Australia - Measures Affecting Importation of Salmon Report of the Panel (Continued) 6.45 Of equal, if not more relevance, Dr. Rodgers indicated, was the recently published (October 1997) report by the New Zealand government entitled "Import health risk analysis: salmonids for human consumption". This was effectively a more comprehensive update to the earlier 1994 Report, since it contained additional factual information on pathogen survival, inactivation and likely tissue levels. In addition, the quantitative risk assessment model had been modified and improved by using alternative distributions for the probabilities arising from expert opinion and the possible number of infected fish imported per tonne of product. It was also a more global report because it considered farmed Atlantic salmon from Norway, wild Pacific salmon from the Pacific North West of America and farmed rainbow trout from Denmark. As a result, two additional inputs were included, namely those of prevalence of infection of harvested fish and fish weight. Taking furunculosis as an example again, one of the conclusions had been that, whereas a qualitative risk assessment indicated that the risk of the introduction of Aeromonas salmonicida (the bacterial causal agent of furunculosis) through importations of the commodity (products) was low, quantification of the risk further demonstrated that such an event was unlikely to occur, particularly given the relatively low annual volumes which were likely to be imported. 6.46 Dr. Wooldridge replied that the New Zealand risk assessment addressed the question of import of the same commodity, ocean caught Pacific salmon from Canada, into New Zealand. As with Australia, it also took into account the effect of evisceration and head removal. It included a qualitative analysis, including an assessment plus management recommendations on the full range of hazards identified, and a quantitative assessment on one particular hazard, Aeromonas salmonicida. 6.47 In terms of the qualitative analysis and assessment, the potential pathway from hazard to unwanted outcome (the introduction of fish disease) was outlined on page 3. This included steps involved in release (i.e. import of infected material) and steps involved in exposure and transmission in New Zealand. The main assessment considered 23 diseases as hazards. These were all diseases included in the Australian risk assessment; the one considered in the Australian assessment but not included in the New Zealand assessment was Whirling Disease (Myxobolus cerebralis), but it was stated earlier in the document that Whirling Disease was already present in New Zealand (page 10). 6.48 Thus the same commodity was being assessed, and it was passing through the same processes in Canada before export. The export or release risks for each disease associated with the export of the same grade, quality and processing of salmon at the point of export would therefore be the same regardless of their intended destination. Up to this point therefore, this assessment was completely relevant both scientifically and technically to the dispute and the Australian Final Report. One would expect broadly the same data to be used, and the conclusions regarding the assessed risk of ‘release’ to be the same. Any discrepancy would be a cause for further investigation. 6.49 Differences in transit conditions, processing in the destination country, local exposure pathways, and susceptibility of local fish might be different, and data, assessment results and conclusions would not necessarily be the same for exposure and likely consequences. However, the basic methodology should be similar, and was therefore technically relevant. In actual fact, despite the possibility of differing exposures etc., the New Zealand qualitative assessment was for the most part similar in content, method and conclusions to that in the draft Australian assessment. 6.50 The New Zealand assessment also gave, for each disease, a very clear summary of the reasons for the conclusion reached, and this was a very useful feature of the assessment. From this it was possible to see which diseases were considered of negligible risk due to their release assessment results, and of the 23 considered, 15 were classified as of negligible risk on this assessment alone. Of the remaining 8, four were considered to be of negligible risk partly because they require specific intermediate hosts not considered present locally, and four partly because they were already present in New Zealand, although in all cases release assessment results also played a prominent part in reaching the conclusions. Nevertheless, the exposure patterns and local disease situation (that is, the data used) could, for these eight, be compared directly with Australia if required, as could that for Whirling disease. 6.51 Dr. Wooldridge also commented on the quantitative assessment as used in the New Zealand Report. In both countries assessments, furunculosis caused by Aeromonas salmonicida, was considered to be the most likely disease to be introduced. The New Zealand assessment reached the conclusion, based on the release assessment result, that "Wild, ocean-caught Pacific salmon from the west coast of North America are unlikely to introduce A. salmonicida when imported as headless, eviscerated fish" (page 35). It went on to say "However, because A. salmonicida is the one pathogen of salmon likely to be present in a high concentration in the muscles of diseased fish, it is the subject of a separate quantitative risk assessment" (page 36). 6.52 The method used in this quantitative assessment was based on Monte Carlo (stochastic) simulation, which allows real life uncertainty and variability to be used in the distributions incorporated into the model. Outputs, rather than being single point values, are also given as distributions, which can be illustrated graphically and analyzed statistically. They therefore give much more information than a single point deterministic model, and are in general the preferred method for such quantitative assessments. 