Impact of Human Papillomavirus Vaccination on Cervical Cytology Screening, Colposcopy, and Treatment

Abstract

The impact of human papillomavirus (HPV) vaccination on cervical screening, colposcopy, and treatment is incompletely understood. In 2004–2005, investigators in the Costa Rica Vaccine Trial randomized 7,466 women aged 18–25 years, 1:1, to receive HPV vaccination or hepatitis A vaccination. The worst-ever cytology diagnosis and the 4-year cumulative proportions of colposcopy referral and treatment by vaccination arm were compared for 2 cohorts. The total vaccinated cohort included 6,844 women who provided cervical samples. The naive cohort included 2,284 women with no evidence of previous HPV exposure. In the total vaccinated cohort, HPV-vaccinated women had a significant (P = 0.01) reduction in cytological abnormalities: 12.4% for high-grade lesions and 5.9% for minor lesions. Colposcopy referral was reduced by 7.9% (P = 0.03) and treatment by 11.3% (P = 0.24). Greater relative reductions in abnormal cytology (P < 0.001) were observed for HPV-vaccinated women in the naive cohort: 49.2% for high-grade lesions and 18.1% for minor lesions. Colposcopy referral and treatment were reduced by 21.3% (P = 0.01) and 45.6% (P = 0.08), respectively, in the naive cohort. The overall impact on health services will be modest in the first years after vaccine introduction among young adult women, even in regions with high coverage.

Understanding the natural history of human papillomavirus (HPV) infection and its relationship to the development of cervical cancer has led to 2 new preventive methods: primary prevention through prophylactic HPV vaccination and secondary prevention through improved screening that incorporates HPV testing. However, there is still discussion regarding the best integration of improved screening in vaccinated cohorts. Simulations have predicted that cervical cancer screening with cytology would be less efficient among vaccinated women due to a decrease in the percentage of vaccinated women who ever develop the main target of screening—that is, cervical intraepithelial neoplasia (CIN) grade 3 or worse (CIN3+) lesions (1, 2). However, direct measures of the performance of cytological screening programs among vaccinated women are scarce, with only 1 ecological report from Victoria, Australia (3) and 2 from phase III HPV vaccine efficacy clinical trials (4, 5).

Administration of the currently available vaccines against HPV, Gardasil (Merck and Company Inc., Whitehouse Station, New Jersey) and Cervarix (GlaxoSmithKline Biologicals, Rixensart, Belgium), before HPV exposure is nearly 100% effective against the types of HPV included in the formulations (4, 6, 7). There is consensus that public health vaccination programs against HPV should be focused primarily on adolescent girls prior to initiation of first sexual intercourse, for which the potential benefit of the vaccines is highest. In some regions, young adult women are also considered for HPV vaccination programs, either to complete interrupted vaccination schedules or to vaccinate those who were not vaccinated as adolescents. However, most women vaccinated during young adulthood are already sexually active and are likely to have already experienced HPV infection with carcinogenic types of HPV, including the vaccine types. With this consideration in mind, we measured the effect of HPV vaccination on cytology diagnosis, colposcopy referral, and treatment proportions of young adult women aged 18–25 years participating in a community-based phase III vaccine trial in Costa Rica. We estimated the impact of vaccination on cervical screening for the total group and for a subset intended to represent women who are unexposed to HPV of the vaccine types.

MATERIALS AND METHODS

Study design and population

During 2004–2005, a total of 7,466 women were enrolled in a community-based double-blind randomized controlled phase III trial (the Costa Rica Vaccine Trial) of a virus-like particle vaccine against HPV types 16 and 18 (Cervarix). The study was carried out among healthy women aged 18–25 years who were living in the province of Guanacaste or in nearby areas of Puntarenas Province. The Costa Rica Vaccine Trial was designed to investigate the efficacy of the vaccine in preventing cervical cancer precursors, defined as persistent infection and/or histologically defined high-grade CIN (CIN3, as well as CIN2, because CIN2+ is the threshold for treatment), with particular interest in evaluating the efficacy of the vaccine in relation to previous HPV exposure, age at vaccination, and the potential impact of the vaccine on health services. The Costa Rica Vaccine Trial recruited 30.5% of all women aged 18–25 years in the study area, representing 59.1% of all potentially eligible women. For eligibility, women had to be healthy, not pregnant, not breastfeeding, and using contraception during the vaccination period. The main exclusion criteria were chronic diseases, history of reaction to vaccines, and history of hepatitis A or vaccination against it. Details regarding the study design and methods have been provided elsewhere (8).

