Year: 2013
Type: editorial
Abstract: Hundreds of observational, epidemiological studies have found that moderate alcohol intake is associated with a lower risk of a range of health outcomes compared to alcohol abstention, but that risk increases with heavy drinking (a distribution commonly known as a ‘U-shaped’ or ‘J-shaped’ curve) 1. Many researchers, clinicians and the lay public now accept as common knowledge that moderate drinking is beneficial to one's health 2, and the association is hypothesized to be due to biological effects on several systems, including reduced blood pressure, increased blood lipoproteins and oxidation of low-density lipoprotein (LDL) cholesterol, among other pathways 3, 4. There are, however, methodological issues and data inconsistencies regarding the relation between alcohol intake and health that require further investigation before the case is closed on moderate drinking. Resolving these outstanding issues and inconsistencies is of substantial public health importance: many individuals are now consuming alcohol, typically red wine, as a prescriptive for better health. If the relation between moderate alcohol consumption and health is more artifactual than causal, then public health recommendations will need to be reconsidered. The first issue raised by Dawson et al. 5, and documented in other studies as well, is the ‘sick quitter’ effect. Dawson et al. document that individuals with incident cardiovascular disease and those rating health as good/fair/poor (versus very good or excellent) at age 55+ were more likely to stop drinking, as well as individuals with liver disease at ages 18–34. Thus, when alcohol intake is measured at a single time-point, non-drinkers will constitute both former drinkers who quit due to poor health and life-time abstainers. Separating former drinkers from life-time abstainers in studies of the effects of alcohol consumption on mortality is thus essential. Indeed, a recent analysis that did this showed that the J-shaped curve disappeared after taking into account former drinkers who quit due to poor health and that the association between alcohol consumption and mortality changed to a simple monotonic function: greater levels of alcohol consumption are associated with higher mortality 6. A second issue is that uncontrolled confounding remains a threat to the validity of epidemiological studies on the relation between moderate alcohol consumption and health. Cross-cultural analyses and novel designs can shed light on this issue because they do not rely on traditional survey-based measures and regression-based control of confounders. The results of such studies are not promising for a causal effect of moderate intake. First, Rimm et al. found that the association between moderate intake and health varies by beverage type across countries 7; the beverage type most associated with health benefits is consistently that which is most normative and prevalent within the country. If there were a biological effect of moderate wine or other beverage intake on improving health, we should observe this phenomenon consistently across countries. Secondly, a meta-analysis of Mendelian randomization studies focusing on genetic variants that regulate alcohol metabolization has demonstrated that blood pressure is increased even among those with genotypes associated with moderate drinking, in contrast to traditional observational epidemiological study design findings 8. These findings call into question the extent to which the protective association of moderate consumption with health is truly causal. A final and related issue is that the protective benefits of moderate alcohol consumption are not evident in all demographic groups. Accumulating evidence indicates that African Americans experience no or little protective effect 9-12 of moderate alcohol consumption on incident cardiovascular disease 10, cardiovascular mortality 12 or all-cause mortality 9, 11. The seminal studies documenting the protective health benefits of moderate alcohol consumption have been assessed predominately in samples of majority white, relatively affluent men in the developed world 1. Studies on the health benefits of moderate drinking among women are limited, although data from the Nurses' Health Study indicate that middle-aged, majority white women also demonstrate the health benefits of moderate drinking 13. If moderate drinking is beneficial for health through biological vascular or other mechanisms, why is it not apparent in all demographic subgroups? This inconsistency must be addressed before the full extent of the relation between moderate alcohol consumption and health is firmly established. In conclusion, there has been substantial concern about the potential intended consequences of promoting a message that moderate alcohol consumption is beneficial for health 2, and the points outlined in this commentary are well known to those skeptical of a causal relation 14. Converging evidence suggests that moderate drinking increases the risk of some cancers 15, thus the unintended consequences of promoting moderate drinking for cardiovascular health in the absence of a true causal link could have an overall detrimental effect on population health. However, the prevailing viewpoint with regard to moderate intake, that it is a health-promoting activity, remains. We have been proved wrong too often about the effects of dietary factors 16, medication and supplements 17, 18, and even salt consumption 19, with regard to potential health benefits or adverse effects based on observational study designs. As illustrated by Dawson et al., the case for moderate drinking should not be closed. Investment in research is paramount at this critical time, before moderate drinking is subsumed under the category of known wisdom. None. This research was supported in part by support from New York State Psychiatric Institute and the Department of Epidemiology at the Mailman School of Public Health (Keyes). The funding organizations had no role in the design and conduct of the study; collection, management, analysis and interpretation of the data; and preparation, review or approval of the manuscript.
Source: Addiction
DOI: 10.1111/add.12137
Authors: Katherine M. Keyes, Richard Miech
Institutions: Columbia University, University of Colorado Denver
Language: English
Subfield: Pathology and Forensic Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W1522120014