Pii: s0959-8049(00)00122-2

European Journal of Cancer 36 (2000) 1473±1478 Recommendations on cancer screening in the European Union Advisory Committee on Cancer Prevention * Received 10 January 2000; accepted 22 February 2000 excluding non-melanoma skin cancer, occurred in the European Union in 1995. Of these, 2% were cervical Screening allows the detection of cancers at an early cancers, 13% breast cancers, 13% colorectal cancers stage of invasiveness or even before they become inva- and 8% prostate cancers. Cervical and breast cancer sive. Some lesions can then be treated more e€ectively constituted 4% and 29%, respectively, of new cancers in and the patients can expect to live longer. The key indi- women, and prostate cancer constituted 14% of new cator for the e€ectiveness of screening is a decrease in cancers in men. All rates presented here are age stan- disease-speci®c mortality or incidence.
dardised with the European Standard Population [3].
Screening is, however, the testing of healthy people for diseases which have so far not given rise to symp- toms. Aside from its bene®cial e€ect on disease-speci®c mortality or incidence, screening might therefore also have some negative side-e€ects for the screened population.
Screening is only one method of controlling cancer.
Healthcare providers should know all the potential Whenever possible primary cancer prevention should be bene®ts and risks of screening for a given cancer site given ®rst priority. When cancer screening is undertaken before embarking on new cancer screening programmes.
it should be o€ered only in organised programmes with Furthermore, for the informed public of today, it is quality assurance at all levels, and good information necessary to present these bene®ts and risks in a way about bene®ts and risks. The bene®ts of a screening that allows the individual citizen to decide on whether programme are achieved only if the coverage is high.
to participate in the screening programmes.
When organised screening is o€ered high compliance The purpose of this document is to give recommen- should, therefore, be sought. Opportunistic screening dations on cancer screening in the European Union.
activities are normally not acceptable as they may not These recommendations address the people, the politi- achieve the potential bene®ts and may cause unneces- cians and the health administrations of the Member States, the European Commission and the European New cancer screening tests should be evaluated in randomised trials before being implemented in routine Principles for screening as a tool for the prevention of chronic non-communicable diseases were published by The reduction in disease-speci®c mortality achieved in the World Health Organisation in 1968 [1] and by the trials depends on the sensitivity of the screening test, the Council of Europe in 1994 [2]. These two documents compliance amongst the invited, the screening fre- form, together with the present state of art in each of quency, the number of screens each person has, the the cancer screening ®elds, the basis for the present completeness of the follow-up and the bene®t of early treatment. The negative side-e€ects in the screened All data on incidence and mortality are quoted from population depend on the sensitivity and the speci®city the recently published EUCAN data covering 1995. An of the test, and on the possible side-e€ects of early estimated number of 1 488 000 new cancer cases, treatment. The ®ndings from trials can be extrapolated to the general population only if the conditions in the trials can be reproduced in the routine healthcare sys- * Corresponding author: E. Lynge, Institute of Public Health, tem. This requires an organisation with a call±recall University of Copenhagen, Blegdamsvej 3, DK-2200 Kùbenhavn N, Denmark. Tel.: +45-35-32-76-35; fax: +45-35-35-11-81.
system and with quality assurance at all levels, and it E-mail address: [email protected] (E. Lynge).
requires an e€ective and appropriate treatment service.
0959-8049/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
Advisory Committee on Cancer Prevention / European Journal of Cancer 36 (2000) 1473±1478 Centralised data systems are needed for the running is thus vital and a high degree of organisation is needed of organised screening programmes. This includes a computerised list of all persons to be targeted by the Nationally organised cervical screening programmes screening programme. It includes also computerised exist in Sweden, Finland, Denmark, The Netherlands data on all screening tests, assessment and ®nal diag- and the UK. A European set of guidelines for cervical noses. Organised screening also implies scienti®c analy- screening was developed in 1993. It provides targets for sis of the outcome of the screening and quick reporting the quality assurance of organised screening pro- of these results to the population and screen providers.
