Critique 104: Moderate alcohol consumption is associated with a lower risk of diabetes — 26 February 2013

Rasouli B, Ahlbom A, Andersson T, Grill V, Midthjell K, Olsson L, Carlsson S.  Alcohol consumption is associated with reduced risk of Type 2 diabetes and autoimmune diabetes in adults: results from the Nord-Trøndelag health study.  Diabet Med 2013;30:56–64.

Authors’ Abstract

Aims We investigated the influence of different aspects of alcohol consumption on the risk of Type 2 diabetes and autoimmune diabetes in adults.

Methods We used data from the Nord-Trøndelag Health Survey (HUNT) study, in which all adults aged ≥ 20 years from Nord-Trondelag County were invited to participate in three surveys in 1984–1986, 1995–1997 and 2006–2008.  Patients with diabetes were identified using self-reports, and participants with onset age ≥ 35 years were classified as having Type 2 diabetes if they were negative for anti-glutamic acid decarboxylase (n = 1841) and as having autoimmune diabetes if they were positive for anti-glutamic acid decarboxylase (n = 140). Hazard ratios of amount and frequency of alcohol use, alcoholic beverage choice, and binge drinking and alcohol use disorders were estimated.

Results Moderate alcohol consumption (adjusted for confounders) was associated with a reduced risk of Type 2 diabetes in men, but not in women (hazard ratio for men 10–15 g ⁄ day 0.48, 95% CI 0.28–0.77; hazard ratio for women ≥ 10 g ⁄ day 0.81, 95% CI 0.33–1.96). The reduced risk was primarily linked to consumption of wine [hazard ratio 0.93, 95% CI 0.87–0.99 (per g ⁄ day)]. No increased risk was seen in participants reporting binge drinking or in problem drinkers. The results were also compatible with a reduced risk of autoimmune diabetes associated with alcohol consumption [hazard ratio 0.70, 95% CI 0.45–1.08 (frequent consumption) and hazard ratio 0.36, 95% CI 0.13–0.97 (2–7 g ⁄ day)].

Conclusions Moderate alcohol consumption associates with reduced risk of both Type 2 diabetes and autoimmune diabetes. A protective effect of alcohol intake may be limited to men. High alcohol consumption does not seem to carry an increased risk of diabetes.

Forum Comments

Background:  The scientific literature has consistently shown that moderate alcohol consumption is associated with a lower risk of type II diabetes mellitus (DM).  In data from prospective studies, the risk of diabetes has generally been found to be about 30% lower for moderate drinkers than for abstainers (Howard et al, Effect of alcohol consumption on diabetes mellitus: a systematic review. Ann Intern Med 2004;140:211-219;  Koppes et al, Moderate alcohol consumption lowers the risk of type 2 diabetes: a meta-analysis of prospective observational studies, Diabetes Care 2005;28:719-725).  Some studies have shown that this reduction in risk is no longer seen among people reporting binge drinking or drinking excessively (Pietraszek et al. Alcohol and type 2 diabetes. A review. Nutr Metab Cardiovasc Dis 2010;20:366-375).

Comments on the present study:  The present study as based on very large population surveys in Norway, with the first examination in 1984-86 having a very high response rate.  The development of diabetes was ascertained at two later examinations, about 10 and 20 years later.  While there were few baseline differences between subjects developing Type II DM (n=1,841) and those developing auto-immune DM (n=140), overall subjects developing DM were much more likely to be obese and have a positive family history than subjects not developing DM.

Subjects consuming some alcohol but on < 1 occasion over the preceding 2 weeks were taken as the referent group.  When assessing the effects of the frequency of alcohol consumption, alcohol consumption on 5-10 occasions in the preceding 2 weeks was associated with a lower risk of Type II DM of 29% (HR 0.71, CI 0.51, 0.99), with no significant effect of more frequent consumption or of reported episodes of intoxication.  For men, consumption of 5-10 drinks over two weeks gave a HR of 0.69 (CI 0.47, 1.02); for women, the HR was very similar, at 0.73, but with wider CIs (0.34, 1.54).  There were many fewer subjects with autoimmune DM, but the adjusted HR for 1-4 drinks/two weeks was 0.70 (CI 0.45, 1.08) and for 5-10 times the HR was 0.72 (0.28, 1.83).  Reported intoxication was associated with a trend towards a lower risk for this type of DM, not a higher risk.

