A comparison of HPV DNA testing and liquid based cytology over three rounds of primary cervical screening: Extended follow up in the ARTISTIC trial

Background The additional sensitivity of HPV testing compared with cytology could permit extended cervical screening intervals. We wished to determine, through a further (third) round of screening in the ARTISTIC trial, the protection provided by a negative baseline HPV screen compared with that of cytology over a 6 year period. Methods Cumulative rates of CIN2 or worse (CIN2+) and CIN3 or worse (CIN3+) were correlated with baseline HPV status and cytology. HPV was detected using the Hybrid Capture 2 (Qiagen) assay for high risk types and genotyped using the Linear Array (Roche) and Papillocheck (Greiner) assays. LBC was performed using ThinPrep (Hologic). Findings Round 3 included 8,873 women of whom 6,337 had been screened in both rounds 1 and 2 and 2,536 had not been screened since round 1. The median duration of follow-up was 72.7 months. The cumulative rate of CIN2+ over three rounds was 3.88% (95%CI 3.59%, 4.17%) overall; 2.39% in round 1, 0.78% in round 2 and 0.74% in round 3. Cumulative rates by baseline status were 20.53% (95%CI 19.04%, 22.08%) for abnormal cytology, 20.12% (95%CI 18.68%, 21.61%) for HPV detection, 1.41% (95%CI 1.19%, 1.65%) for negative cytology and 0.87% (95%CI 0.70%, 1.06%) for a negative HPV test. In HPV negative women aged over 50 the cumulative rate was 0.16% (95%CI 0.07%, 0.34%). Women who were HPV positive/cytology negative at entry had a cumulative CIN2+ rate of 7.73% (95%CI 6.29%, 9.36%) over 6 years, twice the overall rate. Interpretation A negative HPV test was significantly more protective than normal cytology over three rounds. The findings of this extension of ARTISTIC suggest that the screening interval could be extended to 6 years if HPV testing replaced cytology as the primary screening test. Keywords Cervical Screening HPV DNA testing 1 Introduction Screening with cervical cytology at regular intervals has formed the basis for cervical cancer prevention worldwide. In the United Kingdom, the introduction of the national cervical screening programme in 1988 resulted in a marked reduction in cervical cancer deaths. 1 Liquid based cytology (LBC) was implemented between 2003 and 2008 and this has facilitated reflex HPV testing to triage low grade cytology which has undergone limited implementation. Primary cervical screening using HPV DNA testing instead of cytology has been advocated on the basis of greater sensitivity for detection of cervical neoplasia. Randomised trials comparing HPV testing combined with conventional cytology against cytology alone have shown substantially increased detection of high grade cervical intraepithelial neoplasia (CIN2+) in a single screening round. 2–4 In the ARTISTIC trial, which used LBC instead of conventional cytology, a smaller proportion of CIN2+ cases diagnosed in the initial round were in HPV positive women with negative cytology. 5 These trials all showed a reduced incidence of cervical intraepithelial neoplasia grade 3 or worse (CIN3+) in the subsequent screening round, but no difference when the results of two successive screening rounds were summed. Thus both regular cytology and HPV testing identify underlying cervical lesions, but a positive HPV test sometimes leads to earlier diagnosis and also predicts increased risk upto ten years later. 6 The ARTISTIC trial also revealed a very high negative predictive value for HPV testing with fewer than 0.1% of women who were initially HPV negative (HPV–ve) being found to have CIN3+ over two screening rounds. 5 These observations suggest that a negative HPV test confers greater protection over a longer period than negative cytology. We have extended follow up in the ARTISTIC study over a further (third) screening round with two objectives. The first was to compare baseline cytology and HPV testing and typing in terms of detection of CIN2 or worse (CIN2+) over a 6 year period. The second was to use HPV genotyping results on round 1 samples to estimate the potential impact of HPV vaccination on the incidence of abnormal cytology and high grade CIN over three rounds of screening. 2 Methods 2.1 The protocols for rounds 1, 2 and 3 are summarised in Fig. 1 Rounds 1 and 2: The ARTISTIC trial methods and design have been reported for rounds 1 and 2. 