AGRICULTURAL AND FOOD SCIENCE Agricultural and Food Science (2025) 34: 38–50 38 https://doi.org/10.23986/afsci.145477 Comparison of the effects of bio-based and mineral fertiliser use on heavy metals dietary exposure in six European countries María D.R. Domínguez Carrasco1, Tapio Salo2, Riikka Keskinen2 and Johanna Suomi1 1Finnish Food Authority, Helsinki, Finland 2Natural Resources Institute Finland (Luke), Tietotie 4, FI-31600 Jokioinen, Finland e-mail: Johanna.suomi@ruokavirasto.fi This study compared the effects of using bio-based fertilisers (BBFs) of high (BBFH) or low (BBFL) metal content with conventional mineral superphosphate (SP) fertilisation on dietary exposure of arsenic (As), cadmium (Cd), nickel (Ni) and lead (Pb), in Finland, Denmark, France, Germany, Hungary, and Spain. We estimated changes in the metal content of the following crops: wheat, barley, oat, rye, potato, carrot, sunflower seed, and maize following 100- year scenarios of fertilization with the different products. To estimate changes in chronic dietary exposure to met- als via food, we used available national data of food consumption and metal content in target crops. Our results showed low changes to current chronic dietary exposure after using BBFs. Exceedance of the maximum allowed levels of Cd and Pb in food (EU 2023/915) was more rare with low-Cd BBF than with SP mineral fertilizer. Only the BBFH slightly increased the dietary exposure to Cd, although similarly to SP. In conclusion, the studied BBFs did not increase dietary exposure to heavy metals, especially compared with the use of conventional SP mineral fertiliser. Key words: recycling, health, risk, hazard, agriculture, contaminant Introduction Environmental impacts from production and excessive use of virgin synthetic and mineral fertilisers encompass greenhouse gas emissions during Haber-Bosch process (nitrogen-based fertilisers manufacturing), mining impacts (phosphorus extractions), and water systems eutrophication as a consequence of nitrogen and phosphorus over- loading from agricultural soils (Guan et al. 2017, Hussain et al. 2021, Osorio-Tejada et al. 2022, Dias Blanco et al. 2023) social, and also a food sovereignty challenge for several countries. Due to the constantly growing global population, the food and feed production, and in turn, fertiliser needs are further rising (EC 2012, IPCC 2019). To minimize environmental impacts and enhance sustainability of food production, increasing circularity through the development of biobased fertilisers (BBFs) augments. BBFs can be produced from, e.g., treated sewage sludge, animal waste or food production by-products such as bones and feathers or crop residues. Depending on the source and processing methods, BBFs may contain different ecotoxicological hazards (Chojnacka et al. 2020), which can be harmful to human health. For instance, dioxins, perfluorinated alkylated substances (PFAS), endocrine disrupting compounds, drug residues, antimicrobial resistant microbes, or heavy metals (HM) are some examples of these hazards. Though some HMs are essential for living organisms in trace levels, in gen- eral HMs are toxic, especially for humans. Cadmium (Cd), lead (Pb), inorganic arsenic (iAs) and nickel (Ni), target- ed in this paper, are known to cause cancer, bone diseases or disorders in kidney, cardiovascular and neurological systems, and death (EFSA 2009a Cd, EFSA 2010 Pb, JECFA 2011 iAs, EFSA 2020 Ni, WHO 2007). In human body, the retention time of HM can range, depending on the organ, from days (Ni) to decades (Pb in bone, Cd in kidney) (e.g. see EFSA Journal reports for Cd (2012), Pb (2010) or Ni (2020) taking into account new occurrence data, the updated benchmark dose (BMD). The limits of HM content in fertilisers and food products are thus well regulat- ed in Europe by, e.g., Commission Regulations (EU) 2023/915 and (EU) 2019/1009 as well as (EU) 2022/973, and at national level. Although HMs are naturally present in Earth´s crust, human activities such as industry, urban waste, mining and agriculture contribute to HM pollution (Tchounwou et al. 2012, Saleh and Hassan 2022). Fertilisers containing HMs can increase the soil content of HM or enhance the mobility of existing HM reserves in soil. Metals enter in the food chain from soil through accumulation in crops´ tissues, however the type of fertiliser affects the