You are on page 1of 7

Original Research Communications

Caffeine ingestion increases the insulin response to an


oral-glucose-tolerance test in obese men before and after weight
loss
13
Heather J Petrie, Sara E Chown, Laura M Belfie, Alison M Duncan, Drew H McLaren, Julie A Conquer, and
Terry E Graham
ABSTRACT
Background: Caffeine ingestion decreases the insulin sensitivity
index (ISI) for an oral-glucose-tolerance test (OGTT) and decreases
insulin-induced glucose disposal in lean male subjects during a hy-
perinsulinemic clamp.
Objective: We examined the effects of caffeine ingestion on insulin
and glucose homeostasis in obese men before and after a nutrition
and exercise intervention.
Design: Nine sedentary, obese [body mass index (in kg/m
2
): 34.0
1.0] men who had refrained fromexercise and caffeine ingestion for
48 h underwent 2 oral-glucose-tolerance tests (OGTTs). The sub-
jects randomly received caffeine (5 mg/kg) or placebo 1 h before
each OGTT. After a 12-wk nutrition and exercise intervention, dur-
ing which time the subjects avoided dietary caffeine, the OGTTs
were repeated.
Results: The intervention resulted in decreases (P 0.05) in body
weight (8.5 1.5 kg), percentage body fat (2.8 0.7%), and fasting
glucose, insulin, and proinsulin concentrations and increases in the
ISI for the placebo OGTT (P 0.05). Caffeine caused a greater
(P 0.05) OGTT insulin response and a lower (P 0.05) ISI both
before and after weight loss. The proinsulin-insulin ratio indicated
that neither weight loss nor caffeine affected the nature of the cell
secretion of insulin.
Conclusions: A nutrition and exercise intervention improved,
whereas caffeine ingestion impaired, insulin-glucose homeostasis in
obese men. The results are consistent with previous findings that
caffeine ingestion contributes to insulin resistance. Am J Clin
Nutr 2004;80:228.
KEY WORDS Insulin resistance, type 2 diabetes, carbohy-
drate, proinsulin, cell function, hyperinsulinemia, oral-glucose-
tolerance test, exercise, insulin sensitivity index
INTRODUCTION
Obesity is associated with impaired glucose tolerance (IGT)
and type 2 diabetes (14). Common to both IGT and type 2
diabetes are hyperinsulinemia and insulin resistance, which re-
sult from the impairment of insulins action on insulin-sensitive
tissues such as skeletal muscle (5). Insulin resistance can be
characterized by elevated basal seruminsulin concentrations, an
elevated insulin response to glucose ingestion, and a decrease in
glucose uptake in the peripheral tissues, especially skeletal mus-
cle (6, 7). Increased hepatic glucose output and hyperglycemia
generally ensue (5). Many investigators have associated these
metabolic characteristics with visceral adiposity (8-10).
Lifestyle interventions, such as reductions in energy and sat-
urated fat intakes and increases in physical activity, have resulted
in improvements in insulin and glucose control in obese persons
(1113). However, the influence of caffeine on glucose tolerance
and insulin sensitivity has not been studied. Ingestion of caffeine
by lean persons before an oral-glucose-tolerance test (OGTT)
was shown to result in a greater insulin response and no decrease
in the glucose response (14, 15). Studies using hyperinsulinemic
euglycemic clamps have confirmed that caffeine ingestion re-
duces whole-body glucose disposal by 1530% (16, 17) and
glucose uptake in the leg (muscle) by 50% (18). Although it is
known that skeletal muscle is the major tissue that becomes
insulin-insensitive after caffeine ingestion, the mechanism of
action is uncertain. Although many potential mechanisms exist,
considerable evidence indicates that physiologic concentrations
of caffeine antagonize adenosine receptors (1921) both in skel-
etal muscle (2123) and in the central nervous system, with the
latter resulting in an increase in sympathetic activity (15, 20, 21).
The results of studies that have examinedthe effects of chronic
coffee consumption on the risk of type 2 diabetes are inconsis-
tent. van Dam and Feskens (24) reported epidemiologic data
indicating that high coffee consumption is associated with a
reduced risk of type 2 diabetes, even in those with other lifestyle
1
From the Department of Human Biology and Nutritional Sciences, Uni-
versity of Guelph, Canada.
2
Supported by an NSERCCollaborative Health Research Grant, Muscle-
tech Inc. HJP held an NSERC Industrial postgraduate scholarship supported
by the Gatorade Sports Science Institute. SEC held an NSERC Industrial
postgraduate scholarship supported by Muscletech Inc.
3
Reprints not available. Address correspondence to TEG, Department of
Human Biology and Nutritional Sciences, University of Guelph, Guelph,
Ontario, Canada N1G 2W1. E-mail: terrygra@uoguelph.ca.
Received April 17, 2003.
Accepted for publication December 3, 2003.
22 Am J Clin Nutr 2004;80:228. Printed in USA. 2004 American Society for Clinical Nutrition
factors that may increase the risk of type 2 diabetes. In contrast,
a prospective study of Pima Indians by Knowler et al (25) found
no evidence for any relation between coffee consumption and
type 2 diabetes. We elected to examine the acute effects of caf-
feine ingestion on glucose and insulin homeostasis in obese sub-
jects before and after a nutrition and exercise intervention. We
hypothesized that caffeine ingestion before an OGTT would
result in an exaggerated insulin response. It is plausible that
weight loss and increased fitness would improve metabolic ho-
meostasis and decrease the sensitivity to adenosine (26, 27). Thus,
we hypothesized that a nutrition and exercise program would lead
to significant improvements in glucose and insulin control and in
the amelioration of caffeine-induced insulin resistance.
SUBJECTS AND METHODS
Subjects
Potential subjects were screened by a physician to confirm no
history of disease, including diabetes mellitus, and no drug or
alcohol abuse. Eligibility criteria included young (1934 y),
obese [BMI (in kg/m
2
) 30], and sedentary (physical activity
3 times/wk) men with a stable body weight (maximum fluctu-
ation of 2 kg within the previous 2 mo). Subjects were also
required to have a systolic blood pressure 160 mm Hg and a
diastolic blood pressure 100 mm Hg, not be using drugs or
dietary supplements, and to have not partaken in any weight-loss
program within the previous 2 mo. Subjects provided written
consent after receiving detailed verbal and written explanations
of the study, which had received approval from the Human Sub-
jects Committee of the University of Guelph, Ontario, Canada.
Preintervention period
During a 2-wk baseline (preintervention) period, a nutritionist
analyzed the subjects usual diets. Measurements of waist and
hip circumference, height (to the nearest 0.5 cm), and body
weight (tothe nearest 0.1kg) were made, andpercentage bodyfat
was determined by dual energy X-ray absorptiometry. Subjects
walked on a treadmill at progressively increasing speeds and
slopes for the determination of maximal oxygen consumption
(

