Clinical Comparison of Pressure-Pulse andIndicator-Dilution Cardiac Output Determination

ROBERT M. CUNDICK, JR., PH.D., AND REED M. GARDNER, PH.D.

SUMMARY Two clinical studies of cardiac output determination using the pressure-pulse technique are

presented. The Warner pressure-pulse method of estimating cardiac output was compared with the dye-dilution technique in 17 patients. Both the Warner and a variation of the Bourgeois pressure-pulse methodswere compared with thermodilution in 13 patients.The Warner vs dye-dilution comparison resulted in a poor correlation coefficient (r = 0.61). Values for the

Warner equation calibration constant (K) were nonstationary, varying with time from -69 to 135% of the ini-tial value in individual patients. When thermodilution was compared with the two methods, the correlation waspoor (r = 0.58 for the Warner method; r = 0.50 for the modified Bourgeois method).The Warner and a variation of the Bourgeois pressure-pulse methods for monitoring the cardiac output of

critically ill patients with widely varying mean arterial pressures are not sufficiently reliable for clinicaldecision-making.

ACCURATE ESTIMATION of cardiac output fromthe central arterial pressure wave form has beenwidely investigated. A reliable method would offer amajor advantage over other techniques in that con-tinuous, beat-to-beat stroke volume calculationswould be possible. Many methods have beenproposed.'-" Digital computers have significantlyreduced the computation time required to processpulse-pressure data, and have thus focused greaterattention on the feasibility of using pulse-pressuretechniques for continuous intensive care cardiac out-put monitoring. Investigations comparing pressure-pulse methods of cardiac output determination withother measurement techniques have produced mixedresults.6-10

In this paper we present the results of two clinicalstudies. The first is a comparison of the Warnerpressure-pulse method' with simultaneously per-formed dye-dilution curves. The second is a com-parison of the Warner and Bourgeois methods withsimultaneously determined thermodilution cardiacoutputs.

Materials and MethodsDetailed derivations of the Warner and Bourgeois

methods are found elsewhere." 6 The Warner equationused in the studies is:

SV = K (Pmd) (1 + Sa (1)

where SV is stroke volume, K is a calibration con-stant determined by making a simultaneous measure-ment by an independent method, Pmd is the average in-crement in mean pressure, over the arterial bed, fromthe previous diastole to end-systole (referred to as the

From the Department of Medical Biophysics and Computing,University of Utah, LDS Hospital, Salt Lake City, Utah.Address for correspondence: Reed M. Gardner, Ph.D., Depart-

ment of Medical Biophysics and Computing, LDS Hospital, 325Eighth Avenue, Salt Lake City, Utah 84143.

Received April 21, 1979; revision accepted January 30, 1980.Circulation 62, No. 2, 1980.

mean distending pressure [fig. 1]), and Sa and Darepresent the systolic and diastolic areas, respectively.In Warner's original derivation, the calculation of themean time of transmission (t,) of the pressure wave(from a central location to the periphery) required thesimultaneous measurement of both a central and aperipheral arterial pressure. The mean time of trans-mission was later assumed to be constant at 80 msec,"1making measurement of the peripheral arterialpressure unnecessary. The method, as published in1966 and 1968,"l 12 used the square root of mean dis-tending pressure (V/P,m) instead of mean distendingpressure. The method currently used by Warner andassociates is equation 1.The equation derived by Bourgeois and associates5

is:

SV = KA [(Pes - Pd) + SA1 (2)

where SV is stroke volume and KA is a calibrationconstant determined by simultaneously measuringstroke volume by an independent method, Pd is thediastolic pressure preceding the systolic pressure waveof the curve of interest (see fig. 2), P,, is the pressureat end-systole (dicrotic notch), SA is the area underthe curve during systole, and r is the time constantof the exponential decay of the diastolic portion of thepressure contour. Dependent variables are thelogarithms of sample points from the diastolic por-tion of the. pressure curve from the dicrotic notch toend-diastole; time is the independent variable.The first comparative study was between the

