European Heart Journal (1991) 12, 1055-1063
Multicentre study of the determination of peak oxygen uptake and ventilatory threshold during bicycle exercise in chronic heart failure Comparison of graphical methods, interobserver variability and influence of the exercise protocol A. COHEN-SOLAL*, F. ZANNADf, J.-G. KAYANAKISj, P. GUERET§, J. F . AUPETIT|| AND H. KOLSKY^
for the VO2 French Study Group *Cardiology, CHU Bichat, Paris, France; f Clinical pharmacology and cardiology, Nancy, France; \Cardiology, Clinique Paulmy, Bayonne, France; ^Cardiology, CHU Dupuytren, Limoges, France; \\Cardiology, HopitalSaint-Joseph, Lyon, France and\Laboratoires Hoechst, Puteaux, France KEY WORDS: Ventilatory threshold, peak oxygen uptake, respiratory gas analysis, congestive heart failure, exercise testing. Assessment of the ventilatory threshold ( VT) has been proposed to assess exercise tolerance more objectively, particularly in clinical trials, but reproducibility, interobserver variability andfeasibility of the graphical methods for determination of VThave not been properly studied in patients with chronic heart failure (CHF). Fifty-one patients with mild to moderate CHF (mean peak oxygen uptake (VO2): 20-5 ml .min''. kg~') were assessed during two consecutive bicycle exercise tests within 8 days. Two graded exercise protocols were compared with stages of 30 W every 3 min (22 patients) or 10 W\ min (29 patients). VT was determined separately by five trained physicians usingfivedifferent graphical methods. The 'crossing method' (first crossing of the VCO2 and VO2 curves) yielded the highest rate of determination (88%) but tended to overestimate the mean VT. The VE method (disproportionate increase of ventilation relative to VO,) produced the best interobserver agreement (coefficient of variation =78%). Peak VO2 was very highly reproducible in both exercise protocols (relative difference test 2-test 1/test 1 = —0-32% for the 30 W 3 min protocol; +2-18% for the 10 W. min~' protocol). The reproducibility of VT was slightly lower regardless of the graphical method used to determine it (relative differences variedfrom —3-3% to +7-3%). Therefore, peak VO2 appears more suitable than VTfor assessment of exercise tolerance in CHF.
Introduction
to medical therapy in C H F . However, reliability of these
Exercise testing is widely used to evaluate the symptoms and the functional capacity of patients with heart disease, In chronic heart failure (CHF) it is now well documented that measurements of left ventricular function at rest correlate poorly with functional class and exercise tolerance"1, mainly because functional impairment is more dependant on abnormalities of the peripheral circulation than on the degree of alteration of cardiac function'2-3'. Recently, there has been renewed interest in the
measurements in a multicentre study remains an unsolved jssue particularly when different equipment is used. feasibility, interobserver variability, reproducibility of measurements have not yet been extensively studied in P. a t i e n t s W l t l | C H R Moreover, the importance of the exerclse P r o t o c o 1 1 S , n o t a s * e l 1 *hara*e"zfduin t h e s e p a t ' e " t S a s 1S in " normal trained subjects"4-'6'. The purpose of this study was to determine the most reliable graphical method for measuring VT and to compare th?reproduci-
measurement of maximum oxygen consumption (VO2
J1'1^
max)" and ventilatory threshold (VT)' as possible objective, symptom-independent and reproducible criteria of evaluation of cardiorespiratory functional reserve'6"10'.
y , multicentre study, with two different mcremental protocols in patients with CHF.
In normal subjects, the ventilatory threshold can easily be recognized during exercise by respiratory gas analysis'" 1 2 1 . However, recent studies have emphasized important conceptual a n d methodological problems associated with the determination of the ventilatory threshold" 3 1 , Despite those limitations, these variables are increasingly used for assessing the exercise tolerance a n d the response submmed for publican™ on 13 June 1990, and in revised form 21 September 1990. Correspondence: Dr Alain Cohen-Solal, Service de Cardiologie, Hopital Biehat, 46 rue Henri Huchard, 75018 Paris, France.
