CSITJMD

sslii

y>ts

o(avdt

Ahpsr[isTc[wd

asrct

A

Ace

©P

AD

ULT

CA

RD

IAC

linical Outcome of Asymptomatic Severe Aortictenosis With Medical and Surgical Management:mportance of STS Score at Diagnosishierry Le Tourneau, MD, Patricia A. Pellikka, MD, Morgan L. Brown, MD,

oseph F. Malouf, MD, Douglas W. Mahoney, MS, Hartzell V. Schaff, MD, andaurice Enriquez-Sarano, MD

ivision of Cardiovascular Diseases and Section of Biostatistics, Mayo Clinic, Rochester, Minnesota

i[1oaswe0

Aoigrc

Background. The Society of Thoracic Surgeons (STS)core aims at predicting operative mortality in cardiacurgery. The value of this score in predicting short- andong-term survival with medical or surgical managementn patients with asymptomatic severe aortic stenosis (AS)s unknown.

Methods. In a cohort of 694 patients (aged 71 � 11ears) with isolated, asymptomatic severe AS (velocity4 m/s), STS score was calculated at baseline and its link

o survival analyzed. Patients were stratified by STScore less than 4%, 4% to 6.5%, and 6.5% or greater.

Results. The STS score showed no association withperative mortality within 1 year of diagnosis or any time1%, 2.9%, and 6.1%, respectively, by strata; p � 0.08) and

weak association with 1-year survival (p � 0.04). Con-ersely, long-term survival (10-year) was strongly pre-icted by STS score strata (78%, 47%, and 16%, respec-

ively; p < 0.0001). In multivariate analysis, STS score

iladiAvaSou

acptdm

P

SA1

ular Diseases, Mayo Clinic, 200 First St SW, Rochester, MN 55905;-mail: sarano.maurice@mayo.edu.

2010 by The Society of Thoracic Surgeonsublished by Elsevier Inc

ndependently predicted mortality (hazard ratio/1%, 1.151.12 to 1.18], p < 0.0001) or cardiac death (1.21 [1.17 to.25], p < 0.0001). Aortic valve replacement within 1 yearf diagnosis markedly improved survival (adjusted haz-rd ratio, 0.58, p < 0.001). However, benefit of earlyurgery varied according to strata, with no overt benefitith low score (p � 0.83), whereas early surgery consid-

rably improved survival in the intermediate strata (p <.001).Conclusions. For patients with asymptomatic severeS, STS score is a powerful tool for predicting long-termutcome and for selecting patients (particularly those atntermediate risk) who benefit markedly from early sur-ery. Hence, risk-scoring using STS score should beoutinely performed in patients with AS to support thelinical decision-making process.

(Ann Thorac Surg 2010;90:1876–83)

© 2010 by The Society of Thoracic Surgeons

ortic valve stenosis (AS) is frequent, progressive,and predominant in the elderly [1]. Most patients

ave symptoms after diagnosis of severe AS [2], incuroor survival after symptom onset, and are relieved byurgery [2-5]. Hence, aortic valve replacement (AVR) isecommended for symptomatic patients with severe AS6, 7]. Conversely, management of asymptomatic patientss controversial. Risk with medical management is con-idered low, but sudden death remains a concern [2, 6-8].he risk of surgery is low, but rises with age andomorbidity [9, 10]. While some argue for early surgery5, 11], current guidelines acknowledge that watchfulaiting or referral to surgery rely on low level of evi-ence and remain debated [6, 7].The balance between operative and conservative man-

gement risk in individual asymptomatic patients withevere AS is difficult to assess. Thus, simple tools allowingisk estimation at diagnosis would be of major interest forlinical decision making. Scoring systems were developedo predict operative mortality after cardiac surgery [12-14],

ccepted for publication July 21, 2010.

ddress correspondence to Dr Enriquez-Sarano, Division of Cardiovas-

ncluding heart valve surgery [15, 16], and may predictong-term postoperative outcome [17, 18]. However,nalysis was focused on high-risk subsets [16], and noata are available regarding usefulness of risk-scoring

n subsets considered low risk such as asymptomaticS, regarding prediction of short- and long-term sur-ival and regarding impact on outcome with medicalnd surgical management. The Society of Thoracicurgeons (STS) score can be calculated with and with-ut assumptions for coronary lesions and may proveseful for these difficult clinical decisions.We therefore calculated STS score for a cohort of

symptomatic patients with severe AS and no history oforonary disease, and sought (1) to evaluate its value inredicting short-term and long-term outcome irrespec-

ive of the management strategy selected, and (2) toetermine its potential role in the clinical decision-aking process.

