Chronic flavanol-rich cocoa powder supplementation reduces body fat mass in endurance athletes by modifying the follistatin/ myostatin ratio and leptin levels Jose Ángel García-Merino,a Diego Moreno-Pérez,b Beatriz de Lucas,a Maria Gregoria Montalvo-Lominchar,a Elsa Muñoz,c Lara Sánchez,d Fernando Naclerio,e Karen Marlene Herrera-Rocha,f Martha Rocío Moreno-Jiménez,f Nuria Elisabeth Rocha-Guzmánf and Mar Larrosa *a a MAS microbiota group, Faculty of Biomedical and Health Sciences, Universidad Europea de Madrid, Spain. E-mail: mar.larrosa@universidadeuropea.es b Departamento de Educación, Métodos de Investigación y Evaluación, Universidad Pontificia de Comillas, ICAI-ICADE, Cantoblanco, Madrid 28015, Spain c Medical Service, Universidad Europea de Madrid, Spain d Faculty of Biomedical and Health Sciences, Universidad Europea de Madrid, Spain e Department of Life and Sports Sciences, University of Greenwich, Eltham SE9 2TB, UK f Grupo de Investigación en Alimentos Funcionales y Nutracéuticos. Unidad de Posgrado, Investigación y Desarrollo Tecnológico. TecNM/Instituto Tecnológico de Durango, Durango, Mexico Abstract Flavanols-rich cocoa has positive effects on lipid metabolism and might enhance the performance of athletes through an improvement in their body composition. To test this hypothesis a placebo-controlled intervention study in training endurance athletes who received 5 g of cocoa daily (425 mg of flavanols) for 10 weeks was performed. Dietary intake, body composition, exercise performance and plasma levels of follistatin, myostatin and leptin were measured. Cocoa intake significantly reduced body fat percentage (p= 0.020), specifically in the trunk (p = 0.022), visceral area (p = 0.034) and lower limbs (p = 0.004). The reduction in body fat mass was accompanied by an increase in plasma follistatin and a decrease in leptin, while myostatin levels remained unchanged. The intake of cocoa reduced the percentage of body fat of athletes, without any impact on athletes’ performance. The change in fat body composition did not improve athletes’ performance. Introduction A dietary supplement is a manufactured product that is taken by mouth, containing a dietary ingredient, and that claims to improve overall health status. Many athletes (between 60 to 90%) use dietary supplements in an attempt to enhance physical performance, improve body composition or promote recovery from exercise, and these can take several forms including energy drinks, caffeine, vitamins/minerals and amino acids/proteins, among others.1 Cocoa beans from the cocoa tree, Theobroma cacao, are rich in polyphenols (flavonoids) and more specifically in flavanols (procyanidins and catechins) that confer antioxidant and cardioprotective properties.2 Flavanols may play a role in preventing obesity by increasing lipolysis and inducing changes in the lipid profile, and by decreasing the levels of low-density lipoproteins, triglycerides, total cholesterol and inflammatory markers.3,4 On account of its numerous biological effects and its safety profile, cocoa supplementation might improve the performance of and recovery from exercise and counteract some of the physiological effects associated with exercise such as oxidative stress. Indeed, combining cocoa supplementation with exercise has shown promising effects to counteract the exercise- induced oxidative stress and to improve lipid and carbohydrate metabolism in healthy and overweight subjects.5 In addition to flavanol monomers catechin and epicatechin, cocoa also contains oligomeric procyanidins, which together seem to be responsible for the observed positive effects of cocoa. However, it is very difficult to draw general conclusions on the effects of cocoa, because in many of studies the content of these compounds is not specified5 and the composition of cocoa bioactive compounds varies considerably depending on the cocoa genotype, the fermentation degree and drying processes,6 generating cocoa products with very diverse composition.7,8 In a pilot study in six humans, administration of 25 mg day−1 of epicatechin significantly elevated the plasma ratio of follistatin/myostatin and also increased handgrip strength.9 Myostatin is a myokine produced in muscle and adipose tissue that is involved in muscle growth and metabolism, and its activation produces muscle atrophy and promotes fat accumulation.10 Conversely, follistatin promotes the adipose tissue browning and decreases the abdominal fat content in mice, and also counteracts the myostatin blockade of muscle growth.11,12 The loss of white adipose tissue could be beneficial for the body composition of athletes and their physical performance, as it has been described that a decrease in fat mass positively correlates with athletes’ performance13 and predicts race time in recreational marathoners.14 In this regard, we sought to examine whether the chronic intake of a high-flavanols characterized cocoa modifies the body composition of endurance athletes through the alteration of the circulating levels of adipokines, as a primary outcome, and whether it had an impact on exercise performance. Materials and methods Materials Hexane, acetone, formic acid, acetonitrile, methanol solvents and reagents were from J.T. Baker (JT Baker Chemical Company, Phillispburg, NJ, USA). Pure chemical standards were obtained from Sigma-Aldrich (St Louis, MI, USA). Highflavonol cocoa (83 mg g−1) was purchased in the United Kingdom market. Maltodextrin supplement was acquired from Prozis (Esposende, Portugal). Polyphenolic characterization of cocoa supplement As the flavonoid content in cocoa and derived products varies significantly between different manufacturers, we analyzed seven different cocoa powders marketed in Spain and the United Kingdom using the protocol established by Alañón et al.,15 and the cocoa with the highest flavanol content was selected to perform the study. Briefly, cocoa powders were defatted with a series of successive washes in hexane (1 : 10 w/v) in an ultrasonic bath, followed by centrifugation at 10 000 rpm for 10 min at 4 °C to obtain the residual defatted pellets. This procedure was repeated three times. The pellets were –air-dried for 24 h to remove the residual organic solvent, freeze-dried, and then submitted to three repeated extractions with 70% acetone (1 : 10 w/v) in an ultrasonic bath. Supernatants were recovered after centrifugation at 10 000 rpm for 10 min at 25 °C, and concentrated in a roto-evaporator (Büchi Labortechnik AG, Flawil Switzerland). Samples were kept at −80 °C until analysis. Analyses were carried out by ultra-high-performance liquid chromatography (Waters Corp., Milford, MA, USA) coupled with tandem photodiode array detection and electrospray ionization (ESI) triple quadrupole mass spectrometry (MS/MS) (Xevo TQS, Waters Corp., Wexford, Ireland). The chromatography system consisted of a sample manager (6 °C) and a quaternary solvent manager. The polyphenolic compounds were separated on a C18, 150 mm × 2.1 mm, 1.7 µm column (Agilent Technologies, Santa Clara, CA, USA), operated at 30 °C. The elution profile included two solvents: acidified Milli Q water with 7.5 mM formic acid (solvent A) and acetonitrile (solvent B): initial – 3% B, 1.88 min gradient to 9% B; 5.66 min gradient to 16% B; 16.90 min gradient to 50% B; 19.62 min gradient to 3% B; 20.0 min 3% B for column stabilization at a flow rate of 210 µL min−1. Ionization was carried out using methanol as co-solvent with 0.1% formic acid (v/v) at a flow of 5 µL min−1 with the use of an isocratic solvent manager (Waters Corp.). Multiple reactions ionization mode was used for MS/MS assays. ESI operating conditions for the negative ionization mode were as follows: capillary voltage 2.5 kV, desolvation temperature 400 °C, source temperature 150 °C, desolvation gas flow 800 L h−1 and cone gas flow 150 L h−1, collision gas flow 0.13 mL min−1, MS mode collision energy 5.0 eV and MS/MS mode collision energy 20.0 eV. For the identification and quantification of the polyphenolic profile, a mixture of standards (20ng µL−1) was used to determine retention time and m/z values and MS/MS transitions. Multiple reaction monitoring mode was used for samples and tandards. The liquid chromatography and tandem Xevo TQ-S triple quadrupole MS control and data processing was done using MassLinx v. 4.1 Software (Waters Corp.). The cocoa polymerization degree (mDP) was obtained following acid-catalysis in the presence of phloroglucinol using the method previously described by Kennedy and Jones.16 Chromatographic conditions were the same as described above. Experimental design, dietary supplementation and control of the intervention compliance This study was a randomized, parallel-group placebo-controlled trial. The Research Ethics Committee on Drug of the Comunidad de Madrid (CEIm), Spain approved the study (ref.: 07/694487.9/17). All procedures were in accordance with the 1964 Helsinki Declaration and its later amendments. Written informed consent was obtained from all participants. Inclusion criteria were male endurance cross-country athletes, 18–50 years of age, with high physical condition (VO2 ≥ 55 ml (kg min)−1 and BMI 18–25 kg m−2; exclusion criteria were consumption of any kind of nutritional or ergogenic supplement, be vegetarian or vegan, smoke, chronic medication, gastrointestinal surgery, or any diagnosed disease. Subjects were randomized to two equal-size treatment groups for 10 weeks using the RAND function of Excel (Microsoft Office Excel 2019). The cocoa group (CO; n = 22) received 5 g of fat-reduced cocoa containing 425 mg of flavonols and the control group (CT; n = 22) received 5 g of maltodextrin. Both supplements were provided in identical single-dose paper sachets to dissolve in semi-skimmed milk which does not affect cocoa flavonoid bioavailability.17 Maltodextrin is used as a carbohydrate supplement by athletes, however, 5 g of maltodextrin provided insufficient carbohydrate or energy bolus to adversely impact daily macronutrient and energy intake.18 Maltodextrin was not colored of cocoa- flavored and participants were told they were taking two supplements to study their effect on body composition and performance. Sachets were dispensed by postal mail and researchers that collected and analysed the data were blinded. Supplements were consumed during breakfast before exercise tests. During the 10 week intervention period, participants were called by telephone every week to check that they were taking the corresponding supplement. All measurements were recorded before and after the 10 week endurance training intervention. Dietary and training monitoring Dietary habits of participants were recorded at the beginning and after the 10 week intervention period. A Food Frequency Questionnaire (FFQ)19 and