Original Research ajog.

org

GYNECOLOGY
Management of women with human papillomavirus
persistence: long-term follow-up of a randomized
clinical trial
Kristina Elfgren, MD; K. Miriam Elfström, PhD; Pontus Naucler, MD; Lisen Arnheim-Dahlström, PhD; Joakim Dillner, MD

BACKGROUND: Introduction of human papillomavirusebased measures: cumulative incidence of cervical intraepithelial neoplasia grade
screening is ongoing in many countries, given its higher sensitivity and 2 or worse and cervical intraepithelial neoplasia grade 3 or worse.
longer-lasting protection compared with cytology-based screening. RESULTS: Among women who continued to attend and had continuous
However, optimal clinical management of human papillomavirusepositive human papillomavirus persistence, all (40 of 40, 100% [95% confidence
but cytology-negative women is unclear, and additional studies with interval, 91e100%]) developed cervical intraepithelial neoplasia grade 2 or
clinical follow-up are warranted. worse. There were no cases among women who cleared their human
OBJECTIVE: The aim of the current study was to investigate the long- papillomavirus persistence (0 of 35, 0% (95% confidence interval, 0e10%)
term outcomes of the clinical management used in a double-blind, ran- (P < .001). Among women who had had human papillomavirus persistence
domized clinical trial of human papillomavirus screening conducted in the but did not continue with repeated human papillomavirus tests (unknown
context of the routine, organized screening program in Sweden. persistence status), 56% (15 of 27 women) developed cervical intraepithelial
STUDY DESIGN: Among 12,527 women aged 32e38 years enrolled neoplasia grade 2 or worse. Almost all cases occurred within 6 years. The
in the trial, we followed up the 195 women who attended the colposcopy intensive clinical management in the trial appeared to result in diagnoses of
screening who were cytologically normal but persistently human papillo- earlier cervical intraepithelial neoplasia grade 2 or worse but apparently did
mavirus positive (at least 12 months later; median, 19 months) in the not prevent cervical intraepithelial neoplasia grade 2 or worse.
human papillomavirus testing arm (n ¼ 100) or were randomly selected CONCLUSION: Women with human papillomavirus persistence will, in
from the control arm (n ¼ 95). Women in the human papillomavirus testing general, either become human papillomavirus negative or develop cervical
arm were followed up with repeated human papillomavirus testing, cy- intraepithelial neoplasia grade 2 or worse within 6 years, even with
tologies, and colposcopies if persistently human papillomavirusepositive intensive clinical follow-up.
without cervical intraepithelial neoplasia grade 2 or worse. A similar
number of random colposcopies and tests were carried out in the control Key words: cervical intraepithelial neoplasia, colposcopy, human
arm. Women were followed up over 13 years for the main outcome papillomavirus, long term-follow-up, randomized clinical trial

C ytological screening every 3e5

years reduces the cervical cancer
risk by about 80% in countries with HPV-based screening is more sensitive
colposcopy with colposcopically directed
biopsies and endocervical cytology.
In case of normal colposcopy 2 blind
organized screening programs through than cytology in detecting CIN grade 2 cervical biopsies were taken. At the first
early detection and treatment of pre- or worse (CIN2þ), providing a greater study colposcopy, 28 of 100 women with
cursor lesions (cervical intraepithelial and more long-lasting protection against HPV persistence in the intervention
neoplasia [CIN]).1,2 Clinical manage- invasive disease.4-7 However, the optimal arm and 2 of 95 women selected at
ment of women includes referral to clinical management of women who random from the control arm had
colposcopy, colposcopically directed bi- screen HPV positive but cytology nega- CIN2þ.13 If women were not treated for
opsies, surgical removal of histologically tive is unclear.8 Persistent infection with CIN2þ, they were invited for repeat
verified high-grade lesions and post- high-risk (HR) HPV is necessary for the yearly HPV testing and were, if persis-
treatment follow-up.1 development of cervical cancer.9,10 Most tently HPV positive, invited to an addi-
Human papillomavirus (HPV) is infections are transient, but women with tional colposcopy.
the major cause of cervical cancer.3 a persistent infection have a high risk for The aim of the present study was to
development of CIN2þ and cervical evaluate the long-term outcome of the
cancer.11,12 clinical management of these women,
Cite this article as: Elfgren K, Elfström KM, Naucler P,
et al. Management of women with human papillomavirus The HPV screening trial Swedescreen using comprehensive nationwide link-
persistence: long-term follow-up of a randomized clinical was nested in the population-based, ages with the nationwide Swedish cer-
trial. Am J Obstet Gynecol 2017;216:264.e1-7. organized cervical cancer screening pro- vical screening registry.
0002-9378/free
gram in Sweden. The clinical manage-
ª 2016 Elsevier Inc. All rights reserved. ment of HPV-positive, cytology-negative Materials and Methods
http://dx.doi.org/10.1016/j.ajog.2016.10.042 women was a repeat HPV test at least 12 Study population and data
Related editorial, page 206 months later and, if type-specific HPV The Swedescreen population-based ran-
persistence was found, referral to domized clinical trial (RCT) of primary

