Transcript ACE Meeting - American College of Epidemiology

Impact of Mandatory and/or Universal
Immunization for HPV on the Epidemiology
of Cervical Cancer
Melinda Aldrich, Moderator
Eduardo Franco, McGill University
Patricia Buffler, University of California, Berkeley
Annual Meeting of the American College of Epidemiology
International Epidemiological Association
September 17, 2007
Frequently asked questions concerning the
implementation of HPV vaccination
 How long does the protection last?
 Is the vaccine protective in women with prior exposure
to HPV?
 Will cross-type protection last?
 Will other types replace HPV 16 and 18?
 Should men be vaccinated to protect women from
cervical cancer?
 Which groups should be prioritized to receive
vaccination?
 Is screening going to change after HPV vaccination?
American College of Epidemiology
25th Annual Meeting, Ft. Lauderdale
September 15-18, 2007
Roundtable on HPV vaccination and cervical cancer
Co-sponsored by the International Epidemiological Association
Expected impact of HPV vaccines on current
cervical cancer screening practices
Eduardo L. Franco
Professor, Departments of Epidemiology and Oncology
Director, Division of Cancer Epidemiology
McGill University, Montreal, Canada
Points to cover:
 Evidence in favour of HPV vaccination: basic facts
 Is screening needed after vaccination?
 Expected effects of vaccination on the burden of
precancerous lesions and cervical cancer: A plea for
universal vaccination
 Loss of cytology screening performance due to
reduction in lesion prevalence: quantitative and
qualitative effects
 Advantages of HPV testing as primary screening
test followed by cytologic triage
Incidence of Invasive Cervical Cancer
Age-standardized (world population 1960) rates per 100,000 women per year
Source: GLOBOCAN 2002 (Ferlay et al., 2004)
Relative risks for associations between
HPV and cervical cancer in case-control
studies of first generation
NAH: non-amplified DNA hybridization
PCR: polymerase chain reaction
Proportion of a given cancer that is preventable by
elimination of the exposure to the causal agent
Cancer prevention target
HPV infection and cervical
cancer
Smoking and lung cancer
Chronic HBV infection and
liver carcinoma
Alcohol drinking and oral
cancer
Hormone replacement therapy
and endometrial cancer
Environmental tobacco smoke
and lung cancer
Population attributable risk
> 95%
75%-85%
10%-30% (low risk areas)
50%-90% (high risk areas)
25%-70%
15%-50%
15%-20%
(mucosal and cutaneous PVs of humans and primates)
Species A7: HPV
18 and related
Species A10:
HPVs 6, 11
and related
Species A9:
HPV 16 and
related
(cutaneous PVs
of humans)
(cutaneous PVs
of humans)
De Villiers et al., Virology 2004
Vaccine Composition %
53.5
HPV 16
Monovalent: 53.5
Bivalent: 70.7
17.2
HPV 18
6.7
HPV 45
Trivalent: 77.4
2.9
HPV 31
Tetravalent: 80.3
2.6
HPV 33
Pentavalent: 82.9
2.3
HPV 52
HPV 58
HPV 35
Hexavalent: 85.2
2.2
Heptavalent: 87.4
1.4
Octavalent: 88.8
1.3
HPV 59
HPV 56
1.2
HPV 51
1.0
0.7
HPV 39
HPV 68
HPV 73
HPV 82
Nonavalent: 90.1
Decavalent: 91.3
0.6
0.5
0.3
0
20
40
60
80
100
Adapted from: Munoz et al., IJC 2004
Phase IIb Efficacy Results
Endpoints
Incident Infection
HPV 16/18
Persistent Infection
All vaccine related types
CIN 2/3
HPV 16/18
External Genital Warts
HPV 6/11
GSK's Vaccine
Merck's Vaccine
(5.5 yr Follow Up
Gall et al AACR Apr 2007)
(5 yr Follow Up
Villa LL Br J Ca 2006)
