Incremental and decremental cardiopulmonary exercise testing protocols produce similar maximum oxygen uptake in athletes

The aim of the study was to evaluate and compare the maximal oxygen uptake (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dot{\mathrm{V}}$$\end{document}V˙O2max) achieved during incremental and decremental protocols in highly trained athletes. Nineteen moderate trained runners and rowers completed, on separate days, (i) an initial incremental \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dot{\mathrm{V}}$$\end{document}V˙O2max test (INC) on a treadmill, followed by a verification phase (VER); (ii) a familiarization of a decremental test (DEC); (iii) a tailored DEC; (iv) a test with decremental and incremental phases (DEC-INC); (v) and a repeated incremental test (INCF). During each test \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dot{\mathrm{V}}$$\end{document}V˙O2, carbon dioxide production, ventilation, heart and breath rates and ratings of perceived exertion were measured. No differences were observed in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dot{\mathrm{V}}$$\end{document}V˙O2max between INC (61.3 ± 5.2 ml kg−1 min−1) and DEC (61.1 ± 5.1 ml kg−1 min−1; average difference of ~ 11.58 ml min−1; p = 0.831), between INC and DEC-INC (60.9 ± 5.3 ml kg−1 min−1; average difference of ~ 4.8 ml min−1; p = 0.942) or between INC and INCF (60.7 ± 4.4 ml kg−1 min−1; p = 0.394). \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dot{\mathrm{V}}$$\end{document}V˙O2max during VER (59.8 ± 5.1 ml kg−1 min−1) was 1.50 ± 2.20 ml kg−1 min−1 lower (~ 2.45%; p = 0.008) compared with values measured during INC. The typical error in the test-to-test changes for evaluating \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dot{\mathrm{V}}$$\end{document}V˙O2max over the five tests was 2.4 ml kg−1 min−1 (95% CI 1.4–3.4 ml kg−1 min−1). Decremental tests do not elicit higher \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dot{\mathrm{V}}$$\end{document}V˙O2max than incremental tests in trained runners and rowers, suggesting that a plateau in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\dot{\mathrm{V}}$$\end{document}V˙O2 during the classic incremental and verification tests represents the maximum ceiling of aerobic power.

