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 public health related to the presence of perchlorate in food, in particular fruits and vegetables. Perchlorate (ClO4-) is a chemical contaminant which is released in the environment from both natural and anthropogenic sources. Biomonitoring studies show the presence of background levels of perchlorate in the general population, suggesting that it is likely a ubiquitous environmental contaminant. Different sources of contamination were identified. These include the use of fertilisers of natural origin in which perchlorate may be present, such as Chilean nitrate; industrial emissions of perchlorate into the environment, in particular resulting from the use of ammonium perchlorate in solid propellants for rocket and missiles; the natural formation of perchlorate in the atmosphere and surface water, and the formation of perchlorate during the degradation of chlorine-based products such as sodium or calcium hypochlorite. While industrial emissions are not expected to represent a main and widespread source of contamination in Europe, the use of natural fertilisers such as Chilean nitrate may lead to substantial concentrations in fruit and vegetables, due to the efficient uptake via the plant roots. Similarly, plant irrigation with perchlorate-contaminated groundwater can contribute to accumulation of perchlorate in fruit and vegetables. Water disinfection with chlorine-based biocidal products, potentially degrading to perchlorate, could be another notable source of contamination for drinking water and food. Additional sources of contamination, such as the use of chlorine-based products in biocidal applications other than water disinfection and plant protection applications and the natural formation of perchlorate in the atmosphere and in surface water, could marginally contribute to the presence of perchlorate in food and drinking water.
Perchlorate has been reported to occur in a wide range of foods, including vegetables, fruit, milk and dairy products, juice, beer, wine and bottled water. In a recent assessment from 2010, the Joint FAO/WHO Committee on Food Additives (JECFA) reviewed the available occurrence data and observed the highest mean concentrations in vegetables (range of means 4.8–110 µg/kg), fruits (range of means 0.5–28 µg/kg), vegetable and fruit juice (26 µg/kg) and infant formulae (10 µg/kg).
EFSA received the analytical results of 11 675 samples submitted by eight Member States, of which about 20 % were considered to be suspect samples. (Suspect sampling is defined as a selection of an individual product or establishment in order to confirm or reject a suspicion of non-conformity. It is not a random sampling, therefore there is no sample extracted from the population.) The majority of the samples belonged to the food groups ‘Vegetables and vegetable products’ and ‘Fruits and fruit products’. Excluding the suspect samples, the highest mean perchlorate concentrations were observed in turnips (350 µg/kg, upper bound (UB)) and in lettuce, excluding Iceberg-type lettuce (120 µg/kg, UB). For all other ‘Vegetables and vegetable products’ the mean concentration was below 70 µg/kg. For ‘Fruits and fruit products’ the mean concentration ranged from < limit of detection to 12 µg/kg.
In addition to the data submitted to EFSA, occurrence data from the literature on ‘Infant formulae, powder’, ‘Infant formulae, liquid’, ‘Milk and dairy products’, ‘Alcoholic beverages’ and ‘Fruit juices’ and breast milk were used in the exposure assessment.
The CONTAM Panel performed the exposure assessment of perchlorate using a chronic and a ‘short-term’ exposure scenario. The latter scenario was developed to take into account the possibility of being exposed to relatively high levels of perchlorate for a short period, e.g. two to three weeks, considering that higher levels of thyroid iodine uptake inhibition for short periods could induce adverse effects in vulnerable groups of the population, such as breast-fed infants and young children with low iodine intake. Higher exposure over such periods is plausible for people living in areas where local produce contains higher levels of perchlorate.
The CONTAM Panel concluded that adverse effects following a single-day exposure to perchlorate at levels relevant for dietary exposure are not expected in any group of the population and therefore no acute exposure estimation was carried out.
Only the scenario excluding the suspect samples was considered appropriate to characterise the risk from chronic perchlorate exposure.
