It’s high time to seriously question the widely held belief in the health benefits of using non-excessive amounts of alcohol. Reasons are many: the lack of dose-response relationships; the characteristics and lifestyles of today’s abstainers and non-excessive alcohol users; the lack of plausible biological mechanisms; the problems in the classification of drinking groups, and the general limitations of observational studies…

A farmer in remote Sri Lanka, totally out of contact with the English-speaking media and the virtual world, is still more than 95% likely to report with conviction that using ‘non-excessive amounts’ of alcohol is good for his health and likely to lengthen his life. My guess is that the same holds for similar individuals in Venezuela, Vietnam or Vanuatu. How much should we credit national and international health agencies for this remarkable achievement in global public education – in the event that my guess is correct?

My guess is also that people are far less likely to hold with similar conviction that tobacco smoking causes heart disease or leads to premature death. And who should we blame – in the event that my guess is correct here too – for this failure of public education?

You may not think my guesses are accurate. I invite you to test with the next 10 people you meet how strongly they believe the reported benefits of alcohol use versus the harms of smoking and come to your own conclusions. You may find it quite enlightening to test whether the views of farmers (or accountants, teachers or others) differ from those of doctors.

Have we jumped prematurely to the conclusion that associations between better health outcomes in people consuming non-excessive amounts of alcohol – compared to abstainers – reflects a causal link? An article in the journal Addiction suggests that we have. The author, Hans-Olav Fekjaer, provides the most cogently argued case you will likely come across as to why this is so. Quite apart from other causality criteria that aren’t met in the ’alcohol health benefit’ case, he demonstrates the complete lack of plausibility of a biological mechanism underlying the diverse conditions allegedly prevented by alcohol, which include:

Alzheimer’s disease/dementia


Colorectal cancer

Common cold

Coronary heart disease

Diabetes (type 2)


Hearing loss

Intermittent claudication

Liver cirrhosis

Low birth weight, prematurity

Lower urinary tract symptoms (in men)

Metabolic syndrome

Negative child development



Psychiatric disorders

Renal cell cancer

Rheumatoid arthritis

Stroke (ischemic)

These are collectively reflected in better general health status and lower total mortality. No underlying mechanism put forward to explain these associations is more plausible than that they are the result of vector behavior (which may include skewing the alcohol research agenda, selective publication and other things). The second most plausible cause or contributor to the implausible list of ‘protections’ is the fact that abstainers, in the countries where such studies are conducted ad nauseam, are a tiny minority – statistically outside the normal population.

Why have people keen to prove that medicinal properties reside in alcohol not sponsored research on similar associations in countries where abstainers are not a tiny minority? Or is it likely that they may indeed have sponsored many such studies, which have somehow failed to meet the standards required for publication in indexed journals?

Hans-Olav Fekjaer would not find acceptance, in a scientific journal article, for speculation on such matters. But we should continue our own examination of likely underlying causes, as dispassionately as we can. It finally boils down to making an honest assessment of the probability and plausibility of each candidate for the title of  ’cause’ (or causes) underlying the observed associations.

For further reading:

Alcohol—a universal preventive agent? A critical analysis, Hans Olav Fekjær

Article first published online: 1 MAR 2013

DOI: 10.1111/add.12104