MAJOR ARTICLE
Long-Term Persistence of Prevalently Detected
Human Papillomavirus Infections in the Absence
of Detectable Cervical Precancer and Cancer
Philip E. Castle,1 Ana Cecilia Rodrguez,4 Robert D. Burk,3 Rolando Herrero,4 Sholom Wacholder,1 Allan Hildesheim,1
Jorge Morales,4 Greg Rydzak,2 and Mark Schiffman,1 for the Proyecto Epidemiológico Guanacaste (PEG) Group 1Division
of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Department of Health and Human Services,
Bethesda, 2Information Management Services, Silver Spring, Maryland; 3Albert Einstein Cancer Center, Albert Einstein College of Medicine, Bronx, New
York; and 4Proyecto Epidemiológico Guanacaste, INCIENSA Foundation, San José, Costa Rica
Background. Detailed descriptions of long-term persistence of human papillomavirus (HPV) in the absence of
cervical precancer are lacking.
Methods. In a large, population-based natural study conducted in Guanacaste, Costa Rica, we studied a subset of
810 initially HPV-positive women with >3 years of active follow-up with >3 screening visits who had no future
evidence of cervical precancer. Cervical specimens were tested for .40 HPV genotypes using a MY09/11 L1-targeted
polymerase chain reaction method.
Results. Seventy-two prevalently-detected HPV infections (5%) in 58 women (7%) persisted until the end of
the follow-up period (median duration of follow-up, 7 years) without evidence of cervical precancer. At enrollment,
women with long-term persistence were more likely to have multiple prevalently-detected HPV infections (P ,.001)
than were women who cleared their baseline HPV infections during follow-up. In a logistic regression model,
women with long-term persistence were more likely than women who cleared infections to have another newlydetected HPV infection detectable at >3 visits (odds ratio, 2.6; 95% confidence interval, 1.2–5.6).
Conclusions. Women with long-term persistence of HPV infection appear to be generally more susceptible to
other HPV infections, especially longer-lasting infections, than are women who cleared their HPV infections.
INTRODUCTION
Although virtually all cervical cancer and its immediate
precursor lesions are caused by persistent high-risk
(HR)-human papillomavirus (HPV) infection, the
Received 14 July 2010; accepted 23 September 2010.
Potential conflicts of interest: none reported.
The following individuals participated in the PEG Group: M. Schiffman, project
officer, A. Hildesheim, co-project officer, S. Wacholder, statistician, M. Sherman
and D. Solomon, pathologists, at the National Cancer Institute, NIH, Bethesda MD;
M. C. Bratti, principal investigator, R. Herrero, co-principal investigator, A. C.
RodrÍguez, assistant principal investigator, J. Morales, colposcopist, M. Alfaro,
cytopathologist, and D. Guillén, histopathologist, at the Costa Rican Foundation for
Health Sciences, San José, Costa Rica; M. Hutchinson, cytologist, at Women and
Infants' Hospital, Providence, Rhode Island; and J. Schussler, senior programmer
and analyst, at Information Management Services, Inc., Silver Spring, Maryland.
Reprints or correspondence: Philip E. Castle, MD, MPH, Div of Cancer
Epidemiology and Genetics, National Cancer Institute, 6120 Executive Blvd, EPS
Rm 5026, MSC 7234, Bethesda, MD 20892-7234 ([email protected]).
The Journal of Infectious Diseases 2011;203:814–822
Published by Oxford University Press on behalf of the Infectious Diseases Society of
America 2011.
1537-6613/2011/2036-0001$15.00
DOI: 10.1093/infdis/jiq116
814
d
JID 2011:203 (15 March)
d
Castle et al.
converse is not true: not all persistent HR-HPV infections produce cervical precancer, and only the minority of presumably larger precancers ever invades
[1]. When HPV infections were recognized in cervical
screening only by their cytopathic (the mildly abnormal Pap smear) or histopathologic (mild dysplasia
or cervical intraepithelial neoplasia grade 1) correlates, the dogma was that roughly one-half of all infections cleared and that one-fifth to one-third of
infections persisted without progression to highgrade disease [2, 3, 4]. These fractions have proven to
be qualitatively wrong and were based on a lack of
knowledge about sequential, separate HPV infections
(practically indistinguishable microscopically or on
visual examination) that were acquired and cleared
rapidly. Importantly, most HPV infections clear or
become undetectable within a year or two [5, 6, 7, 8].
