Heart 1996;76:326-331 ACE inhibitors unmask incoordinate diastolic wall motion in restrictive left ventricular disease Michael Y Henein, Aham Amadi, Christine O'Sullivan, Andrew Coats, Derek G Gibson Abstract Objective-To assess the effect of ACE inhibition on left ventricular filling and wall motion in patients with a clinical diagnosis of heart failure. Design-Prospective examination of left ventricular systolic and diastolic function using M mode echocardiography and pulsed and continuous wave Doppler before and three weeks after starting an ACE inhibitor. Setting-A tertiary referral centre for cardiac disease equipped with non-inva- sive facilities. Subjects-30 outpatients with a clinical diagnosis of heart failure in whom treat- ment with an ACE inhibitor was started; age 61 (SD 11) years; 27 male; 3 female; 21 healthy controls of similar age. Results-Left ventricular cavity was dilated both at end systole and end dias- tole, and fractional shortening reduced. Although mean isovolumetric relaxation time (IVRT) and transmitral E (early) to A (late) filling velocity (EIA) ratio were not different from normal, a value of 1.0 on the normal frequency plot of the E/A ratio divided the patients bimodally into two groups: 20 patients (group A) with E/A ratio > 1P0 and 10 patients (group B) < 10. In group A patients, IVRT was short as was transmitral E wave decelera- tion time compared to normal (P < 0.001), fulfilling the criteria of restrictive left ventricular physiology. Left ventricular wall motion during IVRT was coordinate and left ventricular end diastolic pressure was raised on the apex- cardiogram (P < 0-001). In group B, E wave deceleration time was longer, relax- ation incoordinate, and apexcardiogram normal. With an ACE inhibitor: in group A, left ventricular dimensions fell at end diastole (P < 0.05) and end systole (P < 0.01) but fractional shortening did not change; long axis total excursion (P < 0.01) and peak rate of shortening (P < 0.05) both increased; IVRT increased (P < 0.001) with the appearance of markedly incoordinate wall motion, minor axis lengthening, and long axis shortening (P < 0.001 for both); A wave amplitude also consistently increased (P < 0.001); finally, transmitral E wave velocity fell and A wave velocity increased. ACE inhibition did not alter any of the left ventricular minor and long axis or transmitral Doppler variables in patients in group B. Conclusions-Patients with a clinical diagnosis of heart failure differ in their presentation and response to ACE inhibi- tion according to baseline haemodynam- ics. In restrictive left ventricular phys- iology, ACE inhibition reduces cavity size and prolongs IVRT, compatible with a fall in left atrial pressure. At the same time, ventricular relaxation becomes very delayed and incoordinate, greatly reduc- ing early diastolic left ventricular filling velocity. Thus ACE inhibition unmasks major diastolic abnormalities in patients with restrictive left ventricular disease. (Heart 1996;76:326-331) Keywords: ACE inhibitor; restrictive left ventricular physiology; diastolic incoordination; heart failure Angiotensin converting enzyme (ACE) inhibitors have been shown to improve exer- cise tolerance and prolong life expectancy in patients with left ventricular disease.' 2 However, the nature of the disturbance of ven- tricular function in patients presenting with the clinical syndrome of heart failure is very variable and is itself likely to have a major effect on drug action. It was our purpose in the present study, therefore, to describe left ven- tricular pathophysiology in a group of patients thought to merit treatment with ACE inhibitors on the ground of a clinical diagnosis of heart failure and to determine its effects on these abnormally functioning ventricles, using non-invasive methods based on echocardiog- raphy. Patients and methods SUBJECTS We studied 30 outpatients in whom a clinical diagnosis of heart failure had been made and an ACE inhibitor was considered to be indi- cated. All the patients were symptomatic with dyspnoea, class II-IV (New York Heart Association), limiting exercise tolerance, along with cardiac enlargement on chest x ray, and ventricular cavity size-determined by echocardiography-increased to 5 8 cm or more (the upper 95% confidence limit of nor- mal) at end diastole. No patient had evidence of structural valve disease, pericardial disease, more than mild functional mitral regurgita- tion, or significant arrhythmia. They were all in sinus rhythm. The drug (captopril, Cardiac Department, Royal Brompton Hospital, London, United Kingdom M Y Henein A Amadi C O'Sullivan A Coats D G Gibson Correspondence to: Dr Derek G Gibson, Cardiac Department, Royal Brompton Hospital, Sydney Street, London SW3 6NP. Accepted for publication 3 May 1996 326 group.bmj.com on July 13, 2011 - Published by heart.bmj.com Downloaded from
