Our study conforms to the CONSORT statement, we provided details of the study population and background; subject selection (inclusion/exclusion criteria); methods of randomization and blinding; description of potential sources of bias; efficacy and safety measures; description of quantitative variables and statistical methods with sample size definition.

Forty-one healthy, college-aged males were chosen for this study. Subjects needed to have had uninterrupted endurance training experience for a minimum of three months and three or more sessions each day. We excluded subjects with the following conditions: neurological, cardiorespiratory, musculoskeletal, endocrine, renal, metabolic and other related disease(s) that would prevent the experimental procedures from being performed, individuals with resting systolic blood pressure >130 mmHg and resting diastolic >90 mmHg, smokers, subjects undergoing pharmacotherapies, sedentary and inadequately active subjects consistent with the International Physical Activity Questionnaire.15

We excluded subjects who stated they used anabolic steroids. Those who presented series of RR intervals with under 95% of sinus beats and those who were unable to complete all stages of the protocols (Figure 1).

Figure 1.

Mean values and respective standard deviation of systolic arterial pressure (A), diastolic arterial pressure (B) and heart rate (C) obtained in the control and caffeine protocols at rest and during recovery (Rec).

Ca: Caffeine; Co: Control; HR Peak: maximum heart rate value during exercise. R1: 5th minute of rec; R2: 10th minute of rec; R3: 15th minute of rec; R4: 20th minute of rec and; R5: 30th minute of rec.

*Co: p<0.05 vs. Rest in Control protocol; *Ca: p<0.05 vs. Rest in Caffeine protocol; **Ca: p<0.05 vs. R1 in Caffeine protocol (n=30 subjects).

(0.17MB).

Informed consent was signed by all subjects before the tests. Experimental procedures were approved by the Research Ethics Committee of University of Pernambuco, Recife, PE, Brazil (Number 3.056.402) and were in agreement with the 466/2012 resolution of the National Health Council of 12 December 2012.

This is a prospective, crossover and controlled trial performed at the University of Pernambuco, Recife, PE, Brazil. The study was registered with ClinicalTrials.gov (NCT03899675).

So as to define potential sources of bias, the descriptive profile of the subjects was defined to characterize the sample, reduce the volatility of the variables, improving reproducibility and physiological interpretation.

Before the start of the experimental procedures, subjects were recognized according to age, mass (kg), height (m), systolic (mmHg) and diastolic arterial pressure (mmHg), waist (cm), abdominal (cm) and hip (cm) circumferences, waist-to-hip ratio and body mass index (BMI).

The experimental procedure was split into three stages with a minimum interval of 48 hours between them, so as to allow the subjects adequate recovery time. On the first day, the physical exercise intensity test was performed based on the 1RM test, which was completed before the other stages, as it was necessary to prescribe the exercise intensity in the others. In the second and third phase of the research, the control and CAF protocols were performed. The subsequent phases were the control and CAF protocols and the order in which they were implemented was established through a randomization process involving the tossing of a coin. The subjects were blinded throughout their protocol and were uninformed about the order of the protocols

Before the start of the first stage, body weight measurements on a digital scale (Welmy, W200/5, São Paulo, Brazil) and height using a stadiometer (ES 2020 - Sanny, São Paulo, Brazil) were recorded.

Training intensity was defined based on the maximal 1RM following the recommendations of the American College of Sports Medicines’ Guidelines for Exercise Testing and Prescription.16 Subjects were coached to test the load in the four proposed physical exercises with 1RM performed in the subsequent exercises:

• 1.

  Leg press 45°,

• 2.

  Extending chair,

• 3.

  Abductor chair and

• 4.

  Squatting.

On subsequent days of data collection, the exercises were performed based on the load of 75% 1RM for the aforesaid exercises, based on the load from the test 75% 1RM previously performed, a total of four sets were completed with a maximum of 10 repetitions for each exercise.

Prior to starting protocol, the HR monitor (Polar RS800CX, Finland) was positioned on the subjects to register HR beat-to-beat, followed by administration of a 300 mg CAF capsule or a placebo;according to the selected protocol. We set the dose of CAF at 300 mg as this is considered safe. The 2015-2020 Dietary Guidelines for Americans allow up to a maximum of 400 mg/day,17 whilst the United States Food and Drug Administration specifies that 300 mg is within the maximum allowed per day.18

Resting phase: Initially, the subjects were seated, at rest, for 5 minutes, whilst recording HR, SAP and DAP values measured in the last minute of the resting phase. Similarly, in this phase, HRV indices were recorded for 5 continuous minutes (rest interval).

Exercise: After these measurements the subjects performed strength exercise at 75% of 1 RM.

