Abstract
the Study of Human Performance (CIPER), Faculdade de Motricidade Humana, Universidade de Lisboa, Cruz Quebrada-Dafundo, 1499-002 Lisboa, Portugal 11 Comprehensive Health Research Centre (CHRC), Universidade deÉvora, 7004-516Évora, Portugal 12 SPRINT Sport Physical Activity and Health Research & Innovation Center, Centro de Investigação e Inovação em Desporto Atividade Física e Saúde, 2001-904 Santarém, Portugal *Correspondence: mario.espada@ese.ips.pt Abstract:Critical speed (CS), the respiratory compensation point (RCP), and the midpoint between gas exchange threshold and maxial oxygen uptake (VO2max) (i.e., 50%∆) have been considered indexes able to demarcate the boundary between the heavy and severe exercise domains. However, the agreement between these indexes—and therefore the validity of using them reciprocally—remains to be reported in running. The current study analyzed the agreement between RCP,50%∆ , and CS. Twelve young runners performed an incremental test to assess VO2max, RCP, and 50%∆, with CS estimated by the linear model of time-limited trials at 90, 95, and 110% of the speed corresponding to VO2max. One-way ANOVA showed no differences when comparing VO2and running speed at CS vs. 50%∆vs. RCP (47.5±4.4 vs.46.6±4.4 vs.47.8±4.5 mLO2·kg −1 ·min −1 ; and13.9±1.3 vs.13.7±1.3 vs. 14.0±1.4 km·h −1 ;p> 0.05 for all comparisons). The bias for 50%∆vs. CS was−0.82±1.55 mLO2·kg −1 · min −1 and−0.23±0.55 km·h −1 , and for RCP vs. CS, it was 0.36±1.21 mLO2·kg −1 · min −1 and 0.05±0.46 km·h −1 . Therefore, the agreement between RCP, 50%∆, and CS in estimating VO2responses and running speed did not preclude their reciprocal similarity in exercise intensity, although the observed individual variability in physiological variables is a constraint on considering these indexes interchangeable. Keywords:critical speed; respiratory compensation point; running; youth Sports2025,13, 167 https://doi.org/10.3390/sports13060167
Sports2025,13, 167 2 of 14 1. Introduction To define the zones for exercise training, the profiles of physiological variables like oxygen uptake (VO2) and blood lactate concentration ([La − ]) during running or cycling have been used to account for the interplay of energy systems (aerobic vs. anaerobic activa- tion) and acid–base balance control (metabolites (H+, Pi) clearance ability) [1,2]. Basically, the transition from an evenly steady state to an uncontrolled rise toward maximal values of VO2and [La − ], passing through an intermediary state of uneven but still controlled responses, characterizes the range of exercise intensities across the moderate (i.e., highly tolerable), heavy (fairly tolerable), and severe (poorly tolerable) domains [1–3]. To demar- cate these domains, the responses of VO2, [La − ], pulmonary ventilation (VE), and gas exchange variables (like carbon dioxide production (VCO2), ventilatory equivalents for O2 and CO2, and pulmonary end-tidal O2and CO2) have been analyzed during incremental or constant-load exercises and applied to parametrize the transitions between physiological profiles in each exercise domain [4,5]. In this context, the onset of additional VCO2release through the buffering process of H+ (due to the increased demand on the glycolytic pathway) drives [La − ], VE, and VCO2to rise more sharply, marking the transition from the moderate to heavy domain. This transition has been indexed by the lactate threshold (LT), the gas exchange threshold (GET), and the first ventilatory threshold (VT1) [4,5]. The heavy domain marks the upper limit of exercise at which blood lactate concentration ([La − ]) and oxygen uptake (VO2) responses [2–4] remain high but stable, i.e., an isocapnic buffering region [5]. In this region, an exercise intensity can also be observed eliciting the maximal balance between the rate of metabolite efflux from the muscle and the rate of metabolite clearance from the blood (known as the maximal lactate steady state, MLSS) [2]. Above the heavy domain, the responses of [La − ] and VO2are projected to their respective maximal values [6,7]; thus, the transition from the heavy to severe domain is characterized by respiratory compensation for metabolic acidosis (i.e., VErises exponentially (hyperventilation) to reduce the
