To investigate the effects of different coffee amounts on dietary intake and appetite feelings in normal-weight and overweight/obese individuals.
To investigate the effects of different coffee amounts on dietary intake and appetite feelings in normal-weight and overweight/obese individuals.
Thirty-three volunteers (16 normal-weight, 17 overweight/obese) participated in three trials: they consumed a standard breakfast along with 200 ml of either coffee with 3 or 6 mg caffeine/kg body weight (Coffee 3 and Coffee 6, respectively), or water. At fasting and at standard time points for the 3 h following breakfast/drink consumption participants recorded their appetite feelings on visual analogue scales. At 180 min, participants consumed an ad libitum meal and the next day they recalled their food intake during the experimental day.
A significant intervention effect was found for the energy intake of the ad libitum meal (P = 0.05) and of the whole day (P = 0.02) only in overweight/obese individuals. Specifically, Coffee 6 resulted in a reduced energy intake during the ad libitum meal compared to Coffee 3 (P = 0.03) and in the total day compared to both water (P = 0.04) and Coffee 3 (P = 0.008). No effect was observed for the appetite feelings.
A moderate coffee amount can effectively reduce energy intake in the following meal and in the total day compared to lower or no coffee intake in overweight/obese participants.
Coffee has received special attention the last decade, as many health benefits have been found to be associated with its consumption . However, limited research has been undertaken with regard to the effect of coffee and/or caffeine consumption on energy intake and body weight control, despite the anecdotal perception of the general public that coffee exerts an appetite suppressing role.
Cross-sectional studies indicate that coffee consumers have lower body mass index (BMI) compared to individuals who do not consume coffee [2, 3]. Epidemiological evidence also indicates a negative association between coffee and/or caffeine consumption, and body weight gain in the long term [4, 5]. However, a study in elderly individuals from the Mediterranean islands found a strong positive dose response relationship of coffee drinking and BMI that was mediated by physical inactivity and smoking habits .
Limited experimental research has been undertaken with regard to the effect of coffee on energy balance and especially energy intake. The majority of the clinical studies have evaluated the effects of caffeine, a coffee constituent, and not those of coffee (which is a mixture of chemical substances) on energy expenditure. According to a recent meta-analysis, including studies dated back to 80s, a dose-response effect of caffeine on 24 h energy expenditure exists with a mean increase of 0.1 kcal mg−1 . Caffeine has also been studied, but in a lesser extent, in relation to its effect on energy intake and appetite related feelings. In specific, Tremblay et al. observed a lower energy intake in the following meal 30 min after 300 mg of caffeine administration only in men . The combination of 800 mg of caffeine with 6 g of red pepper in the experimental daily dietary plan resulted in ∼880 kcal lower daily energy intake compared to the no treatment , whereas 50 and 100 mg of caffeine amplified the hunger-suppressing effect of nicotine .
In relation to coffee, a recent study of our group found no acute effect of coffee containing 3 mg of caffeine/kg of body weight on energy intake and minor changes on appetite-related feelings in nonobese participants compared to decaffeinated coffee and water . It is unknown, however, whether higher amounts of caffeinated coffee may have a different or a dose-response effect on normal-weight or overweight/obese individuals. Therefore, we conducted a randomized, crossover, controlled study to investigate the acute effects of different amounts of caffeinated coffee on dietary intake and appetite feelings, in apparently healthy normal-weight and overweight/obese volunteers.