6.53 The assessment evaluated the risk of four possible product presentations (page 56) but most importantly, the release assessment was explicitly separated from the exposure assessment (e.g. page 58). As with the qualitative assessment, this release assessment was both scientifically and technically applicable to Australia as well as New Zealand. Similarly, although specific import data (in particular estimated import tonnage per annum) and exposure routes and susceptibilities might be different, requiring an exposure model different in detail, the technical methods employed in this quantitative assessment were fully applicable to any other import destination including Australia. 6.54 Dr. Wooldridge noted that a quantitative assessment required also the availability of appropriate data, and the Australian Final Report argued that such data was not available. However, it was clear from the New Zealand assessment that sources of data for the release assessment part, for this disease at least, did exist and in her opinion no convincing argument was made as to why such data could not also have been utilised by Australia. She observed that there was generally much more data in existence for almost any quantitative risk assessment than at first sight seemed likely to be available. However the source and type of such data was not always immediately apparent until a model was under development, when specific data requirements could be fully considered. 6.55 In summary, Dr. Wooldridge considered that all of the technical methodology and a large proportion of the scientific data used in the New Zealand risk assessment was relevant to both the dispute and the Australian Final Report. She recalled that she was not competent to comment upon the accuracy of the fish-based data, but specifically for that used in the release assessments and given the conditions specified, if it was correct for New Zealand, then it was also correct for Australia. Question 3. Canada argues that for a risk assessment to be appropriate it is not enough to claim/prove the possibility of risk. According to Canada, one has to evaluate and give a certain probability of risk. What is your view on this point from a scientific/technical perspective? To what extent, from a scientific/technical point of view, does one have to quantify, or use expressions which qualify, a risk in a risk assessment for the risk assessment to be deemed satisfactory? To what extent can you establish the probability of occurrence of the risk involved when dealing with more than one disease agent? 6.56 Dr. Burmaster indicated that he agreed with Canada on this point. He thought that a quantitative risk assessment using probabilistic techniques was the most appropriate tool for resolving this dispute between the two countries. Although he could not specify sharp boundaries as to what extent it was necessary to quantify, or use expressions which quantify, a risk for a satisfactory risk assessment, he considered that the current dispute certainly fell within the region where a fully quantitative risk assessment was necessary. With regard to the establishment of the probability of occurrence of risk when dealing with more than one disease agent, he referred to his response to Question 4. 6.57 Dr. Rodgers replied that the difference between the possibility of risk and the probability of risk was largely based on the availability of reliable data. When there was insufficient data for a quantitative risk assessment it was necessary to undertake a qualitative analysis instead. Alternatively both types of analysis could be used to compliment each other. However, the more mathematical concept of quantitative risk assessment, which used probability distributions in a predictive way gave an indication of the probability of an event occurring. On the other hand the same probability could not be generated from a subjective qualitative analysis, since only the possible outcome of an event, or series of events, could be generated using this methodology. In other words, quantitative risk analysis attached numerical equivalents to qualitative estimates and assumptions. The quantitative approach was more accurate and ideally should be used (data permitting) to provide a more valid answer to a potentially complex chain of interactions. However, this would mean that all the variables associated with an importation would need to be known with some accuracy in order that the risks could be established mathematically. 6.58 The possibility and the probability of an event occurring both embodied elements of likelihood and risk. However, Dr. Rodgers was not aware of any pre-requisite to use the quantitative method, particularly in view of the lack of data in certain key areas of aquatic animal health. This applied equally when dealing with more than one disease agent, since, although many risk factors were common between different diseases, each disease might have unique factors to consider and each of these would have a variable quantity and quality of usable data. 6.59 Dr. Wooldridge indicated with regard to the question of probability versus possibility, that in her opinion, the requirement of a risk assessment was to evaluate the probability of risk. She identified this as one of the minimum requirements (see response to Question 1). Given the existence of a particular disease agent, one could always construct a possible infection transmission scenario, however improbable, and therefore demonstration of the possibility of successful transmission and disease was not adequate. 