Vous êtes sur la page 1sur 10

American Journal of Epidemiology Vol. 154, No.

8
Copyright © 2001 by the Johns Hopkins University Bloomberg School of Public Health Printed in U.S.A.
All rights reserved

Alcohol Consumption and the Risk of Diabetes Mellitus Kao et al.


Alcohol Consumption and the Risk of Type 2 Diabetes Mellitus
Atherosclerosis Risk in Communities Study

W. H. Linda Kao,1 Ian B. Puddey,2 Lori L. Boland,3 Robert L. Watson,4 and Frederick L. Brancati1,5

Evidence regarding the association between alcohol consumption and type 2 diabetes risk remains
inconsistent, particularly with regard to male-female differences. The authors conducted a prospective study of
type 2 diabetes risk associated with alcohol consumption in a cohort of 12,261 middle-aged participants of the
Atherosclerosis Risk in Communities Study (1990–1998), who were followed between 3 and 6 years. Alcohol
consumption at baseline was characterized into lifetime abstainers, former drinkers, and current drinkers of
various levels. Incident diabetes was determined by blood glucose measurements and self-report. After
adjustment for potential confounders, an increased risk of diabetes was found in men who drank >21
drinks/week when compared with men who drank ≤1 drink/week (odds ratio = 1.50, 95% confidence interval:
1.02, 2.20) while no significant association was found in women. This increased diabetes risk among men who
drank >21 drinks/week was predominantly related to spirits rather than to beer or wine consumption. The relative
odds of incident diabetes in a comparison of men who drank >14 drinks of spirits per week with men who were

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011


current drinkers but reported no regular use of spirits, beer, or wine were 1.82 (95% confidence interval: 1.14,
2.92). Results of this study support the hypothesis that high alcohol intake increases diabetes risk among
middle-aged men. However, more moderate levels of alcohol consumption do not increase risk of type 2
diabetes in either middle-aged men or women. Am J Epidemiol 2001;154:748–57.

alcohol drinking; cohort studies; diabetes mellitus, non-insulin-dependent; prospective studies

Type 2 diabetes mellitus affects over 15 million tion of pancreatitis, disturbance of carbohydrate and glucose
Americans and leads to excess risk of cardiovascular dis- metabolism, and impairment of liver function (5–7).
eases and other morbidities (1–3). Aside from obesity and However, the association between alcohol and diabetes or
physical inactivity, there are few other well-established other diabetes-related physiologic endpoints has been incon-
modifiable risk factors for type 2 diabetes. Alcohol con- sistent in previous small clinical studies (8–16), animal stud-
sumption represents a potentially important, modifiable risk ies (17, 18), and larger cross-sectional epidemiologic stud-
factor of type 2 diabetes, especially given that more than ies (5, 19–24).
half of the adults in the United States are current drinkers Prospective epidemiologic studies have yielded similarly
(4). equivocal results; furthermore, results of these studies sug-
The hypothesized diabetogenic effects of alcohol include gest a potential difference in the association of alcohol and
its contribution to excess caloric intake and obesity, induc- incident type 2 diabetes between women and men (25–32).
The Atherosclerosis Risk in Communities (ARIC) Study
Received for publication March 21, 2000, and accepted for publi- offers a unique opportunity to examine the association
cation May 10, 2001. between alcohol consumption and the risk of type 2 diabetes
Abbreviation: ARIC, Atherosclerosis Risk in Communities. mellitus in a large community-based cohort of middle-aged
1
Department of Epidemiology, Bloomberg School of Public
Health, Johns Hopkins University, Baltimore, MD.
women and men.
2
Department of Medicine, University of Western Australia, Perth,
Australia. MATERIALS AND METHODS
3
Division of Epidemiology, School of Public Health, University of
Minnesota, Minneapolis, MN. Study population
4
Division of Epidemiology, University of Mississippi Medical
Center, Jackson, MS.
5 The ARIC Study is an ongoing prospective study that
Department of Medicine, School of Medicine, Johns Hopkins
University, Baltimore, MD. examines clinical and subclinical atherosclerotic diseases in
Reprint requests to Dr. Frederick L. Brancati, Welch Center for a cohort of 15,792 persons, aged 45–64 years at baseline
Prevention, Epidemiology, and Clinical Research, the Johns Hopkins examination, selected by probability sampling from four US
Medical Institutions, 2024 East Monument St., Suite 2-600, communities. The sampling procedure and methods used in
Baltimore, MD 21205 (e-mail: fbrancat@welch.jhu.edu).
This study was presented in part at the 31st Annual Meeting of the ARIC Study have been described in detail elsewhere
the Society for Epidemiologic Research, Chicago, IL, June 24–26, (33). This analysis was based on information obtained over
1998. 6 years of follow-up, which included two clinic visits sched-

748
Alcohol Consumption and the Risk of Diabetes Mellitus 749

uled at 3 (visit 2) and 6 (visit 3) years after baseline. We history of diabetes was defined as having either biologic par-
excluded participants who reported ethnicity other than ent with diabetes. Body mass index (weight (kg)/height (m)2)
Black or White (n  48), had diabetes at visit 1 (n  1,867), and waist/hip ratio were determined by the anthropometric
had missing exposure information (n  89), had missing measurements taken at the baseline clinic visit.
diabetes information at visit 1 or were lost to follow-up Measurements were made with the participants wearing
before visit 2 of the study took place (n  1,022), and had light-weight, nonconstricting underwear and no shoes. Waist
missing information on the potential confounders (n  315), and hip measurements were taken at the level of the umbili-
as well as persons in whom total caloric intake was within cus and the level of maximal protrusion of the gluteal mus-
the top 1 percent or bottom 1 percent of the entire cohort due cles, respectively. Intrareader and interreader correlations
to probable reporting error (n  190). After these exclu- between repeated waist/hip ratio measures were 0.94 and
sions, 12,261 participants (6,838 women and 5,423 men) 0.91, respectively (34). Physical activity was assessed using
remained. a modified interviewer-administered version of the question-
naire developed by Baecke et al. (35). Results from the
questionnaire concerning physical activity were further con-
Exposure assessment densed to a sport-related physical activity index with scores
ranging from 1 to 5, with 1 indicating the lowest level of
Alcohol consumption was assessed using five questions.
activity and 5 the highest level. Total caloric intake was
The following two questions were used to determine the
derived from an interviewer-administered, modified version
current drinking status of the persons: “Do you presently
of the 61-item food frequency questionnaire developed by
drink alcoholic beverages?” and “Have you ever consumed
Willett et al. (36). Smoking status was categorized into never,
alcoholic beverages?” Persons were classified as lifetime
former, and current smokers. Hypertension was defined by
abstainers if they answered “no” to both questions. Persons
the presence of any of the following: 1) systolic blood pres-
who answered “no” to the first question and “yes” to the sec-
sure of ≥140 mmHg, 2) diastolic blood pressure of ≥90
ond question were classified as former drinkers. Persons

