Introduction

Mercury, a naturally occurring metal found in the Earth’s crust, exists in various chemical forms including elemental mercury, inorganic mercury compounds, and organic mercury compounds, the most notable of which is methylmercury (World Health Organization (WHO), 2011). Methylmercury, formed by the microbial conversion of inorganic mercury in aquatic systems, is the primary concern when assessing the risks of mercury exposure, as it accumulates in the food chain and bioaccumulates in the tissues of fish and other aquatic organisms (Clarkson 2002). Exposure to mercury occurs through multiple routes, including inhalation of elemental mercury vapor, ingestion of contaminated food and water, and dermal contact with mercury-containing products, with the consumption of fish and seafood being a major source of exposure for the population (European Food Safety Authority [EFSA], 2012). Apart from dietary intake, other potential sources of mercury exposure include dental amalgam fillings, occupational settings (e.g., certain industrial processes), and ambient air pollution from natural and anthropogenic sources (Clarkson 2002) (EFSA 2012). Adverse effects of mercury on human health are well-documented. Chronic exposure to methylmercury, primarily through fish consumption, can lead to neurodevelopmental deficits in children, including cognitive impairments, learning disabilities, and delays in motor skills. In adults, exposure to high levels of mercury can result in neurological symptoms, cardiovascular effects, and renal dysfunction (EFSA 2012). To protect human health from exposure to chemical substances, international entities like the Joint FAO/WHO Expert Committee on Food Additives (JECFA) and European Food Safety Authority (EFSA) established health-based guidance values, such as the Tolerable Weekly Intake (TWI) for methyl and inorganic mercury (EFSA 2012; Joint FAO/WHO Expert Committee on Food Additives [JECFA], 2004).

Portugal’s unique cultural and culinary traditions place a substantial emphasis on seafood and fish consumption. In 2021, Portugal ranked sixth globally in terms of average fish and seafood consumption per capita, with 59.41 kg consumed annually, based on FAO data cited by Our World in Data (Our World in Data 2021). This high consumption makes the Portuguese population particularly exposed to mercury and underscores the need to perform exposure and risk assessment studies that provide scientific evidence to support adequate measures for public health protection.

A Total Diet Study (TDS) is a valuable public health tool for assessing the exposure of a population to various chemical substances such as contaminants and nutrients, reflecting real dietary habits. These studies provide a comprehensive view of dietary exposure by analyzing food items as typically consumed. They are essential for designing effective strategies to manage and reduce exposure (Vasco et al. 2021). Initiatives to harmonise TDS methodologies contribute to a more standardised assessment of dietary exposure and improves comparability across countries (EFSA, Food and Agriculture Organization of the United Nations [FAO], WHO 2011; Pité et al. 2018). In summary, mercury exposure might be a significant concern for the Portuguese population, particularly through seafood consumption, and its impact on human health underscores the importance of assessing and managing exposure. The main objective of the present study was to assess the baseline chronic dietary exposure of the Portuguese population to methylmercury and inorganic mercury using a harmonised TDS methodology and to estimate the population’s risk of exceeding the health-based guidance values.

Material and Methods

The first Portuguese TDS described by Vasco (Vasco et al. 2021) was conducted by the Food and Nutrition Department of the National Institute of Health Doctor Ricardo Jorge, IP (INSA) using the harmonised TDS methodology developed during the TDS-Exposure project (Dofkova et al. 2016; Pité et al. 2018).

Food Consumption Data and Food Sampling

The consumption survey characteristics, the definition and preparation of TDS samples and the collection of food items was described in detail by Vasco (Vasco et al. 2021). Briefly, based on the food consumption survey results (Poínhos et al. 2009) a core food list containing 1072 food items, representative of the general Portuguese population diet (male and female from 18 to 93 years old, n = 3529), was classified and mapped by the exposure hierarchy of the FoodEx2 food classification and description system (EFSA 2011) into 20 food groups. For ‘adults and elderly’ (18 to 74 years old) ‘male and female’, from now on named ‘overall population’ (n = 3272), a TDS food list was compiled. This list included 528 food items, covering 95% of the whole diet. These food items were selected based on two criteria applied at food group level: firstly, they had to be the most consumed foods contributing to at least 90% of the general food consumption, and secondly, they were foods known or suspected to be the primary sources of exposure or intake to one or more substances of interest (Vasco et al. 2021). The foods items were aggregated and resulted in 164 different TDS samples. Each TDS sample was classified by FoodEx2 and was composed by twelve subsamples. Potential seasonal and regional variations in chemical occurrence were also considered resulting in 227 TDS samples (2724 subsamples). Foods were collected between April 2014 and March 2016 in the selected shops of the Greater Lisbon region (hyper-or supermarket, discount, local markets for fish, fruits and vegetables, butchers, bakeries, pastries, restaurants and takeaway) selected according to market information. The food items were prepared “as consumed” by the population, respecting the most common food preparation and cooking habits (e.g. washing, peeling, removing inedible parts like bones, scales, shells, seeds, and boiling, grilling, roasting, baking, frying). The edible part of the subsamples were homogenised, pooled into the respective TDS samples that were further aliquoted and stored at – 20 ºC, until analysis (Vasco et al. 2021).