264.e1 American Journal of Obstetrics & Gynecology MARCH 2017

ajog.org GYNECOLOGY Original Research

HPV screening was started in May 1997. testing were collected. All samples were The dimension and border of abnormal
Women aged 32e38 years attending analyzed in the intervention arm as well areas were assessed by application of 5%
organized screening in 5 different regions as a random sample of the control arm. potassium iodine to the ectocervix. All
of Sweden were invited to participate. In HPV DNA testing was conducted using acetowhite areas, including metaplasia
Sweden, women are invited for organized polymerase chain reaction with GP5þ/ (undifferentiated epithelium), inflam-
screening only if they have not had any 6þ polymerase chain reaction primers mation, and neoplasia were considered
recent opportunistic smears. No further and subsequent typing with reverse dot- as abnormal colposcopies and biopsied.
exclusion criteria were used. blot hybridization. The assay correlated The endocervix was sampled with a
Since 1997, the trial participants have well to the HPV assay used in other cytobrush.16
been followed up with registry linkages. clinical validation studies.15 Blind biopsies were taken at 12 o’clock
Using unique personal identification and 6 o’clock on the ectocervix close to
numbers, linkages to the Swedish Na- Colposcopy follow-up the squamocolumnar junction if the
tional Cervical Screening Registry In the intervention arm, women who colposcopy was considered to be normal,
(NKCx) have been performed up to were HPV positive and cytology nega- according to standard protocol at the
December 2012 to identify all histological tive at enrollment were invited for a time.13 Treatment and further manage-
endpoints occurring among study par- second HPV test at least 12 months later ment of the women based on the results
ticipants, regardless of whether the his- (median actual attendance, 19 months; of cytologies and histopathologies fol-
tological endpoint was the result of study interquartile range, 15e27 months). At lowed the routine procedures of the
colposcopies. Until the primary endpoint baseline, 433 women were HPV posi- screening program.
of the study was reached (protection tive, of which 341 were cytology nega-
against CIN2þ in a second screening tive and therefore invited to a second Statistical analysis
round), all histopathologies were re- HPV test. A total of 270 women atten- The 195 women who attended the first
reviewed by an expert pathologist.13 ded the second HPV test, of which 119 study colposcopies were followed up
For the long-term follow-up after this women (44%) had type-specific HPV using registry linkages. The start of the
point, routine histopathological di- persistence and were subsequently follow-up was the date of the first test in
agnoses were used and identified through invited to colposcopy (100 women the RCT. The first instance of histologi-
linkage to the NKCx. The current analysis attended).13 In the control arm, 111 cally confirmed CIN2þ or CIN3þ were
makes use of the re-reviewed histopa- women were selected at random for identified through the study colpos-
thologies from the active study period mock HPV testing and subsequent col- copies and linkages to the NKCx.
and routine histopathological diagnoses poscopies to control for ascertainment Women were censored at the last regis-
that were obtained for lesions occurring bias (95 women attended). The results tered testing date, which is appropriate
outside the study colposcopies and dur- of the first study colposcopies have been for screen-detected disease. There were
ing the long-term follow-up. published.13 no cases of invasive disease found during
Following the first study colposcopy, follow-up in either arm.
Randomization and masking women who had not been treated for Linkages to the Swedish population
In total, 12,527 women were enrolled, CIN2þ were invited for yearly repeat registry found that 191 of 195 women
following informed consent, and ran- HPV tests. Throughout the study, all were alive and resident in Sweden
domized 1:1 to HPV and cytology double cytological abnormalities were managed throughout the follow-up. Two women
testing (intervention arm, n ¼ 6257) or according to the routine procedures used were deceased (both in the intervention
cytology only (control arm, n ¼ 6270). in the Swedish organized screening arm) and 2 women had emigrated. All
Women were randomized using program. Women with continuous type- 195 women were alive and resident in
computer-generated random numbers by specific HPV persistence as well as a Sweden for >5 years. Women were
an independent institute (the Cancer similar number of randomly chosen categorized as having a known persis-
Registry of Stockholm). Further details on control women were invited for addi- tence status (time frame determined as a
the study protocol have been published tional colposcopies (Figure 1). Subse- positive HPV test result plus 1 year for
previously.14 Both the study participants quent colposcopies were performed the analysis), unknown persistence sta-
and the colposcopists were blinded to the according to the same protocol and tus (no HPV test performed 1 year after
women’s HPV status. The trial was un- conducted by the same gynecologists as the last HPV test; in other words, more
blinded in August 2003, after the the first study colposcopies, except that than 1 year since the last HPV test was
completion of the first study colposcopies HPV test results were unblinded. taken), HPV type-specific clearance after
and 3 years after recruitment concluded. The protocol included an examina- persistence, and HPV negativity after
tion of the transformation zone for vis- persistence. Women who switched types
Intervention ibility, maturation, and epithelial were persistently positive for one type
HPV DNA testing changes. Five percent acetic acid was and then at a subsequent test were
For all women, a Papanicolaou smear applied to identify undifferentiated negative for the initial type but positive
and a brush sample for HPV DNA epithelia or inflammation as well as CIN. for a new type.

