Diet, obesity, lifestyle and cancer prevention: Epidemiologic perspectives Graham A Colditz, MD DrPH Niess-Gain Professor Chief, November, 2017
Outline Review evidence on contribution of diet, obesity, lifestyle and in particualr, increasing burden of obesity on cancer Identify: Issues Gaps in knowledge Opportunities
Lifestyle: high income countries Cause % cancer caused Smoking 33 Overweight/obesity 20 Diet 5 Lack of exercise 5 Occupation 5 Viruses 5-7 Family history 5 Alcohol 3 UV/ionizing radiation 2 Magnitude possible reduction Time (yrs) Reproductive 3 Pollution 2 Colditz et al, Sci Transl Med 2012
Lifestyle: high income countries Cause % cancer caused Magnitude possible reduction Smoking 33 75% Overweight/obesity 20 50% Diet 5 50% Lack of exercise 5 85% Occupation 5 50% Viruses 5-7 100% Family history 5 50% Alcohol 3 50% UV/ionizing radiation 2 50% Time (yrs) Reproductive 3 0 Pollution 2 0
Lifestyle: high income countries Cause % cancer caused Magnitude possible reduction Time (yrs) Smoking 33 75% 10-20 Overweight/obesity 20 50% 2-20 Diet 5 50% 5-20 Lack of exercise 5 85% 5-20 Occupation 5 50% 20-40 Viruses 5-7 100% 20-40 Family history 5 50% 2-10 Alcohol 3 50% 5-20 UV/ionizing radiation 2 50% 2-10 Reproductive 3 0 N/A Pollution 2 0 N/A
Time course: lung & total mortality Current smoker: continuing
Medical interventions proven to prevent cancer: high-income Intervention Target Magnitude of reduction Time (yrs) Aspirin Colon mortality 40% 20+ SERMs Breast incidence 40-50% 5+ Salpingo oophorectomy Screening for colorectal cancer Familial breast cancer 50% 3+ Colon cancer mortality 30-40% 10 Viruses - HPV Cervical cancer incidence 50-100% 20+ - Hep B Liver cancer incidence 70-100% 20+ Mammography Breast cancer mortality 30% 10-20 Serial CT lung Lung cancer mortality 20% 6+ Colditz et al, Sci Transl Med 2012
Histologically confirmed cervical abnormalities, Vic, Australia CIN2, CIN3 Brotherton et al MJA 2016
Pancreatic cancer Yachida Nature 2010; Luebeck EG. Nature 2010
Summary of evidence: Adolescent exposures relation to risk of BBD and breast cancer Lifestyle Relative Risk BBD Breast Cancer (premenopausal) Alcohol Peak Growth Velocity Height Nuts Fiber Carotenoids Vegetable protein Family history Physical activity Fruit and veggies Colditz Bohlke Berkey Br Ca Res Treat 2015, Colditz & Bohlke 2014
Alcohol intake, ages 18-22, incident proliferative BBD Alcohol intake (grams/ day) Cases (678) Personyear RR (95% CI) None 155 64,827 1.0 reference 0.1-4.9 193 78,365 1.11 (0.89, 1.38) 5.0-14.9 236 88,310 1.36 (1.09, 1.69) >15 30 9519 1.35 (1.01, 1.81) p, trend <0.01 Liu et al. Pediatrics, 2012
Alcohol before first pregnancy, NHSII Liu, Colditz, Tamimi JNCI 2013
Adolescent fiber & proliferative BBD: NHSII Su et al. Cancer Causes Control 2010
Are we there yet?
Obesity and cancer, time line - 1990s WHO, US Dietary Guidelines, adopt common cut points 2002 IARC report 2008 et seq WCRF report 2016 update of IARC report
IARC 2002 and Calle 2003 Review of evidence on weight obesity and physical activity in relation to cancer Calle: ACS cohort published after the IARC review panel
IARC 2002 Sufficient evidence in humans for cancerpreventive effect of avoidance of weight gain for cancers of the colon, esophagus (adenocarcinoma), kidney (renal cell), breast (postmenopausal), and corpus uteri Translate: Obesity causes cancer IACR Handbooks of Cancer Prevention Vol 6, 2002
Review of Evidence, IARC 2002 Level of Evidence Obesity Risk Increase Associated with Obesity Small (RR 1.09-1.34) Moderate (RR 1.35-1.99) Large (RR 2.0-4.9) Very Large (RR 5.0+) Convincing Colon Breast Esophagus Uterus Kidney Probable
Calle et al 2003
Workgroup reviewed measures of adiposity; animal models; mechanisms; and epidemiologic evidence. Concluded lack of body fatness lowers risk, or obesity causes cancer. NEJM August 25, 2016
Evidence evolving From only a couple of prospective cohorts in 2002, adding ACS mortality in 2003 Now evidence from 30 to 50 or more prospective cohorts Pooled analysis of individual participant data from studies addressing BMI and less common cancers
Why prospective studies and pooled data Measure adiposity and risk of subsequent cancer Avoid weight change due to disease
Individual participant data pooled analysis IPD meta-analyses can improve the quality of data and the type of analyses that can be done and produce more reliable results (Stewart and Tierney 2002). For this reason they are considered to be a gold standard of systematic review. In fact, IPD meta-analyses have produced definitive answers to clinical questions, which might not have been obtained from summary data. Cochrane Handbook Ch 18 and IPD methods
GI Gastric cardia Liver Pancreas Gall bladder
Pancreas More than 20 prospective studies and casecontrol studies indicating a positive doseresponse relation. Observed in the large majority of studies and in both genders. Compared to normal weigh, the RR for overweight was 1.18 (1.03-1.36) and for obesity 1.47 (1.23-1.75), estimated from pooled analysis of 14 cohorts [Genkinger 2011].