Mercury Determination

The determination of total mercury occurrence in TDS samples was performed by a ISO/IEC 17025 accredited method using a direct mercury analyser DMA-80 (Milestone Inc., Shelton, CT, USA), based on the principle of thermal decomposition, amalgamation and atomic spectrometric detection (254 nm). Slight modifications were made, to the analytical DMA-80 programme concerning the temperatures and times, to optimize for the different food matrices (Martins et al. 2013). Two calibration curves with working ranges between 2.5 and 30 ng and 100 and 500 ng were used. The limits of detection (LOD) and quantification (LOQ) were performed on 10 blanks analysed at different days (LOD: 3 × SD and LOQ: 10 × SD of the sample boat blanks) (EN 13804). LOD ranged between 0.07 and 4.6 μg/kg and LOQ between 0.2 and 13.8 μg/kg. About 0.2–0.3 g (exceptions: popcorn 0.02 g; fruit salad 0.83 g) of each TDS sample was directly weighed into the pre-cleaned sample boat and placed into the auto-sampler. The results were expressed as micrograms per kilogram (μg/kg). The measurement precision expressed by relative standard deviation was below 0.70% for repeatability (RSDr) and 1.16% for intermediate precision (RSDR). The participation in Proficiency Test Schemes launched by Accredited Proficiency Testing Providers (Fapas®, York, UK) and the analysis of the certified reference material (MRC) NIST 1566b Oyster tissue were used to evaluate the laboratory Performance (z-scores between − 2.9 and 1.2). The relative uncertainty of the results was 5%.

Occurrence, Exposure Assessment and Risk Characterization

Occurrence results reported below limit of detection and/or limit of quantification (left-censored data) were managed using the upper bound (UB) and lower bound (LB) substitution methods (EFSA 2010; GEMS/Food-EURO 1995), In the UB approach the non-detected results were set to the LOD value and the results between the limit of detection and quantification were set to the LOQ value. In the LB approach the non-detected results were set to zero and results between the limit of detection and quantification were set to the LOD value. For the TDS samples collected in four seasons, the mean occurrence value was used for exposure estimation after applying the left-censored substitution method. For total mercury occurrence data, minimum (min), maximum (max) mean and median values were calculated for each FoodEx2 group. To estimate methylmercury level from total mercury occurrence values, EFSA’s conservative approach was used assuming 100% methylmercury in fish and 80% in seafood other than fish. Inorganic mercury was estimated as 20% of total mercury in fish as well as, 50% in seafood other than fish. In other foods, mercury was considered to be 100% inorganic (EFSA 2012). A semiprobablistic approach was used to perform methylmercury and inorganic mercury dietary exposure assessment by combining the estimated UB and LB occurrence data with the corresponding consumption recorded in the food consumption survey. Occurrence and consumption data were linked by coding both with the FoodEx2 food classification and description system. The exposure calculation was carried out with the Monte Carlo Risk Assessment (MCRA) software using the observed individual mean (OIM) model (Boon et al. 2018). Exposure was estimated for food as measured (TDS sample) and for food as eaten (TDS subsamples) and was expressed as mean, median (P50) and 95th percentile (P95), representing the average and highly exposed population. Exposure estimates, expressed in µg/kg bw/week, were conducted for the ‘overall population’ and segmented by age group (‘adults and elderly’, 18–74 years old; ‘adults’, 18–64 years old; ‘elderly’, 65–74 years old) for both males and females, and separately for males and females. Exposure to methylmercury was also estimated for the sub-population of 18 to 45 years old childbearing age women (905 individuals). A deterministic approach was used to evaluate exposure of the ‘overall population’ to inorganic mercury through the consumption of tap water. Mean tap water consumption was combined with the average maximum total mercury occurrence value (2018 to 2022) obtained from the water quality legal control of the Greater Lisbon municipalities. For risk characterization, estimated exposure values of methylmercury and inorganic mercury were compared with tolerable weekly intakes (TWI) of 1.3 and 4 µg/kg of body weight (bw), respectively, the most recent health based guidance values (HBGV) set by EFSA (EFSA 2012). The percentages of individuals that exceeded the TWI were also estimated. The contribution of each food item to the total exposure was estimated and the main contributors were highlighted.