The mean chronic dietary exposure ranged from 0.04 to 0.20 µg/kg body weight (b.w.) per day (minimum lower bound (LB)-maximum UB) in adolescents and adult age classes. For ‘other children’ and ‘toddlers’, the mean chronic dietary exposure ranged from 0.07 to 0.37 µg/kg b.w. per day and 0.18 to 0.50 µg/kg b.w. per day (minimum LB-maximum UB), respectively. For ‘infants’, only two consumption surveys were available, resulting in mean chronic dietary exposure of 0.13–0.54 µg/kg b.w. per day (minimum LB-maximum UB). The 95th percentile chronic dietary exposure ranged from 0.10 to 0.51 µg/kg b.w. per day (minimum LB-maximum UB) in ‘adolescents’ and adult age classes. For ‘other children’ and ‘toddlers’ the 95th percentile chronic dietary exposure ranged from 0.19 to 0.72 µg/kg b.w. per day and 0.34 to 0.97 µg/kg b.w. per day (minimum LB-maximum UB), respectively. For infants, the 95th percentile chronic dietary exposure could be calculated for only one consumption survey, resulting in exposure of 0.32–0.61 µg/kg b.w. per day (LB-UB).
‘Vegetables and vegetable products’ and ‘Milk and dairy products’ were identified in the chronic exposure assessment as the most important contributors to perchlorate exposure in all age groups.
Based on mean concentrations of perchlorate in breast milk from the USA, the dietary exposure of breast-fed infants with a mean milk consumption ranged from 0.76 to 4.3 μg/kg b.w. per day, and for infants with a high milk consumption ranged from 1.1 to 6.5 μg/kg b.w. per day. The relevance of these data for the European Union is unknown.
For the short-term exposure the CONTAM Panel considered scenarios excluding and including the suspect samples.
For the scenario excluding suspect samples, the estimations of mean ‘short-term’ dietary exposure to high percentile levels of perchlorate ranged from 0.38 to 1.9 μg/kg b.w. per day in adolescents and adult age classes, and from 1.5 to 2.7 μg/kg b.w. per day in toddlers across dietary surveys. For infants, only two consumption surveys were available resulting in an exposure of 1.2 and 1.5 µg/kg b.w. per day. The highest mean ‘short-term’ dietary exposure level of 3.0 μg/kg b.w. per day was estimated in the age class ‘other children’. The 95th percentile ‘short-term’ dietary exposure levels ranged from 0.94 to 4.6 μg/kg b.w. per day in adolescents and adult age classes, and from 3.6 to 6.2 μg/kg b.w. per day in toddlers across dietary surveys. For infants, only one consumption survey was available resulting in an exposure of 3.8 µg/kg b.w. per day. The highest 95th percentile ‘short-term’ dietary exposure level of 7.2 μg/kg b.w. per day was estimated in the age class ‘other children’.
For the scenario including suspect samples, the estimations of mean ‘short-term’ dietary exposure to high percentile levels of perchlorate ranged from 0.54 to 5.0 μg/kg b.w. per day in adolescents and adult age classes, and from 2.2 to 4.4 μg/kg b.w. per day in toddlers across dietary surveys. For infants, only two consumption surveys were available resulting in an exposure of 1.5 and 2.3 µg/kg b.w. per day. The highest mean ‘short-term’ dietary exposure level of 5.3 μg/kg b.w. per day was estimated in the age class ‘other children’. The 95th percentile ‘short-term’ dietary exposure levels ranged from 1.3 to 14 μg/kg b.w. per day in adolescents and adult age classes, and from 4.5 to 9.4 μg/kg b.w. per day in toddlers across dietary surveys. For infants, only one consumption survey was available resulting in an exposure of 5.7 µg/kg b.w. per day. The highest 95th percentile ‘short-term’ dietary exposure level of 18 μg/kg b.w. per day was estimated in the age class ‘other children’.
