Effects of Ethanol Ingestion on Glucose Tolerance and Insulin Secretion in Normal and Diabetic Subjects lames McMonagle To investigate the effect of ethanol on carbohydrate homeostasir in circumstances in which food and ethanol are usually ingested, ethanol was administered hourly in the afternoon prior to the ingestion of a glucose load at 5:00 p.m. in a group of normal subjects and in mild diabetics. In both groups the blood glucose levels following the glucose load were 30-80 mg/ 100 ml lower and the early insulin secretory response ( 15-45 min) was 35%-40% higher after ethanol ingestion. In contrast, ethanol intake had no effect on the
and Philip
Felig
glucagon response to glucose ingestion. These data suggest that ethanol enhances glucose-stimulated insulin secretion. The dampened blood glucose rise observed with ethanol may be related to the augmented insulin response or to decreased gastrointestinal absorption of glucose. In mild diabetic patients, moderate intake of ethanol is without acute deleterious effects on carbohydmte homeostasis and may in some instances improve the blood glucose response to ingested carbohydrate.
R
ECENT INVESTIGATIONS concerning the effect of ethanol on glucose tolerance and insulin secretion have resulted in conflicting data. Metz et al. have shown that, in a group of normal volunteers, the administration of ethanol prior to a glucose load lowers the blood glucose rise and increases the early insulin response.’ Shortly thereafter, Friedenberg and Metz obtained similar results in a group of normal controls and in a group of non-insulin-dependent diabetics.2 Studying patients with a history of diabetes mellitus or a strong family history of diabetes, Phillips found that ethanol pretreatment resulted in a deterioration of glucose tolerance (i.e., an exaggerated blood glucose rise in response to a glucose load) in association with an increased insulin secretory response.3 When ethanol and the glucose load were administered simultaneously to a group of healthy volunteers, Dornhorst observed a deterioration in glucose tolerance in association with an increased insulin secretory response.4 In each of the above studies, the effect of ethanol was determined by giving rather large amounts (approximately 0.5 1 of pure ethanol) or by administering the ethanol throughout the night (between 8 p.m. and 6 a.m.). In addition, the effect of ethanol was only examined on glucose ingested in the morning after an overnight fast. Consequently, it was the objective of this study to examine the effect of ethanol on glucose disposal and on insulin and glucagon secretion in mildly diabetic and nondiabetic volunteers in circumstances which more closely approximate the usual mode of ethanol and food intake.
From the Department of Internal Medicine, Yale University School of Medicine, New Haven, Conn. Received for publication September 26, 1974. Supported in part by USPHS research grant AM 13526 and research career development award AM 70219from the National Institutes of Health. Reprint requests should be addressed to Philip Felig. M.D., Department of Internal Medicine Yale University School of Medicine 333 Cedar Street New Haven, Conn. 06510. o 1975 by Grune & Stratton, Inc.
