\rm{C}}}}}}_{{{{{{\rm{crop}}}}}}}\left( {{{{{{\rm{dwt}}}}}}} \right){{{{{\rm{x}}}}}}\left( 1 - {{{{{\rm{Fwater}}}}}}_{{{{{{\rm{plant}}}}}}} \right)$$

The water content (ca. 93%w/w) in crops (Fwaterplant) can be calculated using existing EUSES defaults.

Aerial deposition

EUSES does not contain the exposure pathway (for MvE) of aerial deposition of particles on crop cover and subsequent partial (metal) foliar uptake by plants. Unlike root metal transfer, less research has been conducted to examine foliar uptake of metals by plant leaves from the atmosphere. It has been demonstrated that foliar metal uptake by plants from the atmosphere does occur [6, 7]. Therefore, in this study aerial deposited on above-ground plant crops is addressed by a model from US EPA [8] that translates a deposition flux (expressed in kg/m2/year) into a plant concentration (kg/kg dwt) and which has been applied before in MvE exposure assessments of metals in a REACH context [9]. It is likely that part of the deposited fraction to edible plant parts is washed off before human consumption. The proportion of metal removed by cleaning depends on many factors, e.g. the type of the sample and the cleaning procedure used. Values ranging from 10 to 96% have been reported for the removal of Pb by washing foodstuff before eating [10]. As this is quite a broad range and the literature is inconclusive, a wash off factor of 50% was used for this study.

Transfer from environment to cattle (meat and dairy products)

The transfer from feedstuff to meat and dairy products is calculated in EUSES by means of the bioaccumulation factor (BAF) using following formula:

$$BAF_i = \frac{{Pb\;concentration}}{D}$$

where i ϵ{meat, milk}

where D represents the daily dietary lead intake, calculated as

$${{{{{\rm{D}}}}}} = {{{{{\rm{feed}}}}}}\;{{{{{\rm{intake}}}}}}\;{{{{{\rm{x}}}}}}\;{{{{{\rm{Pb}}}}}}\;{{{{{\rm{concentration}}}}}}\;{{{{{\rm{in}}}}}}\;{{{{{\rm{feedstuff}}}}}}$$

where: Feed intake: kg_feedstuff/day

Pb concentration in feedstuff: mglead/kg_feedstuff

Pb concentration in meat, milk: mglead/ kgi

Daily dietary lead intake: mglead/day

BAFi: day/kgi

A literature search for BAFs was conducted and quality criteria were used to select the most reliable sources. Results can be found in [11] and are summarised in Table 2. The concentrations in meat and dairy products are subsequently calculated following the model in EUSES [2].

Table 2 Transfer factor (Bioaccumulation factors (BAF)) for animal products (from ref. [11]).
Full size table

Transfer from environment to fish

For the purpose of a secondary poisoning assessment), the bioconcentration factor (BCF) for aquatic species is typically derived for aquatic invertebrates and fish combined, assuming the next trophic level eats a mixed diet of crustaceans, molluscs, annelids, acarids, insects and fish. This BCF originating from the secondary poisoning assessment (1533 L/kg wwt) was used in Tier 1 calculations of the MvE assessment [12]. At the Tier 2 level, a more human diet specific BCF for aquatic species was derived for the purpose of the MvE assessment. This BCF is only based on accumulation in fish species because it can be assumed that freshwater aquatic invertebrates are not (or in rare instances) caught and consumed by the local human population. Within typical environmental concentration range (i.e. between 0.18 µg/L (background concentration) and 15 µg/L (based on the 95th percentile of the PEClocal values), the gathered BCFs for fish ranged between 11 and 143 L/kg wwt (10–90th %) with a median value of 23 L/kg wwt [12].

External human exposure

The human Pb intake is calculated as the sum of the product of intake rates and intake concentrations (for every intake medium):

$$human\;exposure = {\sum} {C_{intake\;medium} \times intake\;rate_{intake\;medium}}$$

with intake medium: food categories, drinking water, dust, soil and air

intake rate = daily amount expressed as mass per day for oral intake media (kg/day) and volume for inhalation. A daily inhalation volume of 20 m³ air for adults and 10 m³ for children is used.

Oral exposure

EUSES applies default daily intakes for 5 broad EUSES food categories (“fish”, “leaf crops including fruit and cereals”, “root crops”, “meat” and “dairy products”), in combination with predicted levels of Pb in those biotic matrices (see Section “Modelling of Pb in the environment”). This approach includes a very conservative assumption that the population in the neighbourhood of a lead battery production/recycling site is exposed to 100 % locally grown food and has a daily intake equal to the highest country-average across European countries for each food category [13].

