Long maximal incremental tests accurately assess aerobic fitness in class II and III obese men.

This study aimed to compare two different maximal incremental tests with different time durations [a maximal incremental ramp test with a short time duration (8-12 min) (STest) and a maximal incremental test with a longer time duration (20-25 min) (LTest)] to investigate whether an LTest accurately assesses aerobic fitness in class II and III obese men. Twenty obese men (BMI≥35 kg.m-2) without secondary pathologies (mean±SE; 36.7±1.9 yr; 41.8±0.7 kg*m-2) completed an STest (warm-up: 40 W; increment: 20 W*min-1) and an LTest [warm-up: 20% of the peak power output (PPO) reached during the STest; increment: 10% PPO every 5 min until 70% PPO was reached or until the respiratory exchange ratio reached 1.0, followed by 15 W.min-1 until exhaustion] on a cycle-ergometer to assess the peak oxygen uptake [Formula: see text] and peak heart rate (HRpeak) of each test. There were no significant differences in [Formula: see text] (STest: 3.1±0.1 L*min-1; LTest: 3.0±0.1 L*min-1) and HRpeak (STest: 174±4 bpm; LTest: 173±4 bpm) between the two tests. Bland-Altman plot analyses showed good agreement and Pearson product-moment and intra-class correlation coefficients showed a strong correlation between [Formula: see text] (r=0.81 for both; p≤0.001) and HRpeak (r=0.95 for both; p≤0.001) during both tests. [Formula: see text] and HRpeak assessments were not compromised by test duration in class II and III obese men. Therefore, we suggest that the LTest is a feasible test that accurately assesses aerobic fitness and may allow for the exercise intensity prescription and individualization that will lead to improved therapeutic approaches in treating obesity and severe obesity.


Introduction
Maximal incremental exercise testing is commonly used in exercise physiology to determine physiological variables, such as peak oxygen uptake ( _ V O 2peak ), peak heart rate (HR peak ) and peak power output (PPO), together with other submaximal metabolic parameters [i.e., lactate (LT) and ventilatory thresholds (VT)]. The accuracy of the determination of these variables during a maximal incremental exercise test is important for exercise prescription and individualization in athletes, sedentary healthy individuals and patients [1]. It has been suggested that a maximal incremental exercise test should last between 8 and 12 minutes with short stage duration (1 min) to elicit _ V O 2peak [2][3][4], whereas longer protocols (~25 min) with long stage duration (3-5 min) report significantly lower _ V O 2peak [5][6][7] and PPO [8,9] and higher HR peak [5,9,10]. However, some studies report no significant differences in _ V O 2peak and HR peak [8,[11][12][13], or in _ V O 2 and HR at VT 1 [7,11], between short and long maximal incremental tests with different stage and time durations, suggesting that both exercises may have a practical relevance and may be useful in exercise intensity prescription and individualization in healthy men [14].
In class II and III obese individuals, exercise intensity prescription and individualization are strongly recommended as part of each patient's multidisciplinary medical and surgical management in order to improve the poor aerobic fitness [15] and thus decrease the mortality risk in this population [16]. However, there are limited indications regarding which test is the most appropriate for the evaluation of aerobic fitness and the subsequent prescription of exercise training programs in class II and III obese individuals [15]. Severe obesity is also specifically characterized by a depressed capacity to oxidize lipids [17], which does not always occur at lower levels of obesity [18]. This decreased fat oxidation may be involved and contribute to the development of insulin resistance in severely obese individuals [17].
Endurance training targeting an exercise intensity (Fat max ) that elicits the maximal fat oxidation (MFO) is appropriate in order to enhance fat oxidation rates and insulin sensitivity in obese individuals [19], highlighting the importance of correctly assessing Fat max . However, this is not possible with an incremental test with short stage duration, which is characterized by a non steady-state condition, but only with an incremental exercise test with longer stage duration (i.e., 5-6 min) during which steady state is reached for each step. Therefore, an incremental exercise test with longer stage duration may be an appropriate test to determine fat oxidation kinetics, MFO and Fat max (metabolic fitness [20]) in obese and severely obese individuals [21]. Although it has previously been suggested that long test duration may affect _ V O 2peak assessment by reaching the limit of exercise tolerance earlier [3], this test has already been used to assess aerobic and metabolic fitness in class I and II obese individuals [22,23]. In these studies, _ V O 2peak seems to be correctly assessed because Fat max (expressed in % _ V O 2peak ) has been found at similar values than those previously reported in obese subjects [19,21,24]. However, Ara et al. [22] and Larsen et al. [23] did not compare their maximal incremental long tests to a maximal incremental short test in order to attest whether long test elicits valid _ V O 2peak . Therefore, this study aimed to compare two maximal incremental tests with different time durations [a maximal incremental ramp test with a short time duration (8-12 min) (S Test ) and a maximal incremental test with a longer time duration (20-25 min) (L Test )] in a group of class II and III obese men. It was hypothesized that the L Test would elicit similar _ V O 2peak , HR peak , _ V O 2 and HR at VT 1 compared to the S Test , suggesting that the L Test is an appropriate test to evaluate aerobic fitness. Moreover, this single test may also lead to simultaneously determine metabolic fitness (i.e., fat oxidation kinetics, MFO and Fat max ) in order to obtain a more complete assessment of physical fitness in class II and III obese men and may aid in the exercise intensity prescription and individualization in this population.

