The allowance levels currently valid for phycotoxins are based mainly on data derived from poisoning incidents. However, these data are seldom accurate and complete, and mainly restricted to acute toxicity. In some cases, the allowance level is also adapted to the limitations of the detection method. For risk assessment purposes, human intake levels of (shell)fish should be standardized.
Currently the toxicological risk evaluation for PSP toxins can only be based on acute toxicity data. Sub-chronic and chronic data for animals as well as humans are not available. Lowest doses causing mild symptoms of PSP in humans vary between 120 and 304 µg/person and lowest doses associated with severe intoxications/fatalities vary between 456 and 576 µg STX/person. In order to protect more susceptible persons (children, elderly, unhealthy) usually an uncertainty factor of 10 is applied for calculation of TDI values for contaminants, based on human data. However, for PSP the calculations are complicated by the following factors: at what levels should the effects be considered as adverse, and what level is the actual NOAEL and LOAEL? On the other hand, since the data on PSP represent many individuals, displaying large differences in susceptibility, an uncertainty factor of 10 may not be needed (Aune, 2001). Most countries apply a tolerance level of 80 µg STX eq/100 g mussel meat. If the consumption of shellfish is estimated to be between 100 and 300 g/meal, a margin of safety of about < 1 to 3.8 toward mild symptoms is present and, more important, a margin of safety of only 1.9 to 7.2 toward serious intoxications or death. These margins are quite small or there is no margin at all.
However, it is neither practical nor realistic to establish a very low tolerance level because the mouse bioassay is currently the most widely used method to determine PSP toxins and the present detection limit of this assay is approximately 40 mg PSP (STX eq)/100 g shellfish. Once more sensitive (and reliable) analytical chemical methods are available, the toxicity figures of STX and derivatives after acute and (sub)chronic exposure should be re-evaluated.
The various toxins in the DSP complex can be divided into three groups namely okadaic acid and the structurally related DTXs, the PTXs and the YTXs.
An EU Working Group on Toxicology of DSP and AZP has recommended allowance levels for these three groups of DSP toxins (EU/SANCO, 2001).
OA and DTXs
In animal experiments, cancer promoting and genotoxic effects of OA and DTXs are seen at relatively high doses and long exposure periods compared with the levels causing diarrhoea in humans shortly after consumption of contaminated shellfish. Consequently it is unlikely that a substantial risk of cancer exists in consumers of shellfish due to these toxins. Therefore, human risk assessment is based on a N(L)OAEL from animal or human data with the use of an uncertainty factor. Human data are preferred when available.
Taking into account all human exposure figures, it can be
concluded that the lowest levels causing diarrhoeic effects in humans vary from
32 to 55 µg OA and/or DTX1. These figures have been derived from Japanese
and Norwegian human data. The effects seem to be restricted to diarrhoea,
vomiting, headache and general discomfort. No serious and irreversible adverse
health effects have been seen at these levels (EU/SANCO, 2001). Current European
Regulations allow maximum levels of OA, DTXs and PTXs together of 160 µg OA
eq/kg edible tissue. If the consumption of shellfish is estimated to be between
100 and 300 g/meal, there is a margin of safety of about < 1 to 3.4 toward
the diarrhoeic effects. These margins are quite small or there is no margin at
all. EU/SANCO (2001) stated that, if the level of OA and DTXs in shellfish is
not higher than 16 µg/100 g shellfish meat, there is no appreciable health
risk at a consumption of
100 g mussel meat/day.
Concerning the PTXs, human toxicity data are not available. Therefore a safe level for humans is based on animal toxicity data. For toxins in the PTX group, data on animal toxicity are only available for PTX2. Effects such as tumour induction and tumour promoting are not known. The LOAEL for PTX2 by oral administration to mice was reported to be 0.25 mg/kg bw based on diarrhoeic effects and effects on the liver. The NOAEL should be estimated by applying a factor of 10 to the LOAEL. To extrapolate the animal data to human risks, a factor of 100 is applied. Thus, by applying an uncertainty factor of 1 000, a safe level of 0.25 µg/kg bw can be calculated for humans ~ 15 µg for an adult weighing 60 kg. EU/SANCO (2001) has recommended an allowance level of 15 µg/100 g shellfish meat. However, if the consumption of shellfish is estimated to be between 100 and 300 g per meal, the allowance level has to be between 5 and 15 µg/100 g edible shellfish tissue.