Patients and methods Fifty-one patients (44 males a n d 7 females), aged 30 t o 76 years, entered the study in three different centres, All h a d mild to m o d e r a t e stable chronic congestive heart failure (class II o r III of the N Y H A classification) a n d exercise limitation d u e t o fatigue o r dyspnoea. T h e aetiology of heart failure was coronary artery disease, primary dilated c a r d i o m y o p a t h y , hypertension a n d valvular regurgitation. All patients h a d an ejection fraction, determined b y angiography Or radionuclide veiltriculo g r a p h y less than 5 0 % o r a n enlarged left ventricle
1
51
bic
V T a d peak
"
° X . ygen U p t a ? ^
V 2
° ] iunng
cle exer cise i n a
0I95-668X/91/I01055+09 $03.00/0
Downloaded from https://academic.oup.com/eurheartj/article-abstract/12/10/1055/440724 by University of Adelaide user on 04 May 2018
© 1991 The European Society of Cardiology
1056
A.Cohen-SolaletaL
(echographic end-diastolic left ventricular diameter > 3 cm . m~2). Patients with myocardial infarction during the previous 2 months, myocardial ischaemia or arrhythmias during exercise, pulmonary disease or exercise intolerance for any reason other than fatigue or dyspnoea were excluded from the study. EXERCISE TESTING
All patients had been previously familiarized with bicycle exercise and measurement of respiratory gas exchanges. Patients continued their medications on the days of exercise tests and all tests were performed in the sitting position at the same time of day at least 3 h after the last meal. Patients were encouraged to exercise until they felt unable to continue. Each patient underwent two exercise tests using the same protocol, within less than 8 days. Patients were divided into two groups: group I patients (n = 22) performed tests beginning at a 30 W load with subsequent increments of 30 W every 3 min; group II patients (n = 29) performed tests beginning at 10 W with increments of 10 W every minute. MEASUREMENTS OF RESPIRATORY GAS EXCHANGES
Respiratory gas analysis was carried out in the same way in each centre, following rules established during preliminary sessions between physicians. Each patient breathed through a three-way low resistance valve (Hans Rudolph) with a clamp placed on the nose. Mixed expired air was sampled by a computer and analysed every 15 s (Centre 1: MMC Horizon II, SensorMedics), or every 30 s (Centre 2: Ergooxyscreen, Jaeger; Centre 3: Oxycon IV, Mijnhardt). Gas analysers were carefully calibrated before each test. The following graphs were computerized and plotted at the end of the experiment: (1) VO2 and carbon dioxide production (VCO2) against time; (2) minute ventilation (VE) against time; (3) respiratory exchange ratio (RER) against time; (4) ventilatory equivalents for O2 (VE/VO2) and for CO2 (VE/VCO2) against time. Peak oxygen uptake (peak VO2) was denned as the highest value of VO2 reached at the end of the exercise. Ventilatory threshold (VT) was determined using five of the current graphical methods'5'8-10' (Fig. 1) to show the VO2 level reached when one of the following occurs: the first crossing of the VO2 and VCO2 curves, when RER = 1 (the crossing method); a disporportionate increase in CO2 production relative to VO2 (the VCO2 method); an increase in the ventilatory equivalent for oxygen relative to the ventilatory equivalent for CO2 (the VE/VO2 method); a non linear increase in respiratory exchange ratio (the RER method); a disporportionate increase in VE relative to VO2 (the VE method). The curves obtained after each exercise test for each patient were analysed in a blinded fashion byfiveseparate observers familiar with the technique of respiratory gas analysis. STATISTICAL ANALYSIS
Results are expressed as mean + standard deviation (SD). The feasibility of the graphic determinations of VT
Downloaded from https://academic.oup.com/eurheartj/article-abstract/12/10/1055/440724 by University of Adelaide user on 04 May 2018
was assessed by: the rate of determination of VT for each single graphical method (rate of determination = number of observations where it was possible to detect a VT x 100/ total number of observations); the proportion of five full determinations for each graphical method (percentage of patients where all five observers could detect a VT using the same single graphical method). Interobserver variability for each graphical method was assessed by the calculation of the coefficients of variations (CV) (CV = standard deviation of the mean VT/mean VT determined by all observers using a single graphical method) and the distribution of the CV by their median and their 75% quartiles (Q75). Variability among the VT values as determined by five different graphical methods was assessed in individual patients and for each method by the deviations from the mean of thefivegraphical determinations. The reproducibility of VT values and peak VO2 was assessed by linear regression analysis, paired Student's t-test and 95% confidence intervals of the mean'1718'. The variation of peak VO2 and VT between two consecutive stress tests was expressed (in percentage) as the mean of the relative difference of the values between the two tests for each patient (test 2-test I/test 1). VT and peak VO2 reproducibility were assessed separately for group I (30 W every 3 min) and group II (10 W . min"') and displayed by linear regression and box-plots'19'.
