© World Health Organization

WHO Technical Report Series, No. 927, 2005

Annex 3
Recommendations for the production and control of
influenza vaccine (inactivated)

Recommendations published by WHO are intended to be scientific and

advisory in nature. The parts of each section printed in type of normal
size have been written in a form, such that, should a national regulatory
authority so desire, they may be adopted as they stand as definitive
national requirements or used as the basis of such requirements. Those
parts of each section printed in small type are comments and recommen-
dations for guidance for those manufacturers and national regulatory
authorities which may benefit from additional information.
It is recommended that modifications be made only on condition that the
modifications ensure that the vaccine is at least as safe and efficacious
as that prepared in accordance with the recommendations set out below.
In order to facilitate the international distribution of vaccine made in
accordance with these recommendations, a summary protocol for the
recording of results of the tests is given in Appendix 1.
The terms “national regulatory authority” and “national control laboratory”
as used in these recommendations, always refer to the country in which
the vaccine is manufactured.

Introduction 101

General considerations 102

Part A. Manufacturing recommendations 104

A.1 Definitions 104
A.2 General manufacturing requirements 107
A.3 Production control 108
A.4 Filling and containers 120
A.5 Control tests on final lot 121
A.6 Records 122
A.7 Retained samples 122
A.8 Labelling 122
A.9 Distribution and transport 122
A.10 Stability testing, storage and expiry date 123 E

99

6952 ECE TEXT 99 8/21/05, 10:57

 Part B. Requirements for national control authorities 123
B.1 General 123
B.2 Release and certification 124
B.3 Clinical evaluation of influenza vaccines 124

Authors 124

Acknowledgements 125

References 125

Appendix 1
Summary protocol for influenza vaccine (inactivated) (master/working
seed lot Type A or Type B) 127

Appendix 2
Reference laboratories 133

Appendix 3
Model certificate for the release of influenza vaccines (inactivated) 134

100

6952 ECE TEXT 100 8/21/05, 10:57

 Introduction
Influenza is a significant cause of morbidity and mortality and has a
major social and economic impact throughout the world. During
major epidemics many people require medical treatment or hos-
pitalization. Excess mortality often accompanies influenza epidemics,
the vast majority of those affected being elderly. Because the elderly
constitute the most rapidly increasing sector of the population
in many countries, the epidemiology of influenza can be expected to
change accordingly, especially in the developed countries. At present,
the only means of influenza prophylaxis generally available is
vaccination.
In 1967, a group of experts formulated requirements for inactivated
influenza vaccine and these were published as an annex to the twen-
tieth report of the Expert Committee on Biological Standardization
(1). During the following 5 years, technical developments in the puri-
fication of the virus suspensions from which vaccines were made, as
well as in the measurement of the virus content, were such that the
potency of whole virus vaccines could be expressed in international
units. Accordingly, an addendum to the requirements was annexed to
the twenty-fifth report of the Expert Committee on Biological Stan-
dardization (2). In its twenty-ninth report (3), the Committee recog-
nized that technical developments had completely altered the method
of measurement of the haemagglutinin content of the vaccines
and that the International Reference Preparation of Influenza Virus
Haemagglutinin (Type A) established in 1967 was no longer appro-
priate for controlling the haemagglutinin content of inactivated
influenza vaccines because it no longer represented the haemag-
glutinin of the prevalent strains. Accordingly, the International
Reference Preparation was withdrawn, and the Committee recom-
mended that the requirements for inactivated influenza vaccine
should be revised. Revised requirements were approved by the WHO
Expert Committee on Biological Standardization in 1978 (4) and
modified in 1990 (5).
Since 1990, there have been significant new developments in methods
of influenza vaccine production resulting from: increased develop-
ment of mammalian cell lines for vaccine production (6); increased
experience in use of adjuvants; and rapid development of reverse
genetics technologies for generation of vaccine viruses. There has also
been considerable effort directed to pandemic planning to ensure that
safe, effective vaccines can be quickly produced in response to a
pandemic emergency. Consequently it has become necessary to revise
the requirements to reflect these new developments. In accordance E

101

6952 ECE TEXT 101 8/21/05, 10:57

 with current WHO policy, the revised document is renamed as
“Recommendations”.

General considerations
Inactivated influenza vaccines have been in widespread use for nearly
60 years. The efficacy of immunization has varied according to cir-
cumstances, but protection rates of 75–90% have been reported. Dif-
ferences in protective efficacy may result from continuing antigenic
variation in the prevalent epidemic strains. Because of this variation,
the composition of inactivated influenza virus vaccine, unlike that of
most viral vaccines, must be kept constantly under review. Accord-
ingly, WHO publishes recommendations concerning the strains to be
included in the vaccine twice annually.
Influenza vaccines usually contain one or more influenza A viruses.
However, because influenza A viruses undergo frequent and pro-
gressive antigenic drift in their haemagglutinin and neuraminidase
antigens, vaccines containing formerly prevalent viruses are expected
to be less protective against virus variants showing antigenic drift than
against the homologous virus. When a new subtype of influenza A
virus bearing new haemagglutinin (and neuraminidase) antigen(s)
appears, it is likely that vaccine containing the antigen(s) of the
influenza A subtype(s) formerly prevalent will be ineffective, so that
a vaccine containing the new pandemic virus will be required.
Changes in the structure of the haemagglutinin and neuraminidase
molecules, which result in changes in antigenicity as new epidemic
strains appear, involve surface residues in the region of the molecule
furthest from the viral envelope. Prediction of future variations is not
possible because the mechanism of selection of antigenic variants,
antigenic drift, is not known and several evolutionary pathways ap-
pear possible. Antigenic shift (i.e. the appearance of epidemic strains
with a new haemagglutinin subtype) is also unpredictable.
Antigenic drift in influenza B virus strains is less frequent than that in
the A strains and antigenic shift is unknown. Although distinct lin-
eages of influenza B may occasionally co-circulate, it is usual for
influenza vaccines to contain only one influenza B strain.
In addition to antigenic drift and shift, there is another type of varia-
tion among influenza viruses, namely the preferential growth of cer-
tain virus subpopulations in different host cells in which the virus is
cultivated. Influenza viruses grown in embryonated eggs often exhibit
E antigenic and biological differences from those isolated and

102

6952 ECE TEXT 102 8/21/05, 10:57

 maintained in mammalian cells. Sequence analysis of the haemag-
glutinin gene of such variants has shown that, typically, virus grown in
mammalian cells differs from virus from the same source cultivated in
eggs only by the substitution of a single amino acid in the haemag-
glutinin molecule.
The WHO Expert Committee on Biological Standardization recom-
mended in its twenty-ninth report (3) that the potency of influenza
vaccines should be expressed in mg of haemagglutinin per ml (or
dose), as determined by suitable immunodiffusion methods. In order
to standardize these methods, reference antigen (calibrated in mg of
the haemagglutinin per ml) and specific anti-haemagglutinin serum,
suitable for use in the assay of the haemagglutinin content of each
component of inactivated vaccines, are prepared and distributed by
reference laboratories (Appendix 2). A new reference antigen and
antiserum is prepared each time it is necessary to introduce a new
virus strain into the vaccines, and these are standardized by interna-
tional collaborative study.
Over the past 20 years there have been many clinical trials of whole
virus, split and subunit influenza vaccines. This has led to the gener-
ally accepted view that one dose of vaccine containing 15 mg of
haemagglutinin per strain per dose, will stimulate haemagglutination-
inhibition antibody levels consistent with immunity in most primed
individuals (7).
There has been much progress in developing influenza vaccines with
adjuvant and some such vaccines are now licensed. The main issues
for vaccine quality are: demonstration of compatibility of the adju-
vant with the antigenic components of the vaccine; proof of consistent
association with vaccine antigens (if appropriate) at time of produc-
tion and throughout shelf-life; effect of adjuvant on vaccine potency
assays; and biochemical purity of adjuvant.
There is a long history of safety for egg-grown vaccines. However it is
known that influenza viruses cultivated in eggs can be contaminated
with other viral agents and there has been a recent example of con-
tamination of a candidate pandemic vaccine virus with avian adenovi-
rus. These recommendations have therefore been revised in view of
the findings with egg-grown viruses, the increasing use of mammalian
cells for virus isolation and vaccine production, and the improved
methods for detecting extraneous agents.
Influenza pandemic alerts occurred in 1997 (H5N1 virus), 1999 (H9N2
virus) and 2003 (H5N1 and H7N7 viruses), when avian influenza
viruses caused serious illness and, on occasion, death in humans. E

