The main findings of the present study are that the regular consumption of green/roasted coffee beverage induced reductions in body weight and fat% in both hypercholesterolemic and normocholesterolemic subjects; however, only in the hypercholesterolemics WC and WC ratios slightly decreased. Thus, this study supports that the effects of coffee are greater in individuals with higher cardiovascular risk and are modified by inter-individual variability. When the antiobesity and cardiometabolic effects of coffee in the hypercholesterolemic group were further studied through factor analysis, only the main factor that included as the variables with higher coefficients WC, DBP, SBP, body weight, WC/hip and WC/height ratios, followed by TG and ghrelin with a negative coefficient, decreased after coffee intake. Therefore, the coffee blend may reduce cardiovascular risk in hypercholesterolemic subjects mainly by inducing changes related to abdominal obesity, blood pressure and body weight.
4.1. Effects of the Coffee Blend on Anthropometric Measurements in Normocholesterolemic and Hypercholesterolemic Subjects
It is important to note that both the normocholesterolemic and hypercholesterolemic subjects had similar body weight at baseline and after the coffee intervention, always within the normoweight range (BMI 18.5–24.9 kg/m2
). However, the hypercholesterolemics presented higher body fat% (2.6% higher) and WC (6.4 cm higher) at baseline than normocholesterolemics. The regular consumption of a green/roasted coffee beverage induced similar reductions in body weight and body fat%, in both groups. In contrast, WC and WC/ratios tended to decrease (not reaching statistical significance) only in the hypercholesterolemics. Thus, the present study supports that hydroxycinnamic acid-rich foods have greater effectiveness in subjects with higher cardiovascular risk [22
]. The regular intake of the coffee blend here studied, which provided 445 mg of CGA/day, lowered WC by 1.2 cm in the hypercolesterolemics (contrary to the normocholesterolemics, with a WC increase of 0.5 cm), a reduction that is higher to that described by Watanabe et al. [6
] in overweight men and women (0.7 cm) after consuming an instant coffee containing high amounts of CGA (369 mg CGA/serving/day) for 12 weeks, versus the control coffee (35 mg CGA/serving/day). Interestingly, in Watanabe [6
], the visceral fat area, measured by computed tomography scanning, total abdominal fat area and body weight also decreased significantly in the overweight men and women. Attending to a recent meta-analysis [23
] on coffee and obesity based on epidemiologic studies, higher coffee intake was associated, although statistically non-significant, with reduced risk of central obesity as defined by WC in men, but not in women, showing a non-significant but positive association. It would have been interesting to look into the differences in WC considering gender in the present study, however, the number of volunteers in the present work was too low to separate men and women subgroups. In this context, acute effects of coffee on postprandial glucose and insulin have also been shown to be modified by sex and overweight/obese status [24
When different factors related to obesity (endocrine, metabolic and nutritional factors) were evaluated to see their relative contribution to the disease in an obese population, it was shown that dietary intake, especially fat intake, was the most important factor contributing to obesity [24
]. In addition, it was concluded that abdominal adiposity distribution may induce hormonal changes and that these two factors combine to favor metabolic variations in plasma lipids, so that fat distribution may condition alterations in lipid metabolism [20
]. Accordingly, the changes in abdominal adiposity observed in the hypercholesterolemics after coffee consumption, as well as the changes in resistin, leptin, and PAI-1 levels [13
], could be associated with the effects on blood lipids, T-C, LDL-C and TG levels, which were significantly reduced only in the hypercholesterolemics [14
]. In contrast, there were no changes in the fat distribution or the aforementioned parameters in the normocholesterolemics, in agreement with previous studies in healthy normoweight volunteers [11
The decrease in body weight observed in both study groups may be mainly attributed to the content of CGA in the coffee blend, according to a recent review by Gökcen and Sanlier [10
] among others [8
]. Green coffee extract may regulate adipogenesis, as well as genes and proteins associated with lipid metabolism in white adipose tissue and liver, leading to loss of body weight. The slimming role of CGA is also supported by Roshan et al. [27
], who described that decaffeinated green coffee supplementation could decrease weight about twice as much as the placebo by inhibiting adipogenesis and by contributing to controlling appetite, which was evaluated through a scored questionnaire. This outcome is in agreement with the present study, which is among the first to point to coffee controlling weight through the appetite-related hormone ghrelin. Similar results were observed after treatment with a high dose of tea extract containing epigallocatechin gallate, which reduced weight and decreased ghrelin levels in women with central obesity compared with a control group [28
All the aforementioned effects may be associated with the consumption of three cups of the green/roasted coffee blend per day, which represents a moderate, realistic coffee consumption rate. It is important to bear in mind that the intake of certain polyphenol and methylxanthine-rich foods were restricted through the study, and energy intake and physical activity were not modified during the study. Therefore, the weight and fat loss here described may be related to the green/roasted coffee blend. Two more components of the beverage might be responsible for the observed effect on body weight (reviewed in [26
]): (i) caffeine, increasing metabolic rate, energy expenditure, lipid oxidation, and with lipolytic and thermogenic activities; (ii) soluble dietary fiber, forming viscous solutions in the gastrointestinal tract that delay transit time and hinder the digestion and absorption of food nutrients and the reabsorption of bile salts.
