doi:10.1016/S0301-5629(02)00791-3

● Clinical Note

LEFT VENTRICULAR MYOCARDIAL FUNCTION IN CONGENITALVALVAR AORTIC STENOSIS ASSESSED BY ULTRASOUND TISSUE-

VELOCITY AND STRAIN-RATE TECHNIQUES

PETER KIRALY ,*† LIVIA KAPUSTA,* JOHAN M. THIJSSEN* and OTTO DANIELS**Children’s Heart Center, University Medical Center Nijmegen, Nijmegen, The Netherlands; and†Children’s

Department, Markusouszky Teaching Hospital, Szombathely, Hungary

(Received 13 September 2002; in final form 10 December 2002)

Abstract—A pilot study was performed to reveal the potentials of new echo Doppler techniques for the detectionof myocardial changes due to congenital valvar aortic stenosis. A total of 24 patients, (age range 0.1 to 17 years),with various degrees of aortic stenosis, and 24 age- and gender-matched, healthy children were enrolled in thisstudy. Conventional echo Doppler, tissue velocity imaging (TVI) and strain-rate imaging (StRI) measurementswere carried out using the apical four-chamber view and transthoracic long-axis view. All patients had normalfractional shortening of the left ventricle (> 28%). Although the sum of septal and ventricular wall thicknesseswas significantly increased in the patients (p < 0.001), only 6 of the 24 patients showed left ventricularhypertrophy. In tissue velocity mode, systolic and early diastolic wall velocity acceleration was significantlyreduced in both views. Peak systolic and early diastolic wall velocities, as well as strain rate values, in thefour-chamber view were significantly reduced in the patient group. The decrease was highest for the strain-ratevalues in all cases. In conclusion, strain rate values at different moments within the heart cycle might becomeimportant parameters in the assessment of myocardial impairment. Further studies are indicated to assess thecorrelation of these parameters with the severity of stenosis, left ventricular hypertrophy and irreversiblemyocardial function changes. (E-mail: L.Kapusta@cukz.umcn.nl) © 2003 World Federation for Ultrasound inMedicine & Biology.

Key Words: Congenital heart disease, Valvar aortic stenosis, Strain rate, Tissue velocity imaging, Ultrasound.

INTRODUCTION

Valvar aortic stenosis (Emmanouilides et al. 1995)causes an increased pressure overload to the left ventri-cle. The severity of stenosis is defined in clinical practiceby the pressure drop over the stenotic valve (pressuregradient) or by the calculated valve opening area. Tocalculate both indices, especially noninvasively, manypitfalls have to be encountered and assumptions have tobe made. Moreover, both parameters are dependent onthe hemodynamic status. So the question arises, if itwould not be more adequate to evaluate directly theimpact of aortic stenosis on the condition of the organthat suffers from this pressure overload (i.e., the myo-cardium).

Doppler measurement of myocardial wall velocities(tissue velocity imaging, TVI) was first proposed by

Isaaz et al. (1989). After the introduction of a 2-D colorversion of this technique (McDicken et al. 1992), itbecame the basis of many clinical applications and clin-ical studies. It promises more adequate measurement ofglobal and regional systolic function of the heart and,which is not less important, more proper assessment ofdiastolic dysfunction in different diseases (Price et al.2000; Nagueh et al. 2001; Vinereanu et al. 2001). Therecently developed strain rate imaging (SRI) technique(Heimdal et al. 1998) seems to add further details to theabove items by overcoming some of the limitations ofthe TVI measurements (Kowalski et al. 2001; Stoylen etal. 2001). In particular, the involvement of the globalheart movements in TVI measurements is greatlyavoided in SRI.

In adult patients with severe aortic stenosis, alter-ations in tissue velocity vs. time patterns are reported inboth systole and diastole (Lindstrom and Wranne 1999).However, myocardial reaction to pressure overload maybe quite different in children, especially of younger ages,

Address correspondence to: L. Kapusta, M.D., Children’s HeartCenter, University Medical Center Nijmegen, 6500 HB Nijmegen, TheNetherlands. E-mail: L.Kapusta@cukz.umcn.nl

Ultrasound in Med. & Biol., Vol. 29, No. 4, pp. 615–620, 2003Copyright © 2003 World Federation for Ultrasound in Medicine & Biology

Printed in the USA. All rights reserved0301-5629/03/$–see front matter

615

because myocardial cells have yet some multiplicationpotential and vascular growing can more flexibly followthe hypertrophy process. For these reasons, in children,the elasticity of the heart wall may be preserved and heartfailure will only develop after a relatively long time(Emmanouilides et al. 1995).