Clinic visits and sample collection

At the enrollment visit and after providing signed informed consent, women had their study eligibility determined. Women were recruited and randomized regardless of past sexual behavior and regardless of cervical HPV, serological, or cytological status. If they were currently sexually active, women underwent a pelvic examination that included collection of exfoliated cervical cells with a Cervex brush (Rovers Medical Devices B.V., Oss, the Netherlands) that was vigorously rinsed in a PreservCyt solution vial (Hologic Inc., Marlborough, Massachusetts) for liquid-based cytological analysis (ThinPrep; Hologic Inc.). Blood was collected from all participants at enrollment, regardless of previous sexual activity, for determination of HPV 16 and HPV 18 antibodies in serum. After sample collection, women were randomized at a 1:1 ratio to receive a 3-dose schedule of Cervarix or hepatitis A vaccine, which was used as the control vaccine. Details have been previously published (8).

All women were invited to continue with yearly follow-up visits in order to complete 4 years of follow-up, regardless of the number of vaccine doses received. Exfoliated cervical cells were collected at all follow-up visits. Clinical management of the study was based primarily on Costa Rican reading of the liquid-based cytological test results; however, for quality control, all slides that were read as abnormal (atypical squamous cells or worse) by the Costa Rican cytopathologist, plus a 10% random sample of all slides, regardless of the reading, were rescreened and reinterpreted in the United States by a cytotechnologist and a cytopathologist. The interpretation agreement for enrollment samples was very good, with a linear weighted κ value of 0.70 (95% confidence interval (CI): 0.68, 0.73) (8). Detection of carcinogenic HPV DNA using the sample collected in PreservCyt solution (Hologic Inc.) was performed on enrollment samples using the hc2 High-Risk HPV DNA Test (Qiagen Gaithersburg Inc.), hereafter called the Hybrid Capture 2 (HC2) test. During follow-up, HC2 testing was done only as a reflex test to triage cytological readings of atypical squamous cells of undetermined significance (ASC-US). An aliquot of all PreservCyt samples was also tested by polymerase chain reaction-based HPV DNA testing at Delft Diagnostics Laboratory (Delft, the Netherlands) using SPF10 primers with a DNA enzyme immunoassay/line probe assay detection system; samples found to be positive by DNA enzyme immunoassay were retested with specific primers for HPV 16 and HPV 18. Very good agreement regarding HPV positivity was found among enrollment samples tested by means of HC2 and polymerase chain reaction (κ = 0.75, 95% CI: 0.73, 0.76) (9).

Compliance with study visits was high, with an overall attendance rate of 80%. Only 7.1% of women discontinued their participation in the study after the enrollment visit.

Colposcopy

If women were found to have minor cytological abnormalities, they were shifted to a 6-month screening interval. Minor cytological abnormalities included the diagnoses of low-grade squamous intraepithelial lesion (LSIL) and ASC-US positive for high-risk HPV DNA as measured by HC2 (ASC-US/HPV+). An unsatisfactory cytology reading was considered equivalent to a minor cytological abnormality. Throughout this analysis, a cytological diagnosis of ASC-US with a negative reflex HC2 test was considered to be within normal limits, while a cytological diagnosis of ASC-US with a positive reflex HC2 test was considered a minor cytological abnormality and combined with LSIL (ASC-US/LSIL), consistent with the consensus guidelines of the American Cancer Society, the American Society for Colposcopy and Cervical Pathology, and the American Society for Clinical Pathology (10).

Women with a cytological reading (by any of the cytopathologists) of high-grade squamous intraepithelial lesion (HSIL), including the diagnosis of “atypical squamous cells, cannot rule out high-grade lesion” (ASC-H), or any abnormal glandular reading were immediately referred for colposcopic evaluation; women with persistent (6 months or longer) minor cytological abnormalities (ASC-US/LSIL) were also referred to colposcopy. The yearly screening visits were interrupted while the women were attending the colposcopy clinic; after colposcopy release, women continued with semiannual interval screening visits until they had completed 3 consecutive visits with cytology results within normal limits, at which time they were shifted back to the yearly screening interval. During the 4 years of follow-up, a total of 2,051 women (27.5%) were referred to colposcopy at some point.

