The Wolfsburg pilot project for risk-adapted prevention of cervical cancer

On February 1, 2006, a pilot project began in Wolfsburg that was intended to examine the suitability of a new concept for preventing cervical cancer for everyday use. In order to evaluate the project, it is essential to understand the special features of cervical cancer, in particular the knowledge gained over the last three decades from basic research on the role of human papilloma viruses (HPV) in the development of this tumor.

> The American Cancer Society has described the introduction of Papanicolaou's exfoliative cytology as one of the greatest successes of the 20th century in the prevention of cancer. In fact, in almost all regions that have introduced cytology-based screening programs, there has been a sometimes dramatic decline in the incidence and mortality of cervical cancer, while in countries without screening there have been no changes. Without prevention, the life risk of cervical cancer is 3-5 %; the RKI states the current life risk for Germany as only 0.9 % (2004). In other words, the preventive care currently practiced in Germany saves up to 2,000 women from developing cervical cancer every month.
The mechanism of action of prevention is the discovery of precancerous lesions and their surgical removal. This interrupts the natural cycle of carcinogenesis and actively prevents the development of cervical cancer. Despite its undoubted success, cytology still has astonishing shortcomings. The sensitivity for the precursor lesions sought (cervical intraepithelial neoplasia grades 2 and 3 = CIN 2/3) is much lower than expected; in a meta-analysis with more than 60,000 women it was only 53%. Every second of the precursors we are looking for is therefore not detected by the one-off swab. As with any alarm system, there are two possible errors in precautionary measures: in addition to failure in the form of a lack of an alarm message in a dangerous situation, there is also the possibility of a false alarm. This means that 2-3% of all healthy women participating in the annual screening will receive an abnormal cytological result. With regular participation, the majority of healthy women will experience such a false alarm at least once in their lives. The clarification of unclear preventive findings and the surgical treatment of precursors causes considerable anxiety in affected women and is associated with serious morbidity in subsequent pregnancies. In a meta-analysis, knife conization, which is common in Germany, led to a significantly increased perinatal mortality and rate of premature births.
In addition to these disadvantages of the cytology method, the German concept also has organizational weaknesses. While in Scandinavia, for example, potential participants are actively invited, in Germany the women concerned have to visit a gynecologist on their own initiative, which results in comparatively lower participation rates. In countries with organized screening, if there are abnormal screening findings, further clarification is carried out in a minimally invasive manner in specialized consultations using colposcopy with biopsies according to clearly regulated patient pathways. Such an approach avoids incorrect treatment and was also cost-effective in randomized controlled studies. In Germany, however, the procedure for abnormal findings depends heavily on the discretion of the supervising gynecologist. This results to a considerable extent in overtreatment but also in the development of actually preventable carcinomas.
Etiology of cervical carcinoma
In contrast to breast cancer, cervical cancer has a largely monocausal etiology, which also explains why no other human cancer can be prevented in a similarly successful manner. The prerequisite for the development of this carcinoma is the infection of the cervix uteri with certain human papilloma viruses (HPV). Prototypes of these so-called high-risk HPVs (HR-HPV) are types 16 and 18, which were first discovered in invasive cervical carcinomas in 1983 and 1984 by the Heidelberg working group led by Harald zur Hausen. In one study, 99.6 % of all invasive cervical carcinomas from 5 continents contained HR-HPV. The four high-risk types HPV 16, HPV 18, HPV 45 and HPV 31 are associated with approximately 80 % of all invasive carcinomas. HPV 16 alone can be detected in half of all carcinomas, while the remaining 20 HR types are significantly less likely to be involved in carcinogenesis are. However, the development of cervical cancer is not the norm but rather the very rare consequence of a cervical HPV infection. Genital HR-HPV infections are common. However, these are mostly transient infections that heal without any consequences within 18 months in around 80 % cases without any clinical symptoms. In Europe, genital HPV infections are most common in the age group between 18 and 25 years with a prevalence of around 30 %; in Germany it drops to less than 10 % in 30 to 35 year olds and to less than 5 % in women the 40th year of life.
Almost all invasive carcinomas and all obligate precancerous lesions are associated with HPV DNA. The average latency period between initial HPV infection and invasive carcinoma is 15 to 30 years. Based on molecular biological and epidemiological studies, the minimum latency period is apparently 8 years. This results in a conclusion that is unique in oncology and relevant to the prevention of cervical cancer: If an infection of the cervix uteri with high-risk HPV is ruled out, then any risk of cervical cancer can be ruled out for the next 7 years. Prevention based on HPV detection could therefore avoid overdiagnosis and overtreatment of HPV-negative participants and, conversely, focus the necessary attention on women who are actually at risk for the presence and development of precancerous lesions and cancer.
