relative errors in the range of 9-12% and 5-7% for breath-by-breath and mixing chamber systems respectively. Relative errors in ˙VCO2 are found in the range of 5-7% for breath-by-breath systems and approximately 4% for mixing chamber systems. These results show that mixing chamber systems are more accurate, although breath-by-breath systems are known for a better temporal resolution [37]. The validity of the error analysis method depends largely on the validity of input errors, as discussed in chapter 5.
Measurements using a metabolic simulator show significant differences be-tween commercially available CPET systems. Relative errors in ˙VO2 and ˙VCO2
as high as 15% are found, whereas relative errors in ˙VE are in the range of 0-10%.
Differences between our results and other validation studies [67, 68] are largely explained by differences in hardware and software settings. This shows that the specific settings of CPET equipment are important for the performance of the system. Since manufacturers regularly perform hardware and software updates, it is difficult to compare the results of different validation studies.
Moreover, the validation of CPET equipment has not been standardized.
Although cardiopulmonary exercise testing is considered as a gold standard for exercise capacity [70], both our error analysis and the metabolic simulator measurements show that CPET results have substantial measurement errors that are in the range of 4-12%. This range of error hampers clinical appli-cation since these errors are well in the clinical relevant range. For example, differences for athletes between first and second place are in the range of 1%, hampering the clinical application of ˙VO2,max testing of athletes. Furthermore, errors in CPET results can influence patient care in clinical situations. CPET has been identified as as a core assessment for patients with heart failure [71]
and risk stratification is based on key CPET parameters dividing patients in prognostic categories. Using such categorization, measurement errors in the range of 4-12% may hamper clinical application and limit standardization of cut-off values. Moreover, the effect of training on exercise capacity is difficult to quantify using CPET equipment, since training effects (10-15%) are in the range of measurement error [15, 16] and biological variability is in the order of 5% [11, 12, 13, 14].
7.2 Conclusion
Validation of CPET is a difficult process, since there is no gold standard for CPET. The Douglas bag method is often used to validate systems, but biological calibration studies with human subjects using the Douglas bag method still have to deal with biological variability issues [11, 12, 13]. Furthermore, the Douglas bag method has its own accuracy problems [19], such as the diffusion of gas through the collection bag and leakage of two-way-valves. Calibration measurements using a metabolic simulator are a good alternative without having to deal with biological variability issues. However, metabolic simulators do not produce warm and humid air which can cause significant accuracy problems in
CHAPTER 7. GENERAL DISCUSSION 7.2. CONCLUSION
the process of gas analysis. Error analysis has proved itself as a powerful method to study the relative and absolute importance of different sources of error.
Our error analysis method shows that the error in ˙VO2 is larger than in V˙CO2. Moreover, the error of breath-by-breath systems is larger than the error of mixing chamber systems. Differences are largely caused by the delay time error δtdelay. In general, the error of the flow sensor δ ˙V and the assumed temperature of the exhaled air TB are important for the accuracy and precision of CPET results.
Measurements using a metabolic simulator show that breath-by-breath sys-tems are less stabile for different values of minute ventilation than mixing cham-ber systems. Moreover, metabolic simulator measurements show that there are significant differences between the accuracy and precision of different commer-cially available CPET systems.
The performance of CPET systems relies for a significant part on the specific software and hardware, which is difficult to validate separately. Moreover, man-ufacturers often release new versions of software and hardware, making making the comparison of different validation studies cumbersome [67, 68]. In general, validation of CPET equipment remains a difficult process, and systems should be considered as new (and recalibrated) when software or hardware changes have been made.
7.2. CONCLUSION CHAPTER 7. GENERAL DISCUSSION
Chapter 8
Recommendations
This chapter discusses recommendations for future research and important siderations for future acquisition processes, as well as recommendations con-cerning the use of CPET in clinical practice.
