Journal of Clinical Virology 59 (2014) 109–114
Contents lists available at ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
UK population based study to predict impact of HPV vaccination Sam Hibbitts a,∗,1 , Amanda Tristram a,1 , Helen Beer b , Jane McRea b , Bryan Rose b , Anne Hauke b , Dave Nuttall b , Nick Dallimore b , Robert G. Newcombe a , Alison Fiander a a b
School of Medicine, Cardiff University, CF14 4XN Wales, UK Public Health Wales, Screening Division, Cervical Screening Wales, 18 Cathedral Road, Cardiff, CF11 9LJ Wales, UK
a r t i c l e
i n f o
Article history: Received 24 August 2013 Received in revised form 20 November 2013 Accepted 3 December 2013 Keywords: Human papillomavirus (HPV) Cervical screening HPV vaccination Cervarix and Gardasil
a b s t r a c t Background: In 2008 a human papillomavirus (HPV) vaccination programme for cervical cancer prevention was implemented in the UK. Surveillance of vaccine uptake, impact on prevalence of HPV infection and cervical cancer incidence were identified as key measures to evaluate the intervention. Objective: To determine baseline HPV prevalence in unvaccinated women and predict impact of HPV vaccination on high-grade cervical disease (CIN2+). Study design: A pseudo-anonymous prospective cohort was sampled on entry to the routine cervical screening programme between March 2009 and November 2010. In total, 13,306 eligible females were identified and high-risk (hrHPV) type specific status determined. Potential impact of prophylactic vaccination on CIN2+ was calculated by applying HPV vaccine clinical trial data to the baseline HPV type-specific data. Results: Of 13,306 samples tested, 3545 (26.6%) were confirmed positive for at least one hrHPV type and 1325 (10%) were positive for low risk HPV. HPV16 was the predominant type detected in cases positive with either single or multiple hrHPV infection(s) (5.2% and 4.7%, respectively). Based on hrHPV typespecific data, Gardasil would have prevented 33.2% HPV16/18 unrelated CIN2+ compared to 47.1% for Cervarix. This difference was not statistically significant. Conclusion: Prior to the introduction of the HPV vaccine, approximately one-quarter of young women were positive for hrHPV and one-tenth positive for HPV16. Post-vaccination, we anticipate a substantial absolute risk reduction in high-grade cervical disease associated with both targeted and non-targeted hrHPV types. There is no significant difference between the two commercially available vaccines in terms of clinical impact. © 2013 Elsevier B.V. All rights reserved.
1. Background Human papillomavirus (HPV) is highly prevalent in young women following onset of sexual activity [1]. Most HPV infections are transient, asymptomatic and regress spontaneously. However, persistent infection with high-risk HPV (hrHPV) is the underlying cause of cervical cancer [2,3] and two hrHPV types, HPV 16 and 18, account for more than 70% of cervical cancer worldwide [4,5]. Cervical screening programmes identify women at increased risk of developing cervical cancer and are estimated to save around 5000 lives per year in the UK [6]. In the UK, cervical screening programmes employ cervical cytology to identify dyskaryotic cells, colposcopic assessment followed by treatment of high-grade cervical intraepithelial neoplasia
∗ Corresponding author. Tel.: +44 029 2074 4713. E-mail addresses: [email protected], [email protected] (S. Hibbitts). 1 Joint first authors. 1386-6532/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jcv.2013.12.002
(CIN2+), to prevent progression to invasive disease. However, cytological screening is not infallible and there are ongoing changes in cervical screening with introduction of HPV testing in a triage setting and as a test of cure post-treatment, because of improved sensitivity and high negative predictive value for CIN2+ [7–10]. Prophylactic HPV vaccines are licensed in over 100 countries with two commercially available: Cervarix (GlaxoSmithKline, UK) targeting HPV 16 and 18; and Gardasil (Merck, USA) targeting hrHPV types 16 and 18 and low risk HPV (lrHPV) types 6 and 11. Both vaccines have demonstrated 100% efficacy for HPV 16/18 associated CIN3+ in women with no evidence of prior exposure to the virus [11]. Varying degrees of cross-protection for individual nonvaccine types have been reported, against persistent infection and CIN2+ [12,13]. UK wide implementation of HPV vaccination for cervical cancer prevention began with Cervarix in September 2008. The vaccination programme targeted girls aged 12–13 years through schools, with a catch-up campaign for girls aged 15–18 years. Government surveillance from September 2008 to 2012 indicated a greater than
110
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
80% uptake in the target group for all three doses of HPV vaccine [14]. In September 2012, Gardasil replaced Cervarix as the programme vaccine. A variety of HPV prevalence studies have been carried out in routine UK screening populations [1,5,15–18]. The studies have differed in sample size, age-range and HPV testing methodologies and did not include cohorts of young women large enough to detect early changes in non-16/18 HPV infection prevalence following the introduction of vaccination.
52; 56; 58; 59; 66 and 68; (iii) samples that were hrHPV positive in the HPV cocktail but negative following genotyping were classified as borderline for hrHPV and re-analysed with repeat genotyping using the 14 individual HRHPV probes. The hr/lrHPV PCR and genotyping PCRs were performed in final volumes of 25 l and 100 l, respectively. PCR cycling conditions were 94 ◦ C 4 min, 40 cycles of 94 ◦ C 30 s, 55 ◦ C 30 s, 72 ◦ C 30 s followed by 72 ◦ C 4 min. Positive (CaSki) and negative (water) DNA extraction, PCR and ELISA controls were included every 94 samples. A positive result was defined as three times background.