Warner pressure-pulse method and dye-dilution per-formed on 17 patients hospitalized for open heart sur-gery at LDS Hospital. One set of measurements wasperformed before surgery. Within the first 2 days aftersurgery, three to six additional sets of measurementswere performed.A set of measurements consisted of doing dye

curves until two cardiac output values, within 5% ofeach other, were obtained and then averaged. Atypical protocol included a measurement set taken 2-4hours after surgery, with additional measurement sets

371

by guest on May 12, 2018

http://circ.ahajournals.org/D

ownloaded from

VOL 62, No 2, AUGUST 1980

E

E

a:CI)

U)

CL 4

.8 1.0TIME (sec)

t

Al1l

1 1,

AI

11

1

l

1.

FIGURE 1. Parameters of the Warner pressure-pulsemethod. Sa, Da, and Pmd represent the systolic area,diastolic area, and mean distending pressure. The symbol t,represents the average transmission time to the periphery.

obtained at 4-hour intervals thereafter. Indocyaninegreen dye (0.8 mg in 2 ml aqueous solvent) wasmanually injected through a central venous catheter.Withdrawal of the blood sample for dye-dilutiondetermination was made through a cannula in theradial artery before surgery and postoperativelythrough a short Teflon catheter placed in the brachialartery after surgery. Dye curves were obtained with aWaters cuvette densitometer (model XC-50B). Theoutput voltage of the densitometer amplifier wassampled for 45 seconds with a 10-bit analog-to-digitalconverter at 6.25 Hz and processed with the dye-dilution program on LDS Hospital's CDC 3300 com-puter.'3

J60-

:20-E e

E

W

80-

CL)

0 .2 .4 .6 .8 1.0 1.2 1.4TIME (sec)

FIGURE 2. Parameters of the Bourgeois method. Pd, Pes,and SA represent the end-diastolic pressure, end-systolicpressure and systolic area, respectively.

A 100-cm long, 20-gauge, fluid-filled Tefloncatheter (CAP INTRAFUSOR- 18, Sorenson Re-search Company) was attached to a pressure trans-ducer for monitoring central arterial pressure.14' 15 Theresulting system had a natural frequency (fn) of 20 Hzand a damping coefficient (f) of 0.25. Pressure trans-ducers used were P23Db, P23ID (Statham Instru-ments), or Bentley 800 (Bentley-Trantec, Inc.). Thecatheter was inserted through the radial artery and ad-vanced to the subclavian artery to record centralarterial pressure. A continuous-flush system14' 15 wasused to keep the catheter patent (CFS INTRAFLO,Sorenson Research Company). The transducer wasconnected to an amplifier system (Model 780-7C,Hewlett-Packard) that provided a continuous oscillo-scopic display. The signal was conditioned using apressure amplifier designed at LDS Hospital.13 Thepressure wave form was sampled by a computerprogram at 200 samples/sec for 45 seconds. ForWarner method calculations, the program selected thefirst 16 artifact-free pressure contours, computed thepressure-pulse parameters on each and calculated theaverages. Parameters displayed included systolicpressure, diastolic pressure, mean pressure, strokevolume, heart rate, cardiac output and total peripheralresistance.The second study made comparisons between the