Results Peak VO2 in our study was only moderately reduced compared to normal subjects (20-2 ml. mn~'. kg" 1 in group I and 20-9 ml. mn~' . kg"' in group II), confirming the mild to moderate cardiac impairment under therapy of our patients.
GRAPHICAL DETERMINATION OF THE VENTILATORY THRESHOLD
Graphical determination of the VT was carried out in the 22 group 1 patients by the five physicians. Heart rate and ventilatory parameters obtained at maximal exercise are reported in Table 1. Values of heart rate and respiratory exchange ratio observed at peak exercise confirm that patients performed maximal exercise, although one patient stopped exercising before a RER value of 1. Among thefivegraphical methods which were tested, the crossing method yielded the highest rate of determination although some patients stopped exercising before a RER value of 1 (Table 2). Indeed at least one observer could detect a VT using this method in 88% of the observations. The other methods had lower determination rates (from 72 to 78%). With the crossing method all five observers detected a VT in 73% of the patients. For the VE and VE/VO2 methods, the rates of determination were lower (59 and 54% respectively). Interobserver variability was different for each graphical method. The VE method produced the best agreement
Oxygen uptake during bicycle exercise in CHF 1057
1-2 r
E
1-2
0 9 -
-
0-9 E
-
0-6
-
0-3
VC02 method \
0-6 o > 0-3 Crossing point 00
1
1
1
1
0-0
1
10
80
80
VE/V0 2 method
60 -
- 60
40 --
- 40
20
- 20
1
|
|
|
O
|
60
I 5
RER method 40
a: 1-0
II
20
VE method
4
6
10
0 5
Time (min)
Figure 1 Example of the graphical methods used for the determination of the ventilatory threshold (arrow). Upper: VO2: oxygen uptake; VCO2: carbon dioxide production; Crossing: first time of crossing of the VO2/t and VCOj/t curves. Middle: VE/VO2 and VE/VCO2: ventilatory equivalent for 0 2 and for CO 2 . Lower: VE: minute ventilation; RER: respiratory exchange ratio.
between observers since the average coefficient of variation was 7-8%. More importantly, 75% of the coefficients of variation for this criteria (75% quartile) were lower than 9-8%. When VT was calculated for each patient as the mean of the values yielded by thefivedifferent graphical methods, we noticed that the VCO2 and the VE methods tended to underestimate the VT (respectively 53% and 55% of the determinations were at least 2-5% lower than the mean values). In contrast, the RER and the crossing methods
tended to overestimate the VT (respectively 51 and 56% of the determinations were at least 2-5% higher than the mean values).
REPRODUCIBILITY OF PEAK VO, AND VT
Since the VE and the crossing methods offered the best compromise for VT determination (highest rate of determination and lowest variability), we subsequently used both for the reproducibility studies. Tables 3 and 4
Downloaded from https://academic.oup.com/eurheartj/article-abstract/12/10/1055/440724 by University of Adelaide user on 04 May 2018
1058 A. Cohen-Solal et al.
Table 1 Exercise parameters at peak exercise in patients of Group 1
PeakVO2 (ml. rain-', kg-')
MaxHR (b.min-)
Percentage of Predicted max HR
yEmax ,rl)
RERmax
57(16) 39-81
112(01) 0-92-1 31
(| m
(/o)
Mean (SD) range
20 2 ( 5 1 )
11-29
141 (20) 106-192
89(15) 69-125
VO2: oxygen uptake; HR: heart rate; VE: external ventilation; RER: respiratory exchange ratio.
