Detection MarfanVascular pathology in Marfan syndrome is caused by abnormal elastic properties68 ini-tially not associated with clinical symptoms. These abnormal properties are most

Heart 1996;75:307-31 1

Detection of abnormal aortic elastic properties inasymptomatic patients with Marfan syndrome bycombined transoesophageal echocardiography andacoustic quantification

Andreas Franke, Eberhard G Miihler, Heinrich G Klues, Katja Peters, Wolfgang Lepper,Gotz von Bemuth, Peter Hanrath

AbstractObjective-To evaluate the potentialvalue of transoesophageal echocardiogra-phy combined with automated borderdetection and acoustic quantification forthe assessment of elastic properties of thethoracic aorta in patients with Marfansyndrome.Subjects-16 patients with Marfan syn-drome and 12 age matched normal con-trols.Methods-Transoesophageal echocardio-graphy was performed in all subjects.Miniimum and maximum diameters ofthedescending thoracic aorta were obtainedfrom M mode images and acoustic quan-tification was used for the on-line evalua-tion ofcross sectional aortic area and peakpositive area changes over time.Compliance, distensibility, and stiffnessindex were calculated using M mode dataand non-invasively measured blood pres-sure and were compared with the indicesderived from acoustic quantification.Results-Aortic dimensions normalised forbody surface area were not statistically dif-ferent between patients and normal con-trols, but there were significant differencesfor all elasticity indices except compliance.Marfan patients had a lower distensibility[4.2 (SD 1.8) v 5 8 (2-1) cm2/dyn, P < 0-05]and a higher stiffness index [9.7 (3.0) v 7-1(1.8), P < 005]. The dynamic indicesderived from the acoustic quantificationwere significantly smaller in Marfinpatients [peak positive area change: 5.1(1.0) v 7-7 (1.7) cm2/s; P < 0-001; and nor-malised peak positive area change: 2-5 (1-2)v 4 0 (0.8) cm2/s respectively, P < 0.001] andwere suitable to discrnninate between nor-mal and abnormal elastic properties.Conclusions-In Marfan syndrome elasticproperties ofthe descending aorta are sig-nificantly different from normal controls,even in the absence ofvessel dilatation. Inaddition to established static indices,indices derived from acoustic quantifica-tion reflect dynamic changes of the crosssectional area for the evaluation ofregional vessel mechanics. The on-lineassessment of peak positive area changeallows differentiation from normal indi-viduals and may be more accurate thanstandard M mode measurements.

(Heart 1996;75:307-31 1)

Keywords: Marfan syndrome; transoesophagealechocardiography; acoustic quantification; aorticelastic properties

Prognosis in Marfan syndrome is stronglyinfluenced by cardiovascular complications.The majority of all deaths in this patientgroup-nearly 75%-result from aneurysms,dissection or rupture of the thoracic orabdominal aorta, or cardiac failure caused bydilatation of the aortic root and concomitantaortic valve regurgitation."-3 Life expectancyhas improved during the last 20 years becauseof successful cardiovascular surgery in many ofthese patients.45

Vascular pathology in Marfan syndrome iscaused by abnormal elastic properties68 ini-tially not associated with clinical symptoms.These abnormal properties are most likely tobe caused by defective synthesis, secretion,and extracellular matrix formation of the fib-rillin protein, which is widely distributed inelastic tissues.9 The gene coding for fibrillinhas been identified on chromosome 15 and isthe site of mutations causing Marfan syn-drome.9 10A non-invasive method for the evaluation of

elastic properties with reproducible resultsshould be clinically helpful in Marfan patients,especially before morphological changesoccur. Transthoracic echocardiographic visu-alisation of the aorta is limited to the ascend-ing aorta and parts of the aortic arch."I Areproducible short axis plane of the ascendingaorta cannot usually be achieved. Adequatevisualisation of the descending thoracic aortaby the transthoracic approach cannot beobtained. Transoesophageal echocardiogra-phy, however, allows optimal visualisation ofthe thoracic aorta in transverse planes withgood image quality" providing detailed infor-mation of heart cycle dependent changes ofaortic diameter and cross sectional area.