5,7 Briefly, women aged 20–64 undergoing routine cervical screening between 2001 and 2003 underwent liquid based cytology and HPV testing, and were randomly allocated in a ratio of 3:1 to have the HPV result revealed and acted upon, or concealed and further management based on cytology alone. Management of women with abnormal cytology was identical on both arms, to the national guidance for the English Cervical Screening Programme, 8 with the exception in round 2 of referral to colposcopy after two rather than three borderline cytology results. Women in the revealed arm with negative cytology who tested HPV positive were invited for repeat HPV testing at 12 months and if still positive could choose between immediate colposcopy or a repeat HPV test at 24 months and colposcopy if still positive. Prior to the study extension to round 3, the protocol was amended to offer exit colposcopy to women in the revealed arm who were cytology negative and persistently HPV positive after the round 2 test, providing information on detection of high grade disease by HPV testing over two screening rounds. The proportion in the revealed arm who reattended at least once within 24 months was 62.6% (1048/1675) in women who were cytology negative and HPV positive in round 1, and 57.3% (339/592) in women who had been cytology negative in round 1 and were cytology negative and HPV positive in round 2. The corresponding proportions in the concealed arm were 3.4% (19/551) in round 1 and 3.2% (6/187) in round 2. Under the original protocol the trial endpoint was high grade disease detected by round 2 cytology. Round 3: The cohort has now been followed beyond the original protocol for a third screening round. During this study extension, the ethics committee required that HPV results were linked to anonymised data, so women on both randomised arms were managed on the basis of cytology according to national guidelines. HPV results in round 3 were ignored until March 2008, when the Manchester Cytology Centre became one of six Sentinel Sites for HPV triage in England. The aim of the Sentinel Sites project was to evaluate the roll-out of HPV triage of low grade cytological abnormalities. Therefore, women with borderline and mild dyskaryosis cytology who tested HPV positive were referred to colposcopy, whilst those who were negative were returned to routine recall. Cytology was performed with ThinPrep (Hologic, Crawley, UK) LBC and HPV testing with Hybrid Capture 2 (Qiagen, Crawley, UK). The HPV result was based on a cut-off of RLU/Co of 1 pg/μl. Three HPV typing assays were used to genotype HC2 positive samples. All HC2 positive LBC samples accrued during rounds 1 and 2 were genotyped using the prototype Roche Line Blot Assay (LBA) as previously described. 5 In order to compare typing assays, two thirds of HC2 positive round 1 archived samples were also tested by the Greiner Bio-one PapilloCheck assay and in round 2 one third of archived HC2 positive samples were tested by the Roche Linear Array (LA). The prototype LBA was replaced with the commercially available LA assay during the extension period of this study, so only 34% of HC2 positive samples in round 3 were tested by the LBA. The remainder were tested using PapilloCheck only (23%), LA only (15%) or both PapilloCheck and LA (28%). In all three rounds, any HPV type detected by any of the assays was included in the analysis. 2.2 Statistical analysis Women were eligible if they were aged 20–64 when they provided a round 1 sample, which was defined as the first cytologically adequate sample after randomisation at entry that gave a satisfactory HPV result. Many women attended for their next routine screen earlier or later than the scheduled three year recall. To minimise exclusions, the round 2 sample was, therefore, defined as the first cytologically adequate sample taken between 26 and 54 months after the date of the round 1 entry sample. The round 3 sample was defined as the first cytologically adequate sample at least 54 months after the round 1 sample, and at least 24 months after the round 2 sample (if there was one). Women were classified histologically at round 1, round 2 and round 3 on the basis of the highest grade of histology within 30 months of the round 1, round 2 or round 3 cytology. Women with CIN2 histology or worse (CIN2+) were excluded from subsequent rounds. The flow of women through three screening rounds is shown in a CONSORT diagram ( Fig. 2 ). Women were followed for cytology until the end of June 2009 and for histology to October 2009. Follow-up was based on cytology and histology reports received by the labs in Manchester and Stockport. Cytology and histology taken outside these areas were not available. HPV prevalence, cytological abnormality and CIN2+ proportions were tabulated by combinations of round, randomisation arm and HPV status at entry or round. HPV status was split further by main HPV type identified: (1) HPV 16 or 18, (2) HPV 31,33,45,52 and 58, (3) other HPV+ve results, including HC2 positive samples where none of the 13 high risk HPV types was identified. These groupings allow reference to the published data for efficacy of the Cervarix HPV 16/18 vaccine which is used in the UK vaccination programme. 