Seventeen normal-weight (of whom nine females) and seventeen overweight/obese (of whom eight females) individuals were recruited for this study. One normal-weight volunteer dropped out without giving a reason after her participation in her first experimental trial, leaving 16 normal-weight individuals participating in the study. Participants' characteristics are presented in Table 1. All of them were apparently healthy and habitual coffee drinkers (≥1 cup of coffee/day). Smokers, athletes, restrainers [total score in the Dutch eating behavior questionnaire  and in the Restraint subscale >3], individuals on a diet, those taking medication or having a chronic or acute disease were excluded from the study. All volunteers completed a semiquantitative food frequency questionnaire (FFQ) for the assessment of their usual caffeine intake over the previous year (including several types of caffeinated and decaffeinated coffee, tea, caffeinated soft drinks, energy drinks and chocolate), a physical activity questionnaire  and a medical history questionnaire. They were not informed about the true purpose of the study to avoid responses associated with preconceived cognitions regarding the effect of coffee on appetite. All of them gave informed written consent and the experimental protocol was approved by the Harokopio University Ethics Committee. The study was undertaken at the Metabolic Unit of the Department of Nutrition and Dietetics of Harokopio University, from March to June 2011.
|Characteristics||Normal-weight participants (N = 16)||Overweight/obese participants (N = 17)||P value|
|Age, y||25.4 ± 6.1||29.2 ± 7.7||0.01|
|Weight, kg||64.0 ± 8.2||87.7 ± 18.6||<0.001|
|Body mass index, kg m−2||21.3 ± 1.9||30.0 ± 3.9||<0.001|
|Physical activity level||1.3 ± 0.2||1.4 ± 0.3||0.07|
|Energy intake in previous day, kcal day−1||2,360 ± 788||2,244 ± 929||0.51|
|Habitual caffeine intake, mg day−1||229 ± 122||329 ± 115||<0.001|
|Caffeine in Coffee 3 interventiona, mg||192 ± 26||263 ± 57||<0.001|
|Coffee in Coffee 3 interventiona, g||5.0 ± 0.7||6.9 ± 1.5||<0.001|
|Caffeine in Coffee 6 interventionb, mg||383 ± 51||526 ± 114||<0.001|
|Coffee in Coffee 6 interventionb, g||10.0 ± 1.3||13.7 ± 3.0||<0.001|
The study had a randomized, crossover design. Each volunteer took part in three daily trials, separated by, at least, 1-week time. Female participants were on the follicular phase of their menstrual cycle during the experiments to avoid variability of their energy intake due to the different phases of the menstrual cycle . Participants' body weight and height were measured on a levelled platform scale and a wall-mounted stadiometer, to the nearest 0.5 kg and 0.5 cm, respectively and BMI was calculated as weight (kg)/height2 (m2). The cut-off point of BMI 25 kg m−2 was taken to categorize participants as normal-weight (<25 kg m−2) and overweight/obese (≥25 kg m−2) .
The day prior each experiment volunteers had to abstain from any caffeine source, alcohol and physical exercise, to sleep at least 7 h and to come at the lab after an overnight fast of 10 h. Furthermore, for the day preceding the second and third experimental day they had to consume similar foods with the day preceding the first experimental day at about the same quantities and hours daily. Participants came to the lab at 9.00 am and they consumed a standardized breakfast snack along with one of the three experimental drinks. The breakfast snack consisted of one slice of white bread, 5 g of butter, and 10 g of white sugar, providing 142 kcal (6.5% of energy from proteins, 62.5% from carbohydrates and 31.0% from lipids). The experimental drinks were 200 ml of either (a) instant coffee with 3 mg caffeine/kg body weight (Coffee 3), (b) instant coffee with 6 mg caffeine/kg body weight (Coffee 6) or (c) water (Control). Volunteers had to consume the breakfast and the drink within 5 min. To record their appetite feelings they completed three 10-cm visual analogue scales (VAS) related to hunger, satiety and desire to eat, in the fasted state (−15 min), immediately after breakfast and drink consumption (0 min) and at the subsequent 15, 30, 60, 90, 120, 150, and 180 min. During this 3-h period participants were interviewed about the previous day's dietary intake using the 24-h recall method. Blood pressure and heart rate were measured at baseline, 60 and 180 min. Blood pressure was measured in the right forearm of the subject in a sitting position at rest using a standard mercury sphygmomanometer. Heart rate was measured at the wrist.
