Consumption of a variety of shellfish and fish causes an increasing number of human intoxications around the world. Diagnosis depends mainly on recognition of specific signs and symptoms and on identification of marine toxins present in remains of the seafood involved. Indicators for effects and exposure are usually not available due to inadequate analytical methods for the sometimes complex algal toxin mixtures. The effects of algal toxins are generally observed as acute intoxications. Health effects of episodic exposure and chronic exposure to low levels of algal toxins are hardly known. The latter effects may go unreported by the affected individual(s) or may be misdiagnosed by physicians.
Monitoring seafood for toxicity is essential to manage the risks. However, there are several limitations in monitoring for toxicity such as the variation in toxin content between individual shellfish, different detection and even extraction methods for the various toxins requiring a decision which toxins one is testing for, and the frequency of sampling to ensure that toxicity does not rise to dangerous levels in temporal or spatial gap between sampling times or locations. Furthermore, the growing harvest of non-traditional shellfish (such as moon snails, whelks, barnacles, etc.) may increase human health problems and management responsibilities.
Monitoring for toxic plankton may possibly overcome some of these problems. However, plankton populations are patchy and ephemeral, it is difficult to make a quantitative correlation between numbers of toxic plankton and levels of toxins in seafood and the amount of toxin per cell can vary widely. Data on the occurrence of toxic algal species may indicate which toxins may be expected during periods of algal blooms and which seafood products should be considered for analytical monitoring. A problem is that certain algal species, which have never occurred in a certain area, may suddenly appear and then rapidly cause problems. The plankton observations are used to focus toxicity testing, but are not in themselves used for regulatory decisions. Moreover, most monitoring and regulatory programmes often are not adequate to meet the expanding threat of new harmful algal blooms. As a result, when new outbreaks occur, the response is often uncoordinated and slow. Harmful algal blooms cannot be predicted and there is little information on bloom initiation.
Toxic blooms are mostly detected by visual confirmation (water discolouration and fish kills), illness to shellfish consumers and/or human respiratory irritation with actual toxicity verified through time-consuming mouse bioassays and chemical analyses in shellfish samples. This after-the-fact strategy is the consequence of the extremely difficult prediction of the occurrence and magnitude of a bloom. To prevent human intoxication, monitoring programmes relying on enumeration and microscopic identification of harmful taxa in water samples generally suffice. However microscopic based monitoring requires a high level of taxonomic skill, usually takes considerable time, and can be highly variable among personnel.
One of the most serious problems is the lack of information on the biology of harmful algae. For example, little is known about the abundance, distribution, population dynamics and physiology of most of the harmful species, both in local waters and elsewhere. Long-term, routine monitoring of phytoplankton and the environment is essential to obtain data necessary to determine even the most elementary ecology of harmful species. Moreover, because bloom dynamics are complex, the factors that determine bloom dynamics of a species in one geographic area may not affect that species in another area, even though the areas are not widely separated. Therefore alternative evaluation systems for predicting bloom occurrences are highly desirable.
In establishing regulatory criteria and limits for marine toxins, various factors play a role such as the availability of survey data, the availability of toxicological data, the distribution of the toxins throughout sampled lots and the stability in the samples, the availability of analytical methods and regulations already in force in several countries. With respect to toxicity, until now only data on the acute oral toxicity both in experimental animals and humans are available for the majority of the marine toxins. However, repeated exposure to lower sublethal dose levels may be a common feature.
Concerning detection methods, there is a general, worldwide need for rapid, reliable and sensitive methods to determine marine toxins in (shell)fish. The present mouse bioassay is not sensitive enough, shows a considerable variation, is time consuming, is vulnerable to interferences and is unethical in terms of animal welfare. Quilliam (1998b) argues for LC-MS as a universal detection method for all marine toxins. This technique has a low limit of detection, high selectivity and the ability to deal with the structural diversity and labile nature of the toxins. In addition, separation of complex mixtures, accurate and precise quantitation, automation and high throughput, legal acceptability for confirmation and structural information of new toxins are possible with this method. Another new approach that seems promising is the development of biosensors with which multiple toxins can be determined simultaneously.
The development and introduction of adequate and efficient analytical methods can be accelerated by providing information in a fast and proper way, for instance by setting up an Internet accessible database. The database should include parameters such as (chemical) names, physical/chemical properties, classification(s), toxic effect(s), sources, habitat, regulatory limits and literature references.
