IJC International Journal of Cancer

Transcription

1 IJC International Journal of Cancer Early detection of CIN3 and cervical cancer during long-term follow-up using HPV/Pap smear co-testing and risk-adapted follow-up in a locally organised screening programme Alexander Luyten 1, Nina Buttmann-Schweiger 2, Katrin Luyten 1, Claudia Mauritz 3, Axel Reinecke-L uthge 4, Martina Pietralla 5, Chris J.L.M. Meijer 6 and Karl Ulrich Petry 1 1 Department of Gynaecology and Obstetrics, Klinikum Wolfsburg, Wolfsburg, Germany 2 Department of Epidemiology and Health Monitoring, German Centre for Cancer Registry Data, Robert Koch-Institut, Berlin, Germany 3 Health Care Management Unit, Deutsche BKK, Wolfsburg, Germany 4 Institute of Pathology, Klinikum Wolfsburg, Wolfsburg, Germany 5 Gemeinschaftspraxis, Wolfsburg, Germany 6 Department of Pathology, Vrije Universiteit Medical Center Amsterdam, Amsterdam, The Netherlands We evaluated compliance with human papillomavirus (HPV) testing and risk-adapted patient pathways and monitored changes in high-grade cervical disease during long-term follow-up. Women aged >30 years attending routine screening for cervical cancer were managed according to results from first-round screening tests (cytology and high-risk HPV; Hybrid Capture 2). Between February 2006 and January 2011, 19,795 of 19,947 women agreed to participate, of whom 4,067 proceeded to a second screening round 5 years after recruitment. Predefined endpoints were compliance, grade 3 cervical intraepithelial neoplasia or cancer (CIN31), new HPV infection, HPV persistence and abnormal smears in round 2. A total of 765 of 19,795 women (3.9%) in round 1 and 41 of 4,067 (1.0%) in round 2 were referred for colposcopy. Compliance rates with colposcopy were 93.1 and 92.7%, respectively, while histological assessment was performed in 680 of 712 (95.5%) and 36 of 38 (94.7%), respectively. CIN31 rates were 172 of 19,795 (0.87%; 95% confidence intervals: ) in round 1 and 2 of 4,064 (0.05%; 95% confidence intervals: ) in round 2; the difference was statistically significant (Fisher s exact test, p < 0.001). After 5 years, the incidence of new HPV infection was 124 of 3,906 (3.2%) and HPV persistence was observed in 22 of 161 (13.7%). Locally organised HPV/cytology co-testing is feasible and acceptable to women. Risk-adapted management rapidly detected a high rate of prevalent CIN31, while the subsequent long-term risk of new high-grade cervical disease was surprisingly low. It remains unclear if this phenomenon is explained by CIN3 mostly occurring early in life or by modifying the natural course of HPV infection with colposcopy and histological assessment. There is compelling evidence from randomised controlled trials and meta-analyses that primary screening for human papillomavirus (HPV) is more sensitive than cytology-based prevention to detect cervical intraepithelial neoplasia grade 3 (CIN3) and prevent cervical cancer. 1 9 Despite limitations, including variance in patient pathways and different Key words: CIN3, cervical cancer screening, human papillomavirus, colposcopy Author contributions: A.L., C.M., M.P., C.J.L.M.M. and K.U.P. contributed to the study design, A.L., N.B., C.J.L.M.M. and K.U.P. contributed to drafting, A.L., K.L., C.M., M.P. and K.U.P. collected, arranged and managed the data, N.B. and K.L. contributed to the statistical analysis. All authors had access to all the data and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors contributed to the revision of the manuscript. The manuscript was edited by Tim Kelly, funded by the authors. K.U.P. is guarantor Conflict of interest: All authors have completed the Unified Competing Interest form at (available on request from the corresponding author) and declare that: Apart from Deutsche BKK s financing the authors did not receive any support for the submitted work. K.U.P. received speaker s honorarium from Roche Diagn., GSK, Becton Dickinson, Qiagen and occasionally was a consultant for Roche Diagn., A.L. received speaker s honorarium from Qiagen, Sanofi-Pasteur-MSD and Becton Dickinson, M.P. received speaker s honorarium from Qiagen and C.J.L.M.M. received speaker s honorarium from Roche Diagn., GSK, Merck, Qiagen and was occasionally consultant for GSK and Qiagen, advisory board member of Qiagen and is minority shareholder of Self-Screen, a spin-off company of Vrije Universiteit Amsterdam. The authors report no other relationships or activities that could appear to have influenced the submitted work Grant sponsor: Deutsche BKK, Willy Brandt Platz, Wolfsburg, Germany DOI: /ijc History: Received 16 Sep 2013; Accepted 8 Jan 2014; Online 12 Feb 2014 Correspondence to: Prof. Dr. Karl Ulrich Petry, Klinikum Wolfsburg, Frauenklinik Schwerpunkt Gyn akologische Onkologie, Sauerbruchstrasse 7, Wolfsburg, Germany, Tel.: , Fax: , gyn@klinikum.wolfsburg.de

