IJCS | Volume 32, Nº4, July/August 2019

377 Belli et al. Speed and grade increment during cardiopulmonary Int J Cardiovasc Sci. 2019;32(4):374-383 Original Article the subjects underwent 10 min of walking-running in order to reach the target HR, followed by 30 min running at target HR. The treadmill was maintained at a constant level during the submaximal exercise sessions and speedwas adjusted tomaintain a stable HR. Rather than applying a constant workload, we chose to maintain a stable HR because the individuals exercised at intensities above the VT1, where the steady state is not established with a constant work rate. 13,28 Moreover, the target HR is frequently used to monitor aerobic training exercise sessions in practice. 4 During the submaximal exercise sessions, HR and gas exchange responses were continuously monitored as described above. Perceived exertion and blood lactate samples were obtained at rest and every 10 min. Blood lactate analysis Twenty-five µL fingertip blood samples were mixed with 50µL of 1% sodium fluoride. This solution was then frozen for later analysis of blood lactate concentration using a dedicated analyzer (YSI 1500-L Sport, Yellow Springs, Ohio, USA). Statistical analysis Based on a previous study, 29 a minimum sample size was estimated to be 7 subjects, using a power of 90% and an alpha of 0.05 to detect a 10% difference (16 bpm) in the prescribed HR between protocols. Two subjects were added to the sample to account for dropouts. We used the Kolmogorov-Smirnov test to assess the normality of variables. Descriptive data are presented asmean (M) and standard deviations (SD) and 95% confidence interval (95% CI). To evaluate intra-observer reproducibility in the detection of VT 1 , VT 2 , and VO 2 max, gas exchange curves of all tests were reviewed by the same blinded investigator twice within a one-week interval. Paired t tests, Pearson’s correlation coefficients, and Bland- Altman analyses 30 were used to assess intra-observer reproducibility. The responses to incremental and submaximal exercise tests were compared using paired t tests for two means and by generalized estimating equations for three or more means. When appropriate, multiple comparisons were evaluated using Bonferroni correction. The calculations for sample size, planning for randomization and Bland-Altman analyses were performed using R 3.0 (Free Software Foundation’s GNU Project), and all other analyses were performed using SPSS 21.0 software (IBM, New York, USA). Results All subjects completed the incremental exercise tests and the exercise sessions without complications. For the intra-observer reproducibility analysis, there were no significant differences in VO 2 (ml.kg -1 .min -1 ) between the two evaluations in terms of detection of VT 1 (“speed”: 27.5 ± 7.4 [95%CI = 21.8; 33.2] vs 26.4 ± 7.2 [95%CI = 20.9; 31.9], p = 0.11; “grade”: 24.9 ± 6.1 [95% CI = 20.2; 29.6] vs 25.9 ± 6.2 [95% CI = 21.1; 30.7], p = 0.16); VT 2 (“speed”: 32.4 ± 9.2 [95% CI = 25.3; 39.5] vs 32.5 ± 9.2 [95% CI = 25.4; 39.6], p = 0.85; “grade”: 29.7 ± 6.5 [95% CI = 24.6; 34.7] vs 30.9 ± 6.9 [95% CI = 25.6; 36.2], p = 0.08), and VO 2 max (“speed’: 35.8 ± 10.8 [95%CI = 27.5; 44.2] vs 35.8 ± 10.8 [95% CI = 27.2; 44.2], p = 0.99; “grade”: 34.7 ± 9.7 [95% CI = 27.2; 42.1] vs 34.7 ± 9.8 [95% CI = 27.2; 42.1], p = 0.99). Intra-observer agreement results are presented in Figure 2. The two protocols exhibited strong correlation coefficients between the first and second evaluations, varying from 0.95 to 1.00 (1 st and 3 rd columns of Figure 2). Likewise, Bland-Altman plots demonstrated values within the acceptable limits of agreement between the first and second evaluations (2 nd and 4 th columns of Figure 2). Table 1 shows the incremental exercise test results performed according to the speed and grade protocol. Resting HR and blood lactate were similar between the protocols. At peak exercise, HR was higher with the speed protocol, while no differences were observed between VO 2 max, VCO 2 max, and VEmax. Peak respiratory exchange ratio was lower with the speed protocol, as well as perceived leg exertion. Maximal blood lactate concentrations and time to maximal blood lactate concentration during recovery were similar with the two protocols. Figure 3 shows VO 2 and HR responses to incremental exercise according to VT 1 , VT 2 , the mean point between VT 1 and VT 2 , and VO 2 max. VO 2 at these 4 intensities was higher with the speed protocol (Figure 3A). When expressed as a percentage of VO 2 max, these differences were also statistically significant (Intensity: P < 0.01; Protocol: P = 0.01; Interaction: P = 0.18). The “speed” protocol resulted in higher HR (bpm) for the VT1 (153 ± 14 [95% CI = 146; 168] vs 144 ± 8 [95% CI = 137; 150], p < 0.01), the VT 2 (176 ± 7 [95% CI = 171; 182] vs 165 ± 8 [95%CI = 159; 171], p < 0.01), themean point between VT 1 and VT 2 (169 ± 9 [95% CI = 162; 176] vs 156 ± 8 [95% CI = 150; 162], p < 0.01), and peak exercise (189 ± 8 [95% CI =