three 24 hour dietary recalls (two weekdays and one weekend day) were used. Data were analyzed using Dietsource software 3.0 (Novartis, Barcelona, Spain) to obtain the dietary intake of carbohydrates, protein, fats, and fiber, and the total energy. Participants were asked to maintain their normal diet throughout the training period. A single dose of cocoa (5 g) provided 13 kcal, 1.2 g of protein, 0.7 g of carbohydrates and 0.5 g of fat whereas 5 g of maltodextrin provided 19 kcal, 0 g of protein, 4.8 g of carbohydrates and 0 g of fat. Anthropometry and body composition Height and body mass were measured with a stadiometer (Asimed T2, Barcelona, Spain) and a balance scale (Ano Sayol SL, Barcelona, Spain), respectively; body mass index (BMI) was calculated as body mass (kg)/height (m2). Body composition was evaluated by dual- energy X-ray absorptiometry (Hologic DXA scan, Hologic Inc., Barcelona, Spain). Quality control calibration procedures were performed on a spine phantom. Subjects wore typical athletic clothing and removed all metal jewelry. They were positioned supine on the device within the borders delineated by the scanning table. The measures of body composition were as follows: total body fat percentage (BF%), estimated visceral adipose tissue (VAT) and fat (%) and fat-free mass (kg) distribution in the trunk and upper and lower body limbs. The interday coefficient of variation was 1.78%. The participants were endurance (cross-country) athletes. The athletes were training during the 10 week intervention with 5 or 6 training sessions per week. They followed a polarized distribution of training time, approximately 77% in zone 1, ∼5% in zone 2, and 17–18% in zone 3, similar to the protocols of Seiler and Kjerland20 and Esteve-Lanao et al.21 The total training load was approximately 43%, 7% and 50% (Z1–Z2–Z3, respectively). The objective load scale training load quantification method was used.22 Participants reported the duration of the training session in each training area individually 2 h after each work session, and the HR data was recorded in each work session (Garmin Forerunner 235, Garmin Ltd, Georgetown, DC, USA). Determination of maximal oxygen uptake, first and second ventilatory threshold, and maximal aerobic speed After a standardized warm-up of 10 minutes of continuous running on a treadmill (H/P/Cosmos Venus, NussdorfTraunstein, Germany) at 60% of their maximum heart rate (HRmax), subjects performed a maximum oxygen consumption test (VO2max) with a gas analyzer (Ultima™ Series, MGC Diagnostic Corporation, St Paul, MN, USA), which was calibrated according to the manufacturer’s instructions. The volume calibration was performed at different flow rates with a 3 L calibration syringe allowing an error <3%. The calibration of gas analyzers was performed automatically using reference values of environmental gases and cylinders (16% O2, 4% CO2). During the progressive test, HR and oxygen consumption values were constantly monitored and associated with the speed at which the first ventilatory threshold, second ventilatory threshold and maximal aerobic speed (VT1, VT2 and MAS, respectively) were identified. Based on the classical 3-phase model of Skinner and McLellan,23 three main zones were differentiated: VT2 (at or beyond VT2 – zone 3). The following variables were determined: oxygen uptake (VO2), pulmonary ventilation (VE), ventilatory equivalents for oxygen (VE/VO2) and carbon dioxide (VE/CO2), and end-tidal partial pressure of oxygen (PETO2) and carbon dioxide (PETCO2). Using the determined variables absolute maximal oxygen consumption (VO2maxABS), relative maximum oxygen consumption (VO2maxREL), maximal aerobic speed (MAS), first ventilator threshold (VT1) and second ventilator threshold (VT2) were calculated as is described elsewhere.21 The protocol started with a slope of 1% at a speed of 10 km h−1, with increments of 0.3 km h−1 every 30 s until volitional exhaustion.21 Finally, ten minutes after the treadmill test, volunteers ran 1 kilometer (t1km) outdoor as fast as possible and the time needed to cover the distance was recorded. Myostatin, follistatin and leptin analyses Before and after 10 weeks of supplementation, blood samples were collected in vacutainer EDTA-tubes and plasma samples were obtained by centrifugation at 3000 rpm for 10 min, and stored at −80 °C until use. Myostatin and follistatin levels were measured using RayBiotech ELISA kits (Norcross, GA, USA) and leptin analysis was performed using a Human Leptin ELISA (Peprotech, Rocky Hill, NJ, USA) according to the manufacturer’s instructions. Statistical analysis Statistical analysis was carried out using SPSS software 21.0 (SPSS, Chicago, IL, USA). Normality of distribution (Shapiro–Wilk test) and homogeneity of variance (Levene’s test) and sphericity (Mauchly’s test) were checked before analyses. When sphericity was violated the Greenhouse–Geisser test was chosen. All data were analyzed with the use of a 2-factor repeated measures ANCOVA with the supplement intake (between) and time (within) as the main variables. The variables VO2maxABS and VO2maxREL were used as covariates in body composition analyses. The Wilcoxon signed-rank test or the Mann– Whitney test were used when normal distribution could not be assumed. The significance was set at p < 0.05. The effect size was calculated using G*Power 3.1.9.4 software. Graphs were built with