Heart Prognosis ofsupravalve aortic stenosis in patients in Liverpool … · Heart 1996;75:396-402 Prognosis ofsupravalve aortic stenosis in 81 patients in Liverpool (1960-1993) DKitchiner,

Heart 1996;75:396-402

Prognosis of supravalve aortic stenosis in 81patients in Liverpool (1960-1993)

D Kitchiner, M Jackson, K Walsh, I Peart, R Arnold

AbstractObjective-To determine the prognosis ofsupravalve aortic stenosis into early adultlife and the factors affecting this progno-sis.Design-81 patients with supravalve aor-tic stenosis were followed for a medianduration of 8.3 (range 1 to 29) years.Patients-40 patients (49.4%) hadWilliams' syndrome, 18 (22-2%) familialsupravalve aortic stenosis, 18 (22.2%)sporadic supravalve aortic stenosis, andfive (6-2%) other syndromes. Nineteenpatients had additional levels of left ven-tricular outflow tract obstruction.Results-47 patients (58%) underwentoperation; 20% within a year of presenta-tion. Multivariable analysis predicted that88% of patients would undergo interven-tion within 30 years of follow up. Thechance of intervention was increased bymore severe aortic stenosis at presenta-tion and the presence of multilevelobstruction in patients with sporadicsupravalve aortic stenosis. Three deathsoccurred before operation and 13 within amonth of operation. Ten (62.5%) of thepostoperative deaths were in patients withmultilevel obstruction. Predicted survival30 years after presentation was 66%. Riskfactors for survival were age and severityof aortic stenosis at presentation. Multi-level obstruction did not emerge as a sig-nificant risk factor for death because ofthe high association with the severity ofstenosis at presentation. 74% of survivorshad mild or insignificant stenosis at fol-low up.Conclusions-Long-term survival isrelated to age and the severity of aorticstenosis at presentation. Most patientswill require intervention, and most sur-vivors will have mild stenosis.

(Heart 1996;75:396-402)

Keywords: aortic stenosis; William's syndrome; familialsupravalve aortic stenosis; prognosis

Supravalve aortic stenosis is the least commontype of left ventricular outflow obstruction. 'Most patients with this condition haveWilliams' syndrome or a positive family his-tory,23 but about 25% of cases are sporadic.4Progression with time has been demon-strated,256 but these reports contain relativelysmall numbers of patients. The aims of this

study were to determine the prognosis ofpatients with supravalve aortic stenosis frompresentation into early adult life and define thefactors influencing that prognosis.

Patients and methodsEighty one patients (52 (64%) were male)with supravalve aortic stenosis who presentedto the Royal Liverpool Children's Hospitalbetween 1 January 1960 and 31 December1992 were included. The severity of stenosis atpresentation was determined from clinicaldata, together with echo/Doppler or cardiaccatheterisation assessment if performed withina year of presentation. The criteria for assess-ing the severity of stenosis were the same asthose used in a previous study.7 Patients withmild aortic stenosis had a short ejection sys-tolic murmur and a normal electrocardiogram.A Doppler velocity of less than 3 m/s or a peaksystolic gradient of less than 40 mm Hg acrossthe left ventricular outflow tract at cardiaccatheterisation was also used as evidence ofmild stenosis. Patients with moderate stenosishad a long ejection systolic murmur and theelectrocardiogram was normal or showed fea-tures of left ventricular hypertrophy. ADoppler velocity of 3-45 m/s or a peak sys-tolic gradient of 40-80 mm Hg across the leftventricular outflow tract was also used as evi-dence of moderate stenosis. Patients withsevere stenosis at presentation all had symp-toms or left ventricular strain pattern on theelectrocardiogram. The level of obstructionwas determined from clinical, echocardio-graphic, angiographic, operative, and post-mortem information. Differentiation betweenlocalised and diffuse supravalve aortic stenosiswas determined by aortic or left ventricularangiography in 52 patients, echocardiographyin 24, and from operation notes in five inwhom angiograms were not available.Survivors were traced to assess their currentclinical status. Evaluation included clinicalexamination, electrocardiogram, and cross-sectional and Doppler echocardiography.

