ObjectivesObservational cohort studies are used to evaluate the effectiveness of screening mammography in women offered screening. Because screening mammography has no effect on causes of death other than breast cancer (BC), cohort studies should show reductions in the risk of BC death substantially greater than possible reductions in the risk of all-cause death. We assessed the risk of BC deaths and of all-cause (or of nonBC) deaths associated with screening mammography attendance reported in cohort studies. Study Design and SettingCohort studies published from 2002 to 2022 on women invited to screening mammography were searched in PubMed, Web of Sciences, Scopus, and in review articles. Random effect meta-analyses were performed using relative risks (RRs) of death between women who attended screening compared to women who did not attend screening. ResultsEighteen cohort studies were identified, nine that reported RRs of BC deaths only, five that reported RRs of all-cause deaths only, and four that reported RRs for both BC deaths and all-cause deaths. The latter four cohort studies reported 12–76 times more all-cause deaths than BC deaths. The random-effect summary of RR for BC mortality in screening attendees vs nonattendees was 0.55 (95% CI: 0.50–0.60) in 13 cohort studies. The summary of RR for all-cause mortality was 0.54 (0.50–0.58) in 10 cohort studies. In the four cohort studies that evaluated both outcomes, the summary of RRs were 0.63 (0.43–0.83) for BC mortality and of 0.54 (0.44–0.64) for all-cause mortality. ConclusionThe similar relative reductions in breast- and all-cause (or nonBC) mortality indicates that screening mammography attendance is an indicator of characteristics associated with a lower risk of dying from any cause, including from BC, which observational studies have falsely interpreted as a screening effect.