Microsoft Word - WHO_CDS_EPR_GIP_2006

© World Health Organization 2006

The designations employed and the presentation of the material in this publication do not imply the
expression of any opinion whatsoever on the part of the World Health Organization concerning the legal
status of any country, territory, city or area or of its authorities, or concerning the delimitation of its
frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may
not yet be full agreement.

The mention of specific companies or of certain manufacturers’ products does not imply that they are
endorsed or recommended by the World Health Organization in preference to others of a similar nature
that are not mentioned. Errors and omissions excepted, the names of proprietary products are distin-
guished by initial capital letters

All reasonable precautions have been taken by the World Health Organization to verify the information
contained in this publication. However, the published material is being distributed without warranty of
any kind, either express or implied. The responsibility for the interpretation and use of the material lies
with the reader. In no event shall the World Health Organization be liable for damages arising from its
use.

Contents

1 Introduction

2 Executive

summary

Recommendations

5 Detection and diagnosis of human H5N1 infections

Serological methods
Seroprevalence studies
RT-PCR
Development of a rapid field test
Detection and diagnosis of human infections in Indonesia
Discussion

8 Protecting humans: vaccines and antiviral drugs

Vaccines
Distinct genetic groups of circulating viruses
Status of vaccine development
Clinical trials
Use of pre-pandemic vaccines
Resistance of H5N1 viruses to antiviral drugs
The quest for new drugs
Discussion

10 Surveillance in birds and other animals: assessing the coming risks

H5N1 surveillance in Europe
Laboratory testing for European outbreaks
Recent evolution of the H5N1 virus in poultry in China
Lessons from recent poultry outbreaks
Discussion

14 Deciphering the virulence and pathogenicity of H5N1 infections in

humans

H5N1 pathogenesis
Avian-like receptors in the human lung
Disease severity
Determinants of virulence and transmissibility
Host range of avian influenza viruses
Studies on animal H5N1 viruses in Australia
Pathogenesis in the duck
Discussion

17 List of participants

WHO working group on influenza research at the human and animal interface

Influenza research at the

human and animal interface

Report of a WHO working group

From 21 to 22 September 2006, WHO convened a working group of 22 laboratory
directors and senior scientists leading research on influenza at the human and animal
interface. The researchers included directors from some of the laboratories in the
WHO H5 reference network, scientists from veterinary medical institutes, and
virologists and microbiologists in countries affected by outbreaks. The research
discussed represented the work of scores of scientists, students and technicians in the
various laboratories over the last few years. The attending scientists were specifically
asked to interpret their latest research in terms of its implications for public health
policy. Although several avian influenza viruses were considered, emphasis was
firmly placed on what is currently known about human infections with the H5N1 virus
and the presence of this virus in poultry, wild migratory birds, and other animals. At
the same time, however, participants recognized that the next pandemic might well
arise from another virus subtype; surveillance at the animal and human interface
should not be restricted to H5N1 viruses.

Discussion focused on four main topics: methods for the detection and diagnosis of
human infections, the use of vaccines and antiviral drugs to protect humans, current
findings from animal surveillance in countries and regions with recent outbreaks, and
factors governing the virulence and pathogenicity of H5N1 viruses. Issues explored
ranged from explanations for the severity of this disease and its tendency to affect
younger people, through the role of migratory birds in virus spread, to the possibility
that genetic factors might influence transmissibility of the virus among humans.
Issues relating to control, including diagnostic limitations in the detection of human
cases, vaccination policies in poultry, and the identification of avian species that act as
vectors for maintaining virus transmission, were also critically assessed. Throughout
the meeting, repeated reference was made to the added complexities arising from the
recent divergence of H5N1 viruses into several distinct genetic groups that are now
circulating in different parts of the world.

Discussions took place within the context of knowledge about the epidemiology and
ecology of influenza A viruses in avian and other animal species that has been
accumulating for more than 40 years. Some of the scientists who pioneered this
research were present. Their perspective on developments over the past decades
helped the group to pinpoint unusual or unprecedented features of the current disease
situation. In the past, research at the human and animal interface has nearly always
been crisis driven; the present severe crisis with H5N1 infections similarly brings a
need for cohesion and urgency in collaborative research efforts. Information was
generously exchanged during the meeting. Evidence presented indicates that the
H5N1 virus is still evolving in animals and humans; much about the disease it causes
remains poorly understood. Nonetheless, the group had little difficulty in agreeing on
the most pressing research needs. It was further acknowledged that the seriousness of
the present situation, including the risk that a pandemic virus might emerge, is not
likely to diminish in the near future.

WHO working group on influenza research at the human and animal interface

Executive summary

The technology for diagnosing human H5N1 infections is mature, but many tests are
complex, some are liable to error, and some can be performed safely only in
biosafety level 3 facilities. A simple, rapid, robust and reliable test, suitable for use
in the field or at the patient’s bedside, is urgently needed.

In humans, much recent research has focused on the factors responsible for the
pathogenicity and transmissibility of the H5N1 virus. Several lines of evidence
suggest important roles for the polymerase genes, though no single gene has yet
been implicated and several genes may be working in tandem. Nor can the
distinctive age profile of this disease be adequately explained at present. A genetic
predisposition for infection is suspected based on data from rare instances of
human-to-human transmission in genetically-related persons. This possibility, if
more fully explored, might help explain why human cases are comparatively rare
and why the virus is not spreading easily from animals to humans or from human to
human.

The development of a pandemic vaccine has become more difficult following the
divergence of circulating viruses into distinct genetic and antigenic groups. To date,
results from clinical trials of candidate pandemic vaccines have not been promising,
as these vaccines confer little protection across the different genetic groups.
International standards, or “benchmarks”, for evaluating the efficacy of vaccines are
urgently needed. Integrated studies of sera from individuals being vaccinated in the
various clinical trials would be equally useful – for industry as well as for national
authorities.

Monitoring for virus resistance to antiviral drugs needs to continue. Although
resistance to amantadine is now widespread, the possibility exists that these resistant
strains may be replaced by fully susceptible strains as the virus continues to evolve.
Innovative work on novel strategies for drug development was welcomed by the
participants, but new drugs will not be on the market for some time to come.