9 Cumulative CIN2+ and CIN3 or worse (CIN3+) rates were estimated as P = 1–(1– P 1 )(1– P 2 )(1– P 3 ) where P i denotes the proportion with disease ( D i / N i ) in round i . Confidence intervals were calculated from Greenwood’s formula for the standard error √(1–P) 2 Σ i ( D i /( N i ( N i – D i ))). 10 All analyses were programmed in STATA 11. 11 2.3 Role of the funding source The funding source reviewed and approved the study design. The data collection, analysis, interpretation and write-up were conducted independently by the authors. 3 Results Of the 24,510 eligible women randomised in round 1, 15,790 (64.42%), and 8873 (36.20%) were screened in rounds 2 and 3, respectively, and 6337/8873 (71.42%) of round 3 women were also screened in round 2. The median time to round 3 was 72.7 months after round 1. The proportions of women with CIN2, CIN3 and CIN2+ by round and randomisation arm are shown in Table 1 . As previously reported, 7 there was a marked fall in the detection rate of high-grade disease between rounds 1 and 2, but no significant change between rounds 2 and 3. The cumulative detection of CIN2+ over three rounds was 3.9% (95%CI 3.6%, 4.3%) in the revealed arm and 3.7% (95%CI 3.2%, 4.3%) in the concealed arm (OR = 0.94, 95%CI 0.78, 1.13, p > 0.1). The cumulative detection rate of CIN2+ and CIN3+ amongst the entire study cohort over three rounds (median follow-up 6 years) according to baseline cytology, HPV status and age at entry is shown in Tables 2 and 3 , respectively. The cumulative CIN2+ rate for women who were both cytology–ve and HPV–ve at round 1 was 0.67% (95%CI 0.51%, 0.87%), compared with 7.73% (95%CI 6.29%, 9.36%) for those who were cytology–ve/HPV+ve at round 1. Amongst women who were cytology +ve and HPV+ve at baseline, the cumulative rate was 37.44% (95%CI 34.91%, 40.04%) compared with 3.24% (95%CI 2.32%, 4.38%) for those who were cytology+ve/HPV–ve. Whether baseline cytology was negative or positive, baseline HPV detection thus increased the CIN2+ risk more than 10-fold. The risk of CIN2+ detection over 6 years was significantly less if the initial screen was HPV negative (0.87%; 95%CI 0.70%, 1.06%) rather than cytology negative (1.41%; 95%CI 1.19%, 1.65%), as shown in Table 2 . The same effect was seen for CIN3+ ( Table 3 ), the risk following a negative HPV test (0.28%; 95%CI 0.18%, 0.40%) being less than half that following negative cytology (0.63%; 95%CI 0.48%, 0.80%). There was no significant difference ( p > 0.1) in cumulative CIN2+ rate over the three rounds between women who were HPV+ve or cytology+ve at entry. The influence of age and HPV status at round 1 on 6 year cumulative detection of CIN2+ is also shown in Table 2 . Among women who were HPV+ve at baseline, the cumulative detection of CIN2+ was 23.53% at age 20–24, 24.30% at 25–34, 15.92% at 35–49 and 6.40% above age 50. Amongst women who were HPV–ve at baseline the cumulative incidence of CIN2+ over 6 years was much lower and fell sharply with age, from 3.04% at age 20–24 to 1.44% in the 25–34, 0.48% at 35–49 and only 0.16% in women aged over 50 years. The 6 year cumulative incidence of CIN3+ in this group of HPV–ve women aged over 50 years was only 0.05% (95%CI 0.01%, 0.19%). The influence of baseline HPV genotype at round 1 on cumulative CIN2+ over three rounds is shown in Table 4 . The dominant effect of type 16 or 16/18 combined is clearly evident, particularly for CIN3+ with a 3-fold greater rate for type 16 (30.35%) compared with types 31, 33, 45, 52 and 58 combined (10.68%). There were 524 women with a new HR-HPV infection in round 2 (i.e. a type which was not present at round 1) and 241 women with a type-specific persistent infection in round 2 (i.e. same type present in both rounds 1 and 2). The corresponding numbers in round 3 with new and persistent HR-HPV infections were 409 and 96, respectively. Among these women, the CIN2+ rates were much lower in women with a new infection (6.9% in round 2 and 8.3% in round 3) than in those with a persistent infection (22.0% in round 2 and 16.7% in round 3). Since 2008, there has been a UK national programme of vaccinating teenage girls with the bivalent Cervarix (GSK) vaccine. The potential impact of HPV vaccination of 12/13 year old girls (almost all of whom are expected to be HPV–ve) on the subsequent incidence of cervical cytology abnormalities can be calculated from Table 5 , which shows cytology and HPV typing results at each round. Reported data on Cervarix vaccine efficacy in women who were HPV DNA negative, indicated 98% efficacy against CIN2+ associated with types 16/18, and cross protection conferring 70% efficacy against CIN2+ associated with types 31/33/45/52/58. 