103

6952 ECE TEXT 103 8/21/05, 10:57

 Experience gained from these events has illustrated that different
strategies for the production and clinical use of vaccine may be
needed in response to a pandemic.
• It may be necessary to generate a vaccine virus from a highly
pathogenic virus by use of reverse genetics.
• Initially, reference reagents for testing vaccine potency may not be
available;
• A vaccine with an adjuvant may be desirable.
• Monovalent vaccine may be preferred to conventional trivalent
vaccine.
• Whole virus vaccine may be preferred to a split or subunit
preparation.
• A different dosing strategy may be needed (i.e. two doses).
It is important to develop WHO recommendations for production
and quality control of the vaccine that reflect the special needs of a
pandemic.
Although the technology is still under development, the use of re-
verse genetics for vaccine virus development is likely to affect
interpandemic and pandemic vaccines alike. This technology involves
transfecting mammalian cells with plasmids coding for influenza
virus genes in order to produce a virus reassortant. Production of
reassortants by reverse genetics is similar in concept to traditional
methods, but there are some quality issues which should be taken into
account.
• The influenza virus haemagglutinin and neuraminidase genes may
be derived from a variety of sources (an egg isolate, an isolate in
cells approved or not approved for human vaccine production,
virus present in clinical specimens).
• The reassortant virus will have been generated in mammalian cells.
• In some countries, reassortants produced by reverse genetics may
be classified as “genetically modified organisms” and vaccine pro-
duction should comply with national Contained Use regulations
(although the final inactivated product will not be a genetically
modified organism).

Part A. Manufacturing recommendations

A.1 Definitions
A.1.1 Proper name
The proper name shall be “influenza vaccine (whole virion,
E inactivated)” or “influenza vaccine (split virion, inactivated)”,

104

6952 ECE TEXT 104 8/21/05, 10:57

 “influenza vaccine (surface antigens, inactivated)” or “influenza vac-
cine (inactivated, adjuvanted)” translated into the language of the
country of use.
The use of the proper name should be limited to vaccines that satisfy the
recommendations specified below.

A.1.2 Descriptive definition

Influenza vaccine is a sterile, aqueous suspension of a strain or strains
of influenza virus, type A or B, or a mixture of these two types, which
have been grown individually in embryonated hen’s eggs or in mam-
malian cells. Four types of influenza vaccine are available:
(i) a suspension of whole virus particles inactivated by a suitable
method;
(ii) a suspension treated so that the virus particles have been
partially or completely disrupted by physicochemical means
(split vaccine);
(iii) a suspension treated so that the preparation consists predomi-
nantly of haemagglutinin and neuraminidase antigens (subunit
vaccine);
(iv) a suspension of inactivated influenza virus particles, split or sub-
unit components formulated with an adjuvant.
The preparation should satisfy all the requirements formulated below.

A.1.3 Choice of vaccine strain

The World Health Organization reviews the world epidemiological
situation twice annually and if necessary recommends new vaccine
strain(s) in accordance with the available evidence.
Such strains, or those antigenically related to them, should be used
in accordance with the regulations in force in the country
concerned.
It is now common practice to use reassortant strains that give high yields of
the appropriate surface antigens. Reassortant strains for vaccine produc-
tion have the surface glycoproteins (haemagglutinin and neuraminidase) of
the recommended reference virus and the internal proteins of a high-growth
donor virus.
These recommendations shall also apply to the subsequent production
and quality control of reassortant vaccine viruses produced by reverse
genetics.
The passage history of the parent and reassortant virus strains should
be approved by the national regulatory authority. E

105

6952 ECE TEXT 105 8/21/05, 10:57

 A.1.4 Reference reagents
WHO reference antigens for strain characterization of influenza virus
are preparations that are antigenically representative of viruses iso-
lated throughout the world. They may be obtained from one of the
WHO Collaborating Centres for Reference and Research on Influ-
enza (see Appendix 2).
Candidate influenza vaccine viruses are preparations antigenically
representative of a virus strain likely to be included in a current
vaccine. They may be wild-type viruses or reassortant viruses with
surface antigens appropriate for the current recommendations for
vaccine strains. They are distributed on demand when a new virus
appears and the likelihood of its spreading throughout the world
makes its inclusion in a vaccine desirable. These preparations may be
obtained from one of the custodian laboratories listed in Appendix 2.
Antigen reagents for standardization of vaccine potency contain a
calibrated quantity of haemagglutinin antigen of influenza virus mea-
sured in mg/ml. The calibrations are performed by international col-
laborative study using single radial immunodiffusion tests (8) with
purified virus of known haemagglutinin antigen concentration. The
reference antigen and antiserum reagents are used to calibrate the
haemagglutinin content of inactivated influenza vaccines by an in
vitro immunodiffusion test. These reference haemagglutinin antigens,
together with the specific antihaemagglutinin sera, may be obtained
for the purpose of such tests from one of the custodian laboratories
listed in Appendix 2.

A.1.5 Terminology
Master seed lot: A quantity of virus, antigenically representative of a
WHO-recommended strain, that has been processed at one time to
assure a uniform composition and is fully characterized. It is used for
the preparation of working seed lots. The master seed lot and its
passage level are approved by the national authority.
Working seed lot: A quantity of fully characterized virus of uniform
composition that is derived from a master seed lot by a number of
passages that does not exceed the maximum approved by the national
regulatory authority. The working seed lot is used for the production
of vaccines.
Cell seed: A quantity of well-characterized cells of human or animal
origin stored frozen in liquid nitrogen in aliquots of uniform compo-
sition derived from a single tissue or cell, one or more of which would
E be used for the production of a master cell bank.

106

6952 ECE TEXT 106 8/21/05, 10:57

 Master cell bank: A quantity of fully characterized cells of human or
animal origin derived from the cell seed stored frozen in liquid nitro-
gen in aliquots of uniform composition, one or more of which may be
used for the production of a manufacturer’s working cell bank. The
testing performed on a replacement master cell bank (derived from
the same clone or from an existing master or working cell bank) is the
same as for the initial master cell bank, unless a justified exception is
made.
Working cell bank (WCB): A quantity of cells of uniform composition
derived from one or more ampoules of the master cell bank, which
may be used for the production cell culture. In normal practice, a cell
bank is expanded by serial subculture up to a passage number (or
population doubling, as appropriate) selected by the manufacturer, at
which point the cells are combined to give a single pool and preserved
cryogenically to form the manufacturer’s WCB. One or more of the
ampoules from such a pool may be used for the production cell
culture.
Production cell culture: A cell culture derived from one or more
ampoules of the manufacturer’s WCB and used for production of the
live influenza virus.
Single harvest: A quantity of virus suspension derived from either a
group of embryonated eggs or a culture of mammalian cells that were
inoculated with the same virus working seed lot, incubated together
and harvested together in one session.
Monovalent virus pool: A pool of a number of single harvests of a
single virus strain processed at the same time.
Final bulk: The finished vaccine prepared from one or more monova-
lent pools present in the container from which the final containers are
filled. It may contain one or more virus strains.
Final lot: A collection of sealed final containers that are homogeneous
with respect to the risk of contamination during filling. A final lot
must therefore have been filled in one working session from a single
final bulk.