2 Luyten et al What s new? This longitudinal cohort study explored the feasibility of a locally organised screening using HPV and cytology co-testing and defined patient pathways for colposcopy referral with the aim to improve detection of CIN3 and prevention of cervical cancer. HPV-based screening for cervical cancer proved feasible in countries without a nationwide screening programme and at local level. HPV/cytology co-testing and risk-adapted colposcopy detected almost all cases of CIN31 at first colposcopy and the future risk of CIN3 and cancer was low. Thus, a one-time, minimally invasive intervention in women with HPV persistency could virtually eliminate the otherwise high risk of cervical cancer. thresholds for colposcopy referral, randomised controlled trials have encouraged nationally organised HPV screening in some countries for women older than 30 years. Setting up new national screening programmes may be challenging, especially in privately organised health systems. 10,11 At the time this study was designed, cervical cancer screening in Germany was based on annual Papanicolaou (Pap) smear testing conducted by gynaecologists in private practice. Although the emergence of efficient tests for highrisk HPV offered the prospect of improving cervical cancer prevention, 5,6,12 15 the value of HPV testing outside national programmes was unknown. In February 2006, in the Wolfsburg region of Germany, the health insurer Deutsche BKK changed its primary cervical cancer screening programme to Pap cytology combined with HPV testing every 5 years for women aged 30 years and older. At this time, it was considered that there was insufficient evidence to move to HPV screening without cytology. Tests for high-risk HPV can identify almost all women at high risk of developing cervical cancer, but the risk of CIN31 is only in the range 3 7% in women with normal cytology and a positive high-risk HPV test Therefore, clinicians need an effective strategy to identify the minority of women in this cohort with CIN Overall, the success of HPV screening, balancing the need to prevent all cases of cervical cancer while avoiding overtreatment, requires an effective colposcopy protocol. Here, we report cross-sectional and longitudinal data from the first German regional pilot project of primary HPV screening, which was started in February 2006 and, in February 2011, extended until 31 January We wanted to demonstrate that the improved prevention of cervical cancer observed in randomised controlled trials can be achieved by a locally organised HPV screening programme with a rigorous, individualised, risk-based follow-up protocol. Studies describing the rationale for our screening pathway and colposcopy protocol have been published previously. 20,21 Material and Methods The WOLfsburg pilot project for better prevention of cervical cancer with Primary HPV SCREENing (WOLPHSCREEN) is based on a contract between the health insurance company Deutsche BKK, all gynaecologists in private practice in Wolfsburg and surrounding areas and Klinikum Wolfsburg. Local ethics committee approval was not required because Deutsche BKK uses the project as their routine screening concept. Study design In Germany, where almost all people are covered by health insurance, it is usual for women to voluntarily visit a gynaecologist in private practice to participate in national screening every year. Between February 2006 and January 2011, 19,947 female members of Deutsche BKK aged 30 years or older (no upper age limit) voluntarily attending routine cervical cancer screening at one of the gynaecologist partners in private practice were invited to participate. Women not attending screening for 2 years were sent reminder letters. This analysis includes screening results up to March 2012 and subsequent colposcopy data up to November Women were allowed to choose whether to participate in the new pilot project or to have annual Pap screening, the existing standard in Germany. A history of hysterectomy was an exclusion criterion. After counselling, participants had to give written consent and to complete a questionnaire about medical history, parity, contraception and smoking. The first screening series (Series 1) comprised data from screening procedures in women recruited between February 2006 and January 2011, whereas Series 2 comprised data from procedures after February 2011 in women originally recruited in 2006 who had completed their 5-year screening interval. In both series, women were assigned and managed according to results from cytology [abnormal Pap smears classified as atypical squamous cells of undetermined significance (ASC-US)/borderline or more] and high-risk HPV tests [Hybrid Capture 2 (HC2)], as shown in Figure 1. To avoid overtreatment, only women with abnormal smears and positive HPV tests or HPV persistency for 1 year were referred for colposcopy as described in Figure 1. In Series 2 we incorporated a novel dual biomarker-based approach (p16/ki-67 dual-stained cytology) to improve the triage of women at risk. 20,21 To minimise the rate of overlooked CIN31, histological assessment was obligatory for women referred for colposcopy. Women in Series 1 who completed 5-year follow-up, remained in the Wolfsburg region and continued membership of Deutsche BKK had their second screening round in Series 2. In the 5-year period between the first and second screening rounds, women were followed up according to the