2. Objectives 3.3. Statistical analysis We aimed to analyse sufficient samples to report individual HPV type specific prevalence in women attending for their first smear, prior to introduction of the HPV vaccination, in order to provide a baseline with which to compare future cohorts and predict the potential impact of HPV vaccines on cervical disease. 3. Study design 3.1. Study population This study was conducted at Cardiff University in collaboration with Cervical Screening Wales (CSW) and was independent of the routine cervical screening programme. Women aged 20–22 years at last birthday, resident in Wales, not offered the HPV vaccine and attending their first call for cervical screening, were considered eligible for inclusion. Sample collection occurred between April 2009 and July 2010 within participating cervical screening laboratories across Wales: Llandough, Singleton, Llandudno, Royal Gwent Hospital, Wrexham Maelor, Ysbyty Glan Clwyd, Withybush, Royal Glamorgan, Princess of Wales, Prince Charles and West Wales General, and border English laboratories Hereford, Chester and Shrewsbury. Cervical samples in liquid based cytology (LBC, BD SurePath, Source Bioscience) were processed by CSW according to the British Society of Clinical Cytology guidelines (Advice for Cytopathology Laboratories, 2004) and patients were managed according to standard operating procedures. Residual LBC samples from eligible cases were flagged, patient identifiers removed and transported to Cardiff University, through the central transport system operating between hospitals in Wales (n = 14,128). CSW confirmed study eligibility and provided clinical data based on a pseudo-anonymous case ID. Data analysis was performed on all samples that complied with the following additional inclusion criteria: first adequate routine cervical smear; complete information available on age and cytology result; adequate HPV test control reading; -globin PCR positive (n = 13,306; 94%). 3.2. Sample storage, DNA extraction, and HPV testing Each LBC sample was washed in 1 ml 10 mM Tris pH 7.4, prior to storage at −80 ◦ C. DNA extraction used a Proteinase K digest with a 100 l cell suspension and 10 l recombinant proteinase K (Roche Diagnostics) incubation at 56 ◦ C for 2 h, then enzyme inactivation at 100 ◦ C for 10 min. Samples were refrigerated at 4 ◦ C for 10 min, centrifuged at 13,000 rpm for 10 min and the supernatant transferred into a 96-well plate and stored at −20 ◦ C until required for further analysis. A control PCR targeting the human -globin gene was performed to determine sample adequacy [15]. The GP5+/GP6+ HPV PCR-ELISA method [19] was performed on all specimens in a 96-well format with minor modifications: (i) PCR-ELISA with cocktails of hrHPV and lrHPV type-specific probes; (ii) PCR-ELISA of all hrHPV-positive samples with genotyping using 14 individual hrHPV probes: HPV 16; 18; 31; 33; 35; 39; 45; 51;
For analyses involving ordinal or quantitative variables (cytology grade, deprivation quintile and age), chi-square for linear by linear association (1 degree of freedom) is reported. To model projected vaccine impact using the baseline cohort, estimates of the proportion of CIN2+ excluding HPV16 and 18, potentially preventable by HPV vaccination were calculated. Odds ratios expressing cross-protection for relevant hrHPV types from the PATRICIA (Cervarix) and FUTURE I/II (Gardasil) vaccine trials were applied to the type-specific baseline data in the 3545-hrHPV positive cases. 100% protection was assumed for the 920 (26%) women with HPV16 and/or 18 infection only. Data were available for 12 (k) nonvaccine hrHPV types in the PATRICIA trial, and 10 (k) non-vaccine hrHPV types in the FUTURE I/II trials, for non-HPV16 and 18 CIN2+ cases. Most odds ratios were in the direction of benefit. Women, who were hrHPV positive in the baseline data series, were assigned a post-vaccination risk of positivity for their hrHPV type below 1, derived from the odds ratio expressing cross-protection for that type derived from the relevant trial. For a few hrHPV types where the odds ratios were in the direction of harm, a similar process was used to assign a low risk of positivity post-vaccination to each type-negative woman. The product of the complements of the post-vaccination risks for all relevant hrHPV types was subtracted from 1 to give the woman’s post-vaccination risk. The projected number of hrHPV positive women was calculated by summation, and then expressed as an absolute risk reduction (ARR) for the total 3545 hrHPV positive women in the baseline series. Confidence intervals were calculated expressing the impact of the finiteness of the trial series, which was the dominant source of imprecision. As an approximation to the Propagating Imprecision (PropImp) algorithm [20], in each analysis the above process using lower and upper [21] confidence limits with was repeated √ z = z0 / k for odds ratios for each of the k non-vaccine HR types included. Using z0 = 1.960 leads to lower and upper 95% limits for the relative risk reduction as shown in Table 3. Comparisons between quadrivalent and bivalent vaccines were run using the same k non-vaccine types √ for both √ vaccines. The PropImp algorithm was used with z1 / k and z2 / k limits for the k non-vaccine types for the two vaccines, then z1 and z2 chosen to maximise interval width subject to z12 + z22 = z02 . 4. Results The analyses reported here are based on 13,306 women that met the inclusion criteria. A social deprivation score (WIMD, revised 2008) was available for 12,803 women (Fig. 1). Sixty one percent (8079 of 13,306) of women attended screening as recommended, aged 20 years. A delay in attending for cervical screening was seen in 39% of women, who attended 12–35 months after their first invitation, aged 21–22 years.
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
111
Fig. 1. Study flow chart. Following routine cytological analysis, residual smear samples were transported to Cardiff University for HPV testing. Screening data was recorded for linkage with the HPV test result.
Fig. 2. HPV type specific infections prevalent in women aged 20–22 years (n = 13,306). In the 13,306 women with a satisfactory HPV test recorded, 3545 (26.6%) had a confirmed HRHPV type reported. 1958 (55%) of the HRHPV positive cases had a single HPV infection detected and 1587 (45%) had multiple HPV types identified. The proportion of cases in the population positive for the HRHPV genotypes tested for are shown and divided into cases with a single and a multiple infection.
High risk HPV infection was confirmed by genotyping in 3545 (26.6%) of the study population. In cases confirmed as hrHPV positive, 1958 (55%) had a single hrHPV type detected and 1587 (45%) had multiple hrHPV types. Ten percent (n = 1325) of the baseline cohort was low risk (lrHPV) HPV positive with 771 of the 1325 (58%) lrHPV cases also positive for hrHPV. The relationship between cytological grade, age, recorded social deprivation score (where available) and HPV status was examined (Table 1). The proportion of women with negative cytology increased significantly with increasing age, from 81.1% at 20 years to 84.3% at 22 years. The proportion with moderate or worse dyskaryosis also increased steadily, from 1.6% in 20 year olds to 2.3% in 22 year olds. No clear relationship between social deprivation score and age at attending first screening was detected. The association between hrHPV prevalence and cytology grade (Table 1) was highly significant (2 = 2456, df = 1, p < 0.0001) with more moderate or worse cytology in women who were hrHPV positive.