STATISTICAL ANALYSISFrequency data are presented as raw counts orpercentages, and differences between groupswere assessed using the X2 8 or Fisher's exacttest9 as appropriate. Continuously distributeddata are given as medians with range, andgroup differences explored using theWilcoxon's rank sum test.8 Seventy per centconfidence limits were used throughout. Timerelated analysis of death and first intervention

Cardiac Unit, RoyalLiverpool Children'sNHS TrustD KitchinerK WalshI PeartR AmoldInstitute of ChildHealth, RoyalLiverpool Children'sNHS Trust, LiverpoolM JacksonCorrespondence to:Dr D Kitchiner, CardiacUnit, Royal Liverpool NHSTrust, Eaton Road,Liverpool L12 2AP.Accepted for publication10 October 1995

396

on January 8, 2021 by guest. Protected by copyright.

http://heart.bmj.com

/H

eart: first published as 10.1136/hrt.75.4.396 on 1 April 1996. D

ownloaded from

http://heart.bmj.com/

Prognosis ofsupravalve aortic stenosis in 81 patients in Liverpool (1960-1993)

Table 1 Age at presentation and incidence of associatedconditions, and distribution ofpatients with supravalveaortic stenosis (SVAS) alone and those with multilevelobstruction

MedianNo of (range) agepatients (months)

Williams' syndrome 40 (49-4) 16 7 (1 6-99-6)Familial SVAS 18 (22-2) 6 9 (0 2-52 3)Sporadic SVAS 18 (22 2) 28-6 (0-1-143 6)Other syndromes* 5 (6 2) 3-9 (0 1-15 4)

SAVAS alone 62 (76 5) 14 6 (0-2-118-5)Multilevel obstruction 19 (23 5) 3-9 (0-1-143 6)

Values in parentheses are percentages. *Noonans (n = 2);Shones (n = 2); rubella (n = 1).

Table 2 Severity of aortic stenosis at presentation relatedto associated conditions, and the presence of supravalveaortic stenosis alone (SVAS) or multilevel obstruction

Insignificantl Moderate Severemild stenosis stenosis stenosis

Williams' syndrome 24 12 4Familial SVAS 9 5 4Sporadic SVAS 7 8 3Other syndromes 1 3 1

SVAS alone 37 20 5Multilevel

obstruction 4 8 7

(surgery or balloon dilatation) after presenta-tion (time zero) were performed actuarially'0and parametrically." Multivariable analyses ofdemographic, clinical, institutional, and mor-phological variables (appendix) were made inthe hazard function domain" and retained inthe multivariable equation when the P valuefor the variable was < 0-1.

ResultsThe median age at presentation was 13-8months with a range of 3 days to 12 years.

Table 3 Distribution of other cardiac lesions in patients with supravalve aortic stenosis(SVAS) related to associated conditions

Williams' Familial Sporadic Othersyndrome SVAS SVAS syndromes

PPS 16 3 5(29 6)PVS 2 3(6-2)VSD it 2 1(4-9)Ao arch obst* + 1 1 3 2tVSD ± PDA(6-2)Total 20 (50) 6 (33) 9 (50) 5 (100)

Values in parentheses are percentages. *One patient had aortic arch interruption. tAdditionalmitral valve disease. $Additional right ventricular outflow tract obstruction.Ao, aortic; obst, obstruction; PDA, patent arterial duct; PPS, peripheral pulmonary artery steno-sis; PVS, pulmonary valve stenosis; VSD, ventricular septal defect.

Table 4 Incidence of surgery and reoperation related to associated conditions in the 47patients who underwent operation, and relation ofsurgery to the presence ofsupravalveaortic stenosis (SVAS) alone or multilevel obstruction

No operation 1 operation 2 operations 3 operations(n = 34) (42-0) (n = 39) (48-2) (n = 7) (8-6) (n = 1) (1-2)

Williams' syndrome 21 19Familial SVAS 7 9 2Sporadic SVAS 6 8 3 1Other syndromes 3 2

SVAS alone 34 27 1Multilevel obstruction 12 6 1

Values in parentheses are percentages.

Table 1 shows the incidence of conditionsassociated with supravalve aortic stenosistogether with age at presentation. The age atpresentation was lower in patients with famil-ial supravalve aortic stenosis than in those withWilliams' syndrome (P < 0 02) or sporadicsupravalve aortic stenosis (P < 0-05).Nineteen patients had additional levels of leftventricular outflow obstruction (multilevelobstruction). Eight had aortic valve stenosis,and 11 had valve and subvalve obstruction.Diffuse supravalve aortic stenosis occurred infive patients (26&3%) with multilevel obstruc-tion, and in only four patients (6 5%) withsupravalve aortic stenosis alone (P < 005).There was no significant difference in the ageat presentation in patients with supravalve aor-tic stenosis alone and those with multilevelobstruction (table 1). The incidence of multi-level obstruction was lower (P < 0.001) inpatients with Williams' syndrome (twopatients, 5 0%) and those with familialsupravalve aortic stenosis (three patients,1 67%) than in patients with sporadicsupravalve aortic stenosis (nine patients, 50%)and those with other syndromes (five patients,100%).There was no difference in the distribution