Higher tier dietary exposure modelling is on one side based on more detailed, narrower food categories, (e.g. category “leaf crop” split into cereal products (grain), leafy vegetables (sensu stricto) and fruits) and on the other side on food category differentiated intake rates, based on recent dietary surveys in the EU.

In order to develop a higher tier (Tier 2) dietary approach for MvE lead exposure, dietary surveys from the EFSA were used as basis. The list of EFSA’s food categories were turned into categories of agricultural products (cfr. categories mentioned in Table 1), to enable the link with the environmental exposure module.

More information can be found in Supplementary information (see Tables S1–S8). The daily intakes selected per food category and corresponding environmental transfer factors can be found in Table 2.

Integrated exposure assessment—modelling Pb levels in blood

In order to convert external oral and inhalation exposure values to predicted Pb levels in blood, a factor converting a daily dose (mg/kg/d) or an air concentration (µg/m³) to the corresponding blood level (µg Pb/l) is needed.

For the conversion of external oral exposure to predicted Pb levels in blood for children, a daily intake of 0.00108 mg/day/kg body weight resulting in increase of 19.48 µg Pb/l blood, was used as basis for the conversion factor of 18037 µg Pb blood/ mg Pb oral intake. These numbers are based on a recent evaluation of dietary intake by EFSA and used in the Annex XV restriction proposal for lead compounds-PVC [14]. For adults, the EFSA 2013 conversion factor of 23809 µg Pb blood/ mg Pb oral intake is used, which is based on a daily intake calculated with the Carlisle and Wade algorithm for adults [15].

For the conversion of external inhalation exposure to predicted Pb levels in blood, the intake values considered by WHO during the calculation of a guideline value for ambient air in Europe are 16 and 19 µg/l for adults and children respectively, at 1 µg Pb/m³. The corresponding conversion factors are thus 16,000 and 19,000 µg/l per mg Pb/m³ for adults and children respectively. These slope factors apply principally for blood Pb levels below or equal to 30 µg/l [16] that are expected in general public rather than occupationally exposed populations.

Data gathering—case study

This tiered modelling approach was applied to a case study of 50 lead battery manufacturing and recycling sites across Europe. Relevant exposure data and input for the modelling tools, has been collected from the producers and recyclers through questionnaires. The questionnaire included inter alia data requests regarding tonnages, releases to the environment, characteristics of stack (emissions), availability of air quality monitoring data, characteristics and land destination of the surroundings, size of nearby population, HBM data in the neighbourhood.

Results and discussion

MvE exposure tiered approach

The framework of the tiered approach for MvE exposure is shown in Fig. 1. The approach starts at Tier 1 level (‘EUSES+’), which is based on EUSES equations and EUSES defaults [2], with the following adaptations:

  • Use of metal transfer factors to model the transfer from soil to leaf and root crops (instead of the current log Kow based equations in EUSES)

  • Inclusion of soil/dust ingestion route

  • Differentiation of exposure calculations according to age categories (two categories: adults and 1–3 years old children)

Fig. 1
figure 1

Tools and methods used at different Tier levels of the framework.

Full size image

At the end of Tier 1 phase, the key dominant pathways for children and adults have been identified and selected for further refinement in Tier 2. Notwithstanding that the dominance of pathways varied across sites, three exposure pathways (ingestion of soil and dust, dietary intake of meat, ingestion of drinking water) appeared to be non-dominant (<10%) across nearly all sites (data not shown). For these exposure pathways, no further refinements were needed.

The pathways of inhalation exposure and dietary exposure (via crops, milk and fish consumption) appeared to be dominant in Tier 1 calculations. It should be noted that the dominance of exposure pathways can be highly variable across sites. For one site, Tier 1 assessment of fish consumption appeared to be the dominant (>90%) pathway, while for another site, inhalation exposure was most important.

A Tier 2 approach for these potential dominant pathways was elaborated (see Fig. 1).

For the inhalation route, shifting from Tier 1 to Tier 2 involved the use of available air monitoring data, and higher tier exposure models. Priority was given to monitoring data over exposure modelling. Air monitoring data, when available, was collected via questionnaires from the companies. The annual average Pb ambient air concentrations reported by the companies were selected because of the relevance of this exposure metric for chronic exposure.