A.2 General manufacturing requirements

The general requirements for manufacturing establishments con-
tained in good manufacturing practices for pharmaceuticals (9) and
biological products (10) should apply to establishments manu
facturing inactivated influenza vaccine, with the addition of the
following: E

107

6952 ECE TEXT 107 8/21/05, 10:57

 Details of standard operating procedures for the preparation
and testing of influenza vaccines adopted by a manufacturer together
with evidence of appropriate validation of the production process,
should be submitted for approval to the national regulatory authority.
Proposals for modification of the manufacturing/control methods
should also be submitted for approval by the national regulatory
authority.
Personnel employed in the production and control facilities should be
adequately trained and protected against accidental infection with
influenza virus (11).
High levels of biological containment are likely to be required for the
production of vaccines for use in pandemic or potential pandemic
situations. WHO will provide advice where appropriate (12) and
national or regional safety guidelines must be followed.
Standard operating procedures need to be developed for dealing with
emergencies involving accidental spillage, leakage or other dissemi-
nation of influenza virus.
The areas where processing of inactivated influenza vaccine takes
place shall be separate from those where work with live influenza
virus is performed.

A.3 Production control

A.3.1 Control of source materials
A.3.1.1 Eggs used for seed virus growth
If the vaccine is to be produced in embryonated eggs, the eggs to be
used should be from closed, specific-pathogen-free, healthy flocks.
The flock should be monitored at regular intervals for specific agents. The
agents monitored may include Mycobacterium avium, fowlpox virus, avian
leukosis virus (ALV) and other avian retroviruses, Newcastle disease virus
and other avian parainfluenza viruses, avian encephalomyelitis virus,
infectious laryngotracheitis virus, avian reticuloendotheliosis virus, Marek’s
disease virus, infectious bursal disease virus, Haemophilus paragallinarum,
Salmonella gallinarum, Salmonella pullorum, Mycoplasma gallisepticum
and Mycoplasma synoviae.
In some countries, all birds are bled when a colony is established, and
thereafter 5% of the birds are bled each month. The resulting serum
samples are screened for antibodies to the relevant pathogens. Any bird
that dies should be investigated to determine the cause of death.

A.3.1.2 Eggs used for vaccine production

If the vaccine is to be produced in embryonated eggs, the eggs should
be from healthy flocks, which are monitored by methods approved by
E local animal health authorities.

108

6952 ECE TEXT 108 8/21/05, 10:57

 As large numbers of eggs are needed for vaccine production, it is not
feasible to use eggs from specific-pathogen-free flocks.
Monitoring of flocks for avian influenza viruses is performed in some
countries.
In both situations (production of vaccine seed and production of
vaccine), the flock must not have been vaccinated with live Newcastle
disease virus vaccine. It is also recommended that eggs be obtained
from young birds.
In countries where use of live Newcastle disease vaccine is mandatory,
vaccination should take place during the first few weeks of the chickens’ life
and well before the use of flocks for supply of eggs.

A.3.1.3 Master cell bank and manufacturer’s working cell bank

If a cell line is used for the manufacture of influenza vaccines, it
should be based on the cell bank system. The national regulatory
authority should approve the master cell bank and should establish
the maximum number of passages (or population doublings) by which
the manufacturer’s WCB is derived from the master cell bank and the
maximum number of passages of the production cultures.
WHO has established a reference cell bank of Vero cells1 characterized in
accordance with the requirements produced in 1996 (13) as modified by
Annex 4 of this report (14). This should not be considered as a master cell
bank for direct use in vaccine production, but may be used to develop a
master cell bank by thorough requalification.

A.3.1.3.1 Identity test

The master cell bank should be characterized according to the WHO

Requirements for the use of animal cells as in vitro substrates for the
production of biologicals (13, 14) as they relate to continuous cell
lines, or to human diploid cells, as appropriate.
The manufacturer’s WCB should be identified by means, inter alia, of
biochemical (e.g. isoenzyme analysis), immunological and cytogenetic
marker tests, and DNA fingerprinting, approved by the national regu-
latory authority.

A.3.1.4 Cell culture medium

Serum used for the propagation of cells should be tested to demon-
strate freedom from bacteria, fungi and mycoplasmas, according to
the requirements given in sections A.5.2 and A.5.3 of the revised

1
Available to manufacturers on application to Quality Assurance and Safety of
Biologicals: Vaccines and Biologicals, World Health Organization, 1211 Geneva 27,
Switzerland. E

109

6952 ECE TEXT 109 8/21/05, 10:57

 Requirements for Biological Substances No. 6 (15) and from infec-
tious viruses. Suitable tests for detecting viruses in bovine serum are
given in Appendix 1 of the Recommendations for Poliomyelitis
Vaccine (Oral) (16).
Where approved by the national regulatory authority, alternative tests for
bovine viruses may be used.
As an additional monitor of quality, sera may be examined for freedom from
phage and endotoxin.
Irradiation may be used to inactivate potential contaminant viruses.
The sources(s) of serum of bovine origin should be approved by the
national regulatory authority. The serum should comply with current
guidelines in relation to animal transmissible spongiform encephalo-
pathies (17).
Human serum should not be used. If human albumin is used, it should
meet the revised Requirements for Biological Substances No. 27
(Requirements for the Collection, Processing and Quality Control of
Blood, Blood Components and Plasma Derivatives) (18), as well as
current guidelines in relation to human transmissible encephalopa-
thies (17).
Manufacturers are encouraged to explore the possibilities of using serum-
free media for production of inactivated influenza vaccine.
Penicillin and other b-lactams should not be used at any stage of the
manufacture.
Other antibiotics may be used at any stage in the manufacture, provided
that the quantity present in the final product is acceptable to the national
regulatory authority. Nontoxic pH indicators may be added, e.g. phenol red
in a concentration of 0.002%. Only substances that have been approved by
the national regulatory authority may be added.
Trypsin used for preparing cell cultures should be tested and found
free of cultivable bacteria, fungi, mycoplasmas and infectious viruses,
especially parvoviruses appropriate to the species of animal used. The
methods used to ensure this should be approved by the national
regulatory authority.
The source(s) of trypsin of bovine origin, if used, should be approved
by the national regulatory authority. Bovine trypsin, if used, should
comply with current guidelines in relation to animal transmissible
spongiform encephalopathies (17).

A.3.1.5 Virus strains

Strains of influenza virus used in the production of inactivated influ-
E enza vaccine should be identified by historical records, which should

110

6952 ECE TEXT 110 8/21/05, 10:57

 include information on the origin of the strains and their subsequent
manipulation.
Only strains that have been isolated in embryonated hen’s eggs,
in cells derived from eggs, or in mammalian cells approved for
human vaccine production under validated laboratory conditions
should be used. The national regulatory laboratory should approve
the virus strain. It is now common practice to use reassortant strains
giving high yields of the appropriate surface antigens. However, it
has been noted that antigenic changes may occur during the deve-
lopment of high-yielding reassortants, and the absence of such
changes should be shown by haemagglutination-inhibition tests
using antibodies to the haemagglutinin of the reassortant and of wild
type-viruses.
Where reassortant strains are used, the parent high-yield strain and
the method of preparing the reassortant should be approved by the
national regulatory authority.
Where reverse genetics is used to generate the reassortant vaccine virus,
the influenza haemagglutinin and neuraminidase genes may be derived
from a variety of sources (egg isolate, mammalian cell isolate or virus in
clinical specimen). The haemagglutinin and neuraminidase genes are
expected to be free of extraneous agents associated with the wild-type
virus by virtue of the recombinant DNA technology employed. The cell
substrate used for transfection to generate the reassortant virus should be
approved for human vaccine production. The derivation of the reassortant
virus should be approved by the national regulatory authority.
Strains of virus suitable for manufacture of a vaccine for use in a pandemic
or potential pandemic should be supplied by procedures agreed by the
WHO Collaborating Centres for Reference and Research on Influenza and
the custodian laboratories for supply of candidate strains (Appendix 1) and
approved by the national regulatory authority.
If any materials of animal (non-avian) origin are used in production,
they should comply with the guidance given in the report of a WHO
Consultation on medicinal and other products in relation to human
and animal transmissible spongiform encephalitis (17) and should be
approved by the national regulatory authority.
Reference strains for antigenic analysis may be obtained from the WHO
Collaborating Centres for Reference and Research on Influenza, or other
custodian laboratories (Appendix 2).

