Br HeartJ Changes myocardial - BMJ

Br Heart J 1992;67:368-76

Changes in myocardial echo amplitude duringreversible ischaemia in humans

David A Lythall, Derek G Gibson, Sudhir S Kushwaha, Michael S Norell,Andrew G Mitchell, Charles J D Ilsley

Department ofCardiology,Harefield Hospital,Middlesex UB9 6JHD A LythallS S KushwahaM S NorellA G MitchellC J D IlsleyDepartment ofCardiology, The RoyalBrompton andNational HeartHospitals,London SW3D G GibsonCorrespondence to:Dr David Lythall,Department of Cardiology,Harefield Hospital,Middlesex UB9 6JH.Accepted for publication19 September 1991

AbstractObjective-This study investigated the

changes in regional myocardial ultra-sonic backscatter, measured as myo-cardial echo amplitude, that occurduring reversible myocardial ischaemiain humans.Design-Left anterior descending

coronary angioplasty was used toproduce reversible myocardial isch-aemia in human subjects. Regionalmyocardial echo amplitude was studiedin the interventricular septum and leftventricular posterior free wall before,during, and after coronary occlusionwith the angioplasty balloon. Wallmotion analysis of the left ventriclewas performed from simultaneous crosssectional echocardiographic imaging.Patients were studied prospectively.Patients-Six patients (mean age 56

(SD 11), range 46 to 69 years) with singlevessel, left anterior descending coronaryartery stenoses, were investigated duringelective coronary angioplasty. A total of11 balloon inflations were studied.Setting-All patient studies were

performed at Harefield Hospital. Echoamplitude analysis was performed at theRoyal Brompton Hospital.Interventions-Angioplasty was per-

formed by the usual procedure atHarefield Hospital for elective coronaryangioplasty. All routine medicationincluding p blockers and calciumantagonists were continued. Inflationpressures were up to 12 atm (1212 kPa)and mean inflation time ranged from 30to 120 (86 (31)) s. In four studies the firstinflation was examined, in three thesecond, in two the third, and in one eachthe fourth and fifth inflations. Echoamplitude and cross sectional echo-cardiographic studies were recordedwith a 3-5 MHz Advanced TechnologyLaboratories (ATL) (720A/8736 series)mechanical sector scanner and an ATLMark HI (860-1 series) echocardiographsystem with 45 dB logarithmic grey scalecompression.Main outcome measures-Regional

echo amplitude was examined in fourregions of the left ventricle-namely, thebasal and mid-septum, and basal andmid-posterior wall. Consecutive enddiastolic and end systolic frames wereanalysed and cyclic variation wasdetermined as the difference between the

level of echo amplitude at end diastoleand at end systole. Measurements weremade before balloon inflation, at peakinflation, and after balloon deflation.Regional wall motion and systolic wallthickening were analysed qualitatively.Results-Before balloon inflation,

cyclic variation in echo amplitude wasnoted in all regions (basal septum, 2-4(SD 11) dB; mid-septum, 2 5 (1-1) dB;basal posterior wall, 3 3 (2 1) dB; mid-posterior wall, 3 9 (1-6) dB). Duringballoon inflation there was a significantfall in cyclic variation to 0 4 (0 9) dB(p < 0 0002) in the mid-septum. This waspredominantly owing to an increase inend systolic echo amplitude from 5 4(2 0) dB to 9 3 (1 9) dB (p < 0 01). Thiswas associated with the development ofsevere hypokinesis or akinesis in themid-septum. No significant changes inecho amplitude occurred in the threeother regions examined. Changes werecompletely reversed after balloondeflation.Conclusions-These results suggest a

causal relation between occlusion of thesupplying coronary artery and bluntingof myocardial echo amplitude cyclicvariation. It is suggested that balloonocclusion produced myocardial isch-aemia. The resultant impairment ofmyocardial contraction then caused ablunting of cyclic variation in echoamplitude. The results of this studyprovide further data about the abilityof quantitative studies of ultrasonicbackscatter to identify alterations in themyocardium duringinjury.

A link between cyclic variation in echoamplitude and cardiac contraction has beenshown.' To assess this relation further ahuman model of reversible myocardialischaemia was used to examine the effect of areversible perturbation of cardiac contractionon cyclic variation of echo amplitude.Percutaneous transluminal coronary angio-plasty of the left anterior descending arteryprovided an ideal model for studying leftventricular performance and function duringshort, controlled, and reversible episodes ofmyocardial ischaemia. Catheter ballooninflation performed during coronary angio-plasty produces virtually total interruption ofregional coronary blood flow. Numerousstudies during coronary angioplasty have

368

on May 3, 2022 by guest. P

rotected by copyright.http://heart.bm

j.com/

Br H

eart J: first published as 10.1136/hrt.67.5.368 on 1 May 1992. D

ownloaded from

http://heart.bmj.com/

Echo amplitude during reversible ischaemia

Table 1 Coronary angioplasty patient characteristics

Total number ofPatient Sex Age inflations studied

1 M 46 12 M 47 13 F 69 24 M 50 25 F 54 36 M 69 2

shown the effect of balloon occlusion on globaland regional left ventricular function usingeither M mode2 or cross sectional echo-cardiography,"5 direct cine left ventriculo-graphy,"5 and intravenous digital subtractionventriculography.9 Similarly left ventricularfilling dynamics have been studied duringcoronary angioplasty with high fidelitymanometer tipped catheters8 and Dopplerechocardiography.'° These studies suggestedthat coronary angioplasty provided an idealmodel in which to study the effect of reversiblemyocardial ischaemia on cyclic variation inecho amplitude. As left anterior descendingcoronary artery angioplasty had been shownto produce changes in segmental shortening inthe distal and mid-intraventricular septum,5cyclic variation in echo amplitude wasexamined in this region of the myocardium ina series of patients with single vessel leftanterior descending stenoses who were under-going coronary angioplasty.

Patients and methodsPATIENTSSix patients with single vessel left anteriordescending coronary artery disease, defined asgreater than 70% luminal stenosis, were inves-tigated during elective coronary angioplasty.No patient had a history of transmural myocar-dial infarction. Two patients were women andpatient age ranged from 46 to 69 years (mean(SD) 56 (11) years). All patients previously haddiagnostic coronary angiography and cine leftventriculography in the 300 right anterioroblique projection performed for the investiga-tion of angina. In all cases left ventriculographywas normal or showed minimal anteriorhypokinesia only. A total of 11 ballooninflations were studied in the six patients; threeinflations in one, two inflations in three, and oneinflation each in the remaining two. The studywas approved by the hospital ethical committeeand all patients gave informed consent. Care

Table 2 Characteristics of angioplasty balloon inflation, electrocardiographic changes,and pain

Inflation Duration of ST segment CardiacPatient No studied inflation (s) change pain

I 1 90 No No2 1 40 Yes Yes3 1 30 Yes Yes3 2 90 Yes Yes4 2 60 Yes No4 4 120 Yes No5 1 90 No No5 3 100 No No5 5 120 No No.6 2 90 No No6 3 120 No No

was taken not to interfere with or prolong thetherapeutic procedure. Table 1 shows charac-teristics of the patients.

PERCUTANEOUS TRANSLUMINAL CORONARYANGIOPLASTY PROTOCOLAngioplasty was performed by the usualprocedure for elective coronary angioplasty atHarefield Hospital. Patients were premedicatedwith intramuscular papavertetum and meto-clopramide. All routine medication including# blockers and calcium antagonists wascontinued.The angioplasty procedure was performed

via the femoral approach. An intravenoussingle dose of 5000 units of heparin wasadministered at the start of the procedure.Non-ionic contrast medium, steerable guidewires, and balloon catheters with diametersvarying from 2-5 to 3-5 mm were used. Inflationpressures were up to 12 atmospheres (1212kPa) and mean inflation time (86 (SD 31))ranged from 30 to 120 s. Table 2 shows that infour studies the first inflation was examined, inthree the second, in two the third, and in oneeach the fourth and fifth inflations.

Continuous electrocardiographic monitor-ing was performed throughout the procedurewith the six limb and augmented limb leads (I,II, III, aVL, aVF, and aVR) being monitoredby a 7 series Seimens Elema Mingograf. Theelectrocardiogram was recorded immediatelybefore and throughout each balloon inflationand after balloon deflation until any changeshad resolved.

ECHO AMPLITUDE AND CROSS SECTIONALECHOCARDIOGRAPHY STUDIESEcho amplitude and cross sectional echo-cardiographic studies were recorded with a3-5 MHz Advanced Technology Laboratories(ATL) (720A/8736 series) mechanical sectorscanner and an ATL Mark III (860-1 series)echocardiograph system with 45 dB logarith-mic grey scale compression. To obtain the leastmodified echo amplitude signal, self correctingcircuits were inactivated to avoid furtherdifferentiation of the original echo signal.Images were simultaneously displayed in greyscale and in colour by a Brompton Encoder(Alltek Hospital Supplies, Shepperton,Middlesex, UK) as they were recorded. On thecolour scale each level of echo amplitude wasrepresented as a single colour in the sequence:cyan, green, yellow, red, magenta, blue, andwhite, the last corresponding to maximumamplitude. The brightness of the displayedsignal was modulated so that within each colourlevel, the luminosity was varied. This approachallowed a continuous range of echo amplitudeto be displayed, and also meant that regions oflow echo amplitude appeared with low bright-ness on the screen. These were less con-spicuous on the final image and thus preservedthe original amplitude perspective."'

All images were acquired with a standardisedprotocol to enable intrapatient and interpatientcomparisons. Patients were studied in thesupine position and long axis parasternal viewswere recorded. Gain settings were standard-

369



j.com/

Br H


ownloaded from


Lythall, Gibson, Kushwaha, Norell, Mitchell, Ilsley

ised: the master gain was increased until theparietal pericardium posterior to the leftventricle was just displayed at the highest levelof the encoder (white). The depth compensa-tion was then set to a linear ramp across theimage at a rate of roughly 2 dB/cm. The gainsettings were not altered thereafter. Imageswere stored on U-matic 3/4 inch video tape in aSony VO 5800PS recorder.Image analysis was performed by a Tandon

PCA desk top computer. Stop frame enddiastolic and end systolic images for analysiswere selected by a video recorder with digitalframe advance (Sony VO 5800PS). Theseframes were transferred from tape to a512 x 256 x 6 bit frame store. Images wereprocessed with an adaptive filtering process forthe reduction of speckle as this decreased echoamplitude variance and allowed smaller samplesizes to be used.2 13 Four areas of interest weredefined, in the interventricular septum (basaland mid-septum), and posterior wall (basal andmid-posterior wall) as defined by the leftventricular anatomical nomenclature ofEdwards et al" and measurements of 256individual picture elements (pixels) were madefrom each area. Four separate measurements,each of 64 pixels, were made and averagedrather than one individual measurement of 256pixels as this allowed the pixel matrix to beplaced within the mid-myocardium and thusavoided including specular echoes from theendocardium or epicardium. A maximum pixelintensity for each image was derived bymeasuring the maximum colour intensity ofthesuperimposed colour scale (usually corres-ponding to 52 levels of grey scale). Thiscorresponded to a value of 45 dB and allowedecho amplitude values (dB) to be assigned (agrey scale level thus corresponding toapproximately 0-86 dB). Individual pixelamplitudes were displayed as a histogram and aseparate digital readout displayed mean,median, and SD values for each regionmeasured. The display of the frequencyhistogram on the image proved useful, as anobviously bimodal form usually meant thatspecular echoes had mistakenly been includedin the area of interest. Consecutive enddiastolic and end systolic frames were analysedand cyclic variation determined as the

Table 3 Mean end diastolic and end systolic echoamplitude values during percutaneous transluminarycoronary angioplasty (mean (SD))

Pre- Peak Post-inflation inflation inflation

Basal septum:ED (dB) 7-5 (1 9) 7-8 (1-4) 7-0 (1-3)ES (dB) 5-1(1-2) 5-9 (1-6) 4-6 (0.8)

Mid-septum:ED(dB) 84(22) 97(21) 79(21)ES (dB) 5-9 (2 0) 9 3 (1.9)* 5-4 (1-2)

Basal PW:ED(dB) 81(2-4) 7.9(17) 72(17)ES (dB) 4-7 (1-0) 5-0 (0-8) 4-2 (0 9)

Mid-PW:ED (dB) 9 1(2-0) 8 5 (2-3) 8 1(2 0)ES (dB) 5-2 (0-9) 5-6 (1-5) 4-6 (1-2)

*p < 0-0001 (ANOVA) for all four groups, p < 0-005 for allpaired groups (Bonferroni's) comparing the mid-septum toother regions. Basal PW, basal posterior wall; ED, enddiastole; ES, end systole; Mid-PW, mid-posterior wall.

difference between the level of echo amplitudeat end diastole and at end systole. Cyclicvariation index was calculated as the ratio:cyclic variation (end diastolic echo amplitudeminus end systolic echo amplitude): end dia-stolic echo amplitude.Images for analysis were recorded at baseline

with the deflated balloon and guide wire acrossthe coronary lesion, at the peak of ballooninflationand after balloon deflation. Inordernotto interfere with the therapeutic procedure andpatient safety and welfare, the timing of theimaging after deflation was by necessityvariable. Seven balloon inflation images wereacquired 15-30 seconds after deflation, but inthe remaining four inflations, images could notbe acquired until five, 10, 20, and 60 minutesafter balloon deflation. In one study images ofthe mid-posterior wall were not suitable foranalysis.

Regional wall motion analysis was per-formed qualitatively. Normal left ventricularwall motion was assessed by the amount ofsystolic myocardial thickening and inwardendocardial motion. Abnormalities of seg-mental wall function were identified by areduction in wall thickening and motion.

ResultsCLINICAL OUTCOME, CHEST PAIN, AND STSEGMENT CHANGESAll six patients underwent successful dilatationof their coronary stenoses. Significant STsegment shifts on the monitored electro-cardiographic leads occurred during fiveballoon inflations. This was accompanied bychest pain consistent with cardiac ischaemiaduring three inflations (table 2). On alloccasions ST segment changes and chest painresolved rapidly after balloon deflation.

WALL MOTION AND THICKENINGABNORMALITIESBefore balloon inflation no significant wallmotion abnormalities were present. Duringeach of the 11 balloon inflations, severehypokinesis or akinesis developed in the mid-septum and distal septum accompanied bydiminished or absent wall thickening. Obvioushyperkinesis or severe hypokinesis of theventricular wall was not seen elsewhere in themyocardium during any balloon inflation,except in the distal mid-posterior wall of twopatients (patients 1 and 2). In the seveninflations where imaging was obtained for thefirst 30 seconds after balloon deflation thehypokinesis or akinesis seen resolved withinthe 30 s after deflation. For the remaining fourinflations, resolution of the wall motionabnormalities was noted at the time of imagingafter deflation.

END DIASTOLIC AND END SYSTOLIC REGIONALECHO AMPLITUDETable 3 shows mean values for regional echoamplitude for all four regions of ventricularmyocardium at both end diastole and endsystole. Tables 4-7 show the individual valuesfor all patients. Values were greatest at end

370 on M

ay 3, 2022 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.67.5.368 on 1 M

ay 1992. Dow

nloaded from



Table 4 Individual echo amplitude (dB) at the basal septum for end diastole (ED)and end systole (ES) during angioplasty

Inflation Pre- Peak Post-Patient Variable studied n inflation inflation inflation

1 ED 1 62 88 63ES 34 71 47CVI 0-45 0 20 0-25

2 ED 1 5 8 5 9 6.8ES 46 40 44CVI 0.20 0-32 0 35

3 ED 1 9 3 10-5 63ES 6-1 8-7 5 2CVI 035 017 017

3 ED 2 8.4 7.6 8.7ES 5.4 5.3 4.4CVI 0 36 0 30 0 50

4 ED 2 5-2 6 1 5.9ES 3 2 5 4 3 6CVI 0-37 011 0 39

4 ED 4 5*4 8-0 4.8ES 4-1 7-6 3-8CVI 0 24 0 05 0.20

5 ED 1 116 90 7.9ES 70 64 38CVI 0 39 0-29 0 52

5 ED 3 8.6 8 1 9 0ES 5-8 5-7 62CVI 0-32 030 0 31ED 5 79 7.7 7.4ES 64 65 5.3CVI 020 015 028

6 ED 2 67 60 7.3ES 5.3 4.3 48CVI 021 029 034

6 ED 3 7.6 7-7 6.7ES 46 3-4 4.3CVI 040 0.55 035

CVI, cyclic variation index.

diastole, and least at end systole. There was nosignificant difference between pre-inflation orpost-inflation values at end diastole or endsystole for the four regions. End systolic valuesin the mid-septum at peak inflation, however,were significantly greater (p < 0 005) thanvalues for the other three regions of myo-cardium examined. For the mid-septal region,systolic values of echo amplitude were greatlyincreased by peak inflation compared with pre-

Table 5 Individual echo amplitude (dB) at the mid-septum for end diastole (ED) andend systole (ES) during angioplasty


1 ED 1 7-8 9-8 6-5ES 40 9.4 5-7CVI 049 004 0.11

2 ED 1 72 88 59ES 56 90 3-7CVI 0-22 -0-03 0-37

3 ED 1 118 10.9 10 2ES 86 87 61CVI 027 020 041

3 ED 2 9-5 11-4 9-8ES 5-7 110 5.9CVI 0-39 0 04 0-40

4 ED 2 50 56 53ES 3.3 6-1 3.4CVI 0-33 -0 09 0 35

4 ED 4 5-3 7-2 5-1ES 4-1 7-4 4-0CVI 0-23 -0 03 0-21

5 ED 1 9.5 11-3 90ES 6-9 11-4 62CVI 0-27 -0-02 0-31

5 ED 3 9-6 9-8 10-6ES 7-7 105 6-9CVI 0-20 -0 07 0-35

5 ED 5 11-2 13-0 9.9ES 9-4 12-4 6-6CVI 0-16 0.05 034

6 ED 2 7-2 10.0 70ES 5-5 9.4 5.7CVI 0-24 007 019

6 ED 3 90 89 8-0ES 4.3 7-5 5-2CVI 0-52 0-16 0-35


inflation (p < 0-01) After balloon deflation,values fell significantly (p < 0 01) and returnedto values not significantly different from pre-inflation values.

CYCLIC VARIATION IN ECHO AMPLITUDETable 8 shows mean values for cyclic variation.Tables 4-7 shows individual values for cyclicvariation index. Values for pre-inflation andpost-inflation were not significantly differentfor the four regions for either the absolute valueof cyclic variation of echo amplitude (enddiastolic echo amplitude minus end systolicecho amplitude) or the cyclic variation index.At peak inflation there was a significant fall inboth these variables in the region of the mid-septum (ANOVA, p < 0 0001) compared withboth pre-inflation (Bonferroni's (comparison ofpairs), p < 0 0002 and p < 0 01 respectively)and post inflation (Bonferroni's, p < 00002and p < 0 01) values. This major decrease incyclic variation was not seen in the other threeregions of myocardium examined (figure).

SUMMARY OF RESULTSDuring balloon inflation the mid-septum anddistal septum of the ventricle became severelyhypokinetic or akinetic. These changes wereaccompanied by significant changes in echoamplitude in the mid-septum. In this region anoticeable increase in end systolic echoamplitude occurred resulting in a fall in cyclicvariation in echo amplitude to 16% of pre-inflation values and a similar fall in cyclicvariation index to 10% of pre-inflation values.These changes paralleled the decrease in wallthickening seen qualitatively. After balloondeflation these changes reverted to pre-inflationvalues as did wall thickening.

DiscussionThis study documents the effects of reversiblecoronary artery occlusion on cyclic variation ofmyocardial echo amplitude. In the mid-septalregion ofthe left ventricular myocardium, afterballoon occlusion of the supplying left anteriordescending coronary artery, regional cyclicvariation in echo amplitude was much blunted,falling to values less than 16% of pre-occlusionvalues. Cyclic variation index, which relates thelevel of cyclic variation to end diastolic levels ofecho amplitude, also declined significantly toless than 10% of pre-occlusion values. Thesechanges were completely reversible afterballoon deflation suggesting a causal relationbetween occlusion of the supplying vessel andthe blunting of cyclic variation. It is postulatedthat balloon coronary artery occlusionproduced myocardial ischaemia in the territoryof the left anterior descending coronary artery.The resultant impairment of myocardial con-traction subsequently caused a blunting ofcardiac cycle dependent variation in back-scattered ultrasound. After balloon deflationand reperfusion these changes were rapidlyreversed with no permanent sequelae. Theevidence for this postulate is discussed further.

371



j.com/

Br H


ownloaded from



Table 6 Individual echo amplitude (dB) at the basal posterior wallfor end diastole(ED) and end systole (ES) during angioplasty


1 ED 1 72 7-6 6-7ES 4-2 6-5 4-6CVI 042 0-15 0-32

2 ED 1 72 8-5 60ES 46 44 41CVI 037 0-48 0-31

3 ED 1 6-6 6-8 79ES 5-1 4-7 3-6CVI 0-22 0-31 054

3 ED 2 6-9 70 6-9ES 4-2 4-5 3-6CVI 040 0-36 0-48

4 ED 2 50 5-7 47ES 3-6 39 32CVI 028 0-31 0-31

4 ED 4 58 58 4-5ES 39 40 33CVI 033 0-31 0-25

5 ED 1 11-7 11-4 90ES 7-1 50 4-6CVI 040 0-56 050

5 ED 3 10-7 9-1 93ES 5-1 5-4 4.3CVI 052 040 054

5 ED 5 11-8 10-0 88ES 40 5-7 60CVI 0-66 044 032

6 ED 2 6-9 7-1 7-4ES 5-7 5*7 5-6CVI 0-17 0-20 0-25

6 ED 3 8-4 8-2 8-7ES 4-1 5-2 3-6CVI 0-52 037 059


EVIDENCE FOR MYOCARDIAL ISCHAEMIAClinicalfindingsAn early manifestation of acute myocardialischaemia is the development of ST segmentchange on the electrocardiogram. Suchchanges were noted during five ofthe 1 1 balloonocclusions in this study. Previous studies inwhich limb lead electrocardiograms only weremonitored have shown similar findings. Hauseret al found ST segment shift in seven of 16patients undergoing left anterior descending

Table 7 Individual echo amplitude (dB) at the mid-posterior wallfor end diastole(ED) and end systole (ES) during angioplasty

Inflation Pre- Peak Post-Patients Variable studied n inflation inflation inflation

1 ED 1 7-5 6-0 6-2ES 4-2 6-2 40CVI 044 -004 035

2 ED 1 8-7 8-8 8-5ES 6-0 92 7-1CVI 033 -004 0-16

3 ED 1 8-9 76 88ES 5.5 4-9 58CVI 0-38 0-36 034

3 ED 2 8-2 70 7.5ES 46 4.5 4.4CVI 045 037 042

4 ED 2 * * *ES * * *CVI * * *

4 ED 4 62 56 41ES 4.7 4.5 4.4CVI 025 0-20 012

5 ED 1 123 12-5 115ES 6-9 6-2 4-8CVI 044 0.50 058

5 ED 3 112 98 90ES 5.7 64 36CVI 0.50 035 060

5 ED 5 10-3 9.9 86ES 5.0 5.9 3.9CVI 052 041 055

6 ED 2 7-4 67 7-6ES 5-2 52 5.4CVI 030 023 028

6 ED 3 10-4 107 93ES 4-1 3.7 38CVI 0-61 066 059

*Not analysable; CVI, cyclic variation index.

coronary angioplasty.' Visser et al found STsegment change in four of 15 patients under-going coronary angioplasty (10 single vessel leftanterior descending, four single vessel rightcoronary artery, and one double vessel rightcoronary artery and left anterior descendingcoronary artery).4 The use of 12 lead electro-cardiographic monitoring greatly increases thefrequency of ST segment change noted. In agroup of 32 patients with mixed single andmultiple vessel disease of whom 17 had angio-plasty of the left anterior descending coronaryartery, Labovitz et al analysed electro-cardiograms to find significant ST segmentchanges in 25 of 31.10 Wohlgelernter et al'detected ST segment changes in 12 of 14patients during angioplasty of the left anteriordescending coronary artery and Norell et al in33 of 52 patients.9 The use of intracoronaryelectrograms, recorded via the tip of the angio-plasty guide wire positioned distal to thecoronary stenosis being treated, greatlyincreased the yield of ST segment changes.'5 16In a study by Jain and Gettes ST segmentchange occurred in 33 of 35 patients under-going left anterior descending coronary angio-plasty but was simultaneously detected in only13 patients by observation of surface leads IIand V5.'16 The absence of ST segment change inover half (six of 11) of the balloon occlusionsexamined in our study does not thereforeexclude acute myocardial ischaemia duringthose occlusions. Similarly the occurrence ofchest pain during only three of 11 ballooninflations was consistent with that previouslyreported. Hauser et al' reported chest painin eight of 16 patients during left anteriordescending coronary angioplasty, and Visser etal4 in five of 15 patients although only 11 ofthese patients underwent left anterior descend-ing coronary angioplasty, the remainder beingright coronary artery lesions. The absence ofchest pain would appear to correlate poorlywith the absence of acute myocardialischaemia, chest pain being the last manifesta-tion of ischaemia.

