Does adequate physical activity attenuate the associations of alcohol and alcohol-related cancer mortality? A pooled study of 54 686 British adults

The potential of physical activity (PA) to attenuate the effects of alcohol consumption on the risks of alcohol-related cancer mortality is unknown. We used data from participants aged 30 years and over in 10 British population-based surveys (Health Surveys for England 1994, 1997, 1998, 1999, 2003, 2004, 2006 and 2008 and the Scottish Health Surveys 1998 and 2003). Alcohol-related cancer mortality included oral cavity, throat, larynx, oesophagus, liver, colorectal, stomach and female breast (conservative definition), and additionally pancreas and lung (broad definition). Alcohol consumption was categorised into six groups based on the UK units/week: (a) never-drinkers, (b) ex-drinkers, (c) occasional drinkers, (d) within guidelines (<14 UK units/week [women]; <21 UK units/week [men]), (e) hazardous (14-35 [women]; 21-49 [men]) and (f) harmful (>35 [women]; >49 [men]). PA was categorised using two dichotomous classifications based on the lower (7.5 Metabolic Equivalent Task [MET]-hours/week) and upper (15 MET-hours/week) recommended limits. Using Cox proportional hazard models, we found a strong direct association between alcohol consumption and mortality risk of alcohol-related cancers, with a significantly higher risk among ex-drinkers (Hazard ratio [HR] = 1.46, 95% confidence interval [CI] = [1.09, 1.94]), drinkers who consumed hazardous (HR = 1.39, 95% CI = [1.06, 1.83]) and harmful amounts of alcohol (HR = 1.62, 95% CI = [1.13, 2.30]) compared to never-drinkers in the fully adjusted model. The increased mortality risks were substantially attenuated when participants in these drinking groups exercised >7.5 MET-hours/week. PA could be promoted as an adjunct risk minimisation measure for alcohol-related cancer prevention.