A.3.1.5.1 Seed lot system

The production of vaccine should be based on a seed lot system. Each

seed lot should be identified as influenza virus of the appropriate
strain by methods acceptable to the national regulatory authority
(section A.1.3). The maximum number of passages between a master
seed lot and a working seed lot should be approved by the national E

111

6952 ECE TEXT 111 8/21/05, 10:57

 regulatory authority. The vaccine should be not more than one pas-
sage from the working seed lot.
Each manufacturer should identify the haemagglutinin and
neuraminidase antigens of the vaccine virus strains by suitable tests
capable of detecting biologically significant variation, as well as cross-
contamination during manipulation.

A.3.1.5.2 Tests on seed lots

Either the master or working seed virus should be shown to be free

from relevant extraneous agents by tests or procedures approved by
the national regulatory authority in accordance with the requirements
of Part A, sections 5.2 and 5.3, of the revised Requirements for
Biological Substances No. 6 (General Requirements for the Sterility
of Biological Substances) (15).
Strategies to ensure freedom from extraneous agents in the final vaccine
may involve a combination of testing the seed virus and validation of the
production process, depending on the substrate used for production. For
egg-derived vaccines, the emphasis should be on a validation process,
whereas for cell-derived vaccines, the emphasis may be on a testing
strategy. The national regulatory authority should approve the strategy
chosen.

Validation strategy
The production process should be validated to demonstrate removal and/or
inactivation of likely potential contaminating agents. Validation may be
performed using appropriate model agents.
If removal or inactivation cannot be demonstrated for a potential
contaminant, a testing strategy should be implemented.

Testing strategy
Cell-derived vaccine
The susceptibility of mammalian cells to various human pathogens should
be taken into account and this information should be used in considering a
list of potential human pathogens to be included in testing for extraneous
agents in seed virus. Pathogens to be considered could include
adenovirus, parainfluenza virus, respiratory syncytial virus, coronavirus,
rhinovirus, enterovirus, human herpesvirus 4 (Epstein–Barr virus), herpes
simplex virus, cytomegalovirus and mycoplasmas.
It is recognized that when a vaccine strain changes, there may be time
constraints that make testing seed viruses for extraneous agents
problematic, and the full results of such testing may not always be available
before further processing. The use of rapid assays (e.g. multiplex
polymerase chain reaction (PCR)) which could be applied within these time
constraints is encouraged.
If an extraneous agent is detected in a seed virus and the mammalian cells
used for production are shown to be susceptible to this agent, the seed
E virus should not be used for vaccine production.

112

6952 ECE TEXT 112 8/21/05, 10:57

 If an extraneous agent is detected in a seed virus and the mammalian
cells are not susceptible to the agent, steps should be in place to ensure
that the contaminating agent is removed and/or inactivated by the
production process. If removal or inactivation of the agent cannot be
demonstrated, appropriate and specific downstream testing at the level
of each inactivated monovalent bulk should be implemented to demon-
strate that any contaminant identified in the seed virus is absent from the
vaccine.
Egg-derived vaccine
A strategy for testing for specific potential contaminating agents may be
needed if removal or inactivation of the agent by the production process
cannot be demonstrated. The use of rapid tests (e.g. PCR) is encouraged.
If the agent is detected, the seed virus should not be used for vaccine
production.
The seed lot should be stored at a temperature lower than -60 °C,
unless it is in the lyophilized form, in which case it should be stored at
a temperature lower than -20 °C.

A.3.2 Production precautions

The manufacture of inactivated influenza vaccines should follow the
relevant guidance on good manufacturing practice (10, Annex 1) and
quality assurance (10, Annex 2) for biological products with the addi-
tion of the following:
• for egg-derived vaccines, only allantoic and amniotic fluids may be
harvested.
• b-Lactam antibiotics should not be used at any stage in the manu-
facture of the vaccine.
Minimal concentrations of other suitable antibiotics may be used.
If vaccines are produced by the splitting of the virus by chemical
means, the splitting conditions and the concentration of the chemicals
used should be approved by the national regulatory authority. If an
adjuvant is used, the concentration of adjuvant should be approved by
the national regulatory authority.

A.3.3 Production of monovalent virus pools

A.3.3.1 Single harvests
For egg-derived vaccine, each strain of virus should be grown in the
allantoic cavity of embryonated hen’s eggs derived from healthy
flocks. After incubation at a controlled temperature, both the allan-
toic and amniotic fluids may be harvested. For mammalian-cell de-
rived vaccine, each strain of virus should be grown in cells approved
for human vaccine production. E

113

6952 ECE TEXT 113 8/21/05, 10:57

 For both cell-derived and egg-derived vaccines, a number of single
harvests of the same strain of virus may be combined to give a
monovalent virus pool. Cell-derived monovalent virus pools should
not be mixed with egg-derived monovalent virus pools.

A.3.3.2 Inactivation of monovalent virus pools

Time of inactivation. To limit the possibility of contamination,
monovalent virus pools should be inactivated as soon as possible after
their preparation. However, if delay is unavoidable, the temperature
and duration of the storage should be validated with respect to
bioburden and quality of the haemagglutinin and neuraminidase
antigens.
Validation with respect to bioburden may be omitted for cell-culture-derived
monovalent virus pools, with the agreement of the national regulatory
authority.
Before monovalent virus pools are inactivated, samples shall be
taken and tested for bacterial and fungal contamination. Limits
for bioburden should be approved by the national regulatory
authority.
Inactivation procedure. The virus in the monovalent virus pools
should be inactivated by a method that has been demonstrated to be
consistently effective in the hands of the manufacturer and has been
approved by the national regulatory authority. For egg-derived vac-
cine, the inactivation process should also have been shown, to the
satisfaction of the national regulatory authority, to be capable of
inactivating avian leukosis viruses and mycoplasmas. If the virus pool
is stored after inactivation, the temperature and duration of the stor-
age should be validated. The inactivation procedure should have been
shown to be capable of inactivating influenza viruses without destroy-
ing their antigenicity.
The usual storage temperature is 5 °C ± 3 °C.
Consideration should be given to investigating whether influenza virus
inactivation also inactivates human or avian pathogens capable of be-
coming extraneous agents. These investigations may be performed using
appropriate model agents e.g.
— egg-derived vaccines: avian leukosis virus, mycoplasma, avian
adenovirus;
— cell-derived vaccines: poliovirus, human immunodeficiency virus,
human adenovirus, parainfluenza virus, minute virus of mice.
If formalin (40% formaldehyde) or b-propiolactone (2-oxetanone) is
used, the concentration by volume should not exceed 0.1% at any
time during inactivation. Other suitable inactivating agents can also be
E used.

114

6952 ECE TEXT 114 8/21/05, 10:57

 Consideration should be given to strategies to limit the entry into the
manufacturing process of potential adventitious agents that may not be
inactivated by the influenza inactivation conditions.

A.3.3.3 Testing of control cells

A cell sample equivalent to at least 500 ml of the cell suspension, at
the concentration employed for vaccine production cultures, should
be used to prepare control cell cultures. In countries with the technol-
ogy for large-scale production, the national regulatory authority
should determine the size of the sample of cells to be examined, the
time at which the control cells should be taken from the production
culture, and how the control cells are maintained.
These control cell cultures should be incubated for at least 2 weeks
and should be examined during this period for evidence of cytopathic
changes. For the test to be valid, not more than 20% of the control
cell cultures may have been discarded for nonspecific, accidental
reasons.
If this examination or any of the tests required in this section show
evidence of the presence in a control culture of any adventitious
agent, the influenza virus grown in the corresponding inoculated cul-
tures should not be used for vaccine production.
Samples not tested immediately should be stored at -60 °C or below.