SigmaPlot V14.0 (Systat Software, Inc. San Jose, CA, USA). Results Cocoa characterization Quantitative analysis showed that the cocoa employed in this study contained mostly flavonoids with 99.6% of flavanols (Table 1) and small amounts of flavonols (quercetin derivatives), the flavanone naringin and the flavone luteolin (data not shown). Within the flavanol group, the major compound was procyanidin B2 (81%), followed by (−) epicatechin (9.5%), catechin (4.58%) and procyanidin B1 (3.79%) (Table 1). The analysis of the degree of polymerization showed that the cocoa employed had an mDP = 2.45. Cocoa also contained 4 mg g−1 of caffeine and 25 mg g−1 of theobromine (indicated by the manufacturer). Characteristics of subjects and dietary intake A total of 44 subjects were enrolled in the study and 32 finished the study: 15 in the CO group and 17 in the CT group (Fig. 1). Seven subjects dropped out from the CO group and 5 from the CT group because of illness (not related to the study), muscular injury, or because they did not follow the training program (Fig. 1). After drop- out, the effect size for the primary outcome (body fat percentage) was computed, obtaining an effect size dz = 1.06 with an α = 0.05 and power (1 − β) =0.095, indicating that observed effect in this parameter is large. The 95% confidence interval was (0.331–2.13). Age and BMI were similar between the two groups (age: CO = 33 ± 7 years; CT = 36 ± 8 years; p = 0.315) (BMI: CO = 21.80 ± 1.30 kg m−2; CT = 22.35 ± 2.13 kg m−2; p = 0.401). Macronutrient data obtained from the FFQ and the three 24 h recalls were not significantly different (data not shown). Both dietary assessment methods indicated that dietary intake was maintained along the study (Table 2). In general, carbohydrate intake was in the lower range of the recommendations of the European Food Safety Authority (EFSA), which recommends an intake of 45 to 60% of total energy intake. The intake of fiber was adequate (above the recommended 25 g d−1).24 Protein intake was within the recommendations for endurance and strength-trained athletes of the American Dietetic Association, the Dietitians of Canada, and the American College of Sports (recommend intake 1.2–2 gper kg body mass per day).25 Fat intake was close to the upper level of EFSA recommendations for fat intake (range 20–35% of total energy intake).24 Body composition No significant changes were found in terms of body mass and body composition along the study before and after 10 weeks of supplementation. However, significant changes were found in several body composition parameters when intragroup analyses were performed (Table 3). In the analysis of body composition, VO2maxABS and VO2maxREL covariables were analyzed as possible factors that could impact the observed results, but no influence of these variables was detected. A decrease in the percentage of total body fat (BF%), visceral fat, as well as the percentage of fat in the lower limbs was observed in the CO group after the intervention (Table 3). By contrast, no significant changes were observed for any of the analyzed fat-free mass parameters. Training load and performance Training load, volume and time spent in training were not significantly different between groups. The 10 weeks training did not produce significant t differences between groups. There was an improvement in VO2maxABS, VO2maxREL, VT1, VT2, MAS and the t1km (p < 0.001) in both groups after the intervention (Table 4). Circulating follistatin, myostatin and leptin Intergroup comparison did not detect any differences before and after supplementation between groups (T0: p = 0.423, p = 0.182, p = 0.187 and T = 10: p = 0.380, 0.987 and 0.418) for myostatin, follistatin and follistatin/myostatin ratio respectively. Time per group interaction was also no significant (p = 0.741, p = 0.242 and p = 0.741) for myostatin, follistatin and follistatin/myostatin ratio. There were no significant changes over time in myostatin levels with either treatment (CO t = 0 = 4.83 ± 2.71 ng mL−1, CO t = 10 weeks = 4.12 ± 1.72 ng mL−1, p = 0.390; CT t = 0 = 4.12 ± 1.72 ng mL−1; CT t = 10 weeks = 3.99 ± 1.42 ng mL−1, p = 0.741); however, follistatin levels increased significantly after 10 weeks of cocoa consumption (Fig. 2A and B). The change in the levels of follistatin also produced a significant change in the follistatin/myostatin ratio in the CO group (Fig. 2C). Lastly, a significant decrease in leptin was detected in the CO group after 10 weeks of supplementation (Fig. 2D) that was also significant when time per group interaction was analyzed (p = 0.030). Discussion Our results show that cocoa improves the body composition of endurance athletes by decreasing body fat levels, without impacting or improving performance. The observed body fat changes could be mediated, at least in part, by the modification of the follistatin/myostatin ratio and by the decrease in leptin levels. Indeed, previous studies in animals have shown that cocoa supplementation counteracts the effects of diets high in fat, preventing weight gain and total fat mass increase, by modulating lipid metabolism.26–28 The effect of cocoa on body composition in a healthy animal model has been addressed very recently and showed that cocoa decreased body mass, BMI, Lee’s index (which evaluates the degree of obesity) and retroperitoneal fat content.29 The positive effects of cocoa on