Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Reproducibility of Fatmax and Fat Oxidation Rates during Exercise in Recreationally Trained Males

  • Ilaria Croci ,

    ilaria.croci@uqconnect.edu.au

    Affiliations The University of Queensland Diamantina Institute, The University of Queensland, Translational Research Institute, Brisbane, Australia, School of Human Movement Studies, University of Queensland, Brisbane, Australia

    ⨯
  • Fabio Borrani,

    Affiliations Institute of Sport Sciences University of Lausanne, University of Lausanne, Lausanne, Switzerland, Department of Physiology, University of Lausanne, Lausanne, Switzerland

    ⨯
  • Nuala Byrne,

    Affiliations Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Australia, Bond Institute of Health and Sport, Bond University, Robina, Australia

    ⨯
  • Rachel Wood,

    Affiliation Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Australia

    ⨯
  • Ingrid Hickman,

    Affiliations The University of Queensland Diamantina Institute, The University of Queensland, Translational Research Institute, Brisbane, Australia, Department of Nutrition and Dietetics, Princess Alexandra Hospital, Brisbane, Australia, Mater Medical Research Institute, Mater Mother’s Hospital, Brisbane, Australia

    ⨯
  • Xavier Chenevière,

    Affiliations Institute of Sport Sciences University of Lausanne, University of Lausanne, Lausanne, Switzerland, Department of Medicine, University of Fribourg, Fribourg, Switzerland

    ⨯
  • Davide Malatesta

    Affiliations Institute of Sport Sciences University of Lausanne, University of Lausanne, Lausanne, Switzerland, Department of Physiology, University of Lausanne, Lausanne, Switzerland

    ⨯

Correction

17 Nov 2014: The PLOS ONE Staff (2014) Correction: Reproducibility of Fatmax and Fat Oxidation Rates during Exercise in Recreationally Trained Males. PLOS ONE 9(11): e114115. https://doi.org/10.1371/journal.pone.0114115 View correction

Abstract

Aerobic exercise training performed at the intensity eliciting maximal fat oxidation (Fatmax) has been shown to improve the metabolic profile of obese patients. However, limited information is available on the reproducibility of Fatmax and related physiological measures. The aim of this study was to assess the intra-individual variability of: a) Fatmax measurements determined using three different data analysis approaches and b) fat and carbohydrate oxidation rates at rest and at each stage of an individualized graded test. Fifteen healthy males [body mass index 23.1±0.6 kg/m2, maximal oxygen consumption () 52.0±2.0 ml/kg/min] completed a maximal test and two identical submaximal incremental tests on ergocycle (30-min rest followed by 5-min stages with increments of 7.5% of the maximal power output). Fat and carbohydrate oxidation rates were determined using indirect calorimetry. Fatmax was determined with three approaches: the sine model (SIN), measured values (MV) and 3rd polynomial curve (P3). Intra-individual coefficients of variation (CVs) and limits of agreement were calculated. CV for Fatmax determined with SIN was 16.4% and tended to be lower than with P3 and MV (18.6% and 20.8%, respectively). Limits of agreement for Fatmax were −2±27% of with SIN, −4±32 with P3 and −4±28 with MV. CVs of oxygen uptake, carbon dioxide production and respiratory exchange rate were <10% at rest and <5% during exercise. Conversely, CVs of fat oxidation rates (20% at rest and 24–49% during exercise) and carbohydrate oxidation rates (33.5% at rest, 8.5–12.9% during exercise) were higher. The intra-individual variability of Fatmax and fat oxidation rates was high (CV>15%), regardless of the data analysis approach employed. Further research on the determinants of the variability of Fatmax and fat oxidation rates is required.

Citation: Croci I, Borrani F, Byrne N, Wood R, Hickman I, Chenevière X, et al. (2014) Reproducibility of Fatmax and Fat Oxidation Rates during Exercise in Recreationally Trained Males. PLoS ONE 9(6): e97930. https://doi.org/10.1371/journal.pone.0097930

Editor: Reury F. P. Bacurau, University of Sao Paulo, Brazil

Received: December 29, 2013; Accepted: April 25, 2014; Published: June 2, 2014

Copyright: © 2014 Croci et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: These authors have no support or funding to report.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Carbohydrate and fat are the two main sources of energy that sustain oxidative metabolism. Their relative utilization during aerobic exercise depends largely on exercise intensity [1], [2]. The whole-body carbohydrate oxidation rate (CHOox) increases with the workload, whereas the whole-body fat oxidation rate (Fox) increases from low to moderate exercise intensities, and then markedly declines at high intensities. The exercise intensity at which the maximal fat oxidation (MFO) rate occurs has been defined as Fatmax [3]. Aerobic exercise training performed at Fatmax has the potential to increase Fox and insulin sensitivity in obese patients [4] and in individuals with metabolic syndrome [5]. In patients with type 2 diabetes, aerobic training targeted at Fatmax was shown to have a greater effect on body composition and glucose control than high intensity interval training [6].

