Alcohol, health, and the heart: implications for clinicians.

The association between light drinking and reduced mortality compared to lifetime abstinence is robust. Confounders, such as unhealthy diet, social class, and social isolation, which are associated both with abstaining and with poorer health, have made it difficult for doctors to assess the meaning of the association, and to know how to respond to individual patients seeking advice on 'drinking for health'. This paper attempts to bring up to date the evidence concerning the confounders. The evidence for a preventive effect is reviewed according to pattern of alcohol use and type of patient. The balance of risk to benefit appears to favour giving medical advice to some patients in middle life with uncontrollable risk factors for coronary heart disease (and probably of ischaemic stroke), who are very infrequent drinkers, to increase slightly the frequency of drinking (not the per session amount) and for some abstainers to consider starting to take alcohol. However, caveats render quite small the number of patients to whom physicians will give this advice. There is also the theoretical risk of a ripple in the population such that more people may move into harmful drinking.