Transcript
Heart 1996;76:326-331
ACE inhibitors unmask incoordinate diastolic wallmotion in restrictive left ventricular disease
Michael Y Henein, Aham Amadi, Christine O'Sullivan, Andrew Coats, Derek G Gibson
AbstractObjective-To assess the effect of ACEinhibition on left ventricular filling andwall motion in patients with a clinicaldiagnosis ofheart failure.Design-Prospective examination of leftventricular systolic and diastolic functionusing M mode echocardiography andpulsed and continuous wave Dopplerbefore and three weeks after starting anACE inhibitor.Setting-A tertiary referral centre forcardiac disease equipped with non-inva-sive facilities.Subjects-30 outpatients with a clinicaldiagnosis of heart failure in whom treat-ment with an ACE inhibitor was started;age 61 (SD 11) years; 27 male; 3 female; 21healthy controls of similar age.Results-Left ventricular cavity wasdilated both at end systole and end dias-tole, and fractional shortening reduced.Although mean isovolumetric relaxationtime (IVRT) and transmitral E (early) toA (late) filling velocity (EIA) ratio werenot different from normal, a value of 1.0on the normal frequency plot of the E/Aratio divided the patients bimodally intotwo groups: 20 patients (group A) withE/A ratio > 1P0 and 10 patients (group B)< 10. In group A patients, IVRT wasshort as was transmitral E wave decelera-tion time compared to normal(P < 0.001), fulfilling the criteria ofrestrictive left ventricular physiology.Left ventricular wall motion during IVRTwas coordinate and left ventricular enddiastolic pressure was raised on the apex-cardiogram (P < 0-001). In group B, Ewave deceleration time was longer, relax-ation incoordinate, and apexcardiogramnormal. With an ACE inhibitor: in groupA, left ventricular dimensions fell at enddiastole (P < 0.05) and end systole(P < 0.01) but fractional shortening didnot change; long axis total excursion(P < 0.01) and peak rate of shortening(P < 0.05) both increased; IVRT increased(P < 0.001) with the appearance ofmarkedly incoordinate wall motion,minor axis lengthening, and long axisshortening (P < 0.001 for both); A waveamplitude also consistently increased(P < 0.001); finally, transmitral E wavevelocity fell and A wave velocityincreased. ACE inhibition did not alterany of the left ventricular minor and longaxis or transmitral Doppler variables in
patients in group B.Conclusions-Patients with a clinicaldiagnosis of heart failure differ in theirpresentation and response to ACE inhibi-tion according to baseline haemodynam-ics. In restrictive left ventricular phys-iology, ACE inhibition reduces cavity sizeand prolongs IVRT, compatible with a fallin left atrial pressure. At the same time,ventricular relaxation becomes verydelayed and incoordinate, greatly reduc-ing early diastolic left ventricular fillingvelocity. Thus ACE inhibition unmasksmajor diastolic abnormalities in patientswith restrictive left ventricular disease.
(Heart 1996;76:326-331)
Keywords: ACE inhibitor; restrictive left ventricularphysiology; diastolic incoordination; heart failure
Angiotensin converting enzyme (ACE)inhibitors have been shown to improve exer-cise tolerance and prolong life expectancy inpatients with left ventricular disease.' 2However, the nature of the disturbance of ven-tricular function in patients presenting withthe clinical syndrome of heart failure is veryvariable and is itself likely to have a majoreffect on drug action. It was our purpose in thepresent study, therefore, to describe left ven-tricular pathophysiology in a group of patientsthought to merit treatment with ACEinhibitors on the ground of a clinical diagnosisof heart failure and to determine its effects onthese abnormally functioning ventricles, usingnon-invasive methods based on echocardiog-raphy.