To determine Fatmax, a submaximal graded exercise test using indirect calorimetry is performed, and data is analyzed with two main steps. First, Fox and CHOox at each stage of the test are calculated from indirect calorimetry measures [oxygen consumption () and carbon dioxide production ()] by means of the stoichiometric equations [7]. Subsequently, Fox values are plotted as a function of exercise intensity and Fatmax is identified with one of the following four commonly used methods: a) the determination of the maximal value of measured Fox reached during each stage of the graded exercise test and identification of the corresponding intensity (measured values approach, MV) [3], [8]–[10], b) the construction of a 3rd polynomial fitting curve (P3) [11], c) the Sine model (SIN) [12] and d) the non-protein “respiratory quotient technique” [13].

Knowledge of the reproducibility of Fatmax is necessary for establishing its usefulness as a parameter for training prescription and for adequately interpreting outcomes from research studies. To date, there has been limited research into the reproducibility of testing Fatmax, and findings to date are conflicting and have methodological limitations [8], [13], [14]. Achten et al. [8] found a coefficient of variation (CV) for Fatmax of 9.6% in 10 endurance athletes tested on three occasions and concluded that Fatmax measurements are reliable. Perez-Martin et al. [13] tested 10 healthy males on two occasions, reported a CV for Fatmax of 11.4% and suggested this to be a “satisfactory result”. Conversely, Meyer et al. [14] studied 21 recreationally trained men and women who completed the test twice. The limits of agreement (LoA) for oxygen consumption at Fatmax corresponded to a heart rate (HR) difference of 35 bpm between the two tests, which lead them to conclude that the intra-individual variability in Fatmax measurements is too large to recommend using this parameter for prescribing exercise training. In addition to coming to different conclusions, these studies had methodological limitations in terms of testing protocol and data analysis approach. Data analysis to determine Fatmax was performed using the MV [8], [14] or the “respiratory quotient technique” [13] approaches. However, Chenevière et al. [12] recently showed that the employment of a mathematical model (SIN) provides a more complete description of the Fatox kinetics as a function of exercise intensity and more accurate Fatmax measures than the “respiratory quotient technique” approach. Secondly, in two of these studies [8], [14], the starting workload of the graded test occurred on average at ∼45% of the maximal oxygen uptake (), therefore not providing information on substrate metabolism at low intensities, while in the other [13], the protocol included a limited number of exercise stages, therefore limiting information for determining Fatmax. Furthermore, the statistical approach to assess reliability used by Achten et al. and Perez-Martin et al. was not comprehensive given that only CVs were reported [15]. Other measures of variability such as the LoA were not calculated.

Crucially, even though Fatmax is calculated from Fox values at each stage of a submaximal graded test, the reproducibility of Fox over a wide range of exercise intensities has not been assessed. Some studies have evaluated the intra-individual variability of the physiological parameters used to determine substrate oxidation (,and respiratory exchange ratio or RER). The authors reported that and were reliable in resting conditions [16], [17] and that CVs for , and RER were lower than 5% in response to each stage of an incremental exercise test [8], [18], [19]. However, while CVs of , and RER are often reported to inform on the variability in substrate oxidation rates, this might be misleading. The relationship existing between those CVs and the variability of Fox and CHOox has not been established.

Limited information is available on the reproducibility of Fatmax and on the reproducibility of CHOox and Fox at each stage of a graded test. It was therefore the aim of this study to assess the intra-individual variability of: a) Fatmax measurements determined using three different data analysis approaches (SIN, P3 and MV), and b) CHOox and Fox at rest and in response to each stage of an individualized graded test. A further aim was to investigate how the CVs of , and RER are related to the CV of Fox.

Methods

Ethics Statement

The study was conducted in accordance with ethical principles of the 1964 World medical Declaration of Helsinki and was approved by the human research ethics committee of the University of Lausanne (Switzerland). All test procedures, risks and benefits associated with the experiment were fully explained, and written informed consent was obtained from all participants.

Subjects

Fifteen healthy, moderately trained male volunteers (see Table 1 for anthropometric and physical characteristics) were recruited to participate in this study. All participants were of normal weight according to the World Health Organization (Body Mass Index<25 kg⋅m−2), non-smokers and disease-free. They were not taking regular medications and were screened for the absence of electrocardiographic abnormalities at rest and during exercise.

thumbnail
Download:
Table 1. Subject characteristics.

https://doi.org/10.1371/journal.pone.0097930.t001

General Design

Each participant completed three test sessions. In the first session anthropometric measurements (i.e., stature, body mass and body composition) were taken and a maximal incremental test on a cycle ergometer was performed. In the remaining two sessions the subjects performed an identical submaximal incremental test (Test 1 and Test 2). The two tests were performed in the morning (start of exercise between 7 and 8 am) after a10-hour overnight fast. They were separated by 3 to 7 days and performed at the same time of day to avoid circadian variance. The volunteers were asked to fill in a 1-day food diary on the day before Test 1 and to repeat this diet before Test 2. Furthermore, participants were asked to refrain from vigorous exercise and alcohol and caffeine consumption in the 24 hours prior to testing. Participants were familiarized with the equipment prior to testing.

