Anaerobic Threshold By Mathematical Model In Healthy And Post-myocardial Infarction Men

Anaerobic Threshold by Mathematical Model in

Healthy and Post-Myocardial Infarction Men

Authors L. D. Novais1, 2_{, E. Silva}1_{, R. P. Simões}1_{, D. I. Sakabe}1, 3_{, L. E. B. Martins}4_{, L. Oliveira}1_{, C. A. R. Diniz}5_,

L. Gallo Jr6_{, A. M. Catai}1

Affiliations Affiliation addresses are listed at the end of the article

Introduction

▼

The anaerobic threshold (AT) is a physiological marker of great importance for the prescription of exercise in healthy and cardiac patients [15]. There are many different methods to determine AT, considering the direct method, i. e., the blood lactate analysis, probably is the most accurate and therefore the “gold standard” for determin-ing the AT. However, among the indirect methods for determining the AT, the ventilatory method (VM) is considered the gold standard [11]. Despite the high reliability of these 2 methods, the analysis of blood lactate is an invasive method and the ventilatory method requires expensive equipment, more common in research environ-ments. However, other methods have also been used, including the observation of pattern changes such as heart rate (HR) [2, 10, 13, 16, 23, 24] and neuromuscular response through the myoelectric signals that reflect the recruitment of the motor unit, which can be measured by sur-face electromyography (sEMG) [18].

The reason for the investigation of alternative methods is the intention to apply models that can be used in physical training and rehabilita-tion environments. The applicarehabilita-tion of mathe-matical models in more accessible variables is a low-cost alternative and can produce reliable results [3, 9, 25, 27].

Conventional protocols are designed to ade-quately determine AT in athletes and healthy individuals. Some studies have shown reduced power and aerobic capacity evaluated by peak V˙O2 and V˙O2 at AT, respectively, in cardiac patients with left ventricular dysfunction, chronic heart failure [26], and coronary disease [14]. Consider-ing the importance of AT determination in aero-bic physical training and its relationship with peak V˙O2, more studies are required to verify the cardiopulmonary and neuromuscular response patterns in the transition from aerobic to anaero-bic metabolism. In addition, the study of these variables and the AT determination in cardiac patients, it is important to establish the correct intensity of effort in a physical exercise program. Nevertheless, although many studies have shown accepted after revision

May 27, 2015 Bibliography DOI http://dx.doi.org/ 10.1055/s-0035-1555776 Published online: October 28, 2015 Int J Sports Med 2016; 37: 112–118 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0172-4622

Correspondence

Dra. Aparecida Maria Catai

Department of Physical Therapy

Cardiovascular Physical Therapy Laboratory

Federal University of Sao Carlos Rodovia Washington Luís, km 235 13565-905 Sao Carlos, SP Brazil Tel.: + 55/16/3361 6707 Fax: + 55/16/33612 081 [email protected] Key words ●▶ exercise ●▶ oxygen uptake ●▶ surface electromyography ●▶ coronary artery disease

Abstract

▼

The aim of this study was to determine the anaerobic threshold (AT) in a population of healthy and post-myocardial infarction men by applying Hinkley’s mathematical method and comparing its performance to the ventilatory visual method. This mathematical model, in lieu of observer-dependent visual determination, can produce more reliable results due to the uni-formity of the procedure. 17 middle-aged men (55 ± 3 years) were studied in 2 groups: 9 healthy men (54 ± 2 years); and 8 men with previous myocardial infarction (57 ± 3 years). All subjects underwent an incremental ramp exercise test until physical exhaustion. Breath-by-breath

ven-tilatory variables, heart rate (HR), and vastus lat-eralis surface electromyography (sEMG) signal were collected throughout the test. Carbon dioxide output (V˙CO2), HR, and sEMG were studied, and the AT determination methods were compared using correlation coefficients and Bland-Altman plots. Parametric statistical tests were applied with significance level set at 5 %. No significant differences were found in the HR, sEMG, and ven-tilatory variables at AT between the different methods, such as the intensity of effort relative to AT. Moreover, important concordance and signif-icant correlations were observed between the methods. We concluded that the mathematical model was suitable for detecting the AT in both healthy and myocardial infarction subjects.

that the mathematical models were sensitive in the detection of AT, these studies were performed in the vast majority with healthy individuals [3, 9, 25, 27], and to our knowledge, there have not been any studies to date that have applied mathemati-cal models to determine the AT in cardiac patients.