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011


mmHg, or 3) current use of antihypertensive medication.
who answered “yes” to both questions were considered cur-
rent drinkers. Among current drinkers, the following three
questions were used to determine the amount and type of Outcome assessment
alcoholic beverage consumed: “How many glasses of wine
do you usually have per week (4-ounce glasses)?”; “How Diabetes mellitus was defined as the presence of any one
many bottles or cans of beer do you usually have per week of the following: 1) fasting glucose of ≥7.0 mmol, 2) non-
(12-ounce bottles or cans)?”; and “How many drinks of hard fasting glucose of ≥11.1 mmol, 3) current use of diabetic
liquor do you usually have per week (1.5-ounce shots)?” medication, or 4) a positive response to the question, “Has a
(One ounce  29.57 ml.) For the primary analyses, ethanol, doctor ever told you that you had diabetes (sugar in the
rather than a specific beverage, was the main independent blood)?” Persons with diabetes or who had unknown dia-
variable. Weekly ethanol consumption was derived from the betes status at baseline were excluded from the prospective
responses to the three beverage questions using the follow- analyses. Persons without diabetes at baseline who met any
ing conversion factors: 4 ounces of wine  10.8 g of of the these conditions at visit 2 or visit 3 were considered
ethanol, 12 ounces of beer  13.2 g of ethanol, and 1.5 incident cases of diabetes (n  239). Persons who met the
ounces of spirits  15.1 g of ethanol. To reclassify current criteria for diabetes at visit 2 but not at visit 3 were nonethe-
drinkers by the number of drinks (nonspecific to type of less considered to have incident type 2 diabetes.
alcohol) consumed per week, we assumed one generic drink
to be equal to 12 g of ethanol. Seven consumption groups
Statistical analysis
were created: lifetime abstainers, former drinkers, current
drinkers who consume <1 drink/week (reference group), All analyses were stratified by gender. The means and fre-
current drinkers who consume 1.1–7 drinks/week, current quencies of potential confounders assessed at baseline were
drinkers who consume 7.1–14 drinks/week, current drinkers calculated for each group of alcohol consumption, and
who consume 14.1–21 drinks/week, and current drinkers analysis of variance and chi-square analysis were used to
who consume more than 21 drinks/week. assess the statistical significance of the differences across
For alcohol-specific analyses, derivation of weekly consumption groups. For continuous variables, p values for
ethanol consumption from the three beverages was not used test for trend among current drinkers were also reported.
(method described in the previous paragraph). Instead, the Incidence rates were calculated for each consumption group
original number of glasses, bottles, and drinks was used to with the use of a person-years approach, and a Poisson
categorize wine, beer, and spirit consumption, respectively. regression model was fitted to determine if the incidence
For each beverage type, persons who identified themselves rates for the seven consumption groups differed signifi-
as current drinkers but reported an average weekly con- cantly. Because the outcome was assessed only at 3-year
sumption of zero servings of that beverage were used as the intervals, the time of follow-up for incident cases was
reference group. assigned to the midpoints between visits, that is, 1.5 and 4.5
Information on age, gender, race, family history of dia- years. For example, persons who were nondiabetic at base-
betes, and education was obtained from home and clinic line but became diabetic at visit 3 were considered to have
interviews conducted at the baseline visit. A positive family been followed for 4.5 years. Nondiabetic persons at both

Am J Epidemiol Vol. 154, No. 8, 2001


750 Kao et al.

visit 1 and visit 2 who were lost to follow-up after visit 2

0.001

<0.001
<0.001
were censored at 3 years after the baseline visit.

trend†

0.10

0.05
0.87

0.95
p
The independent association between baseline alcohol

TABLE 1. Means (SDs*) and frequencies of selected baseline characteristics of 6,838 female ARIC* Study participants by alcohol consumption group, 1990–1998
consumption and subsequent incident diabetes was exam-
ined using logistic regression models. Alcohol consumption

<0.001
0.001

0.001

0.001
<0.001
<0.001
<0.001
<0.001

0.001

0.001

<0.001
overall

0.04
p
was treated both nominally and continuously (for trend
among current drinkers). Multiple linear regression analysis
was also performed to examine the association between

0.914 (0.069)

59.58 (38.58)
25.61 (4.73)

2.16 (0.83)

5.42 (0.56)
baseline alcohol consumption and fasting glucose levels at

53.3 (5.3)
347 (107)

1,793 (630)
visit 2 or visit 3 among persons who did not have diagnosed

58 (0.8)
>21

15.5

19.0
50.0
31.0
19.0

60.3
25.9
13.8
32.8
diabetes by the time of the visit. All statistical analyses were
performed using a SAS statistical package (Cary, North
Carolina) (37).

Current drinkers by alcohol consumption group (drinks/week)

0.878 (0.072)

54.52 (40.25)
24.75 (4.09)

2.41 (0.85)

5.45 (0.47)
14.1–21

53.3 (6.1)
RESULTS

1,490 (512)
130 (1.9)
202 (22)

6.2

7.7
51.5
40.8
16.2

44.6
36.2
19.2
30.8
Baseline characteristics

As shown in table 1, among the 6,838 female ARIC


Study participants, almost one third reported that they were

0.872 (0.078)

56.70 (37.11)
24.76 (4.46)

2.45 (0.77)

5.43 (0.49)
lifetime abstainers, and roughly half identified themselves

52.9 (5.6)
7.1–14

1,545 (502)
395 (5.8)
as current drinkers, of whom roughly half reported an aver-

117 (25)
age weekly consumption of ≤1 drink/week. Both lifetime

11.6

9.1
44.8
46.1
20.8

41.8
31.1
27.1
25.8

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011


abstainers and former drinkers tended to be older, less
physically active, and more obese; they were more similar

* SD, standard deviation; ARIC, Atherosclerosis Risk in Communities; BMI, body mass index (weight (kg)/height (m)2).
to each other than to current drinkers. There was also a

0.874 (0.080)

64.77 (49.00)
26.02 (4.89)

2.48 (0.80)

5.39 (0.47)
greater proportion of Blacks, positive family history of dia-
53.2 (5.6)
1,316 (19.2)
1.1–7

1,487 (522)
betes, and hypertension among lifetime abstainers and for-
40 (20)