3 1410 Early detection of CIN3 and cervical cancer Figure 1. Patient pathways according to predefined risk groups for Series 1 (a) and Series 2 (b). In Series 1, women in Group IV (abnormal Pap/HC2-positive) were referred immediately for colposcopy. In Group II (abnormal Pap/HC2-negative), only cases classified as high-grade squamous intraepithelial/atypical squamous cells-cannot exclude HSIL (HSIL/ASC-H) were referred immediately for colposcopy. Among the remaining women with abnormal Pap tests, cases with persistent atypical squamous cells of undetermined significance/low-grade squamous intraepithelial (ASC-US/LSIL) and/or positive HC2 results were subsequently referred for colposcopy. Women in Group III (normal Pap/HC2-positive) were referred for colposcopy if they had an abnormal Pap test at 6 months or remained HC2-positive at 12 months. In Series 2, management of women in Groups I and IV was the same as Series 1. For women in Group II a repeat Pap smear test after 6 months was not performed because of the observed low risk of CIN21 during the first screening round (see Results). For Group III the repeat Pap smear after 6 months was replaced with immediate p16/ki-67 dual-stained cytology as triage test. 21 Patients with positive dual-stained cytology or with positive HC2 tests after 12 months were referred for colposcopy. protocol shown in Figure 1. Women in Group I had their next screening round 5 years later, although data on cervical histology were collected from hospital and insurance databases. Women in Groups II and III were followed more intensively; in Series 1 the follow-up protocol included repeat cytology after 6 months and HC2 tests after 1 year, and those with positive test results were referred for colposcopy. In Series 2 the protocol was changed to repeat cytology and HC2 after 1 year in Group II and immediate p16/ki-67 dual stain and repeat HC2 after 1 year in Group III. All women in Group IV were immediately referred for colposcopy. After colposcopy, women with CIN1 or less were followed annually with HC2 and cytology tests. When tests normalised, women reverted to the routine screening protocol, but those with persistent abnormalities returned each year for colposcopy. The main focus of this analysis was to compare the first screening round (Series 1) with the second screening round (Series 2). The predefined study endpoints were the rate of compliance with the screening pathway, the overall rates of CIN3 or cancer (CIN31) in the first and second screening rounds, the rate of CIN31 following a colposcopy classified as CIN1 or less and the rates of new HPV infection, HPV persistence and abnormal smears at the second screening round. We chose CIN31 rather than CIN2 or more (CIN21) as a study endpoint because CIN2 is not necessarily a precursor of cervical cancer. In contrast to CIN3, CIN2

4 Luyten et al cases have a moderately high rate of spontaneous regression and poor reproducibility on histology review. Furthermore, neither the insurance database nor the cancer registry contained data on CIN2, so we would not be able to correct our data for cases diagnosed outside the project reliably. Participants who changed their health insurance provider during Series 1 were excluded from the second HPV screening, but were still followed according to their risk group. Endpoints for follow-up were a diagnosis of cervical cancer, hysterectomy for other reasons, leaving the region or declining follow-up procedures. Procedures and data management All participants had a routine pelvic examination with Pap smear followed by a separate cervical sample taken with a Medscan brush for high-risk HPV testing using HC2 (QIA- GEN, Hilden, Germany). In Series 2, a third sample was taken for liquid cytology with p16/ki-67 dual staining using a Cervex brush and ThinPrep (HOLOGIC) transport medium (CINTEC plus, MTM/Roche Diagnostics). All colposcopists were trained according to European Federation for Colposcopy standards ( In brief, colposcopy included the classification of the transformation zone. 22 In cases of Type 1 or 2 transformation zone with visible squamous-columnar junction, colposcopy was expected to be satisfactory and biopsies were taken from all visible lesions even if these were classified only as minor changes. In cases of Type 3 transformation zone, endocervical curettage was mandatory and biopsies were taken from all visible lesions. Screening-specific data were stored in a central database at Klinikum Wolfsburg. There was a continuous central follow-up of abnormal findings. When patients were not managed according to protocol the responsible private practice was informed as a first step. If the patient had changed gynaecologic practice or did not adhere to the patient pathway for other reasons, the coordinator sent a letter detailing the abnormal screening results and an explanation about the proposed clinical management. We used a separate, independent database managed by Deutsche BKK to estimate the participation rate of our target population in the new screening programme and to identify CIN31 cases diagnosed outside the screening protocol. Statistical analysis Exact binomial 95% confidence intervals were calculated based on proportions of women from Series 1 (first screening series, recruited and followed up between