atients and Methods

tudy Patientsretrospective review of our records between 1984 and

995 identified 2,853 patients aged 40 years or more

0003-4975/$36.00doi:10.1016/j.athoracsur.2010.07.070

dvftvaptesgoR

DAdhtlDvww

STuacppbpvAE(a

TTvmaepmwayiotcaDpd

SCaSbtaatoKtlep�absmvaw

R

BBT(iF(5a(atoon

SMpla(prbp

(iaas

1877Ann Thorac Surg LE TOURNEAU ET AL2010;90:1876–83 STS SCORE IN ASYMPTOMATIC AS

AD

ULT

CA

RD

IAC

iagnosed with severe AS defined by a peak systolicelocity of 4 m/s or greater. Exclusion criteria were asollows: (1) cardiac symptoms at the time of first evalua-ion, (2) multivalvular involvement, (3) moderate to se-ere aortic regurgitation, (4) history of clinical coronaryrtery disease, and (5) prior aortic valve intervention orrior cardiac surgery of any cause. Ten additional pa-

ients who denied access to medical records were alsoxcluded from the study. Thus, 694 patients with isolated,evere, asymptomatic AS formed the present studyroup. Data on part of this series were reported previ-usly [2]. The study was approved by the Institutionaleview Board.

oppler-Echocardiographic Measurementsll patients underwent comprehensive Doppler two-imensional echocardiography at diagnosis. Rhythm,eart rate, and blood pressure were recorded. Left ven-

ricular ejection fraction (EF) was calculated. Transvalvu-ar aortic velocity was measured by continuous-waveoppler using multiple windows to obtain maximum jet

elocity. Maximal and mean gradient were calculatedith the modified Bernoulli equation. Aortic valve areaas measured using the continuity equation.

TS Scorehe STS score is based on various risk factors combined,sing a logistic equation to calculate the predicted oper-tive risk. The calculation uses specifically the extent oforonary lesion and need for coronary bypass, so that foratients in whom coronary angiography has not yet beenerformed, a score can be calculated assuming that noypass surgery will be associated with aortic valve re-lacement or that bypass may be required for three-essel disease or for coronary lesions of lesser severity.ll three scores were calculated. Logistic and additiveuropean System for Cardiac Operative Risk Evaluation

EuroSCORE) and Charlson comorbidity scores werelso calculated.

herapeutic Strategy and Follow-Uphe clinical decision regarding indications of balloonalvuloplasty, aortic valve surgery, or initial conservativeanagement was not known at index echocardiogram

nd was made by patients’ personal cardiologists. Afterxclusion of patients who underwent balloon valvulo-lasty (n � 20), patients were stratified according to theanagement approach selected: early surgery performedithin 1 year after diagnosis or initial conservative man-

gement (patients not referred for surgery within the firstear after diagnosis, but who could be referred later if anndication for surgery arose). During follow-up, clinicalutcome was monitored by return visits, mailed ques-ionnaires, and by telephone calls to patient, referringardiologist, and primary care physicians. Informationbout timing and cause of mortality were obtained.eaths were classified as cardiac or noncardiac. Unex-lained and sudden deaths were considered cardiac

eath. m

tatistical Analysisontinuous variables are summarized as mean � SD,nd comparisons between groups were performed usingtudent’s t tests or one-way analysis of variance, followedy Bonferroni tests. Categorical data were compared be-

ween groups using �2 tests. Short-term endpoints werenalyzed using logistic regression. Receiver-operating char-cteristic curve was used to assess the accuracy of STS scoreo predict operative mortality. Long-term endpoints wereverall survival and cardiac death, summarized using theaplan-Meier method and compared using the log rank

est. Univariate and multivariate association with base-ine characteristics used Cox proportional hazards mod-ls. For display purpose, baseline characteristics wereresented stratified by STS score terciles (�4%, 4% to6.5%, �6.5%). Analysis was conducted for STS score

ssuming no bypass surgery required and also assumingypass requirement. Survival after diagnosis was pre-ented stratified by STS score terciles and by manage-ent strategy (surgery or medical follow-up). Aortic

alve surgery was also analyzed as a time varying covari-te in Cox proportional models. A p value less than 0.05as considered statistically significant.

esults

aseline characteristicsaseline characteristics are shown in the left column ofable 1. The AS etiology was degenerative in 501 patients

72%), bicuspid or congenital in 158 (23%), and rheumaticn 38 (5%). Aortic valve area averaged 0.86 � 0.23 cm2.ollow-up, complete in 99.8%, was 3,817 patient-yearsaverage �5 years). The STS score (without bypass) was.8% � 3.5% overall, and was 3.0% � 0.6%, 5.1 � 0.7%,nd 9.7 � 3.6% for the low (�4%, n � 251), intermediate4% to �6.5%, n � 224), and high (�6.5%, n � 219)pproximate terciles, respectively. Baseline characteris-ics stratified by STS score terciles are shown in right partf Table 1. With higher score, patients were older, moreften female, and had more comorbidities, but there waso difference in indexed aortic valve area (p � 0.50).

