non-invasive measurement of cardiac output by a single ... · the absorption rate of acetylene...

Thorax 1984;39: 107-113

Non-invasive measurement of cardiac output by asingle breath constant expiratory techniqueURI ELKAYAM, ARCHIE F WILSON, JOHN MORRISON, PAUL MELTZER, JAMESDAVIS, PAUL KLOSTERMAN, JED LOUVIER, WALTER L HENRY

From the Divisions ofCardiology and Pulmonary Disease, Department ofMedicine, University ofCalifornia,Irvine Medical Center, Orange, California, USA

ABSTRACT A new single breath test has been developed that measures pulmonary blood flow (Qc)and pulmonary tissue volume by using the fact that is proportional to the relationship betweenthe absorption rate of acetylene (C2H2) from the alveolar gas and the rate of change of lungvolume during constant expiratory flow. To make these measurements a bag in bottle system witha rolling seal spirometer, a mass spectrometer, and a minicomputer with analogue to digitalconversion have been used. Qc was compared with cardiac output measured by the thermodilu-tion technique in 20 patients with cardiac disease; some also had mild chronic obstructive pulmo-nary disease. The mean (SD) resting 0c for the group was 5-27 (1.22) I/min and the cardiacoutput measured by thermodilution was 5 30 (1.31) /min. The mean difference between the twoestimations of cardiac output was 0-03 1 and the standard deviation of this difference was 0.76 1.The Qc technique was not successful in patients with an FEV1/FVC less that 60%, but seemed tobe accurate in those with higher FEV1/FVC values. Correction of Qc for the effect of venousadmixture in 14 patients resulted in an average 19% overestimation of cardiac output (6.01(2.52) l/min v 5-05 (1.64) I/min). It is concluded that cardiac output can be accurately measuredin patients with cardiac or mild pulmonary disease. No correction for venous admixture due toventilation-perfusion mismatch was necessary in these patients, presumably because the largebreath used by the technique overcomes most mild ventilation-perfusion maldistribution. Thesefindings, in addition to the non-invasive nature of the technique, suggest potential value for themeasurement of cardiac output in various clinical conditions.

Non-invasive tests of cardiopulmonary functionbased on soluble gas absorption were developedearly in this century.'-3 Widespread application ofthese tests, however, has occurred only in the case ofthe diffusing capacity of the lung for carbon monox-ide (TLco).4 The measurement of pulmonary capil-lary blood flow (Qc) (total cardiac output minusanatomical shunt) has not been widely used, prob-ably because most available techniques are timeconsuming, cumbersome, or of uncertain reliability.The uncertainty is attributable either to lack of afirm theoretical basis, to lack of clinical validation,or to the unknown effects of breathing manoeuvresand regional maldistribution of function on themeasurements.3 5-8

Address for reprint requests: Dr AF Wilson, Department ofMedicine, University of California Medical Center, 101 City DriveSouth, Orange, California 92668, USA.

Accepted 17 October 1983

We have recently introduced a constant expirat-ory flow single breath technique that is based onanalysis of the changes in alveolar gas absorptionwhich occur during individual phases of respiration.This method has a sound theoretical basis8 and willbe shown below to allow rapid and accurate meas-urement of cardiac output. In this study we com-pared pulmonary capillary blood flow measurempentsobtained by the single breath technique with cardiacoutput measured simultaneously by a thermo-dilution technique in patients with cardiovascularand, in some instances, mild pulmonary disorders.