A.3.3.3.1 Tests for haemadsorbing viruses

At the end of the observation period or at the time the virus is

harvested from the production cultures, whichever is the later, at least
25% of the control cells should be tested for the presence of
haemadsorbing viruses using guinea-pig red blood cells. If the cells
have been stored, the duration of storage should not have exceeded 7
days, and the storage temperature should have been in the range of
2–8 °C. In tests for haemadsorbing viruses, calcium and magnesium
ions should be absent from the medium.
This test is usually done using guinea-pig red cells. However, in some
countries the national regulatory authority requires that additional tests for
haemadsorbing viruses should be made in other types of red cell, including
those from humans (blood group O), monkeys and chickens (or other avian
species).
The results of all tests should be read after incubation for 30 min at 0–4 °C
and again after a further incubation for 30 minutes at 20–25 °C. A further
reading for the test with monkey red blood cells should also be taken after
another incubation for 30 min at 34–37 °C.
In some countries the sensitivity of each new batch of red blood cells is
demonstrated by titration against a haemagglutinin antigen before use in
the haemadsorbtion test. E

115

6952 ECE TEXT 115 8/21/05, 10:57

 A.3.3.3.2 Tests on supernatant fluids

A sample of at least 10 ml of the pooled supernatant fluid from the

control cultures collected at the end of the observation period should
be tested in the same cell substrate, but not the same batch, as that
used for production. Additional samples of at least 10 ml should be
tested in both human and monkey cells. The samples should be inocu-
lated into bottles of these cell cultures, in such a way that the dilution
of the supernatant fluid in the nutrient medium does not exceed 1 in
4. The area of the cell sheet should be at least 3 cm2/ml of supernatant
fluid. At least one bottle of each of the cell cultures should remain
uninoculated and serve as a control.
The cultures should be incubated at a temperature of 35–37 °C and
should be observed for a period of at least 2 weeks.
The use of rapid assays (e.g. multiplex PCR), which could be conducted
within the time constraints of the procedure are encouraged.

A.3.3.3.3 Identity test

For vaccines produced in continuous cell culture the control

cells should be identified by means, inter alia, of biochemical (e.g.
isoenzyme analysis), immunological and cytogenetic marker tests
approved by the national regulatory authority.

A.3.3.4 Concentration and purification

The monovalent material should be concentrated and purified by
high-speed centrifugation or other suitable methods approved by the
national regulatory authority, either before or after the inactivation
procedure.
The aim is to separate the virus or viral components from other constituents
in either the allantoic and amniotic fluids or the mammalian cell culture fluids
as efficiently as possible. It is advisable to concentrate and purify the virus
under optimum conditions to preserve its antigenic properties.
Consideration should be given to investigating whether the concentration
and purification steps remove potential extraneous agents.

A.3.4 Control of monovalent virus pools

A.3.4.1 Effective inactivation
The inactivated and purified monovalent virus pool should be shown
not to contain viable influenza virus when tested by a method ap-
proved by the national regulatory authority. Tests for viable virus
should be conducted in eggs for egg-derived vaccine and in the mam-
malian cells used for vaccine production for cell-derived vaccine.
A suitable method for egg-derived vaccine consists of inoculating 0.2 ml of
undiluted monovalent pool and 1 : 10 and 1 : 100 dilutions of the monovalent
E pool into the allantoic cavities of groups of fertilized eggs (ten eggs in each

116

6952 ECE TEXT 116 8/21/05, 10:57

 group), and incubating the eggs at 33–37 °C for 3 days. At least eight of the
ten embryos should survive at each dosage level.
A volume of 0.5 ml of allantoic fluid is harvested from each surviving egg. The
fluid harvested from each group is pooled and 0.2 ml of each of the three
pools is inoculated, undiluted, into a further group of ten fertile eggs.
Haemagglutinin activity should not be detected in these new groups of eggs.
In some countries, alternative methods are used.
In some countries, the requirement that 80% of the embryos should survive
during incubation may be impossible to satisfy. The national regulatory
authority should then specify the requirement to be satisfied.
For mammalian-cell-derived vaccine, methods could be modelled on those
used for egg-derived vaccines with the exception that a mammalian cell
substrate with validated sensitivity should be used. The national regulatory
authority should approve the method used and specify the requirement to
be satisfied.
As testing of residual virus infectivity after inactivation may be problematic
due to aggregation, validation of test sensitivity should be performed.

A.3.4.2 Haemagglutinin content

The content of haemagglutinin in the monovalent virus pool should
be determined by a suitable and approved technique, such as single
radial immunodiffusion. In the test, an influenza reference
haemagglutinin antigen reagent or a national preparation calibrated
against it should be used for purposes of comparison (see section
A.1.4).
For adjuvanted vaccines it should be established whether the adjuvant is
compatible with the antigenic components of the vaccine and whether the
presence of adjuvant interferes with the test for haemagglutinin content. If
there is likely to be interference, this test may be performed before the
addition of adjuvant. The test should be fully validated.
There is evidence to suggest that when cell-derived vaccines are produced
from viruses isolated in eggs, the conventional reference antigen reagents
(egg-derived) are suitable for measurement of haemagglutinin content.
However the use of conventional antigen reagents may not be suitable to
measure the haemagglutinin content of mammalian cell-derived vaccines
when viruses isolated in cells are used for production. As further information
becomes available, advice will be provided by WHO.
During the early stages of production of vaccines for pandemics, there
may be no reference reagents to measure vaccine haemagglutinin
content by conventional methods. It may be necessary to use alternative
estimates of antigen content as advised by WHO and national regulatory
authorities.

A.3.4.3 Presence of neuraminidase

Vaccine should be prepared under conditions that allow retention of
detectable levels of viral neuraminidase for each strain.
In some countries, a test is included for the presence of neuraminidase
enzymatic or antigenic activity. The ratio of haemagglutinin to neurami- E

117

6952 ECE TEXT 117 8/21/05, 10:57

 nidase should be consistent for the particular virus strain and method of
vaccine production used, but the neuraminidases of different strains vary
markedly in their stability during processing.

A.3.4.4 Virus disruption (split vaccines)

Monovalent pools in which the virus has been split by chemical means
should be shown by procedures approved by the national regulatory
authority to consist predominantly of disrupted virus particles.
This test need be performed on only three samples of monovalent pool for
each vaccine strain provided that the test result is satisfactory.

A.3.4.5 Surface antigens (subunit vaccines)

The purity of monovalent pools intended for the preparation of
subunit vaccine shall be determined by polyacrylamide gel electro-
phoresis or by other suitable techniques approved by the national
regulatory authority. Mainly haemagglutinin and neuraminidase
antigens for each strain should be present.
This test need be performed on only three samples of monovalent pool for
each vaccine provided that the test result is satisfactory.

A.3.4.6 Identity
Antigenic specificity may be confirmed by an immunodiffusion or
haemagglutination-inhibition technique using appropriate specific
immune sera. The tests for haemagglutinin content (A.3.4.2) and
presence of neuraminidase (A.3.4.3) also serve as identity tests.
Reference viruses for identity tests may be obtained from reference
laboratories (Appendix 2).
Alternatively antigenic identity may be confirmed by:
— injection of vaccine into mice, chickens or other suitable animals and
demonstration of the production of antibodies to the haemagglutinin of
the influenza virus used to produce the vaccine. In addition,
demonstration of production of antibody to neuraminidase may also be
performed; or
— suitable genetic tests.
With split and subunit vaccines, the identity test may be performed
before virus disruption.

A.3.4.7 Extraneous agents

Cell-derived vaccines
If a contaminating agent is found in the working seed, mammalian
cells are not susceptible to infection by the agent (A.3.1.5.2) and
removal and/or inactivation of the agent by the production process
cannot be demonstrated, monovalent bulks should be tested to ensure
freedom from the agent. The data should be approved by the national
E regulatory authority.