Patients and methodsSUBJECTSWe studied 30 outpatients in whom a clinicaldiagnosis of heart failure had been made andan ACE inhibitor was considered to be indi-cated. All the patients were symptomatic withdyspnoea, class II-IV (New York HeartAssociation), limiting exercise tolerance, alongwith cardiac enlargement on chest x ray, andventricular cavity size-determined byechocardiography-increased to 5 8 cm ormore (the upper 95% confidence limit of nor-mal) at end diastole. No patient had evidenceof structural valve disease, pericardial disease,more than mild functional mitral regurgita-tion, or significant arrhythmia. They were allin sinus rhythm. The drug (captopril,
Cardiac Department,Royal BromptonHospital, London,United KingdomM Y HeneinA AmadiC O'SullivanA CoatsD G GibsonCorrespondence to:Dr Derek G Gibson, CardiacDepartment, RoyalBrompton Hospital, SydneyStreet, London SW3 6NP.Accepted for publication3 May 1996
326
group.bmj.com on July 13, 2011 - Published by heart.bmj.comDownloaded from
scribed were determined by the referringphysician. Echocardiographic measurementswere made before and at least three weeksafter the start of treatment. Between the twoechocardiographic studies other baseline drugswere not altered, including diuretics. Clinicaldetails, aetiology, and treatment are sum-
marised in table 1.Twenty one normal subjects, age 58 (SD
11) years, served as control. None had a his-tory of hypertension, shortness of breath, or
valvar or ischaemic heart disease.
PROCEDURESAll patients were studied at rest. Cross sec-
tional guided M modes were obtained using a
Hewlett Packard Sonos 1500 echocardiographwith a 2-5 MHz transducer interfaced to it.With the patient in the left semilateral posi-tion, M modes of the left ventricular minoraxis were recorded from the parasternal longaxis view with the cursor by the tips of themitral leaflets. M modes of mitral valve leafletsthemselves were also recorded from the sameview, with the cursor at the tips of the leaflets.Long axis M modes of the left and septal seg-
ments were recorded from the apical four-chamber view, with the cursor positioned atthe corresponding angles of the mitral ring.3All records were made during quiet expiration.Transmitral pulsed Doppler records of the for-ward flow velocities were obtained using thesame echocardiograph with the sample volumeat the tips of the mitral valve leaflets, in theapical four-chamber view. Continuous wave
Doppler recordings of mitral regurgitation wasdetected by colour flow and recorded with a
Doptek system (Colchester, UK), using a
2-5 MHz non-imaging transducer. Blood pres-sure was taken with a sphygmomanometer.The apexcardiogram was recorded usingCambridge equipment (time constant 4 s). Anelectrocardiogram and phonocardiogram were
superimposed on all traces. Studies were
recorded photographically at a paper speed of100 mm/s. M mode traces were digitised offline.4From the left ventricular minor axis M
mode trace we measured ventricular dimen-sions using leading edge methodology. Enddiastole was taken as the onset of the Q wave
of the electrocardiogram and end systole as
A2-the first high frequency vibration of theaortic component of the second heart sound-
on the phonocardiogram. Fractional shorten-ing was calculated as the fall in cavity dimen-sion divided by the end diastolic dimension(%). We also measured the total amplitude ofepicardial systolic excursion. Isovolumetricrelaxation was measured as the time intervalfrom the aortic component of the second heartsound to mitral cusp separation on the Mmode trace.From the long axis traces we determined
the total amplitude of excursion. The A wavewas taken as the amplitude of backward ringmotion (towards the atrium) after the P waveof the electrocardiogram. From the digitisedshort and long axis traces we measured peakrates of shortening (in systole) and of length-ening (in diastole) and the time intervalsbetween the Q wave of the electrocardiogramand peak shortening and between A2 of thephonocardiogram and peak lengthening. Fromthe pulsed Doppler traces of the transmitralforward flow velocity we measured peak earlyand late diastolic filling velocities-both fromthe baseline-and hence calculated E/A ratio.We also measured transmitral E wave deceler-ation time and the time interval between A2and the onset of transmitral early diastolic flowusing linear extrapolation to baseline of theacceleration phase. The presence or absenceof functional mitral regurgitation on continu-ous wave Doppler was noted. Systolic anddiastolic blood pressures were recorded andthe relative height of the left ventricular enddiastolic point with respect to the total ampli-tude on the apexcardiogram was measuredand expressed as a percentage.
STATISTICAL ANALYSISBaseline results from all patients are presentedas mean (SD). A normal frequency plot oftransmitral E (early) to A (late) diastolic leftventricular filling velocity ratio (E/A ratio) wasconstructed for the whole group of patientsand departure from a unimodal distributionsought. Baseline results from the two patientgroups defined on this basis were comparedwith normal controls using Student's unpairedt test. Individual results, before and after treat-ment with an ACE inhibitor, were comparedusing paired t test. A P value of less than 0 05was considered significant.
REPRODUCIBILITYThe reproducibility of the digitising processwas assessed in a sample of 20 patients fromduplicate determinations of long axis excur-sion, peak velocities, and timing with respectto the Q wave and A2. Within and betweenobserver values were determined indepen-dently. Reproducibility was assessed as rootmean square (RMS) difference between dupli-cate measurements, and the correspondingvalue of coefficient of variation as the ratioRMS difference/absolute value (table 2).
ResultsCLINICAL DETAILS (TABLES 1 AND 3)The patients studied proved to have a meanage of 61 years, and to be almost all males.
327
group.bmj.com on July 13, 2011 - Published by heart.bmj.comDownloaded from
Intraobserver InterobserverVariables CV (%) CV (%)Total excursion 3-2 4 0Shortening during isovolumetric relaxation 11 5 15Peak shortening rate 4-2 5 4Peak lengthening rate 5 0 6-0Q to onset of shortening 6-0 7 0A2 to onset of lengthening 5-5 6-0
CV, coefficient of variation.
Table 3 Overall left ventricular function
Variables Nornal (n = 21) Patients (n = 30)
Left ventricularEnd diastolic dimension (cm) 4-9 (0-5) 6-9 (1-2)End systolic dimension (cm) 3 3 (0-5) 5-8 (1 5)tFractional shortening (%) 30 (5) 17 (1O)tIsovolumetric relaxation time (ms) 60 (10) 45 (37)Mitral regurgitation (No of patients) 0 22
Transmitral E/A ratio 1-4 (0 4) 1-8 (1-7)
E, early; A, late diastolic left ventricular filling velocities.tp < 0001.
Coronary artery disease and its complicationswere the commonest diagnoses. End diastolicdimension, increased by definition, had amean value of 6'9 cm. In addition, end
.2....... ......................
ECG AV _' "~
PCG _usmIinApex cardiogram
Mitral
Minor axis
-~~~~~~~~~~T
Long axis
Before ACE-I With ACE-I
A composite representing different pressure, flow velocities, and wall motion changesoccurring in early and late diastole from a patient with dominant E wave beforeACEinhibitor and dominantA wave after. Note: (a) the drop in left ventricular end diastolicpressure on the apexcardiogram (top trace) after treatment was associated with a reductionin early diastolic and an increase in late diastoicfiUing velocities; (b) the increase in theisovolumetric relaxation time during which minor axis lengthened and long axis continuedto shorten all through early and mid diastole. ECG, electrocardiogram; PCG,phonocardiogram; A2, second heart sound, which was recorded on all traces and used toalign them.
systolic dimension-5-8 cm-was significantlyincreased and fractional shortening reducedto 17% with respect to normal. The majorityof the patients had mild functional mitralregurgitation as estimated by colour doppler.By contrast, mean values of isovolumetricrelaxation time and transmitral E/A ratio didnot differ significantly from normal, althoughthe scatter was wide. In order to investigatethe possibility that this wide scatter concealedmore than one discrete group within the sam-ple as a whole, we constructed a normal fre-quency plot of values of E/A ratio. Thisdeparted significantly from a unimodal distrib-ution [observed r = 0-871, expected r = 0949(n = 30) for P < 0 01]. The cut off valuebetween the two populations was 10, whichwe used to divide the patients into twogroups.