Methods

THE PATIENTS STUDIEDThe study group (table) was composed of 11 menand 12 women who had undergone diagnostic car-diac catheterisation because of cardiac symptoms.The patients' ages ranged from 20 to 77 years with a

107

on April 3, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.39.2.107 on 1 F

ebruary 1984. Dow

nloaded from

http://thorax.bmj.com/

Elkayam, Wilson, Morrison, Meltzer, Davis, Klosterman, Louvier, HenryData on the 20 patients studied

Age (y) Cardiac andlor pulmonary History of Venous admixtureand sex diagnosis smoking (% shunt)65 F RHD - 11-130 F IHSS + NA40 F No cardiac or pulmonary disease + NA47 F Primary pulmonary hypertension + 18.020 M Non-cardiac pulmonary oedema - 13-050 F RHD and CAD, CABG - NA55 M CAD, CABG (COPD) + 14-253 M CAD + 063 M CAD (COPD) + 13.366 F HTN - NA31 M RHD NA 40-070 F RHD - 44.441 F IHSS + NA47 M CAD, CABG - 16-552 M CAD, (COPD) + 18-262 M CAD, CABG (COPD) + 6-677 F CAD,CABG - 1844 M CAD - NA44 F Coronary spasm 1-442 M Atrial septal defect + 18-2

RHD-rheumatic heart disease; IHSS-idiopathic hypertrophic subaortic stenosis; CAD-coronary artery disease; CABG-recoveringfrom coronary artery bypass graft surgery; COPD-chronic obstructive pulmonary disease; HTN-hypertension; NA-not available;+ heavy cigarette smoking (> 20 cigarettes/day for > 10 years); + light cigarette smoking.

mean of 50 years. Twelve patients had a history ofheavy cigarette smoking (> 20 cigarettes/day formore than 10 years) and two of a lighter tobaccoexposure. Eight patients were non-smokers and inone patient a smoking history was not obtained.Four patients also had chronic obstructive pulmo-nary disease and one was recovering from non-cardiogenic pulmonary oedema. Three patients hadmore severe chronic obstructive disease (FEV,/FVC% less than 60%-see below); in these patientsQc could not be measured reproducibly and theirresults are not included in the data reported. Of theremaining 20 patients, eight were found to havecoronary artery disease, one had coronary spasm,three had valvular heart disease, and one had acombination of coronary and valvular disease. Twopatients had idiopathic hypertrophic subaorticstenosis, one had an atrial septal defect, one hadprimary pulmonary hypertension, and two had noevidence of cardiac disease, although one of themwas found to have systemic hypertension. Routineanalysis of arterial oxygenation was performed aspart of cardiac catheterisation in 14 patients.

SINGLE BREATH CONSTANT EXPIRATORYTECHNIQUETheoryDuring expiratory flow at a constant rate it may beshown8 that

- VE In FA/FA0c at In [(VA + at Vt)/(VAo + alt Vt)]

where FA is the alveolar concentration of acetylene(C2H2) and FAO = FI/FIHe/FAHe. FI is inspiredacetylene concentration; FIHe is inspired helium

concentration. VA is alveolar volume and VAO iS VAat full inspiration and calculated from the formulaVAO = (VI - VD) (FiHe/FAHe), where VI is theinspired volume and VD is dead space. VE iSexpiratory flow rate, Vt is pulmonary parenchymaltissue volume, and ab and at are Bunsen coefficientsfor blood and pulmonary tissue. Since FA and VAcan be constantly measured during expiratory flowand the other indices are constant, the slope andintercept for C2H2 during expiration may be calcu-lated on the basis of linear regression. Pulmonaryparenchymal tissue volume can be measureddirectly9 but because it has only a small effect on thecalculation of Qc8 it can be estimated (3.5 g/cmheight). Since the flow pattern during inspirationmay be considered instantaneous and all calcula-tions emanate from expiratory data, there is no needfor a constant inspiratory flow rate with this method.Similarly, breath holding time need not be carefullycontrolled. It is important, however, that expiratoryflow should be relatively constant.ProcedureTest gas was inspired by the human subjects from abag in bottle (fig 1). After exhalation to nearresidual volume, subjects inspired to near total lungcapacity through a five way valve connected to thebag, which contained a test gas mixture consisting of10% helium, 0-8% acetylene, 0 1%g carbon 18 label-led carbon monoxide, and 21% oxygen, with nit-rogen making up the balance. The gases were moni-tored throughout the test with a Perkin-ElmerMGA 1100 mass spectrometer. After a breath holdof about two seconds, the subjects exhaled at a con-stant rate that varied from 200 to 500 ml/s. At least