Participants
Twenty obese men [body mass index (BMI)!35 kg . m -2 ] without secondary pathologies were recruited to participate in this study (Table 1). Subjects were recruited from the Istituto Auxologico Italiano (Piancavallo, Italy). Subjects with hypertension [blood pressure (BP)>130/90 mmHg], impaired fasting glucose (>6.1 mmol . L -1 ) [25], type 2 diabetes and an abnormal electrocardiogram at rest were excluded. The study was approved by the Ethics Review Committee of the Istituto Auxologico Italiano, Italy. All subjects provided written, voluntary, informed consent before participating. The experiment was conducted according to the Declaration of Helsinki.

Experimental protocol
Subjects performed two maximal incremental tests to exhaustion on a cycle-ergometer (Ebike Basic BPlus, USA) to determine _ V O 2peak , HR peak , peak ventilation ( _ V Epeak ), peak respiratory exchange ratio (RER peak ), PPO and VT 1 ( _ V O 2 , HR and PO) during each of the following tests: 1) a maximal incremental ramp test with a short time duration (8-12 min) (S Test ) in the first session, and 2) a maximal incremental test with a longer time duration (20-25 min) (L Test ) in the second session. This order was fixed because S Test was necessary to individualise the warm-up and increments of the L Test [21].
Maximal incremental ramp test with a short time duration (8-12 min) (S Test ). The S Test was performed at least 2-3 h following the consumption of the last meal. After a 3-min rest period, subjects started with a 5-min warm-up at 40 W, after which the PO was linearly increased by 20 W every minute until exhaustion, which was determined by the inability to maintain a minimum pedalling frequency (i.e., 60 revolutions per min) despite verbal encouragement. This test was used previously [21] and yielded an exercise duration of approximately 10 min.
Maximal incremental test with a longer time duration (20-25 min) (L Test ). The L Test was performed in the morning after a minimum of two days following the S Test . This test was performed in fasted state in order to determine the substrate oxidation. After a standardized 10-min warm-up at 20% PPO reached during S Test , the PO was increased by 10% PPO every 5 min until reaching 70% PPO, or until RER reached 1.0 (adapted from Lanzi et al. [21]). At this point, PO was increased by 15 W every minute until exhaustion as previously defined. From our previous data of a submaximal incremental test with 6 min stage duration [21], we determined that between the fourth and the fifth minute of each stage a steady-state condition was already reached in this population, therefore a protocol with 5 min stage was used to determine substrate oxidation and reduce test duration.

Data analysis and calculation
Gas exchange. _ V O 2 , carbon dioxide production ( _ V CO 2 ) and _ V E were measured continuously using a breath-by-breath online system (V max 229, Sensor Medics, USA). _ V O 2peak , _ V Epeak and RER peak were defined as the highest 10-s mean values recorded before the subject's volitional termination of each test.
Peak heart rate and peak power output. HR was recorded continuously using an HR monitor (Polar RS800, Finland). HR peak and PPO were defined as the highest peak values reached during each test.
Ventilatory threshold 1 and delta efficiency. VT 1 ( _ V O 2 , HR and PO) was determined during each test as described in the literature using Wasserman's ventilatory method [26]. This method consists of visually determining the point at which the _ The estimate of VT 1 was supported using the Beaver ventilatory method [27]. This method consists of visually determining the inflection point of _ V CO 2 with respect to _ V O 2 . Two blinded and independent investigators determined VT 1 . Delta efficiency (DE) was calculated as previously described [28].
Exercise intensity (Fat max ) eliciting maximal fat oxidation. To determine if the L Test is an accurate test to define Fat max and to compare these results to previous findings, Fat max was determined using the SIN model [29], as previously described in this population [21].