Warner and Bourgeois methods and the thermo-dilution-determined cardiac output on 13 patients inthe respiratory intensive care unit (RICU) at LDSHospital. Since thermodilution measurements andcentral arterial pressure monitoring were performedroutinely as a part of the clinical managementprogram, the comparison studies were performedwithout departing from normal patient-care proce-dures. One hundred forty-four sets of simultaneouspressure-pulse and thermodilution measurementswere taken, with a mean of 3.5 measurements per set.A typical protocol included three to four measure-ment sets per day, taken at 3-4-hour intervals. Exceptfor one patient, all were observed for at least 2 days;one of the patients was observed for 3 weeks and onefor 5 weeks. Thermodilution measurements weremade using 10 ml of iced normal saline injectedthrough the right atrial port of a Model 44166 Instru-mentation Laboratories Inc. #7F flow-directed ther-modilution catheter. The resistance change of thecatheter thermistor was measured and converted to anoutput voltage by an amplifier designed and built bythe Department of Medical Biophysics and Comput-ing at the University of Utah. The voltage output ofthe amplifier was sampled at 6.25 samples/sec for 45seconds with an analog-to-digital converter andprocessed by the computer. For low-output states, alonger sampling duration was used to allow the curveto return to baseline. After the computer processed thedata, both the cardiac output value and the time-temperature curve were displayed. Sampled analogdata for pressure, dye-dilution and thermodilution wasstored temporarily on magnetic disc; the data werelater transferred to digital tape for further evaluation.

Pattern recognition of the pressure wave form used

372 CIRCULATION

by guest on May 12, 2018

http://circ.ahajournals.org/D

ownloaded from

PRESSURE-PULSE CARDIAC OUTPUT/Cundick and Gardner

for the Bourgeois method was accomplished by dis-playing the raw sampled pressure-pulse data on a

computer terminal. The incisura, the beginning andend of systole for each beat, were located by adjustinga cursor under human visual control. Ten artifact-free,contiguous pressure wave forms after injection of thethermal indicator were then selected. The computerthen determined Pe,, Pd, and SA (equation 2) byaveraging these values over the 10 beats. The timeconstant r was determined by the computer using themethod for averaging successive diastolic decaysdescribed by Bourgeois.5 The average diastolic decaycurve was then used to determine r as described in theoriginal Bourgeois paper.'6

Results

Five of the 17 patients in the dye-dilution study werefemale; average age was 47.9 years. Five of the 13patients in the thermodilution study were female;average age was 49.0 years.

Cardiac output data from both studies were plottedagainst time. In most cases the pressure-pulse cardiacoutputs were observed to begin diverging from the in-dicator dilution cardiac outputs within a few hours.Long-term drift in the pressure-pulse method had beenobserved before in this clinical setting and had been at-tributed to changes in the cardiovascular system thataltered the calibration constant of the pressure-pulseequation. If the pressure-pulse methods failed as anabsolute measure of changes in cardiac output, thequestion remained whether they could at least beuseful in following trends. To reduce the effect ofdrifts between the pressure-pulse and indicator-dilution methods, the results of both studies wereanalyzed using percent change between the average ofeach measurement set and the average of the preced-ing measurement set. Thus, for each measurement set(in) the percent change from the previous set (in l)would be:

In- In1Percent change = 100 X inn-l (3)

A scatter diagram comparing percent changes in theWarner pressure-pulse method against percentchanges in the dye-dilution cardiac outputs of the firststudy is shown in figure 3. The correlation coefficientwas r = 0.61, with a regression equation of y = 0.69x- 0.59 (52 measurement sets). Scatter diagrams of theresults of the second study comparing thermodilutionto the Warner and Bourgeois pressure-pulse methodsare shown in figures 4 and 5. The correlation co-efficient for the Warner equation was r = 0.58, witha regression equation of y = 0.78x + 2.0 (144measurement sets). The Bourgeois method yielded a

correlation coefficient of r = 0.50, with a regressionequation of y = 0.61x + 2.14 (144 measurement sets).

