Food Research International 89 (2016) 1023–1028

Contents lists available at ScienceDirect

Food Research International

journal homepage: www.elsevier.com/locate/foodres

Long-term consumption of a green/roasted coffee blend positively affects

glucose metabolism and insulin resistance in humans
Beatriz Sarriá ⁎, Sara Martínez-López, Raquel Mateos, Laura Bravo-Clemente
Department of Metabolism and Nutrition, Institute of Food Science, Technology and Nutrition (ICTAN-CSIC), Spanish National Research Council (CSIC), José Antonio Nováis 10, 28040 Madrid, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Protective health effects of coffee could have a widespread impact on public health considering the high intake of
Received 22 July 2015 this beverage in industrialized countries. However, certain of coffee's health effects are contradictory such as
Received in revised form 29 December 2015 those on type 2 diabetes mellitus (T2DM). Green coffee is richer in antioxidant phenols than roasted coffee,
Accepted 31 December 2015
and thus it is likely to be a healthier option. This work evaluated the effects of long-term consumption of green
Available online 4 January 2016
coffee consumption, blended with roasted beans to improve palatability, on different glucose homeostasis
Keywords:
markers as T2DM risk factors. A, randomized, controlled, crossover study was performed in 52 healthy men
Coffee and women who consumed three servings/day of the green/roasted (35:65) coffee blend for 8 weeks during
Type 2 diabetes mellitus the intervention in comparison with not consuming coffee in the control stage. At the beginning and end of
Glucose the coffee and control interventions, blood samples were collected, body weight measured, and dietary records
Insulin and physical activity questionnaires completed. After the coffee intervention, fasting glucose levels and HOMA-
HOMA IR values were significantly lower, whereas QUICKI values were higher showing improved insulin sensitivity.
Glucagon Fasting glucagon levels decreased, which may be associated with the increase in the glucagon-like peptide-1
Glucose-dependent insulinotropic polypeptide
(GLP-1), whereas C-peptide, glucose-dependent insulinotropic polypeptide (GIP), insulin, and HOMA-β were
Glucagon-like peptide-1
not affected. In conclusion, regularly consuming the green/roasted coffee blend may be recommended to prevent
T2DM and reduce cardiovascular risk.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction very limited during fasting conditions, as circulating concentrations of

The worldwide prevalence of type 2 diabetes mellitus (T2DM) is in-
creasing globally and may reach 366 million people by 2030 (Ding,
Bhupathiraju, Chen, van Dam, & Hu, 2014). T2DM is associated with var-
iable degrees of insulin resistance, impaired insulin secretion, moderate
to severe beta-cell apoptosis and increased hepatic glucose production.
Unlike type 1 diabetes mellitus, the onset of T2DM is slow and the met-
abolic abnormalities that lead to hyperglycemia are established long be-
fore overt diabetes. Hepatic glucose production is the main contributor
to fasting plasma glucose concentration and is regulated primarily by
plasma insulin and glucagon concentrations (Abdul-Ghani, Williams,
DeFronzo, & Stern, 2006). In turn, incretin hormones glucagon-like pep-
tide 1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP),
which are secreted in response to food intake, contribute to the regula-
tion of glucagon and glucose-dependent insulin secretion. Both
incretins stimulate insulin secretion, although they exert opposing ef-
fects on glucagon, since GLP-1 suppresses and GIP enhances glucagon
secretion (Yabe et al., 2015). The effects of the incretin hormones are
⁎ Corresponding author at: Institute of Food Science, Technology and Nutrition (ICTAN-
CSIC), José Antonio Novais 10, 28040 Madrid, Spain.
E-mail address: beasarria@ictan.csic.es (B. Sarriá).
GLP-1 and GIP are low (Nolan & Færch, 2012).
Many studies have shown that dietary components or foods affect
postprandial glucose, glucagon, insulin and incretin hormones; howev-
er, less have examined the effects of regular consumption of dietary
components on the fasting concentrations of the glucose metabolism
related biomarkers. Particularly, studies looking into the effects of
micronutrients and phytochemicals on glucose metabolism are scarce.
Advances in understanding the anti-diabetic actions of dietary flavo-
noids have been recently reviewed by Babu, Liu, and Gilbert (2013). In
a previous revision by van Dam (2006), consumption of coffee, rich in
hydroxycinnamic acids and caffeine, was pointed to affect postprandial
glucose metabolism rather than fasting glucose levels. Some human tri-
als have shown that glucose tolerance is reduced shortly after ingestion
of caffeine or caffeinated coffee, suggesting that short-term coffee con-
sumption could increase the risk of diabetes (Olthof, van Dijk, Deacon,
Heine, & van Dam, 2011). However, there is increasing scientific evi-
dence that supports an inverse relationship between coffee consump-
tion and T2DM (Akash, Rehman, & Chen, 2014), being stronger the
association with decaffeinated coffee (Pereira, Parker, & Folsom,
2006); therefore, it has been suggested that the positive, T2DM protec-
tive effects of coffee are associated with non-caffeine compounds.
Coffee contains diterpens (cafestol and kahweol, up to 0.6% of final
weight), and micronutrients, among which outstands magnesium

http://dx.doi.org/10.1016/j.foodres.2015.12.032
0963-9969/© 2016 Elsevier Ltd. All rights reserved.