9 These results, together with the ARTISTIC data correlating HPV types with grades of cytological abnormality shown in Table 5 , suggest that Cervarix would prevent about 68% of cytology graded moderate or worse (HSIL) and 22% of borderline/mild (ASCUS/LSIL) cytology. Of the 757 women with a known HPV result and confirmed CIN2+ over three rounds, 401 (53%) were associated with HPV16 or HPV18 and 197 (26%) with types 31, 33, 45, 52, or 58, suggesting that 70% of CIN2+ cases in ARTISTIC could have been prevented by vaccination. The proportion of CIN2+ cases attributed to these HR-HPV types decreases with age; we estimate that approximately 78% of CIN2+ cases in women aged 20–34 and 65% of CIN2+ cases in women aged 35–49 would be prevented by vaccination. 4 Discussion This extended follow-up of the ARTISTIC cohort to a median of 6 years has focussed on three issues: cumulative CIN2+ rates based on HPV and cytology results at the initial screening round, the influence of different HPV genotypes, and the predicted impact of the national HPV vaccination programme on screening abnormalities. The strengths of the study are (a) long follow-up, (b) the use of LBC and HC2 throughout, both of which are widely used and quality assured products and (c) the conduct of the study within the routine national screening programme. As previously discussed, 7 the reduction from round 1 to round 2 in both abnormal cytology and high-grade histology rates reflected enhanced detection of prevalent disease following the introduction of LBC and retraining of cytology laboratory staff. The change of protocol in round 3 to use HPV triage for low grade cytological abnormality will have reduced the likelihood of CIN remaining undetected within the timeframe of the study. The rate of follow-up from round 1 to round 2 was 64%, similar to the overall rate of adherence of around 65% in consecutive screening rounds in Greater Manchester in the routine programme. Many women who were recruited aged 58 and over at enrolment will have exited screening having reached aged 64 before round 3. The most significant finding of the study is the longer duration of protection provided by initial HPV testing when compared with liquid based cytology, with the cumulative CIN2+ rate over three rounds (about 6 years) following a HPV–ve baseline test (0.87%) being similar to that after two rounds (about 3 years) following a cytology –ve baseline test (0.73%). This was also seen in several European cohorts, 12 and suggests that if HPV testing were the initial screen, screening intervals could be extended from 3 to 6 years without reducing the current level of sensitivity provided by regular cervical cytology. In women aged 50 and over, the interval could perhaps be increased from the current 5 year interval to 10 years given the very small cumulative risk of CIN2+ (0.16%) and CIN3+ (0.05%) for HPV–ve women aged 50–64 years. Women who were HPV+ve/cytology–ve had a 7.73% rate of CIN2+ over 6 years, twice the overall rate for all screened women (3.9%). If primary screening began with HPV testing followed by cytology triage for HPV+ve women, those who were HPV+ve/cytology–ve would require either further triage to identify a higher risk group for referral to colposcopy, or repeat HPV testing after a shorter interval than HPV–ve women. One triage strategy would be to select women with HPV types 16 or 18, which confer a significantly higher risk than other HPV types. Newer HPV tests are offering 16/18 restricted typing. 13 Triage tests using various biomarkers are also being evaluated. In one study, 13 which compared the ability of different tests to identify high grade CIN (albeit in a referral population), CINtec p16 INK4a and HPV RNA for types 16,18,31,33 and 45 (HPV-Proofer, NorChip) achieved far higher specificity than high risk HPV tests and offer potential in this setting. Restricting HPV based primary screening to women aged 30 years and older has been recommended, 14 but such a two tier national programme would create complexity and the prevalence of HPV in young women can be expected to fall following the introduction of teenage HPV vaccination. The analysis of cytological abnormality rates by HPV genotype shows that vaccination would considerably reduce high grade cytology. This has implications for primary cervical screening, which could not maintain current levels of positive predictive value with greatly reduced rates of high grade cytology. One recent modelling study of British women aged 20–29 estimated the HPV vaccination against HPV 16/18 would achieve a 63% reduction in cervical cancer, 51% reduction in CIN3 and a 27% reduction in cytological abnormalities. Even this young age group, however, the impact would not be seen until 2025. 