Results and Discussion

Mercury

Total mercury occurrence levels below the LOD and LOQ represented 77% of the analysed samples, however all TDS samples composed by fish and/or seafood showed occurrence values above LOQ. Data on total mercury occurrence summarised by food group is presented in Table 1. The ‘fish, seafood, amphibians, reptiles and invertebrates’ food group showed the highest total mercury mean occurrence value (0.136 mg/kg) and included the TDS sample with the highest value (0.574 mg/kg, scabbardfish). Other food groups had mean occurrence values equal or below 0.0195 mg/kg.

Table 1 Total mercury content (minimum, maximum, mean and median) of TDS samples expressed by FoodEx2 food group
Full size table

Methylmercury

Negligible differences were found for methylmercury estimates calculated using the UB and LB occurrence values, therefore only the UB estimates are reported.

The estimated dietary exposure of Portuguese ‘overall population’ to methylmercury with the detailed contribution of each TDS sample is presented in Table 2. The corresponding data for ‘male’ and ‘female’ population groups are presented in Tables 3 and 4, respectively. The ‘overall population’ revealed a mean weekly exposure to methylmercury of 1.25 µg/kg bw/week, being lower for ‘adults’ (1.20 µg/kg bw/week) and higher for ‘elderly’ (1.48 µg/kg bw/week) (Table 2). ‘Elderly’ males showed the highest mean methylmercury exposure (1.77 µg/kg bw/week) whereas male ‘adults’ showed the lowest value (1.14 µg/kg bw/week) of all population groups studied (Table 3). Females did not reveal differences in exposure according to age (1.26 µg/kg bw/week ‘adults’ and 1.27 µg/kg bw/week ‘elderly’) (Table 4). The median weekly exposure (P50) ranged from 0 to 0.25 µg/kg bw/week for all sex and age groups, with the highest value observed in male elderly individuals (Table 2, 3, and 4). At the 95th percentile, weekly methylmercury exposure ranged from 4.97 to 5.68 µg/kg bw/week for all sex and age groups except for male elderly individuals, who showed a weekly exposure of 7.29 µg/kg bw/week (Table 2, 3, and 4).

Table 2 Estimated dietary exposure of Portuguese overall population (‘male and female’) to methylmercury
Full size table
Table 3 Estimated dietary exposure of Portuguese male population to methylmercury
Full size table
Table 4 Estimated dietary exposure of Portuguese female population to methylmercury
Full size table

The mean methylmercury exposure for the ‘overall population’ reached 96% of the TWI of 1.3 µg/kg bw/week. However, for elderly males, this exposure surpassed the TWI at 136%. At the 95th percentile, estimated exposures varied from 382% for elderly females to 561% for elderly males compared to the TWI. The proportion of individuals exceeding the TWI ranged from 25.8% for adult males to 36.4% for elderly males, with an overall population average of 27.6%. The food groups ‘fish, seafood, amphibians, reptiles and invertebrates’ and ‘composite dishes’ were the only contributors to methylmercury exposure of all sex and age groups. The main foods as eaten contributing to methylmercury exposure are presented in Table 5. Cod (dried) and hake were the two main contributors (food as eaten) ranging from 16.5 to 26.0% (cod) and from 8.7 to 18.1% (hake).

Table 5 Main foods contributing to methylmercury exposure
Full size table

When consumers only (1659 individuals) were considered, the methylmercury exposure of the ‘overall population’ was estimated as 2.46 μg/kg bw/week (mean) and 1.57 μg/kg bw/week (P50) corresponding to 189 and 121% of the TWI, respectively.