98%
Not Study Aim
(95% CI: 89-100)
P/V: 43/1 events
(N=692, ATP)
100%
96%
(95% CI: 72-100)
P/V: 14/0 events
(N=775, ATP)
(95% CI: 83-100)
P/V: 45/2 events
(N=468, PP)
100%
100%
(95% CI: 33-100)
P/V: 7/0 events
(N=692, ATP)
(95% CI: 16-100)
P/V: 6/0 events
(N=468, PP)
Not Study Aim
100%
(95% CI: <0-100)
P/V: 3/0 events
(N=468, PP)
Phase III Efficacy Results
• After receiving the first dose
• Women: negative for vaccine
specific HPV types at day 0
Persistent HPV 16/18 Infection
6 month
12 month
Efficacy in preventing specific
vaccine type associated
CIN 2/3+
Efficacy in preventing
vaccine specific type associated
Condyloma, VIN, VaIN
GSK’s Vaccine
(15 mo Follow Up
Paavonen Lancet 2007)
N=18,644
Merck’s Vaccine
(3 yr Follow Up
Koutsky et al NEJM 2007
N=12,167
Garland et al NEJM 2007
N=5,455)
80%
(97.9% CI: 70-87)
76%
Not Study Aim
(97.9 % CI: 48-90)
90%
95%
(97.9 % CI: 53-99)
P/V: 21/2 events, TVC
(95% CI: 85-99)
P/V: 62/3 events, USP
100%
98%
(97.9 % CI: 74-100)
P/V: 20/0 events, PPR
(95% CI: 86-100)
P/V: 42/1 events, PPSP
Not Study Aim
95%
(95% CI: 87-99)
P/V: 81/4 events, USP
GSK studies 001 & 007
Sustained seropositivity and
High antibody levels up to 4.5 years
HPV-16
log (ELU/ml)
10000
100%
% seropositive
100%
1000
99.7%
Vaccine HPV-16 IgG
Placebo IgG
99%
99%
100%
100%
100%
17 fold
higher
100
Natural
Infection
10
6%0%
1
month 0
17%
0%
0%
10%
10%
12%
11%
12%
month 7 month 12 month 18 [M25-M32][M33-M38][M39-M44][M45-M50][M51-M53]
HPV-001
Follow-up time (months)
HPV-007
High Correlation Between CVS and Serum Antibodies
According to Age Group for GSK Vaccine
Log ratio (anti-HPV-16/total IgG) in CVS
1.3
Anti-HPV-16
1.2
Age groups
1.1
15-25 years
1.0
26-45 years
0.9
46-55 years
0.8
0.7
0.6
0.5
0.4
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
Log ratio (anti-HPV-16/total IgG) in serum
Presentation TF Schwarz, ACOG 2007
Impact on HPV 16/18-Related CERVICAL CANCER
INCIDENCE with 16/18 Vaccine and Vaccine Lifetime
Coverage Starting at Age 12 Years
5
Incidence per 100,000
No Vaccination
12-yo females
4
12-26 year old females
3
2
1
0
0
10
20
30
40
Years Since Vaccination
Elbasha EH, Dasbach EJ, Insinga RP. Emerg Infect Dis. 2007 Jan;13(1):28-41.
Impact on HPV 16/18-Related CIN 2/3 INCIDENCE
with 16/18 Vaccine and Vaccine Lifetime Coverage
Starting at Age 12 Years
Incidence per 100,000
100
80
No Vaccination
12-yo females
60
12-26 year old females
40
20
0
0
10
20
30
40
Years Since Vaccination
Elbasha EH, Dasbach EJ, Insinga RP. Emerg Infect Dis. 2007 Jan;13(1):28-41.
Public health benefits of different cervical cancer prevention
strategies for Brazil (Goldie et al., Vaccine 2007)
Applied to 70% of the eligible female population is shown. The range represents the
minimum and maximum reductions achieved for each strategy using 50 good-fitting
parameter sets, and the bold line represents the mean reduction achieved.
Risk Factor
Sexual Behaviour
Primary Prevention
HPV Infection and CIN
Cancer Precursor
Secondary Prevention
Incidence Rate
Invasive Cancer
Tertiary Prevention
Death
Mortality Rate
Franco et al., Vaccine, 2006
Is screening needed after
vaccination?
• Yes!!!
– Vaccines protect mostly against HPVs 16 and
18 which cause at most 75% of all cervical
cancers
– Vaccination is for pre-exposure prophylaxis;
most women will continue to rely on screening
Adoption of HPV vaccination:
Which approach to use,
opportunistic or universal?
