Joint
Committee on Vaccination and Immunisation
| Minutes of the HPV Sub-group meeting, Tuesday
23 May 2006 |
In confidence
Attending:
Dr Richard Roberts (CHAIR)
Dr Syed Ahmed
Derinda Fitton
Dr Katherine French
Professor Geoff Garnett
Dr Paul Griffiths
Dr Paul Jackson
Professor Henry Kitchener
Dr Charles Lacey
Julietta Patnick
Health Protection Agency
Professor Liz Miller
Dr John Edmunds
Dr Kate Soldan
MHRA
Dr Mair Powell
Scottish Executive
Dr Elizabeth Stewart
DH
Dr David Salisbury
Dr Dorian Kennedy
Tim Elliot
Dr Karen Noakes (minutes)
MINUTES
1. Announcements and apologies
Members were welcomed to the first meeting of the JCVI subgroup on HPV
vaccine.
Apologies were received from:
Ruanne Barnabus, Leszek Borysiewicz, Jack Cuzick, Gabriel Scally, Andy
Hall and Steven Inglis.
The following members declared interests in GSK or Merck (Sanofi Pasteur
in the UK), the two companies whose products are at an advanced stage
of development.
Richard Roberts non specific, non personal
Syed Ahmed non specific, non personal
John Edmunds non specific, personal
Katherine French specific, non personal
Geoffrey Garnett non personal, non specific
Paul Griffiths personal, non specific
Henry Kitchener non personal, specific
Charles Lacey non personal, specific
Liz Miller non personal, non specific
John Edmunds non personal
2. Introduction by chair and aims of the subgroup.
The chair outlined that the JCVI HPV subgroup had been convened to provide
advice to the main JCVI committee on the evidence on the use of HPV vaccines
and their potential benefit. Advice from the sub-group meetings would
be reported to the main JCVI meeting in November 2006. The subgroup will
sit for at least two meetings (another meeting is scheduled for September
2006). Additional meetings would be dependent on whether further advice
was required from JCVI or if there were outstanding issues. Minutes of
the sub-group meeting would be published on the JCVI website.
The aim of the first meeting was to exchange information on what is already
known on this subject and identify any gaps in the information needed
in order to come to a decision on whether to introduce the vaccine.
3. Epidemiology of HPV
Results from a randomised trial of HPV testing that involved the baseline
screening of 24878 women in Manchester were presented. Cytology samples
were genotyped using the hybrid capture test.
The prevalence of HPV types was examined by age, cytology and histology
at screening entry. The prevalence of all HPV types was highest in the
20-29 year age group (in which prevalence of the commonest type 16 was
10%) and declined thereafter. Types 16 and 18 were the predominant types
across all age groups (38% of positives) but was particularly predominant
in the age group 20-29 years. When high grade cytology samples were examined,
the prevalence of HPV types was more widespread across all ages but there
is still a downward trend with age. The proportion of samples that were
type 16 and 18 increased with severity of cytology and histology.
While the age affect on HPV detection is seen in cytology that is normal
to borderline, there is far less variation in age in samples classified
as moderate and severe because of the high proportion of high grade CIN
histology, 80-90% of which contains HR HPV.
61% of high grade cytology samples were associate with types 16 and 18.
About 50% of CIN3 samples were type 16, 11% type 18, 15% type 31 and 5%
type 45. This data is consistent with published data from other countries
and is likely to be the largest dataset in the UK. The age related incidence
of HPV types is not very different from other countries, including the
US. There may be subtle differences between genotype prevalence (for instance
types 52 and 58 are higher in the Middle East compared to the rest of
the world). Type 16 is the global dominant type.
50-70% HPV infections persist for up to 10 months, falling to 15-20%
after 24 months. This is not necessarily the result of natural history
as some of these women are undergoing intervention. There is a concealed
arm for this study (no intervention) that cannot currently be revealed.
Even so some of the women is this arm will also be treated for disease.
The Manchester dataset also includes information on single and multiple
infections. High grade lesions tend to be associated with single prolonged
infections.
HPV prevalence rates are higher in this study compared with an earlier
dataset collected in Manchester in the early 1990s where 60,000 cytology
samples were screened by PCR (as opposed to Hybrid Capture) with no data
on typing available.
There are no UK data on HPV prevalence rates for females under 15 years
of age. There are some data on prevalence of genital warts in girls, the
incidence rising from 13 years of age. Alternative methods of obtaining
prevalence data was discussed such as testing urine samples. The HPA reported
that two studies to obtain HPV prevalence data from young girls (under
18) were under consideration for funding: serological testing is planned
of samples from the Preston serum bank of unlinked anonymous residual
sera collected from 5-49 year olds of both sexes (broadly representative
of the general population), and collection and testing of residual urine
and/or genital swabs from young girls entering the National Chlamydia
Screening Programme has also been proposed. Other proposed studies involve
older participants, e.g. testing of urines from 18-44 year old participants
in NATSAL (National Survey of Sexual Attitudes and Lifestyles, 2000),
and aim to provide prevalence data by other factors not available from
cervical screening studies, e.g. sexual behaviours.
