The Limited Reliability of Physical Signsfor Estimating Hemodynamicsin Chronic Heart FailureLynne Warner Stevenson, MD, Joseph K. Perloff, MD

The cardiovascular physical examination is used commonly as a basis fordiagnosis and therapy in chronic heart failure, although the relationship betweenphysical signs, increased ventricular filling pressure, and decreased cardiacoutput has not been established for this population. We prospectively comparedphysical signs with hemodynamic measurements in 50 patients with knownchronic heart failure (ejection fraction, .18\m=+-\.06).Rales, edema, and elevatedmean jugular venous pressure were absent in 18 of 43 patients with pulmonarycapillary wedge pressures greater than or equal to 22 mm Hg, for which thecombination of these signs had 58% sensitivity and 100% specificity. Propor-tional pulse pressure correlated well with cardiac index (r=.82), and when lessthan 25% pulse pressure had 91% sensitivity and 83% specificity for a cardiacindex less than 2.2 L/min/m2. In chronic heart failure, reliance on physical signsfor elevated ventricular filling pressure might result in inadequate therapy.Conversely, the adequacy of cardiac output is assessed reliably by pulsepressure. Our results facilitate decisions regarding treatment in chronic heartfailure.

(JAMA 1989;261:884-888)

TIME-HONORED cardiovascularphysical signs of congestive heartfailure include a third heart sound, pul¬monary rales, an abnormal jugular ve¬nous pulse (height and wave form), andperipheral edema.12 These criteria,together with the history and chestroentgenogram, have been useful inidentifying patients with ejection frac¬tions of less than 40% to 50%.3"8 Patientsin whom the diagnosis of heart failure

has been established require constantvigilance to identify hemodynamic dete¬rioration that necessitates adjustmentof therapy. While the relationship be¬tween physical signs and hemodynamicprofile has a firm basis for acute conges¬tive heart failure,9 the chronic state ischaracterized by a host ofcompensatorymechanisms that may cause disparities,such as the absence of rales and periph¬eral edema despite symptomatic eleva¬tion of ventricular filling pressures. Be¬cause the accuracy of physical signs forthe identification of elevated fillingpressure and depressed cardiac outputhas not been established in patientswith chronically depressed left ventric-

ular systolic function, we undertook a

prospective study to correlate physicalsigns with hemodynamic data in 50 con¬secutive patients so diagnosed.PATIENTS AND METHODS

All 50 patients had secure diagnosesof heart failure at the time of electivehemodynamic measurements, whichwere performed as part of an evaluationfor cardiac transplantation (39 patients)or an assessment of the adequacy ofmedical regimens (11 patients). Thirty-seven patients were men, aged 47 ± 15years; 13 were women, aged 39 ±13years. Only patients with heart failureaccompanied by ventricular dilation andejection fractions of less than or equal to.30 (systolic dysfunction) were includedto eliminate those with circulatory con¬

gestion arising from primary restrictionof diastolic filling. Mean ejection frac¬tion was . 18 ± .06. Left ventricular dys¬function resulted from nonischemic di¬lated cardiomyopathy in 36 patients andfrom ischémie heart disease in 14 pa¬tients. The duration of congestivesymptoms ranged from three to 100months (median, 19 months). Medica¬tions being taken at the time of studyincluded digoxin by 42 patients, furose-mide by 42 patients, vasodilators by 28patients, and milrinone by two patients.

Physical examinations were per¬formed in the fluoroscopy suite immedi¬ately before right-sided heart catheter¬ization. Blood pressure was determinedwith a sphygmomanometer and a

From the Division of Cardiology, Department of Medi-cine, UCLA Medical Center, Los Angeles.

Reprints requests to UCLA Medical Center, 10833 LeConte Ave, Rm 47-123 CHS, Los Angeles, CA 90024(Dr Stevenson).

