A Prospective Study of Human Papillomavirus (HPV) Type 16 DNA Detection by Polymerase Chain Reaction and Its Association with Acquisition and Persistence of Other HPV Types
The Journal of Infectious Diseases
A Prospective Study of Human Papillomavirus (HPV) Type 16 DNA Detection by Polymerase Chain Reaction and Its Association with Acquisition and Persistence of Other HPV Types
Kai-Li Liaw () 2 3
Allan Hildesheim 1 2
Robert D. Burk 2 7
Patti Gravitt 1 2
Sholom Wacholder 1 2
M. Michele Manos 2 6
David R. Scott 2 5
Mark E. Sherman 0 2
Robert J. Kurman 0 2
Andrew G. Glass 2 6
Steven M. Anderson 2 4
Mark Schiffman 1 2
0 Department of Pathology, Johns Hopkins Medical Institutions , Baltimore, Maryland
1 Department of Cancer Epidemiology and Genetics, National Cancer Institute , Bethesda
2 Received 1 March 2000; revised 1 September 2000; electronically pub- lished 16 November 2000. University of Pittsburgh , 513 Parran Hall, 130 De Soto St., Pittsburgh, PA 15213
3 Department of Epidemiology, University of Pittsburgh , Pittsburgh, Pennsylvania
4 Laboratory Corporation of America, Research Triangle Park , North Carolina
5 Kaiser Permanente Center for Health Research , Portland, Oregon
6 Kaiser Foundation Research Institute , Oakland, California
7 Departments of Pediatrics, Microbiology and Immunology, and Epidemiology and Social Medicine, Albert Einstein College of Medicine , Bronx , New York
Human papillomavirus (HPV)-16 causes about half the cases of cervical cancer worldwide and is the focus of HPV vaccine development efforts. Systematic data are lacking as to whether the prevention of HPV-16 could affect the equilibrium of infection with other HPV types and thus alter the predicted impact of vaccination on the occurrence of cervical neoplasia. Therefore, the associations of HPV-16 detection with subsequent acquisition of other HPV types and with the persistence of concomitantly detected HPV types were examined prospectively among 1124 initially cytologically normal women. Preexisting HPV-16 was generally associated with an increased risk for subsequent acquisition of other types. HPV-16 did not affect the persistence of concomitant infections, regardless of type. These findings suggest that the prevention or removal of HPV-16 is not likely to promote the risk of infection with other types, a theoretical concern with current vaccination efforts.
Infection with oncogenic genital human papillomaviruses
(HPV) has been established as the main etiologic agent for
cervical neoplasia . Among the >30 genital types of HPV,
HPV-16 is the predominant type found in women with a
diagnosis of cervical neoplasia, especially invasive cancers  and
their precursors, high-grade intraepithelial lesions . Although
it is much less common in the general population, HPV-16
remains among the most prevalent individual HPV types .
Because of its strong causal association with about half the
cervical cancer cases worldwide , current HPV vaccine efforts
have focused on HPV-16 infection .
About 20%30% of HPV-infected women harbor multiple
HPV types [3, 7, 9]. The direct (viral) and indirect (e.g.,
immunologic) interactions among multiple HPV infections of the
cervix are not understood. Specifically, it is unclear how the
equilibrium of other HPV infections might be affected if a
typespecific vaccine successfully prevented HPV-16 infection (by
prophylactic vaccination) or removed it (by therapeutic
vaccination). Particularly, if other cancer-associated types of HPV
act differently in the absence of HPV-16 than in its presence,
the expected results of HPV-16 vaccination on cervical
neoplasia might be affected.
To investigate the association of HPV-16 with other genital
HPV infections, we conducted a prospective study among 1124
originally cytologically normal women whose cervical
specimens were collected and tested for HPV DNA by polymerase
chain reaction (PCR)based methods at 2 different times. In
this study, we examined the subsequent acquisition, clearance,
and persistence of individual types, as well as of various
groupings (including phylogenetic clades and cancer-associated types
versus low-risk types of HPV), in the presence or absence of
HPV-16 DNA at enrollment.