1024 B. Sarriá et al. / Food Research International 89 (2016) 1023–1028

(Ding et al., 2014), but it is also a rich source of polyphenols (up to 11%
of the coffee bean). The main phenolic compounds in green coffee are
hydroxycinnamic acids, mostly 3-, 4-, and 5-caffeoylquinic acids (3-,
4- and 5-CQA), 3,4-, 3,5-, and 4,5-dicaffeoylquinic acids (3,4-, 3,5-, and
4,5-DCQA), and 3-, 4-, and 5-feruloylquinic acids (3-, 4-, and 5-FQA),
among others (Alonso-Salces, Serra, Reniero, & Héberger, 2009).
Roasting drastically degrades and/or transforms green coffee polyphe-
nols (Perrone, Farah, & Donangelo, 2012; Schenker et al., 2002;
Somporn, Kamtuo, Theerakulpisut, & Siriamornpun, 2011) inducing
the formation of Maillard reaction products and quinides. The intake
of green coffee products has increased in recent years as a healthier op-
tion than roasted coffee (Kozuma, Tsuchiya, Kohori, Hase, & Tokimitsu,
2005), although its bitter taste limits green coffee acceptance. Bearing
this in mind, a blend of green and roasted beans can be an alternative
well accepted by coffee consumers with greater health potential than
traditional roasted coffee.
Considering the high intake of this beverage, particularly in industri-
alized countries where it is the largest source of dietary antioxidants
(Tunnicliffe & Shearer, 2008), the health protective effects of coffee
could have a widespread impact on the population health. Moreover,
coffee could be recommended to patients at risk of T2DM as a supple-
mentary therapy in preventing the further progression of the disease
or to prevent the onset in healthy and at risk adults. Therefore, the
aim of this work was to evaluate the effects of long term consumption
of a green/roasted (35/65) coffee blend on glucose homeostasis
markers, as T2DM risk factors, in healthy adults, attempting to under-
stand the mechanisms involved.
2. Experimental methods
2.1. Subjects
This study was conducted according to the guidelines laid down in
the Declaration of Helsinki and all procedures were approved by the
Clinical Research Ethics Committee of Hospital Universitario Puerta de
Hierro Majadahonda in Madrid (Spain). Written informed consent
was obtained from all subjects. Volunteer recruitment was carried out
through placing advertisements in the Universidad Complutense cam-
pus as well as through giving short talks between lectures. The inclusion
criteria were: being non-diabetic (excluded by the results of a glucose
test and health questionnaire), non-vegetarian, non-smoker, non-
pregnant women and men, between 18 and 55 y old, not suffering
from any other chronic pathology and presenting a body mass index be-
tween 20 and 25 kg/m2. None had taken dietary supplements, laxatives,
or antibiotics six months before the start of the study.
Fifty-three subjects initially accepted to participate in the study,
however 52 completed it. Baseline characteristics of the volunteers are
shown in Table 1.
2.2. Study design
intervention, instead of the coffee product, the study participants had
water or an isotonic drink, free of sugar, polyphenols and methylxan-
thines. The soluble green/roasted coffee blend, which was commercial-
ized at the time of the study, was provided by the manufacturing
company in unlabelled, individual sachets containing 2 g of coffee
(equivalent to two teaspoons, quantity that can reasonably be used to
prepare a cup of coffee). The coffee studied contained 85.1 ± 1.6 mg/g
(dry matter) of total hydroxycinnamic acids (mainly chlorogenic acid)
and 20.0 ± 1.8 mg/g (dry matter) of caffeine. Therefore, volunteers
daily consumed 6 g of the coffee blend which provided 510.6 and
120 mg of total hydroxycinnamic acids and caffeine, respectively. The
green/roasted coffee blend was particularly interesting to study because
on the one hand, it is richer in chlorogenic acid than roasted coffee and
thus was expected to be healthier, and on the other hand, the blend
keeps the organoleptic properties of roasted coffee (which green coffee
lacks) that are much appreciated by coffee drinkers, adding to its ac-
ceptability by consumers. From the run-in stage till the end of the
study, foods rich in polyphenols and methylxanthines were restricted.
Hydroxycinnamic acids are abundant in a variety of fruits and vegeta-
bles, such as chard, artichoke, eggplant, broccoli, loquats, tangerines, or-
anges, apricot, cherries, plums, prunes, grapes, raisins, blueberries and
other fruits of the forest. All these foods were constrained, as well as cof-
fee, mate, cocoa, and tea and derived drinks. On the other hand, ferulic
acid and its derivatives are the most abundant hydroxycinnamic acids
found in cereals, thus whole grain products were also restricted along
the study.
2.3. Dietary control and compliance
Subjects were asked to maintain the same dietary habits along the
study. Their dietary intake was regularly evaluated to control any possi-
ble changes. Volunteers were instructed on how to fill in the dietary re-
cords before starting the study. In the run in stage and at the end of the
two intervention periods, volunteers were asked to complete a 72-hour
detailed food intake report, specifying the ingredients and amounts of
food consumed, including serving weights (when possible) or house-
hold measurements. Compliance was controlled by counting the num-
ber of coffee sachets provided to the volunteers before and after the
intervention, as well as by weekly calling the volunteers. In order to as-
sess dietary composition, the program DIAL [Department of Nutrition
and Bromathology I. School of Pharmacy. Complutense University of
Madrid (UCM), Spain] was used.