crop metal content (Shumba et al. 2014). Shumba (2014) found that maize had the highest content of HMs (Cd, copper (Cu), Ni, and zinc [Zn]) after using sewage sludge-based fertiliser. Development of BBF processing and the assessment of HM bioaccumulation in environment and foods will minimize environmental and human health risks of using BBFs in agriculture. Received 3 May 2024 / Accepted 7 February 2025 The Scientific Agricultural Society of Finland ©This is an open access article under the CC BY 4.0 M.D.R. Domínguez Carrasco et al. 39 Cereals and vegetables are the main source of human dietary exposure to HMs, and for example for Finnish adults, their contribution was up to 30% of the total dietary exposure to the HM (Suomi et al. 2020, 2021). According to Khan et al. (2015) up to 90% of total HM human intake have crop origin. In addition, vegetarian, and cereal-based diet might grow as an environmentally friendly alternative to meat-based diet. Hence, HM exposure at popula- tion level might rise unless effective risk management strategies are implemented. Toxicological dose-response studies determine adverse effects (response) caused at certain intake levels (dose). Based on these results, international food and health councils, like Food and Agriculture Organization of the Unit- ed Nations (FAO), World Health Organization (WHO) or European Food Safety Authority (EFSA), develop health- based guidance values for risk assessment. Tolerable daily or weekly intake (TDI and TWI, respectively) values are maximum doses with which no adverse effect is likely over a lifetime exposure. Benchmark dose lower confidence limit (BMDL) values, on the other hand, are determined directly from the dose-response data and they still require a safety factor to arrive at the negligible risk dose. The toxicity of the HM greatly depends on the metal, its species, and on the HM exposure frequency. In humans, the dietary characteristics, the gender and the age, play important role in HM toxicity (Suomi et al. 2021). For in- stance, HM toxicity in children is more severe compared to adults, due to organs formation and regeneration of new cells. In addition, children consume more food per bodyweight than adults do. In this sense, it is important to include age factor in risk assessments of HM dietary exposure. The human HM dietary exposure is a function of the hazard occurrence in food and of the amount and frequency of food consumed (Ranta et al. 2021). The cumulative effects from simultaneous exposure to several chemicals with the same mode of action in the body should also be addressed at a later point, but in a previous study the main component of the cumulative health effects of dietary heavy metals (from the whole diet) was found to be Pb (Suomi and Tuominen 2023). In addition to the way HMs from fertilisers are taken up by plants, the uptake of HMs from plants to cow meat and milk may be of importance. This study is part of Lex4Bio project (Horizon 2020 grant agreement No 818309) whose main aim is to develop BBFs sustainable and safe to use. This paper is done in parallel with the work of Salo et al. 2025 (submitted). While Salo et al. focus on estimation of HM occurrence in crops in 100 years based on a bass mass model, this paper used these results on crop´s HM occurrence and is focussed on the assessment of HM content in BBFs and the as- sociated risk to human health through using BBFs instead of mineral fertilizers. This study compared changes on current HM occurrence in food and HM dietary exposure after using BBFs against traditional mineral fertilisation. The specific aims of this paper are the following: a) To study current occurrence of Cd, Pb, iAs, and Ni in eight crops (wheat, rye, barley, maize, potato, carrot, and sunflower seeds) from the following five countries: Spain, Denmark, Germany, Hungary, and Finland. b) To investigate the effects of BBFs and mineral superphosphate (SP) fertiliser use on the HM occurrence in the target crops and countries in a 100-year time scale. c) To analyse Cd and Pb content in the studied crops against EC regulated maximum levels, after using BBFs and SP fertiliser. d) To study changes in HM dietary exposure after using BBFs at different HM content and compared with SP mineral fertiliser. Dietary exposure of adults (18 to 64 years) and children (3 to 10 years, except Finland 3 to 6 