TS Score and Outcomeortality within 1 year after diagnosis occurred in 35

atients. One-year survival was slightly lower betweenow, intermediate, and high STS score terciles (97%, 95%,nd 92%, respectively; p � 0.04) whereas cardiac deathp � 0.18) did not differ between terciles. Independentredictors of 1-year mortality were aortic velocity (oddsatio 2.27 [1.15 to 4.51] per m/s; p � 0.019) and systoliclood pressure (odds ratio 1.02 [1.01 to 1.04] per mm Hg;� 0.006).For long-term outcome, death any time after diagnosis

under medical or after surgical management) occurredn 289 patients, with 5- and 10-year survival of 71% � 2%nd 49% � 2%. Ten-year survival (Fig 1) was 78%, 47%,nd 16% (p � 0.0001) in the low, intermediate, and high STScore subsets, respectively (hazard ratio 1.16 per 1% incre-

ent, 95% confidence interval: 1.13 to 1.19, p � 0.0001).

ad0i

1r9Sp01

STlvemssc1a

tidtsi

wbspcfspOa

cscwms

T

C

E

a

i

Ae peratp

1878 LE TOURNEAU ET AL Ann Thorac SurgSTS SCORE IN ASYMPTOMATIC AS 2010;90:1876–83

AD

ULT

CA

RD

IAC

Independent predictors of long-term all cause mortalityre listed Table 2. In addition to history of cancer (p � 0.026),iabetes mellitus (p � 0.02), and atrial fibrillation (p �.0002), STS score (hazard ratio 1.15 [1.12 to 1.18] per 1%ncrement; p � 0.0001) independently predicted survival.

Cause of death was cardiac in 117 patients, with 5- and0-year survival free of cardiac death of 86% and 76%,espectively. Ten-year survival free of cardiac death was4%, 81%, and 41% in the low, intermediate, and highTS score groups, respectively (p � 0.0001). Independentredictors of cardiac death were atrial fibrillation (p �.0001) and STS score (hazard ratio 1.21 [1.17 to 1.25] per% increment; p � 0.0001).

TS Score and Management Strategywenty patients (2.9%) underwent balloon aortic valvu-

oplasty an average of 1 year after diagnosis. As balloonalvuloplasty is not a usual therapeutic option in AS, wexcluded these patients from therapeutic strategy assess-ent. Thus, the remaining patients (n � 674) were

tratified according to initial (first-year) managementtrategy (surgical or conservative management). Baselineharacteristics comparisons between initial surgical (n �60) and conservative (n � 514) management strategies

able 1. Baseline Characteristics of the Overall Patient Popul

Overall Low (�4%, n �

linical characteristicsAge, years 71 � 11 61 � 9Male, n (%) 419 (60) 194 (77)BMI, kg/m2 27 � 5 28 � 6Hypertension, n (%) 306 (44) 71 (28)Systolic BP, mm Hg 146 � 24 139 � 24Diabetes mellitus, n (%) 74 (10.7) 6 (2.4)History of cancer, n (%) 152 (22) 48 (19)Atrial fibrillation, n (%) 32 (4.6) 6 (2.4)LV hypertrophy, n (%) 126 (18) 43 (17)Charlson index 3.6 � 2.5 2.3 � 1.9Logistic EuroSCORE (%) 5.6 � 4.9 2.4 � 1.1STS score (%) 5.7 � 3.5 3.0 � 0.6

chocardiographic dataAV velocity, m/s 4.4 � 0.5 4.4 � 0.5AV mean gradient, mm Hg 46 � 11 46 � 11AV area/BSA, cm2/m2 0.46 � 0.11 0.46 � 0.11LV EDD, mm 49 � 6 51 � 6Septum thickness, mm 13.6 � 2.5 14 � 3Left ventricular EF,% 64 � 7 63 � 7LV EF �50%, n (%) 26 (4.6) 10 (4.0)Systolic PAP, mm Hg 37 � 8 32 � 4

p � 0.0001 versus low-risk group. b p � 0.0001 versus intermedintermediate-risk group.

V � aortic valve; BMI � body mass index; BP � blood pressurjection fraction; EuroSCORE � European System for Cardiac Oressure; STS � The Society of Thoracic Surgeons.

re shown Table 3. There are multiple differences be- o

ween initial surgical and conservative groups. However,n a comparison stratified by STS score groups, theseifferences became insignificant (all p � 0.10), showing

hat within each risk strata, baseline characteristics wereimilar whether patients were referred to initial surgicalntervention or to conservative management.