In case of lack of monitoring data, GPM modelled Pb concentrations in ambient air were used. In Tier 2, a conservative assumption was made that the population living in the neighbourhood of a respective site is exposed to levels reported for the monitoring station or the maximum modelled concentrations from the spectrum of wind velocities, wind direction or distances from the GPM output model

For the dietary exposure route, the Tier 2 approach involved a breakdown of the EUSES group ‘Leaf crops including fruit and cereals’ and ‘root crops’ into narrower categories in Tier 2 based on EFSA food consumption data, allowing the combination of category-specific transfer factors, and category-specific differentiation of intake rates (Table 3). Refinement focused on these food categories because exposure is often driven by the relative high contribution of these crops to the dietary intake in the Tier 1 calculations.

Table 3 Man via the environment oral exposure scenario parameters.
Full size table

Tier 2 followed the conservative assumption that populations in neighbourhood of lead battery production or recycling plants are exposed to 100% to locally grown food.

For those sites where the Tier 2 approach resulted in predicted Pb blood levels above 5 µg Pb/L, a Tier 3 assessment was considered. Hereto, additional site-specific information was used to refine the assessment.

The main field for further improvement at Tier 3 for inhalation exposure was the relevance of air quality data in relation to the population in the neighbourhood. In case there were multiple values of annual average Pb ambient air concentration for one site from more than one monitoring point around the site, the highest value was taken forward to calculate the integrated dose in the Tier 2 assessment, irrespective of whether people effectively live and are chronically exposed at this location. In Tier 3, data from monitoring points located at places with no relevance for human inhabitation were excluded. If information available from questionnaires was inconclusive, additional open data sources (e.g. satellite images from Google Maps) were used to investigate the presence and proximity of any residential building in the neighbourhood.

Regarding dietary exposure, the main field of refinement in Tier 3 was the assessment of the proximity and proportion of locally grown food commodities.

Information regarding neighbouring populations (residences) with home gardens, agricultural activities and likely fishing activities in receiving water bodies was retrieved from questionnaires. If information available from questionnaires was inconclusive, above mentioned open data sources were consulted to investigate the home gardens, agricultural land and water bodies in the neighbourhood. If no agriculture or fishing possibilities were present in the influence zone of the site, the additional dietary intake from respectively agricultural products or locally caught fish was considered to be negligible. In case of presence of a local farm in the influence zone, the default assumption of 100% local origin of food was retained. For home gardens and other types of agriculture where dilution in the regional market is likely, a local contribution factor of 30% for home gardens was used, and a 10% contribution of local scale to regional agriculture. The default factor of 30% consumption is based on the upper range of consumption of home garden products as part of total consumption in France and UK [17, 18].

Exposure estimates resulting from tiered exposure modelling framework in the case study of 50 lead battery manufacturing and recycling plants

Regional scale

The impact of lead battery manufacturing and recycling plants at the regional scale, assessed following the Tier 1 approach results in additional Pb levels in blood of 3.7 µg Pb/l blood for 1–3 years children, and 1.6 µg Pb/l blood for adults. Dietary intake from leafy and root crops was the dominant exposure route (ca. 92%), with transfer from roots having a larger contribution compared to aerial deposition of Pb on crops. Consumption of milk products was the second dominant route of exposure, especially for children. Ingestion of Pb via drinking water, soil and dust ingestion and meat contribute each to less than 5% Dominant contribution of food species from vegetable origin was in line with EFSA’s assessment [19].

The contribution of inhalation exposure to Pb was negligible at the regional scale. The Tier 2 approach for dietary exposure, resulted in lower predicted levels (0.14 µg Pb/l blood for 1–3 years children) compared to the Tier 1 approach.

These estimates were low compared to EFSA’s estimated dietary exposure (0.36-1.24 µg/kg bw for average adult consumers, and 0.8–3.1 µg/kg bw for 2 years old children), corresponding to blood levels of 8.6–30 µg Pb/l (adults) and 14.4–56 µg Pb/l (children), following conversion factors mentioned in Section “Integrated exposure assessment—modelling Pb levels in blood”. This is not surprising since the latter reflect all sources of Pb in the diet (including other industrial sources, historical contamination and natural background), while our estimates only model recent contribution of Pb batteries and the associated recycling sector.

Local scale

Predicted exposure for residents living in the neighbourhood of lead battery manufacturing and recycling sites according to the different tiers of the exposure modelling framework is shown in Fig. 2. The predicted additional levels of Pb in blood of 1–3 years old children caused by the lead batteries production and recycling plants varied from small (0.3 µg Pb/l blood − 5th percentile, P5) over 1.52 µg/l Pb/l blood (P50), 3.8 µg/l Pb/l blood (P75) to 9.9 µg/l Pb/l blood (P90 value).