The concept of maximal oxygen uptake ( V O 2max ) was introduced in 1924 after the works of the Nobel laureate Archibald Hill and his colleagues 1 . They observed a linear relationship between workload and oxygen uptake until the V O 2max was reached and proposed that the body has a limited capacity to uptake, transport and/or consumption the oxygen, called physiological ceiling for cardiorespiratory fitness 1 . Since then, V O 2max has been considered one of the most important indicators of endurance capacity 2 and its determination has become one of the most widely used test procedures in experimental and clinical exercise physiology 3 for testing the cardiorespiratory fitness and performance of athletes, the efficacy of training strategies or ergogenic aids and for quantifying the functional predations of chronic diseases 3,4 .
Because V O 2max is an important outcome for both physical performance and health status, test designs that increase the reliability and validity of V O 2max determination have widespread applicability. Since the 1970s, the maximal incremental exercise test has become a popular method for establishing V O 2max 5 . The V O 2max determination typically requires subjects to continue the incremental exercise test until they reach their limit of tolerance 6 . However, Bentley, et al. 7 draw attention to methodological factors that influence physiological parameters such as V O 2max in trained endurance athletes during the incremental exercise protocol. In this sense, the application of criteria to assess whether a 'true' V O 2max was achieved during an incremental test is still widely discussed by exercise physiologists.
The first and primary criterion traditionally used for establishing that a 'true' V O 2max has been reached is when there is no increase in V O 2 in response to an increase in work rate at the end of the incremental test: a plateau www.nature.com/scientificreports/ in V O 2max [8][9][10][11] . Unfortunately, the V O 2max plateau is not always identified in all test individuals 12,13 . Age, modality tested, data processing, physical fitness and test design can influence the incidence of plateaus in V O 2max 9,14,15 . In those instances where a V O plateau is not attained as definitive evidence of V O 2max , investigators commonly elect to substantiate that V O 2max was actually achieved by utilization of secondary criteria assumed to validate V O 2max -heart rate (HR) ≤ 5% of the age-predicted (220-age) maximum, blood lactate concentration ≥ 8 mM, or respiratory exchange ratio (RER) > 1.00, 1.10, or 1.15 16 . Utilization of these criteria can allow for a 30-40% underestimation of the 'true' V O 2max and/or an errant rejection of tests in which subjects had actually achieved their V O 2max 16 . To reduce these limitations, a supramaximal verification test, first proposed by Thoden 17 was used in this century to confirm the 'true' V O 2max 16,[18][19][20][21] . The supramaximal verification test involves a single squarewave bout of exercise performed shortly after the incremental test with a workload higher than the last completed stage of the maximal incremental test. The V O 2max results were compared between phases and consistent V O 2max values in the incremental and verification phases confirms that a 'true' V O 2max has been attained 10,[19][20][21][22] .
Beyond the methodological factors, there also remains considerable debate regarding the factors regulating or limiting V O 2max [23][24][25] . The classical model, based on the studies of Hill, et al. 1 and in accordance with the 'true' V O 2max achieved during incremental and verification tests, proposed that V O 2max is limited by the maximal cardiac output (cardiac limitation) 26 and the diffusional transport of oxygen out of the muscle microcirculation (muscle limitation) 27 . However, at the beginning of this century, Noakes and Marino 28 introduced another theory into this discussion, which states that the cardiovascular system never reaches a limit of work and that V O 2max is regulated, rather than limited, by the number of motor units recruited in the exercising limbs, which is always submaximal. This model proposes that the central nervous system (a central governor) controls the circulation during severe exercise and that there is always both cardiovascular and neuromuscular reserve upon exhaustion during incremental exercise 23 ; however, the brain stops the exercise to prevent catastrophic failure in the body system 28 .
Thus, the 'central governor' model, supported by the experimental design with decremental tests of Beltrami, et al. 29 suggests that the V O 2max achieved during incremental exercise is not the 'true' V O 2max . Beltrami, et al. 29 supported the decremental test design with: (i) an incremental test may cause more anticipatory stress, which may lead to a difference in blood flow response; (ii) a decremental test, with workload progressively easier, might relax brain controls directing the termination of exercise 28,30 ; (iii) evidence suggests that submaximal decremental protocols produce higher-than-expected V O 2 values compared with a similar power output during an incremental protocol 31,32 . In fact, Beltrami, et al. 29 showed that decremental protocols elicit significantly higher V O 2max values (~ 4%) than incremental protocol in subjects involved in running and cross-country skiing training. On the other hand, Taylor, et al. 33 , using similar decremental tests in running and triathlon training subjects, did not report differences in the V O 2max achieved during incremental or decremental tests, and both protocols elicited a similar cardiovascular response. These results showed that V O 2max determination is still challenging, not only from a methodological point of view but also from the factors limiting V O 2max , although it is widely used and well established. Considering the scarce literature on decremental protocols for testing the regulatory factors of V O 2max , the aim of this study was to evaluate and compare the V O 2max achieved during incremental and decremental protocols in trained athletes.