6.60 However, the probability did not have to be expressed quantitatively, and frequently it could not be. In qualitative assessment there was therefore the difficulty of what was meant by the terms used, for example high, medium or low risk, and subjectivity was a potential problem. If there were several hazards being evaluated, it was often possible to distinguish between them and thus class them as high or low within the group being considered. However, in import qualitative risk assessment and when dealing with a disease of serious consequences, one usually looked for something which could be described or classified as a "negligible" risk with respect to the consequence of causing a disease outbreak. That is, given all the data available, the conclusion reached by most people with appropriate expert knowledge would be that the likelihood of the organism passing through every necessary stage in the pathway from potential hazard to unwanted outcome was highly unlikely. It was accepted that there was still a degree of subjectivity even in this classification, and that was why the assessment must be fully transparent. 6.61 Dr. Wooldridge further noted that regarding initial qualitative classification, whatever the initial (unknown) risk, it was often possible to quantify the amount by which that risk could be reduced if certain safeguards are put in place. Thus, given qualitative evidence that a "low" risk initially existed, it might be possible to deduce that a "negligible" risk existed after application of the safeguards and a quantitative assessment of the difference. 6.62 With more than one disease agent, if a quantitative assessment could be undertaken for each disease, then it was possible to quantify the total risk. However, individual quantitative assessments were time consuming and might not be possible. In any event, qualitative assessments were the type generally done initially. 6.63 Whichever method was used, if any one of the diseases being assessed individually was considered by itself to present a risk, at a level agreed by all concerned to be unacceptable to the importing country, then that commodity would not be imported. Given this, it was therefore necessary always to first assess each disease risk separately. A problem only arose when each disease taken individually was assessed as representing a very low level of risk, and when those assessments were qualitative. 6.64 In this situation, if each individual risk had been classed as "negligible" it might be argued with some justification that the overall risk was therefore negligible. However, given some way of further ordering the risks within this "negligible" category, it might still be possible to decide on that disease with the maximum risk. In the approach taken by New Zealand in this situation, the disease with the highest assessed qualitative risk was quantitatively assessed. Although not done by New Zealand, the maximum total risk could then quantitatively be assessed by assuming that each of the other diseases had the same level of risk. This would overestimate the maximum total risk. Variations on this method to more closely model the differences between the diseases might also be possible depending upon data available. 6.65 Dr. Wooldridge additionally noted that it may be (for example due to immunocompromisation) that one disease pre-disposed to another. The effect of this might be difficult to determine, as whilst this might increase the risk of infection with a given disease, it might also increase the total detection rate of disease at inspection. Where there were negligible disease levels for both, this might be assessed as of little relevance either way. 6.66 It was accepted that where a group of diseases were qualitatively classed as other than very low risk, with no additional data, estimating the total risk would be difficult. But the initial proviso above suggests the commodity would in that case be rejected in any event. Further, where a disease was not low risk, then it was often likely to be because it was more prevalent and therefore more was often known about it; if this was the case, there may well data available to quantify that risk. 6.67 Dr. Wooldridge summarized that, although it might not always be possible to fully estimate the total risk when dealing with multiple diseases, it was likely to be possible to come to usable conclusions about the probability of the overall level of risk. Question 4. In the event a sanitary measure deals with several diseases, in terms of risk assessment technique, does one need to assess the likelihood of entry and establishment of each disease agent separately, all in combination, or both? To what extent is it necessary to account for different characteristics of the hosts and disease agents; of the use and disposal characteristics of the products? Should the risk assessment also include an assessment of the likelihood of entry, establishment and spread of the disease agent from imports of products of other species known to carry the same disease(s)? From a scientific/technical point of view, is it valid to undertake a risk assessment either on a disease-by-disease basis or on a product-by-product basis, or could either approach be adopted? What factors should be considered when making such a decision? 6.68 Dr. Burmaster presented the following notation in his response: Nsa = number of fish species of concern in Australia Nsc = number of fish species of concern from Canada Nd = number of fish diseases of concern Np = number of pathways of exposure Ne = number of events to complete one pathway Nw = number of different weather conditions of concern No = number of other factors to include Ntot = total number of variables/distributions to assess Australia might say that it could not undertake a risk assessment for which it must develop probability distributions for Ntot variables: Ntot = Nsa • Nsc • Nd • Np • Ne • Nw • No Dr. Burmaster disagreed with such a position. In his opinion, New Zealand had shown how to reduce the "dimensionality" of the problem to a manageable size by analyzing only certain combinations of Nsa, Nsc, Nd, Np, Ne, Nw, and No that were selected as the "most important" or "dominant" combinations. 