With the introduction of HPV detection into cervical
cancer screening, increasing numbers of women will
be identified with molecularly detected HPV
infection, some with persistent HPV infection—a subset of
which will not have overtly detectable, clinically relevant
disease. Quantifying and describing the frequency and nature
of HPV persistence, by HPV type, will be important to understanding its clinical significance and impact on clinical
practice and management. Here, we investigated in more
epidemiologic detail the occurrence of long-term viral persistence (LTP) in the absence of apparent cervical precancer
and cancer.
Specimen Collection
Two exfoliative cervical specimens were obtained during a single pelvic examination at baseline and all follow-up visits, as
described elsewhere [9, 10]. In brief, the first specimen was used
to make a conventional Papanicolaou smear and was then residual specimen was placed in PreservCyt (Hologic) for production of thin-layer cytology slides. A second cervical specimen
was collected and stored in specimen transport medium (STM;
Qiagen).
HPV DNA Testing
MATERIALS AND METHODS
Study Population
This population-based cohort study, the Proyecto Epidemiologico Guanacaste (PEG), included participants from
Guanacaste, Costa Rica, enrolled during the period from June
1993 through December 1994 with the approval of US National
Cancer Institute and Costa Rican institutional review boards [6,
9, 10]. Detailed methods of cohort recruitment, screening, and
follow-up have been published elsewhere [9, 10].
As previously described [9], a subcohort of sexually active
women (2624 [30.7%] of 8545) (We previously reported
2,626 for this sub-cohort but later found that 2 women had
undergone hysterectomies and therefore were not at risk of
cervical cancer) and 410 virgins were actively observed to explore risk factors for incident cervical intraepithelial neoplasia
grade 2 or more severe (CIN21). This analytic sub-cohort
represented a mixture of higher-risk women (based on enrollment screening results and sexual behavior) and was supplemented by a random sample of the remaining, lower-risk
cohort population.
Women in this actively followed subcohort were initially
observed annually, except for the 492 women with cytologic
low-grade intraepithelial lesion (LSIL) or histologic CIN1 at
enrollment, who were observed at 6-month intervals for increased safety of the participants. Women not included in the
active subcohort were seen at least once for an exit visit 5-7 years
after enrollment. Women who reported themselves as virginal at
enrollment were screened every 6 months after they reported
becoming sexually active.
Throughout the study, women with cytologic, visual, or
Cervigram evidence of high-grade cervical neoplasia underwent colposcopy. At the time of study exit (in most women, 5-7
years after enrollment), to insure the safety of women leaving
the cohort, we referred women to colposcopy for the
following reasons: (1) abnormal (atypical squamous cells of
undetermined significance or LSIL) cytologic interpretation; (2)
a positive Cervigram result during either of the past 2 screening
visits; or (3) persistent HR-HPV infection or infection with
HPV-16 or HPV-18 at either of the 2 most recent screening
visits. Finally, a 6.25% random sample of all women in the
cohort was referred for an exit colposcopy.
The MY09/M11 L1 degenerate primer polymerase chain reaction (MY09/11 PCR) method used to test STM specimens for
HPV genotypes has been described elsewhere [11]. Dot blot
hybridization using HPV genotype-specific oligonucleotide
probes was used to identify .40 individual HPV genotypes [11].
HPV genotypes 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and
68 were considered to be the primary HR-HPV genotypes [12],
and all others were considered noncarcinogenic. HPV genotypes
were also grouped according to phylogenetic species [13].
Pathology
All women with possible high-grade cervical neoplasia at any
time, detected by any screening technique (including nurse
concern on gross examination) were referred to undergo colposcopy and censored from further follow-up. Any women who
received a diagnosis of CIN21 were treated by loop electrosurgical excision procedure or by inpatient surgery, as clinically
indicated. Histology slides were reviewed by one pathologist in
the United States (M.E.S.), and if the assessment was discrepant
with the Costa Rican reading, a second US pathologist (D.S.)
reviewed the case. Only a few very difficult cases with discrepant
diagnoses from all 3 pathologists were adjudicated by joint review, occasionally with consideration of cytologic slides as well
as histology slides.
Statistical Methods
Analyses were conducted 2 ways: with the individual HPV infections and then with the women as the analytic units. We
restricted to our analysis the 2299 women who tested HPV
positive at baseline, limiting the study further to the 1056 HPVpositive women in the active cohort (ie, the intersection of active
follow-up cohort and the cohort of HPV-positive women)
(Figure 1). We excluded women in passive follow-up, with only
an enrollment and exit visit, because the natural history of their
HPV infections was largely unobserved, and therefore, the viral
outcomes of the infection could not be classified. We then excluded women who ever had CIN21 diagnosis during the study.