The hrHPV type specific prevalence in this young cohort of women is shown in Fig. 2. HPV16 was the predominant type detected in both single and multiple HPV infection (n = 696 (5.2%) and n = 627 (4.7%), respectively) and was the only hrHPV type with a higher prevalence in single infections compared to cases with multiple types present. For single infection the second most predominant type detected was HPV51 (1.4%) and in multiple infection the second most predominant type was HPV18 (3.0%). The relationship between hrHPV16, hrHPV18 and other hrHPV types, age, recorded social deprivation score (where available) and cytology was examined (Table 2). Women with severe dyskaryosis and glandular neoplasia had a significantly higher prevalence of ‘HPV16 only’, compared to women with negative/lower-grade cytological abnormalities. The predicted impact HPV vaccination would have on prevention of CIN2+ (excluding HPV16/18) based on the 3545 hrHPV women identified in the Baseline cohort is shown in Table 3. Excluding cases attributed to HPV16 and HPV18, Cervarix would prevent
112
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
Table 1 Cervical cytology results linked with age, social deprivation quintile, and HPV status. Numbers in each category are shown and the percentage of each variable is given in brackets. Variable
Negative
Borderline
Mild
Moderate
Severe dyskaryosis and glandular neoplasia
Total
20 years 21 years 22 years
6551 (81.1) 2564 (82.2) 1775 (84.3)
952 (11.8) 331 (10.6) 183 (8.7)
448 (5.5) 167 (5.4) 100 (4.7)
77 (1.0) 31 (1.0) 27 (1.3)
51 (0.6) 28 (0.9) 21 (0.9)
8079 3121 2106
Quintile 1 (least deprived) Quintile 2 Quintile 3 Quintile 4 Quintile 5 (most deprived)
2065 (81.3) 2069 (80.2) 2133 (82.2) 2051 (82.5) 2149 (82.6)
281 (11.1) 303 (11.7) 285 (11.0) 259 (10.4) 286 (11.0)
137 (5.4) 147 (5.7) 133 (5.1) 136 (5.5) 140 (5.4)
27 (1.1) 36 (1.4) 24 (0.9) 28 (1.1) 16 (0.6)
30 (1.2) 26 (1.0) 20 (0.8) 12 (0.5) 10 (0.4)
2540 2581 2595 2486 2601
hrHPV and lrHPV negative hrHPV only positive lrHPV only positive hr and lrHPV positive
8646 (93.9) 1548 (55.8) 367 (66.2) 329 (42.7)
399 (4.3) 660 (23.8) 139 (25.1) 268 (34.8)
135 (1.5) 395 (14.2) 43 (7.8) 142 (18.4)
17 (0.2) 94 (3.4) 3 (0.5) 21 (2.7)
10 (0.1) 77 (2.8) 2 (0.4) 11 (1.4)
9207 2774 554 771
Table 2 Relationship of hrHPV16, hrHPV18 and other hrHPV types with age, social deprivation score and cytology. Numbers in each category are shown and the percentage of each variable is given in brackets. Variable
hrHPV16 only
hrHPV18 only
hrHPV 16 and hr HPV18
Other hrHPV non-vaccine targeted with or without HPV16 or HPV18 (k = 12)
Total
20 years 21 years 22 years
430 (19.2) 172 (20.8) 94 (19.7)
112 (5.0) 43 (5.2) 16 (3.4)
36 (1.6) 10 (1.2) 7 (1.5)
1662 (74.2) 603 (72.8) 360 (75.5)
2240 828 477
Quintile 1 (least deprived) Quintile 2 Quintile 3 Quintile 4 Quintile 5 (most deprived)
153 (21.6) 164 (22.5) 121 (18.7) 121 (17.7) 121 (18.3)
37 (5.2) 34 (4.7) 25 (3.9) 33 (4.8) 38 (5.7)
9 (1.3) 8 (1.1) 10 (1.5) 8 (1.2) 17 (2.6)
510 (71.9) 524 (71.8) 491 (75.9) 522 (76.3) 485 (73.4)
709 730 647 684 661
Negative Borderline Mild Moderate Severe dyskaryosis and glandular neoplasia
437 (23.3) 128 (13.8) 74 (13.8) 26 (22.6) 31 (35.2)
90 (4.8) 57 (6.1) 18 (3.4) 4 (3.5) 2 (2.3)
28 (1.5) 13 (1.4) 8 (1.5) 3 (2.6) 1 (1.1)
1322 (70.4) 730 (78.7) 437 (81.4) 82 (71.3) 54 (61.4)
1877 928 537 115 88
Table 3 Absolute risk reduction analyses against CIN2+ (excluding HPV16/18) based on hrHPV type frequency data from the baseline cohort and odds ratios expressing cross-protection for bivalent and quadrivalent vaccines from the TVC-naïve cohorts in the PATRICIA and FUTURE I/II trials. Vaccine
Number of hrHPV types included in analysis
Projected absolute risk reduction
95% confidence intervala
Bivalent CERVARIX
12 10 10 10
48.3% 47.1% 33.2% 13.8%
27.0–61.1% 27.5–57.8% 13.6–45.5% −9.4% to 36.5%
Quadrivalent GARDASIL Quadrivalent versus Bivalentb a b
95% confidence intervals calculated by the Propagating Imprecision algorithm express imprecision from the trials series. Differences between absolute risk reductions for the quadrivalent and bivalent vaccines.
48.3% hrHPV cases and Gardasil would prevent 33.2%. However, data for Cervarix is based upon trial data for 12 hrHPV types and Gardasil is based on 10 hrHPV types. When the analysis was restricted to the 10 hrHPV types common to both trials, the difference in protection was 13.8% (95% CI −9.4% to 36.5%) in favour of Cervarix. Given the small number of CIN2+ cases associated with non-vaccine hrHPV types in the vaccine trial data, there are wide confidence intervals in the estimated absolute risk reductions (ARRs) that extend below 0, such that the difference does not approach significance. 5. Discussion This is the largest UK based study (n = 13,306) investigating HPV prevalence in young women (20–22 years) attending their first cervical smear. 26.6% of cases were confirmed positive for at least one hrHPV type and 10% were positive for lrHPV. The majority of women invited for screening aged 20 years responded within the first 12 months (61%). This young cohort had a high proportion of
borderline and mild dyskaryosis compared to all women attending routine cervical screening (20–64 years). These low grade cytological changes may be attributed to the high prevalence of hrHPV in this population and are anticipated following onset of recent sexual activity and potential acquisition of transient HPV infections, with only a minority expected to persist and progress to anogenital neoplasia. The data presented is consistent with this finding and HPV infection was confirmed in a significant proportion of cases with reported borderline and mild dyskaryosis (Table 1). HPV prevalence was also correlated with age (Table 2). The lower hrHPV prevalence in women aged 22 years may indicate that women with a higher self-assessed risk are more concerned and attend for screening sooner than those who perceive their risk as lower. The baseline HPV prevalence data generated in this study was used to model the potential impact of vaccination with Cervarix and Gardasil to determine absolute risk reduction in HPV16/18 unrelated CIN2+. The vaccine modelling method involves the PropImp algorithm. In all analyses reported the PATRICIA and FUTURE I/II vaccine trial data was the dominant source of uncertainty
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
associated with trial size, and this is reflected in the choice of analyses to determine potential protection against high-grade (CIN2+) disease. Modelling was limited by the data reported in the two vaccine trials. To enable an unbiased comparison, the modelling utilised the most comprehensive data available for both vaccines, which restricted analyses to the CIN2+ end-point and TVC-naïve cohort. Assuming protection reported according to the TVC-naïve groups in the PATRICIA and FUTURE I/II trials, with the crossprotection benefit demonstrated by Cervarix a higher percentage of cases would be protected against HPV16/18 unrelated CIN2+ compared to Gardasil, although this difference is not statistically significant. Previous UK studies have investigated the prevalence of hrHPV in the routine screening population (20–64 years) and demonstrated the high prevalence of hrHPV types 16 and 18 [1,5,15–18]. The prevalence of non-16/18 hrHPV types varies between studies and may be attributed to differences between HPV tests utilised, samples types, sample sizes and geographical heterogeneity of the hrHPV type distribution. This is the first study to look specifically at young women having their first cervical smear in order to identify prevalent HPV types upon entry into a cervical screening programme. In this baseline cohort HPV16 was the predominant type in both multiple and single infections. In contrast to other studies, HPV51 was the second most predominant hrHPV type in single HPV infection. Jit et al. [22] used a transmission model to compare efficacy and cost-effectiveness of bivalent and quadrivalent