15.5

11.2
47.0
41.8
20.4

29.2
31.3
39.5
24.1
mer drinkers. Among current drinkers, total caloric intake
and prevalence of hypertension were positively associated
with alcohol consumption, while educational level, family
history of diabetes, and body mass index were inversely

0.880 (0.078)

68.95 (52.24)
26.54 (5.17)

2.41 (0.76)

5.36 (0.47)
associated with alcohol consumption. Persons who drank
53.2 (5.6)
1,736 (25.4)

1,484 (546)
more than 21 drinks per week had a noticeably higher
≤1

1 (3)

waist/hip ratio and total caloric intake and less physical


7.5

10.6
49.9
39.5
22.5

22.4
24.9
52.7
23.2
activity. As expected, smoking and alcohol consumption
were highly associated. Although the fasting serum glucose
level was slightly positively associated with alcohol con-
0.902 (0.076)

86.47 (70.87)
28.32 (6.42)

2.29 (0.76)

5.40 (0.51)
sumption among current drinkers, the fasting serum insulin
drinkers
Former

53.9 (5.6)
980 (14.3)

1,549 (598)

level was inversely associated with alcohol consumption in


this group.
32.6

32.8
42.4
24.8
26.9

29.0
30.5
40.5
35.5

Only about 10 percent of the 5,423 male ARIC Study


participants were lifetime abstainers, and approximately
70 percent were current drinkers (table 2). Among current
0.895 (0.079)

87.48 (64.94)
28.48 (6.14)

2.24 (0.70)

5.39 (0.53)

drinkers, men were drinking more drinks per week than


abstainers
Lifetime

54.1 (5.8)
2,223 (32.5)

1,498 (549)

were women, with the amount consumed by men slightly


† Limited to trend among current drinkers.

higher for each consumption group. The most striking


44.9

28.0
44.1
27.9
26.6

14.7
12.2
73.1
38.6

‡ Based on 6,687 fasting participants.

male-female difference was among the heavier drinkers: 8


percent of men drank >21 drinks/week with a mean intake
of 402 g of alcohol per week, whereas less than 1 percent
Family history of diabetes (%)
Ethanol consumed (g/week)

of women drank >21 drinks/week with a mean intake of


Energy intake (kcal/day)

Serum glucose‡ (mmol)


High school graduate

347 g/week. Despite the disparity in ethanol intake, the


Physical activity score

Serum insulin‡ (pmol)


Characteristic

associations between alcohol consumption and most


Attended college
11 years or less

Hypertension (%)

potential confounders were similar in men and women.


Waist/hip ratio
Education (%)

Smoking (%)
BMI* (kg/m2)

Men who drank >21 drinks/week had the highest


Age (years)

Current
Former
Black (%)

waist/hip ratio and energy intake and the lowest physical


Never
No. (%)

activity score. The proportions of current smokers and


persons with hypertension were also highest in this group
of men.

Am J Epidemiol Vol. 154, No. 8, 2001


Alcohol Consumption and the Risk of Diabetes Mellitus 751

Incident type 2 diabetes and alcohol consumption

<0.001

<0.001
<0.001
trend†

0.79

0.04

0.01
Type 2 diabetes incidence rates of the seven alcohol con-
p

sumption groups differed significantly in women (p < 0.001)


and men (p  0.08) (table 3). In women, the highest inci-
TABLE 2. Means (SDs*) and frequencies of selected baseline characteristics of 5,423 male ARIC* Study participants by alcohol consumption group, 1990–1998

<0.001
0.001

0.001

<0.001
<0.001
<0.001

0.001

0.001
<0.001
<0.001
overall

0.59
0.27
dence rates were found in lifetime abstainers and former
p

drinkers (19.88 and 21.04 per 1,000 person-years, respec-


tively). Among current drinkers, the incidence rates

0.966 (0.052)

72.62 (57.23)
decreased from 11.40 per 1,000 person-years to 6.41 per
26.88 (4.30)

2.46 (0.83)

5.67 (0.59)
1,000 person-years as alcohol consumption increased. The
54.0 (5.7)
402 (182)

1,992 (635)
439 (8.1)
>21

association between alcohol consumption and diabetes risk


19.4

17.3
43.5
39.2
20.0

47.6
42.1
10.3
38.3
appeared inverse and by grade, but the incidence rates of the
higher consumption groups (>14 drinks/week) were based
on very few cases. In men, although the risks for lifetime
Current drinkers by alcohol consumption group (drinks/week)

0.961 (0.050)

abstainers and former drinkers were similar to each other, to

74.03 (51.34)
27.18 (4.20)

2.78 (0.85)

5.68 (0.51)
the reference group (≤1 drink/week), and to their female
14.1–21

54.4 (5.6)

1,795 (586)
313 (5.8)
205 (24)

counterparts (19.16 and 20.69 per 1,000 person-years,


15.0

16.0
31.3
52.7
19.8

32.3
54.0
13.7
29.7

respectively), the risk pattern of current male drinkers was


quite different from that of current female drinkers. Among
male current drinkers, the risk of diabetes appeared to be
0.959 (0.052)

generally flat below 21 drinks/week but increased above the


72.79 (47.58)
27.00 (3.69)

2.74 (0.80)

5.61 (0.48)

threshold of 21 drinks/week. In fact, men who drank >21


54.3 (5.7)
7.1–14

725 (13.4)

1,723 (589)
124 (25)

drinks/week had the highest risk of diabetes (26.56 per


15.6

15.4
35.2
49.4
20.0

27.2
49.4
23.4
30.5

1,000 person-years).

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011


To determine the independent relation of alcohol con-
sumption and incident type 2 diabetes, we adjusted for age,
* SD, standard deviation; ARIC, Atherosclerosis Risk in Communities; BMI, body mass index (weight (kg)/height (m)2).
0.958 (0.052)

race, education, family history of diabetes, body mass index,


77.73 (55.58)
27.24 (3.67)

2.68 (0.83)

5.58 (0.48)

waist/hip ratio, physical activity, total caloric intake, smok-


53.9 (5.8)
1,366 (25.2)
1.1–7

1,742 (621)
43 (22)

ing history, and history of hypertension using logistic


15.8

14.1
37.3
48.6
22.2

24.3
46.9
28.8
26.7

regression. In women, the inverse association between alco-


hol intake and incident diabetes persisted (table 4), but the
trend was no longer statistically significant.
0.960 (0.052)

In men, the lack of association between alcohol con-


82.43 (56.88)
27.29 (3.84)

2.67 (0.82)

5.55 (0.48)

sumption of ≤21 drinks/week and diabetes risk persisted


54.2 (5.8)
815 (15.0)

1,719 (653)

after adjusting for potential confounders. The increased risk


≤1

1 (3)