body composition are believed to be attributed to its flavanols content, particularly the monomeric flavanol epicatechin,30 because of its high bioavailability and potential to exert systemic effects.31 Likewise, the body fat-lowering effects of cocoa have been ascribed to the presence of oligomeric procyanidins,32 which although fairly non-absorbable, can exert their effects at the intestinal level by inhibiting the intestinal absorption of cholesterol and bile acids and modifying the enzymatic activity of α-glucosidase.28 The majority of studies using cocoa or its food derivatives do not carry out a detailed analysis of its composition, which can be key to explain the observed effects. Indeed, there is great variability in cocoa flavonoid content, both in quality and quantity, and there are many factors that influence its composition, for example, the variety and the geographical location of the cocoa beans, and the manufacturing process of the cocoa (fermentation, roasting, alkalization).7,33,34 This variability makes it necessary to perform detailed analyses of the cocoa used in studies. The cocoa supplement used in the present study consisted mainly of procyanidin B2 and approximately 10% epicatechin, with a degree of polymerization of 2.45, in addition to caffeine and theobromine. The epicatechin and catechin contents of the cocoa employed here were 10-fold and 8-fold higher, respectively, than those reported in a study in which 11 cocoa powders were analyzed,7 and the dose used was similar to that employed in other human studies assessing endothelial function and vascular stiffness.35,36 Although studies in animals show a clear effect of cocoa on body fat composition, the relationship in humans is not so clear. The epidemiological data point to an effect of cocoa on fat mass, however, the results of intervention studies are heterogeneous. In a study on 2734 healthy female twins, it was observed that higher intake of flavanols, the principal polyphenols in cocoa, was associated with a lower fat mass independently of genetic or other environmental factors.37 Likewise, in European adolescents, high chocolate consumption was associated with lower central fatness,38 and in a pilot study of normal weight obese women (excessive body fat but a normal BMI), intake of dark chocolate promoted a decrease in the abdomen circumference.39 By contrast, chronic supplementation with cocoa (that provided 451 mg of flavanols per day) along with exercise failed to change body composition in a study of overweight and obese subjects despite improving numerous cardiometabolic risk factors.40 In a recent study carried out in college students who performed light-to-moderate intensity exercise, the consumption of high-cocoa chocolate (508 mg day−1 of polyphenols) for 4 weeks failed to produce any changes in body composition, but the study was not placebo-controlled and its duration was shorter than our present study,41 which might have had an impact on the results. Recently, it has been described that the main flavanols in cocoa, epicatechin, catechin and procyanidin B2, increase beta-oxidation and energy expenditure of white adipose tissue.30 In fact, in a study carried out with a catechin-rich green tea extract, a 1.6% decrease in body fat percentage caused an 11% improvement in performance, probably not due to fat loss itself, but to increased fatty acid oxidation.42 The decrease in body fat mass in our study occurred both in the trunk and the lower limbs, highlighting that visceral fat mass, a cardiovascular risk factor, is a target of cocoa. However, the fat loss in the CO group was not accompanied by an improvement in sports performance, as there were no significant differences between the two groups. In contrast to these results, a chronic study (3 months) using dark chocolate (175 mg of flavanols; 26 mg of epicatechin) found an improvement in cycling exercise capacity (VO2max and power),43 although the study was performed in sedentary people with an initial aerobic capacity much lower than the aerobic capacity of the athletes of our study (24 mL per kg per min VO2max vs. 58 mL per kg per min VO2max) and the supplementation was not accompanied by exercise training. Studies employing an acute intake of cocoa have generally failed to show improvements in physical performance (reviewed in ref. 5). Similar to those results, we did not detect any improvement in performance after cocoa supplementation. Low body fat composition is desirable in athletes to reduce energy expenditure, increase running economy and improve exercise performance. In contrast to our results, the intake of 25 g of cocoa for 7 days was found to improve the physical performance in soccer players by 4% (Cooper test).44 Further, the intake of 25 mg of epicatechin for seven days was found to increase handgrip strength and the plasma follistatin/myostatin ratio in humans.9 Because an increase in this ratio has been associated with an increase in skeletal muscle mass and a decrease in fat accumulation,45 we measured these proteins in plasma to test whether they were potentially responsible for the effect of cocoa on body fat composition (athletes in our study ingested 40 mg day−1 of epicatechin). Our results show that the chronic intake of cocoa increased the plasma ratio of follistatin/myostatin by modifying the levels of follistatin, without changes to myostatin. These