hwabokm,
Vol. 24, No. 5 (May), 1975
625
626
MC MONAGLE
MATERIALS
AND
FELIG
AND METHODS
Sevennondiabetic and five diabetic subjects were studied. Four men and three women between the ages of 19 and 28 (mean 22) comprised the normal nondiabetic group. In this group, there was no history of liver disease, ethanol abuse, or diabetes mellitus. No subject had been taking any medications prior to or during the course of the study. All subjects were within 7% of their “desirable weight” (Metropolitan Life Insurance Company: Statistical Bulletin, 40~3, 1959). None of the female subjects were studied close to (within 3-4 days) or during their menstrual periods. Three men and two women between the ages of 32 and 58 (mean 44) comprised the diabetic group. There was no history of liver disease or ethanol abuse in this group. None of the subjects in this group were on insulin therapy during the study, nor at any time during the 12 mo prior to the study. Four of the five subjects were receiving oral hypoglycemic agents which were discontinued the day before each glucose tolerance test. None of the subjects in this group were taking any other medications prior to or during the course of the study. All subjects were within 12% of their “desirable weight” (Metropolitan Life Insurance Company: 40~3, 1959). Neither female subject was studied close to or during her menstrual period. Both the diabetics and nondiabetics were involved in the same experimental design. For the 3 days prior to each study, the subjects consumed a weight-maintaining diet containing at least 200 g carbohydrate/day. Two studies (ethanol and control) separated by 1 wk were performed on each subject. Each subject thus served as his own control in evaluating the effects of ethanol. During one study, ethanol was administered prior to the ingestion of glucose and during the other study, a low calorie carbonated beverage* replaced the ethanol. The subjects were randomized so that half of them received ethanol first and half received the low calorie soda first. On the day of the study, a standardized lunch containing 700-800 cal and 40% carbohydrate was administered at 12 noon. Beginning at 1 p.m. each subject ingested either: a) 60 ml of ethanol as SO-proof bonded whiskey in 200-250 ml water divided into four equal hourly doses, or b) equal volumes of low calorie soda. One hundred g dextroset dissolved in 10 oz of water was then administered orally at 5 p.m. Venous blood samples for glucose, insulin, glucagon, and ethanol determinations were obtained at 15-30 min intervals during the hour prior to, and for 3 hr after the ingestion of dextrose. Plasma glucose was determined by the glucose oxidase procedure.5 Plasma insulin was measured by radioimmunoassay, employing talc to separate bound from free insulin.6 Pancreatic glucagon was measured in plasma (in the nondiabetics only) by radioimmunoassay, with an antibody (lot 30K) made available by Dr. Roger Unger? The ethanol concentration was measured spectrophotometrically.8 In addition to comparing absolute levels of insulin, the insulin secretory response was expressed as the sum of insulin values over time, as the sum of insulin increments over time, and as the area under the insulin response curve. In evaluating the effect of ethanol on insulin secretion, the initial (15-45 or 60 min) responses as well as the delayed insulin responses were compared. The basis of examining the initial insulin response derives from its usefulness as a sensitive index of minor alterations in insulin secretion.‘,” The paired t test was employed in the statistical analyses.” All values were reported as the mean f standard error. RESULTS
Ethanol Concentration The peak blood ethanol concentration occurred 1 hr after the last ethanol dose and was 29 + 4 mg/lOO ml in the nondiabetics and 35 f 7 mg/lOO ml in the diabetics. All of the subjects tolerated the ethanol well. Some experienced mild transient nausea after its ingestion. None showed evidence of inebriation.
*Cott Diet Ginger (2.22 Cal/6 oz serving), Cott Corporation, tDexto1, Hospital Supply Corporation, McGaw Park, Ill.
New Haven, Corm.
ETHANOL
INGESTION
627
I60
160
z Fig. 1. The effect of ethanol on the plasma glucose response to glucose ingestion in normal subjects. The points plotted are the mean f SEM. The asterisk indicates statistically significant differences (p < 0.025 at 30 min; p < 0.005 at 45 min; p < 0.05 at 60 min; p < 0.05 at 90 min; paired t-test).
140
g d
120
ii 8 3 100
z! 60
1
I
I20
160
I
TIME (mitt)
Normal Subjects
The blood glucose concentrations with and without ethanol treatment are shown in Fig. 1. Ethanol had no effect on the basal glucose concentration but there was a significant depression of the blood glucose rise after oral glucose in the ethanol-treated group. Glucose levels were 30-40 mg/lOO ml lower in the ethanol group at 30-90 min after glucose ingestion. The peak glucose concentration occurred at 45 min in each group. It was 126 =t 17 mg/lOO ml in the ethanol group and 161 + 18 mg/lOO ml in the control group (p < 0.005). Insulin levels are shown in Fig. 2. Ethanol had no significant effect on the basal insulin concentration. There was an increase in the early insulin response
70
-
60
-
20 ’ /l
IO y-m_-_ b-i
’
Control
Fig. 2. The effect of ethanol on the plasma insulin response to glucose ingestion in normal subjects. The points plotted are the mean f SEM. The asterisk indicates statistically significant differences (p < 0.05 at 45 min; p < 0.05 at 75 min; p < 0.01 at 90 min; p < 0.05 at 105 min; p < 0.05 ot 120 min; p < 0.05 at 150 min; paired t test).