We hypothesized that individuals with previous myocardial infarction (MI) have a breaking point in the cardiopulmonary and muscle variables during the exercise protocol and that this point corresponds to the AT, similar to healthy individuals. The aim of this study was to establish the existence of a transition pattern in the incremental exercise test for a population with low physical capacity and by means of a mathematical method applied to HR, carbonic gas production (V˙CO2), and amplitude of electromyography signal as root mean square (RMS) values and correlate this transition with AT determined by ventilatory method through visual identification.

Methods

▼

Study design and population

This was a prospective, cross-sectional, controlled study with 17 middle-aged men (aged between 50 and 59 years) separated into 2 groups: 1) 9 healthy men (HG); and 2) 8 myocardial infarc-tion (MIG). All participants signed a written informed consent and the study protocol was approved by the Human Research Ethics Committee of the University and is in accordance with the ethical standards of the journal [7].

All subjects underwent clinical and physical examinations including resting electrocardiogram (ECG), a maximum exercise test, spirometric test and biochemical tests.

Inclusion criteria for the HG were: apparently healthy status based on clinical examination, no use of any type of medication, tobacco or alcohol. The inclusion criteria for the MIG was: first event of MI with no less than 6 months; no ischemic ST-segment changes during clinical exercise testing with cardiologist; no use of tobacco, alcohol and beta-blockers or calcium channel blockers.

Experimental procedures

The experiments were carried out with controlled temperature and humidity and at the same time of day considering circadian influences on responses. The subjects underwent symptom-lim-ited ramp cardiopulmonary exercise testing (CPX) on an electro-magnetically braked cycle ergometer (Quinton Corival 400, Groninger, Netherlands): 1 min of rest, 4 min of warm-up at 4 W, a continuous incremental phase at 15 W · min − 1_{, and 2 min of} cool-down.

Pulmonary gas exchange measurement: Respiratory gas was

obtained breath-by-breath (CPX-D Breeze Suite 6.4.1, Med-Graphics, St Paul, MN, USA). The V˙O2 peak were expressed as the highest value observed during the last 30 s of exercise [9].

Heart rate measurement: Subjects were monitored using

one-channel heart monitor (ECAFIX TC500, ECAFIX, Sao Paulo, SP, Brazil) and an analog-digital converter Lab PC + (National Instruments, Co., Austin, TX, USA) connected to a microcom-puter. The R-R intervals were recorded from the ECG using sig-nal-processing software [20].

EMG recordings and analysis: the EMG signal from the vastus

lateralis was recorded using a bipolar electrode configuration

with an inter-electrode distance of 20 mm. The electrodes were placed parallel to the muscle fibers over the belly of the muscle at the approximate midpoint between the head of the greater trochanter and the lateral condyle of the femur. A reference elec-trode was placed equidistant to the differential elecelec-trodes [8]. Myoelectrical signals during exercise tests were amplified with band-pass filtering (20 Hz and 500 Hz) and recorded on a digital data recorder. Data were digitized at a sampling frequency of 2 000 Hz and the amplitude of electromyographic signals was reported as RMS (µV).

Determination of AT by ventilatory method: Visual analysis

of ventilatory variables (V˙O2 and V˙CO2) was performed in the moving average graph of 8 breathing cycles. The AT was identi-fied by 3 independent experts using the V-slope method and the AT value was considered as the mean of the data obtained from the 3 independent analyses [9].

Determination of AT by the Hinkley mathematical model: The

Hinkley mathematical model was used to identify the change point on the time line using the maximum likelihood method [3, 9, 27]. This model was applied to the HR, V˙CO2, and RMS index of EMG data for each subject using a mathematical algorithm developed in the statistical software S-Plus (2000 Professional Release 1 version for Windows, 1999).

The model identifies the point at which there is a change in the pattern of response with the variables and thus provides 2 seg-ments of the data series from the maximum likelihood principle, i. e., consists of 2 stages. In the first stage, there is the adjustment of the model for all possible partitions of data without consider-ing the continuity constraint on the point of change. In the sec-ond stage, the partitions that generate inadmissible adjustments are incorporated and new adjustments are made.

●▶ Fig. 1 illustrates the application of the Hinkley method of one

volunteer for the HR (bpm), V˙CO2 (mL · min − 1) and RMS index of EMG (µV) variables, and the change point was identified in each variable.