grammes [8]. Ten centres with cervical screening have, in This analysis is facilitated if the screening database is the past, been ®nancially supported by the Europe Against Cancer programme. These 10 programmes have High quality screening is possible only if the per- recently formed a network focusing on quality assur- sonnel at all levels are adequately trained for their ance, epidemiology and new technologies.
tasks. Performance indicators should be monitored The limited screening resources should be concentrated in the age range of 30±60 years. A large proportion of Ethical, legal, social, medical, organisational and cervical abnormalities will regress to normal if left economic aspects have to be considered before decisions untreated. Screening should therefore de®nitely not start can be made on the implementation of cancer screening.
before the age of 20 years and in many countries prob- Resources, human as well as economic, must be avail- ably not before the age of 30 years. The protective e€ect able in order to assure the appropriate organisation and of screening of women older than 60 years is limited, quality control. Actions have to be taken to ensure dif- especially if these women previously had negative tests.
ferent socio-economic groups equal access to screening.
Screening should be undertaken with a 3±5-year The implementation of a cancer screening programme is interval. Prolonged intervals may be considered in therefore a decision to be made locally, depending on women with a history of negative tests. The bene®t of the disease burden and the healthcare resources.
more frequent screening is very limited and, in addition, Cancer is a leading disease and cause of death it increases the risk of overtreatment of otherwise throughout Europe. European collaboration should facilitate high quality cancer screening programmes and protect the population from poor quality screening.
Pap smears should be the method used in cervical When screening is o€ered it should start at the latest by the age of 30 years and de®nitely not before age 20 In an unscreened population, the incidence of cervical years. The upper age should depend on the available cancer reaches its maximum around the age of 50 years.
resources but should preferably not be lower than 60 In screened populations, the incidence tends to be high- years. Limited screening resources should be con- est for women above the age of 60 years. The incidence centrated in the age range of 30±60 years.
of cervical cancer re¯ects both background risk and Screening intervals should be between 3 and 5 years.
screening activity during the previous decades. The Screening more often than every third year should be highest incidence of cervical cancer is now observed in discouraged. Smear taking in healthy women should be Portugal with 19 per 100 000 and the lowest in Lux- undertaken only in organised screening programmes embourg with 4 per 100 000. Mortality rates are highest with quality assurance at all levels.
in Denmark, Austria and Portugal with 6±7 per 100 000 Cervical cancer screening programmes should be and lowest in Luxembourg and Finland with approxi- organised in accordance with the European guidelines.
3.3.2. To the European Commission and the European A common terminology for histology and cytology Whilst no randomised trials on cervical screening with should be implemented. For the laboratories, a detailed Pap smears were ever carried out, the e€ectiveness of quality control programme should be de®ned based on cervical screening programmes has been demonstrated the existing guidelines and implemented at the national in several countries [4±6]. It is estimated that cervical smears every 3 years can prevent 90% of cervical can- Recommendations for training and quality control cers in a population if all women attend and all detected could be proposed and tested in the network centres. As lesions are adequately followed-up [7]. High compliance di€erent treatment options are currently adopted, Advisory Committee on Cancer Prevention / European Journal of Cancer 36 (2000) 1473±1478 auditing of cases should be performed by a core group develop guidelines for best practice related to breast of clinicians. A concerted e€ort should be made to ®nd screening. The desirable endpoint for each member of the most e€ective methods for follow-up and treatment the network is to establish a co-ordination of the screening activities in their country and to operate a Validation studies of liquid-based and automated service and/or reference centre for these activities.
screening methods with special attention to cost- During its 10 years of existence the network has noted e€ectiveness should be undertaken. Well-designed that population-based screening requires the full sup- studies should be undertaken on the use of human port of national or regional health authorities, and that papilloma virus (HPV) testing as a screening method the decision to start a programme needs to be taken by and/or as a supplementary method in the follow-up of appropriate health authorities. Screening for breast cancer is multidisciplinary and the quality of the whole Studies should be undertaken of recent trends in the process (invitation, diagnosis, assessment of suspicious incidence of cervical cancer in Europe in order to opti- lesions, treatment and follow-up) needs to be ensured mise the lower and upper age limits for screening.
before initiating a programme. Initial and continuous training of all personnel involved is mandatory.