VO
2
max). Subjects were verbally encouraged and

VO
2
max was
accepted when 2 of the following 3 criteria were achieved: no
increase in oxygen consumption (

VO
2
) despite further increases
in treadmill grade, the heart rate met or exceeded the predicted
maximum (220 age), and the respiratory exchange ratio ex-
ceeded 1.1. The relation between

VO
2
and the heart rate was used
to estimate the energy expended during each exercise session of
the nutrition and exercise program.
Two OGTTs were conducted, separated by 72 h. The sub-
jects entered the laboratory in the morning after a 12-h overnight
fast and after having refrained from exercise and caffeine inges-
tion for 48 h. A catheter was placed in the antecubital vein by a
qualified technician and was kept patent with a normal saline
drip. In a randomized, double-blind fashion, the subjects re-
ceived either placebo or caffeine capsules (5 mg/kg body wt) 1 h
(time 0) before ingestion of the 75-g glucose solution (TRUTOL
75; Nerl Diagnostics, East Providence, RI) at time 60. Venous
blood samples were taken at 30, 0, 60, 75, 90, 120, 150, and
180 min.
Nutrition and exercise intervention
The nutrition and exercise intervention was 12 wk in duration.
During this time, subjects were provided with an individualized
target of a daily energy decrement of 1260 kJ (300 kcal), with
2030% of their energy intake derived from fat, and were in-
structed to refrain fromcaffeine in any form. To maximize com-
pliance with the dietary instructions, subjects were educated
about how to avoid dietary caffeine and how to calculate the
amount of energy and the percentage fat from energy in their
foods and beverages. Daily food intake diaries were reviewed
every other day by researchers and every 4 wk by a registered
dietitian. One 3-d food diary was randomly selected every 2 wk
(6 sets in total) to undergo analysis of energy, macronutrient, and
micronutrient intakes by using the Food Processor software pro-
gram(version 7.11; ESHAResearch, Salem, OR). Subjects com-
pleted a 24-h urine collection at 0, 4, 8, and 12 wk for the
assessment adherence to caffeine abstinence. Any subject that
had a positive result for caffeine was removed from the study.
Subjects visited the laboratory every other day throughout the
entire 12-wk intervention, during which time body weight was
measured, food diaries were reviewed, and supervised treadmill
exercise sessions were completed. Heart rates were monitored
during the exercise sessions and were used to estimate when a
2090-kJ (500-kcal) expenditure was achieved. Subjects were in-
structedtoavoidanyexercise other thanthat performedduringtheir
supervised laboratory visits. Overall, the combined dietary changes
and the exercise program were designed to create a total energy
deficit of 15 06017 150 kJ/wk (36004100 kcal/wk) to elicit a
weight loss of 0.5 kg/wk.
Postintervention period
At the conclusion of the 12-wk program, the subjects under-
went dual energy X-ray absorptiometry,

VO
2
max testing and 2
OGTTs (caffeine compared with placebo). The OGTTs were
scheduled to occur no sooner than 72 h after the subjects last
supervisedexercise sessionor their posttesting