Baseline BMI 2135 cases During 846,340 py Forest plot of RR for BMI >30 vs 21-22.9 Genkinhger et al 2011
BMI in early adulthood
Relative risk of ovarian cancer by BMI and HT use Never use HT Ever use Collaborative Group on Epidemiological Studies of Ovarian Cancer (2012) Ovarian Cancer and Body Size: Individual Participant Meta-Analysis Including 25,157 Women with Ovarian Cancer from 47 Epidemiological Studies. PLoS Med 9(4): e1001200. doi:10.1371/journal.pmed.1001200 http://journals.plos.org/plosmedicine/article?id=info:doi/10.1371/journal.pmed.1001200
Evidence, 2016Risk Increase Associated with Obesity Level of Evidence Small (RR 1.09-1.34) Moderate (RR 1.35-1.99) Large (RR 2.0-4.9) Very Large (RR 5.0+) Convincing Probable Ovary Colon Breast Esophagus Thyroid Gastric cardia Kidney Uterus Liver Gall bladder Pancreas Meningioma Multiple myeloma Male breast Fatal prost. Diffuse Large B- cell lymphoma
Time course obesity Increase in childhood adiposity
cancer Pancreatic US incidence 1992 to 2014 Age 20 to 49 Rising incidence 0.9% per year (significant)
cancer Colorectal US incidence 1992 to 2014 Age 20 to 49 Rising incidence 1.7% per year (significant)
Childhood and early adult adiposity Often consistent with adult adiposity and risk Analysis not always clear Methods, correlated variables, and interpretation Challenges in breast cancer Inverse relation with adiposity at ages 5, 10, before menarche Weight gain increases risk How does childhood adiposity reduce risk for life?
Post-menopausal Breast Cancer Risk Pre-menopausal Breast Cancer Risk - 1 0.80 0.98-1 0.66 0.74 1.0 Premenopausal 1.37 (weight change after menopausal to attained) 1.36 (weight change from 18 to attained) Postmenopausal Age (years) Adiposity +kg +kg +kg 0 10 18 30 47 50 Age (years) Reference: Rosner, B., Eliassen, A. H., Toriola, A. T., Chen, W. Y., Hankinson, S. E., Willett, W. C.,... & Colditz, G. A. (2017). Weight and weight changes in early adulthood and later breast cancer risk. International journal of cancer, 140(9), 2003-2014.
Postmenopausal Breast Cancer Risk Pre-menopausal Breast Cancer Risk - 1 0.80-1 0.98 0.66 0.74 1.36 1.0 Premenopausal 1.37 (weight change after menopausal to attained) 1.36 (weight change from 18 to attained) Postmenopausal Age (years) Adiposit y +kg +kg +kg 0 10 18 30 47 50 Age (years) ER-/PR- 0.73 0.70 Breast Cancer Risk(0.55-0.98) (0.46-1.05) Weight change unrelated to risk RR / 30kg Reference: Rosner, B., Eliassen, A. H., Toriola, A. T., Chen, W. Y., Hankinson, S. E., Willett, W. C.,... & Colditz, G. A. (2017). Weight and weight changes in early adulthood and later breast cancer risk. International journal of cancer, 140(9), 2003-2014.
Top priorities to advance the science obesity and cancer Improved (consistent) approaches to modeling weight gain across life course and cancer risk Quantify benefits of weight loss Measures of adiposity Do we have it right, do measures vary by age; race/ethnicity; region of the world
Top priorities to advance the science diet, lifestyle and cancer Improved (consistent) approaches to modeling exposure in time course of cancer development Measures of diet lifestyle in childhood adolescence Can biomarkers in blood banks replace recall of childhood adolescent exposures?
Behavioral, Social, and Policy interventions that impact Cancer Prevention Intervention Target Type of Ix Evidence review Reduce tobacco use Children and Adolescents Smokers to quit Combined Pharmaco/behavioral Ixs Smoke-free policies Tobacco taxes Surgeon General Increase physical activity Individuals and community norms Urban design Stairs and workplace Surgeon General Reduce Obesity Population School & work environment Physical activity Food & beverage IOM report 2012 Limit alcohol intake Population Taxes WHO Reduce UV exp Children, AYA All of above WHO
Will we all have access to driverless cars? What will our cancer risk be? Wall-e Captain