7.7

14.1
36.6
49.3
21.6

20.0
43.4
36.6
27.2

of diabetes seen in men who drank >21 drinks/week also


persisted in the fully adjusted model (odds ratio  1.50, 95
percent confidence interval: 1.02, 2.20) (table 4).
0.962 (0.057)

We tested for effect modification by sex for statistical sig-


84.19 (58.82)
27.38 (4.37)

2.52 (0.81)

5.57 (0.51)

nificance among persons who identified themselves as current


54.8 (5.6)
drinkers

1,096 (20.2)
Former

1,749 (661)

drinkers. In the unadjusted model, the interaction term was


23.9

33.2
37.0
29.8
20.7

28.7
50.8
20.5
30.5

statistically significant (p  0.04). Although this interaction


was no longer statistically significant in the fully adjusted
model (p  0.08), the divergence of the fully adjusted relative
odds of incident diabetes in the higher consumption groups
0.953 (0.053)

85.63 (62.52)
27.26 (4.10)

2.47 (0.78)

5.52 (0.49)

between women and men remained apparent.


abstainers
Lifetime

55.1 (5.5)
669 (12.3)

1,706 (580)

† Limited to trend among current drinkers.


32.4

29.3
36.2
34.5
23.6

12.7
27.6
59.6
34.5

Subsidiary analyses
‡ Based on 5,281 fasting participants.

To test the robustness of our results, three subsidiary


analyses were conducted. The first set of analyses included
Family history of diabetes (%)
Ethanol consumed (g/week)

further adjustment for fasting serum insulin and glucose lev-


Energy intake (kcal/day)

Serum glucose‡ (mmol)


High school graduate

els at baseline because both variables were associated with


Physical activity score

Serum insulin‡ (pmol)


Characteristic

Attended college

alcohol consumption at baseline (table 4). Adjustment for


11 years or less

Hypertension (%)

the fasting serum insulin level at baseline slightly strength-


Waist/hip ratio
Education (%)

Smoking (%)
BMI* (kg/m2)

ened the association between alcohol consumption and the


Age (years)

Current
Former
Black (%)

risk of diabetes in men (male drinkers had higher insulin


Never
No. (%)

levels compared with the reference group at baseline) but


slightly weakened the association in women (female
drinkers at lower insulin levels compared with the reference

Am J Epidemiol Vol. 154, No. 8, 2001


752 Kao et al.

TABLE 3. Incidence of type 2 diabetes in the 12,261 ARIC* Study participants by sex and alcohol consumption groups,
1990–1998

Lifetime Former Current drinkers by alcohol consumption group (drinks/week) p


Variable
abstainers drinkers ≤1 1.1–7 7.1–14 14.1–21 >21 value†

Women
Incident diabetes (no.) 236 108 110 90 18 5 2
Total person-years 11,868 5,133 9,651 7,275 2,196 728 312
Incidence rate 19.88 21.04 11.40 12.37 8.20 6.87 6.41 <0.001
Rate ratio 1.74 1.85 1.00 (reference) 1.08 0.72 0.62 0.56

Men
Incident diabetes (no.) 69 118 74 139 55 32 60
Total person-years 3,602 5,703 4,353 7,352 3,959 1,665 2,259
Incidence rate 19.16 20.69 17.00 18.91 13.89 19.22 26.56 0.08
Rate ratio 1.13 1.22 1.00 (reference) 1.11 0.82 1.13 1.56
* ARIC, Atherosclerosis Risk in Communities.
† p value for overall difference.

group at baseline). In contrast, adjustment for the fasting visits 2 and 3; however, neither association was statistically
serum glucose level at baseline weakened the association significant at the 0.05 level.
between alcohol consumption and the risk of diabetes in The last set of analyses was conducted using continuous
men but strengthened the association in women. Because the serum glucose at either visit 2 or visit 3 as the outcome among

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011


baseline fasting insulin level is hypothesized to be in the current drinkers who did not have diagnosed diabetes, who
causal pathway and the baseline glucose level is part of the were not taking antidiabetic medication at the time of the
outcome, models 2 and 3 of table 4 likely represent overad- visit, and who were nondiabetic at the previous visit. The
justed models. fully adjusted mean fasting glucose levels at either visit 2 or
The second set of analyses examined the association visit 3 were similar among the five consumption groups in
between alcohol consumption at baseline and incident dia- women; however, the adjusted mean fasting glucose levels
betes at visit 2 or at visit 3 separately. This set of analyses were different among the five consumption groups in men. At
was performed to determine if the association between visit 2, the mean glucose levels (mg/dl) for the five consump-
alcohol consumption and diabetes risk was dependent on tion groups in men were 104.8 (≤1 drink/week), 104.4 (1.1–7
follow-up time. The increased risk of diabetes for men who drinks/week), 104.3 (7.1–14 drinks/week), 105.8 (14.1–21
consumed >21 drinks/week at baseline was present at both drinks/week), and 106.5 (>21 drinks/week) (p  0.02 for

TABLE 4. Adjusted relative odds (95% confidence intervals) of incident type 2 diabetes mellitus in a comparison of alcohol
consumption groups in 12,261 ARIC* Study participants by sex, 1990–1998

Alcohol Women Men


consumption
(drinks/week) Model 1† Model 2‡ Model 3§ Model 1 Model 2 Model 3

>21 0.41 (0.10, 1.77) 0.52 (0.12, 2.26) 0.33 (0.08, 1.48) 1.50 (1.02, 2.20) 1.64 (1.10, 2.42) 1.22 (0.80, 1.85)
14.1–21 0.64 (0.25, 1.64) 0.77 (0.29, 2.00) 0.54 (0.20, 1.43) 1.07 (0.68, 1.69) 1.19 (0.75, 1.88) 0.91 (0.56, 1.48)
7.1–14 0.81 (0.47, 1.37) 0.88 (0.52, 1.51) 0.63 (0.36, 1.11) 0.80 (0.55, 1.17) 0.88 (0.60, 1.28) 0.84 (0.56, 1.25)
1.1–7 1.09 (0.80, 1.49) 1.14 (0.84, 1.56) 0.98 (0.71, 1.36) 1.12 (0.82, 1.52) 1.15 (0.84, 1.57) 1.12 (0.81, 1.56)
≤1 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)