vaccines and concluded that the bivalent vaccine may have an advantage in preventing deaths due to cancer. The quadrivalent vaccine provides additional protection against lrHPV types 6 and 11, which are responsible for genital warts and in this respect confers an advantage over the bivalent vaccine in terms of healthcare costs and QALYs [22]. The modelling of vaccine impact using the baseline cohort of women aged 20–22 years, supports the observation that the bivalent vaccine through its reported cross-protection against other hrHPV types would confer an additional benefit in terms of preventing HPV16/18 unrelated CIN2+ in the population. Overall, the carriage of a single infection vis-à-vis a multiple infection did not differ significantly between women aged 20, 21 or 22 years. However, a higher number of moderate to severe dyskaryotic changes were detected in women who had multiple HPV infection. In studies of women of all ages (20–64 years) participating in routine cervical screening this association is not apparent and is likely a result of the screening programme effectively detecting and treating such cases [1]. In conclusion, before introduction of HPV vaccination, approximately one-quarter of women aged 20–22 years were positive for high risk HPV. The predominant types detected were HPV16, HPV18, HPV31, HPV51 and HPV56. With the high rates of HPV51 and HPV56 detected, monitoring of changes in HPV prevalence following the introduction of the HPV vaccination programme should be a priority. Post-vaccination, we anticipate a substantial absolute risk reduction in high-grade cervical disease associated with targeted and non-targeted hrHPV types. Funding The study was funded by Cancer Research Wales and sponsored by Cardiff University who were responsible for research governance. Competing interests SH and AT were on the ROCHE advisory board for HPV testing and ROCHE sponsored SH to attend the IPV conference in 2010 and
113
sponsored SH and AF to attend the IPV conference in 2011. AT and AF were on the GSK HPV vaccine advisory board and AF on the Sanofi Pasteur MSD HPV vaccine advisory board. AT and AF have received sponsorship from both GSK and Sanofi Pasteur MSD to attend HPV conferences and have two grants from GSK and Sanofi Pasteur MSD for research projects on a HPV vaccine and the psychosocial effects of HPV infection. RN has done HPV unrelated consultancy work for GSK. There have been no other financial relationships with any organisations that might have an interest in the submitted work in the previous three years. Ethical approval Dyfed Powys Local Research Ethics Committee approved the study for All Wales recruitment (08/WMW01/69). Acknowledgements The authors gratefully acknowledge all participating cytology labs within Cervical Screening Wales who contributed to the sample identification and the pseudo-anonymisation process. The authors also gratefully acknowledge the research technicians who contributed to the HPV testing, in particular: Joanne Jones, Angharad Edwards and Vasiliki Kiparoglou. Thanks also to advisors Dr Shantini Paranjothy and Dr Malcolm Adams. References [1] Hibbitts S, Jones J, Powell N, Dallimore N, McRea J, Beer H, et al. Human papillomavirus prevalence in women attending routine cervical screening in South Wales, UK: a cross-sectional study. Br J Cancer 2008;99:1929–33. [2] Bosch FX, Schwarz E, Boukamp P, Fusenig NE, Bartsch D, zur Hausen H. Suppression in vivo of human papillomavirus type 18 E6-E7 gene expression in nontumorigenic HeLa X fibroblast hybrid cells. J Virol 1990;64:4743–54. [3] Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah KV, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189:12–9. [4] Munoz N. Human papillomavirus and cancer: the epidemiological evidence. J Clin Virol 2000;19:1–5. [5] Powell N, Boyde A, Tristram A, Hibbitts S, Fiander A. The potential impact of human papillomavirus vaccination in contemporary cytologically screened populations may be underestimated: an observational retrospective analysis of invasive cervical cancers. Int J Cancer 2009;125:2425–7. [6] Peto J, Gilham C, Fletcher O, Matthews FE. The cervical cancer epidemic that screening has prevented in the UK. Lancet 2004;364:249–56. [7] Ronco G, Giorgi-Rossi P, Carozzi F, Confortini M, Dalla Palma P, Del Mistro A, et al. Efficacy of human papillomavirus testing for the detection of invasive cervical cancers and cervical intraepithelial neoplasia: a randomised controlled trial. Lancet Oncol 2010;11:249–57. [8] Cuzick J, Clavel C, Petry KU, Meijer CJ, Hoyer H, Ratnam S, et al. Overview of the European and North American studies on HPV testing in primary cervical cancer screening. Int J Cancer 2006;119:1095–101. [9] Kitchener HC, Gilham C, Sargent A, Bailey A, Albrow R, Roberts C, et al. A comparison of HPV DNA testing and liquid based cytology over three rounds of primary cervical screening: Extended follow up in the ARTISTIC trial. Eur J Cancer 2011;47:864–71. [10] Stoler MH, Wright Jr TC, Sharma A, Apple R, Gutekunst K, Wright TL. High-risk human papillomavirus testing in women with ASC-US cytology: results from the ATHENA HPV study. Am J Clin Pathol 2011;135:468–75. [11] Lehtinen M, Paavonen J, Wheeler CM, Jaisamrarn U, Garland SM, Castellsague X, et al. Overall efficacy of HPV-16/18 AS04-adjuvanted vaccine against grade 3 or greater cervical intraepithelial neoplasia: 4-year end-of-study analysis of the randomised, double-blind PATRICIA trial. Lancet Oncol 2012;13:89–99. [12] Wheeler CM, Castellsague X, Garland SM, Szarewski A, Paavonen J, Naud P, et al. Cross-protective efficacy of HPV-16/18 AS04-adjuvanted vaccine against cervical infection and precancer caused by non-vaccine oncogenic HPV types: 4-year end-of-study analysis of the randomised, double-blind PATRICIA trial. Lancet Oncol 2012;13:100–10. [13] Brown DR, Kjaer SK, Sigurdsson K, Iversen OE, Hernandez-Avila M, Wheeler CM, et al. The impact of quadrivalent human papillomavirus (HPV; types 6, 11, 16, and 18) L1 virus-like particle vaccine on infection and disease due to oncogenic nonvaccine HPV types in generally HPV-naive women aged 16–26 years. J Infect Dis 2009;199:926–35. [14] Wales PH. Vaccine uptake in children in Wales July to September 2012. In: Wales PH, editor. Coverage of Vaccination Evaluated Rapidly (COVER). Cardiff: The National Publich Health Service for Wales; 2012.
114
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
[15] Hibbitts S, Rieck GC, Hart K, Powell NG, Beukenholdt R, Dallimore N, et al. Human papillomavirus infection: an anonymous prevalence study in South Wales, UK. Br J Cancer 2006;95:226–32. [16] Howell-Jones R, Bailey A, Beddows S, Sargent A, de Silva N, Wilson G, et al. Multi-site study of HPV type-specific prevalence in women with cervical cancer, intraepithelial neoplasia and normal cytology, in England. Br J Cancer 2010;103:209–16. [17] Cuschieri KS, Cubie HA, Whitley MW, Seagar AL, Arends MJ, Moore C, et al. Multiple high risk HPV infections are common in cervical neoplasia and young women in a cervical screening population. J Clin Pathol 2004;57:68–72. [18] Kitchener HC, Almonte M, Wheeler P, Desai M, Gilham C, Bailey A, et al. HPV testing in routine cervical screening: cross sectional data from the ARTISTIC trial. Br J Cancer 2006;95:56–61.
[19] Jacobs MV, Snijders PJ, van den Brule AJ, Helmerhorst TJ, Meijer CJ, Walboomers JM. A general primer GP5+/GP6(+)-mediated PCR-enzyme immunoassay method for rapid detection of 14 high-risk and 6 low-risk human papillomavirus genotypes in cervical scrapings. J Clin Microbiol 1997;35: 791–5. [20] Newcombe RG. Propagating imprecision: combining confidence intervals from independent sources. Communications in Statistics - Theory & Methods 2011;40:3154–80. [21] Miettinen O, Nurminen M. Comparative analysis of two rates. Stat Med 1985;4:213–26. [22] Jit M, Chapman R, Hughes O, Choi YH. Comparing bivalent and quadrivalent human papillomavirus vaccines: economic evaluation based on transmission model. BMJ 2011;343:d5775.
