Soft Drinks And Risk Of Obesity Related Cancers

Allison HODGE, Cancer Council Victoria, Australia
BASSETT J. 1 , ENGLISH D. 1, 2

1 Cancer Epidmiology Centre, Cancer Council Victoria, Melbourne, Australia
2 1Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, The University of Melbourne, Melbourne, Australia.

Purpose: To investigate the association between sugar-sweetened (SSD) or diet soft drinks (DSD, artificially sweetened) and obesity related cancers (ORC) in the Melbourne Collaborative Cohort Study (MCCS), in particular the role of body size.
Methods: The MCCS recruited 41,514 adults, aged 40 to 69 years in 1990-94 to identify risk factors for cancer.  A Food Frequency Questionnaire specifically developed for the study was used to collect baseline data on the frequency of consumption of SSD and DSD over the previous 12 months.  Data on incident ORC (oesophagus (adenocarcinoma), gastric cardia, pancreas, colon, rectum, post-menopausal breast (diagnosis > 55 years), endometrium, kidney, aggressive prostate (Gleason score>7) and ovarian) and mortality was obtained by data linkage to the end of 2012.  Cox’s proportional hazard models were fitted, with age as the time metric, to estimate hazards ratios and 95% confidence intervals in 38,621 people free of pre-baseline cancer. 
Results: Increasing intake of SSD and DSD was associated with obesity.  More than 5% of people with a history of diabetes, angina or heart attack (DM_CVD) consumed DSD > 2/day compared with just over 2% in those with no such history.  In a model including 35,435 people with no DM_CVD, adjusting for sex, ethnicity, smoking, physical activity, alcohol consumption, Mediterranean Diet Score, socioeconomic status, and family history of cancer the HR and 95% CI for SSD >2/day versus <1/day was 1.06 (0.93, 1.21) and for DSD 1.27 (1.09, 1.48); for the 3,186 people with DM_CVD, the corresponding HRs were 1.52 (1.08, 2.13) and 1.46 (1.07, 1.98) respectively.  Additional adjustment for BMI changed the HRs minimally.
Conclusions:  In people with no DM_CVD, DSD were associated with an increased risk of ORC, while in those with DM_CVD, positive associations were seen for both DSD and SSD. These associations did not appear to be explained by obesity.