PMC

Impact of choice of reference category on the relationship between alcohol intake and the risk of coronary heart disease.
The solid line shows data from 28 cohort studies (adapted and reproduced with permission from Figure 2 of Corrao et al. 2004. A meta-analysis of alcohol consumption and the risk of 15 diseases. Prev Med, 38: 613-19) and shows the estimated risk-relationship when nondrinkers are used as the reference category. The dashed line shows the same curve with light drinkers (1 g/day) as the reference category. The distances a and b represent the extent that use of nondrinkers as the reference category might lead to overestimation of the benefits of moderate alcohol consumption and overestimation of the level at which alcohol consumption may become cardiotoxic.