Upon entering the colposcopy protocol, women were evaluated by the study colposcopist, and those with acetowhite areas underwent a punch biopsy. Women with histological results of CIN2 or CIN3 were treated with the loop electrosurgical excision procedure. If the histological finding was invasive cancer, women were sent to a referral hospital of Social Security for appropriate treatment. A total of 10 women were referred for further treatment during the study.

All histology slides were read by the study pathologist for clinical management; at the end of follow-up, all slides were read in the United States by a second pathologist. If the opinions of the 2 pathologists differed, a third pathologist read the discordant slides, and the final pathology diagnosis was reached by majority vote.

Statistical analysis

We estimated the impact of vaccination on cytology diagnosis, colposcopy referral, and treatment for 2 cohorts relevant to the integration of vaccination and screening. The total vaccinated cohort (n = 6,844; 3,412 in the HPV arm and 3,432 in the hepatitis A arm) included all women who were randomized into the Costa Rica Vaccine Trial, received the proper vaccine, were sexually active or became sexually active during the 4 years of the trial, and provided cervical samples. Estimates from the total vaccinated cohort show the possible effect of vaccination on national or regional cervical cancer prevention programs if women in this age range (18–25 years) were to be included in a vaccination program and then comprehensively screened. The naive cohort (n = 2,284; 1,143 in the HPV arm and 1,141 in the hepatitis A arm) included the subgroup of women who had no obvious evidence of previous HPV exposure even though they were sexually active or became sexually active during the study. Analyses of the naive cohort were restricted to women who provided cervical samples, were HPV DNA-negative at the cervix for all HPV types tested by polymerase chain reaction at enrollment and at the 6-month visit, and were also negative for HPV 16/18 serology at enrollment (Figure 1). Estimates for the naive cohort are shown as a surrogate measure of what the effect of vaccination on cancer prevention programs would be when adolescent girls who are currently being vaccinated initiate sexual intercourse and begin to be screened for cervical cancer.

Figure 1.

Selection and distribution of participants for the Costa Rica Vaccine Trial, 2004–2010. Biopsy/LEEP by V06, biopsy or loop electrosurgical excision procedure (LEEP) done by visit 6 months after enrollment; CIN2+, cervical intraepithelial neoplasia (CIN) grade 2 or worse; HPV, human papillomavirus; HPV DNA+ by V06, positive for HPV DNA by visit 6 months after enrollment.

The worst cytological diagnosis received during the study period, including enrollment cytology, was determined for each woman using the Bethesda system (11). The proportions of worst cytology readings categorized, as previously described, into normal, ASC-US/LSIL, and HSIL were estimated for each arm and cohort. The differential distribution between study arms was evaluated with a multinomial response regression model; separate models were fitted for each cohort.

The cumulative proportions of women who were ever referred for colposcopic evaluation and ever underwent a loop electrosurgical excision procedure were calculated for each vaccination arm and cohort. The absolute and relative reductions in referral and treatment proportions among the HPV-vaccinated women were estimated, along with 95% confidence intervals. We also analyzed the cumulative colposcopy referral and treatment proportions by time since vaccination for each vaccination arm and cohort. To further evaluate the impact of vaccination among women referred to colposcopy, we estimated the distribution of the different carcinogenic HPV types present among women at the time of referral to colposcopy for each vaccination arm and cohort.

RESULTS

Women who received vaccination against HPV had fewer abnormal cytological diagnoses. Considering the worst cytological diagnosis during the trial for each woman, including enrollment cytology, estimates from the total vaccinated cohort showed that 6.5% of HPV-vaccinated women had HSIL readings as compared with 7.5% in the control arm, for a 12.4% (95% CI: −4.8, 26.8) relative reduction. A smaller relative reduction of 5.9% (95% CI: −2.4, 13.6) was observed for ASC-US/LSIL. The overall difference in diagnosis frequency distribution according to the vaccine received was statistically significant (P = 0.01; Table 1).

Table 1.