This study was performed to assess thepotential value of transoesophageal echocar-diography in combination with automatedborder detection and on-line acoustic quantifi-cation in evaluating the elastic properties ofthe descending thoracic aorta in Marfanpatients and normal controls, in comparisonwith standard measurements of elastic proper-ties (compliance, distensibility, and stiffness

Medical Clinic I andPaediatric Cardiology,Rheinisch-WestfbilischeTechnischeHochschule Aachen,Aachen, GermanyA FrankeE G MiihlerH G KluesK PetersW LepperG von BernuthP HanrathCorrespondence to:Andreas Franke MD,Medical Clinic I (Dept ofCardiology), RWTHAachen, Pauwelsstr 30, D52057 Aachen, GermanyAccepted for publication25 July 1995

307

on January 11, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/H

eart: first published as 10.1136/hrt.75.3.307 on 1 March 1996. D

ownloaded from

Franke, Mfihler, Klues, Peters, Lepper, von Bernuth, Hanrath

index) derived from diameters measured by Mmode echocardiography.

MethodsTwenty four consecutive patients with thediagnosis of Marfan syndrome were studied.All fulfilled the diagnostic criteria for Marfansyndrome. 12-14

Eight of the 24 patients were excluded fromthe study because there were significant mor-phological abnormalities of their nativedescending aorta (two aneurysms, one dissec-tion), because of prosthetic replacement (fourpatients), or because of severe aortic regurgita-tion (one patient). The remaining 16 patientshad a native descending thoracic aorta withoutmorphological abnormalities and formed thefinal study group. The demographic data of allthe patients included and the normal controlsare given in the table. At the time of the studynone of the patients was on cardiovascularmedication or had mitral or aortic regurgita-tion. Two patients had previous cardiovascu-lar surgery with replacement of the aortic valveand the ascending aorta. Thirteen of the 16Marfan patients had a family history of Marfansyndrome.

Demographic and haemodynamic variables of 16 Marfan patients and 12 normal controls.Values are mean (SD) or (range).

Normal subjects Marfan patients P value

n 12 16Sex 5m,7f 7m,9f NSAge (years) 27 (9) (range 20-57) 28 (13) (range 12-57) NSHeight (cm) 174 (6) 188 (9) <0 001Weight (kg) 58 (19) 71 (21) NSBody surface area (m2) 1 77 (0 1) 2-00 (0-17) <0 001Functional classNo limitation 12 14NYHAII 0 2

Heart rate (beats/min) 77 (8) 80 (17) NSBlood pressure (mm Hg)

Systolic 113 (14) 120 (15) NSDiastolic 64 (11) 68 (12) NSMean 80 (12) 86 (12) NSPulse pressure 49 (7) 52 (10) NS

NYHA, New York Heart Association

Figure 1 Cross section of the descending aorta with automated border detection. The areawithin the automatically detected intimal border marked by the thin line is measured on-line and plotted as a graph.

All patients underwent a complete clinicalexamination as well as transthoracic and (aftergiven written informed consent) trans-oesophageal echocardiography (HewlettPackard Sonos 1500, with a 2-5 MHz trans-thoracic and a 5 MHz transoesophageal multi-plane transducer). All echocardiographicstudies were stored on videotape for furtherquantitative off-line analysis.The descending aorta was imaged by trans-

oesophageal echocardiography in an optimisedcross sectional strictly circular plane at thelevel of the left atrium. Great care was taken tovisualise the complete circumference of thedescending thoracic aorta throughout thewhole cardiac cycle.Minimum and maximum diameters of the

descending aorta were determined by conven-tional M mode echocardiography. Non-invasiveoscillometric blood pressure measurements atthe right upper limb were performed every 2min and the average of five measurements dur-ing M mode and acoustic quantification wasused for further calculations. The following vari-ables were calculated based on minimum andmaximum diameters and blood pressure.15-'7

Compliance (cm2/dyn)

C Amax -AminBPmean

Distensibility (cm2/dyn)

D - 2-(dmax -dmin)dmin- (BPsyst- BPdiast)

Stiffness index

SI-ln(BP- Y _BPdiast)(dmax dm.n)/dminAmax; Amn = maximum and minimum crosssectional area.dmax; dmin = maximum and minimum diameter.BPsyst; BPdiast; BPmean = systolic, diastolic, andmean blood pressure.