8.1 Scientific recommendations
Important results of this study are obtained using a theoretical error analysis method. This method does not include errors caused by breath detection algo-rithms and the use of two way valves. Future research can be done to incorporate these errors in the error analysis method. Furthermore, individual calibrations of the gas sensors and flow sensor can be performed to obtain a better estimation for the input errors. This also involves incorporating response time corrections of gas signals [42, 43, 44, 45, 46, 47, 48, 49] and angular momentum corrections of signals from turbine flow sensors [41].
One of the most important limitations of the Vacumed metabolic simulator is the inability to produce warm and humid air. The Max II metabolic calibrator used by the Australian Sports Commission is able to produce warm and humid air and calibration services are offered commercially [17, 72]. Expired air is passed across a waterbath filled with hot water of approximately 60◦C and the air exits fully saturated with water vapor at about 37◦C. Future research could incorporate these techniques in metabolic simulator measurements.
Recently, a novel CPET method has been developed based on respiration chamber technology [73]. The Maastricht Instruments Omnical measures gas fractions of highly diluted expired air, so that fluctuations of gas fractions are minimal and the response time issues of gas analyzers can be neglected. General error propagations theory should be applied to this method to compare the accuracy to other CPET methods.
8.2. ACQUISITION PROCESS CHAPTER 8. RECOMMENDATIONS
8.2 Considerations acquisition process
Measurements using the metabolic simulator show that there are large differ-ences between the error in results of CPET equipment. For high values of minute ventilation ˙VE >100, the two systems with variable orifice flow meters measured ˙VE with larger error than systems with turbine flow sensors. There-fore, it is recommended to measure the ventilation of elite athletes using an CPET systems with turbine flow meter. Moreover, error analysis shows that mixing chamber systems are more accurate in measuring ˙VO2 and ˙VCO2. When temporal resolution is unimportant, a CPET system with mixing chamber is recommended.
8.3 Considerations clinical practice
Previous studies have shown that the outcome of CPET equipment can depend on system specific software and hardware settings [67]. Ideally, CPET systems are re-calibrated when software and/or hardware updates are performed. For example, this can be done using a metabolic simulator.
Error analysis calculations shows that fluctuations in the ambient oxygen concentration can cause significant error when ˙VO2 is calculated. Sufficient ven-tilation of the exercise laboratory should prevent carbon dioxide accumulation and oxygen depletion. Although ventilation can cause a drop (or rise) in ambient temperature, this does not play a significant role according to error analysis cal-culations, especially not when ambient temperature is monitored continuously as most CPET systems do.
Whenever gas fractions are measured, one should know the relative humidity of the gas to determine dry gas fractions. Most CPET systems lead the sampled gas through a Permapure Nafion tube to dry the sample. The sampled gas will not be dried completely, but the vapor pressure of the sample equilibrates with the vapor pressure of ambient air [29]. Since we can measure the vapor pressure of the ambient air, we know the vapor pressure of the gas sample and dry gas fractions can be calculated (section 3.1). However, Permapure Nafion tubes become brown/yellow after many hours of operations and the drying function will be limited. Therefore, Cosmed advises their customers to replace Permapure Nafion tubing after 100 tests. Furthermore, the Permapure Nafion tubing should be dry before a new exercise test is started, since accumulation of water in the Permapure Nafion tubing and possibly the gas sensors due to previous tests will affect the measurement of gas fractions. Drying the Permapure Nafion tubing prior to testing with a ventilator is an easy solution to this problem. To prevent liquid water from entering the sampling tube, the sampling tube should not be mounted downwards on the flow sensor.
References
[1] C.K. Morris, K. Ueshima, T. Kawaguchi, A.Hideg, and V.F. Froelicher. The prognostic value of exercise capacity: A review of the literature. American Heart Journal, 122(5):1423 – 1431, 1991.
[2] S.N. Blair, H.W. Kohl, C.E. Barlow, R.S. Paffenbarger, L.W. Gibbons, and C.A. Marcera. Changes in physicalfitness and allcause mortality -A prospective-study of healthy and unhealthy men. J-AM-A-Journal of the American Medical Association, 273(14):1093–1098, 1995.
