September 16, 2019

expert reaction to conference abstract/poster looking at alcohol consumption in people with type 2 diabetes

An abstract, being presented at the Annual Meeting of the European Association for the Study of Diabetes (EASD), reports that there may be some positive effects to moderate alcohol consumption for people with type 2 diabetes.

Prof Naveed Sattar, Professor of Metabolic Medicine, University of Glasgow, said:

“This report collates only modest amounts of data, even though from randomised trials and it looks only at a small number of risk markers and not hard outcomes, which is what we need to make guideline changes.  The most robust data anywhere, published in 2018, suggest the lower the amounts of alcohol consumed, the lower the risks for multiple heart and stroke related outcomes.  Hence, less, not more alcohol should be the constant message and this includes in people with diabetes.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“It’s always hard, and very often impossible, to assess the quality of a piece of research from the kind of restricted information available about a conference presentation.  That’s particularly true for systematic reviews and meta-analyses, as in this case, because what you need to know is in the details, and the details won’t be available until and unless there’s a proper, peer-reviewed, research paper.  Exactly how did the researchers search for the studies they examined?  How did they assess the quality of the individual studies, and what did they do about it in their analysis?  How consistent were the individual studies in the methods they used, and exactly what they were investigating?  How were the patients chosen in each study?  Exactly what statistical methods were used for the meta-analyses?  We just don’t know any of this yet.  So this might be a good piece of new research, or it might not.  I just can’t say.

“Just a couple of points that might be important.  The individual studies that contributed to this research were randomised clinical trials, where people with diabetes were divided into groups at random and either took alcohol or took something else (such as water, or dealcoholized wine).  This was in the context of various physiological measurements being made, so often very different from the situation of having a drink of wine or beer with your meal at home.  That makes it pretty unclear to what extent the results might apply to everyday life.  (As the press release makes clear, there are in any case other reasons why people with type 2 diabetes need to be careful about their alcohol consumption.)  And maybe it’s important that, of the ten studies that this new research is based on, in eight of them the alcohol was provided in the form of wine, most commonly red wine.  There have been many suggestions in the past that substances other than the alcohol in red wine can have effects on health, though the evidence for that is not clear – but why are the researchers here so sure that the effects they found were due to the alcohol and not something else?  Perhaps they will say in the full research paper if and when it appears.

“The two studies that didn’t use wine (assuming that I identified them rightly from the minimal information given on the poster) do show how unlike everyday life some of these studies were.  They involved the participants drinking 40g of alcohol, that’s 5 UK units, of vodka in one case and in an unstated form in the other.  They did not do that every day.  In one of the studies they did it just once, and in the other it was on special study days at least a week apart, accompanied by a strange special meal of zero-calorie soup, butter, and white bread.  Detailed physiological measurements were taken.  Nothing wrong with that in the context of what the people who ran these studies were trying to find out, but how far does it make sense to combine these with other studies where the participants, in some cases, drank red wine along with normal meals?  I just can’t say without more detailed information.”

Dr Sadie Boniface, Research Coordinator, Institute of Alcohol Studies, said:

“This research poster being presented at a conference is a review of trials of alcohol intake among people with type 2 diabetes, looking at the effects on glucose and lipid metabolism.  By combining the results of 10 trials, the authors found up to a glass of wine a day (or equivalent) was linked to beneficial effects on some biomarkers, but not others.  It’s possible this could be due to the alcohol (ethanol) content, but it could also be due to other compounds in the drinks people were given.  Any benefits would need to be weighed up against all the other health risks of alcohol use, for example the increased risk of cancers even at low levels of drinking.

“This research is unpublished, so has not been through peer review, and only a small amount of information is available; for example we don’t know about the quality of the studies included, meaning they may be subject to important limitations or bias.  This means it’s difficult to draw any conclusions at this stage.”

Dr Bob Patton, drugs alcohol and addictive behaviours research group, University of Surrey, said:

“The authors of this paper pooled the results of 10 randomized controlled trails that explored the effect of alcohol consumption on levels of triglycerides and insulin among people with type 2 diabetes.  They found that for those drinking less than 2.5 units of alcohol per day (in line with current UK guidelines), that these levels were reduced when compared to those drinking above that amount.

“These results will be presented at a conference later this week, and have not been subject to peer review (the scientific process where the work is critically evaluated by experts in the field) or published in a reputable journal.  Therefore any findings should be interpreted with caution.  The work presented has little detail and raises a number of questions around the quality of the data analysed, the calculations undertaken and the limitations of the studies included in the analysis.

“The authors suggest that based upon these findings, recommendations to moderate alcohol consumption for people with type 2 diabetes may need to be reviewed.  In the UK this is not necessary, as the alcohol guidelines are the same for diabetics as for the general population; alcohol consumption is associated with a number of physical and psychological conditions, and drinking within the recommended limits (2 units/day) is the best way to reduce the risk of developing alcohol related harms.”

Abstract/poster title: ‘The effect of alcohol consumption on glucose and lipid metabolism in type 2 diabetes: a systematic review and meta-analysis’ by Y. Chen et al.

This is a conference abstract from the Annual Meeting of the European Association for the Study of Diabetes (EASD), and was under embargo until 23:01 UK time on Monday 16 September 2019.

There is no paper as this is not published work.

Declared interests

Prof Naveed Sattar: “No COI.”

Prof Kevin McConway: “Prof McConway is a member of the SMC Advisory Committee, but his quote above is in his capacity as a professional statistician.”

Dr Sadie Boniface: “I work at the Institute of Alcohol Studies which receives funding from the Alliance House Foundation.  IAS is a small charity based in Westminster and we do research as well as policy work. Our role is to advance the use of the best available evidence in public policy decisions on alcohol. More details and info on how we are funded here: http://www.ias.org.uk/Who-we-are.aspx”

Dr Bob Patton: “No conflicts of interest.”