Clinical Outcome in Aortic Regurgitation with ... - .pdfAortic regurgitation (AR) remains an important cardiac condition,1, 2 although substantial chronic regurgitation can be tolerated

Jane M. Francis, Adrian P. Banning, Jonathan P. Christiansen and Stefan NeubauerSaul G. Myerson, Joanna d'Arcy, Raad Mohiaddin, John P. Greenwood, Theodoros D. Karamitsos,

Clinical Outcome in Aortic Regurgitation with Cardiovascular Magnetic Resonance

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 2012 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation published online August 9, 2012;Circulation.

http://circ.ahajournals.org/content/early/2012/08/09/CIRCULATIONAHA.111.083600World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions:

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

document. Permissions and Rights Question and Answer available in the

Permissions in the middle column of the Web page under Services. Further information about this process isOnce the online version of the published article for which permission is being requested is located, click Request

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Circulation Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

at Ohio State University--Columbus on September 4, 2012http://circ.ahajournals.org/Downloaded from

http://circ.ahajournals.org/content/early/2012/08/09/CIRCULATIONAHA.111.083600

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://circ.ahajournals.org//subscriptions/

http://circ.ahajournals.org/

DOI: 10.1161/CIRCULATIONAHA.111.083600

1

Clinical Outcome in Aortic Regurgitation with

Cardiovascular Magnetic Resonance

Running title: Myerson et al.; Aortic Regurgitation: Outcome with CMR

Saul G. Myerson, MBChB, MD, MRCP, FESC1; Joanna d’Arcy, MBChB, MRCP1;

Raad Mohiaddin, PhD, FRCR, FRCP, FESC2; John P. Greenwood, MBChB, PhD3; Theodoros D.

Karamitsos, MD, PhD1; Jane M. Francis, DCR(R), DNM1; Adrian P. Banning, MBBS, MD,

FRCP, FESC1; Jonathan P. Christiansen, MBChB, MD, FRACP, FACC, FCSANZ4;

Stefan Neubauer, MD, FRCP, FACC, FMedSci1

1Depts. of Cardiology & Cardiovascular Medicine; University of Oxford Centre for Clinical Magnetic Resonance Research, John Radcliffe Hospital, Oxford, 2CMR Unit, Royal Brompton Hospital and the National Heart and Lung Institute, London; 3Multidisciplinary Cardiovascular Research Centre (MCRC) & Leeds Institute of Genetics, Health and Therapeutics, University of

Leeds, Leeds, United Kingdom; 4North Shore Hospital, Auckland, New Zealand

Address for Correspondence:

Dr. Saul Myerson

Dept. of Cardiovascular Medicine

John Radcliffe Hospital

Headley Way

Oxford OX3 9DU, United Kingdom

Tel: +44 1865 222770

Fax: +44 1865 740449

E-mail: [email protected]

Journal Subject Codes: [30] CT and MRI; [124] Cardiovascular imaging agents/Techniques; [38] CV surgery: valvular disease

Karamitsos, MD, PhD ; Jane M. Francis, DCR(R), DNM ; Adrian P. Banning, MMBBBBB S,S, MMD,D

FRCP, FESC1; Jonathan P. Christiansen, MBChB, MD, FRACP, FACC, FCFCFCSASASANZNZNZ444; ;;

Stefan Neubauer, MD, FRCP, FACC, FMedSci1

11DDeDeptp s. oof f f CaCardrdioiolologygygy && CCarardidiovovaascsculularar MMedediciccinnne; UUniniveversitityy y oofof OOxfxforo d CeCentntrere fforor CClilinicacal fMaMagnetic Resesonono aaancecece RRReseseseaeaearcrcrchh,h, JJJohohohnn RRaRadcdclil fffffe HoHoHospitititalalal, , OxOxOxfofordrdr ,, 222CMCMCMRRR UnUnUnititit,, RoRoRoyayayall l BrBrBromomomptptptoon HoHoosps ital andnd thhee NNattiionnal l HeHeHeart t annd d LuLuung IInnnstituuutee,, LoLoondndndononon; 3MuMuMultidiiisccciplpllinnnarry CCaCardrdiiiovvvascccuullar ReReReseseearara chch CCCeneentrtreee ((M(MCRCRCRC)C) &&& LLLeeeeeedsdsds IInnsnstitiitutuutetete ooof f GGGennnettiticscss, , , HeHHealalalththh aaandndnd TTThheerarapepep ututu icics,s,s, UUnnnivveversrsiiityy y ofo

LLeLeeeds,s LLLeeeedsdsd , UUnUniti edd KKKiininggdomm;; 444NNoNortrthh ShShShorree HoHHospspiititaal, AuAuA ckckkllala dndnd, NeNew w ZZZealallana dd d

AdAddd ff CC dd




2

Abstract:

Background - Current indications for surgery in patients with significant aortic regurgitation

(AR) focus on symptoms and left ventricular (LV) dilation/dysfunction. However, prognosis is

already reduced by this stage and earlier identification of patients for surgery could be beneficial.