Statistical analysis
Data are expressed as means±SE for all variables. Normal distribution of the variables was assessed using the Kolmogorov-Smirnoff test. Paired t-tests were used to compare peak and submaximal values between the two different maximal incremental exercise tests. To compare the agreement of the obtained peaks and VT 1 values between the two different maximal incremental exercise tests, Bland-Altman plots were used [30]. The constructed graphs displayed scatter diagrams of the differences plotted against the mean of two measurements. The biases estimated from the mean differences (m À ) were calculated, and 95% limits of agreement were estimated by m À ±1.96 SD. To compare the agreement of the obtained peaks and VT 1 values, we also assessed Pearson product-moment correlation and intra-class correlation (ICC) coefficients. The level of significance was set at p 0.05.

Characteristics of the tests
The duration of the L Test was significantly longer (~2.6-fold) than the S Test (23.2±0.5 and 8.8±0.3 min, respectively; p 0.001). During the L Test , the mean warm-up load was 42±1 W, and the mean increment of the 5-min stage was 21±1 W.
Peak exercise values _ V O 2peak , HR peak and _ V Epeak were similar between the L Test and S Test (Table 2). By contrast, RER peak and PPO were significantly lower in the L Test than in the S Test (Table 2). There was a strong correlation between _ V O 2peak (r = 0.81, p 0.001; Fig 1A), HR peak (r = 0.95, p 0.001; Fig 1C) and _ V Epeak (r = 0.67, p = 0.001; data not shown), as determined by the L Test and S Test , Values are the means SE. _ V _ O 2peak : peak oxygen uptake; HR peak : peak heart rate; _ V _ Epeak : peak ventilation; RER peak : peak respiratory exchange ratio; PPO: peak power output; NS: non significant.  and these data were close to the line of identity. RER peak (r = 0.72, p 0.001; data not shown) and PPO (r = 0.89, p 0.001; Fig 1E) were also strongly correlated between the L Test and S Test , although there was a systematic underestimation in the L Test (i.e., data did not fit with the line of identity). These analyses were also confirmed by Bland-Altman plots (Fig 1B, 1D and 1F) and ICC analyses (Table 3). Biases and 95% limits of agreement for peak values between the L Test and S Test are shown in Table 3.

Ventilatory threshold and delta efficiency values
was similar between the L Test and S Test (Table 2). By contrast, HR VT1 and PO VT1 were significantly lower in the L Test than in the S Test (Table 2). There was a strong correlation between the _ V O 2VT1 (r = 0.72, p 0.001 ; Fig 2A), as determined by the L Test and S Test , and these data were close to the line of identity. HR VT1 (r = 0.67, p = 0.001; Fig 2C) and PO VT1 (r = 0.73, p 0.001; Fig 2E) were strongly correlated between the L Test and S Test , although there was an underestimation in the L Test (i.e., data did not fit with the line of identity). These analyses were also confirmed by Bland-Altman plots (Fig 2B, 2D and 2F) and ICC analyses (Table 3). Biases and 95% limits of agreement for VT 1 values between the L Test and S Test are shown in Table 3. DE was lower during L Test than during S Test (17.8±0.5 and 22.5±0.5%, respectively; p 0.001).

Exercise intensity eliciting maximal fat oxidation
The Fat max during the L Test was found at 50.6±1.9% _ V O 2peak .

Discussion
The results of this study showed that _ V O 2peak , HR peak and _ V O 2VT1 assessments were not compromised by prolonged stage and test duration, suggesting that the L Test is an appropriate test for evaluating aerobic fitness and may be used for prescribing an exercise training regimen in class II and III obese men. There was, however, a significant influence exerted by time duration on PPO, HR and PO at VT 1 . Table 3. Intra-class correlation, biases and 95% limit of agreement of the peak and ventilatory threshold 1 (VT 1 ) values between the maximal incremental test with short (S Test ) and long (L Test ) time duration. Values of bias are the means ± interval confidence (IC). Biases and 95% limits of agreements were estimated with Bland-Altman method. _ V _ O 2peak : peak oxygen uptake; HR peak : peak heart rate; _ V _ Epeak : peak ventilation; RER peak : peak respiratory exchange ratio; PPO: peak power output. * p 0.05 for significant ICC coefficient.