With initial surgical management, surgery indicationas rarely development of symptoms (45 patients, 28%)ut more often progression of AS or of left ventricularystolic dysfunction (56 patients, 35%) and patient/hysician preference (59 patients, 37%). Among initialonservative management, surgery was ultimately per-ormed in 248 patients, with main indication of newymptoms (157 patients, 63%), other cardiac surgery (47atients, 19%), and miscellaneous reasons (44 patients).verall, 130 surgery patients (32%) required coronary

rtery bypass graft surgery.Operative mortality was only 1.9% in the initial surgi-

al group (3 of 160 patients) versus 4.8% predicted by STScore, but this difference did not reach statistical signifi-ance (p � 0.35). Operative mortality for delayed surgeryas low (2.8%, 7 of 248 patients), so that overall operativeortality was 2.5% compared with 4.6% predicted by STS

core (p � 0.40). Stratified according to STS score groups,

and After Stratification According to STS Score

STS Score

p Value)Intermediate

(4% to �6.5% , n � 224)High

(�6.5%, n � 219)

73 � 6a 82 � 6ab �0.0001136 (61)a 89 (41)ab �0.0001

26 � 5a 25 � 5ad �0.0001103 (46)a 132 (60)ad �0.0001151 � 22a 148 � 24c �0.000124 (11)a 44 (20)ad �0.000156 (25) 49 (22) 0.3015 (6.7)c 11 (5.0) 0.0837 (17) 46 (21) 0.413.9 � 2.4a 4.9 � 2.5ab �0.00014.8 � 1.9a 10.4 � 6.1ab �0.00015.1 � 0.7a 9.7 � 3.6ab �0.0001

4.4 � 0.5 4.4 � 0.4 0.6846 � 12 47 � 11d 0.45

0.46 � 0.11 0.45 � 0.10 0.6148 � 6a 46 � 6a �0.000114 � 2 14 � 2 0.8364 � 7 64 � 7 0.337 (3.1) 9 (4.1) 0.84

35 � 6 41 � 8cd 0.002

sk group. c p � 0.05 versus low-risk group. d p � 0.05 versus

BSA � body surface area; EDD � end-diastolic diameter; EF �ive Evaluation; LV � left ventricle; PAP � pulmonary artery

ation

251

ate-ri

e;

perative mortality showed no significant difference ei-

tto0nltaip

gd0

s0aTcosbs

nfilaCibRwlslar

C

TSpTIoanafstfitpmw

TP

A

HDAS

S

1879Ann Thorac Surg LE TOURNEAU ET AL2010;90:1876–83 STS SCORE IN ASYMPTOMATIC AS

AD

ULT

CA

RD

IAC

her for early surgery or surgery at any time (Fig 2), buthere was a trend, with higher STS score, for higherperative mortality in surgery performed at any time (p �.08). However, in multivariate analysis, STS score didot predict operative mortality (p � 0.30), which was

inked to diabetes mellitus (p � 0.014) and atrial fibrilla-ion (p � 0.046). Receiver-operating characteristic curveccuracy was 0.67 for STS score (p � 0.15), confirming thenability of the score to predict operative mortality in theresent study.For short-term (1-year) mortality, therapeutic strate-

ies (initially, surgical versus conservative) showed noifference within score terciles for overall mortality (p �.70) or cardiac death (p � 0.50).

For long-term outcome, survival was higher after initialurgical management (adjusted hazard ratio 0.58, p �.001), irrespective of adjustment by STS score and afterdjustment by all clinical predictors of outcome (Table 4).ime-dependent analysis of AVR performed any timeonfirmed the considerable benefit of surgery in all strataf these asymptomatic patients with severe AS (Table 4,econd line for each stratum). Analyzing the potentialenefit of early surgery (within 1 year of diagnosis),tratification by STS score showed that the benefit was

able 2. Predictors of Long-Term Survival of Asymptomaticatients With Severe Aortic Stenosis

ll-Cause MortalityHazardRatio

95% ConfidenceInterval p Value

istory of cancer 1.30 1.00–1.69 0.026iabetes mellitus 1.52 1.07–2.11 0.02trial fibrillation 2.65 1.64–4.05 0.0002TS score, per 1% 1.15 1.12–1.18 �0.0001

sTS � The Society of Thoracic Surgeons.

ot uniformly distributed to all scoring strata (Table 4,rst line for each strata, and Fig 3). Indeed, risk was very

ow, and there was no difference between the two ther-peutic strategies in the low STS score group (p � 0.83).onversely, survival was higher after early surgery in the

ntermediate STS score group (p � 0.0007) and tended toe higher (p � 0.09) in the high STS score group (Fig 3).esults were similar when patients having symptomsithin the first year were excluded. Hence, mortality was

ow and identical for medical and surgical therapeutictrategies in the low-risk groups. Conversely, there was aarge survival benefit after early surgery for intermedi-te-risk patients, and a borderline benefit for the high-isk patients.

omment

he present study reports the outcome implications ofTS scoring in a large series of initially asymptomaticatients with severe AS treated medically and surgically.he results of our study are paradoxic but important.

ndeed, it is paradoxic that STS score intended to predictperative mortality demonstrated weak links to operativend short-term mortality. Conversely, STS score at diag-osis predicts long-term survival for patients withsymptomatic severe AS. Additionally, STS score is use-ul for risk stratification and management: while earlyurgery is associated with overall improved survival,here is a differential benefit depending on score strati-cation. Contrary to the expectation that low-score pa-

ients are best positioned to benefit from surgery, suchatients incur very low risk irrespective of the manage-ent strategy selected and display no definite benefitith early surgery 10 years after diagnosis. Conversely,