February 2006 and January 2011) and women of the second screening series (Series 2) who continued follow-up after February 2011, including women in whom colposcopy had been performed. Because not all were yet eligible to enter Series 2 and no participant had finished the full 5-year course of Series 2, at the time of reporting, we used a subanalysis to compare the proportion of new CIN3 diagnosed in the first 18 months after recruitment in Series 1 against cases in Series 2 among 4,067 women participating in both rounds. We acknowledge the limitation of missing information on long-term follow-up data. We used the Fisher s exact test for statistical comparisons of the proportions of CIN31 lesions between women of Series 1 and women of Series 2. The p-value was assessed against the nominal 0.05 level of significance. The analyses were conducted using the statistical software R version (R Foundation for Statistical Computing, Vienna, Austria). Potential confounding factors were anticipated in the study design and scrutinised in clinical follow-up documentation. An ageing study population did not allow for correction of age as a confounding factor, and has to be considered in the transferability of results. Selection bias was minimised by efforts to achieve high participation and compliance rates. Further potential sources of bias due to loss to follow-up were addressed by reconciliation of external information sources providing information on women leaving the study population (Deutsche BKK, federal cancer registry), allowing for gross sensitivity estimation. Results Study population, acceptance and participation In total, 19,795 women (mean age years, median age years) participated in Series 1 (Fig. 2), of whom 1,096 (5.5%) were older than 70 years. The observed acceptance was 99.1% in Series 1 and almost all women attending routine screening chose the new screening concept instead of annual Pap smears. The rate of women declining participation in the pilot project ranged from 0.1 to 4.8% between different practices (mean 0.9%). Of the 19,795 women entering Series 1 before January 31, 2011, 4,067 proceeded to Series 2, including 3,925 who had not had a colposcopy and 142 following at least one colposcopy procedure (Fig. 2). A total of 8,308 women were still in follow-up (less than 5 years after entering Series 1), whereas 1,111 had reached the end of follow-up because of cervical cancer, death, hysterectomy or they had left the region. A further 6,256 women were not included in Series 2 because they left Deutsche BKK, but remained under clinical follow-up during their 5-year series. Risk groups and compliance with patient pathways The distribution of women according to risk groups is shown in Table 1. Compliance with colposcopy referrals is shown in Figure 2. In Series 1, 765 of 19,795 women were referred for colposcopy (3.9%). In risk groups II, III and IV, the rate of compliance with colposcopy was 93.1% (712 of 765 referrals). All women in Group IV (abnormal Pap smears and HC2- positive) were referred for colposcopy according to protocol (compliance 95.5%). In Group III (normal cytology and HC2-positive) 536 women were referred for colposcopy (compliance 94.4%). In Group II (abnormal cytology and HC2-negative), 28 women were referred for colposcopy (compliance 50%). Colposcopy was performed off protocol in 19 women in risk group I (normal Pap smears and HC2-negative). In Series 2, the compliance rate was 92.7% (38 of 41).

5 1412 Early detection of CIN3 and cervical cancer Figure 2. Patient flowchart for Series 1 and Series 2. In total, 4,067 women from Series 1 proceeded to Series 2 for the second screening round, comprising 3,865 from Group I, 41 from Group II, 137 from Group III and 24 from Group IV. At the time of reporting, 8,041 women from Series 1 had not yet proceeded to Series 2, but remain on study and will have a second screening round 5 years after their initial screening. Detection of high-grade disease (CIN31) Histological assessment by punch biopsies with (n 5 110) or without curettage (n 5 423) or curettage without biopsies (n 5 98) or immediate excisional treatment (n 5 49) was performed in 680 of 712 (95.5%) women in Series 1 and 36 of 38 (94.7%) in Series 2. Table 1 shows the prevalence of CIN31. In Series 1, 172 cases (0.87%) were diagnosed according to the colposcopy protocol and three more CIN3 were Table 1. Comparison of the absolute risk of CIN31 for Series 1 and Series 2, both overall and according to risk groups Series 1 Series 2 CIN31 CIN31 Statistical significance Total N (%; 95% CI) N % 95% CI Total N (%; 95% CI) N % 95% CI (p-value) Total study 19,795 (100%; ) ,067 (100%; ) <0.001 population Group I 18,318 (92.5%; ) ,882 (95.5%; ) NA Group II 245 (1.2%; ) (1.0%; ) NA Group III 1,031 (5.2%; ) (3.4%; ) <0.001 Group IV 201 (1.0%; ) (0.2%; ) NS Group I 5 Pap normal and HC2-negative; Group II 5 Pap abnormal and HC2-negative; Group III 5 Pap normal and HC2-positive; Group IV 5 Pap abnormal and HC2-positive. Pap normal is a cytology result less than Pap2w (after Papanicolaou), LSIL. p-value against the nominal 0.05 level of significance. Abbreviations: HC2: Hybrid Capture 2; 95% CI: 95% confidence interval.