Mechanisms are needed to monitor the quality of the The di€erent healthcare systems in Europe have made it necessary to ®nd di€erent solutions to common prob- lems. The network has demonstrated the importance of In countries with national population-based cancer high quality radiological examination and the need for a statistics, such as the Nordic countries, the incidence of centralised reading of mammograms taken in a decen- breast cancer has increased during the last four decades.
tralised setting. It has also demonstrated the need for The start of a mammography screening programme is standards on the minimal number of women to be associated with a temporary increase in the incidence of examined in a centre in order to maintain the level of breast cancer, and the European di€erences in breast cancer incidence therefore at present re¯ect both back- European Guidelines for Quality Assurance in Mam- ground risks and screening activities. At present, the mography Screening is a document with minimal and incidence is highest with 120 per 100 000 in The optimal requirements for quality assurance of organised Netherlands, where a screening programme started screening programmes [11]. An updated version will be recently, and lowest in Spain and Greece with 61±63 per 100 000. Breast cancer is rare under the age of 30 years and the incidence increases with age. Breast cancer mortality is highest in Denmark, 38 per 100 000 and lowest in Greece, 23 per 100 000. Mortality rates have increased during the last decades in the majority of Mammography should be the method used in breast European countries, whereas they have been stable or cancer screening. There is at present no convincing evi- decreased slightly in the Nordic countries and in the UK.
dence for the e€ect of screening based on breast self- examination or clinical breast examination.
Women without symptoms of breast cancer should be o€ered mammography examination only in organised Screening for breast cancer with mammography has screening programmes with quality assurance at all been studied in a number of randomised trials. Data levels. When mammography screening is o€ered, only from ®ve Swedish counties showed a 30% decrease in women aged 50±69 years should be invited.
breast cancer mortality amongst women invited to Screening intervals should be 2±3 years.
screening at age 50±69 years [9]. Updated data from Breast cancer screening programmes should be orga- Sweden also indicate a reduction in breast cancer mor- nised in accordance with the European guidelines [11].
tality among women invited to screening at age 40±49 Adverse e€ects of mammography screening in women years [10]. The cost-e€ectiveness is, however, not clear aged 40±49 years may not be negligible, due to the lower predictive value of mammography in this age group, the A European breast cancer screening network was possible detection of non-progressive cancers and the established in 1989 with the aims to provide experience for countries with no breast screening service, to explore Thus, if screening is o€ered to women aged 40±49 methods of implementation into the national health years in some centres or European regions, according to systems, to establish contact for exchange of informa- local resources and quality standards reached in screen- tion between Member States, and most importantly to ing o€ered to older women, the following requirements Advisory Committee on Cancer Prevention / European Journal of Cancer 36 (2000) 1473±1478 are needed: (1) women should be clearly informed about the possible bene®ts and adverse e€ects of screening; (2) organised programmes should be set up in order to Faecal occult blood test, sigmoidoscopy and colon- discourage spontaneous screening in units without ade- oscopy have all been considered as screening tests for quate quality control systems; (3) two-view mammo- graphy with double reading and 12±18 months of The faecal occult blood test is the only test which has interval should be used; (4) data monitoring and proper been extensively evaluated as a screening tool on the population level. Four European trials have been undertaken [12±16]. There are three randomised trials 4.3.2. To the European Commission and European from Funen, Nottingham, UK and Gothenburg, Swe- den and one non-randomised trial from Burgundy, E€orts should be continued to improve breast cancer France. In the last trial people from small areas screening in Europe by promoting exchange of experi- `cantons' were allocated to either the screening or the ence. This may best be achieved by continuation of the control group. Only two screening rounds were under- activities of the European breast cancer screening net- taken in Gothenburg. In Funen, Nottingham and Burgundy, screening was o€ered ®ve times. A recent Updated guidelines should be published at regular meta-analysis of all randomised faecal occult blood test intervals. Quality management should be ensured, trials showed a 16% reduction in colorectal cancer including training and education in business strategy, recruitment, training and retention of quali®ed sta€, Pilot screening programmes with the faecal occult quality assurance providing consumer protection, and blood test will start in two areas in England and Scot- management of political, governmental, economic, land in the year 2000, and pilot projects are under con- social and technical aspects of a programme.
sideration in one area in Austria and one in Spain.