The aim of this pilot study is to evaluate whichtissue velocity and strain rate parameters of the myocar-dium have changed in children with aortic stenosis. Ul-timately, it would be important to look for a possiblerelation between these changes and the conventionalgrading of the severity of stenosis (pressure gradient,hypertrophy). This topic will be the subject of furtherstudy by the authors.

PATIENTS

Study populationA total of 24 patients (16 boys; 9 girls) with non-

corrected, isolated congenital valvar aortic stenosis vis-iting our outpatient clinic at the University Children’sHeart Centre Nijmegen, The Netherlands, were enrolledin this study. Patients with aortic regurgitation (greaterthan grade I) were excluded. The mean age of the pa-tients was 9 � 5 years (range 0.1 to 17 years), meanweight 35 � 21 kg (range 3.2 to 86 kg). As controls, 24age- and gender-matched healthy volunteers were re-cruited. All children were examined according to a strictprotocol, as described in the Methods section. The chil-dren and their parents gave informed consent. This pro-tocol had institutional review board approval.

METHODS

Standard echocardiographyAfter history taking, physical examination and

blood pressure measurement, every patient underwent adetailed standard echo Doppler cardiographic study. Theequipment used was a System Five model (GE/Vingmed;Horten, Norway) that allowed the digital storage of theraw data. Images were acquired with a 3.5-MHz phased-array transducer. Myocardial dimensions were estimatedby M-mode echography in the long-axis view; the frac-tional shortening was then calculated. Hypertrophy wasassessed by measuring the sum of the end-diastolic leftventricular posterior wall (LVPW) and interventricularseptum (IVS) thicknesses in parasternal long-axis view(sum larger than P95 of normal values at same body-weight; as in Kir�ly et al. 1997). The transvalvularpressure gradient at the aortic valve was calculated in theconventional way by the modified Bernoulli equation(e.g., Oates 2001).

For measuring the blood flow velocity in the as-cending aorta, a 2-MHz continuous-wave Doppler trans-ducer (“pencil” ) was used and applied suprasternally and

apically with subjects in left decubitus. For the pressuregradient estimation, the maximum obtainable value wastaken.

Tissue velocity imaging (TVI) of left ventricular wallsOn the same equipment, switching to TVI appli-

cation was available. We recorded colour-coded TVIcine-loops of three consecutive heart cycles with the3.5-MHz transducer from the parasternal long-axisand apical four-chamber views. Hence, two of thethree projections of the left ventricular wall movementvector, the radial and longitudinal ones, could succes-sively be measured. A frame rate of more than 80frames per s was used. The pulse repetition frequency(PRF) for the Doppler measurements was kept as lowas possible (1.5 to 2.0 kHz), resulting in a detectablevelocity range of up to 21 cm/s. A region-of-interest(ROI) of 1 � 1 pixels was used for off-line quantifi-cation of the raw data. This cursor was interactivelytracked through all the frames to keep the measure-ment site at approximately the same position through-out the heart cycle. This was achieved by visuallyusing an anatomical landmark (papillary muscle). Theraw data were digitally stored (by EchoPAC™ GE/Vingmed; Horten, Norway) and evaluated off-line.The images were retrieved for analysis after complet-ing the patient’ s examination.

In the long-axis view, we placed the cursor at theendocardial side of the LVPW, at the level of the tip ofthe mitral leaflet. In the four-chamber view, the morebasal part of the left ventricular free wall was evaluated,while keeping the cursor in the middle of the heart walland during the tracking in the stored cine-loop of images.The estimated time velocity waveform (i.e., single-gatedTVI vs. time waveform, see Fig. 1) was characterised bythree peaks: systolic (S), early (De), and late diastolic(Da) peaks. The descriptions of these and derived pa-rameters are listed in Table 1.