Anthropometric Measurements

Body composition (fat mass and percentage of body fat) was estimated from skin-fold thickness measurements at four sites according to the methods of Durnin and Womersley [20].

Maximal Exercise Test

A maximal incremental test on a cycle ergometer (Ebike Basic BPlus, General Electric, Niskayuna, NY, USA) to determine maximal oxygen uptake () and maximal aerobic power output () was performed. After a 5-min rest period and a 5-min warm-up at 60 W, output was increased by 30 W every minute until volitional exhaustion. was considered as maximal when at least three of the following four criteria were met [21]: 1) a plateauing of (defined as an increase of no more than 2 mL⋅kg−1⋅min−1 with an increase in workload) during the latter stages of the exercise test, 2) an HR>90% of the age-predicted maximum (220-age), 3) an RER>1.1 and 4) an inability to maintain the minimal required pedaling frequency (i.e. 60 rpm) despite maximum effort and verbal encouragement. was calculated as the average over the last 20 seconds of the last stage of the test.

Submaximal Graded Exercise Tests (Test 1 and Test 2)

Test 1 and Test 2 were characterized by two phases: a pre-exercise resting phase (rest) and a submaximal incremental exercise test. They were carried out under identical circumstances with an identical protocol. Data from these two tests were subsequently employed for reliability calculations.

In the pre-exercise resting phase (rest), participants were seated for 30-min on the cycle ergometer and respiratory measures were collected during the last 15-min of this sitting period. Subsequently, a submaximal incremental exercise test to determine whole-body Fox kinetics was performed. After a 10-min warm-up at 20% , the power output was increased by 7.5% every 5-min until RER was >1.0 during the last minute of the stage.

Indirect Calorimetry and Calculations

Oxygen uptake (), carbon dioxide output () and ventilation () were measured continuously using a breath-by-breath system (Oxycon Pro, Jaeger, Würzburg, Germany). Before each test the gas analyzers were calibrated with gases of known concentration (16.00% O2 and 5.02% CO2), and the volume was automatically calibrated at two different flow rates (0.2 L⋅s−1 and 2 L⋅s−1). The HR was recorded continuously using an HR monitor (S810i, Polar Electro OY, Kempele, Finland).

During Test 1 and Test 2, HR and gas exchange data (,) collected during the final 5-min of the pre-exercise resting phase and during the last 2-min of each stage of the submaximal incremental exercise test were averaged and used for calculations. RER was calculated as the ratio between and , while Fox and CHOox were calculated using stoichiometric equations [7], with the assumption that the urinary nitrogen excretion rate was negligible:(1)(2)

(1-RER) was also calculated given that the equation to calculate Fox can be simplified to:(3)

Fox as a function of exercise intensity is reflected by two different linear relationships: a progressive decrease of (1–RER) and a linear increase of as power output is increased. The percentages of total energy expenditure derived from fat (% ENEfat) and CHO (% ENECHO) were calculated [22]:(4)(5)

Data Analysis Approaches to Determine Fatmax

Fox values obtained at each stage of the submaximal graded exercise test (which was terminated when RER was >1) were graphically depicted as a function of exercise intensity. Then, Fatmax and MFO (and subsequently RER, %HRmax at Fatmax, % at Fatmax) were determined using three different data analysis approaches (SIN, MV and P3). The “respiratory quotient technique” was not used in this study since it has been shown to be less accurate than the other methods [12].

SIN model.

The SIN model [12] was used to model and characterize whole-body Fox kinetics:(6)

Dilatation (d), symmetry (s) and translation (t) are the three independent variables representing the main modulations of the curve. K is the constant of intensity and corresponds to (π/100). To fit the experimental data (i.e. Fox rates) and to model the Fox kinetics, the three variables were independently changed using an iterative procedure by minimizing the sum of the mean squares of the differences between the estimated energy derived from fat based on the SIN model and the energy derived from fat calculated from the raw Fox data, as described in a previous study [12]. For each subject, Fatmax was calculated by differentiation of the SIN model equation. The Fatmax zone was determined as the range of exercise intensities with Fox rates within 10% of MFO [3].

P3.

Graphical depiction of Fox values as a function of exercise intensity was performed by constructing a third polynomial curve with intersection at (0;0) [11]. Fatmax was calculated by differentiation of the P3 equation, and corresponded to the intensity at which the value of the differentiated equation was equal to zero.

