Memorandum submitted by the Centre for
Environment, Fisheries and Aquaculture Science (CEFAS) (M12)
EXECUTIVE SUMMARY
1. CEFAS was appointed in 2001 on a four-year
contract to carry out the statutory monitoring programme for shellfish
toxins in England and Wales. This involves testing for toxins
in the flesh of all commercially harvested species, using chemical
assays and a mouse bioassay.
2. Atypical results were observed in tests
for Diarrhetic Shellfish Poisoning (DSP), principally in cockle
samples, and were also observed in Northern Ireland. This problem
coincided with the extension of cockle testing in the surveillance
scheme in 2001, and was drawn to the attention of the FSA in October
2001.
3. Scrutiny of procedures at CEFAS and the
testing laboratories in Scotland and Northern Ireland identified
variations that could affect the likelihood of detecting toxic
responses in the mouse test. We have worked with the Food Standards
Agency, the UK National Reference Laboratory (NRL) and the other
testing laboratories on a unified methodology to ensure comparability.
We also took part in exercises designed to reveal whether atypical
results were due to methodological problems.
4. The independent Makin audit in July 2003
examined the operation of an interim Standard Operating Procedure
(SOP) for the DSP test. We responded to Professor Makin's seven
conclusions about procedures at CEFAS by introducing changes to
local practice and by adopting the unified UK SOP. Subsequent
testing of cockles continued to produce unusual symptoms of intoxication,
though less frequently.
5. There is no evidence to indicate that
atypical toxicity is an artefact of the testing methods or the
way that they are applied. Investigations so far have failed to
attribute the results to the presence of any known toxin.
TOXINS WORK
AT CEFAS
6. Toxins work began at CEFAS (then the
Directorate of Fisheries Research) in 1968 and CEFAS scientists
provided a monitoring programme for Paralytic Shellfish Poisoning
(PSP) for the UK until the early 1990's. Research was conducted
in order to develop analytical alternatives to animal tests, including
High Performance Liquid Chromatography (HPLC) and Liquid Chromatography—Mass
Spectrometry (LC-MS). We also collaborated with Universities to
improve understanding of the environmental triggers for toxic
algal blooms and to provide tissue culture techniques.
7. During most of the 1990's the national
monitoring programme was operated from Aberdeen (the Torry Laboratory,
and subsequently Fisheries Research Services, FRS). CEFAS took
on testing for England and Wales in June 2001 when the Food Standards
Agency let the contract in open competition. In preparation for
this CEFAS gained UKAS accreditation to ISO 17025 for biotoxins
analysis.
8. CEFAS is currently responsible for an
annual programme of testing shellfish taken from harvesting areas
in England and Wales. The contract specifies the range of toxins
to be monitored (PSP, DSP and Amnesic Shellfish Poisoning), the
test methods to be used, the number of samples to be analysed,
the reporting requirements and the quality assurance requirements.
Testing for PSP and DSP is based on protocols for the chemical
extraction of material from shellfish flesh and testing toxicity
on mice, in accordance with the requirements of the European Shellfish
Hygiene Directive (91/492/EEC). Similar responsibilities apply
to FRS in Scotland, and the Department of Agriculture and Rural
Development (DARD) in Northern Ireland.
9. We continue to engage in research, including
collaboration with Irish and Canadian scientists, and recently
using LC-MS we have identified a number of toxins new to the UK
present in coastal waters. We have also developed a screening
method for toxins using insects, which may contribute to reduction
in the use of the mouse test.
OCCURRENCE OF
ATYPICAL RESULTS
10. The biotoxin monitoring programme in
England and Wales was expanded in 2001, to cover all harvesting
areas and all shellfish species. This followed an inspection of
the UK implementation of shellfish hygiene controls by the European
Union Food and Veterinary Office in 1999. Their report criticised
the programme in England and Wales for incomplete coverage.
11. As a consequence, the scale of DSP analysis
increased considerably at the time that CEFAS took over the monitoring.
CEFAS started to observe an unexpectedly high number of positive
results in DSP tests on cockles, associated with unusual clinical
symptoms in mice, suggesting that the results were not due to
classical DSP. The DARD laboratory in Northern Ireland reported
similar findings in cockles. In October 2001 CEFAS wrote to the
FSA detailing these unexpected findings.
12. The cause of the problem has been investigated
by looking for known toxins, by examining whether toxicity could
be an artefact of the method, and by reviewing the quality of
the work carried out by the testing laboratories.
ABSENCE OF
KNOWN TOXINS
13. Initially it was thought that the cause
might be the emergence of a new toxin, Azaspiracid, which had
recently been recognised in the Republic of Ireland and had caused
human health incidents in the UK. Laboratory investigations revealed
traces of azaspiracid in a few samples, but insufficient to explain
the problem.