Infections were classified as LTP without CIN21 if (1) the
same HPV genotype was detected at enrollment and the exit
visit from the study or last visit with a valid PCR result that
occurred >3 years after enrollment, (2) there were >3 positive
tests for that HPV genotype, and (3) there were no consecutive
Long-Term Persistence of Prevalently Detected HPV Infections
d
JID 2011:203 (15 March)
d
815
Figure 1. Consort diagram to outline which women were included in the analyses of long-term human papillomavirus (HPV) persistence, starting with
the subgroup of women who tested HPV positive at baseline by polymerase chain reaction (PCR) regardless of their follow-up, compared with the 2624
women who were undergoing active follow-up). The box at the bottom with gray background and white type highlights the final analytic group. *Includes
29 infections among 18 women with ,3 years of follow-up who cleared infection.
negative results during follow-up. That is, a single intermittent
negative result between 2 positive results for a HPV genotype
was treated as a false-negative result and was recoded as
a positive result. Infections were classified as cleared if there
were consecutive (>2) visits at which the HPV genotype was
not detected and the HPV genotype was never subsequently
detected; in this category, we included 18 women with 29 infections for whom we did not have 3 years of follow-up data
(median duration, 2.4 years) but whose infections apparently
had cleared by this definition. Other infections (n 5 97) not
meeting these criteria (eg, intermittent infections—that is,
those infections that appeared to clear and that then reappeared) were excluded from this analysis because we were
uncertain of how to classify the infections.
816
d
JID 2011:203 (15 March)
d
Castle et al.
The likelihood of developing LTP without CIN21 was estimated for individual HPV genotypes and for groups of
HPV genotypes defined by cancer risk (carcinogenic vs. noncarcinogenic) and phylogenetic species. Binomial 95% confidence intervals (CIs) were calculated.
We also assigned women to hierarchical categories based on
the viral outcome (LTP or cleared) to describe factors associated
with those outcomes. Differences across categories and continuous variables were tested for statistical significance using
Pearson v2 and Kruskal-Wallis tests, respectively.
To model factors related to LTP using conditional logistic
regression, we matched each woman with >1 long-term persistent HPV infection with any women who cleared all infections
on the basis of the following 2 criteria: the number of enrollment
HPV genotypes and the age at enrollment 6 5 years. Odds ratios
(ORs) and 95% CIs were calculated as measures of association.
In model 1, we included as a covariate the number of additional,
newly detected HPV infections of any duration in follow-up. In
model 2, we included as a covariate only those newly detected
HPV infections in follow-up that were more overtly enduring:
those detectable in >3 cervical specimens collected at visits that
spanned >1 year (n.b., similar to our requirements for defining
LTP for testing positive at >3 visits). In addition, models were
adjusted for the following nuisance variables: enrollment age,
number of visits, and the number of new sexual partners during
follow-up.
We also searched our data post-hoc for examples of newly
detected HPV infections that endured for >3 years without
CIN21 and lasted until the end of the study. We required that
women were negative for a HPV genotype with >2 tests before
a newly detected infection could be considered new, or the infection occurred in women who enter the cohort as self-reported
virgins and later became sexually active.
Linux SAS software, version 9.1.3 (SAS Institute), and Stata
software, versions 8.2 and 9.0 (StataCorp), were used for these
analyses. All statistical tests were 2-sided, and P values ,.05 were
considered to be statistically significant.
RESULTS
By our definitions of viral outcomes among women with no
overt cervical disease, 810 women contributed 1403 prevalently detected HPV infections to this analysis. These women
had a mean duration of follow-up of 6.7 years (median, 7.0
years; interquartile range, 7.0-7.1 years). We excluded 74
women (9%) with 97 HPV infections (7%) with ambiguous
patterns (ie, patterns that could not be classified as having
a certain persistent or transient infection[s]). Fifty-eight (7%)
of 810 women had 72 LTP HPV infections without CIN21
(5% of 1403), and 678 (84%) of 810 women had 1234 HPV
infections (88% of 1403) that cleared. Women with single and
multiple infections detected at baseline did not differ in their
likelihood of having LTP (P 5.7). Women classified with LTP
were observed to have infection for a mean duration of 6.5
years (median, 7.0 years; range, 3.5-7.2 years) until their follow-up period ended.
At baseline, women with LTP were significantly older (mean
age, 56.4 years; median age, 58.5 years [range, 46-68 years]) than
those who cleared infection (mean age, 35.8; median age, 32.0
years [interquartile range, 25-43 years; P , .001) (Table 1).
Despite their older age, women with LTP were more likely to
have multiple infections at baseline (63.8% vs 41.3%; P 5 .001)
than were women who cleared their infections.