MARCH 2017 American Journal of Obstetrics & Gynecology 264.e2

Original Research GYNECOLOGY ajog.org

FIGURE 1
CONSORT flow chart of follow-up of women in Swedescreen study

CONSORT flow chart of colposcopic and registry-based follow-up of women in the intervention (HPV typeespecific persistence) and control arms of the
Swedescreen study.
CIN2þ, cervical intraepithelial neoplasia grade 2 or worse; CONSORT, Consolidated Standards of Reporting Trials; HPV, human papillomavirus.
Elfgren et al. Management of women with human papillomavirus persistence. Am J Obstet Gynecol 2017.

Absolute risks for CIN2þ and CIN3þ infection and 95 randomly selected 2 among the randomly selected women
were calculated for the entire follow-up women from the control arm). The in the control arm), the absolute risk for
period (13 years). Cumulative inci- CIN2þ cases detected at the first study CIN2þ over the long-term follow-up
dence proportions of CIN2þ and colposcopy have previously been was 54% (55 of 102, 95% confidence
CIN3þ were estimated using 1 minus described.13 During the long-term interval [CI], 44e63%) and the risk for
the Kaplan-Meier curves. follow-up (up to 13 years), linkages CIN3þ was 32% (33 of 102, 95% CI,
All analyses were completed using were made to the national screening 24e42%). By study arm, the risks for
STATA 11 (StataCorp, College Station, registry NKCx. CIN2þ was 54% in the intervention arm
TX). The Swedescreen study was All 195 women were found to have and 2% in the control arm (log rank
approved by the ethical review board in had additional cytologies in the national value of P < .001).
Stockholm, Sweden (DNR 1996/305). registry (mean number of cytologies, 8; Persistence by type was as follows: 25
The long-term follow-up of the Swe- range, 4e17) and 192 of 195 of the bi- women with HPV 16, 11 with HPV18, 31
descreen study had an additional opsies from the per-protocol colpos- with HPV 31, with 6 HPV 33, 9 with
approval from the ethical review copies were identified in the registry HPV 45, and 20 with other HR HPV (not
board in Stockholm, Sweden (DNR (mean number of cervical histopathol- shown). Among the 25 women with a
2012/780-32). ogies, 2; range, 1e9). An additional 25 type 16 persistence, 68% (17 of 25, 95%
CIN2þ diagnoses occurred in the CI, 48e83%) and 56% (14 of 25, 95%
Results intervention arm, but no further CIN2þ CI, 37e73%) developed CIN2þ and
In total, 195 women with a normal occurred in the control arm. CIN3þ, respectively (not shown).
cytology attended the first study col- Among the 102 women with initial Persistence with HPV16 conferred the
poscopies (100 women in the interven- HR HPV type-specific persistence (all highest risk for CIN2þ followed by HPV
tion arm with persistent HR HPV 100 women in the intervention arm and 18 (7 of 11, 64% [95% CI, 35e85%]),