Measured values.

From the graphical representation of Fox values as a function of exercise intensity, the stage at which the value of measured Fox rates was maximal was determined, and the corresponding intensity was identified [3], [8]–[10], [23].

Theoretical Example to Study how the CVs of and are Related to the CVs of RER and the CV of Fox

In order to investigate how the CV of and are linked to the CVs of parameters informing of substrate utilization (RER, Fox, CHOfat, 1-RER, ENEfat) three theoretical scenarios were created. and values for Test 1 and Test 2 were generated so that CVs of and between Test 1 and Test 2 were ≤3%. A CV of ≤3% for and was chosen in line with results from previous studies [8], [18].

Statistical Analysis

Data are expressed as the means ± standard deviation (SD) for all variables. Intra-individual CVs and LoA were calculated to test the variability between Test 1 and Test 2 for the following measures: a) Fatmax and physiological measures at Fatmax (MFO, RER, % HRmax and % ) determined with three different data analysis approaches (SIN, MV and P3) and b) gas exchange data, HR and substrate oxidation rates at rest and during the first six stages of the submaximal incremental tests (from 20% to 57.5% of ). Intra-individual CVs were calculated for the physiological variables studied in the three theoretical scenarios.

Two-factorial analysis of variance for repeated measures (RMANOVA) was carried out to test for systematic changes in: a) Fatmax, and physiological measures at Fatmax (factor 1: tests, factor 2: data analysis approaches), and b) gas exchange data, HR and substrate oxidation rates (factor 1: tests; factor 2: exercise intensity). For the same outcome measures, one-way RMANOVA was carried out to test for systematic changes in the intra-individual CV at Fatmax.

Bland-Altman scatterplots are presented for Fatmax and MFO determined with SIN, P3 and MV. They show the difference between two corresponding measurements plotted against the mean of the measurements. Reference lines for the mean difference±1.96 SD are given. For all statistical analyses, the level of significance was set at P≤0.05. Statistical analysis was performed with the software SPSS 17.0 for Windows (SPSS, Chicago, IL) and Graph Pad Prism version 5.0 for Mac (GraphPad Software, San Diego, CA).

Results

Fatmax and Physiological Measures at Fatmax Determined with SIN, P3 and MV

Fatmax and physiological measures at Fatmax determined with three data analysis approaches (SIN, P3 and MV) are presented in Table 2. For all parameters, average values (n = 15) obtained from Test 1 and Test 2 were not significantly different (i.e. for Fatmax: P = 0.37 for factor test and P = 0.20 for factor interaction between test and approach), indicating that no habituation or training effects occurred between testing sessions. Average values for Fatmax and related measures obtained with the three different approaches were also not significantly different (i.e. for Fatmax: P = 0.13 for factor approach).

thumbnail
Download:
Table 2. Average values, limits of agreement and CVs for Fatmax and physiological measures at Fatmax determined with three approaches: SIN, P3 and MV.

https://doi.org/10.1371/journal.pone.0097930.t002

On the other hand, the within-individual CVs for Fatmax determined with SIN was 16.4% and tended to be lower (P = 0.10) than with P3 and MV (20.8% and 18.6% respectively). Similarly, the intra-individual CVs of % HR at Fatmax and % at Fatmax determined with SIN were significantly lower than with the other approaches (P = 0.043 and P = 0.05, respectively).

The Bland–Altman scatterplots for Fatmax and MFO (Figure 1) reveal considerable intra-individual variability. The LoA for Fatmax were −2±27% of with SIN, −4±32% with P3, and −4±28% with MV. For MFO they were −0.01±0.25 g/min with SIN, 0.01±0.24 g/min with P3, and 0±0.26 g/min with MV (Table 2). A large between-individual difference in the variability between Test 1 and Test 2 was also seen. Accordingly, the CV at Fatmax ranged from 0 to 48%. For seven subjects it was under 10%, for two subjects it was between 10 and 15%, while for six subjects it was over 20%. However, the size of the difference between Test 1 and Test 2 appeared to be independent of the average value between the two measurements.

thumbnail
Download:
Figure 1. Bland-Altman plots of Fatmax and MFO determined with SIN, P3 or MV. SIN, sine model.

P3, polynomial 3rd degree; MV, measured values; Fatmax, exercise intensity at which maximal fat oxidation rate occurs; , maximal oxygen uptake; MFO, maximal fat oxidation rate; Biases (solid lines) and 95% limits agreement (dashed lines).

https://doi.org/10.1371/journal.pone.0097930.g001

The difference in the HR at Fatmax between test 1 and 2 was <10 bpm in six participants, between 10 and 25 bpm in eight participants and was >25 bpm in one. In both tests, the range of HR frequencies corresponding to the Fatmax zone was broad (it was 38±8 bpm, and ranged from 95±16 to 133±20 bpm).