Statistical analysis

Based on a previous pilot study, the sample size was determined using G * Power (3.1.13 for Windows) comparing values for V˙O2 at the ventilatory anaerobic threshold with the proposed param-eters using the mathematical method applied to the HR, V˙CO2, and RMS index of EMG, and for the current study that would provide sufficient statistical power (β = 0.8) to detect a statistical difference (α = 0.05) was estimated to be 8 subjects in each group. The normal distribution of data was verified by the Shap-iro-Wilk test. Unpaired Student’s t-test was used to compare the anthropometric characteristics between the groups. 2-way ANOVA for repeated measurements was used to compare the variables HR, V˙O2, V˙CO2, V˙E, and RMS between different meth-ods to determine AT and compare relative and absolute values corresponding to intensity of the AT by different methods of identification. The Tukey-Kraemer post-hoc test was used to identify the differences. The correlation of variables at AT between the ventilatory method and Hinkley’s mathematical model was performed using Pearson’s correlation. Additionally, the degree of concordance between the methods used to deter-mine the AT was evaluated by the Bland-Altman concordance analysis [1]. The probability of Type 1 error occurrence was established at 5 % for all tests (α = 0.05).

Results

▼

Study population

●▶ Table 1 lists age, anthropometric characteristics, V˙O₂ peak in

CPX and aerobic classification for both groups, risk factors and medications of the MIG. There were no significant differences between the HG and MIG in relation to age and anthropometric measures. Regarding the V˙O2 peak, as expected, HG was signifi-cantly higher than MIG, and in relation to aerobic classification, all subjects of the HG and 4 subjects of the MIG had a “weak” classification according to the AHA [4], and 4 remaining subjects of the MIG had a “very weak” classification.

Variables at AT obtained by different methods of

identification

The values of HR, V˙O2, V˙CO2, V˙E and RMS at AT obtained from the 2 different methods (VM and Hinkley’s mathematical model) are shown in ●▶ Table 2. The mathematical method applied to HR,

V˙CO2 and the RMS index and the VM showed no significant dif-ference in the values between the methods in both groups (VM vs. mathematical model) and no difference between the 2 groups (HG vs. MIG) for all variables studied in the 2 methods.

Relationship between methods of AT determination

The absolute values (load in W) and relative values (% of maxi-mal power achieved at peak) corresponding to intensity of the AT by different methods of identification are shown in ●▶ Table 3.

No significant differences were found considering the 4 varia-bles that identify the AT and the different groups.

●▶ Table 4 shows the Pearson correlation coefficients and the p

value between Hinkley’s mathematical model (H-HR, H-V˙CO2, and H-RMS) and the VM for AT determination. With respect to HR, V˙O2, and RMS, we observed moderate to strong positive and significant correlations in all variables.

The analysis of concordance between methods to determine the AT was carried out by Bland-Altman plotting, considering the VM as the “gold standard” method in this study; VM vs. H-HR, VM vs. H-V˙CO2, and VM vs. H-RMS were plotted considering the intensity of effort corresponding to AT. As shown in ●▶ Fig. 2a, the

mean of the differences for identifying AT by the VM and H-HR methods was 4.1 ± 12.3 W; between VM and H-V˙CO2 was 2.4 ± 7.5 W ( ●▶ Fig. 2b), and between VM and H-RMS was − 1.1 ± 10.7 W ( ●▶ Fig. 2c). Moreover, there was significant correla-Table 1 Age, anthropometric characteristics, oxygen consumption peak ob-tained in cardiopulmonary exercise testing, aerobic classification, risk factors and medications of volunteers.

HG (n = 9) MIG (n = 8) Age and Anthropometric measures

age, years 54.4 ± 2.9 57.3 ± 3.5

height, cm 166.0 ± 5.0 168.0 ± 6.0

weight, kg 71.8 ± 9.1 80.2 ± 8.6

BMI, kg/m2 _{25.9 ± 2.7 28.3 ± 3.0}

V˙O2peak and Aerobic Classification (AHA)

V˙O2 peak, mL.min − 1 1 579.4 ± 260.9 * 1 301.1 ± 234.6 V˙O2 peak, mL.kg.min − 1 21.6 ± 1.3* 16.2 ± 3.9

weak, n ( %) 9 (100) 4 (50) very weak, n ( %) 0 (0) 4 (50) Risk Factors dyslipidemia, n ( %) 0 4 (50) hypertension, n ( %) 0 5 (62.5) diabetes, n ( %) 0 0 hypo or hyperthyroidism, n ( %) 0 0 Medications anti-arrhythmic, n ( %) – 1 (12.5) ACE inhibitors, n ( %) – 3 (37.5) diuretics, n ( %) – 1 (12.5) anticoagulants, n ( %) – 7 (87.5) lipid lowering, n ( %) – 4 (50)