2.4. Blood samples
Blood samples were drawn after 8–10 h overnight fasting at baseline
and at the last day of the control and coffee intervention. Serum (with-
out anticoagulant) and plasma (EDTA-coated tubes) were separated by
centrifugation and frozen at −80 °C until analysis.

This was a randomized, controlled, crossover study carried out in
free-living people. After a 2 week run in stage, subjects were randomly
assigned to the coffee or control intervention, lasting 8 weeks each,
which were separated by a 2 week washout period. During the coffee in-
tervention, volunteers consumed three times a day the soluble green/
roasted coffee blend, the first at breakfast, the second between breakfast
and lunch, and the third between lunch and dinner. In the control
Table 1
Baseline characteristics of the participants in the study
Women (n = 32)
Age (years) 29.4 ± 9.5
Body mass index (kg/m2) 21.7 ± 2.5
Data represents mean ± standard deviation of mean.
2.5. Diabetes biomarkers and related indexes
Fasting glucose was analyzed using a colorimetric kit (Sprinreact).
Fasting insulin, as well as GIP, GLP-1, C-peptide, and glucagon were ana-
lyzed using the Bio-Rad Multiplex Diabetes kit on Bio-Plex MAGPIX sys-
tem. Using fasting glucose and insulin data, Homeostasis Model
Assessment indexes were calculated to estimate insulin resistance
(HOMA-IR) and beta cell function (HOMA-β) according to the equations
by Matthews et al. (1985): HOMA-IR = [Glucose (mg/dL) × Insulin
(mU/L)] / 405; HOMA-β = [(Insulin (mU/L) × 360) / (Glucose
(mg/dL) − 63)]. Another model to calculate beta- cell function,
Men (n = 20)
the insulin/glucose ratio, was also used (Meier et al., 2001). In addition,
29.8 ± 8.9 the Quantitative Insulin Sensitivity Check Index (QUICKI) was calculat-
24.8 ± 2.7 ed according to the formula by Katz et al. (2000): QUICKI = 1 / [log In-
sulin (mU/L) + log Glucose (mg/dL)].
 B. Sarriá et al. / Food Research International 89 (2016) 1023–1028 1025

2.6. Physical activity and body weight Table 3

Effects of regularly consuming the green/roasted coffee blend on fasting glucose and insu-
lin levels and indexes of insulin resistance/sensitivity (HOMA-IR/QUICKI) and pancreatic
Participants were asked to maintain their usual level of physical ac- function (HOMA-β, insulin/glucose) in healthy subjects.
tivity during the study. Volunteers filled out a questionnaire before
starting the study to evaluate their physical activity including that in- Basal Control Coffee P

volved in their occupation; afterwards this was calculated using the pro- Glucose (mg/dL) 75.47 ± 1.10 a 76.50 ± 1.24 a 72.35 ± 1.1 b 0.007
gram ADN [Department of Nutrition and Bromathology I, School of Insulin (mU/L) 8.20 ± 0.45 8.42 ± 0.45 7.73 ± 0.44 N.S.
HOMA-IR 1.67 ± 0.06 a 1.72 ± 0.05 a 1.47 ± 0.05 b 0.001
Pharmacy, Complutense University of Madrid (UCM), Spain]. At base-
HOMA-β (%) 180.52 ± 18.69 159.63 ± 18.42 179.27 ± 20.71 N.S.
line and the end of each stage of the study, volunteer's body weight Insulin/glucose 0.07 ± 0.01 0.08 ± 0.01 0.07 ± 0.01 N.S.
was monitored using the weighing system BC-418 MA (Tanita [(mU/L)/(mg/dL)]
Corporation). QUICKI 0.36 ± 0.00 a 0.35 ± 0.00 a 0.36 ± 0.00 b 0.008
C-peptide (pg/mL) 924.74 ± 28.88 923.47 ± 28.52 882.47 ± 30.26 N.S.

Data represents mean ± standard error of mean. HOMA-IR: homeostasis model assess-
2.7. Statistical analysis ment of insulin resistance; HOMA-β: homeostasis model assessment of beta cell function;
QUICKI: quantitative insulin sensitivity check index. P values were assessed using the gen-
Data are presented as means ± standard error of the mean, unless eral linear model of variance for repeated measures. a.b Mean values within a row with
specified otherwise. Prior to statistical analysis, normality of distribu- unlike letters were significantly different according to the Bonferroni test.
tion and homogeneity of variance were verified using the Kolmogo-
rov–Smirnov and Levene tests, respectively. The general linear model
In agreement to insulin results, C-peptide concentrations did not
of the variance for repeated measures was used to assess the effects of
change due to coffee consumption. Contrarily, glucagon concentration
consuming coffee, followed by a Bonferroni test to compare the stages
significantly changed along the study, showing lower values after the
in pairs. Statistical significance was set at P b 0.05 and the analysis
control and coffee intervention than at baseline (Fig. 1A). Glucagon con-
was undertaken using the SPSS statistical package (version 21.0, SPSS
centrations were similar to those described by Knop et al. (2013) in
Inc., IBM Company).