When the average amount of alcohol consumed per occasion was assessed, the lowest risk of Type II DM was among subjects reporting an average of 10-14 grams/day (approximately 1 typical drink), with an adjusted HR of 0.56 (CI 0.36, 0.86).  This association gave a HR of 0.48 (0.28, 0.77) for men; for women, the lowest risk for among those reporting ≥ 10 grams/day, with HR 0.81 (0.33, 1.96).  When alcohol intake was considered as quartiles of intake, in comparison with subjects in the lowest quartile (0.01-2 grams/day), subjects reporting 2-7 grams per day had the lowest HR, 0.38 (0.15, 0.98).   In beverage-specific analyses, only wine intake showed a significant effect, with each gram/day of alcohol from wine associated with a HR of 0.93 (0.87, 0.99).

Reviewer Djoussé had some concerns about the precision with which alcohol intake was estimated, given that the investigators calculated the mean alcohol intake by having to make a number of estimations based on the reported number of drinks consumed over a two-week period; also, reported episodes of “intoxication” may have been under-reported.  “I wonder how many people that drank alcohol to ‘intoxication’ level were able to recall that event. An under-reporting could bias the results of heavy drinking towards the null.  I agree with others that residual confounding by other healthful lifestyle factors among wine drinkers cannot be excluded as potential explanation for the better effects attributed to wine.”

Forum member Skovenborg commented: “In the first examination in this prospective study the participants drinking >10 times during the last 14 days reported high rates of low education, physical inactivity and smoking, which may be characteristic of regular drinkers in the Norwegian population.  However, we have reason to suspect unmeasured confounding here, and the authors did not have information about the typical diets in the different categories of alcohol intake.  That might explain why wine is the only beverage associated with significant risk reduction.  In Scandinavia, wine drinking is typically associated with a Mediterranean type of diet (Johansen D et al. Food buying habits of people who buy wine or beer: cross sectional study. BMJ 2006;332:519. doi.org/10.1136/bmj.38694.568981.80).  Thus, it is possible that some of the protection of wine drinkers seen in this study may have related to a Mediterranean type diet, that has been shown to reduce the risk of diabetes and other cardiovascular diseases (Estruch et al, Primary prevention of cardiovascular disease with a Mediterranean Diet. N Engl J Med 2013. DOI: 10.1056/NEJMoa1200303; published 25 February 2013).”

Skovenborg continued: “The study upon which this report is based is an interesting study that included a large community-based sample with a long follow up.  The authors note that the participation rate dropped from 90.3% in the first wave of examinations to 54% in the third wave.  In the first examination, the questions regarding alcohol consumption were rather broad, leaving room for inaccurate consumption categories. In the second examination, the participants reported the number of drinks during the previous two weeks, and the calculated intake of alcohol in grams was then divided by 14 to reach an estimate of the alcohol consumption per day.  However, there is always difficulty in knowing exactly the number of drinking days and the number of drinks per drinking days, a fact that casts doubt on the categories of alcohol consumption and drinking pattern.  The claim that the majority of the participants were light drinkers (with a median reported intake of 2.3 g/day) may not be a valid estimate without knowledge regarding the drinking patterns of the participants.”

Differences in effects by gender and by type of beverage: The apparently greater effect among men than among women is not consistent with results from many other studies.  The reasons for this finding are not known, although the authors stated: “Gender differences could be related to women being more sensitive to the toxic effects of alcohol, experiencing toxicity at half of the dose giving negative symptoms in men.  Alternatively, if women were less likely than men to accurately report their alcohol consumption attributable to social desirability, then this would dilute an association between alcohol consumption and diabetes in women.”  Said reviewer Skovenborg, “Women in this study who reported a moderate consumption of alcohol showed a non-significant reduction of the risk of type 2 diabetes; however, the lack of significance may be a type 1 error due to a small number of diabetes cases among women.”