Methods
Participants. The study included 19 moderate trained men who competed in the junior category at a national or, in some cases, international level (15 distance runners and 4 rowers; age 17.4 ± 1.0 years, body mass 69.2 ± 6.3 kg, height 174.6 ± 4.6 cm and body fat 8.3 ± 1.6%) with at least 6 months of uninterrupted training. The rowers were former runners and they had running as part of their training routine. The participants were injuryfree and did not use any controlled drugs or nutritional supplementation during the experimental protocols. All procedures were approved by the university´s Ethics Committee (51127515.8.0000.5284) and were conducted in accordance with the Declaration of Helsinki. The subjects were kept informed of experimental procedures and risks and signed informed consent before participation in the study. The study followed the STROBE (STrengthening the Reporting of OBservational studies in Epidemiology). All trial sessions took place in one laboratory using the same equipment.
Experimental design. The participants visited the university laboratory on five occasions ( Fig. 1) to complete the study. The sessions were separated by 48-72 h between the first four visits and by 7-10 days between  www.nature.com/scientificreports/ the fourth and the fifth visit. All tests were performed using a motored treadmill (Super ATL, Inbrasport, Porto Alegre, RS, Brazil), a gas analyser (Metalyzer II, Cortex, Leipzig, Germany) that was calibrated before each test according to the manufacturer's instructions, a heart rate monitor (Polar S810i, Kemple, Finland) and a Borg perception effort scale 34 .
On the first visit, the participants performed an incremental test (INC) until volitional exhaustion, and after 15 min a verification phase test (VER) was performed. A familiarization decremental test (DEC) was performed on the second visit and a tailored DEC on the third visit. On the fourth visit, a test with decremental and incremental phases (DEC-INC) was applied. Finally, a repeated incremental test was applied on the fifth visit (INC F ).
The participants were instructed to avoid hard training sessions during the data collection period (15-20 days) and not to ingest caffeine or any kind of stimulant for 6 h before the tests. Tests were scheduled at the same time of the day (afternoon; between 3 and 5 p.m.) and laboratory conditions were stable for the duration of the study (ambient temperature 20-22 °C, relative air humidity 50-60%). All tests were performed on a motor-driven treadmill with a constant inclination of 5% (to ensure that participants do not exceed the capacity of the treadmill). All trial sessions were completed in 15-20 days by each participant. All tests were preceded by a 5-min warmup at 10 km h −1 with no treadmill inclination and participants were verbally encouraged throughout each test.

Incremental test (INC) and verification phase test (VER)
. INC started at 9 km h −1 and 5% grade and the speed was increased by 1 km h −1 every minute until volitional exhaustion. Fifteen minutes after the end of INC in the first session the participants performed VER to confirm the values of V O 2max . VER began at 10 km h −1 and 5% grade for 1 min and then the speed was increase to 1 km h −1 above the maximum speed reached in the previous INC. Participants were instructed to run at that speed for as long as they could. During the interval between INC and VER they were advised to rest or walk. The value of V O 2max was considered the highest after a mean of 20 s (with the data interpolated on a second by second basis and then averaged to retrieve the V O 2max value). INC was repeated on the fifth visit (INC F ) without VER in order to mislead the training effect on V O 2max during the experimental protocol.

Decremental test (DEC). DEC protocols (familiarization and tailored) were established as a function of
the result from the first session (INC + VER). The familiarization DEC was performed at the second visit. The first stage consisted of a graded increase in the speed of the treadmill until it reached that in VER (2 min at 9-10 km h −1 , 1.5 min at 12-13 km h −1 and 30 s at 15.5-16.5 km h −1 ). This graded increase in speed was included to diminish the gap in speed between the warm-up and the high-intensity start of the test. After reaching the VER speed, participants ran for 60% of the time that each had managed during VER (usually around 1 min). After this stage, speed was decreased by 1 km h −1 and maintained for 30 s. The speed in the following stages had consecutive decreases of 0.5 km h −1 that were maintained for 30, 45, 60, 90 and 120 s, respectively. The tailored DEC was performed at the third visit, the duration of the stages being determined by the individual reaction of the participant to the familiarization DEC. Some adjustments (increase or decrease in stage duration) were made to ensure that DEC was longer than 5 min. The value of V O 2max was considered the highest after a mean of 20 s. No physiological data were measured during the familiarization DEC. Instruments and data handling. All physiological data were collected within a fixed time of 10 s, exported from the analyser software into Excel spreadsheets. During INC, data were collected until the end of the final completed stage. Data collected during VER, DEC and DEC-INC were considered up to the final time (at least 30 s) the respective test had been collected. This was done to ensure that comparison between INC and VER was done at two different workloads, as required to define a plateau in V O 2 . The maximum test confirmation criteria for INC and INC F were: (a) the volitional exhaustion and a plateau in V O 2 ; (b) when the plateau in V O 2 was not achieved, the volitional exhaustion and one of the following criteria: heart rate (HR) ≤ 5% of the age-predicted (220-age) maximum or respiratory exchange ratio (RER) > 1.00. A plateau in V O 2 during INC was accepted if the change in V O 2 during the highest 30-s interval between the two final stages of the test was less than half of the normal stage-to-stage change in V O 2 during the initial (linear) parts of the tests for each subject 29 . The average stage-to-stage difference in V O 2 for all participants was calculated as 320 ± 66 ml min −1 , so the plateau phenomenon was defined as a change in V O 2 < 160 ± 33 ml min −1 (or an average of ± 2.2 ml kg −1 min −1 , considering the average body mass of the participants) between the two final stages of the test. The same criterion was used to define a plateau in V O 2 between the V O 2max values measured during INC and VER, because VER was performed at one stage higher than the maximal stage completed during INC 29 .

