A Little Tippling May Be Good for What Ales You

Abstract & Commentary

Synopsis: Moderate alcohol consumption is associated with a decreased incidence of diabetes mellitus and a decrease in heart disease in persons with diabetes.

Source: Howard AA, et al. Ann Intern Med. 2004;140: 211-219.

Alcohol consumption is prevalent in the United States: An estimated 109 million Americans 12 years of age or older drink alcohol.1 Diabetes mellitus affects a large portion of the patients physicians care for. Howard and associates note that physicians are not well informed about how alcohol use affects the risk for or management of their patients with diabetes.

Howard et al conducted a systematic review assessing the effect of alcohol use on the incidence, management, and complications of diabetes mellitus in adults. They reviewed English-language studies in persons 19 years or older and that were identified by searching the MEDLINE database from 1966 to the third week in August 2003.

Two independent assessors reviewed 974 citations to identify all experimental, cohort, or case-controlled studies that assessed the effect of alcohol use on the risk, control, self-management, adverse drug effects or complications. The studies were evaluated on the basis of established criteria.

Thirty-two studies that met the inclusion criteria were reviewed. Compared with no alcohol use, moderate consumption (1-3 drinks/d) is associated with a 33% to 55% lower incidence of diabetes related coronary artery disease. Compared with moderate consumption, heavy consumption (> 3 drinks/d) may be associated with up to a 43% increased incidence of diabetes mellitus. Moderate consumption did not impair glycemic control in persons with diabetes.

Howard et al concluded that moderate alcohol consumption is associated with a decreased incidence of diabetes and heart disease in persons with diabetes.

Comment by Ralph R. Hall, MD, FACP

It is noteworthy that the study used the criteria of the US Preventive Services Task Force for determining internal validity of the studies evaluated.2

Of the 32 included studies 27 were of type 2 diabetes only, 2 were of types 1 & 2. No study assessed the effects of alcohol consumption on self-care, or complications other than coronary heart disease and retinopathy.

Only 6 studies assessed the effect of alcohol on glycemic control. All of these studies were small and had only fair ratings on the evaluation scale. (I was unable to obtain these publications for review.) It would be helpful to know the number and age of patients in the study.

Two of the studies assessed the effect of alcohol consumption on medication-related complications. One rated "good" assessed troglitazone related complications and one rated "fair" studied sulfonylurea complications. There were no studies in patients taking other agents used in the treatment of diabetes.

Recently, Avgaro and colleagues studied subjects with intravenous glucose tests with and without administration of 40 g vodka given in sips over a one-hour period of time.3 They measured insulin sensitivity and beta cell production of insulin. The study demonstrated a 17% and 55% improvement in insulin sensitivity in eight persons with and without type 2 diabetes. Circulating free fatty acids decreased in both sets of subjects but was statistically significant in only the diabetics. There was no measurable increase in beta cell production of insulin. This suggests that alcohol may mediate the improvement in insulin sensitivity and thus be relevant to its cardio-protective effects. These are acute effects, however, and there are no long term studies to indicate what moderate alcohol consumption does to blood pressure or renal function.

There are other studies, recently reported, indicating the potential for prolonged benefit of moderate alcohol consumption. Thamer and associates reported that adiponectin levels, which increase insulin sensitivity, were significantly increased in those who consumed moderate amounts of alcohol.4

Since c-peptide has many favorable effects when administered to type 1 diabetics, the effects of alcohol on type 1 patients, who do not have adequate c-peptide levels may be quite different than in type 2 diabetics who do have c-peptide present.5

The Howard et al review is valuable in that it addresses moderate alcohol intake. Most textbook citations refer primarily to excessive alcohol intake and to hypoglycemia. As Howard et al indicate, there needs to be better studies regarding alcohol and compliance and self management. It seems obvious, however, that diabetics who can’t control their alcohol intake are not likely to control other aspects of their lives.

Dr. Hall, Emeritus Professor of Medicine University of Missouri-Kansas City School of Medicine, is Associate Editor of Internal Medicine Alert.


1. 2001 National Household Survey on Drug Abuse, US Department of Health and Services; 2002. Publication no. SMA 02-3758.

2. Harris RP, et al. Am J Prev Med. 2001;20; 21-35.

3. Avogaro A, et al. Diabetes Care. 2004; 27: 1369-1374.

4. Thamer C, et al. Diabetes Care. 2004;27:1240.

5. Bloomgarden ZT. Diabetes Care. 2004;27:1506-1507.