In the first study, the transition from pre- to post-thoracic surgery was often accompanied by largechanges in cardiac output. Consequently, one reasonfor the improved correlation coefficient in the dye-dilution study over the thermodilution study was that

m0

I-

z

3

100

m0

W

m-

3:

r=.61

0

00 / -

* 0

. *-

0

I DENTITY---REGRESSIONLINE

-60 -20 0 20DYE DILUTION

60 100

FIGURE 3. Data from the dye-dilution study comparingpercent changes in cardiac outputs computed using theWarner pressure-pulse method against percent changes indye-dilution cardiac output values. The regression equationis y = 0.69x - 0.59.

more data points were present at the extremes of car-diac output. When the same data were graphed andthe preoperative mneasurements eliminated, the cor-relation coefficient was reduced from r = 0.61 to r =

0.48.Considering that long-term changes in the pressure-

r - .580

5-p0e

0

a

0 0

. .

IDENTITY- - -REGRESSION

LINE-100 ,

-100 -60 -20 0 20 60 100THERMODILUTION

FIGURE 4. Data from the thermodilution study compar-ing percent changes in cardiac outputs computed using theWarner pressure-pulse method against percent changes inthermodilution cardiac output values. The regression equa-tion is y = 0.78x + 2.0.

373

, /'..

by guest on May 12, 2018

http://circ.ahajournals.org/D

ownloaded from

VOL 62, No 2, AUGUST 1980

r z.50*

%0 00

A.1

IDENTITY---R EGR ESSION

LINE0

'v , , , ,

-100 -60 -20 0 20THERMODILUTION

1 I- -l60 100

FIGURE 5. Same as figure 4, with the Bourgeois pressure-pulse method rather than the Warner method.

pulse calibration constant might cause the pressure-pulse cardiac outputs to drift away from the truevalues, the time between calibration and successivemeasurements was shortened to evaluate the effect, ifany, on the agreement between thermodilution andpressure-pulse cardiac outputs. In the thermodilutionstudy, measurements were excluded where more than5 hours had passed since the previous set. No improve-ment was observed; correlation coefficients decreasedto r = 0.48 from r = 0.58 for all data with the Warnermethod and to r = 0.46 from r = 0.50 with theBourgeois method (59 measurement sets). Correlationcoefficients decreased further when maximum timebetween measurement sets was reduced to 2 hours (r= 0.32 for the Warner method; r = 0.24 for theBourgeois method; 31 measurement sets.)

evaluation of the data from patients in the dye-dilution study was conducted to determine if a rela-tionship existed between change in the constant andchange in the mean pressure. Plots of mean pressurevs measurement number and K vs measurementnumber on the same graph revealed that an inverserelationship between K and mean pressure was pres-ent in 80% of the patients. Linear regressions wereperformed on corresponding values for mean pressureand K in individual patients. Only two patients failedto show appreciable inverse correlation. The averagecorrelation coefficient was r = 0.80. These resultssuggested that the vascular compliance was not con-stant over the mean pressure ranges of the post-operative thoracic surgery patients.A similar analysis was made using data from the

thermodilution study. New values of the K (Warner)and KA (Bourgeois) constants were computed for eachmeasurement set using the thermodilution cardiacoutput values. Constants for each method were plottedagainst mean pressure for the corresponding measure-ments. Correlation coefficients for both the Warnerand Bourgeois methods are listed in table 1 for in-dividual patients. A correlation coefficient was not ob-tained from one patient, in whom only two measure-ment sets were performed. Correlation coefficientswere not as high as for the dye-dilution study.

Perhaps the high inverse correlations between meanpressure and K on the dye-dilution study were notobserved in the thermodilution study because thesepatients were followed longer. When adjacentmeasurement sets were connected with lines on themean pressure vs K and KA plots, the majority of thelines connecting successive points were of negativeslope with a similar slope value. An example for theWarner constant against mean pressure is given infigure 6 for the patient with the most data. A similareffect was noted on all patients for both the Warnerand Bourgeois K values.