Table 2 Feasibility and inlerobserver variability of the graphical determinations of the ventilatory threshold (protocol 30 W every 3 mm) n = 22 Method for determining ventilatory threshold
VCO 2
Determination rate (%) 74 Five full determination rate (%) 32 Quartiles of CV(%) Mean 70 Q75 12-5 Distribution of the deviations from the mean (%) - 2 - 5 % +2S% 29
VE
RER
VE/VO2
Crossing
76 59 7-8 9-8
72 18 7-8 11-8
78 54 9-4 18-4
88 73 8-5 13-7
55 26 19
34 15 51
38 23 39
20 24 56
CV = coefficient of variation; Q = quartiles; VO2: oxygen uptake; VCO2: carbon dioxide production; VE: minute ventilation; RER: respiratory exchange ratio; VE/VO2: ventilatory equivalent for oxygen; Crossing: crossing of the VCO2 and VO2 curves. 8 6
g o
T
4 2
-T
T +
T
1
-4
1
o
1
j_
jo o
-2
-6 -
o
1
o—
-8
•
•
0
A Peak VO2 (lOWmin"')
1
A Peak VO2 (30Wper3min)
A Ventilatory threshold I (lOWmin')
i A Ventilatory threshold (30 W per 3 mm)
Figure 2 Box-plots of the reproducibility between test 1 and test 2 of peak VO2 (A peak VOJ and ventilatory threshold (A ventilatory threshold) for the 10 W. min"1 (left) and the 30 W every 3 min (right) protocol. Difference of the values of peak VO2 and ventilatory threshold between test 2 and test 1 for each patient, expressed in ml. min" 1 . kg"1 of VO2, are plotted on the y axis. The box bounds the first and third quartiles of the data; the horizontal line within the box-plot represents the median of the differences and the cross the mean of the differences.
summarize the reproducibility data corresponding respectively to the 30W every 3min and 10W.min" 1 protocols. Peak VO2 was very highly reproducible since values varied less than 2-5% between the two subsequent tests
Downloaded from https://academic.oup.com/eurheartj/article-abstract/12/10/1055/440724 by University of Adelaide user on 04 May 2018
using the two protocols ( — 0-32% in protocol 10 W . min ' and + 2-18% in protocol 30 W every 3 min). Accordingly, confidence intervals of the mean were very small (Fig. 2) and correlation coefficients were high (0-93 and 0-97) (Figs 3 and 4).
Oxygen uptake during bicycle exercise in CHF 1059
35 -
25
OJ
20
Peak
(ml mm 1
30
15
.••-'£/?'
y = l-02x -0-40
-••-'>
Multicentre study of the determination of peak oxygen uptake and ventilatory threshold during bicycle exercise in chronic heart failure Comparison of graphical methods, interobserver variability and influence of the exercise protocol A. COHEN-SOLAL*, F. ZANNADf, J.-G. KAYANAKISj, P. GUERET§, J. F . AUPETIT|| AND H. KOLSKY^
for the VO2 French Study Group *Cardiology, CHU Bichat, Paris, France; f Clinical pharmacology and cardiology, Nancy, France; \Cardiology, Clinique Paulmy, Bayonne, France; ^Cardiology, CHU Dupuytren, Limoges, France; \\Cardiology, HopitalSaint-Joseph, Lyon, France and\Laboratoires Hoechst, Puteaux, France KEY WORDS: Ventilatory threshold, peak oxygen uptake, respiratory gas analysis, congestive heart failure, exercise testing. Assessment of the ventilatory threshold ( VT) has been proposed to assess exercise tolerance more objectively, particularly in clinical trials, but reproducibility, interobserver variability andfeasibility of the graphical methods for determination of VThave not been properly studied in patients with chronic heart failure (CHF). Fifty-one patients with mild to moderate CHF (mean peak oxygen uptake (VO2): 20-5 ml .min''. kg~') were assessed during two consecutive bicycle exercise tests within 8 days. Two graded exercise protocols were compared with stages of 30 W every 3 min (22 patients) or 10 W\ min (29 patients). VT was determined separately by five trained physicians usingfivedifferent graphical methods. The 'crossing method' (first crossing of the VCO2 and VO2 curves) yielded the highest rate of determination (88%) but tended to overestimate the mean VT. The VE method (disproportionate increase of ventilation relative to VO,) produced the best interobserver agreement (coefficient of variation =78%). Peak VO2 was very highly reproducible in both exercise protocols (relative difference test 2-test 1/test 1 = —0-32% for the 30 W 3 min protocol; +2-18% for the 10 W. min~' protocol). The reproducibility of VT was slightly lower regardless of the graphical method used to determine it (relative differences variedfrom —3-3% to +7-3%). Therefore, peak VO2 appears more suitable than VTfor assessment of exercise tolerance in CHF.
