Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Pulmonary diffusion limitation, V̇ /Q̇ mismatch and pulmonary transit time in highly trained athletes.. Hopkins, Susan R. 1992

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


ubc_1993_spring_phd_hopkins_susan.pdf [ 6.99MB ]
JSON: 1.0076880.json
JSON-LD: 1.0076880+ld.json
RDF/XML (Pretty): 1.0076880.xml
RDF/JSON: 1.0076880+rdf.json
Turtle: 1.0076880+rdf-turtle.txt
N-Triples: 1.0076880+rdf-ntriples.txt
Original Record: 1.0076880 +original-record.json
Full Text

Full Text

PULMONARY DIFFUSION LIMITATION, VA/Q MISMATCH AND PULMONARYTRANSIT TIME IN HIGHLY TRAINED ATHLETES DURING MAXIMAL EXERCISEbySUSAN ROBERTA HOPKINSB. Med. Sci., Memorial University of Newfoundland, 1978M.D., Memorial University of Newfoundland, 1980M.P.E., The University of British Columbia, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTSFOR THE DEGREE OF DOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Interdisciplinary Studies, Medicine/Physical Education/Zoology)We accept this thesis as conforming to the required standardTHE UNIVERSITY OF BRITISH COLUMBIANovember 1992© Susan Roberta Hopkins, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of  Interdi:aeiplinary StudiesThe University of British ColumbiaVancouver, CanadaDate  Fehruary25. 1993DE-6 (2/88)ABSTRACTTo investigate the relationship between pulmonary diffusion limitation, ventilation-perfusion (VA/Q) mismatch, pulmonary transit times (PTT) and pulmonary gas exchangeduring exercise, 10 highly trained male athletes (age=26.4±4.4 years, Height=185.5±5.3cms, Weight=78.2±8.6 kg, V 02max=5.15±0.521-min -1) under went exercise testing atrest (R) and 150W, 300W and maximal exercise (372±22W), corresponding to an oxygenconsumption (V02) of 0.41±0.09, 2.16±0.17, 4.32±0.35 and 5.13±0.50 1-min-1respectively, while trace amounts of six inert gases were infused via a peripheral vein.Arterial blood samples, mixed expired gas samples and metabolic data were obtained.Observed alveolar arterial difference ([A-a]D02(0)was calculated according to the alveolargas equation. Indices of VA/Q mismatch: LogSDi and Log SDa and predicted [A-a]D02([A-a]DO2(p)) were derived from 50 compartment model analysis of retentions andexcretions of the inert gases. Additional indices of '/A/I,) mismatch: DISPR*, DISPE andDISPR*_E and inert gas alveolar difference ([A-a]D, R(A-a)D and E(A-a)D) were obtaineddirectly from the inert gas data. One to two weeks later, the subjects underwent first passradionuclide angiography using a Siemens ZLC wide field of view gamma camera.Following in vitro labeling with 99mTechnecium, 5-10 ml of the subject's blood, containing10-20 mCi of activity, were injected at rest. First pass and post-static data were obtained onan ADAC 3003 computer and cardiac output was calculated using the Stewart Hamiltonequation. PTT was determined using deconvolution and centroid methods. Gatedradionuclide angiography was then performed at rest, 150, and 300W. On a separateoccasion, first pass cardiac outputs and pulmonary transit times were obtained at maximalexercise. Mean arterial partial pressure of 02 (Pa02) decreased significantly from rest to150W , and from 150 to 300W to a low value of 86±9 torn, before increasing to nearresting values at maximal exercise. [A-a]D02(3) increased across each exercise levelshowever only the increase from 150 to 300 W was significant. The overall and perfusion-related indices of VA/Q mismatch showed a significant increase with exercise, mainly as aiiresult of increasing perfusion of areas of high VA/Q [A-a]D02(0 was greater thanpredicted, becoming significant during heavy exercise, indicating diffusion limitation.Cardiac output increased from 6.9±0.9 1-min -1 (R) to 25.2±2.5 1-min-1 at 300W and33.3±3.7 1-min-1 at maximal exercise. End diastolic volume increased from R to heavyexercise (p < 0.001), accompanied by a decrease in end systolic volume (p =0.05). Strokevolume and ejection fraction also increased significantly from R to 300W (p < 0.001).Deconvolution PTT decreased from 9.32±1.41 s at rest to 2.91±0.30 s during max exerciseand was highly correlated with centroid PTT both at rest (r=0.99, p<0.001) and duringmaximal exercise (r=0.96, p<0.001). PTT during maximal exercise was significantlycorrelated with Pa02 (1=0.65, p<0.05) and [A-a]D02(0)_[A-a]D02(p) (r=-0.60, p<0.05).Calculated pulmonary blood volume increased during maximal exercise by 57% over restingvalues to over 25% of total blood volume and when corrected for body surface areacorrelated significantly with Pa02 (r=0.69, p<0.05). There was a significant correlationbetween (A-a)D, PTT, the ventilatory equivalent for CO2 and Pa02 during maximalexercise (r=0.94, p<0.01) allowing prediction of over 80% of the variance in Pa02 betweensubjects. These data indicate that highly trained athletes develop VA/Q mismatchaccompanied by diffusion limitation during maximal exercise. Observed decrease in Pa02during high intensity exercise is the result of a complex interaction between VA/Qmismatch, hypoventilation and diffusion limitation secondary to shortened pulmonarytransit.iiiTABLE OF CONTENTSAbstract ^ iiList of TablesviList of Figures ^ viiList of abbreviations and symbols ^ viiiAcknowledgmentxIntroduction 1Methods ^ 6Baseline data^ 6Subject preparation 7Test protocol 7Multiple inert gas analysis ^ 8Arterial blood gases 9Cardiac output^ 10Pulmonary transit time12Data analysis 13Results ^ 14General data ^ 14Ventilation and metabolic data ^ 15Arterial blood gases and oxygen saturation ^ 16VA/Q inequality ^ 22Diffusion disequilibrium 25Cardiac output and cardiac volumes ^ 26Pulmonary transit time26Pulmonary blood volume ^32Discussion ^ 33Multiple inert gas analysis and exercise ^ 34MIGET data ^35.^.Indices of dispersion and VA/Q mismatch ^ 35Diffusion limitation37Effect of uncertainty in cardiac output on MIGET data ^38Maintenance of steady state conditions 40Blood and plasma volume ^ 41Radionuclide cardiography 42Cardiac function in athletes 43Comparison with previous investigations ^44Effect of increasing exercise intensity on cardiac volumes^44Pulmonary transit times ^47Theory 47Pulmonary transit times and exercise in humans ^49Pulmonary blood volume ^49Relationship of pulmonary transit time and blood volume topulmonary capillary transit time and blood volume ^50Arterial blood gas and metabolic data^ 51Arterial blood measures ^51Mixed venous P02 52Possible mechanisms of exercise induced hypoxemia ^52Summary of findings^ 56References ^57Appendix A: Review of literature 70Introduction^ 70ivRespiration and the respiratory muscles ^ 71Energetics ^72Blood flow to the diaphragm and other respiratory muscles^72Oxygen consumption of respiratory muscles 74Lactate Production ^ 76Theoretical calculations of respiratory muscle VO2 and Q in humansexercising at maximal levels ^ 77Respiratory muscle fatigue 79Histochemical properties of mammalian respiratory muscles ^79Evidence for fatigue80Respiratory drives ^ 81Pulmonary mechanics and expiratory flow limitation ^ 83Pressure-volume relationships^ 83Characteristics of flow inside tubes83Flow-volume relationships 84Pulmonary diffusion and gas exchange 86Factors related to diffusion of oxygen ^ 86Factors related to diffusion of carbon dioxide 87Factors affecting pulmonary gas exchange 88Does the pulmonary system constrain exercise? 92Summary^96Appendix B: Methodological background ^ 97Quantitative radiocardiography 97Red blood cell labeling 97First pass determination of cardiac output ^ 97Gated radionucleide angiography 99Measurement of blood volume^ 100Pulmonary transit time ^ 101Frequency distribution of pulmonary transit times ^ 102Multiple inert gas elimination 105Ventilation and perfusion relationships ^ 105Multiple inert gas elimination theory107Practical aspects ^ 110Dead space 111Shunt ^112VA/Q distributions ^ 113Alveolar arterial difference 114Appendix C: Statistical analyses and raw data^ 116Anova tables ^116Metabolic data ^ 116Blood gas data 118MIGET data for all six gases^ 121Cardiac data125Regression ^ 127Appendix D: Individual Subject Data^ 131vLIST OF TABLES^Table 1.^Subject descriptive data ^ 14Table 2.^Plasma volume and calculated whole blood volume - individual subjectdata^ 15Table 3.^Arterial blood gases, metabolic data, transit times and MIGETsummary data at rest, light, heavy and maximal exercise (i±S D) ^ 18Table 4.^Calculated mixed venous and arterio-venous 02 difference duringrest, light, heavy and maximal exercise. ^ 22Table 5.^(A-a)D, R(A-a)D and E(A-a)D at rest and during light, heavy andmaximal exercise (i ±SD) ^ 24Table 6.^Cardiac output, volumes and ejection fraction at rest and during, light,heavy and maximal exercise (x±- SD) ^ 26Table 7.^Pulmonary transit times by the deconvolution and centroid methods ^28Table 8.^Pulmonary blood volume at rest and during maximal exercise ^32Table 9.^Comparison of MIGET data from present study with otherinvestigators ^ 38Table 10.^Effect of varying cardiac output on residual sum of squares and indicesof dispersion ^ 40Table 11.^Cardiac output and volume indices comparison with other studies ^46Table 12.^Blood flow of respiratory muscles ^ 73Table 13. VO2 of respiratory muscles 75Table 14.^Oxygen cost of unobstructed hyperventilation ^76Table 15.^Alveolar-arterial differences at rest and during exercise 91Table 16.^Blood gas and cardio-respiratory data obtained at 70-100% of V02max ^ 95viLIST OF FIGURESFigure 1.^Pa02, PaCO2 and [A-a]DO2(o) for individual subjects during lightexercise ^ 19Figure 2.^Pa02, PaCO2 and [A-4D02(o) for individual subjects during heavyexercise ^ 20Figure 3.^Pa02, PaCO2 and [A-4D02(o) for individual subjects duringmaximal exercise^ 21Figure 4.^Dispersion indices at rest and during light, heavy and maximalexercise ^ 24Figure 5.^Observed and predicted alveolar-arterial 02 difference ^25Figure 6A.^Raw data and gamma univariate fit for subject one during rest^29Figure 6B.^Frequency distribution of transit times for data in Figure 6A^29Figure 7A.^Raw data and gamma univariate fit for subject one during maximalexercise ^ 30Figure 7A.^Frequency distribution of transit times for data in Figure 7A^30Figure 8.^Pa02 and [A-41)02 (o-p) versus transit time ^31Figure 9.^Maximal expiratory flow volume loop with exercise and maximumvoluntary ventilation manoeuvre ^ 84Figure 10.^The relationship between end capillary P02 and mixed venous P02for lung units of differing VA/Q ^ 107viiLIST OF ABBREVIATIONS AND SYMBOLS(A-a)D^Inert gas alveolar arterial difference area[A-a]D02 Alveolar arterial difference for oxygen(a-v)02 diff^Arterio-venous difference for oxygenDISPR* Index of dispersion for retention of MIGET gasesDISPE^Index of dispersion for excretion of MIGET gasesDISPR*_E Overall index of dispersionDLCO^Diffusing capacity for carbon monoxideE(A-a)D Excretion component of inert gas alveolar arterial difference areaEIH^Exercise induced hypoxemiaFVC Forced vital capacityFEVi^Forced expiratory flow in 1 second1311 Iodine 131 a radioactive isotope of iodineLambda — blood gas partition coefficientLog SDa^Standard deviation of the log normal distribution for blood flowLog SDv.^Standard deviation of the log normal distribution for ventilationMEFV^Maximal expiratory flow volumeMIGET Multiple inert gas elimination techniqueMVV^Maximum voluntary ventilationpH Negative logarithm of hydrogen ion concentrationPaO2^Arterial partial pressure of oxygenPAO2 Alveolar partial pressure of oxygenPa/NV^Retention (R) of a MIGET gas defined as the ratio of arterial tomixed venous partial pressurePbar barometric pressurePE^ partial pressure in expired gasPE/Pir Excretion (E) of a MIGET gas defined as the ratio of mixedexpired to mixed venous partial pressureviiiP102^Partial pressure of inspired oxygenP ,^mixed venous partial pressure13,02^Mixed venous partial pressure of oxygenQ Cardiac output or blood flowR(A-a)D^Retention component of inert gas alveolar arterial difference areaRER Respiratory exchange ratioSa02^Arterial oxygen saturationSF6 Sulfur hexafluorane99mTc^Technetium 99 an meta stable radio-isotope of TechnetiumVAIQ Ventilation-perfusion ratioVc^ Pulmonary capillary blood volumeVCO2 Minute production of carbon dioxideVD^Dead spaceVE Minute ventilationi7E/VCO2^Ventilatory equivalent for carbon dioxidei/E/V02 Ventilatory equivalent for oxygenV02^Minute consumption of oxygenVT Tidal volumeaACKNOWLEDGMENTI would like to take this opportunity to thank everyone who helped in the preparation of thisthesis:Dr. Don McKenzie, my supervisor for his helpful advice, sense of humor, supportand endless proof reading,Dr. Tom Robertson, Dave Frazer, and Thien Tran, their help and patience inanalyzing, programming, printing and interpreting mountains of MIGET data,Dr. Brownie Schoene for his help in providing technical support, creative financing,accommodation and beer; Dr. Rob Glenny and Ron Saxon for their enthusiastichelp in data collection,Dr. Al Belzberg, and the technical staff at St. Paul's Hospital Nuclear Medicine forthe excellent help with the nuclear medicine data,Barry Wiggs, for his help with deconvolution,My subjects who endured VO2 max tests, big needles, radiation and blizzards in thename of Science,My committee members Peter Hochachka, Bob Schutz and Jeremy Road for theiradvice,My parents, Bob and Barbara Hopkins for supporting me in the decision to return toschool,and Trevor Cooper, for providing a shoulder to cry on.The financial assistance of the Medical Research Council of Canada, and the BritishColumbia Lung Association is also gratefully acknowledged.xINTRODUCTIONDissecting the factors that limit maximal exercise performance has been afundamental area of investigation in exercise physiology in both human and animals.Factors such as oxygen (02) transport and the circulatory system, peripheral blood flow,and 02 diffusion and mitochondrial 02 utilization have been discussed (39). The effects ofendurance type physical training on the cardiovascular and musculosketetal system are wellknown and include increases in cardiac output (Q), stroke volume, arterio-venous [(a-v)02]difference, plasma volume and peripheral muscle blood flow (20) all of which contribute tothe increase in maximal oxygen consumption (VO2max). Similarly, effects of training onthe respiratory system have been documented and include: increased respiratory muscleenzyme activities (82, 113), and increased maximal voluntary ventilation (MVV) (150)maximal ventilation (51) maximal sustainable ventilation (47, 91) and flow velocities (16).In the gas exchanging portions of the lung however, development is complete in childhood(113) and a training effect (aside from blood volume or hemoglobin alterations) on gasexchanging areas is unlikely (10, 135). As a result, the lung has attracted attention as apotential constraint to maximal performance particularly at altitude (55, 74, 161, 177, 178)and in highly trained athletes (4, 34, 35, 72, 105, 127) where a supra-normal cardiovascularsystem may unmask the respiratory system's limited capacity to adapt.Exercise induced hypoxemia (EIH) was first reported by Han -op (65) who madearterial blood gas measurements on patients with a variety of clinical problems as well ashealthy normal subjects. Included in these measures were samples taken from a healthyCaucasian male who performed "fifteen minutes of brisk exercise consisting of arm andtrunk movements and vigorous hopping about the room until quite dyspneic." Hedocumented a decline in % saturation of arterial blood (Sa02) from 95.6% at rest to 85.5%following exercise. Sporadic reports of impaired gas exchange during exercise followed(146, 159, 185) however the lung was not widely recognized as a potential limiting factor toexercise performance as the preponderance of evidence indicated that arterial oxygenation1was maintained during exercise (9, 17, 156). More recently, a number of authors (35, 72,126, 127) have confirmed the findings of Harrop (65), Rowell (146), Thompson (159) andothers, stimulating interest in this avenue of exploration.Two causes of the decline in Pa02 with exercise can be outlined, based on alveolarpartial pressure of oxygen (PAO2) and arterial partial pressure of oxygen (Pa02). If Pa02 islow and PAO2 is also low (< 105 torr) this suggests inadequacy of ventilation either as aresult of mechanical limitation of flow or as a result of blunted respiratory drives. If Pa02 islow and PAO2 is high (>110) resulting in very wide alveolar-arterial difference ([A-a]D02)this suggests an inadequacy of gas exchange either as a result of ventilation-perfusion(VA/Q) mismatch or diffusion limitation. In fact, both of these situations have beendescribed in the literature. Dempsey et al., (35) exercised sixteen subjects at 60-90% ofVO2max and reported Pa02 of less than 60 torr associated with little or no alveolarhyperventilation. Administration of normoxic helium-oxygen mixtures improved Pa02 onlyto the extent that PAO2 was increased and it was concluded that the magnitude of thehyperventilatory response to exercise was a major determining factor of the hypoxemiaobserved. In contrast to this, Hopkins and McKenzie (72) found that alveolar P02 washigh in their subjects who developed hypoxemia, with no evidence of inadequateventilation. Both groups raised the possibility of impaired gas exchange as a possibleexplanation for some of the hypoxemia seen in their subjects.As exercise intensity increases, cardiac output to the working muscles alsoincreases, in some highly trained individuals to over 40 I-min-1 (43). The time that the redblood cell spends in the pulmonary vascular bed is directly related to flow and pulmonarycapillary blood volume; as flow increases, expansion of the pulmonary vascular bed bydilation and recruitment (59) becomes of paramount importance in ensuring adequate timefor equilibration of gas exchange. The average resting pulmonary capillary transit time hasbeen estimated to be - 0.75 s (80, 144) falling to about 0.3 s during exercise. Theserepresent mean capillary transit times as a whole, and it is possible that, given regional2differences in pulmonary transit (70), some red cells may travel too rapidly through thepulmonary vasculature resulting in arterial hypoxemia.Transfer of a gas across the pulmonary blood gas barrier can be described by theequation (166) :100 kA ^ax ^t Px(t) = PAX + (P .ix - PA x).e 60^dVc • WIMWx.,where Px is the partial pressure of gas x , t is time, PA and Pv are alveolar and mixedvenous partial pressures of gas x, k is the diffusion coefficient of gas x, A is the cross-sectional area of diffusion, d is the thickness of the blood gas barrier, ax is the solubility ofthe gas x in the blood gas barrier, ox is the solubility of gas x in the blood and MW is themolecular weight of the gas x. The effect of decreasing transit time can be seen byinspection of the above equation. All other factors being equal, as transit time decreases Pxwill fall as a function of ct, such that when t=0 the partial pressure is equal to the mixedvenous partial pressure. In the exercising athlete, the time for equilibration of oxygen isfurther lengthened by low mixed venous P02 (34) and a right-ward shifted oxygen-hemoglobin equilibrium curve secondary to temperature increase, high mixed venousPCO2, and low pH, which in addition to possible rapid pulmonary transit may compromisegas exchange.Two recent technological advances have made possible further detailedinvestigations into diffusion limitation and gas exchange in exercising athletes. The multipleinert gas elimination technique (MIGET) as it is currently in use, was developed by Wagnerand co-workers (45, 167, 168) and is discussed in detail in Appendix B. It utilizes a multi-compartment model of pulmonary gas exchange calculated from data obtained fromelimination of infused inert gases (usually six) based on the relationship:3PAX^Pc i x Pix^PiTz XX + VANwhere PAx is the alveolar partial pressure, NI is the mixed venous partial pressure, Pc'x isthe end capillary partial pressure and Xx is the blood gas partition coefficient of gas x . Inertgases will approach equilibrium across the alveolus faster than oxygen, therefore ifdiffusion dis-equilibrium occurs, the observed [A-41)02 will exceed that predicted frominert gas exchange and will be improved by administration of 100% oxygen. Postpulmonary shunt via the bronchial and thesbian veins will not affect inert gases as they arenot metabolized, however administration of 100% 02 will disproportionately increase the[A-4D02 because it will increase alveolar P02 with less effect on arterial P02 as shuntedblood bypasses the gas exchanging areas. These techniques thus allow the contributingfactors to [A-41)02 to be dissected.Inert gas studies in normal subjects exercising at a VO2 of about 3.01•min -1 (161)and in moderately trained individuals exercising at a VO2 of 4.01•min -1 (63) havesuggested diffusion limitation, secondary to rapid pulmonary transit, as a cause of impairedgas exchange. Right heart to left heart whole lung transit time can be measured usingradioisotopically labeled RBC and a frequency distribution of pulmonary transit times can becalculated (70, 71, 96). The time required for an indicator to flow past an observation pointdown stream from an entry point is related not only to the time it take the bolus to flow pastthe point but also how quickly it arrived there. In this case the gamma camera provides theobservational point to observe the bolus curve derived from labeled RBC traversing the rightventricle and the output curve derived from the left ventricle. Transit time is determined bysubtracting the first moment of the right ventricular curve from the first moment of the leftventricular curve. Deconvolution is a mathematical process by which a frequency4distribution of transit times (a transfer function, h(t)) can be derived from the input (rightventricular) and output (left ventricular) time activity curves.It was the purpose of this study to investigate pulmonary gas exchange and transittimes during exercise in a population of highly trained male athletes exercising near V02max who on the basis of high pulmonary blood flow may develop shortened pulmonarytransit and diffusion limitation.5METHODSBaseline dataSixteen non-smoking healthy male athletes with no prior history of respiratory orcardiac disease underwent preliminary studies. After giving informed consent, a historywas obtained and a physical examination was performed, seeking evidence of cardio-pulmonary abnormality. All subjects were screened for pulmonary disease with pulmonaryfunction tests consisting of forced expiratory volume in 1 second (FEV1), forced vitalcapacity (FVC), peak flow rates, and 12 second maximum voluntary ventilation ( MMV)using a Medical Graphic CAD/Net 2001 metabolic cart equipped with Medical Graphics1070 pulmonary function testing software.Maximal oxygen uptake was determined on an electronically braked cycleergometer (Minjardt KEM-3) equipped with a racing saddle and pedals. After a 10 minutewarm-up (75-100 watts), the subjects rode a progressive exercise test with the workintensity ramped at 30 watt•min-1 until they were unable to continue. Minute ventilation(VE), oxygen consumption (V02 ) and carbon dioxide production (VCO2) were measuredon a breath-by-breath basis and tabulated every 15 seconds with a Medical GraphicsCAD/Net 2001 Metabolic Cart equipped with Medical Graphics 2001 software. Heart ratewas monitored and recorded by cardiac monitor (Lifepak 6) interfaced with the metaboliccart. VO2 max was considered to be the average of the four highest consecutive 15 secondmeasures of oxygen uptake. These results were used to calculate a work-load whichrepresented greater than 90% of VO2 max. Subjects were excluded from further testingunless their VO2 max was greater than 5.0 1-min-1 or 60 ml-kg- 1 .min-1 . On a separateoccasion single breath carbon monoxide diffusing capacity ( DLCO) was obtained at theUBC Hospital Pulmonary Laboratory (7).6Ten subjects fulfilling the entry criteria were transported the following week to theCardio-Pulmonary Laboratory at the Harborview Medical Center (Seattle Wa.).Subject Preparation Prior to the exercise test, an indwelling arterial cannula (Arrow # 20 gauge) wasinserted in the right radial artery. Cannula patency was maintained by frequent flushing withnormal saline to which heparin sodium (2000 u.1-1 ) had been added. Under steriletechnique, and cardiac monitoring, a number 7.5F triple lumen Swan-Ganz catheter (fortemperature monitoring) was introduced into the venous circulation via the left antecubitalvein and positioned in the region of the superior vena cava. Cannula patency was maintainedas described for the arterial cannula. A second venous line (#18 gauge) was inserted via theright antecubital vein into the peripheral venous circulation for infusion of the inert gases.The subjects were instructed to discontinue if any unusual symptoms developed. At leastthree physicians were present at all times during the exercise testing one of whom hadprimary responsibility for the subject. EKG tracing (lead II) was monitored continuously.Test protocol Subjects were seated on the bicycle ergometer previously described and connected toa respiratory circuit. This consisted of a Rudolph (2700) valve connected by large diameterheated tubing to a heated 131 Plexiglas mixing chamber which in turn was connected to apneumotach. Breath by breath analysis of VO2 and VCO2 was obtained by a MedicalGraphics 2000 system equipped with 2001 software similar to the system previouslydescribed for V02 max testing. After a ten minute rest period to allow stabilization ofventilatory data the testing protocol was started. Arterial blood gases, VE, V02, VCO2,heart rate were obtained at rest and every minute at each of the exercise levels. The exerciselevels consisted of five minutes each of light (150w, mean V02 = 42% of V02 max),heavy (300w, mean V02 = 86% of V02 max) and near maximal exercise (>90% VO2max, mean power output = 371±30 watts). Mixed expired gases and arterial blood samples7for inert gas analysis were collected at rest and during the last minute at each exercise level.The subjects rested approximately 5 minutes between each exercise level to allowrecalibration of the instruments.Multiple inert gas analysisSix inert gases (SF6, ethane, cyclopropane, halothane, ether and acetone) dissolvedin 5% dextrose (55, 63, 167, 168, 180) were infused starting 20 min before the start of theexperiment, at a rate in ml•min-1 corresponding to one quarter of the expected VE in1•min-1 (range 5 - 50 ml•min-1). Duplicate arterial (8 ml) and expired gas (30 ml) sampleswere collected in pre-heparinized glass syringes during the last minute at each exercise leveland analyzed by gas chromatography. The time delay between arterial and expired samplesdue to the hoses and mixing chamber was calculated by dividing the mixing chambervolume by the VE and the time of collection of the expired gas sample was adjustedaccordingly. Mixed venous inert gas concentrations were calculated from the Fickequation. Retention (Pa/Pv) and excretion (PE/Nr) values were used to estimate VA/Q andstandard deviation of the log normal distribution of perfusion (LogSDQ. ) and ventilation(LogSDv) distribution were used as an index of VA/Q inequality (168). Retention (R(A-a)D)and excretion (E(A-a)D) components of the inert gas alveolar-arterial difference area ([A-a]D) were also derived directly from the inert gas data (42, 69). Three additional indiceswere derived directly from the data as described by Gale et al., (53). DISP R* analogous tolog SDO and R(A-a)D, DISPE analogous to log SI:or and E(A-a)D, and DISPR*_Eanalogous to (A-a)D were calculated as follows:nDISPR*.E = 100 x^i=1 n8nI(Ri-Rhomoi) 2i=DISPR* = 100 x^1 n0\1 nI(Ehomoi-Ei) 2i=DISPE = 100 x^1 nwhereEhomoi = Rhomoi —Xi+ V• AQTn is the number of gases, Ei and Ri represent excretions and retentions of the gas of interestand Ei is excretion corrected for dead space:Ei —^Ei V D1 - VTPredicted values for Pa02, PaCO2 and [A-41)02, ([A-a]D02(p)) based on thederived VA/Q distribution were then calculated (174, 180). These values, calculated frominert gas data, reflect only that predicted by infra-pulmonary shunt and ventilation perfusioninequality, therefore an indirect estimate of diffusion impairment [A-4D02(o-p), ispossible. It should be noted that the inert gas analysis assumes steady state and the absenceof post-pulmonary shunt.Arterial blood gasesThe arterial samples were anaerobically collected in pre-heparinized glass syringesand were maintained in ice until the test session was complete. Each 2 ml sample wasanalyzed for pH, Pa02 and PaCO2 using a Corning Blood Gas/ pH Analyzer, which iscalibrated daily and after each exercise test against a known standard. Resting samples andthe