For PTX2 seco acid (PTX2-SA), human exposure data are available from a pipi shellfish poisoning event (56 cases of hospitalisation) in New South Wales (Australia) in December 1997 (ANZFA, 2001). According to Quilliam et al. (2000), PTX2-SA may have contributed to the gastrointestinal symptoms, vomiting or diarrhoea in humans (Aune, 2001). Burgess and Shaw (2001) reported that the patients consumed approximately 500 g of pipis containing 300 µg PTX-2SA/kg (~150 µg PTX-2SA/person ~2.5 µg/kg bw for a 60 kg weighing person). A safe level for humans of 0.025 µg/kg bw for PTX-2SA can be calculated by applying an uncertainty factor of 100 (10 for intraspecies differences and 10 for extrapolation from LOAEL to NOAEL) (~1.5 µg/person weighing 60 kg). This means that for PTX2-SA, the allowance level has to be between 0.5 and 1.5 µg/100 g edible tissue at consumption between 100 and 300 g per meal.
For the YTXs, no human data are available. Therefore, a safe level in humans is based on animal data. The NOAEL in mice by acute oral administration was estimated to be 1.0 mg/kg bw based on cardiac effects. A safe level for humans towards acute toxic effects of YTX is calculated to be 10 µg/kg bw by applying an uncertainty factor of 100. For an adult weighing 60 kg, this would mean a safe level of 600 µg YTX. In view of the lack of data on repeated administration and a high uncertainty factor recommended by WHO for a substance that injures cardiac muscles, the calculated safe level for humans given above could be lowered by a factor 6 to 100 µg (EU/SANCO, 2001). EU/SANCO (2001) recommended an allowance level of 100 µg YTXs/100 g shellfish meat. However, if the consumption of shellfish is estimated to be between 100 and 300 g per meal, the allowance level has to be between 33 and 100 µg/100 g edible shellfish tissue.
The generally applied guideline value of 20 mg DA/kg mussels is derived from an ASP incident in Canada (Prince Edward Island) and is taken on by several other countries. The guideline level of 20 mg DA/kg is equal to an intake of 0.03 to 0.1 mg DA/kg bw per person with a body weight of 60 kg assuming that consumption of mussels is between 100 and 300 g/meal. The epidemiological data used to derive the guideline value, revealed mild gastrointestinal effects in humans at 1 mg DA/kg bw. Afterwards the guideline value was supported by acute studies in animals. However, when doses required to cause overt toxicity in animal species were compared, mice and rats appeared to be relatively insensitive compared with monkeys and oral dosing required more toxin (more than 10 times in rodents) to achieve the same effects as i.p. dosing. Rats showed overt effects of DA poisoning at single oral doses of about 80 mg/kg bw, whereas monkeys showed vomiting, gagging and yawning already at 1 mg/kg bw. A single oral dose of 0.75 mg DA/kg bw in monkeys did not induce overt effects. This apparent decreased sensitivity in rodents may be the result of their inability to vomit and/or the finding that the plasma half-lifetime of DA in the rat is about 6 times less than that in the monkey. Comparing the guideline value of 20 mg DA/kg of mussel tissue (~ 0.1 mg/kg bw for humans assuming a consumption of 300 g mussels per meal) with the no-effect dose (0.75 mg/kg bw) in acute oral studies in monkeys, a factor smaller than 10 is between these figures. There is no knowledge of the effects of long-term exposure to low levels of DA. However, short-term animal studies with repeated exposure do not point to altered DA clearance from serum or greater neurotoxic responses than after single exposures.
Reasonable good dose-response data were determined for 10 persons involved in the Canadian incident (elderly people, aged from 60 to 84 years). According to these data the NOAEL is 0.2-0.3 mg DA/kg bw, while the LOAEL was 0.9-2.0 mg DA/kg bw and serious intoxications were recorded at 1.9 to 4.2 mg DA/kg bw. Interestingly, the intake estimates showed surprisingly large consumption of blue mussels, 120 to 400 g mussel meat per person per meal (Aune, 2001). This means that there is a factor two between the NOAEL and the regulatory limit of 20 mg DA/kg mussel meat which is equivalent to 0.1 mg/kg bw for a 60 kg weighing person with a mussel meat consumption of 300 g per meal. Between the LOAEL and the regulatory limit there is a margin of 9 to 20 and between the level of serious effects and the regulatory limit there is a margin of 19 to 42.
Based on the lack of sufficient data on toxicity and the analytical difficulties in determining brevetoxin exposure, risk assessment is not possible. Current risk management (in states on coasts of the Gulf of Mexico) is based on shellfish bed closures at 5 000 G. breve cells/litre with reopening based on determination of PbTx in shellfish at <80 µg/100 g.
EU/SANCO (2001) stated that based on poisoning incidents in Ireland, levels of AZAs causing human intoxication were calculated to be between 6.7 and 24.9 µg. These figures included a reduction in AZA content due to heating of the mussels. New data on heat stability revealed that this reduction of the toxin content due to heating was not justified. Therefore the recalculated range of the lowest observed adverse effect level (LOAEL) appeared to be between 23 and 86 mg per person assuming a maximum consumption of 100 g shellfish per meal. EU/SANCO (2001) applied a safety factor of three to convert the LOAEL to a NAOEL. Based on an intake level of a maximum of 100 g shellfish meat/meal, and the lowest LOAEL divided by three, EU/SANCO (2001) stated that an allowance level of 8 mg AZAs/100 g of shellfish should result in no appreciable risk for human health. To allow for detection by the mouse bioassay a level of 16 µg/100 g was proposed. However at a shellfish consumption of 300 g per meal, a person will consume already an amount of AZAs equal to the LOAEL in humans.
Ofuji et al. (1999b) reported a level for total AZAs in raw mussel meat in poisoning incidents of 1.4 µg/g of meat. At a consumption of 100 to 300 g per meal this means an intake of 140 to 420 µg AZAs/person. As these figures represent an effect level (LOAEL) usually a factor 10 is used for calculation of a NOAEL. This means that the NOAEL is 14 to 42 µg per person assuming a consumption of 100 to 300 g shellfish meat/meal. As a consequence the allowance level in shellfish meat has to be 14 µg/100 g. It has to be noted that no factor of 10 was applied to the NOAEL for intraspecies differences (variation in the human population).
The available animal data on ciguatoxin are not suitable for risk assessment. Therefore, human data derived from poisoning incidents should be used.
Mild CFP symptoms in some persons can be already expected after consuming fish containing the main Pacific ciguatoxin (P-CTX-1) at a level of 0.1 µg/kg. The main Caribbean ciguatoxin (C-CTX-1) is less polar and 10-fold less toxic than P-CTX-1. Assuming a fish consumption of 500 g per meal and a human body weight of 50 kg, this corresponds to 0.001 µg/kg bw (=LOAEL). These figures are derived from a large serving of the least toxic fish causing effects in some people. A level of 0.01 µg/kg bw, which is ten times the level causing mild symptoms in some persons, would be expected to be toxic in most people. By applying an uncertainty factor of 10 (for intraspecies differences) to the lowest level causing mild symptoms in humans (=LOAEL), a safe level of 0.01 µg/kg of fish flesh can be calculated (Lehane, 2000; Lehane and Lewis, 2000). It has to be noted that the usual application of an uncertainty factor of 10 to the LOAEL for calculation of a NOAEL was not performed.
At present the risk assessment of phycotoxins has not been performed in a straightforward way. Risk management and risk assessment have been mixed in the process complicating the procedure. In general, there is a lack of toxicological data particularly on repeated exposure. Epidemiological data mainly existed of poisoning incidents with their inherent limitations. This all cumulated in provisional risk assessments of certain phycotoxins which were not always logic and consistent. For some phycotoxins, even the lack of minimal data has prohibited risk assessment.
If adequate scientific (toxicological, epidemiological and occurrence) data are available, a risk assessment can be performed by applying generally accepted safety or uncertainty factors. An adequate set of animal data will allow the derivation of a no-observed adverse effect level (NOAEL). A safe level for humans can be calculated by applying an uncertainty factor of 100 (10 for interspecies differences and 10 for intraspecies differences) to the NOAEL. If an adequate set of human data is available, a safe level for all humans can be calculated by applying an uncertainty factor of 10 (for intraspecies differences) to the NOAEL, derived from those human data.