the rate of metabolite clearance from the blood (known as the maximal lactate steady state, MLSS) [2]. Above the heavy domain, the responses of [La − ] and VO2are projected to their respective maximal values [6,7]; thus, the transition from the heavy to severe domain is characterized by respiratory compensation for metabolic acidosis (i.e., VErises exponentially (hyperventilation) to reduce the excess of CO2in the blood), which has been indexed by the lactate turn point (LTP), respiratory compensation point (RCP), and second ventilatory threshold (VT2) [4,5]. In running, critical speed (CS) has been recognized and well supported as the exercise intensity demarcating the upper limit of the heavy domain [2,6–8]. Thus, below CS, there is a range of exercise intensities eliciting physiological [1,2,8] and metabolic [2,3,9] responses without inducing sufficient accumulation of metabolites to disturb the acid– base balance [9,10]. However, the determination of CS yields values subjected to protocol influence (i.e., time limit (tLim) between 2 and 15 min) [1–3] and mathematical adjustments (linear and non-linear equations) [2,11] and is often expensive and physically demanding for athletes [12]. In addition, other indexes able to parametrize the zone of heavy but still sustainable exercise, mainly those assessed during incremental exercise tests, have been compared to CS regarding their observed similarities [6,13]. For example, it has been proposed that the exercise intensity at half the difference (i.e.,∆50%) between maximal oxygen consumption (VO2max) and GET would also delimit the upper boundary of exercise intensities with physiological responses still characterizing the heavy domain [14]. Despite all normaliza- tion efforts, information from incremental exercise tests is dependent on the specificity of training status and movement pattern [15–18]. Lansley et al. [19] reported more consistent physiological responses during exercise when the level of intensity was prescribed at a given %∆rather than at %VO2max. Moreover, another study reported that the transition from the heavy to severe domain in running can be observed in a wide range between 40 and 60%∆[20]. However, typical physiological responses of the heavy domain were reported during cycling at 50%∆[14], whereas responses at 60, 70, and 80%∆were typical
at a given %∆rather than at %VO2max. Moreover, another study reported that the transition from the heavy to severe domain in running can be observed in a wide range between 40 and 60%∆[20]. However, typical physiological responses of the heavy domain were reported during cycling at 50%∆[14], whereas responses at 60, 70, and 80%∆were typical
Sports2025,13, 167 3 of 14 of the physiological profile in the severe domain [21]. Therefore, there is enough theoretical support for the assumption that 50%∆might be a reliable index of the transition between sustainable and exhaustive exercises [14]. As already mentioned, another index from the incremental exercise test able to charac- terize the heavy sustainable exercise intensity is RCP [22,23], which has shown similarity to CS [3,11]. The physiological significance of RCP corresponds to an exercise intensity beyond which the mechanisms for controlling the acid–base balance lose the capacity to buffer hydrogen anion production due to the increased demand for ATP resynthesis from anaerobic glycolytic metabolism [4,5,10,11,13]. Thus, the concept of RCP is best aligned with the assumption that it represents an exercise intensity around which the physiological response can transition from the heavy to severe domain [24,25]. Therefore, if CS is an index from a time-limited model able to evidence agreement with threshold indexes of incremental exercise, then CS can be a reliable and feasible index to support high-intensity training planning for young runners. It is notable that the insufficient number of investiga- tions to support the agreement or disagreement between CS and indexes from incremental exercise, particularly RCP and 50%∆, contrasts with the substantial evidence supporting their roles in demarcating exercise zones characterized by submaximal and sustainable physiological responses. In addition, previous studies did not refute that both RCP and CS are indexes eliciting comparable load/speed intensities or physiological responses [11,26], although the interchangeability in physiological responses is still being questioned [5,18]. Thus, the present study aimed to analyze metabolic (VO2) and speed (km·h −1 ) agreement at 50%∆and RCP with CS during treadmill running in trained young runners. The hypothesis was that these indexes would show statistical similarities in terms of oxidative demand. However, when comparing running speeds, differences were expected to reduce compatibility due to the influence of individual anthropometric characteristics on stride length and frequency and their association with running economy [17]. 2. Materials and Methods 2.1. Participants Twelve young male runners (15.7±1.8 years; 1.7±0.1 m; and 57.1±11.7 kg), regularly involved in a running training