When possible, we used the built-in algorithm foraveraging the curves of the three heart cycles. However,the heart rate variability caused, in some cases, morethan 4% difference between the length of the differentQ–Q periods. In these cases, the averaging of the wholecurves would have masked the real shape of the curve(because the tops do not coincide) and, for this reason,the parameters were measured for each heart cycle sep-arately and the mean parameter values were calculatedthereafter. Finally, parameters obtained during a heartcycle with visibly low-quality images were left out ofthis averaging procedure.

Strain-rate calculations

Principle. The main limitation of TVI measure-ment is that it is not able to distinguish between the

616 Ultrasound in Medicine and Biology Volume 29, Number 4, 2003

local movements of the myocardium and the globaldisplacement of the heart within the chest. Strain ratecan be considered as the rate of tissue deformation.The velocity gradient (i.e., the strain rate) is defined as(Fig. 2) the velocity difference between two locationsat the myocardium divided by the distance betweenthem (Heimdal et al. 1998). According to this conven-

Fig. 2. (top, left) TVI velocity vs. time waveform obtained inlong-axis view. S � systolic peak, De � early diastolic peak.Vertical dashed lines drawn at peak times. (bottom, left) Con-ventional M-mode along scan line of TVI velocity measure-ment. Measurement boxes are indicated around peak times.(right, top and bottom) TVI M-mode traces from boxes (left);measurements of strain rates at systolic (SRs) and early dia-

stolic (SRe) time of peak wall velocities are indicated.

Fig. 3. Principle of strain rate measurement: velocity is mea-sured at two locations that are interactively tracked at fixedposition on the heart wall throughout the cardiac cycle, and that

are 10-mm apart.

Fig. 1. (left, bottom) Conventional 2-D echogram in apicalfour-chamber view; (left, top) same, now with color overlay oftissue velocity information. In both images: box in left ventric-ular free wall indicates measurement position of velocity vs.time waveform. (right) example of corresponding tissue veloc-

ity vs. time waveform. (right, bottom) ECG registration.

Table 1. Definition of symbols for tissue velocity and strainrate parameters

Category Definition

SRs and SRe Strain rate as measured at peak ofsystolic wave and of early diastolicwave; unit � 1/s

S, De and Da The tissue velocity values as measuredon the velocity/time curve (see Fig.1). S � the peak systolic, De �peak early diastolic, and Da � peaklate diastolic wave; unit � cm/s

Sacc, Deacc The tissue velocity value divided bythe time from the beginning to thetop of the wave. As a “goldstandard” , we considered as startand end of a wave where the curveleaves and reaches the zero line;unit � cm/s2

tS(%), tDe(%), tDa(%) The time from the beginning of theQRS-complex of the ECG to thetops of the individual waves dividedby the Q–Q duration; presented aspercentages of the Q–Q period; unit� %.

Sdur(%), Dedur(%), Dadur(%) The time duration of the whole S, Deand Da wave divided by the lengthof the Q–Q period, and presented aspercentages of the Q–Q period

Aortic valvar stenosis, TVI and SRI ● P. KIRALY et al. 617

tion, in the four-chamber view, shortening (duringsystole) is related to negative strain rate and length-ening (during diastole) to positive strain rate (longi-tudinal axis of velocity and strain rate). Using thelong-axis view, the same applies to thinning and thick-ening, respectively (radial axis).

Strain rate calculation for the left ventricular wall.In the parasternal long-axis view, first a single gatedtissue velocity vs. time trace was made in a ventricularwall (Fig. 2, top left). The systolic and early and latediastolic peaks were interactively indicated in thistrace. Then, the strain rate measurement was per-formed by using the colour-coded TVI M-mode imageof the left ventricular posterior wall, in which thetimes of the systolic, early and late diastolic peakswere placed by software means (Fig. 2, right). In thisM-mode, the velocities of the endocardial (vendo) andepicardial (vepi) sites of the LV wall were taken, alongwith precise measurement of the distance (�r) be-tween them. Strain rate was calculated as (vendo �vepi)/�r. In this calculation, a linear velocity gradientis implicitely assumed to be present, as was confirmedby Yamada et al. (1999).