Methylmercury exposure of the Portuguese population was assessed by Carvalho (Carvalho et al. 2022) using a different food consumption survey (IAN-AF 2015–2016) and official monitoring occurrence data. The authors reported methylmercury mean exposure of 0.61 µg/kg bw/week for adults and 0.59 µg/kg bw/week for elderly, approximately half of the values obtained in the current study. The percentage of individuals exceeding the TWI was 12.9% for adults and 11.6% for elderly age groups vs 26.6% and 32.8% in the present study. Although, fish consumption by the Portuguese population changed from 74.1 g/day (Dofkova et al. 2016) to 48.1 g/day (Inquérito Alimentar Nacional e de Atividade Física [IAN-AF], 2016), methodological differences may also have affected the exposure estimates. TDS samples were composed of 12 subsamples (food items), prepared as consumed according to food cooking habits, and pooled to provide a robust mean occurrence value; whereas monitoring data were individual values (raw fish), as a consequence the probabilistic approaches used to input concentration values to individual eating occasions varied between the two studies. However, both studies showed a concerning percentage of population exceeding the methylmercury TWI.

TDS methodology was used by different authors to estimate methylmercury exposure of adult populations. Arnich (Arnich et al. 2012) reported a mean exposure to methylmercury of 0.12 μg/kg bw/week and a P95 of 0.43 μg/kg bw/week for the adults French population (18 to 79 years old) corresponding to 9.2% and 33% of the TWI (1.3 μg/kg bw/week), respectively. The 1.6 μg/kg bw/week health-based guidance value set by JECFA (JECFA, 2004) was exceeded by 0.8% of individuals (Arnich et al. 2012). Mean exposure of Chile’s adult population (18 to 65 years old) to mercury from fish and shellfish food group (corresponding to methylmercury) was estimated by Munoz as 0.16 μg/kg bw/week in Santiago city (Muñoz et al. 2005) and 0.048 μg/kg bw/week in Valdivia city (Muñoz et al. 2017), corresponding respectively to 12% and 3.7% of the 1.3 μg/kg bw/week TWI (European Food Safety Authority [EFSA], 2012). Nasreddine (Nasreddine et al. 2006) estimated a mean methylmercury exposure of 0.28 µg/kg bw/week for the urban adult Lebanese population (25 to 54 years old) corresponding to 21.5% of the TWI of 1.3 µg/kg bw/week (EFSA 2012). Methylmercury mean exposure by the Czechia’s adult population was reported as 4.3% of 1.3 μg/kg bw/week TWI corresponding to 0.056 μg/kg bw/week (Sokolov et al. 2022). Exposure of Germany’s adult population (14 to 79 years old) to methylmercury was estimated by Sarvan as 0.072 (mean) and 0.49 (P95) μg/kg bw/week (Sarvan et al. 2021), corresponding respectively 5.5 and 38% of the 1.3 μg/kg bw/week TWI (EFSA 2012). When consumers only were considered, methylmercury exposure was estimated by Kolbaum as 0.28 (mean) and 0.85 (P95) μg/kg bw/week (Kolbaum et al. 2019) and by Sarvan as 0.339 (mean) the later corresponding to 26% of the 1.3 μg/kg bw/week TWI (Sarvan et al. 2021). Marín (Marin et al. 2017) reported a methylmercury average exposure of 0.60 μg/kg bw/week (UB) estimated based on consumption of fish and seafood by the adult population (16 to 95 years old) of Valencia region, Spain. Methylmercury mean exposure of 1.25 μg/kg bw/week estimated in this study was well above the values reported by the previously referred studies that ranged from 0.048 to 0.60 μg/kg bw/week. The same was observed when consumers only mean exposure (2.64 μg/kg bw/week) was compared with German population values.

The considerably higher methylmercury exposure estimates found in the present study as compared to those reported in other countries might be explained by the significantly higher fish and seafood consumption by the Portuguese population and not by the methylmercury occurrence levels of the fish consumed.