VO
2
maxtest inan
effort to eliminate any effect of the previous exercise session on
glucose tolerance.
Blood and urine sampling and analyses
Each venous sample consisted of 14 mL blood; 7 mL was
transferred to a sodium heparintreated tube, and the remainder
was placed in an untreated tube. The blood from the untreated
tube was allowed to clot and was then centrifuged at 1200 g at
room temperature for 12 min. The serum was frozen at 20 C
for subsequent analysis of insulin (Coat-a-Count Radioimmuno-
assay kit; Diagnostic Products Corporation, Los Angeles), pro-
insulin (Human Proinsulin kit; Linco Research Inc, St Charles,
MO), C-peptide (Human C-Peptide Radioimmunoassay kit;
Linco Research Inc), free fatty acids (FFAs) (NEFA kit; Wako
Bioproducts, Richmond, VA), and glycerol (28). Two hundred
microliters of heparin-treated blood was transferred to 1 mL of
0.6 mmol perchloric acid/L and then centrifuged at 2400 g at
room temperature for 3 min. The supernatant fluid was stored at
20 C until analyzed for glucose (29) and lactate (30). The
proinsulin assay has a specificity of 0.1% for human insulin,
100% for intact proinsulin, and 95% for des-31,32 proinsulin.
Thus, the actual measurement is proinsulin-like compounds, but
for simplicity we refer to this as proinsulin. Urinary caffeine and
OBESITY AND CAFFEINE-INDUCED INSULIN RESISTANCE 23
its metabolites were analyzed by reversed-phase HPLCby using
a diode array detector (31).
Calculations and statistics
Areas under the curve (AUCs) for glucose, insulin, C-peptide,
FFAs, glycerol, and lactate were calculated with the trapezoid
method (32) by using the time 0 value as the baseline. The ratio
of proinsulin to insulin was calculated at time 60 (0 min of the
OGTT) and time 90 (30 min of the OGTT). The insulin sensi-
tivityindex(ISI), developedbyMatsuda andDeFronzo(33), was
used to estimate whole-body insulin sensitivity. The ISI, which
provides a number between 0 and 12, is significantly correlated
with whole-body insulin sensitivity as assessed by the hyperin-
sulinemic euglycemic clamp (r 0.73, P 0.0001) (33).
To examine the effect of the 12-wk intervention on energy and
nutrient intakes, paired t tests (SIGMASTAT for WINDOWS,
version 2.03 19921997; SPSS Inc, Chicago) were used to com-
pare the 10-d food diaries collected before the intervention with
the food diaries collected during the intervention.
Pairwise contrast t tests (SAS, version 8 1999; SAS Institute,
Cary, NC) were used to examine the effects of the nutrition and
exercise intervention on

VO
2
max, body weight, body composi-
tion, and fasting blood glucose, lactate, glycerol, FFA, insulin,
proinsulin, and C-peptide concentrations. Pairwise contrast t
tests were used to examine the effect of caffeine ingestion alone
on all fasting blood endpoints (time 0 min compared with time 60
min). The PROC MIXED procedure, followed by Tukeys test
for multiple comparisons, was used to determine whether caf-
feine ingestion had an effect on the ISI and all AUCs both before
and after the intervention and to determine whether the 12-wk
intervention had an effect on the ISI and all AUCs (SAS, version
8 1999; SAS Institute). For all analyses, statistical significance
was accepted at a P value of 0.05. All means are reported with
SEMs.
RESULTS
Seventeen subjects began the nutrition and exercise interven-
tion, but 8 subjects either withdrew from the study because of
illness (n 1), injury (n 2), or time constraints (n 4) or were
removed from the data analysis because of noncompliance (ie,
methylxanthines were detected in the urine; n 1). The 9 re-
maining subjects ranged in age from23 to 34 y (30.6 1.2 y). All
subjects were caffeine consumers and had intakes ranging from
one chocolate bar a week to an occasional maximum of 5 cups
coffee/d.
Compared with values before the 12-wk intervention began,
intakes of total energy, fat, and saturated fatty acids decreased sig-
nificantly; the intake of dietary fiber increased (Table 1); and

VO
2
max increased (Table 2). Mean body weight, BMI, percentage
body fat, fat mass, fat-free mass, and waist and hip circumferences
all significantly decreased after the intervention (Table 2).
Blood glucose concentrations
Bloodglucose concentrations before andafter the intervention
are summarized in Figure 1. Although the 12-wk nutrition and
exercise program did not significantly alter the blood glucose
AUC (Table 3), it significantly lowered the fasting glucose
concentrations from4.6 0.1 to 4.0 0.1 mmol/L. Ingestion of
FIGURE 1. Mean (SEM) blood glucose concentrations in response to
an oral-glucose-tolerance test. Caffeine was administered at 0 min, and 75 g
glucose was administered at 60 min. F, preintervention placebo; E, prein-
tervention caffeine; , postintervention placebo; , postintervention caf-
feine. n 9. There was a significant decrease in fasting blood glucose after
the intervention (P 0.05; PROC MIXED procedure in SAS followed by
paired t tests).
TABLE 1
Dietary analysis of self-reported dietary intakes recorded before and
during the 12-wk intervention
1
Diet component Before intervention After intervention
Energy (kJ) 11046 435 9372 385
2
Fat (% of energy) 31 2 26 1
2
SFAs (% of energy) 11 1 8 0.5
2
MUFAs (% of energy) 10 1 8 0.5
PUFAs (% of energy) 4 0.5 4 0.5
Cholesterol (mg) 290 45 276 24
Carbohydrate (%) 50 3 53 1
Fiber (g) 19 1 22 1
2
Protein (% of energy) 16 1 19 1
1
All values are x SEM; n 9. SFAs, saturated fatty acids; MUFAs,
monounsaturated fatty acids; PUFAs, polyunsaturated fatty acids.
2
Significantly different from before intervention, P 0.05 (paired t
tests).
TABLE 2
Maximum oxygen consumption (V

O
2max
) and body-composition
measurements
1
Before intervention After intervention
V

O
2
max (mL kg
1

min
1
)
36.9 2.1 45.6 2.3
2
Body weight (kg) 103.4 4.7 94.8 4.6
2
BMI (kg/m
2
) 34.0 1.0 31.2 1.0
2
Body fat (%) 29.3 1.2 26.5 1.2
2
Fat mass (kg) 30.5 2.2 25.5 2.4
2
Fat-free mass (kg) 72.9 2.9 69.4 2.6
2
Waist circumference (cm) 104.4 3.4 98.3 2.9
2
Hip circumference (cm) 114.2 2.5 108.4 2.0
2
Waist-to-hip ratio 0.92 0.03 0.91 0.02
1
All values are x SEM; n 9. V