In an apical four-chamber view, care was takenthat the left ventricular free wall was as parallel aspossible with the ultrasound (US) beam. However,because of the small size of the heart in children, theleft ventricular wall is considerably curved. This hasto be considered as a limitation of this method inpediatric patients, as far as the angle-dependency ofthe method is considered. We used the middle part ofthe left ventricular free wall for strain-rate measure-ments in the four-chamber view, which is generallyparallel to the US beam. We placed a curved line alongthe LV free wall, equidistant from the endo- andepicardial sides. The software then calculated aCAMM (curved anatomical M-mode) plot along thisline (Fig. 3). The velocity difference between twopoints along this line, at a distance of 10 mm (forinfants, 6 mm) from each other, was measured in thisM-mode colour tissue velocity plot at two instantsduring the heart cycle: at the moment of the peak ofthe S wave (SRs), and at the peak early diastole (De)wave (SRe). The strain rate at peak late diastolic (Da)wave (SRa) could very often not be estimated and wasnot incorporated in the statistical analysis.

StatisticsWe took the mean of parameters obtained from the

three heart beats of each cine-loop. The Wilcoxon’s logrank test was used for the comparison between the pa-tient group vs. the control group. A p value of 0.05 orless was considered statistically significant. Because a

relatively small group of patients was included in thispreliminary study, the relation between the velocity andstrain rate measurements with the severity of the aorticstenosis was not yet studied.

RESULTS

Conventional echo DopplerThe total of 24 patients could be divided into the

following subgroups, according to the pressure gradient-based severity grading (Keane et al. 1993): intermediate(n � 15), severe (n � 7), and critical (n � 2).

Median pressure gradient of the patients was 43.2mmHg, ranging from 20 to 132 mmHg. The sum of theend-diastolic thicknesses of the IVS and the LVPW wassignificantly increased in the patient group (p � 0.001).However, only six patients showed left ventricular hy-pertrophy. The fractional shortening was, for each pa-tient, higher than the normal value of 28%.

Tissue velocities and strain rate measurements (Table 2)

Systole. The peak systolic (S) tissue velocity in thefour-chamber view (logitudinal contraction) was signif-icantly reduced in the patient group (by 31%), but thelong-axis view (radial thickening) did not show signifi-cant change. The time-to-peak (tS%) value tended to belengthened in both views, but the difference was signif-icant only in the four-chamber view. The acceleration ofthe systolic wave (Sacc) was significantly reduced in bothviews for the patient group (by 38% in both). Strain ratein systole (StRs) was significantly lower in patients in thefour-chamber view, but showed no marked change in thelong-axis view. Comparing the results of systolic StRIand TVI, it appears that the highest level of significanceis found by strain rate, whereas the mean value for thepatient group is almost halved (by 49%).

Diastole. In the four-chamber view, the peak earlydiastolic strain rate (StRe) was significantly decreased inthe patient group. The mean decrease corresponds to41% of the normal value. In the patient group, the peakwall velocity in early diastole (De) and its acceleration(Deacc) were also reduced significantly in the four-cham-ber view (Table 2); the mean decrease was approxi-mately 35%. The De:Da ratio of TVI was found to bedecreased significantly in the patient group in the four-chamber view, which is consistent with the lower earlydiastolic wall velocity (De).

In the long-axis view, only the early diastolic ve-locity accelerations (Deacc) was found to be significantlyreduced for the patients.

No difference was found in any of the tissue veloc-ity and strain rate parameters between patients with orwithout cardiac hypertrophy (as assessed by M-modeechocardiography).