The global picture of influenza viruses in the avian world has changed significantly
since 2002. The massive die-off of migratory birds at Qinghai Lake in mid-2005
was unprecedented, and migratory birds now appear to be contributing to
geographical spread of highly pathogenic virus. Importantly, evidence was
presented for a change in virus shedding patterns, with increased shedding from the
respiratory tract rather than the cloaca. Thus, for surveillance purposes, a
corresponding change in sampling strategies – including both cloacal and
pharyngeal swabs – is called for to get a true picture of the situation. Furthermore,
domestic ducks and geese – and not chickens – have been identified as the true
vectors of disease transmission in poultry. Recently, studies have demonstrated that
the virus is now moving both ways in relay transmission, from poultry to migratory
birds and back again. This finding might help explain some of the continuing
geographical spread.

Continued widespread infections in poultry were viewed as an important on-going
risk for human cases and the related risk of a pandemic. Participants agreed that

WHO working group on influenza research at the human and animal interface

more needs to be done in the animal sector to control this virus. Countries with
adequate resources should continue to make culling their first-choice control
strategy; experiences in Japan and Korea have shown that such an approach, though
costly and disruptive, can ultimately be successful. In countries with limited
resources, however, the group strongly recommended use of widespread poultry
vaccination with appropriate, high quality vaccines accompanied by appropriate
surveillance to detect possible asymptomatic virus circulation. In a related
recommendation, the group suggested that countries with outbreaks should look at
the factors driving continued circulation of the virus, and then use that knowledge to
develop tailored interventions. Baseline data could be used, for example, to identify
seasons of peak virus activity and this information, too, can guide intervention
strategies. Hong Kong used this approach following the 1997 outbreak and found
that live animal markets were maintaining, amplifying, and disseminating the virus.
Intervention at this critical point eventually freed Hong Kong from the virus. More
recently, Viet Nam introduced a policy of mass poultry vaccination; human cases
subsequently ceased. Poultry vaccination is, however, recognized as having some
limitations as a control strategy, and these limitations need to be addressed on an
urgent basis: chicken immunology is much better understood than duck
immunology; ducks react differently to poultry vaccines, yet vaccines tend to be
approved based on protection in chickens only; production quality control and
antigen content of vaccines are not standardized worldwide and sub-optimal
vaccines have been used; in some countries, not all vaccine manufacturing takes
place under the control of national authorities and vaccine efficacy is not always
monitored.

Recommendations

1. Make the development of a simple, robust, and reliable diagnostic test for use in

the field and at the patient’s bedside a high priority. Facilitate industry’s
development of such a test by providing representative panels of viruses and
addressing relevant issues of intellectual property rights.

2. Publish recommended diagnostic tests and methods for their accurate

performance, including an alert to common pitfalls, on the WHO web site, and
develop a schedule and system for regularly updating tests and kits with
appropriate reagents.

3. Investigate the sensitivity with which currently available diagnostic tests are

capable of detecting mild or asymptomatic infections.

4. Establish benchmarks for evaluating the effectiveness of candidate pandemic

vaccines.

5. Integrate data on antibody responses in persons participating in the various

clinical trials of candidate pandemic vaccines.

6. Determine which (if any) animal model provides the best information on cross-

clade protection among H5N1 variants.

WHO working group on influenza research at the human and animal interface

7. Continue to monitor H5N1 virus strains, in humans and avian species, to

determine changing patterns of resistance to antiviral drugs.

8. Investigate factors that may make children and young adults especially

vulnerable to infection.

9. Conduct studies to determine whether a genetic predisposition increases the

likelihood of human infection or of human-to-human transmission among
genetically-related persons.

10. Address and resolve the ethical issues that arise when DNA banks are

established using specimens from deceased patients, family members, close
contacts, and controls.

11. Develop a single, agreed upon system of nomenclature to describe different

phylogenetic, genetic, and antigenic groups of H5N1 viruses globally.

12. In countries experiencing continuing outbreaks in poultry, conduct studies to

identify the factors driving continued transmission of the virus, and plan
interventions accordingly.

13. When culling is impracticable as a control strategy, introduce a policy of

poultry vaccination, accompanied by systematic monitoring, in the interest of
reducing opportunities for human exposures and infections to occur.

14. Standardize antigen content in poultry vaccines and insist on rigorous quality

control worldwide in line with OIE standards.

15. Monitor virus activity in backyard flocks and live animal markets as well as at

commercial farms.

16. Adjust sampling procedures for ducks in line with changes in the currently

recognized pattern of virus excretion, whereby more virus is now being shed
via the respiratory tract than via faeces.

17. Continue to recognize, for the purposes of surveillance and research on

pathogenesis, the potential role of pigs (or other species) as intermediate hosts
in the generation of pandemic viruses.

18. Enhance international collaboration in the surveillance of wild birds and in the

sharing of data from such surveillance efforts.

19. Improve understanding of migratory routes for wild waterfowl and strengthen

collaborative interactions with ornithologists.

WHO working group on influenza research at the human and animal interface

Detection and diagnosis of human H5N1 infections

Serological methods.

A presentation from the US Centers for Disease Control and

Prevention (CDC) described the state-of-the-art in diagnostic testing, including
laboratory confirmation of H5N1 infections based on serological tests. Such testing
becomes important when poor sampling technique, poor specimen quality, or other
problems rule out isolation of the virus or preclude the use of PCR-based tests. For
example, two cases from the 1997 outbreak in Hong Kong were confirmed on the
basis of serological results alone. Serological tests are also used in studies aimed at
screening for mild or asymptomatic illness in population groups with high potential
exposures to the virus, including family members and contacts of patients, poultry
workers and cullers, and workers in live animal markets.

Serodiagnosis depends on the use of paired sera collected at specified dates during
the course of a patient’s illness, as dictated by the kinetics of the antibody response.
Antibody can usually be detected 10 days post-infection; detection of antibody 9
days post-infection is rare. When the collection of paired sera is not possible,
diagnosis can be made using results from a single serum sample evaluated together
with clinical and epidemiological data.

Four methods for serodiagnosis were discussed. Microneutralization is considered to
be the gold standard, but this assay requires use of biosafety level 3 facilities, which
can be an important limiting factor. A second method, horse red cell
haemagglutination inhibition, performs well for H5N1 and can be conducted in a
biosafety level 2 facility. Western blot can be used as a confirmatory assay, but is
not suitable for screening purposes, as it produces too many false-positive results.
Serological methods based on the detection of IgG and IgM have specificity issues
that depend on the patient’s age, thus necessitating the use of age-matched controls.
This method can, however, be useful in children.