Wall motion and systolic wall thickeningCoronary artery ligation in open chested dogscauses echocardiographically detectable sys-tolic wall thinning in the ischaemic regionwithin 30 s.17 18 In our study, decreased systolicwall thickening was noted in the distal and mid-interventricular septum during all 11 ballooninflations studied. Similar findings have beenwell documented by other workers. Hauser etal found new or increased wall motionabnormalities in 14 of 16 patients during leftanterior descending coronary angioplasty.'Hypokinesia was noted within 33 (19 (7)) s ofballoon occlusion in all 14 patients and invari-ably preceded ST segment shift or chest painwhere these occurred. Wall motion returned tonormal within 30 seconds ofballoon deflation in12 of these patients. Alam et al studied eightpatients with left anterior descending singlevessel coronary artery disease during 14balloon occlusions and noted a fall in systolicwall thickening of50% or more in the interven-tricular septum during 13 of these occlusions.2

372



j.com/

Br H


ownloaded from



Table 8 Mean (SD) cyclic variation in echo amplitude and cyclic variation indexduring angioplasty

p ValuePre-inflation Peak inflation Post inflation for all paired(dB) (dB) (dB) groups (Bonife

Cyclic variation in echo ampitudeBasal septum 2 4 (1 1) 1 9 (1 0) 2 4 (1-1) p < 0-005Mid-septum 2-5 (1 1) 0-4 (0-9)* 2-5 (1-2) p < 0-005Basal PW 3-3 (2-1) 2-9 (1-6) 3-0 (1-5) p < 0-005Mid-PW 3-9(1-6) 2 8(2-5) 3-5 (2 0) p < 0-005

Cyclic variation indexBasal septum 0 32 (0 09) 0-25 (0-14) 0-33 (0-11) p < 0-01Mid-septum 0 30 (0-12) 0-03 (0-09)* 0-31 (0 10) p < 0-01Basal PW 0-39 (0-14) 0-35 (0-12) 0-40(0-13) p < 0-01Mid-PW 042 (011) 0-30 (0-21) 0-40(0-18) p < 0-01

*p < 0-0001 (ANOVA) for all four groups, comparing the mid-septum with other regions.PW, basal posterior wall, Mid-PW, mid-posterior wall.

Wall thickening abnormalities began w20 s of balloon inflation in all 13 casesreturned to pre-occlusion levels within 20balloon deflation. Visser et al documesimilar wall motion abnormalities develoin 15 patients during 49 balloon inflationsmean of 8 (3) s after balloon inflatWall motion returned to baseline in all pat19 (8) s after balloon deflation. AWohlgelernter et al reported 14 patientsundergoing angioplasty of an isolatedanterior descending stenosis who all develIsevere hypokinesia (29%) or akinesia (71°the septal and apical region of the left ventby balloon inflation lasting 60 s.5 Recove:pre-occlusion values began within 15(range 6 to 22) and was complete in all patby 70 s after inflation with a mean time tcrecovery of 43 (17) s. Evidence that ischais the cause of the wall motion abnormarecorded echocardiographically was prosby techniques to improve myocardial perftdistal to the occluding angioplasty baltTimmis et al infused arterial blood througjangioplasty catheter into the distal coroartery in 10 patients during 60 s balocclusions.'9 These inflations were comp

*

06

05

x0)

C

040)

0-3.20

C 020)

0)

0-1

Basal PW Mid-septum Basal septum Mid-PV\

Mean (SD) cyclic variation index during coronary angioplasty. During balloon injin the left anterior descending artery there was a significant fall in echo amplitudevariation index in the mid-septal region only. After balloon deflation and the returtsystolic myocardial thickening, cyclic variation index was restored to pre-inflationvalues. *p, < 0-01; PW, posterior wall.

with control occlusions with either no distalinfusion or distal infusion with Hartman'ssolution. All dilatations without blood infusionin the 10 patients produced regional wall

troni) motion abnormalities within 16-0 (3-2) s. Bycontrast only one dilatation with distal bloodinfusion produced wall motion abnormalities.Infusion with Hartman's was not helpful inlimiting wall motion abnormalities suggestingthat oxygen delivery via arterial blood, ratherthan wash out of metabolites, was responsiblefor the benefit. Jaffe et al studied 42patients undergoing coronary angioplasty

Basal using distal transcatheter coronary perfusionwith oxygenated Fluosol DA (20% emulsion),a perfluorocarbon oxygen transport fluid.20

,ithin They found a dramatic fall in left ventricularandm regional wall motion abnormality during per-and fusion with oxygenated Fluosol DA compareds of with controls. It seems that the development ofnlted regional wall motion abnormalities during)ping coronary angioplasty not only reflects myo-4at cardial ischaemia but is also a sensitive and

tion. early marker of it. The development ofregionalients wall motion abnormalities during coronaryigain angioplasty in this study seem to reflecteach underlying myocardial ischaemia in theleft uoped territory supplied by the left anterior descend-

/O) of ing coronary artery.tricle

CYCLIC VARIATION IN REGIONALry to

BACKSCATTERED ULTRASOUND(5) s In 1983 and 1984, Madaras et al and Barzilai et:ients al first reported cyclic variation in backscattered) full myocardial ultrasound. 122 They analysedemia myocardial integrated backscatter as a measureLlities of backscattered ultrasound and examined thisiided in a dog model after occlusion of the leftasion anterior descending artery. Cyclic variation inloon, integrated backscatter was significantly bluntedhthey in ischaemic areas ofmyocardium. These areasloary were previously delineated at open thora-°°d cotomy during brief coronary occlusion. The

zare effects of reperfusion were not studied in thismodel. In a later study using a similar openchest dog model, cyclic variation in integrated

Post backscatter was examined 10 min aftercoronary occlusion and after two hours ofreperfusion. Cyclic variation at the leftventricular apex in the four dogs studied was5-6 (1-4) dB at baseline and fell to 0 4 (1-5) dBafter 10 min of occlusion (p < 0-02). After twohours of reperfusion, values were not signifi-cantly different from baseline at 3-9 (1-2) dB.Cross sectional echocardiographic apical fourchamber views of the apex showed dyskinesis

....... and loss of systolic thickening after 10 min of.... occlusion. This was substantially but not

.. completely restored to baseline values after twohours of reperfusion. Corresponding values forsystolic wall thickening at the apex were;...i-.-.-. baseline, 63-4 (17-7)%; 10 min occlusion, 2-2(11-5)% (p < 0-001); two hours reperfusion,52-9 (26-8)% (NS). Similar effects of ischaemia

v on cyclic variation of backscattered ultrasoundfrom dog myocardium after coronary artery

flation occlusion have been found by several otheryclic workers.2"25n of Cyclic variation in backscattered myocardial

ultrasound has also been found in humans.