BEFORE TREATMENTMinor and long axis function and transmitralDopplerThese are shown in tables 4-6.
In the 20 patients (group A) in whom theE/A ratio was > 1 0, left ventricular cavity sizewas larger at both end diastole and end sys-tole, and fractional shortening was less thanthat in the remaining patients (group B).Though the timing and velocity of systolicshortening of both short and long axes weresimilar, there were major differences in dias-tolic function. Apart from the differences inE/A ratio, used to define the groups, E wavedeceleration time in group A was 75 (20) ms, avalue well within the range previouslydescribed as being characteristic of a restric-tive pattern of filling.5 That in group B wassignificantly longer, though still shorter thannormal. While left ventricular wall motion wascoordinate during isovolumetric relaxation ingroup A, minor axis increased and septal longaxis reduced in group B, indicating a changein left ventricular cavity shape. Finally, rela-tive height of the late diastolic A wave on theapexcardiogram was significantly increased ingroup A, but not in group B.
Effect ofACE inhibition in group AWith treatment, left ventricular dimensions atend diastole and end systole both fell signifi-cantly, though fractional shortening did notchange (table 4). Systolic total excursion andpeak rate of shortening of left ventricular longaxis both increased (table 5). Isovolumetricrelaxation time was greatly prolonged, andwall motion became incoordinate during thisperiod with an increase in left ventricularminor axis, and a fall in long axis, compatiblewith a change in cavity shape. Though therate of early diastolic left ventricular long axislengthening did not change, A wave ampli-tude consistently increased (table 5). Leftventricular filling pattern as shown by pulsedDoppler changed strikingly (table 6).Transmitral peak E wave velocity fell and itsdeceleration time increased. A wave alsoincreased so that the E/A ratio was consis-tently reduced. Finally, the relative height ofthe A wave on the apexcardiogram fell.
328
group.bmj.com on July 13, 2011 - Published by heart.bmj.comDownloaded from
EDD, end diastolic dimension; ESD, end systolic dimension; A2, second heart sound; BP, blood pressure.*P < 0 05; tP < 001; tP < 0001; ap < 0.05 v pre ACE; bp < 0.01 v pre ACE.
Table 5 Long axis function: comparison ofgroups. Values are mean (SD)
-, Abnormal shortening during IVRT; IVRT, isovolumetric relaxation.*P < 0-05; tP < 0-01; tP < 0-005; ITP < 0-001 v normal; ap < 0-05 v pre ACE; bp < 0 005 v pre ACE, cp < 0 001 v pre ACE.
Table 6 Transmitral Doppler and apexcardiogram: comparison ofgroups. Values are mean (SD)
E, early diastolic filling; A, late diastolic filling.*P < 0 05; tP < 0.01; tP < 0.001; ap < 0-001 v pre ACE.
Effect ofACE inhibition in group B (tables 4-6)With the same dose of ACE inhibition givenfor the same period of time, there was no sig-nificant change in any of the measurementsobtained from left ventricular minor or longaxes, transmitral Doppler, or apexcardiogramin the patients in group B.
Comparison between group A afterACEinhibitors and group B before ACE inhibitorsSince ACE inhibition led to the pattern of leftventricular function in patients in group Aresembling that seen under control conditionsin group B, we compared the two states for-mally (tables 4-6). This resemblance appliedto all measurements except left ventricular end
systolic dimension and fractional shortening,which were more reduced in group A, and freewall long axis, where total excursion(P < 0 01), peak shortening (P < 0 01), andpeak lengthening (P < 0 05) velocities werelower with more incoordination during isovol-umetric relaxation, P < 0 01.