264.e3 American Journal of Obstetrics & Gynecology MARCH 2017

ajog.org GYNECOLOGY Original Research

HPV 31 (18 of 31, 58% [95% CI,

FIGURE 2
41e74%]), HPV 33 (3 of 6, 50% [95%
Cumulative proportion of CIN2D in Swedescreen study over 13 years
CI, 44e100%]), HPV 45 (4 of 9, 44%
[95% CI, 19e73%]), and other HR types
(6 of 20, 30% [95% CI, 15e52%]) (not
shown). None of the women in the
control arm who were HR HPV negative
at baseline developed CIN2þ during
follow-up (not shown).
The cumulative incidence proportion
of CIN2þ increased steadily in the
intervention arm in the first 6 years of
follow-up. However, after 6 years, only 3
additional cases of CIN2þ occurred
(Figure 2). For CIN3þ, the pattern was
similar, with only 1 additional case
occuring after 6 years (Figure 3). Among
the 40 women who attended repeat
testing and were known to be continu-
ously HPV persistent, all 40 of 40
developed CIN2þ within 7 years of the The cumulative incidence proportion of CIN2þ in the intervention (gray line) and control (black line)
baseline HPV-positive test result arms of the Swedescreen study over 13 years of follow-up.
(Figure 4); 63% (95% CI, 48e77%) and CIN2þ, cervical intraepithelial neoplasia grade 2 or worse.
98% (95% CI, 89e100%) had developed Elfgren et al. Management of women with human papillomavirus persistence. Am J Obstet Gynecol 2017.
CIN2þ at CIN3þ after 6 years of follow-
up, respectively.
The risks for women who had not If HPV-based primary screening need very long-term clinical follow-up,
attended a yearly repeat test and had an had identified a group of continuously this would have been a significant
unknown persistence status were similar HPV-positive women who do not burden. However, because all of the
to the risk seen for all women who had develop any treatable lesions and may women with continuous HPV
had persistence at baseline (50% devel-
oped CIN2þ within 7 years) (Figure 4).
In this group, 11% (95% CI, 4e31%) FIGURE 3
and 46% (95% CI, 29e66%) developed Cumulative proportion of CIN3D in the Swedescreen study over 13 years
CIN2þ at CIN3þ after 6 years of follow-
up, respectively. None of the women who
had initially had HPV persistence
developed CIN2þ following a negative
HPV test (clearance, n ¼ 31) or
following a type switch (n ¼ 4)
(Figure 4).

Comment
Our results provide evidence of a very
high long-term risk of CIN2þ among
women with continuous HPV persis-
tence, even with an initially negative
cytology and negative colposcopically
directed biopsies. We also provide evi-
dence that women who had had HPV
persistence but have become HPV
negative are not at measurable long-term
The cumulative incidence proportion of CIN3þ in the intervention (grey line) and control (black line)
risk of CIN2þ. We find that women with arms of the Swedescreen study over 13 years of follow-up.
HPV persistence over 7 years either CIN3þ, cervical intraepithelial neoplasia grade 3 or worse.
develop CIN2þ or become HPV Elfgren et al. Management of women with human papillomavirus persistence. Am J Obstet Gynecol 2017.
negative.