118

6952 ECE TEXT 118 8/21/05, 10:57

 Egg-derived vaccines
If removal and/or inactivation of a potential contaminating agent by
the production process cannot be demonstrated, monovalent bulks
should be tested to ensure freedom from the agent. The data should
be approved by the national regulatory authority.

A.3.4.9 Purity of cell-derived vaccines

To monitor consistency in purity, monovalent virus pools derived
from mammalian cell cultures should be tested for the ratio of
haemagglutinin content: total protein. This ratio should be within the
limits approved by the national regulatory authority.
For viruses grown in continuous cell culture, the purified monovalent
pool should be tested for residual cellular DNA. The purification
process should be shown to consistently reduce the level of cellular
DNA to less than 10 ng per human dose. This test may be omitted,
with the agreement of the national regulatory authority, if the manu-
facturing process is validated as achieving this specification.

A.3.4.9 Tests for chemicals used in production

The concentration of each detergent, organic solvent and inactivating
agent remaining in the final vaccine should be determined using meth-
ods approved by the national regulatory authority. These concentra-
tions should not exceed the upper limits specified by the national
regulatory authority. For preservatives, both the method of testing
and the concentration should be approved by the national regulatory
authority.
Alternatively, tests for chemicals may be performed on the final bulk.

A.3.5 Control of final bulk

Final bulks are prepared by mixing and diluting monovalent pools of
the relevant strains. In the preparation of the final bulk, only preser-
vatives or other substances, including diluents, approved by the na-
tional regulatory authority should be added. Such substances should
have been shown by appropriate tests not to impair the safety or
effectiveness of the product in the concentrations used, and should
not be added before samples have been taken for any tests that would
be affected by their presence.
Vaccines for use during pandemics are likely to contain only one strain.

A.3.5.1 Test for content of haemagglutinin antigen

The haemagglutinin concentration in the final bulk should be deter-
mined as described in section A.3.4.2.
This test may be omitted if such a test is performed on each final lot. E

119

6952 ECE TEXT 119 8/21/05, 10:57

 A.3.5.2 Sterility tests
Each bulk should be tested for sterility by a method approved by the
national regulatory authority.
Many countries have regulations governing sterility testing. Where
these do not exist, the revised Requirements for Biological Sub-
stances No. 6 (General Requirements for the Sterility of Biological
Substances) (15) should be satisfied. If a preservative has been added
to the vaccine, appropriate measures should be taken to prevent it
from interfering with the sterility test.

A.3.5.3 Total protein

The total protein content should be not more than six times the total
haemagglutinin content of the vaccine, as determined in the test for
haemagglutinin content, but in any case not more than 100 mg of
protein per virus strain per human dose and not more than a total of
300 mg of protein per human dose.
In some countries, protein stabilizers are added to vaccine. The total
protein content should reflect such additions. For subunit vaccines, a lower
protein content is achievable, i.e. not more than 40 mg of protein per virus
strain per human dose and not more than a total of 120 mg of protein per
human dose.

A.3.5.4 Ovalbumin (egg-derived vaccines)

The ovalbumin content should be not more than 5 mg per human dose.
The amount of ovalbumin should be determined by a suitable tech-
nique using a suitable reference preparation of ovalbumin.
Values of less than 1 mg of ovalbumin per human dose are attainable and
lower limits may be set.

A.3.5.4 Adjuvant content

If an adjuvant has been added to the vaccine, its content should be
determined by a method approved by the national regulatory author-
ity. The amount and nature of the adjuvant should be within the range
shown to be clinically effective and should be approved by the
national regulatory authority.
The formulation of adjuvant and antigen should be stable and consistent.
The purity of the adjuvant should be demonstrated to be within the range
found for vaccine lots shown to be clinically effective.

A.4 Filling and containers

The requirements concerning filling and containers given in Good
manufacturing practices for biological products (10, annex 1, section
4) should apply. Single- and multiple-dose containers may be used. If
the latter are used, a suitable preservative, approved by the national
E regulatory authority, should be incorporated.

120

6952 ECE TEXT 120 8/21/05, 10:57

 A.5 Control tests on final lot
A.5.1 Identity test
An identity test should be performed by a method approved by the
national regulatory authority on at least one container from each final
lot.
The identity of the haemagglutinins in the vaccine should be deter-
mined by an immunological technique, such as immunodiffusion or
haemagglutinin inhibition, using the appropriate specific immune
serum.
In some countries, a test to identify the specific neuraminidase antigens is
also included.

A.5.2 Sterility test

Final containers should be tested for sterility as described in section
A.3.5.2.

A.5.3 Haemagglutinin content

The test for haemagglutinin antigen concentration is performed as
described in section A.3.4.2.
The vaccine should contain in each human dose at least 15 mg of
haemagglutinin of each strain used in the preparation.
In some countries, lower limits may be set, based on clinical experience.
Expression of haemagglutinin antigen content can also reflect uncertainty of
measurement by stipulating that the lower confidence interval (P = 0.95) of
the assay should be not less than 12 mg of haemagglutinin of each strain per
dose.
It may be necessary to formulate vaccine for use in a pandemic to contain
a different haemagglutinin antigen concentration. Advice will be provided
by WHO and the national regulatory authority.

A.5.4 General safety (innocuity) tests

Each filling lot should be tested for unexpected toxicity (sometimes
called abnormal toxicity) using a general safety (innocuity) test ap-
proved by the national regulatory authority.
This test may be omitted for routine lot release once consistency of
production has been well established to the satisfaction of the national
regulatory authority and when good manufacturing practices are in place.
Each lot, if tested, should pass a test for abnormal toxicity.

A.5.5 Endotoxin
A test for endotoxin should be included, e.g. the Limulus amoebocyte
lysate test. E

121

6952 ECE TEXT 121 8/21/05, 10:57

 The permissible level of endotoxin is determined by the national regulatory
authority. It is likely that the permissible level of endotoxin for mammalian-
cell-derived vaccine will be lower than that for egg-derived vaccine.

A.5.6 Inspection of final containers

Each container in each final lot shall be inspected visually, and those
showing abnormalities such as lack of integrity shall be discarded.

A.6 Records
The requirements given in section 8 of Good manufacturing practices
for biological products (10, annex 1) should apply.

A.7 Retained samples

The requirements given in section 9 of Good manufacturing practices
for biological products (10, annex 1) should apply.

A.8 Labelling
The requirements given in section 7 of Good manufacturing practices
for biological products (10, annex 1) should apply, with the addition
of the following information.
The label on the carton, the container or the leaflet accompanying the
container should state:
— that the vaccine has been prepared from virus propagated in
embryonated hen’s eggs or in mammalian cells;
— the type of cell line i.e. monkey, dog, etc. (if appropriate);
— the strain or strains of influenza virus present in the preparation;
— the haemagglutinin content in mg per virus strain, expressed as mg of
haemagglutinin per dose;
— the number of doses, if the product is issued in a multiple-dose
container;
— the influenza season for which the vaccine is intended;
— the method used for inactivating the virus;
— the name and maximum quantity of any antibiotic present in the vaccine;
— the name and concentration of any preservative added;
— the name and concentration of any adjuvant added;
— the temperature recommended during storage and transport;
— the expiry date; and
— any special dosing schedules (e.g. for a pandemic vaccine).
For a pandemic vaccine — special dosing schedules (e.g. two doses).

A.9 Distribution and transport

The requirements given in section 8 of Good manufacturing practices
for biological products (10, annex 1) should apply.