373



j.com/

Br H


ownloaded from



This has been shown by integrated backscatter during balloon inflation. Backscatter fellscatter, 29 grey scale echo amplitude,30 and from 8 6 (3 3) units pre-inflation to 3-2 (1-7)colour encoded echo amplitude.' The effect of units at peak inflation (p < 0 001). They didischaemia on cyclic variation in backscattered not comment on whether this reversed afterultrasound in humans is limited. Milunski et al balloon deflation. They also noted an increase instudied 21 patients in a cardiac care unit within cyclic variation of backscatter in areas of myo-the first 24 hours after the onset of symptoms cardium that were identified as hyperkineticsuggestive of acute myocardial infarction.26 during balloon inflation. In our study theAreas of ischaemic injury were identified by grouped data did not show a significant increasecross sectional echocardiography and electro- in cyclic variation of backscattered ultrasoundcardiography. Cyclic variation of integrated in either basal segments of the left ventriclebackscatter was measured in these areas (where hyperkinesis would be most likely to

and normal control areas of myocardium. occur) during balloon inflation, althoughThrombolytic therapy was given to 17 patients. increases in cyclic variation index of > 10%Cyclic variation in normal regions was 4-8 (0-5) were noted in both areas during two of the 11dB compared with 0-8 (0 3) dB (p < 0-05) in inflations (tables 4, 6).regions with infarction. In the 15 patients who There are several explanations for thewere serially studied and also underwent changes in cyclic variation that occur incoronary angiography to define vessel patency, association with myocardial ischaemia.10 had patent infarct related coronary arteries. Wickline et al have suggested a mechanism toIn these patients the magnitude of cyclic varia- explain cyclic variation in backscatteredtion increased with time (1-3 (0 6) dB to 2-5 myocardial ultrasound (in their case integrated(0'5) dB, p < 0 05) where in those with backscatter).38 They suggested a relationoccluded infarct related vessels no significant between cyclic variation in integrated back-recovery was seen (0 3 (0 3) dB to 0-6 (0-3) dB, scatter and myocardial contractile functionNS). It was concluded that the restoration of which reflects cyclic alterations in myofibrillarcyclic variation in the patent vessel group elastic parameters, with the juxtaposition ofrepresented the beneficial effect of reperfusion. intracellular and extracellular elastic elementsSignificantly the improvement in cyclic varia- that have different intrinsic acoustiction shown in this group was not accompanied impedances providing an appropriately sizedby a noticeable improvement in wall motion. scattering interface at the cellular level.

It is probably incorrect, however, to com- Because acoustic impedance mismatch ispare the brief periods of coronary occlusion partially determined by elastic modulus,during clinical angioplasty with the prolonged changes in local elastic moduli which resultperiods of coronary occlusion in the study of from myocardial elastic elements not obeyingMilunski et al,2' and the animal studies already Hook's law as they are stretched, may alter thementioned in which coronary occlusion times degree of impedance mismatch. With a simplegreater than five min were used. Many inves- Maxwell type muscle model to depict cardiactigators have shown that although partial cell mechanical behaviour, a model employingrecovery of myocardial contractile function an extracellular parallel elastic element andafter brief ischaemia occurs promptly, intracellular series elastic and contractilecomplete recovery may require longer periods elements can be constructed. This modelof reperfusion."' With more prolonged predicts the decrease in backscatteredperiods of myocardial ischaemia (> 5 min) it is ultrasound found during systole. Cardiac cyclelikely that structural changes are occurring. dependent alterations in the degree of localMyofibril stretching after similar periods of acoustic impedance mismatch may thereforeischaemia are well described' as are other elicit concomitant changes in backscatteredultrastructural changes.35 Zhao et al found ultrasound. More recently Rijsterborgh et aldisruption of the collagen matrix of dog confirmed in animal studies that the highermyocardium with accompanying persistent systolic backscatter seen during ischaemiamyocardial dysfunction, after 12 sequential five could be explained by the decrease in wallmin occlusions of the left anterior descending thickness that occurs with ischaemia and not bycoronary artery.' This was unaccompanied by the ischaemia itself, in keeping with theirreversible cellular damage. Therefore it is not influence of myocardial contraction on cyclicsurprising that cyclic variation may not be fully variation.39 Other possibilities include altera-restored after several hours of reperfusion if tions in the relative insonifying angle due toischaemia is prolonged more than a few changing fibre angle and shape during theminutes with resultant damage to cellular and cardiac cycle as it is known that both reflectionextracellular structures. This contrasts with and attenuation of ultrasound alter accordingthe rapid recovery of both myocardial systolic to the direction of fibre insonification.4"function and cyclic variation in echo amplitude Whatever the exact basis of cyclic variation inthat occurred with coronary reperfusion, after backscattered ultrasound, there is no doubt an

balloon deflation during angioplasty, which this important relation with contractile events

study documents. occuring within the myocardium at either a

One other study, by Hajduczki et aP7 has cellular or fibre level. As myocardial ischaemiadocumented changes in backscattered ultra- occurs these contractile events are disruptedsound during coronary angioplasty in humans. and cyclic variation as a result is blunted.They examined patients during left anterior Changes in myocardial blood flow duringdescending coronary angioplasty and showed a coronary occlusion must also be considered as a

substantial change in cyclic variation of back- cause of the blunting of cyclic variation in

374



j.com/

Br H


ownloaded from



backscattered ultrasound, as changes in redblood cell concentration have been shown tonoticeably affect overall levels (as distinct fromcyclic variation) of backscattered ultrasound inexperimental models.4"A3 Wickline et al,however, in a series of experiments on perfusedand non-perfused dog hearts that werecyclically distended and relaxed by anintracavitary ventricular balloon, showed thatmyocardial perfusion with whole blood did notcontribute significantly to the phenomenon ofcyclic variation in backscattered ultrasound.38Importantly, in this study there was an increasein end diastolic echo amplitude in the mid-septum during balloon inflation from 8-4 (2 2)dB to 9.7 (2-1) dB that is significant (p < 005)if the mid-septum alone is considered. Thiswould be consistent with a fall in concentrationof formed blood elements in the myocardiumduring coronary occlusion.43

ConclusionIt is likely that balloon coronary occlusionduring angioplasty results in reversiblemyocardial ischaemia which produces areversible blunting of cardiac cycle dependentbackscattered ultrasound after temporary con-tractile dysfunction induced by ischaemia. Theresults of this study are consistent with thishypothesis and provide further data on theability of techniques involved in thequantitative study of backscattered ultrasoundto clarify alterations in the myocardium duringinjury.