obese (BMI = 32 ± 1 kg/m2) but otherwise healthy men. The decrease
in glucagon concentration may be related with the significant increase
3. Results in GLP-1 levels after the coffee intervention compared to baseline values
(Fig. 1B). Differently, there were not significant changes in GIP levels
Attending to volunteers' reports and to the number of servings (Fig. 1C). Insulin, glucagon, C-peptide, GLP-1 and GIP values were
returned after the intervention, dietary compliance was high. The 72- within the ranges observed by Wang, Zhou, Yaung, Ma, and Geng
hour detailed food intake reports showed that energy, protein, carbohy- (2010).
drate, dietary fiber, lipid and the polyunsaturated/saturated ratio did
not show differences along the study. However, the consumption of
the coffee blend significantly reduced body weight (Table 2). 4. Discussion
Baseline fasting glucose concentration values were within the range
of normality according to the Spanish Society of Clinical Biochemistry There are many epidemiological and prospective cohort studies that
and Molecular Pathology (SEQC) (Table 3), and insulin results were in point to an inverse association between the effects of habitual coffee
agreement with previous studies (Lecoultre et al., 2014; Meier et al., consumption and the risk of T2DM (revised in van Dam, 2006; Ding
2001). The values of HOMA-IR were similar to those earlier observed et al., 2014). In contrast, long term, randomized, controlled interven-
in healthy adults (Acosta, Escalona, Maiz, Pollak, & Leighton, 2002; tions are scarce, although necessary to understand the tolerance to the
Lichnovská, Gwozdziewiczová, & Hrebícek, 2002). Insulin sensitivity re- acute effects of coffee components developed after sustained consump-
sults, estimated through the QUICKI index (Table 3), which keeps a good tion, to elucidate the underlying mechanisms involved in regular con-
correlation with HOMA-IR index as both are calculated from the values sumption and to establish causality of the effects observed in
of fasting glucose and insulin, were similar to those described by Katz epidemiological and cohort studies. Caffeine intake has been associated
et al. (2000) in healthy volunteers. Finally, the values of HOMA-β with an acute reduction in insulin sensitivity due to increased epineph-
index observed were analogous to those reported by Meier et al. rine release, declining this effect after continued intake (van Dam,
(2001) in healthy controls. 2006). When dose-dependency of caffeine metabolism under multiple
Among all the aforementioned glucose homeostasis biomarkers and dosing (4.2 and 12 mg/kg/day of caffeine, divided in 6 doses) was ran-
related indexes, only glucose and HOMA-IR decreased after the coffee domly tested in healthy subjects, complete tolerance to the effects of
intervention, in contrast to insulin sensitivity which increased, being caffeine was developed after 5 days (Denaro, Brown, Jacob, &
the values observed after the coffee intervention significantly different Benowitz, 1991). Conversely, another study in healthy volunteers de-
to basal and control data. scribed that caffeine tolerance development takes longer, as after con-
suming 1 L of coffee/day (caffeine intake was not specified) for
2 weeks, fasting glucose concentrations were higher compared to base-
Table 2 line, whereas after 4 weeks not (van Dam, Pasman, & Verhoef, 2004). It
Reported energy, macronutrient and dietary fiber intakes, PUFA/SFA intake ratio and body is likely that in the present study caffeine tolerance occurred, because
weight.
the dose of caffeine was relatively low (120 mg/day, i.e. 2.1 and
Basal Control Coffee P 1.5 mg/kg/day for women and men, respectively) and the coffee inter-
Intake per day vention was 8 weeks long. Nevertheless, according to observational
Energy (Kcal) 1870.12 ± 72.15 1854.83 ± 63.50 1801.46 ± 60.11 N.S. studies (using caffeinated and decaffeinated coffee) the inverse associa-
Protein (g) 83.81 ± 3.31 77.73 ± 2.93 76.46 ± 2.84 N.S. tion between coffee consumption and risk of T2DM should be attributed
Carbohydrate (g) 181.59 ± 8.21 175.77 ± 6.59 174.25 ± 7.56 N.S.
to coffee's content in chlorogenic acid rather than caffeine (Ding et al.,
Dietary fiber (g) 15.23 ± 0.82 15.99 ± 0.81 15.47 ± 0.75 N.S.
Lipid (g) 77.94 ± 3.16 74.19 ± 2.46 72.20 ± 3.46 N.S. 2014). In agreement, attending to the bioavailability studies in humans
PUFA/SFA 0.42 ± 0.02 0.39 ± 0.02 0.46 ± 0.03 N.S. carried out in our group using the green/roasted coffee
Body weight (kg) 63.01 ± 1.71 62.85 ± 1.73 62.31 ± 1.76 0.024 (Martínez-López, Sarriá, Baeza, Mateos, & Bravo-Clemente, 2014;
Data represents mean ± standard error of mean. P values were assessed using the general Mateos et al., 2015), methylxanthines in the coffee blend may have con-
linear model of variance for repeated measures. tributed to the biological activity of the coffee product observed,
1026 B. Sarriá et al. / Food Research International 89 (2016) 1023–1028
Fig. 1. Effects of consuming the green/roasted coffee blend on the concentration of
A) Glucagon; B) GLP-1: glucagon like peptide-1; and C) GIP: glucose- dependent
insulinotrophic peptide. Data represents mean ± standard error of mean. Different
letters represent statistically significant differences between the stages of the study
(P b 0.05).
although to a lower extent than the phenolic compounds in coffee as the
time circulating in plasma of the phenolic metabolites was higher.