Adverse neurodevelopmental outcomes in infants and young children during pregnancy are associated with methylmercury, therefore, exposure to methylmercury was estimated for 905 Portuguese women of childbearing age (18 to 45 years old). Table 6 shows the estimated dietary exposure to methylmercury of Portuguese childbearing age women, detailing the contribution of each TDS sample to exposure. The estimated mean weekly exposure to methylmercury was 1.13 µg/kg bw/week corresponding to 86.9% of the TWI (1.3 µg/kg bw/week). Whereas, the estimated median weekly exposure (P50) was 0 µg/kg bw/week. At the 95th percentile, methylmercury weekly exposure was 5.56 µg/kg bw/week, approximately four fold the TWI. The percentage of childbearing age women exceeding the TWI was 25.0%. When only women of childbearing age who consumed fish and seafood either classified as ‘fish, seafood, amphibians, reptiles and invertebrates’ or ‘composite dishes’ (409 individuals) were considered (consumers only) the estimated mean weekly exposure was 2.50 µg/kg bw/week (192% of TWI), two times higher than that of all women of childbearing age (Table 6). Hake (20.2%), dried cod (14.5%) and canned tuna (12.8%) were the three main contributors to methylmercury exposure.

Table 6 Estimated dietary exposure of Portuguese of childbearing age women (18 to 45 years old) to methylmercury
Full size table

Carvalho (Carvalho et al. 2022) reported a methylmercury mean exposure of 0.58 μg/kg bw/week in Portuguese women of childbearing age (15 to 49 years old), corresponding to 44.6% of the TWI (1.3 µg/kg bw/week). The percentage of childbearing age women exceeding the TWI was 12.7%. The values reported by Carvalho (Carvalho et al. 2022), using a different methodology, were approximately half of those found in the present study. A methylmercury mean exposure of 0.98 μg/kg bw/week was reported by Ortega-García (Ortega-García et al. 2009) in childbearing age women of Murcia, Spain corresponding to 75.4% of the TWI (1.3 µg/kg bw/week). P95 exposure was estimated at 2.44 μg/kg bw/week (188% of the TWI) (Ortega-García et al. 2009). The mean exposure level estimated in the present study was slightly higher than the one reported by Ortega-García (Ortega-García et al. 2009) whereas P95 exposure level was approximately twice. Arnich (Arnich et al. 2012) reported a methylmercury mean exposure of 0.13 μg/kg bw/week in women of childbearing age (18 to 45 years old) in France, corresponding to 10% of the TWI (1.3 µg/kg bw/week). P95 exposure was estimated at 0.47 μg/kg bw/week (36% of the TWI) (Arnich et al. 2012). The mean and P95 exposure levels reported by Arnich (Arnich et al. 2012) were one order of magnitude lower than those found in the present study.

Consumers-only childbearing age women (15 to 49 years old) of Germany showed a mean exposure to methylmercury of 0.21 μg/kg bw/week corresponding to 16% of the TWI (1.3 µg/kg bw/week) and a P95 exposure of 0.72 μg/kg bw/week, 55% of the TWI (Kolbaum et al. 2019). The mean exposure level estimated in the present study was approximately one order of magnitude higher than the one reported by Kolbaum (Kolbaum et al. 2019).

A biomonitoring study in childbearing women (25 to 44 years old) in Portugal reported a mean blood mercury of 5.9 µg/L (Santiago et al., n.d.) corresponding to a weekly exposure to methylmercury of 0.81 µg/kg bw/week calculated according to the average ratios between intake and blood levels indicated by WHO (WHO 2008). Other biomonitoring studies focusing on specific Portuguese regions and pregnant women (Alves et al. 2017; Caetano et al. 2019; Namorado et al. 2023; Nunes et al. 2014) or childbearing age women (Castaño et al. 2015) presented results of the same order of magnitude.

The referred biomonitoring studies were in line with the methylmercury exposure estimates obtained by the present study and by Carvalho (Carvalho et al. 2022) and support the need of special consideration regarding fish consumption by the Portuguese childbearing age women.

The European Commission (EC) has advised the establishment of specific national guidelines regarding fish and seafood consumption to maximize their health benefits while minimizing the risk of mercury toxicity, specifying which species and how often they should be consumed (European Commission [EC], 2022). Subsequently, recent recommendations for fish and seafood consumption in Portugal have been issued, highlighting pregnant and breastfeeding women, as well as children up to 10 years old, as vulnerable groups (General Directorate for Food and Veterinary [DGAV] et al., 2023). To adhere to these recommendations, the vulnerable population groups are advised to avoid species with high mercury levels such as blue shark, fresh tuna, ling, scabbardfish, small-spotted catshark, smooth-hound, and swordfish, and to consume species with low to medium mercury content three to 4 times a week for optimal health benefits. For the general population a consumption up to 7 times a week regardless of the species is advocated. The recommendations were developed by a working group promoted by the General Directorate for Food and Veterinary which included the National Institute of Health Doctor Ricardo Jorge, the Economic and Food Safety Authority, the Faculty of Nutrition and Food Sciences of the University of Porto, the Portuguese Institute for Sea and Atmosphere and the Public Health Institute of the University of Porto, and were based on a risk–benefit assessment associated with fish consumption by the Portuguese population (Carvalho et al. 2022).