Materials and Methods
Study population. The study was conducted at 7 Kaiser
Permanente gynecology or health appraisal clinics in Portland,
Oregon. To focus on the target population of prophylactic HPV
vaccines, we studied the prospective effects of HPV-16 among
initially cytologically normal women. Therefore, the subjects
included in this study were selected from a cohort of 17,654 women
with normal Pap smear results and no known history of past
cervical neoplasia. Details of this cohort are given elsewhere . In
brief, these cytologically normal women were recruited between
April 1989 and November 1990. At the enrollment visit, in addition
to the routine Pap smear, a cervicovaginal lavage specimen was
collected and stored at 2707C for HPV testing. The women
returned without intervention by study staff, according to the general
recommendations of Kaiser Permanente. Thus, this study
represented a passive follow-up of women who returned at irregular
intervals, with a modal interval of 1 year.
The 1124 women in this study came from 2 sources within the
17,654-woman cohort. The main population consisted of 1035
women from an incident case-control study nested within the cohort
. To achieve greater numbers of initially HPV-infected women
for the persistence study, we added 89 women known to be
HPVpositive at enrollment who happened to have a second visit during
the 18-month recruitment. The 2 sources of study subjects are
shown in figure 1.
Women from the incident case-control study were selected from
among the 1298 of 1649 participants who had complete pairs of
specimens, consisting of the enrollment specimen and a second
lavage specimen collected at the time of selection as a case of
incident neoplasia or matched control. Forty-eight women were
excluded because our pathology review of enrollment and
preenrollment slides revealed prevalent cervical neoplasia. Of note,
second cervical specimens were not collected from women with first
Pap abnormalities diagnosed within 9 months of enrollment. Their
enrollment specimens were not tested. Such cases were suspected
to represent prevalent neoplasia, given that Kaiser Permanentes
policy discouraged Pap tests within 9 months of a previous normal
Despite the lack of closely spaced specimen collections from the
incident case-control study, we wished to address the early natural
history of women who were HPV-positive at enrollment. Therefore,
we added a supplemental group of women who attended the
participating clinics more than once during the 18-month enrollment period
and whose initial enrollment specimens were HPV DNApositive
(n p 166 of 1181 returning twice during enrollment). To study
shortterm viral persistence, the selection of subjects intentionally included
only women testing HPV-positive initially, with confirmed positivity
and adequate HPV typing (n p 123; see below).
When we combined the incident case-control and supplemental
group, we chose the best specimen pair for each woman, as the
one with full typing at a single laboratory with the longest interval
between tests. When these final exclusions were made, 1124 women
were chosen for the study. Of these, 1035 were in the incident
casecontrol study and 89 were in the supplemental study.
The subjects included 208 women with incident, abnormal
cytologic diagnoses at the time of the second specimen collection and
901 women who continued to be cytologically normal. Fifteen
women had inadequate diagnoses. We focused the study on HPV
DNA testing rather than on cytologic diagnosis of HPV infection.
Only 25 of the women with incident cytologic abnormalities had
high-grade intraepithelial lesions, and deleting this small fraction
of the study population did not affect the conclusions.
Detection of HPV DNA. The 2 paired cervical specimens from
each of the 1124 selected women were tested by means of the same
PCR-based method , in 4 collaborating laboratories over
the course of the study. All laboratories were masked to any
information regarding the subjects. In brief, in all laboratories,
cervical lavage specimens were amplified by the L1 consensus primer
pair MY09 and MY11. Amplification products were first
hybridized with a generic HPV probe mixture, to determine overall
positivity. In all laboratories, PCR amplification of a human b-globin
gene fragment was used to determine the integrity of the specimens.
Typing methods varied slightly as noted below. The specimens
from the incident case-control study were tested at 2 laboratories.