Phenolic acids in coffee and their degradation products formed dur-
ing roasting (quinides) may impact glucose absorption in the intestine
through the inhibition of glucose-6-phosphate translocase-1, an en-
zyme known to play a role in intestinal glucose transport, and reducing
the sodium gradient driven apical glucose transport (McCarty, 2005).
According to in vitro (Arion et al., 1997) and animal studies (Herling
et al., 1999), another possible mechanism responsible for the benefits
of chlorogenic acid and its derivatives on glucose metabolism is the de-
crease of hepatic glucose output through inhibition of glucose-6 phos-
phatase. The third mechanism is supported by a metabolic study in
ileostomy patients, suggesting that most of the chlorogenic acid
absorbed in the small intestine is absorbed intact and may be extensive-
ly metabolized in the liver (Olthof, Hollman, & Katan, 2001).
The aforementioned mechanisms may explain the decrease in
fasting glucose levels observed in the present study after regularly con-
suming coffee. This outcome is in agreement with previous observation-
al studies (Hino et al., 2007; Pham et al., 2014) and a clinical trial (van
Dam et al., 2004) but not with other interventions (Kempf et al., 2010;
Rebello et al., 2011). It would have been interesting to carry out an
oral glucose tolerance test or a meal tolerance test to better understand
the underlying physiology beyond the fasting state. According to the lit-
erature, glucose increase in the oral glucose tolerance test is lower in ha-
bitual coffee drinkers than in non-drinkers (Rustenbeck et al., 2014; van
Dam et al., 2004; Yamaji et al., 2004), and there are no changes in fasting
or postprandial glucose and insulin (30 min and 2 h after intake) re-
sponses between consuming 4 or 8 cups of coffee/day for 4 weeks
(Kempf et al., 2010).
In contrast to fasting glucose concentrations, fasting insulin levels did
not change after the coffee intervention, but interestingly C-peptide
levels were reduced 4.7% and thus a positive tendency at insulin produc-
tion can be inferred, which possibly might have been observed if the
study had been longer and/or the intake of coffee higher. The C-peptide
outcome is in accordance with a cross-sectional study that described
that greater intakes of caffeinated and decaffeinated coffee were associ-
ated with lower fasting C-peptide concentrations in healthy women,
being stronger the caffeinated coffee and C-peptide relationship in
obese and overweight women (27 and 20% reduction, respectively)
than in normoweight (11%; Wu, Willet, Hankinson, & Giovannnucci,
2005). Thus, it is possible that in this study, because the participants pre-
sented a BMI under 25 kg/m2, the effect of coffee on C-peptide was
attenuated.
Interestingly, the changes in glucose and insulin observed after the
coffee intervention resulted in a decrease in insulin resistance and in-
crease of insulin sensitivity, according to the HOMA-IR (which reflects
hepatic more than peripheral insulin resistance, Nolan & Færch, 2012)
and QUICKI indexes, respectively. An inverse association between coffee
consumption and HOMA-IR has been reported in epidemiological and
cross-sectional studies carried out in different ethnic populations al-
though not in a Japanese study (Pham et al., 2014), neither in three ran-
domized, controlled trials (Kempf et al., 2010; Ohnaka et al., 2012;
Wedick et al., 2011). However, when the results of the Japanese study
(Pham et al., 2014) were stratified by BMI, an inverse association be-
tween coffee and HOMA-IR was observed in the overweight/obese sub-
jects but not in the normoweight, which could be related to the stronger
inverse association between caffeinated coffee and C-peptide in obese
and overweight women than normoweight (Wu et al., 2005). Perhaps
the inverse association between coffee consumption and HOMA-IR
was attenuated in the present study because subjects' BMI b25 kg/m2,
nevertheless the beneficial effect of coffee on HOMA-IR was observed
and reached the level of statistical significance.
In contrast to HOMA-IR, HOMA-β has been less rigorously evaluated
and is a less robust estimate of beta-cell function (Nolan & Færch, 2012).
Therefore, the insulin/glucose ratio was also calculated. According to
both indexes of beta-cell function, regular consumption of the coffee
blend did not produce changes, which is in accordance with epidemio-
logical studies in healthy subjects (Agardh et al., 2004; Rebello et al.,
2011).
In T2DM, glucagon secretion is enhanced; in fact, hyperglucagonemia
contributes importantly to the hyperglycemia observed in T2DM patients.