Seroprevalence studies

. Results of seroprevalence surveys conducted in Hong

Kong in 1997 were briefly reviewed. Seroprevalence in sampled poultry workers
was significant, at 10%, but was less in cullers (3%), and was 0% in the general
population. More recently, serological testing, using paired sera, of 2,109 cullers in
the Republic of Korea detected H5 antibodies in 9 persons (4 in CDC tests and 5 in
Korean tests); infections were associated with mild or no illness and were acquired
prior to the use of personal protective equipment by cullers. In Thailand, all health
care workers studied were negative. In Viet Nam, all contacts of patients were
negative, but some family members had positive results. All sera submitted from
Djibouti, Nigeria, Kazakhstan, and Mongolia for testing at CDC were negative.

RT-PCR

. RT-PCR tests that amplify gene fragments are a universal assay. They are

highly sensitive and specific, and follow a fairly simple protocol. The risk of cross-
contamination is, however, an important problem. To ensure accurate results, it is
important to run quality controls, to validate primers and probes, and to validate the
many different machines and platforms now in use. False-negative test results can
arise from poor sample quality and inefficient extraction, pointing to the need to
control the quality of samples to determine whether degradation had occurred. The
interpretation of test results must always be made with caution; even with PCR, a

WHO working group on influenza research at the human and animal interface

test result might be negative for influenza A virus, but positive for the H5 virus
subtype, or vice versa.

In the US, plans to stockpile PCR diagnostic kits, as part of pandemic preparedness
aimed at increasing laboratory surge capacity, encountered some initial difficulties
in securing regulatory approval to distribute the kits. Other countries planning to
strengthen laboratory surge capacity might want to anticipate similar regulatory
constraints. Countries planning to stockpile PCR diagnostic kits must also anticipate
the need to update primers in line with the continuing evolution of the virus.

Development of a rapid field test.

A presentation from the University of Hong

described an ongoing joint effort with Xiamen University to develop a rapid test,
based on panels of monoclonal antibodies, capable of detecting H5 antigen under
field conditions. Such a test is urgently needed, as the virus is presently having its
greatest impact in developing countries which must often rely on external diagnostic
verification. The assay aims to detect H5 antibody within 20 minutes and to achieve
reliable results without the need for extensive training. A major problem in the
development of such a test arises from the significant antigenic and genetic diversity
now being observed in circulating H5N1 viruses. Research has, however, identified
some monoclonal antibodies with broad cross-reactivity across many of the genetic
groups. The test shows greater sensitivity when compared with commercial tests and
is achieving good specificity as well.

Detection and diagnosis of human infections in Indonesia

. A presentation from

the Ministry of Health’s National Institute of Health Research and Development in
Jakarta gave a comprehensive account of research challenges and plans in Indonesia,
offering insight into many unresolved questions that continue to cloud
understanding of this disease. The Institute is responsible for coordinating all avian
influenza research relevant to humans within the country, and welcomes outside
collaboration, particularly when such collaboration can result in capacity building
and technology transfer.

In Indonesia, the H5N1 virus has been circulating in poultry since mid-2003.
Outbreaks have occurred in 29 of the country’s 33 provinces. Research is focusing
on the extent of the problem in animals, the epidemiology of human cases, and the
epidemiology of clusters of human cases. During 2006, Indonesia had (by end-
September) confirmed 46 human cases, of which 37 were fatal, resulting in a case
fatality of 80.4%. Most patients were under the age of 30 years. For unknown
reasons, fatality was higher in females. One third of the cases were part of clusters.

Human cases have been reported from 8 provinces. It is not presently understood
why human cases have occurred in only 8 out of the 29 provinces affected by animal
outbreaks or why – given the widespread nature of poultry outbreaks – so
comparatively few human cases have occurred. As in many other countries,
Indonesia sees a need for much greater collaboration between the human health and
veterinary sectors.

Efforts are under way to achieve a better picture of the size of the problem in
humans. As a detection strategy, the country is conducting systematic surveillance
for influenza-like illness. In collaboration with the Netherlands, seroprevalence
studies are being conducted in poultry farmers and cullers. Hospital records will be
examined in a search for retrospective cases of acute respiratory illness. Some 44

WHO working group on influenza research at the human and animal interface

hospitals have been appointed for this study and 8 laboratories have been designated,
but these are not yet fully functional.

Testing of suspected human H5N1 infections is performed rapidly in the national
laboratory, with parallel confirmatory testing performed in the NAMRU-3
laboratory. Three tests are used in Indonesia: RT-PCR, haemagglutinin inhibition
with horse red blood cells as a confirmatory assay, and DNA sequencing. Some 34
viruses have been characterized, and none has shown evidence of reassortment. The
country does, however, face a backlog in the testing of samples arising from active
surveillance for influenza-like illness. Around 1000 samples are now being
submitted for testing each week; testing can face a backlog of as much as 3 months,
occasionally resulting in the delayed confirmation of a human case.

Indonesia sees a clear need to conduct more research, yet lacks the capacity to
undertake all studies on its own and welcomes collaborative efforts.

Discussion

Some participants expressed surprise that seroprevalence studies were detecting so
few cases, especially in close contacts of confirmed cases. Moreover, recent
surveillance studies in poultry continue to find a high prevalence of H5N1 viruses in
live animal markets. Are tests sufficiently sensitive to pick these up, especially if
infections are mild or asymptomatic? Is there something inherently different about
this virus that complicates the detection of antibodies in human sera? Additionally,
in some birds fully protected by vaccination, tests have been unable to detect
antibodies, suggesting that immune mechanisms other than antibodies may be
important. Several participants agreed that careful interpretation of results is needed.

A caution was raised concerning use of the horse red blood cell test as a
confirmatory assay: while this test performs well with the H5 virus subtype, it is not
suitable for detection of human infections with the H7 virus subtype.

Given the widespread presence of the virus in poultry in Indonesia, a question was
raised about the use of poultry vaccination as a control strategy. It was felt that the
public health advantages of doing so could be considerable. Viet Nam, for example,
introduced such a control strategy and subsequently experienced no further human
cases. The meeting was informed that the Indonesian government was presently
weighing the advantages and disadvantages of large-scale poultry vaccination, but
has not yet reached a decision. Such decisions understandably have a high political
profile.