Physiological Measures at Each Stage of a Submaximal Graded Test

The course of average ,, RER, HR, Fox and CHOox in response to two identical submaximal graded test performed on separate days (Test 1 and Test 2) is presented in Figure 2. There was no significant difference between Test 1 and Test 2 in any of the parameters assessed. Average values of most physiological variables (except Fox) significantly increased with exercise intensity (P<0.001).

thumbnail
Download:
Figure 2. Course of average , , HR, RER, Fox and CHOox during two identical submaximal incremental tests (mean and SD).

, maximal aerobic power output; , oxygen uptake; , carbon dioxide production; RER, respiratory exchange ratio; HR, heart rate; Fox, fat oxidation rate; CHOox, carbohydrate oxidation rate. *significantly increases with exercise intensity, §rest significantly different than exercise (20–57.5% ), † significantly different than 57.5% .

https://doi.org/10.1371/journal.pone.0097930.g002

CVs of ,, HR and RER were <10% at rest and <5% during exercise (Table 3). For instance, CVs for , and RER were 7.5%, 9.1% and 3.8% at rest and were, on average, 3.1%, 3.0% and 2.5% during exercise. In contrast, CVs for CHOox and Fox were markedly higher. The CV for Fox was 20.6% at rest and ranged from 24.1 to 49.2% during exercise, while for CHOox, it was 33.5% at rest and ranged from 8.5% to 12.9% during exercise. Interestingly, although the CV of RER was <4% under each condition, the CV of 1-RER was markedly higher (>20%), and was equal to the CV of the % of ENEfat. LoA between Test 1 and Test 2 are presented in Table 4.

thumbnail
Download:
Table 3. Coefficients of variation (%) for respiratory values and substrate oxidation rates in reponse to a submaximal graded exercise test.

https://doi.org/10.1371/journal.pone.0097930.t003

thumbnail
Download:
Table 4. Limits of agreement between Test 1 and Test 2 for respiratory values and substrate oxidation rates in response to a submaximal graded exercise test.

https://doi.org/10.1371/journal.pone.0097930.t004

Theoretical Example to Study how the CVs of and are Related to the CVs of RER and the CV of Fox

Three theoretical scenarios in which the CVs of and were ≤3% are presented in Table 5 and additional results with mathematical explanations are presented in appendix S1. In case scenario 1 and 2, the CVs of Fox from were markedly different (3.1% vs. 38.2%) despite the CVs of and being identical (3%) (see Appendix S1, eq. 7, 8 and 9). Further, the CV of 1-RER was higher than the CV of RER, and was equal to the CV of %ENEfat. This difference was particularly apparent in case 2, where the CV of RER was 6%, while the CVs of Fox and 1-RER were 38.2 and 35.3%, respectively.

thumbnail
Download:
Table 5. Three case scenario in which CV for and are≤3%.

https://doi.org/10.1371/journal.pone.0097930.t005

From the analysis of the three theoretical scenarios (as well as from the analysis of the whole dataset of 15 participants) we also observed that the CV of Fox can be calculated from sum or subtraction of the CV of (1-RER) and the CV of (Appendix S1, eq. 10 and 11). For example, in case 2, the CV of Fox was 38.2% and was the sum of the CVs of 1-RER (35.3%) and (3.0%). In case 3, the CV of Fox was 15.3%, and equaled the CV of 1-RER (15.3%) ± CV (0.0%).

Discussion

In this study we assessed the reproducibility of Fatmax measurements determined with three different data analysis approaches and of CHOox and Fox at rest (while sitting) and in response to each stage of an individualized graded test. We observed that the intra-individual variability of Fatmax was large (CV>16%) regardless of the data analysis approach employed and that Fox at rest and at each stage of a graded test was also variable (CV>20%), despite the CVs of ,and RER being <5%.

The reproducibility of Fox values at each stage of a graded test, despite being a key aspect in the determination of Fatmax, was previously unexplored. In the current study, the CVs found for the parameters from which Fatox is calculated (, and RER) were in line with previous observations. At rest, the CV for RER was 3.8%, which closely mirrors the CV of 3.5% found by Roffey et al. [17]. In the present study the resting assessment was performed with the individuals in a seated position and this needs to be taken into consideration when making comparisons with studies in which resting metabolism was assessed with participants lying supine. During exercise, the average CVs for , and RER were 3.1%, 3.0% and 2.5%, respectively, and were similar or lower than those reported in previous investigations [8], [13], [18], [19]. Despite this, the CVs found for Fox were >20%. This shows that even though CHOox and Fox are calculated from and by means of the stoichiometric equations [7], a low variability in those parameters is not necessarily indicative of low variability in CHOox and Fox.