The HPV assays for specimens from the first third of the women
in that study were performed at Cetus (Emeryville, CA). The
specimens from the remaining women were completed at Albert
Einstein College of Medicine (Bronx, NY). Paired specimens were
tested in the same laboratory batches, to minimize miscellaneous
sources of error.
In both laboratories, type-specific oligonucleotide probes were
used, in addition to the generic probes to identify individual HPV
types . We demonstrated good agreement between the 2
laboratory protocols . The agreement on HPV positivity was
100% among those with single-type infections, whereas, among
those with multiple-type infections, the Cetus protocol tended to
detect more types. At Cetus, the type-specific probes included
HPV6/-11, -16, -18, -26, -31, -33, -35, -39, -40, -42, -45, -51, -52, -53,
-54, -55, -56, -57, -58, -59, -66, -68, -73, and -83, PAP155, and
W13B. Because the probes for HPV-6 and HPV-11 were originally
mixed at Cetus, detection of these 2 types was not differentiated,
and results were labeled as HPV-6/-11. However, partial retesting
data showed that HPV-6 predominated. Sixteen additional
typespecific probes were initially included for this study in the assays
performed at the laboratory at Albert Einstein College of Medicine,
to detect the DNA of HPV-2, -13, -32, -34, -61, -62, -64, -67, -69,
-70, and -72, AE2, AE5, AE6, AE7, and AE8 . As discussed
below, these additional types were eventually not considered in the
HPV testing of specimens for the 89 women in the supplemental
group proceeded as follows. The first specimens from the 1181
women who came in twice during enrollment were screened for
overall HPV positivity at Roche Biomedical (now Laboratory
Corporation of America, Research Triangle Park, NC) with the use of
MY09MY11 consensus primers. Those testing positive (n p
166) and a 10% sample of those with initial negative results were
retested for positivity and typed at Roche Molecular Systems
(Alameda, CA) by means of a reverse dot-blot strip test . The strip
test included HPV-6, -11, -16, -18, -26, -31, -33, -35, -39, -40, -42,
-45, -51 through -59, -66, -68, -73, and -83, PAP155, and W13B.
Because this study depends on HPV typing, only the 123 specimens
that tested positive at both Roche Biomedical and Roche
Molecular Systems were included.
HPV testing was not a proven diagnostic modality at the time
of this investigation. Therefore, the HPV testing results were not
revealed to the clinicians or study participants and did not influence
Statistical methods. HPV detection (overall and type-specific)
in both initial and second specimens was treated as a binary
variable (positive vs. negative) for risk estimation. Analyses were
limited to the 27 types that were typed at all laboratories (although
the phylogenetically related HPV-6 and -11 were not differentiated
in the type-specific analyses). Therefore, we decided to ignore
typespecific infections with any of the 16 additional types assayed only
at Albert Einstein.
Among the 27 HPV types that were included in the analysis, 4
types, including HPV-26, -42, and -57 and W13B, were not detected
in any of the specimens. Thus, the final analytical data set was
restricted to 22 types detected, in addition to HPV-16. We analyzed
the effects of HPV-16 detection at enrollment on the acquisition and
persistence of other nonHPV-16 types. Among women who were
initially negative for specific nonHPV-16 types, we studied
typespecific acquisition versus continued negativity. Among women who
were initially positive for specific nonHPV-16 types, we studied
persistence (positive for a given type at both times) versus clearance.
In addition to the type-specific analyses, we grouped the 22
nonHPV-16 types 2 ways, to achieve greater statistical power. First,
we grouped them according to their genetic relatedness (phylogenetic
clades), for investigation of the influence of HPV-16 on infection
status of these different groups [16, 17] (see also the online listing at
http://hpv-web.lanl.gov). Second, we grouped the types according to
their association with cervical cancer (cancer-associated types vs.
lowrisk types) .