Downloaded From: http://jama.jamanetwork.com/ by a University of Tennessee User on 08/30/2013

stethoscope applied to the right or leftbrachial artery. Care was taken to pre¬vent prolonged cuff inflation to avoidvenous congestion and a spurious de¬crease in pulse pressure.2,3 The diastolicpressure was taken as the level at whichKorotkoff's sounds disappeared. Theproportional pulse pressure was definedas follows:

systolic pressure-

diastolic pressuresystolic pressure

Pulmonary rales were graded from 0to 4. A score of 1 indicates rales over one

quarter ofthe posterior lung fields and 4indicates rales over the entire posteriorlung fields. Central venous pressurewas assessed from the internal and ex¬ternal jugular veins with the patient at30° to 45°. Central venous pressure ele¬vation was graded as 0 if the crests ofneither the internal nor the externaljugular veins were visible above theclavicle in this position and as 4 if thecrests were visible at the earlobe in thisposition and in the full upright position.Peripheral edema was graded as 0 to 4according to the depth of indentation atthe ankle. Third heart sounds were

sought by identifying the left ventricu¬lar impulse in the left lateral decubitusposition and applying the bell of thestethoscope just lightly enough for askin seal.2 Orthopnea (during the previ¬ous week) was graded as 0 to 4, with 0indicating no need for more than one

pillow on a flat bed and 4 indicating atleast one night spent sleeping in a sit¬ting position.

Radiographic information may con¬tribute to the diagnosis of low ejectionfraction and heart failure,3"8 but chestroentgenograms were not included be¬cause this study was designed to deter¬mine the reliability of physical signs forassessment of hemodynamic statuswithout recourse to other diagnosticprocedures.

A #7 F balloon-tipped thermodilutionpulmonary artery catheter was insertedpercutaneously into the right internaljugular vein or the right antecubitalvein, the latter when a sufficiently su¬

pine position was not tolerated. The po¬sition of the catheter tip was confirmedby fluoroscopy, and pressures weremeasured in the right atrium, pulmo¬nary trunk, and pulmonary capillarywedge positions. The right atrial andwedge pressures were recorded on stripcharts in the respiratory midposition.Thermodilution cardiac outputs weredetermined in triplicate with icedinjectate.

Tb establish the maximum sensitivityof the physical signs, all scores greaterthan zero were considered positive. Pul¬monary capillary wedge pressure great-

er than 20 mm Hg and cardiac index lessthan 2.2 L/min/m2 were chosen as con¬servative estimates of hemodynamicabnormalities warranting therapy. Be¬cause no patient had a pulmonary capil¬lary wedge pressure in the range of 19 to21 mm Hg, the categories were dividedinto greater than or equal to 22 mm Hgand less than or equal to 18 mm Hg. Thecorrelations between hemodynamicsand clinical measurements were as¬sessed with linear regression and in2x2 tables, for which significance wasdetermined by two-tailed Fisher's exacttest.

RESULTSThe age, etiology of left ventricular

failure, and physical signs for the 50patients are shown in Table 1. Raleswere present in eight patients, eleva¬tions in mean jugular venous pressureswere present in 25 patients, and periph¬eral edema was present in ten patients;third heart sounds were present in 48patients.

Pulmonary capillary wedge pres¬sures ranged from 8 to 44 mm Hg (Table2) and were greater than or equal to 35mm Hg in 18 (36%) of 50 patients, great¬er than or equal to 25 mm Hg in 38patients (76%), and greater than or

equal to 22 mm Hg in 43 patients (86%).The remaining seven wedge pressureswere 8 to 18 mm Hg. The eight patientswith rales all had pulmonary wedgepressures greater than or equal to 22mm Hg. The presence of rales, howev¬er, identified only 19% of patients withpulmonary wedge pressures greaterthan or equal to 22 mm Hg and 11% ofpatients with pulmonary wedge pres¬sures greater than or equal to 35 mm

Hg. Third heart sounds were present inall but two patients (wedge pressures, 8and 26 mm Hg) and therefore were notspecific for identifying high filling pres¬sures in our study group.

Right atrial pressures were greaterthan or equal to 10 mm Hg in 28 pa¬tients, all of whom had pulmonarywedge pressures greater than or equalto 22 mm Hg. Elevated right-sided fill¬ing pressures could be diagnosed fromeither the jugular venous pressure orfrom edema or from both in 21 (75%) of28 patients with right atrial pressuresgreater than or equal to 10 mm Hg. Allten patients with edema also had abnor¬mally high jugular venous pressures.

Because elevated left ventricular fill¬ing pressures were less evident clini¬cally than the frequently coexisting ele¬vations of right ventricular fillingpressures, the reliability of the physicalexamination for detection of pulmonarycapillary wedge pressure greater thanor equal to 22 mm Hg was assessed by

Table 1.—Clinical Characteristics and PhysicalFindings in 50 Patients With Chronic Heart Failure

Characteristic/Finding Present AbsentCoronary artery disease 16 34Male gender 37 13Rales 8 42Third heart sound 48 2Increased jugular venous

pressure 25 25Peripheral edema 10 40Orthopnea 39 11

Table 2.—Hemodynamic Measurements in 50Patients With Chronic Heart Failure

HemodynamicMeasurement_Mean ± SD RangeEjection fraction .18±.06 .08-.29Heart rate,

beats per min 98±18 64-130Systolic blood pressure,

mmHg 109 + 18 80-150Diastolic blood pressure,

mmHg 82±11 50-100Mean arterial pressure,

mmHg 91±11 60-113

Right atrial pressure,mmHg 12±7 2-38

Pulmonary wedge pressure,mm Hg 30 ± 9 8-44

Cardiac index,L/min/m2 2.1 ±0.8 1.0-4.7

Stroke volume index,cc/m2 22 ±9 10-43

Systemic vascular resistance,dynes-sec-cm_5 1900 + 800 760-4500

using jugular venous pressure and ede¬ma as criteria in addition to pulmonaryrales (Table 3). Elevated jugular venous

pressure was the most sensitive criteri¬on and was present in every patient whohad other physical signs of congestion.However, even using all these criteria,elevated filling pressures could not bediagnosed in 18 (42%) of43 patients withpulmonary capillary wedge pressuresgreater than or equal to 22 mm Hg.Physical evidence specific for conges¬tion was absent in eight (44%) of 18 pa¬tients with pulmonary capillary wedgepressures greater than or equal to 35mmHg.

To assess tolerance for the supineposition during catheterization, allpatients were questioned regardingorthopnea. Orthopnea within the pre¬ceding week was reported by 39 (91%) of43 patients with pulmonary wedge pres¬sures greater than or equal to 22 mm

Hg. Recent orthopnea was absent infour patients with pulmonary wedgepressures greater than or equal to 22mm Hg and in all seven patients withlower pressures.

The proportional pulse pressure cor¬related well with cardiac index (r = .82,P<.001) (Figure), with stroke volume

Downloaded From: http://jama.jamanetwork.com/ by a University of Tennessee User on 08/30/2013

)

Table 3.—Relation Between Physical Examinationand Elevated Pulmonary Capillary Wedge Pressure

Physical Examination

Pulmonary CapillaryWedge Pressure, mm Hg Positive* Negative

s22t 25 18S18 0 7

P<.01

'Evidence of elevated filling pressures with pul¬monary capillary wedge pressure greater than or equalto 22 mm Hg included rales alone, no patients; rales andelevated venous pressure, two patients; rales, elevatedvenous pressure, and peripheral edema, six patients;elevated venous pressure and edema, four patients;elevated venous pressure alone, 13 patients; andedema alone, no patients.

fThe criterion of 22 mm Hg was chosen because fivepatients had pulmonary capillary wedge pressuresranging from 22 to 24 mm Hg and the next highest valuewas 18 mmHg.

index (r=.78, P<.001), and with theinverse of systemic vascular resistance(r = .65, P<.001). There was poor corre¬

lation, however, between cardiac indexor stroke volume index and systolicpressure (r= .36 and r = .45, respective¬ly) or mean arterial pressure (r = .05 andr=.08, respectively). A proportionalpulse pressure less than or equal to 25%identified 29 (91%) of 32 patients withcardiac indexes less than 2.2 L/min/m2(Table 4). The proportional pulse pres¬sure was greater than 25% in 15 (83%) of18 patients with cardiac indexes greaterthan 2.2 L/min/m2. If a cardiac index ofless than or equal to 2.0 L/min/m2 was

selected, the results were 93% and 80%.The use of proportional pulse pressureyielded an 88% predictive accuracy forlow cardiac output.