4.2. Factor Analysis
To further understand the antiobesity and cardiometabolic effects in the hypercholesterolemic group observed in the present work and in previous studies after the intake of the coffee blend [14
], factor analysis was carried out using the baseline values of the variables that showed significant changes due to coffee intake (body weight, body fat percentage, glucose, PAI-1, resistin, leptin, ghrelin, DBP and SBP), both a significant coffee and coffee x group interaction (glucagon, T-C, TG, LDL-C and HOMA-IR) or only a significant coffee x group interaction (WC, WC/hip and WC/height). In addition to the aforementioned variables, energy and lipid intake, which are among the most important factors contributing to obesity [20
], were also included in the analysis, as well as insulin and HDL-C due to the close relationship with the other parameters.
After factor analysis, three factors emerged. F1, the main factor, accounted for 25.5% of the variance and included as the variables with higher coefficients WC, DBP, SBP, body weight, WC/hip and WC/height ratios, followed by TG and ghrelin with a negative coefficient. Therefore, within F1 were modifiable risk parameters that greatly contribute to mortality and major cardiovascular disease. Some of the associations of parameters observed within F1 are well known. Numerous studies have shown that overall obesity, defined by BMI (overall obesity) or WC (central obesity) is associated with hypertension [30
]. Hypertension has been closely associated with obesity-related parameters not only in adults but also in adolescents, so that normal weight and overweight have shown a positive association with healthy and high BP, respectively [32
]. The association between body weight and levels of TG is also in accordance with previous studies in grown-ups [33
]. Alterations in levels of TG have been related to markers of central obesity, measured with simple anthropometric indices (WC and WC/ratios) and with advanced computed tomography techniques [34
]. In contrast, the relationship between body weight and ghrelin has been less studied. It has been reported that ghrelin levels decrease in humans with obesity and metabolic syndrome and increase during weight loss, suggesting that the hormone plays a role in energy adaptation [36
]. In agreement, in patients with anorexia nervosa, weight gain decreases elevated plasma ghrelin concentrations [37
]. The present study supports that ghrelin is inversely related not only with body weight but also with WC and WC ratios (abdominal adiposity).
With respect to F2, which accounted for 13.13% of the variance, the variables that showed higher loadings were related to glucose metabolism: HOMA-IR, resistin, glucose, insulin, leptin and PAI-1. In addition to the well-known markers of glucose metabolism: HOMA-IR, glucose and insulin, resistin, leptin and PAI-1 are also related to type 2 diabetes (T2D). In fact, elevated levels of resistin have been associated with a higher risk of T2D [38
]. Similarly, leptin has also been associated with T2D; in addition to the influence of this hormone on insulin and glucagon levels, leptin might also indirectly regulate glucose metabolism by altering levels of other hormones that are related to glucose metabolism [39
]. Likewise, many studies support the implication of PAI-1 in the development of T2D, a recent meta-analysis of observational studies supports the link between PAI-1 and T2D [40
The third main factor, F3, accounted for 11.79% of the variance and contained variables related to dietary intake. The highest loadings were energy and lipid intake followed by glucagon. According to Garaulet et al. [20
], diet is the factor that explains most variability in obese patients, more than 40% is related to dietary habits, and within the diet factor, fat intake followed by energy intake are the most important factors contributing to obesity [24
]. Accordingly, within F3, the higher loads corresponded to energy (0.894) and lipid intake (0.731). On the other hand, the association of glucagon and intake is well known, as glucagon suppresses appetite and also modulates lipid metabolism, and by the two routes reduces body weight and adiposity. Moreover, glucagon may promote weight loss by stimulation of energy expenditure and thermogenesis [41
When the three factors were comparatively studied, before and after the coffee intervention (Table 4
), a general decrease was observed, although only the main factor, F1, was reduced significantly, in contrast to F2 and F3 that slightly decreased without reaching the level of statistical significance. Attending to the present results, it may be inferred that in hypercholesterolemic subjects, the regular consumption of the green/roasted coffee blend mainly induces changes in abdominal obesity, blood pressure and body weight, followed by triglycerides and inversely with ghrelin. The results here presented are supported by a recent study in humans that shows that CGA may significantly decrease WC and abdominal fat, as well as SBP and appetite, and also may reduce weight marginally [27
Limitations of this study: both the healthy and cardiovascular risk study population were normoweight. The intake of the coffee was not monitored using a coffee intake biomarker. Interindividual variability, particularly gender, was not addressed. The biological pathways associated with the main bioactive components in coffee have been described as the main pathways responsible for the reduction of cardiovascular risk; however, it cannot be discarded that other biological pathways may have been involved. Strengths of the study: the participants were homogenous regarding their lifestyle characteristics, education level and social status. Volunteers’ physical activity and dietary habits were controlled through questionnaires and did not show changes along the study. In addition, they did not smoke. Thus, several characteristics that may influence the individual response were considered. The coffee blend was well-accepted by the volunteers and self-reported compliance was high. Regarding the coffee product, sources of variability such as coffee preparation and adding sugar or dairy products were controlled by instructing the volunteers to prepare the coffee beverage only with water.