The grouping of phylogenetic clades was based on the L1
sequences of HPV, the region that encodes the major capsid protein
 and associates with humoral immune responses to HPV
infection. The HPV types were categorized into 5 major clades,
including clade A9 (the HPV-16 group, including HPV-16, -31, -33,
-35, -52, and -58); clade A10 (the HPV-6/-11 group, including
HPV6, -11, and -55); clade A7 (the HPV-18 group, including HPV-18,
-39, -45, -59, and -68); clade A3 (the HPV-83 group, including
HPV83 and PAP155); and clade A6 (the HPV-53 group, including
HPV53, -56, and -66). Four HPV types were not included in the
phylogenetic analysis. Of the types we studied and detected, clade A8
includes only HPV-40, clade A11 includes only HPV-73, clade A5
contains only HPV-51, and HPV-54 is not yet assigned to a clade.
In the analyses by phylogenetic clade, the coding had the
following logic. In the acquisition analyses, those women who
remained negative to all of the types within a clade throughout the
study were included as the reference group, whereas those who
acquired any of the types within that clade (>1 types) were counted
once as acquisition. For the persistence analyses, those who had
an initial infection with any type within the same clade were
included. Women with infections persisting through the follow-up
were counted once as persistence, regardless of the number of
types within that clade that persisted. Only women who lost all the
initial infections within that clade were counted (once) as
clearance in the persistence analyses. The conclusions were not changed
by varying the definitions, which affected only multiple infections
(e.g., defining persistence as the repeat detection of all types
originally found from a clade).
The risk group analysis followed the same logic as the clade
analysis. Cancer-associated types other than HPV-16 were defined
as types that were present in >1% of cervical carcinomas in a large
international study that is, HPV-18, -31, -33, -35, -39, -45,
-51, -52, -56, -58, -59, and -68. The low-risk types were defined as
HPV-6/-11, -40, -53, -54, -55, -66, -73, and -83 and PAP155. Women
with multiple infections could contribute to both the
cancerassociated and low-risk group analyses.
Because of small numbers in each type-specific analysis, we
estimated the risks for acquisition and persistence associated with
the presence of HPV-16 by use of crude odds ratios (OR) without
any adjustment. However, in an attempt to disentangle the effects
of HPV-16 from confounding by sexual behavior, we also separately
evaluated the risk of acquiring a new type in women reporting
having had 01 or >2 sex partners between the 2 visits. Similarly,
we evaluated older and younger women separately.
To take the time element into consideration, we used
stratification by time period to analyze the persistence of HPV-16 and
other types. We plotted the curves for HPV persistence, assuming
that each subject was measured halfway through the follow-up time
period when they returned. Use of Kaplan-Meier (life-table)
methods instead did not alter the conclusions (data not shown). Women
returning during a time interval contributed to the denominator of
the persistence rate for that interval. Women contributed to the
numerator of the persistence rate for that interval if they still were
positive for the type found at enrollment. We examined the
persistence of the nonHPV-16 types in 2 major groups, according to
the initial presence or absence of HPV-16. Women who did not
initially have HPV-16 were further classified hierarchically as
having either >1 cancer-associated versus only low-risk types. Thus,
for the graph, each woman contributed to only 1 data point.
HPV infection occurs frequently in younger women;
therefore, women selected for this study (median ages, 28 years for
the incident case-control group and 23 years for the
supplemental group) were younger than the unselected members of
the Kaiser Permanente cohort of cytologically normal women
(median age, 34 years). The mean follow-up for the 1124 women
in the entire study was 25.4 months (range, 865 months), with
longer follow-up in the acquisition than the persistence analyses
(mean, 27.7 vs. 16.6 months). The difference in follow-up time
resulted from the dual source of study specimens. As described
previously, to assess short-term persistence, we included in our
study 89 women who attended the participating clinics more
than once during the 18-month enrollment period and whose
initial enrollment specimens were HPV DNApositive.
