Following a request from the European Commission, the EFSA Panel on Contaminants in the Food Chain (CONTAM Panel) was asked to deliver a scientific opinion on the risks to human health related to the presence of inorganic mercury and methylmercury in food. The Panel was asked to consider new developments regarding the toxicity of inorganic mercury and methylmercury since the last opinion of the European Food Safety Authority (EFSA) of 24 February 2004 and to evaluate whether the provisional tolerable weekly intakes (PTWIs) established by the Joint FAO/WHO Expert Committee on Food Additives (JECFA) of 1.6 µg/kg body weight (b.w.) for methylmercury and of 4 µg/kg b.w. for inorganic mercury were considered appropriate. The CONTAM Panel was also asked to assess human dietary exposure, taking into account specific sensitive groups and to consider the non-dietary sources of exposure to mercury species.
Mercury is a metal that is released into the environment from both natural and anthropogenic sources. Once released, mercury undergoes a series of complex transformations and cycles between atmosphere, ocean and land. The three chemical forms of mercury are (i) elemental or metallic mercury (Hg0), (ii) inorganic mercury (mercurous (Hg22+) and mercuric (Hg2+) cations) and (iii) organic mercury. Methylmercury is by far the most common form of organic mercury in the food chain.
This opinion focuses only on the risks related to dietary inorganic mercury and methylmercury exposure and does not assess the nutritional benefits linked to certain foods (e.g. fish and other seafood).
A call for annual collection of chemical contaminant occurrence data in food and feed, including mercury, was issued by EFSA in December 2010. In response, EFSA received 59 820 results on mercury in food from 20 European countries, mainly covering the period from 2004 to 2011. A total number of 59 650 results were described with sufficient detail to be used in the statistical analysis of the respective food groups; 98.2 % of the samples were for total mercury, 1.8 % for methylmercury and three samples for inorganic mercury.
All the 20 food groups available at the first level of FoodEx were covered in the current data collection. The food groups ‘Fish and other seafood’ and ‘Meat and meat products’ dominated the food product coverage with 36.8 % and 17.6 % respectively. These were followed by ‘Grain and grain-based products’ at 7.8 % and ‘Vegetables and vegetable products (including fungi)’ at 7.3 %. More than 60 % of the data were below the limit of detection (LOD) or the limit of quantification (LOQ) (left-censored (LC)) in 11 of the food groups. However, 12 % of the results for ‘Fish and other seafood’, which had the highest values of total mercury in comparison to all other food categories, were LC. The mercury content varied widely among different fish species, and was highest in predatory fish.
Because of the lack of specific information on methylmercury and inorganic mercury data in the database, the exposure assessment (except for human milk) was based on the data submitted for total mercury. The analysed total mercury was converted to methylmercury and inorganic mercury by applying conversion factors based on the methylmercury/total mercury proportion derived from literature data, using a conservative approach. For fish meat, fish products, fish offal and unspecified fish and seafood a conversion factor of 1.0 was used for methylmercury and 0.2 for inorganic mercury. For crustaceans, molluscs and amphibians the conversion factor was 0.8 for methylmercury and 0.5 for inorganic mercury. For all other food categories apart from ‘Fish and other seafood’, total mercury was regarded as inorganic mercury. Because this approach was chosen, total mercury dietary exposure cannot be derived by adding inorganic and methylmercury dietary exposure together. In order to estimate dietary exposure, the consumption data of each individual within the surveys were multiplied by the mean occurrence data for the relevant food categories, resulting in a distribution of exposure, from which the mean and 95th percentile were identified for each survey and age class. For human milk, the mean concentrations of methylmercury and inorganic mercury in a limited number of European studies were used for exposure assessment.
The dietary exposure to methylmercury was based only on the food group ‘Fish and other seafood’ and since there was little difference between the lower bound (LB) and upper bound (UB) exposure estimates, the middle bound (MB) exposures were used. The mean MB methylmercury dietary exposure varied from the lowest minimum of 0.06 μg/kg b.w. per week seen in elderly and very elderly to the highest maximum of 1.57 μg/kg b.w. per week in toddlers. The 95th percentile MB dietary exposure ranged from the lowest minimum of 0.14 μg/kg b.w. per week in very elderly to the highest maximum of 5.05 μg/kg b.w. per week in adolescents. Based on mean concentrations of methylmercury in human milk, the dietary exposure to methylmercury for infants with an average human milk consumption ranged from 0.09 to 0.62 µg/kg b.w. per week and for infants with high milk consumption the dietary exposure ranged from 0.14 to 0.94 µg/kg b.w. per week.