9.1.1 Conclusions related to Paralytic Shellfish Poisoning (PSP)
The tolerance levels set for PSP toxins thus far are largely pragmatic decisions based on intoxication events, and although there are many reported cases of human intoxications due to shellfish toxins, it is difficult to obtain reliable human toxicity data. For example, variations in observed toxicity of PSP toxins to humans may be due not only to variable sensitivity between people, but also to the composition of individual toxins in the samples. Toxin profiles can vary according to the species of shellfish consumed and the area of harvest. In addition, toxic doses are often estimated from left-over toxic seafood. This is not necessarily representative of the ingested food because PSP toxins may be unevenly distributed throughout lots and within individual shellfish, and not all PSP toxins are stable.
It is possible to measure PSP compounds by a number of analytical-chemical methods but they all have some limitations, and they often cannot easily be operated because of the lack of reference materials, although recently some progress has been made in this area. In 2003, certified standards of STX, neoSTX, dc-STX, GNTX 1-4, GNTX 2/3 and GNTX 5 are commercially available. However, they are expensive and mainly available from one source. Yet, their availability significantly improves the quality of the data that are obtained by LC-methodology. The efforts undertaken by the European Commissions SMT Programme have led to shellfish reference materials with certified mass fractions of some of the toxicologically most significant PSP toxins. Despite these positive developments, the analytical situation remains difficult and the lack of pure PSP compounds in sufficient quantities for repeated dose toxicity studies is a limiting factor in the development of reliable risk assessment.
9.1.2 Conclusions related to Diarrhoeic Shellfish Poisoning (DSP)
The variety in biological activities of the DSP toxins may cause some problems. Although PTXs and YTXs are acutely toxic to mice after i.p. injection, their oral toxicity to humans is unknown. Therefore, more toxicological data on PTXs and YTXs have to become available. Furthermore OA and DTX possess tumour promoting activity and OA shows also genotoxic and immunotoxic activity. These effects raise questions as to the human health risks of (sub)chronic exposure to low levels of these compounds. A pressing problem is the lack of sufficient quantities of DSP toxins to perform (sub)chronic animal toxicity studies.
Although mammalian bioassays for DSP toxicity are applied worldwide, there are large differences in performance of, for instance, the mouse bioassay (toxicity endpoint is animal death; no consensus on appropriate observation time) among different countries, resulting in differences in specificity and detectability. A major problem is the fact that the mouse bioassay detects all DSP components and probably also other toxins. However, it is not possible to distinguish between the various toxins whereas specific legal limits for the toxin groups have been established (for instance in the EU). On the other hand, the rat bioassay detects only OA and DTXs because the endpoints in this assay are soft stool, diarrhoea and feed refusal which effects are known to be caused by OA and DTXs only (and AZAs).
Chemical methods (LC) are useful for identification and quantification of selected diarrhoeic toxins (usually OA or DTXs). Recently an LC method for the detection of YTXs was developed, but until now no method for PTXs is available except an LC-MS method; however its performance is not yet satisfactory. Chemical methods are applied as a regulatory tool primarily for confirmation of the results obtained in a bioassay.
None of the many approaches to determine DSP toxins in shellfish has been evaluated in a formal collaborative study according to ISO/IUPAC/AOAC so that the performance characteristics are not fully known. The further development, evaluation and comparison of the various techniques would become significantly easier if reliable reference standards and reference materials (such as lyophilized mussel samples with certified contents of several DSP toxins) could be developed and made available to the scientific community.
9.1.3 Conclusions related to Amnesic Shellfish Poisoning (ASP)
Compared to the paralytic and diarrhoeic shellfish poisons, problems with amnesic shellfish poisons seem to be of a lesser magnitude. Only one confirmed outbreak of ASP causing severe illness in exposed people was reported worldwide, specifically in Prince Edward Island, Canada in 1987. After the first outbreak in Canada, human illnesses (mild and short lived) were only observed in one outbreak, specifically after consumption of contaminated razor clams (from the West Coast of the United States). However, health authorities were not able to confirm that the illnesses were caused by DA. In two outbreaks, the death of cormorants and/or brown pelicans due to the consumption of contaminated anchovies or mackerel was reported indicating that herbivorous fish can act as vectors for DA. In the last few years (1999 to 2002), DA was detected also in shellfish from some European countries.