last sample drawn had hemoglobin and hematocrit determined. Using Kelman's routines9(85-87) Sa02, Pa02 and PaCO2 were also corrected for arterial blood temperaturemeasured at the superior vena cava via the Swan-Ganz catheter.Cardiac outputOne to two weeks later in the Department of Nuclear Medicine at St. Paul's Hospitalcardiac output was measured at each exercise intensity using a combination of first pass andgated radiocardiography. Right ventricle to left ventricle first pass transit time was alsomeasured during rest and near maximal exercise. Plasma volume was measured using 1311labeled albumin (RISA) (76).Ten ml of whole blood was withdrawn from the subject into a pre-heparinizedsyringe via an antecubital vein. The blood was labeled with 99-mTechnetium (99-mTc) usinga standard commercial red blood cell (RBC) labeling kit (152) (Brookhaven NationalLaboratories, Upton, N.Y.). A 5-10m1 aliquot of RBC injectate containing 10-20 mCi ofactivity was obtained. A 20 gauge plastic cannula was inserted in the right median basilicvein and maintained patent with normal saline. The subjects were seated on the bicycleergometer previously described in front of a Siemens ZLC 3700 gamma camera with a widefield of view, medium energy columnator. The subjects were studied in the left anterioroblique position by placing the ergometer at angle to the camera and having the subject leanforward placing his chest directly in contact with the camera and grasping the camera withthe left hand. In all cases, except for one at near maximal exercise, the images obtained wereof excellent technical quality and allowed good separation of the right and left ventricle.Heart rate was recorded using a Lifepak 6 monitor/defibrillator.After stabilization of heart rate to the level observed during resting level of the inertgas experiments, one half the labeled RBC were injected into the dead space of the plasticcannula and flushed into the venous circulation with a rapid bolus of 15 ml of normalsaline. Data were acquired and processed on an ADAC 3003 computer (ADACLaboratories, Sunnyvale, Ca) at 0.5 frames•second -1 during the first pass through the10central circulation. Without changing the position of the camera a 2 minute static image wasobtained after the bolus injection. In the static projection, a large region of interest wasdrawn over the left ventricle using the light-pen and applied to the dynamic view. A firstpass time activity curve and static or equilibrium counts were obtained. Cardiac output (Q)was calculated from the Stewart Hamilton equation:•^TB V Q = Ceq.C(t)dtwhere Ceq represents the static counts from the left ventricle region of interest, TBV isblood volume (from measured plasma volume and hematocrit) and C(t).dt is the area underthe left ventricle first pass curve. The upstroke and down stroke of this curve areextrapolated to zero to exclude counts from the right ventricle and recirculation of blood.The remainder of RBC were then administered and the gated studies wereconducted. Using the same experimental setup as described above, count data was obtainedat a framing rate of 16 frames per R-R interval of the EKG and stored on the hard disk in a64x64 pixel matrix. Data were acquired for two minutes, after the third minute of eachexercise level. Gated ventricular imaging was conducted at rest, 150 watts and 300 watts.Each of the 16 images acquired was then displayed and the left ventricular region of interestwas manually drawn using a light-pen. A background region of interest was drawn laterallyand inferiorly to the left ventricle. Left ventricular ejection fraction was then calculated fromthe difference between background subtracted end-diastolic and end-systolic counts andexpressed as a percentage of end-diastolic counts.Two, 5 ml samples of blood were drawn at the end of each exercise period andimaged in petri dishes for five minutes. The average background subtracted count rate for 5ml of blood was thus obtained. The left ventricular count rate at each exercise level wascorrected for loss by decay of 99mTc using standard tables and the left ventricular end-diastolic volume was obtained by dividing the left ventricular count rate by the 5 ml countrate; cardiac output was then obtained by multiplying by the ejection fraction and the11average heart rate during data acquisition. Correction for attenuation of counts by the chestwall for the gated studies was made by comparing the first pass cardiac output at rest withthe result obtained from the resting gated study, and the results obtained during the gatedstudies at 150 and 300w were adjusted accordingly. Of the 40 gated determinations ofcardiac output and cardiac volumes, 7 were lost for technical reasons. For the purpose ofMIGET calculations, Q was calculated from a linear regression of Q vs VO2 for theremaining subjects. On a separate occasion, at least a week apart to minimize the effect ofresidual radioactivity, the first pass study was repeated at the near maximal exercise level.Pulmonary transit timeRight ventricle to left ventricle RBC pulmonary transit times were calculated fromthe first pass time-activity curves (70, 96). In the static image, large regions of interest weredrawn with the light pen outlining the right and left ventricle. These regions of interest werethen applied to the dynamic images and time activity curves were obtained. In a similarfashion, the quality of the bolus injection was checked by drawing a region of interest in thesuperior vena cava. The numerical data thus obtained was then down loaded directly into aZenith 386 PC for further processing. Each time-activity curve was fitted using linear leastsquares to a gamma (y) function using the method of Starmer and Clark (153) programmedinto a Lotus 123 spread sheet. The transit time across the lungs was then obtained bysubtracting the first moment of the right ventricular time-activity curve from that of the leftventricle (96). To confirm these results and obtain a frequency distribution of transit times(also known as a transfer function) deconvolution analysis was applied to the gamma fittedcurves. This was done using Gauss 386 statistical software. The area underneath both theinput (right ventricular) curve and the output (left ventricular) curve was set equal to 1. Aseries of fifty transfer functions was then generated with a mean transit time equal to thatobtained by the subtraction of first moments, and a variance ranging from that of the inputcurve to that of the output curve. The transfer functions were then convoluted with the input12curve and ridge regression was applied to determine contribution of each output curvederived by this process to the final output curve. This process was repeated with slightalterations in the variance and mean transit times of the transfer functions until a satisfactoryvisual fit was obtained. Pulmonary blood volume was calculated by dividing the cardiacoutput in ml•second4 by the transit time.Data analysisData were analyzed using analysis of variance for repeated measures to determinedifferences in blood gas, cardiac output, stroke volume, indices of dispersion and metabolicdata between the rest, 150w, 300w, and near maximal exercise conditions. A similaranalysis was used to determine differences in end systolic volume, end diastolic volume andejection fraction between rest, 150w and 300w. Where overall significance was obtained,Scheffe's testing was applied post-hoc to determine were these differences occurred.Student's t test was used to evaluate changes in pulmonary transit time and pulmonary bloodvolume between rest and maximal exercise. Linear regression was used to determine therelationship between [A-a]D02(o-p) during maximal exercise and pulmonary transit time,and between blood volumes and Pa02 and [A-4D02(o). Multiple linear regression wasused to examine the relationships between VENCO2 , PTT, (A-a)D, and Pa02.13RESULTSGeneral dataSubject descriptive information is given in Table 1. The high levels attained for VO2max indicate that the subjects were highly trained males. Blood volume indices frommeasured plasma volume and hematocrit are given in Table 2. The subjects exhibited abovenormal values for plasma volume and red cell mass. Plasma volume was as high as 159% ofpredicted based on norms calculated from the subjects height, weight and age. Meanhematocrit was at the low end of the normal range however calculated red cell mass was alsomore than 130% of predicted. In all cases plasma and red cell volumes were outside theupper limit of the normal range. Mean hematocrit for the subjects was 42.6% pre-exerciseand increased to 44.2% post exercise. This change was statistically significant (t=1.98,p<0.05)Table 1. Subject descriptive dataAge(years)Height Weight VO2 max(cm)^(kg)^(1•min-1 )FVC(1)FEY'(1)1 31 188.5 89.7 6.31 7.02 6.732 28 180.7 67.5 4.37 5.52 4.623 23 193.5 79.5 5.07 6.73 6.374 30 181.5 81.0 5.17 5.34 4.685 20 188.4 82.3 5.12 6.47 5.266 25 185.2 76.5 5.36 7.07 6.107 19 183.5 68.5 4.91 6.05 5.168 28 185.1 77.2 4.81 5.94 5.179 29 176.5 68.2 4.84 5.49 4.8410 31 192.5 92.0 5.57 6.02 4.91Mean 26.4 185.5 78.2 5.15 6.17 5.38± SD 4.4 5.3 8.6 0.52 0.63 0.75VE max MW DLCOm,(1•min-1 ) (1-min- 0mm Hg-1 )218.5186.1240.2201.0211.5199.4209.2212.1186.2201.6206.616.3230 44.70198 41.99257 44.98223 42.94219 49.74241 49.90204 43.47178194 34.19208 43.04215.2 43.8823.6 4.6314*Subject failed to complete this portion of the testingTable 2. Plasma volume and calculated whole blood volume - individual subject data.WholeBloodVol.(ml)ml•kg-1 %Predicted PlasmaVol.(ml) ml•kg-1 %Predicted RBCVol.(ml)%Predicted1 8075 90.0 152 5140 57.3 161 2934 1372 5720 84.7 130 3621 53.6 136 2100 1213 6959 87.5 134 4231 53.2 134 2727 1334 6142 75.8 127 3859 47.6 133 2283 1175 7467 90.7 145 4912 59.7 159 2556 1256 6802 88.9 140 4249 55.5 145 2553 1337 6759 98.7 145 4359 63.6 157 2400 1358 6757 87.5 139 3848 49.8 131 2906 1519 5643 82.7 130 3499 51.3 133 2144 12510 8126 88.3 144 5143 55.9 153 2983 131Mean 6845 87.5 139 4286 54.8 144 2559 131±SD 868 5.9 8 606 4.7 12 325 10Ventilation and metabolic dataRespiratory and metabolic data are given in Table 3. Statistical information inbrackets refers to Scheffe's F test unless otherwise stated. Mean blood temperaturemeasured at the superior vena cava was 36.8±0.5 °C at rest and rose significantly at eachexercise level to 38.0±0.7 °C at the end of the heaviest exercise task (omnibusF(3,27)=24.3 P < 0.001).VE increased from 13.4±2.6 1-min -1 at rest to 178.1+16.3 1-min-1at the end of maximal exercise, and was significantly less than peak values (mean=206.61•min-1 ) reached during the initial V02 max test (t=5.76, p<0.001). V02 increased from0.41± 0.09 1-min-1 at rest to 2.16±0.17 1-min -1 , 4.32±0.35 1•min-1 , and 5.13±0.501-min-1 at light (150 watts) , heavy (300 watts) and maximal exercise respectively. VCO2 atthe corresponding exercise levels was 0.36±0.10 1-min-1 , 1.85±0.11 1•min -1 , 4.39±0.351•min-1 and 5.97±0.681-min-1 . The respiratory exchange ratio (RER) was not significantlydifferent between rest and light exercise (0.88E111 rest vs 0.86±0.05 light exerciseF=0.06, p=NS); however it increased significantly from light to heavy exercise (F=7.74,p<0.05) to 1.02±0.11 during heavy exercise indicating that the subjects were close to their15anaerobic threshold during this workload. During maximal exercise RER increasedsignificantly from heavy exercise (F=6.50, p<0.05) to 1.17±0.08. The ventilatoryequivalent for carbon dioxide (VE/VCO2) decreased significantly from rest to light exercise(F=20.5, p<0.05), then increased slightly from light to heavy exercise and from heavy tomaximal exercise, although VE/VCO2 at max was significantly less than resting values(F=7.6, p<0.05). The ventilatory equivalent for oxygen (VE/V02) also decreasedsignificantly from rest to light exercise (F=18.2, p<0.05) and increased from light to heavyexercise (F=5.08, p<0.05) and from heavy to maximal exercise (F=6.26, p<0.05).VE/V02 at rest was not significantly different from maximal exercise (F=0.25,p=NS).There were modest correlations between VENCO2 and Pa02 during heavy andmaximal exercise (r=0.43 and r=0.53 respectively) for these 10 subjects but they were notstatistically significant. There was a similar correlation between VE/V02 and Pa02 duringheavy exercise (r=0.40), again not statistically significant.Arterial blood gases and oxygen saturation Group mean blood gas values are given in Table 3 and individual subject data forPa02, PaCO2 and [A-41)02(o) during light, heavy and maximal exercise are given inFigures 1, 2 and 3. Mean Pa02 was 98±6 torr at rest and decreased significantly (F=36.3,p<0.05) to 91±8 ton during light exercise. Pa02 declined further to 86±9 torr during heavyexercise (F=12.8 compared to rest, p<0.05) and increased to near resting levels (94±8 torr)at the end of maximal exercise. Alveolar arterial difference did not increase significantlyfrom rest to light exercise (F=0.29, p=NS); however the increase in [A-41)02(o) fromlight to heavy exercise was significant (F=27.5, P<0.05). There was no further increase in[A-4D02(o) between heavy and maximal exercise (F=0.14, p=NS); therefore anyimprovement in Pa02 was due to improved alveolar P02. There was little evidence ofhypoventilation at the end of maximal exercise as mean PAO2 calculated by the alveolar gasequation was 121±3 torr. PaCO2 was slightly depressed during the rest measurements16(38±3 torr), likely indicating some anxiety on the part of our subjects; it then rose to morenormal levels during light exercise, before decreasing significantly during heavy andmaximal exercise (F=8.4,p<0.05 and F=6.5 p<0.05 for heavy and maximal exercise vsrest). Blood pH did not change significantly from rest to light exercise, but decreasedsignificantly from light to heavy exercise (F=7.94, p<0.05) and from heavy to maximalexercise (F=15.0, p<0.05). Sa02 decreased significantly over the exercise levels (omnibusF(3,27)= 18.5 p<0.0001) to a mean of 94.2±2.3%.There was considerable inter-subject variability in the blood gas response over thefive minute exercise period at each exercise intensity (see Figures 1, 2 and 3). For example,subject six maintained Pa02 above 98 torr during heavy exercise and above 105 torr duringmaximal exercise. Corresponding [A-41)02(o) was less than 11 ton throughout testing andPA02 was above 110 ton at the end of heavy exercise and 115 torr at the end of maximalexercise, confirming adequate ventilation and gas exchange. In contrast to this, subject 4developed hypoxemia ( Pa02 = 74 ton) during heavy exercise associated with an [A-41)02(o) of 38 ton. At maximal exercise the Pa02 increased to 81 torr as a result ofincreasing PAO2 (123 ton) and [A-a]D02(o) widened further to 42 ton. Subject 5, alsodemonstrated hypoxemia during heavy exercise with Pa02 at the end of exercise of 74 torr,however there is also evidence of hypoventilation, as PaCO2 was elevated above restinglevels and PA02 was 102 ton.17Table 3. Arterial blood gases. metabolic data. transit times and MIGET summary data atrest. light, heavy and maximal exercise (i±SD) Rest Light Exercise(150 watts)Heavy Exercise(300 watts)MaximalExercise(371±30 watts)Heart Rate 70±10 116±9 156±8 166±8Q (1•min-1 ) 6.9±0.9 16.1±1.5 25.3±2.5 33.3±3.7VE (1•min-1) 13.4±2.6 45.5±3.5 119.3±27.3 178.1±16.3i/EATO2 33.5±6.2 21.2±1.5 27.6±6.5 34.9±3.8TEATCO2 38.5±6.8 24.6±1.6 27.1±5.3 30.0±2.8V02 (1•min-1 ) 0.41±0.09 2.16±0.17 4.32±0.35 5.13±0.50VCO2 (1•min-1) 0.36±0.1 1.85±0.11 4.39±0.35 5.97±0.68RER 0.88±0.11 0.86±0.05 1.02±0.11 1.17±0.08PAO2 (torr) 102±8 98±5 111±6 121±3Pa02 (torr) 98±6 91±8 86±9 94±8Sa02 (%) 97.6±0.4 96.7±0.9 95.3±1.8 94.2±2.3[A-a]D02(o) (torr) 4±16 6±8 26±9 27±9[A-a]D02(p) (torr) 7±8 10±4 17±7 17±3[A-a]D02(o-p) (ton) -3±15 -4±9 9±13 10+12PaCO2 (ton) 38±3 41±3 36±4 32±3pH 7.44±0.03 7.41±0.03 7.34±0.05 7.24±0.06Log SDT 1.09±0.55 0.90±0.47 1.09±0.29 1.18±0.14Log SDa 0.38±0.20 0.44±0.12 0.64±0.17 0.78±0.11DISPR* 2.527±2.129 2.827±2.092 4.631±2.291 5.733±1.332DISPE 3.881±3.070 3.918±3.015 5.549±2.712 5.459±1.034DISPR*-E 5.408±4.466 6.090±4.653 9.343±4.509 10.385±2.124Transit time (sec) 9.32±1.41 2.91±0.30Pulmonary bloodvolume (1)1.08±0.17 1.61±0.27181^2^3^4^5Time (min)Figure 1. Pa02, PaCO2 and [A-a]D02(o) for individual subjects during light exercise.-10 0^1^2^3^4^5Time (min)Legend: subject 1=open circle, subject 2=open square, subject 3=open diamond, subject4=closed circle, subject 5=closed square, subject 6=closed diamond, subject 7=diagonalcross, subject 8=horizontal cross, subject 9=open triangle, subject 10=open circle with dot.192^3^4^5Time (min)Figure 2. Pa02, PaCO2 and [A-a]D02(o) for individual subjects during heavy exercise.0^1^2^3^4^5Time (min)Legend: subject 1=open circle, subject 2=open square, subject 3=open diamond, subject4=closed circle, subject 5=closed square, subject 6=closed diamond, subject 7=diagonalcross, subject 8=horizontal cross, subject 9=open triangle, subject 10=open circle with dot.2050401001^2^3^4^5Time (min)Figure 3.Pa02, PaCO2 and [A-MO2(o) during maximal exercise.2^3^4^5Time (min)Legend: subject 1=open circle, subject 2=open square, subject 3=open diamond, subject4=closed circle, subject 5=closed square, subject 6=closed diamond, subject 7=diagonalcross, subject 8=horizontal cross, subject 9=open triangle, subject 10=open circle with dot.021Mixed venous P02 (NO2) and arterio-venous difference (a-i,r02diff) werecalculated from the Fick equation and are given in Table 4. Mean N ,02 was 36±4 torr at restand decreased significantly (F=17.9, P<0.05) during light exercise. A further decreasebetween light and heavy exercise was not statistically significant (F=2.49, p>0.05), norwas the slight increase from heavy to maximal exercise (X NO2 = 15± 6 torr heavy exercisevs 17±7 torr maximal exercise). The a-i/O2diff increased between rest and light exercise (F=3.98, p<0.05), was the same for light and heavy exercise (71±7 and 71±9 ton respectively)and increased further (F=1.69,p>0.05) during heavy exercise to a maximum value of 77±11torr.Table 4. Calculated mixed venous and arterio-venous 02 difference during rest. light,heavy and maximal exercise. SubjectRestNO2^(a-v)02(torr)^(ton)150 wattsNO2^(a-v)02(ton)^(ton)300 wattsNO2^(a-v)02(ton)^(ton)MaximalNO2^(a-v)O2(torr)^(torr)1 30 59 19 77 14 68 11 772 40 48 21 67 24 63 19 723 31 67 17 78 20 71 29 684 36 58 19 64 13 61 19 625 34 65 24 62 13 61 25 656 43 64 24 80 18 82 12 977 32 73 22 76 17 77 19 838 40 59 18 60 14 69 15 789 35 68 22 72 5 88 11 8610 35 61 17 75 8 72 7 82SD.^36±4^62±7^20±3^71±7^15± 6^71±9^17±7^77±11Pv- 02 = mixed venous partial pressure of oxygen, [(a-v)02] = arterio-venous difference foroxygen.'/* ^inequalityConsiderable difficulty was experienced in fitting the data from the present study tothe MIGET model of Wagner et al., (168) as evidenced by high residual sum of squares atx±22rest. There was a significant improvement in the residual sum of squares during heavy andmaximal exercise when compared to rest (F=3.3, and F=3.5, p<0.05) and an acceptable fitwas achieved in the maximal exercise data. The mean of the log normal perfusion (mean ofQ) distribution increased significantly over the exercise levels (omnibus F(3,27)=45.6,p<0.001) as did the mean of the log normal ventilation distribution (mean of V) (omnibusF(3,27)=61.2, p<0.0001). There was a non significant increase in the VA/Q heterogeneity,as measured by log SDQ from rest to light exercise; however the increase from light toheavy exercise and from heavy to maximal exercise was significant (F=6.6, p<0.05 andF=3.14, p<0.05). There was no significant change in the log SCor over the exercise levels(onmibus F(3,27)=1.35, P>0.05).Elimination of SF6 data from the analysis resulted in low residual sums of squares(4.27±2.7 at rest, which increased to a maximum value of 12.2±11.1 at max exercise)indicating adequate model fit. The results of the statistical analyses for the mean Q, mean Vand logSD j were unchanged when the SF6 data were eliminated, however no statisticallysignificant change in log SDQ was found with exercise (omnibus F(3,27)=1.12).The independent indices of VA/Q mismatch DISPR., DISP E and DISPR._E arepresented in Figure 4. There was a significant increase in dispersion as measured by two ofthe three indices. There was a highly significant increase in DISP R* across exercise levels(omnibus F(3,27) =25.8, p<0.0001), as was the increase in DISPR._E (F(3,27)=8.45,p<0.001). The increase in DISP E was not significant (F(3,27)=2.45, p=0.09). The areaunder the curve described by the plot of alveolar arterial difference for the six inert gasesas a function of A, (A-a)D and the retention and excretion components of the area (R(A-a)Dand E(A-a)D) are given in Table 5. DISPR*, R(A-a)D and log SDQ are all measures of theperfusion distribution and therefore should be comparable. Similarly DISP E, E[A-a] and logS1D, are comparable measures of the ventilation distribution and DISPR._ E and (A-a)D arecomparable overall indices of dispersion.2314121086420Table 5. (A-a)D, R(A-a)D and E(A-ajD at rest and during light. heavy and maximalexercise ("X ±SD). Rest 150 Watts 300 Watts Maximal Exercise(A-a)D 0.211±0.171 0.233±0.185 0.360±0.178 0.411±0.098R(A-a)D 0.066±0.060 0.088±0.066 0.146±0.073 0.181±0.043E(A-a)D 0.259±0.341 0.161±0.139 0.201±0.088 0.231±0.047There was a significant increase in (A-a)D and R(A-a)D over the exercise levels(omnibus F(3,27)=9.1, p<0.001 and omnibus F(3,27)=21.0,p<0.0001) paralleling theincreases in the corresponding parameters of the other methods of analysis. There was nosignificant change in E(A-a)D over the exercise levels (omnibus F(3,27)=1.7, p=0.64).Figure 4. Dispersion indices at rest and during light, heavy and maximal exercise. * ** *-100^0^100^200^300^400Work Load (watts)Legend: Closed circles = DISPR* Oa SD), open squares = DISPE (31 ± SD), opendiamonds = DISPR*_E (X ± SD), * = significantly different from rest (p<0.05),** =significantly different from light exercise and rest (p<0.05).244030-10Diffusion disequilibrium [A-4D02 predicted on the basis of the model of Wagner et al.(164), [A-41)02(p),and that not accounted for by the inert gas analysis [A-4D02(o-p) is given in Table 3. [A-4D02(p) is compared to observed [A-41)02 in Figure 5. [A-4D02(p) increasedsignificantly over the exercise levels (omnibus F= 12.2, p<0.0001), paralleling thedispersion indices. There were a non-significant increases in [A-4D02(p) between rest andlight exercise. The increase between light and heavy exercise was significant (F=3.7,p<0.05) and there was no further increase from heavy to maximal exercise. During heavyand maximal exercise [A-41)02(o) was 9-10 tort . greater than that predicted by the inert gasexchange (F=5.3,and F=11.4, p<0.05) suggesting diffusion limitation.Figure 5. Observed and predicted alveolar-arterial 02 difference-100^0^100^200^300^400Work Load (watts)Legend: Open squares=predicted [A-4D02 (mean±SD), closed squares=observed [A-4D02 (mean±SD) * = significantly different than predicted (p<0.05)25Cardiac output and cardiac volumesThe results of the first pass and gated cardiac studies are given in Table 6. Datawere processed by three independent observers and the results were averaged. Thecorrelation was 0.90 between observers and the mean difference in cardiac output betweenobservers was approximately 3% at each exercise level. Cardiac output rose from 7.0±0.91•min-1 at rest to 33.3±3.7 1-min-1 at maximal exercise, accompanied by a significantincrease in ejection fraction from 0.63±0.05 % at rest to 0.76±.05% during light exercise(F=35.8 p<0.05). A further increase in ejection fraction to 0.80±0.04 % at 300 wattsapproached but did not reach statistical significance (F=3.2, p>0.05 compared to 150watts). End diastolic volume increased significantly (omnibus F(2,12)=20.1, p<0.05) fromrest to heavy exercise as did stroke volume (F=22.5, p<0.05). There was also a decrease inend systolic volume between rest and heavy exercise (omnibus F(2,12)=4.4, p<0.05).Table 6. Cardiac output. volumes and ejection fraction at rest and during. light, heavy and maximal exercise.(i±SD). Rest 150 watts 300 watts MaximalQ 7.0± 0.9 16.2 ± 1.5 25.3 ± 2.6 33.3 ±3.6Heart Rate 70±10 116±9 156±8 166±8Stroke Volume 102 ± 21 140 ± 17 163 ± 22 201±23End Diastolic Volume 149 ± 16 184 ± 25 198 ± 22 *End Systolic Volume 55 ± 10 44 ± 12 41 ± 10 *Ejection Fraction 0.63 ± 0.04 0.76 ± 0.05 0.80 ± 0.04 ** = cardiac volumes obtained during gated study only which was not conducted duringmaximal exercise.Pulmonary transit timeFigures 6A and 7A contain representative raw data and gamma univariate fits forregions of interest drawn over the right and left ventricles. Figure 6B and 7B show thefrequency distribution of transit times (transfer function) obtained by the deconvolution26method for the same subject and the output curve obtained when the transfer function isconvoluted with the input curve from the right ventricle. Mean transit times at rest andmaximal exercise obtained by both deconvolution and centroid techniques are presented inTable 7. In one subject, (subject 10) a satisfactory fit was not obtained by deconvolutionanalysis and results are reported for the centroid method only in this individual. Meantransit time at rest was 9.31± 1.45 seconds by the centroid method and 9.32± 1.41 bydeconvolution. These were highly correlated (r = 0.99, p<0.0001). During exercise meantransit times decreased significantly to less than 3 seconds (2.90± 0.35 centroid, 2.91± 0.30deconvolution; t deconvolution =12.3,p<0.001) and these were also highly correlated(r=0.96, p<0.001). The mean duration of the right ventricular curve did not changesignificantly from rest to exercise (4.96 ± 1.37 vs 4.19 ± 0.80 seconds, t = 1.54, p = 0.1)however there was a significant decrease in the mean duration of the left ventricular curvefrom 14.29 ± 1.52 to 7.09 ± 1.06 seconds (t=11.2, p< 0.001). Pulmonary transit time wassignificantly correlated with Pa02 (Figure 8) (r=0.65, p<0.05) during maximal exerciseand [A-a]l:02(o) (r=-0.59, p<0.05). There was also a significant relationship (Figure 8)between [A-a]D02(o-p) and transit time (r=-0.60, p<0.05). When multiple linear regressionwas used to determine the relationship between transit time, i/FATCO2, [A-a]D and Pa02there was a highly significant relationship (R=0.94, R2=0.88, adjusted R2=0.83, p<0.01).Thus at maximal exercise, over 80% of the variance between subjects in Pa02 can beexplained on the basis of transit time, V. A/Q mismatch and the ventilatory equivalent forCO2.27T mon," fly lu en .n n r m X± D1M - I •isRest (s)Centroid^DeconvolutionExercise (s)Centroid^Deconvolution1 10.09 10.19 2.99 2.922 7.24 7.37 2.77 2.783 8.60 8.75 3.18 3.164 7.72 7.45 2.68 2.675 9.12 9.23 2.74 2.996 8.96 8.85 3.76 3.567 8.63 8.73 2.95 2.918 9.80 9.82 2.60 2.589 11.36 11.35 2.68 2.6910 11.75 11.43 2.71 *Mean^9.33 ± 1.45^9.32 ± 1.41^2.90 ± 0.35^2.91 ± 0.30±SD* = unable to fit satisfactory deconvolution analysis28OOFigure 6A. Raw data and gamma univariate fit for subject one during rest4.5Legend: Open triangles = left ventricular counts, open diamonds = right ventricular counts.Solid lines = gamma univariate fit for the right and left ventricular curves.Figure 6B. Frequency distribution of transit times for data in Figure 6A.0 0^5^10^15^20^25^30^35^40Time (seconds)Legend: Solid lines = gamma univariate fit for the right ventricle and left ventricle, dashedand dotted line = frequency distribution of transit times (transfer function), dashed line =resulting output curve when the right ventricular curve is convoluted with the transferfunction.290Co^10Figure 7A. Raw data and gamma univariate fit for subject one during maximal exercise.0^4^a^12^16^20^21Time (s)Legend: Open triangles = left ventricular counts, open diamonds = right ventricular counts.Solid lines = gamma univariate fit for the right and left ventricular curves.Figure 7B. Frequency distribution of transit times for data in Figure 7A.° 0^5^10^15^20Time (seconds)Legend: Solid lines = gamma univariate fit for the right ventricle and left ventricle, dashedand dotted line = frequency distribution of transit times (transfer function), dashed line =resulting output curve when the right ventricular curve is convoluted with the transferfunction.