To further study how the CVs of and are related to the CVs of RER and the CV of Fox, three theoretical scenarios were created. At present, scientific reports as well as companies validating calorimeters tend to draw information on the variability of substrate oxidation from the CVs of , and RER. The results of the theoretical scenarios (Table 5) and the mathematical explanations presented in the appendix S1 illustrate that those CVs do not provide sufficient information on the variability of substrate oxidation rates.

As can be seen in case 2, when the and vary in different directions between two tests (increase in and decrease in or viceversa), the variability of Fox is high. This is because in such conditions, the standard deviation of Fox results from the sum of the standard deviations of and , multiplied by a factor 1.67. Therefore, in addition to the size of the change in and between tests, it is crucial to know whether they change in the same or opposite sense between measurements.

The RER is the ratio between and and, therefore, provides information on the relationship between those measurements. However, in the theoretical scenarios the CV of RER remains low (<6%) also when the variability in Fox is high (>30%), showing that the CV of RER is not a parameter adequately informing on the variability in the proportion of nutrients utilized. This is because the RER is value bounded in an interval separate from zero (0.7–1.0) and therefore the CV is not an adequate measure to assess the variability of RER. On the other hand, the CV of 1-RER appears to be an informative marker on the variability in Fox rates: it provides the same results as the CV of ENEfat, it accounts for a large proportion of the CV of Fox, and it is simple to calculate.

In this study, as well as in other studies investigating the reproducibility of indirect calorimetry measures [8], [13], [14], [16]–[19], the total variation observed between Test 1 and Test 2 is the sum of both biological and equipment variation. It was beyond the scope of this study to assess the relative contribution of each. However, the average variation of the equipment (gas analysis system) used in this study is known. It was assessed using a portable metabolic simulator (which excludes any biological variability) and the average CV for and was 1.9±0.6% and 1.3±0.5% respectively [18].

In addition to investigating the variability in Fox and related parameters at each stage of a graded test, a novel feature of this study was the assessment of the intra-individual variability in Fatmax determined with the SIN model and its comparison with the variability of Fatmax measures obtained using different data analysis approaches. All the approaches to determine Fatmax presented in the literature were compared in this analysis, except the “respiratory quotient technique”, since it has previously been shown to be less accurate [12]. The comparison revealed that the intra-individual CV at Fatmax was higher than 16% with any of the data analysis approaches employed and that there was a relatively small difference between approaches. However, the CVs of Fatmax, % at Fatmax and of % HRmax determined with SIN were lower than with P3 and MV, possibly because the SIN model provides an accurate and more complete description of the Fox kinetics as a function of exercise intensity than the other data analysis approaches. These results support the use of SIN over other approaches in future studies given that it is more reliable and provides more detailed information.

The intra-individual variability of Fatmax and related parameters found in this study was in line with those of Meyer et al. [14]. In the present study the LoA for Fatmax determined with SIN were −2.0±27.7 of , while in the study from Meyer et al. LoA for Fatmax of −3.9±27.7 of were observed. Further, also consistent with the results published by Meyer et al. [14], the within-individual variability was markedly different between individuals. On the other hand, the CV for Fatmax observed in this study, on average, was slightly higher than those reported in other studies [8], [13], [24]. The lower CV found by Achten et al. [8] (9.6%) could be due to the fact that measurements were repeated three times [and the CV generally decreases when the number of measurements increases [25]] and were performed in trained athletes, who may have a less variable response to exercise than individuals with a lower training level. Overall, the differences in the results obtained between studies are difficult to interpret, particularly because most studies only report average results, and do not present “individual responses” and/or ranges. This highlights the need for a better understanding of the determinants of intra-individual variability in Fatmax.

Previous studies in the field considered an intra-individual variability of ±10 bpm in the HR at Fatmax acceptable, since this value reflects a realistic margin in individuals who use HR for the monitoring of training intensity [8], [14]. In the present study this target was met by the majority, but not all, participants. However, the range of intensities at which Fox is within 10% of MFO (Fatmax zone) was broad and this was consistent with previous observations [3], [26]. Therefore, despite its variability, training prescription at Fatmax ensures that high rates of Fox are elicited on different days.

The determination of Fox and Fatmax (and therefore the determination of their variability) is influenced by a number of methodological factors including the exercise test design, the data analysis approach and the pre-test conditions. In this study, a robust methodological approach was employed. The submaximal graded exercise was individualized based upon the results of a maximal test. It started at 20% of and the workload was subsequently increased by 7.5% every 5-min. This ensured the reaching of a steady state [27] and allowed to study Fox at several intensities (participants performed at least six exercise stages with an RER<1). Further, the statistical analysis was carried out in accordance with the recommendations for reliability assessment in sport medicine [15].