Suppression of glucagon secretion is a possible treatment of the disease
(Knop et al., 2013), and drugs that suppress glucagon secretion or antag-
onize the glucagon receptor have already been developed (Christensen,
Bagger, Vilsbøll, & Knop, 2011). Activation of GLP-1 receptors effectively
inhibits glucagon secretion in humans, decelerating gastric emptying,
inhibiting food intake, and elevating insulin secretion (Christensen et al.,
2011). There are incretin-based therapies based on two major classes of
drugs: GLP-1 receptor agonists and DPP-4 inhibitors, which increase
 B. Sarriá et al. / Food Research International 89 (2016) 1023–1028
GLP-1 receptor signaling by means of an exogenous GLP-1 analog and en-
hancement of endogenous GLP-1 levels, respectively. In contrast to the
aforementioned pharmacological treatments based on targeting the
alpha-cell, the effects of dietary compounds on glucagon secretion and
incretin hormones have been less studied. In a recent trial, 30 day resver-
atrol supplementation suppressed postprandial glucagon response with-
out affecting fasting glucagon levels (Knop et al., 2013). In contrast, in the
present work the intake of the green/roasted coffee decreased fasting glu-
cagon levels compared to baseline, which may be related to the increase
of GLP-1 concentration after the coffee intervention. Although the mech-
anism of GLP-1 inhibition on glucagon secretion is not entirely clear, it has
been previously described at normal fasting glucose levels (Christensen
et al., 2011) in agreement with the present study.
In order to control lifestyle confounding factors that may influence
the risk of developing T2DM, such as smoking, age and body mass
index, these factors were considered when the inclusion criteria was
established. The volunteers who carried out this study were non-
smokers and their age and body mass were within a narrow range,
thus minimizing their possible influence. Physical activity and dietary
intake are other confounding factors that were followed and controlled,
as far as possible, through questionnaires. Attending to participants' an-
swers, they were sedentary and maintained their physical activity unal-
tered along the study (data not shown). Concerning dietary intake,
confounding factors known to affect T2DM are energy, polyunsaturat-
ed/saturated fat and dietary fiber intake (Zhang, Lee, Cowan, Fabsitz, &
Howard, 2011), which showed no changes during the study. Partici-
pants were instructed to avoid or use minimum amount of milk with
the coffee, and non-caloric sweeteners were recommended instead of
sugar. All these factors were controlled so the effects on glucose metab-
olism observed could be attributed to the consumption of the green/
roasted coffee blend as far as possible.
5. Strengths & limitations
One strength of this study was its design (randomized, controlled
and crossover) and in addition, each intervention lasted 8 weeks. Con-
founding factors, which may influence the effects of coffee drinking,
such as BMI, physical activity, smoking, and certain other dietary factors
were considered at setting the inclusion criteria and monitored through
questionnaires. Limitations of the study: fasting glucose was measured
at baseline and after the interventions; however, an oral glucose toler-
ance test was not carried out. We relied the absence of diabetes on a
fasting glucose blood test and a self-reported health questionnaire filled
out before the start of the study.
6. Conclusions
This study supports that regularly consuming the green/roasted cof-
fee blend lowers fasting glucose levels and insulin resistance and in-
creases insulin sensitivity. In addition, fasting glucagon levels
decrease, which may be associated with the increase in fasting GLP-1
concentrations. All these effects positively contribute to the prevention
of T2DM and lower cardiovascular risk. Therefore, the use of the green/
roasted coffee blend within a balanced diet could be recommended to
patients at risk of T2DM as a supplementary therapy or to healthy adults
in order to prevent the onset of the disease. Considering the high intake
of coffee in industrialized countries, these results are of interest from a
public health viewpoint.
Acknowledgments
This work was funded by the Spanish Ministry of Science and Inno-
vation (projects AGL2010-18269 and Consolider-Ingenio CSD2007-
00063). The authors thank L. T. Cayuelas and M. Jiménez-Romano for
analyzing the 72 h food records and Nestlé España, S.A. for providing
the coffee and packing it in blind monodosis especially for the study.
1027
S.M.-L. thanks the Spanish National Research Council for her predoctoral
fellowship under the JAE-Pre program funded by the European Social
Fund (JAE-Pre 097. BOE 17-12-2008). All authors revised and approved
the final version of the manuscript. The authors declare no conflicts of
interest.
References
Abdul-Ghani, M. A., Williams, K., DeFronzo, R., & Stern, M. (2006). Risk of progression to
type 2 diabetes based on relationship between postload plasma glucose and fasting
plasma glucose. Diabetes Care, 29, 1613–1618.
Acosta, A. M., Escalona, M., Maiz, A., Pollak, F., & Leighton, F. (2002). Determinación del ín-
dice de resistencia insulínica mediante HOMA en una población de la Región
Metropolitana de Chile. Revista Médica de Chile, 130, 1227–1231.
Agardh, E. E., Carlsson, S., Ahlbom, A., Efendic, S., Grill, V., Hammar, N., et al. (2004). Coffee
consumption, type 2 diabetes and impaired glucose tolerance in Swedish men and
women. Journal of Internal Medicine, 255, 645–652.
Akash, M. S., Rehman, K., & Chen, S. (2014). Effects of coffee on type 2 diabetes mellitus.
Nutrition, 30, 755–763.
Alonso-Salces, R. M., Serra, F., Reniero, F., & Héberger, K. (2009). Botanical and geograph-
ical characterization of green coffee (Coffea arabica and Coffea canephora): Chemo-
metric evaluation of phenolic and methylxanthine contents. Journal of Agricultural
and Food Chemistry, 57, 4224–4235.
Arion, W. J., Canfield, W. K., Ramos, F. C., Schindler, P. W., Burger, H. J., Hemmerle, H., et al.