Fish meat was the dominating contributor to methylmercury dietary exposure for all age classes, followed by fish products. In particular tuna, swordfish, cod, whiting and pike were major contributors to methylmercury dietary exposure in the adult age groups, while the same species, with the addition of hake, were the most important contributors in the child age groups. Dietary exposure in women of child-bearing age was especially considered and found not to be different from adults in general. The dietary exposure estimations in high and frequent consumers of fish meat (95th percentile, consumers only) was in general approximately two-fold higher in comparison to the total population and varied from a minimum MB of 0.54 μg/kg b.w. per week in elderly to a maximum MB of 7.48 μg/kg b.w. per week in other children.
The estimation of dietary exposure to inorganic mercury was based on minimum LB and maximum UB data due to the high proportion of LC data and the large difference between LB and UB concentrations. The mean dietary exposure to inorganic mercury varied from the lowest minimum LB of 0.13 μg/kg b.w. per week in elderly to the highest maximum UB of 2.16 μg/kg b.w. per week in toddlers. The 95th percentile dietary exposure was estimated to be from the lowest minimum LB of 0.25 μg/kg b.w. per week in elderly and very elderly to the highest maximum UB of 4.06 μg/kg b.w. per week in toddlers. Based on mean concentrations of inorganic mercury in human milk, the dietary exposure for infants with an average milk consumption ranges from 0.17 to 1.29 µg/kg b.w. per week and from 0.25 to 1.94 µg/kg b.w. per week for infants with a high milk consumption.
At FoodEx Level 1, ‘Fish and other seafood’, ‘Non-alcoholic beverages’ and ‘Composite food’ were the most important contributors to inorganic mercury dietary exposure in the European population. Dietary exposure to inorganic mercury was driven by high concentrations in the case of fish and other seafood and composite food (where a high proportion of the data were LC), but was more likely driven by high consumption in the case of non-alcoholic beverages.
Non-dietary exposure to methylmercury is likely to be of minor importance for the general population in Europe, but exposure to elemental mercury via the outgassing of dental amalgam is believed to strongly contribute to the internal inorganic mercury exposure.
After oral intake, methylmercury is much more extensively and rapidly absorbed than mercuric and mercurous mercury. In human blood mercuric mercury is divided between plasma and erythrocytes, with more being present in plasma, whereas methylmercury is accumulated to a large extent (> 90 %) in the erythrocytes. In contrast to mercuric mercury, methylmercury is able to enter the hair follicle, and to cross the placenta as well as the blood-brain and blood-cerebrospinal fluid barriers, allowing accumulation in hair, the fetus and the brain. Mercuric mercury in the brain is generally the result of either in situ demethylation of organic mercury species or oxidation of elemental mercury. Excretion of absorbed mercuric mercury occurs mainly via urine, whereas the main pathway of excretion of absorbed methylmercury is via faeces in the form of mercuric mercury. Urinary total mercury might be a suitable biomarker of inorganic (and elemental) mercury, but not for methylmercury exposure. Total mercury in hair and blood are routinely used as biomarkers to assess long term methylmercury exposure. A frequently cited total mercury blood to hair ratio is 1:250, however large variations exist, especially in people with infrequent fish consumption.
A recent developmental study of methylmercury in mice, applying only one low dose, indicated effects on body weight gain, locomotor function and auditory function. A large study in rats showed developmental immunotoxic effects at low doses, and the lower 95 % confidence limit for a benchmark response of 5 % (BMDL05) of 0.01 mg/kg b.w. per day, expressed as methylmercuric chloride (equivalent to 0.008 mg/kg b.w. per day, expressed as mercury) for the specific antibody response in rats was the lowest BMDL. While bearing this in mind, the Panel concluded that experimental animal studies on methylmercury did not provide a better primary basis than the human data for a health-based guidance value.
New data from the Faroe Islands Cohort 1 at children’s age 14 years indicated that the association between prenatal exposure and neurological auditory function was still present at 14 years, but with a smaller impact than at seven years. Reassessment of the data from the Faroese Cohort 1 participants at age seven years indicated that beneficial effects of fish consumption together with imprecision in the measurements of fish consumption and determination of mercury in hair might underestimate the effects of methylmercury.