After the 3-h period participants were offered an ad libitum lunch meal from a buffet, consisting of common Greek diet foods (pasta, tomato sauce, beef, salad, cheese, yogurt, fruit, juice, and chocolate). They consumed as much food as they desired until they felt satiated, within 30 min. Researchers weighted any food participants placed on their plate as well as any leftovers, to record their dietary intake. To monitor a possible decrease in food consumption due to a habituation and a reduced liking of the food, another VAS for meal liking was completed by the participants each time they finished their experimental lunch meal. Finally, the day following the experiment, participants were asked to report their food and fluid intake for the rest of the experimental day—after leaving the lab until they slept—through a telephone recall.
For the analysis of the dietary intake, in terms of energy and macronutrient intake, the Nutritionist Pro, version 2.2 software (Axxya Systems-Nutritionist Pro, Stafford, TX) was used. The software food database was expanded by the analyses of traditional Greek foods and recipes as well as information provided by food companies. For the determination of participants' usual caffeine intake the caffeine content from the various foods and beverages included in the FFQ was calculated using information from the United States Department of Agriculture National Nutrient Database for Standard Reference  as well as the food industry. The physical activity questionnaire was analyzed for participant's energy expenditure of the week prior the completion of the questionnaire and the Physical Activity Level (the ratio of total energy expenditure to basal metabolic rate) was calculated for each participant.
The experimental design was 3 (intervention groups) × 2 (blocks, i.e., normal-weight vs. overweight/obese). Thus, to adequately detect an increase of 0.5 SD in appetite ratings score, a sample of eight participants per intervention group, i.e., a total of 24 participants was considered necessary to achieve 80% statistical power at 5% significance level of two-sided hypotheses. Data were analyzed using the Statistical Package for Social Sciences statistical software package version 18 (SPSS, Chicago, IL). Means for baseline characteristics were compared using the Independed-samples T test and the X2 test. Repeated measures analysis of covariance (participant code as a covariate) was used for the comparisons of the response curves of VAS ratings, by testing for an intervention effect and a time × intervention interaction. The mean values of energy and macronutrient intake during the ad libitum meal and the total day (sum of breakfast, ad libitum and rest of the day intakes) between the three trials were compared through univariate analysis of variance and further adjustment was made for the previous day's energy intake or habitual caffeine intake. Bonferroni post-hoc test was used to compare one intervention to another. Furthermore, blood pressure and heart rate values were compared within interventions at the three time points (baseline, 60 and 180 min), through one-way analysis of variance. All tests were two-sided and differences were considered significant at P < 0.05. Data are presented as mean ± standard deviation (SD), unless otherwise stated.
Normal-weight participants consumed comparable energy in the ad libitum meal in the three interventions (Figure 1A). The same was also evident for their total daily energy intake (Figure 1B). However, in overweight/obese participants, significant differences were revealed in the energy intake during the ad libitum meal (P = 0.05) and the whole day (P = 0.02) (Figure 1C,D, respectively). In specific, the consumption of Coffee 3 resulted in an elevated energy intake of 233 kcal at the ad libitum meal compared to the consumption of water (P = 0.06), whereas Coffee 6 consumption presented no difference compared to water consumption. Furthermore, Coffee 3 ingestion resulted in an increased energy intake of 281 kcal at the ad libitum meal compared to Coffee 6 ingestion (P = 0.03). The consumption of Coffee 6 resulted in 550 kcal lower total daily intake compared to the consumption of water (P = 0.04) whereas Coffee 3 consumption presented no difference compared to water consumption. Coffee 6 induced also a 725 kcal lower total energy intake compared to the consumption of Coffee 3 (P = 0.008). The results remained unchanged when the energy of the previous day or the habitual caffeine intake were used as covariates in the analysis.
Macronutrient analysis revealed a statistical significant difference in carbohydrates consumption in the total day among the three interventions in overweight/obese volunteers (P = 0.008) (data not shown). Coffee 6 consumption induced a significantly higher percentage of energy from carbohydrates compared to water ingestion (7.7%) (P = 0.002). Similar results occurred when the energy of the previous day or the habitual caffeine intake were used as covariates.