6.69 He also believed that New Zealand had demonstrated how to reduce the dimensionality of the problem of different characteristics of the hosts and disease agent, or of the use and disposal characteristics of the products, to a manageable size. Dr. Burmaster indicated that the risk assessment should possibly also include an assessment of the likelihood of entry, establishment and spread of the disease agent from imports of products of other species know to carry the same disease(s). As a hypothetical example, fish disease D could arrive in Australia via the importation of frozen, eviscerated, and beheaded salmon from Canada with probability Pc. The same fish disease could arrive in Australia via the importation of live, ornamental goldfish from Malaysia with probability Pm. If a risk assessment were to show that Pc < Pm, it would not make sense for Australia to ban the importation of the salmon from Canada and not ban the importation of the goldfish from Malaysia. 6.70 When preparing the scope for a quantitative risk assessment, Dr. Burmaster considered that it was essential for the risk assessor to consider both a disease-by-disease basis and a product-by-product basis. He further observed that New Zealand had shown how to reduce the dimensionality of the problem to a manageable size by focusing on certain key combinations of fish species, pathways, events, etc. Along the way, the risk assessor(s) should use expert judgment to rank the importance of different combinations of species, diseases, pathways, etc. as a way to determine which combinations to quantify first. This was always an iterative process. As the risk assessment developed, new insights would develop that would allow the risk assessors to understand and place suitable emphasis on the most important combinations of species, diseases, pathways, etc. 6.71 With regard to the risk assessment technique, Dr. Wooldridge replied that, as detailed in the previous question, one needed first to assess the likelihood of the unwanted outcome (here the establishment of disease) separately for each disease. The outcome of this would determine the next step in the assessment, as detailed previously. 6.72 Hazard identification would automatically involve the initial host characteristics. As indicated in her answer to question 1, it was necessary to outline the steps in the pathway from hazard identification to unwanted outcome. This would include such things as inspection, processing etc. If different disease agents had different probabilities of being detected, inactivated, surviving in waste water, surviving whilst clinging to wrapping paper etc., then these characteristics must be taken into account. For some of these aspects (for example disease detection) the specific host being considered might also be relevant. For some of these characteristics (for example use of wrapping paper), use and disposal characteristics of the product might be relevant. Whenever relevant, such factors must be considered. In addition the susceptibility and disease status of the potential hosts in the importing country were relevant. 6.73 With regard to imports of other products of other species known to carry the same disease, information on the existence of such products was important in two respects. First, a comparison with other products might provide data necessary to assess the probability of exposure, transmission and consequences of a given disease post-entry within a country or region. Secondly, the information on the existence of such products was an important part of the overall risk analysis (as opposed to the risk assessment) and should be sought as part of that analysis. If such products were known to exist, it was relevant to the setting of acceptable levels of risk and therefore to the risk management part of the whole risk analysis. 6.74 Dr. Wooldridge considered that whether a complete risk assessment of any other particular product was required depended on the precise situation; in certain circumstances it might be enough to demonstrate that a particular product containing a given disease agent, and subject to similar use and disposal pathways, had been imported regularly for many years with no detected disease consequences. 6.75 The approach to a risk assessment depended upon the initial risk for which an assessment was required. If the question asked was "What is the risk of an exotic disease being introduced with product X?", then the hazard identification required that all exotic diseases potentially present in product X were identified, and the risk for each assessed. If the initial question asked was "What is the risk of introducing exotic disease Y into country Z?" then the hazard was disease Y and the potential pathways for import which would need to be considered included the initial identification of all products which could potentially contain disease agent Y and were (or might be) imported into country Z. Thus the crucial factor in the initial approach was the question asked (i.e. the risk to be assessed). Either approach was perfectly valid dependent upon the risk to be assessed. The exact question asked depended on the underlying reasons for the risk assessment being undertaken and was likely to be a risk management decision, although input from the risk assessor was frequently sought. Question 5. Australia contends that one cannot compare the risk posed by other aquatic animals or products hosting any or all of the same disease agents with those posed by uncooked salmon in the absence of a risk assessment for those other aquatic animals or products. What is your view? 6.76 Dr. Burmaster referred to his answers to Question 4, above. 