Also, concerning the situation in Indonesia, participants cited suggestive evidence,
largely from the Karo family cluster in North Sumatra, that genetic factors might
influence human susceptibility to infection, as only blood relatives were infected in
that cluster, despite multiple opportunities for the virus to spread to spouses or into
the general community. Indonesia welcomed collaboration in exploring this
possibility, which has been included among the country’s research plans. The
banking of DNA from deceased patients, family members, contacts, and controls
would, however, raise important ethical issues. It was suggested that WHO could
play a leading role in proactively addressing these issues.

WHO working group on influenza research at the human and animal interface

II.

Protecting humans: vaccines and antiviral drugs

Vaccines.

A presentation from the WHO collaborating centre in the United

Kingdom reviewed the status of vaccine development. Results to date have not been
promising. Efforts to develop a vaccine that confers adequate protection have been
greatly complicated by the emergence of genetically and antigenically diverse
viruses that are now simultaneously circulating in different geographical areas.
Vaccines that protected against viruses from clade one demonstrated poor cross-
reactivity for virus subgroups in the second clade.

Distinct genetic groups of circulating viruses.

In August 2006, WHO published its

recommendations for H5N1 candidate vaccine viruses in the Weekly
Epidemiological Record

. This publication also draws attention to the problem

caused by the diversity of viruses currently in circulation. Viruses have recently
evolved into two distinct genetic groups, or phylogenetic clades. Clade 1 viruses
circulated in Cambodia, Thailand, and Viet Nam during 2004 and 2005 and were
responsible for human cases reported in those countries. Clade 2 viruses, which are
genetically and antigenically distinct, initially circulated in poultry in China and
Indonesia during 2004 and the first half of 2005, without causing human cases. In
mid-2005, the epidemiology of clade 2 viruses shifted, circulation of the viruses
increased, and westward spread began, initially in wild birds, then in poultry, then in
sporadic human cases in Turkey, Azerbaijan, Iraq, Egypt, and Djibouti. Beginning
in the second half of 2005, these viruses also caused human cases in Indonesia and
China.

Six distinct subgroups within clade 2 were subsequently distinguished, of which
three also differ in geographical distribution. Of these, one subgroup has continued
to circulate in Indonesia, a second subgroup of the so-called Qinghai Lake-like
viruses has caused outbreaks in Europe, the Middle East, and Africa, and the third
subgroup is circulating mainly in China and, to a lesser extent, in Viet Nam.
Haemagglutination inhibition tests using ferret antisera have, unfortunately,
demonstrated poor cross-reactivity among the different clades and subgroups, thus
complicating efforts to produce a fully protective vaccine.

WHO has also published a phylogenetic tree

showing how the various human and

animal viruses from different geographical areas cluster together in genetic groups.
An exception occurs with the viruses from the Karo family cluster in North Sumatra;
in inhibition tests, these viruses did not react well with other viruses isolated in
Indonesia and showed greater similarity to viruses circulating in China.

Status of vaccine development.

The results of several vaccine trials were reviewed.

In the USA, Sanofi-Pasteur’s candidate vaccine uses a split virus and 90 micrograms
of antigen in a two-dose schedule. In France, trials with a Sanofi-Pasteur alum-
adjuvanted split virus vaccine using 30 micrograms of antigen in a two-dose
schedule suggested that the adjuvant was not having a substantial effect. In Australia,
CSL is developing a split virus vaccine using 15 micrograms of antigen in a two-
dose schedule. Sinovac in China has shown good results with an alum-adjuvanted
whole virus vaccine using 10 micrograms of antigen in a two-dose schedule. In

http://www.who.int/wer/2006/wer8134_35/en/index.html

http://www.who.int/csr/disease/avian_influenza/guidelines/recommendationvaccine.pdf

WHO working group on influenza research at the human and animal interface

Belgium, GlaxoSmithKline has reported results with a candidate vaccine using 3.8
micrograms of virus in a two-dose schedule with AS03 adjuvant. It was also
reported that Omnivest in Hungary is developing a vaccine and evaluating a one-
dose schedule.

Clinical trials.

Several clinical trials of candidate vaccines are under way in Europe,

two are being conducted in the USA, and four are ongoing in Japan. In the USA,
phase I clinical trials, have, unfortunately, demonstrated low antibody response to
viruses outside the same clade as the candidate vaccine virus. In Japan, phase I
clinical trials produced data similar to that in the USA, showing a lack of cross-
protection among the two clades and subgroups within clade two.

Other antigen

sparing strategies

. Intradermal injection as an antigen-sparing

strategy does not look promising and is not likely to be suitable for worldwide use.

Use of pre

pandemic vaccines

. Participants noted that many fundamental questions

underlying the development of an effective pandemic vaccine have not yet been
answered. What is the best adjuvant for boosting the response? Does one antigen
dominate in influencing antigen recognition? What are the benchmarks for assessing
an adequate level of protection? In the absence of scientific answers to these
questions, concern was expressed that national policy decisions about which
vaccines to stockpile may be premature, despite the understandable desire of
governments to invest now in some means of protecting their populations in the
event of an influenza pandemic.

Resistance of H5N1 viruses to antiviral drugs

. A presentation from the Hong

Kong University summarized the results of surveillance for drug-resistant strains of
H5N1 virus. For amantadine, which is the second-choice antiviral drug, clade 2
viruses are more sensitive than clade 1 viruses. In Indonesia, however, (where clade
2 viruses are circulating) the prevalence of resistance to amantadine is
approximately 50%. For 2006, Hong Kong data showed 28.8% (17/15) amantadine
resistance among Indonesian isolates and 20% (15/75) for Chinese isolates. For the
most part, the subgroup of Qinghai Lake-like viruses showed susceptibility to
amantadine. In Viet Nam, from 50% to 100% of viruses studied in 2003 and 2004
showed a mutation associated with amantadine resistance, but this figure dropped to
10% in 2005. Amantadine resistance is also being observed in viruses responsible
for seasonal influenza. It is not known if resistant strains of these viruses will persist
or be replaced by fully susceptible strains. Resistance to the neuraminidase inhibitor,
oseltamivir – presently the first-choice antiviral drug – has been observed in a few
patients, and that finding is of concern. Further, surveillance studies also indicate a
low prevalence of resistance mutations to oseltamivir in avian isolates, especially in
2005 and 2006.