618 Ultrasound in Medicine and Biology Volume 29, Number 4, 2003

DISCUSSION

Summarising the findings above, we may state thatmany parameters provided by tissue velocity and strainrate techniques proved to be influenced by the pressureoverload of the left ventricle. The systolic and earlydiastolic peak wall velocities, their accelerations and thesystolic and early diastolic strain rate values tend todecrease, but the late diastolic (artrial contraction) peakvelocity and the measured time intervals proved not to bemuch influenced by pressure overload. We found thatparameters concerning the longitudinal wall movementsmeasured in four-chamber view were more affected inpatients with valvar aortic stenosis compared with theradial movement parameters measured in long-axis view.Furthermore, the decrease of the systolic strain rate pa-rameters of the patients group was higher and statisti-cally more significant than that of the TVI parameters.This finding underlines the intrinsically higher accuracyof strain rate measurement due to the neutralisation ofthe influence of the global heart movement (velocity).

In aortic stenosis, deterioration of the global systolicfunction of the left ventricle occurs mainly in severecases. None of our patients with severe or critical aorticstenosis showed a fractional shortening below the normalrange. However, reduction of systolic and early diastolicpeak wall velocity and its acceleration was observed forthe longitudinal vector component of the systolic heartwall movement. This finding was coherent with reportsof TVI mitral annulus displacement studies in aorticstenosis in adults (Lindstrom and Wranne 1999; n � 10).Similar findings were reported in TVI studies of systemicand hypertensive hypertrophy (Oki et al. 2000, n � 20;Stoylen et al. 2001, n � 26 and Vinereanu et al. 2001,n � 15).

The decrease of early diastolic peak wall velocityand strain rate might reflect a decreased ability forrelaxation of the left ventricular myocardium in thepatients with AS. In other words, the left ventricularwall is assumed to become stiffer due to aortic valvarstenosis. According to our findings, presence of LVhypertrophy had no relation to this observation. Thiscould indicate either that myocardial changes occurearlier than is detected by the conventional method, orthat wall velocity-related diastolic dysfunction (infour-chamber view) is not due to hypertrophy, butreflects transient (e.g., haemodynamic) changes or sec-ondary myocardial damage. The first possibility maybe substantiated by the finding that the sum of heartwall thicknesses of the patients is systematically ap-proximately 2 mm larger than of their controls, al-though their wall thickness increases significantly withincreased age. Further investigations based on age-compensated thickness are, therefore, indicated.

Currently, the authors are enrolling a greater numberof patients in the study and following those destined forsurgical or catheter intervention. In this way, it mightbecome possible to distinguish between reversible andirreversible myocardial changes, and to answer the ques-tions that this pilot study has raised.

The overall conclusion from this pilot study may bethat, in particular, the strain rate technique could becomea valuable tool for noninvasive detection and follow-upof myocardial function disturbance due to congenitalvalvar aortic stenosis.

Acknowledgments—This work was supported by a collaborative re-search grant by the Association for European Paediatric Cardiology(AEPC). The authors thank Martin van Wijk, MSEE, for his assistancein the statistical analysis of the data.

Table 2. Results of measurements in tissue velocity and strain rate waveforms (mean � SD)

Parameter Unit

Long axis view Four-chamber view

Controls Patients p Controls Patients p

SRs s�1 2.1 � 1.2 2.0 � 1.2 NS �4.7 � 1.1 �2.4 � 1.2 0.0001S cm/s 4.0 � 1.0 3.7 � 1.3 NS 9.0 � 2.1 6.2 � 2.4 0.0002Sacc cm/s2 75 � 51 46 � 19 0.04 192 � 74 118 � 90 0.002Sdur(%) % 38.0 � 6.3 38.5 � 6.1 NS 38.0 � 6.6 36.7 � 6.8 NStS(%) % 17.6 � 4.9 20.8 � 5.2 NS 13.8 � 4.5 17.6 � 6.7 0.03Sre s�1 �6.5 � 3.5 �6.2 � 3.4 NS 8.5 � 1.2 5.0 � 2.1 0.0001De cm/s �10.3 � 3.7 �9.1 � 4.2 NS �15.6 � 3.2 �10.1 � 4.0 0.0001Deacc cm/s2 225 � 89 156 � 72 0.003 216 � 32 140 � 50 0.0001Dedur(%) % 15.4 � 4.7 19.3 � 4.0 NS 17.9 � 2.9 17.8 � 4.1 NStDe(%) % 57.0 � 8.5 56.9 � 8.7 NS 61.2 � 9.5 57.2 � 10.0 NSDa cm/s �1.0 � 0.7 �1.3 � 1.4 NS �2.7 � 0.8 �3.4 � 1.5 NSDadur(%) % 10.0 � 4.5 10.5 � 4.5 NS 10.1 � 2.1 10.1 � 3.0 NStDa(%) % 82 � 24 91 � 4 NS 92 � 3 93 � 3 NSS/De – 0.4 � 0.2 0.6 � 0.6 NS 0.6 � 0.2 0.8 � 0.7 NSDe/Da – 19 � 15 14 � 11 NS 6.2 � 2.6 3.5 � 1.9 0.001