Fig. 2: Predicted additional total daily Pb intake for human  arising from industrial activities near industrial facilities according to differnt tiered levels.
figure 2

Upper left figure: data for children exposure around lead battery production sites. Upper right figure: data for adults around lead battery production sites. Lower left figure: data for children exposure around battery recycling sites. Lower right figure: data for adults around battery recycling sites. SBP systolic blood pressure (thresholds derived EFSA, 2010).

Full size image

Also at Tier 2 and Tier 3 of the modelling approach, a large variability in exposure across sites was found: a 100-fold difference in predicted additional exposure due to industrial activities in the lead battery production sector and a near 40-fold difference in the lead battery recycling sector.

Dominance of exposure routes varied largely across sites. In general, ingestion of soil and dust, dietary intake of meat, ingestion of drinking water appeared to be non-dominant (<10%) while inhalation of air, consumption of crops and milk are more important exposure routes. The predicted Pb levels in crops are dominated by soil to root transfer, rather than via direct aerial deposition on crop leaves. The relative contribution of exposure routes for the P50 exposure of the lead battery production and recycling sector is shown in Fig. 3. Shifting from Tier 1 to Tier 2 approach resulted on average in a factor 20-fold reduction in predicted Pb exposure compared to the Tier 1 calculations. (see Fig. 2).

Fig. 3: Relative contribution of exposure routes to median internal Pb levels.
figure 3

Upper left figure: data for adults around lead batteries manufacturing sites. Upper right figure: data for adults around lead recycling sites. Lower left figure: data  for children  around lead batteries manufacturing sites. Lower right figure: data for children around lead reycling sites.

Full size image

The route and sources of exposure involved in the reduction of predicted exposure at Tier 2 compared to Tier 1, was rather diverse and varied largely across sites. For some sites, where EUSES modelled Pb air concentrations dominated exposure in Tier 1, the reduction was mainly realised by replacing EUSES modelled air data by air monitoring data or GPM modelled air concentrations. For other sites where the dietary exposure dominated exposure in Tier 1, a reduction was due to a reduced dietary intake based on crop and food commodity specific transfer factors, and/or a shift in BCF fish by considering only edible fish (instead of BCF fish from the environmental assessment).

It was investigated to perform Tier 3 modelling for sites with Tier 2 predicted additional Pb levels exceeding 5 µg/l (n = 12). For two of these sites input from companies was insufficient to allow Tier 3 modelling, for the other 10 sites Tier 3 assessment was applied. For these sites, shifting from Tier 2 to Tier 3 approach resulted in a further reduction in predicted Pb exposure of 2–30% (impacting mainly the lead battery manufacturing sector). The following refinements were applied in Tier 3: consumption of locally caught freshwater fish at the discharge point was confirmed as not relevant (1 site) and improvement in the inhalation exposure assessment (9 sites). For example, for one site, the Tier 2 assessment was based on the maximum GPM modelled concentration around the site corresponding to a distance <100 m distance the stacks. Based on questionnaires and checks using satellite images, it appeared that the closest inhabitation was at 2.2 km distance from the site. A refined modelling accounting for this inhabitation relevant location, resulted in an exposure 50% lower than the Tier 2 calculated exposure for this site.

For some sites, a further refinement of exposure predictions could be realised if air quality monitoring data in the neighbourhood would become available. Lastly, for two sites, the Tier 3 predicted exposure was still high (e.g. predicted levels of additional >12 µg Pb/l children), even when based on monitoring data in relation to relevance of nearby population. In these cases, (1) biomonitoring of local populations could be helpful to investigate the impact of industrial activities on the body burden of residents, (2) further follow-up of time trends in environmental exposure and the related MvE is advised and (3) additional risk managements measures could be considered at the site.

Comparison with human biomonitoring data

Regional scale

A recent overview of HBM data on Pb levels in blood in the European population was made in the framework of the HBM4EU project [20] and summarised here:

Surveys measuring blood lead levels in the general population have been conducted in several countries since the early 1980s. After phasing out lead from petrol in Europe, general population blood lead levels fell drastically. Results of blood lead level surveys conducted during the past two decades among the general population (including children) are available for 16 European countries. A decreasing trend in blood lead level of children could be observed with the phasing out of leaded petrol in various countries [20].

However, unfortunately, there are few data on current blood lead levels found in the general population in European countries today. Only 7 countries (Belgium, Germany, Denmark, Kosovo, Poland, Slovenia and Sweden) were found to have reported general population blood lead levels after 2014 [20].