Fig 1. Overall survival stratified by The Soci-ety of Thoracic Surgeons (STS) score terciles:less than 4% (heavy line); 4% to 6.5%(dashed line); and 6.5% or greater (light line).Five-year and 10-year survival rates are indi-cated. Note the markedly different survivalaccording to each STS score tercile.

urgery improves survival for intermediate- and high-

rpi

ac

RTtawTtsm[aAr

dSssa

Fwagib

TA(

C

E

R

ABSvn

TAM

AS

A

C

C

1880 LE TOURNEAU ET AL Ann Thorac SurgSTS SCORE IN ASYMPTOMATIC AS 2010;90:1876–83

AD

ULT

CA

RD

IAC

isk patients. Thus, STS score, appropriately interpreted,rovides useful prognostic information and should be

ntegrated into the clinical decision-making process for

ig 2. Observed operative mortality (OM) after surgery performedithin 1 year of diagnosis (white bars) and after surgery performedt any time (gray bars) compared with The Society of Thoracic Sur-eons (STS) score-predicted mortality (black bars). Patients are strat-fied by STS score tercile. Operative mortality at any time tended to

able 3. Baseline Characteristics in 674 Asymptomaticortic Stenosis Stratified According to Initial Surgical

Surgery Within 1 Year) or Conservative Management

SurgeryWithin1 Year

(n � 160)

ConservativeManagement

(n � 514) p Value

linical characteristicsAge, years 69 � 10 72 � 11 0.01Male, n (%) 91 (57) 322 (62) 0.20BMI, kg/m2 26 � 4 27 � 6 0.39Hypertension, n (%) 64 (40) 229 (45) 0.32Systolic BP, mm Hg 145 � 25 147 � 24 0.57Diabetes mellitus, n (%) 8 (5.0) 65 (12.6) 0.005History of cancer, n (%) 27 (17) 122 (24) 0.08Atrial fibrillation, n (%) 6 (3.8) 25 (4.9) 0.67

chocardiographic dataAV velocity, m/s 4.6 � 0.5 4.3 � 0.4 �0.0001AV mean gradient, mm

Hg53 � 13 45 � 10 �0.0001

AV area/BSA, cm2/m2 0.40 � 0.10 0.48 � 0.11 �0.0001AV area �0.6 cm2,

n (%)17 (10.6) 16 (3.1) �0.0001

LV EDD, mm 48.9 � 6.5 48.8 � 6.3 0.87Septum thickness, mm 13.9 � 2.6 13.5 � 2.4 0.17LVEF,% 63.8 � 7.3 63.7 � 7.1 0.88LVEF �50%, n (%) 7 (4.4) 18 (3.5) 0.63Systolic PAP, mm Hg 40 � 11 37 � 7 0.30

isk scoresSTS score, % 4.8 � 2.1 6 � 3.7 0.0002Logistic EuroSCORE, % 4.5 � 2.9 5.9 � 5.4 0.002

V � aortic valve; BMI � body mass index; BP � blood pressure;SA � body surface area; EDD � end-diastolic diameter; Euro-CORE � European System for Cardiac Operative Evaluation; LV � leftentricle; LVEF � left ventricular ejection fraction; PAP � pulmo-ary artery pressure; STS � The Society of Thoracic Surgeons.

re higher with higher STS score (p � 0.08).

symptomatic patients with severe AS to select thoseonsidered for early intervention.

isk Scoringhe STS score aims at predicting cardiac operative mor-

ality [14, 16]. In our AS population, STS score displayedn insignificant link to operative mortality, and no or aeak association with short-term mortality (1 year).hus, in that regard, STS score was disappointing, but

hat may be in part due to the low operative andhort-term mortality observed in our study. Operativeortality, in our series, tended to be lower than predicted

16, 19], a difference possibly linked to volume of surgicalctivity [20]. Thus, it is essential not to contraindicateVR purely based on high score, as acceptable operative

isk may be achieved in high-volume centers [16].However, for long-term outcome, STS score was not

isappointing. Recent studies suggested that the Euro-CORE may predict outcome beyond the immediateurgical phase, whether coronary artery bypass grafturgery [17], AVR [21, 22], or mitral valve surgery [18]re performed. However, sparse data are available

able 4. Impact of Surgical Therapeutic Strategy onll-Cause Mortality and Cardiac Death in Multivariate Coxodel

djustment andtratification

STS Score

Hazard Ratio 95% CI p Value

ll-cause mortalityOverall

Surgery within 1 year 0.58 0.41–0.80 0.0008Surgery at any time 0.39 0.29–0.53 �0.0001