620
MC MONAGLE
AND
FEUG
Table 1. Influence of Ethanol on Insulin Response in Normal Subjects Statistical Significance Sum of the insulin concentrations (pU/ml)
15-45 min
134&
187 f 27
p < 0.05
60-90
175 f 37
119zk30
p < 0.05
257 f 38
155 f 33
p < 0.005
15-45 min
114*
13
158 f 23
p < 0.05
60-90
155 zt 40
86 f 24
231 f 41
105 f 29
105-180 Sum of the increments of insulin W/ml)
105-180 Area under insulin response curve (pU x min/ml) Insulin:Glucose ratio (pU/mg)
min min min min
O-60 min 60-180
min
O-60 min 60-180
min
16
p < 0.025 p < 0.0025
2460 f
283
3233 f 468
p < 0.05
6598a
1261
4485 k 817
p < 0.025
0.307 f 0.035
0.473 f 0.071
p < 0.0125
0.437 f
0.356 f
NS
0.071
0.046
to the ingestion of an oral glucose load after ethanol intake. In the ethanol group, the insulin concentration was significantly higher at 45 min (p < 0.05) and was significantly lower from 75 to 150 min (p < 0.05). The peak insulin concentration was the same in both groups (66 f 12 pU/ml), but the peak occurred 60 min earlier in the ethanol study (45 min as compared to 105 min in the control study). The insulin secretory response was also expressed as the sum of the plasma insulin concentrations. As seen in Table 1, the initial insulin secretory response (15-45 min) was 40% higher in the ethanol group (p < 0.05). Expressed as the sum of increments (i.e., the summation of the increases of insulin over basal levels), the early insulin secretory response (15-45 min) was significantly higher in the ethanol group (p < 0.05). When the insulin secretory response was expressed as the total area under the insulin response curve from zero to 60 min and from 60 to 180 min, the early insulin response was higher in the ethanol group (p < 0.05). As noted above, glucose levels in the ethanol-treated group were reduced. To evaluate the insulin response as a function of a glycemic stimulus, Seltzer I2 introduced the concept of the “insulinogenic index” which is a ratio of the insulin increment to the glucose increment. In Table 1, the insulin:glucose ratio (1:G ratio) was defined as the total area circumscribed by the insulin curve divided by the total area circumscribed by the glucose curve (over a given time interval). Over the first 60 min, the I:G ratio was significantly higher in the ethanol group (p < 0.0125). Glucagon levels are shown in Fig. 3. Basal glucagon levels tended to be higher in the ethanol group but this difference was not significant. Glucose suppression of glucagon was demonstrated in both groups. All postglucose ingestion glucagon levels fell 25%-40% below basal values in both the ethanol and control groups (p < 0.05). Although glucagon levels were significantly higher in the ethanol group at 45, 75, and 90 min, the mean maximal decline in glucagon was comparable in the two groups (control 57.3 f 7.7 pg/ml; ethanol 64.3 f 10.0 pg/ml). Diabetics
Glucose levels are shown in Fig. 4. Ethanol had no effect on the basal glucose concentration but ethanol did dampen the blood glucose rise following the in-
629
ETHANOL INGESTION
- too E \ x 60 -
Fig. 3. The effect of ethanol on the plasma glucagon response to glucase ingestion in normal subjects. The points plotted are the mean f SEM. The asterisk indicates statistically significant differences (p < 0.05 at 45 min; p < 0.025 at 75 min; p < 0.01 at 90 min; paired t test).