(1997). Chlorogenic acid and hydroxynitrobenzaldehyde: New inhibitors of hepatic
glucose 6-phosphatase. Archives of Biochemistry and Biophysics, 339, 315–322.
Babu, P. V., Liu, D., & Gilbert, E. R. (2013). Recent advances in understanding the anti-diabetic
actions of dietary flavonoids. Journal of Nutritional Biochemistry, 24, 1777–1789.
Christensen, M., Bagger, J. I., Vilsbøll, T., & Knop, F. K. (2011). The alpha-cell as target for
type 2 diabetes therapy. The Review of Diabetic Studies, 8, 369–381.
van Dam, R. M. (2006). Coffee and type 2 diabetes: From beans to beta-cells. Nutrition,
Metabolism, and Cardiovascular Diseases, 16, 69–77.
van Dam, R. M., Pasman, W. J., & Verhoef, P. (2004). Effects of coffee consumption on
fasting blood glucose and insulin concentrations: Randomized controlled trials in
healthy volunteers. Diabetes Care, 27, 2990–2992.
Denaro, C. P., Brown, C. R., Jacob, P. I. I. I., & Benowitz, N. L. (1991). Effects of caffeine with
repeated dosing. European Journal of Clinical Pharmacology, 40, 273–278.
Ding, M., Bhupathiraju, S. N., Chen, M., van Dam, R. M., & Hu, F. B. (2014). Caffeinated and
decaffeinated coffee consumption and risk of type 2 diabetes: A systematic review
and a dose–response meta-analysis. Diabetes Care, 37, 569–586.
Herling, A. W., Burger, H., Schubert, G., Hemmerle, H., Schaefer, H., & Kramer, W. (1999).
Alterations of carbohydrate and lipid intermediary metabolism during inhibition of
glucose-6-phosphatase in rats. European Journal of Pharmacology, 386, 75–82.
Hino, A., Adachi, H., Enomoto, M., Furuki, K., Shigetoh, Y., Ohtsuka, M., et al. (2007). Habit-
ual coffee but not green tea consumption is inversely associated with metabolic syn-
drome: An epidemiological study in a general Japanese population. Diabetes Research
and Clinical Practice, 76, 383–389.
Katz, A., Nambi, S. S., Mather, K., Baron, A. D., Follmann, D. A., Sullivan, G., & Quon, M. J.
(2000). Quantitative insulin sensitivity check index: A simple, accurate method for
assessing insulin sensitivity in humans. The Journal of Clinical Endocrinology and
Metabolism, 85, 2402–2410.
Kempf, K., Herder, C., Erlund, I., Kolb, H., Martin, S., Carstensen, et al. (2010). Effects of cof-
fee consumption on subclinical inflammation and other risk factors for type 2 diabe-
tes: A clinical trial. The American Journal of Clinical Nutrition, 91, 950–957.
Knop, F. K., Konings, E., Timmers, S., Schrauwen, P., Holst, J. J., & Blaak, E. E. (2013). Thirty
days of resveratrol supplementation does not affect postprandial incretin hormone
responses, but suppresses postprandial glucagon in obese subjects. Diabetic
Medicine, 30, 1214–1218.
Kozuma, K., Tsuchiya, S., Kohori, J., Hase, T., & Tokimitsu, I. (2005). Antihypertensive effect
of green coffee bean extract on mildly hypertension subjects. Hypertension Research,
28, 711–718.
Lecoultre, V., Carrel, G., Egli, L., Binnert, C., Boss, A., MacMillan, E. L., et al. (2014). Coffee
consumption attenuates short-term fructose-induced liver insulin resistance in
healthy men. The American Journal of Clinical Nutrition, 99, 268–275.
Lichnovská, R., Gwozdziewiczová, S., & Hrebícek, J. (2002). Gender differences in factors
influencing insulin resistance in elderly hyperlipemic non-diabetic subjects.
Cardiovascular Diabetology, 1, 4–13.
Martínez-López, S., Sarriá, B., Baeza, G., Mateos, R., & Bravo-Clemente, L. (2014). Pharma-
cokinetics of caffeine and its metabolites in plasma and urine after consuming a sol-
uble green/roasted coffee blend by healthy subjects. Food Research International, 64,
125–133.
Mateos, R., Gómez-Juaristi, M., Martínez-López, S., Baeza, G., Amigo-Benavent, M., Sarriá,
B., & Bravo-Clemente, L. (2015). Bioavailability of hydroxycinnamate derivatives after
consuming a soluble green/roasted coffee blend by healthy humans. Poster presented in
the international. EuroFoodChem XVIII, Madrid: Congress.
Matthews, D. R., Hosker, J. P., Rudenski, A. S., Naylor, B. A., Treacher, D. F., & Turner, R. C.
(1985). Homeostasis model assessment: Insulin resistance and β-cell function from
fasting plasma glucose and insulin concentrations in man. Diabetologia, 28, 412–419.
McCarty, M. F. (2005). A chlorogenic acid-induced increase in GLP-1 production may me-
diate the impact of heavy coffee consumption on diabetes risk. Medical Hypotheses,
64, 48–53.
Meier, J. J., Hücking, K., Holst, J. J., Deacon, C. F., Schmiegel, W. H., & Nauck, M. A. (2001).