The echocardiographic system used for quanti-tative integrated backscatter imaging has beendescribed elsewhere.'8 In brief, backscatterdata from each A-line are integrated (over3 2 Ms) and then used for on-line, real timeimage reconstruction. Time gain settings werecarefully optimised to improve the visualisa-tion of endothelial aortic borders. Lateral gaincompensation was not used to avoid artificialunderestimation of the aortic cross sectionalarea. A region of interest was manually tracedaround the descending thoracic aorta.Integrated software then calculates the identi-fied area along the internal aortic border anddisplays the results as trace and digital data onthe monitor (fig 1). For the present study theminimum and maximum cross sectional area(cm2) and peak positive area change (cm2/s)were calculated as an average from five con-secutive heart cycles derived from the area anddA/dt graphical recording (fig 2).The normalised peak positive area change (per

second) was defined as peak positive area changedivided by the minlimum cross sectional area:

ddt = dAdtmaxAmin

308

on January 11, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/H

eart: first published as 10.1136/hrt.75.3.307 on 1 March 1996. D

ownloaded from

Abnormal aortic elastic properties in Marfan syndrome

For comparison, 12 age matched normal sub-jects without cardiovascular diseases or con-

- , _ _,2nective tissue disorders underwent the sameprotocol after giving their informed consent.

STATISTICAL ANALYSISData were expressed as mean (SD). Anunpaired Student t test and a bivariate linearcorrelation analysis as well as a xI test wereused where appropriate to study the relationbetween different echocardiographic and

^__._ haemodynamic variables. A P value of less_ -'WS than 0 05 was considered significant.

ResultsElm_. DEMOGRAPHIC AND HAEMODYNAMIC VARIABLES

There were no significant differences betweenMarfan patients and controls in age, weight,heart rate, systolic and diastolic blood

(A) pressure, or pulse pressure, though there wasa slight but not statistically significanttendency for larger pulse pressure in Marfanpatients (table). Marfan patients were signifi-

_--I,,- i-cantly taller than controls, with correspondinglarger body surface area.

4 - ..~- M MODE BASED INDICES_W ^ There were no significant differences in

minimum and maximum aortic diameters and_ ~_ cross sectional areas normalised for body

surface (fig 3), even though absolute dimen-sions were larger in Marfan patients [mini-mum diameter: 1-9 (SD 0-5) v 1-6 (02) cm;maximum diameter: 2-1 (0-5) v 1-8 (0 1) cm;

____________________ minimum cross sectional area: 2-6 (1-8) v 2-0(0-5) cm2; maximum cross sectional area: 3-1(1-8) v 2-5 (0 5) cm2; P < 0 05 for alldifferences].

There was no difference in compliancebetween both groups (fig 4A), but Marfan-I-

(B)Figure 2 (A) Descending thoracic aorta of a normal subject with simultaneous acoustic NS P<0.05 P<0.05quantification and on-line display of cross sectional area (upper line) and its changes overtime (dA/dt, lower line). Peak positive area change (dA/dtm,) as a distinct maximum is Amarked as the cross sectional area increases with ariving pulse wave. Minimum andmaximum cross sectional areas are also marked (Ami, Amaj). (B) Imaging of the thoracicaorta in a patient with Marfan syndrome. Note the flattened graph of cross sectional areaand the low maxima of the dAldt graph compared to the normal control.

54T56 4.25.8 771

(2) (17) (8) (2.1) (0) (1.8)2.0 _ NS NS NS NS2.0~~~~~~~~Compliance Distensibility Stiffness index(10 cm2dyn-) (cm2dyn-1

1.5 Marfan patients, n=16_; Controls, n=12L|

TT_lTP<0 001 ~ P<0.0011.0 10T 9 8-B f7 T

0-5 I5 1. 1 1r1

1 1.0 1 0-9 1 1.4 m 11 Tt F0 7(7(3)7(0.1) (053)1(0- 1)4(009)(0-2)J (0.9) (0-2)Max diameter Min diameter Max area Min area

(cm/m ) (cm/m ) (cm /m ) (cm /m2

Max dA/dt Norm dA/dt

(cm_2s-__(s21Marfan patients, n=16_; Controls, n=12 = (

Figure 4 (A) Comparison of both groups for elasticityFigure 3 Aortic dimensions ofMarfan patients and controls, normalisedfor body surface indices based onM mode measurements. (B) Elasticityarea. indices based on the results ofautomated border detection.

309

on January 11, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/H

eart: first published as 10.1136/hrt.75.3.307 on 1 March 1996. D

ownloaded from

Franke, Muhler, Klues, Peters, Lepper, von Bernuth, Hanrath

patients had a significantly lower distensibility[4-2 (1-8) v 5-8 (2.1) cm2/dyn, P < 0 05] and ahigher stiffness index than normals [9-7 (3-0)v7 1 (1-8),P< 005].

INDICES DERIVED FROM ACOUSTICQUANTIFICATIONQualitative individual analysis of the on-lineregistered cross sectional area graph revealed aflattened curve for Marfan patients comparedto normals (fig 2B).