$ :: 2
60 -
d 40
-
20 -
OL
I
L -30
/o
IOO~romr deztrose
I l30
I 60
I
I
I
I
lgg
+t20
*I30
l1gg
TIME (min)
gestion of glucose. Glucose levels were 40-80 mg/lOO ml lower in the ethanol group at 90-180 min after glucose ingestion. The two curves were significantly different from 105 to 180 min. Insulin levels are shown in Fig. 5. Ethanol had no significant effect on the basal insulin concentration. As in the normal subjects, insulin levels were significantly greater at 15 and 30 min after ethanol treatment (p < 0.025). The peak insulin concentration also occurred earlier in the ethanol study (30 min versus 90 min). Expressed as the sum of the plasma insulin concentrations or as the area
TIME
(min)
Fig. 4. The effect of ethanol on the plasma glucose response to glucose ingestion in mild diabetics. The points plotted are the mean f SEM. The asterisk indicates statistically significant differences (p < 0.0125 at 105 min; p < 0.005 at 120 min; p < 0.005 at 150 min; p < 0.025 at 180 min; paired t test).
630
MC MONAGLE
AND FEUG
to glucose ingestion in mild diabetics. The points plotted ore the mean * SEM. The asterisk indicates statistically significant differences (p < 0.025 at toogram*drltrou
TIME (min)
under the insulin response curve (Table 2), the early insulin response (15-45 or 60 min) was 35% higher in the ethanol group (p < 0.025). The 1:G ratio was also significantly higher in the ethanol group over the first 60 min (p < 0.0125) as well as the last 120 min (p < 0.01). DISCUSSION
The data presented here indicate that ethanol intake dampens the blood glucose rise and increases the early insulin response to the ingestion of an oral glucose load. Glucose levels were 30-80 mg/lOO ml lower after ethanol ingestion. The early insulin secretory response was 35x--40% higher in the ethanol-treated groups. These effects were observed in both normal controls and in mild diabetics in a setting in which food and ethanol are commonly consumed. The concept that a particular substance can enhance glucose-mediated insulin secretion is not new. Metz previously indicated the ability of ethanol to lower the blood glucose rise and to increase the early insulin response to a glucose load.’ The data presented here support the observations of Metz and extend Table 2. Influence of Ethanol on insulin Response in Diabetic Subjects Control Sum of the insulin concentrations (aU/m~)
(~U/ml)
15-45 min
49.8 f 10.8
68.4 f
12.8
p < 0.025
59.0 f 13.3
69.8 f
16.6
NS
76.4 i
74.6 i
13.9
NS
min
(&I x min/ml) InsulimGlucose ratio W/ma)
18.0
15-45 min
10.8 f 6.4
24.0 f
10.0
O.OS
and Philip
Felig
glucagon response to glucose ingestion. These data suggest that ethanol enhances glucose-stimulated insulin secretion. The dampened blood glucose rise observed with ethanol may be related to the augmented insulin response or to decreased gastrointestinal absorption of glucose. In mild diabetic patients, moderate intake of ethanol is without acute deleterious effects on carbohydmte homeostasis and may in some instances improve the blood glucose response to ingested carbohydrate.
R
ECENT INVESTIGATIONS concerning the effect of ethanol on glucose tolerance and insulin secretion have resulted in conflicting data. Metz et al. have shown that, in a group of normal volunteers, the administration of ethanol prior to a glucose load lowers the blood glucose rise and increases the early insulin response.’ Shortly thereafter, Friedenberg and Metz obtained similar results in a group of normal controls and in a group of non-insulin-dependent diabetics.2 Studying patients with a history of diabetes mellitus or a strong family history of diabetes, Phillips found that ethanol pretreatment resulted in a deterioration of glucose tolerance (i.e., an exaggerated blood glucose rise in response to a glucose load) in association with an increased insulin secretory response.3 When ethanol and the glucose load were administered simultaneously to a group of healthy volunteers, Dornhorst observed a deterioration in glucose tolerance in association with an increased insulin secretory response.4 In each of the above studies, the effect of ethanol was determined by giving rather large amounts (approximately 0.5 1 of pure ethanol) or by administering the ethanol throughout the night (between 8 p.m. and 6 a.m.). In addition, the effect of ethanol was only examined on glucose ingested in the morning after an overnight fast. Consequently, it was the objective of this study to examine the effect of ethanol on glucose disposal and on insulin and glucagon secretion in mildly diabetic and nondiabetic volunteers in circumstances which more closely approximate the usual mode of ethanol and food intake.