Said reviewer Finkel: “Alcohol’s apparent protection from diabetes, a decrease in risk of 30 percent, at least in men, is impressive.  It is said that we have 26 million diabetics in the US alone, mostly Type 2: the magnitude of benefit would, therefore, be enormous. The sex difference is also of great interest, and still puzzling.”

Mechanisms of alcohol’s effects on diabetes:  The investigators of this study stated that improvement in insulin sensitivity, immune mechanisms, and inflammation might explain the lower risk of diabetes among moderate drinkers.  Commenting on their finding of greater protection from wine, they also pointed out that wine drinking may also be related to healthier behaviours compared with drinking of spirits, particularly with regard to diet and socio-economic factors.

Reviewer Van Velden noted that in their own studies of the influence of both red and white wine on health, “fasting glucose decreased and insulin increased in ‘normal’ people, but such findings were not found in people with the metabolic syndrome.  This suggests that while alcohol does seem to play a protective role in diabetes, it may not be present in overweight people and people with genetic risk factors for cardiovascular disease such as APO E4 polymorhism, and familial dylipidemias.”

Added reviewer Orgogozo:  “The present paper confirms the higher risk of diabetes associated with increased weight, and many papers have shown that moderate alcohol consumption may be associated with a modest decrease in weight.  This could also explain part of the inverse correlation between alcohol and diabetes.  (It would be interesting to know if there was a difference between this association for wine drinkers, often found to have a healthier life style, and for consumers of other types of alcoholic beverages.)”

Forum Summary

A population-based longitudinal study in Norway showed a lower risk of developing type II diabetes, and a tendency for less auto-immune diabetes, among subjects reporting moderate alcohol consumption.  The lowest risk was among subjects reporting that they consumed alcohol 5-10 times during the previous two weeks or reporting a daily consumption of 10-15 grams of alcohol per day (slightly less than one to a little over one drink per day, using 12 g of alcohol per day as a “typical drink”).

The overall results from this study support most previous research showing a rather large (approximately 30% or more) reduction in the risk of Type II diabetes mellitus to be associated with moderate drinking.  In this study, the strongest protective effect related to the consumption of wine, rather than beer or spirits.  Unlike some previous research, the present study did not show an increased risk of diabetes among binge drinkers, heavier drinkers, those reporting episodes of intoxication, or those giving positive responses on a CAGE questionnaire (a measure of an alcohol use disorder).  As it is likely that alcoholics were under-represented in the cohort, the ability to show the effects of very heavy drinking may have been limited.  

While the numbers of subjects with auto-immune diabetes were small, there was a tendency seen for a similar reduction in risk with moderate drinking as with the much more common Type II diabetes.  Overall, this study strongly supports much previous research showing that moderate alcohol consumption is associated with a lower risk of developing diabetes.

                                                                         *                     *                      *

Contributions to this critique by the International Scientific Forum on Alcohol Research were provided by the following members:

Erik Skovenborg, MD, Scandinavian Medical Alcohol Board, Practitioner, Aarhus, Denmark

Gordon Troup, MSc, DSc, School of Physics, Monash University, Victoria, Australia

Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA

Pierre-Louis Teissedre, PhD, Faculty of Oenology – ISVV, University Victor Segalen Bordeaux 2, Bordeaux, France

Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway

Dominique Lanzmann-Petithory,MD, PhD, Nutrition/Cardiology, Praticien Hospitalier Hôpital Emile Roux, Paris, France

David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa

Luc Djoussé, MD, DSc, Dept. of Medicine, Division of Aging, Brigham & Women’s Hospital and Harvard Medical School, Boston, MA, USA 

Jean-Marc Orgogozo, MD, Professor of Neurology and Head of the Neurology Divisions, the University Hospital of Bordeaux, Pessac, France

R. Curtis Ellison, MD, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA