CA1327631C - System for non-invasive detection of changes of cardiac volumes and aortic pulses - Google Patents

Abstract

ABSTRACT

A method and an apparatus therefor for monitoring cardiac function in an animal or human subject including the steps of: placing a first movement detecting transducer on the torso, said transducer overlying at least part of two diametrically opposed borders of the heart or great vessels:
generating a signal indicative of the movement of the torso portion subtended by the transducer, said signal including a cardiac component comprising at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform: and assessing cardiac function by monitoring changes in said ventricular volume waveform or said aortic pressure pulse waveform.

Description

13276~1 lo~ ~

BACI~GRO~ND OF ~S INV~i~ION

Field o~ the Invention This invention pertains to non-invasive monitors, and more particularly to non-invasive monitoring of cardiac function.

Prior Ar~ ¦
Al~hough the electrocardiogram ~EXG) has been the primary non-invasive device for continuously monitoring activity of the heart lo in clinical medicine, it reflects solely electrical activation of cardiac muscle and provides no information on the mechanical characteristics of ~he cardiac pump. Consequently, the EKG may show normal or near normal waveforms in the presence of greatly i~paired blood pumping capacity of the heart. Conversely, the EXG
~avefora ~ay bQ abnormal desp~t~ norm~l or near normal pumpinq action. In terEs of lifo support, adequate circulation of blood froo the heart to the tissue8, as reflectQd by the blood pumping capacity of the heart, is of paramount i~portance.

Ob~iously, non-inva~ive techniques for monitoring the blood pu~plng capacity of the heart are preferrad over invasive ones.
Nevertheless, invasive cardia¢ ~onitoring techniqu~s, because of their per¢eived greater ac¢ura¢y and ability to provide continuous ~onitoring, ¢ontinue ~o be enployed in, for example, critically 47\S300~A.06 ~
O,lg89 ~ I

ill patients. Invasive techniques generally have as their basis a catheter, such as a Swan-Ganz catheter, placed ~uch that its tip lays within the pulmonary artery. This provides contlnuous recording of pressures in the pulmonary artery, and in certai~
instances pressures in the right ventricle, right atrium and indirectly the le~ atrium (pulmonary capillary wedge pressure).
In~ection of inert dye or cold saline fr~m the catheter allows discrete measurements of cardiac output by dye dilution method or thermoailution, respectively. Alternatively, sampling blood for oxygen content in the pulmonary artery and a system~c artery together with ~easurement of oxygen consumption permits calculation of cardiac output by the F~c~ principle.

However, insertion of a cardiac catheter into the body may be ha~ardous. Its usa can lead to death, which occurs in 1% of cases, and ~orbidity, which occurs in 33% of cases, as a result of inf~c~ion and~or damage to thQ hQart valves, cardiac arrhythmias, and pul~onary thromboemboli~. Errors of technique, measurement, ~udgmQnt ~nd interpretation are common. It has been estimated that on~-~alf ~illion Swan-Gan~ cathet~rs ussd in the United States in 1986 resultQd in tho deat~ of ~8 many as 1000 or more patients.
Further~or~, cardiac c~tbeters cannot be kept in place for more than ~ fe~ days owing to hazards fro~ infection. m ey ar~ also oostly and labor intensivQ since cathstsrized pat$ents reguire 2S intensiv~ car~ unit$ ~hich cost two to fivQ times more than .

~7\53000A.06 ~DRCH 20, 1989 3 standard semi-private beds. In addition, health care worXers face the risk of AIDS acquired vlrus and hepatitis virus as a result of exposure to blood of the infected patient during catheter introduction and subsequent maintenance.

Noreover, cardiac catheters do not directly provide measurement of change ~n ventricular volume. While such measurements can be indirectly obtained in con~unction with in~ection of radiopaque dye and roentgenographic imaging, this technique is time-consuming and co~tly, and dangerous hypotension and bradycardia ~ay be induced by the dye. Further, the number o~
studies in a given patient is limited by the hazards of x-ray exposure and radiopaque dya in~ection~.

Angiograpbic techniques provide tha most widely accepted mQans for ~easuring ventricular volumQs. They allow cslculation of the extent ~nd velocity of wall shortening ~nd of regional a~norealities of w~ll ~otion. When they ~re combined with ~e~ure~ent of pressure, both ventricular compliance and afterload 20 ~ ., t~e force~ acting ~ithin thQ wall that oppose shortening) c~n be de~Qr-ined. When the rQsults ara expressed in units oorrectad for ~u~cl~ lengtb or circumferancas of the ventricle, co~pari80n8 can be madQ betweQn individuals with widely dif~ering hcart si~es.
2S ` -47\53000A.06 ~ARCH 20, 1989 4 ~ S;~ S~.~~-~r,~

Cineangiography provides a large nu~ber of sequential observations per unit of time, typically 30 to 60 frames per second. Although contrast mat~rial can be in~ected into the pulmonary artery and left atrium, the left ventricle is outlined more clearly when aye is directly in~ected into the ventricular cavity. Therefore, the latter approach is used in most patients, except in those with severe aortic regurgitat~on in whom the contrast material may be in~ected into the aorta, with the resultant reflux of contrast material outlining the left ventricular cavity.

.
In~ection of a contrast agent does not producQ hemodynamic changes (except for premature beats~ until approximately the sixth beat after in~ection. T~e hyperosmolarity produced by the contrast aq~nt increases the blood volume, which begins to raise preload and heart rate uithin 30 seconds of the in~ection, an effQct that may per~ist for as long as two ~ours. Therefore, this technique cannot be utilised for repetitive mQasurements within a short time span.
Further, contrast agents also depress contractility directly, though newer nonionic agen~s have bQen fo~nd useful for minimizing tbese adversc effects.

I~ calculating ventricular volumes or dimensions from angiogra~s, it i8 essential to take into account and apply appropriate correction factors for magnification as well as ~ .
~7\53000A.06 ~ARCH 20, 1989 5 s"~ "~
'`..``'.;''.'.."'''`".,'''`'',' `'"~'.''`~'''''''` ~ ;'' ' ''""

distortion produced by nonparallel x-ray beams. In order to app~y these correction factors, care must be taXen to determin~
accurately the tube-to-patient and tube-to-film distance~.
Correction is best accomplished by filming a calibrated grid at S the position of ~he ventricla. Thus, angiographic methods do not have wide clinical application owing to their complexity, safety cons~derations, invasiveness, and side effectæ o~ the contrast agents.

The importance of measuring changes of ventricular volume was well expressed by Davila in a symposium on measurement of left ventricular volume. He pointed ou~ that the description of the functional mechanics of thQ left ventricl~ reguires measurement of force, strain and velocity ~ratc of strain). Pr~ssure, a standard measure~ent in cardiac catheteri8ation laboratories, critical care un$~ and operating rooms, is not n~cessarily dependent on shape (geo~atry) or size ~volum~) of th~ ventricle. Howev~r, force and strain ~ust bQ expressed in r~lation to geometry and size of the fluid container.
In tho same symposiu~, Chap~an et al described a cin~ang$ograp~ic method for measuring ventricular volume. These ~or~rs also took into account the shortcomings of their method and ~ade thQ followin~ obs~rvations: "The ideal system for following change in ventricular volum~ is obviously one which is ~7\5300QA.06 ~ARCR 20, 1989 6 ,: .

` ~
1 321~3 1 fully applicabl~ to the free-living organism, which requires no in~ection of any sort, and which can be used repeatedly over long periods of ti~e without danger or discomfort to tha 6ub~ect. Such a system, if it ever becomes available, can hardly be based on roentgenologic principles. But until some entirely different principle emerges and is applied, the roentqenologic principle is indispensable. n A further requirement for an ideal system would be a minimum of physician or technician time for utilizing such technology and interpreting the results.

Because of the obvious advantage of non-invasive techniques over invasive ones, a continuing search has been made for xeliable non-invasive ~ethods of assessing cardiac performance. Such ~ethods are needed particularly in detecting serial changes in ~ardiac function and in ~valuating both acute and chronic effects of interventions such as drug therapy and cardiac operations. The five principal non-invasive ~ethods for assessing cardiac perfor~ance are: systolic time intQrvals, M-mode and two-di ensional e~h w ardiography, radionuclide angiography, gated co~puterized tomography tCT scanning), and gated magnetic resonance ~aging (NRI). All but tbe first of thesQ are ~lternatives to ~ngiography for measurement of ventri~ular volumes and/or di~ensions and therQfore per~it the non-invasiv~ estimation of e~ection phase indices. Other than in patients with obstruct~on to left ventricular outflo~, wall stress (afterload) can be ~7\53000A.06 ~ARCH 20, 1989 7 3 ~ A

estimated from a combination of systemic arterial pressur~, ; ventricular radius, and wall thickness. All four non-invas~ve imaglng methods allow est~mation of ventr~cular systolic and diastolic volumes; none, however, is satisfactory for continuous s or near-continuous ~onitoring of critically ill patients.

Systolic time intervals have been usually obtained with the combinat~on of an external transducer on the carotid artery in the neck to display its pulsations, a microphon~ over the heart to record heart sounds, and the electrocardiogram. T~is technique has ne~er enjoyed wide popularity because of both technical and p~ysiologic reasons: (1) reliable, reproducible recordings are difficult to obtain, (2) prominent internal ~ugular venous pulsations in the horizontal body posture ~ay be superimposed on the carotid artery pulsations rendering interpretation of the c~rotid arterial ~aveform difficult, (3) accurate recording of ~eart sounds ~ay be difficult to obtain particularly in pat~ents ~ith obesity or emphysema, (~) 8y8tolic ti~e intervals are sens~ti~e to many pharmacologic and hemodynamic influences including changes in left ventricular preload and afterload which ~ay introduce misleading values, (5) changes in duration of systolic time interv~ls can be influenced by pat~ent posture and ti~e o~ day ~hen recordings are mad~, (6) carotid puls~ contours to calculate systolic time intervals can be difficult to interpret in patients ~ith aortic valve disease, and l7) presence of .
~7\5300QA.06 ~U~RCH 20, 1989 8 1 ~2763 1 congestive heart fa~ lure can either normalize abnormal value~ or make normal values abnormal.

Echocardiography involves ultrasonic imaging of ventricular s wall motion to monitor cardiac function. With this technigue, the dynamics of ventricular wall contraction and the internal dimensions of the cardiac chambers can be recorded. The apparatuses used for echocardiography encompass a wide variety of increasingly sophisticated and computer-aided imaging and analysis systems. The transducer placements on the chest require the services of skilled technicians and incorrect placements lead to ~isleading information. Furthermore, theso systems are quite expensive, not readily porta~le, reguir~ that the patient be studied in the left lateral dQcubitus posture, and ar~ not intended for continuous monitoring of critically ill patients throughout the day or during exercise.

In addition to thQ forQgoing drawback8, Qchocardiography has s~vQral inherent limitations. For ex~mple, ~11 ultrasonie beams havQ a defined breadth and height compara~le to the size of the cry~tal transduc OE face. B~yond its focal point, t~e beam's cross-sQctional arQa enlargQs in dirQct proportion to th~ distance from thQ tr~nsducQr face. Therefore, in N-mode (single transducer) chocardiography, two laterally separated structures may appear in 2S d~rect anteroposterior r~lationship.

~7\S~OOOA.06 -MARCH 20, 1989 9 ... , :, . . . . .. .

Two-d~'mensional electrocardiographiC techniques also producf~tt distortions, whic~ increase with increasing d~stance between the target and the central beam axis. In these instru~fents, axial S resolution (1-2 m~t) is superior to lateral resolution (4-s mm3f.
Because of the complex nature by which two-dimensional images are generated, artifacts may appear as intracardiac masses to the casual observer. ~urther, delineation of the endocardiu~t of the lef~ ventricle in its entirety is achieved only 70 to 80~ of the time. Also, respiratory interference limits the ability to obtain continuous beat to beat recordings, particularly during exercise.

Attempts have also been made to defftermine left ventricular end-diastolic and end-systolic volumes from dimensions derived from echocardiograp~y. These hawQ met with variablf~ffsuccess, depending on the patient population studied and whether N-mode or two di~fen~tional echo techniqueffs were employed. M-mode dimfensionst are U8'~d ~o calculats! left ventricular volu~Q through an application Or the angiographic concept of thQ left ventricle as an ellipsoid. t Ha~effver, N-~od~f~ echocardiograp~ffy allows measurement of only one ~ `
Iffafft ~entricular dimension, the septalposterolateral dimension, ~hich is viewed at t~feff level o~ the chordaef tendineae. j ~-~
Consefguently, to calculate volu~fe from this ~ingle dimension, the following ~fssu~Fftions are madsf: ~1) thfeff ventricle being examined ~5 do-- iD f~ct approxlnat~ th g ometry of an ellip-oid, both ln ~7\53000A.06 ~ARCH 20, 1989 10 1 ~
:

diastole and systole: ~2) the septal-posterolateral dimension measured coincides with the minor axis of the ellipsoid; (3) the orthogonal minor axis i8 equal to the measured minor axis: and (4) the major axis is twice the length of the minor axes. While good s correlations between angiographic and echo left ventricular volumeshave been obtained, correlations are poor in patients who have asynergetic v~ntricular wall motion, which occurs in patients with coronary artery disease in whom damaged areas of the le~t ventricular wall do not move in phase with the normal areas. Also, because ventricular volume curves as a function of time cannot be derived without utilisation o~ several assumptions and approximations, they are not usually reported.

Two-dimensional echocardiography offers considerable advantage for ~sti~ation of left ventricular volume because it allows direct ~easure ent of all three hemiaxes on the ~llipsoid model and also allows application of other volum~ formulations, sucb aæ Simpson's rule~ Studies have shown that correlations between echocardiographic and an~iographic volu~es are substantially ~provod whQn two-dimen~ional ~ethods are u~ed, and good corr~lations have been obta~ned Qven in the presence of ventr~cular asynergy. The greatest disadvantage to quantitative two-dimensional echocardiography is the inability to obtain technically satisfactory i~ages in all patients and the labor involved in analy~ing the studies. This technique, as with the M-mode, does 47\53000A.06 MARCH 20, ~989 11 1 327~1 not readily provide dynamic changes of ventricular volume over time.

Echocardiography has also been employed to est~mate the velocity of ventricular circumferential fiber shortening (Vcf).
This echo maasurement is analogous to the derivative of change ~n ventr~cular volume during systole and ~erves as a measure of ventricular contractility. Its application in M-mode echocardiography a~sumes that the left ventricular internal dimension is measured at the midventricular level. The mean rate of shortening is deter~ned by dividing the calculated circumference expression by the left ventricular e~ection time (~T), which aay be measured from the concomitant carotid pulse tracing or from tha time duration of echocardiographic aortic valve opening. Peak Vcf can be si~ilarly derived by extrapolation from ~he ~axi~um systolic slopa of poæterior and septal walls. Vcf is inaccurata in patients ~ith asynergetic movement of t~e left vantricle ~s in patients ~ith ischQmic heart disease.

~ean velocity of circuuferential fiber shortQninq (Vcf) can be detarein~d simply fro~ ~e~surements of end-diastoli¢ and end-systolic di~ensions by e¢hoc~rdiogr~phy, CT scanning, or NRI.
Since the ~entricle i8 ~pproximately circular at its minor axis the circu~ference i8 egual to diameter (~). Mean Vc~ tin circu~ferQnce/sec) i8 therefore the difference between end-47\53000A.06 HARCH 20, 1989 ~2 diastolic and end-systolic circumference ~n cm) divided by th~
product of the duration of ejection (in sec) and the end-dia~tol~c circumference. Values of vc~ obta~ned by echocardlography compare closely with those determined from c~neangiograms.

Echocardiography has also been employed to estimate stroke volume (SV), which is the difference ~etween end-dia~tolic volume and end-systolic volume. This technique suffers from th~ inherent lack of accuracy in volume estimations and, clinically, stroke volume varies widely with different physiologic c~rcumstances such as body size, heart rate, posture and exercise. It is, therefore, not as useful a measurement as contractility. Nevertheless, provided that sub~ects wit~ l~ft ventricular asynergy are excludQd from analysis, fair correlations have bQen reported between stroke volume derived from M-mod~ echocardiographic and two dimensional ec~o techniques on the one hand, and both thermodilution and angiographic stroke volume measurements on th~ other.

AnothQr non-invasivQ techniquQ i8 the apex cardiogram which 20 i8 obtained by employing a transducer over the maximal cardiac iDpulse on thie ~nterior surfac~ of thQ left hemithorax in combination ~ith thQ electrocardiogram. Thi8 technique i~ of li~ited usefulness for several reasons. In particular, the recording of the npex cardioqram i8 strongly affe¢ted by the ch~r~cteristics of t~Q recording tr~nsducor and coupling of the ~7\S3000A.06 I~ARC~ 20~ 1989 13 . -,',. ' ' ' ,: " ' `' -',.'" ~ .'' ' ' ';'' "` ';' ~` '' ' "'' 1; ' ' '' ` ' ' ', ', , . .i". ''' ` ' ' ' transducer to the skin surface. ~n the absencQ o~ a palpable cardiac impulse on the chest, which may occur in patients with emphysema, the apex cardiogram cannot b~ obtained. Moreover, interpretation of the apex cardiogram waveform for hemodyna~ic measuremen~s is even more problematic than systol~c time intervals.

Another non-invasive device for monitoring cardiac function in the kinetocardiograph. ~his device records localized chest wall ~ovements with a transducer consisting of a small metal arm attached to a flat end piece which directly contacts the chest wall. Motion of the metal arm is transmitted to a bellows, connected to a piesoelectric or strain gauge transducer.

The bellows and picXup are mounted from a crossbar over the b~d, and the end piece can b~ plac~d perpendicular to any location on the chest. The a~plified signal, denoted the kinetocardiogram (KCG~, is obtained during breath holding at end-expiration. The RCG measures low frequency inward and outward c~est movements, ~hich range from 5 ~icrons in the lQft axilla to 200 microns directly over the precordiu~.

~ inetocardiography differs from apex c~rdiography in which out~ard ~ovenent~ arQ a~centuated by àn ~ir displace~ent funnel tran~ducer placed over tho apex o* the hea~t (a position where pulsations can bQ felt by the examiner). ~or example, the KCG

~7\53000A.06 NARCH 20, 1989 14 -senses true displacements of the precord~um because of it3 external crossbar frame of reference, whereas the apex cardlogram senses relative rib cage interspace ~otion. Also, the xca is su~ficiently sensitive so that records can be obtained from many points over the precordium and not ~ust at the apex as with the apex cardiograph.

KCG recordings in humans wer~ initially described in locations where the precordial electrocardiographic electrode leads were conventionally positioned. In these locations, the XCG generally depicts inward motion of the chest wall following t~e Q~S wave of thQ electrocardiogra~ followed by ~ large number of low freguency vibrations superimposed upon an upward, outward motion. The investigators who initially described tha KQG attributed the chest ~ovements to a combination o~ thQ following factors: (1) movements due to tho cardiac impact against the chest wall, (2) changes in tho intrathoracic blood volu Q a~ the result of e~ction or filling of the haart, (3) impact of ~lood ~n the graat vess~ls aqainst the chast wall and ~4) positional and shape ohanges of the contracting and relaxing heart. Traoings of KQG over the anterior and post~rior rib caqe reveal: (1) a caroto~uqular type of pulse tracing in the infraclavicular area (attribu~ed by the investigators to a ~ixed arterial venous pulse transmitted fro~ the subolaYian or axillary blood vessels~, (2) with th~ sub~ect prone, ~ wav~form confiquration s~ilar posteriorly to the V~
~lectro¢ardiographio electrode placement position, and (3) with 47\53000A.06 ~ARCH 2Q, lg89 15 upright posture, a smaller amplitude, no~sy opposite deflection signal at a posterior position corresponding to the anterior XCG
signal. The investigators attributed these findings to a combination of the factors listed above.

The KCG depicts precordial outward systolic bulges in approximately 66~ of patients with known myocardial infarctions.
The largest outward motion is found most often at the V3 electrocardiographic electrode placement position. Outward precordial bulges occur during exercise in about 30% of patients who develop anginal pain.

Although the KCG appears to provide useful information on the ~echanical prop4rties Or heart muscle, it has never received widespread clinical acceptance. This is probably because of: (1) the unwieldy transducer to patient interface; ~2) restriction of patient movemant and need for breathholding during recording; (3) noisy, often uninterpretable signals; t4) requirQment of a great dQal of skill to interpret recordings from different locations on the rib cagQ; and ~s) lack of quantitation of the KCG waveforms ~ith respect to change~ of ~en~ricular volume events obtained from analysi~ of the recordings.

Another non-invasive device for monitoring cardiac function i8 the cardio~ymograph (C~G). This device, available from ~7\53000A.06 NARCH 20, 1989 16 1~ 1 327631 Cardiokinetics~ Seattle, Washington, consiæts of a circular, ~lat capacitive plate mounted in a plastic ring strapped to the chest.
Tissue motion beneath the transducer distorts an induced electromagnetic field which in turn alters t~e frequency of the oscillator plate. This change of freguency is converted to a chanqe of voltage proportional to the chest wall motion at the transducer site and then displayed as an analog waveform. ~he CXG
provides waveforms during breathholding quite similar in appearance to the kinetocardiogram. It depicts left ventricular wall motion abnormal~ties ~ust like the KCG and therefore can be used to improve the diagnostic accuracy of exercise testing as an additional marker of myocardial ischemia.

The cardiokymogram suffers fro~ the same limitations as the kinetocardiogra~, namely, ~1) an unwieldy transducer to patient int~rfac~; (2) reætrict~on of patient ~ove~ent and need for breath~olding durinq recording; (3) noi~y, often uninterpretable signals; (~) requirement of a great deal of skill to interpret recording~ fro~ different locations on the rib cage: and (5) lack of quantitation of the CKG waveforms with respect to changes of ~entricular volume events abtained from analysi~ of the recordings.

Electroky ography and radarkymography are still other techniques for non-invasively monitoring cardiac function. The ~otions of the borders of the cardiovascular s~adow obtained with 47\53000A.06 ~ARCH ~0, 1989 17 roentgen rays can be v~sualized directly on A fluoroscope by uslng a photomultiplier tube to give a phasic analog slgna~ from cyclic variations in light produced by movement of the underlying heart border (electrokymography), or from a video monitor o~ the fluoroscopic image and similar tracking technology (radarkymoqraphy). A graphic record of the segmental motion on the left heart border provides recordings which cloæely resemble the contour curve of changes in left ventricular volume over time.

Such technology can be utilized to diagnose localized segmental dys~unction of the ventricular wall. For example, radarkymography has been used to diagnose ventricular wall abnormalities, including asynergistic and akinetic motion, associatQd with acute myocardial infarction. Radarkymography co~pares f~vorably with left ventricular cineangiography in the d~agnosis of ~syn~rgistic myocardial contraction.

Ho~ovQr, rad~rkymography and ~lectrokymoqraphy can be used only ~her~ ~n int~rfac~ is ~isuali~ed between the cardiac silhouette and ad~acQnt ~truCtures. Poor visualization is encountered in pul~onary fibrosis, pulmonary edema, pleural f~brosis and bony distortion~ of the rib cage. Dyspneic patients ~r~ difficult to study sinc~ Qxtraneous ~otions of the hear~
cau8ed by rospiration introduce artifacts. Finally, both methods ~7\53000A.06 ~ARCH 20, 1989 18 subject the patient to exposure to Roentgen rays and this hazard prevents their us~e in situations requiring long ter~ monitoring.

A still further non-invasive technique for monitoring cardiac function is impedance cardiography. It has long been recognized that the passage of a high freguency, low electrical current signal between electrodes placed on the heart or directed through the heart across the intact thorax produces changes of el~ctrical impedance which varies directly with the lenq~h and inversely with the cross-sectional area of the conductor.

In impedance card~ography, detection of localized motion of the heart is hiqhly dependsnt upon the placement of the electrodes.
To circumvent th~ problems of electrode placement, the entire thorax i8 treated as a con~uctor by placing exciting and receiving ~lectrodes at the upper and lower borders of the thorax. This p4r~its esti~ation of the ~aqnitude of cardiac stroke volume as the difference in i~pedance batween systol~ and diastole. Absolute values of cardiac stroXe volume (amount of blood e~ected by the heart per beat) are obtained by incorporating the rate of change of inped~nce (an index of the velocity differences in pulsa volume) into an empirically derived equation. It is the derivative ~avefora of torso impedanco th~t forms the basis for its ~easure~ent by the co~ercial device, the Minnesota impedance cardiograph, for calculating cardiac output.

~7\5~000A.06 ~ARCH 20, 1989 19 ~` 1 327631 Although impedance cardiogra~s were initially recorded during breathholding to eliminate impedance change~ superimposed ~y respiration, it has been found that ensemble-averaging o~ torso impedance waveforms using the R-wave of the electrocardiogram as a trigger pulse provides comparable waveforms during normal respiration in healthy sub~ects at rest and exercise and in critically ill patients.

Because changes of transthoracic electrical impedance to detact changes of cardiac volume are highly dependent on electrode placement, segmental changes of cardiac volumes and accurate reproduction of volume contours ovar time cannot readily be recorded with such eechnology. On the other hand, treating all changes of he~odynasics of the entire thorax as a single conductor appQ~rs to provide reasonabla estimates of stroke volume of the ~eart.

le ~as also long been recognized that heart motion produces gas flow within the lungs, though the mechanism of this phenomenon ~a~ pu~zled inVQ8tiqator8 for many years. One of the earliest r~searchers suggested that each heart contract~on sent a volume of blood out of th~ thorax and the consequent negative pressure inside thQ affectivity rigid container caused an inflow at the mouth.
~S Althouqh this ~aspir~ting~ effect of the heart was subsequently ~7\53000A~06 RCH 20, 1989 20 ~` 1 327~)3 1 well documented, the observation that the flow pulses were al~o present in open-chest animal preparations pointed to other mechanisms.

Cardiogenic flow pulses have been attributed to direct beating of the heart against the pulmonary parenchy~a. Although artifactually induced vascular pressure pulses produce flow oscillations in the airways, these oscillations can still be seen in an airway of a lobe to which the lobar branch of the pulmonary ar,tery has been entirely obstructed. Furthermore, in~ection of 25-50 ml of saline into the canine pericardial sac markedly diminishes all cardiogenic oscillations within intrapulmonary conducting airways despite the presence of normal pulmonary arterial pulsations. These observations suggest that neither pul~onary vascular pulsations nor volume changes of the heart, which should not be affected by a small pericardial effusion, were responsible for cardiogeni~ flow oscillations.

The heart has an irregular shapQ and contracts with a twisting ~ction; this results in a forceful thrust to some parts of the ~d~oining lung, ~hereas other parts follow the inward movement of the ~yocardiua. It is these localized transient inflations and doflations which appear to produce intrapulmonary to-and-fro flow oscillations. Pericardial fluid tends to ~ake the external surface of the pericardial sac ~ore spherical 80 that rotation or twisting ~7\53000A.06 ~ARCH 20, 1989 21 -' I 32763 1 of the heart no longer produces a thrust aqainst t~e lung, thereby diminishing cardiogenlc oscillations of the air columns.

The actual redistribution of the flow pulse among intrapulmonary airways originating from the heart depends upon relative impedance of the airways. Its magnitude depends upon the force and acceleration of the cardiac movement. However, apart from the heart movement, intrapulmonary factors must also influence the pattern and extent of transmission of the pressure impulse and the ~onal volume changes that it causes. Thus, whether a zone ad~acent to the heart deflates or not, giving rise to a flow pulse in the airways subtending it, depends upon its t~e constant. The smaller lts co~pliance and resistance, the mor~ likely ~t is to respond to thQ cardiogenic pressure impulse by emptying. In contrast, if thQ time constant i8 high (e.g., due to increased a~rway resistance), minimal emptying occurs during the time of the pressure cycle, resulting in smaller or absent flow pulses in the air~ay~. `
,.:
The preceding discussion accounts for a nuiber of experimental observations regarding rQcording8 of expired gas flow. Thus, alt~ough cardiogenic oscillations appear on recordings of continuous expired gas concentrat~ons in most normal sub~ects, patients ~it~ emphy~ema may not demonstrate this phenomenon.
Abs~nc~ o~ cardiogeni¢ oscillation8 has been observed in patients ~7\53000A.06 ~ARCH 20, 1989 a2 - ~ 1 327631 with bronchial asthma, with oscillations reappearing after partial relief of the bronchial obstruction. Lung disease oscillations are not seen in the trachea unless they are also present w~thin the lobar airways.

Luisada in 1942 reviewed the historical background for the designation, "pneumocardiogram~, and defined it as the recording of pressure changes whi¢h occur in the air passages of the lung as a consequence of the heart beat. HQ noted that graphic recordings of ~his phenomenon were published as early as 1861 in animals and in humans ~n 1876. He utilized a pressure sensing transducer from one nostril while the sub;ect breathed normally and employed electronic filtering to eliminate the slower respiratory waves.
H~ attributed the four positive and five negative deflections of t~e resulting complex waveform to the following events: 1) auricular contraction: 2) papillary muscle contraction; 3) first ventricular ~ave; 4) perip~eral pulse; 5) second ventricular wave;
6~ semilunar valve closure; 7) first diastolic wave; 8) tricuspid valvQ opening; and 9) sècond diastollc wave. He believed that the mult~ple waveforms present in the pneumocardiogram were due to the difference between v~nou~ inflow to, and arterial outflow from, the thorax.

Blair and Wedd in 1939 m-asured rib cage movements from a site b~low the sternum by r~cording pressure changes within a bellows ~7\5300QA.06 NARCH 20, 1989 23 pneumograph manufactured by the Harvard Apparatus Company. The cardiogenic oscillations recorded during breathholding were attributed by the authors to excessive outflow of blood from the chest over inflow into the chest. They ~alculated this volume to be 30 ml by assuming that the recording below the sternum was representative of the entire thorax.

Cardiogenic oscillations during breathholding have also been observed on analog signals from device~ which display the tot~l external movements of the respiratory system. Such oscillations were noted by ~ee and Dubois in 1955 who enclosed a subject within an airtiqht chamber, the body plethysmograph. The subject first breathheld after inspiring air and small oscillations of pressure (calibrated as a volume~ were sensed from the body plethysmograph with a sensitive pressUrQ gauge. Thase oscillations were ~ttr~buted to the heartbeat, but no significance was attached to t~e resulting complex waveforms by Lee and ~uboi~ or by the present inventor. After thQ recording waQ obtained whila breathholding on ~ir, the sub~ect inspired nitrous oxidQ (N20~, a solublQ gas, which ~s ta~en up by the pulmon~ry capillary blood flow.

In 1961, Wasserman and Comroe modified the body plQthysmographic techniqu~ of Lee and Dubois by substituting the sub~Qct's own tborax for the rigid body plethysmograph. Change in Qpiroaetric voluoe then reflected the exchange of gas molecules 47\53000A.06 NARCH 20, 1989 24 between alveoli and blood as long as thoraclc volume remained constant. The latter was an important requirement of the method.
Accord~ngly, to continuously monitor any movements of the chest or abdomen which would invalidate this requiremènt, two mercury in rubber strain gauqes were placed around the rib cage and upper abdomen and connected together to permit analog recording of circumferential movements of the combined rib cage and abdominal compartments.

Wasserman and Comroe believed that the cardiogenic oscillations observed with their ~ethod reflected changes in thoracic blood volume. They did not consider the oscillations to be related to changes in ventricular volume. The present inventor accepted the interpretation given by Wasserman and Comroe to the cardiogenic oscillations observed with the~r technique and used ~a8ser~an and Comroe's results in a review paper on ~easurement of c~rdiac output by alveolar gas exchange.

In 1965, Bosman and L~e utili~ed a body plethysmograph -flo~et~r mQthod ~to study the effects of cardiac contraction upon ch~nge~ in lung gas volumes during bre~thholdinq both with the glottis open and closed.~ They reported ~nd depicted curves with ~ultipla ri~es and falls fro~ the body plethysmograph and pneu~otachoqraph. T~QY interpreted these complex waveforms as ~howing an exces of aortic outfl~w over venous inflow to the ~7\53000A.06 HARCH 20, 1989 25 -` -- 1 327631 ~-thorax during systole and a reverse during diastole. Using more sophisticated technology, their work confirmed the ~indings of Bla~r and Nedd.

~7\5300QA.06 ~URCN 20, 1989 26 : ' suMMAay OF ~ INV~N~ION

The present invention, which ls sometimes referred to here1n as the thoracocardiograph or TCG, is ~ased upon the discovery that during breathholding, small oscillations detected ~y sensors placed on the rib cage (RC) and a~dominal (AB) surfaces and ordinarily used to monitor breathing patterns closely resemble ventricular vol~me curves and aortic pressure pulses depending upon their respective placements on these surfaces. m ese sensors include those which ~easures changes of rib cage and abdominal dimensions, such as the respiratory inductive plethysmograph which measures changes in cross-sectional area; the inductance circum~erential transducer which measures partial cross-sectional area; t~e mercury in silastic strain gaugQ~ bellows pneumograph, and differential linear transformer which measure circumference and partial circu~ference; magnetor~t~rs whic~ measur~ diameters; and partitioned pressure, voluo~ and c~pacitance body plethysmographs ~hich ~easur~ volu~e~.

~7\53000A.06 ..
MARCH 20, 198g 27 -~3RI~!:F DE8CRIP~N OF ~ DRAl~NG8 In the drawings:

Fig. 1 is a diagrammatic representation showing the placement of wide band (left panel) and narrow band ~right panel) transducers about the human torso:

Fig. 2 is a graphic r~presentation showing waveforms derived in accordance with the present invention in the supine (left panel) and standing ~right panel) postures;

Fig. 3 is a graphic representation similar to the left panel in Fig. 2;

Fig. 4 is a recordin~ from a sQmirecumbent normal sub~ect using ense~ble averaging to display an averagsd vascular pulse and ~ vontricular volume curvQ with their corrQsponding derivatives;
~ ig. 5 ~8 a graphic representation showing the use of curve fitting techniques to extract cardiogenic waveforms from raw data d~rivQd in accordance with the prQsent invention:

~7\5300QA.06 ~ARCH 20, 1989 28 1 3~7631 Fig. 6 compares waveforms derived using a narrow band sensor and a single bellows pneumograph:

Fig. 7 compares waveforms derived with ~ narrow band sensor and a surface inductive plethysmograph;

Fig. 8 shows waveforms dQrived, for different horizontal postures, using a wide band sensor at tbe lower rib cage and anot~er wide band sensor at the mid-abdominal level;
Fig. 9 illustrates waveforms derived using wide band sensors at the upper and lower rib cage placements and showing the effect of lung volu~e on ventricular volume curves;

Fig. 10 shows the effect of the Valsalva maneuver on ventricul~r volume curves ~derived in accordance with the present invention;

Fig. 11 depicts wa~eforDs s~owing the effect of exercise on 2~ stro~e VOlUOe8;
''.

Fig. 12 depicts waveforJs showing the effect of exercise on stro~e volume with the sub~ect in t~e supinQ posture;

~7\S~OOOA.06 ~ARCH 20, lg89 ~ 29 -;,.....

Fig. 13 shows the effect of amyl nitrite on ventricular volume curves derived in accordance with the present invention:

Fig. 14 is a rscording of a carotid arterials waveform and a S left ventricular volume curve in a subject with ischemic heart disease;

Fig. 15 is a waveform derived in accordance with the present invention and showing dyskinetic motion of a ventricle segment resulting from pulmonary hypertension:

Fig. 16 is a graphic representation of a comparison of stroke volume measure~ents derived in acCordance with the present invention and in accordance with the thermodilution method;
Fig. 17 d~picts waveform~ derived using narrow band sensors in accordance with the present invention and showing ventri~ular volu~Q curves derived fro~ dogs with various band placements and body posture~;
Fig. 18 is a graphic represent~ation comparing stro~e volume as detQr~inQd in accord~nce with the present invention and as derivQd using i~pedance cardiography;

~7\53000A.06 NARCH 20, 1989 30 . '' ~ ' t ,' . '. ~` . ' ~ i ' ~ 1327631 Fig. 19 iæ a graphic representation of cardiac output a~
derived from impledance cardiography upon in~ection with terbutaline and saline:

Fig. 20 is a graphic representation for comparison with Fig.
19 and showing cardiac output as derived in accordance with the :
present invention upon in~action of terbutaline and saline:

Fig. 21 is a series of recordings showing the effect of externally pacing the right ventricle on the ventricular volume curve of a mechanically ventilated, anesthetized dog: and Fig. 22 compares wav~forms measurQd during breathing of room ~ir (left panel) and an hypoxic mixture tright panel).
;:`

~7\53000A.06 ~ARCR 20, 1989 31 .
DETAILED DE8CRIPT~ON_OF ~REFBRR~D E~ODIMENT~

I have discovered that dur~ng breathholding, cardiogenic oscillations derived from sensors placed on the rib cage and abdomen surfaces for display~ng breathing movements differ in waveformiconfiguration depending upon the location enclosed by thè
vertical heiqht of the sensor. The invention will be particularl~y described with reference to the respiratory inductive plethiysmograph and its associated sensors, though as noted above, the present invention may be practiced with other devices used for ~easuring dimensional changes at the rib cag~ and abdomen.

The resp~ratory inductive plethysmograph is commercially a~ailaU e fro~ Non-Invasi~e Monitoring Syst~ms, Inc. ~NIMS) under tbe trade na~es Respigraph and Respitrak and is des¢ribed in United State~ Patent No. ~,308,872~ Basically, this a~paratus cooprisQs t~o coils of Teflon-insulated wire SQWn onto elastic cloth bAnds encircling tbe rib caqe and abdomen. The leads from tha ~ir are connected to LC oscillator ~odules, or preferably a sharea nodule, sucb that the inductance of the wires comprise~ the induct~nce element of the oscillator. Changes in the cross-~ectional area of tbe rib cage and abdominal compartments result ~n changes in tho inductance of the wires and hence changes in the * ~ '` ''` ' `
, 32 :

3~ ~, `' ', ', oscillation frequency of the oscillator. The result~ng signal~ for the rib cage and abdominal compartments are demodulated and displayed as analog voltage signals. In respiration applications, these signals can be calibrated and summed to reflect absolute tidal volume.

Fig. 1 shows placements of sensors employed with the respiratory inductive plethysmograph. The left-hand panel in Fig.
I illustrates the placement of commercially available wide band (WB) sensors, ~0 cm in heiqht, on the upper and lower rib cage tRC) and mid-abdomen (AB). In the usual application of this device to non-invasively monitor breathing patterns, the sensor shown at the upper rib cage closely depicts the placomen~ for respiratory~
monitoring.
For purposes of the present invention, the respiratory inductiv plethysmograph was used with modified sensors. In particular, sensors as employed in the pre~ent invention were only 2.5 o in height, such that each sensor subtended a narrower portion of the torso than the commercial wide band sensors shown ln th~ left-hand panel in ~ig. 1. The narrow band (NB) sensors us~d ~ith t~ present invention are shown in the rlght-hand panel in Fig. 1. Th~ xiphoid procQss of the gternum has been takQn as the arbitrary point of reference for placement for the NB sensor~
e~ployed ~ith the present invention, as it is an easily recogn~zed ~7\5300Qa.06 ~ARCH 20, 1989 33 -`` 1 327631 anatomic location which demarcates the caudal limlt o~ the bony thoracic cage in the midline from the cranial l~mit of the soft tissues of the abdomen. While the invention will be described herein in conjunction with the NB sensors, it will be apparent as this description progresses that sensors of any height may be employed, depending upon the information being sough~.

Fig. 2 shows, for a normal adult, waveforms traced from polygraph recordings of the electrocardiogram (ERG) and the analog lo voltage signal from a narrow band sensor employed with a respiratory inductive plethysmograph as taken during sequential breathholds. Tha QRS complex of the EKG, labeled R in Fig. 2, ~arks electrical activation of the ventricles of the heart, which precedes ventricular muscular contraction. As is well known, contraction of tbe ventricle causes ventricular volume to decrease as blood is e~ected (systole) from the ventricles into the thoracic aorta and pul~onary artery.

As indicated in Fig. 2, a single narrow band sensor was moved Qith~r abovQ or below the xiphoid process at 2.5 cm intervals b~fore each sequential bre~thhold. As also indicated in Fig. 2, th~ trial was repeatQd in both tbe supinQ and standing positions, left and right-hand panels in Fig. 2, respectively~ The uppermost crani~l border of the rib cage in a normal adult 187 cm in height, 2S uhose waveforms are dep~ct~d in Fig. 2, was situated ~25 cm abov~
, .
~7\53000A.06 ~RCH 20, 1989 34 i the xiphoid process. The multiple tracings at the xiphoid process denote repetitiv~tracings of sequential breathholds ~rom polygraph recordings at this site and demonstrate good reproduc~bility of the measurement. of course, the tracing~ at the xiphoid process were taken with a narrow band sensor disposed between th~ "-2.5cm~ and ~2.5cm~ positions, i.e. over the xiphoid process, the xiphoid process tracings being shown between the "~22.5cm" and "+25cm" in Fig. 2 simply as a matter of convenience.

By making the recordings during breathholding, waveform deflections duè to respiration are eliminated. Accordingly, it is known that the waveforms depicted in Fig. 2 are duQ to physiological changes ~nrelatad to breathing. BecausQ the changes in rib cage and abdominal di~ensions reflected by the signals shown in Fig. 2 arQ of considerably lower ampl~tude than those resulting fro~ respiration, the qain of the respiratory inductive plethys~o~raph employed in generating the waveforms, thQ
Respigraph, ~as ad~usted to about ten to twenty times the gain setting for respiration applications.
The narrow band ~ensors were placed such that recordings ~era ulti~ately obtained fro~ almost all horizontal cross-sectional regions of the rib cage surf~ce in the supine and stand~ng postur~s. The configurations of ~he rQsulting wavQforms as d~picted in Fig. 2, which I have found to be relatQd to specific s3000a.06 ~ARCH 20, 1989 35 .

cardiovascular structures, distinctly differed d~pend~ng upon th~
cross-sectional location subtended by the band. Thus, the cardiogenic signals from bandc disposed at the level of the xiphoid process show a rapid decrease ~n volume (systole) following the R
wave of the EKG, which reached its nadir shortly before or after termination of the T wave of the EKG, depending upon the precise location of the band. As shown, the diastolic phase of ventricular muscular relaxation is ~arked, at the xiphoid process, by a brief initial rapid increase in ventricular volume to a more gradual rise before reaching a peak plateau coincident with the next R wave.
This plateau continues slig~tly past the R wave before the downstroke of systole repeats itself.

T~e configurations of the cardiogenic oscillations shown in lS Fig. 2 are, as s~own, extremely dependent upon location of the sensor both in the supine and standing posituresi. The waveforms taken at the xiphoid as depicted in Fig. 2 closely resemble the ventricular volume wavefor~ as measured by sensor~ of length, di~ wtQr or volume surgically installed on tha ~earts of dogs, or froo a cardio~eter enclosing the isolated heart.

Still refQrring to Fig. 2, the band placed +25 ¢m above the xiphoid, i.e. at thQ uppermost portion of the sternum, depicts an upgoing deflQction following the R wave rather than a downgoing deflection as detected at the xiphoid process. It more closely ~7\53000A.06 ~ARCH 20, 1989 36 resembles the waveform of the descending aortic pressure pulse as detected in the prior art using other techniques. From +2.5 to ~17.5 cm above the xiphoid, the amplitudes of the signals in the supine posture diminish but still resemble ventricular volume curves. There are less marked variations of amplitude in the standing posture. For example, the amplitude of the waveform recorded with a band placed ~lo cm above the xiphoid in the standing position is approximately equivalent in a~plitude to the wave~orm at the xiphoid process. As also seen in F~g. 2, the timing of the systolic downstroke following the R wave and its slope varies among the recordings taken at different locations above the xiphoid.

The wave~orms of cardiogenic oscillations in the supine posture show an initial upward systolic deflection at the xiphoid location which is more pronounced -2.5 and -5 cn below the xiphoid.
This upward deflection denotes th~ period of isovolumetric contraction, a wQll docu~ented phenom~non. At location-~ b~low the xiphoid, the ~id-anterior sections of the band lie on the abdominal surface but the lateral and posterior sections overly the rib ca~e.
Therefore, chang~s in left ventricular volume are primarily recorded at these locations because the cardiac apex of the lef~
ventricular wall is located at the lowermost portion of the rib cage. FUrther, the slope of systolic e~ection appears to be `~
~5 stQepQx at theæe locations bQlow the xiphoid than a~ove $t. Thls ~7\53000A.06 ~ `
~ARCH 20, 19~9 37 32763~
is consistent with prior art observations that aplcal segments display a higher velocity of contract~on than basilar segments.

As is well known, during isovolumetric ventr~cular contraction immediatiely after electrical activation of the heart muscle, shortening of the long axis predominates such that the heart bacomes more spherical and the transverse diameter toward the apex actually increases. This phenomenon accounts for the brie~, often quite prominent, upward systolic deflections of isovolumetric contraction at the xiphoid, -2.5, and -5 cm band locations, and the diminution or absence of an upward deflection at this same point in tim~ in the waveforms in locations fro~ ~2.5 to +17.5 cm above the xiphoid. This is consistant with the o~servation in canines that the isovolumetric contraction of the left vèntricle varies in prominence ~epending upon the location uhere thiQ dimensional gauges ar~ surgically installed. The circumferential and length waveforms froa the canine laft ventricle as reported in the literature display pro~inent isovolu~Qtrio contraction whioh is stri~ingly si~ r to the human isovolumetric contraction waveform from bands 2~ pl~cedi fro~ -2.5 to -10 cm below the xiphoid (See Figs. 1 and 2).
ThQ upward isovolu~etric deflectioni~ are less mar~ed in the st~nding posture presu~ably bec~us~ greater longitudinal orient~tion of the heart due to gr~vity produces ~ lesser spherical c~rd~ac sh~pe at the onset of ~ystol~ than in the supine posture.

~7\53000A.06 NARCH 20, 1989 38 ` 1 32763 1 The timing sequence in Fig. 2 is consistent with ~luoroscopic imaqing of the ~eart in which the observer perceives a wave o~
~uscular contraction from the cardiac base to apex. Similar timing of the initial changes in ventricular volume with systole has also been described with dimensions recorded during biplane coronary cineangiograms. Fig. 2 also shows that the amplitude of the change in ventricular volume is less at the cranial than the caudal portions of the rib cage. Since the base of the heart i8 located more cranially than the apex, the finding of lesser chanqes of volume is consistent with the conclusion that the band measures the horizontal sector of cardiac volume changes subtended by the height of the band. So, if the atria and ventricles lie anatomically in the same horizontal plane at a particular rib cage location, summation of such signals would be expected. And, indeed, su~mation of the ventricular and atrial volume curves as reported in the literature is consistent witb the wavefoxms observed at position~ ~12.5 to +17.5 Q above the xiphoid as shown in Fig. 2.
m us, in theso ~avoforms, tho downstrokQ o~ systole i8 moro gradual at the basQ of the heart than the apex because the atria are in their diastolic period and rising in volume thereby cancelling in part the ventricular systolic volume amplitude. Further, at the nadir of ventricular systole, the upward rounded curve represents tho predo~inant peak of atrial diastole.

47\53000A.06 NARCH 20, 1989 3 Fig. ~ shows tracings from thQ bands 2.5 CQ in height in the same subject whc,se waveforms are depicted in Fig. 2, but taken one week later at locations ranging from 15 cm below to 10 ~m above the xiphoid process. As seen from a comparison of Figs. 2 and 3, S the appearance of the waveforms is cons~stent for recordings ta~en at t~e same locations, but one weeX apart, evidencing good reproducibility of thQ results. Referring to Fig. 3, the bands placed -12.5 cm and -15 cm below the xiphoid on the abdominal surface show deflections more closely resembling the abdominal aortic pressure pulse. It should be noted that the -15 cm location was 2.5 cm above the umbilicus.

Although the description thus far is based on waveforms generated during breathholding, the display of the averaged wavQform at any location Can also be ob~ained during breathing by the uell known technique of ensem~le-averaging us~ng the R wave of the ERG or the upstroke of a systemic arterial pulse obtained non-invasively or invasively as a trigger to di8play solely the ~emodyna~ic signals while eliminating the breat~ing waveform. Fig.
~0 ~ shows the ventricular volume curve together with the elQctrocardiograph and also the descending aortic pressure pulse fro~ the upper rib cago ~it~ thQ electrocardiogram using an average of 50 heart bQats. In FIG. 4, starting vith the left panel, from top to botto~, t~Q first panel shows the carotid arterial waveform:
the se¢ond panel shows t~e carotid arterial waveform derivative;

~7\5~000A.0~
MARCH 20, 1989 40 : .

the third panel shows the ventricular volume curve from TCG just below the xiphoid process; and the fourth panel shows the derivative of TCG. On the right, from top to bottom, the first panel shows the descending thoracic aortic pulse o~tained from TCG
just above nipple level on the RC; the second panel shows its corresponding derivative; and the third and fourth panels show, respectively, the left ventricular volume curve from TCG ~ust below the xiphoid process, and the corresponding derivative. The finding of an aortic pressure pulse on the recordings shown on the right lo side of thi~ Figure demonstrates heterogenicity of cardiogenic oscillations from different thoracic sites. The hatched line depicts ~he EKG; the lowest panel displays the second derivative of the E~G.

The preceding descrip~ion of varied waveform configurations of cardiogenic oscillations obtained with external sensors placed on the rib cage and abdomin~l surfaces accounts for the inconsistencies ~nd ~isinterpretations regarding previous recordings of these signals. Thus, the signal from a whole body plethysmogr~ph represents the sum of both positive and negative deflQc~ions from the rib cage added to positive deflections from t~e abdominal compart~ent. Simi~ar mixing of signals is displayed on the su~ ~ignal from the rib cage and abdominal signals utilizing the respiratory inductive plQthysmograph or bellows pneumograph in ~hich tr~nsducers are placed upon both the rib cage and abdominal ~7\53000A.06 ~RCH 20, 1989 41 ` 1 327631 surfaces. And in a previous study using a slngle bellows pneumograph placed just ~elow the sternum, the authors interpreted the waveform assuming that this location was representative of the cardiogenic oscillations of the entire thorax rather than reflecting cardiovas~ular events localized to their recording SitQ.

Fig. 5 illustrates a further technique for obtaining waveforms in accordance with the present invention during breathing.
~eferring to Fig. 5, the irregular waveform in the upper tracing shows the signal detected during breathing from a single narrow band sensor connected to a respiratory inductive plethysmograph, with the band positioned at tha xiphoid, wbich indicates that the band is positioned over the ventriclas. This raw signal includes both a larger amplitude respiration component, and a smaller one due to cardiac function, tha latter b~inq the one of interest here.
To rQmove the signal component resultin~ from respiration, the raw signal in the upper tracing of Fig. 5 is matched, using a conventional CUr~'Q fitting equation with a cubic spline over sequQntial cyclas, each of which comprisas two cardiac beats. I~
this curv~e fit, dapicted as the discontinuous "~mooth" wavefor~,in tho upper tracing, i8 subtracted from the raw signal, the tracing depicted in the lower panel results, the discontinuities in the lo~r tracing rQsulting from the cur~e fitting technique described above,, though thes~ discontinuities may be eliminated by employing conventional ig~,oothing techn,iques to ad~acent cur~e fits as will ~,7\53000A.06 NAR~H 20, 1989 42 t . - . ... ..

be apparent to those of ordinary skill in the art. Similarly, th~
"noise" on the lower tracing may be eliminated ~y high ~requency filtering. Even with the discontinuities and noise on the tracing in the lower panel, ~t may be seen that the lower tracing corresponds to ventricular volume curves as published in the literature. The lowermost tracing in the lowar panel is simply the EKG. The removal of the respiratory waveform to prov~de beat to ~eat display of the cardiogeni~ oscillatlons may also be carried out by other digital adaptive filtering technigues.

Since compliance of the rib cage remains constant during brief recording periods, change in amplitudes of the ventricular volume curve should provide accurate trending of relative changes in stroke volume as well as ventricular contractility and relaxation char~cteristics. The product of stroke volume and heart rate represent~ cardi~c output and relative trends of the latter are ~lso av~lable. Al~o, ti~ing of systole and diastole slopes of various portions of the ventricular volume curve, and various volu~e8 ~8 ~ ratio to stro~ volu~e, should allow comparisons a~ong dif*erent sub~ects and trend plots over time in a s~ngle subject.
Finally, ~easure~ent of the absolute value of stroke volume by independent ~ethods 8uch ~8 dy~ diiution, thermal dilu~ion, i~ped~nce cardiography, radionucleide scans of the heart, 2-D
echocardiography, ~ngiography, etc., allows one to set the initial 2S amplifier gains ~or the external sensor used in the present 47\S300QA.06 MARCH 20, 1989 43 '' ,~',~ - `~ ,-,,, ; ~ , " ;~ ~ ~t ;~ "~ "~ "~ " "~

1 32763 t invention to be equlvalent to the values of stroke volume obtained by the preceding methods.

It has not been possible to calibrate the ventricular volume curve to absolute volumes independent of another method for obtain~ng absolute values of stroke volume. However, it is possible to compare amplitude of cardiogenic oscillations from one site on the rib cage to another at a reference location. Thus, in a ~eries of experiments involving six nor~al sub~ects, a band 2~5 cm in he~ght ~as placed horizontally immediately below the xiphoid proc~ss and designated reference (REF) because solely the left ventricle is transected anatomically at this s~te. Other bands were placed 3 cm b~low REF, and 3, 6, 9 and 12 cQ above REF, and at the u~bilical leYel~ The electrical gain of respiratory axcursions of these bands uas ad~usted to be equivalent to the band at R~F and the a~plitude o~ their cardiao ~av~forms was compared to the cardiac ~avQform of the REF band. In supine, semi-recumbent and saatad postures, at R~F, 3 c~ bQlow and 3 cm above it, the card~ac ~a~efor~s ~ad the contour o~ ventricular volume curves.
Nore cepbalad, ~avafor~s tended to have compleY oscillations. At the ~ighest rib caga leYQl and umbilicu8, waveforms resembled descending ~ortic pressure pulæe~. Amplitudes of waveforms were generally s~aller at the ~6 and ~9 cm sites compared to the REF
~and in all postures, vi~. 41% to 70% of REF tp < .01). There was ~5 no correlation between amplitudes of cardiac and corresponding ~7\5300QA.OC
~ARCH 20, lg89 44 1 32763 1 -~
respiratory wave~orms (r = -.14). Thus, this method o~ amplltude analysis should permit a study for obtaining nor~al values and b~
capable of diagnosing hypfokinetic ventricular segments ~decreased motion) as might occur in patients with ischemic heart disease.
As stated earlier, TCG appears to reflect changes in cross-sectional area of the cardiovascular structures underlying the transducer. Since respiratory airflow and regional lung expansion ~ay be altered by different density gases filling the lungs, we investigated whether or not t~e TCG waveform was influenced by this factor. In addition to T~f~G for measurement of changes of stroke volumQ ~SV), systolic afnd diastolic timing and volumffffff events, PEP~LVET was obtained as a carotid systol~c time interval (S~
Six normal ~fn breathed (1) air, (2) 20%f~ and 80% ~ffffffff~ and (3) 20%
~2 and 80% SFf5 for 5 minutQs and 3 ~easures of TCG and STI'~ were fcarried out ovar another 5 ~finutes. Ther~sf werQ no differences aaong thef 3 gas ~ixtures ~hose densities varied 12-fold, in heart r~tef, SV, PEP/LVET, peak ffaf~ection rate/SV, and time of R-wave to pe~ ff3ction ratef. Therefore, thi~ confir~s that T~f~G measurement of ventricular function is unaffected by changes in physical composition of gases ~ithin thfefflungs. This i8 additional evidence that 'rQG displays changes in volume of underlying cardiovascular structures.

~fffff7\53000A~ 06 ~AR'CH 20, 1989 45 ~hile the present inventlo~ has thus far been descr~bed based on measurements taken with a respiratory inductive plethysmograph using narrow and wide bands, other externally placed respiratory monitoring devices can be employed to record changes of cardiac volumes and aortic pressure pulses. Fig. 6 illustrates this point.
The waveforms shown in Fig. 6 were obtained by placlng narrow band~
connected to a respiratory inductive plethysmograph at the xiphoid and +25 cm above it, and a single bellows pneumograph (BP) +7.5 cm above the xiphoid process. NIP denotes the recording from a necX inductive plethysmograph which provides a non-invasive wa~eform of the carotid arterial pressure pulse as described in ; u.s, Patent Nos~ 4,452,252`and 4,456,~15. The EXG is also shown in ~ig. 6. ~ig. 6 shows that the waveform from thQ bellows pneumograph IBP) closely resembles the ventricular volume curve obtained using the resp~ratory ~nductive plethysmograph.

~eferring to F~g. 7, a recording ta~en with a surface inductive plethysmograph (SIP) placed on the rib cage over the left bord~r of the heas~ ~8 shown together with the EKG and a recording taken Wi~h the Sespiratory inductive ple~hysmograph at the xiphoid. As described in Canadian Patent No. 1,216,635, the entire content of ~hich i8 hereby incorporated ~y reference, the SIP measures changes of surface cross-sectional area underneath the ~5 ~irQ loop of the transducer. As seen in Fig. 7, the SIP also .. -'..~.~
~` `,'.:. ': .

- 1327631 ~_ .
provides a record~ng depicting ventricular volume changes, though - the waveform appears slightly distorted compared to tha - - corresponding waveforms obtained from cross-sectional sl~ces around , :
~-` the r~b cage as recorded with the respiratory inductive - 5 plethysmograph using a band placed at the xiphoid process.

Fig. 8 depicts, for different horizental postures, ventricular volume waveforms recorded with a wide band placed at the lower rib cage placement shown in the left panel in Fig. 1, and the abdominal aortic pressure pulse recorded with a wide band placèd at the mid-abdominal level shown in Fig. 1. The EKG is also shown in Fig. 8. The ventricular volume curves show similar configurations among the variouæ postures, but there is accentuation of the iso~olumetric contraction period of the systolic part of the ventricular volume curve in the left lateral decubitus and the prone postures. Sliqht alterations in configuration with changes of posture are not unexpected since the heart is free to rotate and elongate in the rib cage as a function of gravity. ThQ sQctor of the heart subtended by the externally plac~d band or a similar external monitoring device would change ~f th~ ~eart became orientQd in a different plane~ The abdominal ~ortic pressur~ pul~e tracing i8 clearly rQcognizable in the supine posture and completQly ~bsent in tha prone posture. This is probably because the æupine posture permits maxi~um transmission of thc ~ortic pulse through th~ more compliant anterior abdominal ~7\S3QOOA.06 HARCH 20, 1989 47 wall whereas in the prone posture~ aortic pressure pulse transmissions to the anterior wall are highly da~ped leaving only the back and the sides of the abdomen for transmission o~ vascular oscillat~ons and the large amount of muscle mass present in the back and sides of the abdomen causes compliance (increased stiffness) of these regions which damps the aortic pulse pressure waveforms. Since compliance of the entire rib cage is much higher tban tbe heavily muscled lower back, satisfactory recordings of ventricular volume curves are obtained in all horizontal postures.
Devices e~her tban the respiratory inductive plethysmograph --whicb are utilised to measure breathing patterns by changes of partial circu~ferences ar~ conventionally placed on thQ anterior surface of the rib cage and abdomen compartments. TbesQ include th~ bellows pneumograph, mercury in silastic ~train gauges, and thQ linear differential transformer. They are incapable of ~ccuratoly monitoring breathing movements in tho pron~ position b~cause ~otion of the anterior surface of the transducer on the rib c~ge i~ restricted owlng to tbe interposition of the transducer ~Qt~een t~ssues of the rib cage and the horizontal surface of the b~d. Since thes~ devices do not generally provide accurate a~ure~ of lateral and posterior motion, thQy cannot display ~entricular volume curves ~hen the sub~ect assumes the prone postur~. Magnetometers, which are conventionally placed to measure ~5 changes o~ anteropostQrilor diameters of the rib caga and abdomen * T~

;~``,) .. , compartments with respiration, do not produc~ accurate representations of changes of respiration nor ventricular volume when the subject assumes the lateral decubitua postures owing to exclusion of lateral rib cage movemQnts which go undetected by tho transducer.

Fig. 9 depicts the effect of lung volume on ventricular volume curves obtained in accordance with the present invention. In each of the four panels, recordings taken with a respiratory inductive lo plethysmograph using a wide band at the upper and lower rib cage placements in the left-hand panel of Fiq. 1 are sbown together with the EKG. In Fig. 9, near ~C indicates near total lung capacity signifying that the subject inspired a deep breath almost to the limit of vital capacity and breathheld with a closed airway at this lS lung volume level. FRC signifies functional residual capacity, i.Q. lung volu~o at the end of nor~al expiration, ~nd ~bet~een FRC
TLC~ in Fig. 9 signifies a moderately deep inspiration followed ~y a breath~old at this level. RV connotes residual volume, i.e.
lung volumQ after full expiration, and ~near ~Vn in Fig. 9 indicates breathholding at a lung volume near the lower limits of vital capacity. ~Between FRC and RV~ signifies breathholding after a ~odorately deep eKpiration.

As seen in Fig. 9, the configuration of the diastolic slope o~ tho ventricular volu~ curve is altered by the lung volume level ~7\5~000A.06 MARCH 20, 1989 49 suc~ that the terminal slope is flat at the high lung volumes and slopes upward at low lung volumes. Furthermore, the slope of initial ventricular systole is more gradual at the high lung volume (nnear TLC" and ~between FRC & TLCn) than the steeper slopes at the low lung volumes ("near RV" and "between FRC & RV"). There are minimal differences in the amplitudes of ventricular volume curves at the various lung volume levels except for a ælight increase at the level "between FRC & RV". These data suggest that myocardial contractility is increased during breathholding at low lung volumes compared to high lung volumes as expressed by the ~ore rapid slope of systole at the low lung volume level. Furthermore, the flat ~lope o~ the terminal diastolic curve suggests that at the high lung volume levels ventricular compliance is decreased compared to ventricular compliance at the low lung volume levels. In the lattQr situation, the terminal curve slopes upward. This further suggests t~at the primary ventricular volume measured by the band ~t the lower ri~ cage placement is the left ventricular volume Qinc~ it is known that both diminishad myocardial contractility and lo~ered left ventricular compliance occur at increasing lung volume le~el~.

Referring to Fig. 10, the Qffect of the Valsalva maneuver on ventricular volume curves derived in accordance with the present invention is ~hown. The Vals~lva maneuvQr consists of straining against eitber a closed glottis or ~n occluded airway. Fig. 10 47\53000A.06 NARCH 20, 1989 50 `' ' ` ; ' ' ' ' ` ' ` ' ` ' ' ` ~ ".~ ! . ~, - 1 32763~ ,-depicts such a ~aneuver with wide bands placed at the upper and lower rib cage pllacements depicted in the left-hand panel ln Fig.
1. Waveforms derived from the neck lnductive plethysmograph (NIP) for recording carotid arterial pressur~ pulses and the bellows pneumograph (BP) placed ~7.5 cm above xiphoid for recording ventricular volume are also displayed, as is the ERG. During the Valsalva maneuver, the pressure at the mouth rose to about 60 cm H20. The amplitudes of the ventricular volume waveform at the lower rib cage placement and the thoracic aortic pressure puls~ at t~e upper rib cage placement showed a marked fall in amplitude during tha Valsalva maneuver, as did the NIP and BP recordings.
The slope of systol~c e~ect~on of the ventricular volume curve during the Valsalva maneuver markedly diminished. The stroke volume during thQ Valsalva maneuver for the band at the lower rib caga placement fell to 67% of t~e baseline, and rose 29% above baseline upon release of the Valsalva ~aneuver. There was a conco~itant rise in theii carotid arterial pressure pulse recorded wit~ NIP, but the wavefor~ of BP failad to disclose this rise. The findings in Fig. 10 are si~ilar to those obtained with le~t ventricular angiography in which the fall in stroke volume derived froa ventricular volume measure~eints fQll from 35 to 75% of the basQlinQ during the straining p~riod. Si~ilar decline of stroke ~olu~e (53%) ha~e also been obtained using an intracardiac i~pedance catheter in the right ventricle. Echocardiography ~easurements o~ ventricular Yolu~e8 in both healthy sub~ects and ~7\53000A.06 MARCH 20, 1989 51 -~` 1 327631 patients with congestive heart failure demonstrate similar reductions in st:roke volume during V~lsalva maneuvers.

The effect of exercise upon stroke volume depends upon the body posture in which exercise i8 carried out. In normal adults, utilization of dye dilution technigues f~r cardiac output allows calculation of stroke volume by dividing cardiac output by heart rate. During walking on a treadmill, one prior art study showed that stroke volume had an initial larqe rise with light exercise, i.e. heart rate rose from a baseline of 87 b/m to 115 b/m and stroke volume increased 69~. 8troke volu~e continued to rise slightly with more str~nuous exercise up to a maximum of 84% above basQline at a heart rate of 171 b/m. On the other hand, in supine bicycle exercise, stroke volume increased only 6% during mild exercise, froa a baseline heart rate of 71 b/m to 119 b/m. With ~oderate ~xercise, heart rate rose to 127 b/m but stroke volume increased only 13% over baseline. Referrinq to Fig. 11, in a nor~al adult instru~ented with narrow bands connected to a raspiratory inductive pl~thysmograph and ~eated on a bicycle, during breath~olding imm~diately after terminating an exercise load, stroke volume increased from 35 to 65% over baseline on the band placed 2.5 c~ above the xiphoid, while heart rate increased fro~ 54 b/n up to 125 b/~. Referring to Fig. ~2, in the supine po~turc, the ris~ of stroke volu~e with exercise measured with the b~nd was ~uch les8 than the seated posture, ~ounting to 32% over ' 47\S3000A.06 NARCH 20, 1989 52 .
, .

~V~k~ .~r~,;;;",x~

1 32763~
baseline, while heart rate-increased fro~ 67 b/m up to 116 b/m.
The increases of stroke volume exceed those reported for supine bicycle exercise using the dye dilution method for cardiac output but are in agreement for the difference in response of stro~e volume to exercise in the seated and supine positions. As shown, the rate of both systolic ejection and diastolic filling of the ventricle as measured with the bands narkedly increased w~th exercise.

Amyl nitrite/ a vaporized compound at room temperature which is administered by nasal inhalat~on, produces an immediate fall in systemic vascular resistancQ associated with secondary altera~ions of cardiac hemodynamics. Fig. 13 shows the effect of this drug on the vantricular volume curve in a supine normal adult as refl~cted by measurements taken with a respiratory inductive plQthysmograph with a ~ide band plac~ at the lower rib cage placement and on the abdominal aortic pressure pulse as reflected by ~Qasure~ants taken with a wide band at the mid-abdominal placement. Fifteen ~econds after ~nhalation of amyl nitrite, ~tro~e voluma increased 39~ above bas~line and heart rate rose from thQ baseline of S~ b/m to 84 b/~. Nyocardial contractility ar~edly increased as indicated by the morQ rapid 810pe of systolic e~ection aftQr amyl nitritQ. There was also a more rapid r~se in fill~ng during thQ di~stolic portion of the ventricular volume ~5 CUrVQ. ThQ increased rata of myocardial con~ractility was also ~7\5300QA.06 ~ARCH 20, 1989 53 present 27 seconds after amyl nitrite administration when the heart rate had slowed to 67 b/m. Thirty seconds after amyl nitrite administration, the heart rate was slower than baseline at 48 b/m and myocardial contractility returned to baseline value. Thu5, measurement of ventricular volume curves with the respiratory inducti~e plethysmography in accordance with the invention provides a beat by beat recording during breathholding of the alterations expected from 8 drug which increases myocardial contractility and cardiac output. of course, t~is information could also be derivled during breathing by employing ensemble averaging or ths curve fitting technigue as more fully described above.

~cute myocardial infarction with/without subsequent healing uay produce paradoxical, dyskinetic or akinatic motion of the in~ured seg~ent of the ventricle. In addition, silent ischemia may ~180 induce such changes. ~ig. 14 depicts a dysXinetic ventricular volu~Q curve in a patient ~ith ische~ic heart disease. More particularly, FIG. 1~ sho~s a rQcording of a carotid arterial ~vQfor~ and left ventricular volume curve in a patient with ische~ic heart disease BPs ~ systolic blood pressure; BPd =
diastolic bload pressur~; P~Pu ~ pre-e~ection period uncorrected for pu18~ wavafor~ delay: ~VET ~ left ventricular e~ection time;
dP~dt ~ ~aximum ratQ of rise of carotid arterial waveform.
Ventr~cular wall dy~kinesis is shown in the third recording from tho top. Note that the time from the R wave o~ the EKG to Peak ~7\S3000A.06 MARCH 20, 1989 54 '.'''''''"

` ` - 1 32763 1 --Ejection Rate (PE~) is markedly prolonged to 520 ~s. Identical findings were obtained with echocardiography. Dyskinetic motion also may be present in patients with pulmonary hypertension $n whom right ventricular enlargement is present. This event has been detected with bands located 5 cm above the reference band (placed just below the xiphoid process) which showed a normal left ventricular volume curve contour (Fig. 15).

Cardiac output can be measured by the thermal dilution method lC using a Swan-Ganz cardiac catheter whose tip is placed within the pulmonary artery. Stroke volume is calculated by dividing the cardiac output by heart rate. This value can be used to calibrate the systolic portion of the ventricular volume curve (obta~ned with a narrow band sensor or si~ilar transducer) to an absolute volume value in thQ baselinQ period. Thereafter, this valuQ can be utili2ed for all subsaquant calculations of stroke volume from the ventr~cul~r volume curves to ascertain both the absolute volume v~ri~tions an~ to astablish the validity of the non-invasive ~Qasure~ent. Tho ~ccuracy of the latter depends upon the ~ssunption that the ventricle can b~ considered as moving with one deqreo of freedom but tbis assumption can only be proven by comparing thQ thermodilution method (or other ~ardiac output ~thod) to the ~easure~en~g madQ with the non-invasive TCG
technique. This Qxperiment was carried out in 8~X anesthetized aogs. BasQline values w~re obtained by gimultaneous collection of 47\53000A.06 NARCH 20, 1989 55 ~ 1 327631 narrow band derived ~TCG) and thermodilution values o~ stroXe volume and cardiac output. The animal~ were then g~ve 50 ml infusions of a 10% dextran 40 solution every 15 minutes with repeated simultaneous measurements at each t~me interval until S cardiac output by the thermodilution method no longer increased with further dextran 40 infusions. Stroke volume by the thermodilution method rose to a maximum of 40% above the baseline value. In 46 paired comparisons, 87i~ of stroke volume values based on the TCG fell within 20% of stroke volume measurements based on lo thermodilution values ~Fig. 16). Thierefore, TCG appears to provide an accurate measure of changes of stroka volume and cardiac output in anesthetised dogs.

Because the rib cage of ~he dog is highly compliant much like thie rib cage of a human baby, a study was undertaXen to determine if it ~ould be possible to obtain satisfactory recordings of ventricular volume curves usinq bands on the r~b cage and abdomen in dogs. Fig. 17 depicts recordings o~ bands placed around the uppQr rib cag~ of the dog ~ust underneath the axilla with ~ariations of place~ent approx~mately 1 to 2 cm upward or downward.
T~e left panels depict recordings when the dog was in th~ left latQral decubitu~ position on a flat table. The upper left panel ~ho~s a typical ventricular volume curve along with a superimposed rounded upward wavQ of atrial diastole at the nadir of ventricular systole. In the lower lQft panQl, the band has been moved downward 47\53000A.06 ;
MARCH 20, 1989 56 ~

1 327631 ~
about 1 cm and the wave of atrial diastole i8 now elim~nated.
There is absence! of the isovolumetric contraction phase o~ the ventricular volume curve in the upper left panel as indicated by the tracing labeled "RIP-upper RC (NB) n, which connotes a narrow band placed at the upper rib cage and connected to a respiratory inductive plethysmograph, but a prominent upward deflection on the right upper panel. The bands on the mid-upper abdomen were considerably dependent upon their placement SitQ. When the dog was placed on a V shape table to support the body in a different orientation, the mid-abdominal band on the upper right panel showed a typical wave~orm of the abdominal aortic pressuxe pulse with an easily recognizable dicrotic notch. Other band placements also gave ~bdominal aortic pressure pulses.

It is clear from th~ foregoing description that external ~onitoring with non-invasive sensors which measure rib cage and abdo~inal ~ovements are capable of recording cardiovascular events ~n ~aults, babies and animals. With appropriate sensors and data processing, recording of segmental ventricular volume curves and ~ortlc pressure pulsQs on a beat by beat basis during breathholding is po~sible. Average wave~orms can be obtained during breathing throuq~ thQ technique o~ enæemble-av~raging, using as a trigger th~ QRS complex of th~ electrocardioqram or the upstroke of a systemic ~rteri~l pulse recorded either non-invasively or inv~sively. Alternatively, the curve fitting techniques and ~7\53000A.06 ~UURCH 20, 1989 s7 adaptive digital filtering techniques may be employed to extract the cardiogenic w~aveforms from the respiratory waveform. Further, waveforms of ventricular volumes and aortic pulses at different cardiac cycle lengths, and at variou~ points in the lung volume level and airflow, can also be obtained. The technology described herein carries ma~or implications in terms of physiologic, pharmacologic and clinical understanding of cardiac function and diagnosis of heart disease in adults, babies, and animals.

It is impractical to measure stroke volumes and cardiac output with the invasive thermodilution technique in normal human subjects in order to establish the validity of the thoracocardiogram (TCG) for measuring changes of stroke volume (SV) and cardiac output ~oo). There is a~ple evidencQ in the literature that ~e impedance cardiograph can measur~ changes of stro~e volume and cardiac output if body posture is fixed. This device uses an empirical eguation and an assu~pt~on that the thorax can be treated as ~ single conductor to deriv~ values of strok~ volume. ThQ waveform from thi8 d~viCQs rQsemblQs an aortic puls~ and is opposite in polarity to th~ ventricular volume wavQforo obtained with the thoracocardi~graph. Valu~s of strokQ volume ~easured by the t~oracocardiogra~ tSY(TCG~] were compared to SV measured by i pQdanc~ car*iograph ~IC) in six normal sQmirQcumbent men after subcutanQous in~ect~on of 0.25 ~g terbutaline to cause increased SV. On another day, 1 ~1 of salinQ was in~ected subcutaneously to ~7\53000A.06 HARCH 20, 1989 58 1 327631 ~

serve as a control. Data fro~ TCG and IC were collected every fiv~
minutes during a 30 minute baseline and 90 minutes after in~ection.
Maximum ~ncreases of SV and CO after terbutaline were 27~ and 50%, respectively; sv and co were not alterea by the saline in~ection.
In 288 paired comparisons, 91S of SV(TCG~ values fell within 20%
of SV(IC) ~Fig. 18). Further, there was no statistically significant dif~erences between IC and TC~ derived cardiac outputs at any tims poin~ in the preceding investigation (COmparQ Figs. 19 and 20). These data indicate that TCG derived ventricular volu~ne curves accura~ely estimate changes of stroke volume and cardiac output.

The configuration of tbe ventricular volume curv~ provides indices of systolic and diastolic function of the heart. For systolic function, these intervals and volum~ ratios WerQ compared to systolic ti~e intervals which are w~ll known timing measures of th~ carotid arterial pulse, to est~blish rolationships for systolic function between these two different measurements of cardiac contractility. In the first series of experi~ents, effects of terbutaline ~ere lnvestigated.

Terbutaline has been purported to be a beta-2 adrenergic agonist but its administration is associated with a marked, ~ustained incr~ase of cardiac output ~CO). Tha latter is attributed to sy~temic vasodilation and possibly anhanced 47\53000A.06 NARCH 20, 1989 59 f 1 327631 ~

ventricular cont:ractility. See Chest, volu~e 68, pages 616 et seq., 1975. ~o further characterize its action, several non-invasive cardiovascular monitoring techniques were employed. ~he left ventricular volume curve (LW C) was displayed as an av~raged cardioqenic oscillation with the thoracocardiograph (TCG). The respiratory signal was eliminated by an ensemble-averaging method.
In addition to systolic and diastolic volumeæ from LW C (TCG), ot~er parameters were measured: (1) heart rate (HR) by EKG; (2) blood pressure (BP) by cuff auscultation: (3) systolic time 10 intervals from the carotid pulse obtained by cominqation of an inductive plethysmograp~ band around neck and th~ EKG; (4) dP/dt of carotid pulse after calibration with BP: and (5) e~ection fraction (EF) by equation utilising PEP/LVET (PEP - pre-e~ection period and LVET ~ left ventricular e~ection ti~e). In a 2 day 15 crossover study, 5 no~mals receiv~d terbutaline .25 mg subcutaneously or salin~ and data were analyzed at baseline and peak rQsponsQ, 10-20 ~inutes ~fter in~ection. Compared to saline, terbutaline produced significsnt rises over basel~ne in HR (20%), LVE~I (8%) tLVETI - left ventricular e~ection time index], EF
20 (16%), strokQ volumQ (28t~, cardiac output (54%), peak e~ection rat~ ~PER) (61%), dP/dt (70%) and left ventricular stroke work (27%). Terbutaline significantly decreased diastolic BP (9%), PEPI
(20%), PEP/LVET (31%) R to PER time (13%) tR - R wave and PER -peak e~ction wave~ and peripheral vascular resistance (43%).
25 Early diastolic filling flows, volumes and timing were not altered.

~7\53000~.06 NARCH 20, 1989 60 The simplest and most consistent parameter of the 8y8tolic portion of the L W C was shortening of R-PER time, easily recognlzed points on the EKG and TCG waveforms, respectively.

The systolic amplitude of the ventricular volume curve can be utilized to estimate trends in stroke volume, and in con~unction with an invasive technique 8uch a8 thermal dilution or dye d~lution, or a non-invasivQ technique such as impedance cardiography or echocardiography, etc., the ventricular volume waveform can be calibrated to an absolute volume. Measurements can be obta~ned with the standard (w~de) or narrow band transducers of tha respiratory inductive plethysmograph, but alternatively other non-invasive sensing devices that have been utili2ed for ~easuring breathing movements can al80 capture cardiovascular ~vents as a function of th~ height subtended by their ~ransducers.
Th~ rQspiratory inductive plethysmograp~ is preferablQ to other 8UCh dovices because it can provide accurate display of ventricular ~olu~e curves indepQndQnt of postural changes in the horizontal plane, ~heraas oth~r rQspiratory monitorinq devices generally cannot accur~taly record ven~ricular volu~ curves in the prone or lateral d~cubitus postures. The product of heart rate times stroke voluo~ equals cardiao output. Nith the invention, cardiac output ~easure~ents can be obtained at rest and exercise in both normal ~nd disQas~d sub~Qcts. These measurQments can be recorded during `~
br~athholding on a beat by beat basis or during breathing with ~7\S3000A.06 HARCH 20, 1989 61 ~
:.

display of the aVeragQ ventricular volume curve by ensemble-averaging using the QRS wave of the electrocardiogram or upstroke of a systemic arterial pulse as a trigger, or by using curve fitting or adaptive digital filtering techniques as described above to extract the cardiogenic waveform from the respiratory waveform.
Further, ventricular volume curves can be recorded in sleeping humans or animals during spontaneous central apneas, obstructive apneas or during breathing. This should permit analysis of the effect of such entities as Obstructive Sleep Apnea Syndrome on cardiac performance and guide therapeutic interventions. `

The electronics of any of the sensors mentioned above can be ~iniaturised such that thQy can be incorporated into a tape recorder, compact disc, etc.- Holter monitoring device to carry out ambulatory monitoring for both electrical activation of the haart through the electrooardiogra~ and its mechanical response as detected by the present invention for non-invasively recording ventricular volume waveforms and aortic pressure pulse. This ~hould be u~Qful in characteri8ing cardiao arrhythmias and the effects of cardiac ischemia on cardiac performance.

Fig. 21 illustratQs the ~ffQcts of externally pacing the right ventricle on thQ ventricular volu~e curve o~ ~ mechanically ventilated, anesthetized dog. ThQ data were obtained by ensemble-~veraging using the QRS wave of the ERG as the trigger signal. The .
~7\s3000a.06 ~ARCH 20, 1989 62 - ` W 1 327631 --uppermost tracing depicts th~ unpaced ~nUPn) EKG and ventricular volume curve; lhere the heart rate was 112 beats/minute. At ventricular pacing (nVPn) heart rates below the unpaced heart rate (75 b/m, 92b/m and 107 b/m), the ventricular volume waveform has a similar appearance and timing relation to the QRS as the unpaced recording. However, at hig~er ventriculflr pacing rates (123 b/m, 132 b/m and 184 b/m), there is a ~arked delay relative to the QRS
due to paradoxic (dyskenesis) of the ventricular segment subtended by the inductive plethysmographic transducer. The ventricular lo volume curves of FIG. 21 were obtained with a band sensor placed at the xip~oid process of the dog.

The recording of both the electrocardiogram and the ventricular volume waveform miqht help to differentiate suprsventricular tachycardia with aberrant electr~cal conduction fro~ ventricular tachycardia in which the QRS complex is indistinguishable. These two arrhythmias require different modes Or Janage~en~, a~ the ventricular tachycardi~ is an immediate life-threatening cardiac arrhyth i~ whereas ~upraventricular tachycardia with aberrant conduction is not. The differentiation may be possibl- by t~ree ueans: 1) recording of atrial diastole from a b~nd placed 10 to 15 c~ above xiphoid ~Figs. 2 and 3) in a manner analogous to ~ugular vQnous pulsation~ for recording of regular atrial contractions during supraventricular tachycardia, 2) obsQrving abnor~al ventricular uaveforms with timing and phase .
~7\5300QA.06 ~$RCH 20, 1989 63 -` ~ 1 32763 1 abnormalities along with dyskinetic wall motion as in ventricular pacing, or 3) loss of the isovolumetric contraction period of the ventricular volume curve.

Although continuous electrocardiographic recording over 8 to 24 hours with a tape recorder (Holter ~onitoring) is often utilized to detect transient cardiac ischemia (impaired blood supply to the ventricular muscle) as reflected by alterations of the ST segment and T wave of the electrocardiogram, it is well recognized that t;he usefulness of this approach is limited by artifacts and the major applicability of Holter monitoring resides in diagnosis of cardiac arrhythmias. However, segmental abnormalities of ventricular wall ~otion may preced~ electrocardiographic abnor~alities. Thus, the combination of electrocardiographic Holter recording and segmental ventricular volume waveforms with bands or other devices for sensing rib cag~ movements should improve the diagnostic accuracy of detecting ~yocardial ischemia both in patients who have chest pain and those ~ho do not ~silant ische~ia).

~easurement of changes of cardiac output in patients who are critically ill s~ould help to guide therapeutic decis~ons, either t~rough th~ use of appropriate doses of drugs and fluids, or with surgical interventions. The utilisation of ventricular volume ~aveforms to estl~ate ~troke volume in crit~cally ill patients provides information r~garding the effects of intravenous fluid ~7\s3000a.06 MARCH 20, 1989 C4 327 63 ~ ^

challenge, i.e., if intravenous fluids are given and cardlac output increases, then the therapeutic intervention probably is appropriate. On the other hand, if intravenous ~luid is administered and cardiac output remains the same or fall~, then the fluid challenge is probably inappropriate. This algorithm might diminish the utilization of invasive swan Ganz catheters placed in the pulmonary artery which are also used to ascertain whether fluid challenges are appropriate through cardiac output measurement and pulmonary arterial and left atrial ~wedge) pressure recordings.
The employment of such technology carries ma~or risks to the patient, including death, and to the health care worker the risk of viral hepatitis and AIDS because of exposure to blood products.
Non-invasive monitoring in accordance with the present invention poses no hazards to the patient nor health care workers while still providing similar hemodyna~ic information. The ventricular volume ~ave~orms along with the electrocardiogram can be obtained at the b~ds~de or trans itted to a video-based central station computeri~ed display for data processing either through hard wire connections or teleoetry.
Although cardiac output is an important parameter in quiding ~anage~ent of patients, trending of systemic oxygen delivery (D02) D~y be ~ ~ore valuable test. D2 i8 d~fined as the product of c~rdi~c output and arterial oxygen content. It signi~ies the ~olu~e of oxygenated blood delivered to the tissues. A fall in D2 ~7\53000A.06 ~ARCH ao, 1989 65 327631 ^
produced either by decreased cardiac output, decreased arterial oxygen content or both can cause tissue ischemia and tissue death.
since arterial oxygen capacity is a function of the amount of hemoglobin in the blood, viz. 1.34 ml of oxyqen can combine with 1 gm of hemoglobin, one can calculate oxygen content by multiplying the oxygen capacity of the blood by arterial oxygen saturation.
The latter can be obtained non-inva~ively by means of a commercially available device, the pulse oximeter. If hemoglobin content of the blood is stable, then relative changes in D2 can be obtained by multiplying arterial oxygen saturation by cardiac output. Thus, trends of D2 can be monitored non-invasively using pulse oximetry and TCG. -To illustratQ the importance f D2 measurements, consider the effects of breathinq a hypoxic mixture. It has been reported that breathing a hypoxic mixturQ ~F,02 ~ 0.1) for 7 to 20 minutes incraases heart rate (HR) 24%, stroke volume (SV) 16% and cardiac output tco) 38% compared to room air ~6 publications, 64 normals);
CO ~as ~easurQd by indicator dilution techniques. I extended such observations by ~dminis~ering graded hypoxio ~ixtures for 12 minutes, viz. Flo~ of .17, .15, .12 and .10 to 7 normals to ostablish dose-responsiveness for cardiac performance and oxygen delivery (DC~ ~ CaC~ x C0). SV and CO were measured with the thoracocardiograph (TCG). In addition, oxygen saturation (SaO2) fro~ pulse oxi~try, ~ection fraction IEF) from an eguation 47\53000A.06 HARCH 20, 1989 ~6 32763~ ^
involving PEP/L'VET, and minute ventilation (Vl) from RIP were obtained. The table below lists mean SaO2 and fractional changes of other parameters compared to Fl02 - .21 (SaO2 - 96%). In the table, an n*n denotes a statistically significant difference from Fl02 = .21.
S~ SaO2 ~B sv co ~E R-~R vl Do2 . 17 90* 1 . 03 l.ol 1.031.02 .97* 1.06 .98 ~.15 88* 1.05* 1.04 1.07l.oS* .96~ 6 99 lo .12 76* 1.19* 1.10* 1.30*1.08* .90* 1.22* 1 0~
. 10 67* 1 . 27* 1 . 19* 1.48*1.08* .87* 1.20* 1.~3 In the table, F,02 = fractional concentration of oxygen in gas mixture ~roo~ air = .21); SaO2 = arterial oxygen saturation; HR =
heart rato; sv = stroke volume; Co = cardiac output; ~F 3 e~ection fraction; R-PER ~ interval from R wave of EKG to peak e~ection rat~
on TCG ventricular volume curva ~Fig. 22): V~ ~ ~inute ventilation:
and DC~ ~ gystemic oxygen delivery. Change~ of HR, SV and co at F~02 ~ 0.1 agrae well with prior reported values. In normals, C0 ros~ proportionally 80 t~at DC~ wa8 maintained constant with brief graded decre Qnt~ of Sao~ Thi8 illustrates the importance of con~idcring D02 rat~er than Co alone. Not surprisingly, there were no untoward sycptQ~s in thesQ sub~ects despite falls in SaO2 to ~alues ~8 lo~ as 55t. Est~ation of D02 with decr~ased SaO2 in nor al and diseased states over prolonged time intervals needs lnvestigation sincc D02 ultimately dQtermines tissue viability.

47\53000A.06 NARC~ 20, 1989 67 ~ 327631 In contrast to grade hypoxia and terbutaline administration experiments, heal~-up tilting of normal sub~ects produces decreased cardiac output and decreased cardiac contractility. The amplitude of the TCG derived ventricular volume curve may not accurately reflect the fall in stroke volume owing to changes in the volume-motion coefficient of the rib cage with major chanqes of body posture as in changing from supine to upright postures. However, the configuration of the curve is altered in an expected way and provides useful information on contract~lity, viz. instead of a shortening of the R-P~R interval as in hypoxia and after terbutaline in~ection, head-up tilting causes the R-PER interval to lengthen, a finding consistent with decreased card~ac contractility.

The monitoring of trends in cardiac output during anesthesia using thQ non-invasivQ sensor placed upon the surface of th~ rib cagQ in patiQnts undergoing peripheral or abdo inal (i.e. non-chest rel~tQd) surqical oporations provides a valuable measurQ of cardiac perfor~ancQ. It i8 well known that anosthatic agents and surgical intQrvQntions often doletQriously affect cardiac output.

~valuation of appropriate cardiac pacing rates and tho effects Or different pacing sequencQs on stroke volumo is an important considoration in cardiac pacemaker therapy. This can be accomplis~ed by analysi# of beat to beat stro~o volume estimations ~7\53000A.06 MaRCH 20, 1989 68 1 32763~
from ventricular volume waveforms obtained with external sensors placed on the rib cage. In addition, control of optimal pacing rates through a servo loop can be accomplished by ~onitoring stroke volume to reset the pacing rate for optimal stroke volume performance during exercise. This has already been carried out on a research basis with an intràcardiac placed catheter for beat to beat changes of cardiac impedance.

Tha monitoring of ventricular volume curves should also be useful in evaluating changes of cardiac output in sub~ects confined to inaccessible environments such as the magnetic resonance imaging device, space capsules, diving bells, diving suits, high and low pressure chambers etc.

Measurements of stroke volume during various mechanical ~entilatory aodalities should help to establish mechanical ventilator settings which least deleteriously affect cardiac output~ The ventricular ~olum~ waveform ~asured with external sensor~ on the rib cage can be obtained during mechanical ventilation by the ensemble-averaging, curve fitting and other adap~ive digital filtering t~chniques as described above to extract thQ cardiac ~a~efore.

The Valsalva aneuver viz. straining with a closed glottis decreases stro~e volumo as shown above from measurements of the ~7\53000A.06 ~ -~ARCH 20, 1989 69 ` -^ " 1 327631 ~

ventricular waveform in a normal sub~Qct. The stroke volume normally increaæes after the straining maneuver 18 halted and the glottis is opened. Such a response may not occur ~n patients with heart disease and therefore the maneuver may help to differentiate normal subjects from patients with heart disease.

In addition to using t~e respiration signal for monitoring breathing patterns in babies with near SIDS (Sudden Infant Death Syndrome), monitoring of stroke volume and cardiac output from non-invasive determinations of ventricular volume curves as described abOVQ should aid in the early detection of cardiac abnormalities since it is known that bradycardia is orten associated with apneas in these babies.

SincQ thQ invention provides a mechanical indication of cardiac p~rformance, it uill be usQ~ul in est~blishing a timely diagnosis of death fro~ cardiac standstill Qven though electrical acti~ity of the heart ~ay still be present.

~0 Th~ rapidity of ~entricular emptying a~ a measure of ~yocardial contractility can be obtained as the slope of the v~ntricular volu~Q ~avQfore fro~ the external ~ensing device placsd on th~ rib cage during systole or by taking an electrical analog or digital derivati~e o~ this w~vefor~. The slope of rapid filling 2S ~or the ventricular volu~e curve at the end of isovolumetric .~:
~7\53000A.06 ~ARCH 20, 1989 70 1 327631 ~--relaxation provides a measure of the mechanical characteristics of ventricular muscle. The slope of late diastole provide~ a ~easure as to whether the heart is filled, has limited diastolic reserve, or has a great deal of diastolic reserve as indicated by a upward sloping deflection of this portion of the curve. All the situations discussed in the preceding sections, regarding cardiac output and stroke volume, apply for the importance of analyzing the configuration of the ventricular volume waveform to asssss cardiac performance.
The configurations of the ventricular volume and aortic pressure pulses may be abnormal in patients with heart disease at rest, exercise, sleep, and with environmental stresses, e.g.
temperature, humidity, etc. The waveform of the ventricular volume curv~ in patients with valvular heart disease has distinctive characteristics. For example, in patients with aortic stenosis the rate of systolic e~ection of the ventricular volume curve is diminished wherea~ in p~tients with mitral stenosis the rate of diastolic filling i8 diminished. The upstroke of the aortic pre8SUre pU18~ iS also di~inished in aortic stanosis. Patients ~ith coronary artery diseas~ may have limited ventricular wall motion due to ventricular compliance and have slow filling of diastole. Patients wit~ con~tricted pericarditis or restrictive myocardiopathy ~ay show diastolic plateaus as a result of these ~5 de~ects.

~7\53000A.06 HARCH 20, 1989 71 ` 1 327631 A long flat diastolic plateau has been observed in the ventricular volume curve obtained with the present invention in a patient with pulmonary edema, a pulmonary arterial wedge pressure s of 27 mmHg, and an enlarged heart on the chect roentgenogram. This type of waveform presumably indicates ventricular distention and might serve as a non-invasive monitor ô left atrial pressure in such patients.

Abnormial ventricular ~otion takes place with stunned myocardium after myocardial ischemia secondary to occlusion of a coronary vessel or with therapeutic angioplasty in which brief occlusion of the coronary artery supplying a region of ventricular ~uscle produces abnormal wall motion of this part. Indeed, abnormal wall function during myocardial ischemia precedes Qlectrocardiographic abnormalities and is a more sensitive diagnostic sign. Acute ~yocardial infarction produces abnormal ~ntricular volume waveiforms which may be reversed by ad~inistration of thro~bolytic agents. Thi~ phenomenon is best stud~ed ~ith segmental sen~ors over a large height of the rib cage r~ther than a ~ide band sensor enclosing the entire ventricle since s~al~ regions of abnormal motion might be missed under these circu~stanc . The configuration of the ventricular volume curve during the Valsalva maneuver in which systolic e~ection and ~troke volume are ~arkedly diminished in norm~l sub~ects, and is followed ~7\53000A.06 ~ARCH 20, 1989 72 ~ 32763~
by an increase o~ these parameters after release of straining, may not occur in patients with heart disease and thus offers criter~a for distingu~shing normals from patients with heart diseas~.
Furthermore, c~anging the configuration of the ventricular volume curve by tilting the sub~ect from the supine to upright postures and vice-versa produces characteristic alterations in thQ
configuration of the ventricular volume waveform. For example, in the standing posture, the terminal diastolic portion of the ventricular volume curve normally slopes upwards whereas in the supine posture terminal diastole has a flat plateau. This signifies that the heart is well filled in the supine but not the -upright posture, which might not occur in patients with heart disease.

With narrow band external sensors, ventricular volume ~avefor~s at different portions of the ventricle can be recorded suc~ that tim~ng and motion analysis between the segment~ can be carried out. This s~ould prove useful in asses~ing the effects of acute ische~ia and ~yocardial infarction on configuration of the ventricular volumQ ~avefor~ since it is well known that ventricular ~all ~otion i8 impaired in these circumstances. This can result in dyskinetic, akinetic or ~ypokinetic motion of segmental portions of the ventricular wall with consequent abnormalities of the segmental ventricular volu~e waveforms. ~sing the non-invasive nethod of the invention with sensor~ on the rib cage to display ~7\53000A.06 -~ARCH 20, 1989 73 , . . . ~ .

1 32763~ ~
segmental ventricular volume waveforms should make poss~ble the diagnosis of such abnormalities and to ascertaln the effectivenes~
of treatment either with intravenous administration of thrombolytic agents or angioplasty of the appropriate coronary artery.
Furthermore, long term periodic follow-up with the non-invasive technology o~ the invention should help in establishing the efficacy of treatment. For example, the effect of coronary artery bypass grafts on segmental ventricular volume curves can be determ~ned post operatively; if follow-up evaluations show new segments of abnormal wall motion different from the baseline establis~ed after surgery, then diagnosis of restenosis of the coronary artery might be suspected.

Analysis of segmental ventricular volume configuration with such interventions as card~ac pacing, exercisa, Valsalva maneuver, tilt, ~nd drug i~dministrat~on, etc. should enhanc~ its d~agnostic ffQctivQnQss. The effec~s of anesthesla agents on ventricular volu~e ~avefor s should halp to guid~ decisions on cardiovascular sti~tus during surgery. Finally, ambulatory Holter monitoring usin~ thQ olQctrocardiogr~p~ and segmental ventricular wavaform analy8i8 ~it~ sQparation of curves into` histograms of cardiac lQngths and QlQc~rical abnormaliti~s such as the ST-T wave deprQssions or lnversions can be utilised to correlat~ electrical and ~Qchanical Qvants during arrhythmias and periods o~ potent~al a5 ~yocardial i8che~ia.

47\53000A.06 NARCH ao, 1989 74 ~ 327~

In conjunction with invasive catheterization of the left ventricle, ventricular pressure-volume curves can be constructed to attain a definitive understanding of ventricular performance.

With an a~ray of external transducers placed on the abdomen of a pregnant wo~an and recording of the fetal electrocardiograph, it s~ould be possible to recognise and distinguish the sensor which contains the waveform of ventricular volume by ensemble-averaging or adaptive digital filtering methods. The latter techniques should aliminate maternal respiratoxy and cardiovascular pulsations leaving only th~ ventricular volume curves of the fetus. ~hls measure would help to diagnose fetal cardiac distress by display of both electrocardiographic and ventricular volume waveform muscle ~bnor-~litie~ and provido early id~ntification of fetal distress ~hich ~ight requira obstetrical interventions.

In con~unction ~ith tha ventricular volume curve, the analysls of the thoracic aortic and tho abdominal aortic pres~ura pulses should provide useful information on di~gnosis of valvular heart dise~s~ such as aortic stanosis and ~ convenient non-invasiva means -~
to follo~ tha outco~e aft~r surgical v~lvular repair. Thus, the upstro~o of thQ aortic pressure curve will diminish with aortic stenosis. Abnormal aortic pressure pulses occur with stable and ~7\S3000A.06 -~ARCH 20, 1989 75 ` 1 32763 1 ~
dissecting aneurysms of the thoracic and abdominal aorta and 8hould ~elp in establishing their diagnosis.

In sum, the utilization of the non-~nvas~va method of th~
invention for recording ventricular volume waveforms either globally or segmentally together with analysis of aortic pressure pulses is an important advance in clinical and research cardiology.
The electrocardiogram has served a highly useful purpose as an indicator o~ normal and abnormal electrical activity of the heartbeat, but provides no information on the mechanical responses to electrical activation. The invention described herein is the first known to continuously non-in~asivQly monitor mechanical performance of the heart by display of s~gmental characteristics.
It is also thQ first known invention to quantitativQly continuously ~onitor chanqes in stroke volume. Further, the same external transducer for card~ac ~onitoring can be utilized to non-invasiv~ly, continuously monitor the br~athing pattern. Several of th~ ~any applications that such a ~afe, non-invasive diagnostic tool ~ill acco~plish have been describ~d above. Obviously, many oth~r application~ ~ill com~ to mind in th~ futur~, and accordingly t~a abov description should ba constru~d as illustrative and not in a l~iting sens~, thQ 8COp~ of the invention being defined by th~ follow~ng cl~i~.

~\53000A.06 ~ARCH 20, 1989 76 ~

Claims (60)

1. A method for monitoring cardiac function in an animal or human subject comprising:
placing a first movement detecting transducer on the torso, said transducer overlying at least part of two diametrically opposed borders of the heart or great vessels;
generating a signal indicative of the movement of the torso portion subtended by the transducer, said signal including a cardiac component comprising at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform;
and assessing cardiac function by monitoring changes in said ventricular volume waveform or said aortic pressure pulse waveform.

2. The method of claim 1, wherein said movement detecting transducer comprises a conductor disposed on said torso portion for movement therewith, movement of said torso portion resulting in corresponding changes in the self-inductance of said conductor.

3. The method of claim 2, wherein said conductor extends about said torso portion and subtends a finite height.

4. The method of claim 3, wherein said height is about 2.5 cm.

5. The method of claim 1, comprising the step of removing a respiration component from said signal.

6. The method of claim 5, wherein said respiration component removing step comprises performing said assessing step during breathholding for avoiding changes in said signal due to respiration.

7. The method of claim 5, wherein said respiration component removing step comprises ensemble averaging said signal for removing the respiration component.

8. The method of claim 1, wherein said respiration component removing step comprises subtracting a curve fit from said signal for removing the respiration component.

9. The method of claim 1, wherein said respiration component removing step comprises adaptive digital filtering of said signal for removing the respiration component.

10. The method of claim 5, further comprising high pass filtering said signal for removing noise.

11. The method of claims 1 or 3, wherein said transducer is disposed at or near the xiphoid process and wherein said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform is a segmental ventricular volume waveform.

12. The method of claims 1 or 3, wherein said transducer is disposed at or near the uppermost portion of the sternum or the abdomen, and wherein said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform is a segmental aortic pressure pulse waveform.

13. The method of claims 1 or 3, wherein said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform is a segmental ventricular volume waveform, and wherein said assessing step further comprises monitoring the amplitude of said ventricular volume waveform for monitoring stroke volume.

14. The method of claims 1 or 3, wherein said assessing step further comprises monitoring changes in the slope, derivative of slope, or duration of said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform.

15. The method of claim 13, and further comprising monitoring the heart rate of said subject; and monitoring changes in cardiac output by monitoring changes in the product of said heart rate and said stroke volume.

16. The method of claim 13, and further comprising measuring the absolute value of stroke volume by an independent method, and adjusting the level of said signal to indicate said absolute value, whereby said signal indicates absolute stroke volume.

17. The method of claim 1, further comprising placing at least one additional movement detecting transducer on the torso, said additional transducer also overlying at least part of two diametrically opposed borders of the heart or great vessels; generating a signal indicative of the movement of the torso portion subtended by said at least one additional transducer, said signal generated by said at least one additional transducer including a cardiac component comprising at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform; and wherein said cardiac function assessing step comprises monitoring changes in said waveforms generated by said first transducer and said at least one additional transducer.

18. The method of claim 17, wherein said assessing step further comprises comparing the waveforms of said signals generated by said transducers.

19. The method of claim 17, wherein one of said movement detection transducers is selected as a reference, extracting the respiration components, of the signals generated by said transducers; adjusting the gain of the respiration component of each movement detection transducer other than said reference transducer to equal the gain of the respiration component of said reference transducer; and assessing cardiac function by comparing the amplitude of said cardiac component of said reference transducer to the amplitude of said cardiac component of at least one other transducer.

20. The method of claim 19, further comprising obtaining the relative amplitudes of said cardiac component of said signals generated by said transducers; repeating the steps of claim 19 on a known normal; obtaining the relative amplitudes of said cardiac component of said signal generated by said transducers when on said known normal; and comparing said relative amplitudes for said subject to said relative amplitudes for said known normal for assessing cardiac function of said subject relative to said normal.

21. The method of claim 1, wherein said movement detecting transducer is a bellows pneumograph, a mercury in silastic strain gauge, an inductive circumferential *TM

transducer, a differential linear transformer, or a surface inductive plethysmograph.

22. The method of claim 1, further comprising generating an EKG signal for said subject, and wherein said assessing step further comprises monitoring changes in the timing of said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform relative to said EKG.

23. The method of claim 15, further comprising monitoring arterial oxygen saturation; and monitoring systemic oxygen delivery (DO2) trends by monitoring trends in the product of cardiac output and arterial oxygen saturation.

24. The method of claim 1, further comprising measuring ventricular pressure; generating a signal indicative of ventricular pressure, wherein said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform is a segmental ventricular volume waveform; and constructing ventricular volume-ventricular pressure curves from said signals.

25. The method of claim 5, further comprising monitoring said respiration component.

26. The method of claim 1, wherein said subject is pregnant, and further comprising disposing said movement detecting transducer for detecting a cardiac component of said fetus; and removing a respiration component and the cardiac component of said subject from said signal.

27. The method of claim 26, further comprising removing said respiration component of said subject by monitoring during breathholding.

28. The method of claim 1, wherein said two diametrically opposed borders are the left and right borders of the heart.

29. The method of claims 17, 18, 19 or 20, wherein said transducers subtend an entire dimension of the heart.

30. The method of claim 29, wherein said dimension is the height of said heart from the most apical segment to the most basilar segment

31. An apparatus for monitoring cardiac function in an animal or human subject comprising:
a first movement detecting transducer for disposition on the torso, said transducer adapted for overlying at least part of two diametrically opposed borders of the heart or great vessels;
means for generating a signal indicative of the movement of the torso portions subtended by the transducer, said signal including a cardiac component 82 .

comprising at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform; and means for assessing cardiac function by monitoring changes in said ventricular volume waveform or said aortic pressure pulse waveform.

32. The apparatus of claim 31, wherein said movement detecting transducer comprises a conductor for disposition on said torso portion for movement therewith, with movement of said torso portion resulting in corresponding changes in the self-inductance of said conductor.

33. The apparatus of claim 32, wherein said conductor is adapted to extend about said torso portion and subtend a finite height.

34. The apparatus of claim 33, wherein said height is about 2.5 cm.

35. The apparatus of claim 31, further comprising means for removing a respiration component from said signal.

36. The apparatus of claim 35, wherein said means for removing the respiration component from said signal comprises means for ensemble averaging said signal.

37. The apparatus of claim 35, wherein said means for removing the respiration component from said signal comprises means for subtracting a curve fit from said signal.

38. The apparatus of claim 35, wherein said means for removing the respiration component from said signal comprises means for adaptive digital filtering said signal.

39. The apparatus of claim 35, further comprising means for high pass filtering said signal for removing noise.

40. The apparatus of claims 31 or 33, wherein said transducer is adapted for disposition at or near the xiphoid process and wherein said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform is a segmental ventricular volume waveform.

41. The apparatus of claims 31 or 33, wherein said transducer is adapted for disposition at or near the uppermost portion of the sternum or the abdomen, and wherein said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform is a segmental aortic pressure pulse waveform.

42. The apparatus of claims 31, or 33, wherein said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform is a segmental ventricular volume waveform, and wherein said assessing means further comprises means for monitoring the amplitude of said ventricular volume curve for monitoring stroke volume.

43. The apparatus of claims or 31 or 33, wherein said assessing means further comprises means for monitoring changes in the slope, derivative of slope, or duration of said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform.

44. The apparatus of claim 42, and further comprising means for monitoring the heart rate of said subject; and means for multiplying said heart rate by stroke volume for monitoring cardiac output.

45. The apparatus of claim 42, and further comprising independent means for measuring the absolute value of stroke volume; and means for adjusting the level of said signal to indicate said absolute value, whereby said signal indicates absolute stroke volume.

46. The apparatus of claim 31, further comprising at least one additional movement detecting transducer for disposition on the torso, said transducer overlying at least part of two diametrically opposed borders of the heart or great vessels; means for generating a signal indicative of .

the movement of the torso portion subtended by said at least one additional transducer, said signal generated by said at least one additional transducer including a cardiac component comprising at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform; and wherein said means for assessing cardiac function comprises means for monitoring changes in said waveforms generated by said first transducer and said at least one additional transducer.

47. The apparatus of claim 46, wherein said assessing means further comprises means for comparing the waveforms of said signals generated by said transducers.

48. The apparatus of claim 46, wherein one of said movement detection transducers is selected as a reference, further comprising means for extracting the respiration components of the signals generated by said transducers;
means for adjusting the gain of the respiration component of each movement detection transducer other than said reference transducer to equal the gain of the respiration component of said reference transducer; and means for assessing cardiac function by comparing the amplitude of said cardiac component of said reference transducer to the amplitude of said cardiac component of at least one other transducer.

49. The apparatus of claim 48, further comprising means for obtaining the relative amplitudes of said cardiac component of said signals generated by said transducers, whereby cardiac function may be assessed by comparing said relative amplitudes for said subject to relative amplitudes obtained with a known normal.

50. The apparatus of claim 31, wherein said movement detecting transducer is a bellows pneumograph, a mercury in silastic strain gauge, an inductive circumferential transducer, a differential linear transformer or a surface inductive plethysmograph.

51. The apparatus of claim 31, further comprising means for generating an EKG signal for said subject, and wherein said assessing means further comprises means for monitoring changes in the timing of said at least a segmental ventricular volume waveform or a segmental aortic pressure pulse waveform relative to said EKG.

52. The apparatus of claim 44, further comprising means for monitoring arterial oxygen saturation; and means for monitoring systemic oxygen delivery (DO2) trends by monitoring trends in the product of cardiac output and arterial oxygen saturation.

53. The apparatus of claim 31, further comprising means for measuring ventricular pressure; means for generating a signal indicative of ventricular pressure, wherein said at least segmental ventricular volume waveform or a segmental aortic pressure pulse waveform is a segmental ventricular volume waveform; and means for constructing ventricular volume-ventricular pressure curves from said signals.

54. The apparatus of claim 35, further comprising means for monitoring said respiration component.

55. The apparatus of claim 31, wherein said subject is pregnant, wherein said movement detecting transducer is adapted to be disposed for detecting a cardiac component of said fetus, and further comprising means for removing a respiration component and the cardiac component of said subject from said signal.

56. The apparatus of claim 31, further comprising means adapted to be worn by said subject for recording the transducer signal for accommodating ambulatory monitoring.

57. The apparatus of claim 46, further comprising means adapted to be worn by said subject for recording said transducer signals for accommodating ambulatory monitoring.

58. The apparatus of claim 31, wherein said two diametrically opposed borders are the left and right borders of the heart.

59. The apparatus of claims 45, 46, 47 or 48, wherein said transducers are adapted to subtend an entire dimension of the heart.

60. The apparatus of claim 59, wherein said dimension is the height of said heart from the most apical segment to the most basilar segment.