• What must not happen:
– To adopt only opportunistic vaccination and maintain
cervical cancer screening based on frequent cytology,
which benefit only women with health care access
• If this happens:
– Resources will be wasted and there may be no
further reductions in morbidity and mortality from
cervical cancer
– Lack of equitable access to prevention
Expected short-term outcomes
Settings with organized or opportunistic Pap
screening:
• Reductions of case loads of ASC, LSIL, and HSIL
to be triaged or managed; reductions of
colposcopy referrals
• Plausible estimates: 40% for those vaccinated
against 16/18 and 50% for those protected
against 6/11/16/18
Franco et al., Vaccine, 2006
Expected short-term outcomes
• Reductions in case loads a function primarily of
two factors:
– Uptake of HPV vaccination by successive
cohorts of adolescents and young women
– Time it will take for protected women to reach
screening age
• Impact on case loads initially minimal for women
vaccinated between the ages of 11 and 18 years
Franco et al., Vaccine, 2006
Expected long-term outcomes
Settings with organized or opportunistic Pap
screening:
• Reduction of cervical cancer burden unlikely to be
observed for at least a decade because of the
latency for averted HSILs to progress to invasive
lesions
• Potential problems with opposite effects:
– 1) Lack of equitable access to benefit: High vaccine uptake
may happen among women who will eventually be
compliant with screening recommendations
– 2) Non-compliance with screening because of perception
that vaccine is fully protective
Franco et al., Vaccine, 2006
Expected long-term outcomes
Lack of equitable access to benefit:
• Like mothers, like daughters…
– Young women who are vaccinated are likely to
comply with screening later in life
– Initial enthusiasm with reduction in cervical
abnormalities; however, because of their high
compliance with screening these women would not
be likely to develop cervical cancer
– Non-vaccinated women less likely to be screened 
their lesions will progress undetected  cytology
surveillance oblivious to their occurrence until cancer
is diagnosed
Franco et al., Vaccine, 2006
Loss of screening performance
due to vaccination
• As successive cohorts of women are
vaccinated:
– Reduction in prevalence of cytological abnormalities
– End result: decrease in positive predictive value of
cytology
– Increase in false positive rates will lead to nonrigorous diagnostic work-up
– Impact on cytotechnician training and quality
assurance
1.00
0.90
Predictive Value
0.80
0.70
0.60
Lesion No lesion
0.50
0.40
Lesion No lesion
Test +
51
198
Test +
51
18
Test -
49
9702
Test -
49
882
Total
100
9900
Total
100
900
0.30
0.20
PPV
0.10
NPV
0.00
0
0.1
0.2
0.3
0.4
0.5
Prevalence
Assumptions: constant 51% sensitivity and 98% specificity (as per Nanda et al., 2000)
Franco et al., Vaccine 2006
Possible qualitative changes in
Pap cytology performance
• Sensitivity will be negatively affected:
– Today’s typical case load: approximately 10% of all
smears contain abnormalities that are serious enough
to merit slide review
– Reduction in lesion prevalence  fatigue will set in
given expectation that abnormalities will be rare 
smears may not be read as thoroughly  more false
negatives
– End result: further decline in the PPV of cytology
– (some of the lowest estimates of Pap sensitivity are in
frequently screened, low risk populations of developed
countries)
Franco et al., Vaccine 2006
Possible qualitative changes in
Pap cytology performance
• But specificity may suffer as well…
– Decrease in signal-to-noise ratio of cytology  due to
rarity of squamous abnormalities and koilocytotic atypias
(the signal) inflammatory changes or reactive atypias
(the noise) may be overcalled
– Could be aggravated by cytotechnician’s fear that
relevant abnormalities will be missed
– Heightened awareness of the potential for false-negative
diagnoses may lead to more false-positive reports 
loss in specificity
– End result: further decline in the PPV of cytology
Franco et al., Vaccine 2006
Joint effects of changes in sensitivity, specificity, and
lesion prevalence on the PPV of a screening test
Positive Predictive Value
1.0
1.0
Prev=
50%
1.0
Prev=
20%
0.8
0.8
0.8
0.6
0.6
0.6
0.4
0.4
0.4
0.2
0.2
0.2
0.0
0.0
0.0
1.0 0.0
0.8
0.2
0.4
0.6
0.8
Prev=
5%
1.0
1.0 0.0
0.8
0.2
0.4
0.6
0.8
Prev=
1%
1.0 0.0
0.8
0.6
0.6
0.6
0.4
0.4
0.4
0.2
0.2
0.2
0.0
0.0
0.0
0.0
0.2
0.4
0.6
0.8
1.0 0.0
0.2
0.4
0.6
0.8
Prev=
10%
0.2
0.4
0.6
0.8
1.0
0.4
0.6
0.8
1.0
Prev=
0.5%
1.0 0.0
0.2
Sensitivity
Specificity: red: 95%, blue: 85%, and green: 75%
Graphs represent decreasing hypothetical situations of lesion prevalence: Africa and Latin
America: 10%-20%, Western countries: 5%-10%, Triage: 50%
Franco et al., Vaccine, 2006
Quantitative and qualitative
penalties on the PPV of cytology
• In consequence:
– Cytology laboratories will tend to err on the side
of conservatism to decrease risk of malpractice