Former drinkers 1.10 (0.81, 1.49) 1.09 (0.79, 1.48) 1.18 (0.85, 1.65) 1.06 (0.77, 1.47) 1.06 (0.76, 1.46) 1.06 (0.75, 1.51)
Lifetime abstainers 1.10 (0.84, 1.43) 1.08 (0.82, 1.43) 1.15 (0.86, 1.54) 1.14 (0.79, 1.65) 1.09 (0.75, 1.58) 1.18 (0.80, 1.76)

p for overall¶ 0.53 0.82 0.10 0.13 0.13 0.56


p for trend# 0.11 0.35 0.02 0.16 0.04 0.65
* ARIC, Atherosclerosis Risk in Communities.
† Model 1 adjusted for age (≤49 vs. 49.1–53.9 vs. 54–58.9 vs. ≥59), race (binary), education (11 years or less vs. high school graduate
vs. attended college), family history of diabetes (binary), body mass index (quartiles), waist/hip ratio (quartiles), physical activity score (con-
tinuous), total energy intake (quartiles), smoking history (never vs. former vs. current), and history of hypertension (binary) using logistic
regression.
‡ Model 2 = model 1 + fasting serum insulin level (quartiles) at baseline.
§ Model 3 = model 2 + fasting serum glucose level (quartiles) at baseline.
¶ p for overall is the p value of the test for overall significance of alcohol consumption groups on a nominal scale.
# p for trend is the p value of the test for linear trend among current drinkers.

Am J Epidemiol Vol. 154, No. 8, 2001


Alcohol Consumption and the Risk of Diabetes Mellitus 753

overall difference). The corresponding numbers from visit 3 current drinkers but reported no regular weekly consump-
were 100.0, 101.2, 101.2, 101.2, and 102.8 (p  0.06 for tion of wine, beer, or spirits. Table 5 shows the adjusted rel-
overall difference). At both visits 2 and 3, men who drank >21 ative odds of diabetes from multiple logistic regression,
drinks/week at baseline had the highest fasting serum glucose which adjusted for intakes of other types of alcoholic bever-
level at subsequent visits. ages; that is, nine dummy variables were included in the
regression to adjust for intake of other alcoholic beverages,
Consumption of specific alcoholic beverages and the in addition to the potential confounders mentioned previ-
risk of type 2 diabetes ously. In the fully adjusted model, persons who drank >14
drinks of spirits per week were still at higher risk for devel-
To assess whether the discrepancy between women and oping diabetes (odds ratio  1.82, 95 percent confidence
men might have been explained by differences in alcoholic interval: 1.14, 2.92). There was no statistically significant
beverage preference, we characterized the beverage preference association between wine intake and the risk of diabetes at
of current drinkers. Figure 1 shows the mean percentage of the conventional 0.05 level.
ethanol contributed by specific alcoholic beverages for each One possible explanation for the excess risk associated
consumption group (total alcohol intake) in women and men. with spirits might be differences in underlying health behav-
In the reference groups (≤1 drink/week), although all persons iors as compared with wine drinkers, particularly with
identified themselves as current drinkers, only 11 percent of regard to diet. To investigate this possibility, we examined
the women and 8 percent of the men reported regular weekly the correlation between the percentage of ethanol con-
alcohol consumption (the remaining persons in the reference tributed by each alcoholic beverage and measures that were
group identified themselves as current alcohol drinkers but thought to indicate healthy behavior. In general, the
reported no regular weekly alcohol consumption), and in those increased percentage of ethanol from wine was associated
persons who reported regular use of alcohol, wine accounted with healthier behaviors, such as increased physical activity
for 100 percent for ethanol consumed. In general, the contri- and total intake of dietary fiber and decreased total fat

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011


bution to total alcohol intake from wine was roughly two times intake. Although these associations were statistically sig-
higher in women than in men, the contribution from beer was nificant, the magnitude of the correlations was small ( 0 r 0 <
about two times higher in men than in women, and the contri- 0.2). Additional adjustment for dietary intakes of carbo-
bution from spirits was similar between women and men. hydrate, total fat, and protein did not alter the association
Next, we examined the association between incident type between a high intake of spirits (>14 drinks/week) and the
2 diabetes and each of the three alcoholic beverages, spirits, risk of diabetes in men (results not shown).
beer, and wine, separately. Since the amount of alcohol con-
sumed by female ARIC Study participants was relatively DISCUSSION
small, analysis by alcohol type was limited to the male par-
ticipants. In each beverage analysis, the reference group was These data support the following conclusions. First, after
composed of male participants who identified themselves as adjustment for potential confounders, middle-aged women

FIGURE 1. Mean percentage of ethanol contributed by specific alcoholic beverage in 12,261 participants by sex and consumption group,
Atherosclerosis Risk in Communities Study, 1990–1998.

Am J Epidemiol Vol. 154, No. 8, 2001


754 Kao et al.

TABLE 5. Adjusted* relative odds (95% confidence intervals) of incident type 2 diabetes mellitus in
5,423 male ARIC† Study participants by alcoholic beverage type and alcohol consumption groups,
1990–1998

Servings of
alcohol Spirits Beer Wine
per week‡

>14 servings/week 1.82 (1.14, 2.92) 1.19 (0.77, 1.85) 0.63 (0.08, 4.95)
8–14 servings/week 1.02 (0.67, 1.56) 0.74 (0.48, 1.14) 0.84 (0.29, 2.44)
1–7 servings/week 1.00 (0.77, 1.29) 0.88 (0.68, 1.13) 1.10 (0.82, 1.50)
None 1.00 (reference) 1.00 (reference) 1.00 (reference)

p for overall 0.11 0.27 0.86


* Adjusted for age (≤49 vs. 49.1–53.9 vs. 54–58.9 vs. ≥59), race (binary), education (11 years or less vs. high
school graduate vs. attended college), family history of diabetes (binary), body mass index (quartiles), waist/hip
ratio (quartiles), physical activity score (continuous), total energy intake (quartiles), smoking history (never vs. for-
mer vs. current), history of hypertension (binary), spirits consumption (none vs. 1–7 servings/week, 8–14 serv-
ings/week, >14 servings/week), wine consumption (none vs. 1–7 servings/week, 8–14 servings/week, >14 serv-
ings/week), and beer consumption (none vs. 1–7 servings/week, 8–14 servings/week, >14 servings/week) using
logistic regression.
† ARIC, Atherosclerosis Risk in Communities.
‡ One serving of spirits is one 1.5-ounce glass, one serving of beer is one 12-ounce can, and one serving of
wine is one 4-ounce glass. One ounce = 29.57 ml.

and men who consumed a moderate amount of alcohol edge, this study is the first study to present a more compre-

