The evidence gap between alcohol consumption and oesophageal squamous cell carcinoma in east Africa – Authors' reply

We welcome the comments by Jin Zhang and colleagues on our study,¹ which reported the strong role of alcohol consumption in oesophageal squamous cell cancer (ESCC) in east Africa. Our findings, which showed that alcohol consumption was a risk factor for ESCC, are consistent with the IARC monographs.² In Kenya and Tanzania, we found that a large proportion of male patients with ESCC was attributable to alcohol consumption. Public health actions to reduce harmful alcohol consumption will benefit not only cancer prevention but also prevention of other NCDs and alcohol-associated harms in society.

Zhang and colleagues highlight investigations to further refine the association between alcohol consumption and risk of ESCC. Many of these suggestions are already in the pipeline, given that the ESCCAPE study is contributing to a genome-wide association study of ESCC in Africa.³ Genetic influences on alcohol metabolism—notably polymorphisms of ADH and ALDH2 genes—are known to influence concentrations of acetaldehyde and risk of ESCC, and have been well studied in Asian populations, but not in African populations. We have not been able to investigate such effect modification without genotyping because there is no known accessible proxy, such as self-reported alcohol flushing, in African populations. Although interactions between genes and the environment will be studied, investigating the primary effect of a main exposure is an important first step. A genetic-based precision component is unlikely to benefit or be feasible for primary prevention strategies tackling alcohol; however, such a component might help to refine the identification of individuals at high risk for early detection.

Zhang and colleagues speculate that genetic differences between ethnic groups might contribute to heterogeneity in the magnitude of the relationship between alcohol consumption and risk of ESCC. We do not expect genetic effects to be the driving factor behind the absence of association in Malawian men because any genetic modification to the effect of alcohol on ESCC risk would probably affect both sexes, and a positive association was seen in Malawian women. Therefore, we considered reporting biases to be a more plausible explanation for the sex differences. Effect modification of the relationship between alcohol consumption and risk of ESCC by tobacco use was presented and was found to be present in Kenya, as has been previously published.⁴ However, the pattern of alcohol consumption and tobacco use is such that the power to detect interactions is reduced because there are few smokers who have never consumed alcohol.

The priorities in our analytical approach acknowledge that the research needs, priority questions, and status of existing knowledge in aetiological cancer epidemiology are not homogenous worldwide. Building a solid foundational knowledge on major environmental and lifestyle risk factors remains a necessary research priority to support cancer control plans in sub-Saharan Africa. Risk refinements will be made as research investments, which include prospective studies and improved exposure assessments and biobanking, are made.

We declare no competing interests.