■■•Figure 8. Pa02 and [A-ajD02 (o-p) versus transit time.110 —105 —100 —95 —90 —85 —80 ^^2.4^2.6 2.8^3^3.2^3.4^3.6Transit Time (seconds)40 —30 •20 —   1 0 — • ••0 — •-10 —  -20 I I I I I I2.4 2.6 2.8^3^3.2^3.4^3.6Transit Time (seconds)31Pulmonary blood volumePulmonary blood volume and pulmonary blood volume index obtained at rest andduring maximal exercise are presented in Table 8. Pulmonary blood volume increasedsignificantly during exercise (t=6.1, p<0.001) by over 50%. Pulmonary blood volume wassignificantly correlated with whole blood volume at rest (r=0.67, p<0.05) and there was asimilar trend during exercise although the relationship did not attain statistical significance(r=0.52, p=0.06). Resting pulmonary blood volume index (pulmonary blood volume/BSA)correlated significantly with resting [A-a]D02 (r=-0.65, p<0.05). Exercising pulmonaryblood volume index correlated significantly with Pa02 (r=0.69, p<0.01) and [A-a]D02(o)(r=-0.57, p<0.05). Pulmonary blood volume index correlated significantly with [A-a]D02(o-p) during exercise (r=-0.68,p<0.05), as did whole blood volume (in ml•kg -1) (r=-0.61, p<0.05).Table 8. Pulmonary blood volume at rest and during maximal exercisePulmonaryBloodVolume (1)RestPulmonaryBloodVolumeIndex% ofTotalBloodVolumeMaximal exercisePulmonary^PulmonaryBlood BloodVolume (1)^VolumeIndex% ofTotalBloodVolume% increaseinpulmonarybloodvolume1 1.24 0.58 15 1.67 0.78 21 402 0.74 0.40 13 1.28 0.69 22 733 0.95 0.45 13 1.82 0.87 26 974 1.06 0.52 18 1.42 0.70 23 305 1.17 0.56 15 1.79 0.86 22 426 1.21 0.61 18 2.05 1.03 32 777 0.90 0.48 13 1.82 0.96 27 1078 1.08 0.54 16 1.54 0.76 23 449 1.11 0.60 20 1.18 0.64 21 610 1.33 0.61 17 1.53 0.69 19 11MEAN 1.08 0.53 16 1.65 0.82 24 57±SD0.18 0.07 2 0.30 0.15 4 3432DISCUSSIONSeveral authors have reported hypoxemia and arterial desaturation during short termheavy or maximal exercise which increases with physical conditioning (146) and is morecommon in highly trained athletes (184). In this population the reported incidence ofhypoxemia during exercise may be as high as 52% (126).The etiology of the hypoxemia continues to attract considerable debate. Dempseyand co-workers (35) exercised sixteen highly trained athletes at 70-90% of VO2 maxbreathing different gas mixtures. At this exercise intensity, several subjects exhibited littlealveolar hyperventilation and the authors concluded that "the magnitude of thehyperventilatory response was a major determinant of the hypoxemia seen in our athletes".The effect of hypoventilation is to reduce alveolar P02 and thus the arterial P02 without aneffect on [A-a]D02. Additional evidence in support of inadequacy of ventilation as acausative factor include the observations that during MVV testing and exercise, peakexpiratory flows may approach or exceed those defined by the maximal expiratory flow-volume curve (67, 78, 118) and the observation that administration of He:02 mixturesincreases ventilation, decreases PaCO2, and improves arterial oxygenation, without altering[A-a])02 (35, 36, 170). Humans also exhibit entrainment to a variable extent with running,walking, rowing and cycling (13, 158) and this has been implicated as a cause ofhypoxemia in galloping horses (34). Aside from the mechanical factors described above,hypoventilation could also be caused by blunted respiratory drives (26, 106, 148).Pulmonary diffusion limitation secondary to shortened pulmonary transit representsan attractive alternate hypothesis to hypoventilation as a cause of exercise inducedhypoxemia. Multiple inert gas studies have shown evidence of diffusion limitation in mencapable of sustaining VO2 — 4 I-min -1 , however since MIGET studies have not been madein very highly trained athletes who exhibit exercise induced hypoxemia, the arguments arelargely theoretical. The consistent observation that [A-a]D02 widens with increasingexercise intensity, (35, 63, 72, 161, 182) offers indirect support to this idea.33Multiple inert gas analysis and exerciseThe multiple inert gas analysis technique is based on the observation that theretention of a gas in the blood is related to the solubility of the gas (X) and the VA/Qdistribution. By using gases that bracket the solubility of 02 and CO2 and measuringarterial and mixed expired concentrations, it is possible to estimate shunt, dead space and theshape of the VA/Q distribution. It is also possible using the derived VA/Q distribution topredict the behavior of 02 and CO2. Assumptions that are made in the multiple inert gasanalysis are: 1. there is steady state gas exchange, 2. the lung units are arranged in paralleland behave as independent compartments 3. ventilation and perfusion is non-pulsatile, 4.diffusion dis-equilibrium and extra-pulmonary shunt are absent. It is the exploitation of thislast point that has formed the basis of detection of diffusion dis-equilibrium, as anyobserved [A-a]D02 that exceeds that predicted from the MIGET VA/Q distribution is likelydue to diffusion limitation or extra-pulmonary shunt. The advantages of the MIGETtechnique are that a change in Pi02 which may alter the VA/0 distribution is not required,and that the trace amounts of gases infused are not sufficient to alter concentrations of thephysiologic gases.Despite more than ten years experience with this method, studies involving humansare few, and are mostly confined to normals rather than highly trained athletes. Gledhill etal., (60) studied five male subjects at rest and during light exercise (V02 1.8 1-min -1) andfound an increase in V A/Q as measured by log SDV and log SDQ, associated with awidening of [A-a]D02 which was improved by the breathing of a high density gas. Incontrast , Derks (37) found no changes in dispersion in his subjects, again at levels ofexercise less than 2.01•min-1 . The next studies, published in 1985 as companion papers byTorre-Bueno et al.(161), and Gale et al. (55), contributed substantially to knowledge of theeffects of exercise and altitude on ventilation and perfusion mismatch and diffusionlimitation. Nine subjects were studied at various exercise levels up to a V02 of almost 31•min-1 , at sea-level and simulated altitude corresponding to 5,000, 10,000 and 15,000 feet.34At sea-level there was a trend towards worsening of VA/Q relationships with exercisealthough the results did not reach statistical significance. Resting VA/Q did not increase withaltitude, although the combination of exercise and altitude did produced significantdeterioration in the VA/Q relationships (55). There was no evidence of diffusion dis-equilibrium at rest for any altitude although there was evidence for diffusion limitationduring exercise at altitudes above 10,000 ft. There was a suggestion that, in subjects capableof higher levels of exercise, diffusion limitation might be present although small "n"hampered definitive conclusions (161).This last observation was addressed further in a paper by Hammond et al., (63)who was able to exercise moderately trained athletes to a VO2 of almost 4.0 1-min -1 .Evidence of diffusion disequilibrium was found at the highest exercise intensity (300 watts),as the average measured [A-0)02 exceeded that predicted from the derived VA/Qdistribution by more than 12 ton. These results were confirmed by Bebout et al., (12) whofound evidence of diffusion limitation during exercise that was worsened by altitude andimproved by two weeks of acclimatization. The authors concluded that the effect ofacclimatization was to lower cardiac output at any given exercise level and improve diffusionlimitation through an effect on pulmonary transit time.MIGET data.Indices of dispersion and VAN mismatchSignificant increases in overall indices of dispersion and in the indices of dispersionrelated to the blood flow distribution were observed with exercise regardless of the indexused. There was little change between rest and light exercise, marked increases betweenlight and heavy exercise and little further change between heavy and maximal exercise. Nosignificant change was apparent in any of the indices of dispersion related to ventilation.These data indicate increasing V. A/Q mismatch predominantly related to heterogeneity ofblood flow. When blood flow was examined with respect to areas of low (V. A/Q <0.1),35normal (0.1<VA/Q < 10) and high (VA/Q >10) VA/Q (40) it was evident that theheterogeneity in flow resulted from significantly increasing flow to areas of high V. A/Q.rather than areas of shunt or low V. A/Q . Perfusion of areas of high VA/Q accounted forover 10% of blood flow during maximal exercise compared with less than 2% at rest. Atmaximal exercise perfusion of areas of low V. A/Q accounted for less than 1% of blood flowand intrapulmonary shunt was not found. Similar patterns of dispersion of the perfusionindices have been reported resulting from hypoxic pulmonary vasoconstriction in lobarpreparations in the dog (41) which is exacerbated in oleic acid induced pulmonary edema(40). Difficulty in interpreting these data stems from an overall right shift of thecompartmental ventilation versus log VA/Q. ^as a result of an approximately twenty-fold increase in VE compared to a six-fold increase in Q. During exercise, therefore,perfusion to areas of low V. A/Q may be obscured by this shift and will be manifest by anincrease in log SDO and other perfusion related indices.At rest and during exercise to a VO2 of 4.01-min-1 the mean of the Q anddistribution and log SDO compare favorably with those at comparable exercise levelsobtained during similar studies (12, 63, 147) (Table 9). The mean of the V distribution andlog SDa are higher at all exercise levels and are due to the recovery of areas with apparentlyvery high VA/Q distributions. In some cases the excretion of acetone (the gas of highestsolubility) exceeded the retention, violating laws of mass balance. This apparent paradoxcan likely be explained on the basis of an complex interaction between airways heating andcooling, as it was much more pronounced at maximal exercise. Areas of very high VA/Qhave been reported in many experimental situations including high frequency ventilation(68, 108, 138), and have been considered to be artifactual. No alveolar zones with such ahigh VA/Q seem anatomically likely, but gas exchange of soluble gases by the airways ispossible (162). Gas solubility and transport is related to airways temperature, mucoustemperature, water content and thickness. During high intensity exercise there will be bothheating of the airways secondary to increased bronchial blood flow and increases in core36temperature, as well a cooling secondary to hyperventilation. It is possible that excretion ofthe soluble gases deposited in the airway mucosa and mucous earlier during the experimentmay be enhanced by the increase in body temperature. The indices of ventilationheterogeneity E(A-a)D, DISPE and log SDv. would be most likely to be affected and do notexplain the significant increases in the perfusion related indices.The increase in VA/Q heterogeneity with exercise, manifest by increasing [A-a]D02(p) is greater than reported in other studies (12, 63) and is not an effect of the higherexercise intensity achieved since [A-a]D02(p) was higher even at the submaximalworkloads. The reasons for this are not apparent but may reflect an unique characteristic ofthis highly trained subject population.Diffusion limitation The MIGET analysis of pulmonary gas exchange accounts for V. A/Q mismatch andintra-pulmonary shunt. [A-a]1)02 (o-p) therefore represents that portion of the [A-a]D02which is not accounted for by those factors and represents diffusion limitation orextrapulmonary shunt. A 1% extra-pulmonary shunt would result in a fall in Sa02 of about0.7% and a decrease in Pa02 of about 7 torr accounting for most of the difference betweenobserved and predicted values of [A-a]D02. The issue of extrapulmonary shunt thereforebecomes crucial in the detection of diffusion limitation. Since extrapulmonary shunt was notmeasured in this study this possibility canot be dismissed. Previous work (161) duringexercise at sea level and altitude has shown minimal (< 0.18 %) shunt, although thesemeasurements are extremely difficult to make. If it is assumed that the subjects in thepresent study are similar with respect to extrapulmonary shunt then a 0.18% shunt in thepresent study would account for approximately 1 ton of the [A-a]D02 (o-p).Although the [A-a]D02(o-p) demonstrated a correlation of 0.60 with pulmonarytransit time over 60% of variance between subjects is not accounted for by PTT. There aremany possible explanations and making a direct comparison is fraught with difficulties. The37comparison is only as good as the measures used to make it. Pulmonary transit time is onlyan indirect indicator of pulmonary capillary transit time and therefore the relationshipdescribed above reflects uncertainty in this measure as well as uncertainty related to thedetection of diffusion limitation by MIGET.Table 9. Comparison of MIGET data from present study with other investigatorsif021•min-1Mean Q Mean V Log SD(*) Log SM DISPR* DISPE DISPR*_E StudyRest1.00±0.231.11±0.170.63±0.091.91±0.991.28±0.200.68±0.230.38±0.200.35±0.050.28±0.131.09±0.550.42±0.100.26±0.042.53±2.131.08±0.290.68±0.433.88±3.071.07±0.290.59±0.275.41±4.471.99±0.511.16±0.61p122.46 4.21 0.44 0.90 2.83 3.92 6.09±0.62 ±1.50 ±0.12 ±0.47 ±2.09 ±3.01 ±4.65 P3.16 3.91 0.44 0.46 1.64 1.38 2.722-3 ±0.66 ±0.83 ±0.05 ±0.05 ±0.25 ±0.17 ±0.32 12.79 2.98 0.34 0.31 0.93 0.82 1.48±0.69 ±0.77 ±0.07 ±0.03 ±0.28 ±0.28 ±0.54 22.94 7.27 0.64 1.09 4.63± 5.55 9.34±0.62 ±3.67 ±0.17 ±0.29 2.29 ±2.71 ±4.51 P>3-4.5 * * 0.55±0.17 0.49±0.20 * * * 33.33 3.96 0.58 0.36 1.44 0.96 2.14±0.50 ±0.61 ±0.30 ±0.09 ±0.81 ±0.47 ±1.10 24.44 11.82 0.78 1.18 5.73 5.45 10.39>5.0 ±0.53 ±2.90 ±0.11 ±0.14 ±1.33 ±1.03 ±2.12 p* = not reported. P= present study, 1= study of Schaffartzik et al., group 1 subjects (147),2= Hammond et al., (63), 3= Bebout et al., (12)Effect of uncertainty in cardiac output on MIGET dataSmall non significant differences were noted in heart rate between measures madeduring the cardiac output determinations and the inert gas measures, at rest and submaximal38exercise. During maximal exercise, the two measures were nearly identical. On average,heart rate was 6-7 beats per minute higher during the nuclear medicine studies, resulting inan error of 12% at rest, 5.7% during light exercise and 4.4% during heavy exercise. Thelikely explanation , relates to small differences in posture during the two studies; during theinert gas studies the subjects were free to assume the posture that they were the mostcomfortable with, usually in a relaxed position holding on to the bars of the ergometer. Incontrast, during the cardiac output determinations the subjects were required to hold on tothe gamma camera, an awkward position at best, which likely changed their pedalingefficiency slightly.Although cardiac output was not measured simultaneously with the MIGETdeterminations, uncertainty of the cardiac output measurements is unlikely to affect the inertgas calculations with respect to the recovered VA/Q. ^or the predicted [A-a]D02.In theory, uncertainty in measurement of cardiac output, would be most apparent in theinsoluble gases (small X ) as the mixed venous levels (P■f) of the inert gases are calculatedfrom arterial and mixed expired (PE) levels:Pi = Pa +PEVEA.QTFor highly soluble gases, X. is large and Pa dominates the right hand side of the equation.For insoluble gases such as SF6, the term (pE•E)/( )vQT) dominates the equation anduncertainty in QT will be transmitted directly to calculations of retention and excretion data.^.and from there into the calculation of the VA/Q distribution.This issue has been addressed by Wagner et al., (169) who compared recoveredVA/Q distributions, residual sum of squares (model fit), [A-a]D02 (p), and Pa02 fordetection of diffusion limitation in ten patients at rest using measured cardiac output and asensitivity analysis over assumed cardiac outputs of 2 to 121•min -1 . Little effect of assumedcardiac output was observed on most of the variables of interest. Log SD(*) was insensitive39to large changes in cardiac output, although slightly more effect was observed on log SDv.The predicted values of Pa02, PaCO2 and [A-0)02 were also insensitive to assumedcardiac output and residual sum of squares was unaltered. These findings were alsoconfirmed in the present study. Table 10 gives the residual sum of squares (RSS) and meanQ, mean V, logSDa and log SDV for a poorly fitting data set varying cardiac outputs from15 to 40 1.min-1 .Table 10. Effect of varying cardiac output on residual sum of squares and indices ofdispersion 1•min-1RSS^mean ofdistribution mean of Vdistribution Log SDO Log SIDT15 114 6.52 12.35 0.558 1.02520 113 4.83 9.92 0.535 1.15325 112 3.84 8.40 0.515 1.25330 111 3.17 7.35 0.496 1.13435 110 2.71 6.52 0.486 1.39840 109 2.37 5.90 0.473 1.457The mean of the ventilation and perfusion distributions are sensitive to alterations incardiac output but are not of great importance since it is dispersion that determines VA/ Qmismatch. The indices of dispersion log SDa and log ST:07 as described by Wagner et al.,(169) are relatively insensitive to changes in cardiac output; for an increase in cardiac outputof over 150%, log SDa decreased by 15% while log SDI*/ increased by 14%. It can beconcluded that the pattern of the VA/Q relationship recovered from inert gas data is notaffected by uncertainties in cardiac output, and cannot explain the difficulties in fitting theSF6 data.Maintenance of steady state conditionsSteady state conditions in the strictest sense cannot be attained at exercise levelsabove the anaerobic threshold and mean ventilation changed from 164.2 to 178.11-min -1over the data collection during maximal exercise. This was an increase of 7.8%, which is40relatively small and unlikely to affect the results as the very high levels of blood flow andventilation ensure rapid equilibration rates.Blood and plasma volume Methods for determination of blood and plasma volume are based on measurementof dilution of a known amount of a tracer material. The use of radiolabelled human serumalbumin (RISA) is currently the accepted method for routine measurement of plasmavolume in human subjects (76), although the use of albumin as an intravascular markerlikely overestimates the volume of distribution by a factor of over 5% when compared tolarger macromolecules such as fibrinogen (15). The RISA method for measurement ofplasma volume compares favorably to measurements made with Evans blue dye, carbonmonoxide gas, and 99mTc labeled erythrocytes (160), although, as would be expected fromthe preceding statement, RISA measurements agree closest with Evan blue dye measureswhich also uses albumin as the carrier molecule for the tracer. When red cell mass is notmeasured independently, but estimated from venous hematocrit, the accuracy of thedetermination is dependent on correction of venous hematocrit to whole body hematocrit(76).The subjects were both plasma volume expanded and had increased red blood cellmass. These findings have been documented in the past by a number of authors. Cross-sectional studies indicate that highly trained men and women have blood volumes that areapproximately 25% larger than sedentary subjects (38, 88) with the proportionate increase inplasma volume being greater than the proportionate increase in red cell mass (21). Thesubjects had plasma volumes 44% greater than predicted for normal subjects although themean value of 87.5 ml•kg-1 whole blood volume for this study compares favorably withvalues reported by Kjellberg et al., (88) and others (21, 38). Little is known about themechanisms of the changes in hematological parameters, however it is postulated thatincreases in circulating plasma proteins are an important factor in chronic hypervolemia, as41are renal mechanisms involving renin-angiotensin-aldosterone, and vasopressin (31). Thishypervolemia may offer advantages with respect to thermoregulation and hemodynamicsand has shown a strong correlation with VO2 (31). Changes in blood volume associatedwith chronic endurance exercise may account for up to one half of the difference in strokevolume between trained and untrained men (73) and a decrease in plasma and blood volumewith detraining has been shown to decrease both VO2 max and stroke volume in enduranceathletes (32).Radionuclide cardiography First pass and gated radionuclide cardiography offer many advantages over othermethods for determination of cardiac output and cardiac volumes. The method is relativelynon-invasive requiring only an intravenous line and serial measurements can be mademaking these techniques suitable for exercise studies, particularly those involving more thanone exercise level. A combination of first pass and gated determination of cardiac outputwas chosen for the following reasons: First pass techniques also allowed simultaneousdetermination of right-heart to left heart pulmonary transit times, and measurement ofcardiac output at rest by both methods enabled the subsequent gated determinations to becorrected for attenuation of counts by the chest wall without on relying assumptions aboutventricular geometry or depth. We used the gated technique during light and heavy exerciseto minimize radiation exposure to the subjects, and first pass was only done during maximalexercise because of concern regarding motion artifact and data acquisition time.First pass radionuclide cardiography has been compared with more traditionalmethods of cardiac output determination both at rest and exercise and shows goodagreement with values obtained by thermodilution (83), contrast ventriculography and gatedequilibrium radionuclide angiography (122). Gated equilibrium radionuclide angiography,similarly has shown excellent agreement with contrast ventriculography (122, 151), direct42Fick method (111) and thermodilution (33). Thus both radionuclide methods can be usedwith confidence that they are valid means of determining cardiac output.Cardiac function in athletesThe term "athletes heart" has been used to describe the alteration in cardiac structureand function which accompany regular physical conditioning. In recent years, M-Modeechocardiography and gated and first pass radionuclide angiography have allowed a moreaccurate picture of the effects of exercise on the heart. Maron (103) extensively reviewed 28echocardiographic studies in athletes from a variety of athletic backgrounds. Regular athletictraining produces an increase in left ventricular mass, even when corrected for body size andincreases in left ventricular cavity dimension especially in endurance athletes. Leftventricular end-systolic dimension is usually increased, and many studies report increases inposterior left ventricular wall as well as ventricular septal thickness (103). Enlargement ofthe left atrium is also commonly reported, as is an increase in right ventricular mass. Theeffects of training on the heart has been examined longitudinally (94, 95, 115, 134) and theavailable data indicate that the nature of alterations in cardiac structure are somewhatdependent on the type of training stimulus, with strength athletes exposed to predominantlya pressure load tending to show an increase in left ventricular thickness while those exposedto endurance training, and hence a volume load, showed an increase in ventricular volume.In a longitudinal study, Rerych et al., (134) studied collegiate level swimmers withfirst pass radionuclide angiography and found an increase in resting stroke volume, enddiastolic volume, end systolic volume and a decrease in ejection fraction after six months ofswim training. During exercise, maximal cardiac output was increased predominantly as aresult of increases in end diastolic volume, while ejection fraction was unchanged from pre-training levels. Fagard et al., (46) studied highly trained amateur and professional cyclistswith echocardiography in both the competitive and rest seasons and compared them torecreationally active controls. During the resting season the athletes had smaller heart size43and significant decreases in total left ventricle end systolic diameter, due to reductions inwall thickness than in the competitive season. When compared to controls the cyclists hadmuch higher ratios of wall thickness to internal dimensions in both the competitive seasonand the resting season suggesting that both volume and pressure (presumably due toisometric upper body exercise during cycling) were responsible for the increase in cardiacdimensions.Comparison with previous investigations The data from the present study are similar to those obtained by direct Ficktechnique, radioangiography and echocardiography (Table 11). At rest, the cardiac outputwas slightly higher in both the present study and the other radioangiographic study (134),but resting heart rate and cardiac output would be expected to be the most influenced byextraneous variables such as subject arousal etc. The resting stroke volume index from thepresent study is comparable to the mean value for the other studies cited. During exercise ata VO2 of 1.5-21•min-1 and 'T02 of 3-41•min-1 the present data again are very similar tothose previously reported. During maximal exercise, although cardiac output is very similarto that obtained in the other studies, stroke volume index is higher. This may possibly beexplained by the study design. Collecting clean first pass data was of paramount importancefor measurement of pulmonary transit times and motion artifact was of great concern. Itwas elected to inject the labeled red blood cells at the end of 3 minutes of exercise and datawere acquired for 150 seconds after injection. At this work intensity heart rate clearly hadnot fully stabilized and it may be possible that stroke volume may have fallen and heart raterisen, with redistribution of blood flow as the exercise continued, which would not beevident during the data collection.Effect of increasing exercise intensity on cardiac volumes Exercise physiology text books (20) report that stroke volume levels off withincreasing exercise intensity, possibly as a result of decreased diastolic filling time and44reduced end diastolic volume. Indeed, some authors have reported a decrease in enddiastolic volume with increasing exercise (57) while others have reported a decrease, nochange or an increase in end diastolic volume (49, 124). A decrease in stroke volume withincreasing exercise intensity was not found in any of the subjects, and only one subject (#9)showed a plateau in stroke volume between heavy and maximal exercise. These data wouldtend to support the recent observations by Gledhill et al., (61) who found increasing strokevolume with increasing exercise in endurance trained athletes without a shortening of leftventricular ejection time compared to sedentary controls, leading the authors to conclude thatenhanced preload and Frank-Starling mechanism were important factors for maintenance ofstroke volume in athletic subjects.45Table 11. Cardiac output and volume indices comparison with other studies Exercise Cardiac End Diastolic End Systolic Stroke Studyintensity Index Volume Volume Volumeindex Index Indexmi•m-2 ml•m-2 ml•m-2 ml•m-2Rest 3.4±0.4 73±8 27±5 51±10 present R-G3.4±0.5 85±20 22±6 64±15 Rerych R-FP,(134)3.0±0.9 85±14 37±11 48±15 Ginzton E (57)2.4±0.03 38±6 Hermannsen D(66)iT02 1.5-2.0 8.0+0.7 91±13 22±7 69±8 present R-G1•min-18.8±1.7 90±17 24±7 66±13 Ginzton E (57)8.2±1.3 67±9 Ekblom F (43)iT02 3-4 12.5±1.2 98±11 21±5 81±11 present study1•min-1 R-G13.8±1.3 90±7 Ekblom F (43)11.1±1.5 66±9 Hermannsen D(66)i[02 4+ 16.5±1.8 100±12 present study1-min-1 R-FP16.3±4.4 104±19 14±7 90±20 Rerych R-FP(134)18±0.6 95±4 Ekblom F (43)R-FP = Radioangiography first pass method, R-G = radioangiography gated method, E=echocardiography, F = Fick method, D= Dye dilution technique.46Pulmonary transit timesTheoryThe mean transit time for a well mixed indicator to flow through a specific volumeat a given flow rate is described by the relationship: transit time = volume/flow. The timerequired for an indicator to flow past an observation point down stream from an entry pointis related not only to the time it takes the bolus to flow past the point but also how quickly itarrived there. Transit time is the time that a bolus remains in a compartment if it is injecteddirectly and instantaneously into the compartment. The first moment describes not only thetime that the indicator is in the compartment but also how quickly or slowly it arrived there.The first moment therefore represents the summation of all transit times up to that point. If itwere possible to deliver indicator material instantaneously into the compartment of interestthe first moment would be the same as the transit time. Transit time of a compartment can bedetermined by subtracting the first moment of the bolus from that of the output curvederived from the compartment.In the case of pulmonary transit times, the bolus or input curve is derived from atime activity curve of the right ventricle and the output curve is derived from the leftventricle; transit time is determined by subtracting the first moment of the right ventricularcurve from the first moment of the left ventricular curve. This method is referred to as thecentroid method. Deconvolution