Linear regression showed no correla¬tion between the amount of pulmonarycapillary wedge pressure elevation andthe amount of cardiac index reduction(r=.22, not significant). However, theseven patients with pulmonary capil¬lary wedge pressures less than or equalto 18 mm Hg all had cardiac indexesgreater than 2.2 L/min/m2 (Table 5), asdid 11 (26%) of 43 patients with highpulmonary capillary wedge pressures.Right atrial pressure correlated withcardiac index (r=.53, P<.01), and 25(89%) of 28 patients with right atrialpressures greater than or equal to 10mm Hg had cardiac indexes less than 2.2L/min/m2, as did seven (32%) of 22 pa¬tients with right atrial pressures lessthanl0mmHg(P<.001).COMMENT

This prospective study comparing thecardiovascular physical examination tohemodynamic measurements in pa¬tients with known chronic heart failure(characterized by systolic dysfunctionand ventricular dilation) showed thatleft ventricular filling pressure fre-

60-

40-

0)Q.

0.«

gtaoQ-O

PPP, =£25%20-

" i ^1

CI, «2.2

Cardiac Index, L/min/m2

Relationship between cardiac index (CI) and proportional pulse pressure (PPP) in 50 patients with chronicheart failure.

quently exceeded 22 to 25 mm Hg de¬spite the absence of specific evidence inthe physical signs. However, cardiac in¬dexes less than or equal to 2.2 L/min/m2were predicted reliably by the propor¬tional pulse pressure measured at thebedside.

While other studies have reportedthe value of clinical information in diag¬nosing left ventricular dysfunction,3"8the purpose of this study was not todistinguish normal from abnormal ven¬tricular function but to assess hemody¬namic status among patients known tohave low left ventricular ejection frac¬tions. The primary hemodynamic ab¬normalities caused by left ventricularsystolic dysfunction are elevation of leftventricular filling pressures and de¬pression of cardiac output, which do notnecessarily occur together. The reliabil¬ity of the physical signs for identifyingmarked to severe elevations of ventric¬ular filling pressure and reductions incardiac output has not been establishedpreviously in this population, for whominferences drawn from physical signsare important for optimal therapy withvasodilators and diuretics, especially inthe outpatient setting.

A sudden elevation of pulmonary ve¬nous pressure causes rales due to ex¬travasation of fluid into the alveoli,10 butchronic exudation of fluid is associatedwith an increase in lymphatic drainageso that the alveoli remain relatively dryand rales are absent. In our patients in

Table 4. —Relation Between Proportional PulsePressure and Cardiac Index in 50 Patients WithChronic Heart Failure

Proportional PulsePressure, %

Cardiac Index, ,- -.

L/mln/m2_«25_>25_«2.2 29 3*>2.2 3 15

•Three patients with cardiac index of 2.0 to 2.17L/min/m2; P<.0001.

Table 5.—Hemodynamic Status for 50 PatientsWith Chronic Heart Failure

Cardiac Index, L/min/m2Pulmonary Capillary '

Wedge Pressure, mm Hg* «2.2_>2.2a22 32 11«18 0 7

*No patients with pressure of 19 to 21 mm Hg;P<.001.

whom symptomatic heart failure hadbeen present for at least three months,pulmonary rales were rarely present,even when the pulmonary capillarywedge pressures were greater than or

equal to 35 mm Hg. However, accumu¬lation of fluid in the alveolar-capillaryinterstitium leads to decreased lungcompliance and dyspnea,11 which alsocan result from an increase in physiolog¬ical dead space due to ventilation-perfu¬sion mismatching.12 Clinical recognitionof this chronic stage of pulmonary ede-

Downloaded From: http://jama.jamanetwork.com/ by a University of Tennessee User on 08/30/2013

ma is difficult13 and should not depend onthe presence of rales.

Third heart sounds have been consid¬ered reliable signs of heart failure andelevated ventricular filling pressure2"8and are helpful for the initial diagnosisof ventricular failure. Nevertheless, inour patients already known to have se¬

verely reduced ejection fractions, thirdheart sounds were too common to bespecific for identifying major elevationsof left ventricular filling pressure.