MARCH 2017 American Journal of Obstetrics & Gynecology 264.e4

Original Research GYNECOLOGY ajog.org

(13 years) and the fact that all women

FIGURE 4
had additional tests on follow-up, iden-
Cumulative proportion of CIN2D among women with type-specific
tified using linkages to a nationwide and
persistence
comprehensive screening registry. The
population coverage of the screening
program in Sweden at the 3-5 year rec-
ommended intervals is 82% and the 6
year population coverage is 91%.17
Among women who have previously
attended screening, the future coverage
tends to be higher than in the general
population, and the fact that all women
in the trial group could be followed up is
therefore not surprising.
Limitations of the study include the
fact that routine histopathological re-
sults were used after the first study col-
poscopy. Also, several women dropped
out from the yearly HPV tests offered,
resulting in the fact that it is unknown
whether their HPV persistence
continued or whether these women had
The lines represent women with a known persistence status (black line; n ¼ 40), unknown cleared their HPV. The high dropout rate
persistence status (dashed line; n ¼ 27), HPV type switch (clearance of the previously persistent and the limited benefits obtained with
type; dotted line; n ¼ 4), and HPV negativity after persistence (dash-dot line; n ¼ 31). The last 2 lines intensive follow-up with yearly HPV
are entirely overlapping because there were no events detected during follow-up for both the latter 2 testing and extended gynecological in-
groups. vestigations further suggest that a less
CIN2þ, cervical intraepithelial neoplasia grade 2 or worse; HPV, human papillomavirus.
intensive clinical follow-up (eg, repeat
Elfgren et al. Management of women with human papillomavirus persistence. Am J Obstet Gynecol 2017.
HPV testing at 3 and 6 years with referral
only if continuous HPV persistence) is
warranted.
persistence developed CIN2þ and women had attended intensive clinical A note of caution is needed, however,
because no cases were seen among follow-up or not. Too intensive manage- because the safety of such intervals is
women with HPV clearance, continued ment (yearly repeat tests) may result in the impacted by the risk of overlooked high-
follow-up of women with HPV persis- result that women drop out of the follow- grade lesions at colposcopy18 and the
tence after 7 years does not appear to be up, which, considering the very high quality of the HPV testing performed
required. Although our numbers are CIN2þ risks, is potentially unsafe. More because it cannot be assumed that all
small, it seems that HPV 16 carries a conservative management strategies with HPV tests have the same predictive
particularly high risk for CIN2þ and repeat tests at 3 and 6 years, but during ability as the HPV test used in the Swe-
CIN3þ when persistent, suggesting that which every effort is made to ensure a high descreen RCT. Evidence from labora-
women with HPV 16 persistence could attendance rate, may be preferable. tories participating in the global HPV
be a priority. The Swedescreen RCT was nested LabNet HPV genotyping proficiency
The fact that the CIN2þ lesions within the population-based cervical panel studies indicates that laboratory
continued to appear for 7 years during screening program in Sweden. There- performance has improved overall,19 but
intensive clinical follow-up, suggests that fore, the study participants are repre- continued quality assurance and moni-
intensive clinical follow-up of women sentative of women attending screening, toring of HPV tests used in screening is
with HPV persistence is not able to pre- and the results should be generalizable needed. The very high risks for CIN2þ
vent CIN2þ lesions from developing and to the general screening population. seen among women with continuous
merely resulted in a somewhat earlier Because randomly selected women from persistence after 6 years suggest that
diagnosis of these lesions (cf, Figure 4). the control arm underwent the same diagnostic conization could be consid-
This suggests that ensuring that women study procedures of repeated HPV ered if the colposcopy is not adequate
with HPV persistence are not lost to testing, endocervical cytology, colpos- after continuous persistence for 6 years
follow-up should be prioritized over the copy, and colposcopy-directed biopsies, or more.
intensity of the clinical management. verification bias is likely to be limited. HPV persistence has been confirmed
There were no cases of invasive cancer Additional strengths of the study to be consistently and significantly
in the cohort, regardless of whether the include the extensive length of follow-up associated with the development of