Oxygen uptake
The data of the present investigation show that _ V O 2peak was statistically similar and showed good agreement between the L Test and S Test (correlation coefficients and Bland-Altman plot analyses). These results are in line with previous studies, which reported a similar _ V O 2peak between short (~10 min) and long (~25 min) maximal incremental tests with different stage and time durations [8,9,11] in healthy normal-weight individuals, suggesting that the dogmatic view that maximal incremental tests should last between 8 and 12 min to elicit _ V O 2peak [2][3][4] should be reconsidered [14]. Additionally, _ V O 2VT1 also showed good agreement with respect to the L Test and S Test , and these results are in line with previous studies that showed that _ V O 2VT1 was independent of exercise test duration [3,7]. However, our results contrast with previous studies that reported different _ V O 2peak between short and long maximal incremental tests [3,4,7]. The reason for this discrepancy is unclear but may be due to different factors, such as different exercise test protocols (e.g., stage vs. ramp increments). Furthermore, previous studies compared normal and highly trained subjects, whereas this is the first study comparing individuals with a high degree of obesity (BMI ! 35 kg . m -2 ). The Bland-Altman plot analysis of _ V O 2peak was similar to previous studies, which reported a mean bias of 0.1 L . min -1 [8,9], with 95% limits of agreement between 0.4 and -0.6 L . min -1 [8] (which was considered good agreement) between short and long maximal incremental tests with different stage and time durations in well trained triathletes. However, for some individuals (n = 3), the difference in _ V O 2peak between the S Test and L Test was greater (0.41, 0.53 and 0.57 L . min -1 ) than the mean bias (Fig 1A and 1B). This result suggests that these subjects presented with consistently lower _ V O 2peak during the L Test compared to the S Test . Interestingly, these three individuals completed only one or two stages of 15 W . min -1 increments after having completed four stages of 5 min (i.e., 30,40,50 and 60% PPO reached during the S Test ), whereas the other subjects completed up to five 5-min steps (until 70% PPO) or as many as five 1-min steps. It is therefore possible that a premature fatigue state of some subjects may explain the lower _ V O 2peak obtained during the L Test [8], suggesting that envisaging a 5-min rest before starting increments of 15 W . min -1 during the L Test may be a reasonable approach of eliciting _ V O 2peak , as previously described [22].
Heart rate HR peak was also statistically similar and showed very good agreement between the L Test and S Test . Although some studies reported higher HR peak during prolonged incremental exercise tests [5,9,10] (most likely linked to higher body temperatures or increased skin blood flow compared to parameters observed during short incremental exercise tests [3]), other studies suggested that HR peak may not be affected by stage and exercise test duration [8,[11][12][13]31]. Additionally, our results are similar to others [8], who reported a mean bias of 3 bpm, with 95% limits of agreement between 6 and -12 bpm between short and long maximal incremental tests with different stage durations in well trained triathletes. On the other hand, contrary to Weston et al. [7], HR VT1 was lower during the L Test compared to the S Test . However, the HR VT1 mean bias was~9 bpm (~5%) between the two tests, and it may be within the range of day-to-day HR variability [32]; therefore, it may be useful in prescribing an appropriate training regimen.

Power output
In line with previous studies [7][8][9], our results show the significant influence of protocol time duration on PPO, findings similar to those of Bishop et al. [9], who reported a mean bias of 34.4 W, with 95% limits of agreement between 59.7 and 9.0 W between short and long maximal incremental tests with different stage and time durations in moderately active females. Interestingly, the results of the present study and those of Bishop et al. [9] show that PPO demonstrated good correlations with respect to short and long maximal incremental tests, although a systematic underestimation of PPO in prolonged exercise was noted (Fig 1E), also attested by lower ICC coefficient. Similarly, as previously reported [7], PO VT1 was also significantly lower during the L Test . The higher PO VT1 noted during the S Test may be related to the physiological lag time between the increase in work rate and gas exchange responses, leading to an overestimation of VT 1 when expressed as a work rate (PO VT1 ) but not when expressed as metabolic units ( _ V O 2VT1 ) [7]. Moreover, although not measured, it is possible that the higher PPO observed during the S Test was related to lower blood lactate concentrations during the S Test compared to the L Test , allowing subjects to attain a higher PO before suffering from local muscle fatigue [7,11]. Additionally, the _ V O 2 slow component for exercises above the VT 1 [33] may be undetectable until the end of testing during rapidly-incremental ramp tests [34] but has sufficient time to be expressed during prolonged exercise tests [35], which may explain the lower PPO but similar _ V O 2peak and the lower DE noted during the L Test .

