Elsevier

The Lancet

Volume 364, Issue 9430, 17–23 July 2004, Pages 249-256
The Lancet

Articles
The cervical cancer epidemic that screening has prevented in the UK

https://doi.org/10.1016/S0140-6736(04)16674-9 Get rights and content

Summary

Background

Recent reports suggest that the reduction in mortality achieved by the UK national cervical screening programme is too small to justify its financial and psychosocial costs, except perhaps in a few high-risk women.

Methods

We analysed trends in mortality before 1988, when the British national screening programme was launched, to estimate what future trends in cervical cancer mortality would have been without any screening.

Findings

Cervical cancer mortality in England and Wales in women younger than 35 years rose three-fold from 1967 to 1987. By 1988, incidence in this age-range was among the highest in the world despite substantial opportunistic screening. Since national screening was started in 1988, this rising trend has been reversed.

Interpretation

Cervical screening has prevented an epidemic that would have killed about one in 65 of all British women born since 1950 and culminated in about 6000 deaths per year in this country. However, these estimates are subject to substantial uncertainty, particularly in relation to the effects of oral contraceptives and changes in sexual behaviour. 80% or more of these deaths (up to 5000 deaths per year) are likely to be prevented by screening, which means that about 100000 (one in 80) of the 8 million British women born between 1951 and 1970 will be saved from premature death by the cervical screening programme at a cost per life saved of about £36 000. The birth cohort trends also provide strong evidence that the death rate throughout life is substantially lower in women who were first screened when they were younger.

Introduction

The death rate from cervical cancer has fallen in many countries since the introduction of systematic cytological screening.1 In England and Wales, the rate in women younger than 65 years has halved since 1988, when the UK national call–recall system began. However, this reduction does not provide a reliable estimate of the effect of screening. Cervical cancer rates in this country had been falling in older women since the 1960s for reasons independent of screening, and widespread opportunistic screening was already preventing many deaths by the late 1970s, particularly in younger women.2

Infection with one of the oncogenic HPV (human papillomavirus) types is a necessary cause of cervical cancer, as HPV DNA is detectable in virtually all cervical cancers.3 The changes in sexual behaviour and widespread use of oral rather than barrier contraceptives, which began in the 1960s in Britain, led to rises in prevalence of many sexually transmitted diseases. Genital HPV infection must also have become much more frequent since the 1960s. Infection with oncogenic HPVs could not be diagnosed reliably until the late 1980s, but the recorded incidence of genital warts, which are usually caused by HPV6 or HPV11—rose six-fold in England and Wales between 1971 and 1994.4 Analysis of stored serum samples in the Nordic countries showed a major rise in seroprevalence of HPV and other sexually transmitted diseases from the 1960s to the 1980s.5

HPV prevalence is underestimated in cross-sectional studies, since most infections are only transiently detectable.6 A population-based randomised trial of HPV testing in addition to cytology has recruited more than 25000 women attending for routine screening in Manchester, UK, between 2001 and 2003. HPV prevalences by Hybrid Capture 2 in women aged 20–29, 30–39, 40–49, and 50–64 years were 33%, 15%, 9%, and 6%, respectively.7 After 1 year, 59% of cytologically normal HPV carriers had cleared the infection, implying that two tests a year apart would detect oncogenic HPV in about half of all women younger than 30 years.7 In a random sample of over 6000 women attending for routine cytological screening in Manchester, UK, between 1987 and 1993,8 the corresponding age-specific prevalences were 20%, 10%, 5%, and 3%, respectively,9 by MY0911 consensus primer PCR.10

The cervical cancer death rate in women aged 20–34 years in England and Wales rose from 0·7 per 100000 (163 deaths) in 1963–67 to 2·2 per 100000 (605 deaths) in 1983–87. A large rise in HPV prevalence is presumably the major cause of this three-fold increase, although other infections and cofactors, particularly oral contraceptive use, may have contributed.8, 11, 12 However, in women older than 50 years, the death rate fell by almost a third over the same period, continuing a long-term decline that began earlier in the 20th century. These opposite underlying trends, and the fact that screening young women cannot affect their death rate in old age until about 50 years later, complicate interpretation of the reduction in mortality since inception of the screening programme. The number of lives saved by screening can be estimated only from a projection of future numbers of cervical cancer deaths that would have occurred with no screening; and a credible projection must be derived from analysis of past changes in mortality.