Research should be encouraged on the impact of Annual faecal occult blood tests are o€ered as part of screening on breast cancer mortality, progression of the German cancer screening activities.
mammography-detected lesions, ethical questions, More complex faecal occult blood tests, especially population acceptance, method of invitation, cost-e€ec- immunological tests, have been developed [18,19]. They tiveness and psychosocial e€ects. These research activ- are more sensitive, but their speci®city at the population ities should address mammography screening both level is not well established. The e€ectiveness of ¯exible below the age of 50 years, in the age range of 50±69 sigmoidoscopy as a screening tool is currently being years and from the age of 70 years onwards. Support tested in randomised trials in England and Italy [20,21].
should be given to the development of appropriate data A system should be set up for accreditation on a European level of screening programmes applying to become reference centres in the breast cancer screening As colorectal cancer is a major health problem in many European countries faecal occult blood screening should be seriously considered as a preventive measure.
The decision on whether or not to embark on these screening programmes must depend on the availability of the professional expertise and the priority setting for If screening programmes are implemented they should For men the highest incidence of colorectal cancer use the faecal occult blood screening test and colon- incidence is found in Ireland, Austria and Denmark oscopy should be used for the follow-up of test positive with 58±61 per 100 000 and the lowest in Greece with 25 cases. Screening should be o€ered to men and women per 100 000. For women the highest incidence is found aged 50 years to approximately 74 years. The screening in Denmark, The Netherlands and Ireland with 40±43 per 100 000 and the lowest in Greece with 19 per Other screening methods such as immunological tests, 100 000. Mortality rates for men are highest in Denmark ¯exible sigmoidoscopy and colonoscopy can at present and Ireland with 35±36 per 100 000 and lowest in Greece not be recommended for population screening.
with 13 per 100 000. For women the mortality is highest in Denmark with 27 per 100 000 and lowest in Greece 5.3.2. To the European Commission and Parliament with 9 per 100 000. Despite advances in diagnostic Guidelines should be developed both at the European techniques and treatment the 5-year survival rates and national levels on quality assurance of faecal occult Advisory Committee on Cancer Prevention / European Journal of Cancer 36 (2000) 1473±1478 E€orts should be continued to improve faecal occult [27,28], potential overdiagnosis [29], quality of life and blood tests. They must be carefully evaluated at a interval cancers. An update of the international co- population level before being proposed in organised operation will be published soon. A comprehensive screening programmes with a special attention to cost- review on prostate cancer screening has been published e€ectiveness. The e€ectiveness of ¯exible sigmoidoscopy as a screening tool should be evaluated in randomised As long as randomised studies have not shown an advantage on prostate cancer mortality or related qual- ity of life, screening for prostate cancer is not recom- The highest incidence of prostate cancer is observed in Finland, 101 per 100 000 being four times higher than in 6.3.2. To the European Commission and European Greece, 24 per 100 000. This pronounced di€erence between European countries may re¯ect di€erences in The European randomised trial should be completed.
medical procedures, in addition to variation in exposure to risk factors. This is supported by a smaller variation in mortality, being highest in Sweden, 36 per 100 000 and lowest in Greece, 17 per 100 000.
Prostate cancer is predominantly a disease of older Decisions on the implementation of cancer screening age, and due to increasing longevity the number of cases programmes should be made within the frame of the is expected to increase over the coming years [22]. Part general priority setting on the use of healthcare resources.
of the presently observed increase in incidence in some Cancer screening should only be o€ered to healthy European countries is most likely due to opportunistic people if the screening is proven to decrease the disease- screening with the prostate speci®c antigen (PSA).
speci®c mortality or incidence, if the bene®ts and risks are well known, and if the cost-e€ectiveness of the screening is acceptable. At present, these screening The e€ect of screening on prostate cancer mortality has not been documented. Rectal examination has been . Pap smear screening for cervical abnormalities part of the annual health check-up o€ered in Germany starting at the latest by the age of 30 years and since the 1970s, but apart from this prostate cancer de®nitely not before the age of 20 years.
screening has not been an accepted policy in Europe.