Women with LTP were observed for a slightly shorter duration than were women who cleared their infections (P 5.04), but
there were no significant differences in the number of follow-up
visits (P 5.2). There were no significant differences in the
number of self-reported new sexual partners (P 5.9) or number
of new HPV infections detected during follow-up (P 5.5) by
viral outcome.
Among the common HPV genotypes in the study population
(ie, those for which there were >20 infections detected at
baseline), the most likely to establish LTP by our definition were
HPV-71 (n 5 96; LTP: 15.6%; 95% CI, 9.0%-24.5%]), HPV-16
(n 5 82; LTP: 12.2%; 95% CI, 6.0%-21.3%]), and HPV-45 (n 5
37; LTP: 8.1%; 95% CI, 1.7%-21.9%]). Grouped by phylogenetic
species, a14 (which included only HPV-71 here; HPV-90 and
HPV-106 were not detected), a9 (which included HPV-16; n 5
331; LTP: 6.0%; 95% CI, 3.7%-9.2%]), and a3 (n 5 290; LTP:
5.2%; 95% CI, 2.9%-8.4%]) were the most likely to establish
long-term persistence (Figure 2). By contrast, no infections with
HPV genotypes in a5 species, which includes HR–HPV-51, established long-term viral persistence without CIN21 (96 [LTP,
0%; 95% CI, 0.0%-3.8%]).
There was no significant difference in the likelihood of longterm viral persistence between carcinogenic and non–HR-HPV
genotypes (5.2% vs 5.0%, respectively; P 5 .9, by the Fisher exact
test) in this group of women in which women with CIN21 were
excluded.
Placed into context, among all women in the actively observed subcohort, women who tested positive for HR-HPV at
baseline (n 5 636) were twice as likely to have an outcome of
a diagnosis of CIN21 during follow-up than an outcome of
LTP (without CIN21) (61 vs 27). Conversely, women who
tested positive for non–HR-HPV at baseline (n 5 431) were
4-fold less likely to have an outcome of CIN21 during
follow-up than an outcome of LTP (without CIN21) (7 vs 31;
P , .001).
Because women who had LTP without CIN21 were very
different from those who cleared infections, we matched 47
(81%) of 58 women with LTP without CIN21 to 480 (71%) of
678 women whose infections cleared on the number of enrollment HPV genotypes and the enrollment age (65 years). This
subset of 47 women with LTP had a total of 53 long-term persistent HPV infections (42 women had single LTP, 4 had 2 LTPs,
and 1 had 3 LTPs). Because matching was not restricted in any
way (eg, the number of control subjects matched to case patients
and the number of times a control subject was used in a match),
the cases had 1–163 matched control subjects, and each control
subject was matched with 1–9 case patients. The 12 case patients
with LTP for whom we could not find matches were older
(median age, 68 vs 54 years; P , .001) and had more HPV
genotypes detected at baseline (median, 4 vs 2; P , .001) than
women with LTP for who we matched to women who cleared
their infections.
In the conditional logistic model (Table 2) that adjusted on
the enrollment age, the number of follow-up visits, and whether
Long-Term Persistence of Prevalently Detected HPV Infections
d
JID 2011:203 (15 March)
d
817
Table 1. Baseline Characteristics of Women Whose Human Papillomavirus (HPV) Infections Cleared versus Those Whose HPV
Infections Persisted
Enrollment
Clearance
Persistence
(n 5 678)
(n 5 58)
P
Smoking
Never
603
(88.9)
46
(79.3)
26
49
(3.8)
(7.2)
9
3
(15.5)
(5.2)
Former
Current
.002
Oral contraceptive use
Missing
(0.1)
0
(0.0)
Never
214
1
(31.6)
38
(65.5)
Former
299
(44.1)
16
(27.6)
Current
164
(24.2)
4
(6.9)
0-1
2
67
265
(9.9)
(39.1)
4
10
(6.9)
(17.2)
3-5
173
(25.5)
7
(12.1)
>6
173
(25.5)
37
(63.8))
165
(24.3)
16
(27.6)
91
(13.4)
5
(8.6)
212
(31.3)
15
(25.9)
210
(31.0)
22
(37.9)
,.001
Parity
,.001
Age at first sexual experience
,16
16
17-18
>19
Lifetime no. of sex partners
1
238
(35.1)
26
(44.8)
2-3
266
(39.2)
27
(46.6)
>4
174
(25.7)
5
(8.6)
0-1
635
(93.7)
56
(96.6)
>2
43
(6.3)
2
(3.4)
.5
.007
Recent no. of sex partners
Income, Colones
Missing
,15,000
31
(4.6)
8
(13.8)
152
(22.4)
26
(44.8)
15,000 to ,25,000
187
(27.6)
9
(15.5)
25,000 to ,35,000
133
(19.6)
5
(8.6)
>35,000
175
(25.8)
10
(17.2)
.6
,.001
Marital status at baseline
Married or living together
Divorced, separated, widowed, or single
HPV infections at baseline
476
(70.2)
36
(62.1)
202
(29.8)
22
(37.9)
Single
398
(58.7)
21
(36.2)
Multiple
280
(41.3)
37
(63.8)
.2
.001
NOTE. Data are no. (%) of subjects, unless otherwise indicated. Kruskal-Wallis tests were used to test for differences between viral outcomes for continuous
variables, Fisher’s exact tests were used for two category variables, and Pearson’s chi-square for were used for three or more category variables.