Physical signs of high right ventricu¬lar filling pressure usually were evidentwhen the right atrial mean pressure was

greater than 10 mm Hg. Variations inloading conditions of the right ventricleare reliably transmitted directly intothe systemic venous system, which canbe assessed visually in the jugularpulse.2 Elevated right atrial pressure,in this setting of left ventricular failure,was always accompanied by elevatedleft ventricular filling pressure, al¬though the converse was not true in thisand previous studies.914 Elevated jugu¬lar venous pressure and peripheral ede¬ma, therefore, were specific but not sen¬sitive for high left ventricular fillingpressure. The correlation between highright ventricular filling pressure andlow cardiac output may reflect the effectofpoor left ventricular function on rightventricular performance, impairment ofleft ventricular filling by right ventricu¬lar distention, or the importance ofrightventricular function as a determinant ofcardiac output, at least during physicalexercise.1616

Exertional dyspnea may result morefrom inadequate tissue oxygen deliverythan from pulmonary venous conges¬tion,17 but orthopnea results from posi¬tional volume changes that increase leftventricular filling pressure and extra-vascular pulmonary fluid.18 In our

study, a history of recent orthopnea wasthe most reliable clinical indicator ofhigh left ventricular filling pressure.

Cuff blood pressure has been used inthe diagnosis of a low ejection fraction4but previously has not been evaluatedspecifically for assessment of cardiacoutput in patients with chronically lowejection fractions. In our patients, nei¬ther the systolic blood pressure nor themean arterial pressure alone predictedcardiac index or stroke volume index,which were predicted by the propor¬tional pulse pressure. The pulse pres¬sure reflects the complex interplay be¬tween stroke volume, heart rate, aorticdistensibility, and peripheral vasculartone and their relative contributions toincident and reflected waves.19"21 Whilemeasurement of cardiac output frompulse pressure was first suggested in1904,22 even the introduction of multiple

adjustment factors did not allow precisequantitation of cardiac output frompulse pressures for a varied popula-tion.28"25

In this study of simple physical as¬sessment of hemodynamic status, pulsepressure was used primarily to identifythe presence of severely reduced cardi¬ac index rather than to estimate an ex¬act value. As many patients with severeheart failure achieve their best cardiacoutputs at systolic blood pressures be¬low 100 mm Hg,26 pulse pressure as a

proportion of total systolic pressure was

analyzed. A proportional pulse pressureless than or equal to 25% identified 91%of patients with cardiac indexes lessthan or equal to 2.2 L/min/m2, while ahigher pulse pressure identified 83% ofpatients with higher cardiac indexes.For comparison, proportional pulsepressures derived from blood pressurerecordings in 42 patients with non-cardiac disease (without hypertension)and Fick cardiac outputs of 6.3 ±0.4L/min/m2 were 39% ± 8%, with only onevalue less than or equal to 25% in an

early study on cardiac output measure¬ment.25 It should be emphasized that thecurrent study validates proportionalpulse pressure only for patients withknown chronic dilated left ventricularfailure (ejection fraction, .18±.06) andmay not be applicable to patients withother cardiac diseases, includinghypertension.

The role of the physical examinationin the management of chronic heart fail¬ure is coupled closely to the hemody¬namic goals of therapy. Pulmonary cap¬illary wedge pressure usually can belowered to near-normal levels by di¬uretic and vasodilator therapy in thechronically dilated ventricle withoutcompromising cardiac output.26 Suchtherapy not only improves symptoms ofcongestion but also reduces the left ven¬tricular dilation and mitral régurgita¬tion that increase ventricular work andfurther contribute to decompensation.27A cardiac index of 2.0 to 2.2 L/min/m2represents a reasonable estimate of theminimum necessary for adequate organperfusion with normal hemoglobin leveland resting oxygen consumption.28 Theefficacy of many vasodilator agents forincreasing lower cardiac outputs hasbeen shown.29,30 Pulmonary wedge pres¬sure less than 18 mm Hg and cardiacindex greater than 2.2 L/min/m2, there¬fore, represent conservative hemody¬namic goals for this patient population.