108



j.com/


ebruary 1984. Dow

nloaded from


Non-invasive measurement ofcardiac output by single breath technique

TESTGAS

VIDEO DISPLAYand

PRINTERFig 1 Schematic diagram ofequipment used to measure pulmonary capillary blood flowby a single breath technique. RV-residual volume; TLC-total lung capacity.

two seconds of expiratory flow were necessary toobtain sufficient data for accurate computation.Correction was made for the presence and effect ofwater vapour during expiration (assumed to be47 mm Hg (6.3 kPa)). Inspiratory and expiratoryvolumes as well as expiratory flow rate were moni-tored by connecting a 12 litre rolling seal spirometerto the bottle and to the expiratory line through aStarling resistor, which was used to maintain a con-stant expiratory flow. Gas and volume data weremonitored on line with a digital computer (DigitalElectronics Corporation, PDP 11/23). A typicalgraphical display of the data on volume and threegas concentrations is shown in figure 2. Alveolarvolume (VA) was determined by subtracting thevolume expired from the alveolar volume at endinspiration (VA0). A linear regression of ln FA/FA0(C2H2) v In [(VA + at Vt)/(VA0 + at Vt)] was com-

puted and displayed. A typical graphic display ofthese calculated data is shown in figure 3. Since datapoints were usable only after dead space washoutand before closing volume, the computer operatorselected the limits of the regression data to be usedfor calculation. From the slope and regression of thedata on expired acetylene (5c was calculated. Simi-larly, TLco was calculated from the carbon monox-ide regression. Reported values are averages of atleast two determinations. Usually three or moredeterminations were made.

CARDIAC OUTPUT DETERMINATIONA thermodilution Swan-Ganz catheter was insertedin each patient through the femoral vein as a part ofcardiac catheterisation. Cardiac output was deter-mined by the thermodilution technique with the useof 5% dextrose in water as the indicator. Cardiac

109



j.com/


ebruary 1984. Dow

nloaded from


Elkayam, Wilson, Morrison, Meltzer, Davis, Klosterman, Louvier, Henry

01

I I

I II II I

INSPIRATION I BREATHI HOLD II I

10 15 jEXPIRATION I

TIME (S)Fig 2 Diagrammatic representation ofdisplay of the rawdata used by the computer to calculate pulmonary capillaryblood flow, pulmonary tissue volume, and diffusingcapacity. He-helium concentration; C2H2-acetyleneconcentration; C'80-I8C labelled carbon monoxide.

output determinations were performed in triplicate(less than 10% variation) and computations wereaccomplished with a bedside computer (EdwardsLaboratory Model 9520A). The data reported areaverages of three determinations.

z4 ~~~C2H2

~ ~ ~ C

In [(VA + atVt)/(VAO + atVt)JFig 3 Relationship between acetylene concentration (lnFAIFAd and alveolar volume plus pulmonary tissue volume(In [(VA +atV)/(VA. +atVt1)) used in the calculation ofpulmonary capillary blood flow.

In 10 patients the determination of Qc by thesingle breath technique was performed first and wasimmediately followed by cardiac output determina-tion based on the thermodilution technique. In theother 10 patients the thermodilution cardiac deter-mination was made first. All studies were performedwith patients in the supine position.

CALCULATION OF VENOUS ADMIXTUREThe venous admixture was calculated from the stan-dard shunt formula'0 in the 14 patients in whomarterial blood gas data were available. Alveolaroxygen tension was calculated from the alveolar gasequation'0 and assumed to be equal to pulmonarycapillary oxygen tension, from which pulmonarycapillary oxygen saturation was calculated. Thearteriovenous oxygen content difference wasassumed to be 4 5 ml oxygen/100 ml blood.

STATISTICAL ANALYSISRegression analysis was performed on the values ofpulmonary capillary blood flow and cardiac output.To assess the effect of correcting Qc for venousadmixture on the comparison with cardiac output,data are expressed as means and standard deviations(in parentheses).