Data are presented as mean ± standard deviation. HG = healthy group; MIG = myo-cardial infarction group; BMI = body mass index; AHA = American Heart Association; V˙O2 = oxygen consumption; n = number of subjects; % = percentage in relation to number of subjects of group; ACE = angiotensin conversion enzyme. * Significant difference in relation to MIG (Student t-test)

Fig. 1 Graphics representation of change point applied to heart rate (HR) (graphic a), carbonic gas production (V˙CO2) (graphic b) and elec-tromyography signals reported as root mean square values (RMS EMG) (graphic c) of one representative volunteer (healthy subject).

Healthy a b c Time (s) Time (s) Time (s) HR (bpm ) RMS EMG (uV) 140 130 120 110 100 90 1600 1400 1200 1000 800 600 400 120 100 80 60 500 600 700 800 500 600 700 800 500 600 700 800 VC O2 (mL/min)

tion between VM and H-HR (r = 0.59, p = 0.01), VM and H-V˙CO2 (r = 0.83, p = 0.01), and between VM and H-RMS (r = 0.75, p = 0.01; ●▶ Fig. 2d, e, f respectively). It is noteworthy that, in the

comparative analysis between the methods, only one subject of HG presented in H-RMS value outside the concordance interval, overestimating in 28 W (16 % of the maximum power achieved) compared to that obtained in the VM. On the other hand, only one subject of MIG had overestimated value in 21 W (27 % of peak power) and 13 W (14 % of peak power) in H-HR H-V˙CO2 respectively, compared to the VM.

Discussion

▼

Summary of findings

The main finding of this study was that no significant differences were observed between the VM (i. e., the gold standard method

in this study) and the Hinkley method for determining the AT; moreover, the concordance and correlations between the meth-ods were good. Nevertheless, it is important to stress that Hin-kley method can be applied not only to ventilatory variables such as V˙CO2, but also other variables like HR and RMS index. These results indicate that the mathematical model is suitable for detecting the AT in both healthy and patients with MI with low physical capacity.

Anthropometric variables, V˙O

2

peak, aerobic

classification, risk factors and medications

No difference between the groups in relation to age and anthro-pometric measures suggests that the groups are homogeneous, this is important because several studies have shown that fac-tors such as age, gender and anthropometric characteristics, should influence the physiological variables [4]. Regarding the aerobic classification and V˙O2 peak, all subjects had low physical capacity (weak and very weak classification according to the AHA) and MIG had lower values of V˙O2 than HG. As expected, the subject of MIG showed higher prevalence of risk factors for car-diovascular disease, however, blood lipid levels and blood pres-sure were well controlled by medications in MIG. Despite being a medication widely used in patients with myocardial infarction, none of subjects were using beta-blockers that interferes in car-diac autonomic modulation and consequently could influence the mathematical adjustments used to determine the AT by HR.

Variables at AT by different methods of identification

The similarity of the values of the variables at AT in both meth-ods ( ●▶ Table 2), such as the positive relationship between the

2 methods used for determining the AT, suggest that there is a relationship between cardiorespiratory, metabolic and neuro-muscular responses, and that these integrated responses are clearly perceived mainly from the intensity corresponding to AT, at which aerobic metabolism is supplemented by anaerobic metabolism for energy production. From this intensity, subse-quent adjustments are possibly related to metabolic changes that occur in the muscles during exercise, which are detected by receptors (type III and IV fibers) that are capable of sending this Table 3 Comparison of relative and absolute power values corresponding to

anaerobic threshold by different methods.

Hinkley mathematical model

VM H-HR H-V.CO2 H-RMS HG relative values (%) 45 ± 6 44 ± 7 43 ± 4 47 ± 7 absolute values (W) 64 ± 11 63 ± 10 61 ± 10 67 ± 15 MIG relative values (%) 47 ± 10 42 ± 9 46 ± 10 46 ± 10 absolute values (W) 58 ± 16 51 ± 14 57 ± 14 57 ± 16

Data are presented as mean ± SD. HG = healthy group; MIG = myocardial infarction group; VM = ventilatory method; H-HR = Hinkley mathematical model applied to heart rate data; H-V˙CO2 = Hinkley mathematical model applied to carbonic gas pro-duction data, H-RMS = Hinkley mathematical model applied to electromyographic signals reported as root mean square values. No significant differences, considering the methods to identify the anaerobic threshold and the different groups (2-way ANOVA for repeated measures)

Table 4 Correlation coefficients between mathematical model and ventila-tory method (VM) of anaerobic threshold determination.