Contents lists available at ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
UK population based study to predict impact of HPV vaccination Sam Hibbitts a,∗,1 , Amanda Tristram a,1 , Helen Beer b , Jane McRea b , Bryan Rose b , Anne Hauke b , Dave Nuttall b , Nick Dallimore b , Robert G. Newcombe a , Alison Fiander a a b
School of Medicine, Cardiff University, CF14 4XN Wales, UK Public Health Wales, Screening Division, Cervical Screening Wales, 18 Cathedral Road, Cardiff, CF11 9LJ Wales, UK
a r t i c l e
i n f o
Article history: Received 24 August 2013 Received in revised form 20 November 2013 Accepted 3 December 2013 Keywords: Human papillomavirus (HPV) Cervical screening HPV vaccination Cervarix and Gardasil
a b s t r a c t Background: In 2008 a human papillomavirus (HPV) vaccination programme for cervical cancer prevention was implemented in the UK. Surveillance of vaccine uptake, impact on prevalence of HPV infection and cervical cancer incidence were identified as key measures to evaluate the intervention. Objective: To determine baseline HPV prevalence in unvaccinated women and predict impact of HPV vaccination on high-grade cervical disease (CIN2+). Study design: A pseudo-anonymous prospective cohort was sampled on entry to the routine cervical screening programme between March 2009 and November 2010. In total, 13,306 eligible females were identified and high-risk (hrHPV) type specific status determined. Potential impact of prophylactic vaccination on CIN2+ was calculated by applying HPV vaccine clinical trial data to the baseline HPV type-specific data. Results: Of 13,306 samples tested, 3545 (26.6%) were confirmed positive for at least one hrHPV type and 1325 (10%) were positive for low risk HPV. HPV16 was the predominant type detected in cases positive with either single or multiple hrHPV infection(s) (5.2% and 4.7%, respectively). Based on hrHPV typespecific data, Gardasil would have prevented 33.2% HPV16/18 unrelated CIN2+ compared to 47.1% for Cervarix. This difference was not statistically significant. Conclusion: Prior to the introduction of the HPV vaccine, approximately one-quarter of young women were positive for hrHPV and one-tenth positive for HPV16. Post-vaccination, we anticipate a substantial absolute risk reduction in high-grade cervical disease associated with both targeted and non-targeted hrHPV types. There is no significant difference between the two commercially available vaccines in terms of clinical impact. © 2013 Elsevier B.V. All rights reserved.
1. Background Human papillomavirus (HPV) is highly prevalent in young women following onset of sexual activity [1]. Most HPV infections are transient, asymptomatic and regress spontaneously. However, persistent infection with high-risk HPV (hrHPV) is the underlying cause of cervical cancer [2,3] and two hrHPV types, HPV 16 and 18, account for more than 70% of cervical cancer worldwide [4,5]. Cervical screening programmes identify women at increased risk of developing cervical cancer and are estimated to save around 5000 lives per year in the UK [6]. In the UK, cervical screening programmes employ cervical cytology to identify dyskaryotic cells, colposcopic assessment followed by treatment of high-grade cervical intraepithelial neoplasia
∗ Corresponding author. Tel.: +44 029 2074 4713. E-mail addresses: [email protected], [email protected] (S. Hibbitts). 1 Joint first authors. 1386-6532/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jcv.2013.12.002
(CIN2+), to prevent progression to invasive disease. However, cytological screening is not infallible and there are ongoing changes in cervical screening with introduction of HPV testing in a triage setting and as a test of cure post-treatment, because of improved sensitivity and high negative predictive value for CIN2+ [7–10]. Prophylactic HPV vaccines are licensed in over 100 countries with two commercially available: Cervarix (GlaxoSmithKline, UK) targeting HPV 16 and 18; and Gardasil (Merck, USA) targeting hrHPV types 16 and 18 and low risk HPV (lrHPV) types 6 and 11. Both vaccines have demonstrated 100% efficacy for HPV 16/18 associated CIN3+ in women with no evidence of prior exposure to the virus [11]. Varying degrees of cross-protection for individual nonvaccine types have been reported, against persistent infection and CIN2+ [12,13]. UK wide implementation of HPV vaccination for cervical cancer prevention began with Cervarix in September 2008. The vaccination programme targeted girls aged 12–13 years through schools, with a catch-up campaign for girls aged 15–18 years. Government surveillance from September 2008 to 2012 indicated a greater than
110
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
80% uptake in the target group for all three doses of HPV vaccine [14]. In September 2012, Gardasil replaced Cervarix as the programme vaccine. A variety of HPV prevalence studies have been carried out in routine UK screening populations [1,5,15–18]. The studies have differed in sample size, age-range and HPV testing methodologies and did not include cohorts of young women large enough to detect early changes in non-16/18 HPV infection prevalence following the introduction of vaccination.
52; 56; 58; 59; 66 and 68; (iii) samples that were hrHPV positive in the HPV cocktail but negative following genotyping were classified as borderline for hrHPV and re-analysed with repeat genotyping using the 14 individual HRHPV probes. The hr/lrHPV PCR and genotyping PCRs were performed in final volumes of 25 l and 100 l, respectively. PCR cycling conditions were 94 ◦ C 4 min, 40 cycles of 94 ◦ C 30 s, 55 ◦ C 30 s, 72 ◦ C 30 s followed by 72 ◦ C 4 min. Positive (CaSki) and negative (water) DNA extraction, PCR and ELISA controls were included every 94 samples. A positive result was defined as three times background.