of severity of aortic stenosis at presentationrelated to the various associated conditions (P= 0-12). However, patients with more severestenosis at presentation had an increased ten-dency (P < 0-001) to have multilevel obstruc-tion (table 2). Other cardiovascular lesionswere found in 40 patients (49A4%) (table 3).There was no difference in the incidence ofother cardiac lesions in patients withsupravalve aortic stenosis alone (n = 29;46 8%) and those with multilevel obstruction(n = 11; 58%). With one exception, aorticarch obstruction occurred only in patientswith multilevel obstruction. Aortic regurgita-tion was found at presentation in threepatients (3 7%) who all had additional aorticvalve stenosis.The median duration of follow up in sur-

vivors was 8-3 (range 1-29) years. Ninepatients were lost to follow up after a medianduration of 11 (range 4-1-18-4) years. Aorticregurgitation was found in four patients withadditional aortic valve stenosis at follow up.Infective endocarditis was not encountered.

INTERVENTIONForty seven patients (58%) underwent opera-tion for supravalve aortic stenosis (table 4). Themedian age at first operation was 4 9 (range0-1-20-9) years and there was no difference inrelation to the various associated conditions. Allpatients underwent cardiac catheterisationbefore surgery and the peak systolic gradientacross the left ventricular outflow tract was amedian of 72 (range 42-120) mm Hg. Thepeak systolic gradient was more than50 mm Hg in all but four patients who hadsymptoms or evidence of left ventricular strainpattern on the electrocardiogram. Threepatients underwent balloon dilatation of the leftventricular outflow tract. This was successful inonly one patient with isolated supravalve steno-

397



/H


ownloaded from


Kitchiner, Jackson, Walsh, Peart, Arnold

g 90

(62)00

(D60E0 (41

40E30-

20

@ 10

Dlo0 5 10 15 20 2

Interval since presentation (ye,

Free from intervention (%)

Interval Life table Predicted(years) (%) parametric

(%)

1015202520

804327131313

ars)

813929231612

sis who had previously had surgery. Some 20%of patients underwent intervention within a

year of presentation. Actuarial and hazardanalysis predicted that 88% of patients were

likely to require relief of obstruction within 30years of presentation (fig 1).

Multivariable analysis of potential risk fac-tors (appendix) indicated that more severe

aortic stenosis at presentation was thestrongest determinant of the need for surgical

_100go40 5378 0 50)

Interval since presentation (years)

No of interventions

Aortic stenosis Actualgrade

Mild 13Moderate 25Severe 9

Predicted P value

14.2 0.6921-8 0.1410.9 0.05

intervention. Figure 2 shows that mostpatients who presented with moderate or

severe stenosis underwent operation within thefirst few years after presentation, whilepatients who presented with mild obstructiongradually progressed to require interventionover a number of years. Multivariable analysisalso indicated that the presence of multilevelobstruction in patients with sporadicsupravalve aortic stenosis was a significant(P = 0 002) risk factor for intervention. Thepresence of Noonan's syndrome also provedsignificant (P < 0 05) in an effort to explainthe intervention experiences of the twopatients with this condition as both had addi-tional severe right ventricular outflow obstruc-tion which contributed to their earlymanagement. Factors not significant in rela-tion to intervention were the sex of the patient,syndromes other than Noonan's syndrome,and the presence of aortic regurgitation at pre-

sentation.The reoperation rate was 17%. The inci-

dence was higher (P < 0 01) in patients withmultilevel obstruction than in those withsupravalve aortic stenosis alone (table 4). Themedian duration between first and secondoperations was 4-6 (range 0 5-8 2) years, andthis was similar in patients with the variousassociated conditions. Table 5 shows theresults of surgery.

