In this study we found little evidence that the dietary TAC had an impact on ovarian cancer risk, based on antioxidant content values from the USDA ORAC database and the FRAP-based Antioxidant Food Database. In line with these findings, most individual micronutrients evaluated were also not found to be associated with ovarian cancer, with the exception of selenium. We found a statistically significant 60 % decreased risk of ovarian cancer for women in the highest tertile of selenium intake from food sources compared to the lowest. In contrast, we observed significant increased risk for users of all types of antioxidant supplements examined, compared to non-users. However, no increased risks were observed for the total amount of antioxidant intake when supplement amounts were combined with dietary intake.
Similar to our results, most studies investigating dietary vitamin C and ovarian cancer risk report no association [14–23]. However, some studies have found significant decreased risks [24–28]. The only two cohort studies evaluating vitamin C reported no association, including an analysis of the Nurse’s Health Study, which identified 301 ovarian cancer cases during a follow-up of 16 years . Additionally, a meta-analysis of case-controls studies found no association, reporting a pooled OR of 0.98 (95 % CI: 0.95-1.01) for additional daily intake of 30 mg vitamin C . Only one other study has investigated supplement use, and in contrast to our results found a significantly decreased risk of almost 50 %, and also reported a decreased risk for supplements combined with dietary intake .
We also found no association with vitamin E intake. Similarly, of all previously conducted studies evaluating vitamin E and ovarian cancer risk [14–18, 21–24, 26, 29–31], most found no association. Only a few case–control studies reported decreased risks [17, 24, 26, 31]. In contrast to our findings, the only two studies investigating supplement use reported strong inverse associations [18, 30].
Consistent with our findings, several cohort studies [14, 15, 22, 23] and some case–control studies [16, 18, 32] found no association with beta-carotene from food sources. Also, a meta-analysis of case–control studies reported no association (pooled OR: 0.98; 95 % CI: 0.96-0.99) for additional daily intake of 500mcg beta-carotene . However, an inverse association with beta-carotene has been reported in other studies [17, 26, 27, 29, 31], and non-significant decreased risks have also been reported [21, 33]. To our knowledge, our study is the first to investigate beta-carotene supplement use and ovarian cancer risk.
In support of our findings, a nested case–control study investigating the serum selenium status of ovarian cancer patients compared to healthy controls found that increasing serum selenium was associated with a decreased risk of ovarian cancer (OR: 0.23; 95%CI: 0.1-0.9 for highest tertile compared to lowest; p for trend = 0.02) . Das and Ma (1986) reported similar findings, with a reverse correlation between serum selenium concentration and ovarian cancer incidence. Additionally, Sieja (1998) found ovarian cancer patients had significantly lower serum selenium concentrations compared to controls (p < 0.05), approaching critical levels, particularly during chemotherapy treatment. However, other studies investigating selenium intake from dietary and supplement sources [14, 18, 21], including an analysis of the Women’s Health Initiative , found no association with ovarian cancer risk. Tung et al.  assessed selenium intake among a multiethnic group of patients from Hawaii and Los Angeles and Fleischauer et al.  had a smaller number of cases (n = 169), and used both community and hospital-based controls. Differences in the study populations may partially explain the conflicting results.
Numerous studies have indicated that selenium may have anticancer properties. Human studies have shown a decrease in the incidence of prostate, lung, and colorectal cancers [37, 38]. Additionally, animal studies have shown that, at high doses, selenium compounds reduce tumor yields, inhibit cell growth, and stimulate programmed cell death in cultures. In ovarian cancer cells, it has been demonstrated that selenium compounds inhibit the synthesis of nucleic acids (Rzaeva, 1985). Selenoprotein deficiency has also been found to be present in certain types of cancer. The proposed mechanism by which selenium may have a protective effect on cancer is mainly through its antioxidant properties. Selenium provides antioxidant protection against the effect of reactive oxygen species on cancer initiation and promotion . Geographic variation in selenium concentrations exists, however no study has investigated whether ovarian cancer risk differs in areas with varying selenium concentrations.
Our conflicting findings of a protective effect associated with dietary selenium intake versus an increased risk associated with supplement intake are unclear. Of note, the increased risk observed with selenium supplements only reached statistical significance after further adjustment for physical activity and smoking status. It is possible that related unmeasured confounders may be influencing the association. Studies of supplement intake and relation with cancer risk have been conflicting. A recent systematic review concludes that evidence does not support selenium supplementation in the prevention of cancer . Other reports have suggested that certain supplements may increase the risk of cancer, as has been recently observed with vitamin E supplementation and prostate cancer risk .
Only one other study has developed a total antioxidant score in ovarian cancer patients, using data from the Teacher’s Health Study in California . However, this study only included antioxidant values from fruits and vegetables, identified from variable literature sources, and did not assess supplement use. Similar to our results, no association was observed.
This study may be subject to the limitations of case–control studies, such as recall bias and selection bias. Particularly with our finding of increased risk with supplements, it is possible that cases may have reported use differently than controls. We also had a low participation rate. Potential non-participation bias was assessed by comparing characteristics of women who participated in this study to cases who did not in the New Jersey State Cancer Registry during the study time period. Cases that consented tended to be younger; however the racial/ethnic distribution and distributions by histology, stage and grade were similar. We did not have information on women who did not end up participating as controls. However, the distribution of risk factors in this study is similar to that reported in other studies, which gives us reassurance in the validity of our data.
In conclusion, this study found a strong inverse association of selenium from food sources and ovarian cancer risk, while selenium supplement intake was associated with increased risk. However, we did not find any significant association between TAC intake and ovarian cancer risk. Supplement use in this study was found to be associated with increased ovarian cancer risk for all supplements studied. However, combined amounts of antioxidants from diet and supplement sources were not associated with increased risks. As several reports have recently raised concerns about the safety of vitamin supplements in recent years, including the 2007 WCRF/AICR Report which warned cancer patients and survivors against taking certain supplements , these findings warrant further investigation to better understand the role of selenium from foods and supplements on ovarian cancer risk.