6 Luyten et al Table 2. Details of patients with a diagnosis of cervical cancer No. Age at diagnosis Risk group Pap smear at recruitment Time since last normal Pap smear (years) Histology FIGO stage Normal <1 Adenosquamous Ib Normal <1 Adenocarcinoma Ib Normal <1 SCC Ia Normal <1 SCC Ia Normal <1 SCC Ia Normal <1 SCC IVb Normal <1 Adenocarcinoma Ia Normal <1 SCC Ib LSIL 1 SCC Ib LSIL 1 SCC Ib ASC-H 6 SCC IIb ASC-H 5 SCC Ia HSIL >10 SCC Ib HSIL >10 SCC Ia HSIL 1 SCC Ia Cancer 6 SCC Ia Cancer 9 SCC Ib Cancer 1 Adenocarcinoma Ib Cancer 7 SCC IVa Cancer 27 SCC Ia1 In cases with stage Ia1 the depth of invasion was between 1 and 2.5 mm. Abbreviations: SCC: squamous cell carcinoma; LSIL: low-grade squamous intraepithelial lesion; HSIL: high-grade squamous intraepithelial lesion; ASC-H: atypical squamous cells cannot rule out a high-grade lesion. detected outside the project according to data from Deutsche BKK. Among CIN31 cases there were 13 cases of adenocarcinoma in situ (ACIS) and 19 invasive cervical cancers. One further case of CIN31, a Stage Ib adenosquamous cervical cancer, occurred in Group I; the patient was referred immediately for colposcopy because of a visible lesion with bleeding at clinical examination. No case of CIN31 was detected in Pap smear-positive/hc2-negative women (Group II). The highest prevalence of CIN31 in Series 1 was observed in Group IV: 76 of 192 (39.6%) women who underwent colposcopy (76 of 201 referred for colposcopy; 37.8%). The highest absolute number of CIN31 (n 5 95) in Series 1 was in women with normal cytology and positive HC2 results (Group III, comprising 1,031 women). In Series 1, there were 45 CIN2 cases in Group IV (121 CIN21 overall), 91 in Group III (186 CIN21) and two in Group II. Among 20 cases of invasive cervical cancer (Table 2), only five had a clinical or cytological diagnosis of cancer at referral, whereas eight had normal Pap smears at recruitment, six of whom were diagnosed only because of HPV persistence. A significantly smaller risk for CIN31 was observed in Series 2 compared with Series 1 [2 of 4,067 (0.05%) versus 172 of 19,795 (0.87%), respectively; p < 0.001]. This corresponds to a relative risk reduction of 94.4%. The 4,067 women included in this analysis were among the first screening participants recruited in 2006 who returned on schedule in 2011 for Series 2 and could have an overall lower risk for CIN31 than less punctual women. To exclude such a bias, we compared the risk for CIN31 observed in Series 1 up to November 2007 with the corresponding risk in Series 2 up to November 2012 in this identical cohort; we also observed a significantly smaller risk in Series 2 [2 of 4,067 (0.05%) versus 23 of 4,067 (0.57%); p < 0.05], although this approach underestimates the magnitude of risk reduction because patients with cancer had reached their endpoint in Series 1 and, therefore, were excluded from this cohort. The two cases of CIN3 in Series 2 were observed in women with newly acquired HPV infections. Risk of new HPV infection, HPV persistence and abnormal smears in Series 2 Among 41 HC2-negative women who had abnormal Pap smears in Series 1, just one subject still had an abnormal Pap smear in Series 2 (2.4%). Among 3,906 women who tested HC2-negative in Series 1, 124 tested HC2-positive in Series 2 (Table 3). The incidence of high-risk HPV is therefore 3.2%. In Series 2, HPV persistence over 5 years was observed in 22 of 161 HC2-positive women (13.7%) from Series 1. The true

7 1414 Early detection of CIN3 and cervical cancer Table 3. Risk group distribution among 4,067 women from Series 1 (February 2006 to January 2011) who proceeded to Series 2 (February 2011 to March 2012) Risk group after second screening round in Series 2 I II III IV Total Risk group after first screening round in Series 1 I 3, ,865 II III IV Total 3, ,067 Group I 5 Pap normal and HC2-negative; Group II 5 Pap abnormal and HC2-negative; Group III 5 Pap normal and HC2-positive; Group IV 5 Pap abnormal and HC2-positive. Pap normal is a cytology result less than Pap2w (after Papanicolaou), LSIL. Abbreviation: HC2: Hybrid Capture 2. natural persistence rate may be higher because overall 294 women (1.49% of the total screening population) from Groups III and IV underwent oncological/excisional treatment and a further 386 had colposcopy assessment with biopsies and/or endocervical curettage during Series 1. Among 138 women with HPV clearance in Series 2, 45 underwent excisional treatment during Series 1. Figure 3 shows the absolute numbers of women diagnosed with CIN31 at first colposcopy or subsequent follow-up examinations between 2006 and A total of 159 CIN31 cases were diagnosed at first colposcopy and/or by subsequent excisional procedures. When the diagnosis at first