The quest for new drugs.

The worrisome H5N1 situation has greatly increased

interest in the development of antiviral drugs, as it would be unwise to rely on the
present limited range of therapeutic and prophylactic options. Clinical trials of
additional neuraminidase inhibitors, including zanamivir and peramivir, are
presently under way. In Japan, trials are under way for so-called “long-lasting”
neuraminidase inhibitors capable of achieving very high blood plasma levels after a
single-dose treatment via inhalation. Recent research on the structure of
neuraminidase has found a slightly different catalytic site among different virus

WHO working group on influenza research at the human and animal interface

subtypes including the N1, pointing the way towards the structure-based design of
new drugs that might provide broader-based therapy.

Research from the University of Wisconsin described the use of a novel peptide to
inhibit virus attachment as a new approach to drug development. In

in vitro

studies,

the peptide, which is not specific to virus subtype, inhibits virus binding to cells in a
dose-dependent manner. In mice, the peptide prolonged survival time when
challenged with H5N1 viruses. The potential for side effects of such peptides in
humans needs to be addressed.

Discussion

Concerning pandemic vaccines, participants saw a great need for benchmarks for
the evaluation of candidate vaccines. On such an important matter, it is unwise to
leave assessments of appropriate vaccines to competing manufacturers. Could WHO
advise the world on norms and standards for good vaccines? Would broader cross-
protection conferred by a candidate vaccine make it the superior product? What is
the most appropriate animal model for assessing protection? Participants also saw a
need for integrated studies of sera from vaccinated individuals arising from the
various clinical trials, and agreed that governments should not rush to place orders
for pre-pandemic vaccines when so many fundamental scientific questions are still
outstanding.

III.

Surveillance in birds and other animals: assessing the
coming risks

H5N1 surveillance in Europe.

A presentation from the National Influenza Centre

in the Netherlands summarized surveillance efforts in Europe for avian influenza
viruses in poultry, wild birds, and other animals. For wild birds, collaboration with
ornithologists has provided a wealth of useful information; flyways and migration
patterns are extremely complex. Different wild and domestic species are now known
to respond to the virus in different ways. These differences are clearly important for
the identification of sentinel and reservoir species and the performance of risk
assessment, particularly concerning the potential for a recurrent westward spread of
the virus. In addition, patterns of virus excretion are known to vary according to the
species, and this, too, can affect modes and risks of transmission, particularly as
transmission occurs via water sources.

From 1959 to 2002, surveillance of influenza A viruses in avian species revealed
few major ecological changes. The situation is now very different. The Qinghai
Lake incident, which began in China in late April 2005 and resulted in the death of
some 6000 migratory birds, was followed by a progressively westward spread of
highly pathogenic H5N1 virus. In terms of geographical spread of the virus, mallard
ducks are now regarded as the “champion” vectors; mute swans are highly
susceptible birds that are thought to serve as sentinels, but probably not as vectors of
virus transmission. The prevalence of highly pathogenic H5N1 infection is now
known to be higher in dabbling ducks than in diving ducks. Studies have shown that

WHO working group on influenza research at the human and animal interface

ducks shed more H5N1 virus via the respiratory tract than via faeces; as a sampling
technique, there is growing evidence to suggest that pharyngeal swabs are more
important than cloacal swabs, which might fail to detect infections.

The surveillance network in Europe includes parts of Africa and Asia. Studies of
viruses isolated from poultry in Nigeria have found evidence for three separate
introductions of the virus; this finding argues against spread of the virus through
trade or smuggling from a single source. In other situations, both illegal and normal
trade in birds can play a role in virus spread.

For wild birds, the risk of infection is governed by such factors as the number of
animals, their geographical origin, and their behaviour, including their contact with
free-ranging domestic birds. Relay transmission, in which the virus moves both
ways between wild and domestic birds, may also be an important mechanism for
geographical dispersion of the virus, as it allows, via sequential series of infections,
for virus transmission to occur over long distances. Surveillance has further
determined that highly pathogenic H5N1 viruses show a high rate of mutation;
reassortment with other avian influenza viruses has also been documented.

Other animal species that have been infected with the H5N1 virus include zoo tigers
in Thailand, domestic cats in Europe and elsewhere, and a stone marten and a mink
in Europe. Studies show that cat-to-cat transmission can occur in both domestic cats
and tigers. While the participants considered that avian species remain by far the
most important source of human infections, they also felt that the role of domestic
cats needs to be further investigated, given their very close association with humans.
In experimentally infected non-human primates or macaques, no evidence of
disseminated disease has been found – a finding that appears to be similar to what is
seen in human infections.

Laboratory testing for European outbreaks.

These findings were further

elaborated by a second presentation on surveillance in Europe, provided by the
UK’s Central Veterinary Laboratory Agency at Weybridge. This laboratory has
done much of the confirmatory testing for avian outbreaks in Europe. Test results
further confirm the significance of mute swans as a sentinel species. Surveillance
has found some highly pathogenic H5N1 virus in live, but healthy, wild birds,
including gulls. Again, findings indicate that relay transmission is now occurring
between wild and domestic birds. Rural areas, where wild and domestic birds can
easily mingle, are considered to be of particular concern.

Europe experienced its first outbreaks during late 2005 and early 2006, with mainly
wild birds affected and only a few, mostly well-contained, poultry outbreaks
documented. The winter season at that time was especially severe, resulting in more
frozen water areas than usual, which might explain the high concentrations of birds
in some areas, such as northern Germany. Animal vaccination policy in Europe has
become more open, and some countries are now using vaccination as a control
strategy. Experimental work with sub-lethal doses of H5N1 virus shows that birds
can still shed virus in the absence of signs of illness. Experimental work in avian
species has further shown that highly pathogenic H5 viruses are more lethal than
viruses of the H7 subtype. An analysis of diagnostic proficiency in European
laboratories was undertaken. Overall, performance was good; poor performance was

WHO working group on influenza research at the human and animal interface

often due to the use of commercial diagnostic kits, which did not always produce
reliable results.