6.77 Dr. Wooldridge recalled her answer to question 4, indicating that there were circumstances where it was valid to compare the risk posed by other aquatic animals or products hosting any or all of the same disease agents with those posed by uncooked salmon, even in the absence of a risk assessment for those other aquatic animals or products. In any risk assessment, the more information available the less uncertainty in the final result. A full risk assessment for these other products might well markedly reduce uncertainty in the risk assessment under consideration; less than a full risk assessment was likely to reduce uncertainty by a lesser amount. Nevertheless such information as was available might be adequate to reduce uncertainty to negligible levels for a given situation under consideration; much depended upon the level of similarity. Such information should not be ignored. Question 6. Canada contends that for any given disease agent, the consequences of the disease becoming established in an importing country would be the same regardless of the original imported source. In your view, is this statement technically/scientifically correct? 6.78 Dr. Burmaster responded that the statement was correct; he could not think of a counter-example to this principle. 6.79 Dr. Rodgers replied that this statement was probably generally correct, although there were degrees of severity for some disease agents and the consequences of disease establishment could vary depending on the nature of the pathogen and the genetic basis of the indigenous species. The consequences might be limited depending on the efficiency of the responsible monitoring service and subsequent early detection. The presence or absence of a susceptible host species and unfavourable local conditions might also limit the spread, since the death of only a few fish could go unnoticed. The costs of control or eradication would vary too for the same reasons. The nature of the original import would have a bearing on the possible consequences because the risks for dead (processed) product would be different compared to that from live fish imports. The economic consequences might also be different, since there would be a potential negative effect on the local industry following the introduction of highly competitive cheaper fish product from a guaranteed aquaculture source. On the other hand, a finite resource, such as wild ocean-caught fish, might have less impact because it might not be able to compete as aggressively in terms of cost or quality. This of course might occur the other way round depending on the overheads attached to bringing each commodity to the market. 6.80 It was also relevant to consider that factors other than the presence of a pathogen were important in the consequences of disease entry, because some diseases were more serious in certain areas than in others (e.g. infectious haematopoietic necrosis (IHN), V. ordalii and Piscirickettsia salmonis). On the other hand, certain diseases only occurred in a limited area, such as infectious salmon anaemia (ISA) in Norway, Oncorhynchus masou virus in Japan, epizootic haematopoietic necrosis (EHN) virus in Australia and Ceratomyxa shasta on the Pacific northwest coast. It might be that these diseases could not establish themselves in other regions because the necessary combination of environmental factors, susceptible species and husbandry practices did not exist elsewhere. The opposite might also be true in that they could become very successful at establishing themselves in a new environment and be extremely virulent. The reason for their current "containment" might rely partly on efficient monitoring and subsequent movement controls and partly on the fact that development of aquaculture with the exact conditions necessary for their establishment had not yet occurred in other areas or countries. 6.81 Dr. Wooldridge replied that once a given disease was established country-wide in an importing country, the consequences from that point on would be the same whatever the original imported source or manner of establishment. However, any regional variation in disease establishment might conceivably affect the consequences, in particular the short to medium term consequences, and regional variation in disease establishment might be source-related. Long-term consequences would converge if the disease then became established country-wide. 6.82 This could be illustrated by assuming that Product A, containing agent K was imported into region W of country Z, but the product had no further widespread distribution in that country. Region W contained no susceptible animals of economic importance, although it did contain susceptible animals of no economic importance. One possible consequence was that disease K became established locally in Region W of importing country Z. In this scenario there were short-term biological consequences but no short-term economic consequences. Given these circumstances it might also be an undetected consequence. 6.83 To continue the example, one could assume that Product B, a different source but also containing disease agent K, was imported only into Region X of country Z, a region where there were susceptible animals of economic importance. Local establishment of disease K would then most probably result in both short-term biological and economic consequences. That is, given a different product source with a different import distribution, some of the short-term consequences of disease establishment would be different because the manner of initial establishment was different. Whether, given time, the disease would then spread from either region to become established country-wide would depend upon many country specific factors including, for example, animal migration patterns, agent virulence, climate, geographical barriers, human working patterns and so on. Therefore, when using information derived from the importation of other products or products from other sources in a current risk assessment, information on the ubiquity and history of that comparison import was of particular relevance. The distribution and transmission of fish diseases Question 7. Which disease agents (or strains of disease agent) have been identified as present in Canadian (i) adult, wild, ocean-caught Pacific salmon; (ii) adult wild freshwater-caught Pacific salmon; (iii) adult Pacific Salmon cultured in seawater on the Pacific coast; (iv) adult Atlantic salmon cultured in seawater on the Pacific coast; (v) adult Atlantic salmon cultured in seawater on the Atlantic coast? Please comment on the responses provided by Australia and Canada in this regard and in particular, the discrepancies in the answers provided by parties (7 October submissions "Responses to Questions", Question 2). In your view, what degree of scientific confidence is there of detecting the occurrence of diseases or disease agents given existing methods of monitoring and surveillance? 