by ecancer reporter Clare Sansom

Alcohol drinking is one of the best known risk factors for the development of cancer after tobacco use.

About 3.6% of all cancer cases worldwide can be attributed to alcohol use, proportionally more in men than in women, and alcohol has been associated with increased risk of cancer at many different sites. 

Most studies of alcohol and cancer risk, however, concentrate on excessive drinking: either drinking regular large amounts or occasional, hazardous “binge drinking”.

Light alcohol drinking – generally defined as one drink a day or less on average – has not historically been associated with an increase in risk.  However, a large proportion of the population of most countries can be classified as light drinkers, so even a small increase in risk due to light drinking would equate to a considerable increase in the worldwide cancer burden.

Vincenzo Bagnardi from the University of Milan-Bicocca, Milan, Italy and his colleagues have now conducted an extensive meta-analysis of all published, statistically rigorous case-control or cohort studies of the association between light drinking and cancer.

The analysis was restricted to those studies of cancer types and sites that have at some point been linked to any level of alcohol use: the oral cavity and pharynx, larynx, oesophagus, liver, colorectum and breast. For the same reason, studies of oesophageal cancer were limited to those of the squamous cell subtype.

The researchers identified a total of 222 unique papers that fit all the study criteria; these reported a total of 247 data series of cancer incidence at specific sites. More than half these papers were published in 2001 or later. Some studies included participants who reported occasional alcohol consumption in the reference category of non-drinkers. The pooled studies included about 60.000 cancer cases in non-drinkers and about 92,000 in light drinkers. A total of 110 studies (50%) concerned breast cancer, with these studies involving about 75% of the total cancer cases.

The results of the meta-analysis were recorded in terms of relative risk (RR). Small but significant increases in cancer risk in light drinkers were seen for squamous cell carcinoma of the oesophagus (RR = 1.30; 95% confidence interval (CI) 1.09–1.56), for cancers of the oral cavity and pharynx (RR = 1.17; 95% CI 1.06–1.29) and for female breast cancer (RR = 1.05; 95% CI 1.02–1.08). Interestingly, however, when the results were stratified by geographical area, the increase in relative risk of oesophageal cancer remained significant only in Asian populations. No statistically significant increase in risk was seen for any of the other cancer types studied.

Bagnardi and his co-workers estimated the approximate attributable fraction and number of deaths each year from these cancers that could be assumed to be due to light drinking. The numbers of yearly deaths were estimated as 4980 for oropharyngeal cancer; 23844 for SCC of the oesophagus; and 4909 for female breast cancer, and the attributable fractions ranged from 0.95% for breast cancer to 4.87% for oesophageal cancer in men.

The risk at an individual level, however, is likely to be significantly modulated by genetic differences. The carcinogenic effect of alcohol on the upper parts of the gastro-intestinal tract is thought to arise largely from the presence of acetaldehyde, which is a toxic metabolite of alcohol. The researchers considered that the difference in the relative risk of osophageal cancer between Asian and other populations was very likely to be due to the increased prevalence in Asians of mutations in genes involved in the metabolism of alcohol and acetaldehyde, particularly aldehyde dehydrogenase 2.  

Bagnardi and his colleagues concluded that the literature supported a small but significant increase in cancer risk with light drinking both in breast cancer and in cancers of the regions of the upper gastro-intestinal tract that come into direct contact with alcohol. This increased risk, despite its small size, is responsible for a relatively large number of cancer cases and, therefore, a significant public health problem.

Reference 

V. Bagnardi, V., Rota, M., Botteri, E. and 14 others (2012). Light alcohol drinking and cancer: a meta-analysis. Annals of Oncology, published online ahead of print 21 August 2012.