Introduction
to medical therapy in C H F . However, reliability of these
Exercise testing is widely used to evaluate the symptoms and the functional capacity of patients with heart disease, In chronic heart failure (CHF) it is now well documented that measurements of left ventricular function at rest correlate poorly with functional class and exercise tolerance"1, mainly because functional impairment is more dependant on abnormalities of the peripheral circulation than on the degree of alteration of cardiac function'2-3'. Recently, there has been renewed interest in the
measurements in a multicentre study remains an unsolved jssue particularly when different equipment is used. feasibility, interobserver variability, reproducibility of measurements have not yet been extensively studied in P. a t i e n t s W l t l | C H R Moreover, the importance of the exerclse P r o t o c o 1 1 S , n o t a s * e l 1 *hara*e"zfduin t h e s e p a t ' e " t S a s 1S in " normal trained subjects"4-'6'. The purpose of this study was to determine the most reliable graphical method for measuring VT and to compare th?reproduci-
measurement of maximum oxygen consumption (VO2
J1'1^
max)" and ventilatory threshold (VT)' as possible objective, symptom-independent and reproducible criteria of evaluation of cardiorespiratory functional reserve'6"10'.
y , multicentre study, with two different mcremental protocols in patients with CHF.
In normal subjects, the ventilatory threshold can easily be recognized during exercise by respiratory gas analysis'" 1 2 1 . However, recent studies have emphasized important conceptual a n d methodological problems associated with the determination of the ventilatory threshold" 3 1 , Despite those limitations, these variables are increasingly used for assessing the exercise tolerance a n d the response submmed for publican™ on 13 June 1990, and in revised form 21 September 1990. Correspondence: Dr Alain Cohen-Solal, Service de Cardiologie, Hopital Biehat, 46 rue Henri Huchard, 75018 Paris, France.
Patients and methods Fifty-one patients (44 males a n d 7 females), aged 30 t o 76 years, entered the study in three different centres, All h a d mild to m o d e r a t e stable chronic congestive heart failure (class II o r III of the N Y H A classification) a n d exercise limitation d u e t o fatigue o r dyspnoea. T h e aetiology of heart failure was coronary artery disease, primary dilated c a r d i o m y o p a t h y , hypertension a n d valvular regurgitation. All patients h a d an ejection fraction, determined b y angiography Or radionuclide veiltriculo g r a p h y less than 5 0 % o r a n enlarged left ventricle
1
51
bic
V T a d peak
"
° X . ygen U p t a ? ^
V 2
° ] iunng
cle exer cise i n a
0I95-668X/91/I01055+09 $03.00/0
Downloaded from https://academic.oup.com/eurheartj/article-abstract/12/10/1055/440724 by University of Adelaide user on 04 May 2018
© 1991 The European Society of Cardiology
1056
A.Cohen-SolaletaL
(echographic end-diastolic left ventricular diameter > 3 cm . m~2). Patients with myocardial infarction during the previous 2 months, myocardial ischaemia or arrhythmias during exercise, pulmonary disease or exercise intolerance for any reason other than fatigue or dyspnoea were excluded from the study. EXERCISE TESTING
All patients had been previously familiarized with bicycle exercise and measurement of respiratory gas exchanges. Patients continued their medications on the days of exercise tests and all tests were performed in the sitting position at the same time of day at least 3 h after the last meal. Patients were encouraged to exercise until they felt unable to continue. Each patient underwent two exercise tests using the same protocol, within less than 8 days. Patients were divided into two groups: group I patients (n = 22) performed tests beginning at a 30 W load with subsequent increments of 30 W every 3 min; group II patients (n = 29) performed tests beginning at 10 W with increments of 10 W every minute. MEASUREMENTS OF RESPIRATORY GAS EXCHANGES