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011


(1–14 drinks/week) were not at higher risk of developing hensive examination of the associations between different
type 2 diabetes compared with their counterparts who drank alcoholic beverages and the risk of diabetes.
little alcohol (≤1 drink/week). Second, after adjustment for Both the present study and a previous ARIC Study publi-
potential confounders and established diabetes risk factors, cation by Vitelli et al. (39) reported an inverse cross-
middle-aged men who consumed a substantial amount of sectional relation between the baseline alcohol consumption
alcohol (>21 drinks/week) were about 50 percent more and the baseline fasting serum insulin level in both men and
likely to develop type 2 diabetes compared with their coun- women, and yet the present study also found a direct cross-
terparts who drank ≤1 drink/week. Third, this increased risk sectional association of the baseline alcohol consumption
of diabetes among men who drank >21 drinks/week was with the baseline fasting serum glucose level and a direct
predominantly related to spirits consumption. The risk of prospective relation with the risk of diabetes during follow-up
diabetes among middle-aged men who drank >14 drinks of in men. It is not clear why these contradictory associations
spirits per week was about 80 percent higher than the risk of exist. One hypothesis for these apparently contradictory
men who were current drinkers but did not drink any spirits. results in men is that the inverse association between the
There are two major limitations to this study. First, our baseline alcohol consumption and the baseline fasting serum
analyses of heavy consumption and alcohol beverage types insulin level could be indicative of the beginning of the
were limited to men only. This study did not have sufficient deterioration of beta-cell function experienced by pre-
power to estimate the relative odds of diabetes at higher lev- diabetic persons rather than insulin sensitivity; hence, a
els of alcohol intake in women, specifically consumption of direct association of the baseline alcohol consumption and
≥14 drinks/week. Second, like most previous studies in this the baseline fasting serum glucose level and incident dia-
field, our assessment of alcohol was likely suboptimal for betes during follow-up was observed.
several reasons. The assessment was done at a single time The divergent results from epidemiologic studies of alco-
point and did not evaluate drinking pattern. Participants, hol consumption and the risk of diabetes are reflected in
especially those in the heavier drinking groups, may have nonepidemiologic studies, which have also reported both
underreported their consumption level because the questions diabetogenic and nondiabetogenic effects of ethanol on
regarding alcohol use were administered by an interviewer. physiologic endpoints associated with diabetes. In vitro
Nevertheless, this study had several strengths. First, this studies indicate that exposure of beta cells to alcohol is asso-
is the only community-based, prospective study of the asso- ciated with decreased insulin secretion (40, 41). On the
ciation between alcohol and the risk of type 2 diabetes with other hand, hepatic oxidation of ethanol results in an
a sample that included Whites, Blacks, women, and men. increased ratio of reduced nicotinamide adenine dinu-
Second, the present analyses distinguished among lifetime cleotide to nicotinamide adenine dinucleotide, which can
abstainers, former drinkers, and current drinkers who drink lead to impairment in gluconeogenesis (42). In physiologic
minimally. It has been suggested that persons who are not studies, acute ethanol ingestion has been shown to improve
currently drinking may be abstaining from drinking alcohol glucose tolerance in normal subjects and insulin sensitivity
because of health reasons, and that the use of such persons in diabetic subjects (11, 16); however, acute ethanol admin-
as the reference group could potentially lead to an artifactual istration has been shown to cause a reduction in the glucose
protective effect of alcohol use (38). Third, to our knowl- disposal rate, cause acute insulin resistance in nondiabetic

Am J Epidemiol Vol. 154, No. 8, 2001


Alcohol Consumption and the Risk of Diabetes Mellitus 755

subjects, and worsen insulin resistance in diabetic patients, risk of diabetes seen in ARIC Study male participants who
while chronic alcohol consumption has also been implicated drank >21 drinks/week is consistent with results reported by
to induce beta-cell dysfunction (10, 14, 15). Similarly con- two of these studies, the Rancho Bernardo Study and the
tradictory results arise from animal studies. In rats, acute San Antonio Heart Study (25, 31). However, an inverse
alcohol administration can attenuate basal and hormone- association was found in the other two studies, the Health
induced glucose utilization by skeletal muscles while, in Professionals’ Follow-up Study and the British Regional
obese mice, chronic ethanol consumption can lower blood Heart Study (26, 30). Given the unique nature of the Health
glucose and insulin (17, 18). The exact relation between Professionals’ cohort, it is conceivable that these persons
alcohol and diabetes-related endpoints may depend on the had different beverage preference and drinking patterns
nature of alcohol exposure, that is, acute versus chronic and from those of ARIC Study men or that they had healthier
alcohol administration versus alcohol ingestion. behaviors, which were not measured and were associated
Since 1965, eight prospective studies have examined the with increased consumption of alcohol. Furthermore, the
association between alcohol consumption and the risk of type association between the waist/hip ratio and alcohol con-
2 diabetes (25, 26, 28–32, 43, 44). Five are comparable with sumption, which was positive in the ARIC Study, was
the present study, while significant differences in the study unknown in the Health Professionals’ Study and therefore
methodology and study population exist between the other was not included in their multivariate analyses. The main
three studies and the present study. Of these three, two difference between the present study and the British
showed no association between alcohol intake and the risk of Regional Heart Study may lie in the populations studied.
diabetes in men even at >21 drinks/week (29, 43). We believe The prevalence of coronary heart disease was reported to be
that two methodological issues could explain the discrepancy 24 percent in the British Regional Heart Study but only
between our study and these two studies. First, both previous about 8 percent in male ARIC Study participants at baseline.
studies lacked information on potential confounders, such as Because the lower risk of diabetes seen in moderate drinkers
the waist/hip ratio, education, and hypertension, all of which was more apparent and significant only in men with preex-