Bourgeois et al.5 based their pressure-pulse methodon detecting changes in the time constant of the

DiscussionData from the dye-dilution measurements of the

first study were used to "recalibrate" the Warnerequation for each measurement set to determine whatchanges were occurring in the calibration constant ofthe pressure-pulse equation. Values for the Warnerequation calibration constant (K) were found to benonstationary, varying with time; maximal errorswere -69 and 135% of the initial value in individualpatients. A basic assumption of the Warner pressure-pulse equation is that the relation between change involume and the corresponding change in pressure inthe arterial tree is linear. The ratio of change involume to change in pressure is the vascular com-

pliance, characterized by K in the Warner equation. Inthe derivation of the Warner equation, K is assumedto be constant in the normal physiologic pressurerange. The finding that K is not constant in the clinicalsituation conflicted with this assumption. Further

TABLE 1. Correlation Coefficients for Pressure-Pulse MethodsPlotted Against Mean Pressure

Patient r (Warner) r (Bourgeois)1 +0.03 -0.352 -0.63 -0.653 -0.71 +0.074 -0.99 -0.945 +0.12 -0.026 -0.78 -0.887 -0.31 -0.208 +0.06 -0.119 -0.94 -0.1210 -0.35 -0.2311 -0.70 -0.5812 -0.57 -0.38

6(

Ch0ILLXU,

0Liicc

0a:2

K J | |~~~~~~~~~~~~~~~~

374 CIRCULATION

by guest on May 12, 2018

http://circ.ahajournals.org/D

ownloaded from

PRESSURE-PULSE CARDIAC OUTPUT/Cundick and Gardner

2.8-

2.4-

X 2.0-I-9

en _

1.6-

1.2-

0.8-60 80 100MEAN PRESSURE (mmHg)

FIGURE 6. Scatter diagram showing the value computedfor the Warner constant for each new measurement setplotted against mean pressure. The lines connect successivemeasurements.

diastolic decay of the aortic pressure wave form. Theyfound that the best results were obtained when the tipof the arterial catheter was positioned in a segment ofthe thoracic aorta that gave nearly perfect exponentialdecay in the diastolic wave form. They found this "op-timal" location about 4 cm cephalad of the dorsal in-sertion of the diaphragm in the dog. Dr. Edward L.Bond and Dr. Reed M. Gardner at our institutioncould not find the optimal location for the tip of thearterial monitoring catheter. All arterial monitoringcatheters used in the present studies had their tips inthe subclavian artery. Since this is not the optimallocation described by Bourgeois and associates, it maybe partially responsible for the poor performance ofthe Bourgeois method.

Although several clinical trials of the Warner andother variants of the pressure-pulse method have beenreported,"s 7 8, 10-12 we are unaware of other clinicalreports testing the Bourgeois method. The Bourgeoismethod has been tested under a wide variety ofphysiologic conditions, but all the experimental resultspublished have been from acute, short-term animal

models. Based on the changes in the shape of patientpressure wave forms over time in our critically illpatients, we are uncertain that the so called optimallocation of the catheter will remain fixed. Only carefulexperimentation on several subjects will answer thisquestion.

ConclusionsFurther research on pressure-pulse methods is

warranted, although these results show that thepressure-pulse methods studied were not reliable in-dicators of change in cardiac output. Calculation ofparameters other than cardiac output may be of value.For example, in the derivation of the Warner equationfrom the Windkessel model, the calibration constant(K) represents the compliance of the aorta and largearteries.5 In these studies the arterial pressure waveform was being monitored while cardiac output wasbeing measured by indicator dilution. The arterialwave forms and the cardiac output value were used tocalculate a new value of K for each measurementprocedure. If K represents arterial compliance, thisprotocol could provide a convenient method formonitoring arterial compliance clinically. However,the validity of this method for measuring compliancewould have to be established by comparison withalternative methods.Our data show that use of the Warner and

Bourgeois pressure-pulse methods under the con-ditions presented do not provide sufficiently reliabledata for clinical decision-making in the care ofcritically ill patients. Further research is required ifthe accuracy of the equations is to be improved or ifthe equations are to be adapted for monitoringvariables other than cardiac output. The methods maybe reliable when used on patients who have morestable mean arterial pressures than those we studied.