O
2
max, maximumoxygen consump-
tion.
2
Significantly different from before intervention, P 0.05 (paired t
tests).
24 PETRIE ET AL
caffeine did not elicit any significant changes in blood glucose
concentrations before administration of the glucose beverage
during either OGTT nor was the AUC for blood glucose signif-
icantly affected.
Serum insulin concentrations
Insulin concentrations before and after the intervention are
summarized in Figure 2. The intervention resulted in a decrease
(P0.05) infastingseruminsulinconcentrationfrom89.56.5
to 53.4 3.6 pmol/L. Although the 36%decrease in the AUCfor
insulin (Table 3) in the placebo conditions was not significant
(P0.08), the ISI increased significantly as a result of the 12-wk
intervention (Table 3). There was a significantly greater insulin
response to the OGTT in the caffeine group than in the placebo
group both before and after the intervention (Table 3).
Serum C-peptide concentrations
Fasting serumC-peptide concentrations decreased (P 0.05)
after the intervention. The C-peptide response to the OGTTs,
before and after the intervention, are illustrated in Figure 3.
There was no significant effect of caffeine on the AUC for
C-peptide (Table 3). However, similar to the observation for
insulin, there was a trend (P 0.07) for a decrease (35%) in the
C-peptide response to glucose ingestion after the 12-wk program
(Table 3).
Serum proinsulin concentrations
After the 12-wk intervention, the proinsulin concentration de-
creased (P 0.05) at the 60-min time point (Table 4), but the
proinsulin-insulin ratio did not change significantly before or
after the intervention. There were no significant differences in
proinsulin concentrations between the placebo and caffeine
groups before the OGTT. During the OGTT, the proinsulin con-
centration increased (P0.05) 30 min after all 4 treatments, and
the proinsulin-insulin ratio increased (P 0.05). The interven-
tion did not significantly affect this ratio, but it was lower (P
0.05) with caffeine.
Blood lactate concentrations
There was no significant effect of the 12-wk intervention on
the blood lactate AUC(Table 3), yet there was a small (P0.05)
FIGURE 2. Mean (SEM) serum insulin concentrations in response to
an oral-glucose-tolerance test (OGTT). F, preintervention placebo; E, pre-
intervention caffeine; , postintervention placebo; , postintervention caf-
feine. n 9. There was a significant decrease in fasting seruminsulin (0 min)
after the intervention (P0.05). The insulin response during the OGTTwas
significantly higher in the caffeine group than in the placebo group, both
before and after the 12-wk intervention (P0.05; PROCMIXEDprocedure
in SAS, followed by paired t tests); this was confirmed in the area under the
curve analysis, as presented in Table 3.
FIGURE 3. Mean (SEM) serum C-peptide concentrations in response
to an oral-glucose-tolerance test (OGTT). F, preintervention placebo; E,
preintervention caffeine; , postintervention placebo; , postintervention
caffeine. n 9. There was a significant decrease in fasting serum C-peptide
(0 min) after the 12-wk intervention (P 0.05; PROCMIXEDprocedure in
SAS, followed by paired t tests).
TABLE 3
Area under the curve data for glucose, insulin, C-peptide, lactate, free fatty acids, and glycerol and the insulin sensitivity index in response to an oral-
glucose-tolerance test in subjects who received caffeine or placebo before and after a nutrition and exercise intervention
1
Substrate
Before intervention After intervention
Placebo Caffeine Placebo Caffeine
Glucose (mmol L 2 h) 234 62
a
258 41
a
191 52
a
242 61
a
Insulin (nmol L 2 h) 51.14 9.93
a
58.41 7.6
b
32.50 4.36
a
48.62 5.90
b
C-peptide (nmol L 2 h) 1573 187
a
1522 157
a
1027 178
a
1335 266
a
Lactate (mmol L 2 h) 18 7
a
24 13
a
21 7
a
39 9
a
Free fatty acids (mmol L 2 h) 33.01 5.02
a
54.62 7.63
b
21.85 4.31
c
49.60 8.35
b
Glycerol (mol L 2 h) 2684 1142
a
3226 862
a
1982 290
a
5259 1366
a
ISI 5 1
a
4 1
b
8 1
c
6 1
d
1
All values are x SEM; n 9. Means in a rowwith different superscript letters are significantly different. P0.05 (PROCMIXEDprocedure followed
by Tukeys test for multiple comparisons).
OBESITY AND CAFFEINE-INDUCED INSULIN RESISTANCE 25
decrease in fasting blood lactate concentrations (data not shown)
by the end of the 12-wk program. Ingestion of caffeine did not
elicit a significant change in fasting blood lactate concentrations
or in the blood lactate response to the OGTT.
Serum free fatty acid and glycerol concentrations
Fasting serumFFAs (Figure 4) and glycerol (data not shown)
concentrations increased (P 0.05) in the hour after caffeine
ingestion. This elevation did not significantly differ before or
after the intervention (P 0.05). At time 60, when the OGTT
began, serum insulin concentration began to rise, which coin-
cided with decreases in serumFFAs and glycerol concentrations
to below fasting concentrations. Because of the suppression of
lipolysis by insulin, the AUC for FFAs and glycerol were nega-
tive (Table 3). The AUC for FFA was significantly more pro-
nounced during the caffeine trials, both before and after the
intervention, but this finding appeared to result from the initial
elevation in FFAconcentrations at 60 min, because subsequently
