The incidence of melanoma has continued to rise over the last few decades [1]. Although many explanations have been posited, such as increased screening, detection, and UV exposure [2], it is essential to examine non-UV–related risk factors contributing to its continued rise.

We reviewed published data on obesity, smoking, and alcohol consumption trends worldwide to understand human behaviors and their relationship to melanoma. We collected data for the three risk factors from the Global Health Observatory (GHO), World Health Organization (WHO), published in 2010, as these were the most comprehensive available data with minimal changes in trends noted in the following years. We also collected data for melanoma incidence and mortality from the Global Cancer Observatory (GCO), WHO, published in 2018, as they were the most currently available data. Compiled data were displayed using choropleth maps with color gradients to visualize variations across geographic areas (Figure 1A). Subsequently, each country’s data were plotted, and Spearman rank correlation coefficient (R) was calculated for melanoma incidence and mortality with each risk factor. The choropleth map of each risk factor showed similar patterns to melanoma incidence and melanoma mortality. The statistical analysis depicted a positive correlation (with a positive R) between melanoma incidence/mortality and all risk factors (obesity, smoking, and alcohol consumption). Among them, alcohol consumption showed the strongest positive correlation with both melanoma incidence (R=0.72; P<.001) and mortality (R=0.59; P<.001). Because individuals with lighter skin color (eg, Caucasians) have a higher melanoma incidence, this correlation data might implicate that alcohol consumption is high in countries with lighter skin color, such as European ancestry. To address whether the correlation between alcohol consumption and melanoma incidence is skin color dependent or independent, we reanalyzed the data by continent (Figure 1B). A positive correlation still existed between alcohol consumption and melanoma incidence in Europe, Asia, and Africa (Multimedia Appendix 1, Supplementary Table S1). In particular, the strongest correlation (R=0.68; P<.001) was observed in European countries with exclusively lighter skin color (1-12 or 12-14 of skin color numbers, per a human skin color distribution map in the second figure by Barsh [3]), suggesting that the correlation between alcohol consumption and melanoma incidence is likely to be skin color independent. A positive correlation was also observed between alcohol consumption and melanoma mortality in all continents.

To understand how genetic risk factors have a role in the observed correlation between alcohol consumption and melanoma incidence/mortality, we examined the correlation of aldehyde dehydrogenase 2 (ALDH2) rs671 polymorphism with both alcohol consumption and melanoma outcomes. ALDH2 is an alcohol-metabolizing enzyme, and its allelic variation affects alcohol detoxification [4]. The correlation analysis showed that it was the wild-type ALDH2 allele that was strongly positively correlated with melanoma incidence (R=0.70; P<.001) and mortality (R=0.74; P<.001; Table 1). On the other hand, the allelic variants had a modest to strong negative correlation with melanoma incidence (R=–0.70 to –0.51; P<.001) and mortality (R=–0.73 to –0.57; P<.001). Possible explanations for the opposing effect of ALDH2 polymorphism include that individuals with risk alleles consume less alcohol than wild-type individuals, which is consistent with our data showing a positive and a negative correlation to alcohol consumption in wild-type individuals and allelic variants, respectively.

Overall, our findings highlight an association between alcohol and melanoma outcomes globally. The association was observed not only with melanoma incidence but also with its mortality. We also found a potential involvement of the alcohol-related gene ALDH2. Limitations of our analyses include unavailability of the population statistics for some risk factors by some countries, binary questionnaire of alcohol use without reflecting the quantity of alcohol consumption, and country-based analysis rather than individualized data. To determine whether our cross-country data support individual-level conclusions at individual levels, individual-level studies, such as a recent one [5], need to be conducted. Furthermore, our findings do not necessarily indicate causation from alcohol, and other factors might be involved, including skin/hair color, ethnicity, geography, economy, and lifestyle. Further investigation is warranted to verify these associations at individual levels and elucidate alcohol’s effects on melanoma outcomes by eliminating potential confounding factors such as skin/hair color genotypes.