Pre-test conditions included a 10-hour overnight fast and 24 hours of standardization in diet and physical activity prior to each submaximal graded exercise test. This level of standardization was adopted because it appears to be the most commonly employed approach in our research field [3], [8], [28]–[32] and because more rigorous standardization is difficult to achieve both in out-clinic and research settings. Despite the standardization adopted, in some individuals a high intra-individual variability in Fatmax and related variables was found, suggesting that a longer period of standardization (≥2 days prior to testing) might be needed to improve the reproducibility of those measures. However, while more strict pre-test standardization leads to greater internal validity, it also leads to poorer external validity (i.e harder translation of the results into practice). More generally, while the validity of using a graded exercise tests to determine Fatmax has been reported in a number of studies [3], [13], [33], [34], a recent study questions the usefulness using this approach to prescribe training in populations such as highly trained athletes [35]”.

A number of questions on the reproducibility of substrate metabolism during exercise are still to be answered. Further research is required to: a) describe how standardization in physical activity and diet prior to testing impact on reliability of measurements, b) study the determinants of the variability in CHOox and Fox and c) explore the reproducibility in Fox in other cohorts including overweight and untrained individuals.

In summary, we have shown here that the intra-individual variability in Fatmax is high (CV>16%) and is highly variable between individuals, regardless of the data analysis approach employed. The intra-individual variability at rest and in response to an individualized graded test is high for Fox measures (CV>20% for Fox) although it is low for , and RER (CV<5%). The CV of (1-RER) appears to be a more representative measure of the variability in substrate oxidation than CV of RER. Training prescription at Fatmax can be useful clinically given that, despite its variability, it results in Fatox rates within 90% of MFO on different days. In a research setting, differences in Fatmax and Fox within and between groups can be detected as long as a sufficiently large number of participants is recruited. Further research in this area is required.

Acknowledgments

The authors would like to thank Aleš Neubert for mathematical inputs, and Graeme Macdonald and John Prins for helpful suggestions and criticisms.

Author Contributions

Conceived and designed the experiments: IC FB NB RW IH XC DM. Performed the experiments: IC FB XC DM. Analyzed the data: IC FB XC DM. Contributed reagents/materials/analysis tools: IC FB NB RW IH XC DM. Wrote the paper: IC. Revised the manuscript for important intellectual content and approved the final version: IC FB NB RW IH XC DM.