For all age groups, the ‘short-term’ exposure estimates under the scenario in which suspect samples were included were higher compared to the scenario in which suspect samples were excluded. The use of the suspect samples increased the ‘short-term’ mean exposure estimates in infants and toddlers by up to 64 % and the 95th percentile by up to 53 %. For all other age groups, the inclusion of suspect samples resulted in ‘short-term’ mean exposure estimates that were from 32 to 163 % higher and the 95th percentile exposure estimates that were from 21 to 227 % higher compared to the scenario excluding the suspect samples.
Perchlorate is readily and extensively absorbed from the gastro-intestinal tract in humans and rats. Following absorption, it is widely distributed in the body with the highest concentrations occurring in the thyroid, and it is rapidly excreted mainly in the urine as unchanged parent compound. Following the repeated exposure of rodents to perchlorate, findings included changes in thyroid hormones and thyroid stimulating hormone (TSH) levels and thyroid weight increases. Histopathological findings were also observed in the thyroid (colloid depletion, follicular cell hypertrophy and hyperplasia) and mammary gland (mild atrophy, atypia of the lobular epithelium, scattered foci of marked hyperplasic activity). Thyroid tumours were observed following chronic exposure in rats and mice.
In the thyroid of humans and rodents, perchlorate competitively inhibits the uptake of iodine via the sodium-iodide symporter (NIS). Iodine uptake in the thyroid is a key step in the synthesis of thyroid hormones, and its inhibition may result in the disruption of thyroid hormone synthesis and consequently disruption of the homeostasis of the hypothalamic-pituitary-thyroid axis, leading eventually to the development of hypothyroid symptoms. In comparison with rats, healthy adult humans have lower thyroid hormone turnover rates and larger reserves of iodinated thyroglobulin, allowing them to compensate for reduced hormone synthesis in the thyroid. Due to these differences in thyroid hormone physiology, the data from toxicological studies in rats are of limited use for extrapolating to humans. Human fetuses, neonates and individuals with low iodine intake or genetically predisposed to develop hypothyroidism are potentially more susceptible to the effects of exposure to perchlorate.
In humans, severe iodine deficiency as a result of insufficient iodine intake or sustained exposure to goitrogenic substances, such as perchlorate, at levels that induce depletion of the thyroid hormone stores can result in hypothyroidism. However, a mild to moderate iodine deficiency can lead to the development of toxic multinodular goitre and can result in hyperthyroidism.
Potassium perchlorate has been used for the medical treatment of hyperthyroidism at doses ranging from 400 mg/day to 2 000 mg/day (corresponding to 4–20 mg perchlorate ion/kg b.w. per day for a 70-kg person) administered for prolonged periods to control thyroid hormone levels. Adverse effects have been reported following application of potassium perchlorate at doses ≥ 400 mg/day, with evidence suggesting a direct relationship between the incidence and severity of the effects and the treatment dose and duration. Studies on healthy adult volunteers repeatedly exposed to 0.007–0.5 mg perchlorate ion/kg b.w. per day for two weeks, as well as information from occupational studies, showed no correlation between the exposure to perchlorate and any adverse effects or changes in the thyroid hormone levels, even at exposure levels associated with a substantial inhibition of thyroid iodine uptake. Retrospective epidemiological studies at the general population level, including the most sensitive subjects, showed contradictory results and found no clear association between exposure to perchlorate and an increased incidence of thyroid dysfunction. The CONTAM Panel concluded that these studies were not of use for the risk assessment of perchlorate.
The CONTAM Panel noted that in its evaluation JECFA considered the inhibition of thyroid iodine uptake of 50 % as the benchmark response (BMR), with the justification that both short-term and chronic exposure to perchlorate in healthy adult volunteer studies had shown that such a level of inhibition is not associated with any changes in TSH or thyroid hormone levels.