a woman reported a new sexual partner during follow-up,
women with LTP were less likely to have had >2 sexual partners
in their lifetime at the time of enrollment, compared with
women who cleared their infections (OR for model, 0.52 [95%
CI, 0.28-0.96]; OR for model 2, 0.48 [95% CI, 0.25-0.90]).
Women with LTP were significantly more likely to have a new
infection detected at >3 time points (OR for model 2, 2.6; 95%
CI, 1.2-5.6). Covariates such as smoking, parity, and oral
818
d
JID 2011:203 (15 March)
d
Castle et al.
contraceptive use, considered cofactors for cervical cancer
development [14], were not associated with LTP without
CIN21.
Finally, among the 810 women included in this subcohort, we
observed 753 newly detected HPV infections, of which 20 persisted (in 19 women; median age, 48 years) for >3 years (median
duration, 4.3 years) to the end of the study. There were no
significant differences in the distribution of HPV genotypes by
Figure 2. Percentage of prevalently-detected human papillomavirus (HPV) infections that demonstrated long-term HPV persistence by groups of HPV
genotypes defined by phylogenetic species or cancer risk. Binomial 95% confidence intervals for groups of HPV genotypes are shown as vertical bars.
phylogenetic species or groups between prevalent and incident
LTP without CIN21 (P 5.3).
DISCUSSION
Here we described in detail the occurrence of long-term HPV
persistence without detectable precancer, >3 years after being
prevalently detected and therefore possibly already of much
longer duration than newly appearing infections. Women with
long-term HPV persistence of their prevalently detected infections were much older than women with other viral outcomes, as we reported previously with less-stringent criteria (ie,
testing positive for the same HPV genotype at baseline and 5 or 6
year anniversary visit) [15]. Some older women may be unable
to control or clear their HPV infection(s) because of age-related
immune senescence [16, 17]. We also observed an inverse
relationship between lifetime number of sexual partners and the
likelihood of having a long-term persistent HPV infection. We
believe this to be a chance association, although we cannot rule
out that increased past exposure, as measured by number of
sexual partners, might have increased the likelihood of exposure
and therefore immunity. Most interestingly, we observed that
women with long-term viral persistence were more likely to have
HPV coinfections, especially those of longer (>1 year) duration,
even after adjustment for sexual behaviors, suggesting an unmeasured susceptibility factor.
We note that there was a similar number of HPV-positive
women whose subsequent HPV patterns were classified as ambiguous and were excluded from definition of long-term viral
persistence as those women who were classified as having LTP.
On average, these women with ambiguous viral patterns were
younger than those with long-term viral persistence and were
slightly older than women who were classified as clearing their
HPV infections (data not shown). Speculatively, some of the
women with ambiguous viral patterns may also have LTP but
because they are younger and perhaps have more robust immune responses to the HPV infection, detection of their persistent infection was intermittent due to the infection being
occasionally suppressed a viral levels less than the limits of detection. Thus, our rigorous definition for persistence may have
led to identifying primarily older women and therefore certain
HPV genotypes preferentially found in cervical specimens in
older women [18].
We observed some interesting patterns of long-term HPV
persistence related to HPV phylogeny. First, we never observed
long-term persistence for almost 100 infections by HPV genotypes in the a5 species, which contain 1 certain HR-HPV genotype (HPV-51) and several borderline HR-HPV genotypes
Long-Term Persistence of Prevalently Detected HPV Infections
d
JID 2011:203 (15 March)
d
819
Table 2. Conditional Logistic Regression Model to Calculate
Odds Ratios (OR) and 95% Confidence Intervals (95%CI) as
Measures of Association of Covariates with Long-term HPV
Persistence without Disease (LTP) versus Clearance.