Reassessments of the 4.5 years results and the 10.5 and 17 years follow up studies from the Main Cohort in the Seychelles Child Developmental Study have not revealed any consistent association between prenatal mercury exposure and neurodevelopmental endpoints. Results from the smaller Nutrition Cohort in the Seychelles Child Developmental Study indicated an association between prenatal mercury exposure and decreased scores on neurodevelopmental indices at 9 and 30 months after adjustment for prenatal blood maternal n-3 long-chain polyunsaturated fatty acids (n-3 LCPUFAs). An apparent no-observed-effect level (NOEL) at a mercury level of approximately 11 mg/kg maternal hair was observed. No statistically significant associations between prenatal mercury exposure and developmental endpoints were found at the five years follow up of the study. However, a positive association between maternal prenatal n-3 LCPUFAs, in particular docosahexaenoic acid, and preschool language scores was reported from the five years follow up.
The reported associations between methylmercury exposure and cardiovascular disease were addressed by JECFA in their update in 2006 (FAO/WHO, 2007), and additional studies have become available. The importance of taking the beneficial effects of fish consumption into account when studying cardiovascular outcomes of methylmercury has become evident. Although the observations related to myocardial infarction, heart rate variability and possibly blood pressure are of potential importance, they are still not conclusive. Consequently, after carefully considering other endpoints than neurodevelopmental outcomes, and in particular cardiovascular disease, the CONTAM Panel concludes that associations between methylmercury exposure and neurodevelopmental outcomes after prenatal exposure still form the best basis for derivation of a health-based guidance value for methylmercury.
The mean of the apparent NOEL from the Seychelles nutrition cohort at 9 and 30 months (11 mg/kg maternal hair) and the BMDL05 from the Faroese cohort 1 at age seven years (12 mg/kg in maternal hair), resulting in 11.5 mg/kg maternal hair, was used as the basis for derivation of a health-based guidance value. By application of a maternal hair to maternal blood ratio of 250, the maternal hair mercury concentration with no appreciable adverse effect was converted into a maternal blood mercury concentration of 46 μg/L. Using a one-compartment toxicokinetic model, the value of 46 µg/L in maternal blood was converted to a daily dietary mercury intake of 1.2 µg/kg b.w. A data-derived uncertainty factor of 2 was applied to account for variation in the hair to blood ratio. In addition, a standard factor of 3.2 was applied to account for interindividual variation in toxicokinetics, resulting in a total uncertainty factor of 6.4. A tolerable weekly intake (TWI) for methylmercury of 1.3 µg/kg b.w. expressed as mercury, was established. The Panel noted that this TWI provides a margin of about 40 compared to the BMDL05 for the reduction in antibody response in rats.
The mean dietary exposure across age groups does not exceed the TWI for methylmercury, with the exception of toddlers and other children in some surveys. The medians of 95th percentile dietary exposures across surveys are close to or above the TWI for all age groups. High consumers of fish meat may exceed the TWI by up to approximately six-fold. Unborn children constitute the most vulnerable group for developmental effects of methylmercury exposure, and pregnant women can be present in the group of high and frequent fish consumers. Biomonitoring data on blood and hair concentrations indicate that in the general European population, methylmercury exposure is generally below the TWI. However, higher concentrations in blood and hair are also observed, confirming higher dietary exposure in some population groups. Exposure to methylmercury above the TWI is of concern, but if measures to reduce methylmercury exposure are considered, the potential beneficial effects of fish consumption should also be taken into account.
The critical target for toxicity of inorganic mercury is the kidney. Other targets include the liver, nervous system, immune system, reproductive and developmental systems. Having considered the experimental animal data on inorganic mercury, including some recent studies not reviewed by JECFA in its evaluation of 2010, the Panel agrees with the rationale of JECFA in setting a health-based guidance value using kidney weight changes in male rats as the pivotal effect. Based on the BMDL10 of 0.06 mg/kg b.w. per day, expressed as mercury and an uncertainty factor of 100 to account for inter and intra species differences, with conversion to a weekly basis and rounding to one significant figure, the Panel established a TWI for inorganic mercury of 4 µg/kg b.w., expressed as mercury.
The estimated exposure to inorganic mercury in Europe from the diet alone does not exceed the TWI. Inhaled elemental mercury vapour from dental amalgam, which after absorption is converted to inorganic mercury, is an additional source that is likely to increase the internal inorganic mercury exposure; thus the TWI might be exceeded.
The CONTAM Panel recommends to develop certified reference materials and proficiency testing schemes for inorganic mercury in foodstuffs other than fish and seafood. Further effort should be made to increase the number of methylmercury and inorganic mercury data in all food groups that contribute significantly to overall exposure. In order to decrease the uncertainty in the point of departure derived from the epidemiological studies, more reliable definition of the dose response taking confounding factors into account is needed. Future studies should elucidate the relevance of additional endpoints, such as immunological and cardiovascular endpoints.