264.e5 American Journal of Obstetrics & Gynecology MARCH 2017

ajog.org GYNECOLOGY Original Research

high-grade lesions as described in a In conclusion, the evidence of a very with human papillomavirus 16 before the
2008 meta-analysis.20 However, long- high 7 year risk of developing CIN2þ development of cervical carcinoma in situ.
Cancer Res 2000;60:6027-32.
term persistence of HPV infections in among women with continuous HPV 11. Evander M, Edlund K, Gustafsson A, et al.
the absence of cytological abnormal- persistence, even with an initially Human papillomavirus infection is transient in
ities has also been reported.21 In that negative cytology, indicates that young women: a population-based cohort
study, no comment on management follow-up for this length of time is a study. J Infect Dis 1995;171:1026-30.
was given.21 A population-based study priority. n 12. Ho GY, Bierman R, Beardsley L, Chang CJ,
Burk RD. Natural history of cervicovaginal
in Colombia focusing on women who papillomavirus infection in young women. N Engl
were cytology negative at baseline and Acknowledgments J Med 1998;338:423-8.
examining prevalent and incident HPV We thank the participating women and the past 13. Elfgren K, Rylander E, Radberg T, et al.
infections and risk for high-grade le- members of the Swedescreen study group Colposcopic and histopathologic evaluation of
including researchers, health care practitioners, women participating in population-based
sions showed that HPV type 16
and data managers. The funding sources of this screening for human papillomavirus deoxy-
persistence carried a particularly high study had no involvement in the study design; ribonucleic acid persistence. Am J Obstet
risk for CIN2/3.22 the collection, analysis, and interpretation of the Gynecol 2005;193:650-7.
Data from the Guanacaste data; the writing the report; or the decision to 14. Naucler P, Ryd W, Tornberg S, et al. Human
population-based cohort study reflects submit the paper. The study was registered papillomavirus and Papanicolaou tests to screen
(Clinicaltrials.gov [number NCT00479375]).
our finding that very few infections for cervical cancer. N Engl J Med 2007;357:
1589-97.
persist without the development of a References 15. Jacobs MV, Snijders PJ, Voorhorst FJ, et al.
CIN2þ lesion and highlights the 1. Arbyn M, Anttila A, Jordan J, et al. European Reliable high risk HPV DNA testing by polymerase
importance of duration of infection.23 guidelines for quality assurance in cervical can- chain reaction: an intermethod and intramethod
Compared with our publication of the cer screening. 2nd ed. Luxembourg: Office for comparison. J Clin Pathol 1999;52:498-503.
results from the first study colpos- Official Publications of the European Commu- 16. Jacobs MV, Snijders PJ, van den Brule AJ,
nities; 2008. Helmerhorst TJ, Meijer CJ, Walboomers JM.
copies,13 this long-term follow-up A general primer GP5þ/GP6(þ)-mediated
2. International Agency for Research on Can-
provides further evidence on the cer. Cervix cancer screening/IARC handbooks PCR-enzyme immunoassay method for rapid
clinical management of cytology- of cancer prevention. Vol 10. Lyon (France): detection of 14 high-risk and 6 low-risk human
negative and persistently HPV- International Agency for Research on Cancer; papillomavirus genotypes in cervical scrapings.
positive women. 2005. J Clin Microbiol 1997;35:791-5.
3. Walboomers JM, Jacobs MV, Manos MM, 17. Nationellt Kvalitetsregister för Cervixcancer-
Additionally, this analysis provides 13 prevention. Förebyggande av livmoderhalscancer
et al. Human papillomavirus is a necessary
years of follow-up for this subcohort of cause of invasive cervical cancer worldwide. i Sverige: Verksamhetsberättelse och Årsrapport
the Swedescreen trial, demonstrating the J Pathol 1999;189:12-9. 2014 med data till och med 2013. Stockholm
long-term risk for CIN2þ by persistence 4. Mayrand MH, Duarte-Franco E, Rodrigues I, (Sweden): Nationellt Kvalitetsregister för Cervix-
status and in the context of an intensive et al. Human papillomavirus DNA versus Papa- cancerprevention; 2014.
nicolaou screening tests for cervical cancer. 18. Luyten A, Buttmann-Schweiger N, Luyten K,
clinical follow-up of women with HPV et al. Early detection of CIN3 and cervical cancer
N Engl J Med 2007;357:1579-88.
persistence. In our material, we found 5. Bulkmans NW, Rozendaal L, Voorhorst FJ, during long-term follow-up using HPV/Pap
that women who clear HPV persistence Snijders PJ, Meijer CJ. Long-term protective smear co-testing and risk-adapted follow-up in a
were not at a measurable risk of CIN2þ. effect of high-risk human papillomavirus testing locally organised screening programme. Int J
Although this suggests that these women in population-based cervical screening. Br J Cancer 2014;135:1408-16.
Cancer 2005;92:1800-2. 19. Eklund C, Forslund O, Wallin KL, Dillner J.
may be safely returned to routine Global improvement in genotyping of human
6. Ronco G, Dillner J, Elfstrom KM, et al. Efficacy
screening, reappearance of an HPV of HPV-based screening for prevention of inva- papillomavirus DNA: the 2011 HPV LabNet In-
infection is possible, and treated women sive cervical cancer: follow-up of four European ternational Proficiency Study. J Clin Microbiol
may still be at risk for reinfection and randomised controlled trials. Lancet 2014;383: 2014;52:449-59.
repeated CIN2þ.24 524-32. 20. Koshiol J, Lindsay L, Pimenta JM, Poole C,
7. Elfstrom KM, Smelov V, Johansson AL, et al. Jenkins D, Smith JS. Persistent human papillo-
Whereas following up women with mavirus infection and cervical neoplasia: a sys-
Long term duration of protective effect for HPV
HPV persistence (having an increased negative women: follow-up of primary HPV tematic review and meta-analysis. Am J
risk for high-grade lesions) is more screening randomised controlled trial. BMJ Epidemiol 2008;168:123-37.
easily facilitated by an organized pro- 2014;348:g130. 21. Castle PE, Rodriguez AC, Burk RD, et al.
gram, this issue can be solved in 8. Katki HA, Kinney WK, Fetterman B, et al. Long-term persistence of prevalently detected
Cervical cancer risk for women undergoing human papillomavirus infections in the absence
context-specific ways, depending on the of detectable cervical precancer and cancer.
concurrent testing for human papillomavirus and
clinician or health care unit responsible cervical cytology: a population-based study in J Infect Dis 2011;203:814-22.
for care and the design of medical chart routine clinical practice. Lancet Oncol 2011;12: 22. Munoz N, Hernandez-Suarez G, Mendez F,
information systems. In low-resource 663-72. et al. Persistence of HPV infection and risk of
settings, baseline testing is possible, 9. Munoz N, Castellsague X, de Gonzalez AB, high-grade cervical intraepithelial neoplasia in a
Gissmann L. Chapter 1: HPV in the etiology cohort of Colombian women. Br J Cancer
but persistence testing poses additional 2009;100:1184-90.
of human cancer. Vaccine 2006;24(Suppl 3):
challenges because follow-up over time S3/1-10. 23. Rodriguez AC, Schiffman M, Herrero R, et al.
may be limited by health care 10. Ylitalo N, Josefsson A, Melbye M, et al. Longitudinal study of human papillomavirus
infrastructures. A prospective study showing long-term infection persistence and cervical intraepithelial neoplasia

