Ovarian cancer and diet: from nutrients to lifestyle

Abstract Ovarian cancer is the 6th most common cancer in women worldwide and mortality from this cancer is high, because early diagnosis is difficult (Sankaranarayanan et al. 2006). Thus, identification of modifiable factors contributing to its aetiology is important in reducing the burden of a very fatal form of women’s cancer. The overall aim of this thesis was to investigate the association between diet and ovarian cancer risk within the context of a framework ‘from nutrients to lifestyle’. The study examined dietary and lifestyle factors that had not previously been investigated comprehensively in four main areas: nutrients (retinol, beta-carotene, vitamins E, C, and B), foods (fruit, vegetables, dairy products, eggs, meat and liver), diet patterns and lifestyle (smoking, alcohol, and body-mass-index). The study used the data from two population-based case-control studies of women aged 18-79 years conducted in Australia 10 years apart; the Survey of Women’s Health (SWH, 1990-1993) involves 683 cases and 777 controls, and the Australian Ovarian Cancer Study (AOCS, 2002-2005) includes 1329 cases and 1397 controls. Cases were recruited from gynaecological oncology treatment centres and controls were selected at random from the electoral roll. Detailed information on non-dietary risk factors was obtained using a questionnaire and dietary information was obtained via a semi-quantitative food frequency questionnaire. Multivariable-adjusted odds ratios (ORs) (adjusting for age, parity, oral contraceptive use, education, and energy intake) for ovarian cancer risk were estimated separately for each study with logistic regression modelling and weighted pooled risk estimates for the two studies were calculated using fixed-effects models. Principal components analysis of around 40 food groups was performed to identify dietary patterns in each study separately. There was an increased risk of ovarian cancer associated with retinol intake in both studies (combined OR for the highest vs. lowest quartile =1.42, 95%CI 1.19-1.69), while intake of beta-carotene was inversely related to cancer risk (combined OR=0.80, 95%CI 0.67-0.96). The associations between retinol and beta-carotene and risk appeared to be independent. Liver was also associated with an increased risk; however this association seemed to be explained by the high levels of retinol in liver. Dairy products and eggs are other good sources of retinol, but no clear relations were seen and additional adjustment for saturated fat further attenuated the associations. High vitamin E intake (combined OR=0.73, 95%CI 0.61-0.87) was associated with a decreased risk, but no overall association was seen for vitamin C (combined OR=1.06, 95%CI 0.89-1.27). It is noteworthy that vitamin C seemed to be more beneficial for current smokers than for never/past smokers. For the B vitamins, a significant inverse association was apparent only for niacin intake (combined OR=0.69, 95%CI 0.58-0.82). While there was no association between cancer risk and sources of niacin such as total meat and red meat, other niacin-rich foods such as poultry (combined OR=0.77, 95%CI 0.66-0.89) and fish (combined OR=0.83, 95%CI 0.71-0.97) were significantly inversely associated with risk. In contrast to poultry and fish, high consumption of processed meat was associated with a 24% increase in risk (combined OR=1.24, 95%CI 1.06-1.45). Total fruit (combined OR=0.75, 95%CI 0.60-0.94) and total vegetables (combined OR=0.69, 95%CI 0.52-0.92), specifically cruciferous vegetables (combined OR=0.79, 95%CI 0.63-0.98) and green leafy vegetables (combined OR=0.79, 95%CI 0.67-0.94), were associated with a modestly decreased risk, whereas the inverse association between red/yellow vegetables and risk did not quite reach statistical significance (combined OR=0.84, 95%CI 0.64-1.08). High fruit intake was, like vitamin C intake, somewhat more beneficial for current/past smokers than for never smokers. Exclusion of women who took dietary supplements did not substantially change the observed associations between nutrients and risk. Three major eating patterns were identified: ‘snacks and alcohol’, ‘fruit and vegetable’, and ‘meat and fat’. Significant inverse associations between the snacks and alcohol pattern and risk were attenuated after further adjustment for white/red wine intake in both studies; it thus appeared that the observed association was at least partly due to wine intake. A significant association between the fruit and vegetable pattern and risk was seen only in the more recent study. A diet characterized by high meat and fat was associated with an increased risk of ovarian cancer, although the observed association was stronger in the SWH There was no evidence that the associations between diet pattern and cancer risk varied by women’s lifestyle. Although there was some variation in the analyses stratified by the histologic subtype of ovarian cancer, no consistent patterns of variation were observed for either study. These findings provide additional evidence that a healthy diet defined by high intake of fruit, vegetables, particularly cruciferous and green leafy vegetables, white meat and fish and low in meat and fat, especially processed meats might be beneficial against ovarian cancer