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011


were significantly associated with both alcohol and diabetes isting coronary heart disease versus men without evidence
in our study. Second, these studies had data on current drink- of coronary heart disease in the British Regional Heart
ing status only and so had to combine former drinkers, life- Study, one cannot exclude the possibility that the reference
time abstainers, and occasional drinkers into a single refer- group in the British Regional Heart Study (current occa-
ence group. In the present study, former drinkers and lifetime sional drinkers) was contaminated with persons with preva-
abstainers had the highest unadjusted risk of diabetes; thus, lent coronary heart disease who may have been advised to
such misclassification of exposure status may lead to an over- reduce their alcohol consumption.
estimation of the risk of diabetes in the reference group and Several reasons can possibly explain the differences in
produce a bias toward the null. The last of these three studies risk noted between women and men in this study. First,
found a protective association of moderate alcohol among the women and men may have differing responses to dietary
elderly (44). Differences in the age distribution of the two questionnaires. Previous studies have shown differences
populations and, hence, in the inherent risks for developing between men and women in their attitudes toward alcohol
diabetes and differences in the characterization of alcohol use use (45). If women who consumed high amounts of alcohol
(moderate use was defined as 0.5–<1 ounce/day) may explain were more likely to develop diabetes but were also more
the discrepancy with our results. likely to underreport the amount of alcohol they consumed,
Of the five studies that were comparable with the present then this could result in an apparent protective effect of alco-
study regarding methodology and population, three studies hol on diabetes risk. We have no knowledge of previous
presented results regarding women (25, 28, 31, 32), and in studies’ using questionnaires similar to ours that indicated
these studies, either an inverse or no association between that such underreporting occurs; however, it has been
alcohol intake and diabetes risk was reported. The Nurses’ reported that the sensitivity of instruments designed to
Health Study demonstrated a significant inverse association detect problem drinking, such as the CAGE questionnaire (a
between light-to-moderate alcohol consumption, irrespec- four-item test with questions on Cutting down, Annoyance
tive of beverage type, and the risk of type 2 diabetes (28, at criticism, Guilty feelings, and use of Eye openers), is
32). The remaining two smaller studies showed that lower in college women than men (46). Second, a different
increased alcohol intake was not associated with diabetes beverage preference between women and men may explain
risk but was strongly associated with a decreased body mass some of the observed interaction. Our data showed that the
index (25, 31). This strong inverse association, as indicated increased diabetes risk in men in the highest alcohol intake
by previous ARIC studies (34), may account for the lack of group was related to the consumption of spirits. In this
association between alcohol and the risk of diabetes. In our study, women were about as likely to consume spirits as
study, women with moderate alcohol consumption were not men, but women were also more likely to consume wine and
more likely to develop diabetes than were their counterparts less likely to consume beer. Alcoholic beverage preference
who drank occasionally, even after adjustment for body is associated with demographic and health behavior-related
mass index and the waist/hip ratio. characteristics. Persons who prefer wine are likely to be
Of the five studies that were comparable with the present women, temperate, nonsmokers, better educated, and free of
study regarding methodology and population, four pre- symptoms or risk of illness while persons who prefer spirits
sented results regarding men (25, 26, 30, 31). The increased are likely to be men, heavier drinkers, less educated, and

Am J Epidemiol Vol. 154, No. 8, 2001


756 Kao et al.

afflicted with symptoms or risk factors of major illnesses, betes. Diabetes Metab Rev 1993;9:129–46.
and persons who prefer beer had intermediate traits (24). In 6. Manolio TA, Savage PJ, Burke GL, et al. Association of fast-
ing insulin with blood pressure and lipids in young adults. The
addition, we demonstrated in our study that wine consump- CARDIA Study. Arteriosclerosis 1990;10:430–6.
tion was statistically associated with healthier behaviors 7. Perry IJ, Wannamethee SG, Shaper AG. Prospective study of
while spirits and beer consumption were not. Therefore, we serum gamma-glutamyltransferase and risk of NIDDM.
cannot exclude the possibility that the differences in bever- Diabetes Care 1998;21:732–7.
age preference may simply represent differences in 8. Phillips GB, Safrit HF. Alcoholic diabetes. Induction of glu-
cose intolerance with alcohol. JAMA 1971;217:1513–19.
lifestyles, which we were not able to adequately measure 9. Dornhorst A, Ouyang A. Effect of alcohol on glucose toler-
and adjust in our regression model. Third, this difference ance. Lancet 1971;2:957–9.
could simply be a type II error due to the small number of 10. Avogaro A, Valerio A, Miola M, et al. Ethanol impairs insulin-
women who drank >14 drinks/week. mediated glucose uptake by an indirect mechanism. J Clin
Endocrinol Metab 1996;81:2285–90.
The results of the present study support the hypothesis 11. McMonagle J, Felig P. Effects of ethanol ingestion on glucose
that high alcohol intake (≥21 drinks/week) predicts type 2 tolerance and insulin secretion in normal and diabetic subjects.
diabetes mellitus among middle-aged men, specifically men Metabolism 1975;24:625–32.
who drink more than 14 drinks of spirits per week. 12. Yki-Jarvinen H, Koivisto VA, Ylikahri R, et al. Acute effects
However, more moderate levels of consumption (<21 of ethanol and acetate on glucose kinetics in normal subjects.
Am J Physiol 1988;254(2 Pt 1):E175–80.
drinks/week) do not appear to increase the risk of type 2 13. Yki-Jarvinen H, Nikkila EA. Ethanol decreases glucose uti-
diabetes in middle-aged men and women. Our results, along lization in healthy man. J Clin Endocrinol Metab 1985;61:
with those of previous studies, suggest that strategies for the 941–5.
prevention of type 2 diabetes need not target moderate alco- 14. Shelmet JJ, Reichard GA, Skutches CL, et al. Ethanol causes
acute inhibition of carbohydrate, fat, and protein oxidation and
hol consumption. In contrast, men who drink >14 drinks of insulin resistance. J Clin Invest 1988;81:1137–45.
spirits per week should be advised of the increased risk of 15. Patto RJ, Russo EK, Borges DR, et al. The enteroinsular axis
diabetes associated with heavy alcohol consumption. and endocrine pancreatic function in chronic alcohol con-

