ABSTRACT
BACKGROUND: A previous meta-analysis of 4 Swedish randomized controlled trials demonstrated a reduction in breast cancer mortality but not overall mortality with screening mammography.1 Critics raised concerns about the study methods and validity of the results. This article reported the results from a new meta-analysis of the Swedish studies with longer follow-up. The authors also defined and defended their methods in detail.
POPULATION STUDIED: The 4 Swedish trials were combined to include a study group of 247,010 women aged 40 to 74 years. A total of 129,750 women received mammograpy and 117,260 were controls. Some trials randomized individuals, whereas others used a quasirandomization method in which clusters of people were randomized by day of birth or geographic area. Women were excluded if they had a diagnosis of invasive cancer prior to randomization. Overall, 4001 women younger than 40 years and 14,959 women older than 75 years were excluded.
STUDY DESIGN AND VALIDITY: The meta-analysis combined the results of 4 Swedish randomized control trials and extended the follow-up from previously reported data. The primary end point was breast cancer mortality as recorded in the Swedish Cause of Death Registry. The original files were obtained and records were linked to the 6 regional oncologic centers as well as the Swedish Cause of Death Register. The end date of follow up was December 31, 1996. Two statistical models were developed by the researchers to allow for better comparison between different trials.
OUTCOMES MEASURED: The primary outcome measured was long-term effects on mortality, including age-specific and trial-specific effects. Mortality was separated into breast cancer mortality, cumulative breast cancer mortality by age group, and total mortality.
RESULTS: Using the “evaluation model,” breast cancer mortality was reduced 21% with invitation to mammography screening (relative risk [RR] = 0.79, 95% confidence interval [CI], 0.70–0.89). This risk was similar across all age groups. The cumulative breast cancer mortality per 100,000 women was decreased in each trial and at each age of entry when screening was started. The absolute reduction for all women aged 40 to 74 years at entry was 136 per 100,000 at 18 years after randomization. Overall, total mortality was not affected by mammography screening (RR = 0.98, 95% CI, 0.96–1.00), but the 50- to 59-year and 60- to 69-year age groups did show a mildly significant benefit (RR = 0.95, 95% CI, 0.92–0.98, and RR = 0.94, 95% CI, 0.91–0.97, respectively).
This study confirms screening mammography’s role in the reduction of breast cancer-related deaths. The effects of mammography on breast cancer mortality reduction have persisted after long-term follow-up of previously supportive data. These effects are age dependent and seem to benefit women aged 55 to 69 years. What this analysis failed to demonstrate, however, was a significant reduction in overall mortality. That is, despite being diagnosed with breast cancer through mammography, these women still have a similar risk of dying from any cause compared with those who were not screened. Until legitimate data are presented that dispute the long-term benefits of breast cancer screening, however, mammography persists as a valuable tool in reducing mortality related to breast cancer. The National Cancer Institute continues to recommend screening mammograms every 1 to 2 years starting at age 40.2