Highly significant differences were obtainedbetween both groups for peak positive areachange and normalized peak positive areachange (fig 4B). Peak positive area change waslower in Marfan patients than in normals [5 1(1.0) v 7-7 (1-7) cm2/s, P< 0-001]. Thedifference between the mean values of bothgroups was even higher for normalised peakpositive area change [2-5 (1-2) v 4 0 (0 8) s-',P < 0o001].

LINEAR CORRELATIONSThere was a good linear correlation betweenthe cross sectional area calculated from Mmode derived diameters and the arearegistered on-line by acoustic quantifi-cation (Pearson's p correlation coefficientr = 0-82, P < 0-001). With higher ageminimum and maximum diameters as well ascross sectional areas increased (for both:r = 0-60, P < 0-001), distensibility as well asnormalised peak positive area changedecreased (r = -0 50, P < 0-01) and stiffnessindex increased (low correlation, r = 0 4;P < 0 05). There was a low positivecorrelation between systolic blood pressureand maximum diameters (r = 0 45, P < 0.01)and a negative correlation between systolicpressure and distensibility (r = - 0 55,P < 0 005). These correlations apply for thegroup of Marfan patients as well as to thenormal controls.

There was no statistically significant influ-ence of heart rate, diastolic blood pressure,pulse pressure, or previous operation on any ofthe diameters, areas, or elasticity indices. Eventhe absolute diameters and cross sectionalareas had no statistical influence on the corre-sponding elasticity indices.

DiscussionPathomorphological and functional abnormali-ties of the thoracic and abdominal aorta aremajor determinants of the clinical course andprognosis in Marfan syndrome. Therefore, it isimportant to be able to obtain reliable, repro-ducible, and non-invasive methods for theearly detection of abnormalities in vascularelastic properties. Transoesophageal echocar-diography is of proven benefit in the cross sec-tional visualisation of the thoracic aorta ascompared to transthoracic echocardiography."

Until recently measurements of the aorticelastic properties were based either ontransthoracic M mode derived calcula-tions, 161719 on invasive haemodynamicmeasurements,1620 or on magnetic resonanceimaging (MRI).21-23

PREVIOUS STUDIES ON VASCULAR ELASTICPROPERTIESTransthoracic and transoesophageal M modemeasurements have been used to evaluate theelastic properties of the aortic wall, which hasproved to be age dependent in normal individ-uals. Lower distensibility and a higher stiffnessindex, calculated from these measurements,have been observed in patients with arterialhypertension, coronary artery disease,17 24 25and Marfan syndrome.6

Three recent papers have described thepotential value of MRI for evaluating elasticvascular properties and have confirmed thelower distensibility of the thoracic aorta inpatients with coronary artery disease2' andMarfan syndrome.2223

Intravascular ultrasound techniques havealso been used to determine aortic wall stiff-ness index in animals20 and humans.26 As forMRI, calculations were based on manual trac-ing of the minimum and maximum cross sec-tional area. The invasive character of thismethod limits its application, especially forserial investigations.The accuracy of the calculations of elastic

property indices with all the methodsdescribed above depends on the reliability ofvessel diameter or area measurements. In con-trast to ventricular measurements, maximumand minimum diameters or areas of thedescending aorta are not equal to end systolicand end diastolic diameters or areas; they haveto be chosen manually. Obviously it is a prob-lem to select the true minimum and maximumin a circular structure with small area changes.For M mode measurements, heart cycledependent motion of the descending thoracicaorta leads to sideshift artefacts causing anunderestimation of vessel diameters, becausethe M mode beam may not be located cen-trally throughout the whole cardiac cycle. Anadvantage of the M mode technique, however,is the higher time resolution as compared toacoustic quantification.

PREVIOUS STUDIES USING A COMBINATIONOF ACOUSTIC QUANTIFICATION ANDTRANSOESOPHAGEAL ECHOCARDIOGRAPHYRecently introduced automated border detec-tion and acoustic quantification allow theautomatic detection and calculation of themaximum and minimum areas of any cardio-vascular structure as well as graphical displayof the continuous area change during the heartcycle (dA/dt). First studies using this tech-nique focused on the estimation of left ventric-ular performance.'8 Gerber et al used acousticquantification of the thoracic aorta in combi-nation with transoesophageal echocardiogra-phy for the cross sectional analysis of aorticcross sectional area and for the calculation ofcompliance and distensibility. They found anage dependent decrease of compliance anddistensibility as well as slightly increased aorticdiameters in healthy subjects.27Our study for the first time uses the combi-

nation of transoesophageal echocardiographyand acoustic quantification in Marfan patientsfor the evaluation of elastic properties. The