colposcopy was CIN1 or less and women were followed without invasive treatment, the rate of subsequent CIN3 was relatively low. Just 13 of 418 women (3.1%) in this cohort were diagnosed with CIN3 during follow-up and only five of these 13 cases were classified as newly developed lesions, while the remaining eight cases were most likely overlooked at first colposcopy. Discussion We conducted this longitudinal cohort study to explore the feasibility of a locally organised screening using HPV and cytology co-testing and defined patient pathways for colposcopy referral to achieve a more sensitive detection of CIN3 and a better prevention of cervical cancer than achieved with annual Pap screening. In agreement with previous studies, 1 9,19 WOLPHSCREEN showed that HPV/cytology co-testing reliably identifies the population at risk and following our protocol we were able to detect 92.4% of CIN31 lesions (159 of 172) in the first screening round at first colposcopy with almost obligatory histological assessment. All referred women were followed up until they tested negative for HPV or developed CIN21. In the HC2-positive risk groups III and IV, we found a relatively high rate of CIN31 cases at first colposcopy referral (9.2%; 95% confidence intervals: and 37.8%; 95% confidence intervals: , respectively). At the next screening round, 5 years later, the majority of women from risk groups III (86%) and IV (79%) were classified as risk group I (HC2-negative and normal cytology), as Figure 3. Risk for CIN31 during follow-up between 2006 and The graph shows absolute numbers of women in whom CIN31 was detected at first colposcopy or subsequent follow-up examinations. First colposcopy includes subsequent treatment, e.g., conisation 3 months later. Group I 5 Pap normal and HC2-negative; Group II 5 Pap abnormal and HC2-negative; Group III 5 Pap normal and HC2-positive; Group IV 5 Pap abnormal and HC2-positive. Pap normal is a cytology result less than Pap2w (after Papanicolaou), LSIL. FU: follow-up; HC2: Hybrid Capture 2. the result of either surgical intervention or spontaneous regression. Furthermore, we found very few true new CIN3 lesions after 5 years in Series 2 that were exclusively explained by newly acquired HPV infections. Probably as a result of efficient communication channels within a relatively small screening project, our clinical followup data showed very high compliance rates with patient pathways, even for women who left Deutsche BKK, and we are confident that almost all cases of CIN31 were diagnosed in our institution. For the period from 2006 to 2012, the cancer registry of Lower Saxony confirmed that there was no further case of cervical cancer among participants in the area and the health insurance database showed no case among participants who left the region. It is a weakness of our trial that it was impossible to identify the true target population without a nationwide screening register. We wanted to avoid screening hysterectomised women, an issue reported by the Centers for Disease Control, 23 but Deutsche BKK s database did not contain reliable information about this exclusion criterion. Furthermore, participants frequently changed their

8 Luyten et al health insurance provider or moved to other regions. In 2008 we estimated the participation rate to be 87.2%. 24 When we sent out invitation letters to 4,318 non-attendees and asked for reasons for not participating, 1,216 replied, including 964 women who had a history of hysterectomy. Other limitations of this study were that, at the time of reporting, data from the second screening round were available for only 20% of subjects from the first screening round and, as a pilot project, WOLPHSCREEN had no randomisation. Compared with a meta-analysis of longitudinal trials, 2 our 5-year cumulative incidence of CIN31 for HC2-positive women in Groups III and IV (0.869%; 172 of 19,795) is in the upper range of these studies. Data on the risk of CIN31 in subsequent HPV screening rounds are available from a metaanalysis of four randomised controlled trials, which showed a relative risk of 0.43 ( ). 2 In comparison, there was a much steeper decrease of CIN31 cases observed in WOLPHSCREEN, which may be explained in part by the higher prevalence of HPV and CIN31 in our study compared with studies in countries with a systematic national screening programme. Previous trials either gave no details or had lower rates of histological assessment, while almost all of our participants who were transferred to colposcopy had a histological diagnosis because assessment for Type 3 TZ and for minor changes was mandatory. Only women with original squamous and normal columnar epithelium did not have biopsies. Curettages and/or colposcopically guided biopsies seem to have a therapeutic effect and a protective effect for loop insertion has also been reported. 