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011


Further research should investigate the potential effect mod- sumers: evidence for early beta-cell hypofunction. Mt Sinai J
ification by sex and alcoholic beverages in the association Med 1993;60:317–20.
16. Facchini F, Chen YD, Reaven GM. Light-to-moderate alcohol
between alcohol and the risk of type 2 diabetes. intake is associated with enhanced insulin sensitivity. Diabetes
Care 1994;17:115–19.
17. Spolarics Z, Bagby GJ, Pekala PH, et al. Acute alcohol admin-
istration attenuates insulin-mediated glucose use by skeletal
muscle. Am J Physiol 1994;267(6 Pt 1):E886–91.
18. Alqatari M, Shih MF, Taberner PV. Chronic ethanol consump-
ACKNOWLEDGMENTS tion ameliorates the maturity-onset diabetes-obesity syndrome
in CBA mice. Alcohol Alcohol 1996;31:89–99.
The ARIC Study is carried out as a collaborative study 19. Mayer EJ, Newman B, Quesenberry CP Jr, et al. Alcohol
supported by contracts N01-HC-55015, N01-HC-55016, consumption and insulin concentrations. Role of insulin in
N01-HC-55018, N01-HC-55019, N01-HC-55020, N01-HC- associations of alcohol intake with high-density lipoprotein
55021, and N01-HC-55022 from the National Heart, Lung, cholesterol and triglycerides. Circulation 1993;88(5 Pt 1):
2190–7.
and Blood Institute. F. L. B. was supported by an 20. Barrett-Connor E, Schrott HG, Greendale G, et al. Factors
Established Investigator Grant from the American Heart associated with glucose and insulin levels in healthy post-
Association (Dallas, Texas). W. H. L. K. was supported by menopausal women. Diabetes Care 1996;19:333–40.
NIH training grant T32HL07024-23. 21. Jose GM, Klatsky AL, Siegelaub AB, et al. Serum glucose lev-
els and alcohol-consumption habits in a large population.
The authors thank the staff and participants in the ARIC Diabetes 1977;26:780–5.
Study for their important contributions. 22. Razay G, Heaton KW, Bolton CH, et al. Alcohol consumption
and its relation to cardiovascular risk factors in British women.
BMJ 1992;304:80–3.
23. Selby JV, Newman B, King MC, et al. Environmental and
behavioral determinants of fasting plasma glucose in women. A
matched co-twin analysis. Am J Epidemiol 1987;125:979–88.
REFERENCES 24. Klatsky AL, Friedman GD, Armstrong MA. The relationships
between alcoholic beverage use and other traits to blood pres-
1. Harris MI, Flegal KM, Cowie CC, et al. Prevalence of diabetes, sure: a new Kaiser Permanente study. Circulation 1986;73:
impaired fasting glucose, and impaired glucose tolerance in U.S. 628–36.
adults. The Third National Health and Nutrition Examination 25. Holbrook TL, Barrett-Connor E, Wingard DL. A prospective
Survey, 1988–1994. Diabetes Care 1998;21:518–24. population-based study of alcohol use and non-insulin-
2. Wingard DL, Barrett-Connor E. Heart disease and diabetes. In: dependent diabetes mellitus. Am J Epidemiol 1990;132:902–9.
National Diabetes Data Group, ed. Diabetes in America. 26. Perry IJ, Wannamethee SG, Walker MK, et al. Prospective
Bethesda, MD: National Institutes of Health, 1995:429–48. study of risk factors for development of non-insulin dependent
3. Geiss LS, Herman WH, Smith PJ. Mortality in non-insulin- diabetes in middle aged British men. BMJ 1995;310:560–4.
dependent diabetes. In: National Diabetes Data Group, ed. 27. Medalie JH, Papier CM, Goldbourt U, et al. Major factors in
Diabetes in America. Bethesda, MD: National Institutes of the development of diabetes mellitus in 10,000 men. Arch
Health, 1995:259–82. Intern Med 1975;135:811–17.
4. Berkelman RL, Ralston M, Herndon J, et al. Patterns of alco- 28. Stampfer MJ, Colditz GA, Willett WC, et al. A prospective
hol consumption and alcohol-related morbidity and mortality. study of moderate alcohol drinking and risk of diabetes in
MMWR CDC Surveill Summ 1986;35:1SS–5SS. women. Am J Epidemiol 1988;128:549–58.
5. Avogaro A, Tiengo A. Alcohol, glucose metabolism and dia- 29. Feskens EJ, Kromhout D. Cardiovascular risk factors and the

Am J Epidemiol Vol. 154, No. 8, 2001


Alcohol Consumption and the Risk of Diabetes Mellitus 757

25-year incidence of diabetes mellitus in middle-aged men. 1267–73.


The Zutphen Study. Am J Epidemiol 1989;130:1101–8. 39. Vitelli LL, Folsom AR, Shahar E, et al. Association of dietary
30. Rimm EB, Chan J, Stampfer MJ, et al. Prospective study of composition with fasting serum insulin level: the ARIC Study.
cigarette smoking, alcohol use, and the risk of diabetes in men. Nutr Metab Cardiovasc Dis 1996;6:194–202.
BMJ 1995;310:555–9. 40. Tiengo A, Valerio A, Molinari M, et al. Effect of ethanol,
31. Monterrosa AE, Haffner SM, Stern MP, et al. Sex difference in acetaldehyde, and acetate on insulin and glucagon secretion in
lifestyle factors predictive of diabetes in Mexican-Americans. the perfused rat pancreas. Diabetes 1981;30:705–9.
Diabetes Care 1995;18:448–56. 41. Holley DC, Bagby GJ, Curry DL. Ethanol-insulin interrela-
32. Colditz GA, Manson JE, Stampfer MJ, et al. Diet and risk of tionships in the rat studied in vitro and in vivo: evidence for
clinical diabetes in women. Am J Clin Nutr 1992;55:1018–23. direct ethanol inhibition of biphasic glucose-induced insulin
33. The Atherosclerosis Risk in Communities (ARIC) Study: design secretion. Metabolism 1981;30:894–9.
and objectives. The ARIC investigators. Am J Epidemiol 1989; 42. Arky RA, Veverbrants E, Abramson EA. Irreversible hypo-
129:687–702. glycemia. A complication of alcohol and insulin. JAMA 1968;
34. Duncan BB, Chambless LE, Schmidt MI, et al. Association of 206:575–8.
the waist-to-hip ratio is different with wine than with beer or 43. Hodge AM, Dowse GK, Collins VR, et al. Abnormal glucose
hard liquor consumption. Atherosclerosis Risk in Communities tolerance and alcohol consumption in three populations at high
Study investigators. Am J Epidemiol 1995;142:1034–8. risk of non-insulin-dependent diabetes mellitus. Am J
35. Baecke JA, Burema J, Frijters JE. A short questionnaire for the Epidemiol 1993;137:178–89.
measurement of habitual physical activity in epidemiological 44. Gurwitz JH, Field TS, Glynn RJ, et al. Risk factors for non-
studies. Am J Clin Nutr 1982;36:936–42. insulin-dependent diabetes mellitus requiring treatment in the
36. Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility elderly. J Am Geriatr Soc 1994;42:1235–40.
and validity of a semiquantitative food frequency question- 45. Neve RJ, Lemmens PH, Drop MJ. Gender differences in alco-
naire. Am J Epidemiol 1985;122:51–65. hol use and alcohol problems: mediation by social roles and
37. SAS Institute, Inc. SAS user’s guide: statistics, version 6.11 gender-role attitudes. Subst Use Misuse 1997;32:1439–59.
ed. Cary, NC: SAS Institute, Inc, 1989. 46. Werner MJ, Walker LS, Greene JW. Screening for problem
38. Shaper AG, Wannamethee G, Walker M. Alcohol and mortality drinking among college freshmen. J Adolesc Health 1994;15:
in British men: explaining the U-shaped curve. Lancet 1988;2: 303–10.

Downloaded from aje.oxfordjournals.org by guest on April 29, 2011

Am J Epidemiol Vol. 154, No. 8, 2001

Vous aimerez peut-être aussi