the FFA concentrations decreased similarly during both the caf-
feine and placebo trials (Figure 4). Although the fasting serum
FFA concentration did not change as a result of the 12-wk pro-
gram (P 0.05), the serum FFA response to the placebo OGTT
was more pronounced as a result of the 12-wk program (Table 3
and Figure 4). No significant effect of caffeine on the AUC for
serum glycerol, before or after the 12-wk intervention, was ob-
served (Table 3).
DISCUSSION
The purpose of the present study was to examine the effects of
caffeine on glucose and insulin homeostasis in an obese, seden-
tarypopulationbefore andafter a 12-week, nutritionandexercise
weight-loss intervention. We hypothesized that caffeine inges-
tion before an OGTT would result in an exaggerated insulin
response compared with placebo. We also hypothesized that the
weight-loss program would lead to significant improvements in
glucose tolerance and an amelioration of the caffeine-induced
deterioration in glucose tolerance.
Our first hypothesis was supported. Before the 12-wk inter-
vention the estimated ISI was low and yet caffeine ingestion
decreased it significantly by 15%, and the insulin AUC was
significantly greater during the OGTT. Previously, studies (14,
15) of lean subjects have shown that caffeine ingestion followed
byanOGTTledtosignificantlygreater insulinresponses without
an accompanying decrease in blood glucose concentrations and
a similar relative decrease in the ISI. This hyperinsulinemia is
highly suggestive of insulin resistance, and studies using hyper-
insulinemic clamps and caffeine or theophylline have confirmed
that these methylxanthines depress glucose disposal (1618).
The present study showed that obese men, a population that is
already prone to insulin resistance, respond to caffeine similarly
before and after the intervention, even after weight loss and
increased aerobic fitness.
In contrast with previous work (14, 15), the serum C-peptide
response to the OGTTdid not significantly increase as a result of
caffeine ingestion. These prior studies examined lean, moder-
ately active men, whereas the present study examined sedentary,
obese men. C-peptide and insulin are secreted in equimolar
amounts and, hence, their presence in the serumusually changes
in parallel (34). Obese persons can experience elevated serum
insulin concentrations because of both increased insulin secre-
tion and decreased insulin clearance. Decreased hepatic binding
andextractionof insulinoftencharacterize insulinresistance (35,
36). This could explain the lack of concordance between the
serum C-peptide and insulin responses to caffeine ingestion in
this group.
Caffeine could increase the insulin response by stimulating a
more rapid absorption of glucose (37). This is highly unlikely;
caffeine was ingested 60 min before glucose ingestion, and caf-
feine absorption has been shown to be complete within 45 min of
ingestion (38). Furthermore, if enhanced absorption was the ma-
jor cause of the greater insulin concentration, the blood glucose
response would be reduced.
FIGURE 4. Mean ( SEM) serum free fatty acid (FFA) concentrations
in response to an oral-glucose-tolerance test. F, preintervention placebo; E,
preintervention caffeine; , postintervention placebo; , postintervention
caffeine. n 9. There was a significant increase in FFA concentrations 60
min after caffeine ingestion (at 0 min) (P 0.05; PROC MIXED procedure
in SAS, followed by paired t tests). The magnitude of the increase was not
significantly different after the 12-wk intervention.
TABLE 4
Serum proinsulin concentrations and the proinsulin-insulin ratio in
response to an oral-glucose-tolerance test in subjects who received
caffeine or placebo before and after a nutrition and exercise intervention
1
Proinsulin Proinsulin:insulin
pmol/L
60 min
Before intervention
Placebo 7.7 0.1 0.11 0.01
Caffeine 9.9 0.8 0.12 0.01
After intervention
Placebo 6.1 0.4
2
0.14 0.02
Caffeine 7.3 0.8
2
0.12 0.01
90 min
Before intervention
Placebo 30.0 5.5 0.50 0.01
Caffeine 28.2 4.0 0.46 0.01
3
After intervention
Placebo 17.6 3.0
2
0.50 0.01
Caffeine 19.8 2.6
2
0.40 0.01
3
1
All values are x SEM; n 9.
2
Significantly different from the corresponding value before interven-
tion, P 0.05.
3
Significantly different from the corresponding placebo value, P
0.05 (paired t tests).
26 PETRIE ET AL
It has been reported that pharmacologic concentrations of caf-
feine could directly stimulate the cell secretion of insulin (39).
To examine whether caffeine could alter the manner of cell
secretion, we measured proinsulin concentrations. Normal cell
secretion is composed of 98% insulin and 2% proinsulin-like
compounds (4042). However, because of differences in clear-
ance rates, the proinsulin-insulin ratio is commonly 0.100.15
in both lean and obese subjects (4043), and it is dramatically
elevated in type 2 diabetes because of secretion of immature
vesicles (44, 45). Fritsche et al (40) recently concluded that the
ratio at rest, and particularly that at 30 min of an OGTT, is an
accurate reflection of cell secretion. To the best of our knowl-
edge, caffeines possible actions on these processes have never
been examined. Caffeine had no significant effect on the