310

on January 11, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/H

eart: first published as 10.1136/hrt.75.3.307 on 1 March 1996. D

ownloaded from

Abnormal aortic elastic properties in Marfan syndrome

results for standard indices (distensibility,stiffness index) are in concordance with thosepreviously reported by Hirata et aP6 andshowed a significant difference betweenMarfan patients and normal controls. We alsofound a negative correlation between age anddistensibility, whereas absolute minimum andmaximum diameters or cross sectional areashad no influence on any of the elasticityindices.The indices that we have now introduced

(peak positive area change and normalisedpeak positive area change)-derived fromacoustic quantification and reflecting thedynamic process of area changes-probablydiscriminate better than standard indicesbetween normal individuals and Marfanpatients.

LIMITATIONSMeasurements of the present study were lim-ited to the descending thoracic aorta.Neighbouring structures such as the leftatrium, pulmonary artery, and right ventriclemake it too difficult to place the region ofinterest on the ascending aorta without inter-ference, so that acoustic quantification andon-line measurements of the cross sectionalarea cannot be performed with the same preci-sion and reliability in the ascending as they canin the descending aorta. There is a problem oflimited recognition of intimal borders near tothe transducer. This will have more effect onthe distance measurements by M mode thanon the area measurements using acousticquantification.The integrated backscatter technique is lim-

ited to a time resolution of about 30 to 35frames per second. It also requires 3-2 /is perframe for the integration of all A-line data,causing a slight time delay. This delay, how-ever, is not relevant for the present study,because measurements of minimum and maxi-mum diameters and areas were not based onECG triggering.

FUTURE DIRECTIONSCombined transoesophageal echocardiogra-phy and acoustic quantification for the evalua-tion of elastic properties of the aorta includingindices such as maximum and normalised areachange over time might be a good tool for thelong term control of Marfan patients. Thismethod seems to be suitable for measuring theinfluence and postulated benefit of ,B adrenergicblockade28 on the elastic properties of vesselwalls and might have a prognostic valueregarding evolving cardiovascular complica-tions. To prove this, further investigationswith a larger group of patients and follow upstudies are necessary.

1 Murdoch JL, Walker BA, Halpern BL, Kuzma JW,Mckusick VA. Life expectancy and causes of death in theMarfan's syndrome. N EnglJ Med 1 972;286:804-8.

2 Roberts WC, Honig HS. The spectrum of cardiovasculardisease in the Marfan's syndrome: a clinico-morphologic

study of 18 necropsy patients and comparison to 151 pre-viously reported necropsy patients. Am HeartJ 1982;104:115-35.

3 Marsalese DL, Moodie DS, Vacante M, Lytle BW, GillCC, Sterba R, et al. Marfan's syndrome: natural historyand long-term follow-up of cardiovascular involvement. JAm Coll Cardiol 1989;14:422-8.

4 Silverman DI, Burton KJ, Gray J, Bosner MS, KouchoukosNT, Roman MJ, et al. Life expectancy in the Marfan'ssyndrome. Am Jf Cardiol 1995;75:157-160.

5 Gott VL, Cameron DE, Reitz BA, Pyeritz RE. Currentdiagnosis and prescription for the Marfan's syndrome:aortic root and valve replacement. J Cardiac Surg1994;9(suppl 2): 177-81.

6 Hirata K, Triposkiadis F, Sparks E, Bowen J, Wooley CF,Boudoulas H. The Marfan syndrome: abnormal aorticelastic properties. JAm Coll Cardiol 199 1;18:57-63.

7 Reed CM, Fox ME, Alpert BS. Aortic biomechanical prop-erties in pediatric patients with the Marfan syndrome,and the effects of atenolol. Am _J Cardiol 1993;71:606-8.

8 Savolainen A, Keto P, Hekali P, Nisula L, Kaitila I,Viitasalo M, et al. Aortic distensibility in children withthe Marfan's syndrome. Am _J Cardiol 1992;70:691-3.

9 Pereira L, Levran 0, Ramirez F, Lynch JR, Sykes B, et al.A molecular approach to the stratification of cardio-vascular risk in families with Marfan's syndrome. N Engl _JMed 1994;331:148-53.