14. Subsequent investigation funded by the
FSA demonstrated that the toxicity findings could be independently
replicated by an international DSP expert—Professor Yasumoto
of Japan. However neither we, nor others, were able to document
by LC-MS the presence of any known DSP toxins in samples showing
the atypical response. We were similarly unable to find any correlation
of cockle toxicity with high concentrations of known toxic phytoplankton
or zooplankton species in the water column. Cockle extracts were
sent to international experts specialising in particular toxins
but they too were unable to determine the presence of any known
algal toxin.
15. In order to explore other possibilities,
cockle extracts were tested for unusually high levels of trace
metals, heavy metals and other substances (free fatty acids) that
might have caused interference in the mouse test, but concentrations
were well below those that would have toxic effects. As yet, no
specific agent extracted from cockle flesh has been identified
as the cause of the atypical response.
EXTRACTION AND
TESTING METHODOLOGY
16. The extraction of DSP toxins from shellfish
tissue is a simple procedure involving acetone extraction, centrifugation
or gravity filtration, partition to ether, water wash of the ether
fraction, evaporation, take up of the residue in surfactant and
injection to a mouse. This procedure has been carried out regularly
on shellfish other than cockles, without producing frequent atypical
results. Clearly there is something specific in cockle samples
that causes the problem.
17. Between June and October 2003 the testing
laboratories moved towards a common "interim" extraction
protocol for DSP, leading to the development and adoption in November
2003 of a unified UK SOP specified by the NRL. This protocol has
elements of all approaches, and incorporates new requirements
to ensure minimal solvent carry-over into final extracts. An optimised
UK protocol is being developed by the Central Science Laboratory
(CSL), focussing on the extraction and particle removal stages
of the procedure.
18. The industry has postulated that carry-over
of solvents into the final extract might cause the atypical results,
possibly involving synergism with histamines derived from cockles.
We confirmed that effects were not specific to the use of di-ethyl
ether as the partitioning solvent—positive results were
also seen when dichloromethane was used. It is not possible to
completely remove all traces of solvent, but the methodology introduced
in the unified UK protocol is designed to ensure that there is
minimal carry-over. Monitoring of solvent levels using Gastec
is now routine, and sample extracts are used only if no solvent
is detected. In addition, CSL made measurements of solvent levels
in CEFAS, DARD and FRS extracts before the introduction of the
new protocol. This showed that some high concentrations were found
in samples that gave no mouse response, whereas some samples that
produced an atypical response had low concentrations. This demonstrates
that there is no simple link between solvent carry-over and toxicity.
19. Details of the methods used for acetone
extraction and removal of solids are likely to affect the range
of compounds extracted from samples. The current unified UK SOP
uses two acetone extractions and a gravity filtration step, whereas
the previous CEFAS/DARD protocols employed two extractions with
acetone followed by centrifugation, while FRS made one acetone
extraction followed by filtration. Nevertheless, all protocols
derived from the published method of Yasumoto 1984 as required
by European legislation.
20. Sequential extractions with acetone
draw out a wider polarity group of compounds than a single extraction,
while compounds loosely bound to the shellfish matrix are more
likely to be recovered by centrifugation than by gravity filtration.
Therefore, the previous CEFAS/DARD method would be expected to
extract additional material than the FRS method, and the unified
UK SOP might be expected to extract fewer compounds than the CEFAS/DARD
methods but more than the FRS method. Results obtained since the
change in methods are consistent with this prediction. Both CEFAS
and DARD have reported a much lower incidence of atypical symptoms
of intoxication,and it seems that the concentration of whatever
is responsible for the mortality is now much lower in the extract
from cockles. Conversely in Scotland there are preliminary indications
of an increase in the reporting of (typical) symptoms in the DSP
assay.
QUALITY ASSURANCE
IN CEFAS
21. At CEFAS, quality assurance for the
biotoxin assays is covered by internal systems and by UKAS accreditation
to the international standard ISO 17025. This entails:
— A Quality Manual identifies the individuals
responsible for specific tasks and trained in their use.
— A series of Standard Operating Procedures
describe the tasks in a clear, step by step approach. SOPs are
written to be fit for purpose and carried out by competent trained
staff.
— An internal audit team led by scientists
on a different site perform regular checks that the content of
the SOPs is aligned with scientific practice and examine training
records. There have been 20 internal audits covering toxin work
in three years.
— An external audit team from UKAS
perform regular checks that the testing performed is of a quality
consistent with ISO 17025. All aspects of procedures are audited
and testing is witnessed in practice in the laboratory. UKAS report
on performance and require non-conformities to be addressed within
a set timetable. There have been six audits of toxin work in the
past three years, which confirmed that the quality of CEFAS work
complies with international standards.
— A Home Office audit team carry out
inspections on the animal house and ensure that practice is aligned
with the terms of Project Licenses.
— Standardisation is overseen by the
National Reference Laboratory to ensure that the science carried
out in the UK is aligned with practice dictated by the Shellfish
Hygiene Directive.