6.84 Dr. Rodgers responded that some diseases reported to occur in the above categories include: (i) adult wild ocean-caught Pacific salmon: bacterial kidney disease (BKD), Kudoa thyrsites, Parvicapsula sp., plasmacytoid leukaemia (marine anaemia); (ii) adult wild freshwater-caught Pacific salmon: BKD, Ceratomyxa shasta , enteric redmouth disease (ERM), flexibacteriosis, furunculosis, infectious haematopoietic necrosis (IHN), Loma salmonae, plasmacytoid leukaemia (marine anaemia), proliferative kidney disease (PKD); (iii) adult Pacific salmon cultured in seawater on the Pacific coast: BKD, Parvicapsula sp., piscirikettsiosis; (iv) adult Atlantic salmon cultured in seawater on the Pacific coast: IHN, pancreas disease (PD), piscirikettsiosis, viral haemorrhagic septicaemia (VHS): (v) adult Atlantic salmon cultured in seawater on the Atlantic coast: BKD, ERM, furunculosis, infectious salmon anaemia (ISA), Vibrio salmonicida. Dr. Rodgers noted that he had not been able to undertake a complete literature survey for each disease in each of the above categories due to the time available to respond to the questions. Neither had it been possible, for the same reason, to solicit additional information, in the form of personal communication, directly from other sources. Consequently, the above list of diseases was not exhaustive and it had not always been possible to identify which maturity state any particular disease had been isolated from (i.e. juvenile or adult). In addition, the identification of a particular disease agent did not imply that it occurred regularly, since only an occasional occurrence may have been reported in the literature. Other diseases, such as some parasitic (e.g. protozoal, etc.) or bacterial (e.g. vibriosis) conditions, occurred on a more global scale and would be expected to be ubiquitous. 6.85 There were discrepancies in the responses from both countries in regard to the question concerning the presence of certain disease agents in Canadian salmon. The answer from Canada was more detailed in the sense that it further subdivided the categories of salmon into the above five groups. As a result, this was more accurate, useful information and consequently had probably led to the majority of differences between the two responses. For instance, the presence of furunculosis in wild-caught Pacific salmon was acknowledged by both countries but Canada qualified the answer by stating that it did not, however, occur in ocean caught fish, only in freshwater caught fish and cultured stocks. The same was true for enteric redmouth disease, which Canada contended did not occur in ocean caught Pacific salmon, nor in farmed salmon (Atlantic and Pacific species) on the Pacific coast. 6.86 In other responses Australia indicated unknown or uncertain presence, due to lack of definitive scientific work or difficulty in detection, by the use of a question mark (i.e. Yes? or No?). In all these cases, Canada supplied a definitive "yes" or "no". The main discrepancies related to conflicting responses concern the presence of vibriosis (V. anguillarum and V. ordalii) in wild-caught Pacific salmon, Pacific salmon anaemia virus (EIBS) in Pacific salmon, viral haemorrhagic septicaemia (VHS) in farmed Atlantic salmon, proliferative kidney disease (PKD) in wild-caught Pacific salmon and infectious salmon anaemia (ISA) in farmed Atlantic salmon on the east coast. Canada specifically stated that EIBS and PKD were not known to occur in any of the five categories of salmon. Although there were a few reports in the scientific literature of these two diseases occurring in Pacific salmon in Canada, it was most probable that they referred to juvenile fish rather than adults and consequently the statement that they did not occur in adult fish would be true. As regards vibriosis in wild-caught Pacific salmon, the condition was likely to be widespread in the marine environment but in this particular case it was not clear which specific references had been used to prove or disprove this point. The occurrence of VHS in Canadian Atlantic salmon, which was not mentioned in the Australian response, represented 1995 published information which presumably the Australian case was not aware of at the time. The presumptive diagnosis of ISA in Canadian farmed Atlantic salmon was verbally reported in an international conference of fish pathologists in September 1997, but the current status of ISA was not listed by Canada in its responses to questions. However, this was perhaps not surprising since it was an addition to the Australian Final Report (in place of Kudoa thyrsites) and appeared for the first time in the responses to panel questions. Therefore, it was most likely that Canada was unaware that it had to comment on the current status of this disease. 