Epidemiology of Diet and Melanoma — Letter

In a recent review, Yang and colleagues summarized epidemiological studies on diet and melanoma (1). We thank the authors for drawing attention to this topic given the rising incidence of melanoma worldwide. We would like to add several observations regarding the relationship between folate/folic acid and melanoma.

The authors described a meta-analysis of 13 randomized controlled trials, which found no association between folic acid supplementation and melanoma risk (2). However, the results should be interpreted with caution since few melanoma cases (n=126) were included and follow-up was relatively short (mean trial duration was 5.2 years). The authors described another meta-analysis of 3 trials, which found an inverse association between folic acid supplementation and melanoma risk (3). However, even fewer melanoma cases (n=38) were included, and all three studies were included in the former meta-analysis (2). Therefore, although these two meta-analyses were published in the same year with different results, the larger meta-analysis may be more appropriate for interpretation (2).

The authors also mentioned a prospective cohort study from France that reported a positive association between dietary folate intake and overall skin cancer risk (hazard ratio for top versus bottom tertile: 1.79; 95% confidence interval: 1.07–2.99) (4). However, of the 144 skin cancer cases included, only 20 were melanoma. Thus, it is difficult to extrapolate the findings to melanoma specifically, since other types of skin cancers may have driven the association. Additionally, the study could have been strengthened if the analysis had further accounted for citrus consumption, a potential confounder. Citrus fruits are major sources of dietary folate, and we found that citrus consumption was positively associated with melanoma and nonmelanoma skin cancer risk in two prospective studies (5,6). As the direction of confounding is likely positive and biased away from the null, adjustment may attenuate or even nullify any association.

In fact, we recently found a positive association between dietary folate intake and melanoma risk, and the magnitude of effect was attenuated after adjusting for citrus consumption (7). We, however, found that total folate intake (which includes dietary and supplemental intake) was not associated with melanoma risk. Thus, findings regarding dietary folate intake may be due to other components found in folate-rich foods, rather than folate per se. Furocoumarins in citrus fruits may in part be responsible for the positive association due to their potential photocarcinogenic properties.

The association between folate intake and melanoma risk remains unclear. Since new randomized trials are unfeasible given the potential adverse effects of folate on melanoma, further observational studies are warranted.