Re: Alcohol intake in relation to non-fatal and fatal coronary heart disease and stroke: EPIC-CVD case-cohort study
Error in conclusion from the EPIC-CVD study
Ricci et al (BMJ 2018; 361:k934) report results relating alcohol intake to CVD from a multicentre case-cohort study which they argue “strengthens the evidence for policies to reduce alcohol consumption”. The main results, shown in Table 2, show that increasing alcohol consumption by 12 g/day is associated with a 6% (95% CI 4% to 8%) reduction in non-fatal CHD risk and an 8% (1% to 15%) reduction in fatal CHD risk, based on, respectively, 9307 and 1699 events. In contrast such an increase is associated with a 4% (2% to 7%) increase in non-fatal stroke and a 5% (-2% to 13%) increase in fatal stroke risk, based on somewhat fewer events, 5855 non-fatal and 733 fatal. For the two diseases combined, the larger reductions in risk for the more common disease clearly outweigh the smaller reductions for the less common one. Bearing in mind that some previous studies (cited in the section “Comparisons with other studies” ) showed a positive association of alcohol with CHD, and the extremely large number of cases in the present study, the results surely “weaken” not “strengthen” the evidence for alcohol-reduction policies. How could the authors have reached the opposite conclusion?
Competing interests: No competing interests