DAL was supported as a junior research fellow by the BritishHeart Foundation, London, United Kingdom.

Lythall DA, Gibson DG,Ilsley CD, Yacoub MH. Changesin myocardial echo amplitude related to the cardiac cycle.Eur Heart J 1989;10(suppl):220.

2 Alamn M, Khaja F, Brymer J, Marzelli M, Goldstein S.Echocardiographic evaluation of left ventricular functionduring coronary artery angioplasty. Am J Cardiol 1986;57:20-5.

3 Hauser AM, Gangadharan V, Ramos RG, Gordon S,Timmis GC, Dudlets P. Sequence of mechanical,electrocardiographic and clinical effects of repeatedcoronary artery occlusion in human beings: echo-cardiographic observations during coronary angioplasty.JAm Coll Cardiol 1985;5:193-7.

4 Visser CA, David GK, Kan G, et al. Two-dimensionalechocardiography during percutaneous transluminalcoronary angioplasty. Am Heart J 1986;111:1035-41.

5 Wohlgelernter D, Cleman M, Highman HA, etal. Regionalmyocardial dysfunction during coronary angioplasty:evaluation by 2-dimensional echocardiography and 12lead electrocardiography. J Am Coll Cardiol 1986;7:1245-54.

6 Doorey AJ, Mehmel HC, Schwartz FX, Kubler W.Amelioration by nitroglycerine of left ventricularischaemia induced by percutaneous transluminal coronaryangioplasty: assessment by haemodynamic variables andleft ventriculography. JAm Coll Cardiol 1985;6:267-74.

7 Serruys PW, Wijns W, Van Den Brand M, et al. Leftventricular performance, regional blood flow, wall motionand lactate metabolism during transluminal angioplasty.Circulation 1984;70:25-36.

8 Piscione F, Hugenholtz PG, Serruys PW. Impaired leftventricular filling dynamics during percutaneous trans-luminal angioplasty for coronary artery disease. Am JCardiol 1987;59:29-37.

9 Norell MS, Lyons JP, Gershlick AH, et al. Assessment ofleft ventricular performance during percutaneous trans-luminal coronary angioplasty: a study by intravenousdigital subtraction ventriculography. Br Heart J 1988;59:419-28.

10 Labovitz AJ, Lewen MC, Kern M, Vandormael M,Deligonal U, Kennedy HL. Evaluation of left ventricularsystolic and diastolic dysfunction during transientmyocardial ischaemia produced by angioplasty. JAm CollCardiol 1987;lO:748-55.

11 Logan-Sinclair RB, Wong CM, Gibson DG. Clinicalapplication of amplitude processing of echocardiographicimages. Br Heart J 1981;45:621-7.

12 Bamber JC, Daft C. Adaptive filtering for reduction ofspeckle in ultrasonic pulse-echo images. Ultrasonics 1986;Uan):41-4.

13 Massey RJ, Logan-Sinclair RB, Bamber JC, Gibson DG.Quantitative effects of speckle reduction on cross sectionalechocardiographic images. Br Heart J 1989;62:298-304.

14 Edwards WD, Tajik AJ, Seward JB. Standardized nomen-clature and anatomic basis for regional tomographicanalysis of the heart. Mayo Clin Proc 1981;56:479-97.

15 Freidman PL, Shook TL, Kirschenbaum JM, Selwyn AP,Ganz P. Value ofthe intracoronary electrogram to monitormyocardial ischaemia during percutaneous transluminalcoronary angioplasty. Circulation 1986;74:330-9.

16 Jain A, Gettes LS. Patterns of ST-segment change duringacute no-flow myocardial ischaemia produced by balloonocclusion during angioplasty of the left anterior descend-ing coronary artery. Am J Cardiol 1991;67:305-7.

17 Pandian NG, Kieso RA, Kerber RE. Two-dimensionalechocardiography in experimental coronary stenosis. II.Relationship between systolic wall thinning and regionalmyocardial perfusion in severe coronary stenosis.Circulation 1982;66:603-1 1.

18 Pandian NG, Kerber RE. Two-dimensional echocardio-graphy in experimental coronary stenosis. I. Sensitivityand specificity in detecting transient myocardialdyskinesis: comparison with sonomicrometers. Circula-tion 1982;66:597-602.

19 Timmis AD, Crick JCP, Griffin B, Sowton E. Prevention ofregional myocardial ischaemia during PTCA-the effectsof infusing arterial blood. Br Heart J 1989;61:628 (abstr).

20 Jaffe CC, Wohlgelernter D, Cabin H, et al. Preservation ofleft ventricular ejection fraction during percutaneoustransluminal coronary angioplasty by distal transcatheterperfusion of oxygenated Fluosol DA 20%. Am Heart J1988;115:1 156-64.

21 Madaras El, Barzilai B, Perez JE, Sobel BE, Miller JG.Changes in myocardial backscatter throughout the cardiaccycle. Ultrasonic Imaging 1983;5:229-39.

22 Barzilai B, Madaras El, Sobel BE, Miller JG, Perez JE.Effectsofmyocardial contraction on ultrasonic backscatterbefore and after ischemia. Am JPhysiol 1984;247:478-83.

23 Wickline SA, Thomas LJ III, Miller JG, Sobel BE, PerezJE. Sensitive detection of the effects of reperfusion onmyocardium by ultrasonic tissue characterization withintegrated backscatter. Circulation 1986;74:389-400.

24 Sagar KB, Rhyne TL, Warltier DC, Pelc L, Wann LS.Intramyocardial variability in integrated backscatter:effects of coronary occlusion and reperfusion. Circulation1987;75:436-42.