122

6952 ECE TEXT 122 8/21/05, 10:57

 A.10 Stability testing, storage and expiry date
A.10.1 Stability testing
Adequate stability studies form an essential part of vaccine deve-
lopment. The stability of the vaccine in its final form and at the
recommended storage temperatures should be demonstrated to the
satisfaction of the national regulatory authority on final containers
from at least three lots of final product.
Stability data may be presented to the national regulatory authority after use
of vaccine. Accelerated stability studies may be used.
In some countries, vaccine haemagglutinin content should comply with final
product specifications (see A.5.3) at the expiry date.
The formulation of vaccine antigens and adjuvant (if used) must be
stable throughout its shelf-life. Acceptable limits for stability should
be agreed with national authorities.
When any changes are made in the production process that may affect
stability of the products, the vaccine produced by the new method
should be shown to be stable.

A.10.2 Storage conditions

Inactivated influenza vaccine should be stored at a temperature of
2–8 °C.
If other storage conditions are used, they should be fully validated and
approved by the national regulatory authority.

A.10.3 Expiry date

The expiry date should be fixed with the approval of the national
regulatory authority, and should take account of the experimental
data on stability of the vaccine.
In general, the expiry date should not exceed 1 year from the date of issue
by the manufacturer because the strains used in one year’s vaccine may
not be appropriate the next year.

Part B. Requirements for national control

authorities
B.1 General
The general recommendations for control laboratories, contained in
the Guidelines for national authorities on quality assurance of bio-
logical products should apply (10, Annex 2).
The national regulatory authority should give directions to manufac-
turers concerning the influenza virus strains to be used, the E

123

6952 ECE TEXT 123 8/21/05, 10:57

 haemagglutinin content, whether or not neuraminidase is present,
and the recommended human dose.

B.2 Release and certification

A vaccine lot shall be released only if it fulfils national requirements
and/or Part A of these Recommendations. A statement signed by the
authorized official of the national regulatory authority should be pro-
vided at the request of the manufacturing establishment and should
certify that the lot of vaccine in question satisfies all national require-
ments as well as Part A of these Recommendations. The certificate
should state the number under which the lot was released by the
national controller, and the number appearing on the labels of con-
tainers. Importers of influenza vaccine (inactivated) should be given a
copy of the official national release document. The purpose of the
certificate is to facilitate the exchange of inactivated influenza vaccine
between countries.
An example of a suitable certificate is given in Appendix 3.

B.3 Clinical evaluation of influenza vaccines

In the case of a new manufacturer, the national regulatory authority
should assess the safety and immunogenicity of the vaccine by arrang-
ing for studies in human volunteers of one or more of the lots of
vaccine that have satisfied the above-mentioned requirements. Such
studies shall include the assessment of the immune responses and
adverse reactions in various age groups.
In the case of a significant change in the manufacturing process, clinical
studies may also be required by the national regulatory authority.
Some national authorities require a limited clinical evaluation for licensing
purposes whenever a new vaccine strain is introduced.

Authors
The first draft of these revised Recommendations for Influenza Vaccine (Inacti-
vated) was prepared by the following WHO consultants:
Dr R. Dobbelaer, Biological Standardization, Louis Pasteur Scientific Institute of
Public Health, Brussels, Belgium; Dr R. Levandowski, Center for Biologics Evalua-
tion and Research, Food and Drug Administration, Rockville, MD, USA; and Dr J.
Wood, National Institute for Biological Standards and Control, Potters Bar, Herts.,
England.
A second draft was prepared after an informal WHO Consultation held in Ferney-
E Voltaire, France (10–11 July 2003), attended by the following participants:

124

6952 ECE TEXT 124 8/21/05, 10:57

 Dr T. Colegate, Influenza Technical Affairs Manager, Powderject Pharmaceuticals,
Liverpool, England; Dr R. Dobbelaer, Head, Biological Standardization, Louis
Pasteur Scientific Institute of Public Health, Brussels, Belgium; Dr G. Grohmann,
Therapeutic Goods Administration, Woden ACT, Australia; Dr A. Hampson, WHO
Collaborating Centre for Reference and Research on Influenza, Victoria, Australia;
Dr O. Kistner, Director Virology, Baxter BioScience, Baxter Vaccine AG, Orth/
Donau, Austria; Dr R. Levandowski, Center for Biologics Evaluation and Research,
Food and Drug Administration, Rockville, Rockville, MD, USA; Professor N.V.
Medunitsin, Director, Tarasevic State Research Institute for Standardization and
Control of Medical Biological Preparations, Moscow, Russian Federation; Dr G.
Schild, Emeritus Professor, Imperial College School of Medicine, London, England;
Dr J. Shin, Senior Research Scientist, Division of Viral Products, Biological Evalu-
ation Department, Korea Food and Drug Administration, Seoul, Republic of Korea,
Dr M. Tashiro, Department of Virology III, National Institute of Infectious Diseases,
Tokyo, Japan; Dr T. van der Stappen, National Institute of Public Health
and the Environment, Bilthoven, the Netherlands; Dr R. Winsnes, Norwegian Medi-
cines Agency, Section for Vaccines, Oslo, Norway; Dr J. Wood, Division of Virol-
ogy, National Institute for Biological Standards and Control, Potters Bar, Herts.,
England; Dr Zhou Tiequn, Deputy Director, Department of Viral Vaccines, National
Institute for the Control of Pharmaceutical and Biological Products, Temple of
Heaven, Beijing, People’s Republic of China.

Acknowledgements
Acknowledgements are due to the following experts for their useful and advice,
following comments received on the second draft of these recommendations:
Dr A. Hampson, WHO Collaborating Centre for Reference and Research on
Influenza, Parkville, Victoria, Australia; Dr M. Tashiro, National Institute of Infectious
Diseases, Tokyo, Japan; Dr J. Wood, National Institute for Biological Standards
and Control, Potters Bar, Herts., England.

References
1. WHO Expert Committee on Biological Standardization. Twentieth report.
Geneva, World Health Organization, 1968, Annex 2 (WHO Technical Report
Series, No. 384).
2. WHO Expert Committee on Biological Standardization. Twenty-fifth report.
Geneva, World Health Organization, 1973, Annex 1 (WHO Technical Report
Series, No. 530).
3. WHO Expert Committee on Biological Standardization. Twenty-ninth report.
Geneva, World Health Organization, 1978, Annex 3 (WHO Technical Report
Series, No. 626).
4. WHO Expert Committee on Biological Standardization. Thirtieth report.
Geneva, World Health Organization, 1979, Annex 3 (WHO Technical Report
Series, No. 638). E

125

6952 ECE TEXT 125 8/21/05, 10:57

 5. WHO Expert Committee on Biological Standardization. Forty-first report.
Geneva, World Health Organization, 1991, Annex 2 (WHO Technical Report
Series, No. 814).
6. Brown F et al. Inactivated influenza vaccines prepared in cell culture.
Developments in Biological Standardization, 1999, 98:1–214.
7. Prevention and control of influenza: recommendations of the Advisory
Committee on Immunization Practices (ACIP). Morbidity and Mortality
Weekly Report, 1999, 48:1–28.
8. Wood JM et al. An improved single-radial-immunodiffusion technique for
the assay of influenza haemagglutinin antigen: application for potency
determinations of inactivated whole virus and subunit vaccines. Journal of
Biological Standardization, 1977, 5:237–247.
9. WHO Expert Committee on Specifications for Pharmaceutical Preparations.
Thirty-second report. Geneva. World Health Organization, 1992, Annex 1
(WHO Technical Report Series, No. 823).
10. WHO Expert Committee on Biological Standardization. Forty-second report.
Geneva, World Health Organization, 1992 (WHO Technical Report Series,
No. 822).
11. Biosafety guidelines for personnel engaged in the production of vaccines
and biological products for medicinal use. Geneva, World Health
Organization, 1995 (WHO/CDS/BVI/95.5).
12. WHO interim biosafety risk assessment. Production of pilot lots of
inactivated influenza vaccines from reassortants derived from avian
influenza viruses. Geneva, World Health Organization, 2003 (WHO/CDS/
CSR/RMD/2003.5).
13. WHO Expert Committee on Biological Standardization. Forty-seventh report,
1998. Geneva, World Health Organization, 1998, Annex 6 (WHO Technical
Report Series, No. 878).
14. WHO Expert Committee on Biological Standardization. Fifty fourth report,
2003. Geneva, World Health Organization, 2005, Annex 4 (WHO Technical
Report Series, No. 927).
15. WHO Expert Committee on Biological Standardization. Twenty-fifth report.
Geneva, World Health Organization, 1973, Annex 4 (WHO Technical Report
Series, No. 530).
16. WHO Expert Committee on Biological Standardization. Fifty-first report.
Geneva, World Health Organization, 2002, Annex 1 (WHO Technical Report
Series, No. 904).
17. WHO Guidelines on Transmissible Spongiform Encephalopathies. Geneva,
World Health Organization, 2003 (WHO/BCT/QSD/03.01).
18. WHO Expert Committee on Biological Standardization. Forty-third report.
Geneva, World Health Organization, 1994 (WHO Technical Report Series,
No. 840).

126

6952 ECE TEXT 126 8/21/05, 10:57

 Appendix 1
Summary protocol for influenza vaccine
(inactivated) (master/working seed lot Type A or
Type B)
The model summary protocol that follows is provided as general
guidance to manufacturers. It is not intended to constrain them in
the presentation of data relevant to the complete review of the
quality control tests performed on the vaccine. It is important to
note that satisfactory test results do not necessarily imply that the
vaccine is safe and effective, since many other factors must be taken
into account, including the characteristics of the manufacturing
facility.
Name and address of manufacturer
Laboratory reference no. of lot
Date when the processing was completed

Information on manufacture
Virus used to inoculate eggs or cells for the manufacture of the lot:
(a) strain and substrain
(b) passage level
(c) source and reference no.
(d) remarks
Results of sterility test
Results of tests for extraneous agents
Results of tests on adjuvant (if any)
Conditions of storage

Monovalent virus pool Type A or Type B

Name and address of manufacturer
Laboratory reference no. of virus pool
Virus used to inoculate eggs or cells.
(a) master seed strain and source
(b) passage level of master seed
(c) working seed lot, reference no. and source
Date of inoculation E

127

6952 ECE TEXT 127 8/21/05, 10:57

 Date of harvesting allantoic or amniotic fluids or cell culture fluids
Storage conditions before inactivation
Date of inactivation
Time of inactivation
Method of inactivation
Concentration of inactivating agent
Storage conditions after inactivation
Concentration/purification procedure
Antibiotics used during preparation, if any
Identification of adjuvant added, if any

Tests on monovalent pool1

Test for absence of viable influenza virus
No. of eggs or cell culture vessels inoculated
Incubation time and temperature
Date of test
Results

Determination of haemagglutinin content

Method
Date of determination
Results

Tests for presence of neuraminidase (if performed)

Method
Date of test
Results

Virus disruption (for split vaccine)

Method
Date
Results

1
If there are more than four virus pools in the monovalent pool, the relevant data should
E be given on a separate sheet.

128

6952 ECE TEXT 128 8/21/05, 10:57

 Purity (for subunit vaccine)
Method
Date
Results

Purity (for cell-derived vaccine)

Method
Date
Results

Identity tests
Method
Date of test
Results

Test for extraneous agents (if performed)

Method
Date
Results

Final bulk
Name and address of manufacturer
Identification of final bulk
Identification of monovalent virus pool used to prepare final bulk
Date of manufacture

Control of final bulk

Preservative(s) added and concentration
Any other substances added and concentration

Determination of haemagglutinin content

Method
Date of determination
Results E

129

6952 ECE TEXT 129 8/21/05, 10:57

 Sterility
Date of test
Results

Total protein content

Method
Date of test
Results

Ovalbumin content (egg-derived vaccines)

Method
Date of test
Results

Test for residual DNA (if performed)

Method
Date
Results

Test for adjuvant (if performed)

Method
Date
Results

Tests for chemicals used

Date of tests
Results

Final lot
Identity test
Method
Date of test
E Results

130

6952 ECE TEXT 130 8/21/05, 10:57

 Sterility
Method
Date of test
Results

Determination of haemagglutinin content

Method
Date of determination
Results

Innocuity (if performed)

No. and species of animals
Doses injected
Period of observation
Date of test
Results

Endotoxin content
Method
Date of test
Results

Inspection of final container

Results

Other tests
Additional comments (if any)
A sample of a completed final container label and package insert
should be attached.

Certification by producer
Name of head of production of the final vaccine
Certification by head of the quality assurance department
taking overall responsibility for production and control of the final
vaccine:
I certify that lot no . . . of influenza vaccine (inactivated), whose num-
ber appears on the label of the final container, meets all national E

131

6952 ECE TEXT 131 8/21/05, 10:57

 requirements1 and satisfies Part A of the Requirements for Biological
Substances No. 17, revised 1990.
Signature:
Name (typed):
Date:

Certification by the national controller

If the vaccine is to be exported, provide a copy of the certificate from
the national regulatory authority as described in section B.2, a label of
a final container, and a leaflet of instructions to users.

1
If any national requirement(s) is (are) not met, specify which one(s) and indicate why
E release of the lot has nevertheless been authorized.

132

6952 ECE TEXT 132 8/21/05, 10:57

 Appendix 2
Reference laboratories
WHO collaborating centres for reference and research
on influenza
Centers for Disease Control, Atlanta, GA, USA
National Institute for Medical Research, Mill Hill, London, England
National Institute for Infectious Disease, Tokyo, Japan
WHO Collaborating Center for Reference and Research on
Influenza, Melbourne, Australia

Custodian laboratories for candidate influenza vaccine

viruses and antigen reagents for vaccine potency
National Institute for Biological Standards and Control, Potters Bar,
Herts., England
Center for Biologics Evaluation and Research, Food and Drug
Administration, Bethesda, MD, USA
Therapeutic Goods Administration, Canberra, Australia
National Institute for Infectious Diseases, Tokyo, Japan
Centers for Disease Control, Atlanta, GA, USA
National Institute for Medical Research, Mill Hill, London, England
WHO Collaborating Centre for Reference and Research on Influ-
enza, Melbourne, Australia

Laboratories performing calibration of haemagglutinin

content
National Institute for Biological Standards and Control, Potters Bar,
Herts., England
Center for Biologics Evaluation and Research, Food and Drug
Administration, Bethesda, MD, USA
Therapeutic Goods Administration, Canberra, Australia
National Institute for Infectious Diseases, Tokyo, Japan

133

6952 ECE TEXT 133 8/21/05, 10:57

 Appendix 3
Model certificate for the release of influenza
vaccine (inactivated)1

The following lots of influenza vaccine (inactivated) produced by

_________2 in _________3 whose numbers appear on the labels of the
final containers, meet all national requirements,4 Part A of the Rec-
ommendations for Influenza Vaccine (Inactivated) (revised 2003)5
and the recommendatons for good manufacturing practice and qual-
ity assurance for biological products.6

Lot Number Date of last Expiry lot

potency test by number
manufacturer

As a minimum, this certificate is based on examination of the manu-

facturing protocol.
The number of this certificate is:
The Director of the National Control Laboratory (or Authority as
appropriate):7
Name (typed):
Signature:
Date:

1
To be provided by the national regulatory authority of the country where the vaccines
have been manufactured, on request by the manufacturer.
2
Name of manufacturer.
3
Country.
4
If any national requirement(s) is (are) not met, specify which one(s) and indicate why
release of the lot(s) has nevertheless been authorized by the national regulatory
authority.
5
Published in WHO Technical Report Series, No. 927, 2005, Annex 3 and with the
exception of the provisions on shipping, which the national regulatory authority may not
be in a position to control.
6
Published in WHO Technical Report Series, No. 822, 1992, Annexes 1 and 2.
E 7
Or his or her representative.

134

6952 ECE TEXT 134 8/21/05, 10:57