similarities in terms of oxidative demand. However, when comparing running speeds, differences were expected to reduce compatibility due to the influence of individual anthropometric characteristics on stride length and frequency and their association with running economy [17]. 2. Materials and Methods 2.1. Participants Twelve young male runners (15.7±1.8 years; 1.7±0.1 m; and 57.1±11.7 kg), regularly involved in a running training program for at least two years, participated in this study. Assessments were conducted during the third microcycle of the base training period. The participants were selected after an initial screening of 20 young athletes to exclude those who were recently injured or sick, those experiencing pain or discomfort during the training, and those who were not endurance-trained runners. All participants obtained permission from their guardians and signed an informed consent form acknowledging the procedures. All research procedures were conducted following the Declaration of Helsinki and the ethical standards in sport and exercise science research [27], and were previously approved by the local University Ethics Committee (CAEE: 36936714.6.0000.5398). 2.2. Experimental Design Figure sessions: (1) a progressive ramp test to determine GET, RCP, 50%∆, and VO2max, along with their respective intensities (vGET, vRCP, v50%∆, and vVO2max, respectively); and (2) threeconstant-velocity exercise bouts to voluntary exhaustion at 90%, 95%, and 110% of vVO2max, which were performed for CS prediction. A 24 h interval was maintained between each exercise bout [8]. All tests were conducted on a motorized treadmill (HP/Cosmos Pulsar, Nussdorf-Traunstein, Germany) with a fixed 1.0% incline [26] in an indoor environment with controlled temperature (21–23 ◦ C) and air humidity (50–60%). Participants were instructed to avoid exhaustive training, refrain from drinking alcoholic
Sports2025,13, 167 4 of 14 and caffeinated beverages the day before the assessment, and arrive in a fed and hydrated state [7,8]. Figure 1.An example of estimating a participant’s physiological and speed variables in the study. Panel (A) depicts the maximal incremental test for determining GET, RCP, 50%∆, and VO 2max. Panel (B) illustrates the VO 2response during the predictive trials for CS. Panels (C,D) show the determination of CS using the v-tLim −1 and d-t Limmodels, respectively. 2.3. Maximal Incremental Test During the maximal incremental test (Figure, Panel A), the speed progressed by 1.0 km·h −1 ·min −1 , starting from 7.0 km·h −1 [28] until voluntary exhaustion. VO2was breath-by-breath sampled throughout the test (QuarkPFTergo, Cosmed, Rome, Italy). The O2and CO2concentration analysis system was calibrated before each test using ambient air and a gas with known O2and CO2concentrations, and the turbine analysis was calibrated using a three-liter syringe, according to the manufacturer’s recommendations. VO2values were smoothed by a three-second filter and averaged every six seconds. [La − ] was analyzed at rest and in the first minute after test completion, and its analysis was performed using the enzymatic method (YSL 2500STAT, Yellow Spring, CO, USA) from 25µL of arterial blood diluted in 50µL of 1% NaF solution [11]. VO2max was considered the highest value smoothed by a 30 s moving average. VO2max was determined as the lowest speed in the incremental test that elicited the maxi- mum VO2elevation [17]. GET was determined following Whipp’s [4] recommendations based on the responses of VE·VCO2 −1, VE·VO2 −1, PETCO2, and PETO2. It involved observ-
Sports2025,13, 167 5 of 14 ing an increase in the responses of VE·VO2 −1and PETO2without a change in the response of VE·VCO2 −1and PETCO2. Identification of the metabolic fatigue threshold at 50%∆was performed using VO2at GET and VO2max(50%∆= GET + [(VO2max− GET)×0.5]) [14,16]. The corresponding speed was determined by trend fitting between running speed (km·h −1 ) and its VO2during the incremental test. GET, RCP, 50%∆, and VO2maxdeterminations were conducted independently by three experienced researchers [29]. 2.4. Critical Speed (CS) Determination The predictive trials for CS (90%, 95%, and 110% of vVO2max) (Figure, Panel B) were randomly performed with a minimum interval of 24 h [8]. The value of tLimwas recorded in seconds and associated with the prediction speed using the speed vs. time- limited (v-tLim −1 ) model (Figure, Panel C) and the distance vs. time-limited (d-t Lim) model (Figure, Panel D) (Equations (1) and (2)) [ 12]. The selection of the CS value for each participant was based on the smallest standard error of the estimate (SEE) as the criterion [30]. (a) Speed vs. inverse of time to exhaustion (v-1·tLim −1 ) v=D ′ × 1 (t) +CS (1) (b) Total distance vs. time (D-tLim) d=D ′ ×t+CS (2) wherev= running speed;D ′ = amount of work from bioenergetic reserves;t= the time limit (tLim) of predictive trials; andCS= critical speed. The VO2corresponding to CS was determined using the trend relationship between VO2and running speed during the incremental test [7,30]. 2.5. Statistical Analysis The sample size was previously estimated using G*Power, considering a security level of 95% (Z1-α/2 = 1.960) and power of 85% (Z1-β= 1.036), as well as a high correlation level of 0.80 [31]. The estimated size was 10 participants, which was increased by 20% (N = 12) to avoid statistical underpower with participants withdrawing. Data were expressed as mean±SD with a 95% confidence interval (CI95%). Outliers and normality were checked using a 1.5·IQR range (interquartile range = Q3–Q1) and the Shapiro–Wilk test. Homogeneity, variance, and differences between 50%∆, RCP, and CS (VO2and km·h −1 ) analysis were examined using the Levene test and one-way ANOVA, supplemented