HR (bpm) VM Mathematical model r p H-HR 0.85 0.0001 H-V˙CO2 0.91 0.0001 H-RMS 0.75 0.0005 V˙O2(mL.kg.min − 1) VM Mathematical model r p H-HR 0.85 0.0001 H-V˙CO2 0.91 0.0001 H-RMS 0.75 0.0005 RMS (µV) VM Mathematical model r p H-HR 0.85 0.0001 H-V˙CO2 0.91 0.0001 H-RMS 0.75 0.0005

HR = Heart rate; V˙O2 = oxygen consumption; RMS = root mean squared; H-HR = Hin-kley mathematical model applied to HR data; H-V˙CO2 = Hinkley mathematical model applied to V˙CO2 data; H-RMS = Hinkley mathematical model applied to electromyo-graphic signals reported as root mean square values (Pearson Correlation)

Table 2 Values at anaerobic threshold obtained by the ventilatory and Hinkley mathematical methods.

Hinkley mathematical model

VM H-HR H-V.CO2 H-RMS HG (n = 9) HR, beat/min 95 ± 12 93 ± 10 95 ± 10 97 ± 8 V˙O2, ml/min 777 ± 140 774 ± 128 751 ± 135 813 ± 173 V˙O2, ml/kg/min 11 ± 2 11 ± 2 10 ± 2 11 ± 2 V˙CO2, ml/min 760 ± 128 760 ± 113 730 ± 128 806 ± 192 V˙E, l/min 24 ± 5 24 ± 4 23 ± 4 25 ± 5 RMS, µV 1.2 ± 0.2 1.1 ± 0.2 1.1 ± 0.2 1.2 ± 0.2 MIG (n = 8) HR, beat/min 92 ± 17 89 ± 20 89 ± 15 100 ± 25 V˙O2, ml/min 752 ± 113 728 ± 147 726 ± 93 894 ± 284 V˙O2, ml/kg/min 10 ± 2 9 ± 2 9 ± 1 11 ± 4 V˙CO2, ml/min 739 ± 108 714 ± 178 707 ± 101 943 ± 369 V˙E, l/min 26 ± 3 26 ± 5 26 ± 4 33 ± 11 RMS, µV 0.9 ± 0.2 0.8 ± 0.1 0.9 ± 0.2 0.8 ± 0.2

Data are presented as mean ± standard deviation. HG = healthy group; MIG = myocar-dial infarction group; VM = ventilatory method; H-HR = Hinkley mathematical model applied to heart rate (HR) data; H-V˙CO2 = Hinkley mathematical model applied to carbonic gas production (V˙CO2) data, H-RMS = Hinkley mathematical model applied to electromyographic signals reported as root mean square (RMS) values; V˙O2 = oxy-gen consumption, V˙E = minute ventilation. No significant differences, considering the different methods to identify the AT in both groups and no significant difference between groups (2-way ANOVA for repeated measures)

information to the central nervous system [17], which then responds by adjusting cardiorespiratory function in an attempt to maintain homeostasis [17]. However, our results are in accordance with other studies that evaluated the responses pat-tern of cardiac, ventilatory, metabolic [13, 21, 23, 24] and muscle variables [6, 18] during adjustment to exercise in aerobic to anaerobic metabolism transition.

Furthermore, another important factor that can justify the inte-gration of the responses of cardiac, ventilatory and muscle vari-ables is that at higher loads (from AT intensity) the recruitment of motor units increases in order to provide greater muscular strength to sustain the increased demand imposed by exercise; then, there is initially greater recruitment of oxidative fibers and thereafter, due to the depletion of aerobic metabolism, more type IIx fibers (glycolytic) are recruited in an attempt to meet the increased load, which promotes metabolic changes and stimulates the central nervous system to promote adjustments in the ventilation and cardiac autonomic nervous systems [6, 21, 22, 24].