is a mathematical process by which a frequency distributionof transit times (a transfer function, h(t)) can be derived from the input (right ventricular)and output (left ventricular) time activity curves. The assumption is made that since both theinput curve and output curve can be described by a gamma function (153) that h(t) is also agamma function. A series of 20 to 50 curves is then generated with mean transit timessimilar to that obtained by the centroid method and convoluted with the input curveproducing a series of unique output curves. Ridge regression is then applied to determinethe contribution of the curves obtained to the actual output curve and the final transferfunction is obtained. These data when convoluted with the input curve produces a derived47output curve. The process is repeated until a satisfactory result is obtained. It is important tonote that the transit time obtained from either method represents the delay of the bolusthrough pulmonary arteries, arterioles, capillaries, venules, veins, left atrium and leftventricle and does not just represent pulmonary capillary transit time. Incomplete mixing ofthe bolus can lead to either over or underestimation of times, therefore is preferable tomeasure transit time downstream from the site of injection. Other sources of error includepoor bolus technique and cross contamination of time activity curves from overlyingstructures in the chest.Pulmonary transit times have been measured in humans and animals in a variety ofexperimental conditions. There is a clear gravitational difference in regional transit timeswith the shortest transit times at the base and longest times at the apex of the lung (70, 96).As pulmonary blood flow increases, recruitment of blood volume prevents a drasticdecrease in transit time (70). In animals, very close agreement has been obtained betweenradioisotopic methods and dye techniques (44). Similar support for the validity of thistechnique has been established in humans: MacNee et al.,(96) compared pulmonary regionaltransit times obtained by centroid and deconvolution methods pre-operatively with thoseobtained infra-operatively in five patients undergoing lung resection for carcinoma. Excellentagreement between in vivo and in vitro techniques was found with the mean differencebetween methods less than 0.5 seconds. There were no significant differences betweenthose results obtained by deconvolution (mean over all regions = 4.83 sec) and centroid(mean = 4.53 sec) methods.48Pulmonary transit times and exercise in humansExercise results in a decrease in transit time from resting values of about 5 secondsto less than 2.5 seconds in normal subjects (77). The effect of exercise training onpulmonary transit time have also been demonstrated. Rerych et al., (134) reported anincrease in pulmonary transit times in collegiate swimmers after six months of intensiveexercise training, likely reflecting increases in both total blood volume and pulmonary bloodvolume. Resting values in that study and the present one were greater by more than a factorof two, compared to sedentary subjects. Mean pulmonary transit time during exercise was2.8± 0.3 seconds, a result very similar to that of the present study.Pulmonary blood volumePulmonary blood volume has also been measured during a variety of clinicalsituations and during maximal exercise. Radionuclide methods for determination ofpulmonary blood volume have correlated closely with dye dilution techniques (44), but it isimportant to recognize that both techniques measure the volume of blood between the pointof input of the marker and the downstream measurement. Guintini et al., (58) reportedpulmonary blood volumes in patients with a variety of cardiopulmonary conditions and in 5normal men. Pulmonary blood volume at rest was 293±50 ml•m -2 comprisingapproximately 10% of total blood volume and increased with mild exercise. Slightly higherresting values were obtained by Iskandrian et al., (77), with a non significant increaseduring exercise. Exercise training has been shown to produce an increase in pulmonaryblood volume of approximately 10% at rest and 50% during maximal exercise (134)compared to pre-training values. Maximal exercise in trained individuals led to an 80%increase in pulmonary blood volume index from 465 ml•m -2 at rest to 772 ml•m-2 duringmaximal exercise, similar to values of 530 ml•m -2 at rest and 820 ml•m-2 obtained duringthe present study.49Relationship of pulmonary transit time and blood volume to pulmonary capillary transit timeand blood volumeA central issue with respect to the interpretation of the present data concerns therelationship between whole lung transit time and blood volume and pulmonary capillarytransit time and blood volume, since it is the latter two factors which affect pulmonary gasexchange. Unfortunately data addressing this question are sparse and can only give anestimate of these relationships. Backmann and Hartung(8) measured whole lung bloodvolume in cadavers and then estimated the arterial and venous contributions by the injectionof a very viscous fluid that did not enter the capillaries. They estimated that 53% of thewhole lung blood volume or 270±50 ml was in the pulmonary capillaries and suggested thatthis was an upper limit of the measure, as small arterioles and venules would be includedusing this technique. In contrast to this, studies that calculate capillary blood volume fromdiffusing capacity (64, 171) estimate values of less than one half of that value (about 100ml) and may underestimate the true value, as it is a functional rather than an anatomicalmeasure. If these values for blood volume are applied to the resting data from the presentstudy, calculated pulmonary capillary transit time would be between 1.15 and 2.34 secondsand whole lung transit times would be between 4 and 8 times greater than capillary transittimes. During exercise, there are even less data on which to base calculations. Warren et al.,(171) measured pulmonary capillary blood volume at 215 ml using diffusing capacityduring exercise at a VO2 of greater than 4.01•min-1 . Using this information and againaccepting a value of 270 ml as the upper anatomical limit of pulmonary capillary bloodvolume, these values would give an estimated pulmonary capillary transit time for thepresent study of 0.39-0.49 seconds and whole lung transit time would be some 6 to 7.5time greater than the capillary transit time. It is important to recognize that these representaverage values for transit time and because of the skewed nature of both the pulmonary andpulmonary capillary transit curves a significant portion of the cells may have very shorttransit times.50Arterial blood gas and metabolic dataArterial blood measuresOur subjects exhibited Pa02 s ranging from 74 to 100 ton (mean Sa02 of 95.3%)during heavy exercise and from 81 to 109 ton (mean Sa02 of 94.2%) during maximalexercise. These values are higher than the previous values reported by this laboratory (72)and that of Dempsey et al.,(35) however it is likely that this is as a result of the differentmeans of exercising the subjects; running in the previous studies versus cycling in thepresent one. Cycling was chosen as the method of exercising subjects in the present studybecause it facilitated the MIGET data collection, and because measurement of pulmonarytransit times, cardiac output and cardiac volumes is not currently possible using otherexercise forms due to motion artifact. The relatively high values for pH in the present studyof 7.24±0.06, compared with the previous study value of 7.21±0.06 in a very similargroup of subjects exercising at the same relative intensity, likely are as result of the smallermuscle mass involved in cycling. This factor may also contribute to the higher values forPa02. Three subjects from the 1989 study participated in the current one: all had highervalues for Sa02, Pa02 and lower values for pH than in the study involving running. Nocorrection was made for temperature in the previous study, and it is likely that the degree ofhypoxemia in those subjects was overestimated. A temperature increase of about 1°C wasseen after 5 minutes of exercise at VO2 max in the current study, which may underestimatethe increase in temperature seen with running.The subjects were screened for EIH during the initial VO2 max testing with a pulseoximeter and recorded a mean saturation of 92.2±2.18%. This suggest that either theexercise tests were not of sufficient duration to elicit maximal hypoxemia or that the readingsof the pulse oximeter were unreliable during maximal exercise.51Mixed venous P02Mixed venous blood gas data is calculated from the Fick equation and small errors inmeasurement of cardiac output and VO2 will be transmitted directly to the NO2 calculation.This may be more of a problem with the current study as cardiac output and VO2 were notmeasured simultaneously, as they were in other studies using similar methodology (63,161). This is not likely to affect [A-a]D02(p) (63) or indicators of diffusion limitation, [A-a]D02(o-p). Bearing this in mind, the Pir02 data from the present study should beinterpreted with some caution. At rest, the calculated mean NO2 is similar to valuesreported by several other studies employing both direct measurement and calculation fromthe Fick equation (28, 89, 139, 161). During light exercise, the values are similar to thosereported by Cerretelli et al., (28) although they are approximately 6 torr lower than valuesreported by Torre-Bueno et al., (161). During heavy and maximal exercise, calculated NO2was substantially lower than reported for normal subjects exercising at sea level, althoughvalues lower than 17 torr have been reported for subjects exercising at simulated altitude(161) and breathing hypoxic gas mixtures (139).Possible mechanisms of exercise induced hypoxemiaAlthough marked hypoxemia was not seen in the majority of the subjects in thisstudy the individual variation makes it possible to speculate as to causes of exercise inducedhypoxemia in highly trained subjects. The data presented in this paper is consistent withhypoventilation, VAX) mismatch and diffusion limitation as viable mechanisms in thegenesis of EIH, both in groups of subjects and in individuals, although clearly onemechanism may predominate over the other depending on exercise intensity and individualvariability. This is perhaps best illustrated by comparing subject 4 to subject 5. Subject 5,who hypoventilated to the point of CO2 retention at 300 watts, despite a Pa02 of 74 ton,also had evidence for diffusion limitation both at 300 watts and at maximal exercise, withLA-all302(o-p) of 6 and 17 ton respectively. It is not likely that the source of this subject's52hypoventilation was purely mechanical in nature as PaCO2 decreased to 31 torr and PAO2rose to 123 ton at the end of maximal exercise, indicating increased effective alveolarventilation, when given the appropriate stimulation. Subject 4 had little evidence ofhypoventilation as PaCO2 was 35 ton or less during heavy and maximal exercise, howeverthe inert gas data suggests marked diffusion limitation with [A-a]D02(o-p) greater than 30ton at both exercise levels.From the data in the present study, it is now possible to present an explanation ofthe conflicting findings of the previous study (72) and that of Dempsey et al., (35).Dempsey exercised very highly trained individuals at 70-90% of VO2 max whereas thesubjects in the previous study were exercised at 100% of VO2 max. This is evidenced bysimilar oxygen consumption between the two groups but higher VCO2 and lower pH for thesubjects in the study of Hopkins and McKenzie. It is proposed that three main mechanismscontribute to exercise induced hypoxemia. During heavy exercise, hypoxemia results fromrelative hypoventilation, VA/Q mismatch and diffusion limitation. During maximal exercise,diffusion limitation and VA/Q mismatch continues and may in some individuals worsen, butarterial oxygenation may improve, as ventilation increases in response to stimuli such asmarked acidosis. In this study, as in the previous work, there is little evidence forhypoventilation at maximal exercise in young, highly trained subjects. All of the subjectshad significantly lower values for Pa02 and higher values for PaCO2 at the heavy exerciselevel when compared to maximal exercise. In addition, maximal ventilation during theexercise test was significantly less than during the VO2 max determination, indicating thatat least under some circumstances, our subjects were capable of higher levels of ventilation.It seems unlikely that mechanical restriction of ventilation is an important factor in thesesubjects and suggests that the relative hypoventilation and lower levels of arterial oxygenobserved during heavy exercise may have been related to complex interactions between thework of breathing and drives to breathe.53The relationships between whole blood volume, pulmonary blood volume indexand [A-a]DO2(o-p) suggest that in some endurance athletes with EIH that this problemcould be compounded by dehydration and or cardiovascular drift which would lower bloodvolume and possible affect pulmonary blood volume and pulmonary transit. These resultsmust be interpreted with caution, recognizing that due to the small sample size these subjectsmay not be representative of the population at large. There is no direct evidence in thepresent study to support this hypothesis, and there are no published papers that directlyexamine the effect of volume contraction on exercising pulmonary blood volume and EIH.However, a recent study in this laboratory examined the effect of administration of a singledose of furosemide on EIH using ear oximetry (163). Plasma volume was not measureddirectly but changes in plasma volume were estimated from changes in hematocrit.Although there was considerable intersubject variability, there was a strong correlationbetween the difference in the change in plasma volume between placebo and furosemideadministration and the change in saturation between the two conditions (r=0.75, p<0.01).That is, the subjects who had the greatest decrease in plasma volume with furosemide andexercise also had the greatest decrease in Sa02 when compared to the placebo condition.Furosemide has many systemic effects aside from that of diuresis, and it is difficult to makefirm conclusions however this would suggest that the role of plasma volume and pulmonaryblood volume in the genesis of EIH is worthy of further investigation.Several authors have described alterations in a number of indirect indicators ofbarrier function post-exercise including increases in residual volume, decrease in diffusingcapacity for carbon monoxide (DLCO), and decreases in transthoracic electrical impedancethat are consistent with pulmonary edema (22, 23, 104, 129, 130). Additionally Hammondet al., (63) described widened [A-0)02 and VA/Q inequality following heavy exercise andsuggested that a structural change in the lung could account for their findings. Despite this,little further investigative work has been done in this area, particularly in athleticpopulations. Vaughan et al., (164) failed to find an increase in lung water (measured by the54indicator dilution method) with sustained exercise, although an initial increase in lung waterwas attributed to a redistribution of blood flow, however, their subjects were exercising at awork load of less that 150 watts and a cardiac output of less that 20 1-min -1 . An interestingstudy by Gallagher et al., (56) examined chest radiographs in five males following a V02max test. No evidence of pulmonary edema was found, however this study suffered fromthe same problem as the one of Vaughn et al., the subjects studied were normal non-athleticmales who have not been shown to have evidence for inadequate gas exchange duringexercise. Pulmonary capillary failure has been demonstrated in the rabbit lung at 40 mmHgof transmural pressure and it has been estimated in human lungs, during exercise of lessthan 4.0 1-min-1 , that transcapillary pressures of 36 mmHg would be likely, leaving littlesafety margin (179). The consequences of such capillary failure in humans would beincreased capillary permeability at the lower end of the spectrum and frank pulmonaryhemorrhage with higher pressures.Recently Schaffartzik et al., (147) have argued that sustained increases in log SDOpersisting during recovery from exercise, during exercise and normobaric hypoxia isindicative of pulmonary edema. This subgroup of subjects had significant impairment ofpulmonary function after exercise compared to controls and also had lower arterial saturationand pH associated with exercise. In the present study, it was possible to define twosubgroups with respect to log SDQ however, there were no significant differences betweengroups with respect to pH, Sa02, Pa02 or other measured parameters. It is not possible tocomment on the possibility of pulmonary edema in the present study as no inert gassamples were taken during recovery and the overall shift in the VA/Q distribution mayobscure alterations in blood flow distribution towards zones of low VA/Q . Also thescreening process used in selecting subjects excluded any who had post-exercise cough,hem optysis or decrease in pulmonary flows which could be indicative of pulmonary edema.55Summary of findings This paper describes VA/Q mismatch, diffusion limitation and a variable decrease inPa02 in highly trained athletes during exercise. The decrease in Pa02 is less thanpreviously described for similar athletic populations and may reflect the small muscle massused in cycling exercise. The increase in V A/Q mismatch was largely due to increases in theperfusion related indices of dispersion and does not rule out pulmonary edema as aplausible explanation for these findings. The [A-a]D02(o-p) was significantly correlatedwith pulmonary transit time and Pa02 during maximal exercise could be predicted on thebasis of pulmonary transit time, VF/VCO2 and ^as indicated by (A-a)D. [A-a]1)02(o-p) was correlated with both pulmonary blood volume index and whole blood volumesuggesting that blood volume expansion seen in highly trained athletes and the ability toexpand pulmonary blood volume may provide important defenses against pulmonarydiffusion limitation.56REFERENCES1. Akabas, S.R., A.R. Bazzy, H. Reichman, G.G. Haddad. Training with inspiratoryflow resistive (IFR) loads increases the oxidative capacity of the sheep diaphragm.Am. Rev. Respir. Dis. 1985;131:A307.2. Andersen, P., B. Saltin. Maximal perfusion of skeletal muscle in man. J. Physiol.(Land.) 1985;366:233-249.3. Anholm, J.D., R.L. Johnson, M. Ramanathan. Changes in cardiac output duringsustained maximal ventilation in man. J. Appl. Physiol. 1987;63(1):181-187.4. Anholm, J.D., J. Stray-Gundersen, M. Ramanathan, R.L. Johnson. Sustainedmaximal ventilation after endurance exercise in athletes. J. Appl. Physiol.1989;67(5):1759-1763.5. /strand, P.-O., T.E. Cuddy, B. Saltin, J. Stenberg. Cardiac output duringsubmaximal and maximal work. J. Appl. Physiol. 1964;19(2):268-274.6. Aubier, M., N. Viires, G. Syllie, R. Mozes, C. Roussos. Respiratory musclecontribution to lactic acidosis in low cardiac output. Am. Rev. Respir. Dis.1982;126:684-652.7. Ayres, L.N. Carbon monoxide diffusing capacity. In: Wilson AF, ed. Pulmonaryfunction testing: indications and interpretation. New York: Grune and Stratton,1985:8. Backmann, R., W. Hartung. Differentiating measurements of blood volumes inisolated human lungs. International Symposium on Pulmonary Circulation. Prague:Karger, 1969: 327-337.9. Barr, P.O.-2., M. Beckman, H. Bjurstedt, J. Brismar, C.M. Hesser, G. Matel.Time course of blood gas changes provoked by light and moderate exercise in man.Acta Physiol. Scand. 1964;60:1-17.10. Bartlett, D. Post-natal growth of the mamalian lung : influence of exercise andthyroid activity. Resp. Physiol. 1970;9:50-57.11. Bartlett, R.G., H.F. Brubach, H. Specht. Oxygen cost of Breathing. J. Appl.Physiol. 1958;12(3):413-424.12. Bebout, D.E., D. Storey, J. Roca, et al. Effects of altitude acclimatization onpulmonary gas exchange during exercise. J. Appl. Physiol. 1989:67(6):2286-2295.13. Bechbache, R.R., J. Duffin. The entrainment of breathing frequency by exerciserhythm. J. Physiol. 1977;271:553-561.14. Bender, P.R., B.J. Martin. Maximal ventilation after exhausting exercise. Med. Sci.Sports Exercise 1985; 17(1): 164-167.^15.^Bent-Hansen, L. Initial plasma disappearance and distribution volume of•[1311]albumin and 125 I]fibrinogen in man. Acta Physiol Scand 1989;136:455-461.5716. Bertolin, J.F., J. Caries, A. Tellac. Assessment of the ventilatory performance ofathletes using the maximal expiratory flow-volume curve. Int. J. Sports Med.1986;7:80-85.17. Bjurstedt, H., 0. Wigertz. Dynamics of arterial oxygen tension in response tosinusoidal work load in man. Acta Physiol. Scand. 1971;82:236-249.18. Bradley, M.E., D.E. Leith. Ventilatory muscle training and the oxygen cost ofsustained hyperpnea. J. Appl. Physiol. 1978;45(6):885-892.19. Brancatisano, A., T.C. Amis, A. Tully, L.A. Engel. Blood flow distribution withinrib cage muscles. J. Appl. Physiol. 1991;70(6):2559-2565.20. Brooks, G.A., T.D. Fahey. Exercise Physiology: Human Bioenergetics and itsApplications.Toronto: John Wiley & Sons, 1984:726.21. Brotherhood, J., B. Borozovic, L.G.C.E. Pugh. Hematological status of middle-and long-distance runners. Clin. Sci. MoL Med. 1975;48:139-145.22. Buono, M.J., S.H. Constable, A.R. Morton, T.C. Rotkis, P.R. Stanford, J.H.Wilmore. The effects of an acute bout of exercise on selected pulmonary functionmeasurements. Med. Sci. Sports Exerc. 1981;13:290-293.23. Buono, M.J., J.H. Wilmore, F.B. Roby. Indirect assessment of thoracic fluidbalance following maximal exercise in man. J. Sports Sci. 1983;1:217-226.24. Bye, P.T., S.A. Esau, K.R. Walley, P.T. Macklem, R.L. Pardy. Ventilatorymuscles during exercise in air and 02 in normal men. J. Appl. Physiol.1984;56:464-471.25. Bye, P.T.B., G.A. Farkas, C. Roussos. Respiratory factors limiting exercise. Ann.Rev. Physiol. 1983;45:439-451.26. Byrne-Quinn, E., J.V. Weil, I.E. Sodal, G.F. Filley, R.F. Grover. Ventilatorycontrol and the athlete. J. Appl. Physiol. 1971;30(1):91-98.27. Casaburi, R., J. Daly, J.E. Hansen, R.M. Effros. Abrupt changes in mixed venousblood gas composition after the onset of exercise. J. Appl. Physiol.1989;67(3):1106-1112.28. Cerretelli, P., R. Sikand, L.E. Farhi. Reajustments in cardiac output and gasexchange during onset of exercise and recovery. J. Appl. Physiol. 1966;21(4):1345-1350.29. Coast, J.R., R.A. Jensen, S.S. Cassidy, M. Ramanathan, J. R. L. Johnson.Cardiac output and 02 consumption during inspiratory threshold breathing. J. Appl.Physiol. 1988;64(4):1624-1628.^30.^Coates, J., G. Johnson, A. MacDonald. Breathing frequency and tidal volume:relationship to breathlessness. In: Porter R, ed. Breathing: Herring-BreuerCentenary Symposium. London: Churchill, 1970: 297-314.5831. Covertino, V.A. Blood volume: Its adaptation to endurance training. Med. SciSports Exerc 1991;23(12):1338-1348.32. Coyle, E.F., M.K. Hemmert, A.R. Coggan. Effects of detraining on cardiovascularresponses to exercise: role of blood volume. J. Appl. Physiol. 1986;60(1):95-99.33. Dehmer, G.J., B.G. Firth, L.D. Hillis, P. Nicod, J.T. Willerton, S.E. Lewis.Nongeometric determinations of right ventricular volumes form equilibrium bloodpool scans. Am. J. Cardiol., 1982;49:79-84.34. Dempsey, J.A. Is the lung built for exercise? Med. Sci. Sports Exerc.1986;18(2):143-155.35. Dempsey, J.A., P.G. Hanson, K.S. Henderson. Exercised-induced arterialhypoxaemia in healthy human subjects at sea level. J. Physiol. (Lond)1984;355:161-175.36. Dempsey, J.A., D. Pegelow, R. Fregosi. Mechanical vs chemical determinants ofhyperventilation in heavy exercise (Abs.). Med. Sci. Sports Exerc. 1982;14(suppl.):131.37. Derks, C.M. Ventilation-perfusion distribution in young and old volunters duringmild exercise. Bull. Eur. Pathophysiol. Respir. 1980;16:145-154.38. Dill, D.B., K. Braithwaithe, W.C. Adams, E.M. Bernauer. Blood volume ofmiddle distance runners: effects of 2300 m altitude and comparison with non-athletes. Med. Sci. Sports Exerc. 1974;6:1-7.39. DiPrampero, P.E. Metabolic and circulatory limitations to V02 max at the wholeanimal level. J. Exp. Biol. 1985;115:319-331.40. Domino, K.B., F.W. Cheyney, B.L. Eisenstein, M.P. Hlastala. Effect of regionalalveolar hypoxia on gas exchange in pulmonary edema. J. Appl. Physiol.1992;145:340-347.41. Domino, K.B., B.L. Eisenstein, F.W. Cheney, M.P. Hlastala. Pulmonary bloodflow and ventilation-perfusion heterogenity. J. Appl. Physiol. 1991;71(1):252-258.42. Domino, K.B., M.P. Hlastala, B.L. Einstein, F.W. Cheney. Effect of regionalalveolar hypoxia on gas exchange in dogs. J. Appl. Physiol. 1989;67(2):730-735.43. Ekblom, J., L. Hermansen. Cardiac output in athletes. J. Appl. Physiol.1968;25:619-625.44. Ellis, J.H., P.P. Steele. Comparison of pulmonary blood volume in dogs byradiocardiography and dye dilution. J. Appl. Physiol. 1974;37(4):570-574..^.45.^Evans, J.W., P.D. Wagner. Limits on VA/Q distributions from analysis ofexperimental inert gas elimination. J. Appl. Physiol. 1977;42:889-898.5946. Fagard, R.J., A. Aubert, R. Lysens, J. Staessen, L. Vanees, A. Amery.Noninvasive assessment of seasonal variations in cardiac structure and function incyclists. Circulation 1982;67(4):896-901.47. Fairbarn, M.S., K.C. Coutts, R.L. Pardy, D.C. McKenzie. Improved respiratorymuscle endurance of highly trained cyclists and the effects on maximal exerciseperformance. Int J Sports Med 1991;12(1):66-70.48. Farkas, G.A., C. Roussos. Histochemical and biochemical correlates of ventilatorymuscle fatigue in emphysematous hamsters. J. Clin. Invest. 1984;74:1214-1220.49. Fisman, E.Z., G. Frank, E. Ben-Ari, et al. Altered left ventricular volume andejection fraction responses to supine dynamic exercise. J. Am Coll Cardiol1990;15(3):582-588.50. Fixler, D.E., J.M. Atkins, J.H. Mitchell, L.D. Horowitz. Blood flow torespiratory, cardiac, and limb muscles during graded exercise. Am. J. Physiol.1976;231(5):1515-1519.51. Folinsbee, L.J., E.S. Wallace, J.F. Bedi, S.M. Horvath. Exercise respiratorypattern in elite cyclists and sedentary subjects. Med. Sci. Sports Exerc.1983;15(6):503-509.52. Freedman, S., N. Cooke, J. Moxham. Production of lactic acid by respiratorymuscles. Thorax 1983;38:50-54.53. Fregosi, R.F., M. Sanjak, D.J. Paulson. Endurance training does not affectdiaphragm mitochondrial respiration. Resp. Physiol. 1987;67(2):225-237.54. Fritts, H.W., J. Filler, A.P. Fishman, A. Courand. The efficency of ventilationduring voluntary hyperpnea. J. Clin. Invest. 1959;38:1339-1348.55. Gale, G.E., J. R. Torre-Bueno, R. E. Moon, H. A. Saltzman and P. D. Wagner.Ventilation-perfusion inequality in normal humans during exercise at sea level andsimulated altitude. J. Appl Physiol. 1985;58(3):978-988.56. Gallagher, C.G., W. Huda, M. Rigby, D. Greenberg, M. Younes. Lack ofradiographic evidence of interstitial pulmonary edema after maximal exercise innormal subjects. Am. Rev. Respir. Dis. 1988;137:474-476.57. Ginzton, L.E., R. Conant, M. Brizendine, M.K. Laks. Effect of long-term highintensity aerobic training on left ventricular volume during maximal upright exercise.J Am Coll Cardiol 1986;14(2):364-371.58. Giuntini, C., M.L. Lewis, A.S. Luis, R.M. Harvey. A study of the pulmonaryblood volume in man by quantitative radiocardiography. J. Clin. Invest.1963;42(10):1963.59. Glazier, J.B., J.M.B. Hughes, J.E. Maloney, J. West. Measurements of capillarydimensions and blood volume in rapidly frozen lungs. J. AppL Physiol.1969;26:65-76.6060. Gledhill, N., A.B. Froese, F. J. Buick and A. C. Bryan. Va/Q inhomogenity andAaDO2 in man during exercise: Effect of SF6 breathing. J. Appl Physiology1978;45(4):512-515.61. Gledhill, N.L., D. Cox, V. Jamnik. Stroke volume of endurance athletes does notplateau and the related cardiac time intervals are paradoxical. Med. Sci. SportsExerc. 1992 ;24(5(Suppl.)):S29.62. GOrski, J., Z. Naimiot, J. Giedrojc. Effect of exercise on metabolism of glycogenand triglycerides in the respiratory muscles. Pfliigers Archiv. 1978;377:251-254.63. Hammond, M.D., G.E. Gale, S. Kapitan, A. Ries, P.D. Wagner. Pulmonary gasexchange in humans during exercise at sea level. J. Appl. Physiol.1985;60(5):1590-1598.64. Harris, P., D. Heath. The measurement of blood volume in the lungs. In: Humanpulmonary circulation its form in health and disease. Edinburg, London and NewYork: Churchill Livingstone, 1977: 105-117.65. Harrop, G.A. The oxygen and carbon dioxide content of arterial blood in normalindividuals and in patients with anemia and heart disease. J. Exp. Med.1919;30:241-257.66. Hermansen, L., B. Ekblom and B. Saltin. Cardiac output during submaximal andmaximal treadmill and cycle exercise. J. Appl. Physiol. 