Further work:
1. HK to produce a detailed summary of age specific rates from Manchester
data.
2. HK to look at HPV prevalence data in other countries compared to the
available UK data and prepare a statement.
3. The group to consider whether there are any outstanding data requirements
and consider how to obtain this information.
4. Vaccine efficacy studies JCVI/HPV(06)2
Both manufacturers (GSK and Sanofi Pasteur MSD) had submitted in confidence
unpublished data on results from their clinical trials. One vaccine is
a quadrivalent vaccine (adjuvanted with aluminium hydroxyphosphate sulfate)
which provides protection against types 6,11,16 and 18 (Sanofi Pasteur).
The other vaccine is a bivalent vaccine (adjuvanted with aluminium hydroxide
and 3-deacylated monophosphoryl lipid A (MPL)) providing protection against
types 16 and 18 (GSK). The data between the two manufacturers is not directly
comparable as they use different ELISAs.
Both vaccines, as a 3 dose primary immunisation course provide high initial
antibody levels that plateau out somewhere between 12-18 months. The vaccination
schedule for the GSK product is 0,1 and 6 months. The schedule for the
Sanofi Pasteur product is 0, 2 and 6 months. GSK also have data to show
that there is a good immune response after a 2-dose immunisation course.
Titres were higher in the 10-14 year age group compared with 15-24 years,
and amongst boys, but the significance of this is not clear.
Continued vaccine efficacy has been shown for up to 4.5 (GSK) or 5 (Sanofi
Pasteur) years. It is reasonable to assume protection for up to 10 years
based on the experience of other vaccines.
High vaccine efficacy has been observed for the vaccines for those who
were PCR negative and seronegative on day 1 (used in per protocol and
'intention to treat' analysis of the clinical trials).
There is protection in individuals PCR positive but seronegative. There
is no protection if the individual is PCR positive and seropositive (chronic
infection). Those who were PCR negative but seropositive (i.e. had cleared
the infection) also had high vaccine efficacy in preventing recurrence.
If infected with one type at the time of vaccination, there was evidence
that individuals still gained a high level of protection to the other
HPV types in the vaccine. There are some data to suggest that there may
be some reduction in progression of lesions in those who already have
CIN (vaccine efficacy of 27%) but the confidence intervals are not significant
for this group. Efficacy against type 6 and 11 related external genital
lesions in women was noted for the quadrivalent vaccine: data on efficacy
in men is expected during 2007/2008.
Some cross-protection data are available for the GSK product. Vaccine
efficacy of 55% (12-78%) against type 31 which is related to type 16,
and 94% (63-100%) vaccine efficacy against type 45 which is related to
type 18. The group were not aware of any cross protection data that had
been presented for the Sanofi Pasteur product. There are reasons to expect
a better prospect of cross protection with the vaccine product that is
conjugated to MPL (the GSK product). GSK has good data to show that using
MPL as an adjuvant enhances immunogenicity. The group noted that a paper
has been published looking at the efficacy of an HPV vaccine without an
adjuvant which produced a good T-helper response, suggesting that any
adjuvant-associated boost to immunogenicity may not be necessary for adequate
type-specific protection. The group noted that a paper has been published
looking at the efficacy of an HPV vaccine without an adjuvant which produced
a good T-helper response, suggesting that any adjuvant-associated boost
to immunogenicity may not be necessary for adequate type-specific protection.
The group noted that there was no correlate of protection for HPV vaccine.
The regulators have agreed to extrapolate the clinical trial vaccine efficacy
data for young women to adolescents based on antibody neutralising data
for the two groups. There is no vaccine efficacy data for boys but there
are immunogenicity data that infers that there would be protection for
boys.
In terms of vaccine effectiveness in the real world, a small proportion
of girls may already be infected with one or more of the HPV types. In
the Merck study, there were numerous Modified Intention to Treat (MIT)
cases where individuals became PCR positive (infected with HPV) after
the first dose of vaccine or who were protocol deviation failures. They
were still protected from other HPV vaccine types.
The group discussed whether data indicated boosters of HPV vaccine may
be required in the future. The vaccine is similar to Hepatitis B vaccine
which also protects against a persistant infection. Once primed with HPV
vaccine, boosting may occur through natural infection. This doesn't normally
happen as HPV evades detection by the immune system.
