Alcohol and ischaemic heart disease risk—finally moving beyond interpretation of observational epidemiology

Studies involving Mendelian randomization appear to be casting doubt on the idea that moderate alcohol consumption is protective against coronary heart disease. However, these studies make certain assumptions that may not be warranted. While the methodology is opening up new possibilities with regard to addressing this important question, there is still much work to do before we can be confident in the conclusions. A recent Mendelian randomization analysis 1 challenged the current view that light to moderate consumption of alcohol is associated causally with lower risk of ischaemic heart disease (IHD 2-6). Mendelian randomization has several advantages: (i) it can be considered as almost experimental, thus being associated with stronger control than usual observational studies in epidemiology 7; (ii) its argumentation relies on more or less drinking among drinkers, only thus avoiding the problems of different types of abstention and their differential risks 8, 9 and their measurement problems 10; and (iii) it deals in part with the measurement problem of single versus multiple observations across the life-time, as it is plausible that the genetic constellation has an impact on alcohol intake at not only one time-point. Furthermore, Holmes and colleagues' study is based on a meta-analysis of several studies, thus avoiding errors linked to specific operationalizations. Figure 1 (based on Roerecke & Rehm 4) shows the typical J-shaped curve found in epidemiological investigations. The logic of Holmes and colleagues 1 is as follows: in area (a), (lower than the nadir), any increase in average consumption is linked to a decrease in IHD risk, and in area (b) (higher than the nadir) the reverse is true and the increase in IHD risk accelerates with increasing alcohol intake. Change in ischaemic heart disease mortality risk by average alcohol intake in men. Source: adapted from Roerecke & Rehm 5 The conclusion of Holmes and colleagues 1 is based on their comparison of alcohol intake and risk of IHD by carriers versus non-carriers of the ADH1B rs1229984 A-allele in different strata of average alcohol intake. This allele is part of the primary pathway of alcohol metabolism 11, and is associated with a flushing response, lower levels of usual alcohol consumption and blood ethanol levels 12. The existence of a protective effect implies that less drinking below the nadir should result in an increase of IHD risk whereas, empirically, allele carriers in that category had lower but not higher IHD risks 1. Unfortunately, there is still some doubt as to whether these assumptions are true. First, several risk factors for IHD were associated with allele status across strata of average alcohol intake 1: i.e. the effect could be mediated by other risk factors than alcohol consumption. Secondly, with respect to mediation of the effect via alcohol consumption, there are at least three dimensions of alcohol shown to impact IHD risk: average volume 3, 4, heavy drinking occasions 16 and regularity versus irregularity of drinking 5, 6, 17. There is a complex interaction between these dimensions, and simply focusing on one dimension may lead to problematic results. Holmes and colleagues showed that carriers had both lower average consumption and a lower probability of binge drinking across studies. This becomes particularly important when looking at light to moderate drinkers below the nadir [area (a)]. Based on current epidemiological evidence 16, ceteris paribus, we would expect carriers among light to moderate drinkers to show lower IHD risk because of the lower probability of binge drinking 16. Thirdly, the described analysis was based on aggregate data 1, and we do not know if the average level of alcohol intake of the allele carriers within the strata of average consumption was indeed lower than that of the non-carriers. In order to increase our confidence in the conclusions of the Mendelian randomization analysis the assumptions should be controlled for, at best with a pooled analysis of the underlying individual data. However, despite the large sample size in Holmes and colleagues' study, there might be problems with statistical power to investigate these limitations thoroughly because allele carriers are rare in many European countries. Mendelian randomization studies in populations with more variation in these genotypes may thus be a useful addition to the evidence base of the association between alcohol intake and IHD risk. Finally, let us add some observations for the wider picture regarding the relationship between alcohol and IHD. Unless new confounders are derived theoretically, repeating cohort analyses on average alcohol intake and IHD incidence is probably of limited value (see Rehm & Roerecke 14 for a listing of major confounders already tested). After more than 100 studies in this area with fairly stable associations we will probably not be able to resolve questions of causality with more descriptive epidemiological studies and by discussing their results 18, 19, but only by new and better designs such as, but not limited to, Mendelian randomization studies or more comprehensive experimental trials on biomarkers. None other than acknowledged above. The research leading to these results has been conducted in the context of the European Community's Seventh Framework Programme under grant agreement no. 266813—Addiction and Lifestyles in Contemporary Europe—Reframing Addictions Project (ALICE RAP). Participant organizations in ALICE RAP can be found at http://www.alicerap.eu/about-alice-rap/partners.html. The views expressed here reflect only the author's and the European Union is not liable for any use that may be made of the information contained therein.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.