proinsulin-insulin ratio before the OGTT, and during the OGTT
it resulted in a significantly lower ratio than the corresponding
placebo value. If caffeine was interfering in the processing of
proinsulin or enhancing the secretion of immature vesicles, then
the secretion of proinsulin and the ratio should increase. The
decreased ratio strongly suggests that caffeine did not alter cell
secretion negatively. Rather, the data support the previous find-
ing (1418) that caffeine ingestion resulted in insulin resistance.
The 12-wk intervention significantly improved aerobic capac-
ity and decreased body weight, body fat, and fasting blood glu-
cose (13%), serum insulin (40%), C-peptide (33%), and proin-
sulin (21%) concentrations. During the placebo OGTTs, the
AUC for insulin decreased 37% (nonsignificant), and the ISI
increased 69% (significant) after the intervention. A further in-
dication of improved carbohydrate homeostasis was the obser-
vation that, before the intervention, 2 of the 9 subjects were
classified as having IGT (4) and neither subject fell into this
classification after the intervention. The proinsulin-insulin ratio
of these 2 subjects was lower than the group mean both before
and after intervention, which suggests that cell secretion was
not the problem. Rather, they may have been insulin resistant.
Several reports have shown that insulin resistance is not associ-
ated with an increase in this ratio (41, 43, 46).
The improvements in insulin sensitivity after the intervention
could be attributable to the various lifestyle improvements made
bythe subjects. The subjects significantlydecreasedtheir intakes
of dietary energy, total fat, and saturated fat and significantly
increased their intakes of fiber, all of which were previously
found to correlate with improved glucose tolerance and insulin
sensitivity (4751). The significant reductions in body fat are
also considered to be contributing factors (12, 13, 52). The sub-
jects also improved their aerobic fitness and refrained from in-
gestion of caffeine for 12 wk. Subjectively, there was no relation
between the degree of previous regular caffeine-consumption
habits and the degree of acute caffeine-induced insulin resis-
tance. All of the aforementioned factors could have affected the
results, and further studies are now required to address which of
the factors is most important.
We reject our hypothesis that a 12-week lifestyle intervention
would ameliorate the caffeine-induced insulin resistance ob-
served in the obese group studied. The effects of caffeine on
glucose, insulin, and C-peptide responses to the OGTT were not
significantly altered over time.
The caffeine dose (5 mg/kg) in this obese population was
equivalent to 34 cups of brewed coffee. However, it is im-
portant to note that the present study used pure caffeine rather
than a caffeine-containing beverage such as coffee and that the
caffeine challenge was given after a 2-d withdrawal from caffeine.
Coffee and other caffeinated products are composed of a myriad of
compounds. In fact, caffeine constitutes only 2% of coffees
chemical profile (53). The role of chronic coffee consumption is
controversial; a recent epidemiologic study reported that a high
coffee consumption is associated with a reduced risk of type 2
diabetes despite a high BMI, smoking, alcohol use, a low physical
activity, andanunfavorable diet (24). Incontrast, Knowler et al (25)
foundnoevidenceof arolefor coffeeconsumptionintype2diabetes
in a longitudinal population-based study. One should also realize
that there currently are novel caffeine-containing products such as
energy drinks and alcoholic beverages being marketed. Well-
controlled, prospective intervention studies are required to address
the long-termeffects of these various caffeinated products. It is not
known whether one habituates to the effects of caffeine on insulin
action, but it is noteworthy that, if habituation occurs, our present
data showthat it must be reversed within 48 h in subjects that have
consumed caffeine for many years.
The present study showed that acute caffeine ingestion signif-
icantly dampened whole-body insulin sensitivity in obese, non-
diabetic persons. A12-wk weight-loss program, which consisted
of the adoption of a healthy diet, regular aerobic exercise, and
caffeine withdrawal led to improvements in insulin sensitivity
but was unable to ameliorate the caffeine-induced deterioration
in insulin sensitivity. There was no evidence that caffeine di-
rectly affects cell secretion; rather, it appears that it directly or
indirectly causes peripheral insulin resistance, which in turn
stimulates greater insulinsecretion. The implications for caffeine
use by those who have preexisting hyperinsulinemic and hyper-
glycemic conditions, such as those with type 2 diabetes, remains
uncertain.
We thank the subjects for their valuable participation, our research assis-
tants, the staff of the Health and Performance Center and Human Nutraceu-
tical Research Unit at the University of Guelph, and Premila Sathasivamand
Patrick Sheridan for their technical expertise, time, and assistance.
HJP assisted with the development of the research study protocol, re-