Recovery phase: At the end of the activity, the subjects were again located seated at rest, and were observed for 30 minutes. HR, SAP and DAP values were recorded within five, 10, 15, 20, and 30 minutes of recovery. HRV indices were recorded for 30 minutes of recovery with six intervals of 5 minutes each between them: Rec1 (0 to five minutes), Rec2 (five to 10 minutes), Rec3 (10 to 15 minutes), Rec4 (15 to 20 minutes), Rec5 (20 to 25 minutes) and Rec6 (25 to 30 minutes).” To avoid measurement errors in the assessed parameters, a single rater performed all measurements throughout the whole experiment.

The control phase was completed on the second day of collection, in which subjects did not ingest CAF. The datasets were collected in the standardized way, respecting the times established for the analysis of cardiac activity. Nevertheless, before the tests the subjects were questioned about the use of CAF. At the end of the first day of testing (control), subjects received a capsule containing 300 mg of CAF and in accordance with the earlier guidelines, were instructed to ingest the capsule 45 minutes before the next day of procedure, giving the subjects adequate time for digestion and absorption of the CAF.

SAP and DAP were checked indirectly using a stethoscope (Littman Classic II, Saint Paul, USA) and aneroid sphygmomanometer (Welch Allyn Tycos, New York, USA) placed on the subjects’ left arm.19

HRV analysis was performed throughout the entire experimental protocol using an HR monitor (Polar RS800CX, Finland), equipment validated formerly to record HR.

The recordings contained the required 256 consecutive stable RR intervals, which underwent digital filtering accompanied by manual filtering to eliminate artifacts. Only series greater than 95% of sinus HR were included in the study. Analysis of HRV linear methods in the time domain and the frequency were applied. In the time domain indexes square root of the average of the square of the differences between normal RR intervals adjacent (RMSSD) and standard deviation of the average of all normal RR intervals (SDNN) were necessary. The Poincaré plot indexes: standard deviation of the instantaneous rate variability the rhythm (SD1) and standard deviation of long-term continuous RR interval variability (SD2)and Geometric Index: Triangular index (RRTri)and Triangular interpolation of NN interval histogram (TINN) were also calculated.10

For HRV frequency domain analysis, the spectrum components low frequency and high frequency, in ms2 and in normalized units, extracted using the Fast Fourier Transform (Blackman & Tukey, FFT) were used. The frequency bands required for each component were: low frequency (LF: 0.04 to 0.15 Hz) and high frequency (HF: 0.15 to 0.40 Hz). We decided not to evaluate the LF/HF ratio as it has been demonstrated to be theoretically flawed and empirically unsupported to represent the sympathovagal balance.20

HRV analysis software (Kubios HRV®, Biosignal Analysis and Medical Image Group, Department of Physics, University of Kuopio, Finland)21 was used to scrutinize the linear indices in the time and frequency domains and to perform the Poincaré plots.

To describe sample size, a sample calculation was completed based on the study by Gonzaga et al.3 considering the RMSSD index. The assumed significant difference magnitude was 12 ms, in view of a standard deviation of 16.2 ms, with alpha risk of 5% and beta of 80%. A total of 28 subjects were required for the study.

Regarding the data analysis, we enforced the descriptive statistics on the characterization of the sample and the results were reported as mean values, standard deviation, minimum and maximum. Comparisons of the HRV indices and cardiovascular parameters between CAF and Control protocols; moments were achieved via two-way repeated measures analysis of variance. For analysis of the moments (rest versus recovery), we implemented the Bonferroni post-test for parametric distribution or post-test Dunn for non-parametric distribution. Statistical significance was set at p<0.05 (or <5%) throughout.

Computations were completed using Minitab version 13.20 (Minitab, PA, USA), GraphPad Instat version 3.01, 1998 (GraphPad Software, Inc., San Diego, California, USA) and IBM SPSS Statistics version 22.0 (SPSS Inc., Chicago, IL, USA).

We did not find changes between the groups by examining the resting HR, however, in the CAF group, HR was superior in all recovery period intervals. After the 5th minute of recovery, systolic blood pressure in the CAF cohort was found to be significantly higher when compared to the control group (CAF vs Control: SAP 148.00±15.68 vs 134.00±16.82 p=<0.05) (Figure 1).