Definition in Table 1; NS � not significant (p � 0.05).

Aortic valvar stenosis, TVI and SRI ● P. KIRALY et al. 619

REFERENCES

Emmanouilides GC, Riemenschneider TA, Allen HD, Gutgesell HP,eds. Moss and Adams. Heart disease in infants, children and ado-lescents. Baltimore, MD: Williams and Wilkins, 1995.

Heimdal A, Stoylen A, Torp H, Skaerpe T. Real-time strain rateimaging of the left ventricle by ultrasound. J Am Soc Echocardiogr1998;11:1013–1019.

Isaaz K, Thompson A, Ethevenot G, et al. Doppler echocardiographicmeasurement of low velocity motion of the left ventricular wall.Am J Cardiol 1989;64:66–75.

Keane JF, Driscoll DJ, Gersony WM. Second natural history study ofcongenital heart defects. Results of treatment of patients with aorticvalvar stenosis. Circulation 1993;87(Suppl.1):16–27.

Kiraly P, Kapusta L, van Lier H, Otten-Hofman A, Daniels O. Naturalhistory of congenital aortic valvar stenosis: An echo and Dopplercardiographic study. Cardiol Young 1997;2:188–193.

Kowalski M, Kukulski T, Jamla F, et al. Can natural strain and strainrate quantify regional myocardial deformation? A study in healthysubjects. Ultrasound Med Biol 2001;27:1087–1097.

Lindstrom L, Wranne B. Pulsed tissue Doppler evaluation of mitralannulus motion: A new window to assessment of diastolic function.Clin Physiol 1999;19:1–10.

McDicken WN, Sutherland GR, Moran CM, Gordon LN. ColourDoppler velocity imaging of the myocardium. Ultrasound Med Biol1992;18:651–654.

Nagueh SF, Bachinski LL, Meyer D, et al. Tissue Doppler imagingconsistently detects myocardial abnormalities in patients with hy-pertrophic cardiomyopathy and provides a novel means for an earlydiagnosis before and independently of hypertrophy. Circulation2001;104:128–130.

Oates C. Cardiovascular haemodynamics and Doppler waveforms.London, UK: Greenwich Medical Media, 2001.

Oki T, Mishiro Y, Yamada H, et al. Detection of left ventricularregional relaxation abnormalities and asynchrony in patients withhypertrophic cardiomyopathy with the use of tissue Doppler imag-ing. Am Heart J 2000;139:497–502.

Price DJ, Wallbridge DR, Stewart MJ. Tissue Doppler imaging: Cur-rent and potential clinical applications. Heart 2000;84(Suppl. 2):11–18.

Stoylen A, Slordahl S, Skjelvan GK, Heimdal A, Skaerpje T. Strain rateimaging in normal and reduced diastolic function: Comparison withpulsed Doppler tissue imaging of the mitral annulus. J Am SocEchocardiogr 2001;14:264–274.

Vinereanu D, Florescu N, Sculthorpe N, et al. Differentiation betweenpathologic and physiologic left ventricular hypertrophy by tissueDoppler assessment of long-axis function in patients with hyper-trophic cardiomyopathy or systemic hypertension and in athletes.Am J Cardiol 2001;88:53–58.

Yamada H, Oki T, Tabata T, et al. Assessment of systolic left ventric-ular myocardial velocity profile and gradient using tissue Dopplerimaging in patients with hypertrophic cardiomyopathy. Echocar-diogr 1999;16:775–783.

620 Ultrasound in Medicine and Biology Volume 29, Number 4, 2003