As presented in table 1, women infected initially with
HPV16 were, in general, more likely to acquire another type of HPV
subsequently (overall OR, 1.9; 95% confidence interval [CI],
0.74.7). Specifically, an initial infection with HPV-16
significantly increased the risk for the subsequent acquisition of an
infection with HPV-18, -39, -45, -59, or -66. On the other hand,
none of the women initially infected with HPV-16 acquired an
infection with HPV-33, -35, -40, -54, -58, or -83 or PAP155,
although some of these types (i.e., HPV-35 and -58) were
moderately common in the study population.
No clear pattern was observed for the effect of HPV-16 on
type-specific persistence of other types, although the small
numbers of concurrent infections for specific combinations were too
small to generate any reliable estimates. The risks for
persistence of HPV-18, -31, -51, -58, and -83 were increased in the
presence of preexisting HPV-16, although none of the risks was
statistically significant. In contrast, none of the initial infections
with HPV-35, -39, -40, -52, -55, -56, or -73 or PAP155 persisted
if an infection with HPV-16 had also been detected initially.
Table 2 shows the effects of the initial HPV-16 infection on
the acquisition and persistence of HPV, grouped by genetic
relatedness in the sequences of the HPV L1 region. The patterns
we observed were not consistent. In clade A9, the risk for
acquisition of a nonHPV-16 but phylogenetically related HPV
type was nonsignificantly decreased in the initial presence of
HPV-16 (OR, 0.3; 95% CI, 0.02.4). However, HPV-16 was
strongly associated with an 8-fold risk (OR, 7.9; 95% CI,
3.617.2) for subsequent acquisition of HPV-18, -39, -45, -59,
or -68 (clade A7). This result was consistent with the significant
associations between HPV-16 and subsequent infection with
HPV-18, -39, -45, and -59 in table 1.
For comparison with the effect of HPV-16, we assessed the
association of some other common types of HPV with the
acquisition of additional types in their same phylogenetic clades
(table 2). Therefore, in clades A9, A7, and A6, respectively, we
also examined the effects of HPV-31, -18, and -53 on the
subsequent infection status of phylogenetically related types. In
contrast to the result seen for HPV-16 and other members of
clade A9, HPV-31 was positively, albeit nonsignificantly,
associated with acquisition of other A9 types (OR, 1.6; 95% CI,
0.47.0). Similarly, HPV-18, a member of clade A7, was weakly
associated with the acquisition of the other members of the
same clade (OR, 2.0; 95% CI, 0.315.6). Moreover, in clade
A6, infection with HPV-53, just like HPV-16, was associated
with an 3-fold risk (OR, 2.6; 95% CI, 0.89.0) for a subsequent
acquisition of any of the types in this clade.
None of the risks for the persistence of HPV in the various
clades was significantly altered in the presence or absence of
HPV-16, or other types, at enrollment.
In addition to the effects on HPV infection status in different
clades, we investigated the impact of HPV-16 infection on
acquisition and persistence of nonHPV-16 types grouped by their
strength of association with cervical cancer (i.e.,
cancer-associated types vs. low-risk types) . In table 3, the risk for
acquisition of either a cancer-associated type (OR, 2.6; 95% CI,
1.25.7) or a low-risk type (OR, 2.7; 95% CI, 1.16.7) was
increased in the presence of an initial HPV-16 infection. The
risk for the persistence of a cancer-associated or a low-risk type
was unchanged by initial presence of HPV-16.
Figure 2 shows the probability of the initial infection with
non-16 HPV persisting over the time of follow-up in the initial
presence and absence of HPV-16, as well as the persistence of
HPV-16 itself. HPV-16 persisted longer than the other type
groups. HPV-16 persistence was unchanged in the women who
concurrently had other types as well (data not shown).
Moreover, HPV-16 appeared to have little effect on the persistence
of other types detected initially. It is noteworthy that the
persistence of cancer-associated types was similar to that of
lowrisk types, when HPV-16 was not detected initially.
Relatively few women were examined at intervals of 127
months. Anecdotally, among women with HPV-16 initially who
returned later than 27 months, 3 of 7 remained positive for
HPV-16. The 2 women with longest HPV-16 persistence, 42 and
47 months, had histologically confirmed high-grade lesions.