Respiratory gas analysis was carried out in the same way in each centre, following rules established during preliminary sessions between physicians. Each patient breathed through a three-way low resistance valve (Hans Rudolph) with a clamp placed on the nose. Mixed expired air was sampled by a computer and analysed every 15 s (Centre 1: MMC Horizon II, SensorMedics), or every 30 s (Centre 2: Ergooxyscreen, Jaeger; Centre 3: Oxycon IV, Mijnhardt). Gas analysers were carefully calibrated before each test. The following graphs were computerized and plotted at the end of the experiment: (1) VO2 and carbon dioxide production (VCO2) against time; (2) minute ventilation (VE) against time; (3) respiratory exchange ratio (RER) against time; (4) ventilatory equivalents for O2 (VE/VO2) and for CO2 (VE/VCO2) against time. Peak oxygen uptake (peak VO2) was denned as the highest value of VO2 reached at the end of the exercise. Ventilatory threshold (VT) was determined using five of the current graphical methods'5'8-10' (Fig. 1) to show the VO2 level reached when one of the following occurs: the first crossing of the VO2 and VCO2 curves, when RER = 1 (the crossing method); a disporportionate increase in CO2 production relative to VO2 (the VCO2 method); an increase in the ventilatory equivalent for oxygen relative to the ventilatory equivalent for CO2 (the VE/VO2 method); a non linear increase in respiratory exchange ratio (the RER method); a disporportionate increase in VE relative to VO2 (the VE method). The curves obtained after each exercise test for each patient were analysed in a blinded fashion byfiveseparate observers familiar with the technique of respiratory gas analysis. STATISTICAL ANALYSIS
Results are expressed as mean + standard deviation (SD). The feasibility of the graphic determinations of VT
Downloaded from https://academic.oup.com/eurheartj/article-abstract/12/10/1055/440724 by University of Adelaide user on 04 May 2018
was assessed by: the rate of determination of VT for each single graphical method (rate of determination = number of observations where it was possible to detect a VT x 100/ total number of observations); the proportion of five full determinations for each graphical method (percentage of patients where all five observers could detect a VT using the same single graphical method). Interobserver variability for each graphical method was assessed by the calculation of the coefficients of variations (CV) (CV = standard deviation of the mean VT/mean VT determined by all observers using a single graphical method) and the distribution of the CV by their median and their 75% quartiles (Q75). Variability among the VT values as determined by five different graphical methods was assessed in individual patients and for each method by the deviations from the mean of thefivegraphical determinations. The reproducibility of VT values and peak VO2 was assessed by linear regression analysis, paired Student's t-test and 95% confidence intervals of the mean'1718'. The variation of peak VO2 and VT between two consecutive stress tests was expressed (in percentage) as the mean of the relative difference of the values between the two tests for each patient (test 2-test I/test 1). VT and peak VO2 reproducibility were assessed separately for group I (30 W every 3 min) and group II (10 W . min"') and displayed by linear regression and box-plots'19'.
Results Peak VO2 in our study was only moderately reduced compared to normal subjects (20-2 ml. mn~'. kg" 1 in group I and 20-9 ml. mn~' . kg"' in group II), confirming the mild to moderate cardiac impairment under therapy of our patients.
GRAPHICAL DETERMINATION OF THE VENTILATORY THRESHOLD
Graphical determination of the VT was carried out in the 22 group 1 patients by the five physicians. Heart rate and ventilatory parameters obtained at maximal exercise are reported in Table 1. Values of heart rate and respiratory exchange ratio observed at peak exercise confirm that patients performed maximal exercise, although one patient stopped exercising before a RER value of 1. Among thefivegraphical methods which were tested, the crossing method yielded the highest rate of determination although some patients stopped exercising before a RER value of 1 (Table 2). Indeed at least one observer could detect a VT using this method in 88% of the observations. The other methods had lower determination rates (from 72 to 78%). With the crossing method all five observers detected a VT in 73% of the patients. For the VE and VE/VO2 methods, the rates of determination were lower (59 and 54% respectively). Interobserver variability was different for each graphical method. The VE method produced the best agreement
Oxygen uptake during bicycle exercise in CHF 1057
1-2 r
E
1-2
0 9 -
-
0-9 E
-
0-6
-
0-3
VC02 method \
0-6 o > 0-3 Crossing point 00
1
1
1
1
0-0
1
10
80
80
VE/V0 2 method
60 -
- 60
40 --
- 40
20
- 20
1
|
|
|
O
|
60
I 5
RER method 40
a: 1-0
II
20
VE method
4
6
10
0 5
Time (min)
Figure 1 Example of the graphical methods used for the determination of the ventilatory threshold (arrow). Upper: VO2: oxygen uptake; VCO2: carbon dioxide production; Crossing: first time of crossing of the VO2/t and VCOj/t curves. Middle: VE/VO2 and VE/VCO2: ventilatory equivalent for 0 2 and for CO 2 . Lower: VE: minute ventilation; RER: respiratory exchange ratio.