Distribution of Worst-Ever Cytological Test Results in the Total Vaccinated Cohort, by Vaccine Arm, and the Absolute and Relative Reductions in Rates in Each Diagnostic Category, Costa Rica Vaccine Trial, 2004–2010*

Cytological Test Result	HPV Vaccine		Hepatitis A Vaccine		Rate Reduction, %
	No.	%	No.	%	Absolute	95% CI	Relative	95% CI
HSILa	223	6.5	256	7.5	0.9	–0.33, 2.17	12.4	−4.84, 26.82
ASC-US/LSILb	1,040	30.5	1,112	32.4	1.9	–0.74, 4.57	5.9	−2.37, 13.55
Normalc	2,149	63.0	2,064	60.1
Total	3,412	100	3,432	100

Abbreviations: ASC-US, atypical squamous cells of undetermined significance; CI, confidence interval; HPV, human papillomavirus; HSIL, high-grade squamous intraepithelial lesion; LSIL, low-grade squamous intraepithelial lesion.

* P = 0.01 (multinomial response regression model).

^a HSIL in situ; “atypical squamous cells, cannot rule out high-grade lesion” (ASC-H); and atypical glandular cells of undetermined significance.

^b ASC-US with a positive reflex hc2 High-Risk HPV DNA Test (Qiagen Gaithersburg Inc., Gaithersburg, Maryland) result and LSIL.

^c Within normal limits; inflammatory and atypical squamous cells of undetermined significance with a negative reflex hc2 High-Risk HPV DNA Test result.

Significantly fewer cytological abnormalities were also found for the naive cohort (P = 0.0002; Table 2). Compared with the total vaccinated cohort, in the naive cohort, a larger relative reduction in abnormal cytological results (HSIL and ASC-US/LSIL readings) was observed. When enrollment cytological findings were not taken into consideration, increases in the absolute and relative reductions in the proportion of abnormal cytological results were observed for the total vaccinated cohort (for HSIL, relative reduction = 20.3% (95% CI: 2.1, 35.2); for ASC-US/LSIL, relative reduction = 8.9% (95% CI: 0.3, 16.7)), with an overall statistically significant distribution difference (P = 0.0005). The absolute and relative reductions were minimally modified for the naive cohort (for HSIL, relative reduction = 48.6% (95% CI: 19.4, 67.8); for ASC-US/LSIL, relative reduction = 17.7% (95% CI: 0.5, 31.9)), with an overall difference of P = 0.0003.

Table 2.

Distribution of Worst-Ever Cytological Test Results in the Naive Cohort, by Vaccine Arm, and the Absolute and Relative Reductions in Rates in Each Diagnostic Category, Costa Rica Vaccine Trial, 2004–2010*

Cytological Test Result	HPV Vaccine		Hepatitis A Vaccine		Rate Reduction, %
	No.	%	No.	%	Absolute	95% CI	Relative	95% CI
HSILa	28	2.5	55	4.8	2.4	0.82, 3.76	49.2	20.30, 68.14
ASC-US/LSILb	197	17.2	240	21.0	3.8	0.21, 7.30	18.1	1.09, 32.18
Normalc	918	80.3	846	74.2
Total	1,143	100	1,141	100

Abbreviations: ASC-US, atypical squamous cells of undetermined significance; CI, confidence interval; HPV, human papillomavirus; HSIL, high-grade squamous intraepithelial lesion; LSIL, low-grade squamous intraepithelial lesion.

* P = 0.0002 (multinomial response regression model).

^a HSIL in situ; “atypical squamous cells, cannot rule out high-grade lesion” (ASC-H); and atypical glandular cells of undetermined significance.

^b ASC-US with a positive reflex hc2 High-Risk HPV DNA Test (Qiagen Gaithersburg Inc., Gaithersburg, Maryland) result and LSIL.

^c Within normal limits; inflammatory and atypical squamous cells of undetermined significance with a negative reflex hc2 High-Risk HPV DNA Test result.

Regarding colposcopy referral, HPV-vaccinated women were referred to colposcopy less often than women in the control arm (28.7% as compared with 31.1%), with a χ² P value for difference of 0.03 and a relative reduction of 7.9% (95% CI: −0.45, 15.5) in the total vaccinated cohort. In the naive cohort, the proportion of colposcopy referral by vaccination arm showed a relative reduction of 21.3% (95% CI: 3.9, 35.7) among women who received the HPV vaccine (Table 3). In both cohorts, the proportions of women making colposcopy visits were similar between study arms during the first 2 years after vaccination; the fourth year showed a qualitatively bigger difference, although it was not statistically significant (Figure 2).

Table 3.