Relationship between methods of AT determination

As shown in ●▶ Table 3, the AT was between 51 and 57 W in the

MIG and 61 and 67 W in the HG, with no differences between the methods of analysis. In addition, methods using the Hinkley mathematical model proved to be consistent with the VM, with good concordance between the VM and the Hinkley models ( ●▶ Fig. 2a, b, c) and good correlations ( ●▶ Fig. 2d, e, f). In this

sense, the change in response patterns of cardiorespiratory and

muscle variables that occur in the given time of incremental exercise may be detected by the mathematical model used in this study. Hinkley’s mathematical model is based on the maxi-mum likelihood method for detecting a change point in the behavior pattern of a series of data collected during exercise, therefore regardless the variable analyzed, the model detects the breakage of linearity during the test. However, it is important to emphasize that although there is no significant difference between the ventilatory method and the Hinkley model, besides the good correlation and concordance between the methods in determining the AT, it is important to consider that when observed individually, 2 individuals (1 of HG and 1 of MIG) had load values in Watts out of the concordance interval, with the results obtained with the Hinkey method overestimated com-pared to the ventilatory method, however these individuals rep-resent only 11 % of the sample.

Crescêncio et al. [3] studied a bi-segmental mathematical model for AT determination using the residual sum of squares method. A good correlation between a linear-linear fitting and visual method applied to V˙CO2 data was found in healthy subjects. The mathematical model applied underestimated AT values; how-ever, there were no significant differences in relation to the gold standard method and our results are consistent with this study. Other studies with cyclists [5] and rowers [16] also found agree-ment between the point of deflection of the HR and the AT. When studying paraplegic athletes and non-paraplegics, researchers found a change point in the HR response pattern during the incremental protocol, suggesting that the HR

break-Fig. 2 Bland and Altman plotting shows the concordance between VM and H-HR (graphic a), between VM and H-V˙CO2 (graphic b) and between VM and H-RMS (graphic c). Bias = mean of the differences between the means and ± 1.96 = con-cordance limit of 95 %. On the right side, the significant and positive relationship between VM and H-HR (graphic d), between VM and H-V˙CO2 (graphic e) and between VM and H-RMS (graphic f). (○) Healthy group; (●) Myocardial infarction group; VM = ventilatory method; H-HR = Hinkley mathematical model applied to heart rate (HR) data; H-V˙CO2 = Hinkley mathematical model applied to carbonic gas production (V˙CO2) data, H-RMS = Hinkley mathematical model applied to electromyographic signals reported as root mean square (RMS) values. 40 90 80 70 60 50 40 30 20 90 20 30 20 10 0 –10 –20 –30 15 10 5 0 –5 –10 –15 80 70 60 50 40 30 20 90 80 70 60 50 40 30 20 +1.96SD 28.3 Mean 4.1 –20.1 90 90 80 70 60 50 40 30 90 100 80 70 60 50 40 30 20 30 40 50 60 70 80 90 100 80 70 60 50 80 70 60 50 40 30 20 30 40 80 70 60 50 20 30 40 Average of VM and H-HR (W) Average of VM - H-RMS (W) Average of VM - H-VCO2 (W) H-HR (W) H-RMS (W) H-HVCO2 (W) R=0.59 p=0.01 R=0.83 p=0.01 R=0.75 p=0.01 –1.96SD 17.1 +1.96SD 20.0 +1.96SD –22.1 –1.96SD Mean –1.96SD 2.4 Mean –1.1 –12.4 BIAS=4.1 W; SD=12.3W BIAS=2.4 W; SD=7.5W BIAS=–1.1 W; SD=10.7W VM - H-HR (W ) VM - H-RMS (W )V M - HV CO2 (W ) VM (W ) VM (W ) VM (W ) 30 20 10 0 –10 –20 –30 a b c d e f

point observed during continuous ramp arm exercise can be used to determine the AT. This method, however, tends to over-estimate threshold values [19]. In contrast, Higa et al. [9] studied young and postmenopausal healthy and sedentary women dur-ing continuous incremental cardiopulmonary testdur-ing and found that the mathematical linear regression bi-segmental model can be used for AT identification, showing a good correlation with the gold standard method.

In the present study, we used RMS index of electromyography signal based on many studies that used time domain parameters of EMG for evaluation of change points which occurs due to the additional recruitment of type II fibers during the exercise [18], and no significant differences were found between the Hinkley mathematical model applied to RMS index and the ventilatory method. This finding confirms that there is interconnection between the muscular and respiratory systems and as previ-ously described, the additional recruitment of type II fibers, pro-vokes metabolic changes that result in increased ventilation in attempts to maintain homeostasis [6].

Limitations of this study

A significant limitation of this study was the measurement of the EMG signal that was recorded from only one lower limb muscle, the vastus lateralis. However, the study by Hug et al. [12] showed that the EMG threshold was identified for the vastus lat-eralis and femoral biceps, while other muscles involved in the act of pedaling did not always show breakpoints in the RMS index.