25 Milunski MR, Mohr GA, Wear KA, Sobel BE, Miller JG,Wickline SA. Early identification with ultrasonicintegrated backscatter of viable but stunned myocardiumin dogs. JAm Coll Cardiol 1989;14:462-71.

26 Vered Z, Barzilai B, Mohr GA, et al. Quantitative ultrasonictissue characterization with real-time integrated back-scatter imaging in normal human subjects and in patientswith dilated cardiomyopathy. Circulation 1987;5:1067-73.

27 Milunski MR, Canter CE, Wickline SA, Sobel BE, MillerJG, Perez JE. Cardiac cycle dependent variation ofintegrated backscatter is not distorted by abnormal myo-cardial wall motion in human subjects with paradoxicalseptal motion. Ultrasound Med Biol 1989;15:311-7.

28 Milunski MR, Mohr GA, Perez JE, et al. Ultrasonic tissuecharacterization with integrated backscatter: Acute myo-cardial ischaemia, reperfusion, and stunned myocardiumin patients. Circulation 1989;80:491-503.

29 Masuyama T, Nellessen U, Schnittger I, Tye TL, HaskellWL, Popp RL. Ultrasonic tissue characterization with areal time integrated backscatter imaging systemin normaland aging human hearts. J Am Coll Cardiol 1989;14:1702-8.

30 Olshansky B, Collins SM, Skorton DJ, Prasad NV. Varia-tion of left ventricular myocardial gray level on two-dimensional echocardiograms as a result of cardiaccontraction. Circulation 1984;6:972-7.

31 Heyndrickx GR, Millar RW, McRitchie R, Maroko PR,Vatner S. Regional myocardial function and electro-physiological alterations after brief coronary arteryocclusion in conscious dogs. J Clin Invest 1975;56:978-85.

32 Kloner RA, DeBoer LWV, DarseeJ,IngwallJS, BraunwaldD. Recovery from prolonged abnormalities, of caninemyocardium salvaged from ischaemic necrosis bycoronary reperfusion. Proc Natl Acad Sci USA 1981;78:7152-6.

33 Weiner JM, Apstein CS, Arthur JH, Pirzada FA, Hood WB.Persistence ofmyocardial injury following brief periods ofcoronary occlusion. Cardiovasc Res 1976;10:678-86.

34 Jennings RB, Ganote CE. Structural changes in myo-cardium during-acute ischaemia. Circ Res 1974;34135(suppl III):156-72.

35 Kloner RA, Ganote CE, Whalen DA, Jennings RB. Effect ofa transient period of ischemia on myocardial cells.II. Finestructure during the first few minutes of reflow. Am JPathol 1974;74:399-422.

36 Zhao M, Zhang H, Robinson TF, Factor SM, SonnenblickEH, Eng C. Profound structural alterations of theextracellular collagen matrix in postischemic dysfunc-tional ("Stunned") but viable myocardium. J Am CoilCardiol 1987;10:1322-34.

37 HajduczkiI, Jaffe M, Areeda J, et al. Ultrasonic backscatterand 2D echocardiographic wall motion during PTCA(abstract). Circulation 1988;78:(supplII):442.

375



j.com/

Br H


ownloaded from



38 Wickline SA, Thomas LJ III, Miller JG, Sobel BE, PerezJE. A relationship between ultrasonic integrated back-scatter and myocardial contractile function. J Clin Invest1985;76:2151-60.

39 Rijsterborgh H, Mastik F, Lancee CT, Sassen LMA,Verdouw PD, Roelandt J, et al. The relative contributionsof myocardial wall thickness and ischaemia to ultrasonicmyocardial integrated backscatter during experimentalischaemia. Ultrasound Med Biol 1991;17:41-8.

40 Aygen M, Popp RL. Influence of the orientation of myo-cardial fibres on echocardiographic images. Am J Cardiol

1987;60: 147-52.41 Shung KK, Siegelmann RA, Reed JM. Scattering of

ultrasound by blood. IEEE Trans Biomed Eng 1976;23:460-7.

42 Shung KK, Reid JM. Ultrasonic scattering from tissues.Proceedings of the IGEE Ultrasonics Symposium 1977;77(CH1264-ISU):230-3.

43 Mimbs JW, Bauwens D, Cohen RD, O'Donnell M, MillerJG, Sobel BE. Effects of myocardial ischemia onquantitative ultrasonic backscatter and identification ofresponsible determinants. Circ Res 1981;49:89-96.

PLANTS IN CARDIOLOGY

362

q)

S:

FagarineIn 1932 G Stuckert was the first to isolate a newalkaloid, fagarine, from the Argentinian plantFagara coco (Rutaceae); and with A Sartori heshowed that it had a depressant action on themyocardium of rabbits. Further work at theUniversity of Cordoba by Moisset de Espanesand others showed that fagarine raised thethreshold for atrial and ventricular fibrillationin response to faradic stimulation, and that itdecreased the incidence of ventricular fibrilla-tion after coronary ligation in dogs. In allthese experiments it was more effective thanquinidine. ThenA Taquini tried its effect in sixpatients with atrial flutter or fibrillation whowere resistant to quinidine. In all of themintramuscular fagarine restored sinus rhythmwithin 30 minutes (Science 1945;102:69-70).In 1948 David Scherf showed that fagarinereliably reverted atrial fibrillation induced byaconitine in dogs (Proceedings of the SocietyforExperimental Biology and Medicine 1948;67:59-60).The genus Fagara has been merged with

Zanthoxylum and the name of the original Fcoco (Gill.) Engl. is now Z coco Gill ex Hookand Am. The illustration is of a closely relatedspecies. The South African plant Z capensis, a"fever tree". is used medicinally.The family Rutaceae is widespread,

especially in the tropics, and it includesPilocarpus microphyllus, the source of pilo-carpine. Citrus fruits belong to this family.Rutaceae is the fourth family of plantsdescribed in Plants in Cardiology with anti-arrhythmic properties-quinidine, procaine,and lignocaine all being derived from otherfamilies. It would be interesting to knowwhether fagarine is still under investigation.

A HOLLMAN

ZanthoxylumfagaraTomlinson TB. The biologyof trees native to tropicalFlorida. Allston: HarvardUniversity Print Office,1980:362.reproduced with permission

376



j.com/

Br H


ownloaded from


Documents

Br HeartJ Changes myocardial - BMJ