Reduction in the Rate of Referral for Colposcopy Among Vaccinated Women, by Vaccine Type and Cohort, Costa Rica Vaccine Trial, 2004–2010

Cohort	% HPV-Vaccinated	% Hepatitis A-Vaccinated	Rate Reduction				P Valuea
			Absolute		Relative
			%	95% CI	%	95% CI
Total vaccinated	28.7	31.1	2.5	−0.14, 5.04	7.9	−0.45, 15.54	0.03
Naive	15.0	19.0	4.1	0.68, 7.38	21.3	3.91, 35.68	0.01

Abbreviations: CI, confidence interval; HPV, human papillomavirus.

^a χ² test for differential distribution between study arms.

Figure 2.

Proportion of vaccinated women referred to colposcopy, by time since vaccination, Costa Rica Vaccine Trial, 2004–2010. A) Total vaccinated cohort; B) naive cohort. Hep A, hepatitis A; HPV, human papillomavirus.

The distribution of HPV types among women referred to colposcopy in the naive cohort who received vaccination against HPV showed, at the time of referral, a scarcity of the HPV types included in the vaccine and of those types for which partial protection has been previously described. A similar effect was observed in the total vaccinated cohort, although it was less evident and was restricted to the types included in the vaccine formulation (Figure 3).

Figure 3.

Distribution of carcinogenic types of human papillomavirus (HPV) among vaccinated women at colposcopy referral, Costa Rica Vaccine Trial, 2004–2010. A) Total vaccinated cohort; B) naive cohort. Hep A, hepatitis A.

Regarding proportion of treatment, HPV-vaccinated women had fewer loop electrosurgical excision procedures done (5.0% as compared with 5.7% among women in the placebo arm), with a relative reduction in treatment rate of 11.3% (95% CI: −8.9, 27.8) in the total vaccinated cohort. The naive cohort showed a bigger relative reduction of 45.6% (95% CI: −9.3, 73.9), with a broad confidence interval (Table 4). In the total vaccinated cohort, the proportion of women treated in the control arm was higher than that for the HPV-vaccinated women, but only in the fourth year (Figure 4). In the naive cohort, HPV-vaccinated women underwent less treatment; however, the difference in treatment by arm was restricted to year 4 (Figure 4). To understand whether the overall difference in treatment proportions was due to having fewer women referred to colposcopy, we estimated the treatment rate restricted to women who underwent colposcopy, but we found only a small residual difference in treatment proportion between study arms for both cohorts (data not shown); thus, the decrease in treatment among vaccinated women was largely mediated by the decrease in referral to colposcopy.

Table 4.

Reduction in the Rate of Treatment Among Vaccinated Women, by Vaccine Type and Cohort, Costa Rica Vaccine Trial, 2004–2010

Cohort	% HPV-Vaccinated	% Hepatitis A-Vaccinated	Rate Reduction				P Valuea
			Absolute		Relative
			%	95% CI	%	95% CI
Total vaccinated	5.0	5.7	0.6	−0.46, 1.73	11.3	−8.91, 27.80	0.24
Naive	1.1	1.9	0.9	−0.13, 1.75	45.6	−9.34, 73.90	0.08

Abbreviations: CI, confidence interval; HPV, human papillomavirus.

^a χ² test for differential distribution between study arms.

Figure 4.

Proportion of vaccinated women treated with a loop electrosurgical excision procedure, by time since vaccination, Costa Rica Vaccine Trial, 2004–2010. A) Total vaccinated cohort; B) naive cohort. Hep A, hepatitis A; HPV, human papillomavirus.

DISCUSSION

In this community-based phase III HPV vaccine trial in Costa Rica, the overall 4-year effect of vaccinating young adult women aged 18–25 years against HPV 16/18 on subsequent screening, colposcopy referral, and treatment proportions was modest, with a significant reduction in abnormal cytology readings, a borderline-significant reduction in colposcopy, and a nonsignificant reduction in treatment rates among HPV-vaccinated women. We confirmed that among sexually active women, the impact of vaccination on cytology diagnosis, colposcopy referral, and treatment proportions was more evident among those with no obvious evidence of previous exposure to HPV, as expected. We also confirmed that most women aged 18–25 years, if tested at a given time, have evidence of previous HPV exposure (as shown by the relatively small size of the naive cohort vs. the total vaccinated cohort). Careful cost-benefit analyses should be carried out when considering expanding HPV vaccination programs to include young adult women.