between observers since the average coefficient of variation was 7-8%. More importantly, 75% of the coefficients of variation for this criteria (75% quartile) were lower than 9-8%. When VT was calculated for each patient as the mean of the values yielded by thefivedifferent graphical methods, we noticed that the VCO2 and the VE methods tended to underestimate the VT (respectively 53% and 55% of the determinations were at least 2-5% lower than the mean values). In contrast, the RER and the crossing methods
tended to overestimate the VT (respectively 51 and 56% of the determinations were at least 2-5% higher than the mean values).
REPRODUCIBILITY OF PEAK VO, AND VT
Since the VE and the crossing methods offered the best compromise for VT determination (highest rate of determination and lowest variability), we subsequently used both for the reproducibility studies. Tables 3 and 4
Downloaded from https://academic.oup.com/eurheartj/article-abstract/12/10/1055/440724 by University of Adelaide user on 04 May 2018
1058 A. Cohen-Solal et al.
Table 1 Exercise parameters at peak exercise in patients of Group 1
PeakVO2 (ml. rain-', kg-')
MaxHR (b.min-)
Percentage of Predicted max HR
yEmax ,rl)
RERmax
57(16) 39-81
112(01) 0-92-1 31
(| m
(/o)
Mean (SD) range
20 2 ( 5 1 )
11-29
141 (20) 106-192
89(15) 69-125
VO2: oxygen uptake; HR: heart rate; VE: external ventilation; RER: respiratory exchange ratio.
Table 2 Feasibility and inlerobserver variability of the graphical determinations of the ventilatory threshold (protocol 30 W every 3 mm) n = 22 Method for determining ventilatory threshold
VCO 2
Determination rate (%) 74 Five full determination rate (%) 32 Quartiles of CV(%) Mean 70 Q75 12-5 Distribution of the deviations from the mean (%) - 2 - 5 % +2S% 29
VE
RER
VE/VO2
Crossing
76 59 7-8 9-8
72 18 7-8 11-8
78 54 9-4 18-4
88 73 8-5 13-7
55 26 19
34 15 51
38 23 39
20 24 56
CV = coefficient of variation; Q = quartiles; VO2: oxygen uptake; VCO2: carbon dioxide production; VE: minute ventilation; RER: respiratory exchange ratio; VE/VO2: ventilatory equivalent for oxygen; Crossing: crossing of the VCO2 and VO2 curves. 8 6
g o
T
4 2
-T
T +
T
1
-4
1
o
1
j_
jo o
-2
-6 -
o
1
o—
-8
•
•
0
A Peak VO2 (lOWmin"')
1
A Peak VO2 (30Wper3min)
A Ventilatory threshold I (lOWmin')
i A Ventilatory threshold (30 W per 3 mm)
Figure 2 Box-plots of the reproducibility between test 1 and test 2 of peak VO2 (A peak VOJ and ventilatory threshold (A ventilatory threshold) for the 10 W. min"1 (left) and the 30 W every 3 min (right) protocol. Difference of the values of peak VO2 and ventilatory threshold between test 2 and test 1 for each patient, expressed in ml. min" 1 . kg"1 of VO2, are plotted on the y axis. The box bounds the first and third quartiles of the data; the horizontal line within the box-plot represents the median of the differences and the cross the mean of the differences.
summarize the reproducibility data corresponding respectively to the 30W every 3min and 10W.min" 1 protocols. Peak VO2 was very highly reproducible since values varied less than 2-5% between the two subsequent tests
Downloaded from https://academic.oup.com/eurheartj/article-abstract/12/10/1055/440724 by University of Adelaide user on 04 May 2018
using the two protocols ( — 0-32% in protocol 10 W . min ' and + 2-18% in protocol 30 W every 3 min). Accordingly, confidence intervals of the mean were very small (Fig. 2) and correlation coefficients were high (0-93 and 0-97) (Figs 3 and 4).
Oxygen uptake during bicycle exercise in CHF 1059
35 -
25
OJ
20
Peak
(ml mm 1
30
15
.••-'£/?'
y = l-02x -0-40
-••-'>
