Thirty-two elderly volunteers, of both genders, completed the study. No participants participated in sports or physical training on a regular basis prior to the interventions. No participants showed lesions or disease-associated conditions prior to the study. All signed informed consent forms in agreement with the Brazilian Ethics Committee that had been approved by the National Health Council under the number of CAAE 44907715.2.0000.5653 on 27/07/2015 and followed all ethical and legal requirements. We followed all conventions of the Helsinki Declaration. Baseline volunteer characteristics are displayed in Table 1.

Body weights and heights were obtained using a mechanical scale (Filizolla, Brazil) with a precision of 0.1 kg. On the day of collection, the subjects were instructed to wear light clothing and to repeat the same clothing at all times of evaluation. They were placed on top of the scale barefoot, with their arms at their sides, with their eyes fixed on the horizon and standing still until the measurements could be confirmed as previously described protocols [10, 17–19]. All measurements were performed in the morning.

(1)

BMI Equation.

Four clinical consultations were performed with the Physician who attended to the patients, maintaining their monthly routines. These consultations occurred immediately before the intervention period, 4^th, 8^th, and at the end of twelve weeks of intervention. In these consultations, routine monthly clinical investigations were performed, including blood pressure, heart rate, cardiac auscultation, complete history, orthopedic investigations and other medical investigations. These procedures were performed by the same Physician.

These consultations served as support for the control of health conditions and for reporting adverse clinical events that might be related to the intervention. Clinical, orthopedic and cardiological investigations were performed to determine possible malignant events during and after the intervention period. All consultations were performed in a blind scheme where the Physician does not know who performed or not the exercise to avoid tendencies of the outcomes in the clinical exam.

All tests of this section were selected due to three aspects. First security to the application to elderly subjects, second, the convenience, facility and reproducibility and their application, and three, because all these tests showed an adequate precision and confidence, which were displayed in different previous manuscripts [ 20–23].

In order to evaluate muscular endurance, three tests were performed: one for upper limbs, one for the abdomen and one for lower limbs. The test directed for the upper limbs was arm extension with floor support (Push-Ups). The volunteer was placed in a prone position, with his or her hands resting on the ground, with a distance of 10 to 20 cm from the shoulder line, with the toes facing forward. The positioning of the hands on the ground was not above the shoulder line and, in the initial position of the movement, the subject’s face allowed for an adequate alignment between the trunk and the legs. The application of the test to females was modified only by the support of their knees on the ground. The other procedures were performed for both genders. In order to determine performance, the maximum number of correct repetitions in one minute was recorded [ 21 ].

Regarding abdominal resistance, after placing the subject in a supine position on the ground, the following exercise was performed: flexion of the trunk, hip and knees, embracing the legs with the arms; in order to determine performance, the maximum number of repetitions in one minute was recorded [ 22,23].

The execution of the jump, an eccentric phase, was carried out before it began. After the jump, the subject was instructed to remain with his or her foot at the drop site upon landing. The distance from the most posterior point of the heel to the first marking was measured with the millimeter measuring tape and considered to be the distance obtained in the jump. The jumps were performed three times and the best jump with each limb was used for statistical purposes in a horizontal flat surface [ 20 ]. They could use their upper limbs (MMSS) to aid the effort.

To determine resistance of the lower limbs, an isometric test was selected. The subjects were placed with their backs supported on the wall. Isometric hip, knee and ankle flexion were performed at 90º with the aid of a goniometer. The time was recorded for performance effect. Subjects who were able to maintain this position for longer than one minute performed the test with only one lower limb after a 10-minute recovery period; during this second test, the extra limb was crossed on top of the limb that was resting on the ground, and soon thereafter the subject switched sides [ 20 ].

To determine their respiratory capacity, the 6-minute walk test (TC6M) was performed outdoors. The subjects walked at their own pace and alone, as far as possible during the six minutes; knowing the possible cardiorespiratory changes that might arise, they were allowed to walk slowly, stop, relax when necessary, and begin walking again when they felt ready to resume the exercise. Each individual was encouraged to wear light clothing and not to exercise prior to the test. The following equations determine the level of walking distance predicted for each test performed on the patient:

Distance TC6M (m) = (7.57 x height cm) - (5.02 x age) - (1.76 x weight Kg) – 309 m

Distance TC6M (m) = (2.11 x height cm) - (2.29 x weight kg) - (5.78 x age) + 667.6

Equation 1: Equation used to calculate maximum VO₂.

It should be emphasized that the anthropometric and performance evaluations were only performed 60 days after the beginning of the intervention, because the literature is categorical in stating that 30 days are insufficient for anthropometric and performance measurements to have changed enough to show differences, even in sedentary people.