From the Department of Internal Medicine, Yale University School of Medicine, New Haven, Conn. Received for publication September 26, 1974. Supported in part by USPHS research grant AM 13526 and research career development award AM 70219from the National Institutes of Health. Reprint requests should be addressed to Philip Felig. M.D., Department of Internal Medicine Yale University School of Medicine 333 Cedar Street New Haven, Conn. 06510. o 1975 by Grune & Stratton, Inc.
hwabokm,
Vol. 24, No. 5 (May), 1975
625
626
MC MONAGLE
MATERIALS
AND
FELIG
AND METHODS
Sevennondiabetic and five diabetic subjects were studied. Four men and three women between the ages of 19 and 28 (mean 22) comprised the normal nondiabetic group. In this group, there was no history of liver disease, ethanol abuse, or diabetes mellitus. No subject had been taking any medications prior to or during the course of the study. All subjects were within 7% of their “desirable weight” (Metropolitan Life Insurance Company: Statistical Bulletin, 40~3, 1959). None of the female subjects were studied close to (within 3-4 days) or during their menstrual periods. Three men and two women between the ages of 32 and 58 (mean 44) comprised the diabetic group. There was no history of liver disease or ethanol abuse in this group. None of the subjects in this group were on insulin therapy during the study, nor at any time during the 12 mo prior to the study. Four of the five subjects were receiving oral hypoglycemic agents which were discontinued the day before each glucose tolerance test. None of the subjects in this group were taking any other medications prior to or during the course of the study. All subjects were within 12% of their “desirable weight” (Metropolitan Life Insurance Company: 40~3, 1959). Neither female subject was studied close to or during her menstrual period. Both the diabetics and nondiabetics were involved in the same experimental design. For the 3 days prior to each study, the subjects consumed a weight-maintaining diet containing at least 200 g carbohydrate/day. Two studies (ethanol and control) separated by 1 wk were performed on each subject. Each subject thus served as his own control in evaluating the effects of ethanol. During one study, ethanol was administered prior to the ingestion of glucose and during the other study, a low calorie carbonated beverage* replaced the ethanol. The subjects were randomized so that half of them received ethanol first and half received the low calorie soda first. On the day of the study, a standardized lunch containing 700-800 cal and 40% carbohydrate was administered at 12 noon. Beginning at 1 p.m. each subject ingested either: a) 60 ml of ethanol as SO-proof bonded whiskey in 200-250 ml water divided into four equal hourly doses, or b) equal volumes of low calorie soda. One hundred g dextroset dissolved in 10 oz of water was then administered orally at 5 p.m. Venous blood samples for glucose, insulin, glucagon, and ethanol determinations were obtained at 15-30 min intervals during the hour prior to, and for 3 hr after the ingestion of dextrose. Plasma glucose was determined by the glucose oxidase procedure.5 Plasma insulin was measured by radioimmunoassay, employing talc to separate bound from free insulin.6 Pancreatic glucagon was measured in plasma (in the nondiabetics only) by radioimmunoassay, with an antibody (lot 30K) made available by Dr. Roger Unger? The ethanol concentration was measured spectrophotometrically.8 In addition to comparing absolute levels of insulin, the insulin secretory response was expressed as the sum of insulin values over time, as the sum of insulin increments over time, and as the area under the insulin response curve. In evaluating the effect of ethanol on insulin secretion, the initial (15-45 or 60 min) responses as well as the delayed insulin responses were compared. The basis of examining the initial insulin response derives from its usefulness as a sensitive index of minor alterations in insulin secretion.‘,” The paired t test was employed in the statistical analyses.” All values were reported as the mean f standard error. RESULTS
Ethanol Concentration The peak blood ethanol concentration occurred 1 hr after the last ethanol dose and was 29 + 4 mg/lOO ml in the nondiabetics and 35 f 7 mg/lOO ml in the diabetics. All of the subjects tolerated the ethanol well. Some experienced mild transient nausea after its ingestion. None showed evidence of inebriation.