data are in agreement with a recent human study by Schwarz et al. who reported no changes in muscle gene expression of myostatin after aerobic and anaerobic exercise training and supplementation with 200 mg daily of epicatechin for 4 weeks.46 By contrast, skeletal muscle myostatin expression in sedentary men was found to decrease after acute and long-term exercise.47 We found no changes in serum myostatin levels in our study, although physical improvements were detected in both intervention groups. These results are, however, tempered by the fact that we measured myostatin in blood and not in muscle. The changes of the follistatin/myostatin ratio were accompanied by a decrease in plasma leptin levels in the CO group. Leptin is a hormone secreted by adipose tissue that promotes lipolysis and depresses lipogenesis. Accordingly, the decrease in the levels of leptin after 10 weeks of supplementation could be a consequence of the loss of body fat. Indeed, the results of an in vitro study showed that a cocoa polyphenol powder decreases fat accumulation and the release of leptin in adipocytes,48 and similar effects have been reported for theobromine.49 The in vivo effects of cocoa consumption on leptin and associated mechanisms are not yet clear. It has been observed that cocoa powder, cocoa extract and epicatechin reduces body fat mass and leptin levels in a rat model of obesity,30 whereas in healthy rats cocoa failed to modify leptin levels after 3 weeks of supplementation, although a tendency for a decrease was observed.50 To the best of our knowledge, our study is the first in humans to demonstrate an effect of chronic consumption of cocoa on leptin levels. Consistent with our results, it has been observed that a diet enriched in flavonoids (provided by dark chocolate, dehydrated apple and green tea; giving an approximate daily dose of epicatechin of 425 mg), added to antihypertensive treatment, decreases leptin levels in young hypertensive patients.51 By contrast, acute intake of dark chocolate failed to modify leptin levels in postmenopausal women.52 A limitation of our study is that we have not carried out measurements in short-term periods to study the behavior of the determined parameters over time. Considering together our present findings and those of previous studies, it is clear that more work is needed to clarify the effect of cocoa on adipose and muscle tissue in humans, and this would likely benefit from the use of standardized preparations or individual cocoa components, to determine the underlying mechanisms of the observed effects. Similarly, regarding exercise performance, future research is necessary to assess whether cocoa supplementation, through its effect on body fat, could be beneficial in other exercise modalities. Conclusions Our study shows that consumption of a cocoa extract rich in polyphenols can lower total body and visceral fat mass in endurance athletes, whose fat composition is already intrinsically low. The change in body composition did not equate to an improvement in exercise performance. The modification of the body fat could be mediated by an increase of follistatin and therefore of the follistatin/myostatin ratio concomitant with a decrease in circulating leptin levels. Author contributions Jose Ángel García-Merino: investigation, Diego Moreno-Pérez: investigation, Beatriz de Lucas: investigation and writing, Maria Gregoria Montalvo-Lominchar: investigation, Elsa Muñoz: resources, Lara Sánchez: resources, Fernando Naclerio: supervision, Karen Marlene Herrera-Rocha: investigation, Martha Rocío Moreno-Jiménez: investigation, Nuria Elisabeth Rocha-Guzmán: investigation, Mar Larrosa: conceptualization, methodology, funding acquisition, supervision, writing – original draft. Conflicts of interest There are no conflicts to declare. Acknowledgements This research was funded by the Ministry of Economy and Competitiveness, Spain, project AGL2016-77288-R, Ramón y Cajal program grant (2012_11910), and by the Ministry ofEducation, Culture and Sports Spain (FPU grant 2016/01110). The authors acknowledge the study participants. We thank Dr Kenneth McCreath for critical reading of the manuscript. References 1 J. J. Knapik, R. A. Steelman, S. S. Hoedebecke, K. G. Austin, E. K. Farina and H. R. Lieberman, Prevalence of Dietary Supplement Use by Athletes: Systematic Review and MetaAnalysis., Sports Med., 2016, 46, 103–123. 2 C. Ferri, G. Desideri, L. Ferri, I. Proietti, S. Di Agostino, L. Martella, F. Mai, P. Di Giosia and D. Grassi, Cocoa, blood pressure, and cardiovascular health., J. Agric. Food Chem., 2015, 63, 9901–9909. 3 N. Parsaeyan, H. Mozaffari-Khosravi, A. Absalan and M. R. Mozayan, Beneficial effects of cocoa on lipid peroxidation and inflammatory markers in type 2 diabetic patients and investigation of probable interactions of cocoa active ingredients with prostaglandin synthase-2 (PTGS-2/ COX-2) using virtual analysis., J. Diabetes Metab. Disord., 2014, 13, 30. 4 O. A. Tokede, J. M. Gaziano and L. Djousse, Effects of cocoa products/dark chocolate on serum lipids: a metaanalysis., Eur. J. Clin. Nutr., 2011, 65, 879–886. 5 L. Decroix, D. D. Soares, R. Meeusen, E. Heyman and C. Tonoli, Cocoa Flavanol Supplementation and Exercise: A Systematic Review, Sports Med., 2018, 48, 867–892. 6 B. Albertini, A. Schoubben, D. Guarnaccia, F. Pinelli, M. Della Vecchia, M. Ricci, G. C. Di Renzo and P. Blasi, Effect of fermentation and drying on cocoa polyphenols., J. Agric. Food Chem., 2015, 63, 9948–9953. 