A high-intensity interval training program was applied three days a week, with an average duration of 40 minutes per meeting. All exercises were performed without equipment, using the subject’s body weight alone. Each exercise set was done against the clock at a ratio of 1:2, between activity and rest. The exercise intensity was set up at a minimum of 80%, or more, during the stimuli and twice the time to passive recovery to cardiac frequency decrease. During the recovery, the heart rate reached low levels of about 55% of the max cardiac rate. The control of the cardiac frequency was performed by utilization of the FTI model Polar heart monitor with an observation of the heart frequency immediately after the stop of the challenge for all subjects.

Daily, a standardized 5-minute warm-up with stretching exercises and calisthenics was performed for 5 minutes. Subsequently, the main work out was performed, consisting of forward-backward running, squats, push-ups, sit-ups, jumping jacks, walking lunges, dips, six-point support planks, pulse lunges, reverse lunges, skipping, mountain climbers, arm and leg raises, side steps, direction-changing footwork and climbing/descending steps (25 cm).

Finally, a 5-minute cooldown was carried out to promote calmness. Each exercise was performed for 30, 45 or 60 seconds for recovery, always 2x greater than the stimulus time. In addition, all classes had approximately 10 minutes of stimulation. During all the moments of rest between one stimulus and another, the subject was free, yet encouraged to remain quietly awaiting the next stimulus in order to characterize a passive rest.

The descriptive data were expressed as mean ± standard deviation. To determine the normality of the data, the Shapiro–Wilks test was performed, which determinate that the data are non-parametric type. Then, the Kruskal–Wallis with Dunn’s posterior test was used to determine the difference between the days of experiments. Additionally, the effect size was calculated by Choen’s D test, indicating that when the calculated differences (d ² ) are between 0.01 and 0.14, the effect is defined as little, from 0.15 to 0.34 defined as a medium effect, and 0.35 more, the effect was considered large. All hypothesis tests had a significance of 5%. The entire statistical procedure was performed using the Prism Stat 5.0 program.

Fig. (1). Effect of High-Intensity Intermittent Training on the physical performance of adults. Elderly subjects were divided into a Control Group (CG, n = 16) and an Experimental Group (EG, n = 16) that underwent twelve weeks of high-intensity intermittent training. Before (P1), and after (P2) the interventions, the endurance of the Upper Limbs (RML MMSS) (Fig. A), the endurance of Lower Limbs (RML MMII) (Fig. B), the endurance of Abdominal muscle (RML ABD) (Fig. C), Lower Limb Strength (LLS) (Fig. D), and Maximal Aerobic Power (Fig. E). Kruskal–Wallis, with Dunn’s posterior test, was used with a significance of 5% to indicate the possible differences between all moments of data acquisition. Finally, the effect size between the variables was calculated by Choen’s d test. (A, B, C, D and E **** = p >0.0001 EG-P2 vs CG-P1, CG-P2 and EG-P1; ns= non-significant).

Four clinical consultations were performed with the Physician immediately before the intervention period, four, eight, and at the end of the intervention to serve as support for the control of health conditions and for reporting orthopedic and cardiological adverse clinical events. All investigations in all data points showed no events reported.

With respect to the localized muscular endurance of the upper limbs (Fig. 1A), for the CG at baseline (P1), the value was 15.75 ± 2.43 and at P2, the value was 15.5 ± 2.67 (p >0.05). For the EG at P1, the value was 17.88 ± 1.88 and at P2, it was 27.75 ± 2.65 (p <0.0001) for all comparisons. Cohen’s d showed a medium-size effect (d²= 0.33).

With respect to the localized muscular endurance of the lower limbs (Fig. 1B), for the CG at baseline (P1), the value was 165.1 ± 2.16 and at P2, it was 164.8 ± 2.37 (p >0.05). For the EG at P1, the value was 168 ± 2.33 and at P2, it was 174 ± 2.72 (p <0.0001) for all comparisons. Cohen’s d showed a medium-size effect (d²= 0.32).

With respect to the localized muscular endurance of the abdominal muscle (Fig. 1C), for the CG at baseline (P1), the value was 75.88 ± 1.55 and at P2, it was 75.87 ± 1.12 (p> 0.05). For the EG at P1, the value was 76 ± 1.19 and at P2, it was 79.13 ± 0.99 (p <0.0001) for all comparisons. The Cohen’s d showed medium-size effect (d²= 0.27)

In terms of lower limb muscle power (Fig. 1D), for the CG at baseline (P1), the value was 162.3 ± 1.03 and at P2, it was 162.3 ± 0.7 (p >0.05). For the EG at P1, the value was 162.6 ± 1.18 and at P2, it was 167.8 ± 1.44 (p <0.0001) for all comparisons. The Cohen’s d showed medium-size effect (d²= 0.29)

In terms of aerobic power (Fig. 1E), for the CG at baseline (P1), the value was 23.68 ± 1.54 and at P2, it was 23.85 ± 1.61 mL/Kg/min (p >0.05). For the EG at P1, the value was 24.01 ± 1.39, and at P2, it was 32.91 ± 1.91 mL/Kg/min (p <0.0001) for all comparisons. The Cohen’s d showed large size effect (d²= 0.53)