In ancillary analyses, we explored whether age or sexual
behavior might explain the association of HPV-16 initial detection
with increased risk of acquisition of other HPV types. We
divided women into older versus younger (median age, 27 years).
Overall, HPV acquisition was more likely for younger (20.3%
Risk for HPV type,
previous status for HPV-16
OR (95% CI)
OR (95% CI)
risk for HPV type,
previous status for
selected HPV type
OR (95% CI)
OR (95% CI)
for any type) than for older women (5.5%). The 9 older women
who were HPV-16positive at enrollment resembled younger
women in that their acquisition rate was 22.2%. Therefore, the
relative increase in acquisition associated with the presence of
HPV-16 was seen among the older women (OR, 5.2; 95% CI,
1.026.6) but not among the younger women (OR, 1.0; 95%
CI, 0.33.0). The lack of effect of HPV on the persistence of
other types was seen in both age strata, although viral
persistence of any type other than HPV-16 was slightly more likely
among the older women (46.3% vs. 37.3%, not statistically
With regard to sexual behavior, we divided women according
to the number of sex partners they reported having between
enrollment and the collection of the second specimen (when a
questionnaire was administered). For 104 women, including all
the women in the supplemental group, this information was
missing. The presence of HPV-16 at baseline was strongly and
significantly predictive of subsequent acquisition of both
highrisk and low-risk groups among women reporting 0 or 1 partner
(OR, 5.6 and 6.3, respectively). The associations resulted from
very low acquisition rates among HPV-negative women with
01 partner (the low-risk group analogous to the older women
in the age analysis), not from especially high acquisition among
HPV-positive women with 01 partner. Women reporting >2
partners since enrollment had much higher acquisition rates
than did those with 01 partner, regardless of HPV-16 status
(analogous to the younger women in the age analysis).
Accordingly, among women with more partners, the relative risks
of HPV acquisition associated with HPV-16 detection were
weak and nonsignificant.
As a final point, we investigated the impact of the clearance
of HPV-16 on the acquisition of other types. We found that,
among the 61 women with HPV-16 initially, those without
HPV16 in the second specimen had the same risk of acquisition of
another type of HPV at that second time as did those who
remained HPV-16positive (OR, 1.0; 95% CI, 0.33.1).
In this large-scale prospective study, we examined the
association of an initial infection with HPV-16, the most common
cancer-associated type , with the subsequent acquisition of
another type and with the subsequent persistence of other HPV
types detected concomitantly. We found that an infection with
Risk for HPV types,
previous status for HPV-16
OR (95% CI)
OR (95% CI)
NOTE. CI, confidence interval; 2, negative; OR, odds ratio; 1, positive. Pluses and minuses in parentheses
refer to status for particular HPV type at enrollment/status at subsequent testing.
a HPV-18, -31, -33, -35, -39, -45, -51, -52, -56, -58, -59, and -68.
b HPV-6, -11, -40, -53, -54, -55, -66, -73, and -83 and PAP155.
HPV-16 is associated with a generally increased risk of
subsequent HPV acquisition but does not affect the subsequent
persistence of concomitant HPV infections.
Despite the large size of the study, specific type combinations
were uncommon, limiting statistical power, particularly for the
examination of persistence. Nor could we look at certain more
subtle points of interest. For example, our findings suggested
that the persistence or clearance of HPV-16 did not affect the
risk of acquiring another new HPV type. To gain statistical
power, we created groupings based on phylogenetic relatedness
and known associations with risk of cancer. However, forcing
possibly artificial groupings of clades and risk categories might
have obscured details of type-type interactions. Still, the broad
patterns seemed clear.