Model 1
OR
95% CI
Model 2
OR
95% CI
Marital Status (Enrollment)
Single (ref)
1.0
––
1.0
––
Separated or Divorced
1.1
0.33–3.5
1.1
0.31–3.4
Widowed
1.0
0.26–4.0
0.99
0.25–3.9
Single
0.45
0.16–1.3
0.51
0.18–1.4
1.0
0.48
––
0.25–0.90
Lifetime Number of Sexual Parners (Enrollment)
1
2 or more
1.0
0.52
––
0.28–0.96
Number of Infections (Follow-Up)
0
1.0
––
1 or more
1.6
0.83–3.0
Number of Infections, 3 or more positive visits (Follow-Up)
0
1.0
1 or more
2.6
––
1.2–5.6
NOTE. Women who cleared to their infections were matched to women
with LTP on the number of HPV genotypes detected at baseline and
enrollment age +/-5 years. The variables were mutually adjusted for each
other, and also adjusted for enrollment age, number of visits in follow-up, and
number of new sexual partners reported during follow-up. Model 1 included
a variable for any new HPV infection detected after baseline; Model 2 included
a variable for any new HPV infection detected after baseline that tested
positive at 3 or more visits. Ref = reference group.
that certainly can cause precancerous lesions [12, 19, 20]. Second, non–HR-HPV genotypes were equally capable of persisting
as HR-HPV genotypes, once those infections that developed
into precancerous lesions were excluded; inclusion of precancerous lesions in our definition of persistent infection would
not change our observations appreciably, except that HPV-16
would be by far the most persistent HPV genotype (data not
shown) [21]. Almost 16% of all HPV-71 infections demonstrated long-term persistence, but unlike HPV-16 infections,
these infections carry no risk of cancer. Third, HPV-16, the most
carcinogenic HPV genotype, can persist for many years without
detectable evidence of causing disease (see Table 3a for an anecdotal example).
The observed patterns raise some important biological
questions about HPV persistence. First, is viral persistence by
HR-HPV genotypes leading to the development of cervical
precancer the same phenomenon as viral persistence in the absence of disease? Or is persisting HR-HPV infection without
detectable cervical disease really a microscopic precancerous
lesion that has gone undetected? Second, what is the physical
state—episomal, integrated, or both—of the persisting viral
infection by the HR-HPV genotype? What is the viral methylation patterns related to viral outcomes? And third, do non–HRHPV genotypes persist by the same mechanism(s) as HR-HPV
genotypes? The answers to these questions may provide insights
820
d
JID 2011:203 (15 March)
d
Castle et al.
to poorly understood or unknown aspects of the natural history
of HPV infection, including whether and how often HPV can
establish a viral latent state and the molecular switches leading to
progression.
Within the same subcohort of women who were actively
observed and who lacked detectable CIN21, we found a few
examples of newly appearing infections that persisted for >3
years (data not shown; n.b., these are a subset of newly detected
HPV infections defined for model 2). The distributions of HPV
genotypes within phylogenetic species for the newly detected
infections and prevalently detected infections were roughly
similar. In general, the newly detected infections occurred in
younger women than the prevalently detected infections. Some
newly appearing infections occurred in women with prevalently detected, long-term persistent infection (Table 3a).
Whether these newly detected infections represent truly new
infections or reactivation of latent infections cannot be readily
determined.
The limitations of this analysis of prevalently-detected
women with LTP (without CIN21) are that the data are both
left and right censored—that is, we do not know the history of
the infection prior to enrollment and the outcome of the infection after exiting the study. Some women were excluded
from this analysis because they received a diagnosis of CIN21
but could have had other infections that persisted without
causing CIN21. We cannot be certain when these women first
acquired these infections in the cervix or vagina and, therefore, how long these infections were present before enrollment, only that they endured for a median of 7 years after
enrollment; anecdotally, 22 (79%) of 28 long-term persistent
HPV infections (in 25 women) were still detected during the
conduct of ancillary studies, a median of 2.2 years after the end
of the main cohort study. We also do not know longer-term
risks of cervical precancer or cancer following this phenomenon of LTP.
We recognize that the choice of PCR primers may have
influenced our result. Despite efforts to make primers systems
to target a broad range of HPV genotypes for uniform amplification, different primers undoubtedly have different HPVgenotype specificity [22] that may have influenced our findings.
Finally, with the introduction of HR-HPV DNA testing into
cervical cancer screening in older women (ie, those aged 30–35
years and older), most women with evidence of persistent HPV
infection will be at high risk of CIN21 but a small percentage of
women will be identified as persistently HPV positive for many
years without CIN21. In all likelihood, most of these women
who are repeatedly positive for any HR-HPV genotype will
actually have long-term, HPV genotype-specific viral persistence
(ie, long-term persistence is a clinical reality albeit uncommon
one).