Low score STS score �4%Surgery within 1 year 1.04 0.51–1.97 0.91Surgery at any time 0.43 0.21–0.87 0.02

Intermediate score STS score 4% to �6.5%Surgery within 1 year 0.38 0.22–0.64 0.0002Surgery at any time 0.33 0.21–0.54 �0.0001

High score STS score �6.5%Surgery within 1 year 0.67 0.36–1.16 0.16Surgery at any time 0.46 0.29–0.72 0.0007

ardiac deathOverall

Surgery within 1 year 0.54 0.29–0.92 0.022Surgery at any time 0.31 0.19–0.51 �0.0001

Low score STS score �4%Surgery within 1 year 2.6 0.73–9.04 0.13Surgery at any time 0.42 0.10–1.70 0.22

Intermediate score STS score 4% to �6.5%Surgery within 1 year 0.32 0.11–0.78 0.01Surgery at any time 0.33 0.15–0.72 0.005

High score STS score �6.5%Surgery within 1 year 0.58 0.19–1.74 0.22Surgery at any time 0.28 0.14–0.58 0.0006

I � confidence interval; STS � The Society of Thoracic Surgeons.

egarding the value of STS score for predicting long-

tstsrppsc

RAI[AenpH

toiassnbuaeeasi

lbdhmteassatfrtpddssmiegpfiyriAim

POtafrspod

Fgms(lip

1881Ann Thorac Surg LE TOURNEAU ET AL2010;90:1876–83 STS SCORE IN ASYMPTOMATIC AS

AD

ULT

CA

RD

IAC

erm survival [14, 16]. In our study of severe AS, STScore was a good and independent predictor of long-erm survival and cardiac death with medical andurgical management. Hence, STS score, beyond itsole as a presurgery tool, is a powerful and incrementalredictor of long-term survival among asymptomaticatients with severe AS. Thus, we believe that STScore should be measured and used as part of thelinical decision-making process.

isk-Scoring and Decision Making insymptomatic AS

n severe AS, symptoms are predictors of poor outcome3, 4] and are regarded as a class I recommendation forVR [6, 7]. Aortic valve replacement can also be consid-red for asymptomatic patients with depressed EF, ab-ormal exercise test, or with critical AS when the ex-ected operative mortality is low (�1%) [6-8, 23, 24].

ig 3. Overall survival after diagnosis compared between early sur-ery (within 1 year of diagnosis [solid line]) or initial conservativeanagement (with possible later valve replacement [dashed line])

tratified by tercile of The Society of Thoracic Surgeons (STS) scorelow risk, upper graph; intermediate risk, middle graph; high risk,ower graph). Note the considerable survival benefit of surgery forntermediate-risk patients and the borderline benefit for high-riskatients.

owever, the management of most asymptomatic pa- [

ients remains difficult, and current guidelines are basedn a low level of evidence [6, 7]. Although close monitor-

ng with surgery delayed until symptoms develop is oftendvised [4, 6, 8, 25], some argue for early surgery in thisetting [5, 11], particularly when the benefit/risk ratio ofurgery is high [6, 26]. Therefore, it is essential to identifyew risk markers that would provide insights into thisenefit/risk ratio. Our intended goal was to identify,sing STS score, the patients at low surgical risk (ie, withssumed high benefit/risk ratio) who may benefit fromarly surgery. In that regard, results are surprising, albeitncouraging. Encouraging is the link between score levelnd long-term benefit of management strategies, buturprising is that benefit of AVR was not where wenitially expected it.

Indeed, in the low-score group, survival was excel-ent with both strategies of treatment and they did notenefit from early surgery. Conversely, in the interme-iate-score group, long-term survival was clearlyigher after early surgery than with medical manage-ent, without excess operative-related mortality. In

his intermediate-risk group, the survival benefit ofarly surgery was high (mortality reduction by 60%pproximately; Table 4) and sustained. In the high STScore group, there was a nonsignificant trend for lowerurvival after surgery during the first year, and there-fter survival curves crossed, and ultimately, long-erm survival tended to be higher with early referralor surgery. This result was not expected in patients atelatively high risk for surgery. The higher postopera-ive versus medical survival may be related to thereoperative underestimation of symptoms in an el-erly population with decreased activity and poorlyefined functional capacity [27]. Older patients withevere AS are rarely referred for surgery, even afterymptom onset [27, 28], and tend to fare poorly underedical management. Thus, in clinical practice, it is

mportant at diagnosis to balance risks and benefits ofarly AVR, carefully weighing whether expected lon-evity allows potential benefit of surgery, and for thaturpose STS score is a useful tool. Conversely, ourndings suggest little urgency in offering AVR toounger asymptomatic patients with a low score. Theseisk assessment considerations should participate inndividualized clinical decision making regardingVR, taking into account compliance with close mon-

toring, comorbidity, and center-specific operativeortality achieved.