DiscussionThe syndrome of heart failure is well recog-nised clinically, although formal definitionshave proved very unsatisfactory.6-9 Thepatients identified in the present study weresimilar to those described in western publishedreports as having the syndrome: they were
329
group.bmj.com on July 13, 2011 - Published by heart.bmj.comDownloaded from
elderly, predominantly males, and sufferingfrom the complications of coronary artery dis-ease. By definition, left ventricular cavity sizewas enlarged, and significant valve disease andrhythm disturbances were absent. Echo-cardiographic analysis of these patientsshowed, however, that they were not a homo-geneous group, but were bimodally distributedin terms of their diastolic function. The major-ity (group A) had an abnormally short isovolu-metric relaxation time, and a dominant Ewave, on transmitral Doppler, with a reduceddeceleration time.'0 In the remainder (groupB), isovolumetric relaxation time was long,wall motion incoordinate, E wave amplitudereduced or absent altogether, and the left ven-tricular long axis showed a series of major dis-turbances of relaxation." Since the effects ofany drug in disease are likely to depend on thebaseline characteristics of the patients towhom it is given, we elected to investigate thetwo groups separately.
In patients in group A, oral administrationof an ACE inhibitor had clear cut effects. Leftventricular cavity size fell with no change insystolic shortening velocities, isovolumetricrelaxation time increased, and transmitral Ewave amplitude fell. Left ventricular wallmotion during diastole became strikingly inco-ordinate, with complementary changes inmajor and minor ventricular axes during theisovolumetric relaxation period. At the same
time, the incidence of functional mitral regur-
gitation fell, and the amplitude of long axischanges during atrial systole increased. Bycontrast, in group B patients similar doses ofthe same drugs given for a similar period oftime caused no measurable effect on left ven-tricular haemodynamics. Not only were thepatients bimodally distributed with respect totheir baseline characteristics, therefore, butthis dichotomy was directly reflected in theresponse to treatment.
Several mechanisms have been suggested toexplain the beneficial effects ofACE inhibitionon exercise tolerance and prognosis. Thesehave included a reduction in left ventricularafterload,"2 lowering of left atrial pressure,"improvement of cardiac pump function result-ing from arteriovenous dilatation,'4 or remod-elling of the heart without improving intrinsicpump function.'5 We found no change in arte-rial blood pressure. At least part of theresponse seen in our study in the group Apatients can be explained in terms of a fall inleft atrial pressure. The very short values ofisovolumetric relaxation time present in con-
trol circumstances suggest that it was initiallyraised, and its increase, along with a change inthe pattern of transmitral flow from a domi-nant E wave to a dominant A wave, would becompatible with a substantial fall in early dias-tolic atrioventricular pressure gradient. A
reduction in left atrial pressure would alsoexplain the fall in left ventricular end diastolicdimension in the absence of any change inshortening rate. In addition, left ventricularlong axis relaxation became very incoordinate,with striking prolongation of tension, oftenthroughout early diastole and diastasis, until
the onset of atrial systole. At the same time theminor axis enlarged before any detectabletransmitral flow, suggesting that the termina-tion of wall tension within the myocardiumhad become asynchronous and that the cavityshape was changing.
These results might be explained in morethan one way. It is possible that ACE inhibi-tion has a primary action in inducing incoordi-nate left ventricular diastolic function, thusexplaining why isovolumetric relaxation timeincreased, the E wave was suppressed, andlong axis tension development was prolonged.However, there is no experimental basis forsuch an action. Alternatively, one can notethat disturbances to ventricular physiology ingroup A after ACE inhibition became verysimilar to those in group B before treatment,particularly with respect to isovolumetricrelaxation time and filling characteristics,though different in the extent. Further, ACEinhibition in the latter group was withouteffect on ventricular mechanics. It seemstherefore, that the underlying physiology in allpatients approximated to that seen in group B,but superimposed on this, in group A patients,was an additional factor sensitive to ACE inhi-bition. We cannot define this additional factorprecisely, but a major component of it appearsto be a high left atrial pressure. Increased enddiastolic ventricular stiffness and abnormali-ties of ventricular relaxation are both widelyheld to increase left atrial pressure. However,passive ventricular compliance-determinedlargely by myocardial collagen-is unlikely tochange with short term drug administration,while ventricular relaxation became