Primary HPV screening for
Prevention of cervical cancer
Numerous cohort studies and randomized controlled studies have now been published that demonstrate the accuracy of more precise prevention of cervical cancer through HPV testing at a high level of evidence. An international cohort study with more than 24,000 participants and an observation period of more than six years showed that the combined use of cytology and HPV testing allowed an almost perfect risk assessment. More than a third of the women with abnormal cytology and a positive HPV test developed CIN3 or an invasive carcinoma. For women who only had a positive HPV test at the start of the study, there was a moderate long-term risk of 11%, whereas the corresponding risk was for HPV-negative participants less than one percent. Overall, a negative HPV test ruled out the risk of CIN3/CA disease for six years more reliably than an unremarkable cytological smear for one year.
The publications of six randomized controlled trials (RCTs) with more than a quarter of a million participants and follow-up observations of up to eight years confirmed that HPV testing can better detect precancerous lesions. In a meta-analysis, the detection rate of precancerous lesions and cancer was increased by 50% for the first screening examination; in the second screening round five years later, only half as many CIN3 were found in women previously examined using an HPV test as in the cytological comparison arm. This can only be explained by the fact that HPV testing actually detects almost all existing neoplasms and five years later only newly formed precursors are found at intervals. In an RCT, this already led to a significant reduction in new cases of cervical cancer in the HPV arm compared to the conventional arm. A large RCT in India with more than 130,000 participants, which examined whether a one-time screening test could reduce the number of new cases and deaths from cervical cancer, showed a significant reduction in mortality after 8 years only for the HPV test.
Given the high level of evidence, WHO and IARC classify HPV screening as at least as good as previous cytological screening, and the EU sees the superiority of the HPV test as a primary screening method in its current recommendations, but only in organized screening programs for women aged 30 and over and when the smear intervals are extended to at least 5 years. The higher medical quality of HPV screening if this requirement is met is therefore undisputed, although before it is implemented in daily practice it is still unclear whether such a screening concept would be cost-effective in all European health systems, whether it would be accepted by the participants, and whether it would lead to uncertainty and, especially in Germany, whether annual gynecological examinations could be neglected.
The Wolfsburg pilot project
The first European pilot project started in Wolfsburg in February 2006. It is based on a contract between the German BKK and all practicing gynecologists and the Wolfsburg Clinic. Participation in the project is recommended for female insured persons of Deutsche BKK who are older than 30 and still have their uterus. After appropriate information in practice, the women affected either agree to participate in writing or can decide to maintain the old concept with annual smears. Women who do not take part in preventative care will be contacted.
The further procedure is clearly regulated between the participating doctors and the clinic, depending on the preventive findings. The data collected in a questionnaire and the preventive findings are stored in a central database maintained in the clinic to monitor and control the agreed patient paths. If the cytology is normal and the HPV test is negative (only high-risk types with Hybrid Capture 2), the next smears will be taken after five years, but continued participation in the remaining annual screening is recommended. If there are abnormal findings, the affected women are referred to the dysplasia consultation either immediately (abnormal cyto and positive HPV test) or, if abnormalities persist, after 6 or 12 months.
Participation and adherence to patient pathways
After four years, more than 18,393 women took part in the project, meaning the participation rate in the city of Wolfsburg was around 90% of the target population. Unfortunately, it is not possible to accurately determine the participation rate due to changes in residence and health insurance, as well as uncertain information about possible uterine operations. According to data collection by the German BKK, the practicing gynecologists and a survey of the participants by letter, the proportion of hysterectomized women in the target population is approximately 21% (18-24%). Given the uncertainties about the actual proportion of women who have already had their uterus removed, it is not possible to determine the participation rate more precisely. However, this is in any case very high and exceeds the target of 85 % set in organized prevention programs. The reasons may have been better information provided by the participating gynecologists, but also a written invitation to all BKK members who had not yet enrolled after two years. Not even 1% of the target group decided to maintain the standard preventive care with annual cytological smears, although this proportion fluctuated between 0 and almost 5% between the individual practices.
A specially set up telephone hotline to calm participants who might be upset about screening results proved to be unnecessary; the mass panic feared by some due to positive HPV results did not materialize. This also corresponds to the observations of an English study on the value of the HPV test in primary care.
93.1% of the participants were HPV negative and had unremarkable cytology. Only 0.9% of all participating women had a positive HPV test in combination with an abnormal cytology, while 1.2% of all participants were HPV negative but had abnormal cell smears and, conversely, 4.8% had a positive HPV test with an unremarkable screening smear .
The specified patient pathways were adhered to to a high degree. 91% of the participants with a positive HPV test and abnormal cytology were referred immediately as planned, and 82% of the women with abnormal cytology but a negative HPV test were further clarified in the practice and, if necessary, in the clinic as planned. Even the somewhat complicated algorithm for women with a positive HPV test but an inconspicuous smear was completed according to plan by almost 79%. The reason for the very good rates, even compared to studies, may have been the central data collection and quality control in the clinic. The participating practices were informed about all patients who did not follow the procedure, and the participants were informed in writing about the purpose of the planned examination.