otential Limitationsur outcome study is not a clinical trial testing therapeu-

ic strategies for asymptomatic severe AS patients. It aimst assessing STS score to define subsets that may benefitrom an aggressive AVR strategy. A clinical trial, war-anted in view of the potential benefit of early surgery,hould preferentially target intermediate- and high-riskatients, rather than low-risk patients for whom benefitf surgery would be at best marginal. Severe AS wasefined (aortic velocity �4 m/s) according to guidelines

2, 23, 29], but questions remain regarding the best

df

recboavp1siE

cm

ddincds

sAwppiisscdea

DF

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

1882 LE TOURNEAU ET AL Ann Thorac SurgSTS SCORE IN ASYMPTOMATIC AS 2010;90:1876–83

AD

ULT

CA

RD

IAC

efinition of severe AS [30] and should be addressed inuture studies.

The STS score involves invasive measurements thatequire assumptions in a medical population. How-ver, no patient had a history of coronary disease andalculations were also performed with added risk ofypass surgery, but these calculations did not changeur results. Moreover, stratification of patients bydditive EuroSCORE terciles (0 to 4, 5 to 6, �6) pro-ides similar results. Indeed, EuroSCORE is an inde-endent predictor of survival (hazard ratio 1.37 [1.31 to.45] per 1 point increment; p � 0.0001), and earlyurgery is associated with a clear improved survival inntermediate-risk (p � 0.0004) and high-risk (p � 0.008)uroSCORE groups.Initial surgical management was selected with a 1-year

utoff, but with a 6-month cutoff, it remained superior toedical management (p � 0.0001).Patients were asymptomatic, rarely with ventricular

ysfunction or comorbidity, had no known coronaryisease, prior cardiac interventions, or multivalvular

nvolvement, and have long been considered the “be-ign” end of the AS spectrum [29]. Therefore, weannot extend our results to patients with severelyepressed left ventricular function or patients withevere comorbidity.

In conclusion, in this large population of asymptomaticevere AS, STS score is a useful risk assessment tool.lthough STS score was paradoxically weakly associatedith operative and short-term mortality, it powerfullyredicted long-term survival irrespective of the thera-eutic strategy. Moreover, scoring provided a risk strat-

fication useful for defining patients’ subsets (particularlyntermediate-score groups) for whom early surgery sub-tantially improves long-term survival. Hence, risk as-essment using STS score should be performed duringlinical evaluation, and should be part of the clinicalecision-making process, involving consideration ofarly surgery for patients with severe AS, even if they aresymptomatic at diagnosis.

r Le Tourneau was supported by a grant from the Frenchoundation of Cardiology.

eferences

1. Otto CM, Lind BK, Kitzman DW, Gersh BJ, Siscovick DS.Association of aortic-valve sclerosis with cardiovascularmortality and morbidity in the elderly. N Engl J Med1999;341:142–7.

2. Pellikka PA, Sarano ME, Nishimura RA, et al. Outcome of622 adults with asymptomatic, hemodynamically significantaortic stenosis during prolonged follow-up. Circulation 2005;111:3290–5.

3. Horstkotte D, Loogen F. The natural history of aortic valvestenosis. Eur Heart J 1988;9(Suppl E):57–64.

4. Ross J, Braunwald E. Aortic stenosis. Circulation 1968;38:61–7.

5. Lund O. Preoperative risk evaluation and stratification oflong-term survival after valve replacement for aortic steno-

sis. Reasons for earlier operative intervention. Circulation1990;82:124–39.

6. Bonow RO, Carabello BA, Kanu C, et al. ACC/AHA 2006guidelines for the management of patients with valvularheart disease: a report of the American College of Cardiol-ogy/American Heart Association Task Force on PracticeGuidelines. Circulation 2006;114:e84–231.

7. Vahanian A, Baumgartner H, Bax J, et al. Guidelines on themanagement of valvular heart disease: the task force on themanagement of valvular heart disease of the EuropeanSociety of Cardiology. Eur Heart J 2007;28:230–68.

8. Rosenhek R, Maurer G, Baumgartner H. Should earlyelective surgery be performed in patients with severebut asymptomatic aortic stenosis? Eur Heart J 2002;23:1417–21.

9. Edwards FH, Peterson ED, Coombs LP, et al. Prediction ofoperative mortality after valve replacement surgery. J AmColl Cardiol 2001;37:885–92.

0. Ambler G, Omar RZ, Royston P, Kinsman R, Keogh BE,Taylor KM. Generic, simple risk stratification model forheart valve surgery. Circulation 2005;112:224–31.

1. Rapaport E. Natural history of aortic and mitral valve dis-ease. Am J Cardiol 1975;35:221–7.

2. Nashef SA, Roques F, Michel P, Gauducheau E, LemeshowS, Salamon R. European system for cardiac operative riskevaluation (EuroSCORE). Eur J Cardiothorac Surg 1999;16:9–13.