1970;29(1):82-86.67. Hesser, C.M., D. Linnarsson, L. Fagraeus. Pulmonary mechanics and work ofbreathing at maximal ventilation and raised air pressure. J. Appl.Physiol.1981;50(4):747-753.68. Hlastala, M.P., D.D. Ralph, A.L. Babb. Influence of gas physical properties onpulmonary gas exchange. Adv. Exp. Med. Biol. 1988;227:33-38.69. Hlastala, M.P., H.T. Robertson. Inert gas elimination characteristics of the normaland abnormal lung. J. Appl. Physiol. 1978;44(2):258-266.70. Hogg, J.C., B.A. Martin, S. Lee, T. McLean. Regional differences in erythrocytetransit in normal lungs. J. Appl. Physiol 1985;62:1236-1243.71. Hogg, J.C., T. McLean, B.A. Martin, B. Wiggs. Erythrocyte transit and neutrophilconcentration in the dog lung. J. Appl. Physiol. 1988;65(3):1217-1225.72. Hopkins, S.R., D.C. McKenzie. Hypoxic ventilatory response and arterialdesaturation during heavy work. J. Appl. Physiol. 1989;67(3):1119-1124.73. Hopper, M.K., A.R. Coggan, E.F. Coyle. Exercise stroke volume relative toplasma-volume expansion. J. Appl. Physiol. 1988 ;64 (1) :404-408.74. Houston, C.S., J.R. Sutton, A. Cymerman, J.T. Reeves. Operation Everest II: manat extreme altitude. J. Appl. Physiol. 1987;63(2):887-882.75. Ianuzzo, C.D., M.J. Spaulding, H. Williams. Exercise-induced glycogen utilizationby the respiratory muscles. J. Appl. Physiol. 1987;62(4):1405-1409.6176. International Committee for Standardization in Hematology. Recommended methodsfor measurement of red-cell and plasma volume. J. Nucl. Med. 1980;21:793-800.77. Iskandrian, A.S., A.-H. Hakki, S.A. Kane, B.L. Segal. Changes in pulmonaryblood volume during upright exercise. Chest 1982;82(1):54-58.78. Jensen, J.I., S. Lyager, O.F. Pedersen. The relationship between maximalventilation, breathing pattern and mechanical limitation of ventilation. J. Physiology1980;309:521-532.79. Johnson, R.L., M. Reid. Limits of oxygen transport to the diaphragm. Am. Rev.Respir. Dis. 1979;119:113-114.80. Johnson, R.L.J., W.S. Spicer, J.M. Bishop, R.E. Forster. Pulmonary capillaryblood volume, flow and diffusing capacity during exercise. J. Appl. Physiol.1960;15:893-902.81. Jones, N.L., K.J. Killian, D. Stubbing. The thorax in exercise. In: Roussos C, ed.Thorax. New York: Marcel Dekker, 1985:82. Keens, T.G., V. Chen, P. Patel, P. O'Brien, H. Levison, C.D. Ianuzzo. Cellularadaptations of the ventilatory muscles to a chronic increased respiratory load. J.Appl. Physiol. 1978;44(6):905-908.83. Kelbwk, H., L. Heslet, K. Skagen, 0. Munck, J. Godtfredsen. First passradionuclide cardiography for determination of cardiac output: evaluation of animproved method. Mt. J. Cardiol. 1989;23:79-85.84. Kelbxk, H., H. Jensen, P.V. Madsen, 0. Munck, J. Godtredsen. Determination ofcardiac output by first passage radiocardiography. Scand J. Clin Lab Invest1987;47:469-474.85. Kelman, G.R. Digital computer procedure for the conversion of oxygen tension intosaturation. J. Appl. Physiol. 1966;21:1375-1376.86. Kelman, G.R. Digital subroutines for the conversion of PCO2 into blood CO2content. Resp. Physiol. 1967;3:111-116.87. Kelman, G.R. Computer programs for the production of 02-0O2 diagrams. Resp.Physiol. 1968;4:260-269.88. Kjellberg, S.R., U. Rudhe, T. Sjostrand. Increase of the amount of hemoglobin andplasma volume in connection with endurance training. Acta Physiol. Scand.1949;19:146-151.89. Lang, F., S. Daum. Mixed venous P02 during exercise in patients with differentcardiopulmonary function. Respiration 1983;44(3):161-170.90.^Leiberman, D.A., J.A. Faukner, A.B. Craig, L.C. Maxwell. Performance andhistochemical composition of guinea pig and human diaphragm. J. Appl. Physiol.1973;34(2):233-237.6291. Leith, D.E., M. Bradley. Ventilatory muscle strength and endurance training. J.Appl. Physiol. 1976;41(4):508-516.92. Levinsky, N.G. Acidois and Alkalosis. In: Thorn GW, Adams RD, Braunwald E,Isselbacher KJ, Petersdorf RG, ed. Harrison's Principals of Internal Medicine.Eighth ed. New York: McGraw-Hill, 1977: 377.93. Loke, J., D.A. Mahler, J.A. Virgulto. Respiratory muscle fatigue after marathonrunning. J. Appl. Physiol. 1982;52:821-824.94. Longhurst, J.C., A.R. Kelly, W.J. Gonyea, J.H. Mitchell. Echocardiographic leftventricular masses in distance runners and weight lifters. J. Appl. Physiol.1980;48:154-162.95. Longhurst, J.C., A.R. Kelly, W.J. Gonyea, J.H. Mitchell. Chronic training withstatic and dynamic exercise: cardiovascular adaptation and response to exercise. CircRes 1981;48(suppl 1)(I):171-8.96. MacNee, W., M. B. A, B. Wiggs, A.S. Belzberg, J.C. Hogg. Regional pulmonarytransit times in humans. J. Appl. Physiol. 1989;66:844-50.97. Mador, M.J., F.A. Acevedo. Effect of respiratory muscle fatigue on subsequentexercise performance. J. Appl. Physiol. 1991;70(5):2059-2065.98. Manohar, M. Blood flow to the respiratory and limb muscles and to abdominalorgans during maximal exertion in ponies. J. Physiol. (Lund) 1986;377:25-35.99. Manohar, M. Costal vs. crural diaphragmatic blood flow during submaximal andnear-maximal exercise in ponies. J. Appl. Physiol. 1988;65(4):1514-1519.100. Manohar, M. Inspiratory and expiratory muscle perfusion in maximally exercisedponies. J. Appl. Physiol. 1990;68(2) :544-548 .101. Manohar, M., T.E. Goetz, L.C. Holste, D. Nganwa. Diaphragmatic 02 and lactateextraction during submaximal and maximal exertion in ponies. J. Appl. Physiol.1988;64(3):1203-1209.102. Manohar, M., T.E. Goetz, D. Nganwa. Costal diaphragmatic 02 and lactateextraction in laryngeal hemiplegic ponies during exercise. J. Appl. Physiol.1988;65(4):1723-1728.103. Maron, B.J. Structural features of the athlete heart as defined by echocardiography.J Am Coll Cardiol 1986;7(1):190-203.104. Maron, M.B., L.H. Hamilton, M.G. Maksud. Alterations in pulmonary functionconsequent to competitive marathon running. Med. Sci. Sports Exerc.1979;11(3):244-249.105. Martin, B., M. Heintzelman, H.I. Chen. Exercise performance after ventilatorywork. J. Appl. Physiol. 1982;52(6):1581-1585.63106. Martin, B., J.V. Wiel, K.E. Sparks, R.E. McCullough, R.F. Grover. Exerciseventilation correlates positively with ventilatory chemoresponsiveness. J. Appl.Physiol. 1978;45(4):557-564.107. Martin, B.A., J.M. Stager. Ventilatory endurance in athletes and non athletes. Med.Sci. Sports Exercise 1981;13(1):21-26.108. McEvoy, R.D., N.J.H. Davies, F.L. Manino, et al. Pulmonary gas exchangeduring high frequency ventilation. J. Appl. Physiol. 1982;52(5):1278-1287.109. McKerrow, C.B., A.B. Otis. Oxygen cost of hyperventilation. J. Appl. Physiol.1956;9:375-379.110. Mead, J. Control of respiratory frequency. J. Appl. Physiol. 1960;15:325-336.111. Melin, J.A.,W.Wijns, A. Robert, M. Nannan, P. De Coster, C. Beckers and J. M.Detry. Validation of radionuclide cardiac output measurements during exercise. J.NucL Med 1985;26(12):1386-1393.112. Milic-Emili, J., J.M. Petit, R. Deroanne. Mechanical work of breathing duringexercise in trained and untrained subjects. J. Appl. Physiol. 1962;17:43-46.113. Moore, K.L. The Developing Human.Philadelphia: W. B. Saunders Company,1974114. Moore, R.L., P.D. Gollnick. Response of ventilatory muscles of the rat toendurance training. Pfliigers Arch. 1982;392:268-271.115. Morganroth, J., B.J. Maron, W.L. Henry, S.E. Epstein. Comparative leftventricular dimentions in trained athletes. Ann. Int. Med. 1975;82:521-524.116. Muche, T.I. Skeletal muscle blood flow in exercising dogs. Med. Sci. Sports Exerc.1988;20(5(suppl.)):s104-s108.117. NHLBI, w. summary. Respiratory muscle fatigue: report of the respiratory musclefatigue workshop group. Am. Rev. Respir. Dis. 1990;142:474-480.118. Olafsson, S., R.E. Hyatt. Ventilatory mechanics and expiratory flow limitationduring exercise in normal subjects. J. Clin. Invest. 1969;45:564-573.119. Otis, A.B. The work of breathing. Physiol. Rev. 1954;34:449-458.120. Otis, A.B., W.O. Fenn, H. Rahn. Mechanics of breathing in man. J. Appl. Physiol.1950;2:592-603.121. Pan, L.G., H.V. Forster, G.E. Bisgard, S.M. Dorsey, M.A. Busch.Cardiodynamic variables and ventilation during treadmill exercise in ponies. J. Appl.Physiol. 1984;57:753-759.122. Pfisterer, M.E., D. R. Ricci, G. Schuler, S. S. Swanson, D. G. Gordon, K. E.Peterson and W. L. Ashburn. Validity of left ventricular ejection fractions measuredat rest and peak exercise by equilibrium radionuclide angiography using shortacquisition times. J. Nuc. Med. 1979;20:484-490.64123. Pfisterer, M.E., B. A, S.M. Swanson, R. Slutsky, V. Froelicher, W.L. Ashburn.Reproducibility of ejection-fraction determinations by equilibrium radionuclideangiography in response to supine bicycle exercise: concise communication. J. NuclMed. 1979;20:491-495.124. Pouliner, L.R., G.J. Dehmer, S.E. Lewis, R.W. Parkey, C.G. Bloomquist, J.T.Willerton. Left ventricular performance in normal subjects:A comparison ofresponses to exercies in the supine and upright positions. Circulation1980;62(3):528-534.125. Powers, S., S. Dodd, J. Woodyard, R. Beadle, G. Church. Alterations inhemoglobin saturation during incremental arm and leg exercise. Br. J. Sports Med.1984;18:212-216.126. Powers, S.K., S. Dodd, J. Lawler, et al. Incidence of exercise induced hypoxemiain elite endurance athletes at sea level. Eur. J. Appl. Physiol. 1988;58:298-302.127. Powers, S.K., J. Lawler, J.A. Dempsey, S. Dodd, G. Landry. Effects ofincomplete pulmonary gas exchange on V02 max. J. Appl. Physiol1989;66(6):2491-2495.128. Powers, S.K., J. Lawler, D. Criswell, et al. Endurance -training-induced cellularadaptations in respiratory muscles. J. Appl. Physiol. 1990;69(5):2114-2118.129. Rasmussen, B.S., B. Hanel, K. Jensen, B. Serup, N.H. Secher. Decrease inpulmonary diffusion capacity after maximal exercise. J. Sports Sci. 1984;4:185-188.130. Rassmussen, B.S., P. Elkjaer, B. Juhl. Impaired pulmonary and cardiac functionafter maximal exercise. J. Sports Sci. 1988;6:219-228.131. Rebuck, A.S., E.J.M. Campbell. A clinical method for assessing the ventilatoryresponse to hypoxia. Am. Rev. Resp. Dis. 1974;109:345-350.132. Rebuck, A.S., A.S. Slutsky. Measurement of ventilatory responses to hypercapniaand hypoxia. In: Hornbein TF, ed. Regulation of Breathing, Part II. New York:Marcel Dekker Inc., 1981: 745-773.133. Reid, M.B., R.L. Johnson. Efficiency, maximal blood flow and aerobic workcapacity of canine diaphragm. J. Appl. Physiol. 1983;54(3):763-772.134. Rerych, S.K., P. M. Scholz, D. C. Sabiston and R. H. Jones. Effects of exercisetraining on left ventricular function in normal subjects: a londitudinal study byradionuclide angoigraphy. Am J. Cardiol 1980;45:244-252.135. Reuschlein, P.L., W.G. Reddan, J.F. Burpee, J.B.L. Gee, J. Rankin. The effect ofphysical on the pulmonary diffusing capacity during submaximal work. J. Appl.Physiol. 1968;24:152-158.136. Robertson, C.H., G.H. Foster, R.L. Johnson. The relationship of respiratoryfailure to the oxygen consumption of, lactate production by, and distribution ofblood flow among respiratory muscles during increased inspiratory resistance. J.Clin. Invest. 1977;59:31-42.65137. Robertson, C.H., M.A. Pagel, J. R. L. Johnson. The distribution of blood flow,oxygen consumption and work output among the respiratory muscles duringunobstructed hyperventilation. J. Clin. Invest. 1977;59:43-50.138. Robertson, H.T., R.L. Coffey, T.A. Standout, W.E. Truog. Respiratory and inertgas exchange during high-frequency ventilation. J. Appl. Physiol. 1982;52:683-689.139. Roca, J., M.C. Hogan, D. Storey, et al. Evidence for tissue diffusion limitation ofVO2 max in normal humans. J. Appl. Physiol. 1989;67(1):291-299.140. Rochester, D.F., G. Bettini. Diaphragmatic blood flow and energy expenditure inthe dog. J. Clin. Invest. 1976;57:661-672.141. Rochester, D.F., A.M. Briscoe. Metabolism of the working diaphragm. Am. Rev.Respir. Dis. 1979;119:101-106.142. Rochester, D.R. Measurement of diaphragmatic blood flow and oxygenconsumption in the dog by the Kety-Schmidt Technique. J. Clin. Invest.1974;53:1216-1225.143. Roncoroni, A.J., H.A. Adrogué, C.W. DeObrutski, M.L. Marchisio, M.R.Herrera. Metabolic acidosis in status asthmaticus. Respiration 1976;33:85-94.144. Roughton, F.J.W. The average time spent in the pulmonary capillary and its relationto the rates of CO uptake and elimination in man. Am. J. Physiol. 1943;143:621-633.145. Roughton, F.J.W., R.E. Forster. Relative importance of diffusion and chemicalreaction rates in determing the rate of exchange of gases in the human lung, withspecial reference to true diffusing capacity of pulmonary membrane and volume ofblood in the lung capillaries. J. Appl. Physiol. 1957;11(290-302).146. Rowell, L.R., H.L. Taylor, Y. Wang and W. S. Carlson. Saturation of arterialblood with oxygen during maximal exercise. J. Appl. Physiol. 1964;19(2):284-286.147. Schaffartzik, W., D.C. Poole, T. Derion, et al. VA/Q distribution during heavyexercise and recovery in humans: implications for pulmonary edema. J. Appl.Physiol. 1992;72(5):1657-1667.148. Schoene, R.B. Control of ventilation in climbers to extreme altitude. J. Appl.Physiol. 1982;53(4):886-896.149. Shephard, R.J. The oxygen cost of breathing during vigourous breathing. Quart. J.Exper. Physiol 1966;51:336-360.150. Shephard, R.J. The maximum sustained voluntary ventilation in exercise. Clin.Sci.. 1967;32:189-197.66151. Slutsky, R., J. Karliner, D. Rocci, et al. Left ventricular volumes by gatedequilibrium radionuclide angiography: a new method. Circulation 1979;60(3):556-564.152. Smith, T.D., P. Richards. A simple kit for the preparation of 99m Tc-labeled redblood cells. J. NucL Med 1975;17(2):126-131.153. Starmer, C.F., D.O. Clarke. Computer computations of cardiac output using thegamma function. J. Appl. Physiol. 1970;28:219-220.154. Staub, N.C., J.M. Bishop, R.E. Forster. Importance of diffusion and chemicalreactions in 02 uptake in the lung. J. Appl. Physiol. 1962;18:673-680.155. Stockley, R.A. The contribution of the reflex hypoxic drive to the hyperpnoea ofexercise. Respirat. Physiol. 1978;35:79-87.156. Suskind, M., R.A. Bruce, M.E. McDowell, F.W. Lovejoy. Normal variations inend-tidal and arterial blood carbon dioxide and oxygen tensions during moderateexercise. J. Appl. Physiol. 1950;3:282-290.157. Suzuki, S., J. Suzuki, T. Okubo. Expiratory muscle fatigue in normal subjects. J.Appl. Physiol. 1991;70(6):2632-2639.158. Szal, S.A., R.B. Schoene. Ventilatory response to rowing and cycling in eliteoarswomen. J. Appl. Physiol. 1989;647(1):264-269.159. Thompson, J.M., J. A. Dempsey, L. W. Chosy, N. T. Shahidi and W. G.Reddan. Oxygen transport and oxyhemoglobin dissociation during prolongedmuscular work. J. Appl. Physiol. 1974;37(5):658-664.160. Thomsen, J.K., N.F. Anderson, K. Bulow, A. Devantier. Blood and plasmavolumes determined by carbon monoxide gas, 9mTc-labelled erythrocytes, 125I-albumin and T 1824 technique. Scan. J . Clin. Lab. Invest. 1991;51:185-190.161. Torre-Bueno, J.R., P.D. Wagner, H. A. Saltzman, G. E. Gale and R. E. Moon.Diffusion limitation in normal humans during exercise at sea level and simulatedaltitude. J. Appl. Physiol. 1985;58:989-995.162. Tsu, M.E., A.L. Babb, E.M. Sugiyama, M.P. Hlastala. Dynamics of soluble gasexchange in the airways: II. Effects of breathing conditions. Resp. Physiol.1991;83:261-276.163. Turner, P.S. The effect of furosemide on exercise induced hypoxemia in highlytrained athletes. [M.P.E.]. University of British Columbia, 1992.164. Vaughn, T.R., E.M. DeMarino, N.C. Staub. Indicator dilution lung water andcapillary blood volume in prolonged heavy exercise in normal men. Am. Rev.Respir. Dis. 1976;133:757-162.165. Viires, N., M. Aubier, A. Rassidakis, C. Roussos. Regional blood flowdistribution in dog during induced hypotension and low cardiac output. J. Clin.Invest. 1983;72:935-947.67166. Wagner, P.D. Diffusion and chemical reaction in pulmonary gas exchange. Physiol.Rev. 1977;57:257-312.167. Wagner, P.D., P.F. Naumann, R.B. Laravuso. Simultaneous measurement of eightforeign gases in blood by gas chromotography. J. Appl. Physiol. 1974;36:600-605.168. Wagner, P.D., H.A. Saltzman, J.B. West. Measurement of continuous distributionsof ventilation-perfusion ratios:theory. J. Appl. Physiol. 1974;36(5):588-599.169. Wagner, P.D., C.M. Smith, N.J.H. Davies, R.D. McEvoy, G.E. Gale. Estimationof ventilation-perfusion inequality by inert gas elimination without arterial sampling.J. Appl. Physiol. 1985;59(2):376-383.170. Ward, S.A., B.J. Whipp, C.S. Poon. Density dependant air flow and ventilatorycontrol during exercise. Resp. Physiol. 1982;49:267-277.171. Warren, G.L., K.J. Cureton, W.F. Middendorf, C.A. Ray, J.A. Warren. Redblood cell pulmonary capillary transit time during exercise in athletes. Med. Sci.Sport Exerc. 1991;23(12):1353-1361.172. Wasserman, K., B.J. Whipp, J.A. Davis. Respiratory physiology of exercise:metabolism, gas exchange and ventilatory control. In: Widdicombe JG, ed.International Review of Physiology, Physiology 111. Baltimore: University ParkPress, 1981:173. Weinstein, W.B., W.M. Smoak. Technical difficulties in 99mTc labeling oferythrocytes. J. NucL Med. 1970;11:41-42.174. West, J.B. Ventilation-perfusion inequality and overall gas exchange in computermodels of the lung. Resp. Physiol. 1969;7:88-110.175. West, J.B. State of the art. Ventilation-perfusion relationships. Am. Rev. Respir.Dis. 1977;116:919-943.176. West, J.B. Respiratory physiology-the essentials. (3rd ed.) Baltimore, MA:Williams & Wilkins, 1985:85-127.177. West, J.B., Si. Boyer, D.J. Graber, et al. Maximal exercise at extreme altitudes onthe summit of Mount Everest. J. App!. Physiol. 1983;55(3):688-698.178. West, J.B., P.H. Hackett, K.H. Maret, et al. Pulmonary gas exchange on thesummit of Mount Everest. J. Appl. Physiol. 1983;J. Appl. Physiol.; 55(3):678-687.179. West, J.B., K. Tsukimoto, 0. Mathieu-Costello, R. Prediletto. Stress failure inpulmonary capillaries. J. Appl. Physiol. 1991;70(4):1731-1742.180. West, J.B., P.D. Wagner. Pulmonary gas exchange. In: West JB, ed.Bioengineering Aspects of the Lung. New York: Marcell Dekker, 1977: 361-457.Lung Biology in Health and Disease;181. West, J.B., P.D. Wagner. Ventilation-perfusion relationships. In: Crystal RG, WestJB, et al., ed. The Lung: Scientific Foundations. New York: Raven Press Ltd.,1991: 1289-1305.68182. Whipp, B.J., and K. Wasserman. Alveolar arterial gas pressure differences duringexercise. J. Appl. Physiol. 1969;27:361-365.183. Whipp, B.J., J.A. Davis. Peripheral chemoreceptors and exercise hyperpnoea.Med. Sci. Sports Exerc. 1979;11(2):204-212.184. Williams, J., S. Powers, M. Stuart. Hemoglobin desaturation in highly trainedathletes during heavy exercise. Med. Sci. Sports Exerc. 1986;18:168-173.185. Young, I.H., J. Woolcock. Changes in arterial blood gas tensions during unsteady-state exercise. J. Appl. Physiol. 1978;44(1):93-96.69APPENDIX AREVIEW OF LITERATUREERAT'UREIntroductionA debate of the factors providing a limit to maximal oxygen consumption (VO2max)has occupied exercise physiologists for decades. An increase in VO2 max is associated withhyperoxia and infusion of red blood cells and a decline in VO2max is observed in anemiaand hypoxia. This information is generally interpreted that oxygen delivery to the workingmuscle provides the major limit to maximal oxygen uptake (39). In humans, increases inVO2max have been considered to be a result of both increases in cardiac output (Q) andoxygen extraction (a-v02).As the first step in the supply of oxygen to the working muscle, the respiratorysystem has not been thought to limit maximal performance. In most individuals, arterialoxygen tension is maintained during heavy exercise (20), and maximal capacity to ventilateduring non-exercising conditions greatly exceeds observed exercise ventilation at maximalexercise (67, 118). Recently, reports of arterial hypoxemia in individuals exercising atmaximal levels (35, 125, 126, 146, 159) and respiratory muscle fatigue in marathon runnersafter prolonged exercise (93) have focused attention on the respiratory system as a possiblelimiting factor to performance in individuals capable of maintaining very high levels ofwork.In our laboratory (72), we have documented hypoxemia with arterial P02 as low as68 torr in athletes exercising for five minutes at VO2max. These subjects exhibited onlypartial respiratory compensation for the metabolic acidosis of exercise; despite a mean pH of7.21, mean pCO2 levels of 36.4 ton, indicate only partial respiratory compensation of theacidosis. This has been observed by other authors in highly trained individuals (35) withless trained subjects showing compensatory hyperventilation to a pCO2 of approximately 30ton (34). In resting individuals, respiratory compensation for a metabolic acidosis couldresult in a pCO2 as low as 15 ton (92). In highly trained athletes who have ample70stimulation (acidosis, hypoxia) to breathe, why are such profound disturbances inhomeostasis tolerated without adequate compensation?Recently some authors (25, 34), have recognized several factors, specificallyenergetics, muscle fatigue, respiratory drives and gas exchange, which may limit maximalexercise performance. The following review will discuss the ways in which the respiratorysystem could limit aerobic performance, with particular emphasis on the respiratory systemduring short term, near maximal, exercise.Respiration and the respiratory muscles:During quiet breathing the primary muscle of inspiration is the diaphragm, which asit contracts, increases thoracic volume. Abdominal viscera are displaced, leading toprotrusion of the anterior abdominal wall. The quadratus lumborum acts in synergy with thediaphragm, opposing the the tendency to elevate the 12th rib. The scalene muscles exhibitactivity even during quiet inspiration, and act to fix or elevate the thoracic inlet. Expiration ispassive, and mainly due to elastic recoil of the lungs.As ventilation increases, more of the respiratory muscle mass becomes active and atmaximal exercise almost all of the muscles of the chest and back may be important. Ininspiration, in addition to the muscles previously described, the sternocleidomastoid,external intercostal, serratus posterior superior and possibly portions of trapezius arerecruited. Expiration becomes active, with strong contractions of the oblique and transversemuscles acting to compress the abdomen, increase infra-abdominal pressure, and displacethe diaphragm upwards. In addition, the internal intercostal muscles and serratus posteriorinferior act to assist expiration. The muscles of the back also have a role in respiration,acting as a counterbalance to the flexion induced by the contraction of the abdominalmuscles.71EnergeticsMany investigators have not considered respiratory muscles to receive a substantialfraction of Q or V02 during exercise. At high levels of ventilation if oxygen consumptionby, and blood flow to, the respiratory muscles were substantial, this could limit maximalexercise by direct competition between the working muscle and respiratory muscles.Blood flow to the diaphragm and other respiratory musclesInvestigation into the blood flow and metabolism of the respiratory muscles hasfocused mainly on the diaphragm, however, as exercise intensity increases other muscles ofrespiration become progressively more important. Data collection in humans is hampered bymethodology; most studies have relied on data obtained from dogs, ponies, sheep and rats.Estimates of blood flow of the various respiratory muscles are summarized in Table12 and have been made during rest, inspiratory loads and exercise (19, 50, 98-100, 137,140, 142). At rest, dog respiratory muscle comprising about 3-4% of body weight,receives about 1.5% of cardiac output (136, 137). During inspiratory resistive work, thisvalue rises to greater than 10% of cardiac output. Blood flow to the diaphragm increases by275% and 500% during mild and moderate exercise, which is roughly double the increaseobserved in gastrocnemius muscle (50). Intercostal muscle blood flow increases two tothree times, similar to values for working skeletal muscle (50, 141). Blood flow to therespiratory muscles under work stress varies with the species investigated and type of stressapplied. As outlined in the table below the resting blood flow of the diaphragm is about 150% of intercostal muscle. During maximal exercise, these values may be higher than 260 and130 ml•min-1 .100g-1 for diaphragm and intercostal muscle respectively. Blood flow valuesfor the diaphragm are similar to maximal values reported for locomotor muscles inexercising dogs (239 ml•min-1 .100g-1 )(116), horses (135-237 ml-min-1 .100g-1) (98) andisolated working quadriceps muscle (240 ml•min -1 .100g-1 ) in humans (2).72StudyTable 12. Blood flow of respiratory musclesQdi (ml•min-1 ' Mg- 1 )^Qi (ml-min-1 ' 100g- 1 )-icESD^/Z±SDRest^Stress^Rest^StressReid and Johnson, 20±5(1983)(133)Rochester and 18±7 52±30Bettini,(1976) (140)Rochester,(1974) 22±6(142)Robertson etal.,(1977a)(136)8±2 207±41 10 59Brancatisano et al., 7.3±0.8§ 9.1±1.4§(19) 8.2±0.4* 6.2±0.4*Robertson etal.,(1977b)(137)9±1 33±2Fixler etal.,(1976)(48)16±3 96±18 15±4 43±18Manohar etal.,(1988a)(99)12-13 151-245 7±2 119±9Manohar,(1986)(98) 11±3 261±23 6±1 131±10Manohar, (1990) - 30 -330 - 18 150-170(100)Viires,(1983)(165) -14 -50 -8-9 -16-21Commentsinspiratory load dogsinspiratory load dogsdogsinspiratory load dogsinspiratory load dogshypercapnia dogsmoderate exercisedogsmaximal exerciseponiesmaximal exerciseponiesmaximal exercisetrained poniescardiac tamponadedogsQdi = diaphragmatic blood flow,^intercostal blood flow, * = internal intercostalmuscle, § = external intercostal muscle.73Manohar et al.,(99) originally used the data obtained in their investigations in poniesto argue that the amount of cardiac output diverted to respiratory muscles during maximalexercise is relatively small, based on the measurement of diaphragm blood flow, andexcluding other muscles of respiration. More recent investigations (100), in exercisingponies suggest that blood flow for the entire respiratory mass is actually much higher andmay be close to 15% of the total cardiac output during maximal exercise. In humans, cardiacoutput has been shown to increase by 50% over resting conditions in subjects breathingthrough an inspiratory resistance of 50-60% of maximum inspiratory pressure (29) and by76% in isocapnic hyperventilation at maximal sustainable ventilation >150 1-min -1 (3). Nodirect measures of respiratory muscle blood flow have been made.Oxygen consumption of respiratory musclesIn animal models, oxygen consumption of the diaphragm has been shown to belinearly related to the work of breathing (136, 137). Like skeletal muscle, demands forincreased oxygen are met by both increased flow and oxygen extraction (79, 101, 136,137). In dogs breathing through inspiratory flow resistance, oxygen extraction becomesmaximal at low work rates; increasing demand for oxygen is met by increasing blood flow(136, 137, 140). In ponies exercising at heavy and maximal levels increasing 02 demandsare met by both increments in perfusion and oxygen extraction (102).Direct measures of respiratory muscle oxygen consumption have been made in dogsin a wide variety of experimental conditions (Table 13) (133, 136, 137, 140, 142). Duringmechanical ventilation, '102 of the diaphragm is less than 1 ml•min-1 .100 g-1 (137, 142)and doubles during quiet breathing (136, 137, 140, 142). During hypercapnichyperventilation, oxygen consumption triples (142), although this level of hyperventilationis less than would be expected during heavy exercise. When resistance to inspiratory flow is74used to stress the respiratory muscles, diaphragmatic VO2 increases to about 15 ml-min-1.100g-1 (7.1-22.6) (133, 136, 140).Table 13. VO2 of respiratory musclesStudy^VO2di^VO2i^Comments(ml•min-1 -100g-1)^(ml-min-1-100g-1)i±sd i±sdManohar,(1986)(98)Manohar etal.,(1988b)(101)Reid and Johnson,(1983)(133)Rochester,(1974)(142)Rochester andBettini,(1976)(140)Robertson et al.,Rest1.60.41.2±0.31.4±0.61.7Stress58.645.112.0±2.82.8±0.97.1±4.322.6Rest0.9Stress29.6 horses max exercisehorses max exerciseinsp. flow resistancedogshypercapnia dogsinsp. flow resistancedogsinsp. flow resistance1977(136)^ dogsVO2 di = oxygen consumption diaphragm; VO2 i = oxygen consumption intercostal muscleData during heavy and maximal exercise have been obtained in the pony (98, 101);V02 of the diaphragm reaches as high as 31 ml-min -1 -100g-1 during heavy and 45-58ml•min-1.100g-1 during maximal exercise. Oxygen extraction by the diaphragm increasesprogressively as work rate increases, but has not been shown to reach a value where itwould limit 02 consumption, although observed blood flow is close to maximal.Direct measures of oxygen consumption of the diaphragm or other muscles ofrespiration have not been made in humans. Instead most investigators have measured theincrease in VO2 which