[3] K. Wasserman, J.E. Hansen, D.Y. Sue, B.J. Whipp, and R. Casaburi. Prin-ciples of exercise testing and interpretation. Lea & Febiger, second edition, 1994.
[4] W.D. McArdle, F.I. Katch, and V.L. Katch. Exercise physiology: energy, nutrition, and human performance. Williams & Wilkins, third edition, 1996.
[5] K. Albouaini, M. Egred, A. Alahmar, and D.J. Wright. Cardiopulmonary exercise testing and its application. Heart, 93:1285–1292, 2007.
[6] S.S. Gropper, J.L. Smith, and J.L. Groff. Advanced nutrition and human metabolism. Wadsworth Cengage Learning, 2008.
[7] K. Wasserman. Lactate and related acid base and blood gas changes during constant load and graded exercise. Canadian Medical Association Journal, 96(12):775–783, Mar 1967.
[8] V.F. Froelicher and J. Myers. Manual of exercise testing. Mosby, third edition, 2007.
[9] J. Myers, N. Buchanan, D. Walsh, M. Kraemer, P. McAuley, M. Hamilton-Wessler, and V.F. Froelicher. Comparison of the ramp versus standard ex-ercise protocols. Journal of the American College of Cardiology, 17(6):1334 – 1342, 1991.
[10] E.A. Ashley, J. Myers, and V. Froelicher. Exercise testing in clinical medicine. The Lancet, 356(9241):1592 – 1597, 2000.
REFERENCES REFERENCES
[11] C. Taylor. Some properties of maximal and submaximal exercise with refer-ence to physiological variation and the measurement of exercise tolerance.
American Journal of Physiology – Legacy Content, 142(2):200–212, 1944.
[12] V.L. Katch, S.S. Sady, and P. Freedson. Biological variability in maximum aerobic power. Medicine & Science in Sports & Exercise, 14(1):21–25, 1982.
[13] L.E. Armstrong and D.L. Costill. Variability of respiration and metabolism:
Responses to submaximal cycling and running. Research Quarterly for Exercise and Sport, 56(2):93–96, 1985.
[14] G. Atkinson, R.C.R. Davison, and A.M. Nevill. Performance characteris-tics of gas analysis systems: What we know and what we need to know.
International Journal of Sports Medicine, 26(1):S2–S10, 2005.
[15] H.M. Kemps, G. Schep, W.R. de Vries, S.L. Schmikli, M.L. Zonderland, E.J.M. Thijssen, P.F.F. Wijn, and P.A. Doevendans. Predicting effects of exercise training in patients with heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. American Journal of Cardiology, 102(8):1073–1078, 2008.
[16] H. Kemps, W. de Vries, S. Schmikli, M. Zonderland, A. Hoogeveen, E. Thi-jssen, and G. Schep. Assessment of the effects of physical training in pa-tients with chronic heart failure: the utility of effort-independent exer-cise variables. European Journal of Applied Physiology, 108:469–476, 2010.
10.1007/s00421-009-1230-3.
[17] C.J. Gore and Australian Sports Commission. Physiological tests for elite athletes. Human Kinetics, 2000.
[18] D.J. Macfarlane. Automated metabolic gas analysis systems: A review.
Sports Medicine, 31(12):841–861, 2001.
[19] R.J. Shephard. A critical examination of the douglas bag technique. The Journal of Physiology, 127(3):515–524, 1955.
[20] A. Huszczuk, B.J. Whipp, and K. Wasserman. A respiratory gas exchange simulator for routine calibration in metabolic studies. European Respiratory Journal, 3(4):465–468, 1990.
[21] C.J. Gore, P.G. Catcheside, S.N. French, J.M. Bennett, and J. Laforgia.
Automated VO2max calibrator for open-circuit indirect calorimetry sys-tems. Medicine and Science in Sports and Exercise, 29(8):1095–1103, 1997.
[22] W.L. Beaver, N. Lamarra, and K. Wasserman. Breath-by-breath mea-surement of true alveolar gas-exchange. Journal of Applied Physiology, 51(6):1662–1675, 1981.