The CONTAM Panel concluded that the chronic adaptive changes to compensate for a sustained inhibition of thyroid iodine uptake could lead to long term effects such as the development of multinodular toxic goitre, in particular in populations with mild to moderate iodine deficiency. The CONTAM Panel concluded that a prolonged 50 % inhibition of thyroid iodine uptake by exposure to NIS inhibitory chemicals such as perchlorate may lead to goitre and multinodular toxic goitre, even if short-term exposure does not alter thyroid function tests. Although the consequences of thyroid iodine uptake inhibition below 50 % is unclear, the CONTAM Panel performed benchmark dose (BMD) modelling on the thyroid iodine uptake inhibition using human dose-response data from the Greer et al. (2002) study, applying a benchmark response of 5 %, which is the default value for continuous data. The CONTAM Panel selected the lowest 95 % lower confidence limit for the BMD response of 5 % extra risk (BMDL05) of 0.0012 mg/kg b.w. per day as the reference point and established a total daily intake (TDI) of 0.3 µg/kg b.w. per day by applying an uncertainty factor of 4 to allow for inter human differences in toxicokinetics. No additional uncertainty factors were considered necessary to allow for intraspecies differences in toxicodynamics and for the short duration of the human study.
No data are available on the acute toxic effects of perchlorate in humans. In adults, a single treatment with 1 000 mg potassium perchlorate (10 mg perchlorate ion/kg b.w. for a 70-kg person) is used for diagnostic practice without any adverse effect reported. Amongst the vulnerable subpopulations, potential acute effects of perchlorate have been suggested for fetuses and infants, because they lack the reserve capacity that exists in adult humans and because of the key role of thyroid hormones in fetal and neonatal neurological development. The CONTAM Panel noted that a single-day acute exposure to perchlorate at levels found in food and drinking water is unlikely to cause adverse effects on human health, including the more vulnerable groups of the population. The CONTAM Panel concluded that the establishment of an acute reference dose for perchlorate is not warranted.
The CONTAM Panel considered whether it was possible to identify a level of short-term exposure that would not be expected to cause adverse effects. The CONTAM Panel concluded that short-term exposure for two to three weeks to perchlorate, at levels that are high enough to result in a severe depletion of the thyroid iodine depot, would be critical in breast-fed infants and young children. This would be a particular risk in the case of mild to moderate iodine deficiency. However, such a depletion would be associated with compensatory increases in the activity of the thyroid iodide transporter, and no data are available to evaluate in detail how large the doses of perchlorate would be necessary for such depletion. Therefore, the CONTAM Panel could not establish a short-term health-based guidance value for these populations.
The estimated mean chronic dietary exposure levels for adolescents and the adult age groups did not indicate a health concern when compared with the TDI of 0.3 µg/kg b.w. per day. At the estimated 95th percentile chronic dietary exposure there was an exceedance of the TDI for some surveys. In the younger population groups (‘infants’, ‘toddlers’ and ‘other children’), the TDI was exceeded for both mean (in some surveys) and 95th percentile (in the majority of surveys) exposure estimates. In addition, the estimated exposure for breast-fed infants largely exceeded the TDI. However, the relevance to the European Union is unknown since the estimates for human milk are based on limited literature data from the USA.
Overall, the CONTAM Panel concluded that the chronic dietary exposure to perchlorate is of potential concern, in particular for the high consumers in the younger age groups of the population with mild to moderate iodine deficiency. Furthermore, it is possible that exposure to perchlorate is of concern for infants breast-fed by iodine-deficient mothers and in the short-term for young children with low iodine intake.
The CONTAM Panel recommended that more data should be collected on the occurrence of perchlorate in food in Europe, especially for infant formula, and milk and dairy products. The CONTAM Panel identified the need for biomonitoring data for perchlorate and the associated iodine status in Europe, including data on urine and breast milk, and noted that additional data on the level and duration of thyroid iodine uptake inhibition that has an impact on thyroid hormone levels in the vulnerable subpopulation groups would improve the risk assessment. There is a need for a better understanding of the contribution of various dietary factors to the overall thyroid iodine uptake inhibition.