Elsevier

The Lancet

Volume 361, Issue 9367, 26 April 2003, Pages 1405-1410
The Lancet

Articles
Mammography service screening and mortality in breast cancer patients: 20-year follow-up before and after introduction of screening

https://doi.org/10.1016/S0140-6736(03)13143-1Get rights and content

Summary

Background

The long term effect of mammographic service screening is not well established. We aimed to assess the long-term effect of mammographic screening on death from breast cancer, taking into account potential biases from self-selection, changes in breast cancer incidence, and classification of cause of death.

Methods

We compared deaths from breast cancer diagnosed in the 20 years before screening was introduced (1958–77) with those from breast cancer diagnosed in the 20 years after the introduction of screening (1978–97) in two Swedish counties, in 210 000 women aged 20–69 years. We also compared deaths from all cancers and from all causes in patients diagnosed with breast cancer in the 20 years before and after screening was introduced. In the analysis, data were stratified into age-groups invited for screening (40–69 years) and not invited (20–39 years), and by whether or not the women had actually received screening. We also analysed mortality for the 40–49-year age-group separately.

Findings

The unadjusted risk of death from breast cancer dropped significantly in the second screening period compared with the first in women aged 40–69 years (relative risk [RR] 0·77 [95% CI 0·7–0·85]; p<0·0001). No such decline was seen in 20–39 year olds. After adjustment for age, self-selection bias, and changes in breast-cancer incidence in the 40–69 years age-group, breast-cancer mortality was reduced in women who were screened (0·56; 0·49–0·64 p<0·0001), in those who were not screened (0·84 [0·71–0·99]; p=0·03), and in screened and unscreened women combined (0·59 [0·53–0·66]; p<0·0001). After adjustment for age, self-selection bias, and changes in incidence in the 40–49-year age-group, deaths from breast cancer fell significantly in those who were screened (0·52 [0·4–0·67]; p<0·0001); and in all women, screened and unscreened combined (0·55 [0·44–0·7] p<0·0001) but not in unscreened women (p=0·2). In both 40–69-year and 40–49-year age-groups, reductions in deaths from all cancers and from all-causes in women with breast cancer were consistent with these results.

Interpretation

Taking account of potential biases, changes in clinical practice and changes in the incidence of breast cancer, mammography screening is contributing to substantial reductions in breast cancer mortality in these two Swedish counties.

Introduction

Results from several randomised controlled trials1, 2 and their long-term follow-up3, 4 lend support to the effectiveness of mammographic screening in reducing deaths from breast cancer. However, of continuing importance is whether the reduction in deaths noted in trials can be achieved in routine service screening. Concern has also been expressed that classification of cause of death might be biased in favour of screening,5, 6 despite findings to the contrary.7 Furthermore, although there is favourable evidence from individual trials8, 9 as well as results from meta-analyses,10 some continue to doubt the effectiveness of mammographic screening in women aged 40–49 years.11

We addressed these issues by studying deaths from breast cancers diagnosed in the years before and after the introduction of mammographic screening in two Swedish counties. Since national or regional screening programmes have only been in existence for around 10–12 years, data from long-term follow-up of tumours diagnosed in service screening programmes have not been available. Likewise, estimates of the effect of these programmes on the rate of deaths from breast cancer cannot differentiate between tumours diagnosed in screened and unscreened cohorts.12

Although some have argued that the only proper assessment of the effect of breast-cancer screening is the comparison of all-cause mortality in the entire invited population with that in controls,5, 6 this comparison makes little sense, either from a methodological standpoint, or in terms of the expected effect of a successful screening programme.13, 14, 15 First, screening in itself cannot be expected to affect the rate of deaths from any cause in women who do not develop breast cancer; second, only in breast cancer cases is there the possibility of an error in deciding cause of death. Furthermore, because deaths from breast cancer represent a higher proportion of all deaths in women aged 40–69 years than they do at later ages, a substantial reduction in breast cancer mortality in this age-group ought to be evident in a comparison of all cause mortality.

Thus, we assessed death rates from breast cancer and from other causes in patients with breast cancer by analysis of data from 20 years of observation before and after the introduction of screening to assess the effect of service screening on deaths from breast cancer and all-cause mortality.

Section snippets

Data sources

We used data from the Regional Oncology Centres in Uppsala and Linköping, Sweden, to identify all primary breast cancers diagnosed in women aged 20–39 years, and 40–69 years, in Östergötland and Dalarna (formerly called Kopparberg) counties, for 1958–77, and 1978–97. Vital status of these patients and, if applicable, cause of death were also obtained from the Regional Oncology Centres, and confirmed by the National Cause of Death Register. We used data from Statistics Sweden to estimate the

Results

Table 1 shows the number of breast cancers diagnosed in the two periods by age at diagnosis, and deaths in these women, from breast cancer, all cancers, and all causes. Table 2 shows the overall rates of breast-cancer incidence and mortality from incident cases in the two periods, with unadjusted relative risks and relative risks adjusted for age, increased incidence estimated from those not exposed to screening in the latter period (1978–97), and self-selection bias for attendance or

Discussion

In women aged 40–69 years who were invited to screening, breast cancer mortality fell significantly more than in the those aged 20–39 years. After adjustment for age, changes in incidence, and self-selection bias, there was a significant 44% reduction in breast cancer mortality in women aged 40–69 years who were exposed to screening, and a significant 41% reduction in the overall population of women aged 40–69 years—ie, women both exposed and not exposed to screening. That mortality reductions

References (33)

  • I Andersson et al.

    Reduced breast cancer mortality in women under age 50: updated results from the Malmö Mammographic Screening Program

    J Natl Cancer Inst Monogr

    (1997)
  • RE Hendrick et al.

    Benefit of screening mammography in women aged 40–49: a new meta-analysis of randomized controlled trials

    J Natl Cancer Inst Monogr

    (1997)
  • J Ringash

    Preventive health care, 2001 update: screening mammography among women aged 40–49 years at average risk of breast cancer

    CMAJ

    (2001)
  • RG Blanks et al.

    Effect of NHS breast screening programme on mortality from breast cancer in England and Wales, 1990–8: comparison of observed with predicted mortality

    BMJ

    (2000)
  • L Tabar et al.

    All-cause mortality among breast cancer patients in a screening trial: support for breast cancer mortality as an endpoint

    J Med Screening

    (2002)
  • TR Church et al.

    Re: all-cause mortality in randomized trials of cancer screening

    J Natl Cancer Inst

    (2002)
  • Cited by (0)

    View full text