Quantifying the regurgitation may help, but there is limited data on its link with outcome.

Cardiovascular magnetic resonance (CMR) can accurately quantify AR, and we examined

whether this was associated with the future need for surgery.

Methods and Results - 113 patients with echocardiographic moderate or severe AR were

monitored for up to 9 years (mean 2.6±2.1 years) following a CMR scan, and the progression to

symptoms or other indications for surgery was monitored. AR quantification identified outcome

with high accuracy: 85% of the 39 subjects with regurgitant fraction >33% progressed to surgery

(mostly within 3 years) compared to 8% of 74 subjects with regurgitant fraction 33%

(p<0.0001); area under the curve on receiver operating characteristic analysis 0.93 (p<0.0001).

This ability remained strong on time-dependent Kaplan-Meier survival curves. CMR-derived

end-diastolic volume (LVEDV) >246ml had good, though lower, discriminatory ability (AUC

0.88) but the combination of this with regurgitant fraction provided the best discriminatory

power.

Conclusions - High degrees of CMR-quantified AR were associated with the development of

symptoms or other indications for surgery. Quantifying AR showed slightly better discriminatory

ability than ‘Gold-standard’ CMR ventricular volume assessment. This could provide a new

paradigm for the timing of surgical intervention but requires confirmation in a clinical trial.

Key words: aortic regurgitation; aortic valve replacement; cardiovascular magnetic resonance imaging; outcome; prognosis

monitored for up to 9 years (mean 2.6±2.1 years) following a CMR scan, and thehe ppprororogrgrgresesessisisiononon tto

ymptoms or other indications for surgery was monitored. AR quantification identified outcome

withth hhhigigighhh acaca cucuc raaacycycy:: 85% of the 39 subjects with h rereregugg rgitant fraction >>333333%% progressed to surgery

mmmoosostly withhiniin 333 yyeaearsrsrs))) cococompmpmparararededed tttooo 8%8%8% oof ff 747474 ssubbbjeects s s wiwiwiththth rrregeggururu gigigitatatanntnt ffractcttioioionn n 33333 % % %

ppp<0<0<0.0. 000 1));; ararareaea uuundndeer ttthe cccururrvevev oon n n rereeceeeiviveerer ooopeperrarattitingngg ccchaharraractctererrissttic aannaalalysyssiss 00.99333 (p(p(p<<0..0000001))).

ThThThisisis aaabibibilililitytyty rrremememaiaiaineneneddd stststrororongngng ooonnn tititimememe d-d-depepepenenendededentntnt KaKaKaplplplananan M-M-Meieieiererer sssurururvivivivavavalll cucucurvrvrveseses. CMCMCMRR-R-dedederiririveveveddd




3

Background

Aortic regurgitation (AR) remains an important cardiac condition,1, 2 although substantial chronic

regurgitation can be tolerated for many years, with patients remaining asymptomatic. Aortic

valve replacement is usually reserved for when symptoms or significant left ventricular (LV)

dilation or dysfunction occur,3, 4 but prognosis is already reduced by this stage.5-7 Earlier surgery

has been advocated,3, 8 but it is also important to avoid the increased risks associated with

premature surgery. Optimising the timing of surgery in these patients can therefore be difficult.

Quantifying the AR could be valuable for guiding management, especially in asymptomatic

patients with significant regurgitation, and might be used for the early identification of patients

requiring aortic valve surgery.

Cardiovascular magnetic resonance (CMR) is able to directly quantify aortic

regurgitation with high accuracy and reproducibility, using the technique of phase contrast

velocity mapping.9-11

As CMR also provides highly accurate measurements of left ventricular (LV) mass,

volumes and function12-15 (and is considered the ‘Gold-standard’ for measuring these), it would

appear to be an ideal technique for the assessment of aortic regurgitation, but the utility of CMR

to guide clinical management has not been evaluated. We sought to examine whether CMR

quantitation of aortic regurgitation and LV indices could identify which asymptomatic patients

with significant aortic regurgitation were likely to progress to symptoms or other established

indications for surgery in the near future. We also aimed to compare the CMR quantitation of

aortic regurgitation and LV volume/function indices for their relative predictive ability.

Methods

Subjects and follow up

equiring aortic valve surgery.