7 C. Andres-Lacueva, M. Monagas, N. Khan, M. IzquierdoPulido, M. Urpi-Sarda, J. Permanyer and R. M. LamuelaRaventos, Flavanol and flavonol contents of cocoa powder products: influence of the manufacturing process., J. Agric. Food Chem., 2008, 56, 3111– 3117. 8 S. Langer, L. J. Marshall, A. J. Day and M. R. A. Morgan, Flavanols and methylxanthines in commercially available dark chocolate: A study of the correlation with nonfat cocoa solids, J. Agric. Food Chem., 2011, 59, 8435–8441. 9 G. Gutierrez-Salmean, T. P. Ciaraldi, L. Nogueira, J. Barboza, P. R. Taub, M. C. Hogan, R. R. Henry, E. Meaney, F. Villarreal, G. Ceballos and I. RamirezSanchez, Effects of (-)- epicatechin on molecular modulators of skeletal muscle growth and differentiation., J. Nutr. Biochem., 2014, 25, 91–94. 10 B. Deng, F. Zhang, J. Wen, S. Ye, L. Wang, Y. Yang, P. Gong and S. Jiang, The function of myostatin in the regulation of fat mass in mammals, Nutr. Metab., 2017, 14, 2–7. 11 S.-J. Lee, Y.-S. Lee, T. A. Zimmers, A. Soleimani, M. M. Matzuk, K. Tsuchida, R. D. Cohn and E. R. Barton, Regulation of muscle mass by follistatin and activins., Mol. Endocrinol., 2010, 24, 1998–2008. 12 R. Singh, M. Braga and S. Pervin, Regulation of brown adipocyte metabolism by myostatin/follistatin signaling., Front. Cell Dev. Biol., 2014, 2, 60. 13 M. R. Esco, M. V. Fedewa, Z. S. Cicone, O. A. Sinelnikov, D. Sekulic and C. J. Holmes, Field-Based Performance Tests Are Related to Body Fat Percentage and Fat-Free Mass, But Not Body Mass Index, in Youth Soccer Players., Sports, 2018, 6, 105. 14 B. Knechtle, Barandun, Knechtle, Klipstein, Rüst, T. Rosemann and R. Lepers, Running speed during training and percent body fat predict race time in recreational male marathoners, Open Access J. Sports Med., 2012, 51. 15 M. E. Alañón, S. M. Castle, P. J. Siswanto, T. CifuentesGómez and J. P. E. Spencer, Assessment of flavanol stereoisomers and caffeine and theobromine content in commercial chocolates, Elsevier Ltd, 2016, vol. 208. 16 J. A. Kennedy and G. P. Jones, Analysis of proanthocyanidin cleavage products following acid-catalysis in the presence of excess phloroglucinol., J. Agric. Food Chem., 2001, 49, 1740–1746. 17 E. Roura, C. Andrés-Lacueva, R. Estruch, M. L. Mata-Bilbao, M. Izquierdo-Pulido, A. L. Waterhouse and R. M. LamuelaRaventós, Milk does not affect the bioavailability of cocoa powder flavonoid in healthy human, Ann. Nutr. Metab., 2008, 51, 493–498. 18 A. Naderi, E. P. de Oliveira, T. N. Ziegenfuss and M. T. Willems, Timing, Optimal Dose and Intake Duration of Dietary Supplements with Evidence-Based Use in Sports Nutrition., J. Exercise Nutr. Biochem., 2016, 20, 1–12. 19 J. Vioque and J. Quiles, Encuesta de nutrición y salud de la Comunidad Valenciana, 2003. 20 K. S. Seiler and G. O. Kjerland, Quantifying training intensity distribution in elite endurance athletes: is there evidence for an ‘optimal’ distribution?, Scand. J. Med. Sci. Sports, 2006, 16, 49–56. 21 J. Esteve-Lanao, C. Foster, S. Seiler and A. Lucia, Impact of training intensity distribution on performance in endurance athletes., J. Strength Cond. Res., 2007, 21, 943–949. 22 J. Esteve-Lanao, D. Moreno-Pérez, C. A. Cardona, E. Larumbe-Zabala, I. Muñoz, S. Sellés and R. Cejuela, Is marathon training harder than the Ironman training? An ECO- method comparison, Front. Physiol., 2017, 8, 1–8. 23 J. S. Skinner and T. H. Mclellan, The transition from aerobic to anaerobic metabolism, Res. Q. Exercise Sport, 1980, 51, 234–248. 24 European Food Safety Agency, TECHNICAL REPORT Dietary Reference Values for nutrients Summary report, DOI: 10.2903/sp.efsa.2017.e15121. 25 N. R. Rodriguez, N. M. DiMarco and S. Langley, Position of the American Dietetic Association, Dietitians of Canada, and the American College of Sports Medicine: Nutrition and athletic performance., J. Am. Diet. Assoc., 2009, 109, 509–527. 26 Y. Gu, S. Yu, J. Y. Park, K. Harvatine and J. D. Lambert, Dietary cocoa reduces metabolic endotoxemia and adipose tissue inflammation in high-fat fed mice., J. Nutr. Biochem., 2014, 25, 439–445. 27 N. Matsui, R. Ito, E. Nishimura, M. Yoshikawa, M. Kato, M. Kamei, H. Shibata, I. Matsumoto, K. Abe and S. Hashizume, Ingested cocoa can prevent high-fat dietinduced obesity by regulating the expression of genes for fatty acid metabolism., Nutrition, 2005, 21, 594–601. 28 Y. Yamashita, M. Okabe, M. Natsume and H. Ashida, Prevention mechanisms of glucose intolerance and obesity by cacao liquor procyanidin extract in high-fat diet-fed C57BL/6 mice, Arch. Biochem. Biophys., 2012, 527, 95–104. 29 M. Camps-Bossacoma, M. Garcia-Aloy, S. Saldana-Ruiz, T. Cambras, R. Gonzalez- Dominguez, A. Franch, F. J. PerezCano, C. Andres-Lacueva and M. Castell, Role of Theobromine in Cocoa’s Metabolic Properties in Healthy Rats., J. Agric. Food Chem., 2019, 67, 3605–3614. 30 G. Rabadan-Chávez, L. Quevedo-Corona, A. M. Garcia, E. Reyes-Maldonado and M. E. Jaramillo-Flores, Cocoa powder, cocoa extract and epicatechin attenuate hypercaloric diet-induced obesity through enhanced β-oxidation and energy expenditure in white adipose tissue, J. Funct. Foods, 2016, 20, 54–67. 31 G. Borges, J. I. Ottaviani, J. J. J. van der Hooft, H. Schroeter and A. Crozier, Absorption, metabolism, distribution and excretion of (-)-epicatechin: A review of recent findings., Mol. Aspects Med., 2018, 61, 18–30. 