Methods of analysis for DA are rather straightforward and less complex than those for paralytic and diarrhoeic shellfish poisons. One chemical method for DA in mussels (LC with UV detection) has been successfully validated in a formal collaborative study, whereas another (improved) method is currently subject to a collaborative study. Certified reference materials and calibrants are readily available.
9.1.4 Conclusions related to Neurologic Shellfish Poisoning (NSP)
When humans are exposed to brevetoxins, different exposure routes are possible; the oral route via consumption of contaminated shellfish, the inhalatory route via exposure to aerosolised brevetoxins, and the dermal route via direct contact with contaminated seawater. The effects of the various exposure routes on humans are difficult to assess because toxicity data for brevetoxins are limited. Some acute studies in mice and data from poisoning cases in humans and (marine) mammals are available but acute dermal and inhalation studies are lacking, as well as oral, dermal and inhalation studies with repeated exposure of laboratory animals. Therefore reliable hazard assessment is not possible.
Pure toxins and toxin metabolites would be needed to be able to carry out toxicity studies. In addition, analytical reference materials would be needed to further develop and improve the analytical methodology and to allow analytical quality assurance of monitoring laboratories. Currently the various obstacles on the way to reliable assessment of brevetoxin occurrence and exposure further hamper risk assessment and thus the establishment of meaningful regulations.
Despite these problems, regulations for NSP toxins in shellfish are in force in a few countries, specifically the USA, Italy and New Zealand based on the mouse bioassay. The action level is 20 MU/100 g shellfish flesh (~80 mg PbTx-2/100 g shellfish flesh).
9.1.5 Conclusions related to Azaspiracid Shellfish Poisoning (AZP)
One cause for concern is the lung tumours found in mice after repeated doses of 20 µg AZA/kg bw and higher. This finding should be confirmed in experiments with larger numbers of mice and longer exposure periods (Ito et al., 2002).
The current allowance level has to be revised as new data become available. However, the lack of supply of pure toxins is a serious obstacle to all kinds of studies. The production of pure toxins, in turn, depends on the availability of large amounts of toxic mussels. Development of rapid detection methods such as LC-FLD, ELISA and functional assays should be explored.
9.1.6 Conclusions related to Ciguatera Fish Poisoning (CFP)
Ciguatera poisoning mainly occurs in tropical regions throughout the world and is sporadic in Europe, particularly in the Northern European countries. Therefore, a regular analytical check on the presence of ciguatoxins in imported large predatory fish from endemic areas is considered adequate in countries which are not an endemic area for CFP.
A few specific regulations exist for ciguatoxins. A positive finding in a fish would remove that fish from sale. In some cases, restrictions are placed on the sale of fish of certain species or size from a given area, with no testing of the toxin. The larger a fish is, the older it probably is, and the more toxin it has probably accumulated. Reef carnivores considered being regular ciguatoxin carriers are often banned from sale as a matter of principle. The hazard is linked to the accumulation in the food chain of a toxin, which is impossible to link with any algal bloom. Cell counting of plankton will not predict when a fish has accumulated ciguatoxins or not (Boutrif and Bessey, 2001).
Based on the preceding conclusions, the following recommendations are presented:
1. Data on bloom development with respect to hydrographic and climatic conditions, and nutritional status of the water column are needed.
2. Toxicity studies on effects after repeated exposure to marine toxins should be performed.
3. Chemical analytical techniques capable of separating, identifying and quantifying individual marine toxins should be further developed
4. As alternatives to rodent assays, assays have to be developed to be used when uncharacterized bloom events occur. Emphasis on the use of in vitro techniques where blooms have been characterized should reduce the use of test systems with live animals.
5. To facilitate fast application of adequate analytical methods for marine biotoxins, a database should be developed including basic data on marine biotoxins such as chemical structures, physical/chemical properties and analytical methods.
6. Both for the submission of toxicity data and for the development and validation of analytical techniques, the production of pure toxin standards and certified reference material are required.
7. Formal risk assessments of the marine biotoxins should be performed by recognized international bodies - such as the Joint FAO/WHO Committee on Food Additives (JECFA) and the European Food Safety Authority (EFSA) - and should be based on sound scientific data of toxicity and exposure. In the absence of sufficient data, an expert consultation could be considered in order to explore the possibilities for adequate risk assessment which should be the basis for meaningful regulations.