Cardiovascular magnetic resonance (CMR) is able to directly quantify aortic

eegugugurgrgrgitititatatatioioionn wiiththth hhigh accuracy and reproducibbbilililittyy, using the teechchc niiquququeee of phase contrast

vvelooocic ty mappipingngg.99-19-1111

AsAs CCCMRMRMR aallslsoo prprovovo ididesese hhhigigighlhlhlyy aacaccucucurararateee mmmeeaeassurrrememmenenentsss ooof f f leeeftftft vveenenttrtriciccululu arara (((LVLVLV)) mamamassss, ,

volumes and d fufufuncncn tititionono 121212-151515 (((ananand d isisis cconononsiiidededererered d d thththe e e ‘GGGololo d-d-stststaaandndndararrd’dd fororor mmmeaeaeasususuririr ngngng ttthehehesesese),)) it would




4

Patients at least 18 years of age were recruited from four high-volume CMR centres in Oxford,

London, Leeds (UK) and Auckland (New Zealand). All asymptomatic patients with moderate or

severe chronic aortic regurgitation on echocardiography by standard (semi-quantitative)

assessment16 were eligible for inclusion and had a baseline CMR scan. Exclusion criteria

included the presence of other significant valve disease or clinical and/or angiographic evidence

for coronary disease.

Patients had a baseline CMR scan and were followed for up to 9 years. Those who

remained asymptomatic and under conservative management were designated the ‘conservative’

group, while those that developed symptoms or other established indications for surgery3 were

designated the ‘crossover’ group, with the decision for surgery taken as the point of censoring.

All clinical decisions were taken by the treating physician. In Oxford, patients participated in a

research study, with annual CMR scans, and clinical decisions were made without knowledge of

the CMR data. In the other three centres, study patients were identified from the clinical CMR

databases (though were initially diagnosed with echocardiography) and clinicians had access to

the CMR data. Events were only counted however if the reason for aortic valve surgery was for

established indications (primarily symptoms, excess LV dilation or LV dysfunction). Patients

undergoing aortic valve replacement for indications outside the established criteria3 (which do

not include CMR assessment), or when surgery was primarily for other surgery (e.g. aortic root

replacement) were considered to be in the conservative group but censored at the time of surgery.

In addition, a minimum period of two months was required between the CMR scan and the

decision for surgery, to avoid the potential bias of patients having a CMR scan ‘en-route’ to

surgery that had already been planned.

A third group was also included to compare CMR parameters with both the conservative

designated the ‘crossover’ group, with the decision for surgery taken as the poinnt t t ofoff cenenensososoririringngn ..

All clinical decisions were taken by the treating physician. In Oxford, patients participated in a

eeseseearararchchch ssstututudydyd , wiwiwitthth annual CMR scans, and clinnnicici aaal decisions wwerere e mamaaddede without knowledge of

hhhe CCMR dataa. InInIn ttthee ooththhererr tthrhrhreeee cccenenntrtreses,, stuududyyy paaatiieenttss wwwererre idideeenttitififiededed ffrorommm tththee clclc inininicici alal CCCMRMRMR

dadatatatabababaseses ss (t(t(thohohougugghh wewerrre iininitiialala lylyy dddiaiaiagngnososo ededed wwwititth hh eecechhohoccacardrdrdioioiogrgrrapapa hhyhy))) anana ddd cclclininnicici iaiaiansnss hhhadad aacccccesesss ttoto

he CMR datataa.. EvEvEvenenentstss wwwerrreee onono lylyy cccououountttededed hhhowowoweveve ererer iifff tthehee rreaeaeasososon n n fofof r r r aoaoaortrtrticicic vvvalala veveve sssurururgegegeryr was for ff




5

and crossover groups. This group included patients already due for aortic valve replacement (the

‘surgical’ group), having developed established indications for surgery.3

The research study was approved by the Oxfordshire Central Research Ethics Committee

(Project code C02.020) and the Waitemata District Health Board “Knowledge Centre” in New

Zealand (Project number RM0980711302); all research subjects gave written informed consent.

CMR scanning

All scans were performed on 1.5 Tesla scanners (either Siemens Avanto [Siemens Medical

Solutions, Erlangen, Germany] or Philips Achieva scanners [Philips Healthcare, Best, The

Netherlands]) and analysed in each centre using the manufacturers’ software (Siemens Argus and

Philips ViewForum respectively) for both volumes and flow. All images were electrocardiogram

(ECG)-gated and most were obtained during an 8-16 second breath-hold to remove cardiac

motion due to the respiratory cycle. Subjects underwent a left ventricular function study as

previously described,14 consisting of a stack of contiguous short axis cine images from base to

apex, from which left ventricular end-diastolic and end-systolic volumes (LVEDV and LVESV

respectively) and mass were measured. Each value was also indexed to body surface area. Cine

image sequences were steady-state free precession (Siemens ‘TrueFISP’ or Philips ‘balanced fast

field echo’); temporal resolution 45msec; echo time 1.40-1.54msec; repetition time 2.80-

3.08msec; field of view 380x380mm; flip angle 50-60o).