SURVIVAL

There were 16 deaths (19-8%) during thestudy period. Fourteen patients had severe

aortic stenosis and two with Noonan's syn-

drome had additional severe pulmonary valveand peripheral pulmonary artery stenosis. Themortality was lower (P < 0 001) in patientswith isolated supravalve aortic stenosis than inthose with multilevel stenosis (table 5). Fourpatients with Williams' syndrome (10%), fivewith familial supravalve aortic stenosis(27-7%, three with sporadic supravalve aorticstenosis (16-7%), and four with other syn-

dromes (80%) died. There were no sudden,unexpected deaths. Two patients with severe

isolated supravalve aortic stenosis died beforeoperation could be performed and one withWilliams' syndrome and severe kyphoscoliosisdeclined surgery. Thirteen patients diedwithin a month of operation. Three of thesepatients had severe isolated supravalve stenosiswith profound left ventricular hypertrophy.Ten patients with multilevel obstruction diedafter operation. Six of these had supravalve,valve, and subvalve obstruction, one had

Figure 2 Freedom from intervention after presentation in81 patients with supravalve aortic stenosis stratified byseverity of aortic stenosis at presentation. Kaplan-Meierestimates offreedom from intervention are depicted as in fig1 for patients with mild (0), moderate (A) and severe

aortic stenosis (O) at presentation. The predicted timerelatedfreedom from intervention for each group ( ),obtained by averaging the predicted patient-specificfreedom from first intervention estimates derivedfrom thesolution to the multivariable equation for each member ofthe group, and 70% confidence intervals (- -) are shown.The table gives results from an internal validation of themultivariable equation in terms of the association betweenactual and predicted number of interventions for eachstratum.

Table 5 Current clinical status after surgery in patientswith supravalve aortic stenosis (SVAS) alone and in thosewith multilevel obstnrction, and median age at initialoperation

SVAS Multilevelalone obstruction

No stenosis 4Mild stenosis 13 4Moderate stenosis 6 2Hospital mortality

(death < 30 days) 3 10Lost to follow up 2 3Median (range) age at 60-5 42-7first operation (months) (9-4-250 9) (0-8-229-4)

Figure 1 Freedom fromfirst intervention afterpresentation in 81 patientswith supravalve aorticstenosis. This graphcontains two depictions.The first results from anactuarial (Kaplan-Meier)analysis in which eachintervention (0) ispositioned actuarially alongthe y axis and the timeintervention along thex axis. Bars represent 70%confidence intervals and thenumbers in parenthesesindicate the number ofpatients followed up at thatinterval after intervention.The second is a graphicalrepresentation of thesolution to the equation(nomogram) resultingfromthe parametric analysis offreedom from intervention,in which a continuous pointestimate ( ) and the70% confidence (- -)limits are shown. The tablegives the time relateddepictions.

398



/H


ownloaded from



Figure 3 Survival afterpresentation in 81 patientswith supravalve aorticstenosis. The generalformof the depictions (actuarialand parametric) are as infig 1, with death at anytime after presentation (0)in terms of survival.

0-

(I)

4-

i>21Cn

Survival rate (%)

Interval(years)110

15202530

0

a)

4-

0 a

._0.

._0

cL

5 10 15 20 25

Interval since presentation (years)

Life Predictedtable (%) parametric (%)

95 9377 7977 6571 7171 6971 66

30

supravalve and valve stenosis, and three hadadditional surgery for right ventricular outflowtract obstruction. Four of the patients withmultilevel obstruction died at reoperation andthree of these operations included attemptedaortic valve or root replacement. The opera-tive mortality was lower (P < 0-005) inpatients with supravalve aortic stenosis alone(10X7%) than in those with multilevel obstruc-tion (53%).An actuarial analysis of death at any time

after presentation indicated that survivaldeclined gradually over time (fig 3) to plateauat 71%, 17 years after presentation. Evidencefrom the parametric model suggests that theproportion of patients alive after 30 years offollow up is unlikely to differ much from thisvalue. Multivariable analysis determined thatrisk factors for death were more severe aorticstenosis at presentation, younger age at pre-

sentation, and the presence of Noonan's syn-

drome (appendix). Multilevel obstruction didnot emerge as a significant risk factor because

Figure 4 Graph showingrisk adjusted time relatedsurvival after presentationwith supravalve aorticstenosis according toseverity of aortic stenosis.The depiction is anomogram of a specificsolution of themultivariable equation(appendix) where 1 15years (13-8 months) is thevalue enteredfor age atpresentation and no forpresence ofNoonan'ssyndrome.