25,26 Therefore, we attribute the low rate of subsequent CIN3 to the high rate of minimally invasive diagnostic procedures and the intense followup, 21 resulting in an immediate detection and treatment of prevalent CIN21 cases combined with modification of the natural course of persisting HPV infections. These findings suggest that once CIN21 has been completely cleared from the at-risk population then the future risk of CIN3 in women with CIN1 or persistent HPV infection should be even lower than reported in randomised controlled trials and other projects. 1 9,19 Our colposcopy protocol may be just one of many possible minimally invasive interventions. As we did not observe differences in performance between individual colposcopists, the effectiveness in preventing cervical cancer appears to depend on wellstandardised procedures rather than on individual skills. In conclusion, we demonstrated that a locally organised screening program is feasible with a very low rate of CIN31 cases in the second screening round 5 years after using HPV/cytology co-testing and risk-adapted colposcopy at recruitment. Further improvements in cervical cancer prevention will probably rely more on better colposcopy quality assessment References and compliance with patient pathways rather than on better screening or triage tests. Previously, we have emphasised the importance of colposcopy quality assurance to improve the success of screening. 20 However, the ability to significantly decrease new CIN31 cases may create a new challenge for subsequent screening rounds as it could potentially lead to increased rates of unnecessary procedures in HPV-positive women without lesions. We are aware that HPV-only testing with further extended screening intervals 12 may show a better cost-effectiveness than co-testing and some aspects of patient pathways could be optimised. For example, the follow-up protocol for women in Groups II and III could be simplified by replacing repeat cytology after 6 months and HC2 after 1 year with a single intervention at 12 months using one cervical liquid-based medium for cytology and HPV testing. In addition, there is scope to improve the screening protocol by refining the use of biomarkers such as p16/ki67 and adapting to new challenges in the era of widespread HPV vaccination. However, it remains important to recognise that the development of the WOLPHSCREEN protocol, which involved a major change from annual Pap screening to 5- yearly screening based on HPV tests, depended upon achieving consensus among multiple stakeholders. As the number of new HPV infections was moderately low and the occurrence of new CIN3 almost absent after 5 years in our population, we hypothesise that, in most cases, relevant high-risk HPV infections must be acquired early in life and HPV screening at age years will detect the majority of clinically relevant high-risk HPV-related diseases for a lifetime. Acknowledgements The project was supported by Hologic with free Cervex brush samplers and ThinPrep transport vials. The authors thank G unter Unger, Hannover for providing cancer registry data about cervical cancer cases in Wolfsburg region, Bernd Braun, Klinikum Wolfsburg for performing Hybrid Capture testing, Theodora Vasileva and Sarah Scherbring as members of the Wolphscreen colposcopy team and Stefan Lorenz, Birgit Dziuk and Holger S oldner from Deutsche BKK for their assistance. Last but not least they thank all participating gynaecologists for their superb collaboration. WOLPHSCREEN members with more than 250 recruited participants (in alphabetic order): Henryk Czuczwara (Wolfsburg), Gerta Djatschenko (Wolfsburg), Sonja Engisch (Wolfsburg). Maren von Freymann (Wolfsburg), Elisabeth Grunwald (Wolfsburg), Ilge Hettmer-Bothmann (Gifhorn), Mohamed Jifi (Wolfsburg), Beate K astner (Wolfsburg), Britta Kayser (Gifhorn), Heidrun Kieslinger (Wolfsburg), Rainer K onigsmann (Wolfsburg), Daniela Levi (Wolfsburg), Andreas Metzner (Wolfsburg), Dieter M uller (Wolfsburg), Lutz M uller (Gifhorn), Nelli M uller (Wolfsburg), Julia Nicolai (Wolfsburg), Achim Reinartz (Helmstedt), Frauke Schmidt (Wolfsburg), H.-P. Schulze (Wolfsburg), Doris Suerdieck (Braunschweig), Klaus Theiler (Wolfsburg), Bettina Wanninger (Gifhorn), Ingrid Werner (Gifhorn) and Gerhard W olbern (Wolfsburg). Deutsche BKK did not review the study design and were not involved in data collection, analysis or writing the paper. 1. Anttila A, Kotaniemi-Talonen L, Leinonen M, et al. Rate of cervical cancer, severe intraepithelial neoplasia, and adenocarcinoma in situ in primary HPV DNA screening with cytology triage: randomised study within organised screening programme. BMJ 2010;340:c Arbyn M, Ronco G, Anttila A, et al. Evidence regarding human papillomavirus testinginsecondarypreventionof

