Risk of diabetes and alcohol consumption
Original article:
Alcohol consumption and the risk of diabetes by body mass index levels in a
cohort of 5,636 Japanese. Watanabe M, Barzi F, Neal B, Ueshima H,
Miyoshi Y, Okayama A, Choudhury SH. Diabetes Res Clin Pract 2002; 57: 191–7.
Summary
In this study the authors followed 5636 employees of a Japanese insurance
company for a mean of 5.7 years. The mean age of the employees was 44 years
(range 35–68 years), 28% were female, 47% were regular smokers and 78%
regularly consumed alcohol.
During the follow-up period, 264 subjects developed diabetes, diagnosed by a
fasting plasma glucose ≥7.8 mmol/l. At baseline
all had fasting plasma glucose <7.8 mmol/l and no history of chronic
pancreatitis either at the start of the study or during follow-up. Alcohol
consumption was self-reported as no alcohol consumption, alcohol consumption
≤21 g/day, >21 to ≤42
g/day or >42 g/day. On the basis of baseline BMI the participants were
classified into low (≤22.0 kg/m2), middle
(22.1–24.9 kg/m2) and high (≥25.0 kg/m2)
categories of BMI (Table I).
Table I: Baseline characteristics of participants
overall and by BMI (Japan, 1991–1999).
In the overall analyses of outcome, factors that were significantly
associated with risk of diabetes were: female gender (RR = 0.57), baseline
fasting plasma glucose (RR = 6.98) and current tobacco use (RR = 1.50). In
the analyses, total cholesterol, HDL cholesterol, triglycerides and systolic
blood pressure were also significantly associated with risk of diabetes.
Among alcohol drinkers, compared with non-drinkers, the relative risk of
diabetes was 3.19 in the low BMI group but lower in the middle (0.41) and
high (0.74) BMI groups. These results were not changed when risk was
adjusted by excluding those with higher, but non-diabetic, fasting plasma
glucose levels at the start, or by correcting for possible confounders
(Table II).
Table II: Association of alcohol consumption with
risk of clinical diagnosis of diabetes or fasting plasma glucose
≥7.8 mmol/l overall and by BMI (Japan, 1991–1999).
Comment
This is an interesting report with a very surprising finding: that regular
alcohol consumption is associated with a higher risk of diabetes in lean
individuals but seems to protect against diabetes in the those with slightly
higher BMI. How can this be possible? It may, of course, be a chance
occurrence, as the authors suggest.
By European standards the body weights were rather low, and in the group
with the negative effect of alcohol on development of diabetes, BMI was at
or below 22 kg/m2. In this group 44% were women, whereas in the two higher
BMI groups fewer than 20% were women.
One would have wished for separate analyses for men and women, or at least
some comments with regard to gender. In a more recent publication from Japan
on males only, the authors found little difference in men below or above BMI
23.2 kg/m2 [1].
In the present study the baseline values were for mean BMI 22.9 kg/m2, total
cholesterol 5.3 mmol/l, HDL cholesterol 1.4 mmol/l, blood pressure 122/78
mmHg and mean age 44 years. Perhaps the Japanese are becoming more like
Europeans, at least when it comes to total cholesterol, if not BMI. Current
evidence for both men and women indicates that moderate alcohol consumption
probably has some positive health effects, which include prevention of type
2 diabetes. In information to the public and to people with diabetes, I have
expressed the following: middle-aged and elderly people without
contraindications to drinking may have a moderate health benefit of regular,
light-to-moderate alcohol consumption. A general view is that this
represents one to two units of alcohol daily, or almost daily (e.g. a glass
of wine). There seems to be some additional benefit from choosing wine, but
other kinds of alcoholic beverages also seem to have an effect.
There are no documented health benefits from binge drinking, and
particularly not for people with diabetes. For young people there are no
documented health benefits of alcohol use. On the contrary, the young are
often the victims of the unwanted and harmful effects of alcohol abuse, such
as road traffic accidents, violence, social problems, addiction and birth
defects.
Reference
1. Nakanishi N, Suzuki K, Tatara K. Alcohol consumption and risk for
development of impaired fasting glucose or type 2 diabetes in middle-aged
Japanese men. Diabetes Care 2003; 26: 48–54.
Summary and Comment:
Jak Jervell, Oslo, Norway