cruited the subjects, cocoordinated the study, comanaged the research assis-
tants, performed the laboratory analyses, collected and interpreted data, and
wrote the manuscript. SEC assisted with the development of the research
study protocol, recruited the subjects, cocoordinated the research study,
comanaged research assistants, recruited the subjects, performed laboratory
analyses and collected data. LMB cocoordinated the research study, coman-
aged the research assistants, recruited the subjects, performed the laboratory
analyses, and collected data. AMD assisted with the supervision of data
collection, assistedwiththe data analysis andinterpretation, andassistedwith
the manuscript preparation. DHM established the assay for proinsulin, con-
ducted the analyses, and assisted with the data interpretation. JAC assisted
with the study design and supervised the data collection and analysis. TEG
designed the study, supervised the research, directed the data analysis and
interpretation, and assisted with the manuscript preparation. TEG was the
primary investigator of a research contract fromMuscletech Inc; he received
no personal financial support from the company.
REFERENCES
1. BirminghamCL, Muller JL, Palepu A, Spinelli JJ, Anis AH. The cost of
obesity in Canada. Can Med Assoc J 1999;160:4838.
2. Cairney J, Wade TJ. Correlates of body weight in the 1994 national popu-
lation health survey. Int J Obes Relat Metab Disord 1998;22:58491.
3. WHO. Obesity: preventing and managing the global epidemic. Report of
a WHO Consultation. World Health Organ Tech Rep Ser 2000;894.
4. WHO. Prevention of diabetes mellitus. Report of a WHO Study Group.
World Health Organ Tech Rep Ser 1994;844.
5. Anderson JW, Geil PB. Nutritional management of diabetes mellitus. In:
OBESITY AND CAFFEINE-INDUCED INSULIN RESISTANCE 27
Shils ME, Olson JA, Shike M, eds. Modern nutrition in health and
disease. Philadelphia: Lea & Febiger, 1994:125984.
6. DeFronzoRA, GunnarssonR, BjorkmanO, OlssonM, WahrenJ. Effects
of insulin on peripheral and splanchnic glucose metabolism in
noninsulin-dependent (type II) diabetes mellitus. J Clin Invest 1985;76:
14955.
7. Olefsky JM, Nolan JJ. Insulin resistance and non-insulin-dependent
diabetes mellitus: cellular and molecular mechanisms. Am J Clin Nutr
1995;61:S980S6.
8. Despres JP, Nadeau A, Tremblay A, et al. Role of deep abdominal fat in
the association between regional adipose tissue distribution and glucose
tolerance in obese women. Diabetes 1989;38:3049.
9. Nicklas BJ, Rogus EM, Colman EG, Goldberg AP. Visceral adiposity,
increased adipocyte lipolysis, and metabolic dysfunction in obese post-
menopausal women. Am J Physiol 1996;270:E728.
10. Pascot A, Despres JP, Lemieux I, et al. Contribution of visceral obesity
to the deterioration of the metabolic risk profile in men with impaired
glucose tolerance. Diabetologia 2000;43:112635.
11. Finnish Diabetes Prevention Group. Prevention of type 2 diabetes mel-
litus by changes in lifestyle among subjects with impaired glucose tol-
erance. N Engl J Med 2001;344:134350.
12. Diabetes Prevention Program Research Group. Reduction in the inci-
dence of type 2 diabetes with lifestyle intervention or metformin. NEngl
J Med 2002;346:393493.
13. Ross R, Dagnone D, Jones PJH, et al. Reduction in obesity and related
comorbid conditions after diet-induced weight loss or exercise-induced
weight loss in men. Ann Intern Med 2000;133:92103.
14. Graham TE, Sathasivam P, Rowland M, Marko N, Greer F, Battram D.
Caffeine ingestion elevates plasma insulin response in humans during an
oral glucose tolerance test. Can J Physiol Pharmacol 2001;79:55965.
15. Thong FS, Graham TE. Caffeine-induced impairment of glucose toler-
ance is abolished by beta-adrenergic receptor blockade in humans.
J Appl Physiol 2002;92:234752.
16. Greer F, Hudson R, Ross R, Graham TE. Caffeine ingestion decreases
glucose disposal during a hyperinsulinemic-euglycemic clamp in sed-
entary humans. Diabetes 2001;50:234954.
17. Keijzers GB, Galan BED, Tack CJ, Smits P. Caffeine can decrease
insulin sensitivity in humans. Diabetes Care 2002;25:3649.
18. Thong FSL, Derave W, Kiens B, et al. Caffeine-induced impairment of
insulin action but not insulin signaling in human skeletal muscle is
reduced by exercise. Diabetes 2002;51:58390.
19. Thong FSL, GrahamTE. The putative roles of adenosine in insulin- and
exercise-mediated regulation of glucose transport and glycogen metab-
olism in skeletal muscle. Can J Appl Physiol 2002;27:15278.
20. Fredholm BB. Adenosine, adenosine receptors and the actions of caf-
feine. Pharmacol Toxicol 1998;76:93101.
21. Graham TE. Caffeine and exercise: metabolism, endurance and perfor-
mance. Sports Med 2001;31:785807.
22. Vergauwen L, Hespel P, Richter EA. Adenosine receptors mediate syn-
ergistic stimulation of glucose uptake and transport by insulin and con-
tractions in rat skeletal muscle. J Clin Invest 1994;93:97481.
23. Han D-H, Hansen PA, Nolte LA, Holloszy JO. Removal of adenosine
decreases the responsiveness of skeletal muscle transport to insulin and
contractions. Diabetes 1998;47:167175.
24. van Dam RM, Feskens EJM. Coffee consumption and risk of type 2
diabetes mellitus. Lancet 2002;360:14778.