3. Roques F, Michel P, Goldstone AR, Nashef SA. The logisticEuroSCORE. Eur Heart J 2003;24:881–2.

4. Ad N, Barnett SD, Speir AM. The performance of theEuroSCORE and the Society of Thoracic Surgeons mortalityrisk score: the gender factor. Interact Cardiovasc ThoracSurg 2007;6:192–5.

5. Roques F, Nashef SA, Michel P. Risk factors for earlymortality after valve surgery in Europe in the 1990s: lessonsfrom the EuroSCORE pilot program. J Heart Valve Dis2001;10:572–8.

6. Dewey TM, Brown D, Ryan WH, Herbert MA, Prince SL,Mack MJ. Reliability of risk algorithms in predicting earlyand late operative outcomes in high-risk patients undergo-ing aortic valve replacement. J Thorac Cardiovasc Surg2008;135:180–7.

7. Toumpoulis IK, Anagnostopoulos CE, DeRose JJ, SwistelDG. European system for cardiac operative risk evaluationpredicts long-term survival in patients with coronaryartery bypass grafting. Eur J Cardiothorac Surg 2004;25:51– 8.

8. Heikkinen J, Biancari F, Satta J, et al. Predicting immediateand late outcome after surgery for mitral valve regurgitationwith EuroSCORE. J Heart Valve Dis 2007;16:116–21.

9. Grossi EA, Schwartz CF, Yu PJ, et al. High-risk aortic valvereplacement: are the outcomes as bad as predicted? AnnThorac Surg 2008;85:102–7.

0. Goodney PP, O’Connor GT, Wennberg DE, Birkmeyer JD.Do hospitals with low mortality rates in coronary arterybypass also perform well in valve replacement? Ann ThoracSurg 2003;76:1131–7.

1. Walther T, Rastan A, Falk V, et al. Patient prosthesismismatch affects short- and long-term outcomes afteraortic valve replacement. Eur J Cardiothorac Surg 2006;30:15–9.

2. Monin JL, Monchi M, Kirsch ME, et al. Low-gradient aorticstenosis: impact of prosthesis-patient mismatch on survival.Eur Heart J 2007;28:2620–6.

3. Otto CM, Burwash IG, Legget ME, et al. Prospective study ofasymptomatic valvular aortic stenosis. Clinical, echocardio-graphic, and exercise predictors of outcome. Circulation1997;95:2262–70.

4. Lancellotti P, Lebois F, Simon M, Tombeux C, Chauvel C,Pierard LA. Prognostic importance of quantitative exerciseDoppler echocardiography in asymptomatic valvular aortic

stenosis. Circulation 2005;112:I377–82.

2

2

2

2

2

3

N

TMetdbw

c(

1883Ann Thorac Surg LE TOURNEAU ET AL2010;90:1876–83 STS SCORE IN ASYMPTOMATIC AS

©P

RD

IAC

5. Braunwald E. On the natural history of severe aortic steno-sis. J Am Coll Cardiol 1990;15:1018–20.

6. Carabello BA. Timing of valve replacement in aorticstenosis. Moving closer to perfection. Circulation 1997;95:2241–3.

7. Bouma BJ, van Den Brink RB, van Der Meulen JH, et al. Tooperate or not on elderly patients with aortic stenosis: thedecision and its consequences. Heart 1999;82:143– 8.

8. Iung B, Cachier A, Baron G, et al. Decision-making in elderly

otice From the American Boa

written) examination.

ei

BC2

2010 by The Society of Thoracic Surgeonsublished by Elsevier Inc

patients with severe aortic stenosis: why are so many deniedsurgery? Eur Heart J 2005;26:2714–20.

9. Pellikka PA, Nishimura RA, Bailey KR, Tajik AJ. The naturalhistory of adults with asymptomatic, hemodynamically sig-nificant aortic stenosis. J Am Coll Cardiol 1990;15:1012–7.

0. Hachicha Z, Dumesnil JG, Bogaty P, Pibarot P. Paradoxicallow-flow, low-gradient severe aortic stenosis despite pre-served ejection fraction is associated with higher afterloadand reduced survival. Circulation 2007;115:2856 – 64.

AD

ULT

CA

rd of Thoracic Surgery

he 2011 Part I (written) examination will be held ononday, November 21, 2011. It is planned that the

xamination will be given at multiple sites throughouthe United States using an electronic format. The closingate for registration is August 15, 2011. Those wishing toe considered for examination must apply online atww.abts.org.To be admissible to the Part II (oral) examination, a

andidate must have successfully completed the Part I

A candidate applying for admission to the certifyingxamination must fulfill all the requirements of the Boardn force at the time the application is received.

Please address all communications to the Americanoard of Thoracic Surgery, 633 N St. Clair St, Suite 2320,hicago, IL 60611; telephone: (312) 202-5900; fax: (312)02-5960; e-mail: info@abts.org.

Ann Thorac Surg 2010;90:1883 • 0003-4975/$36.00