L Test and exercise training prescriptions
It has been established that monitoring _ V O 2 and HR during effort is the most commonly used method of prescribing and individualizing exercise training to determine exercise intensity (expressed in % _ V O 2peak and %HR peak ). Moreover, training target zones are also usually defined based on % _ V O 2peak and %HR peak to individualize exercise training regimens and to determine the effects of a training session [32,36]. In obese individuals, the individualization concept of training plays a pivotal role in weight management, particularly in reducing cardiovascular risk and the risk of developing secondary pathologies [37]. Indeed, it has been demonstrated that various forms of training for which exercise intensity was individualized at a target %HR peak (corresponding to VT 1 [38], moderate intensities [39,40] and high-intensities [40][41][42]) determined by a short (~10 min) maximal incremental test may improve health-related outcomes (i.e., _ V O 2peak , muscle oxidative capacity, lipid profiles and insulin sensitivity) in this population.
From a clinical standpoint, as our results show good agreement in HR and _ V O 2 between the L Test and S Test : we believe that the L Test is also an appropriate test for evaluating aerobic fitness and for prescribing exercise training regimens in class II and III obese men. Additionally, compared to short incremental tests, prolonged incremental exercise may also be used to assess fat oxidation kinetics, MFO and Fat max in obese and severely obese individuals [21]. Indeed, it has been previously demonstrated that individualized Fat max training may significantly increase muscle oxidative capacity, as well as fat oxidation rates during exercise and insulin sensitivity in obese individuals [19,43], highlighting its clinical relevance in the treatment of obesity [37] and the importance of correctly assessing Fat max as a function of measured _ V O 2peak [44]. However, to reduce the number of times that subjects have to report to the laboratory before starting training, it is preferable that only one test be performed. Therefore, we suggest that a prolonged incremental exercise test that starts with a 10-min warm-up at 40 W, followed by 20 W increments every 5 min until reaching 120-140 W (i.e., 4 or 5 stages), followed by 15 W increments every minute until exhaustion would be a feasible and accurate test for assessing aerobic fitness and prescribing an exercise training regimen in class II and III obese men.

Methodological considerations
Some methodological limitations arose from the study and need to be further addressed. Firstly, the subjects always completed the S Test first and the L Test second. Although a randomised counterbalanced test order would have been preferable, in our study design we need to firstly conduct the S Test with regard to determine the correct PO for the warm-up and for the 5-min stage increments during the L Test in order to individualise each protocol and obtain enough points to assess fat oxidation kinetics, MFO and Fat max in our subjects [29]. Moreover, through this study design, we were able to develop a single test protocol specific to class II and III obese men that accurately and simultaneously assess aerobic and metabolic fitness (see above for details). In this line, Fat max seems to be accurately assessed during L Test because has been found at similar values (~51% _ V O 2peak ) than those previously reported in this population [19,[21][22][23][24]. However, further investigations are needed to confirm this claim. Secondly, as we focused primarily on _ V O 2peak and not on _ V O 2max , our results may also have been affected. However, it was recently suggested that _ V O 2peak may also be indicative of a true _ V O 2max in both lean [45] and obese individuals [1]. Additionally, previous studies have already compared _ V O 2peak between two different maximal incremental tests with different stage and test durations in normalweight individuals [7][8][9]11]. Moreover, the observed agreement in HR peak and _ V O 2peak with respect to the L Test and S Test suggests that these measurements are reproducible with different tests in class II and III obese men. However, the lower RER obtained during the L Test may be related to the depletion of bicarbonate reserves as a result of increased time spent above VT 1 [10], suggesting that the use of RER as an indicator of maximal effort in the setting of prolonged incremental tests should be reconsidered.
In summary, we demonstrate that _ V O 2peak , HR peak and _ V O 2VT1 assessments were not compromised by prolonged test durations in class II and III obese men. Therefore, we suggest that the L Test is a feasible and accurate maximal incremental test and may be used to evaluate aerobic and metabolic fitness and to prescribe exercise training regimens to improve therapeutic approaches used to treat obesity and severe obesity.