The four hemodynamic profiles de¬fined for acute myocardial infarction9were adapted to our patients withchronic heart failure. Although highpulmonary capillary wedge pressuregenerally accompanies low cardiac in-

dex, the converse is not always true.There were seven patients with the firstprofile, cardiac index greater than 2.2L/min/m2 and pulmonary wedge pres¬sure less than or equal to 18 mm Hg,consistent with adequate therapy. Pul¬monary capillary wedge pressure was

high without low cardiac index in 11patients (second profile). Unlike pa¬tients with acute infarction,9 none ofour

patients with chronic heart failure pre¬sented with low filling pressure and lowcardiac index (third profile). Pulmonarycapillary wedge pressure was high withlow cardiac index in 32 patients (fourthprofile). Despite the use of digoxin,diuretics, and vasodilators in most pa¬tients prior to evaluation, 43 (86%) of 50patients referred thus had hemodynam¬ic profiles with targets for further ther¬apy with diuretics, vasodilators, orboth.

Physical evidence of pulmonary con¬

gestion (rales) was specific for profile 2or 4, but rare. Detection of elevatedmean jugular venous pressure wasmore common, but still not very sensi¬tive. Narrow proportional pulse pres¬sure indicated profile 4 (or 3). When thehemodynamic profile and targets fortherapy can be identified from the phys¬ical examination, therapy often can beadjusted effectively in the outpatientsetting, without invasive quantitationof left ventricular filling pressures orcardiac output. However, the frequen¬cy of undetected high left ventricularfilling pressures, with or without lowcardiac index, indicates that patientswith low ejection fractions who describedyspnea and particularly orthopneashould not be doubted merely becausespecific physical signs ofpulmonary andsystemic venous congestion are absent.These patients may require invasivestudy to detect and treat severely ele¬vated left ventricular filling pressures.In addition, patients in whom empirictherapy based on pulse pressure, ve¬nous pressure, or orthopnea is limitedby hypotension or declining renal func¬tion may benefit from insertion of an

indwelling pulmonary arterial catheterto facilitate more precise design oftreatment. Initial administration of in¬travenous diuretics and vasodilatorstailored to reduce pulmonary capillarywedge pressure and systemic vascularresistance often allows the subsequentdesign of effective oral therapy in pa¬tients previously considered clinicallyrefractory.31

By prospective comparison of cardio¬vascular physical signs with hemody¬namic measurements in patients withchronic congestive heart failure from di¬lated cardiomyopathy or end-stage cor¬

onary artery disease, our study has

Downloaded From: http://jama.jamanetwork.com/ by a University of Tennessee User on 08/30/2013

shown that marked to severe elevationsof left ventricular filling pressure fre¬quently are undetected, while the ade¬quacy of resting cardiac output is as¬sessed reliably by the proportionalpulse pressure. These observations al¬low more effective adjustment of outpa¬tient medical therapy for chronic heartfailure and more judicious use of inva¬sive hemodynamic monitoring in thispopulation.

This study was supported in part by the Strei¬sand/American Heart Association Endowment andwas carried out during Dr Stevenson's tenure as aclinician-scientist of the American Heart Associa¬tion, Greater Los Angeles Affiliate.

We wish to express appreciation to the staff ofthe UCLA 4 West Special Procedures Room, su¬

pervised by Richard Pandolfi, and to thank DianeMiller and Michelle Moravec for secretarialassistance.

References

1. Mattingly TW: Clinical and hemodynamic fea-tures of primary myocardial disease. Trans AmClin Clim Assoc 1958;70:132-139.2. Perloff JK: The arterial pulse, in Physical Ex-amination of the Heart and Circulation. Philadel-phia, WB Saunders Co, 1982, p 50.3. Mattleman SJ, Hakki A-H, Abdulmassih AS, etal: Reliability of bedside evaluation in determiningleft ventricular function: Correlation with left ven-tricular ejection fraction determined by radionu-clide ventriculography. J Am Coll Cardiol 1983;1:417-420.4. Cease KB, Nicklas JM: Prediction of left ven-tricular ejection fraction using simple quantitativeclinical information. Am J Med 1986;81:429-439.5. Echeverria HH, Bilsker MS, Myerburg RJ, etal: Congestive heart failure: Echocardiographic in-sights. Am J Med 1983;75:750-755.