10090

- 801-

°l 704-, 60_ 50.> 402 30

0 201000 5 10 15 20 25

Interval since presentation (years)30

Survival (%) according to aortic stenosis gradeInterval Mild (%) Moderate (%) Severe (%)

(years)1 99.4 96.110 97.6 85.020 96.4 78.030 95.3 72.9

76.633.819.112.0

0

=Moderate

- ==-

5 10 15 20 25 30Interval between presentation and follow up

(years)

Distribution of aortic stenosis grade at follow up(%)Interval No Mild Moderate(years) stenosis stenosis stenosis

1 3 62 355 4 67 2910 6 71 2320 10 76 1430 18 74 8

Figure 5 Relation of interval between presentation andfollow up with the grade of aortic stenosis severity in the 65surviving patients at lastfollow up evaluation. The actualproportions ofpatients with no (0), mild (0), ormoderate (A) stenosis (appendix) at the mid-points of 10year intervals are represented. Nomograms of theparametric predictions of the prevalences of stenosis grade( ) and 70% confidence intervals (--- -) are shown.The table gives percentage ofpatients in each aortic stenosisgrade (total 100%) at each time interval.

of a close association with the severity of aorticstenosis at presentation. Factors insignificantin relation to survival were the sex of thepatient, the presence of syndromes other thanNoonan's syndrome, and the presence of aorticregurgitation at presentation. Figure 4 showsthe clear separation in time related survival inpatients with mild, moderate, and severe aor-tic stenosis at presentation when adjusted forthe presence of the risk factors of age at pre-sentation and Noonan's syndrome. The datain fig 4 indicate that 30 year survival is pre-dicted to be 95 3% for such a patient present-ing with mild stenosis, 72-9% for one withmoderate stenosis, and 12-0% for one withsevere stenosis.

FOLLOW UPFigure 5 shows the distribution of the severityof aortic stenosis in survivors at follow up.With increasing duration of follow up, therewas a reduction in the proportion of patientswith moderate stenosis and an increase in thenumber with no stenosis. The proportion ofpatients with mild stenosis remained fairly

Table 6 Current status ofpatients with supravalve aorticstenosis (SVAS) alone and those with multilevelobstruction *

SVAS Multilevelalone obstruction

Mild stenosis 20 (32 2)Moderate stenosis 7 (11-3)Surgery and survived 23 (37-1) 6 (31-6)Lost to follow up 6 (9 7) 3 (15 8)Died 6 (9 7) 10 (52 6)

Values in parentheses are percentages. *Each patient appearsonce.

399



/H


ownloaded from


Kitchiner, J3ackson, Walsh, Peart, Arnold

constant at between 60 and 75%. Table 6shows the current clinical status of all patients.

DiscussionSupravalve aortic stenosis is a rare condition.Only careful follow up of patients over a longperiod provides an insight into its prognosis.This study comprised the largest group ofpatients followed from presentation. Mostpatients were traced and re-examined, and theduration of follow up in survivors allowed pre-dictions of the prognosis into adult life.

INTERVENTIONAlthough only 47 patients (58%) underwentintervention during the study period, hazardanalysis indicated that more patients are likelyto undergo intervention the longer the dura-tion of follow up (fig 1). Patients with otherlevels of left ventricular outflow tract obstruc-tion have also shown progression withtime,1 7 12 but this has not previously been con-firmed in a large group of patients withsupravalve aortic stenosis. The emergence ofincreasing severity of aortic stenosis at presen-tation as a risk factor for intervention, as foundin this study, has also been documented inpatients with other levels of left ventricularoutflow tract obstruction.' 7In general, thepresence of multilevel obstruction was closelyrelated to more severe aortic stenosis at pre-sentation, but this was not the case in patientswith sporadic supravalve aortic stenosis.Therefore, for the purposes of predicting thelikely time intervention, the occurrence ofmultilevel obstruction in patients with spo-radic supravalve aortic stenosis emerged as asignificant risk factor, in addition to the infor-mation provided by the severity of aorticstenosis at presentation. Noonan's syndromeis a very rare association with supravalve aorticstenosis'3 and the atypical nature of the twopatients with this syndrome was confirmed inthe requirement for additional interventions.These two patients had additional severe pul-monary valve stenosis and intervention wasalso combined with relief of this obstruction.

SURVIVALMortality was highest in patients with multi-level obstruction and this indicates the seriousnature of this particular condition. Difficultiesin the management of patients with multilevelobstruction are reflected in the high postopera-tive mortality in this group. Hazard analysispredicted that 66% of the study group wouldbe alive after 30 years of follow up, and multi-variable analysis indicated that the severity ofaortic stenosis at presentation significantlyinfluenced survival. Figure 4 shows the pre-dicted risk ofpremature death for patients withan identical distribution of risk factors-that is,risk adjusted, in all respects other than theseverity of aortic stenosis. Therefore it repre-sents the true effect of the severity of aorticstenosis at presentation on survival. A naturalorder emerged in terms of premature deathwith increasing severity of stenosis at presenta-tion, which has also been found in patients