Even though the Costa Rica Vaccine Trial was a phase III clinical trial, it represented an adequate setting for evaluating the public health impact of screening and vaccinating young adult women against HPV, given its community-based design; one-third of all women aged 18–25 years who were living in the study area were recruited (8). The total vaccinated cohort excluded only virginal women for whom we had no cytology results during the study and the very few women who received mixed vaccine types (n = 622 in total), thus providing a good surrogate measure for the general population. As expected in this age range, most women in the total vaccinated cohort (66.2%) had evidence of previous HPV exposure at study entry. The naive cohort was defined to be a highly restricted group that included only women without evidence of previous HPV exposure, providing, as closely as we could generate, a target group of HPV-naive women for whom vaccination could have the most impact. The findings for the naive cohort provide an estimate of what can be expected in terms of cervical cancer screening programs when the adolescents who are currently being vaccinated reach young adulthood and are screened annually.

The 4 years of follow-up of the total vaccinated cohort allowed us to calculate the impact of vaccination on cytology at the population level by directly comparing the frequency of cytological diagnosis by severity, after determining the worst-ever result for each woman, and study arm. First, we confirmed that among women in the age range 18–25 years, regardless of vaccination status, cytological evidence of HPV infection (LSIL or HPV DNA-positive ASC-US) is very common, especially if women are screened frequently. A total of 31.4% (95% CI: 30.3, 32.6) of the women had ASC-US/LSIL as their worst-ever cytology result, while an additional 7.0% (95% CI: 6.4, 7.6) had HSIL readings at some point during the trial. In the total vaccinated cohort, the relative reduction in HSIL and ASC-US/LSIL readings among HPV-vaccinated women was statistically significant but modest, since, as expected, many lesions and/or infections were already present at vaccination, even if enrollment cytology results are not taken into consideration. In the naive cohort, the reductions were more evident, especially for HSIL, because of the high vaccine efficacy against incident HPV 16/18 infections (the main causal types) and partial efficacy against other types shown for this vaccine (12, 13). The impact of HPV vaccination on cytology was expected based on analogous findings from the Papilloma Trial Against Cancer in Young Adults, better known as the PATRICIA Study, as described by Lehtinen et al. (4).

Colposcopy will also be affected as a consequence of decreased real CIN2+ cases. However, in this group of young adult women with comprehensive screening, the reduction in CIN2+ diagnosis among those sent to colposcopy was not significantly different by arm. The 4-year follow-up time might have been too short to show a significant difference. The proportion of women referred to colposcopy was lower in the HPV-vaccinated arm and fewer required treatment, especially in the last years of follow-up. The distribution of HPV types at the time women were referred to colposcopy differed by study arm, especially in the naive cohort, where a marked reduction was observed for the other HPV types for which partial vaccine protection has been shown (7, 12, 13). Based on previous knowledge of the natural history of HPV infection, an increase in the observed difference in referral and treatment proportions in future years would be expected, particularly among women without evidence of previous HPV exposure (14).

Although women included in the Costa Rica Vaccine Trial were in the age range being considered for catch-up vaccination (18–25 years), they were comprehensively screened for cervical HPV and lesions and at an earlier age than is recommend in most countries. Therefore, our results, examining screening starting immediately after vaccination, might underestimate the true effect vaccination would have on cervical cancer screening programs if these women were tested at older ages, by which time more incident abnormalities might develop in the unvaccinated group. Evidence from 2 other phase III clinical trials that included more selected populations also showed a modest impact on cytology, colposcopy referral, and treatment proportions in the first years after vaccination, with an increasing impact after the first 2 years (4, 5). The impact on cytology and colposcopy and treatment proportions would have been different, though difficult to predict, if a different screening strategy had been used—for example, using molecular detection as the primary screening method with extended screening intervals rather than cytology every year. Similarly, in a population with a different baseline prevalence of abnormal cytology readings, the absolute impact of vaccination would be different.

This study provides direct evidence of the impact on abnormal cytology, colposcopy referral, and treatment proportions that a public health vaccination strategy for young adult women would yield in the initial years following vaccination. Overall, the impact of vaccination among young adult women will be modest in a screening program if women are tested during the first few years after vaccine introduction, even in areas with high vaccine uptake. Any program aiming to measure the effect of vaccination on screening programs will need to incorporate long-term surveillance.