years) from Finland, Denmark, Germany, France, Hungary and Spain was studied. Despite the lack of current occurrence data, a rough estimate was possible for France, which was not included in aim (a). Due to better data availability, HM exposure of Finnish adults was assessed with a more precise method than the other population groups. Materials and methods All data harmonization, analyses, and graphs used for this paper, explained below, have been done using RStudio (Posit team 2023). Agricultural and Food Science (2025) 34: 38–50 40 Effects of fertilisation on the current heavy metal occurrence in European crops To study the effects of mineral and BBFs fertilisation in HM occurrence in European crops, we first surveyed the current situation in five European countries participating in the Lex4Bio project (excluding France, see explanations in following subsection). Thereafter, the relative change from this current level induced by 100-years of fertiliza- tion with a BBF high or low in metals or conventional SP was calculated based on a mass balance model (report- ed separately by Salo et al. 2025). Finally, this proportional change in content was used to estimate the changes in current density distribution of HM occurrence data and to estimate changes in dietary exposure (explained in the second section of Methods). Current occurrence of HM in European crops To study the current HM occurrence values in European crops we used data (from 2000 to 2021) of Cd, Pb, iAs, and Ni content in wheat, oat, barley, rye, maize, sunflower seeds, potato, and carrot (hereinafter LexCrops) origi- nated and analysed in Finland, Spain, Germany, Hungary, and Denmark. Finnish data from years 2000 to 2017 and 2020 were provided by the Finnish Cereal Committee (VYR) and previously collected for national risk assessment reported in Suomi et al. (2020, 2021). The main bulk of occurrence data (2017–2021) was requested from EFSA via Public Access to Documents (hereinafter, PAD-data) (regulation (EC) No 1049/2001). These data had been sub- mitted to the EFSA Data Warehouse by the corresponding EU Member States in the context of a call for continu- ous collection of chemical contaminants occurrence data in food and feed. To clarify, the food samples in the en- tire dataset used in this study were taken and analysed in the context of national monitoring or, to a much lesser extent, research by national authorities. We requested from data providers the authorisation to scientifically use their data included in the PAD-dataset (request PAD 2022/144). Spain, Germany, Denmark, Hungary, and Finland granted the permission, and only those results were used in analysis of occurrence data, although codifying coun- tries´ names for the sake of confidentiality. Sample-level results were available at the EFSA food coding system FoodEx2 levels L1 to L3, mainly at the more detailed levels. HM content analyses were done in accredited laboratories at each country. In total, the dataset used comprised 10882 samples of Pb, 9105 of Cd, 6012 of total As, 441 of iAs, and 5715 of Ni (see Table 1s in Supplementary ma- terial). Analysis methods and number of hazard-food pairs analysed varied among countries and years (Table 1s). However, these variabilities did not affect our statistical analyses cause the aim of this project was not to compare occurrence data geographically or chronologically. Arsenic was mostly analysed as total arsenic instead of inorganic and organic arsenic independently. We estimat- ed missing measurements of inorganic arsenic as the 70% of the total arsenic reported, which is the factor also used by EFSA (EFSA 2009a) for plant-based foods. In addition to numerical occurrence data, a part of the data was left-censored, which means that their HM con- tent was lower than limit of quantification (LOQ) (range summarised in Table 1s) and/or limit of observation (LOD) (values not shown). A traditional way of handling left-censored data is using lower bound (LB) and/or upper bound (UB) scenarios, where values below the limit of reporting (LOD or LOQ) are included as zero values (LB) or as val- ues equal to the limit of reporting (UB). After data cleaning (see Data Cleaning and uncertainties, in Supplementary material), the lower bound values were calculated. The mean and range of contents for each food-hazard pair in each target country are summarized