32 M. R. Dorenkott, L. E. Griffin, K. M. Goodrich, K. A. Thompson-Witrick, G. Fundaro, L. Ye, J. R. Stevens, M. Ali, S. F. O’Keefe, M. W. Hulver and A. P. Neilson, Oligomeric cocoa procyanidins possess enhanced bioactivity compared to monomeric and polymeric cocoa procyanidins for preventing the development of obesity, insulin resistance, and impaired glucose tolerance during high-fat feeding., J. Agric. Food Chem., 2014, 62, 2216–2227. 33 N. A. Febrianto and F. Zhu, Intravariety Diversity of Bioactive Compounds in Trinitario Cocoa Beans with Different Degrees of Fermentation., J. Agric. Food Chem., 2019, 67, 3150–3158. 34 N. A. Febrianto and F. Zhu, Diversity in Composition of Bioactive Compounds Among 26 Cocoa Genotypes., J. Agric. Food Chem., 2019, 67, 9501–9509. 35 C. Heiss, R. Sansone, H. Karimi, M. Krabbe, D. Schuler, A. Rodriguez-Mateos, T. Kraemer, M. M. Cortese-Krott, G. G. C. Kuhnle, J. P. E. Spencer, H. Schroeter, M. W. Merx and M. Kelm, Impact of cocoa flavanol intake on agedependent vascular stiffness in healthy men: a randomized, controlled, double-masked trial., Age, 2015, 37, 9794. 36 R. Sansone, A. Rodriguez-Mateos, J. Heuel, D. Falk, D. Schuler, R. Wagstaff, G. G. C. Kuhnle, J. P. E. Spencer, H. Schroeter, M. W. Merx, M. Kelm and C. Heiss, Cocoa flavanol intake improves endothelial function and Framingham Risk Score in healthy men and women: a randomised, controlled, double-masked trial: the Flaviola Health Study., Br. J. Nutr., 2015, 114, 1246–1255. 37 A. Jennings, A. MacGregor, T. Spector and A. Cassidy, Higher dietary flavonoid intakes are associated with lower objectively measured body composition in women: evidence from discordant monozygotic twins., Am. J. Clin. Nutr., 2017, 105, 626–634. 38 M. Cuenca-Garcia, J. R. Ruiz, F. B. Ortega and M. J. Castillo, Association between chocolate consumption and fatness in European adolescents., Nutrition, 2014, 30, 236–239. 39 L. Di Renzo, M. Rizzo, F. Sarlo, C. Colica, L. Iacopino, E. Domino, D. Sergi and A. De Lorenzo, Effects of dark chocolate in a population of normal weight obese women: A pilot study, Eur. Rev. Med. Pharmacol. Sci., 2013, 17, 2257– 2266. 40 K. Davison, A. M. Coates, J. D. Buckley and P. R. C. Howe, Effect of cocoa flavanols and exercise on cardiometabolic risk factors in overweight and obese subjects, Int. J. Obes., 2008, 32, 1289–1296. 41 M. Nishiwaki, Y. Nakano and N. Matsumoto, Effects of regular high-cocoa chocolate intake on arterial stiffness and metabolic characteristics during exercise., Nutrition, 2019, 60, 53–58. 42 J. D. Roberts, M. G. Roberts, M. D. Tarpey, J. C. Weekes and C. H. Thomas, The effect of a decaffeinated green tea extract formula on fat oxidation, body composition and exercise performance, J. Int. Soc. Sports Nutr., 2015, 12, 1–9. 43 P. R. Taub, I. Ramirez-Sanchez, M. Patel, E. Higginbotham, A. Moreno-Ulloa, L. M. Román-Pintos, P. Phillips, G. Perkins, G. Ceballos and F. Villarreal, Beneficial effects of dark chocolate on exercise capacity in sedentary subjects: Underlying mechanisms. A double blind, randomized, placebo controlled trial, Food Funct., 2016, 7, 3686–3693. 44 J. A. Gonzalez-Garrido, J. R. Garcia-Sanchez, S. GarridoLlanos and I. M. Olivares- Corichi, An association of cocoa consumption with improved physical fitness and decreased muscle damage and oxidative stress in athletes., J. Sports Med. Phys. Fitness, 2017, 57, 441–447. 45 M. Nakatani, M. Kokubo, Y. Ohsawa, Y. Sunada and K. Tsuchida, Follistatin-derived peptide expression in muscle decreases adipose tissue mass and prevents hepatic steatosis., Am. J. Physiol. Endocrinol. Metab., 2011, 300, E543–E553. 46 N. A. Schwarz, Z. J. Blahnik, S. Prahadeeswaran, S. K. McKinley-Barnard, S. L. Holden and A. Waldhelm, (-)-Epicatechin Supplementation Inhibits Aerobic Adaptations to Cycling Exercise in Humans., Front. Nutr.,2018, 5, 132. 47 M. Hjorth, S. Pourteymour, S. W. Gorgens, T. M. Langleite, S. Lee, T. Holen, H. L. Gulseth, K. I. Birkeland, J. Jensen, C. A. Drevon and F. Norheim, Myostatin in relation to physical activity and dysglycaemia and its effect on energy metabolism in human skeletal muscle cells., Acta Physiol., 2016, 217, 45–60. 48 C. L. Campbell, E. A. Foegeding and G. K. Harris, Cocoa and Whey Protein Differentially Affect Markers of Lipid and Glucose Metabolism and Satiety, J. Med. Food, 2016, 19, 219–227. 49 Y. J. Jang, H. J. Koo, E.-H. Sohn, S. C. Kang, D.-K. Rhee and S. Pyo, Theobromine inhibits differentiation of 3 T3-L1 cells during the early stage of adipogenesis via AMPK and MAPK signaling pathways., Food Funct., 2015, 6, 2365–2374. 50 M. Massot-Cladera, J. Mayneris-Perxachs, A. Costabile, J. R. Swann, À. Franch, F. J. Pérez-Cano and M. Castell, Association between urinary metabolic profile and the intestinal effects of cocoa in rats, Br. J. Nutr., 2017, 117, 623–634. 51 M. M. de Jesús Romero-Prado, J. A. Curiel-Beltrán, M. V. Miramontes-Espino, E. G. Cardona-Muñoz, A. RiosArellano and L. B. Balam-Salazar, Dietary Flavonoids Added to Pharmacological Antihypertensive Therapy are Effective in Improving Blood Pressure, Basic Clin. Pharmacol. Toxicol., 2015, 117, 57–64. 52 C. E. Marsh, D. J. Green, L. H. Naylor and K. J. Guelfi, Consumption of dark chocolate attenuates subsequent food intake compared with milk and white chocolate in postmenopausal women, Appetite, 2017, 116, 544–551. TABLES FIGURES