Results

The close relationship between cardiac output asdetermined by the thermodilution technique andpulmonary capillary blood flow as determined by thesingle breath technique is shown in figure 4. Therelationship between the techniques can bedescribed by the equation Qc = -0-23 + 1*06 car-diac output +0 70 I/min. Mean (SD) cardiac outputwas 5-30 (1.31) 1/min and pulmonary capillary bloodflow 5-27 (1-22) I/min. The mean difference bet-ween techniques was 003 (0.76) I/min. The 95%confidence limits for the estimation of the (ther-modilution) cardiac output by the single breathmethod were about ± 1.49 1.The mean coefficient of variation for the

acetylene measurements was 8-7% (4.4%), whichmay be contrasted with 7.0% (6.4%) for the ther-modilution measurements. The mean value of ven-ous admixture during the breathing of room air wasfound to be 15*5% (13.0%) in the 14 patientsstudied. Correction of Qc for venous admixture inthese 14 patients resulted in considerable overesti-mation of the data by comparison with the cardiacoutput (6.01 (2.52) /min v 5-05 (1-64); p =< 0-05). No significant difference was found bet-ween the uncorrected values of Qc (4.89 (1.58) Vmin) and the cardiac output as measured by thethermodilution technique (5.05 (1-64) I/min).

110



j.com/


ebruary 1984. Dow

nloaded from


Non-invasive measurement ofcardiac output by single breath technique

Fig 4 Relationship between cardiac outputas determined by the thermodilutiontechnique and pulmonary capillary bloodflow in the 20 patients. The line of identity isshown.

:0

2 3 4 5 6 7 8(C,H,) PULMONARY CAPILLARY BLOOD FLOW (6c)-(L/MIN)

In 12 patients pulmonary function values wereavailable. In three of these patients, the ln - lngraphs were not reproducible and were clearlyunphysiological. These three patients all hadFEV1/FVC% values of less than 60% and were notincluded in the comparison data of figure 4. Theother patients in whom pulmonary function wasevaluated had FEV,/FVC% values ranging from66% to 86%; their data were included in theanalysis. Within this small sample of nine there wasno correlation between FEV,/FVC% values andthermodilution/Oc measurement differences.

Discussion

Assessment of cardiac output is essential to theunderstanding of cardiocirculatory function and themanagement of various clinical disorders. Theavailability of an accurate and convenient non-invasive technique for measuring cardiac outputshould therefore be of great clinical importance.Most non-invasive methods of measuring cardiacoutput have considerable inherent limitations andare not in widespread use. For example, M modeechocardiography has been shown to be useful in theassessment of size and function of the left ventricle,especially in normal individuals."I 12 The accuracy ofthis technique, however, decreases appreciably as

the ventricle becomes larger and when segmentalwall motion abnormalities develop.'3 Radionuclideventriculography has recently been shown capableof producing an accurate assessment of left ventricu-lar volumes and cardiac output.'415 The exposure toradioactive material associated with this techniqueis, however, an important limitation when repeatedmeasurements are needed. Preliminary studies haverecently shown that transcutaneous Doppler aorticblood flow measurements made with the aid ofechocardiography can provide an accurate assess-ment of cardiac output.'6 17 Further studies areneeded to evaluate this method in disease states.The use of soluble gas techniques for the determi-

nation of pulmonary capillary blood flow and car-diac output is not new. Previous attempts using thesolubility characteristics of acetylene, carbon mon-oxide, and helium have met with varying degrees ofsuccess. A multiple single breath technique was firstdescribed by Cander and Forster in 1959.3 Later arebreathing method was developed by Sackner etal.5 The multiple single breath method is technicallydifficult to perform since each breath must haveidentical inspiratory and expiratory flow rates, andidentical beginning and ending lung volumes. Anadditional possible source of error is that Q, maychange during the hour or so necessary to completethe measurements. Although Cander and Forsterfou-nd good agreement between the average esti-

10-

9-

8-

7.-

6-

5-

z

a0

0C-)

0-)

wI

4

3

2-

1 -

ill

.