. Mammography screening for breast cancer in Opportunistic screening is, however, increasing. In the USA, the incidence of prostate cancer has almost dou- . Faecal occult blood screening for colorectal cancer bled from 1986 to 1992 to decline again from 1992. This is most likely due to PSA screening [23]. A slight decline No other screening test should be o€ered healthy in prostate cancer mortality started in American men in people before these tests have been shown to decrease the 1992, but the decline is so far without a conclusive disease-speci®c mortality or incidence. Once the e€ec- tiveness of a new screening test has been demonstrated, The European Randomised Study of Screening for evaluation of modi®ed tests (e.g. alternative tests to the Prostate Cancer (ERSPC) was initiated in 1994 in two and faecal occult blood analysis or interpretation of cervical later in seven EU countries. It is the purpose of the study specimens) may be possible using surrogate endpoints.
to test a 20% reduction in prostate cancer mortality Potentially promising screening tests should be eval- after two screens in men followed-up for 10 years. The uated in randomised controlled trials, as is currently the study aims at randomising 192 000 men to the screening or control groups. In November 1999, 170 000 men have been randomised. Final results are expected in 2008.
ERSPC has joined forces with the Prostate, Lung, . Mammography screening for women aged 40±49 Colon, Ovary (PLOC) screening study of the US National Cancer Institute [26]. The US study will . Flexible sigmoidoscopy for colorectal cancer.
include 63 625 men. A common analysis has been plan- Pap smear screening for cervical abnormalities, mam- ned. In the meantime, the collected data o€er excellent mography screening for women aged 50±69 years and opportunities for evaluation of the screening test faecal occult blood screening for colorectal cancer in Advisory Committee on Cancer Prevention / European Journal of Cancer 36 (2000) 1473±1478 subjects aged 50±74 years should be o€ered only in randomised study for detection of colorectal cancer by faecal organised screening programmes with quality assurance occult blood testing. Results of 68 308 subjects. Scand J Gastro- at all levels, and good information about bene®ts and 15. Tazi MA, Faivre J, Dassonville F, Lamour J, Milan C, Durand G. Participation in faecal occult blood screening for colorectal cancer in a well de®ned French population: results of ®ve screen- ing rounds from 1988 to 1996. J Med Screen 1997, 4, 147±151.
16. Faivre J, Tazi MA, Milan C, Lejeune C, Durand G, Lamour J.
Controlled trial of faecal occult blood screening for colorectal The document was ®nalised after discussion at the cancer in Burgundy (France). Results of the ®rst 9 years. Gastro- Conference on Screening and early Detection of Cancer, 17. Towler B, Irwig L, Glasziou P, Kewenter J, Weller D, Silagy C. A systematic review of the e€ects of screening for colorectal cancer using the faecal occult blood test, Hemoccult. Br Med J 1998, 18. Saito H, Soma Y, Koeda J, et al. Reduction in risk of mortality from colorectal cancer by fecal occult blood screening with 1. Wilson JMG, Jungner G. Principles and Practice of Screening for immunochemical hemagglutination test. A case±control study.
Disease. Public Health Papers 34. Geneva, World Health Orga- 19. Castiglione G, Zappa M, Grazzini G, et al. Immunochemical vs 2. Council of Europe: Committee of Ministers. On Screening as a guaiac faecal occult blood tests in a population-based screening Tool of Preventive Medicine. Recommendation no. R (94) 11.
programme for colorectal cancer. Br J Cancer 1996, 74, 141±144.
Strasbourg, Council of Europe, 1994.
20. Atkin W, Cuzick J, Northover JMA, Whynes D. Prevention of 3. http://iarc.fr. International Agency for Research on Cancer.
colorectal cancer by once-only sigmoidoscopy. Lancet 1993, 341, Cancer Incidence Data Bases. EUCAN 1995 (retrieved 13 October 21. Senore C, Segnan N, Rossini FP, et al. Screening for colorectal 4. Hakama M. Trends in the incidence of cervical cancer in the cancer by once only sigmoidoscopy: a feasibility study in Turin, Nordic countries. In Magnus K, ed. Trends in Cancer Incidence.
Italy. J Med Screen 1996, 3, 72±78.