Reduced insulinotropic effect of gastric inhibitory polypeptide in first-degree rela-
tives of patients with type 2 diabetes. Diabetes, 50, 2497–2504.
1028 B. Sarriá et al. / Food Research International 89 (2016) 1023–1028
Nolan, J. J., & Færch, K. (2012). Estimating insulin sensitivity and beta cell function: Per-
spectives from the modern pandemics of obesity and type 2 diabetes. Diabetologia,
55, 2863–2867.
Ohnaka, K., Ikeda, M., Maki, T., Okada, T., Shimazoe, T., Adachi, M., et al. (2012). Effects of
16-week consumption of caffeinated and decaffeinated instant coffee on glucose me-
tabolism in a randomized controlled trial. Journal of Nutritional Biochemistry, 207426.
Olthof, M. R., Hollman, P. C., & Katan, M. B. (2001). Chlorogenic acid and caffeic acid are
absorbed in humans. Journal of Nutrition, 131, 66–71.
Olthof, M. R., van Dijk, A. E., Deacon, C. F., Heine, R. J., & van Dam, R. M. (2011). Acute ef-
fects of decaffeinated coffee and the major coffee components chlorogenic acid and
trigonelline on incretin hormones. Nutrition and Metabolism, 8, 10.
Pereira, M. A., Parker, E. D., & Folsom, A. R. (2006). Coffee consumption and risk of type 2
diabetes mellitus: An 11-year prospective study of 28,812 postmenopausal women.
Archives of Internal Medicine, 166, 1311–1316.
Perrone, D., Farah, A., & Donangelo, C. M. (2012). Influence of coffee roasting on the incor-
poration of phenolic compounds into melanoidins and their relationship with antiox-
idant activity of the brew. Journal of Agricultural and Food Chemistry, 60, 4265–4275.
Pham, N. M., Nanri, A., Kochi, T., Kuwahara, K., Tsuruoka, H., Kurotani, K., et al. (2014). Cof-
fee and green tea consumption is associated with insulin resistance in Japanese
adults. Metabolism, 63, 400–408.
Rebello, S. A., Chen, C. H., Naidoo, N., Xu, W., Lee, J., Chia, K. S., et al. (2011). Coffee and tea
consumption in relation to inflammation and basal glucose metabolism in a multi-
ethnic Asian population: A cross-sectional study. Nutrition Journal, 2 (10:61).
Rustenbeck, I., Lier-Glaubit, Z. V., Willenborg, M., Eggert, F., Engelhardt, U., & Jörns, A.
(2014). Effect of chronic coffee consumption on weight gain and glycaemia in a
mouse model of obesity and type 2 diabetes. Nutrition and Diabetes, 4, e123.
Schenker, S., Heinemann, C., Huber, M., Pompizzi, R., Perren, R., & Escher, R. (2002). Im-
pact of roasting conditions on the formation of aroma compounds in coffee beans.
Journal of Food Science, 67, 60–66.
Somporn, C., Kamtuo, A., Theerakulpisut, P., & Siriamornpun, S. (2011). Effects of roasting
degree on radical scavenging activity, phenolics and volatile compounds of Arabica
coffee beans (Coffea Arabica L. cv. Carimor). International Journal of Food Science and
Technology, 46, 2287–2296.
Tunnicliffe, J. M., & Shearer, J. (2008). Coffee, glucose homeostasis, and insulin resistance:
Physiological mechanisms and mediators. Applied Physiology, Nutrition, and Metabolism,
33, 1290–1300.
Wang, Q. S., Zhou, H., Yaung, D., Ma, L., & Geng, W. (2010). www.bio-rad.com/webroot/
web/pdf/lsr/literature/Bulletin_5985A.pdf
Wedick, N. M., Brennan, A. M., Sun, Q., Hu, F. B., Mantzoros, C. S., & van Dam, R. M. (2011).
Effects of caffeinated and decaffeinated coffee on biological risk factors for type 2 di-
abetes: A randomized controlled trial. Nutrition Journal, 13 (10:93).
Wu, T., Willet, W. C., Hankinson, S. E., & Giovannnucci, E. (2005). Caffeinated coffee, decaf-
feinated coffee, and caffeine in relation to plasma C-peptide levels: A marker of insu-
lin secretion in U.S. women. Diabetes Care, 28, 1390–1396.
Yabe, D., Kuroe, A., Watanabe, K., Iwasaki, M., Hamasaki, A., Hamamoto, Y., et al. (2015).
Early phase glucagon and insulin secretory abnormalities, but not incretin secretion,
are similarly responsible for hyperglycemia after ingestion of nutrients. Journal of
Diabetes and its Complications, 29, 413–421.
Yamaji, T., Mizoue, T., Tabata, S., Ogawa, S., Yamaguchi, K., Shimizu, E., et al. (2004). Coffee
consumption and glucose tolerance status in middle-aged Japanese men. Diabetologia,
47, 2145–2151.
Zhang, Y., Lee, E. T., Cowan, L. D., Fabsitz, R. R., & Howard, B. V. (2011). Coffee consumption
and the incidence of type 2 diabetes in men and women with normal glucose toler-
ance: The Strong Heart Study. Nutrition, Metabolism, and Cardiovascular Diseases, 21,
418–423.