2. Objectives 3.3. Statistical analysis We aimed to analyse sufficient samples to report individual HPV type specific prevalence in women attending for their first smear, prior to introduction of the HPV vaccination, in order to provide a baseline with which to compare future cohorts and predict the potential impact of HPV vaccines on cervical disease. 3. Study design 3.1. Study population This study was conducted at Cardiff University in collaboration with Cervical Screening Wales (CSW) and was independent of the routine cervical screening programme. Women aged 20–22 years at last birthday, resident in Wales, not offered the HPV vaccine and attending their first call for cervical screening, were considered eligible for inclusion. Sample collection occurred between April 2009 and July 2010 within participating cervical screening laboratories across Wales: Llandough, Singleton, Llandudno, Royal Gwent Hospital, Wrexham Maelor, Ysbyty Glan Clwyd, Withybush, Royal Glamorgan, Princess of Wales, Prince Charles and West Wales General, and border English laboratories Hereford, Chester and Shrewsbury. Cervical samples in liquid based cytology (LBC, BD SurePath, Source Bioscience) were processed by CSW according to the British Society of Clinical Cytology guidelines (Advice for Cytopathology Laboratories, 2004) and patients were managed according to standard operating procedures. Residual LBC samples from eligible cases were flagged, patient identifiers removed and transported to Cardiff University, through the central transport system operating between hospitals in Wales (n = 14,128). CSW confirmed study eligibility and provided clinical data based on a pseudo-anonymous case ID. Data analysis was performed on all samples that complied with the following additional inclusion criteria: first adequate routine cervical smear; complete information available on age and cytology result; adequate HPV test control reading; -globin PCR positive (n = 13,306; 94%). 3.2. Sample storage, DNA extraction, and HPV testing Each LBC sample was washed in 1 ml 10 mM Tris pH 7.4, prior to storage at −80 ◦ C. DNA extraction used a Proteinase K digest with a 100 l cell suspension and 10 l recombinant proteinase K (Roche Diagnostics) incubation at 56 ◦ C for 2 h, then enzyme inactivation at 100 ◦ C for 10 min. Samples were refrigerated at 4 ◦ C for 10 min, centrifuged at 13,000 rpm for 10 min and the supernatant transferred into a 96-well plate and stored at −20 ◦ C until required for further analysis. A control PCR targeting the human -globin gene was performed to determine sample adequacy [15]. The GP5+/GP6+ HPV PCR-ELISA method [19] was performed on all specimens in a 96-well format with minor modifications: (i) PCR-ELISA with cocktails of hrHPV and lrHPV type-specific probes; (ii) PCR-ELISA of all hrHPV-positive samples with genotyping using 14 individual hrHPV probes: HPV 16; 18; 31; 33; 35; 39; 45; 51;
For analyses involving ordinal or quantitative variables (cytology grade, deprivation quintile and age), chi-square for linear by linear association (1 degree of freedom) is reported. To model projected vaccine impact using the baseline cohort, estimates of the proportion of CIN2+ excluding HPV16 and 18, potentially preventable by HPV vaccination were calculated. Odds ratios expressing cross-protection for relevant hrHPV types from the PATRICIA (Cervarix) and FUTURE I/II (Gardasil) vaccine trials were applied to the type-specific baseline data in the 3545-hrHPV positive cases. 100% protection was assumed for the 920 (26%) women with HPV16 and/or 18 infection only. Data were available for 12 (k) nonvaccine hrHPV types in the PATRICIA trial, and 10 (k) non-vaccine hrHPV types in the FUTURE I/II trials, for non-HPV16 and 18 CIN2+ cases. Most odds ratios were in the direction of benefit. Women, who were hrHPV positive in the baseline data series, were assigned a post-vaccination risk of positivity for their hrHPV type below 1, derived from the odds ratio expressing cross-protection for that type derived from the relevant trial. For a few hrHPV types where the odds ratios were in the direction of harm, a similar process was used to assign a low risk of positivity post-vaccination to each type-negative woman. The product of the complements of the post-vaccination risks for all relevant hrHPV types was subtracted from 1 to give the woman’s post-vaccination risk. The projected number of hrHPV positive women was calculated by summation, and then expressed as an absolute risk reduction (ARR) for the total 3545 hrHPV positive women in the baseline series. Confidence intervals were calculated expressing the impact of the finiteness of the trial series, which was the dominant source of imprecision. As an approximation to the Propagating Imprecision (PropImp) algorithm [20], in each analysis the above process using lower and upper [21] confidence limits with was repeated √ z = z0 / k for odds ratios for each of the k non-vaccine HR types included. Using z0 = 1.960 leads to lower and upper 95% limits for the relative risk reduction as shown in Table 3. Comparisons between quadrivalent and bivalent vaccines were run using the same k non-vaccine types √ for both √ vaccines. The PropImp algorithm was used with z1 / k and z2 / k limits for the k non-vaccine types for the two vaccines, then z1 and z2 chosen to maximise interval width subject to z12 + z22 = z02 . 4. Results The analyses reported here are based on 13,306 women that met the inclusion criteria. A social deprivation score (WIMD, revised 2008) was available for 12,803 women (Fig. 1). Sixty one percent (8079 of 13,306) of women attended screening as recommended, aged 20 years. A delay in attending for cervical screening was seen in 39% of women, who attended 12–35 months after their first invitation, aged 21–22 years.
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
111
Fig. 1. Study flow chart. Following routine cytological analysis, residual smear samples were transported to Cardiff University for HPV testing. Screening data was recorded for linkage with the HPV test result.
Fig. 2. HPV type specific infections prevalent in women aged 20–22 years (n = 13,306). In the 13,306 women with a satisfactory HPV test recorded, 3545 (26.6%) had a confirmed HRHPV type reported. 1958 (55%) of the HRHPV positive cases had a single HPV infection detected and 1587 (45%) had multiple HPV types identified. The proportion of cases in the population positive for the HRHPV genotypes tested for are shown and divided into cases with a single and a multiple infection.
High risk HPV infection was confirmed by genotyping in 3545 (26.6%) of the study population. In cases confirmed as hrHPV positive, 1958 (55%) had a single hrHPV type detected and 1587 (45%) had multiple hrHPV types. Ten percent (n = 1325) of the baseline cohort was low risk (lrHPV) HPV positive with 771 of the 1325 (58%) lrHPV cases also positive for hrHPV. The relationship between cytological grade, age, recorded social deprivation score (where available) and HPV status was examined (Table 1). The proportion of women with negative cytology increased significantly with increasing age, from 81.1% at 20 years to 84.3% at 22 years. The proportion with moderate or worse dyskaryosis also increased steadily, from 1.6% in 20 year olds to 2.3% in 22 year olds. No clear relationship between social deprivation score and age at attending first screening was detected. The association between hrHPV prevalence and cytology grade (Table 1) was highly significant (2 = 2456, df = 1, p < 0.0001) with more moderate or worse cytology in women who were hrHPV positive.