Recent evolution of the H5N1 virus in poultry in China

. A presentation from the

University of Hong Kong reviewed findings from a surveillance network that covers
some 4 billion birds in China. These findings indicate that the virus has become
endemic and is continuing to evolve. The Z genotype of the virus remains dominant.
In migratory birds, the Qinghai Lake outbreak changed the epidemiology in mid-
2005, and this change signalled the progressive westward spread of Qinghai Lake
viruses and their descendants. Surveillance over the past 24 months showed a peak
in virus activity during the month of January, followed by a decline in virus activity
in April. The situation is severe and not yet fully under control. During the past year,
more than 1,300 H5N1 isolates have been obtained from poultry in southern China.
Prevalence is higher in domestic ducks and geese than in chickens, and highly
pathogenic virus was being found in apparently healthy birds. Most of these isolates
belonged to the dominant Z genotype. In southern China, viruses related to the
Fujian-like lineage of the Z genotype were found in 80% of the isolates and that
figure has recently risen to 95%, indicating that, in southern China, Fujian-like
viruses are replacing other virus lineages and becoming the dominant lineage within
the genotype. To manage this situation, the whole poultry population will need to be
vaccinated, accompanied by monitoring of effectiveness of the vaccination
programs.

Lessons from recent poultry outbreaks.

Japan experienced an outbreak, beginning

in January 2004, of highly pathogenic H5N1 avian influenza in poultry in three
areas, and this outbreak – which was controlled – has been well documented. In
June 2005, a second outbreak, caused by a low pathogenic H5N2 virus, began and
continued through April 2006, resulting in the culling of some 5.68 million birds.
Interestingly, the closest virus on the phylogenetic tree was from Nicaragua, a
finding that could not be readily explained. Studies demonstrated that the H5N2
virus is highly adapted to chickens; it grows very well in chick embryos.
Serosurveillance of poultry has continued in Japan, and the virus is no longer being
detected. Japan has also been working region-wide to improve diagnostic capacity
in Asian countries, as many have no biosafety level 3 facilities or sequencing
capability. A panel of viruses for diagnostic testing has been developed and training
courses in animal influenza have been conducted. A new website cataloguing
Japan's collection of low pathogenic avian influenza strains available to researchers
was presented and discussed.

In Kazakhstan

– a large country with many different ecological zones and wetlands

criss-crossed by migratory routes – a system for surveillance, mainly in wild birds,
has been in place since 1979. In 2005, the country experienced an outbreak of
highly pathogenic H5N1 avian influenza on a goose farm, an open facility allowing
opportunities for wild and domestic birds to mingle. In March and April of 2006, the
surveillance system detected highly pathogenic H5N1 virus in 80 wild birds found
dead. All viruses isolated from these birds were highly pathogenic, resulting in
100% mortality. Characterization of the viruses showed their similarity to Qinghai
Lake-like strains isolated from birds from Angola, Nigeria, and the Russian
Federation. Work was also described regarding Kazakhstan’s efforts to develop an
ISCOM-based poultry vaccine.

WHO working group on influenza research at the human and animal interface

The Republic of Korea was the first country, in the current wave of outbreaks, to
report highly pathogenic H5N1 in poultry. That outbreak began on 10 December
2003 and continued for 14 weeks. Since 21 March 2004, no H5N1 virus has been
detected in the country. The outbreak was costly and highly disruptive. At that time,
Korean animal authorities implemented strong control measures without vaccination,
including culling within a 3 km radius, strict movement controls, disinfection
activities, intensive surveillance, and biosecurity and public awareness activities in
line with the contingency plan. The government adopted a policy of 100%
compensation for lost birds. Altogether, some 5.6 million birds were destroyed at a
cost of more than US$ 150 million. As part of risk assessment for a resurgence or
reintroduction of highly pathogenic avian influenza, active surveillance has focused
on duck farms and wild migratory birds, including both winter and summer
migrations. Serological surveillance is used on duck farms. In addition, passive
surveillance is conducted for wild birds found dead.

Discussion

Participants expressed concern about the detection of H5N1 birds this spring in
Kazakhstan at the Caspian Sea and reports of dead birds in recent weeks, as that
finding suggests a possible repeat of the pattern of westward spread of the virus that
caused so much alarm in late 2005 and early 2006. As outbreaks at Qinghai Lake
have been reported again in 2006, countries located along the autumn migratory
routes will need to be vigilant.

Questions were raised about the unusual outbreak of H5N2 avian influenza in Japan.
As the H5N2 strain was used in poultry vaccines, could the illegal use of possibly
sub-standard vaccine have been a source of the outbreak? The group was informed
that Japanese authorities investigated that possibility, but were unable to reach a
conclusion. Similar to results with H5N1 virus infections in previous studies, the
Japanese H5N2 virus was found to infect pigs (miniature pigs) in experimental
studies, but virus shedding was limited and of short duration.

As Korea was the first country to report a poultry outbreak during the most recent
H5N1 outbreak, questions were also raised about the source of that outbreak. The
meeting was informed that the Korean virus is very similar to H5N1 viruses that
circulated in southern China during 2002.

Participants emphasized that the potential role of pigs in the emergence of a
pandemic virus should not be forgotten. Surveillance for influenza viruses in pigs
can yield important information about the prevalence and ecology of these viruses;
such an approach can also provide early evidence of a reassortant virus with
pandemic potential.

WHO working group on influenza research at the human and animal interface

IV.

Deciphering the virulence and pathogenicity of H5N1
infections in humans

H5N1 pathogenesis.

A presentation from the University of Hong Kong considered

how the virus infects humans and why the resulting disease can be so severe. The
situation with human infections was described as paradoxical: thousands upon
thousands of people are likely to have been exposed through contact with poultry,
yet fewer than 300 human infections have been detected; conversely, 30% of the
initially infected human patients in Hong Kong in 1997 had no apparent contact
with birds. Overall, the virus continues to show inefficient spread, both from
animals to humans and from human to human. Though the virus replicates in
humans, infected people do not transmit it easily to others. Clusters of cases have
occurred, but transmission has not been sustained and appears to be confined to
genetically-related persons.

What factors might be at work? Genetic susceptibility of the host? Host resistance
factors? The significant species barrier observed clearly involves receptor specificity;
however, other genes unrelated to receptor interactions might be involved. Equally
possible, multiple genes might be working in tandem to govern human susceptibility
to infection.

Avian-like receptors in the human lung.