suits
– Safequard: to maintain unnecessarily frequent
screening visits as policy to provide protection
against false-negatives
• Conclusion: costly and ineffective way of
combining screening to vaccination
Age standardized incidence of invasive cervical cancer and
coverage of screening, England, 197195
(Quinn et al., BMJ 1999; 318: 9048)
Cervical Cancer Mortality Rates, Crude and Adjusted for Age,
per 100,000 Women in Brazil (1979-99)
6.00
5.00
Rate
4.00
3.00
2.00
1.00
0.00
Year
Population World Standard; Sources: System of Information on
Mortality - SIM/DATASUS/MS; Foundation IBGE; Division of
Epidemiologia and Vigilância - CONPREV/INCA/MS
Crude rate
Age-standardized rate
Women who have sex
with HPV-infected men
(within weeks to months
some will develop)
Pap
Cytology
HPV
Testing
Detected
with low
sensitivity
Detected
with high
sensitivity
Detected
with
moderate
sensitivity
Detected
with high
sensitivity
HR-HPV infection
(within months some will
develop)
Persistent HR-HPV
infection
(within months to years
some will develop)
HG cervical lesions
(within months to years
some will develop)
Cervical cancer
Perceived
as cause
of low
specificity
CCCaST Study: First Screening Round Results*
Indices
Sensitivity
Specificity
PPV
NPV
Screening test
Pap
HPV
Pap
HPV
Pap
HPV
Pap
HPV
Estimate (95%CI)
55.4 (33.6-77.2)
94.6 (84.2-100)
96.8 (96.3-97.3)
94.1 (93.4-94.8)
4.1 (2.4-6.8)
7.8 (5.4-11.3)
97.6 (96.1-98.6)
99.8 (98.8-100)
* 10,171 women in Montreal and St. John’s, aged 30-69 years,
randomized to Pap or HPV as primary screening method;
estimates corrected for verification bias (Mayrand et al., NEJM in press)
Why is HPV testing an attractive option for
cervical cancer screening?
•
•
•
•
•
More sensitive than the Pap test
More “upstream” in the carcinogenic process, thus
enabling a longer safety margin for screening intervals
Can be automated, centralized, and be quality-checked
for large specimen throughput
May be more cost-effective than cytology if deployed for
high volume testing, such as in primary screening
A more logical choice for screening women vaccinated
against HPV infection
Need for assessing the basis of screening
programs following vaccination
• Pap cytology will not be the same if left as primary
test
• Solution: HPV testing as primary screening test
followed by cytologic triage:
– HPV testing more “upstream” than cytology  longer latency
safety window
– HPV testing more sensitive and not prone to the vagaries of a
test based on subjective interpretation
– HPV testing less likely to vary in sensitivity and specificity as a
function of decreasing prevalence in infections and lesions
– Cytology will perform better in the artificially high lesion
prevalence when triaging HPV+ women
Franco et al., Vaccine, 2006
Other benefits from the HPV-Pap
screening algorithm
• Dividend: A surveillance system integrated with
vaccination registries to monitor vaccine efficacy,
duration of protection, and cross-protection
• Rational approach to assuage concerns that
frequency of screening must not be changed to
avoid missing lesions caused by other oncogenic
HPV types
• Improved detection of glandular lesions
• Potential for using self-collected cervical samples
• Cytology too important to be used as screening test;
it should be reserved for diagnostic triage
Franco et al., Vaccine 2006
The case for synergy in prevention modalities
1. Screening will have to continue in the HPV
vaccination era
2. Opportunistic (as opposed to universal) vaccination
will create (further) inequity in access to benefit
3. Cytology screening performance will degrade
following vaccination
4. HPV infection surveillance will be needed postvaccination
5. Proposal: reformulate screening as an integrated
approach complementing vaccination
Policy and Ethical Aspects of
Universal HPV Vaccination
• People generally know very little about HPV and
vaccination.
• When given information, most (approximately
75%) support a vaccine that prevents cancer
(provided it is effective and safe).
• Advisory Council on Immunization Practices in
the US recommends HPV vaccination for
various groups, including women up to 26 years
old. (Texas first state to make HPV vaccination in girls
legally compulsory for students entering sixth grade;
other states considering such legislation. Legislation
subsequently rescinded in Texas.)
Policy and Ethical Aspects of
Universal HPV Vaccination
• HPV vaccination not the technological “silver
bullet”.
• “Decisions about HPV vaccine will be made in
the context of organised opposition to childhood
vaccines, allegations that vaccine risks are
downplayed, mistrust towards physicians and
drug manufacturers, disagreements over
childrearing and sexuality, and inaccurate
information on the internet.”
(Bernard Lo, BMJ. May 13, 2006.)
Policy and Ethical Aspects of
Universal HPV Vaccination
• Opposition of bio-ethicists to compulsory
vaccination programs in general.