with other levels of left ventricular outflowtract obstruction.' 7

The age of the patient at presentation wasalso shown to modify survival significantly,with younger patients having a poorer out-come. Similar findings have been documentedin patients with other levels of left ventricularoutflow tract obstruction.' 7 13 Patients withsupravalve stenosis and Noonan's syndromehad a particularly poor survival record, proba-bly because of their associated cardiac condi-tions. Patients with more than one level of leftventricular outflow tract obstruction arereported to have a worse prognosis than thosewith a single level of stenosis,14 15 but multivari-able analysis has not previously been per-formed in relation to patients with supravalveaortic stenosis. Perhaps surprisingly, multilevelobstruction did not emerge as a significant riskfactor for death. The severity of aortic stenosisat presentation was found to correlate closelywith the presence of multilevel obstruction.The significantly higher mortality in patientswith multilevel obstruction than in patientswith supravalve aortic stenosis alone is thusaccounted for by their differences in the sever-ity of aortic stenosis at presentation.

Multilevel obstruction is more common inpatients with supravalve aortic stenosis than inthose aortic valve stenosis. It occurs in20-45% of patients with supravalve aorticstenosis,4 16 17 whereas less than 10% of patientswith aortic valve stenosis have other levels ofobstruction.' 16 18 Diffuse supravalve stenosis isreported in 15-24% of patients,2 17 18 but waslower in the current study. These patients gen-erally have a worse prognosis than those withlocalised obstruction,2 19 20 but again, diffusestenosis did not emerge as a risk factor fordeath or intervention in this study due to ahigh degree of association with other identifiedrisk factors.

FOLLOW UPThe proportion of patients with mild stenosisremained reasonable constant (fig 5).Although progression of stenosis occurred withtime in patients who presented with mildstenosis, others who presented with moderateor severe stenosis underwent operation whichresulted in mild residual stenosis at follow up.Other patients had complete relief of theobstruction, so the number of patients with nostenosis also increased with time. The declinein the proportion of patients with moderatestenosis with time results from intervention, asno patient improved without it.

CONDITIONS ASSOCIATED WITH SUPRAVALVEAORTIC STENOSISPatients with supravalve aortic stenosis com-monly have associated conditions. Williams'syndrome was the most common conditionassociated with supravalve aortic stenosis andit occurred more frequently than previouslyreported.2202' No patient with Williams' syn-drome required reoperation. Patients withWilliams' syndrome had a relatively goodprognosis with a 10% mortality, as found inanother study.6

400



/H


ownloaded from



Familial supravalve aortic stenosis occurredalmost half as frequently as Williams' syn-drome, with an incidence similar to that inother series.222 In a study by Johnson et al,23the mean age at presentation was 14 years.With the use of cross sectional and Dopplerechocardiography,523 early diagnosis is possi-ble. Patients in the current study presented at ayounger age than those with Williams' syn-drome, probably because the family historyencouraged early cardiac assessment.The incidence of sporadic supravalve aortic

stenosis (22%) was a little lower than in otherseries,222 possibly because of careful clinicalassessment for features of Williams' syndromeand a family history of supravalve aorticstenosis. The relatively older age at presenta-tion in this group may have been becausepatients with Williams' syndrome or a familyhistory were likely to undergo cardiac exami-nation at a younger age. Reoperation wasmore common in patients with sporadicsupravalve aortic stenosis. This reflects thesignificantly higher incidence of multilevelobstruction in patients with sporadic supra-valve aortic stenosis than in with Williams'syndrome or familial supravalve aortic steno-sis and inadequate relief of stenosis at initialoperation.

Rubella syndrome is known to be associatedwith supravalve aortic stenosis,24 and patientswith Shone's syndrome have left ventricularoutflow obstruction25 which can includesupravalve aortic stenosis. The association ofNoonan's syndrome with supravalve aorticstenosis is extremely rare. The incidence ofthese syndromes was low but all had multilevelobstruction. In some ways they were not rep-resentative of patients with supravalve aorticstenosis. However, multilevel obstruction didoccur in other patients with supravalve aorticstenosis, so it was considered that thesepatients should not be excluded. Patients withthese syndromes presented at a young age, allhad other cardiac lesions, and all requiredsurgery for supravalve aortic stenosis. Theyhad the highest mortality of any subgroup,which consisted entirely of early postoperativedeaths. Noonan's syndrome emerged as a sig-nificant (P < 002) risk factor for interventionand death because of unsuccessful attempts torelieve the coexistent severe right ventricularoutflow tract obstruction.The incidence and distribution of other car-

diac lesions was similar to that in other pub-lished series.6 '7 The most common lesion wasperipheral pulmonary artery stenosis, whichoccurred in patients with Williams' syndromeand those with the familial types of supravalveaortic stenosis, but was also found in patientswith sporadic supravalve aortic stenosis. Thehigher incidence of aortic arch obstruction inpatients with multilevel obstruction empha-sises the diffuse nature of left ventricular out-flow tract obstruction.