References1. Warner HR, Swan HJC, Connolly DC, Tompkins RG, Wood

EH: Quantitation of beat-to-beat changes in stroke volumefrom the aortic pulse contour in man. J Appl Physiol 5: 495,1953

2. Jones WB, Hefner LL, Bancroft WH Jr, Klip W: Velocity ofblood flow and stroke volume obtained from the pressure pulse.J Clin Invest 38: 2087, 1959

3. Kouchoukos NT, Sheppard LC, McDonald DA, Kirklin JW:Estimation of stroke volume in the dog by a pulse contourmethod. Circ Res 24: 611, 1970

4. Guier WH, Friesinger GC, Ross RS: Beat-by-beat strokevolume from aortic pulse pressure analysis. IEEE TransBiomed Eng 21: 285, 1974

5. Bourgeois MJ, Gilbert BK, von Bernuth G, Wood EH: Con-tinuous determination of beat-to-beat stroke volume from aor-tic pressure pulses in the dog. Circ Res 39: 15, 1976

6. Herd JA, LeClair NR, Simon W: Arterial pressure pulse con-tours during hemorrhage in anesthetized dogs. J Appi Physiol21: 1864, 1966

7. Alderman EL, Branzil A, Sanders W, Brown BW, HarrisonDC: Evaluation of the pulse-contour method of determiningstroke volume in man. Circulation 46: 546, 1972

8. Rutherford BD, Gardner RM, Warner HR: Comparison ofdyedilution curve and arterial pulse contour analysis for determina-tion of stroke volume in critically ill patients. Chest 62: 376,1972

9. Cibulski AA, Lehan PH, Hellems HK: Pressure methods for

375

by guest on May 12, 2018

http://circ.ahajournals.org/D

ownloaded from

VOL 62, No 2, AUGUST 1980

estimating right and left ventricular stroke volumes. Am JPhysiol 225: 1460, 1973

10. Starmer CF, McHale PA, Cobb FR, Greenfield JC Jr: Evalua-tion of several methods for computing stroke volume from cen-tral aortic pressure. Circ Res 33: 139, 1973

11. Warner HR: Role of computers in medical research. JAMA196: 944, 1966

12. Warner HR, Gardner RM, Toronto AF: Computer basedmonitoring of cardiovascular functions in postoperativepatients. Circulation 37 (suppl II): II-68, 1968

13. Gardner RM, Hoecherl W: Instrumentation for computerized

heart catheterization. IEEE Trans Biomed Eng 18: 60, 197014. Gardner RM, Warner HR, Toronto AF, Gaisford WD:

Catheter-flush system for continuous monitoring of centralarterial pulse wave form. J Appl Physiol 29: 911, 1970

15. Gardner RM, Schwartz R, Wong HC, Burke JP: Percuta-neous indwelling radial artery catheters for monitoring cardio-vascular function. N Engl J Med 290: 1227, 1974

16. Bourgeois MJ, Gilbert BK, Donald DE, Wood EH:Characteristics of aortic diastolic pressure decay with applica-tion to the continuous monitoring of changes in peripheralvascular resistance. Circ Res 35: 56, 1974