References

  1. 1. van Loon LJ, Greenhaff PL, Constantin-Teodosiu D, Saris WH, Wagenmakers AJ (2001) The effects of increasing exercise intensity on muscle fuel utilisation in humans. J Physiol 536: 295–304.
  2. 2. Romijn JA, Coyle EF, Sidossis LS, Gastaldelli A, Horowitz JF, et al. (1993) Regulation of endogenous fat and carbohydrate metabolism in relation to exercise intensity and duration. Am J Physiol 265: E380–391.
  3. 3. Achten J, Gleeson M, Jeukendrup AE (2002) Determination of the exercise intensity that elicits maximal fat oxidation. Med Sci Sports Exerc 34: 92–97.
  4. 4. Venables MC, Jeukendrup AE (2008) Endurance training and obesity: effect on substrate metabolism and insulin sensitivity. Med Sci Sports Exerc 40: 495–502.
  5. 5. Dumortier M, Brandou F, Perez-Martin A, Fedou C, Mercier J, et al. (2003) Low intensity endurance exercise targeted for lipid oxidation improves body composition and insulin sensitivity in patients with the metabolic syndrome. Diabetes Metab 29: 509–518.
  6. 6. Maurie J, Brun J-F, Jean E, Romain A-J, Mercier J (2011) Comparison of Square-Wave Endurance Exercise Test (SWEET) training and endurance training targeted at the level of maximal lipid oxidation in type 2 diabetics. Science & Sports 26: 92–96.
  7. 7. Frayn KN (1983) Calculation of substrate oxidation rates in vivo from gaseous exchange. J Appl Physiol 55: 628–634.
  8. 8. Achten J, Jeukendrup AE (2003) Maximal fat oxidation during exercise in trained men. Int J Sports Med 24: 603–608.
  9. 9. Achten J, Venables MC, Jeukendrup AE (2003) Fat oxidation rates are higher during running compared with cycling over a wide range of intensities. Metabolism 52: 747–752.
  10. 10. Venables MC, Achten J, Jeukendrup AE (2005) Determinants of fat oxidation during exercise in healthy men and women: a cross-sectional study. J Appl Physiol 98: 160–167.
  11. 11. Stisen AB, Stougaard O, Langfort J, Helge JW, Sahlin K, et al. (2006) Maximal fat oxidation rates in endurance trained and untrained women. Eur J Appl Physiol 98: 497–506.
  12. 12. Cheneviere X, Malatesta D, Peters EM, Borrani F (2009) A mathematical model to describe fat oxidation kinetics during graded exercise. Med Sci Sports Exerc 41: 1615–1625.
  13. 13. Perez-Martin A, Dumortier M, Raynaud E, Brun JF, Fedou C, et al. (2001) Balance of substrate oxidation during submaximal exercise in lean and obese people. Diabetes Metab 27: 466–474.
  14. 14. Meyer T, Folz C, Rosenberger F, Kindermann W (2009) The reliability of fat. Scand J Med Sci Sports 19: 213–221.
  15. 15. Atkinson G, Nevill AM (1998) Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med 26: 217–238.
  16. 16. Gasic S, Schneider B, Waldhausl W (1997) Indirect calorimetry: variability of consecutive baseline determinations of carbohydrate and fat utilization from gas exchange measurements. Horm Metab Res 29: 12–15.
  17. 17. Roffey DM, Byrne NM, Hills AP (2006) Day-to-day variance in measurement of resting metabolic rate using ventilated-hood and mouthpiece & nose-clip indirect calorimetry systems. JPEN J Parenter Enteral Nutr 30: 426–432.
  18. 18. Carter J, Jeukendrup AE (2002) Validity and reliability of three commercially available breath-by-breath respiratory systems. Eur J Appl Physiol 86: 435–441.
  19. 19. Michallet AS, Tonini J, Regnier J, Guinot M, Favre-Juvin A, et al. (2008) Methodological aspects of crossover and maximum fat-oxidation rate point determination. Diabetes Metab 34: 514–523.
  20. 20. Durnin JV, Womersley J (1973) Total body fat, calculated from body density, and its relationship to skinfold thickness in 571 people aged 12–72 years. Proc Nutr Soc 32: 45A.
  21. 21. Wood RE, Hills AP, Hunter GR, King NA, Byrne NM (2010) Vo2max in overweight and obese adults: do they meet the threshold criteria? Med Sci Sports Exerc 42: 470–477.
  22. 22. Brandou F, Savy-Pacaux AM, Marie J, Bauloz M, Maret-Fleuret I, et al. (2005) Impact of high- and low-intensity targeted exercise training on the type of substrate utilization in obese boys submitted to a hypocaloric diet. Diabetes Metab 31: 327–335.
  23. 23. Croci I, Byrne NM, Choquette S, Hills AP, Chachay VS, et al. (2013) Whole-body substrate metabolism is associated with disease severity in patients with non-alcoholic fatty liver disease. Gut 62: 1625–1633.
  24. 24. Gmada N, Marzouki H, Haboubi M, Tabka Z, Shephard RJ, et al. (2011) Crossover and maximal fat-oxidation points in sedentary healthy subjects: methodological issues. Diabetes Metab 38: 40–45.
  25. 25. Hopkins WG (2000) Measures of reliability in sports medicine and science. Sports Med 30: 1–15.
  26. 26. Cheneviere X, Malatesta D, Gojanovic B, Borrani F (2010) Differences in whole-body fat oxidation kinetics between cycling and running. Eur J Appl Physiol 109: 1037–1045.
  27. 27. MacRae HS, Noakes TD, Dennis SC (1995) Role of decreased carbohydrate oxidation on slower rises in ventilation with increasing exercise intensity after training. Eur J Appl Physiol Occup Physiol 71: 523–529.
  28. 28. Achten J, Jeukendrup AE (2004) Relation between plasma lactate concentration and fat oxidation rates over a wide range of exercise intensities. Int J Sports Med 25: 32–37.
  29. 29. Brandou F, Dumortier M, Garandeau P, Mercier J, Brun JF (2003) Effects of a two-month rehabilitation program on substrate utilization during exercise in obese adolescents. Diabetes Metab 29: 20–27.
  30. 30. Tolfrey K, Jeukendrup AE, Batterham AM (2010) Group- and individual-level coincidence of the ‘Fatmax’ and lactate accumulation in adolescents. Eur J Appl Physiol 109: 1145–1153.
  31. 31. Aucouturier J, Rance M, Meyer M, Isacco L, Thivel D, et al. (2009) Determination of the maximal fat oxidation point in obese children and adolescents: validity of methods to assess maximal aerobic power. Eur J Appl Physiol 105: 325–331.
  32. 32. Kang J, Rashti SL, Tranchina CP, Ratamess NA, Faigenbaum AD, et al. (2009) Effect of preceding resistance exercise on metabolism during subsequent aerobic session. Eur J Appl Physiol 107: 43–50.
  33. 33. Takagi S, Sakamoto S, Midorikawa T, Konishi M, Katsumura T (2014) Determination of the exercise intensity that elicits maximal fat oxidation in short-time testing. J Sports Sci 32: 175–182.
  34. 34. Crisp NA, Guelfi KJ, Licari MK, Braham R, Fournier PA (2012) Does exercise duration affect Fatmax in overweight boys? Eur J Appl Physiol 112: 2557–2564.
  35. 35. Schwindling S, Scharhag-Rosenberger F, Kindermann W, Meyer T (2013) Limited Benefit of Fatmax-Test to Derive Training Prescriptions. Int J Sports Med.