* 0



j.com/


ebruary 1984. Dow

nloaded from


Elkayam, Wilson, Morrison, Meltzer, Davis, Klosterman, Louvier, Henrymated values of cardiac output obtained by theirmethod and the average values of cardiac outputmeasured by the Fick method, as reported by otherinvestigators, no direct comparison was performed.3The rebreathing method described by Sackner

et al is more clinically applicable than the multiplesingle breath technique since it permits rapid meas-urements5 but it involves such rapid, deep breathingthat appreciable changes of Qc can result during thetest. Furthermore, alveolar volumes change rapidlyduring the test and can be never known with preci-sion, a fact that may introduce an error in calcula-tion of Qc Although Sackner et al reported goodagreement between Oc as determined by the rebrea-thing technique and cardiac output as measuredsimultaneously by the dye dilution method in anaes-thetised dogs,5 the clinical usefulness of thesefindings in man has not been assessed.The single breath constant expiratory technique

used in the present study has some other practicaladvantages over the other two soluble gas techni-ques. The method is not dependent on reproductionof breath size and beginning lung volume; all theexpired gas data are used and the single breath madein performing the test probably does not change thesupine Oc.1-8 Initial clinical evaluation of the techni-que showed it to be reproducible9 and the valuesobtained were close to those expected. The presentstudy shows good agreement between the pulmo-nary capillary blood flow as measured by thismethod and cardiac output as determined by thethermodilution technique in patients with cardiacdisorders, patients with histories of heavy smoking,and several patients with chronic obstructive lungdisease.

Uncertainty about the accuracy of measurementsof cardiac output with any soluble gas technique inpatients with pulmonary disease is mainly attribut-able to the unknown effects of appreciable regionalmaldistribution of ventilation. Data from the threepatients with moderately severe lung disease indi-cate that the technique is not applicable to patientswith more than mild obstruction. The techniqueproved to be accurate, however, in patients withmild chronic obstructive pulmonary disease andsome venous admixture. The results suggest that thebreathing manoeuvre, which requires inspiration tonear total lung capacity, probably overcomes mild tomoderate ventilation-perfusion maldistribution.Recalculation of Oc after correction for venousadmixture resulted in gross overestimation of thecardiac output as measured by the thermodilutiontechnique. The small difference (3%) betweenmeasured cardiac output and uncorrected Qc is ingood agreement with the known anatomical venousadmixture.'9

Recently Denison and colleagues20 have reportedthe use of a single breath technique based on calcu-lation of Qc during expiration over small volumechanges. This approach has been contrasted with theother techniques used for the related measurementof diffusing capacity.8

Several factors may limit the clinical applicationof the single breath technique for the assessment ofcardiac output. One is the need for expensiveequipment such as the mass spectrometer and com-puter. Another is the need to wait 15-20 minutesbetween determinations for washout of acetylene.Furthermore, the method is inaccurate in patientswith more than mild pulmonary functional altera-tions.We have shown that measurement of pulmonary

capillary blood flow by a new single breath techni-que can provide an accurate estimate of cardiac out-put in patients with cardiovascular or mild chronicobstructive pulmonary disease or both. Correctionfor the effect of venous admixture using arterialoxygen values is unnecessary. The accuracy, repro-ducibility, and simplicity of the single breath con-stant expiratory procedure suggest that it may proveto be a useful technique for the non-invasive deter-mination of cardiac output in normal subjects as wellas in patients with cardiac or mild pulmonary dis-ease.

This work was supported in part by a grant from theNational Heart, Lung and Blood Institute (PHS-HL-1670C-2) and the American Heart Association.

References

'Krogh A, Linhard J. Measurement of blood flow throughthe lungs of man. Skand Arch Physiol 1912;27:100-25.

2 Bauman H, Grollman A. fiber der theoretetischen undpraktischen Grundlagen und die klinische Zuverlas-sigkeit der Acetylenmethode zur Bestimmung desMinutevolumens. Klin Med 1930;115:41-53.