• Tension between the utilitarian philosophy of
public health versus emphasis on cardinal
principles of beneficence, non-maleficence,
justice and autonomy.
• Ethical aspects of mandatory vaccination
programs for sexually transmitted infections
more pronounced than for casually transmitted
infections. (Not a clear mandate for universal
protection with a sexually transmitted disease.)
Policy and Ethical Aspects of
Universal HPV Vaccination
• Utilitarianism looks at the rightness or
wrongness of a decision based on its
consequences.
• Utilitarian ethics supports universal vaccination
at age 11-12 years, but as a theory it is
problematic as the main basis for the HPV public
policy decision. (Likelihood of some injustice against innocent
persons, may compromise personal integrity, may trample individual
liberty and conscience, and may be used to justify harm to minorities
in order to satisfy the preference of the majority.)
Policy and Ethical Aspects of
Universal HPV Vaccination
• Beneficence (“doing good”) is
straightforward as HPV vaccination
reduces cervical cancer.
• Theoretical harms are proposed under
non-maleficence category, e.g. decreases
in safer sex practices and inappropriate
decreased use of cytological screening
due to confusion.
Policy and Ethical Aspects of
Universal HPV Vaccination
• Concept of justice requires equitable
treatment and that the vaccine should be
available to all who need it. (There are
currently a large number of uninsured children and
adults and insurance programs for children and adults
without vaccine coverage.)
• Concept of justice also implies that those
at risk for the disease itself are those that
incur the risks of the vaccine.
Policy and Ethical Aspects of
Universal HPV Vaccination
• Inequities existing in access to current
cervical cancer screening will affect
access to HPV vaccination as well.
• Opportunistic vaccination efforts (as
opposed to universal or mandatory
access) will exacerbate these inequities.
Policy and Ethical Aspects of
Universal HPV Vaccination
• Principle of autonomy indicates that
persons should make their own choices
and decisions reflecting the concept of
inherent worth of the individual.
• In summary, HPV vaccination would
clearly be recommended due to
beneficence, but not required for school
entry due to autonomy, justice and
hypothetical non-maleficence concerns.
Policy and Ethical Aspects of
Universal HPV Vaccination
• Concern that proposals for universal
vaccination may be premature.
• Impact of HPV vaccination on the natural
history of HPV infection, i.e. the
chickenpox analogy.
Policy and Ethical Aspects of
Universal HPV Vaccination
• Lack of clarity regarding goals of HPV
vaccination programs – Is the goal
prevention of cervical cancer or prevention
of HPV infection?
• If the goal is to reduce HPV infection,
should boys be vaccinated as well as girls,
i.e. “herd immunity”?
Policy and Ethical Aspects of
Universal HPV Vaccination
• Concern that HPV vaccination programs
will promote sexual promiscuity and/or
undermine abstinence-based programs.
• Parental versus child/adolescent consent
with respect to prevention of diseases
related to adolescent’s sexuality and
reproductive health. (For adolescents under
age 18, vaccinations generally require informed
consent from both parents and the adolescent.)
Policy and Ethical Aspects of
Universal HPV Vaccination
• Questions about adequate justification for
universal or mandatory vaccine, i.e., Is
there an epidemic that needs to be
addressed? Lack of understanding about
the importance of prevention.
• Need to address existing perception of
HPV infection as benign. (Analogy with HBV
vaccination is useful, but may not be perfect because of
differences in transmission of HBV.)
Policy and Ethical Aspects of
Universal HPV Vaccination
• Growing societal concerns about the safety of
vaccines.
• Increasing number of parents/children who are
opting out of mandatory vaccination programs
for reasons other than religious beliefs, and the
impact of reduced vaccination levels on
population immunity.
• Concern about increase in number of
vaccinations now required and potential impact
on compliance with existing vaccination
programs.
Policy and Ethical Aspects of
Universal HPV Vaccination
• Concerns about influence of commercial
interests on the promotion of HPV
vaccination.
• Concerns about the effectiveness of the
available vaccines (What will happen in actual
practice?).
• Uncertainties about the impact of HPV
vaccination on the natural history of HPV.
Policy and Ethical Aspects:
Conclusions
• Only certain way to determine long term
effects of HPV vaccination will be to follow
vaccinated adolescents and women for
several decades.
• Implies a commitment to long term
surveillance.
Policy and Ethical Aspects:
Conclusions
• Societal concerns emphasize the need for
transparent policies that acknowledge
disagreements and uncertainties regarding
HPV vaccination.
• And policies that will build trust and
support for HPV vaccine in the context of
all programs to promote adolescent health.