even moreabnormal with ACE inhibition. Two addi-tional possibilities require more consideration.Mild functional mitral regurgitation was com-mon, though not invariable, in group Apatients, and its incidence declined somewhatwith treatment. In addition, left atrial pressuremay rise due to an increase in central bloodvolume and thus reflect the properties of theperipheral circulation, explaining its responseto a vasodilator. Regardless of the mechanismof the increase in left atrial pressure, webelieve that a high early diastolic atrioventricu-lar pressure gradient leads to an apparentlynormal pattern, not only of filling, as is wellrecognised,5 but also of wall motion, by over-riding disturbances of relaxation. Abnormalrelaxation only becomes apparent when it isunmasked as atrial pressure falls, although theeffects of these disturbances on ventricular fill-ing in patients with restrictive physiology canreadily be appreciated. The greater effects ofACE inhibition-seen in our study-com-pared with those of nitrates16 may reflect theeffect of chronic compared with acute adminis-tration, since it is established that their benefi-cial effects on exercise may require severalweeks to appear.'2Our study had obvious limitations. In order
to reflect the clinical use ofACE inhibition, wedid not fix drug or dose nor did we attempt toestablish rigid entry criteria to define heart fail-ure, since it was our aim to study both thenature of ventricular disease and drug use as it
330
group.bmj.com on July 13, 2011 - Published by heart.bmj.comDownloaded from
ACE inhibitors in restrictive left ventricular disease
occurs in cardiological practice. The dosesused were initial ones, and were thus low.Once a patient is established on the drug, it isstandard practice to increase the dose until anoptimal clinical response is achieved. Thismay explain persisting differences betweengroup A after treatment with an ACE inhibitorand group B. The study was an exploratoryone, and thus designed to be non-invasive. Itwould clearly have been of interest to havecorrelated direct measurements of left atrialpressure with the echocardiographic changesthat we have described. However, no mecha-nism other than raised left atrial pressure hasever been found to underlie an abnormallyshort resting isovolumetric relaxation time. Itwould also have been of interest to have deter-mined the time relations of the change in fill-ing pattern during the period between our twosets of observations and also to exercise toler-ance. In common with previous clinical trialsof the effects of ACE inhibition,217 18 we didnot consider patients with atrial fibrillation,although they form a significant proportion ofthose eligible for treatment. Finally, all ourpatients were already on diuretic treatmentand the dose was not altered for the study, sowe are unable to distinguish a primary effect ofACE inhibition on ventricular loading condi-tions from a secondary one potentiatingdiuretic action. Both, though, are likely tocontribute to the effects of ACE inhibition inclinical practice.Our observations seem to have a number of
general consequences. They illustrate the var-ied nature of left ventricular disease encom-passed within the clinical syndrome of heartfailure, and underline difficulties that arisewhen any rigorous definition is attempted.The dependence of the response to ACE inhi-bition on baseline physiology stresses the sig-nificance of allowing for the protean nature ofventricular disease in assessing drug effects. It isclear that an intervention documented asincreasing exercise tolerance and improvingprognosis may make measurements of left ven-tricular relaxation more abnormal. Whetherthe changes we observed contribute directly tothe known therapeutic effects of ACE inhibi-tion we cannot say, although we note that alow E/A ratio has been linked both to a greaterexercise tolerance'9 and a better prognosis20than a high one. We have previously noted thebimodal distribution of the E/A ratio.2' Its gen-esis is unclear, but its apparent stability seemsworthy of note. Simple control system theorysuggests a positive feedback mechanismwhereby the additional factor we haveassumed to be its basis aggravates the condi-tions that give rise to it in the first place, thusleading to a bistable system. Finally, ourresults raise the possibility that among thelarge numbers of patients with the clinical
diagnosis of heart failure, there may be consid-erable variation in response to ACE inhibition,a significant minority showing no effect at all.As long as ACE inhibition is thought to bewithout important side effects, this findingwould be of little practical consequence otherthan reducing the overall cost-effectiveness oftreatment, but should any clinically significantinteraction be demonstrated, for example withaspirin,22 then the case for more discriminatingdrug use would be clear.
1 Captopril Multicenter Research Group. A placebo-con-trolled trial of captopril in refractory chronic congestiveheart failure. JAm Coll Cardiol 1983;2:755-63.