Cervical screening expanded steadily in Britain since the 1960s. Women were often screened when they attended obstetric or family planning clinics, and sometimes when visiting their general practitioner. This opportunistic screening led to inefficient focus on young women. In 1981, 1·7 million of the 3 million Papanicolaou smears taken in England and Wales were from women younger than 35 years:13 at this time, 5·3 million women were aged 20–34 years. A small proportion of smears were from women younger than 20 years and some were repeats of abnormal or inadequate smears. Nevertheless, these data suggest that by 1981, about a quarter of all women in England and Wales aged 20–34 years were being screened each year. However, some were screened annually whereas others were missed, and many older women were never screened. A computerised local recall system was therefore initiated in 1985, and in 1988 the national programme was launched. All women aged 20–64 (recently changed to 25–64) years receive an invitation every 3–5 years, according to local policy, and those with abnormal smears are followed up to ensure that they are retested. Screening coverage rose sharply from 40–60% in 1989 to 80% in 1992,14 and by 2001, 85% of women younger than 60 years and 78% of women 60–64 years of age had been screened within the past 5 years.15

Comparison of death rates since 1988 with projected rates in the absence of screening suggests that by 1997, about 1300 lives per year were already being saved.16 In this analysis, we have extended these projections to estimate the eventual effect of the national screening programme on future mortality. British cervical cancer mortality trends are especially informative because cervical cancer was reliably distinguished from uterine cancer in mortality statistics by the early 1950s. Between 30% and 90% of uterine cancer death certificates did not specify the site (cervix or corpus) in most European countries until the 1960s or later.1 However, in women younger than 45 years, almost all fatal uterine cancers originate in the cervix. Reliable international comparisons up to age 44 years can therefore be based on mortality from all cancers of the uterus.

Section snippets

Methods

We obtained death rates for all cancers of the uterus up to age 44 years for the 38 countries in Europe (including the former USSR), North America, and Australasia with populations exceeding 1 million for which complete data were available in the WHO database for 1985–99.17

Every year, the Office for National Statistics publishes cervical cancer death rates in England and Wales. We grouped data from 1953–57 to 1983–87 into 5-year periods and 5-year age-groups from 20–24 to 80–84 years, to give

Role of the funding source

The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Birth cohort trends

figure 1 and the table show cervical cancer death rates between 1953 and 1987 in England and Wales, before the national screening programme began. Actual rates are very different in the various periods, but the shape of the age-distribution is roughly the same in every birth cohort born around or before 1952 (figure 1), rising sharply to age 45 years and remaining virtually constant from age 50 to 85 years. A simple age and birth-cohort model gives a reasonably good fit (table). The dotted line

Discussion

Cervical cancer mortality projections based on birth-cohort trends suggest that the introduction of effective national screening prevented an epidemic that would—without any screening—have culminated in British rates being among the highest in the world at all ages, with about 11000 invasive cancers and 5500 deaths per year in England and Wales by 2030. Inclusion of Scotland and Northern Ireland would inflate our projection of the eventual maximum to at least 6000 deaths and 12000 cases in the

References (42)

  • AB Miller

    The (in)efficiency of cervical screening in Europe

    Eur J Cancer

    (2002)
  • P Sasieni et al.

    Changing rates of adenocarcinoma and adenosquamous carcinoma of the cervix in England

    Lancet

    (2001)
  • AE Raffle et al.

    Detection rates for abnormal cervical smears: what are we screening for?

    Lancet

    (1995)
  • DM Parkin et al.

    The impact of screening on the incidence of cervical cancer in England and Wales

    Br J Obstet Gynaecol

    (1985)
  • JMM Walboomers et al.

    Human papillomavirus is a necessary cause of invasive cervical cancer worldwide

    J Pathol

    (1999)
  • I Simms et al.

    Epidemiology of genital warts in England and Wales: 1971 to 1994

    Genitourin Med

    (1997)
  • HC Kitchener et al.

    The ARTISTIC Trial: A Randomised Trial In Screening To Improve Cytology

    (Feb 20–26, 2004)
  • JM Deacon et al.

    Sexual behaviour and smoking as determinants of cervical HPV infection and of CIN3 among those infected: a case-control study nested within the Manchester cohort

    Br J Cancer

    (2000)
  • Peto J, Gilham C, Deacon J, et al. Cervical HPV infection and neoplasia in a large population based prospective study:...
  • CL Peyton et al.

    Comparison of PCR- and hybrid capture-based human papillomavirus detection systems using multiple cervical specimen collection strategies

    J Clin Microbiol

    (1998)
  • Office for National Statistics

    Social trends, no 14

    (1984)
  • Cited by (644)

    • Topical delivery of drugs in the treatment of high-grade cervical squamous intraepithelial lesions: A meta-analysis

      2024, European Journal of Obstetrics and Gynecology and Reproductive Biology
    • Principles of cancer screening

      2024, Surgery (United Kingdom)
    View all citing articles on Scopus
    View full text