No significant intervention effect or a time × intervention interaction was found for the ratings in the three VAS scales in normal-weight or overweight/obese participants (Figure 2). However, in overweight/obese participants only, post hoc analysis revealed higher satiety ratings after Coffee 3 consumption at 15 and 60 min compared to water consumption (P = 0.02 and 0.04, respectively) and at 15 min compared to Coffee 6 consumption (P = 0.03) (Figure 2E). The VAS ratings for meal liking were not different between the interventions.
Data for blood pressure and heart rate at baseline, 60 and 180 min following breakfast/beverage ingestion are presented in Supporting Information Table 1. No statistical significant difference was observed within the three interventions neither in normal-weight nor in overweight/obese participants.
Coffee is one of the most frequently consumed beverages . This is the first study that examined a potential dose effect of coffee on energy intake and appetite related feelings in normal-weight and overweight/obese participants. We found that a moderate coffee intake results in lower energy consumption in the subsequent meal as well as in the total day, compared to lower or no intake, only in overweight/obese participants. Thus, if a minimum appetite suppressant amount exists, this is different between overweight/obese and normal-weight individuals; otherwise, a cut-off amount exists only for overweight/obese people. Potential mechanisms that may explain these hypotheses could be related to the appetite hormones and glucose metabolism markers . It is known that overweight/obese individuals have different metabolic and/or hormonal profiles compared to normal-weight individuals [19, 20] and thus, they may respond differently to dietary factors, like coffee, that can affect appetite hormones and glucose metabolism [21, 22]. Another potential explanation might be a potentially different rate of caffeine metabolism between the two groups, if the results are mainly attributed to caffeine . It remains to be investigated whether higher amounts of coffee and/or caffeine may alter the energy intake of normal-weight individuals, although it is also of question whether higher coffee amounts could be easily drunk as a single dose. Further research is also necessary to evaluate whether the effects observed after Coffee 6 ingestion in the current study, still exist when coffee is administered in two to four doses during the day.
The present study does not indicate a dose-response relationship between coffee drinking and energy intake or appetite-related feelings. A potential explanation might be that by increasing the amount of coffee there is also a potential alteration of the equilibrium between the actions of the different coffee compounds. Thus, it could be hypothesized that when a small coffee amount is consumed, i.e., Coffee 3, then the dominant effect comes from the compounds that promote energy intake, but when a higher coffee amount is consumed, i.e., Coffee 6 then the balance is changed in favor of the compounds that prevent energy consumption. However, further research is necessary to confirm such a hypothesis and to examine which compounds may promote energy intake, as caffeine, chlorogenic acids and mannooligosaccharides are known to promote a negative energy balance and/or induce a reduction in body weight .
The crossover, randomized design of the study is one of the study strengths. The small coffee amount given in the experiment represented the usual intake and the experimental conditions, with the breakfast accompanying coffee consumption in the morning, imitated everyday life conditions. We included men and women and we controlled for women's menstrual cycle to extend our results to both sexes. On the other hand, the results cannot be generalized to other population groups, i.e., nonhealthy, smokers, noncoffee consumers, as the volunteers participated in the study were selected according to specific criteria. Lastly, Coffee 6 was not easily consumed by most of the volunteers, although no major adverse effects were mentioned.
In conclusion, there was no effect of coffee on dietary intake and appetite related feelings in normal-weight individuals; however, in overweight/obese coffee was found to exert an effect as consumption of a moderate amount in the morning, providing 6 mg of caffeine/kg body weight (equivalent to two to four cups), significantly reduced energy intake in the lunch, compared to a lower or no intake, and this effect was maintained during the rest of the day. Further research is needed to elucidate responsible mechanisms and the potential of using coffee drinking as an eating behavior included in weight reducing diets.
We would like to thank Mrs Antigoni Tsiafitsa for her technical assistance.