*Cott Diet Ginger (2.22 Cal/6 oz serving), Cott Corporation, tDexto1, Hospital Supply Corporation, McGaw Park, Ill.
New Haven, Corm.
ETHANOL
INGESTION
627
I60
160
z Fig. 1. The effect of ethanol on the plasma glucose response to glucose ingestion in normal subjects. The points plotted are the mean f SEM. The asterisk indicates statistically significant differences (p < 0.025 at 30 min; p < 0.005 at 45 min; p < 0.05 at 60 min; p < 0.05 at 90 min; paired t-test).
140
g d
120
ii 8 3 100
z! 60
1
I
I20
160
I
TIME (mitt)
Normal Subjects
The blood glucose concentrations with and without ethanol treatment are shown in Fig. 1. Ethanol had no effect on the basal glucose concentration but there was a significant depression of the blood glucose rise after oral glucose in the ethanol-treated group. Glucose levels were 30-40 mg/lOO ml lower in the ethanol group at 30-90 min after glucose ingestion. The peak glucose concentration occurred at 45 min in each group. It was 126 =t 17 mg/lOO ml in the ethanol group and 161 + 18 mg/lOO ml in the control group (p < 0.005). Insulin levels are shown in Fig. 2. Ethanol had no significant effect on the basal insulin concentration. There was an increase in the early insulin response
70
-
60
-
20 ’ /l
IO y-m_-_ b-i
’
Control
Fig. 2. The effect of ethanol on the plasma insulin response to glucose ingestion in normal subjects. The points plotted are the mean f SEM. The asterisk indicates statistically significant differences (p < 0.05 at 45 min; p < 0.05 at 75 min; p < 0.01 at 90 min; p < 0.05 at 105 min; p < 0.05 ot 120 min; p < 0.05 at 150 min; paired t test).
620
MC MONAGLE
AND
FEUG
Table 1. Influence of Ethanol on Insulin Response in Normal Subjects Statistical Significance Sum of the insulin concentrations (pU/ml)
15-45 min
134&
187 f 27
p < 0.05
60-90
175 f 37
119zk30
p < 0.05
257 f 38
155 f 33
p < 0.005
15-45 min
114*
13
158 f 23
p < 0.05
60-90
155 zt 40
86 f 24
231 f 41
105 f 29
105-180 Sum of the increments of insulin W/ml)
105-180 Area under insulin response curve (pU x min/ml) Insulin:Glucose ratio (pU/mg)
min min min min
O-60 min 60-180
min
O-60 min 60-180
min
16
p < 0.025 p < 0.0025
2460 f
283
3233 f 468
p < 0.05
6598a
1261
4485 k 817
p < 0.025
0.307 f 0.035
0.473 f 0.071
p < 0.0125
0.437 f
0.356 f
NS
0.071
0.046
to the ingestion of an oral glucose load after ethanol intake. In the ethanol group, the insulin concentration was significantly higher at 45 min (p < 0.05) and was significantly lower from 75 to 150 min (p < 0.05). The peak insulin concentration was the same in both groups (66 f 12 pU/ml), but the peak occurred 60 min earlier in the ethanol study (45 min as compared to 105 min in the control study). The insulin secretory response was also expressed as the sum of the plasma insulin concentrations. As seen in Table 1, the initial insulin secretory response (15-45 min) was 40% higher in the ethanol group (p < 0.05). Expressed as the sum of increments (i.e., the summation of the increases of insulin over basal levels), the early insulin secretory response (15-45 min) was significantly higher in the ethanol group (p < 0.05). When the insulin secretory response was expressed as the total area under the insulin response curve from zero to 60 min and from 60 to 180 min, the early insulin response was higher in the ethanol group (p < 0.05). As noted above, glucose levels in the ethanol-treated group were reduced. To evaluate the insulin response as a function of a glycemic stimulus, Seltzer I2 introduced the concept of the “insulinogenic index” which is a ratio of the insulin increment to the glucose increment. In Table 1, the insulin:glucose ratio (1:G ratio) was defined as the total area circumscribed by the insulin curve divided by the total area circumscribed