The hrHPV type specific prevalence in this young cohort of women is shown in Fig. 2. HPV16 was the predominant type detected in both single and multiple HPV infection (n = 696 (5.2%) and n = 627 (4.7%), respectively) and was the only hrHPV type with a higher prevalence in single infections compared to cases with multiple types present. For single infection the second most predominant type detected was HPV51 (1.4%) and in multiple infection the second most predominant type was HPV18 (3.0%). The relationship between hrHPV16, hrHPV18 and other hrHPV types, age, recorded social deprivation score (where available) and cytology was examined (Table 2). Women with severe dyskaryosis and glandular neoplasia had a significantly higher prevalence of ‘HPV16 only’, compared to women with negative/lower-grade cytological abnormalities. The predicted impact HPV vaccination would have on prevention of CIN2+ (excluding HPV16/18) based on the 3545 hrHPV women identified in the Baseline cohort is shown in Table 3. Excluding cases attributed to HPV16 and HPV18, Cervarix would prevent
112
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
Table 1 Cervical cytology results linked with age, social deprivation quintile, and HPV status. Numbers in each category are shown and the percentage of each variable is given in brackets. Variable
Negative
Borderline
Mild
Moderate
Severe dyskaryosis and glandular neoplasia
Total
20 years 21 years 22 years
6551 (81.1) 2564 (82.2) 1775 (84.3)
952 (11.8) 331 (10.6) 183 (8.7)
448 (5.5) 167 (5.4) 100 (4.7)
77 (1.0) 31 (1.0) 27 (1.3)
51 (0.6) 28 (0.9) 21 (0.9)
8079 3121 2106
Quintile 1 (least deprived) Quintile 2 Quintile 3 Quintile 4 Quintile 5 (most deprived)
2065 (81.3) 2069 (80.2) 2133 (82.2) 2051 (82.5) 2149 (82.6)
281 (11.1) 303 (11.7) 285 (11.0) 259 (10.4) 286 (11.0)
137 (5.4) 147 (5.7) 133 (5.1) 136 (5.5) 140 (5.4)
27 (1.1) 36 (1.4) 24 (0.9) 28 (1.1) 16 (0.6)
30 (1.2) 26 (1.0) 20 (0.8) 12 (0.5) 10 (0.4)
2540 2581 2595 2486 2601
hrHPV and lrHPV negative hrHPV only positive lrHPV only positive hr and lrHPV positive
8646 (93.9) 1548 (55.8) 367 (66.2) 329 (42.7)
399 (4.3) 660 (23.8) 139 (25.1) 268 (34.8)
135 (1.5) 395 (14.2) 43 (7.8) 142 (18.4)
17 (0.2) 94 (3.4) 3 (0.5) 21 (2.7)
10 (0.1) 77 (2.8) 2 (0.4) 11 (1.4)
9207 2774 554 771
Table 2 Relationship of hrHPV16, hrHPV18 and other hrHPV types with age, social deprivation score and cytology. Numbers in each category are shown and the percentage of each variable is given in brackets. Variable
hrHPV16 only
hrHPV18 only
hrHPV 16 and hr HPV18
Other hrHPV non-vaccine targeted with or without HPV16 or HPV18 (k = 12)
Total
20 years 21 years 22 years
430 (19.2) 172 (20.8) 94 (19.7)
112 (5.0) 43 (5.2) 16 (3.4)
36 (1.6) 10 (1.2) 7 (1.5)
1662 (74.2) 603 (72.8) 360 (75.5)
2240 828 477
Quintile 1 (least deprived) Quintile 2 Quintile 3 Quintile 4 Quintile 5 (most deprived)
153 (21.6) 164 (22.5) 121 (18.7) 121 (17.7) 121 (18.3)
37 (5.2) 34 (4.7) 25 (3.9) 33 (4.8) 38 (5.7)
9 (1.3) 8 (1.1) 10 (1.5) 8 (1.2) 17 (2.6)
510 (71.9) 524 (71.8) 491 (75.9) 522 (76.3) 485 (73.4)
709 730 647 684 661
Negative Borderline Mild Moderate Severe dyskaryosis and glandular neoplasia
437 (23.3) 128 (13.8) 74 (13.8) 26 (22.6) 31 (35.2)
90 (4.8) 57 (6.1) 18 (3.4) 4 (3.5) 2 (2.3)
28 (1.5) 13 (1.4) 8 (1.5) 3 (2.6) 1 (1.1)
1322 (70.4) 730 (78.7) 437 (81.4) 82 (71.3) 54 (61.4)
1877 928 537 115 88
Table 3 Absolute risk reduction analyses against CIN2+ (excluding HPV16/18) based on hrHPV type frequency data from the baseline cohort and odds ratios expressing cross-protection for bivalent and quadrivalent vaccines from the TVC-naïve cohorts in the PATRICIA and FUTURE I/II trials. Vaccine
Number of hrHPV types included in analysis
Projected absolute risk reduction
95% confidence intervala
Bivalent CERVARIX
12 10 10 10
48.3% 47.1% 33.2% 13.8%
27.0–61.1% 27.5–57.8% 13.6–45.5% −9.4% to 36.5%
Quadrivalent GARDASIL Quadrivalent versus Bivalentb a b
95% confidence intervals calculated by the Propagating Imprecision algorithm express imprecision from the trials series. Differences between absolute risk reductions for the quadrivalent and bivalent vaccines.
48.3% hrHPV cases and Gardasil would prevent 33.2%. However, data for Cervarix is based upon trial data for 12 hrHPV types and Gardasil is based on 10 hrHPV types. When the analysis was restricted to the 10 hrHPV types common to both trials, the difference in protection was 13.8% (95% CI −9.4% to 36.5%) in favour of Cervarix. Given the small number of CIN2+ cases associated with non-vaccine hrHPV types in the vaccine trial data, there are wide confidence intervals in the estimated absolute risk reductions (ARRs) that extend below 0, such that the difference does not approach significance. 5. Discussion This is the largest UK based study (n = 13,306) investigating HPV prevalence in young women (20–22 years) attending their first cervical smear. 26.6% of cases were confirmed positive for at least one hrHPV type and 10% were positive for lrHPV. The majority of women invited for screening aged 20 years responded within the first 12 months (61%). This young cohort had a high proportion of
borderline and mild dyskaryosis compared to all women attending routine cervical screening (20–64 years). These low grade cytological changes may be attributed to the high prevalence of hrHPV in this population and are anticipated following onset of recent sexual activity and potential acquisition of transient HPV infections, with only a minority expected to persist and progress to anogenital neoplasia. The data presented is consistent with this finding and HPV infection was confirmed in a significant proportion of cases with reported borderline and mild dyskaryosis (Table 1). HPV prevalence was also correlated with age (Table 2). The lower hrHPV prevalence in women aged 22 years may indicate that women with a higher self-assessed risk are more concerned and attend for screening sooner than those who perceive their risk as lower. The baseline HPV prevalence data generated in this study was used to model the potential impact of vaccination with Cervarix and Gardasil to determine absolute risk reduction in HPV16/18 unrelated CIN2+. The vaccine modelling method involves the PropImp algorithm. In all analyses reported the PATRICIA and FUTURE I/II vaccine trial data was the dominant source of uncertainty
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
associated with trial size, and this is reflected in the choice of analyses to determine potential protection against high-grade (CIN2+) disease. Modelling was limited by the data reported in the two vaccine trials. To enable an unbiased comparison, the modelling utilised the most comprehensive data available for both vaccines, which restricted analyses to the CIN2+ end-point and TVC-naïve cohort. Assuming protection reported according to the TVC-naïve groups in the PATRICIA and FUTURE I/II trials, with the crossprotection benefit demonstrated by Cervarix a higher percentage of cases would be protected against HPV16/18 unrelated CIN2+ compared to Gardasil, although this difference is not statistically significant. Previous UK studies have investigated the prevalence of hrHPV in the routine screening population (20–64 years) and demonstrated the high prevalence of hrHPV types 16 and 18 [1,5,15–18]. The prevalence of non-16/18 hrHPV types varies between studies and may be attributed to differences between HPV tests utilised, samples types, sample sizes and geographical heterogeneity of the hrHPV type distribution. This is the first study to look specifically at young women having their first cervical smear in order to identify prevalent HPV types upon entry into a cervical screening programme. In this baseline cohort HPV16 was the predominant type in both multiple and single infections. In contrast to other studies, HPV51 was the second most predominant hrHPV type in single HPV infection. Jit et al. [22] used a transmission model to compare efficacy and cost-effectiveness of bivalent and quadrivalent vaccines and concluded that the bivalent vaccine may have an advantage in preventing deaths due to cancer. The quadrivalent vaccine provides