Thirteen healthy, well-trained, recreational male athletes took part in the study. All the subjects were informed of the purpose and demands of the study before giving their written consent to participate. The protocol was in accordance with the Declaration of Helsinki for research of human subjects and was approved by the Balearic Islands Clinical Investigation Ethics Committee (reference IB 2399/14 PI).
Participants were enrolled after fulfilling all inclusion criteria and possessing none of the exclusion criteria. Subjects could be included if they were currently healthy, consumed caffeine regularly in any form (coffee, tea, soda, supplements, etc.), were between 20 and 45 years old, and engaged in at least 8 h of training or competition per week. Exclusion criteria were: smokers, use of analgesic or anti-inflammatory drugs within the previous 2 weeks, habitual alcohol consumers, and presence of injury or illness incompatible with the sport practice or caffeine intake. Fourteen subjects were initially recruited, but one of them was dropped from the study, as he had only performed the main exercise in the placebo condition due to lack of availability. Table 1 shows the main characteristics of participants in the study. Participants in the study did not consume caffeine supplements, and they reported a habitual caffeine intake similar to the one that would result from 1 to 2 cups of coffee.
A randomized, crossover, double-blinded study of supplementation with caffeine was performed. At 8–10 days before the beginning of the study, each subject performed a continuous incremental exercise test on a treadmill (H-P Cosmos, Pulsar 3, Nussdorf-Traunstein, Germany) until volitional exhaustion to determine their maximal oxygen uptake (VO2max). The participants warmed up for 3 min at 4 km/h before starting the test. The test began at 6 km/h, and the participant work rate was increased by 1 km/h every minute until exhaustion. Each test was ended when added work did not increase or decrease the oxygen consumption, and the resulting value was recorded as the VO2max. Expired gas was continuously analyzed (H-P Cosmos, Jaeger-MasterScreen CPX, Nussdorf-Traunstein, Germany), and heart rates (HR) were measured continuously using short-range radio telemetry (Polar Beat; Polar Electro, Oy, Finland). Level running speeds equivalent to 70% VO2max were subsequently calculated via regression equations from the VO2-running speed relationship. Before the exercise test, participants’ height, weight and skinfolds required to measure body fat mass using the Faulkner equation were measured (triceps, abdominal, subscapular, and suprailiac).
Questionnaires to determine habitual nutritional and caffeine intake, as well as the amount of physical activity performed, were filled out while the participant was in the laboratory (by the participant himself) or were given to be completed and turned in on the first main trial day. Habitual physical activity was determined by using the International Physical Activity Questionnaire (IPAQ) , thus providing quantitative information on training loads in metabolic equivalent (MET)-h/week. Habitual nutritional intake was assessed using a self-reported 7-day food record. Mean energy, carbohydrate, protein, and lipid contents for each participant’s diet were calculated using commercial software (Nutrisalud; CSG Software, Huesca, Spain) based on Spanish food composition tables. Furthermore, habitual caffeine intake was measured with a self-reported caffeine consumption questionnaire used by our group [16, 17].
Finally, each subject was given a comprehensive list of caffeine-containing foods and drinks and was instructed to abstain from these products during the 24 h preceding each exercise trial. The subjects were not allowed to use analgesic or anti-inflammatory drugs during the study protocol, and they were instructed to maintain their habitual diet, to refrain from alcohol intake and not to participate in any sporting activity during the 48 h preceding each main experimental trial.
For the main exercise trials, on two occasions 1 week apart, the subjects reported to the laboratory at the same hour and were randomly assigned to either the caffeine or placebo trial in a crossover random block design. The subjects were then required to empty their bladder before body mass (in shorts only) was recorded. After sitting quietly for 10 min, an initial resting blood sample was obtained from an antecubital vein by venipuncture. Following blood sampling, in the caffeine trial subjects were given 6 mg·kg− 1 body mass of caffeine dissolved in 200 ml of fruit juice drink; in the placebo trial, the subjects were given the same volume of only the fruit juice. Both drinks were matched to be similar in taste and appearance. A 24-h recall was performed to ensure that nutritional intake was similar before each trial (results not shown) and within the participants’ habitual nutritional intake. After resting quietly in the laboratory for 1 h and a short warm-up, subjects began running on the treadmill at the speed equivalent to 70% VO2max for 60 min. Heart rates were continuously recorded, and expired air was analyzed every 20 min thereafter for control of intensity. An additional venous blood sample was obtained at immediately post-exercise before body mass (in shorts only) was recorded again. A final venous blood sample was obtained at 2 h post-exercise. For all samples, 12 ml of blood was collected, and samples were obtained with the subject in a seated position. For both trials, subjects could drink water ad libitum, with the water intake during the test measured. No other fluid or food intake was allowed until the blood sample had been collected at 2 h post-exercise. Laboratory conditions were 22.0 ± 0.6 °C and 54.2 ± 6.1% relative humidity.