6. Eagle KA, Quertermous T, Singer DE, et al:Left ventricular ejection fraction: Physician esti-mates compared with gated blood pool scan mea-surements. Arch Intern Med 1988;148:882-885.7. Marantz PR, Tobin JN, Wassertheil-Smolker S,et al: Relationship between left ventricular systolicfunction and congestive heart failure diagnosed byclinical criteria. Circulation 1988;77:607-612.8. Harian WR, Oberman A, Grimm R, et al: Chron-ic congestive heart failure in coronary artery dis-ease: Clinical criteria. Ann Intern Med 1987;86:133\x=req-\138.9. Forrester JS, Diamond G, Chatterjee K, et al:Medical therapy of acute myocardial infarction byapplication of hemodynamic subsets. N Engl J Med1976;295:1356-1362.10. Staub NC, Nagano H, Pearce ML: Pulmonaryedema in dogs, especially the sequence offluid accu-mulation in lungs. J Appl Physiol 1967;22:227-240.11. Straub NC: Pulmonary edema. Physiol Rev1974;54:679-700.12. Sullivan MJ, Higginbotham MB, Cobb FR: In-creased exercise ventilation in patients with chron-ic heart failure: Intact control despite hemodynam-ic and pulmonary abnormalities. Circulation1988;77:552-559.13. Szidon JP, Pietra GG, Fishman AP: The alveo-lar-capillary membrane and pulmonary edema. NEngl J Med 1972;286:1200-1204.14. Mangano DT: Monitoring pulmonary arterialpressure in coronary-artery disease. Anesthesiolo-gy 1980;53:364-370.15. Baker BJ, Wilen MW, Boyd CM, et al: Relationof right ventricular ejection fraction to exercisecapacity in chronic left ventricular failure. Am JCardiol 1984;54:596-599.16. Barber D, DiSessa T, Child JS, et al: Hemody-namic responses to isolated increments in heartrate by atrial pacing after a Fontan procedure. AmHeart J 1988;115:837-841.17. Fink LI, Wilson JR, Ferraro N: Exercise ven-tilation and pulmonary artery wedge pressure inchronic stable congestive heart failure. Am J Car-diol 1986;57:249-253.18. Perera GA, Berliner RW: The relation of pos-tural hemodilution to paroxysmal dyspnea. J Clin

Invest 1943;22:25-28.19. O'Rourke MF: The arterial pulse in health anddisease. Am Heart J 1971;82:687-702.20. Pepine CJ, Nichols WW, Conti R: Aortic inputimpedance in heart failure. Circulation1978;58:460-465.21. Abrams J: Arterial pulse and pressure. Cardio-vasc Rev Rep 1985;6:1055-1073.22. Erlanger J, Hooker DR: Studies in blood pres-sure. Johns Hopkins Hosp Bull 1904;15:179-180.23. Remington JW, Noback CR, Hamilton WF, etal: Volume elasticity characteristics of the humanaorta and prediction of the stroke volume from thepressure pulse. Am J Physiol 1948;153:298-308.24. Brotmacher L: Evaluation of derivation of car-diac output from blood pressure measurements.Circ Res 1957;5:589-593.25. Starr I: Clinical tests of the simple method ofestimating cardiac stroke volume from blood pres-sure and age. Circulation 1954;9:664-681.26. Stevenson LW, Tillisch JH: Maintenance ofcardiac output with normal filling pressures in pa-tients with dilated heart failure. Circulation1986;74:1303-1308.27. Stevenson LW, Bellil D, Grover-McKay M, etal: Effects of afterload reduction on left ventricularvolume and mitral regurgitation in severe conges-tive heart failure secondary to ischemic or idio-pathic dilated cardiomyopathy. Am J Cardiol1987;60:654-658.28. Grossman W, Braunwald E: High cardiac out-put states, in Braunwald E (ed): Heart Disease.Philadelphia, WB Saunders Co, 1984, p 807.29. Cohn JN, Franciosa JA: Vasodilator therapy ofcardiac failure. N Engl J Med 1977;297:27-31, 254\x=req-\258.30. Chatterjee K, Parmely WW: Vasodilator ther-apy for acute myocardial infarction and chronic con-

gestive heart failure. J Am Coll Cardiol 1983;1:133-153.31. Stevenson LW, Dracup KA, Tillisch JH: Effi-cacy of medical therapy tailored for severe conges-tive heart failure in patients transferred for urgentcardiac transplantation. Am J Cardiol, in press.

Downloaded From: http://jama.jamanetwork.com/ by a University of Tennessee User on 08/30/2013