22. In addition, a special audit of all
three testing laboratories was organised by the FSA in the summer
of 2003 to address concerns about the atypical DSP test results.
The report by Professor Hugh Makin included a set of overall findings
and recommendations, and seven conclusions specific to his observations
of procedures at CEFAS. We have responded to these by introducing
changes to local practice, and by collaboration with the NRL,
the FSA and the other testing laboratories to introduce the unified
UK SOP. Some appropriate changes (standardisation of the assessment
of toxic symptoms, and the use of positive and negative controls
in the mouse test) await further discussion between the FSA and
the Home Office to take account of animal welfare issues.
23. Of Prof. Makin's specific conclusions
relating to CEFAS systems, three relate to the style and authorship
of SOPs. While this identifies scope for improvement, there are
no implications for the integrity of the data produced. Another
conclusion refers to the risk of potential overload if large numbers
of samples exceed the laboratory's capacity. This has never happened,
and has no relevance to the occurrence of atypical test results.
24. The remaining three conclusions could
potentially influence results. One relates to the lack of negative
"controls" (ie extra tests on shellfish material known
to be free of toxin) and positive controls (ie a "clean"
sample to which a standard amount of toxin is added) in the DSP
mouse test. We agree that such controls are important in routine
assay work. In this case, there are practical difficulties, particularly
in relation to the detection of unknown toxins, and the Home Office
does not currently permit the use of additional animals for control
tests. This issue remains under discussion.
25. However, substantial alternative evidence
provides some assurance that the testing reliably detects DSP.
First, the method has been developed through research involving
experimental controls. Second, a large majority of samples of
most species in the monitoring programme give negative results,
which demonstrates that the method itself does not systematically
cause symptoms of toxicity. Third, where sample size allows, typical
DSP responses are confirmed by chemical analysis using LC-MS,
showing that the method is capable of detecting known DSP toxins.
26. The sixth conclusion deals with the
presence of fluid in the final extracts, which was also observed
during the audit of DARD. It is well recognised that extracts
do not always evaporate to dryness, and the risks of solvent carry-over
have been fully addressed in the new UK protocol, as described
above.
27. Finally, he documented a list of deviations
between written procedures and practices observed on the occasion
of the audit visit. Only a few of these were reported by internal
auditors or by UKAS. Many were a result of the wording of the
SOPs, or delays in updating the documents, rather than scientifically
inappropriate practices. The remainder were points of specific
detail that could not have caused significant errors in the results.
Nevertheless, we have taken steps to ensure that written procedures
and actual practice now correspond exactly.
28. Therefore, we believe that deviations
from current protocols are minor, and in the context of the Makin
report, have no significant implications regarding the scientific
conclusions reached on causation of the atypical DSP results.
Prof Makin concluded that no evidence emerged that the atypical
response is a methodological or procedural artefact.
IMPROVEMENTS TO
CURRENT METHODOLOGY
29. In addition to developing the unified
UK SOP for the current method, it is desirable to move towards
approaches that do not involve the mouse test, although there
are constraints on how easily changes could be introduced.
30. An amendment to the legal framework
(Commission Decision 2002/225/EC) clears the way for implementation
of non-mammalian test methods for DSP. It defines a number of
toxins associated with DSP and lays down maximum permitted levels
for each toxin in shellfish. The mouse bioassay is specified as
the reference method but alternative or complementary methods
are permitted providing they detect all specified analogues and
provide an equivalent level of health protection. Important caveats
are that standards must be available before chemical analysis
is possible and that methods should be validated according to
international protocols. Significantly, where test results are
discrepant between methods, the mouse test shall be considered
to give the definitive result. Thus it might be possible to reduce
the scale of mouse testing, if an alternative was used as a preliminary
screening assay. To completely replace the mouse test would require
amendments to the existing legislation.
31. The most advanced alternative is liquid
chromatography—mass spectrometry (LC-MS). A significant
obstacle to its wide scale adoption has been the unavailability
of a full suite of standards for the analogues specified in Commission
Decision 2002/225/EC. However, standards are now, or will shortly
become available, for representative analogues from each group
in the DSP complex. This is sufficient coverage to begin validation
of LC-MS as an alternative assay. Also, LC-MS could be considered
as a primary screen with positive samples tested using the mouse
test or another biological assay for confirmation.
32. LC-MS alone would not be capable of
detecting unknown toxins. Other biological assays are therefore
required. Options include cell culture, enzyme reactions, and
tests on whole animals. Invertebrates are more acceptable and
more tractable for standardised toxicity test protocols than mammals.
In some cases the biochemical modes of action are known to be
similar, and insects have been shown to be good surrogates for
assessing certain effects on human health. CEFAS has been exploring
the use of an assay for PSP and DSP using the American Cockroach
that would also be capable of detecting the effects of unknown
toxins. This work has produced encouraging preliminary results,
soon to be published.
January 2004
| 