[23] J.S. Ultman and S. Bursztein. Analysis of error in the determination of respiratory gas exchange at varying FIO2. Journal of Applied Physiology, 50(1):210–216, 1981.
REFERENCES REFERENCES
[24] P.H. Breen. Importance of temperature and humidity in the measure-ment of pulmonary oxygen uptake per breath during anesthesia. Annals of Biomedical Engineering, 28(9):1159–1164, 2000.
[25] G.L. Ruppel. Manual of pulmonary function testing. Mosby, eighth edition, 2003.
[26] J.E. Cotes. Lung function - Assesment and application in medicine. Black-well Scientific Publications, fifth edition, 1993.
[27] W.L. Beaver. Water vapor corrections in oxygen consumption calculations.
Journal of Applied Physiology, 35(6):928–931, 1973.
[28] J.W. Severinghaus. Water vapor calibration errors in some capnometers:
Respiratory conventions misunderstood by manufacturers? Anesthesiology, 70(6):996–998, 1989.
[29] X. Ye and M.D. LeVan. Water transport properties of nafion membranes:
Part i. single-tube membrane module for air drying. Journal of Membrane Science, 221(1-2):147 – 161, 2003.
[30] C.G. Douglas. A method for determining the total respiratory exchange in man. The Journal of Physiology, 42(Suppl):xvii–xviii, 1911.
[31] W.L. Beaver, K. Wasserman, and B.J. Whipp. On-line computer analysis and breath-by-breath graphical display of exercise function tests. Journal of Applied Physiology, 34(1):128–132, 1973.
[32] J.H. Wilmore and D.L. Costill. Semiautomated systems approach to the assessment of oxygen uptake during exercise. Journal of Applied Physiology, 36(5):618–620, 1974.
[33] J.H. Wilmore, J.A. Davis, and A.C. Norton. An automated system for assessing metabolic and respiratory function during exercise. Journal of Applied Physiology, 40(4):619–624, 1976.
[34] C.B. Cooper and T.W. Storer. Exercise testing and interpretation: a prac-tical approach. Cambridge University Press, first edition, 2001.
[35] V.C. Madama. Pulmonary function testing and cardiopulmonary stress testing. Delmar Publishing Team, 1998.
[36] R. Eston and T. Reilly. Kinanthropometry and Exercise Physiology Labo-ratory Manual: Tests, Procedures and Data: Volume 2, Physiology. Rout-ledge, second edition, 2001.
[37] K. Roecker, S. Prettin, and S. Sorichter. Gas exchange measurements with high temporal resolution: The breath-by-breath approach. International Journal of Sports Medicine, 26(S 1):S11–S18, 2005.
REFERENCES REFERENCES
[38] J.H. Wilmore and D.L. Costill. Adequacy of the haldane transformation in the computation of exercise vo2 in man. Journal of Applied Physiology, 35(1):85–89, 1973.
[39] D. Jones. Biomedical Sensors. Momentum Press, 2010.
[40] O.F. Pedersen, T.R. Rasmussen, S.K. Kjaergaard, M.R. Miller, and P.H.
Quanjer. Frequency response of variable orifice type peak flow meters: re-quirements and testing. European Respiratory Journal, 8(5):849–855, 1995.
[41] M.P. Yeh, T.D. Adams, R.M. Gardner, and F.G. Yanowitz. Turbine flowmeter vs. fleisch pneumotachometer: a comparative study for exercise testing. Journal of Applied Physiology, 63(3):1289–1295, 1987.
[42] R. Arieli and H.D. Van Liew. Corrections for the response-time and delay of mass spectrometers. Journal of Applied Physiology, 51(6):1417–1422, 1981.
[43] H. Noguchi, Y. Ogushi, I. Yoshiya, N. Itakura, and H. Yamabayashi.
Breath-by-breath VCO2 and VO2 required compensation for transport de-lay and dynamic response. Journal of Applied Physiology, 52(1):79–84, 1982.