9 1416 Early detection of CIN3 and cervical cancer cervical cancer. Vaccine 2012;30 (Suppl. 5): F88 F Kitchener HC, Gilham C, Sargent A, et al. A comparison of HPV DNA testing and liquid based cytology over three rounds of primary cervical screening: extended follow up in the ARTIS- TIC trial. Eur J Cancer 2011;47: Leinonen M, Nieminen P, Kotaniemi-Talonen L, et al. Age-specific evaluation of primary human papillomavirus screening vs conventional cytology in a randomized setting. J Natl Cancer Inst 2009; 101: Mayrand MH, Duarte-Franco E, Rodrigues I, et al.; Canadian Cervical Cancer Screening Trial Study Group. Human papillomavirus versus Papanicolaou screening tests for cervical cancer. N Engl J Med 2007;357: Naucler P, Ryd W, T ornberg S, et al. Human papillomavirus and Papanicolaou tests to screen for cervical cancer. N Engl J Med 2007;357: Rijkaart DC, Berkhof J, Rozendaal L, et al. Human papillomavirus testing for the detection of high-grade cervical intraepithelial neoplasia and cancer: final results of the POBASCAM randomised controlled trial. Lancet Oncol 2012; 13: Ronco G, Giorgi-Rossi P, Carozzi F, 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: Sankaranarayanan R, Nene BM, Shastri SS, et al. HPV screening for cervical cancer in rural India. N Engl J Med 2009;360: Roland KB, Soman A, Benard VB, et al. Human papillomavirus and Papanicolaou tests screening interval recommendations in the United States. Am J Obstet Gynecol 2011;205:447.e1 447.e Saraiya M, Berkowitz Z, Yabroff KR, et al. Cervical cancer screening with both human papillomavirus and Papanicolaou testing vs Papanicolaou testing alone: what screening intervals are physicians recommending? Arch Intern Med 2010;170: Arbyn M, Sasieni P, Meijer CJ, et al. Chapter 9: Clinical applications of HPV testing: a summary of meta-analyses. Vaccine 2006;24 (Suppl. 3):S3/ 78 S3/ Davies P, Arbyn M, Dillner J, et al. A report on the current status of European research on the use of human papillomavirus testing for primary cervical cancer screening. Int J Cancer 2006;118: Ronco G, Segnan N, Giorgi-Rossi P, et al. Human papillomavirus testing and liquid-based cytology in primary cervical screening: results at recruitment from the New Technologies for Cervical Cancer randomized controlled trial. J Natl Cancer Inst 2006;98: Bulkmans NW, Berkhof J, Rozendaal L, et al. 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 2007;370: Petry KU, Menton S, Menton M, et al. Inclusion of HPV testing in routine cervical cancer screening for women above 29 years in Germany: results for 8466 patients. Br J Cancer 2003;88: Castle PE, Fetterman B, Poitras N, et al. Five-year experience of human papillomavirus DNA and Papanicolaou test contesting. Obstet Gynecol 2009;113: Thrall MJ, Russell DK, Facik MS, et al. High-risk HPV testing in women 30 years or older with negative Papanicolaou tests: initial clinical experience with 18-month follow-up. Am J Clin Pathol 2010;133: Katki HA, Kinney WK, Fetterman B, et al. Cervical cancer risk for women undergoing concurrent testing for human papillomavirus and cervical cytology: a population-based study in routine clinical practice. Lancet Oncol 2011;12: Petry KU, Luyten A, Scherbring S. Accuracy of colposcopy management to detect CIN3 and invasive cancer in women with abnormal screening tests: results from a primary HPV screening project from 2006 to 2011 in Wolfsburg, Germany. Gynecol Oncol 2013;128: Petry KU, Schmidt D, Scherbring S, et al. Triaging Pap cytology negative, HPV positive cervical cancer screening results with p16/ki-67 dual-stained cytology. Gynecol Oncol 2011;121: Walker P, Dexeus S, De Palo G, et al.; Nomenclature Committee of the International Federation for Cervical Pathology and Colposcopy. International terminology of colposcopy: an updated report from the International Federation for Cervical Pathology and Colposcopy. Obstet Gynecol 2003;101: Centers for Disease Control and Prevention (CDC). Cervical cancer screening among women by hysterectomy status and among women aged 65 years United States, MMWR Morb Mortal Wkly Rep 2013;61: Luyten A, Scherbring S, Reinecke-L uthge A, et al. Risk-adapted primary HPV cervical cancer screening project in Wolfsburg, Germany experience over 3 years. J Clin Virol 2009;46 (Suppl. 3):S5 S Castellsague X, Dıaz M, Vaccarella S, et al. Intrauterine device use, cervical infection with human papillomavirus, and risk of cervical cancer: a pooled analysis of 26 epidemiological studies. Lancet Oncol 2011;12: Stoler MH, Vichnin MD, Ferenczy A, et al. The accuracy of colposcopic biopsy: analyses from the placebo arm of the Gardasil clinical trials. Int J Cancer 2011;128: