Elsevier

The Lancet

Volume 359, Issue 9310, 16 March 2002, Pages 909-919
The Lancet

Articles
Long-term effects of mammography screening: updated overview of the Swedish randomised trials

https://doi.org/10.1016/S0140-6736(02)08020-0Get rights and content

Summary

Background

There has been much debate about the value of screening mammography. Here we update the overview of the Swedish randomised controlled trials on mammography screening up to and including 1996. The Kopparberg part of the Two-County trial was not available for the overview, but the continuation of the Malmö trial (MMST II) has been added. The article also contains basic data from the trials that have not been presented before.

Methods

The trials (n=247 010, invited group 129 750, control group 117 260) have been followed up by record linkage to the Swedish Cancer and Cause of Death Registers. The relative risks (RR) for breast cancer death and mortality were calculated for the invited and the control groups. The trial-specific as well as the age-specific effects were analysed. RRs were calculated by the density method, with total person-time experience of the cohort by time interval of follow-up as a basis for estimating mortality rates. We calculated weighted RRs and 95% CI with the Mantel-Haenszel procedure.

Findings

The median trial time–the time from randomisation until the first round was completed for the control group or if the control group was not invited, until end of follow-up–was 6.5 years (range 3.0-18.1). The median follow-up time, the time from randomisation, to the end of follow-up, was 15.8 years (5.8-20.2). There were 511 breast cancer deaths in 1864 770 women-years in the invited groups and 584 breast cancer deaths in 1 688 440 women-years in the control groups, a significant 21% reduction in breast cancer mortality (RR=0.79, 95% CI 0.70-0.89). The reduction was greatest in the age group 60-69 years at entry (33%). Looking at 5-year age groups, there were statistically significant effects in the age groups 55-59, 60-64, and 65-69 years (RR=0.76, 0.68, and 0.69, respectively). There was a small effect in women 50-54 years at randomisation (RR=0.95). The benefit in terms of cumulative breast cancer mortality started to emerge at about 4 years after randomisation and continued to increase to about 10 years. Thereafter the benefit in absolute terms was maintained throughout the period of observation. The age-adjusted relative risk for the total mortality was 0.98 (0.96-1.00).

Interpretation

The advantageous effect of breast screening on breast cancer mortality persists after long-term follow-up. The recent criticism against the Swedish randomised controlled trials is misleading and scientifically unfounded.

Introduction

Service-screening for breast cancer occurs in several countries with the aim to decrease breast cancer mortality. The scientific basis for these programmes are the randomised screening trials. There are seven such studies, four from Sweden. The Swedish trials have a similar design: they were all population-based and compared invitation to breast screening with mammography alone versus no invitation. These Swedish trials differed from the other trials (the Health Insurance Plan of Greater New York, the Edinburgh trial, and the Canadian National Breast Screening Study), which all evaluated mammography combined with breast self-examination, clinical breast examination, or both. Moreover, the New York and the Canadian trials were not population-based.

The Swedish Cancer Society initiated an overview of the Swedish trials in the late 1980s. The objective was to validate the results from the individual trials through a method that was common to all trials, including a blind review of all deaths among breast cancer cases by an independent endpoint committee. Another objective was to increase the statistical power. The first results of the overview were published in 19931 and an-update focusing on the age group 40-49 in 1997.2 Concerns raised about the validity of the results from the trials.3 include inappropriate exclusions, poor randomisation, and the excess total mortality in women invited to screening.

Our aim here was to extend the follow-up and to analyse the age-specific and trial-specific effects on breast cancer mortality, to describe the randomisation procedures in more detail, and to assess the quality of the cluster randomisation used in Östergötland. The Kopparberg part of the Two-County (WE) trial was not available for analysis but the continuation of the Malmö trial (MMST II) was added.

Section snippets

Methods

Details of the Swedish mammography screening trials have been published (table 1), and are summarised below.

Malmö trial

The Malmö Mammographic Screening Trial (MMST) included women in the city of Malmö from October, 1976. In the first part (MMST I) women born between 1908 and 1932 were randomised with individual stratification by year of birth. Women were invited to screen-film mammography alone, in the first two rounds with two views (craniocaudal and oblique) and in subsequent rounds with either two views or the oblique view alone depending on the parenchymal pattern. A single oblique view was used for women

Number of women randomised, trial time, and follow-up time

Our analysis was based on the follow-up of 247 010 women, 129 750 of whom were invited to mammography screening and 117 260 of whom were controls. 4001 women below the age of 40 and 14 959 women from Östergötland aged 75 and above were excluded. Age distribution by trial is in table 2

Median trial time and range are in table 3. Trial time was defined as time from date of randomisation until the control groups completed the first round of screening. In trials in which the control groups were not

Discussion

Our aim was to elucidate some issues that have been raised in recent reviews of the Swedish trials. In addition, we wanted to assess the long-term effects on mortality, including age-specific and trial-specific effects. Our latest overview, which is unbiased and unconfounded for study design, confirms and extends previous results. Our main observation was that the benefit of screening was maintained several years after the trials had been closed. In general, the benefit in absolute terms

References (15)

There are more references available in the full text version of this article.

Cited by (0)

View full text