July 27, 2017

expert reaction to study looking at drinking alcohol and risk of diabetes

In new research published in Diabetologia scientists report that alcohol drinking frequency is associated with a reduced risk of diabetes and that consumption of alcohol over 3-4 days per week is associated with the lowest risk of diabetes.

Dr Graham Wheeler, Bayesian Medical Statistician, UCL, said:

“Whilst this large study has found an association between moderate weekly alcohol consumption and a reduced risk of diabetes, this alone does not prove a causal link.

“Establishing a biological mechanism for how this protective effect might work is key to understanding the findings of these types of study.

“In the Danish study, participants were asked to recall drinking habits only once.  So participants may under- or over-report their true alcohol consumption.  We also don’t know how their drinking habits changed as they were followed up.

“Researchers looked at the association between diabetes onset and lots of different categories of drinking behaviour, which increases the chance of claiming at least one association is statistically significant, when actually it isn’t.

“Whilst drinkers may want to raise a glass upon hearing this news, alcohol has been linked to the increased risk of alcoholic hepatitis, liver cirrhosis and several cancers.  Further research will help us piece together the complex relationship between alcohol consumption and diabetes.”

Dr Amitava Banerjee, Senior Clinical Lecturer in Clinical Data Science and Honorary Consultant Cardiologist, UCL, said:

“This new Danish study claims that frequent alcohol consumption is associated with a reduced risk of diabetes in both men and women.  There are other studies which have suggested that frequent alcohol consumption is not associated with lower risks of diabetes, but the evidence is mixed, partly due to varying definitions of drinking patterns and lack of good prospective studies.

“On the basis of this study, more country-specific, well-designed studies need to be done to look at the detailed associations between alcohol and diabetes over the long-term.  Current guidelines to restrict weekly drinking to 14 units per week and to spread your drinks throughout the week are unchanged by this study.  When considering alcohol intake, we should take into account all of the possible long-term health consequences, not just diabetes.

“In common with previous studies in this area, there are at least four reasons for great caution in interpreting the results.  First, in general people do not self-report their alcohol or dietary intake accurately (they tend to under-report), and there is a lack of scalable tools outside research to gather the best data from the real world.  We are not told if the alcohol questionnaire in this study has been previously used or validated.  Moreover, change in alcohol consumption over time was not assessed.

“Second, it is always difficult to know how representative of a population the sample is.  For example, only 14% of people invited to complete the questionnaire did so.  Were heavier drinkers who had diabetes less likely to complete the questionnaire about drinking?   Incidence rates of diabetes and drinking patterns vary considerably by country so the generalisability to other populations or countries is difficult to know.

“Third, as noted by the authors, “confounding may contribute to the favourable associations often associated with moderate alcohol consumption and high drinking frequency”, because infrequent drinkers drink more on drinking days and are more likely than frequent drinkers to have cardiovascular risk factors, including diabetes.

“Fourth, diabetes is a chronic condition which develops over years and 4.9 years of median follow-up is insufficient to accurately capture the full long-term association with alcohol consumption.”

Prof. Nick Finer, Honorary Consultant Endocrinologist and Bariatric Physician, University College of London Hospitals, and member of the Society for Endocrinology, said:

“This study confirms previous work showing that moderate alcohol intake (particularly wine) is associated with a lower risk of developing diabetes.  It adds to previous work by suggesting that ‘frequent’ drinking – that is consuming alcohol on 3-4 days each week – was also protective compared to drinking 1 or less days each week.  While this might suggest binge-drinking was harmful, the study did not find this, perhaps because the incidence of reported binge drinking was low.

“There are caveats, as always, in interpretation of these studies based on surveys.  They only report associations and while the associations persist even after taking into account possible confounding factors the study does not provide any evidence that increasing alcohol intake or consumption frequency will reduce the risks of diabetes.  Also, the study was in a Danish population with a relatively healthy lifestyle compared to the UK, with a low prevalence of obesity, (10% compared to the current UK 25%), itself a major risk factor for type 2 diabetes, so it may not immediately translate to a UK population.  Alcohol is also of course a potentially important source of calories.  Diabetes prevention efforts rightly focus on helping people achieve a healthy weight and to take adequate physical activity.”

* ‘Alcohol drinking patterns and risk of diabetes: a cohort study of 70,551 men and women from the general Danish population’ by Charlotte Holst et al. published in Diabetologia on Thursday 27 July 2017.

Declared interests

Dr Graham Wheeler: “I am employed by UCL, have a visiting researcher position at the MRC Biostatistics Unit at the University of Cambridge, am a Fellow of the Royal Statistical Society, and a voluntary research committee member for Chiltern Music Therapy, a not-for-profit organisation providing music therapy services.  I have no financial conflicts of interest.”

Dr Amitava Banerjee: “I have no conflicts of interest.”

Prof. Nick Finer: “Prof Finer declares that he is employed by Novo Nordisk, a manufacturer of diabetes and weight loss medications.”