Perspectives and practical relevance

To our knowledge, this was the first study to identify the AT by mathematical models in myocardial infarction subjects. In addi-tion, this study clearly shows that a simple and low-cost math-ematical tool applied in a population of healthy and myocardial infarction subjects was able to determine the AT using the EMG threshold and HR threshold as an alternative method and this fact has great practical relevance and potential impact to area of sports medicine.

Conclusion

▼

The aerobic to anaerobic metabolism transition, changes in car-diopulmonary and muscular variable response patterns occur at similar moments during dynamic exercise, allowing the use of mathematical model to indirectly detect AT using HR and the RMS index, thus providing a non-invasive and low-cost method that can be used in healthy and myocardial infarction subjects.

Acknowledgements

▼

This study was funded by FAPESP (01/07427-2), CAPES (PNPD 23038.006927/2011-92), and CNPQ (311938/2013-2). The authors wish to thank Anielle C. M. Takahashi and Fabiana M. Darezzo for technical assistance.

Conflict of interest: The authors have no conflict of interest to

declare.

Affiliations

1 _{Department of Physical Therapy, Cardiovascular Physical Therapy} Laboratory, Nucleus of Research in Physical Exercise, Federal University of Sao Carlos – UFSCar, Sao Carlos, Sao Paulo, Brazil

2 _{Department of Physical Therapy, Federal University of Triangulo Mineiro,} Uberaba, Minas Gerais, Brazil

3 _{Department of Physical Therapy, Integrated College Einstein of Limeira,} Limeira, Sao Paulo, Brazil

4 _{Laboratory of Exercise Physiology, Faculty of Physical Education, Estadual} University of Campinas, Campinas, Sao Paulo, Brazil

5 _{Department of Statistics, Federal University of Sao Carlos, Sao Carlos, Sao} Paulo, Brazil

6 _{Department of Clinical Medicine, Laboratory of Exercise Physiology, Division} of Cardiology, Faculty of Medicine of Ribeirao Preto, University of Sao Paulo, Ribeirao Preto, Brazil

References

1 Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 307–310

2 Bunc V, Hofmann P, Leitner H, Gaisl G. Verification of the heart rate threshold. Eur J Appl Physiol 1995; 70: 263–269

3 Crescêncio JC, Martins LEB, Murta LO Jr, Antloga CM, Kozuki RT, Santos MDS, Marun Neto JA, Maciel BC, Gallo L Jr. Measurement of anaerobic threshold during dynamic exercise in healthy subjects: Comparison among visual analysis and mathematical models. Comput Cardiol 2003; 30: 801–804

4 Fletcher GF, Balady G, Blair SN, Blumenthal J, Caspersen C, Chaitman B, Epstein S, Sivarajan Froelicher ES, Froelicher VF, Pina IL, Pollock ML. Statement on exercise: benefits and recommendations for physical activity programs for all Americans. A statement for health profes-sionals by the Committee on Exercise and Cardiac Rehabilitation of the Council on Clinical Cardiology, American Heart Association. Cir-culation. 1996: 94: 857–862

5 Grazzi G, Mazzoni G, Casoni I, Uliari S, Collini G, Heide Lv, Conconi F. Identification of a VO2 deflection point coinciding with the heart rate deflection point and ventilatory threshold in cycling. J Strength Cond Res 2008; 22: 1116–1123

6 Hanon C, Thépaut-Mathieu C, Hausswirth C, Le Chevalier JM. Electro-myogram as an indicator of neuromuscular fatigue during incremen-tal exercise. Eur J Appl Physiol 1998; 78: 315–323

7 Harriss DJ, Atkinson G. Ethical standards in sport and exercise science research: 2014 update. Int J Sports Med 2013; 34: 1025–1028 8 Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G. Development of

rec-ommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol 2000; 10: 361–374

9 Higa MN, Silva E, Neves VFC, Catai AM, Gallo L Jr, Silva de Sá MF. Com-parison of anaerobic threshold determined by visual and mathemati-cal methods in healthy women. Braz J Med Biol Res 2007; 40: 501–508 10 Hofmann P, Bunc V, Leitner H, Pokan R, Gaisl G. Heart rate thresh-old related to lactate turn point and steady-state exercise on a cycle ergometer. Eur J Applied Physiol 1994; 69: 132–139

11 Hopker JG, Jobson SA, Pandit JJ. Controversies in the physiological basis of the anaerobic threshold and their implications for clinical cardio-pulmonary exercise testing. Anaesthesia 2011; 66: 111–123 12 Hug F, Laplaud D, Savin B, Grelot L. Occurrence of electromyographic

and ventilatory thresholds in professional road cyclists. Eur J Appl Physiol 2003; 90: 643–646