Decremental and incremental test (DEC-INC
Considering the stage when V O 2max was reached for each test, minute ventilation (VE), breathing rate (BR), respiratory exchange ratio (RER), heart rate (HR) and rating of perceived exertion (RPE) were evaluated.

Statistical analysis.
Results are expressed as the mean ± standard deviation (SD). The normality of the data was tested and confirmed by Shapiro-Wilk's test (p > 0.05), which allowed the use of parametric statistics.

Results
The plateau phenomenon in V O 2 was observed in 13 participants during INC, 7 participants during VER and 12 participants during INC F . When the INC and VER protocols were combined, 17 participants achieved the plateau phenomenon. Figure

Discussion
The results of the present study showed that, regardless of the test performed, the V O 2max achieved in each test was very similar and its difference was smaller than the value of the plateau phenomenon between the last two stages of INC or even the measurement error in test-to-test changes. That is, contrary to the study by Beltrami,et al. 29 no significant differences were found regarding the V O 2max values between INC and DEC. Moreover, the test proposed by us with decremental and incremental stages (DEC-INC) also showed no difference from INC. Thus, although more recently the theory of the plateau phenomenon has been criticized 37 , the results of the present study show that the plateau obtained at the time of exhaustion during a traditional test (i.e. INC) may infer the maximum ceiling of cardiorespiratory capacity and thus remain in exercise, as proposed by the classical theory 1,38,39 .
In an attempt to elucidate the validity of a cardiopulmonary test to assess maximal cardiorespiratory capacity, in addition to the plateau phenomenon, a series of validation criteria have been proposed in the literature, such as expected values of HR, RER and blood lactate concentrations 3 . In addition to those already cited, in order to improve the reliability of determining V O 2max , VER has been proposed 18 . This test requires the participant to perform the exercise for as long as possible at a constant intensity above that achieved in INC. Therefore, even with increased intensity, if the V O 2max value is not higher than that found during INC it is assumed that the maximum aerobic power ceiling was reached during both tests 10,40,41 . From this application, the frequency with which the plateau was identified increased, strengthening the theory that discusses this phenomenon 41 (when the INC and VER protocols were combined, 17 of the 19 participants achieved the plateau phenomenon. Evidence using VER as an INC validity criterion has controversial results. At some point VER will identify different responses to INC depending on a series of factors (participants, tests performed, equipment used, etc.). Murias, et al. 42 used VER (105%) to analyse the effectiveness of INC on a ramp format in young and old individuals. No significant differences were found between the two tests in the two populations, thus, although VER may be very seductive in finding different values, in that study it did not present any additional validation. Furthermore, Bhammar, et al. 43 when comparing obese and non-obese children, found that the V O 2max values in VER (105% of v V O 2max during INC) were higher by approximately 6% and 10%, respectively. Barker, et al. 44 also found that supramaximal testing at 105% of the power output achieved during ramp exercise did not increase the V O 2max achieved compared to the ramp test, thus suggesting the achievement of a true V O 2max during the initial ramp test for young people. However, the authors also concluded that the adherence to commonly used secondary criteria to validate a maximal effort (expected values of HR, RER and blood lactate concentration) in young people can result in either a submaximal V O 2max or a rejection of a participant's V O 2max score despite a plateau being evident. Finally, Beltrami, et al. 29 29 to break with the traditional tests for determination of aerobic power and to evaluate the validity of the traditional INC in determining the maximum ceiling of cardiorespiratory fitness. Beltrami,et al. 29 demonstrated that in DEC, their participants achieved significantly higher V O 2max values (4.4%) than those found when applying the traditional INC. Although the study was cited as being innovative as a function of protocol and approach, previous studies, even without the use of DEC as a tool for determining V O 2max , are essential for understanding the phenomenon behind the oxygen uptake response to this stimulus 31,33,[45][46][47] . The results of the present study show no differences between INC and DEC in the V O 2max achieved, or in the test developed by our research group, DEC-INC, which sought an increase