Re: light drinking has positive public health consequences

Abstract

We read with interest the letter by Steichen et al., who discussed the results of our meta-analysis [1.Bagnardi V. Rota M. Botteri E. et al.Light alcohol drinking and cancer: a meta-analysis.Ann Oncol. 2013; 24: 301-308Abstract Full Text Full Text PDF PubMed Scopus (257) Google Scholar]. We reported that light drinking increases the risk of some types of cancer, and that this might have important consequences on the public health. The authors of the letter pointed out two limitations: (i) the increased incidence of some cancers in light drinkers will not necessarily translate into increased attributable mortality and (ii) even increased cancer mortality in light drinkers will not inevitably translate into increased overall mortality. With regard to the first point, classic incidence studies in epidemiology to determine relative risk comprise both the onset of disease and mortality as end points, especially in disease categories with high-case fatality such as cancer. Therefore, we assumed that the relative risk of cancer incidence could be used for the estimation of the alcohol attributable cancer mortality. In other words, we assumed that light alcohol drinking does not affect cancer survival. Of course, this assumption could be wrong, but de facto the association of light alcohol drinking with both cancer stage at diagnosis and cancer prognosis is still unknown. In other disease categories, there has been a separation between the relative risk functions for onset and mortality, and it turned out that the respective risks were not always the same [2.Rehm J. Taylor B. Mohapatra S. et al.Alcohol as a risk factor for liver cirrhosis—a systematic review and meta-analysis.Drug Alcohol Rev. 2010; 29: 437-445Crossref PubMed Scopus (436) Google Scholar]. Quite opposite to the results of our study were the findings of another meta-analysis published in this journal, which reported that the relative risk of overall cancer mortality is reduced by 9% in light drinkers when compared with occasional/non-drinkers [3.Jin M. Cai S. Guo J. et al.Ann Oncol. Alcohol drinking and all cancer mortality: a meta-analysis. Ann Oncol.. 2013; 24: 807-816Google Scholar]. We believe that the protective effect of light drinking on overall cancer mortality could be subject to several biases. The most important is that light drinking can be part of a generally healthy lifestyle which is protective against many types of cancer: light drinkers might smoke less, be more educated and health-aware, and follow a healthier diet than non-drinkers and heavy drinkers [4.Ruidavets J.B. Bataille V. Dallongeville J. et al.Alcohol intake and diet in France, the prominent role of lifestyle.Eur Heart J. 2004; 25: 1153-1162Crossref PubMed Scopus (110) Google Scholar]. Despite these potential biases, a strong positive dose–response relationship between alcohol and cancer risk had been reported for all the cancer sites we considered in our meta-analysis [5.Bagnardi V. Blangiardo M. La Vecchia C. et al.A meta-analysis of alcohol drinking and cancer risk.Br J Cancer. 2001; 85: 1700-1705Crossref PubMed Scopus (466) Google Scholar]. Moreover, while it is hard to find plausible biological carcinogenic mechanisms to explain the protective effect of alcohol on cancer risk, more and more robust explanations have been proposed to sustain its detrimental effect [6.Poschl G. Seitz H.K. Alcohol and cancer.Alcohol Alcohol. 2004; 39: 155-165Crossref PubMed Scopus (459) Google Scholar, 7.Seitz H.K. Stickel F. Molecular mechanisms of alcohol-mediated carcinogenesis.Nat Rev Cancer. 2007; 7: 599-612Crossref PubMed Scopus (796) Google Scholar]. Regarding the second limitation raised by Steichen et al., many epidemiologic studies have consistently shown that light-to-moderate consumers of alcohol have a reduced risk of all-cause mortality in comparison with non-drinkers and heavy drinkers, mainly due the protective effect of alcohol consumption against coronary heart diseases [8.Corrao G. Rubbiati L. Bagnardi V. et al.Alcohol and coronary heart disease: a meta-analysis.Addiction. 2000; 95: 1505-1523Crossref PubMed Scopus (700) Google Scholar]. Therefore, we agree with the authors of the letter that the existing balance between harmful and beneficial effects of alcohol, particularly at low doses, still needs to be clarified. However, we did not attempt to estimate the effect of light drinking on total mortality. Obtaining the best estimate of the effect of light drinking on cancer risk, which is what we have aimed at doing, is an important but by no means exclusive component of the separate exercise of estimating the effect of this habit on general mortality. The authors have declared no conflicts of interest.