Minoxidil: Right Atrial Cardiac Pathologyin Animals and in Man

J. T. SOBOTA, M.D., W. B. MARTIN, M.D., R. G. CARLSON, D.V.M., PH.D.,AND E. S. FEENSTRA, D.V.M., PH.D.

SUMMARY Minoxidil is a clinically effective vasodilator with a low order of toxicity, except for a canine-specific lesion of the right atrium of the heart of dogs after they are given daily doses of 1 mg/kg or greater for30 days. In these experiments the canine right atrial lesion was generally located adjacent to the second of thethree atrial branches of the right coronary artery and had grossly red or yellowish thickening of the atrial wall.Microscopically there was extravasation of a few intact red cells, atrophy of myocardial cells adjacent to thesubepicardium, the presence of mast cells and macrophages containing hemosiderin with subsequent prolifera-tion of angioblasts and connective tissue cells. Cytoplasmic loss in myocardial cells of the left ventricularpapillary muscle identical in all respects to that seen after other hypotensive agents and fi-adrenoceptoragonists was also observed in minoxidil-treated rodents and dogs. Seventy-nine of 158 patients who died secon-dary to severe hypertension but who received minoxidil were autopsied. Six cases had significant right atrialpathology not related to gross and histologic changes found in the right atrial lesions of dogs. A prospectivestudy of the right atrium of normotensive and hypertensive patients not treated with minoxidil shows an in-creasing frequency of histologic changes, such as myocardial fibrosis and hydropic vacuolization that arerelated to age and the presence of hypertension.

MINOXIDIL (Loniten, Upjohn) is a chemicallyunique (fig. 1), orally active vasodilator drug thatprofoundly and consistently reduces mean bloodpressure in rats, miniature pigs, rhesus monkeys, nor-motensive mongrel and beagle and hypertensivemongrel dogs.' Several clinical trials have reconfirmedits efficacy;2 3 a controlled trial by Dargie et al.4showed that minoxidil, in combination with pro-pranolol and diuretics, lowers the blood pressure ofhypertensive patients resistant to treatment with largedoses of standard antihypertensive drugs.The therapeutic importance of chronically ad-

ministered minoxidil in man stimulated comprehen-sive toxicologic studies. In general, the drug has littletoxicity when given intraperitoneally or orally and

From the Upjohn Company, Kalamazoo, Michigan.Presented in part at the Brook Lodge Symposium, August,

Michigan, entitled, "Minoxidil (Loniten): A New Drug for SevereHypertension," November 8, 1979.Address for correspondence: Joseph T. Sobota, M.D., The Up-

john Company, Kalamazoo, Michigan 49001.Received June 16, 1978; revision accepted January 28, 1980.Circulation 62, No. 2, 1980.

causes minimal pharmacologic effects except in the-cardiovascular system. The drug does produce hir-sutism in humans, an annoying side effect, especiallyin women. Teratology studies in pregnant rats andrabbits were negative. Daily doses of 7 mg/kg of bodyweight in monkeys and 30 mg/kg in rats older than 1year caused no significant organ abnormalities.5

Toxicity in canines is also of a low order, but aspecies-specific lesion of the right atrium of the heartoccurs with single daily doses of 1 mg/kg or greatergiven for 1 month. A nonspecific myocardial lesionhas also been described in rodents and dogs brieflytreated with minoxidil, identical to those induced byhydralazine, other hypotensive agents and isopro-terenol.6' 7 This lesion consists of a loss of cytoplasmin myocardial cells in the left ventricular papillarymuscle.The usefulness of minoxidil in hypertensive disease

in man had to be carefully evaluated against the possi-ble occurrence of right atrial lesions similar to those inthe beagle and the nonspecific left ventricular papil-lary muscle change. Little information is available ongross and microscopic right atrial pathology. Infarctsand mural thrombi of the atria in humans have beenstudied by Cushing et al.,8 and by Soderstrom;9

376 CIRCULATION

by guest on May 12, 2018

http://circ.ahajournals.org/D

ownloaded from

R M Cundick, Jr and R M Gardnerdetermination.

Clinical comparison of pressure-pulse and indicator-dilution cardiac output

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1980 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/01.CIR.62.2.371

1980;62:371-376Circulation. 

http://circ.ahajournals.org/content/62/2/371the World Wide Web at:

The online version of this article, along with updated information and services, is located on

  http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer information about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. FurtherEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculationpublished in Requests for permissions to reproduce figures, tables, or portions of articles originallyPermissions:

by guest on May 12, 2018

http://circ.ahajournals.org/D

ownloaded from