in V O 2max with increasing oxygen uptake intensity after its decrease. We did not discard methodological differences between our study and the research of Beltrami, et al. 29 , mainly the difference in the gas analyzer and the difference between the procedures for collection and analysis of respiratory data. These results are in agreement with those obtained by Taylor, et al. 33 , who briefly compared the values of V O 2max , cardiac output and systolic volume obtained by means of INC and DEC in very well-trained triathletes and runners, a population similar to that of the present study and also of Beltrami, et al. 29 . Taylor, et al. 33 also found no significant differences between INC and DEC in maximal oxygen uptake values (57.29 ± 8.94 ml kg −1 min −1 vs. 60.82 ± 8.49 ml kg −1 min −1 , respectively) or in cardiocirculatory values. They also concluded that DEC was not capable of causing higher values of maximal oxygen uptake but may be an alternative for this population. We do not share the same view, as DEC exhibits the same possible obstacles as INC (i.e. the need to establish initial loads, the duration of stages and the decrease in intensity) and is an 'open-ended' test 48 . In addition, due to the need to achieve high intensities in order to decrease intensity later, the error in high intensity selection is much higher than the error in intensity selection for INC. Thus, we have not assumed DEC to be a good alternative for determining V O 2max , although it may be explored in the future.
Even with the emergence and acceptance of VER as a tool to try to validate the maximum exercise condition maintained by an individual, the debate about other factors that permeate the phenomenon of V O 2max remains intense. Among the most common theory is that the cardiovascular system limits V O 2max in a maximal or supramaximal situation 23,49 . According to Elliott, et al. 23 the evidence is controversial. For instance, some studies show that cardiac output stabilizes at maximal and supramaximal loads, whereas others found that cardiac output does not stabilize in athletes. Barker  We cannot fail to mentioned that the results for identifying or not the ceiling effect found by different studies can be partially explained by methodological differences (e.g.; chronological age, sex, training level, training status, type of exercise, sports modality, ergometer, criteria for interruption and confirmation of the maximum test, criterion for plateau, motivation strategy, exercise protocol, V O 2 sampling frequency, statistical analysis). Studies that used breath by breath analysis have a greater dispersion of V O 2 data, which makes it difficult to identify a plateau. On the contrary, data collected every 10 s and analyzed in an average of 30 s (as in our study) decrease the variability of V O 2 data, facilitating the identification of a plateau. On the other hand, there are few studies that make clear the criteria for interrupting the maximum test, as well as the criteria used to confirm the maximum test. This can really generate discrepant results and must be taken into consideration when comparisons are made.
According to Hill's classical theory (1924), running and rowing athletes reached the maximum ceiling of aerobic power by conventional testing (similar V O 2max in INC, VER, DEC, DEC-INC and INC F ), in contrast to other studies 23,29 , that the V O 2max reached during conventional tests was lower than unconventional decremental tests. It is important to highlight that these results were achieved by trained athletes, with the possibility of different results in untrained individuals, since the participation of the different mechanisms behind the onset of fatigue/exhaustion is different depending on the training state. According to this study, the verification phase after the incremental protocol for athletes it's not necessary; however, since not all athletes reach the V O 2max plateau during the incremental protocol, we recommend the verification phase when the excellence is sought (the chances that all athletes reach the plateau in V O 2max increases considerably when INC and VER are combined. Future research should aim to design a DEC protocol for non-athletic populations.
In summary, decremental tests do not elicit higher V O 2max than incremental tests in trained runners and rowers, suggesting that a plateau in V O 2 during the classic incremental and verification tests represents the maximum ceiling of aerobic power.

Data availability
The views expressed in this publication are those of the author(s) and not necessarily those of the funders. All authors had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. Data sharing statement Data are not publicly available, but applications for data sharing can be made. For enquiries, please contact NMFS (nunosfrade@gmail.com).