2 The CONSENSUS Trial Study Group. Effects of enalaprilon mortality in severe congestive heart failure: results ofthe Cooperative North Scandinavian Enalapril SurvivalStudy (CONSENSUS). N Engl J Med 1987;316:1429-35.
3 Jones CJH, Raposo L, Gibson DG. Functional importanceof the long axis dynamics of the human left ventricle. BrHeart3 1990;63:215-20.
4 Gibson DG, Brown D. Measurement of instantaneous leftventricular dimension and filling rate in man, usingechocardiography. Br HeartJ 1 973;35:1141-9.
5 Appleton CP, Hatle LK, Popp RL. Demonstration ofrestrictive ventricular physiology by Doppler echocardiog-raphy. JAm Coll Cardiol 1988;11:757-68.
6 Lewis T. Diseases of the heart, 1st ed. London: Macmillan,1933:151.
7 Paul Wood. Heart failure. In: Diseases of the heart and circu-lation. London: Eyre and Spottiswoode, 1950:154-93.
9 The Task Force on Heart Failure of the European Society ofCardiology. Guidlines for the diagnosis of heart failure.Eur HeartJ 1 995;16:741-5 1.
10 Henein MY, Gibson DG. Abnormal subendocardial func-tion in restrictive left ventricular disease. Br HeartJt 1994;72:237-42.
11 Henein MY, Gibson DG. Suppression of left ventricularearly diastolic filling by long axis asynchrony. Br Heart Y1995;73:151-7.
12 Ader R, Chatterjee K, Ports T, Brundage B, Hiramatsu B,Parmley W. Immediate and sustained hemodynamic andclinical improvement in chronic heart failure by an oralangiotensin-converting enzyme inhibitor. Circulation1980;61:931-7.
13 Levine TB, Franciosa JA, Cohn JN. Acute and long termresponse to oral converting-enzyme inhibitor, captopril,in congestive heart failure. Circulation 1980;62:35-41.
14 Fitzpatrick D, Nicholls MG, Ikram H, Espiner EA.Haemodynamic, hormonal, and electrolyte effects ofenalapril in heart failure. BrHeart3 1983;50:163-9.
15 Baur LHB, Schipperheyn JJ, Baan J, Van Der Laarse A,Buis B, Van Der Wall EE, et al. Influence of angiotensinconverting enzyme inhibition on pump function and car-diac contractility in patients with chronic congestiveheart failure. Br HeartJY 199 1;65:137-42.
16 Stoddard MF, Pearson AC, Kern MJ, Ratcliff J, MrosekDG, Labovitz AJ. Influence of alteration in preload onthe pattern of left ventricular diastolic filling as assessedby Doppler echocardiography in humans. Circulation1989;79: 1226-36.
17 The SOLVD investigators. Effect of enalapril on survival inpatients with reduced left ventricular ejection fractionand congestive heart failure. N Engl J Med 1991;325:293-302.
18 The SOLVD investigators. Effect of enalapril on mortalityand the development of heart failure in asymptomaticpatients with reduced left ventricular ejection fractions. NEnglJ Med 1992;327:685-91.
19 Nishimura RA, Abel MD, Housmans PR, Warnes CA,Tajik AJ. Mitral velocity curves as a function of differentloading conditions: evaluation by intraoperative trans-oesophageal Doppler echocardiography. Y Am SocEchocardiogr 1989;2:79-87.
20 Shen WF, Tribouilloy C, Rey J-L, Baudhuin J-J, Boey S,Dufosse H, et al. Prognostic significance of Doppler-derived left ventricular diastolic filling variables in dilatedcardiomyopathy. Am HeartJ7 1992;124:1524-33.
21 Ng KSK, Gibson DG. Relation of filling pattern to dias-tolic function in severe left ventricular disease. Br HeartJY1990;63:209-14.
22 Cleland JGF, Bulpitt CJ, Falk RH, Findlay IN, OakleyCM, Murray G, et al. Is aspirin safe for patients withheart failure? Br Heart3 1995;74:215-9.
331
group.bmj.com on July 13, 2011 - Published by heart.bmj.comDownloaded from