Inorganic Mercury

The estimated dietary exposure of Portuguese overall population (‘male and female’) to inorganic mercury is presented in Table 7. Corresponding data for ‘male’ and ‘female’ population groups are presented in Tables 8 and 9, respectively.

Table 7 Estimated dietary exposure of Portuguese overall population (‘male and female’) to inorganic mercury
Full size table
Table 8 Estimated dietary exposure of Portuguese male population to inorganic mercury
Full size table
Table 9 Estimated dietary exposure of Portuguese female population to inorganic mercury
Full size table

Mean exposure to inorganic mercury for the ‘overall population’ was 0.30 µg/kg bw/week (0.30 for males and 0.33 for females) using the LB approach, and 0.37 µg/kg bw/week (0.36 for males and 0.40 for females) using the UB approach. Among other sex and age group combinations, mean exposure ranged from 0.29 µg/kg bw/week (for male adults and female elderly) to 0.40 µg/kg bw/week for elderly males at LB, and from 0.35 µg/kg bw/week (for male adults and female elderly) to 0.46 µg/kg bw/week for elderly males at UB. At the 95th percentile, inorganic mercury weekly exposure of the ‘overall population’ was 1.26 (LB) and 1.27 (UB) µg/kg bw/week. Across all population groups, the mean estimated exposure to inorganic mercury ranged from 7.1 to 10.0% (LB) and from 8.7 to 11.5% (UB) of the TWI of 4 µg/kg bw/week. At the 95th percentile, this exposure ranged from 25.8 to 37.2% (LB) and from 27.3 to 38.7% (UB). The TWI was exceeded by 3.4% (LB) and 3.5% (UB) individuals of the ‘overall population’. The food groups ‘fish, seafood, amphibians, reptiles and invertebrates’ and ‘composite dishes’ contributed 78 to 88% for inorganic mercury exposure of all sex and age groups. Detailed data on the contribution of each TDS sample to inorganic mercury exposure are presented as supplementary information (Supplementary Table S1, S2, S3). The main foods as eaten contributing to inorganic mercury exposure are presented in Table 10. Cod (dried) and hake were the two main contributors (food as eaten) ranging from 11.2 to 20.0% and from 6.7 to 12.3%, respectively.

Table 10 Main foods contributing to inorganic mercury exposure
Full size table

Arnich (Arnich et al. 2012) reported for the French adults a dietary (except seafood) inorganic mercury mean exposure of 0.042 μg/kg bw/week (LB) and 1.26 μg/kg bw/week (UB) and 95th percentiles of 0.182 (LB) and 2.03 (UB) μg/kg bw/week. The TWI was not exceeded by any individual. The mean exposure level estimated in the present study was approximately ten times higher at LB and three times lower at UB than the one reported by Arnich (Arnich et al. 2012), whereas at P95 it was seven times higher at LB and half at UB. The disparity in results may be due to methodological differences between the two studies, such as: (i) the inclusion of seafood TDS samples in the present study using EFSA’s conservative approach (EFSA 2012); (ii) percentage of left censored data; and (iii) analytical limits.

The population of Germany (14 to 80 years old) showed a mean exposure to inorganic mercury of 0.02 (LB) and 0.07 (UB) μg/kg bw/week corresponding to 1 and 2% of the TWI (4 µg/kg bw/week), respectively, and a P95 exposure of 0.07 (LB) and 0.14 (UB) μg/kg bw/week, corresponding to 2 and 4% of the TWI (4 µg/kg bw/week), respectively (Kolbaum et al. 2019). The mean exposure level estimated in the present study was approximately five times higher than the one reported by Kolbaum (Kolbaum et al. 2019) and seven times higher at P95. However, Kolbaum study did not include fish and seafood samples in the estimation of inorganic mercury exposure and considered a different age group.