10 Dietz HC, Pyeritz RE, Hall BD, Cadle RG, Hamosha A,Schwartz J. The Marfan's syndrome locus: confirmationof assignment to chromosome 15 and identification oftightly linked markers at 15ql 5-q21-3. Genomics 1991;9:355-61.

11 Simpson IA, de Belder MA, Treasure T, Camm AJ,Pumphrey CW. Cardiovascular manifestations ofMarfan's syndrome: improved evaluation by trans-oesophageal echocardiography. Br Heart _J 1993;69:104-8.

12 Pyeritz RE, Mckusick VA. The Marfan's syndrome: diag-nosis and management. N Engl J Med 1 979;300:772-7.

13 Beighton P, de Paepe A, Dunks D, Finidori G, Gedde-Dahl T, Goodman R, et al. International nosology of her-itable disorders of connective tissue, Berlin 1986. Am JfMed Genet 1988;29:581-94.

14 Pyeritz RE. Diagnosis of Marfan's syndrome. In: RoycePM, Steinmann B, eds. Connective tissue and its heritabledisorders: molecular, genetic and medical aspects. New York:Wiley-Liss, 1993:437-68.

15 Hirata K, Triposkiadis F, Sparks E, Bowen J, BoudoulasH, Wooley CF. Marfan's syndrome: cardiovascular phys-ical findings and diagnostic correlates. Am Heart J71992;123:743-57.

16 Stefanadis C, Stratos C, Boudoulas H, Kourouklis C,Toutouzas P. Distensibility of the ascending aorta: com-parison of invasive and non-invasive techniques inhealthy men and in men with coronary artery disease. EurHeartJ 1990;11:990-6.

17 Stefanadis C, Wooley CF, Bush CA, Kolibash AJ,Boudoulas H. Aortic distensibility abnormalities in coro-nary artery disease. Am _J Cardiol 1987;59: 1300-4.

18 Perez JE, Waggoner AD, Barzilai B, Melton HE, Miller JG,Sobel BE. On-line assessment of ventricular function byautomatic boundary detection and ultrasonic backscatterimaging. JAm Coll Cardiol 1992;19:313-20.

19 Gamble G, Zorn J, Sanders G, MacMahon S Sharpe N.Estimation of arterial stiffness, compliance and distensi-bility from M-mode ultrasound measurements of thecommon carotid artery. Stroke 1994;25:11-6.

20 Slordahl SA, Piene H, Linker DT, Vik A. Segmental aorticwall stiffness from intravascular ultrasound at normal andsubnormal aortic pressure in pigs. Acta Physiol Scand199 1;143:227-32.

21 Mohiaddin RH, Underwood SR, Bogren HG, Firmin DN,Klipstein RH, Rees RSO, et al. Regional aortic compli-ance studied by magnetic resonance imaging: the effectsof age, training and coronary artery disease. Br Heart Jf1989;62:90-6.

22 Savolainen A, Keto P, Hekali P, Nisula L, Kaitila I, ViitasalM, et al. Aortic distensibility in children with Marfan'ssyndrome. Am Jf Cardiol 1992;70:691-3.

23 Adams JN, Brooks M, Redpath TW, Smith FW, Dean J,Gray J, et al. Aortic distensibility and stiffness index mea-sured by magnetic resonance imaging in patients withMarfan's syndrome. Br HeartJ 1995;73:265-9.

24 Hirai T, Sasayama S, Kawasaki T, Yagi S. Stiffness of sys-temic arteries in patients with myocardial infarction.Circulation 1989;80:78-82.

25 Pasierski TJ, Binkley PF, Pearson AC. Evaluation of aorticdistensibility with transoesophageal echocardiography.Am HeartJ_ 1992;123:1288.

26 Heintz B, Walkenhorst F, Gillessen T, Dorr R, Krebs W,vom Dahl J, et al. In vivo characterization of segmentalelastic properties of the aortic tree from intravascularultrasound. Cardiol Elderly 1994;2:127-32

27 Gerber TC, Foley DA, Greenleaf JF, Tajik J, Seward JB.Assessment of age-dependant changes of aortic distensi-bility with a new acoustic quantification system [abstr]. JAm Coll Cardiol 1993;21:298A.

28 Shores J, Berger KR, Murphy EA, Pyeritz RE. Progressionof aortic dilatation and the benefit of long-term beta-adrenergic blockade in Marfan's syndrome. NEnglJMed1994;330:1335-41.

311

on January 11, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/H

eart: first published as 10.1136/hrt.75.3.307 on 1 March 1996. D

ownloaded from