in Table 2s, the results are shown only for cases where the number of samples was eight or higher. The current distribution of HM occurrence in LexCrops was considered the starting point for analyses of BBFs and SP effects on content, explained in the following sections. However, due to the lack of individual-level food con- sumption data for target population, except for Finnish adults, we were unable to use the current HM occurrence data in the analyses of dietary exposure changes in other cases than for Finnish adults. HM occurrence data was used in BIKE model to estimate the HM exposure of Finnish adults (explained below in this article). Calculation of HM content change factor induced by fertilisers use To assess the effects of BBFs on the crop HM content and further on current dietary exposure, we conducted 100- year mass balance modelling for several crops in conditions of Spain, Denmark, Hungary, Finland, Germany and France (Salo et al. 2025). To compare with BBFs effects, we also used typical mineral superphosphate fertiliser M.D.R. Domínguez Carrasco et al. 41 (SP) of each LexCountry. Results of SP uses were compared with lower (L) and higher (H) HM concentrated BBFs. In addition, we used a very high Cd-concentrated BBFE, close to the EC maximum level, just to see the effect on the Cd content in wheat (Table 3s). Results and details of mass-balance model will be reported in a separate paper (Salo et al. 2025). In brief, the met- al inputs to and losses from a representative cropland in each country were estimated using best available pub- lished literature values or values calculated from experimental data of the Lex4Bio project. The input values for the model included total metal content in soil, aerial deposition, annual runoff and erosion, average liming and phosphorous fertilisation rates, metal contents in lime and fertilisers, average yield level of the target crops, soil- to-plant transfer factor and soil/water partition coefficient. For mineral fertilizer, typical HM contents in superphos- phate of each country were used, whereas for BBFs the calculations were carried out with 2–4 varying contents adopted from screening of European BBF products under the Lex4Bio project. Wheat, barley, oat, rye, maize, and sunflower seed were modelled as monocultures, whereas potato and carrot were rotated with barley. The model produced annual estimates of HM mass balance, annual HM loss through leaching and erosion, HM content in runoff water, and soil, and yield uptake and content at current state (S0) and after 100-years (S100). The relative change in HM content (Sch) with time (i.e., S100/S0) was considered the change factor induced by the BBF. Sch (re- sults in Table 3s) was further multiplied by current HM occurrence and dietary exposure data to estimate changes in content and foodborne exposure respectively in LexCountry´s crops and population. Effects of BBFs and SP in current HM occurrence data in European crops The relative content change factor Sch estimated with HM mass-balance model was multiplied by current occur- rence HM data at lower bound (explained in Current occurrence of HM in European crops). To visualize the effects of BBFs and SP in HM occurrence, we plotted the current density curve (Before in Fig. 1s) overlapping with density curves obtained after using Sch for BBFH or BBFL (Fig. 1s). We analysed the number of cases in which the current contents of Cd and Pb (C in Table 3) exceeded the Europe- an Commission maximum levels in foods (EC_ML) regulated in (EU) 2023/915, pooling data of all countries. After multiplying current HM content values by Sch at each fertiliser use scenario, we repeated calculations of the num- ber of cases exceeding Cd and Pb EC_ML after using BBFs (H, L in Table 3) and SP, followed by estimations of the relative proportion (in %) of cases after BBFs and SP uses out of the current situation (Table 3). Effects of fertilization on current dietary exposure to heavy metals in European population Current dietary exposure assessments Due to the different source and quality of food consumption data, estimations of HM dietary exposure through food products derived from LexCrops (hereinafter Lexfood) were different for Finnish population compared of that for the rest of the cases. In addition, we estimated separately HM exposure for Finnish adults using BIKE model (explained in Finnish adults´ dietary exposure assessment before and after BBFs, BIKE model). Dietary exposure is compared with health-based guidance values summarized in Table 1. Exposure below TWI or TDI value is of negligible risk. For Pb, the margin of exposure between the BMDL value and total exposure should be at minimum 10 for negligible risk, and for iAs, at least 1000 due to its carcinogenic properties. In all exposure assessments, we assumed that neither the food consumption behaviour of the population groups nor the contents in other components of the diet than Lexfoods would change over the studied period. Table 1. Toxicological reference values used in this paper. T(W/D)I is Tolerable (weekly or daily respectively) intake, and BMDL(i) is the benchmark dose at lower confidence limit for a (1 or 0.5% respectively) of cases. Heavy metal Toxicological reference value and its type Reference Cd TWI=2.5 µg kg-1 bw week-1 EFSA (2009) Pb BMDL01= 0.5 µg kg-1 bw d-1 EFSA (2010) Ni TDI = 13.5 µg kg-1 bw d-1 EFSA (2020) iAs BMDL0.5=3.0 µg kg-1 bw d-1 FAO/WHO (2011) Agricultural and Food Science (2025) 34: 38–50 42 For Spain, Germany, France, Hungary, and Denmark we used the total dietary exposure reported by EFSA (EFSA 2009a&b, 2012, 2020, and 2021 for Cd, Pb, Ni and iAs respectively, see Table 2a) in adults (18 to 64 years) and children (3 to 10 years) based on general food categories instead of ingredient levels. We first estimated the con- tribution of Lexfood categories to the total dietary exposure, considering that all ingredients of studied food cat- egory contributed equally to the 100% of the EFSA reported total dietary exposure. To estimate the contribution of Lexfoods to the total dietary exposure we first collected consumption information of food containing Lexfood as ingredients, from EFSA Food Consumption Database (URL https://www.efsa.europa.eu/en/data-report/food-con- sumption-data). EFSA Food Consumption database contains data on nationally conducted surveys, uploaded to EFSA by the national health authorities, the data producers of surveys. Only the statistical indicators at age- and gender- groups are available as open access, and these were used by us. We selected food consumption values from the same surveys that EFSA used in the dietary exposure reports for each HM. We selected food categories from L1 to L5 in the FoodEx2 coding system, out of which L1 is the roughest and L5 the most detailed categoriza- tion. From L1 we selected the following food groups: a) grains and grain-based products, containing wheat, oat, barley, rye and maize; b) starchy roots or tubers and products thereof, sugar plants, containing potato; c) vegeta- bles and vegetable products, including carrot; and d) legumes, nuts, oilseeds and spices, which included sunflow- er seeds. Finally, we estimated the content of LexCrop in L5 categories using recipes found in internet and choos- ing those mixed foods either with more than 60% of LexCrop or with high consumption in corresponding age- or country-groups. Finally, the relative Lexfood out of the total food consumed was multiplied by the total HM dietary exposure to estimate the current Lexfood dietary exposure at each age- and country-group. For Finnish adults (25 to 64 years) and Finnish children (3 to 6 years) we used dietary exposure values calculated in the national risk assessments Suomi et al. (2020) and (2015) respectively, the main results of which were published in Suomi et al. (2018) and (2021). In the national risk assessments, the food consumption data were collected in surveys FinDiet 2007 and FinDiet 2012 (adults, whole country) and DIPP (children, regional). For the current study, we used the older 2007 consumption data to estimate the exposure from other food sources, as it was entirely calculated to ingredients. For comparison, Table 2b also includes the corresponding EFSA estimates for Finland with these food consumption data. EU-level data used by EFSA usually gives a more conservative estimate than national data, and the difference in iAs results in Table 2b is caused by the uncertainty from using conversion fac- tors (total As to iAs) in the national assessment. Table 2a. Current chronic dietary exposure to HM from the whole diet according to EFSA’s reports (EFSA 2012 Cd; EFSA 2012 Pb; EFSA 2021 iAs; EFSA 2020 Ni). Values are lower bound estimates (