Have you seen this paper?

GBD 2016 Alcohol Collaborators  (2018) Alcohol use and burden for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016.
Lancet; 392: 1015–35  http://dx.doi.org/10.1016/S0140-6736(18)31310-2

This wide reaching study, which was based on pre-existing work, aimed to address known biases in that work. Among other study outcomes, the authors carried out a new meta-analysis to assess the dose response risk of alcohol for 23 diseases, including cancer. The level of alcohol consumption at which an individual’s total attributable risk of ill health increased was also estimated. In 2016, one third of the world’s population were drinkers (have drunk some alcohol in the last year) and, although there was a wide variation by location, prevalence was  found to be higher in more affluent areas and in males more than females. The amount that was drunk daily was also higher in more affluent areas. In 2016, 2.8 million deaths were attributed to alcohol use globally. Disease burden attributable to alcohol was found to be lower in  women than men and in areas of higher deprivation compared to areas of less deprivation. Some protective effects for ischaemic heart disease and diabetes were found for both men and women especially in more affluent areas but, for the over 50s, cancer accounted for 27% of women’s and 19% of men’s alcohol attributable deaths. In more affluent areas cancer was the main attributable burden of disease for both sexes. The level of alcohol to minimise the overall risk of ill health was 0 (95% CI 0.0 – 0.8) standard drinks daily with a steadily increasing relative risk of ill health with increasing consumption. The protective effects of alcohol in heart disease and diabetes were offset by the increasing risk of cancer with increasing consumption and would only be significant in populations where diabetes and ischaemic heart d disease comprised more than 60% of total deaths in a population.