[44] J. H. Bates, G. K. Prisk, T. E. Tanner, and A. E. McKinnon. Correcting for the dynamic response of a respiratory mass spectrometer. Journal of Applied Physiology, 55(3):1015–1022, 1983.
[45] B.E. Shykoff and H.T. Swanson. A model-free method for mass spectrom-eter response correction. Journal of Applied Physiology, 63(5):2148–2153, 1987.
[46] M.J. Turner and S. Culbert. Apparatus to measure the step and frequency responses of gas analysis instruments (respiratory gas analysis). Physiolog-ical Measurement, 14(3):317, 1993.
[47] L. Wong, R. Hamilton, E. Palayiwa, and C. Hahn. A real-time algorithm to improve the response time of a clinical multigas analyser. Journal of Clinical Monitoring and Computing, 14(6):441–446, 1998.
[48] A.D. Farmery and C.E.W. Hahn. Response-time enhancement of a clini-cal gas analyzer facilitates measurement of breath-by-breath gas exchange.
Journal of Applied Physiology, 89(2):581–589, 2000.
[49] M.A. Garcia, V.H. Nascimento, and J.C.T.B. Moraes. Improvement of gas exchange measurements in ergospirometry with the use of optimal filters for O2 and CO2 transducers. Revista Brasileira de Engenharia Biom´edica, 23(3):263–275, 2007.
REFERENCES REFERENCES
[50] R.L. Hughson, D.R. Northey, H.C. Xing, B.H. Dietrich, and J.E. Cochrane.
Alignment of ventilation and gas fraction for breath-by-breath respiratory gas exchange calculations in exercise. Computers and Biomedical Research, 24(2):118 – 128, 1991.
[51] D.N. Proctor and K.C. Beck. Delay time adjustments to minimize errors in breath-by-breath measurement of vo2 during exercise. Journal of Applied Physiology, 81(6):2495–2499, 1996.
[52] P. Webb. Air temperatures in respiratory tracts of resting subjects in cold.
Journal of Applied Physiology, 4(5):378–382, 1951.
[53] P. Cole. Recordings of respiratory air temperature. The Journal of Laryn-gology & Otology, 68(5):295–307, 1954.
[54] J.E.C. Walker and R.E. Wells Jr. Heat and water exchange in the respira-tory tract. The American Journal of Medicine, 30(2):259 – 267, 1961.
[55] P. H¨oppe. Temperatures of expired air under varying climatic conditions.
International Journal of Biometeorology, 25:127–132, 1981.
[56] L.E. Sandals. Oxygen uptake during middle-distance running. PhD thesis, University of Gloucestershire, 2003.
[57] P. Palange, S.A. Ward, K-H. Carlsen, R. Casaburi, C.G. Gallagher, R. Gos-selink, D.E. O’Donnell, L. Puente-Maestu, A.M. Schols, S. Singh, and B.J.
Whipp. Recommendations on the use of exercise testing in clinical practice.
European Respiratory Journal, 29(1):185–209, 2007.
[58] A. Mezzani, P. Agostoni, A. Cohen-Solal, U. Corr ˜A˘a, A. Jegier, E. Kouidi, S. Mazic, P. Meurin, M. Piepoli, A. Simon, C. Van Laethem, and L. Van-hees. Standards for the use of cardiopulmonary exercise testing for the functional evaluation of cardiac patients: a report from the exercise physi-ology section of the european association for cardiovascular prevention and rehabilitation. Journal of Cardiovascular Risk, 16(3):249–267, 2009.
[59] T. Meyer, T. Georg, C. Becker, and W. Kindermann. Reliability of gas exchange measurements from two different spiroergometry systems. Inter-national Journal of Sports Medicine, 22(8):593–597, 2001.
[60] J.R. Taylor. An introduction to error analysis: the study of uncertainties in physical measurements. University Science Books, 1997.
[61] Cosmed, Italy. Cosmed Quark CPET brochure, 2009.