On the whole, CAF impaired HRV recovery after resistance exercise. Significant differences were observed in the RMSSD (REC 5 vs. REC1** and REC 5 vs. REC 2*** p<0.001; REC 6 vs. REC1** and REC 6 vs. REC2*** p<0.001), SDNN (REC 5 vs. REC1** and REC 5 vs. REC 2*** p<0.001; REC 6 vs. REC1** and REC 6 vs. REC2*** p<0.001) (Figure 2), TINN (REC 4 vs. REST* p<0.001); REC 6 vs. REC1** p<0.0001 and REC 6 vs. REC2*** p<0.001) (Figure 3), SD1 (REC5 vs. REC1** p>0.0001 and REC 5 vs. REC2*** p>0.0001), SD2 (REC 5 vs. REST* p>0.0001; REC 5 vs. REC 1** p>0.0001 and REC 5 vs. REC2*** p>0.0001) and REC 6 vs. REC 1** p<0.0001 and REC 6 vs. REC 2*** p<0.0001) (Figure 4), LF (REC 5 vs., REC1** p=0.0005 and REC 5 vs., REC2*** p=0.0004) and REC 6 vs. REC1** p = 0.004 and REC 6 vs. REC 2 p = 0.0004), HF (REC 6 vs. REST* p=0.007) indices between the control group and the CAF group. The results of this study indicate that CAF in combination with exercise strength can impair vagal recovery for a period of 30 minutes (Figure 5).

Figure 2.

Mean values and respective standard deviations of time domain HRV indices obtained in the control and caffeine protocols at rest and during recovery (Rec).

Ca: Caffeine; Co: Control; RMSSD: root-mean square of differences between adjacent normal RR intervals in a time interval; SDNN: standard deviation of all normal RR intervals; ms: milliseconds;

*Co: p<0.05 vs. Rest in Control protocol; *Ca: p<0.05 vs. Rest in Caffeine protocol; **Co: p<0.05 vs. Rec 1 in Control protocol; **Ca: p<0.05 vs. R1 in Caffeine protocol; ***Ca: p<0.05 vs. Rec 2 in Caffeine protocol (n=30 subjects).

(0.13MB).

Figure 3.

Mean values and respective standard deviations of geometric HRV indices obtained in the control and caffeine protocols at rest and during recovery (Rec). Ca: Caffeine; Co: Control; ms: milliseconds; RRTri: triangular index; TINN: triangular interpolation of RR interval histogram; *Co: p<0.05 vs Rest in Control protocol; *Ca: p<0.05 vs. Rest in Caffeine protocol; **Ca: p<0.05 vs. Rec 1 in Caffeine protocol; ***Ca: p<0.05 vs. Rec 2 in Caffeine protocol (n=30 subjects).

(0.12MB).

Figure 4.

Mean values and respective standard deviations of the Poincaré HRV indices obtained in the control and caffeine protocols at rest and during recovery (Rec).

Ca: Caffeine; Co: Control; ms: milliseconds; SD1: standard deviation of the instantaneous variability of the beat-to-beat heart rate; SD2: standard deviation of long-term continuous RR interval variability. *Co: p<0.05 vs. Rest in Control protocol; *Ca: p<0.05 vs. Rest in Caffeine protocol; **Co: p<0.05 vs. Rec1 in Control protocol; **Ca: p<0.05 vs. Rec1 in Caffeine protocol; ***Ca: p<0.05 vs. Rec2 in Caffeine protocol (n=30 subjects).

(0.12MB).

Figure 5.

Mean values and respective standard deviations of frequency domain HRV indices obtained in the control and caffeine protocols at rest and during recovery (Rec). LF: low frequency; HF: high frequency; ms: milliseconds; Ca: Caffeine; Co: Control; *Co: p<0.05 vs Rest in Control protocol; *Ca: p<0.05 vs Rest in Caffeine protocol; **Co: p<0.05 vs Rec1 in Control protocol; **Ca: p<0.05 vs Rec1 in Caffeine protocol; ***Ca: p<0.05 vs Rec2 in Caffeine protocol (n=30 subjects).

(0.13MB).

HRV reduction is perceptible in the decrease of SDNN, as observed in the CAF group even 30 minutes after recovery (REC 6). CAF similarly delayed the recovery of the frequency domain indices LF and HF, where we found significant differences between the groups. Slow recovery of HF establishes that CAF delivers a reduction in parasympathetic conduction for a period of 30 minutes after exercise cessation (Rest vs. REC 6).

The RRtri, TINN and SD2 geometric indices were affected by CAF intake prior to exercise, meaning that its use enables a lower ANS adaptation to the stimulus of physical exercise, interpreted by an overactivity of the sympathetic response to exercise. From this viewpoint, in the CAF group, TINN showed delayed recovery until the last moment (REC 4, 5 and 6). The recovery of SD2 in the CAF group was late; this parameter was significantly reduced during the final intervals of the recovery period (REC 5 and REC 6).