Recent research on receptor expression

in the human respiratory tract was reviewed. Work has shown that receptors for
avian-like viruses can be found in the epithelia deep in the human lung. This finding
helps explain several features of human infections – if the virus lodges deep in the
lung, that would explain the severity of the disease, the inefficient spread from
animals to humans, and the virtual absence of human-to-human transmission. The
paradox was not, however, fully resolved, since data were also presented showing
that the H5N1 virus can also cause infection in cells from nasopharyngeal biopsies
of the upper respiratory tract, like normal influenza. It is furthermore known that
normal human H3N2 influenza viruses can cause infection deep in the lungs. In
H5N1 infection, higher viral loads have been detected in the pharynx than seen with
normal influenza, but the high prolonged virus replication observed in H5N1
infection might simply reflect a lack of pre-existing immunity.

The question remains open: what is needed for the H5N1 virus to transmit
efficiently from one human to another? Overall, the genetic controls of cross-species
infectivity and transmissibility of influenza viruses are complex and not yet fully
understood. In experimentally infected ferrets, very inefficient animal-to-animal
transmission was shown to occur. In further ferret experiments using artificially
reassorted viruses, the efficiency of animal-to-animal transmission improved
somewhat. These experiments suggest that a reassortment of genes between
mammalian and avian viruses may not alone have a great impact on transmissibility
with the viruses tested to dated. This finding does not, however, exclude the
possibility that adaptive mutation alone or reassortment together with adaptive
mutation may give rise to a pandemic virus.

Disease severity

. The disease caused in humans by the H5N1 virus was described

as fundamentally different from that caused by normal influenza. In H5N1 infection,
the disease syndrome typically shows progressive primary viral pneumonia, acute
respiratory distress, marked leukopenia and lymphopenia, and (in some cases)

WHO working group on influenza research at the human and animal interface

diarrhoea and liver or renal dysfunction. What might explain this severity? Some
limited findings suggest that the virus might cause disseminated infection, affecting
multiple organs. In some patients with a fatal outcome, virus has been detected in
faeces, serum, and blood plasma. However, respiratory pathology remains the
primary cause of death. Additional data presented support the hypothesis that severe
disease is based on induction of a “cytokine storm”; it was pointed out, however,
that this remains a “chicken-and-egg” dilemma – does an overwhelming level of
cytokinemia result in, or from, extensive tissue damage and disease?

Determinants of virulence

and transmissibility

. A mutation in the PB2 gene, at

position 627, has been shown to influence pathogenicity in mice, but that finding
has not consistently correlated with severity of infection among viruses isolated
from patients in Indonesia and Viet Nam. Investigation of the role of internal
polymerase genes is continuing. It appears clear, however, that the external HA and
NA genes are not the sole drivers of disease severity. Likewise, transmissibility of
the virus may ultimately prove to be a genetically complex trait. One especially
important question that was discussed is whether the H5N1 virus is likely to retain
its present high lethality should it acquire an ability to spread easily from person to
person, and thus start a pandemic. Should the virus improve its transmissibility by
acquiring, through a reassortment event, internal human genes, then the lethality of
the virus would most likely be reduced. However, should the virus improve its
transmissibility through adaptation as a wholly avian virus, then the present high
lethality could be maintained during a pandemic.

Host range of avian influenza viruses

. A presentation from the WHO

collaborating centre in the United Kingdom explained ongoing research aimed at
identifying specific genes in influenza viruses that influence their ability to infect
different species. These experiments, which used a chloramphenicol
acetyltransferase reporter system to map the contribution of individual amino acids
to polymerase function, had confirmed the role of a Lys mutation at position 627 on
the PB2 gene in regulating the species-specific activity of internal polymerase genes.
In these experiments, neither the HA nor NA external genes showed a role in
determining the ability of influenza viruses to infect human cells. Interferon
induction and response were shown to contribute to the host range of avian
influenza viruses; however, the control of these effects depends on genes other than
just the NS1. Additional studies described the creation of chicken/human chimeric
cell lines with variable amounts of chicken genetic components, which may be
found to be useful in understanding the role of host genetic factors in species
specificity.

Studies of animal H5N1 viruses in Australia

. A presentation from the Australian

Animal Health Laboratory reviewed a range of recent studies. Concerning avian
influenza in birds, in the north, surveillance has found Newcastle disease virus but
no avian influenza viruses whatsoever. It is somewhat puzzling that Qinghai Lake
viruses have migrated north and westwards but not towards the south. The
laboratory has also conducted work on experimentally infected birds showing that
the severity of disease is dose-dependent. In ducks, studies indicate that the
inoculation dose of H5N1 virus could be titrated down to a point where birds
become infected and seroconvert, but do not develop clinical illness. In ferrets,
experimental infection at low doses induced pulmonary disease; higher doses

WHO working group on influenza research at the human and animal interface

induced systemic disease. Ferrets have also been used to test vaccines and have
demonstrated improved antibody response with an adjuvanted vaccine. Isolates from
Cambodia, Indonesia, and Malaysia have been characterized; some Indonesian
viruses have shown mutations associated with resistance to oseltamavir.

Pathogenesis in the duck

. A presentation from the WHO collaborating centre for

animal influenza in Memphis, Tennessee concentrated on pathogenicity in ducks.
As mentioned previously, mallard ducks are thought to be a major force in
maintenance of viruses in the wild, but the highly pathogenic phenotype does not
remain stable; instead, there appears to be rapid selection for a virus that is non-
pathogenic for ducks, but still highly pathogenic for chickens and, presumably, for
humans as well. In both Viet Nam and Thailand, ducks were infected but appeared
to remain healthy. From 1997 to 2002, most virus shedding in ducks occurred via
faeces. More recently, however, most virus shedding has occurred via the
respiratory tract; this route of virus shedding does, however, continue to cause
contamination of water sources used by the birds. In domestic poultry, ducks shed
virus for up to 17 days and pheasants shed virus for up to 40 days.

Concerning mechanisms of pathogenicity, experimental studies show that five
amino acid changes transform low pathogenic viruses into highly pathogenic viruses;
pathogenicity is associated with plaque size in MDCK cells. In ducks, mutations in
internal polymerase genes rather than the HA appear to control high pathogenicity.
It is not known, however, exactly which of the polymerase genes work to increase
pathogenicity. Finally, it was pointed out that chickens are clearly not normal hosts
for influenza A viruses; only viruses of the H5 and H7 subtypes replicate efficiently
in chickens but this situation is changing: H9N2 viruses are endemic in chickens
across Eurasia.