accompany an increase in VE. The estimate in the total cost ofventilation at maximal levels of exercise varies from 3% (112) to as much as 25% (54) of75VO2 max. This variation is in part due to different means of increasing VE: exercise,hyperventilation, CO2 and inspiratory flow resistance, as well as difficulty in obtainingsteady state CO2 and accurate VO2 measures (81) . The results of these investigations aresummarized in Table 14. The relationship of oxygen cost to ventilation is not a linearfunction but rather a series of curves (11, 109, 119) with a rapidly increasing slope as theupper limit of ventilation is approached (109). The oxygen cost of a particular ventilationvaries with respiratory rate (11) and most investigators agree that healthy subjectsspontaneously select the tidal volume and respiratory rate that minimizes respiratory work(110, 120). At any metabolic level the VO2 of respiratory muscles is less in trained thanuntrained subjects, reflecting the lower level of pulmonary ventilation for any given oxygenuptake (112).Table 14. Oxygen cost of unobstructed hyperventilation02 cost calculation Cost of 120 Cost of 170STUDY studied 1•min' 1 VEVE(ml•min-1 )(ml -min- 1)Anholm et al., (1987)(3)50-220 1 376).049[VE" 241 546Bradley and Leith,(1978)(18) 103-2501 -682+8.31 VE 315 813Shephard,(1966)(149)90-1301 4.3VE 516Bartlett et al.,(1958)(11)20-2001 estimated from curve -300= 343-900x=753Lactate ProductionNo evidence of net diaphragmatic lactate production is seen in ponies during exerciseat maximal levels (V02 >120 1 -min-1 ) (101) even when inspiratory work ofbreathing is increased by laryngeal hemiplegia (102). Similar findings are observed in the76dog when animals are subjected to an inspiratory flow resistance (137). In low cardiacoutput states, net diaphragmatic lactate production is observed only when the mixed venousP02 falls below 20 ton (6). Indirect evidence suggests that other respiratory muscles maycontribute to net lactate production. During cardiac tamponade, spontaneously breathingdogs have been shown to have blood lactate levels roughly double that of mechanicallyventilated dogs (-7 vs 3 mmo1•1-1)(165). In humans, Roncoroni et al.,(143) documentedmetabolic acidosis in almost 40 % of patients in status asthmaticus, which could not beattributed to poor gas exchange in these individuals leading to the conclusion thatrespiratory muscles in the presence of severe airway obstruction may increase blood lactatelevels via anaerobic glycolysis. In healthy humans small (-1 mmo1•1 -1) increases in bloodlactate have been recorded in isocapnic hyperventilation at approximately 70% of maximumbreathing capacity (52).Theoretical calculations of respiratory muscle V02 and Q in humans exercising at maximallevelsDifficulty in making direct measures of respiratory muscle metabolism in humansubjects has restricted research in this area. It is also difficult to compare the work indifferent animal species because of different methods of inducing stress on the respiratorymuscles (hypercapnia, inspiratory loading, maximal exercise, cardiac tamponade) as well asinterspecies differences. Nonetheless a certain pattern arises: During resting conditions,oxygen consumption of the diaphragm is remarkably constant across species , and isgenerally less than 2 ml•min-1 -100g-1. During maximal exercise this value increases toabout 45-60 ml•min-1 .100g-1 in horses (98, 101). Data are not available for other animalspecies. The values for intercostal muscle are roughly half of these values (<1 and -30ml•min-1 .100g-1 (98)) both at rest and during exercise. Diaphragmatic blood flow at rest isalso similar between species (-10-30 ml•min-1 -100g-1) and has been reported higher than300 ml•min-1.100g-1 in ponies (99). Again the values for intercostal muscle blood flow77during resting and exercising conditions are about half (-10 and as high as 130ml•min-1 .100g-1) of the values reported for the diaphragm.This information can be used to calculate theoretical values of blood flow andoxygen consumption for human respiratory muscle during maximal exercise, provided thefollowing assumptions are made:1.The total respiratory muscle mass in humans is - 4 kg.2. The diaphragm is roughly 15 % of the total respiratory muscle mass or 600g.3. The other 3400g, which includes intercostal muscle, is relatively homogeneous,allowing calculations to be made for the entire respiratory mass.4. Meaningful comparisons can be made between animal species and humans.The pony is the mostly completely studied species and the only one in whichsystematic observations have been made at heavy and maximal exercise. Therefore datafrom these investigations are used in the following calculations.Qdi^= 260 ml•min-1 .100g-1 x 600 g^=^1560 ml•min-1Qi^= 120 ml•min-1 .100g-1 x 3400 g^4080 ml•min-15640 ml-min-1V02 di = 59 ml•min-1 .100g-1 x 600 g^354 ml•min-1V02 i = 30 ml•min-1 .100g-1 x 3400 g^1020 ml•min-11374 ml•min-1Assuming a VE of 150 1-min-1 , V02 of 4.6 1-min-1 (72) and Q of 27 1-min-1 (5)then these figures represent 30% of VO2 and 20% of Q for exercising humans at maximalwork loads. Clearly these numbers are substantial and support the idea that respiratorymuscles may be in direct competition with skeletal muscle for oxygen and cardiac output.The resistance of the diaphragm to lactate production has also been clearlydemonstrated in a number of species and under a wide variety of experimental conditions.78However the diaphragm constitutes less than 20% of the total respiratory muscle mass andno data is available for the other muscles of respiration under exercising conditions. There isno reason to expect that intercostal, abdominal and other muscles of respiration shouldbehave differently than other exercising skeletal muscle. Any increase in ventilation, ratherthan increasing pH and lessening the impact of metabolic acidosis, may actually worsen itthrough increasing respiratory work. This line of thought, while highly speculative, is notnew. Otis (119) argued that a critical ventilation could be reached above which any increasein VO2 would go entirely to respiratory muscles and estimated this ventilation at 1401•min-1. It is logical to consider that exercise ventilation represents an optimum level ofventilation balancing respiratory muscle oxygen consumption and blood flow against thedemands of skeletal muscle and pH homeostasis. The level of ventilation achieved is acompromise between supply and the cost of supplying ventilation.Respiratory muscle fatigueRespiratory muscle fatigue is commonly seen in chronic obstructive lung disease, orin acute medical situations such as adult respiratory distress syndrome, where pulmonaryedema increases respiratory work. In the normal exercising human, respiratory muscleshave been considered to be fatigue resistant under most conditions, however reports ofdecrements in respiratory muscle strength after prolonged exercise (93) and decrease inperformance following respiratory work (97, 105), indicate that this is not the case.h f m o m 11^1"The histochemical and biochemical characteristics of respiratory muscles have beeninvestigated in both animals and humans. The diaphragm of most mammals are composedof varying percentages of the three skeletal muscle fiber types. Biopsy specimens takenduring thoracotomy in humans show the composition of the diaphragm to be 55% SG, 21%FOG and 24%FG fibers (90). The oxidative capacity of diaphragm is greater than othermixed skeletal muscle, and closely resembles an intermediate between skeletal and cardiac79muscle (62, 114). The oxidative capacity can be trained by imposing inspiratory resistiveloads (1, 82) and in induced emphysema (48). The effect of endurance exercise on therespiratory muscles is controversial, with some authors documenting an increase in therespiratory capacity of the rat diaphragm and intercostal muscles with endurance training(114, 128) and no change reported by others (53) using an almost identical protocol.After endurance exercise, depletion of muscle glycogen stores has been found inboth diaphragm and intercostal muscle (62, 75, 114) as well as triglyceride depletion in thediaphragm (62).This evidence points to the diaphragm as a muscle with high aerobiccapacity, relying on aerobic glycolysis and fatty acid oxidation to offset the metabolic cost ofwork. Other muscles of respiration have not been as extensively evaluated, but wouldappear to have characteristics in common with other skeletal muscle.Evidence for fatigueMuscular fatigue has been defined as a reversible reduction in force generatingcapacity which is relieved by rest (117). Task failure is defined as the inability to maintain orcontinue the force required to perform a particular task. For the respiratory system this canbe defined as the failure of the respiratory muscles to generate a given pleural pressure. Theunique characteristics of diaphragm muscle with ability to maintain very high oxidativecapacity, renders this muscle relatively resistant to fatigue compared with skeletal muscle(172) however, as a correlate to the histochemical data, several studies have shown thathigh levels of ventilation cannot be maintained indefinitely (14, 24, 107). A decline in thestrength of the ventilatory muscles at the end of a marathon race, with a fall in maximuminspiratory and expiratory mouth pressures and transdiaphragmatic pressures suggests thatthese considerations may be of practical concern (93). Reduced time to exhaustion has beenobserved during short-term maximal exercise after 150 minutes of maximal ventilation (105)and after inspiratory threshold loading (97). Expiratory muscle fatigue has also beenreported as a result of increased expiratory work in normal subjects, persisting for up to an80hour after cessation of expiratory muscle loading (157). Expiratory loading also inducesinspiratory muscle fatigue and these findings may have important implications for the athletewith exercise induced asthma.Taken together the histochemical and whole body information point to therespiratory system as a system that is fatigue resistant under healthy, non-exercisingconditions but that will demonstrate substrate depletion and a decline in performance withlevels of exercise that are not uncommon in today's active society.Respiratory drivesAfter the preceding discussion, the advantages of a blunted ventilatory response toexercise to an athlete appear obvious. If at any given level of exercise the ventilation isreduced, the individual will be less likely to develop respiratory muscle fatigue, will requirea smaller fraction of total cardiac output and V02 diverted to respiratory muscles and willexperience less dyspnoea.Although controversial, most authors would agree that the peripheralchemoreceptors contribute to exercise ventilation to a substantial degree. The estimatedcontribution of the hypoxic ventilatory response to exercise ventilation ranges from 16 to30% of the total VE (106, 155, 183). Resting hypoxic and hypercapnic drives are positivelyrelated (131) and also correlate with drives measured during moderate exercise (106) andwith exercise ventilation (106, 132). During maximal levels of exercise the relationshipbetween respiratory drives and exercise ventilation is less clear and it is likely that, duringexercise near maximal levels, other stimuli to ventilate over-ride any contribution by hypoxicdrives (35, 72). The relationship between hypercapnic drives and ventilation duringmaximal or very heavy exercise has not been determined.A link has been postulated between outstanding endurance performance andchemoreception (106) as some studies have shown endurance athletes to have bluntedresponses to hypoxia and hypercapnia, compared to other athletes and sedentary controls81(26, 106, 148) . It is tempting to speculate that these blunted respiratory drives may berelated to exercise induced hypoxemia. Dempsey et al., (35) felt that hypoventilation was amajor contributing factor to the hypoxemia seen in the athletes studied. The athlete with ablunted respiratory response to hypoxia, would ventilate less in response to the hypoxicstimulus with substantial savings in the cost of maintaining a high level of ventilation at theexpense of less than optimal arterial oxygenation and presumably 02 delivery. A morerecent study (72), failed to demonstrate a relationship between hypoxemia in maximalexercise and resting hypoxic drives and implicated widening alveolar-arterial differences asbeing the most important contributor to the hypoxemia seen in their subject population.Further investigation is needed to determine the relationship between exercising hypoxic andhypercapnic responses and hypoxemia in exercise, however it would appear that the role ofrespiratory drives in determining exercise ventilation is a minor one during maximalexercise, although at less intense levels of exertion, respiratory drives modify the respiratoryresponse to submaximal exercise.82Pulmonary mechanics and expiratory flow limitationPressure-volume relationshipsIn order for air to flow through a tube a difference in pressure must exist betweenthe two ends. In the lung, this pressure difference is generated at rest by the contraction ofthe diaphragm in inspiration, and by passive elastic recoil during expiration. During quietbreathing, infra-alveolar pressure becomes slightly negative with respect to atmosphericpressure during inspiration and slightly positive during expiration (± 1 torr). Duringexercise inspiration is aided by the external intercostal muscles and other accessory musclesof inspiration. Expiration becomes active, with contraction of the muscles of the abdominalwall and internal intercostal muscles contributing to driving pressure.Characteristics of flow inside tubes At low flow rates air flow in a tube is in smooth streams parallel to the the walls ofthe tube. This phenomenon is termed laminar flow and the flow rate is related to the lengthand diameter of the tube, the driving pressure and the viscosity of the flowing material. Asflow rates increase, turbulent flow becomes more likely. The viscosity of the materialbecomes relatively unimportant, however driving pressure must increase to maintain flow asgas density increases. The complexity of the branching system, and irregularity of thesurface in the lung makes laminar flow unlikely in much of the bronchial tree. In most partsof the lung, transitional flow predominates with eddy formation at the branch points of thebronchial tree.Airways resistance to flow can be modified by disease states and chemical agents.The effect of exercise is modest bronchodilation secondary to sympathetic nervous systemstimulation. Intuitively, it seems logical that much of the resistance to flow should be locatedin the very small airways, however direct measures have shown that the majority of the drop83••• pressure takes place over the first 10 generations of bronchial branching (176). Theexplanation for this is likely the large number of small airways relative to larger airways.Flow-volume relationshipsMaximal expiratory flow-volume (MEFV) curves have been used for many years asa clinical assessment of pulmonary function. At rest normal tidal breathing falls well withinthe limits defined by the MEFV curve, however in certain circumstances these limits may beapproached or even exceeded (Figure 9). During progressive exercise the demands forincreased ventilation are met by an increase in tidal volume, to about 50% of forced vitalcapacity (30, 51) and breathing frequency.Eiguis. 9. Maximal expiratory flow volume loop with exercise and maximum voluntaryventilation manoeuvre15  ^15MVV10^ 105 5,•••■•.Cep^0 ^05 5.210^ 1015  ^150 1 2 3 4 5 6 0 1 2 3 4 5 6Volume (1)A(adapted from Jensen et al., 1980 (78))Solid line = maximal expiratory flow volume loop. Dotted line = tidal volume duringmaximal voluntary ventilation (A) and exercise (B)The ventilation at maximal exercise is approximately 75% of that during maximalvoluntary ventilation manoeuvres (MVV) (67) and this information has been interpreted asindicating that mechanical factors do not limit ventilation during exercise. During very heavyexercise, there is some evidence to suggest that this may be inaccurate. Peak expiratory84flows have been measured, that approach or even exceed maximal expiratory flow measuredduring MEFV loops (67, 78, 118). Several explanations are possible: Firstly, during MVVmeasurements subjects breathe at higher lung volumes and therefore have higher expiratoryflow (118); Secondly, during exercise, maximum flow may be increased by increasedelastic recoil pressure; Thirdly, during exercise there may be decreased resistance to flow asa result of bronchodilation secondary to increased sympathetic or decreased parasympathetictone (78). Finally, an increase in ventilation, alveolar P02 and arterial oxygenation is seenwith the administration of He:02 mixtures during heavy exercise, and is associated with adecrease in PaCO2 (35, 36, 170). In interpreting the information gained from administrationof low density gases caution must be exercised to avoid attributing all of the improvement tothe effects on the MEFV curve as there will also be a reduction in the work of breathing forany given level of ventilation, with attendant interaction with respiratory muscle blood flowand metabolism.A final factor which should be considered is entrainment, the alteration ofbreathing frequency to be in step with exercise rhythm. One to one entrainment in gallopinghorses is well known (34, 121), and has been proposed as a factor causing hypoxemia inthese animals (34). During exercise in humans, entrainment is much less complete anddepends on the form of exercise. During walking about half of normal subjects are partiallyor completely entrained, increasing to about 80% during running. Entrainment duringcycling is found in about 30% of subjects (13) and is apparent in oarswomen, at low workintensities (158). From the preceding arguments it is clear the respiratory mechanics havethe potential to limit exercise ventilation, and by inference maximal exercise throughentrainment, expiratory flow limitation or by interaction with the work of breathing.85Pulmonary diffusion and gas exchangeIn the preceding pages some of the factors that may contribute to a respiratoryconstraint to exercise have been discussed. Diffusion and gas exchange may interact withthese factors on several levels as well as directly contributing to limitation of performance.For example, if through hypo-ventilation or mechanical limitation of flow, alveolar P02 isreduced, then the driving pressure for diffusion is also reduced and may lead to a decline inarterial oxygen content and subsequent decrease in oxygen delivery. It is not the purpose ofthis review to completely discuss the physics of respiratory gas exchange as it has been thesubject of several excellent reviews. The reader is referred to the work of Peter Wagner(166) for a more in-depth discussion of the points discussed below.As air travels from the atmosphere to the alveolus it becomes fully saturated withwater vapor and in the alveolus becomes mixed with alveolar air containing lowerconcentration of oxygen and a higher concentration of carbon dioxide. It is usually assumedthat alveolar gas is fully saturated with water which gives a partial pressure of 47 torn at abarometric pressure of 760 torr and 37°C. Under resting conditions the gas tensions ofalveolar air are approximately P02 = 105 toff, PCO2 = 40 toff, PH2O = 47 ton and PN2= 568 ton.Once a gas has arrived at the alveolus it must pass through alveolar fluid, alveolarepithelium, epithelial basement membrane, interstitial space, capillary basement membraneand through capillary endothelium into the blood.Factors related to diffusion of oxygenThe rate of combination of oxygen with hemoglobin is not instantaneous but ratheroccurs as a finite rate. The uptake of oxygen can be considered in two steps (145): 1.Diffusion across the alveolar membrane and plasma to the red blood cell and 2. Combinationwith hemoglobin. The rate of combination of hemoglobin with oxygen (given by 0) is a86function of hemoglobin saturation, being greatest at low saturation and approaching zero atsaturations above 80% (154). The overall transfer from the alveolus to the red cell anddiffusing capacity of the lung (DL) can be related by:1 _ 1 +  1 DL DM OV cWhere Vc is capillary blood volume, and DM is the diffusing capacity of the alveolarmembrane. DM is close in value to Vc, therefore for 0 greater than 1, the major resistance touptake would be that determined by combination of hemoglobin with oxygen. Most of thechanges in P02 and 02 content take place in the first 0.2 second of the red blood cell transitin the pulmonary capillary when saturation is low and 8 is high, and rate is predominantlydetermined by DM. Later as saturation rises, 0 decreases and the chemical reactionresistance predominates. The amount of oxygen dissolved in plasma is linearly related topartial pressure and is about 0.3 m1•100 m1 -1 , less than 2% of that bound to hemoglobinand it is generally neglected in order to simplify calculations. The rate of rise in P02 isaffected by simultaneous CO2 exchange: as capillary PCO2 falls the oxygen hemoglobinequilibrium curve is shifted towards the left and more oxygen is taken up by hemoglobin,delaying the rise in P02.Factors related to diffusion of carbon dioxideAlthough the diffusion across alveolar membrane is approximately 20 times greaterfor carbon dioxide than for oxygen, the diffusion rate for CO2 may be similar, or evenslower, than for 02. The reason for this becomes apparent when the chemical reactionswithin the blood necessary for CO2 exchange are examined. Transfer of CO2 in the blood isin three forms: dissolved CO2, HCO3-, and bound to protein (carbamino compounds),primarily hemoglobin. In arterial blood, HCO3 - is the predominant form of CO2 transport87(90%), with 5% bound to proteins and 5% as dissolved CO2. Plasma CO2 accounts forabout 2/3 of the total CO2, with the remainder present intracellularly, bound to hemoglobin.As CO2 diffuses into the alveolar gas, it is replaced by plasma carbamino-0O2 and by theconversion of HCO3 - and H+ to CO2 and water. In the plasma, this reaction is notcatalyzed by carbonic anhydrase, as it is within the red blood cell and therefore proceedsmuch (-104) more slowly. As HCO3 - from the plasma is consumed, it is replaced by redcell HCO3- which is exchanged with chloride ion (chloride shift). CO2 also diffuses out ofthe red blood cell where is has been produced by the much more rapid carbonic anhydrasecatalyzed conversion of HCO3 -. Deoxy-hemoglobin acts as a buffer for H+, as oxygencombines with hemoglobin H+ is released, which facilitates the conversion of bicarbonate tocarbon dioxide and water (the Haldane effect). Although for CO2 DM is 20 times that for02, the slowness of chemical reactions, particularly the chloride shift, with a 0/ 2 of 0.15seconds renders the time for CO2 diffusion equilibrium, if anything, slower than that for02. Due to the absence of carbonic anhydrase in the plasma, the plasma concentrations ofCO2, H+, and HCO3 - are not in equilibrium at the end of the pulmonary transit. This leadsto a post-capillary increase in pH and CO2 and a fall in HCO3 -.Factors affecting pulmonary gas exchange In the intact organism gas exchange can be modified by a number of factors.1. Shunt: In the normal human, small amounts of blood by-pass the pulmonary capillariesand therefore do not participate in gas exchange. This includes blood from the bronchialarteries and coronary veins that drains directly into the left ventricle via the thesbian veins.Normally shunt accounts for less than 2% of cardiac output.2. Hypoventilation: The level of alveolar 02 and CO2 is determined by a balance betweensupply and removal. Hypoventilation from any cause will impair pulmonary gas exchange88by decreasing the driving pressure across the alveolar membrane for both carbon dioxide(decreased removal) and oxygen (decreased supply).3.Ventilation-perfusion (VA/Q) mis-match: If ventilation is abolished in a particular area ofthe lung then the blood passing through this portion of the lung is essentially shuntedthrough the lung; no gas exchange takes place, and no change in gas content is observed. Ifon the other hand perfusion tends towards zero, the gas content of that blood will resemblethat of alveolar gas. Ventilation and perfusion are not uniform throughout the lung, theeffects of gravity render the bases better perfused than the apices of the lungs and the apicesrelatively better ventilated than the bases. (VA/Q) mis-match will tend to depress Pa02 asmore blood will be poorly oxygenated.4. Reduction of alveolar surface area: Through disease states, such as chronic obstructivepulmonary disease, alveolar surface area can be reduced thus reducing the available surfacearea for gas exchange. When advanced, this will be manifest in decreased diffusingcapacity for CO and impaired gas exchange during exercise.5. Diffusing capacity: The diffusing capacity of the lung can diminished by pulmonarydisease. In order to produce abnormalities of gas exchange it must be reduced to 20-25% ofnormal.6. Altitude: The effect of altitude is to decrease total barometric pressure and therefore thepartial pressure of inspired oxygen. Some compensation is made by increasing alveolarventilation (and decreasing alveolar CO2) however the overall effect is to decrease PAO2,and the driving pressure for oxygen across the alveolar membrane.7. Transit time: Under normal conditions, partial pressure equilibrium is reached after about0.25 seconds of gas exchange. The average transit time of red cells in the pulmonarycapillaries is obtained by the ratio of capillary blood volume to cardiac output (Vc/Q) orabout 0.75 seconds for a person at rest with a Vc of 75 ml and Q of 6000 ml•min -1 (80).During exercise, Vc may increase by a factor of 2 while cardiac output may increase to 301•min-1 giving an average transit time of 0.30 seconds. Cardiac output in excess of 40891•min-1 has been reported in some elite athletes (43) which would reduce average transittime to 0.23 seconds using these calculations. It should be pointed out that these numbersrepresent average transit times, and some red cells will travel faster than these values.Capillary flow is not uniform but rather is pulsatile, which may reduce transit time further.8. Exercise: The effects of exercise on diffusion of oxygen across the alveolar membranecan be seen in Table 15 . At rest, the normal alveolar-arterial difference is about 7 to 10 ton(60, 161, 182). During light exercise, gas exchange may improve (182), as a result ofimproved ventilation-perfusion relationships. As exercise intensity increases, gas exchangedeteriorates and reported mean [A-a]D02 ranges from 11 to 41 ton at VO2 of 2.7 1-min-1and above.90Table 15. Alveolar-arterial differences at rest and during exerciseStudy V02 Q PAO2 Pa02 [A-a]DO21-min-1 1•min-1 ton- torr torrGledhill et al., 1978 0.29±.03 5.9±1.6 97.8±2.2 86.9±1.6 10.1±2.5(60)Whipp and 0.32±.07 105.0±1.2 97.0±1.9 7.4±1.9Wasserman, 1969(182)Torre-Bueno et al., 0.35±.04 6.6±1.3 111* 102.7±10.8 8.6±5.41985 (161)Whipp and 1.65±.45 102.0±4.6 98.0±5.5 3.8±2.9Wasserman,1969(182)Gledhill et al., 1978 1.84±.07 15.5±3.1 102.9±2.2 87.5±2.7 15.5±1.6(60)Torre-Bueno et al., 2.71±.53 21.7±3.4 113.0* 89.9±3.8 23.1±7.51985 (161)Whipp and 3.31±.58 108.0±6.4 97.0±3.73 10.8±3.6Wasserman, 1969(182)Hammond et aL, 3.97±.29 24.9±3.2 113.7* 90.7±8.2 23.0±8.01986 (63)Hopkins and 4.54±.45 119.0±1.5 78.0±8.6 41.0±7.7McKenzie,1989 (72)*= Standard deviation not reportedPulmonary gas exchange has been studied using multiple inert gas techniques, at restand during exercise at sea level and simulated altitude to 15,000 feet, allowing thecontributing factors to [A-0)02 to be dissected. At sea level, ventilation-perfusionrelationships worsen with exercise, with the major factor contributing to [A-4)02 being91VA/(5 mis-match at V02 up to 3.01•min-1 . At rest, the contribution of post-pulmonaryshunt to [A-4D02 is undetectably small and does not increase with exercise, simulatedaltitude or exercise at simulated altitude (161). Of particular interest is the effect of exerciseon the diffusion component of [A-4D02. Gale et al. (55), at exercise corresponding to aV02 greater that 3.01•min-1, found a trend toward greater [A-4D02 than could bepredicted from inert gas data, and suggested diffusion limitation was the likely cause.Similar results were obtained at V02 = 2-3 1-min-1 at simulated altitude of 5000 ft.Statistical conclusions were hampered by the small "n" of the study.In an attempt to clarify this issue, Hammond et al., (63) studied gas exchange inmen at various exercise intensities up to VO2 - 4.01•min -1 (essentially maximum). Duringthe very heavy exercise levels, [A-41)02 was measured at 23.0±8.0 torr of which 10.7±7.8torr could be predicted by inert gas data. Administration of 100% 02 did not alter the "/A/4).relationships. They observed that [A-4D02 increased linearly with VO2. Analysis of dataprovided by Torre-Bueno et al., (161) yields a correlation between cardiac output and [A-MO2 of r = 0.68. As cardiac output increases pulmonary transit time will decrease fromresting levels, therefore at the higher levels of exercise it is possible that some individualswill exhibit increased [A-4D02 because of shortened pulmonary transit time and diffusiondis-equilibrium. No inert gas studies have been made on highly trained athletes capable ofhigh cardiac output, and the possibility for shortened pulmonary transit and significantarterial hypoxemia.Does the pulmonary system constrain exercise? In the preceding pages, the evidence that the pulmonary system may constrainmaximal exercise performance has been explored in five areas: energetics, fatigue,respiratory drives, mechanics and diffusion. These arguments will now be applied to dataobtained from several studies (Table 16) in an attempt to integrate this information.92Group one subjects are sedentary individuals who are unlikely to have a respiratorylimitation to performance. This is borne out by the blood gas data indicating that the arterialP02 and saturation are maintained and the metabolic acidosis of exercise is relatively wellcompensated (pH = 7.31). At a ventilation of 80 1-min -1 the oxygen cost of breathing isabout 12% of the total VO2 (from mean of Table 14) and flow volume loops are unlikely tobe mechanically constraining, confirmed by maintenance of alveolar P02. Bluntedrespiratory drives are unlikely in sedentary subjects such as these and have not contributedto significant hypoventilation as indicated by pH and alveolar P02 and CO2 status. Theobserved alveolar-arterial difference is only slightly greater than that predicted from multipleinert gases therefore they are not likely to be diffusion limited (161). The predicted meanpulmonary transit time based on a capillary blood volume of 150 ml (80) is 0.41 seconds,well within that which will allow full 02 equilibration.Group Two subjects, again show minimal evidence for pulmonary constraint;Alveolar P02 is maintained as is arterial P02 (91 torr) and Sa02 (94.5 %). Hypoventilationsecondary to fatigue, blunted drives or mechanics are unlikely to be a factor as evidenced bymaintenance of alveolar P02, although the pH values indicate less compensation for themetabolic acidosis of exercise. In these subjects measured alveolar-arterial differences areapproximately 12 ton greater than predicted for inert gas data (63) suggesting somediffusion limitation. Calculated mean pulmonary transit time is 0.36 seconds, closer to the0.25 seconds required for full equilibration.Some interesting comparisons can be made between groups three, four and five whoare all highly trained individuals exercising at a VO2 greater that 4 1.min-1 , with similarlevels of ventilation and predicted cardiac output, but quite different alveolar and blood gasdata. Group three is able to maintain arterial saturation above 92 % and P02 above 80 bymaintaining a very high alveolar P02 in the face of widening alveolar-arterial differences toabout 37 ton. VCO2 is very high in these subjects as evidenced by R values greater than1.15 and hypocapnia to 37 torr despite this level of ventilation. Dempsey et al., (35) have93suggested that to maintain a PaCO2 of 30 torr in subjects with a VCO2 of 5 - 6 1•min-1 therequired ventilation is — 240 1-min -1, a level unlikely to be sustained even by highly trainedindividuals. Group four subjects, despite very similar ventilatory data, are unable tomaintain arterial P02 and Sa02 and have extremely high alveolar-arterial differences. Inertgas data has not been obtained in these individuals but it is tempting to speculate that thetrend in [A-aJDO2 due to diffusion observed at VO2 = 3.0 1-min -1 and statisticallysignificant difference at 4.0 1-min -1 , may continue to increase at higher intensity exercise. Ifa cardiac output of 31 1•min-1 (43) can be assumed for these subjects, predicted meanpulmonary transit time would be less than 0.30 seconds and almost half of the RBC transittimes would be less that that required for full oxygen equilibration.Why group three and four subjects, who are identical in so many respects, shouldbehave so differently with respect to Pa02 and [A-a]D02 is unclear but perhaps someanswers can be found in inspection of the Fick equation. Whole body VO2 is the product ofblood flow (in this case cardiac output) and the arterio-venous difference, therefore twostrategies are available to increase V02: oxygen delivery can be increased, (increase in Q)or a-v difference can be increased (increased peripheral extraction). It is possible that groupthree subjects despite similar V02, may have a lower cardiac output and maintain VO2 bysuperior peripheral extraction. The net result would be a longer pulmonary transit time andlower [A-aJDO2 due to diffusion.Group five subjects are different again, and clearly have evidence of hypoventilationas evidenced by the inability to maintain alveolar P02 greater than 108 tom, as well as awidened alveolar-arterial difference indicative of possible diffusion limitation. pH is higherand PaCO2 lower, most likely due to lower intensity exercise - 90% vs 100% of VO2 max.It is likely that in these subjects, mechanical factors may also play a role in the genesis ofhypoxemia. It is also possible that at this submaximal exercise level that there may be a rolefor respiratory drives, which is not evident during exercise at 100% of VO2max.94T b e^B1 BOO •^• „.^a^•^-sibirtts I. d • t^0-100'0 ofGroup 1 2 3 4 5Training status Untrained Moderately Highly Highly HighlyTrained Trained Trained TrainedV02 (1•min-1) 2.71 3.97 4.45 4.65 4.81VCO2 (1•min-1 ) 2.75 4.19 5.27 5.42 4.62R 1.05 1.04 1.18 1.17 0.960 (1•min-1 ) 27.1 24.9 ? ? ?PAO2 (ton) 113 114 119 119 108Pa02 (torr) 90 91 82 71 75[A-4D02(0)(torr) 23 23 37 48 33[A-4D02(0(torr) 22 11 ? ? ?Sa02 (%) 95.6§ 94.5§ 93.0 89.9 91.9PaCO2 (ton) 37 35 37 37 33pH 7.31 7.24 7.21 7.20 7.29n 9 8 7 5 16§ = Sa02 calculated from normogram; 1 = data from Gale et al., 1985 (55) and Torre-Bueno et al., 1985 (161); 2 = Hammond et al., 1986 (63); 3,4 = Hopkins and McKenzie,1989 (72) and unpublished data, 5 = Dempsey et al., 1984 (35), respiratory exchange ratio;0= cardiac output (1•min-1); VE = minute ventilation (1•min-1 ); PAO2 = alveolar 02 (ton);Pa02 = arterial 02 (ton); [A-4D02(0) = observed alveolar-arterial difference (ton); [A-41)02(0 = predicted alveolar-arterial difference (ton); Sa02 = arterial hemoglobinsaturation (%); PaCO2 = arterial CO2 (ton).95SummaryThis review has identified several factors by which the respiratory system mayconstrain exercise performance. In sedentary or moderately trained individuals, the datadoes not favor evidence of pulmonary limitation and it seem likely that these individuals areconstrained by other factors, such as 02 delivery and extraction. In highly trainedindividuals exercising at high intensity, the picture is different, with falling Pa02, widening[A-4)02 and inability to maintain acid-base homeostasis suggesting a pulmonaryconstraint. The relative contribution of various factors is unclear, but the balance of theevidence favours mechanical limitation of ventilation and diffusion limitation at the lung asthe most attractive possibilities.96APPENDIX B:METHODOLOGICAL BACKGROUNDQuantitative radiocardiographyRed blood cell labelingLabeling red blood cells with 99mTechnecium provides a stable intravascular markerthat has wide spread clinical applications, particularly in blood flow measurements. Simpleincubation of the erythrocytes with 99mTc does not provide a satisfactory result because ofloss of radioactivity with washing of the cells (173). The addition of stannous citrate actingas a reducing agent, allows the 99mTc-pertechnetate to cross the red cell membrane,improving the stability of the label, while still providing labeling efficiency of over 90%(152). Commercial kit preparations (152) have the advantage of allowing the preparation ofsmall quantities of labelled cells with high yields with a minimal amount of handling thusreducing the risk of contamination. The kit consists of a sterile reagent tube containing 100units of sodium heparin, 2.6 1.4 stannous citrate (1.0 .tg stannous ion), sodium citrate andanhydrous dextrose. About 6 ml of whole blood is added to the vial followed, after mixingwith 4 ml of sterile saline. After centrifuging, 2 ml of erythrocytes are withdrawn and addedto a sterile vial containing 99mTc-pertechnetate and incubated for five minutes. The resultingmixture is assayed for radioactivity prior to injection for RBC label studies.First pass determination of cardiac outputAfter injection into a flowing stream eventually all of a tracer, although diluted, willeventually pass by an observational point down-stream and the amount of indicator (givenby I) can be calculated by00I= FoiC i(t)dt97where F is flow and C i(t) is the observed concentration at time t. In tracers confined to thevascular space the equilibrium concentration (Ceq) times volume of distribution (Vd) willequal the total amount of tracer injected andCOCeq•Vd= F SC I(Odt.0This method can be applied to all tracers, but for radioisotopes where concentrationscannot be measured directly by an external counter it is necessary to relate concentration tocount-rate by allowing for the counting efficiency of the detector for the tracer material inthe volume of interest. After summation the following equation can be derived:E•Vd dF_Awhere E is the observed equilibrium count rate, Vd is the volume of distribution and A is thearea under the first pass time activity curve after correction for re-circulation.To determine cardiac output this analysis is applied to a first pass time activity curvederived from gamma camera imaging of a region of interest in the left ventricle, following abolus injection of 99mTc-pertechnetate label red cells. The equilibrium count rate ismeasured in the same region of interest after complete mixing of the tracer has occurred.Sources of error with this method arise from poor bolus technique at the time of injection,difficulties in extrapolation of the down slope of the first pass curve, dead time count lossesby the detector system, and incomplete mixing of the tracer. Volume of dilution of the tracermust also be determined and can also be measured using radioisotopes (76). Determinationof cardiac output by first pass radiocardiography has shown excellent correlation with othermethods (83, 84) and has the advantages of being non-invasive, able to evaluate separatelythe right and left ventricle and provide information about pulmonary transit times.98Gated radionucleide angiography: Gated radionucleide angiography is based on the principle that if cardiac volumescan be determined in end-systole and end-diastole, and if heart rate is known, cardiac outputcan be determined from the product of stroke volume and heart rate. After injection andequilibration of labeled erythrocytes, data acquisition is performed by dividing the R-Rinterval into a discrete number of frames (usually 16), and using the QRS complex as a gate.A series of images are acquired that provides a dynamic picture of cardiac function. Bymeasuring the number of counts in the left ventricle over the sixteen frames end-systolic andend-diastolic counts can be determined and left ventricular ejection fraction can then becalculated from the difference between background subtracted end-diastolic and end-systoliccounts and expressed as a percentage of end-diastolic counts.Samples of blood are drawn at the end of each data acquisition and imaged in petridishes for five minutes. The average background subtracted count rate for 5 ml of blood isobtained, and the left ventricular count rate is corrected for loss by decay of 99mTc usingstandard tables. Left ventricular end-diastolic volume is obtained by dividing the leftventricular count rate by the 5 ml count rate; cardiac output is obtained by multiplying thestroke volume by the average heart rate during data acquisition, usually 2 to 3 minutes. Theadvantage of gated studies are that sequential evaluations are possible and that dynamicinformation about cardiac wall motion can be obtained. The major disadvantage is thatmovement artifact becomes more likely as exercise intensity is increased and a correctionfactor must be applied to correct the ventricular counts for attenuation by the structures ofthe chest. This last problem can be over come for sequential studies if a first passdetermination of cardiac output is performed prior to the acquisition of gated information.The ventricle depth for the gated studies is the set with the first pass information obtained atthe identical workload so that the two studies agree and the result is applied to subsequentdeterminations of cardiac output. Gated equilibrium ventriculography has been shown to bea reliable and valid means of measuring cardiac output and ventricular volumes (111, 122,99123). Data obtained by this method correlates well with that obtained by contrast and first-pass ventriculography (122).Measurement of blood volume: Plasma volume can be measured relatively simply and accurately by the use ofradioisotopes. Currently, the use of radioiodine ( 1311) labelled serum albumin (RISA) is themethod recommended by The International Committee for Standardization in Haematology(76). Briefly, 5m1 of radioiodine labelled human serum albumin is injected intravenouslyand the time of injection recorded. At 10, 20 and 30 minutes from the time of the originalinjection 5m1 of blood is withdrawn and plasma volume is calculated as:S•D•V Pv —Powhere S is the concentration of radioactivity in a prepared standard solution, D is thedilution of the standard, V is the volume of RISA injected, and 130 is the concentration ofradioactivity in the sample extrapolated back to time zero, based on the radioactivity countsin the three measured samples. This method can also used with only a single sample taken at10 minutes. Total blood volume can then be calculated from the hematocrit (H y) and theplasma volume:PvBV —11 - 0.9H vwhere 0.9 is a correction factor for relating whole body hematocrit to venous hematocrit.100Pulmonary transit time: Theory: The mean transit time for a well mixed indicator to flow through a specificvolume at a given flow rate is described by the relationship: transit time = volume/flow. Thetime required for an indicator to flow past an observation point down stream from an entrypoint is related not only to the time it take the bolus to flow past the point but also howquickly it arrived there. Transit time is the time that it takes a bolus to remain in acompartment if it is injected directly and instantaneously into the compartment. The firstmoment describes not only the time that the indicator is in the compartment but also howquickly or slowly it arrived there. The first moment therefore represents the summation ofall transit times up to that point. If it were possible to deliver indicator materialinstantaneously into the compartment of interest the first moment would be the same as thetransit time. Transit time of a compartment can be determined by subtracting the firstmoment of the bolus from that of the compartment.To measure pulmonary transit times, the bolus or input curve is derived from alabeled RBC time activity curve of the right ventricle, the output curve is derived from theleft ventricle and transit time is determined by subtracting the first moment of the rightventricular curve from the first moment of the left ventricular curve. This method is referredto as the centroid method. Deconvolution is a mathematical process by which a frequencydistribution of transit times (a transfer function, h(t)) can be derived from the input (rightventricular) and output (left ventricular) time activity curves. It is important to note that thetransit time obtained from either method represents the the delay of the bolus throughpulmonary arteries, arterioles, capillaries, venules, veins, left atrium and left ventricle anddoes not just represent pulmonary transit time.101Frequency distribution of pulmonary transit times: A frequency distribution (h(t)), such that the transformation of the input curve by thetransfer function produces the output curve can be written as:i(t)*h(t) = o(t)where i(t) is the input curve, o(t) is the output curve and * represents convolution or acomplex transformation. The process of determining the shape of h(t) is termeddeconvolution.Several methods have been used to determine the shape of h(t). For example,o(t) = di('c)h(t-t)thwhere i is a variable used only for integration. If this is the case then a numerical solutioncan be sought by assuming that the integral can be approximated by a sum. Then theapproximation of the output curve is given by:o(nAt) = /i(sAt)h(n-s+1)At)where n is an integer which varies between 1 and m (m is the number of data points in theinput and output functions) At is the time interval between the two data points and s is theinteger variable for the summation. Knowing o(nAt) and i(nAt) it is possible to solve for thefrequency distribution of transit times. Unfortunately this method is extremely unstable inthe initial part of the input curve where it is required to extrapolate towards zero. Thisuncertainty is then propagated throughout the solution.Fourier analysis which describes h(t) in terms of a Fourier transform (3) has alsobeen used for calculation of transit times.3 [o(t)] = 3 [i(t)*h(t)]= 3 [i(t)] • 3[h(t)]and solving for h(t):102h(t) = z [ 3 [0(t)] L5 [i(t)]this analysis, however also suffers from the same problem as solution one, namelypropagation of error.More recently, deconvolution analysis using multiple fitting functions has been usedto determine h(t). This has the advantage that knowledge of h(to) is not crucial and thereforeeliminates the problem of propagation of error. The first step in this process is to fit asmooth function to i(t) and o(t). This involves elimination of noise from movement artifactduring data collection, re-circulation of tracer and background radiation. Application of agamma function to indicator dilution curves gives an excellent fit (153). The family ofcurves given by this model are described byC(tn )= k(tn -ta)cce -(tn-ta)/13 'where k, a and 13 are arbitrary ,ta and to are times of appearance and nth time respectivelyand C(tn) is the indicator concentration at tn. This function can be linearized (see (153) fordetails) and then weighted least squares procedure can be used to fit the preceding equationto the input and output curves.The area under the curve can then be determined using00JC(t)dt = to + 13 2(a+1).0Since for the determination of transit times the absolute area under the curve is notimportant, the area under both the input curve and the output curve can be set equal to 1 (allindicator that appears in the right ventricle must at some time appear in the left ventricle).The next step is to determine many h(t)s, each with an area under the curve equal to 1,making the assumption that since both the input curve and output curve are described by a yfunction that h(t) is also y distributed. A family of twenty to fifty curves (ie h i (t),h2(t)...h50(t) is then generated with a mean transit time for these curves that is equal to the103difference between the first moments of the input and output curves and variances that rangefrom the variance of the input curve to the variance of the output curve. Each of these curvesis then convoluted with the input curve to produce a unique output curve each of whichcontributes to the final output curve:o(t) = i(t)*h(t)=i(t)*[hi(t) + h2(t) +^h5o(t)]Given many I n(t) it should be possible to be able to approximate h(t). If the area under h(t)is set equal to 1 then:o(t) = i(t)*[ahi(t) + bh2(t) + ch3(t)...]where a + b + c = 1ando (t) = ai(t)*hi(t) + bi(t)*h2(t) + ci(t)*h3(t)...Multiple regression analysis is then applied to determine the contribution of each ofthese curves to h(t). Unfortunately, each of the generated curves are highly correlated withone another (a situation termed multicollinearity) resulting in either a highly unstable result,such as negative numbers for a,b,c,... (since i(t) = 1 and o(t) = 1 the area under each h n(t)must also equal one and a,b,c,... must add up to 1 and vary between 0 and 1). Anotherproblem is the refusal of least squares regression software packages to complete theregression since every curve is very similar. In order to overcome these difficulties,constrained ridge regression is used to determine the relative contribution of a,b,c... to thefinal solution. Ridge regression acts on the assumption that the least squares estimation ofthe regression coefficients tends to be too large and therefore applies a shrinkage factor tothe least squares estimator biasing it towards zero. The amount of shrinkage that ridgeregression applies to each regression coefficient is proportional to to the coefficient's104variance; it is assumed that the greater the variance the less the certainty that the estimatedcoefficient reflects the true value.The final step in this process involves convoluting the input curve with the derivedtransfer function. If the resulting o(t) is the same as the measured o(t) then the transferfunction must be correct and the analysis is complete. If the result is unacceptable then theentire process is repeated until a good fit is obtained.A further problem associated with deconvolution analysis revolves around thedefinition and quantification of a "good fit" of the derived o(t) with the actual o(t). Thecurves under discussion represent complex mathematical functions which are difficult toassign numerical values to determine goodness of fit. If a numerical scale is assigned, afurther problem arises with quantification. For example a numerical score of 7 may fit betterthan 10, but how much better? The usual method relies on manual observation to determinegoodness of fit. While not entirely satisfactory it should be noted that small changes in fitare associated with minimal, if any, changes in transit time.Incomplete mixing of the bolus can lead in either over or underestimation of times,therefore is is preferable to measure transit time downstream from the site of injection. Othersources of error include poor bolus technique and cross contamination of time activitycurves from overlying structures in the chest.Multiple inert gas eliminationVentilation and perfusion relationships The partial pressure of 02, CO2 and N2 in any gas exchanging unit is determined bythree major factors: the ventilation perfusion ratio, the composition of the inspired gases andthe composition of mixed venous blood. If CO2 is the gas of interest the following equationcan be derived:iTCO2 =•CO2 •K105where VCO2 is CO2 output, VA is alveolar ventilation, PACO2 is alveolar PCO2 and K is aconstant. Similarly the loss of CO2 from the capillary blood can be described by :VCO2 = O(CiTc02 - Cc'co2)Where Q is blood flow, C\-Tc02 is mixed venous CO2, and Cc CO2 is end capillary CO2.Under steady state conditions these equations must be the same andVA •PA CO2•K = O(CVCO2 Cc' CO2)rearranging:VA CVCO2 Cc ' CO2PA CO2•KThe effect of changing inspired oxygen concentration on arterial blood gases isdetermined by the ventilation perfusion ratio. For lung units with VA/Q > 1 end capillary02 is high regardless of F102, whereas lung units with VA/Q - 0.1 end capillary 02 willincrease rapidly as the F102 is increased. Those with V A/Q < .01 show little response toincreasing 02 unless the inspired 02 is greater than 50%. The effect of changing the mixedvenous P02 is the most marked in lung units with VA/Q between 0.1 and 1, and endcapillary 02 will decrease rapidly when the mixed venous P02 falls (Figure 10). This isespecially important when one considers the effect of exercise on mixed venous P02 asvalues of about 25 torr are not uncommon (27) in normal subjects and in athletes exercisingnear max (V02 - 5.0 1-min-1 , Q - 33 1-min-1) this can be calculated to be less than 20 tonby the Fick equation. Alterations in the inspired P02 can also alter the VA/Q relationshipby causing hypoxic vasoconstriction with low F102, or collapse of low VA/Q units as 02is absorbed with high F102 .106Figure 10. The relationship between end capillary P02 and mixed venous P02 forlung units of differing ,T,A/()1 60'E^12080n.400VA/Q =10VA/Q=1.2112(^vA/Q 0.1.21^4kvAio =0.01VA/Q =0.001Ar-0^10 20 30 40 50 60 70Mixed Venous P01 (Ton)From West, 1977 (175)Multiple inert gas elimination theory: An inert gas can be defined as one that obeys Henry's law, that is, that at constanttemperature the solubility of the gas in a liquid is directly proportional to the pressure the gasexerts on the liquid. The diffusion of gas through tissues is described by Fick's law ofdiffusion. This states that the rate of transfer of a gas through a tissue or membrane isproportional to the difference in partial pressure of the gas on either side of the tissue. Thiscan be written as :F = K(Pi - P2)(1)107Where F is flow of the gas across the membrane, K is a constant and P 1 and P2 are partialpressures of the gases on either side of the membrane. The constant K is related to thesurface area that is in contact with the gas, the solubility of the gas in the tissue, and isinversely proportional to the thickness of the membrane and the square root of themolecular weight of the gas of interest (166). Fick's law when applied to a finite volume ofblood (a "slug") makes several assumptions: the blood is perfectly mixed, with ahomogeneous capacity for the gas, there is no axial diffusion and the slug of blood isflowing at a uniform rate. The preceding equation can be expressed in terms of diffusingcapacity:Vx(t) = -Dx[PAx-Px(t)](2)Where Vx is the flow rate and Dx is the diffusing capacity for gas x, PAx is the alveolarpressure and Px(t) is the instantaneous capillary pressure of gas x. This equation can alsobe expressed as a rate in change in partial pressure in pulmonary capillary blood:Vx(t) = OrPx(t)i-V4where Px(t) is the rate of change of capillary partial pressure, Vc is the capillary bloodvolume and R is the solubility of the gas in blood.Diffusing capacity can also be expressed in terms of its component parts:A ^ax ^Dx =^• , " "VMWx(3)(4)108where k is the diffusion coefficient of gas x in the blood-gas barrier, A is the cross-sectionalarea over which diffusion is occurring, d is the thickness of the blood-gas barrier, ax is thesolubility of the gas x in the blood-gas barrier and MW is the molecular weight of the gas.Equation 3 and 4 can be combined with equation 2 and integrated with respect totime, giving the following:100 kA ax^tPx(t) = PAx + (P vx- PAx)-e L 60 dVc PxVMWx_(5)Where t = time and Pir is mixed venous Px. Although complicated, equation 5 can be usedto make the following points: Absolute solubility of a gas is less important than the ratio ofsolubilities in blood and alveolar wall tissue, making the ratio of ax to 13x an importantregulator of equilibrium rate. In the same fashion, if two gases are equal in other factorsthey will differ in their equilibration rates on the basis of molecular weight.The Fick principle can also be applied to an inert gas (x):V•PIx - VA•PAx = AxQ(Pc'x - Pix )where VI is the inspired ventilation in the lung unit, VA is expired ventilation in the lung unitof interest, Pix , Pc'x and Pv- x are inspired, end capillary and mixed venous concentrationsof gas x , and Xx is the blood gas partition coefficient:xx = R x (Pbar -PH2O body temp)/100For most inert gases Piz can be considered to be zero and and in steady state the equationsimplifies to :PAX^P c ' x^Xx•^•PV x^Piix^X x + VA/Q109This analysis has considered VA/Q relationships in one lung unit or a perfectlyhomogeneous lung, but it can be applied to gas exchange in a heterogeneous lung dividedinto N compartments each representing a given ^ratio by summation (69, 168):P exp x A,x^ -^ VAi •x^j=1^(Xx + VA/05 E P art xPAT xXx x + VA/45where P exp x is the mixed expired partial pressure of the gas and P art x is the mixedP exp x^ P art xarterial partial pressure. ^ has been termed the excretion (E) of gas x and ^PV- x Pv-xis referred to as the retention (R) of gas x.Practical aspectsAlthough theroretically it would be possible to apply this model to a single gas ofknown X , it is preferable to use six to eight gases of varying X. A gas mixture of 20%SF6 (X - 0.005), 20% cyclopropane (X - 0.5) and 60% ethane (X - 0.1) is bubbled understerile