25. Knowler WC, Pettit DJ, Saad MF, Bennett PH. Diabetes mellitus in the
Pima Indians: incidence, risk factors and pathogenesis. Diabetes Metab
Rev 1990;6:17.
26. Carey GB, Sidmore KA. Exercise attenuates the anti-lipolytic effect of
adenosine in adipocytes isolated from miniature swine. Int J Obes Relat
Metab Disord 1994;18:15560.
27. CareyGB. Cellular adaptations infat tissue of exercise-trainedminiature
swine: role of excess intake. J Appl Physiol 2000;88:8817.
28. Lowry OH, Passanneau JV. Enzymatic analysis: a practical guide.
Towowa, NJ: Humana Press, 1993:1712.
29. Bergmeyer HU, Bernt SF, Stork H. D-glucose determination with hex-
okinase and glucose-6-phosphate dehydrogenase. In: Bergmeyer HU,
ed. Methods of enzymatic analysis. New York: Academic Press, 1974:
1196201.
30. Gutmann OI, Wahlefeld AW. L-()-Lactate determination with lactate
dehydrogenase and NAD. In: Bergmeyer HU, ed. Methods of enzymatic
analysis. New York: Academic Press, 1974:146475.
31. Tang BK, Zubovitis T, Kalow W. Determination of acetylated caffeine
metabolites by high-performance exclusive chromatography. J Chro-
matogr 1986;375:1703.
32. Allison DB, Paultre F, Maggio C, Mezzitis N, Pi Sunyer XFP. The use
of areas under curves in diabetes research. Diabetes Care 1995;18:24550.
33. Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained fromoral
glucose tolerance testing. Comparison with the euglycemic insulin
clamp. Diabetes Care 1999;22:146270.
34. Brodsky EG, Devlin JT. Hormone and nutrient interactions. In: Shils
ME, Olson JA, Shike M, eds. Modern nutrition in health and disease.
Philadelphia: Lea & Febiger, 1994:60318.
35. Bjorntorp P. Portal adipose tissue as a generator of risk factors for
cardiovascular disease and diabetes. Arteriosclerosis 1990;10:4936.
36. Roden M, Price TB, Perseghin G, et al. Mechanism of free fatty acid-
induced insulin resistance in humans. J Clin Invest 1996;97:285965.
37. Van Nieuwenhoven MA, Brummer RJM, Brouns F. Gastrointestinal
function during exercise: comparison of water, sports drink, and sports
drink with caffeine. J Appl Physiol 2000;89:107985.
38. Arnaud NJ. Metabolism of caffeine and other components of coffee. In:
Garattini S, ed. Caffeine, coffee and health. New York: Raven Press,
1993:4396.
39. Bruton JD, Lemmens R, Shi CL, et al. Ryanodine receptors of pancreatic
beta-cells mediate a distinct context-dependent signal for insulin secre-
tion. FASEB J 2003;17:3013.
40. Fritsche A, Madaus A, Stefan N, et al. Relationships among age, proin-
sulinconversion, andbeta-cell functioninnondiabetic humans. Diabetes
2002;51:S2349.
41. Nagi DK, Knowler WC, Mohamed-Ali V, Bennet PH, Yudkin JS. Intact
proinsulin, des 31,32 proinsulin, and specific insulin concentrations
among nondiabetic and diabetic subjects in populations at varying risks
of type 2 diabetes. Diabetes Care 1998;21:12733.
42. Reaven GM, Chen YD, Hollenback CB, Sheu WH, Ostrega D, Polonsky
KS. Plasma insulin, C-peptide, and proinsulin concentrations in obese
and nonobese individuals with varying degrees of glucose tolerance.
J Clin Endocrinol Metab 1993;76:448.
43. Wang PW, Abbasi F, Carantoni M, Chen YD, Azhar S, Reaven GM.
Insulin resistance does not change the ratio of proinsulin to insulin in
normal volunteers. J Clin Endocrinol Metab 1997;82:32214.
44. Shiraishi I, Iwamoto Y, Kuzuya T, Matsuda A, Kumakura S. Hyperin-
sulinemia in obesity is not accompanied by an increase in serum proin-
sulin/insulin ratio in groups of human subjects with and without glucose
intolerance. Diabetologia 1991;34:73741.
45. Polonsky KS, Gumbiner B, Ostrega D, Griver K, Tager H, Henry RR.
Alterations in immunoreactive proinsulin and insulin clearance induced
by weight loss in NIDDM. Diabetes 1994;43:8717.
46. Roder ME, Dinesen B, Harling SG, et al. Intact proinsulin and beta-cell
function in lean and obese subjects with and without type 2 diabetes.
Diabetes Care 1999;22:60914.
47. Lovejoy J, DiGirolamo M. Habitual dietary intake and insulin sensitivity
in lean and obese adults. Am J Clin Nutr 1992;55:11749.
48. Marshall JA, Bessessen DH, Hamman RF. High saturated fat and low
starch and fibre are associated with hyperinsulinemia in a non-diabetic
population: the San Luis Valley Diabetes Study. Diabetologia 1997;40:
4308.
49. Oakes ND, Cooney GJ, Camilleri S, ChisholmDJ, Kraegen EW. Mech-
anisms of liver and muscle insulin resistance induced by chronic high-fat
feeding. Diabetes 1997;46:176874.
50. Regensteiner JG, Mayer EJ, Shetterly SM, et al. Relationship between
habitual physical activity and insulin levels among nondiabetic men and
women. San Luis Valley Diabetes Study. Diabetes Care 1991;14:106674.
51. SonnichsenAC, Richter WO, Schwandt P. Benefit fromhypocaloric diet
in obese men depends on the extent of weight-loss regarding cholesterol,
and on a simultaneous change in body fat distribution regarding insulin
sensitivity and glucose tolerance. Metabolism 1992;41:10359.
52. Rice B, Janssen I, Hudson R, Ross R. Effects of aerobic or resistance
exercise and/or diet on glucose tolerance and plasma insulin levels in
obese men. Diabetes Care 1999;22:68491.
53. Spiller MA. The chemical components of coffee. In: Siller GA, ed.
Caffeine. New York: CRC Press, 1998:97161.
28 PETRIE ET AL

You might also like