6.87 Dr. Rodgers observed that there was much reliance in both submissions on personal communications with respected fish pathologists and these should be accepted at face value, since it was not possible to corroborate all this information within the current time frame. However, such information was usually based on more recent data and was usually unpublished, although it could contain an element of subjective opinion and assumption. The Canadian information was broken down into more relevant groups, since they had greater access to data from research projects, veterinary reports and monitoring and surveillance programmes, which were not always readily available to the general scientific community. Consequently, this should represent the most accurate up-to-date picture for the diseases of concern, since information in the published scientific literature could be quickly out of date and many historical published reports were not followed up by subsequent more recent studies. 6.88 The degree of scientific confidence in disease detection with existing methods of monitoring relied on the lowest limits of detection for each specific test or series of identification methods. Currently, there were some difficulties with the isolation of some fish viruses (e.g. PD, ISA, etc.) or bacterial strains (e.g. the lengthy incubation time for Renibacterium salmoninarum , the causal agent of BKD). In addition, there was a cut-off sensitivity point for most diagnostic methods which leads to the carrier state being very difficult to detect, except with the most sensitive of tests (e.g. PCR). However, even these tests had limits of sensitivity, which were albeit very low, and they did not always rely on culture as a prerequisite for identification. Nevertheless, these deficiencies in sensitivity for existing methods were generally accepted in terms of both their supportive science (pending improvements) and the legislative policies that stipulated their use in regular surveillance programmes. 6.89 Regardless of the reason for carrying out a disease examination there was a need for some form of standardized sampling methodology which would allow a reasonably accurate assessment of the health status of a particular stock to be made. While local conditions in respect of the cultured or wild species and the presence of specific diseases might vary considerably, the principles of such a sampling methodology tended to remain constant. For a correct disease diagnosis it was often necessary to demonstrate the presence of the pathogen by culture, followed by some form of confirmatory serological test. However, there were certain measures that could be employed to increase the chances of detecting the presence of a pathogen. For clinical outbreaks it was usually prudent to sample moribund fish exhibiting clinical signs but at the same time selecting certain target tissues. Testing fish with the minimum or absence of symptoms (e.g. the "asymptomatic carrier" state) would normally be biased towards any fish with possible clinical signs, selecting the correct age of fish (e.g. at least one-month old fry in the case of infectious pancreatic necrosis (IPN) or screening five-month old fish for whirling disease) and sampling during periods when the disease was most likely to be detected (e.g. February-April in Europe for BKD). Examination of fish with no suspicion of disease for screening purposes would also consider these factors but in addition would sample a sufficient number of fish for effective sampling. This number was usually based on the work of Ossiander and Wedemeyer (1973), as stipulated in the OIE Diagnostic Manual for Aquatic Animal Diseases. This study indicates the minimum sample size for each lot of fish that provides a 95 per cent confidence of including infected specimens in the fish sampled, assuming a minimum prevalence of infection equal to or greater than 2 per cent, 5 per cent or 10 per cent. For instance, 150 out of 100,000+ fish need to be sampled at the 2 per cent level. This is reduced to 30 out of 100,000+ fish at the 10 per cent level. Testing for statutory purposes was generally conducted at the 95 per cent level of confidence with an assumed incidence of disease of 2 per cent. Occasionally, there might be justification for increasing the level of confidence from 95 per cent to 99 per cent (i.e. 225 fish at the 2 per cent level) for a particular examination. However, any subsequent increase in sample size gives a negligible increase in the statistical probability of finding an infected specimen. 6.90 Despite attempts by various international organisations, such as EIFAC, FAO and OIE, to discuss the possibility of producing a uniform system of monitoring, none really existed on a world-wide scale. Individual countries (e.g. Australia, Canada and the United States) or whole trading blocks, such as the European Union, had developed their own baseline rules relating to importations and these were well formed for those countries which had a heavy involvement in aquaculture or had an important sport fishery. In this context the receiving country or trading block tended to set the conditions of importation. 6.91 Dr. Rodgers noted that one other aspect to consider was that regularly tested stocks were normally considered as a lesser risk than occasionally, or untested stocks or products, since regular monitoring provided a background database of information over time. For instance, it would be considered that ova from wild salmonids would present a greater disease risk than ova from farmed stocks since the latter could be tested regularly throughout their lifetime. Estimates of prevalence data were usually based on surveillance information obtained from fish health authorities in potential exporting countries but such data were generally limited to those countries that had statutory monitoring in place. This type of data, however, was actually designed to overestimate the prevalence of infection in commercially harvested fish, since the monitoring programme targeted fish from which an infectious agent was most likely to be isolated, such as spawning fish, fry or fingerlings and fish exhibiting signs of infectious disease, as mentioned above. Unfortunately, this was rarely the case for wild populations of fish because regular monitoring programmes did not normally exist, unless diagnosis was related to the occurrence of large, noticeable mortalities. However, sampling returning anadromous salmonids in their freshwater phase was occasionally the exception. Post-harvesting testing for disease in fish destined for human consumption was very rarely undertaken. Question 8. What aquatic animals, other than salmonids, are known to be carriers of any, several or all of the disease agents subject to this dispute? Please comment on the responses provided by Australia and Canada in this regard (7 October submissions "Responses to Questions", Question 3). Would the epidemiological factors relevant to disease transmission be the same in all cases? 