The relationship of alcohol consumption with atrial fibrillation (AF) is inconsistent, and little is known about the impact of alcohol intake on AF incidence in older people with cardiovascular diseases (CVD) or diabetes. The purpose of this analysis is to assess the association between alcohol consumption and incident AF among older individuals with CVD or diabetes mellitus with end organ damage. A cohort of individuals (n=30,430) without AF at baseline, aged ≥55 years, with established CVD or high risk diabetes mellitus, and participating in two randomized trials, the ONTARGET or TRANSCEND (median follow-up of 56 months), were included in the analysis. Alcohol intake was classified as low (Group 1, <1 drink/week, referent), medium (Group 2, 1-14 drinks/week for women and 1-21 drinks/week for men) and high (Group 3, >14 drinks/week for women or >21 drinks/week for men) consumption, according to the World Health Organization recommendations. Multivariable Cox regression was used to model the relationship between alcohol intake and incident AF. There were 2093 new onset AF events in these patients. The age- and sex-standardized incidence rates were 14.47, 17.29 and 20.89 events per 1000 person years in Group 1, 2 and 3, respectively with an additional 6.42 events per 1000 person years in high compared to low alcohol consumers. The unadjusted and adjusted hazard ratios of incident AF associated with alcohol intake were shown in the Table. A significant linear dose response relationship was found (p for trend <0.0001 and =0.0019, respectively in the unadjusted and full-adjusted models). In conclusion, moderate alcohol intake is associated with an increased incidence of AF in older individuals with CVD or diabetes mellitus and in a linear fashion with increasing consumption. Recommendations made about the protective effects of moderate alcohol intake in patients at high risk of CVD should be tempered with this new finding.