12) to avoid statistical underpower with participants withdrawing. Data were expressed as mean±SD with a 95% confidence interval (CI95%). Outliers and normality were checked using a 1.5·IQR range (interquartile range = Q3–Q1) and the Shapiro–Wilk test. Homogeneity, variance, and differences between 50%∆, RCP, and CS (VO2and km·h −1 ) analysis were examined using the Levene test and one-way ANOVA, supplemented by Fisher’s LSD test. The normality, independence, and homoscedasticity of the residuals for the regression analyses between 50%∆, RCP, and CS were assessed using the Shapiro–Wilk, Durbin– Watson, and Breusch–Pagan tests, respectively. Influential residuals or leverage points were identified through Cook’s distance. For agreement analysis, the difference between the means of variables was checked using the two-tailed Student’st-test for a single sample, and the proportional bias of differences was assessed through linear regression analysis between the mean (independent variable) and the difference (dependent variable) of the measurement. An analysis of dispersion was performed using the sample-adjusted coefficient of determination (R 2 adj.) and SEE, whilst agreement was analyzed with Bland– Altman plots [32] and the standard error of the mean (SEM = SD ÷ √ n). The effect size for ANOVA and R 2 adj. was calculated using eta squared (η 2) consid- ering threshold values as follows: <0.04 [trivial], 0.04–0.24 [weak], 0.25–0.63 [medium], and≥0.64 [strong] [33]. The correlations between the variables were assessed using the
Sports2025,13, 167 6 of 14 coefficient of correlation (r) derived from the coefficient of determination (R 2 ), with which the sample power was determined (Equation (3)) [31]. Z 1−β= √ n−3 1 2 Ln θ 1+r 1−r ι −Z 1− α 2 (3) whereZ 1−βprovides the coefficient for determining the sample power by the bicaudal normal distribution of the valuer[31]. The IBM SPSS Statistics software (version 27, 2021, IBM Co., Ltd., Armonk, NY, USA) was used to conduct statistical analyses. Statistical significance was determined atα< 0.05. 3. Results The value of VO2maxreached 53.0±5.1 (CI95%= 49.8–56.2) mLO2·kg −1 · min −1 , and the corresponding speed reached 16.1±1.7 (CI95%= 15.0–17.2) km·h −1 . The lower limit of the heavy domain was indexed by GET, corresponding to 76.2±4.2% VO2max (70.1±6.2% vVO2max ), and CS marked the upper limit at89.6±3.1% VO2max (86.5±3.6% vVO2max ). TheVO2responses during predictive trials for CS reached 53.0±4.9 (110% vVO2max),53.9±4.4 (95% vVO2max), and56.2±7.2 mLO2·kg −1 ·min −1 (90% vVO2max),providing greater prediction rigor for the v −1 · tLim −1 fit withR 2 = 0.98±0.03 (SEE = 2.89±1.69% and CI95%= 0.25–0.55 km·h −1 ). Figure speed values of the indexes used in the upper limit of the heavy domain. Another variable for sustainable exercise was RCP, which was situated at 90.2±2.8% VO2max (86.8±3.7% vVO2max). Both 50%∆(88.0±3.7% VO2maxand 85.1±3.1% vVO2max) and RCP showed no significant differences when compared to CS, respectively, regarding VO2(F [2,35]= 0.227, η 2= 0.014 [trivial],p= 0.65 andp= 0.84, respectively) and running speed (F [2,35]= 0.151, η 2 = 0.009[trivial],p= 0.68 andp= 0.92, respectively). Similarly, there was no difference in RCP compared to 50%∆for VO2(p= 0.51) and km·h −1 (p= 0.61), confirming the hypothesis regarding the statistical similarity between them. Figure 2.Individual values (circles), mean±SD (line and whiskers), and 95% confidence interval (gray box) of VO 2 − (brown) and running speed (light brown) associated with 50%∆, RCP, and CS. The regression and agreement analyses between 50%∆, RCP, and CS regarding VO2 and running speed are shown in Figures, respectively. The agreement analyses in rel- ative values of VO2and km·h −1 are shown in Figure. Both 50%
whiskers), and 95% confidence interval (gray box) of VO 2 − (brown) and running speed (light brown) associated with 50%∆, RCP, and CS. The regression and agreement analyses between 50%∆, RCP, and CS regarding VO2 and running speed are shown in Figures, respectively. The agreement analyses in rel- ative values of VO2and km·h −1 are shown in Figure. Both 50% ∆(F [1,10]= 72.13 ,p< 0.01) and RCP (F [1,10]= 128.6,p< 0.01) were able to account for the variance in VO2at CS with a potential of 86.6% [strong] and 92.1% [strong], respectively (FigureA,C), and a bias(t [11]=−1.830 ,p= 0.94 and F [1,11]= 0.003,p= 0.96; t [11]= 1.046,p= 0.32 and F [1,11]= 0.065,p= 0.80) of−0.82±1.55 and 0.36±1.21 mLO2·kg −1 · min −1 (SEM = 3.0 and2.4 mLO2·kg −1 ·min −1 or 6.9 and 4.9%, respectively, FigureB,D) or −1.74±3.54
Description
This study analyzes the agreement between physiological indexes in running speed at different fatigue thresholds in young runners.