6.92 Dr. Rodgers replied that it had not been possible to undertake a comprehensive literature search to determine which aquatic animals, other than salmonids, were known to be carriers of any, several or all of the disease agents subject to this dispute. However, an examination of the lists provided by both Australia and Canada indicated that they were comprehensive and authoritative. The list from Australia tended to be more general in content and used phrases such as salmonids, cyprinids or a wide range of salmonids, whereas the list from Canada was more detailed since it listed the relevant species separately. However, the Canadian list was incomplete because it did not deal with all the diseases, although the original question concerned "... any, several or all of the diseases ..." and there did not seem to have been a specific requirement to detail all species for all the diseases. Nevertheless, since the Canadian list did not include Piscirickettsia salmonis, Renibacterium salmoninarum (BKD), Vibrio ordalii, herpesvirus salmonis type 1, Pacific salmon anaemia virus, salmon leukaemia virus, salmon pancreas disease virus (PD), Enterocytozoon salmonis, Loma salmonae, Ceratomyxa shasta, Henneguya salminicola, Myxobolus cerebralis, Parvicapsula spp., proliferative kidney disease (PKD), rosette agent and infectious salmon anaemia (ISA), the Australian list was more useful in this respect. Neither list included Kudoa thyrsites, which to all intents and purposes was widely distributed and had a broad host range. 6.93 Disease transmission combined many biological, behavioural and environmental factors that were interrelated. The epizootiological factors relevant to disease transmission would not necessarily be the same for each disease, since they were complex and numerous, although the general aspects were common. Fish diseases manifested themselves as a result of a variety of circumstances, such as genetic background, nutrition, stress, injury or cohabitation. The relationship between disease epizootics among fish and fish stocking densities was important, since overcrowding of fish was often followed by infections of opportunist bacteria, fungi or parasites. Non-infectious disease that occurred as a result of fish culture mis-management could also be prevalent in these circumstances. Some fish pathogens were a constant and natural part of the environment, usually without causing disease problems and mortality (e.g. certain parasites). There was a unique relationship between fish, their pathogens and their environment which meant that a balance normally existed between the three factors, with the fishÿs immune system playing an active role in maintaining this balance. However, if there was an alteration in one or more of the environmental characteristics then there might be a shift in the balance to the benefit or detriment of either the fish or the pathogen. Another relevant disease-causing factor was the introduction of a potential pathogen into an already stable environment. The introduction of different types, strains or novel pathogens could upset the natural balance of a fish population. Indigenous fish developed a degree of immunity to relevant pathogens in their environment but they may never have encountered new strains. 6.94 The reasons for a disease outbreak were varied, representing complex interactions between the host and the disease-causing situation. In the cases of host-pathogen interactions, the onset of disease represented a decrease in the resistance of the host such as during reproductive stages or due to environmental stress, poor husbandry conditions, in conjunction with an increase in numbers and/or virulence of the pathogen. However, it was not clear what was the maximum number of cells which constituted a genuine infection, since there was a big difference in the number of cells required to establish clinical disease in aquatic animals. This might range from 1 (e.g. crayfish plague) to 100 (e.g. furunculosis) or more (e.g. BKD). 6.95 Dr. Rodgers stated that it was generally accepted that four major groups of disease might be identified in terms of epizootiology. These included sporadic diseases, which occurred sporadically in comparatively few numbers of a population; epizootics, which were large-scale outbreaks of communicable animal disease occurring temporarily within limited geographical areas; panzootics, which occurred over large areas; and enzootics, which persisted or re-occurred as low level outbreaks in certain areas. 6.96 It had been suggested that dense populations of fish would maintain a given level of diseased individuals, regardless of whether the populations were shoals in the sea or aquacultured stocks. However, there was a generally accepted lack of information about the occurrence of disease in wild fish. Most studies related to the potential effects of pollution on free-living marine fish and the incidence of disease in returning anadromous salmonids. Much less information was available on disease transmission between wild and aquacultured fish and vice versa. To continue with Australia - Measures Affecting Importation of Salmon: Question 9

Agreements  |  Disciplines | Trade Policy Developments | Countries | Trade & SMEs Home | What's New | SiteMap | Resources | Search | Disclaimer