The predicted contribution of Pb in human blood arising from emissions from lead battery manufacturing and recycling for the regional scale was 0.15 µg Pb/l for children (1–3 years) and 0.06 µg Pb/l for adults. This value is about 1% of total Pb blood levels, according to available recent monitoring data for children <7 years across Europe (e.g. new-borns in Belgium (2012–2015): mean 9.5 µg Pb/l; 1–6 years old children in France (2008–2009): 14.9 µg Pb/l; 3–6 years old children in Poland (2013): 24.7 µg Pb/l); and less than 1% compared to recent monitoring data for adults across Europe (e.g. adults in Spain (2007–2010): mean 24 µg/l; adults in Sweden (2004–2014; 25–35 years men: 11.1 µg/l; 50–60 years men: 15.1 µg/l; 25–35 years women: 9.7 µg/l; 50–60 years women: 13.1 µg/l; adults in Italy: GM 19.9 µg/l) (source: [20]).

Local scale

For several hotspots near (past) Pb industrial activities like mining, refining and use of Pb substances, levels of Pb in blood of children are known to be elevated compared to similar populations in non-industrial sites [3, 21], but little publicly available blood lead data is reported for populations living in the vicinity of lead battery manufacturing and recycling facilities in Europe.

Two sites included in our case study reported Pb levels in children living in the neighbourhood of their facility: one site reported Pb monitoring data in school children on average 12–13 µg Pb/l (data from 2006 to 2012); and a second site reported as a mean value of 27 µg Pb/l based on 105 children 2.5–6 years old (data from 2009).

The predicted contribution of Pb in human blood arising from lead battery recycling for those sites was about 4 and 8% compared to the recent available biomonitoring data provided for children.

Data on blood lead levels in workers prior to occupation were more abundant (19 companies) and may reflect non-occupational exposure of adults in the local neighbourhood, assuming that employees live rather close to the company. The majority of the resulted data are reported in exposure ranges or reporting against a limit of quantification. The highest reported values were a range of 20–163 µg/l and a range of 10–150 µg/l (P90: 58 µg/l); while on the other side of the ranges, also low values (e.g. 10.2 µg/l) have been reported. Thus, also for adults, only for two sites a comparison of predicted versus monitoring data was possible because for all other sites, the reporting ranges were too wide. The predicted contribution of Pb in human blood arising from lead battery manufacturing for those two sites was 19 and 25% compared to the HBM data of new employees for these sites.

For these two sites the data indicate that the contribution of Pb in human blood arising from lead battery manufacturing or recycling is at the local scale rather very small (<1%) to significant (25%) compared to other sources of Pb exposure. Additionally, the levels of the available non-occupationally exposed population blood lead data from the local sites are in the same ranges as those reported for the regional scale based on HBM4EU data [20].

These conclusions cannot be expanded to other sites (and thus is not representative for the sector. Therefore, interpretation of comparisons of predicted levels at the local scale with available HBM data must be done with caution.

Reflections regarding comparison of predicted versus measured levels of Pb in blood

The comparison between predicted levels and HBM data is complicated by the fact that whereas lead in food and beverages is the primary source of non-occupational lead exposure, it is difficult to accurately apportion the source of this lead. It may originate from on-going anthropogenic activity. However, other sources may originate from legacy uses such as closed industrial facilities, old lead piping/solders for potable drinking water, gasoline, paint, toys, consumer articles, cookware or from the burning of fossil fuels for power generation etc [3]. Many of these uses are now restricted in the EU or cannot be further controlled through chemical use regulations like REACH.

Finally, modelling approaches are a simplification of the complex real-life processes (routes of exposure, human behaviour affecting exposure factors, internal dosimetry, etc.), and sometimes lead to significant over- or underpredicting of exposure because of various reasons [22]. Therefore further model verification is well advised before it is used for regulatory decision making.

The predictivity of the tier 2 intake model used for this case study was assessed by entering Pb monitoring data from Belgium in different environmental matrices (air, water, soil, diet); thus reflecting all likely sources of lead present in the environment. This resulted in predicted Pb in human blood of 11.7 µg Pb/l (children), compared with available HBM monitoring data of 16.6 µg/l for children of ages 2.5–6 years. This gives us some confidence in the model which was used for the prediction of Pb in this case study.

The tiered approach framework for MvE exposure assessment used in this study can be used in the context of fulfilling REACH exposure and risk assessment requirements, both for lead and potentially for other metal substances where the default approaches) are not applicable or fail to demonstrate safe use. It can also be used in contexts where a realistic exposure estimates, rather than a precautionary safe use demonstration is warranted, such as for impact assessment for socio-economic cost of exposure analysis.