[62] E.M. Winter, A.M. Jones, British Association of Sport, and Exercise Sci-ences. Sport and exercise physiology testing guidelines: the British Asso-ciation of Sport and Exercise Sciences guide. Number v. 2 in Sport and Exercise Physiology Testing Guidelines: The British Association of Sport and Exercise Sciences Guide. Routledge, 2006.
REFERENCES REFERENCES
[63] U. Boutellier, T. Kundig, U. Gomez, P. Pietsch, and E.A. Koller. Respira-tory phase detection and delay determination for breath-by-breath analysis.
Journal of Applied Physiology, 62(2):837–843, 1987.
[64] VacuMed, USA. Operating & service manual; Motorized syringe No. 17050 with metabolic calibration kit No. 17054, 2005.
[65] J.M. Bland and D.G. Altman. Statistical methods for assessing agreement between two methods of clinical measurement. The Lancet, 327(8476):307 – 310, 1986.
[66] Ø. Foss and J. Hall´en. Validity and stability of a computerized metabolic system with mixing chamber. International Journal of Sports Medicine, 26(7):569–575, 2005.
[67] H. Rosdahl, L. Gullstrand, J. Salier-Eriksson, P. Johansson, and P. Schantz.
Evaluation of the oxycon mobile metabolic system against the douglas bag method. European Journal of Applied Physiology, 109:159–171, 2010.
[68] M. Kusch, U. Hoffmann, and A. Vogt. Quality control of various ergospiro-metric measuring parameters by means of a gas exchange simulating sys-tem. Austrian Journal of Cardiology, 10(9):405, 2003.
[69] F. Prieur, T. Busso, J. Castells, R. Bonnefoy, H. Benoit, A. Geyssant, and C. Denis. A system to simulate gas exchange in humans to control quality of metabolic measurements. European Journal of Applied Physiology and Occupational Physiology, 78:549–554, 1998.
[70] J. Roca and R. Rabinovic. Lung function testing, chapter Clinical exer-cise testing, pages 146–165. European respiratory monograph. European Respiratory Society Journals Limited, 2005.
[71] R. Arena, J. Myers, and M. Guazzi. Cardiopulmonary exercise testing is a core assessment for patients with heart failure. Congestive Heart Failure, 17(3):115–119, 2011.
[72] Australian Institute of Sport. http://www.ausport.gov.au/ais/sssm/quality assurance/services/indirect calorimetry.
[73] P.F.M. Schoffelen, K.R. Westerterp, W.H.M. Saris, and F. TenHoor. A dual-respiration chamber system with automated calibration. Journal of Applied Physiology, 83(6):2064–2072, 1997.
Appendix A: Accuracy and precision
Figure 1: Schematic drawing showing (a) , (b) , (c) and (d) [60].
Measurements performed with small error are accurate and precise. How-ever, there is a difference between accuracy and precision. Accurate measure-ments have a small systematic error, whereas precise measuremeasure-ments have a small random error. This is shown schematically in figure 1. Figure 1a shows a precise and accurate measurement, with a small systematic error and a small random error. Figure 1b shows a precise but accurate measurement, where the random error is small but the systematic error is large. Figure 1c is a schematic drawing of an accurate but inprecise measurement, with a small systematic and large random error. Finally, figure 1d shows an inaccurate an inprecise measurement, with large systematic and random errors.
APPENDIX . APPENDIX A: ACCURACY AND PRECISION
Appendix B: Error propagation
Suppose quantity y is computed with the multi variable function f () from n variables, x1, ..., xn
y = f (x1, ..., xn). (1)
Since variables x1, ..., xn are the result of measurements and are subject to errors depending on the accuracy and precision of the measurement, a quantity xi is stated as xi± δxi. This means that the best estimate of the measurement is xi and that it is almost certain that the true value lies between xi+ δxi and xi− δxi, where δx is called the error or uncertainty. Because variables x1, ..., xn are subject to uncertainties, the computed quantity y will also be subject to uncertainty δy. The uncertainty δy can be calculated according to general error propagation theory [60]
and the error found is never larger than the ordinary sum
δy ≤