As previously mentioned, tap water was excluded from the TDS food list and, therefore, a deterministic approach was used to evaluate its contribution to inorganic mercury exposure of the Portuguese population. The exposure of the ‘overall population’ to inorganic mercury through the consumption of tap water was estimated at 0.006 µg/kg bw/week (0.15% of TWI), corresponding roughly to 1.6% of the total inorganic mercury exposure.

Strengths and Limitations

TDSs are very useful to assess chronic exposure to chemicals because they allow covering the whole diet with limited costs by using food sample pooling. Furthermore, foods are analysed after being prepared according to common consumption habits, thus, the chemicals in the foods analysed are closely representative of what is actually ingested by the consumer. The use of a TDS harmonised methodology provides data that facilitates assessing time trends and country comparison of the dietary exposure of populations to chemical substances. Nevertheless, uncertainties are present at the different steps of exposure assessment and may lead to over or underestimation. The consumption data used in this study was gathered in 2009, and it is possible that food habits have evolved since then. However, the consumption of fish and seafood per year per capita, which is the primary contributor to mercury exposure, remained relatively stable from 2009 (57.17 kg) to the years when the food collection took place, from 2014 (53.22 kg) to 2016 (57.33 kg), according to Our World in Data (Our World in Data 2021). Another uncertainty arises from the fact that only one day of consumption was recorded, potentially resulting in an overestimation of the higher percentiles of chronic exposure. Regarding the choice of foods to analyse, using the food group approach when applying the selection criteria to include foods in the TDS food list avoided the loss of some less consumed foods that could contribute considerably to exposure. If foods were ranked according to the amount consumed across the whole diet, some less consumed foods would never be selected even if they were important contributors to exposure. Occurrence data used to perform the exposure assessment estimation was based on total mercury analytical determination and a high left-censored data rate was observed for food groups other than those containing fish and seafood. To further refine the exposure assessment to mercury, analytical improvements would be desirable, such as lowering LODs for total mercury determination and/or preferably conducting speciation analysis.

Conclusions

The mean estimated baseline exposure of the Portuguese overall population (adults and elderly) to methylmercury and inorganic mercury did not exceed the respective HBGV. However, the percentage of individuals exceeding the HBGV was 27.6% for methylmercury and 3.5% for inorganic mercury (UB approach). According to these results the possibility of adverse health effects cannot be excluded. The methylmercury mean and P50 exposure results of the overall population evidence a distribution markedly skewed to the right with only half of the individuals being exposed. Consequently, when focusing on consumers only, the mean and P50 values were observed at 189% and 121% of the Tolerable Weekly Intake (TWI), respectively. The findings are particularly of concern for women of childbearing age, as their exposure to methylmercury poses risks of adverse neurodevelopmental outcomes to the child during pregnancy. The previous findings pose a double concern on the one hand the very high exposure to methylmercury of fish and seafood consumers, as discussed before, and on the other hand the potential low intake of omega 3 fatty acids by the non-consumers, with particular relevance to women of childbearing age. Omega 3 fatty acids has been associated with a reduction of coronary heart disease mortality in adults and with neurodevelopmental improvement in infants and young children derived from mother’s fish consumption during pregnancy (FAO 2011) (EFSA 2015).

Future assessment of methylmercury exposure by the Portuguese population and particularly of the vulnerable groups (pregnant and lactating women and children up to ten years old) identified in the fish and seafood consumption recommendations issued recently in Portugal will be needed to evaluate the impact of this risk management measure.

Despite the TDS methodology being recommended as the most cost-effective means to evaluate chronic exposure to food contaminants and the international efforts to harmonize procedures, discrepancies persist in published data, making comparisons challenging. Variations arise in food list preparation, food classification, food sampling, and reporting exposure results. Harmonisation initiatives such as those by FAO/WHO/EFSA (EFSA et al. 2011) and TDS-Exposure (Pité et al. 2018) focused mainly on TDS food list preparation and sampling, and the FoodEx2 food classification system was proved adequate by the TDS Exposure project (Dofkova et al. 2016) to harmonise food and sample lists across countries. However, future agreement on minimum requirements to report exposure assessment results would be valuable to allow for improved international comparison, namely food grouping (food classification), assumptions on occurrence (elements speciation), population age groups, sex, descriptive statistics and overall population vs consumers-only.