The group agreed that there was sufficient data to make an informed decision
on the efficacy of HPV vaccine. Safety data were noted. Further data on
cross protection and co-administration may be made available in the future.
Further work:
Sanofi Pasteur plan to submit further information on their product to
the Department before the next meeting.
5. Modelling the cost effectiveness of HPV vaccine JCVI/HPV(06)3
Modelling work from Imperial College on population impact of vaccination
strategies was presented. Two models have been predominantly used to model
the impact of HPV vaccine on the 4 HPV types. They have been calibrated
with US data.
There are still issues to be resolved within the model and questions
to be answered, including age of vaccination, duration of protection and
the impact of the programme on men and women. Currently, life long type
specific immunity and the rates of HSIL (High-Grade Squamous Intraepithelial
Lesion) and LSIL (Low-Grade Squamous Intraepithelial Lesion) are not able
to fit to the model. The assumptions that are made on the heterogenicity
of risk factors also significantly alter the model.
Additional data is required on age specific rates, UK sexual behaviour,
a better understanding is needed of how quickly some cases move from low
to high grade dysplasia. Rates of hysterectomy should also be included
in the model.
The currently model looks at vaccinating both boys and girls at age 12,
15 or 18. The model is based on US data on age related sexual behaviour.
In Finland, the average age of sexual debut is older but this doesn't
make much difference to the modelled outcomes. Given these unresolved
issues the model as it currently stands highlights that:
The screening assumptions used in the model based on US data are that
80% individuals are screened (from age 21 years) and that the screening
is 60% sensitive. In England, screening starts later at 25 years and the
interval is three yearly. Coverage at this age is 50% and increases with
age. On average, attendance is 82%. This highlights a problem of recruiting
a new generation to the screening programme. Starting age is lower in
Scotland and Wales. This is a GP led programme and under the current QOF
arrangements, GPs receive target payments based on a deduction of those
who declined to be screened.
Further work:
A full sensitivity model for the UK needs to be developed in order to
make calculations of cost-effectiveness. This may not be available before
the next meeting.
A paper critically reviewing 4 published cost-effectiveness studies was
presented. All are based on a US population where vaccination is more
attractive because of a less cost effective starting point in terms of
frequency of cervical screening. The design of three of the studies used
a static Markov cohort model and one (Taira at al) uses a dynamic disease
transition model. Only one study (Taira et al) looks at vaccinating males
in addition to females. Three studies (not. Kulasingam et al) adapt the
progression and screening assumption of Myers at al. Each paper assumes
vaccination of 12 year old girls. All use QALYs as an outcome measure,
all papers use the IM report for their primary data.
Vaccine coverage makes no difference to the cost effectiveness of HPV
vaccination in the static models. In the transmission model, the findings
are similar to the model used at Imperial. The vaccination of boys is
only cost effective if both coverage and vaccine efficacy is low.
The results from the models are extremely sensitive to the discount rates
used. None of the models have been validated, none of the studies compare
models results to the actual data. It was noted that this is currently
being looked at in the Goldie model and will be published in the future.
The HPA have compared the Myers model with UK screening data. In order
to do this, LSIL/HSIL has to be translated to CIN. In the Myers model,
high grade CIN and cancer is overestimated by around 50% compared with
the observed UK data. The Myers model infection rates and progression
rates both appear to be wrong when the cancer data are examined.
The Goldie model is the only study to have looked at the effect of type
replacement. The Taira model has adapted the SIS (Susceptible, Immune
Suceptible) model to a SIR (Susceptible, Immune, Recovered) model. When
this model is used herd immunity becomes beneficial.
The group did not consider that the models discussed in the review paper
were currently able to provide the information required for JCVI to make
a decision on whether the HPV vaccination programme would be cost effective.
At the current time the outputs of the models were not comparable to real
data.
It was suggested that the priority should be development of progression
models shown to fit UK data. Transmission dynamic model should then look
at the benefits of herd immunity and the vaccination of boys. Models should
also look at the benefits of introducing a catch-up campaign. There may
be an argument for the vaccination of boys as part of a catch-up but not
as part of the routine programme. It would also be useful to conduct a
study to look at the 'quality of life' as the QALY assumptions in the
published studies are poorly supported by actual data, only one unpublished
study of expert opinion which generally assume worse quality of life than
either public or patient measures.
The group commented that it would be important to determine how many
cases of cervical cancer in the UK are from women who are from overseas
and have not been screened in the UK. The group also discussed suitable
measures of R0 and whether following natural immunity whether individuals
return to being susceptible.
Further work:
Both Imperial College and the Health Protection Agency to determine
further work that needs to be carried out on the models using UK data.