3Cander L, Forster RE. Determination of pulmonaryparenchymal tissue volume and pulmonary capillaryblood flow in man. J Appl Physiol 1959;14:541-50.

4Ogilvie CM, Forster RE, Blakemore DS, Morton JW. Astandardized breath holding technique for measure-ment of the diffusing capacity of the lung for carbonmonoxide. J Clin Invest 1957;36:1.

s Sackner MA, Greeneltch D, Heiman MS, Epstein S,Atkins N. Diffusing capacity, membrane diffusingcapacity, capiliary blood volume, pulmonary tissuevolume and cardiac output measured by a rebreathingtechnique. Am Rev Respir Dis 1975;111:157-65.

6Armitage GH, Arnott WM. Effect of voluntary hyper-pnoea on pulmonary blood flow. J Physiol1949;109:64-9.

7Michaelson ED, Sackner MA, Johnson RL jun. Vertical

112



j.com/


ebruary 1984. Dow

nloaded from


Non-invasive measurement ofcardiac output by single breath techniquedistribution of pulmonary diffusing capacity and capil-lary blood flow in man.J Clin Invest 1973;52:359-69.

8 Martonen TB, Wilson AF. Theoretical basis of singlebreath gas absorption tests. J Math Biol1982;14:203-20.

9 Morrison JT, Wilson AF, Vaziri MD, Brunsting L, DavisJ. Determination of pulmonary tissue volume, pulmo-nary capillary blood flow and diffusing capacity of thelung before and after hemodialysis. Int J Artif Organs1980;3:259-62.

:0 Rahn H, Fenn WO. A graphical analysis of the respirat-ory gas exchange, the 02-CO2 diagram. WashingtonDC: American Physiological Society, 1955.

" Pombo JF, Troy BL, Russell RO jun. Left ventricularvolumes and ejection fraction by echocardiography.Circulation 1971;43:480-90.

12 Fortuin NJ, Hood WP jun, Sherman E, Craig E. Deter-mination of left ventricular volumes by ultrasound.Circulation 1971;4:575-84.

13 Linhart JW, Mintz GS, Segal BL, Kawai N, Kotler MN.Left ventricular volume measurement by echocardiog-raphy: fact or fiction? Am J Cardiol 1975;36:114-8.

4 Slatsky R, Karliner J, Rice D, Kaiser R, Pfisterer M. Leftventricular volumes by gated equilibrium radionuclideangiography: a new method. Circulation 1979;60:555-64.

5 Konstam MA, Wynne J, Holman BL, Brown EJ. NeillJM, Kozlowski RT. Use of equilibrium (gated)radionuclide ventriculography to quantitate left ven-tricular output in patients with and without left-sidedvalvular regurgitation. Circulation 1981;64:578-85.

16 Darsee JR, Walter PF, Nutter DO. TranscutaneousDoppler method of measuring cardiac output. II-Non-invasive measurement of transcutaneous Dop-pler aortic blood velocity integration and M-modeechocardiography. Am J Cardiol 1980;46:613-8.

17 Gardin JM, Tobis JM, Debestani A, et al. Superiority oftwo-dimensional measurement of aortic vessel diame-ter in Doppler echocardiographic estimates of leftventricular stroke volume. Clin Res 1983;31:8A (abs-tract).

18 Vermeire P, Butler J. Effect of respiration on pulmonarycapillary blood flow in man. Circ Res 1968;22:299-308.

19 Melemgaard K. The alveolar-arterial oxygen difference:its size and components in normal man. Acta PhysiolScand 1966;67:10-20.

20 Denison DM, Davies NJH, Meyer M, Pierce RJ, ScheidP. Single-exhalation method for study of lobar andsegmental lung function by mass spectrometry in man.Respir Physiol 1980;42:87-99.

113



j.com/


ebruary 1984. Dow

nloaded from

non-invasive measurement of cardiac output by a single ... · the absorption rate of acetylene...

Documents