13 Karapetian GK, Engels HJ, Gretebeck RJ. Use of heart rate variability to estimate LT and VT. Int J Sports Med 2008; 29: 652–657

14 Kodama S, Saito K, Tanaka S, Maki M, Yachi Y, Asumi M, Sugawara A, Totsuka K, Shimano H, Ohashi Y, Yamada N, Sone H. Cardiorespiratory fitness as a quantitative predictor of all-cause mortality and cardio-vascular events in healthy men and women: a meta-analysis. JAMA 2009; 301: 2024–2035

15 Meyer T, Görge G, Schwaab B, Hildebrandt K, Walldorf J, Schäfer C, Kin-dermann I, Scharbag J, KinKin-dermann W. An alternative approach for exercise prescription and efficacy testing in patients with chronic heart failure: a randomized controlled training study. Am Heart J 2005; 149: 926e1–926e7

16 Mikulic P, Vucetic V, Sentija D. Strong relationship between heart rate deflection point and ventilatory threshold in trained rowers. J Strength Cond Res 2011; 25: 360–366

17 Rowell LB. Reflex control of the circulation during exercise. Int J Sports Med 1992; 13: S25–S27

18 Sbriccoli P, Sacchetti M, Felici F, Gizzi L, Lenti M, Scotto A, DeVito GJ. Non-invasive assessment of muscle fiber conduction velocity during an incremental maximal cycling test. Electromyogr Kinesiol 2009; 19: 380–386

19 Schmid A, Huonker M, Aramendi JF, Kluppel E, Barturen JM, Grathwoh D, Schmidt-Trucksass A, Berg A, Keul J. Heart rate deflection compared to 4 mmol x l(-1) lactate threshold during incremental exercise and to lactate during steady-state exercise on an arm-cranking ergometer in paraplegic athletes. Eur J Appl Physiol 1998; 78: 177–182 20 Silva E, Catai AM, Trevelin LC, Guimarães JO, Silva LP Jr, Silva LMP,

Oliveira L, Milan LA, Martins LEB, Gallo L Jr. Design of a computerized system to evaluate the cardiac function during dynamic exercise. In: Proceedings of World Congress of Medical Physics and Biomedical Engineering 1994; 39a: 409–409

21 Simões RP, Mendes RG, Castello V, Machado HG, Almeida LB, Baldissera V, Catai M, Arena R, Borghi-Silva A. Heart-rate variability and blood-lactate threshold interaction during progressive resistance exercise in healthy older men. J Strength Cond Res 2010; 24: 1313–1320 22 Simões RP, Bonjorno JC Jr, Beltrame T, Catai AM, Arena R, Borghi-Silva A.

Slower heart rate and oxygen consumption kinetic responses in the on- and off-transient during a discontinuous incremental exercise: effects of aging. Braz J Phys Ther 2013; 17: 69–76

23 Simões RP, Castello-Simões V, Mendes RG, Archiza B, Santos DA, Machado HG, Bonjorno JC Jr, Oliveira CR, Reis MS, Catai AM, Arena R, Borghi-Silva A. Lactate and heart rate variability threshold during resistance exercise in the young and elderly. Int J Sports Med 2013; 34: 991–996 24 Simões RP, Castello-Simões V, Mendes RG, Archiza B, Dos Santos DA,

Bonjorno JC Jr, de Oliveira CR, Catai AM, Arena R, Borghi-Silva A. Iden-tification of anaerobic threshold by analysis of heart rate variability during discontinuous dynamic and resistance exercise protocols in healthy older men. Clin Physiol Funct Imaging 2014; 34: 98–108 25 Soler AM, Folledo M, Martins LEB, Lima-Filho EC, Gallo L Jr. Anaerobic

threshold estimation by statistical modelling. Braz J Med Biol 1989; 22: 795–797

26 Spruit MA, Wouters EF, Eterman RM, Meijer K, Wagers SS, Stakenborg KH, Uszko-Lencer NH. Task-related oxygen uptake and symptoms dur-ing activities of daily life in CHF patients and healthy subjects. Eur J Appl Physiol 2011; 111: 1679–1686

27 Zamunér AR, Catai AM, Martins LE, Sakabe DI, Da Silva E. Identification and agreement of first turn point by mathematical analysis applied to heart rate, carbon dioxide output and electromyography. Braz J Phys Ther 2013; 17: 614–622