22. Boyle P, Maisonneuve P, Napalkov P. Geographical and tem- 5. LaÈaÈra E, Day N, Hakama M. Trends in mortality from cervical poral patterns of incidence and mortality from prostate cancer.
cancer in the Nordic countries: association with organised Urology 1999, 46(Suppl. 3A), 47±55.
screening programmes. Lancet 1987, i, 1247±1249.
23. Hankey BF, Feuer EJ, Clegg LX, et al. Cancer surveillance series: 6. Sasieni PD, Adams J. E€ect of screening on cervical cancer mor- interpreting trends in prostate cancer Ð part I: evidence of the tality in England and Wales: analysis of trends with an age period e€ects of screening in recent prostate cancer incidence, mortality, cohort model. Br Med J 1999, 318, 1244±1245.
and survival rates. J Natl Cancer Inst 1999, 91, 1017±1024.
7. IARC Working Group on Evaluation of Cervical Cancer 24. Feuer EJ, Merrill RM, Hankey BF. Cancer surveillance series: Screening Programmes. Screening for squamous cervical cancer: interpreting trends in prostate cancer Ð part II: cause of death duration of low risk after negative results of cervical cytology and misclassi®cation and the recent rise and fall in prostate cancer its implication for screening policies. Br Med J 1986, 293, 659± mortality. J Natl Cancer Inst 1999, 91, 1025±1032.
25. Etzioni R, Legler JM, Feuer EJ, Merrill RM, Cronin KA, Han- 8. Coleman D, Day NE, Douglas G, et al. European guidelines for key BF. Cancer surveillance series: interpreting trends in prostate quality assurance in cervical cancer screening. Eur J Cancer 1993, cancer Ð part III: quantifying the link between population pros- tate-speci®c antigen testing and recent declines in prostate cancer 9. NystroÈm L, Rutqvist LE, Wall S, et al. Breast cancer screening mortality. J Natl Cancer Inst 1999, 91, 1033±1039.
with mammography: overview of Swedish randomised trials.
26. Auvinen A, Rietbergen JBW, Denis LJ, SchroÈder FH, Prorok PhC for the International Prostate Cancer Screening Trial Eva- 10. Larsson L-G, Andersson I, Bjurstam N, et al. Updated overview luation Group. Prospective evaluation plan for randomised trials of the Swedish randomised trials on breast cancer screening with of prostate cancer screening. J Med Screen 1996, 3, 97±104.
mammography: age group 40±49 at randomisation. Monogr Natl 27. Beemsterboer PMM, Kranse R, de Koning HJ, Habbema JDF, SchroÈder FH. Changing role for 3 screening modalities in the 11. European Commission. European Guidelines for Quality Assur- European Randomised Study of Screening for Prostate Cancer ance in Mammography Screening. 2nd edn. Bruxelles, European (Rotterdam). Int J Cancer 1999, 84, 437±441.
28. SchroÈder FH, Van der Maas P, Beemsterboer PMM, et al. Digi- 12. Hardcastle JD, Chamberlain JO, Robinson MHE, et al. Random- tal rectal examination (DRE) Ð its value in the diagnosis of ised controlled trial of faecal-occult-blood screening for colo- prostate cancer. J Natl Cancer Inst 1998, 90, 1817±1823.
rectal cancer. Lancet 1996, 348, 1472±1477.
29. SchroÈder FH, Cruijssen-Koeter I van der, Kranse R, et al. Pros- 13. Kronborg O, Fenger C, Olsen J, Jùrgensen OD, Sùndergaard O.
tate cancer detection at low values of prostate speci®c antigen Randomised study as screening for colorectal cancer with faecal- occult blood test. Lancet 1996, 348, 1467±1471.
30. SchroÈder FH. Prostate cancer. In Kramer BS, Gohagan JK, 14. Kewenter J, Brevenge H, Engaras B, Haglind E, Ahren C.
Prorok PC, eds. Cancer Screening: Therapy and Practice. New Results of screening, rescreening and follow-up in a prospective York, Marcel Dekker Inc, 1999, 461±514.

Source: http://www.spa.lt/en/ARBYN-EuJCancer2000_ViennaRecommendations.pdf