Adiposity, exercise, diet, and outcomes after breast cancer diagnosis: World Cancer Research Fund (WCRF)/American Institute for Cancer Research (AICR) systematic literature review and meta-analysis.


Session type:

Margarita Cariolou1, Katia Balducci1, Nerea Becerra-Tomás1, Leila Abar1, Dagfinn Aune1, Georgios Markozannes1, Neesha Nanu1, Ana Rita Vieira1, Edward L Giovannucci2, Marc J Gunter3, Alan Jackson4, Ellen Kampman2, Viv Lund2, Kate Allen5, Nigel Brockton6, Helen Crocker5, Daphne Katsikioti5, Deirdre McGinley-Gieser6, Panagiota Mitrou5, Martin Wiseman5, Anne McTiernan7, Steven Clinton2, Teresa Norat1, Konstantinos K Tsilidis1, Doris S.M Chan1
1Imperial College London, 2Other, 3International Agency for Research on Cancer (IARC), 4University of Southampton, 5World Cancer Research Fund International, 6American Institute for Cancer Research, 7Fred Hutchinson Cancer Research Center



Despite improvements in breast cancer therapy and survival, evidence-based cancer survivorship-specific recommendations regarding diet, exercise and weight management have not been established. Our aim was to summarise peer-reviewed literature since the previous WCRF/AICR 2014 Continuous Update Project report and evaluate what the latest evidence shows.


We searched PubMed and Embase until 31/12/2018 for randomised controlled trials (RCTs) or longitudinal observational studies. We conducted linear and non-linear dose-response meta-analyses and investigated between-study heterogeneity and small-study effects. An Expert Panel judged the evidence using the pre-defined WCRF/AICR survivorship criteria.


Overall, 332 publications were identified. Meta-analyses were performed only for observational studies, as very few relevant RCTs were identified. Each 5 kg/m2 BMI increment was associated with a 7% (95%CI:4%-9%, I2=58%, studies=55) and 9% (95%CI:5%-13%, I2=61%, studies=39), higher risk of all-cause, breast cancer-specific mortality, respectively. The risk of second primary cancer development was 14% higher (95%CI:7%-21%, I2=5%, studies=9). Evidence was graded as “strong” to support a probable causal relationship. Each 10 MET-h/week increment of recreational physical activity was associated with a 16% and 9% lower risk of all-cause (95%CI:8%-24%, I2=84%, studies=7) and breast cancer-specific mortality (95%CI:2%-16%, I2=50%, studies=6), respectively. Non-linear analyses revealed risk reductions for all-cause and breast cancer-specific mortality with increasing exercise, reaching a plateau at ~20 MET-h/week. Evidence was graded as “limited-suggestive”. A 10 g/day increment of fibre intake was associated with a 14% lower all-cause mortality (95%CI:1%-26%, I2=0%, studies=3). Every 10 nmol/L increment in serum 25(OH)D concentration was associated with an 8% and 6% lower risk of all-cause (95%CI:5%-11%, I2=0%, studies=5) and breast cancer-specific mortality (95%CI:1%-10%, I2=24%, studies=5); this evidence was judged as “limited-suggestive”. Our analyses did not detect associations between any other dietary exposures and breast cancer prognosis, and these thus received a “limited-no-conclusion” grading.


Our findings indicate that adiposity (“strong” evidence) and physical activity (“limited-suggestive” evidence) could influence breast cancer prognosis. However, residual confounding, reverse causation and other potential biases must be considered. Large, well-designed observational studies with repeated measurements, and relevant RCTs with high compliance are necessary to clarify the impact of body composition, exercise and diet on outcomes after breast cancer diagnosis.

Impact statement