The clinical implications and best patient management for
the approximately 1 in 20 HPV-positive women who, by our
Table 3. A Case with Prevalently- and Incidentally-detected (HPV-83 and HPV-81, respectively) Long-term Human Papillomavirus (HPV)
Persistence (A) and a Case of Long-term HPV-16 Persistence (B), Both without Evidence of Clinical Disease
A. Enrollment age: 69 years
Time since
Visit
Costa Rica
Johns Hopkins
Thin-Prep
Thin-Prep
Cervigram
PCR Test
Colposcopy
enrollment (months)
Type
Pap Result
Pap Result
US
Costa Rica
Result
Results
Impression
Screen
Screen
Reactive
Koil. Atyp
Missing
Reactive
Normal
0
13
ASC-US
ASC-US
Negative
Negative
66,71,83
59,71,82,82v,83
19
Screen
Reactive
Normal
Negative
81,83,89
28
Screen
Reactive
ASC-US
Negative
83
36
Screen
Reactive
ASC-US
Negative
32,81,83
43
Screen
Reactive
ASC-US
Negative
32,81,83
50
Screen
Reactive
Reactive
Negative
81,83
58
Screen
Reactive
Normal
66
84
Screen
Screen
Reactive
Reactive
Normal
Normal
98
Colpo
Reactive
Reactive
Negative
32,81,83
Negative
Negative
32,81,83
32,81,83
Normal*
B. Enrollment Age: 54 Years
Time Since
Visit
Costa Rica
Johns Hopkins
Thin-Prep
Thin-Prep
Cervigram
PCR Test
Colposcopy
Enrollment (months)
Type
Pap Result
Pap Result
US
Costa Rica
Result
Results
Impression
0
Screen
2
Colpo
13
Screen
Normal
ASC-US
Reactive
Negative
16
Reactive
Reactive
Normal
Negative
16
Normal
16,45
16
25
Screen
Inadequate
Normal
Negative
38
Screen
Normal
Normal
Negative
16
50
Screen
Reactive
Normal
Negative
16,59,61,67
61
84
Screen
Screen
ASC-US
Reactive
Normal
Normal
Negative
Negative
16,84
16,84
95
Colpo
103
Postcolp
105
Colpo
Reactive
Normal
Normal
Normal
Normal
NOTE. Bold face type highlights the prevalently-detected long-term persistent HPV genotype; underlined type highlights the incidentally-detected long-term
persistent HPV genotype. *Punch biopsy was diagnosed as normal. ASC-US, atypical squamous cells of undetermined significance; Koil. Atyp, koilocytotic atypia;
Reactive, reactive changes.
definition, will have long-term viral persistence without evidence of cervical disease among those with an index positive
HPV test are uncertain. In places where HR-HPV testing has
been introduced, long-term persistence without precancer has
been observed (Dr. Walter Kinney, Kaiser Permanente Northern California; personal communication) but not quantified.
Although there are certainly anecdotal experiences of persistent
mild cytological abnormalities, there are sparse reports in the
literature regarding the outcome of this phenomenon, which is
almost certainly the result of long-term HPV viral persistence.
One study [23] that excised the cervical tissue of 102 women
following >4 borderline or mildly abnormal smears and no
colposcopic evidence of high-grade disease found only 11 cases
(11%) of undiagnosed CIN2/3. Thus, although immediate
excision provides the greatest safety, almost 90% of women
would be overtreated. Another management approach would be
a more aggressive colposcopic exam, including 4-quadrant
microbiopsies and endocervical curettage to maximize disease
ascertainment without excision [24].
Funding
National Institutes of Health (N01-CP-21081, N01-CP-33061, N01-CP40542, N01-CP-50535, N01-CP-81023, intramural program, CA78527 to
RB). The Guanacaste cohort (design and conduct of the study, sample
collection, management, analysis and interpretation of the data) for the
enrollment and follow-up phases were supported [in part] by the Intramural Research Program of the National Cancer Institute, National
Institutes of Health, Department of Health and Human Services.
References
1. McCredie MR, Sharples KJ, Paul C, et al. Natural history of cervical
neoplasia and risk of invasive cancer in women with cervical intra-
Long-Term Persistence of Prevalently Detected HPV Infections
d
JID 2011:203 (15 March)
d
821
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
822
epithelial neoplasia 3: a retrospective cohort study. Lancet Oncol 2008;
9:425–34.
Ostor AG. Natural history of cervical intraepithelial neoplasia: a critical
review. Int J Gynecol Pathol 1993; 12:186–92.
Nasiell K, Roger V, Nasiell M. Behavior of mild cervical dysplasia
during long-term follow-up. Obstet Gynecol 1986; 67:665–9.