MARCH 2017 American Journal of Obstetrics & Gynecology 264.e6

Original Research GYNECOLOGY ajog.org

grade 2/3: critical role of duration of infection. Departments of Laboratory Medicine (Drs Elfström and Drs Lisen Arnheim-Dahlström and Dillner have
J Natl Cancer Inst 2010;102:315-24. Dillner) and Medical Epidemiology and Biostatistics (Drs received grants from Merck/SPMSD for unconditional
24. Kreimer AR, Schiffman M, Herrero R, et al. Arnheim-Dahlström and Dillner), Karolinska Institutet, studies. The other authors report no conflict of interest.
Long-term risk of recurrent cervical human and Department of Medicine, Solna, Karolinska University Presented at the 30th International Papillomavirus
papillomavirus infection and precancer and Hospital (Dr Naucler), Karolinska Institutet, Stockholm, Conference and Clinical and Public Health Workshops,
cancer following excisional treatment. Int J Sweden. Sept. 17e21 2015, Lisbon, Portugal, and at the 26th
Cancer 2012;131:211-8. Received Aug. 8, 2016; revised Oct. 28, 2016; International Papillomavirus Conference with Clinical and
accepted Oct. 31, 2016. Public Health Workshops, July 3e8, 2010, Montreal,
This study was supported by the Swedish Cancer Canada.
Author and article information Society (2098, 3824), Europe Against Cancer Corresponding author: Joakim Dillner, MD. joakim.
From the Department of Clinical Science, Intervention, (SPC2002475), and the European Union 7th framework [email protected]
and Technology, Karolinska University Hospital program excellence project in comparative effectiveness
(Dr Elfgren), Karolinska Institutet, Huddinge, Stockholm, research (CoheaHr number 603019).

264.e7 American Journal of Obstetrics & Gynecology MARCH 2017