conditions into a bag of 5% dextrose followed by injections of 2 m1.1 -1 enflurane (X- 2) , 0.5 m1-1- ldiethyl ether (A, - 12) and 6 m1.1- lacetone (X - 300). This mixture isinfused into a vein and after a period of equilibration arterial blood and mixed expired gassamples are simultaneously obtained in ungreased gas tight syringes (167). The gases in theblood are extracted into the gaseous phase by the addition of helium and then analyzedusing electron capture (in the case of SF6) and flame ionization (in the case of the other110five gases) gas chromatography. The results (Peqm - equilibrium pressure) are correctedback to the original concentrations in blood (Po) using the following equation:Vg VIPt) = Peqm • [1.0 +where k is the barometric pressure-saturated water vapour pressure/100, S is the solubilityof the gas in blood (measured directly by a second equilibration) and V1 and Vg are thevolume of liquid and gas in the sample respectively. Mixed venous concentrations of the gasare then calculated using the Fick equation for inert gases.Pi = Pa + PEVE XQTVE is measured in the usual fashion, QT is total cardiac output and X is obtained from themeasured solubility:— Sx • 1(Pbar-PH20)00Dead spaceRetentions (R) and Excretions (E) are calculated as described previously. Anenormous amount of information can be obtained from this analysis (69, 181). The inertgas dead space (VDIG) represents ventilation that does not come into contact with perfusionand results in a decreased excretion ratio of the gas.VDIG — RhomoWhere Rhomo and Ehomo are retentions and excretions of the homogeneous lungrespectively. Physiological dead space represents both inert gas dead space and thatresulting from excess ventilation to a lung unit. Physiological dead space ratio can becalculated as:Rhomo-Ehomo111Pa PE- —VD Pa-PE Pi Pv R-EVT — P a^Pa — RPvWhere R and E are retention and excretion in the heterogeneous lung. The alveolar deadspace which is physiological dead space minus inert gas dead space can also be determineddirectly from retention and excretion data:, Pa homoPE homo VDalv (Pa-Pa homo)^+ (PE homo-PE)VT^ PaThe effect of VA/Q heterogenity is to increase the retention and decrease the excretion of agas.ShuntShunt refers to that portion of blood flow that bypasses gas exchanging areas andcan be divided into intra-pulmonary and extra-pulmonary shunt. The effect of either is toincrease the retention of any gas without an effect on the excretion of a gas. Note that theinert gas analysis does not consider extra-pulmonary shunt. Venous admixture (Qva)includes both pure intra pulmonary shunt and that from over perfusion of lung units and canbe calculated by:Qva PA - Pa Pa - PA QT PA - Pv Pi - PAAlveolar partial pressure can be calculated using mixed expired partial pressure from:PE = VDIGPA 1 VTand the preceding equation becomes:Qva R - E' where E' — ^•^ 1 -— 1 - E '^VDIG QT VT112VA/Q distributionsThe VA/Q distribution can be estimated from the inert gas data using a linear leastsquares regression with enforced smoothing (69, 168) and from this model, predicted Pa02and PaCO2 can be calculated. Diffusion limitation for inert gases and 02 can be determinedfrom this model. Diffusion limitation for inert gases would manifest as retention of gaseswith high molecular weight (SF6 and enflurane), while that of 02 would be evident as amuch lower Pa02 than predicted from the model.The degree of VA/Q mismatch is generally evaluated from two types of indices ofdispersion (55) one derived from the model described above and the other derived directlyfrom the retention and excretion data. The log standard deviations of blood flow (log SDQ)and ventilation (log SDI>) are calculated as the square root of the second moment about themean for both blood flow and ventilation (174). Normally this value is about 0.3-0.4 with0.6 being at the upper limit of normal in young healthy subjects (181), and representing analveolar-arterial difference of about 5 torn. In addition the curves resulting from the graph ofventilation of blood flow versus ventilation/perfusion ratio are centered on VA/ Q of 1 anddo not have very high or low VA/Q, or areas of shunt.More recently Gale et al., (55) derived three indices of dispersion: DISPR*analogous to log SDQ, DISPE analogous to log SNT and DISPR*_E an overall index ofdispersion.DISPR*-E0\1 nI(Ri-Et)2== 100 x^i 1 n0\/ n1(Ri-Rhomot)2i=1 DISPR*= 100 x1130\1 nI(Ehomot-Et)2i=DISPE = 100 x^1 nwhereEhomoi = Rhomoi Xi  Xi+ VAQTand Ei and Ri represent excretions and retentions of the gas of interest and Et is excretioncorrected for dead space:Ei — Ei V DVTThese have been shown to correlate well with log SDQ and log SlDr (55), although it shouldbe noted that DISPR unlike log SDa includes shunt (VA/Q = 0) and DISPE includes areasof dead space (VA/Q = 00) which is not true of log Spy. •Alveolar arterial differenceFor a distribution of ventilation and perfusion ratios the partial pressure of mixedexpired gas is given by:PA — j=Ny, PAjj=1and the mixed mixed arterial partial pressure is given by :j=NIPaj*ajPa _ j=1 j=Nj=1j=NIPAj*VAjj=1114As the distribution of ventilation and perfusion broadens the PAO2 and Pa02 move furtheraway from each other; in the perfectly homogeneous lung without shunt, or VA/Q mis-match they would be equal. The alveolar-arterial difference tends to zero as A, tends to zeroor infinity, and [A-a]D02 is the greatest for gases of intermediate solubility. Reducing theVA/Q has the most effect on gases of low solubility, while raising the VA/Q has the mostpronounced effect on gases of high solubility (180).115APPENDIX C STATISTICAL ANALYSES AND RAW DATA* = significant at p < 0.05Anova tablesMetabolic dataTemperatureSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 4.167*Between subjects 9 2.367 0.0371 Rest vs Heavy exercise 4.167*Within subjects 30 Rest vs Maximal exercise 24.000*treatment 3 24.33 0.0001 Light vs Heavy 0residual 27 Light vs Maximal 8.167*total 39 Heavy vs Maximal 8.167*VentilationSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 9.280*Between subjects 9 0.085 0.9997 Rest vs Heavy exercise 101.007*Within subjects 30 Rest vs Maximal exercise 244.312*treatment 3 296.57 0.0001 Light vs Heavy 49.054*residual 27 Light vs Maximal 158.360total 39 Heavy vs Maximal 31.140*i-Q2Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 58.312*Between subjects 9 0.032 1 Rest vs Heavy exercise 293.176*Within subjects 30 Rest vs Maximal exercise 426.225*treatment 3 524.67 0.001 Light vs Heavy 89.988*residual 27 Light vs Maximal 169.234*total 39 Heavy vs Maximal 12.410*116'CO2Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 25.803*Between subjects 9 0.028 1 Rest vs Heavy exercise 188.121*Within subjects 30 Rest vs Maximal exercise 364.961*treatment 3 439.65 0.0001 Light vs Heavy 74.638*residual 27 Light vs Maximal 196.680*total 39 Heavy vs Maximal 28.997*ilEic•2Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 18.191*Between subjects 9 1.064 0.42 Rest vs Heavy exercise 4.043*Within subjects 30 Rest vs Maximal exercise 0.245treatment 3 28.248 0.0001 Light vs Heavy 5.082*residual 27 Light vs Maximal 22.657*total 39 _ Heavy vs Maximal 6.279*ilvilco2Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 20.476*Between subjects 9 0.746 0.66 Rest vs Heavy exercise 13.818*Within subjects 30 Rest vs Maximal exercise 7.616*treatment 3 23.298 0.0001 Light vs Heavy 0.653residual 27 Light vs Maximal 3.117*total 39 _ Heavy vs Maximal 0.917117Respiratory exchange ratioSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.580Between subjects 9 0.633 0.76 Rest vs Heavy exercise 6.460*Within subjects 30 Rest vs Maximal exercise 25.928*treatment 3 37.573 0.0001 Light vs Heavy 7.740*residual 27 Light vs Maximal 28.435*total 39 Heavy vs Maximal 6.504*131gxegaLgUaSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 1.288Between subjects 9 0.481 0.88 Rest vs Heavy exercise 15.624*Within subjects 30 Rest vs Maximal exercise 61.266*treatment 3 72.959 0.0001 Light vs Heavy 7.940*residual 27 Light vs Maximal 44.787*total 39 Heavy vs Maximal 15.012*PaCO2Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 2.908Between subjects 9 0.941 0.51 Rest vs Heavy exercise 1.425Within subjects 30 Rest vs Maximal exercise 14.074*treatment 3 31.565 0.0001 Light vs Heavy 8.404*residual 27 Light vs Maximal 29.777*total 39 Heavy vs Maximal 6.543*118Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 1.435Between subjects 9 0.577 0.83 Rest vs Heavy exercise 6.551 *Within subjects 30 Rest vs Maximal exercise 27.046*treatment 3 48.435 0.0001 Light vs Heavy 14.119*residual 27 Light vs Maximal 40.093*total 39 Heavy vs Maximal 6.976*P,122Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 3.634*Between subjects 9 4.319 0.0011 Rest vs Heavy exercise 12.774*Within subjects 30 Rest vs Maximal exercise 1.214treatment 3 13.581 0.0001 Light vs Heavy 2.782residual 27 Light vs Maximal 0.647total 39 Heavy vs Maximal 6.111*1A-AllX22(o)Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.292Between subjects 9 1.41 0.22 Rest vs Heavy exercise 33.422*Within subjects 30 Rest vs Maximal exercise 37.938*treatment 3 65.419 0.0001 Light vs Heavy 27.647*residual 27 Light vs Maximal 31.574*total 39 Heavy vs Maximal 0.143119Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 1.228Between subjects 9 1.173 0.12 Rest vs Heavy exercise 7.45*Within subjects 30 Rest vs Maximal exercise 15.915*treatment 3 18.555 0.0001 Light vs Heavy 2.629residual 27 Light vs Maximal 8.301*total 39 Heavy vs Maximal 1.587Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 17.943*Between subjects 9 0.34 0.95 Rest vs Heavy exercise 33.802*Within subjects 30 Rest vs Maximal exercise 27.379*treatment 3 41.473 0.0001 Light vs Heavy 2.490residual 27 Light vs Maximal 0.993total 39 Heavy vs Maximal 0.338fa:iX22Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 3.985*Between subjects 9 2.983 0.01 Rest vs Heavy exercise 4.075*Within subjects 30 Rest vs Maximal exercise 11.02*treatment 3 11.262 0.0001 Light vs Heavy 0.001residual 27 Light vs Maximal 1.751total 39 Heavy vs Maximal 1.692120MICiET data for all six_gasesResidual sum of squaresSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 1.054Between subjects 9 2.821 0.02 Rest vs Heavy exercise 3.357*Within subjects 30 Rest vs Maximal exercise 3.466*treatment 3 4.612 0.01 Light vs Heavy 0.649residual 27 Light vs Maximal 0.698total 39 Heavy vs Maximal 0.001Mean of Q distributionSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 8.065*Between subjects 9 0.133 0.98 Rest vs Heavy exercise 14.191*Within subjects 30 Rest vs Maximal exercise 44.748*treatment 3 45.611 0.0001 Light vs Heavy 0.860residual 27 Light vs Maximal 14.818*total 39 Heavy vs Maximal 8.540*Mean of V distributionSource Degrees offreedomF- test P value Comparison Scheffe F-testBetween subjectsWithin subjectstreatmentresidualtotal930327390.8161.1870.600.0001Rest vs Light exerciseRest vs Heavy exerciseRest vs Maximal exerciseLight vs HeavyLight vs MaximalHeavy vs Maximal2.91715.917*54.555*5.206*32.243*11.537*121Log SDSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.551Between subjects 9 1.536 0.91 Rest vs Heavy exercise 10.258*Within subjects 30 Rest vs Maximal exercise 24.488*treatment 3 31.046 0.0001 Light vs Heavy 6.054*residual 27 Light vs Maximal 17.963*total 39 Heavy vs Maximal 3.048*Loa SDSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.579Between subjects 9 3.61 0.005 Rest vs Heavy exercise 0.002Within subjects 30 Rest vs Maximal exercise 0.131treatment 3 1.347 0.28 Light vs Heavy 0.600residual 27 Light vs Maximal 1.262total 39 Heavy vs Maximal 0.122ISP *Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.168Between subjects 9 4.428 0.0013 Rest vs Heavy exercise 8.232*Within subjects 30 Rest vs Maximal exercise 19.110*treatment 3 25.757 0.001 Light vs Heavy 6.048*residual 27 Light vs Maximal 15.696*total 39 Heavy vs Maximal 2.257122pISPSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.001Between subjects 9 4.094 0.001 Rest vs Heavy exercise 1.992Within subjects 30 Rest vs Maximal exercise 1.884treatment 3 2.453 0.089 Light vs Heavy 1.949residual 27 Light vs Maximal 1.841total 39 Heavy vs Maximal 0.108Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.111Between subjects 9 3.773 0.003 Rest vs Heavy exercise 3.694*Within subjects 30 Rest vs Maximal exercise 5.909*treatment 3 8.449 0.0004 Light vs Heavy 2.524residual 27 Light vs Maximal 4.400*total 39 Heavy vs Maximal 0.259WA-MDSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.631Between subjects 9 2.899 0.0136 Rest vs Heavy exercise 8.095*Within subjects 30 Rest vs Maximal exercise 16.684*treatment 3 20.998 0.001 Light vs Heavy 4.205*residual 27 Light vs Maximal 10.825*total 39 _ Heavy vs Maximal 1.536123RA-MDSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.517Between subjects 9 1.709 0.1304 Rest vs Heavy exercise 0.176Within subjects 30 Rest vs Maximal exercise 0.043treatment 3 0.567 0.6418 Light vs Heavy 0.090residual 27 Light vs Maximal 0.262total 39 Heavy vs Maximal 0.045(A-a)DSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.0780Between subjects 9 3.809 0.0027 Rest vs Heavy exercise 3.555*Within subjects 30 Rest vs Maximal exercise 6.443*treatment 3 9.094 0.0003 Light vs Heavy 2.581residual 27 Light vs Maximal 5.106*total 39 Heavy vs Maximal 0.4261-A-a1DOziplSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.745Between subjects 9 1.326 0.2652 Rest vs Heavy exercise 7.900*Within subjects 30 Rest vs Maximal exercise 7.969*treatment 3 12.124 0.0001 Light vs Heavy 3.793*residual 27 Light vs Maximal 3.840*total 39 Heavy vs Maximal 0.0001124Source Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 0.0530Between subjects 9 4.593 0.0013 Rest vs Heavy exercise 4.029*Within subjects 30 Rest vs Maximal exercise 4.609*treatment 3 9.679 0.0002 Light vs Heavy 5.001*residual 27 Light vs Maximal 5.646*total 39 Heavy vs Maximal 0.0200Cardiac_dat4Stroke volumeSource Degrees offreedomF- test P value Comparison Scheffe F-testRest vs Light exercise 8.813*Between subjects 9 0.488 0.87 Rest vs Heavy exercise 22.487*Within subjects 30 Rest vs Maximal exercise 59.456*treatment 3 62.601 0.001 Light vs Heavy 3.245*residual 27 Light vs Maximal 22.487*total 39 Heavy vs Maximal 8.813*End diastolic volumeSource Degrees offreedomF- test P value Comparison Scheffe F-testBetween subjectsWithin subjectstreatmentresidualtotal614212201.1121.4210.400.0001Rest vs Light exerciseRest vs Heavy exerciseLight vs Heavy10.429*20.124*1.777125End ystolic volumeSource Degrees offreedomF- test P value Comparison Scheffe F-testBetween subjectsWithin subjectstreatmentresidualtotal614212202.024.400.130.37Rest vs Light exerciseRest vs Heavy exerciseLight vs Heavy_2.2724.069*0.26Ejection fractionSource Degrees offreedomF- test P value Comparison Scheffe F-testBetween subjectsWithin subjectstreatmentresidualtotal614212200.62165.3380.710.0001Rest vs Light exerciseRest vs Heavy exerciseLight vs Heavy35.118*59.657*3.232126RegressionVENCO2 vs Pa02 during heavy exerciseDegrees of Freedom R^R2^Slope^Intercept9^0.43 0.19 0.266 4.233Source^Degrees of freedom F test^P (one tailed)RegressionResidualTotal 1891.85 0.11VE N CO2 vs Pa02 during maximal exerciseDegrees of Freedom R^R2^Slope^Intercept9^0.53^0.28^0.184 12.657Source^Degrees of freedom F test P (one tailed)Regression^1^3.18^0.06Residual^8Total 9VE /V02 vs Pa02 during heavy exerciseDegrees of Freedom R^R2^Slope^Intercept9^0.402^0.161^0.3 1.895Source^Degrees of freedom F test P (one tailed)Regression^1^1.539^0.12Residual^8Total 9.^.ILIILCA22 vs PaO2 during maximal exerciseDegrees of Freedom R^R2^Slope^Intercept9^0.11^0.13^0.055 29.721Source^Degrees of freedom F test P (one tailed)Regression^1^0.109^0.37Residual^8Total 9127Degrees of Freedom R^R2^Slope^Intercept9^0.648^0.420^17.9^42.243Source^Degrees of freedom F test P (one tailed)RegressionResidualTotal 1895.795^0.02Transit, time rest centroid method vs deconvolutionDegrees of Freedom R9^0.993R2^Slope^Intercept0.987^0.963 0.339Source^Degrees of freedom F test P (one tailed)Regression^1^605.419^0.0001Residual 8Total^9Transit time during exercise centroid method vs deconvolutionDegrees of Freedom R^R2^Slope^Intercept9^0.955 0.899^1.152^-0.432Source^Degrees of freedom F test P (one tailed)Regression^1^72.437^0.0001Residual 8Total^9Pulmonary transit time vs Pa02 at maximal exercisePulmonary transit time vs 1A-a1D02(o) at maximal exerciseDegrees of Freedom R^R2^Slope^Intercept9^0.591^0.35^-17.861^78.643Source^Degrees of freedom F test P (one tailed)Regression^1^4.301^0.04Residual 8Total^9128Degrees of freedom F test^P (one tailed)Source0.656Degrees of Freedom R9R2 Slope InterceptSource Degrees of freedom F test^P (one tailed)6.040 0.0201RegressionResidualTotal89Degrees of Freedom R^R2^Slope^Intercept9^0.63 0.39 -23.814 78.304Regression^1^5.283 0.025Residual^8Total 9Exercising pulmonary blood index vs whole blood volume ant-kg-110.43 14.379^-458.816Degrees of Freedom R^R2^Slope^Intercept9^0.651^0.423^-0.088^51.175Source^Degrees of freedom F test P (one tailed)RegressionResidualTotal 1895.871^0.02Pulmonary transit time vs1A-a1D02_(o-p) Pulmonary blood volume index vs fA-a1D02(o) at restPulmonary blood volume index vs fA-a1D02(o) at maximal exerciseDegrees of Freedom R^R2^Slope^Intercept9^0.569^0.34^-8.367^1024.178Source^Degrees of freedom F test P (one tailed)Regression^1^3.84^0.04Residual^8Total 9129Pulmonary blood volume index vs Pa02 at maximal exerciseDegrees of Freedom R^R2^Slope^Intercept9^0.689 0.475 11.073 -242.825Source^Degrees of freedom F test^P (one tailed)Regression^1^7.23 0.01Residual^8Total 9Pulmonary blood volume index vs IA-alDO2 (o-p) at maximal exerciseDegrees of Freedom R^R2^Slope^Intercept9^0.679^0.46^-0.058^56.446Source^Degrees of freedom F test P (one tailed)Regression^1^6.836^0.015Residual^8Total 9Multiple regression to predict Pa02 at maximal exerciseDegrees of Freedom R9^0.94R20.844Parameter Value Partial F P valueIntercept -25.224inert gas (A-a)D 28.356 5.064 0.0327transit time 19.141 26.926 0.0012VE/VCO2 1.714 2.503 0.0028130APPENDIX D: INDIVIDUAL SUBJECT DATA.*=missing data or not collected at this exercise levelSubject 1.Workload (watts)^R 150 300 420Temp. (°C) 36.9 37.0 37.0 37.8HCT 40 45PaCO2 (ton) 36 36 36 31PAO2 (torr) 89 102 112 120Pa02 (torr) 89 96 82 85[A-a]D02(o) (torr) 0 6 30 35[A-a]D02(p) (ton) 5 8 10 15[A-a]D02(o-p) (ton) -4 -2 19 20pH 7.45 7.46 7.41 7.24Sa02 % 97.2 97.6 96.5 93.813;z02 (ton) 30 19 14 11VE (1•min-1 ) 15.743.4 105.4 195.4638 2001 4191 6192VO2 0.min-1)430 1683 4316 7112V. CO2 (1•min-1)24.6 21.7 25.1 31.6VE/V0236.5 25.8 24.4 27.5VE/VCO27.3 15.2# 26.6# 34.4Q (1•min-1)* * * *End diastolic volume (ml)End systolic volume (ml) * * * *Stroke volume (ml) 126 143 186 206Ejection fraction * * * *Transit time deconvolution (s) 10.19 2.92Transit time centroid (s) 10.09 2.99Pulmonary blood volume (1) 1.234 1.694Mean residual sum of squares 99.6 69.05 94.65 33.30.955 2.145 2.82 4.31Mean of a1.045 2.79 3.46 7.575Mean of V0.24 0.37 0.44 0.63LogSD Q0.42 0.80 0.47 0.91LogSD VDISP R*-E 1.052 3.292 2.657 6.439DISP R* 0.476 1.447 1.400 3.422DISP E* 0.559 2.151 1.478 3.522Inert Gas A-a area 0.039 0.122 0.087 0.237131Subject 2Workload (watts) R 150 300 350Temp. (°C) 35.5 36.7 36.7 36.9HCT 45 48PaCO2 (toff) 33 43 30 29PA02 (torr) 108 102 124 123Pa02 (toff) 88 88 87 91[A-a]1302(o) (torr) 20 14 37 32[A-a]D02(p) (toff) 12 9 22 16[A-0302(o-p) (torr) 8 5 15 16pH 7.48 7.40 7.32 7.29Sa02 % 98.3 97.0 96.4 96.4NO2 (toff) 40 21 24 19VE (1•min-1 ) 11.9 44.8 133.1 175.4V02 (1•min-1) 273 1867 3668 4494VCO2 (1-min-1) 240 1834 4720 5431ifEti102 43.59 24.00 36.2939.0349.58 24.43 28.20 32.30VF_JVCO26.1 13.9 24.7 27.6Q (1•min-1)End diastolic volume (ml) 132 159 190 *End systolic volume (m1) 45 43 31 *Stroke volume (ml) 87 116 158 161Ejection fraction 0.67 0.74 0.83 *Transit time deconvolution (s) 7.37 2.78Transit time centroid (s) 7.24 2.77Pulmonary blood volume (1) 0.743 1.277Mean residual sum of squares 9 37 19 220.89 4.53 2.26 3.98Mean of a1.61 5.50 5.30 8.66Mean of V0.64 0.47 0.77 0.67LogSD Q1.33 0.44 1.26 1.13LogSD VDISP R*-E 5.211 1.750 6.273 8.547DISP R* 1.809 0.969 3.098 4.371DISP E* 3.997 0.926 3.764 4.847Inert Gas A-a area 0.215 0.059 0.241 0.327132Subject 3Workload (watts) R 150 300 390Temp. (°C) 36.5 36.7 37.0 37.6HCT 48 48PaCO2 (torr) 38 41 35 31PAO2 (ton) 103 92 113 124Pa02 (torr) 98 95 91 96[A-a]1302(o) (ton) 5 -3 22 28[A-a]I302(p) (ton) 4 7 16 16[A-a]D02(o-p) (ton) 2 -10 6 12pH 7.47 7.42 7.37 7.27Sa02 % 97.9 97.4 96.7 95.8I3-702 (ton) 313 17 2920VE (1.min-1) 17.1 45.0 177.7 195.8V02 (I.min-1) 520 2393 4367 5176VCO2 (1•min-1) 468 1889 4499 665732.9 18.8 40.7 37.8VEJV0236.5 23.8 39.5 29.4VEJVCO2Q (1.min-1) 6.5 14.4 21.9 34.6End diastolic volume (m1) 121 158 178 *End systolic volume (ml) 42 36 37 *Stroke volume (ml) 79 122 140 197Ejection fraction 0.66 0.77 0.80 *Transit time deconvolution (s) 8.75 3.16Transit time centroid (s) 8.60 3.18Pulmonary blood volume (1) 0.94 1.83Mean residual sum of squares 66 67 35 720.92 2.16 2.62 4.73Mean of Q1.25 2.68 5.61 10.31Mean of V0.24 0.29 0.51 0.69LogSD Q0.82 0.65 1.29 1.16LogSD vDISP R*_E 3.213 2.283 8.303 8.071DISP R* 1.054 0.989 3.718 4.214DISP E* 2.469 2.511 5.283 4.381Inert Gas A-a area 0.126 0.086 0.321 0.328133Subject 4Workload (watts)Temp. (C)HCTPaCO2 (torr)PAO2 (ton)Pa02 (ton)[A-a]D02(o) (ton)[A-a]1302(p) (ton)[A-a]D02(o-p) (ton)pHSa02 %K02 (ton)VE (1•min-1)V02 (1.min-1)VCO2 (1 min-1)VE/V02VE/VCO2Q (1•min-1)End diastolic volume (m1)End systolic volume (ml)Stroke volume (ml)Ejection fractionTransit time deconvolution (s)Transit time centroid (s)Pulmonary blood volume (1)Mean residual sum of squaresMean of QMean of VLogSD QLogSD VDISP R*-EDISP R*DISP E*Inert Gas A-a areaR 150 300 36037.3 37.5 38.1 38.840 4636 39 35 30115 101 113 12494 83 74 8121 18 39 431 4 6 1319 14 33 307.4497.17.4195.57.2990.77.1888.136 19 13 1918.5 43.4 114.3 183.7548 2086 4654 5178611 1835 4895 658633.76 20.81 24.56 35.4830.28 23.65 23.35 27.898.5 15.7# 29.0# 31.9#* * * ** * *147 141 187 197* * *7.45 2.677.72 2.681.075 1.422198 125 86 861.19 3.15 2.91 4.561.23 3.55 4.83 11.950.18 0.28 0.40 0.730.18 0.39 0.97 1.180.430 1.261 5.597 10.1240.230 0.668 2.480 5.5250.237 0.706 3.653 5.4190.014 0.045 0.221 0.405134Subject 5Workload (watts) R 150 300 400Temp. (C) 36.7 37.0 36.8 38.3HCT 43.0 44.0PaCO2 (ton) 41 46 44 33PA02 (torr) 103 94 103 122Pa02 (ton) 99 86 74 93[A-a]D02(o) (ton) 4 8 29 29[A-a]D02(p) (ton) 11 12 23 15[A-a]D02(o-p) (ton) -7 -4 6 14pH 7.48 7.45 7.38 7.20Sa02 % 97.9 96.8 94.8 93.5P■;02 (ton) 34  24 2513VE (1.min-1) 9.9 36.4 75.8 147.64141913 3974 5291V02 (1•min-1)VCO2 (1•min-1) 372 1621 3752 607223.91 19.03 19.07 27.9VE/V0226.61 22.46 20.2 24.31VEIVCO27.6 17.8 24.4 36.0Q (1•min-1)End diastolic volume (ml) 154 185 185End systolic volume (ml) 45 33 31 *Stroke volume (m1) 108 152 154 227Ejection fraction 0.71 0.84 0.84 *Transit time deconvolution (s) 9.23 2.99Transit time centroid (s) 9.12 2.74Pulmonary blood volume (1) 1.162 1.719Mean residual sum of squares 27.1 19.7 17.2 61.90.91 2.38 1.99 3.38Mean of 61.67 3.40 4.16 11.30Mean of VLogSD Q 0.42 0.51 0.61 0.791.32 0.66 0.86 1.35LogSD vDISP R*-E 6.791 3.634 7.120 12.974DISP R* 2.366 1.843 3.495 6.760DISP E* 5.017 2.128 4.307 7.292Inert Gas A-a area 0.261 0.132 0.264 0.508135Subject 6Workload (watts) R 150 300 370Temp. (T) 37.1 37.5 37.8 39.0HCT 43.0 42.0PaCO2 (ton) 40 40 38 33PAO2 (ton) 109 105 109 119Pa02 (ton) 107 104 100 110[A-4302(0) (ion) 2 1 9 9[A-a]D02(p) (torr) 3 13 22 21[A-a]D02(o-p) (torr) -1 -13 -12 -12pH 7.41 7.39 7.30 7.20Sa02 % 97.8 97.3 95.5 94.7602 (ton.) 43 24 18 1212.4 44.1 105.4 159.2VE (1•min-1)371 1938 4350 4911V02 (1•min-1)VCO2 (1•min-1) 363 1753 4135 517133.42 22.76 24.23 32.4234.16 25.16 25.49 30.79VE/VCO28.2 16.3 21.9 34.5Q (1•min-1)End diastolic volume (ml) 165 167 186 *End systolic volume (m1) 66 39 56 *Stroke volume (m1) 99 128 130 197Ejection fraction 0.60 0.77 0.77 *Transit time deconvolution (s) 8.85 3.56Transit time centroid (s) 8.96 3.76Pulmonary blood volume (1) 1.217 2.105Mean residual sum of squares 79.3 16.7 16.8 2.20.66 1.86 3.36 4.58Mean of Q1.43 5.03 10.11 14.19Mean of V0.30 0.63 0.83 0.90LogSD Q1.65 1.39 1.22 1.16LogSD VDISP R*-E 5.216 11.457 11.900 11.974DISP R* 1.588 5.062 6.421 6.847DISP E* 4.155 7.509 6.522 6.051Inert Gas A-a area 20.9 0.214 0.442 0.462136Subject 7Workload (watts) R 150 300 370Temp. (°C) 36.9 36.9 36.9 38.1HCT 41.0 43.0PaCO2 (toff) 41 44 37 32PAO2 (toff) 107 99 114 123Pa02 (torr) 105 98 94 106[A-4302(o) (toff) 2 1 20 17[A-a]D02(p) (torr) 5 10 16 19[A-a]D02(o-p) (toff) -3 -9 3 -2pH 7.40 7.38 7.31 7.16Sa02 % 97.8 97.2 96.6 95.1P;702 (toff) 32 22 17 1915.3 49.3 105.8 184.0VE (1•min-1)454 2424 4108 4912VO2 (1•min-1)436 2165 4218 5713VCO2 (1•min-1)33.7 20.34 25.75 37.46VE/V0235.09 22.77 25.08 32.21VEJVCO26.2 18.2 26.8# 37.6Q (1•min-1)End diastolic volume (ml) 146 233 223End systolic volume (ml) 57 67 39Stroke volume (m1) 89 166 184 229Ejection fraction 0.62 0.72 0.83Transit time deconvolution (s) 8.73 2.91Transit time centroid (s) 8.63 2.95Pulmonary blood volume (1) 0.897 1.836Mean residual sum of squares 190.2 88.5 18.8 15.41.20 1.72 3.04 4.72Mean of Q4.15 6.34 11.84 13.39Mean of0.40 0.53 0.77 0.85LogSD QLogSD V 1.82 1.70 1.42 1.10DISP R*-E 12.570 13.778 14.567 11.072DISP R* 4.484 5.630 7.338 6.316DISP E* 9.327 9.482 8.529 5.663Inert Gas A-a area 0.513 0.550 0.574 0.437137Subject 8Workload (watts)Temp. (°C)HCTPaCO2 (toff)PAO2 (toff)Pa02 (toff)[A-a]D02(o) (toff)[A-a302(p) (toff)[A-a]D02(o-p) (toff)PHSa02 %PC/02 (toff)VE (1•min-1)V02 (1•min-1)VCO2 (1.min-1)VE/V02VE/VCO2Q (1•min-1)End diastolic volume (m1)End systolic volume (ml)Stroke volume (m1)Ejection fractionTransit time deconvolution (s)Transit time centroid (s)Pulmonary blood volume (1)Mean residual sum of squaresMean of QMean of VLogSD QLogSD VDISP R*-EDISP R*DISP E*Inert Gas A-a areaR 150 300 34037.2 37.3 37.5 38.343.0 41.042 43 36 3097 95 110 12299 78 83 93-2 17 27 291 20 25 20-3 -3 2 97.4097.27.3894.67.3395.07.2594.3404 18 151412.0 43.4 107.6 169.6350 2163 4337 4673299 1813 4218 532434.29 20.06 24.81 36.2940.13 23.94 25.51 31.866.6 15.5 22.8 35.7145 185 188 *63 59 48 *81 126 139 2090.56 0.68 0.74 *9.82 2.589.80 2.601.079 1.541* 4.4 8.6 5.0* 4.72 1.96 4.77* 11.43 7.23 14.63* 0.82 0.75 0.881.00 1.58 1.151.898 9.993 16.805 11.9995.684 5.688 7.632 6.9161.393 5.015 10.811 5.9990.076 0.373 0.640 0.472138Subject 9Workload (watts)Temp. (°C)HCTPaCO2 (ton)PAO2 (ton)Pa02 (ton)[A-a]1302(o) (ton)[A-a]1302(p) (ton)[A-a]D02(o-p) (ton)pHSa02 %Pv02 (ton)VE (1•min-1)V02 (1.min-1)VCO2 (1-min-1 )VE/V02VE/VCO26 (1•min-1)End diastolic volume (ml)End systolic volume (ml)Stroke volume (m1)Ejection fractionTransit time deconvolution (s)Transit time centroid (s)Pulmonary blood volume (1)Mean residual sum of squaresMean of QMean of VLogSD QLogSD VDISP R*-EDISP R*DISP E*Inert Gas A-a areaR 150 300 32037.1 37.5 37.5 38.243.0 45.036 40 33 3193 94 114 120103 94 93 101-10 0 21 1926 9 24 13-36 -8 -3 67.41 7.38 7.27 7.2397.6 96.5 95.6 95.3035 22 5 1110.2 49.3 139.1 158.1439 2476 5006 4749307 1947 4915 521323.23 19.91 27.79 33.2933.22 25.32 28.3 30.335.9 18.4 26.4# 26.4#164 19159 43105 148 165 1650.64 0.8011.35 2.6911.36 2.681.117 1.1812.1 50.1 11.1 6.60.89 2.32 4.04 5.282.76 5.97 13.56 17.200.76 0.53 0.86 0.921.34 1.38 1.22 1.4113.412 9.924 13.115 12.3626.105 4.387 7.246 7.2778.445 6.469 6.939 6.0010.480 0.391 0.513 0.497139Subject 10Workload (watts) R 150 300 390Temp. (°C) 36.9# 37.1# 37.3# 38.2#HCT 40.0 40.0PaCO2 (ton) 39 41 39 38PA02 (ton) 95 92 101 113Pa02 (ton) 96 92 80 90[A-a]D02(o) (ton) -1 0 21 23[A-a]D02(p) (ton) 3 11 9 18[A-a]D02(o-p) (ton) -3 -10 12 5pH 7.44 7.42 7.39 7.36Sa02 % 97.5 96.9 95.2 95.1013■7•02 (ton) 35 17 8 712.4 48.9 106.2 182.7VE (l min-1)V02 (1•min-1) 354 2264 4562 5569VCO2 (1•min-1 ) 283 1791 4026 626135.03 21.6 23.28 32.81VE/V0243.82 27.3 26.38 29.18VE/VCO2-^-1Q ( ) 7.0 16.4 28.5 33.8End diastolic volume (m1) 165 198 235End systolic volume (ml) 63 36 48 *Stroke volume (ml) 102 161 187 222Ejection fraction 0.62 0.81 0.79 *Transit time deconvolution (s) 11.43Transit time centroid (s) 11.75 2.15Pulmonary blood volume (1) 1.352 1.211Mean residual sum of squares 113 49.3 20.6 12.61.42 1.88 3.39 4.38Mean of Q2.08 2.59 6.53 11.83Mean of V• 0.26 0.37 0.58 0.86LogSD Q0.91 0.68 1.11 1.20LogSD vDISP R*-E 4.290 3.532 7.090 10.284DISP R* 1.469 1.587 3.483 5.683DISP E* 3.217 2.279 4.205 5.412Inert Gas A-a area 0.171 0.130 0.275 0.418140


Citation Scheme:


Usage Statistics

Country Views Downloads
United States 27 5
United Kingdom 15 1
Russia 14 0
China 9 11
Iraq 2 0
Pakistan 1 0
Egypt 1 0
France 1 0
Finland 1 0
City Views Downloads
Tamworth 14 0
Unknown 12 11
Penza 6 0
Washington 6 0
Cleveland 5 0
San Francisco 4 0
Mountain View 4 0
Guangzhou 3 0
Shenzhen 3 11
Ashburn 3 0
Beijing 3 0
Verona 2 0
Sulaymaniyah 2 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items