by the glucose curve (over a given time interval). Over the first 60 min, the I:G ratio was significantly higher in the ethanol group (p < 0.0125). Glucagon levels are shown in Fig. 3. Basal glucagon levels tended to be higher in the ethanol group but this difference was not significant. Glucose suppression of glucagon was demonstrated in both groups. All postglucose ingestion glucagon levels fell 25%-40% below basal values in both the ethanol and control groups (p < 0.05). Although glucagon levels were significantly higher in the ethanol group at 45, 75, and 90 min, the mean maximal decline in glucagon was comparable in the two groups (control 57.3 f 7.7 pg/ml; ethanol 64.3 f 10.0 pg/ml). Diabetics
Glucose levels are shown in Fig. 4. Ethanol had no effect on the basal glucose concentration but ethanol did dampen the blood glucose rise following the in-
629
ETHANOL INGESTION
- too E \ x 60 -
Fig. 3. The effect of ethanol on the plasma glucagon response to glucase ingestion in normal subjects. The points plotted are the mean f SEM. The asterisk indicates statistically significant differences (p < 0.05 at 45 min; p < 0.025 at 75 min; p < 0.01 at 90 min; paired t test).
$ :: 2
60 -
d 40
-
20 -
OL
I
L -30
/o
IOO~romr deztrose
I l30
I 60
I
I
I
I
lgg
+t20
*I30
l1gg
TIME (min)
gestion of glucose. Glucose levels were 40-80 mg/lOO ml lower in the ethanol group at 90-180 min after glucose ingestion. The two curves were significantly different from 105 to 180 min. Insulin levels are shown in Fig. 5. Ethanol had no significant effect on the basal insulin concentration. As in the normal subjects, insulin levels were significantly greater at 15 and 30 min after ethanol treatment (p < 0.025). The peak insulin concentration also occurred earlier in the ethanol study (30 min versus 90 min). Expressed as the sum of the plasma insulin concentrations or as the area
TIME
(min)
Fig. 4. The effect of ethanol on the plasma glucose response to glucose ingestion in mild diabetics. The points plotted are the mean f SEM. The asterisk indicates statistically significant differences (p < 0.0125 at 105 min; p < 0.005 at 120 min; p < 0.005 at 150 min; p < 0.025 at 180 min; paired t test).
630
MC MONAGLE
AND FEUG
to glucose ingestion in mild diabetics. The points plotted ore the mean * SEM. The asterisk indicates statistically significant differences (p < 0.025 at toogram*drltrou
TIME (min)
under the insulin response curve (Table 2), the early insulin response (15-45 or 60 min) was 35% higher in the ethanol group (p < 0.025). The 1:G ratio was also significantly higher in the ethanol group over the first 60 min (p < 0.0125) as well as the last 120 min (p < 0.01). DISCUSSION
The data presented here indicate that ethanol intake dampens the blood glucose rise and increases the early insulin response to the ingestion of an oral glucose load. Glucose levels were 30-80 mg/lOO ml lower after ethanol ingestion. The early insulin secretory response was 35x--40% higher in the ethanol-treated groups. These effects were observed in both normal controls and in mild diabetics in a setting in which food and ethanol are commonly consumed. The concept that a particular substance can enhance glucose-mediated insulin secretion is not new. Metz previously indicated the ability of ethanol to lower the blood glucose rise and to increase the early insulin response to a glucose load.’ The data presented here support the observations of Metz and extend Table 2. Influence of Ethanol on insulin Response in Diabetic Subjects Control Sum of the insulin concentrations (aU/m~)
(~U/ml)
15-45 min
49.8 f 10.8
68.4 f
12.8
p < 0.025
59.0 f 13.3
69.8 f
16.6
NS
76.4 i
74.6 i
13.9
NS
min
(&I x min/ml) InsulimGlucose ratio W/ma)
18.0
15-45 min
10.8 f 6.4
24.0 f
10.0
O.OS
