THE GOTHENBURG BREAST SCREENING TRIAL - FIRST RESULTS ON MORTALITY, INCIDENCE, AND MADE OF DETECTION FOR WOMEN AGES 39-49 YEARS AT RANDOMIZATION

BACKGROUND, The effect of mammography screening on breast carcinoma mo rtality in women ages < 50 years remains unclear. METHODS. A randomize d trial of invitation to breast carcinoma screening with mammography w as performed in the city of Gothenburg, Sweden. The purpose was to est imate the effect of mammographic screening on breast carcinoma mortali ty in women ages < 50 years. Randomization was initially by day-of-bir th cluster (18% of subjects), and subsequently by individual (82% of s ubjects). Between September 1983 and April 1984, 11,724 women ages 39- 49 years were randomized to the study group. This group was invited to mammographic screening every 18 months. Two-view mammography was used at each screen unless the density of the breast at the previous scree n indicated that single view was adequate. Fourteen thousand two hundr ed and seventeen women in the same age range were randomized to a cont rol group that was not invited to undergo screening until the fifth sc reen of the study group (between 6 and 7 years after randomization). W omen with breast carcinoma diagnosed up to the time immediately after the first screen of the control group were followed for death from bre ast carcinoma until the end of December 1994. RESULTS, A 45% reduction in mortality fi om breast carcinoma was observed in the study group c ompared with the control group (relative risk [RR] = 0.55, P = 0.035, 95% confidence interval [CI], 0.31-0.96). A conservative estimate base d on removal of the tumors detected at the first screen of the control group gave a mortality reduction of 44% (RR = 0.56, P = 0.046, 95% CI , 0.31-0.99). In both cases, the effect was statistically significant. CONCLUSIONS. Mammographic screening can reduce mortality from breast carcinoma in women ages < 50 years. The mortality reduction can be sub stantial if high quality mammography is used and an 18-month interscre ening interval is strictly adhered to. [See editorial on pages 2035-9, this issue.] (C) 1997 American Cancer Society.