The generally increased risks for acquisition of another type
of HPV among the women who had a preexisting infection with
HPV-16 suggest the common mode of transmission. It is probable
that a previous infection with HPV-16 is merely a marker of
sexual exposure that, in turn, represents more exposure to other
types of HPV. HPV-16 detection increased the risk for the
acquisition for other HPV types among women with 01 partner
and among older women, both low-risk groups. Because we could
not differentiate when a single sex partner during follow-up was
a new one, we could not rule out the possibility that HPV-16 at
enrollment was somehow associated with high-risk new sex
partners, which could also produce a positive association.
Alternatively, the detection of HPV-16 (or any other type) may
imply that the woman is immunologically susceptible to HPV
infections generally, although the age and sex partner data argue
for a behavioral explanation.
The observations regarding type-specific HPV acquisition
were reflected in the grouped analysis by phylogenetic clades.
The risk for the acquisition of a particular HPV type in most
phylogenetic clades was elevated in the presence of a preexisting
infection with HPV-16.
As an exception, the risk of acquiring a new infection with
nonHPV-16 types in clade A9 (HPV-31, -33, -35, -52, or -58)
was not increased among those who already had an infection
with HPV-16 (OR, 0.3; 95% CI, 0.02.4), which belongs to that
clade. Before any biologic interpretations are made, chance could
explain the singular observations for the combination of
HPV16 and other types within clade A9. With that definite caveat,
there is a theoretical possibility that a previous infection with
HPV-16 could continuously elicit immunologic responses (most
likely cell-mediated responses) that reduce the subsequent point
prevalence of HPV types that are genetically most related to
HPV-16. Such protective cross-reactivity could alter the impact
of an HPV-16 vaccine. However, arguing against
cross-protection, we observed that HPV-31 was associated with a slightly
increased risk of infection with related types in clade A9, thereby
leaving the HPV-16 finding isolated and unexplained.
Regarding viral persistence, we found that an initial infection
with HPV-16 did not affect the subsequent persistence of the
concomitant infection with other types detected initially.
Although our study population was large and included additional
HPV-positive subjects, !300 had any type of HPV infection
initially, resulting in unstable estimates for type-specific persistence
analyses. Despite limited numbers, the general pattern clearly
suggested a null effect on persistence. Analyses by phylogenetic
clade or by association with cervical cancer supported our
conclusion of unaltered risks for the persistence of a non-16 HPV
in the presence of a preexisting infection with HPV-16.
The time analysis shown in figure 2 indicated that persistence
of other types is not noticeably affected by the initial presence
of HPV-16 at any given point during the follow-up. The longer
persistence of HPV-16, compared with other types
(cancerassociated and low-risk alike), suggests that the earlier findings
of a higher percentage of persistence for cancer-associated types
may be influenced predominantly by HPV-16 , although
one recent study did not reach this conclusion .
In summary, our acquisition data suggest that infection with
HPV-16 is associated with an increased risk of subsequent
acquisition of another type of HPV, through sexual behavior or
immunology-related mechanisms. The findings (with the
possible exception of the data regarding clade A9) do not support
a field effect hypothesis that the first established HPV infection
on the cervix will reduce infection with other genetically related
HPV types. The persistence data imply that concomitant
infections with HPV-16 and other types, once established on the
cervix, will not affect each other subsequently. Overall, our
findings suggest that the prevention or removal of HPV-16 is
not likely to promote the risk of infection with other types, a
theoretical concern with current vaccination efforts [8, 18]. On
the basis of particularly the persistence data, it appears that
HPV types tend to act as independent sexually transmitted
We thank Brenda Rush, Leilani Wilson, Chris Eddy, Pat Werlein
and Pauline Love (Kaiser-Permanente); Diane Cadell (Westat, Inc.);
Tracy Zhang (Cetus); W. Qu, G. Jiang, Q. Lin, and Y. Cruz (Albert
Einstein college of Medicine); and Julie Buckland (IMS) for technical
expertise. We also thank the physicians and nurse practitioners of
Kaiser-Permanente for specimen collection.