additional protection against lrHPV types 6 and 11, which are responsible for genital warts and in this respect confers an advantage over the bivalent vaccine in terms of healthcare costs and QALYs [22]. The modelling of vaccine impact using the baseline cohort of women aged 20–22 years, supports the observation that the bivalent vaccine through its reported cross-protection against other hrHPV types would confer an additional benefit in terms of preventing HPV16/18 unrelated CIN2+ in the population. Overall, the carriage of a single infection vis-à-vis a multiple infection did not differ significantly between women aged 20, 21 or 22 years. However, a higher number of moderate to severe dyskaryotic changes were detected in women who had multiple HPV infection. In studies of women of all ages (20–64 years) participating in routine cervical screening this association is not apparent and is likely a result of the screening programme effectively detecting and treating such cases [1]. In conclusion, before introduction of HPV vaccination, approximately one-quarter of women aged 20–22 years were positive for high risk HPV. The predominant types detected were HPV16, HPV18, HPV31, HPV51 and HPV56. With the high rates of HPV51 and HPV56 detected, monitoring of changes in HPV prevalence following the introduction of the HPV vaccination programme should be a priority. Post-vaccination, we anticipate a substantial absolute risk reduction in high-grade cervical disease associated with targeted and non-targeted hrHPV types. Funding The study was funded by Cancer Research Wales and sponsored by Cardiff University who were responsible for research governance. Competing interests SH and AT were on the ROCHE advisory board for HPV testing and ROCHE sponsored SH to attend the IPV conference in 2010 and
113
sponsored SH and AF to attend the IPV conference in 2011. AT and AF were on the GSK HPV vaccine advisory board and AF on the Sanofi Pasteur MSD HPV vaccine advisory board. AT and AF have received sponsorship from both GSK and Sanofi Pasteur MSD to attend HPV conferences and have two grants from GSK and Sanofi Pasteur MSD for research projects on a HPV vaccine and the psychosocial effects of HPV infection. RN has done HPV unrelated consultancy work for GSK. There have been no other financial relationships with any organisations that might have an interest in the submitted work in the previous three years. Ethical approval Dyfed Powys Local Research Ethics Committee approved the study for All Wales recruitment (08/WMW01/69). Acknowledgements The authors gratefully acknowledge all participating cytology labs within Cervical Screening Wales who contributed to the sample identification and the pseudo-anonymisation process. The authors also gratefully acknowledge the research technicians who contributed to the HPV testing, in particular: Joanne Jones, Angharad Edwards and Vasiliki Kiparoglou. Thanks also to advisors Dr Shantini Paranjothy and Dr Malcolm Adams. References [1] Hibbitts S, Jones J, Powell N, Dallimore N, McRea J, Beer H, et al. Human papillomavirus prevalence in women attending routine cervical screening in South Wales, UK: a cross-sectional study. Br J Cancer 2008;99:1929–33. [2] Bosch FX, Schwarz E, Boukamp P, Fusenig NE, Bartsch D, zur Hausen H. Suppression in vivo of human papillomavirus type 18 E6-E7 gene expression in nontumorigenic HeLa X fibroblast hybrid cells. J Virol 1990;64:4743–54. [3] Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah KV, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189:12–9. [4] Munoz N. Human papillomavirus and cancer: the epidemiological evidence. J Clin Virol 2000;19:1–5. [5] Powell N, Boyde A, Tristram A, Hibbitts S, Fiander A. The potential impact of human papillomavirus vaccination in contemporary cytologically screened populations may be underestimated: an observational retrospective analysis of invasive cervical cancers. Int J Cancer 2009;125:2425–7. [6] Peto J, Gilham C, Fletcher O, Matthews FE. The cervical cancer epidemic that screening has prevented in the UK. Lancet 2004;364:249–56. [7] Ronco G, Giorgi-Rossi P, Carozzi F, Confortini M, Dalla Palma P, Del Mistro A, et al. Efficacy of human papillomavirus testing for the detection of invasive cervical cancers and cervical intraepithelial neoplasia: a randomised controlled trial. Lancet Oncol 2010;11:249–57. [8] Cuzick J, Clavel C, Petry KU, Meijer CJ, Hoyer H, Ratnam S, et al. Overview of the European and North American studies on HPV testing in primary cervical cancer screening. Int J Cancer 2006;119:1095–101. [9] Kitchener HC, Gilham C, Sargent A, Bailey A, Albrow R, Roberts C, et al. A comparison of HPV DNA testing and liquid based cytology over three rounds of primary cervical screening: Extended follow up in the ARTISTIC trial. Eur J Cancer 2011;47:864–71. [10] Stoler MH, Wright Jr TC, Sharma A, Apple R, Gutekunst K, Wright TL. High-risk human papillomavirus testing in women with ASC-US cytology: results from the ATHENA HPV study. Am J Clin Pathol 2011;135:468–75. [11] Lehtinen M, Paavonen J, Wheeler CM, Jaisamrarn U, Garland SM, Castellsague X, et al. Overall efficacy of HPV-16/18 AS04-adjuvanted vaccine against grade 3 or greater cervical intraepithelial neoplasia: 4-year end-of-study analysis of the randomised, double-blind PATRICIA trial. Lancet Oncol 2012;13:89–99. [12] Wheeler CM, Castellsague X, Garland SM, Szarewski A, Paavonen J, Naud P, et al. Cross-protective efficacy of HPV-16/18 AS04-adjuvanted vaccine against cervical infection and precancer caused by non-vaccine oncogenic HPV types: 4-year end-of-study analysis of the randomised, double-blind PATRICIA trial. Lancet Oncol 2012;13:100–10. [13] Brown DR, Kjaer SK, Sigurdsson K, Iversen OE, Hernandez-Avila M, Wheeler CM, et al. The impact of quadrivalent human papillomavirus (HPV; types 6, 11, 16, and 18) L1 virus-like particle vaccine on infection and disease due to oncogenic nonvaccine HPV types in generally HPV-naive women aged 16–26 years. J Infect Dis 2009;199:926–35. [14] Wales PH. Vaccine uptake in children in Wales July to September 2012. In: Wales PH, editor. Coverage of Vaccination Evaluated Rapidly (COVER). Cardiff: The National Publich Health Service for Wales; 2012.
114
S. Hibbitts et al. / Journal of Clinical Virology 59 (2014) 109–114
[15] Hibbitts S, Rieck GC, Hart K, Powell NG, Beukenholdt R, Dallimore N, et al. Human papillomavirus infection: an anonymous prevalence study in South Wales, UK. Br J Cancer 2006;95:226–32. [16] Howell-Jones R, Bailey A, Beddows S, Sargent A, de Silva N, Wilson G, et al. Multi-site study of HPV type-specific prevalence in women with cervical cancer, intraepithelial neoplasia and normal cytology, in England. Br J Cancer 2010;103:209–16. [17] Cuschieri KS, Cubie HA, Whitley MW, Seagar AL, Arends MJ, Moore C, et al. Multiple high risk HPV infections are common in cervical neoplasia and young women in a cervical screening population. J Clin Pathol 2004;57:68–72. [18] Kitchener HC, Almonte M, Wheeler P, Desai M, Gilham C, Bailey A, et al. HPV testing in routine cervical screening: cross sectional data from the ARTISTIC trial. Br J Cancer 2006;95:56–61.
[19] Jacobs MV, Snijders PJ, van den Brule AJ, Helmerhorst TJ, Meijer CJ, Walboomers JM. A general primer GP5+/GP6(+)-mediated PCR-enzyme immunoassay method for rapid detection of 14 high-risk and 6 low-risk human papillomavirus genotypes in cervical scrapings. J Clin Microbiol 1997;35: 791–5. [20] Newcombe RG. Propagating imprecision: combining confidence intervals from independent sources. Communications in Statistics - Theory & Methods 2011;40:3154–80. [21] Miettinen O, Nurminen M. Comparative analysis of two rates. Stat Med 1985;4:213–26. [22] Jit M, Chapman R, Hughes O, Choi YH. Comparing bivalent and quadrivalent human papillomavirus vaccines: economic evaluation based on transmission model. BMJ 2011;343:d5775.