Referral rates, diagnostics and therapy
Only 2.9% of all participants required a referral for colposcopy. In the four-year period, a total of 142 women were diagnosed with CIN3 or cervical cancer; all but one case were HPV positive. In contrast, exactly 50% (n=71) of these patients had normal cytological findings of classes Pap I/II and were only referred later due to abnormal control smears after 6 or 12 months (n=35) or only because of persistent HPV detection. The cytological smear revealed suspicion of CIN3 or cancer in only 17 of the 142 people suffering from CIN3 and carcinoma. In 124 patients, participation in the project led to an earlier diagnosis and thus possibly to better prevention of carcinomas. What is particularly striking is the high proportion of precursors of adenocarcinoma of the uterine cervix. Several epidemiological studies have shown that this subtype of cervical cancer is obviously not prevented, or at least not significantly so, by current preventive measures. Approximately 20% of all invasive cancers are adenocarcinomas, while in routine care, almost 99% of CIN3 are precursors of the more common squamous cell carcinoma. In the Wolfsburg project, 12 of 124 CIN3 were precursors of adenocarcinoma. It can be expected that a number of women in this group in particular have been spared from developing cancer that could not have been prevented with the usual preventative measures.
If only the 18 women who had invasive carcinomas were considered, then 7 cases were diagnosed in women who had not taken part in preventive care for more than 3 years. The remaining 11 cases must be classified according to international criteria as interval cancers and thus failures of cytological screening, since the women had regularly taken part in routine screening before enrolling in the Wolfsburg project. 6 of these 11 cases were diagnosed in women with again normal cytology based solely on their positive HPV test.
The defined patient pathways ensured a very good selection of those who were actually sick and over-treatment of healthy people was avoided. However, the proportion of women who only had an HPV infection without precancerous lesions was relatively high, at 36% (195 of 535) of all participants referred for colposcopy.
The women already suffering from cervical cancer were treated in accordance with current oncological guidelines; CIN3 were treated on an outpatient and minimally invasive basis using laser conization or snare resection.
It is also noteworthy that the participation rate in the annual routine gynecological examination was higher than for those insured by other health insurance companies for whom annual cytosmears were retained. While 82% of the Deutsche BKK insured persons who were included in the project in 2006 returned to the same practice at least once for a routine examination by the end of 2008, this was only the case for just under 60% of the insured persons from other health insurance funds.
Despite all the advantages, further improvements appear to be possible in the pilot project. The risk of CIN3 in participants with a negative HPV test, even with an abnormal cytological smear, is so low (no case of CIN3 or Ca among 209 women) that the previous six-monthly checks seem excessive.
However, the greatest need for optimization is among women with normal cytology and a positive HPV test. On the one hand, these are the big winners, as it is precisely in this group that sick women were found who would otherwise have slipped through the preventive safety net. At the beginning of the project, however, we assumed that the cytological control smears would detect the majority of CIN3+ cases after 6 months and that only a few patients would have to be referred for clarification after one year due to a high spontaneous healing rate of approx. 60%. However, the spontaneous healing rate actually observed was only 42 % and more than half of all CIN3+ cases were not detected by the control cytology. This led to a diagnosis delay of 12 months in three patients with invasive carcinomas and, on the other hand, to referrals of women for further clarification who only had an HPV infection without any cell changes. Through accompanying scientific studies, immunocytochemical test procedures were identified that allow early identification of those who are actually sick with good sensitivity and specificity. There is a further opportunity for improvement at this point for the extension of the project that is due to begin in February 2011.
Several international studies and two German HTA reports have shown that HPV-based prevention is more economically efficient than the previous standard if the intervals between examinations are extended to at least two years and the participants are at least between 25 and 35 years old exhibit. The cost effectiveness of structured programs is also very well documented for the use of colposcopy to clarify abnormal preventive findings instead of the diagnostic conization that is common in Germany. In principle, the Wolfsburg pilot project is certainly cost-effective. However, bonus payments were made to the practicing gynecologists to ensure a high level of recruitment and acceptance. This was a sensible step for the implementation of the medically sensible project, which should not be fundamentally questioned in the future. The exact amount of such a bonus payment is currently to be clarified by a health economic study, the results of which are still pending.
Even if there is political will, a change in the German precautionary strategy in line with the pilot project is likely to be made more difficult by the lack of infrastructure to date. In addition to an organizational structure for invitations and quality assurance, a network of colposcopy consultation hours with sufficient expertise would also have to be established. <<

Summary:
Pilot project of the German BKK
Background:
In randomized trials, human papillomavirus (HPV) detection was more effective than cytology-based standard screening in organized screening of women over 30 years of age in preventing cervical cancer. However, so far there is no evidence that this superiority can also be transferred to routine preventive care.
Methods:
In the Wolfsburg pilot project, female insured persons from Deutsche BKK who are older than 30 years are divided into risk groups using HPV tests and cytology. For the majority with an assumed zero risk (HPV negative/cyto normal), the next smears will not be taken for another 5 years; all others, depending on the risk, will either be referred for minimally invasive diagnostics or checked in the practice.