6. Licensing issues
Both companies had made it known that they had filed for EMEA approval
and expected a decision in 2006 and early 2007. Data on vaccine efficacy
in warts for men would not be available to 2008. The licensure of the
vaccine would be dependent on the Companies' commitment to assess vaccine
effectiveness. . The WHO are developing a biological standard for these
vaccines, primarily for countries that have less experience and for pre-qualification
for purchase by UNICEF. Both manufacturers have declared an intention
to seek pre-qualification for UNICEF. The group discussed whether it was
likely that additional HPV types would be added to the existing HPV vaccine.
It would not be possible to do placebo-vaccine trials of vaccines containing
additional types once HPV vaccine is introduced.
7. Cervical screening programme
An update on the screening programme and future considerations was provided.
The UK programme is currently converting from smear testing to liquid
based cytology. Scotland has already completed this process. It is possible
that HPV testing will be introduced as part of a triage process in the
next few years. The initial screening test may be unable to routinely
detect HPV type. With respect to how the screening programme will respond
to the possible introduction of HPV vaccine in the UK, this would impact
once the newly vaccinated cohorts pass into the screening process which
will probably be 10-15 years later. In the future the programme may move
towards cervical cancer control rather than screening.
Typing will be important in a post vaccination era and typing kits are
already available. The cost of treating genital warts given in the paper
was noted to be much higher than published data.
Experience in Scotland, where a targeted Hepatitis B vaccination campaign
was carried out on a similar age group, suggests that higher vaccine uptake
can be achieved compared to screening rates. The introduction of HPV vaccine
may have a role in reducing inequalities in cervical cancer prevention
due to uptake of screening.
8. Potential routine immunisation programme - considerations
The UK have considerable experience running immunisation programmes and
campaigns in schools, the longest of these being the BCG programme which
required 2 interventions. The programmes were well implemented by highly
trained school nurses. Other examples include rubella vaccine for girls
at age 13 years; the MR campaign in 1994 where both sexes were vaccinated
from age 5 to 15 years and the introduction of Meningitis C vaccine between
1999 and 2001 where pupils aged 5 to 18 years were vaccinated. Experience
from these campaigns has shown that the highest vaccine coverage is achieved
in the last year of primary school. High vaccine coverage is still seen
in secondary school but this falls as the children get older. During the
MR campaign it was noted that it takes longer to vaccinate children in
primary school compared to secondary school. The time at which you choose
to vaccinate pupils is also important, generally vaccination programmes
are not welcome in the summer term. Also the vaccination schedule choice
may have some impact on its successful implementation and the programme
would ideally be delivered within a single academic year. In Scotland,
during the targeted hepatitis B campaign, the best coverage was achieved
in early secondary school where vaccination could be tied in with health
education.
Although the routine schools BCG programme has ceased and, in effect,
freed up school nursing time, school nursing services in some areas are
currently under resourced so the introduction of an HPV school programme
would be challenging.
The vaccine uptake achieved during school campaigns should be put into
the models used to estimate the impact and cost effectiveness of an HPV
programme.
It was noted that a 2-dose schedule may be easier to implement. Although
correlates of protection are being extrapolated to younger age groups,
it would be a bigger jump to move from 3 to 2 doses and extrapolate level
of protection after a 2 dose course. During the clinical trials, both
vaccines showed an incremental increase in antibody levels after each
dose of the 3-dose course. Manufacturers have no plans to look at a 2-dose
schedule. If considering the option of a 2-dose schedule, it would also
be necessary to consider the potential impact on long term protection
(which has yet to be established for a 3 dose course).
The group also discussed work that has been undertaken to look at parents'
attitudes to HPV and HPV vaccination. A Manchester population based survey
looking at parents attitudes to HPV vaccine has been published and the
Department had carried out qualitative research involving discussion groups
of parents of 8-10 year olds. Both studies found that most parents have
not heard of HPV and were not aware of the role of HPV in cervical cancer.
Parents had concerns about offering a vaccine that protects against a
sexually transmitted infection to children at a young age and the sexual
heath issues that could arise.
It was noted information on impact of vaccination on sexual activity
would be relevant to addressing parents' questions and concerns.
9. AOB
The group agreed that there were a number of uncertainties around the
data presented at the meeting and whether the data was adequate would
have to be examined before going to the main JCVI meeting.
Further work:
Before the next subgroup meeting, members would have to decide which
pieces of further research were critical to decision making and which
pieces of research were of interest in the long term but did not need
to be completed before a recommendation could be made on HPV vaccine.
Date of next meeting- - 22 September 2006
10. Papers from ACIP Meeting 21 February 06 JCVI/HPV(06)6
These papers were provided for information.
|
You
are here: 