Nasiell K, Nasiell M, Vaclavinkova V. Behavior of moderate cervical
dysplasia during long-term follow-up. Obstet Gynecol 1983; 61:609–14.
Plummer M, Schiffman M, Castle PE, Maucort-Boulch D, Wheeler
CM. A 2-year prospective study of human papillomavirus persistence
among women with a cytological diagnosis of atypical squamous cells
of undetermined significance or low-grade squamous intraepithelial
lesion. J Infect Dis 2007; 195:1582–9.
Rodriguez AC, Schiffman M, Herrero R, et al. Rapid clearance of human papillomavirus and implications for clinical focus on persistent
infections. J Natl Cancer Inst 2008; 100:513–7.
Richardson H, Kelsall G, Tellier P, et al. The natural history of typespecific human papillomavirus infections in female university students.
Cancer Epidemiol Biomarkers Prev 2003; 12:485–90.
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.
Bratti MC, Rodriguez AC, Schiffman M, et al. Description of a sevenyear prospective study of human papillomavirus infection and cervical
neoplasia among 10 000 women in Guanacaste, Costa Rica [Description
de un estudio prospectivo de siete anos sobre la infeccion por el virus del
papiloma humano y el cancer cervicouterino en 10 000 mujeres de
Guanacaste, Costa Rica]. Rev Panam Salud Publica 2004; 15:75–89.
Herrero R, Schiffman MH, Bratti C, et al. Design and methods of
a population-based natural history study of cervical neoplasia in a rural
province of Costa Rica: the Guanacaste project. Rev Panam Salud
Publica 1997; 1:362–75.
Castle PE, Schiffman M, Gravitt PE, et al. Comparisons of HPV DNA
detection by MY09/11 PCR methods. J Med Virol 2002; 68:417–23.
Bouvard V, Baan R, Straif K, et al. A review of human carcinogens–part
B: biological agents. Lancet Oncol 2009; 10:321–2.
d
JID 2011:203 (15 March)
d
Castle et al.
13. Bernard HU, Burk RD, Chen Z, van Doorslaer K, Hausen H, de Villiers
EM. Classification of papillomaviruses (PVs) based on 189 PV types
and proposal of taxonomic amendments. Virology 2010; 401:70–9.
14. Castellsague X, Munoz N. Chapter 3: cofactors in human papillomavirus carcinogenesis–role of parity, oral contraceptives, and tobacco
smoking. J Natl Cancer Inst Monogr 2003; 31:20–8.
15. Castle PE, Schiffman M, Herrero R, et al. A prospective study of age
trends in cervical human papillomavirus acquisition and persistence in
guanacaste, costa rica. J Infect Dis 2005; 191:1808–16.
16. Garcia-Pineres AJ, Hildesheim A, Herrero R, et al. Persistent human
papillomavirus infection is associated with a generalized decrease in
immune responsiveness in older women. Cancer Res 2006; 66:11070–6.
17. Einstein MH, Schiller JT, Viscidi RP, et al. Clinician’s guide to human
papillomavirus immunology: knowns and unknowns. Lancet Infect Dis
2009; 9:347–56.
18. Castle PE, Jeronimo J, Schiffman M, et al. Age-related changes of the
cervix influence human papillomavirus type distribution. Cancer Res
2006; 66:1218–24.
19. Smith JS, Lindsay L, Hoots B, et al. Human papillomavirus type distribution in invasive cervical cancer and high-grade cervical lesions:
a meta-analysis update. Int J Cancer 2007; 121:621–32.
20. Castle PE, Cox JT, Jeronimo J, et al. An analysis of high-risk human
papillomavirus DNA-negative cervical precancers in the ASCUS-LSIL
Triage Study (ALTS). Obstet Gynecol 2008; 111:847–56.
21. Schiffman M, Herrero R, Desalle R, et al. The carcinogenicity of human
papillomavirus types reflects viral evolution. Virology 2005; 337:76–84.
22. Gravitt PE, Viscidi RP. Measurement of exposures to human papillomavirus. In Rohan TE, Shah KH eds: Cervical cancer: from etiology to
prevention, 1st ed. Netherlands: Springer, 2004; 119–42.
23. Pisal N, Sindos M, Mansell ME, Freeman-Wang T, Singer A. Persistent
minor smear abnormalities: is large loop excision of the transformation
zone (LLETZ) the solution? J Obstet Gynaecol 2003; 23:528–30.
24. Pretorius RG, Zhang WH, Belinson JL, et al. Colposcopically directed
biopsy, random cervical biopsy, and endocervical curettage in the diagnosis of cervical intraepithelial neoplasia II or worse. Am J Obstet
Gynecol 2004; 191:430–4.