Infective endocarditis occurs in patientswith supravalve aortic stenosis'8 but was notencountered in this study. Sudden death didnot occur in any patient in this study but it hasbeen reported, in relation to severe supravalve

aortic stenosis and obstruction of coronaryostia in patients with mild stenosis.1826

ConclusionsPatients with severe supravalve aortic stenosispresent early and undergo intervention at ayoung age. Patients who present with moder-ate or mild stenosis fare better, but most willrequire intervention. Most survivors will havemild stenosis.

The authors thank Eugene H Blackstone for guidance in theuse of the parametric analysis of time related events. MJ is sup-ported by a grant from the Cardiac Fund, Royal LiverpoolChildren's NHS Trust.

1 Kitchiner DJ, Jackson M, Malaiya N, Walsh K, Peart I,Arnold R. The incidence and prognosis of left ventricularoutflow obstruction in Liverpool (1960-1991). Br HeartJ7 1994;71:588-95.

2 Flaker G, Teske D, Kilman J, Hosier D, Wooley C.Supravalvular aortic stenosis. A 20-year clinical perspec-tive and experience with patch aortoplasty. Am J7 Cardiol1983;51:256-60.

3 Jones KL, Smith DW. The Williams elfin facies syndrome. Anew perspective. Jf Pediatr 1975;86:718-23.

4 Pansegrau DG, Kioshos AM, Durnin RE, Kroetz FW.Supravalvular aortic stenosis in adults. Am J Cardiol1973;31:637-41.

5 Wren C, Oslizlok P, Bull C. Natural history of supravalveaortic stenosis and pulmonary artery stenosis. J Am CollCardiol 1990;15:1625-30.

6 Zalzstein E, Moes CAF, Musewe NN, Freedom RM.Spectrum of cardiovascular anomalies in Williams-Beuren syndrome. Pediat Cardiol 1991 ;12:219-23.

7 Kitchiner DJ, Jackson M, Walsh K, Peart I, Arnold R. Theincidence and prognosis of congenital aortic valve stenosis;Liverpool (1960-1990). BrHeartJ 1993;69:71-9.

8 Armitage P, Berry G. Statistical methods in medical research.2nd ed. Oxford: Blackwell Scientific Publications, 1987:371-421.

9 Fisher RA. Statistical methods and scientific reference. 3rd ed.New York: Hafner Press, 1973:71-8.

10 Kaplan EL, Meier P. Nonparametric estimation fromincomplete observations. Am Stat Assoc J 1958;53:457-81.

11 Blackstone EH, Naftel DC, Turner ME Jnr. The decompo-sition of time varying hazard into phases, each incorpo-rating a separate stream of concomitant information. JAm Stat Assoc 1986;81:615-24.

12 Walker SH, Duncan DB. Estimation of the probability ofan event as a function of several independent variables.Biometrika 1967;54: 167-79.

13 Keane JF, Dirscoll DJ, Gersony WM, Hayes CJ, Kidd L,O'Fallon WM, et al. Second natural history study of con-genital heart defects. Results of treatment of patients withaortic valvar stenosis. Circulation 1993;87(suppl I): 11-27.

14 Sreeram N, Kitchiner D, Smith A. Spectrum of valvarabnormalities in Noonan's syndrome: a pathologic study.Cardiol Young 1994;4:62-6.

15 Fisher RD, Mason DT, Morrow AG. Results of operativetreatment in congenital aortic stenosis. Pre- and postop-erative hemodynamic evaluations. I Thorac CardiovascSurg 1970;59:218-24.

16 Kirklin JW, Barratt-Boyes BG. Congenital discrete sub-valvular aortic stenosis. In: Kirklin JW, Barratt-BoyesBG, eds. Cardiac surgery. New York: John Wiley andSons, 1986:972-1012.

17 Sharma BK, Fujiwara H, Hallman L, Ott DA, Reul GJ,Cooley DA. Supravalvar aortic stenosis: a 29-year reviewof surgical experience. Ann Thorac Surg 1991;51:1031-9.

18 Peterson TA, Todd DB, Edwards JE. Supravalvularaortic stenosis. J Thorac Cardiovasc Surg 1965;50:734-41.

19 Bernhard WF, Keane JF, Fellows KE, Litwin SB, GrossRE. Progress and problems in the surgical managementof congenital aortic stenosis. J Thorac Cardiovasc Surg1973;66:404-19.