15 Another study estimated a 76% reduction in lifetime risk of cervical cancer for 12 year old girls vaccinated in the UK against HPV 16/18. 16 In conclusion, these ARTISTIC follow-up results, together with other published data showing the high negative predictive value of HPV testing, suggest that changing primary screening from cytology to HPV testing would offer advantages in terms of increased screening intervals and increased sensitivity. Conflict of interest statement HCK is Chair of the Advisory Committee in Cervical Screening (ACCS). The views expressed in this report are those of the author(s) and in no way represent views of the ACCS or the Department of Health. All authors received funding from the NIHR HTA to conduct the study. Contributors HCK, JP, SM, MD and CR all contributed to the study design. HCK, JP and CG contributed to drafting. JP, CG and CR contributed to the statistical analysis. AS, AB and AT did the virological analyses. MD and JM supervised and co-ordinated the cytological activity. HCK acted as chief investigator. RA acted as trial co-ordinator. All authors contributed to revision of the manuscript. Funding The ARTISTIC trial was funded by the NIHR Health Technology Assessment Programme. Trial registration ISRCTN: 25417821. Acknowledgements We are grateful to Yvonne Hughes and Dr Kate Morgan and the staff at both the Manchester Cytology Centre and Stepping Hill cytology laboratory for their commitment to the trial. This project was funded by the NIHR Health Technology Assessment Programme (Project number 98/04/99) and will be published in full in the Health Technology Assessment journal, 2011. See the HTA programme website for further project information. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Department of Health This work was supported by the NIHR Manchester Biomedical Research Centre . References 1 J. Peto C. Gilham O. Fletcher F. Matthews The cervical cancer screening epidemic that screening has prevented in the UK Lancet 364 2004 249 256 2 P. Naucler W. Ryd S. Tornberg Human papillomavirus and Papanicolaou tests to screen for cervical cancer N Engl J Med 357 16 2007 1589 1597 3 N.W. Bulkmans J. Berkhof L. Rozendaal Human papillomavirus DNA testing for the detection of cervical intraepithelial neoplasia grade 3 and cancer: 5-year follow-up of a randomised controlled implementation trial Lancet 370 9601 2007 1764 1772 4 G. Ronco P. Giorgi-Rossi F. Carozzi Efficacy of human papillomavirus testing for the detection of invasive cervical cancers and cervical intraepithelial neoplasia: a randomised controlled trial Lancet Oncol 11 3 2010 249 257 5 Kitchener HC, Almonte M, Dowie R et al., ARTISTIC: a randomised trial of HPV testing in primary cervical screening. Health Technol Assess 2009; 13(51). 6 S. Kjaer E. Hogdall K. Frederiksen The absolute risk of cervical abnormalities in high-risk human papillomavirus-positive, cytologically normal women over a 10-year period Cancer Res 66 21 2006 10630 10636 7 H.C. Kitchener M. Almonte C. Thomson HPV testing in combination with liquid-based cytology in primary cervical screening (ARTISTIC): a randomised controlled trial Lancet Oncol 10 7 2009 672 682 8 NHS Cervical Screening Programme, Colposcopy and Programme Management – Guidelines for the NHS Cervical Screening Programme, NHS Cancer Screening Programmes: Sheffield, 2004. 9 J. Paavonen P. Naud J. Salmeron Efficacy of human papillomavirus (HPV)-16/18 AS04-adjuvanted vaccine against cervical infection and precancer caused by oncogenic HPV types (PATRICIA): final analysis of a double-blind, randomised study in young women Lancet 374 9686 2009 301 314 10 Greenwood M, The natural duration of cancer. Reports on Public Health and Medical Subjects. Her Majesty’s Stationery Office: London; 1926. p. 1–26. 11 STATA 11, StataCorp: College Station, TX; 2009. 12 J. Dillner M. Rebolj P. Birembaut Long term predictive values of cytology, human papillomavirus testing in cervical cancer screening: joint European cohort study BMJ 337 2008 a1754 13 A. Szarewski L. Ambroisine L. Cadman Comparison of predictors for high-grade cervical intraepithelial neoplasia in women with abnormal smears Cancer Epidemiol Biomarkers Prev 17 11 2008 3033 3042 14 Franceschi S, Denny L, Irwin KL et al., Eurogin 2010 roadmap on cervical cancer prevention . Int J Cancer 2011; [Epub ahead of print]. 15 J. Cuzick A. Castanon P. Sasieni Predicted impact of vaccination against human papillomavirus 16/18 on cancer incidence and cervical abnormalities in women aged 20–29 in the UK Br J Cancer 102 5 2010 933 939 16 M. Kohli N. Ferko A. Martin Estimating the long-term impact of a prophylactic human papillomavirus 16/18 vaccine on the burden of cervical cancer in the UK Br J Cancer 96 1 2007 143 150