1. Human papillomaviruses . IARC Monogr Eval Carcinog Risks Hum 1995 ; 64 : 277 - 82 .
2. Bosch FX , Manos MM , Munoz N , et al. Prevalence of HPV DNA in cervical cancer in 22 countries . J Natl Cancer Inst 1995 ; 87 : 796 - 802 .
3. Bauer HM , Ting Y , Greer CE , et al. Genital human papillomavirus infection in female university students as determined by a PCR-based method . JAMA 1991 ; 265 : 472 - 7 .
4. Ho GYF , Bierman R , Beardsley L , Chang CJ , Burk RD . Natural history of cervicovaginal papillomavirus infection in young women . N Engl J Med 1998 ; 338 : 423 - 8 .
5. Kjaer SK , van den Brule AJC , Bock JE , et al. Determinants for genital human papillomavirus (HPV) infection in 1000 randomly chosen young Danish women with normal pap smear: are there different risk profiles for oncogenic and nononcogenic HPV types? Cancer Epidemiol Biomarkers Prev 1997 ; 6 : 799 - 805 .
6. Cuzick J , Sasieni P , Davies P , et al. A systematic review of the role of human papillomavirus testing within a cervical screening programme . Health Technol Assess 1999 ; 3 : 1 - 131 .
7. Herrero R , Hildesheim A , Bratti C , et al. A population-based study of human papillomavirus infection and all grades of cervical neoplasia in rural Costa Rica . J Natl Cancer Inst 2000 ; 92 : 464 - 74 .
8. Schiller JT , Lowy DR . Papillomavirus-like particles and HPV vaccine development . Semin Cancer Biol 1996 ; 7 : 373 - 82 .
9. Franco EL , Villa LL , Sobrinho JP , et al. Epidemiology of acquisition and clearance of cervical human papillomavirus infection in women from a high-risk area for cervical cancer . J Infect Dis 1999 ; 180 : 1415 - 23 .
10. Liaw KL , Glass AG , Manos MM , et al. Detection of human papillomavirus DNA in cytologically normal women and subsequent cervical squamous intraepithelial lesions . J Natl Cancer Inst 1999 ; 91 : 954 - 60 .
11. Hsing AW , Burk RD , Liaw K , et al. Inter-laboratory agreement in a PCRbased human papillomavirus DNA assay . Cancer Epidemiol Biomarkers Prev 1996 ; 5 : 483 - 4 .
12. Ting Y , Manos M. Detection and typing of genital human papillomaviruses . In: Innis M, Gelfand D , Sninsky J , White T, eds. PCR protocols: a guide to methods and application . San Diego: Academic Press , 1990 : 356 - 67 .
13. Hildesheim A , Schiffman MH , Gravitt P , et al. Persistence of type-specific human papillomavirus infection among cytologically normal women in Portland, Oregon . J Infect Dis 1994 ; 169 : 235 - 40 .
14. Qu W , Jiang G , Cruz Y , et al. PCR detection of human papillomavirus: comparison between MY09/MY11 and GP51 primer systems . J Clin Microbiol 1997 ; 35 : 1304 - 10 .
15. Gravitt PE , Peyton CL , Apple RJ , Wheeler CM . Genotyping of 27 human papillomavirus types by using L1 consensus PCR products by a singlehybridization, reverse line blot detection method . J Clin Microbiol 1998 ; 36 : 3020 - 7 .
16. Van Ranst M , Kaplan JB , Burk RD . Phylogenetic classification of human papillomaviruses: correlation with clinical manifestations . J Gen Virol 1992 ; 73 : 2653 - 60 .
17. Tachezy R , Van Ranst MA , Cruz Y , Burk RD . Analysis of short novel human papillomavirus sequences . Biochem Biophys Res Commun 1994 ; 204 : 820 - 7 .
18. Kirnbauer R , Booy F , Cheng N , Lowy DR , Schiller JT . Papillomavirus L1 major capsid protein self-assembles into virus-like particles that are highly immunogenic . Proc Natl Acad Sci USA 1992 ; 89 : 12180 - 4 .