Blood sampling and measurements
Seated venous blood samples were collected in suitable vacutainers with ethylenediaminetetraacetic acid (EDTA) or heparin as the anticoagulant. Hematocrit and hemoglobin were determined in the EDTA sample using a hematology analyzer (Horiba ABX Pentra 60, Diagnostics) for estimating plasma volume . Within 30 min after blood collection, plasma was obtained by centrifugation (15 min, 1000×g, 4 °C) of the EDTA-blood samples. These plasma samples were stored at − 70 °C until measurements were performed. Concentrations of cytokines (IL-10, IL-6, IL-8, IL-1ra, IL-4, IL-1β, IL-12 and IFN-γ, caffeine, adrenaline, cortisol and cAMP were measured in the EDTA plasma samples.
Whole heparinized blood was incubated with 10 ng/ml lipopolysaccharide (LPS, Escherichia coli serotype 055:B5; Sigma, St Louis, MO, USA) dissolved in culture medium (RPMI-1640 Medium, Sigma, St Louis, MO, USA), or with the same volume of culture medium (spontaneous production) for 24 h at 37 °C. Immediately after incubation, samples were centrifuged at 1000 g for 10 min to obtain the supernatants. Aliquots were stored at − 70 °C until assay. Concentrations of IL-10, IL-6, IL-8, IL-1ra and TNF-α were determined in these culture supernatants. Monocyte counts were used to normalize cytokine production (difference between cytokine concentration in stimulated and unstimulated cultures) on a per cell basis .
Caffeine was measured in plasma by an HPLC method as previously described . Adrenaline and cortisol were measured in plasma using commercially available enzyme-linked immunosorbent assay kits (Abnova Corporation, Taiwan) and (Elabscience Biotechnology Co, Ltd) respectively, with a spectrophotometric microplate reader (PowerWavei; BioTek, Winooski, VT). cAMP concentrations were determined in plasma using commercially available enzyme-linked immunosorbent assay kits (Arbor Assays) with a spectrophotometric microplate reader (PowerWavei; BioTek, Winooski, VT).
Cytokine concentrations were determined in plasma and in culture supernatants using commercially available enzyme-linked immunosorbent assay kits (Invitrogen, Carlsbad, CA, USA), with a spectrophotometric microplate reader.
Statistical analysis was carried out using the Statistical Package for Social Sciences (IBM SPSS Statistics 23.0 for windows). The results are expressed as the means and standard deviations (S.D.), and p < 0.05 was considered statistically significant for all analysis. All the data were tested for their normal distribution (Shapiro-Wilk test). If the data were not normally distributed, statistical analysis was carried out on the logarithmic transformation of the data (cytokine, adrenaline, and cortisol concentrations). Exercise-related parameters (weight lost, maximum and average HR) under both conditions, placebo and supplemented, were compared using a t-test for unpaired data. Changes in blood variables during the study were analyzed using a time (pre-exercise, post-exercise, recovery) x condition (caffeine, placebo) within-between subjects ANOVA with repeated measures. Any significant F ratios subsequently shown were assessed using post hoc comparisons with Holm-Bonferroni correction for multiple comparisons applied to the unadjusted p value. Multiple linear regression models for IL-10 and IL-6 (logarithmic transformed) were also analyzed. Independent variables considered were time (pre-exercise, post-exercise, recovery), condition (caffeine, placebo) and adrenaline concentrations. The IL-10 model also included IL-6 as an independent variable.