Discussion

Questions were raised about the extent to which H5N1 viruses spill over into the
bloodstream of infected humans and whether this is necessarily associated with
tissue damage. As very few autopsies have been performed, such questions remain
difficult to answer. Reference was, however, made to evidence from two pregnant
patients in China where virus was detected in the placenta.

Several participants sought explanations for the strong propensity of the H5N1 virus
to infect children and young adults, causing severe disease with high case-fatality.
In persons older than 50 years, infections were much less common Exposure history
might be one explanation, as children tend to treat birds as pets or play in areas
frequented by poultry. That hypothesis was not, however, considered adequate to
explain all cases. Are immunological factors at work? Some evidence suggests a
broader antibody response in the elderly. Could previous exposure to other influenza
viruses help explain the distinctive age profile? Cytokine expression can also be
related to age, and this might be another hypothesis to explore.

Concerning the potential high lethality of a wholly avian pandemic virus, some
modelling studies have suggested that pandemic spread could not be fully sustained
in the presence of very high mortality. All such matters remain difficult to predict.

WHO working group on influenza research at the human and animal interface

List of participants

Dr Vladimir E. Berezin

Acting Director
Institute of Microbiology and Virology
Ministry of Education and Science
Kazakhstan

Dr Terry Besselaar

National Institute for Communicable
Diseases, Vaccine Preventable
Virus Infectious Unit
Gauteng, South Africa

Dr Ian H. Brown

Central Veterinary Laboratory
Agency-Weybridge
Surrey, United Kingdom

Dr Christianne J.M. Bruschke

Project Manager
World Organization for Animal
Health OIE
Paris, France

Dr Honglin Chen

Department of Microbiology, Faculty
of Medicine
The University of Hong Kong
Hong Kong SAR, People’s Republic
of China

Dr Nancy Cox

WHO Collaborating Centre for
Reference and Research on
Influenza
National Centers for Infectious
Diseases
Centers for Disease Control and
Prevention
Atlanta, GA, USA

Dr Gwenaëlle Dauphin

OFFLU Liaison Officer and
Laboratory Expert, Animal Health
Service
Food and Agriculture Organization
(FAO)
Rome, Italy

Dr William Dundon

OIE, FAO and National Reference
Laboratory for Newcastle Disease
and Avian Influenza
Instituto Zooprofilattico Sperimentale
delle Venezie
Venice, Italy

Dr. B. C. Easterday

Dean and Professor Emeritus
School of Veterinary Medicine
University of Wisconsin
Madison, WI, USA

Dr Yi Guan

Department of Microbiology
The University of Hong Kong, State
Key Laboratory of Emerging
Infectious Diseases
Hong Kong SAR, People’s Republic
of China

Dr Alan Hay

WHO Collaborating Centre for
Reference and Research on
Influenza
National Institute for Medical
Research
London, United Kingdom

Dr Youn-Jeong Lee

Researcher, Avian Disease Division
Ministry of Agriculture & Forestry
National Veterinary Research &
Quarantine Service
Anyang City Gyeonggi-do
Republic of Korea

Dr John McCauley

National Institute for Medical
Research, Division of Virology
Mill Hill
London, United Kingdom

Dr Christopher Olsen

Professor of Public Health,
Department of Pathobiological
Sciences School of Veterinary
Medicine
University of Wisconsin-Madison
Madison, WI USA

Prof. Albert D. M. E. Osterhaus

Institute of Virology
Erasmus Universiteit
National Influenza Centre
Rotterdam, The Netherlands

Dr Malik Peiris

Professor
Department of Microbiology
The University of Hong Kong,
Faculty of Medicine, Queen Mary
Hospital
Hong Kong SAR, People’s Republic
of China

Dr Yoshihiro Sakoda

Hokkaido University
Graduate School of Veterinary
Medicine
Sapporo City, Japan

WHO working group on influenza research at the human and animal interface

Dr Stacy Schultz-Cherry

University of Wisconsin-Madison,
Department of Pathobiological
Sciences and Department of Medical
Microbiology and Immunology
Madison, WI, USA

Dr Endang Sedyaningsih

Senior Researcher
National Institute of Health Research
and Development
Jakarta, Indonesia

Dr Paul Selleck

Australian Animal Health Laboratory
Geelong, Australia

Dr Triono Soendoro

Jakarta
Indonesia

Dr Masato Tashiro

Director, WHO Collaborating Centre
for Reference and Research on
Influenza
Department of Viral Diseases and
Vaccine Control
National Institute of Infectious
Disease
Tokyo, Japan

Dr Robert Webster

WHO collaborating Center for
Studies on the Ecology of Influenza
in Animals, Virology Division
Department of Infectious Disease
St Jude Children’s Research
Hospital
Memphis, TN, USA

Dr Xu Zhen

Beijing
China

Secretariat

Dr Michael J. Ryan

Director, Department of Epidemic
and Pandemic Alert and Response

Dr Keiji Fukuda

Coordinator, WHO Global Influenza
Programme Department of Epidemic
and Pandemic Alert and Response

Dr Maurizio Barbeschi

Office for Alert and Response
Operations Department of Epidemic
and Pandemic Alert and Response

Dr Amina Chaieb

Office for Alert and Response
Operations Department of Epidemic
and Pandemic Alert and Response

Dr Alice Croisier

WHO Global Influenza Programme
Department of Epidemic and
Pandemic Alert and Response

Dr Daniel Lavanchy

Office for Alert and Response
Operations Department of Epidemic
and Pandemic Alert and Response

Dr Suzie Lyons

Office for Alert and Response
Operations
Department of Epidemic and
Pandemic Alert and Response

Dr Elizabeth Mumford

WHO Global Influenza Programme
Department of Epidemic and
Pandemic Alert and Response

Dr Mike Perdue

WHO Global Influenza Programme
Department of Epidemic and
Pandemic Alert and Response

Dr Kaat Vandemaele

Office for Alert and Response
Operations Department of Epidemic
and Pandemic Alert and Response

Dr Wenqing Zhang

WHO Global Influenza Programme
Department of Epidemic and
Pandemic Alert and Response