20 Rastelli GC, McGoon DC, Ongley PA, Mankin HT,Kirklin JW. Surgical treatment of supravalvular aorticstenosis. Report of 16 cases and review of literature. JThorac Cardiovasc Surg 1966;51:873-82.

21 Keane JF, Fellows KE, LaFarge CG, Nadas AS, BernhardWF. The surgical management of discrete and diffusesupravalvar aortic stenosis. Circulation 1976;54: 112-7.

22 Martin EC, Moseley IF. Supravalve aortic stenosis. BrHeartJ7 1973;35:758-65.

23 Johnson LW, Fishman RA, Schneider B, Parker FB,Husson G, Webb WR. Familial supravalvular aorticstenosis. Report of a large family and review of the litera-ture. Chest 1976;70:494-500.

24 Weyman AE, Caldwell RL, Hurwitz RA, Girod DA, DillonJC, Feigenbaum H, et al. Cross-sectional echocardio-graphic characterization of aortic obstruction. 1. Supra-valvular aortic stenosis and aortic hypoplasia. Circulation1 978;57:49 1-7.

401



/H


ownloaded from


Kitchiner, Jackson, Walsh, Peart, Arnold

25 Varghese PJ, Izukawa T, Rowe RD. Supravalvular aorticstenosis as part of rubella syndrome, with discussion ofpathogenesis. Br Heart J 1969;31:59-62.

26 Shone JD, Sellers RD, Anderson RC, Adams P, LilleheiCE, Edwards JE. The development complex of "para-chute mitral valve" supravalvular ring of left atrium,subaortic stenosis and coarctation of the aorta. Am JCardiol 1963;11:714-25.

AppendixVARIABLES ENTERED INTO MULTIVARIABLE ANALYSES OF

FREEDOM FROM FIRST INTERVENTION AND DEATH AT

ANY TIMEDemographic age at presentation (with transforms), sex.Clinical aortic stenosis severity at presentation (withtransforms), presence of aortic regurgitation, presenceof Noonan's, rubella, Shone's or Williams' syndrome,and familial supravalve aortic stenosis.Institutional interval from first patient presenting toRoyal Liverpool Children's Hospital to date of presen-tation (with transforms).Morphology level of obstruction: diffuse or localisedsupravalve aortic stenosis. These factors were consid-ered as isolated potential risk factors and as interactiveterms (product of two potential risk factors).

coefficients, standard deviations, and p values (inparentheses) of the multivariable risk factor equation:T= 1, a = 1, y = 1, i7 = 0265, intercept= 000055946,aortic stenosis severity at presentation - 1, multipliedby interval from first patient presenting to RoyalLiverpool Children's Hospital to date of presentation'0 010 (0-001) (P < 0-0001), presence of multilevelobstruction (in patients with sporadic supravalve aorticstenosis) 1-436 (0 389) (P = 0-002), presence ofNoonan's syndrome 1-586 (0 703) (P = 0 033).

FREEDOM FROM DEATH AT ANY TIME AFTERPRESENTATIONParameter estimates: single declining hazard phase ,3 =002312534, T = 1, a = 1, y = 1, t = 0488.

Shaping parameter estimates and the regressioncoefficients, standard deviations, and p values (inparentheses) of the multivariable risk factor equation:T = 1, a = 1, y = 1, a7 = 0-610, intercept = 0-00054268,aortic stenosis severity at presentation 1-899 (0 40)(P < 0-0001), age at presentation (natural logarithmictransform) -0-386 (0-12) (P = 0001), presence ofNoonan's syndrome 2-222 (0 89) (P = 0-013).

FREEDOM FROM FIRST INTERVENTION AT ANY TIME

AFTER PRESENTATIONParameter estimates: single declining hazard phase /3 =

0-04210502, = 1, a = 1, y = 1, q =0651.Shaping parameter estimates and the regression

'This variable was created for each patient in the study fromthe transformation of severity of aortic stenosis (by subtracting1) and multiplying it with the interval from first patient to eachpatient's date of presentation, and was necessary to reflect theincreasing probability of intervention with experience of thehospital in correcting supravalve aortic stenosis.

402



/H


ownloaded from


Documents

Heart Prognosis ofsupravalve aortic stenosis in patients in Liverpool … · Heart 1996;75:396-402 Prognosis ofsupravalve aortic stenosis in 81 patients in Liverpool (1960-1993) DKitchiner,