Forward and regurgitant aortic flow were quantified using through-plane phase-contrast

velocity-mapping. This involves placing an image slice perpendicular to the direction of flow in

the aortic root, and measuring the velocity of flow through the image plane within each voxel.

From the resulting images, a region of interest identifying the aortic root is defined, and flow is

integrated for the whole cardiac cycle to provide forward and regurgitant flow through the aortic

Philips ViewForum respectively) for both volumes and flow. All images were eleleeectrtrtrocccararardididiogogograram

ECG)-gated and most were obtained during an 8-16 second breath-hold to remove cardiac n

momotititiononon dddueueue tttooo thheee rrerespiratory cycle. Subjects unnnddderrrwent a left venenntricucuulalalar r function study as

prevvvioi usly desscrcribibededd,,1411 ccconononsisiiststtining g g oofof aa ssttackkk oof cooonnttiguououousus sshhhorortt axxxisis ccininee imimmaagageses ffroroommm bababasese tooo

apapexexex, frfrf omomm wwwhihichchh lllefeft t veveentntriccuculalalarrr enenend-d-didid asasastototoliccc ananandd enend-d-d sysysysststolollicici vvvooolululummmess (L(LLVEVEVEDVDVV aaanddd LLLVEVEESVVV

espectivelyy) ) ) ananand d mamamasssss wwwererre e e mememeasasasururredede ... EaEaEachchch vvvalala uuue e e wawaw ss alala sososo iindndndexexe ededed ttto o o bobobodydyd sssurururfafafacecece aarea. Cine




6

valve per cardiac cycle. The image plane was placed ~0.5cm above the aortic valve at end-

diastole, but maintaining a position in the aortic root throughout the cardiac cycle (Figure 1).

Imaging closer to the valve reduces the underestimation of regurgitation that can occur,17, 18 and

although increased turbulent flow can occur close to the valve,19 we have not found this to be a

problem in practice. If significant turbulence or aliasing was seen in the velocity image, the

acquisition was repeated a few millimetres further from the valve, and/or with a higher velocity

window. The original flow sequences9-11 acquire data over many cardiac cycles, taking

approximately two minutes with patients breathing freely, while newer magnetic resonance

sequences20 can acquire flow data within a single breath-hold (12-16 heart beats). Others have

suggested that the older free-breathing techniques may be more accurate for flow

quantification,19 as the newer sequences may be more prone to background flow offset errors

from the faster switching of magnetic field gradients (which can potentially cause significant

errors in flow quantification), but this has not been systematically examined. Free-breathing flow

sequences were used in Oxford and Leeds, while breath-hold flow sequences were used in the

other two centres. In all centres, the potential for background flow offset errors was reduced19 by

i) ensuring all flow sequences were acquired with the region of interest in the image slice located

at the isocentre of the magnet to minimise any inhomogeneities in the magnetic field, and ii)

using retrospective ECG-gating for all flow sequences, which also helps to ensure coverage of

the entire cardiac cycle. Image parameters: temporal resolution 25-55msec; echo time 2.6-

3.2msec; repetition time 4.3-7.8msec; field of view 320x320mm; velocity window 2.5-4.0m/sec;

signal averages: 1 for breath-hold sequences, 3 for free-breathing sequences; typical acquisition

time 12-16 seconds for breath-hold sequences, 2-3 minutes for free-breathing sequences. From

these images, forward and regurgitant aortic flow were measured by integrating the flow in each

uggested that the older free-breathing techniques may be more accurate for floww w

quantification,19 as the newer sequences may be more prone to background flow offset errors

frromomm ttthehehe fffasasasteteter swswwiititchc ing of magnetic field graddieieienntts (which can pppoto enntititiaalally cause significant

errrooorsr in flow qquuuannntififficici atatatioioion)n)n), bbubut t thththisi hhaaas nooot beeenn sysssteteemamaatiicacallllly exxamammiinnededd. FFrFreeeee-b-bbrerer aaathihihingngg ffflllow

eequququenenencecess wewewerere uuusseed d iinin OOxfxfoorord dd ananand d d LeLeeededdss,s, wwwhihihilelel bbrrereatatth-h--hohoholdldd fffloloow w sesesequququeenncecees s wewew rerere uuuseeed inin tthehehe

other two centnttrereres.s. IIIn n n alala l l l ceentntn rrreses,,, thththeee popootetetentntntiaiaiall fofoforr bababackckc grgrrouououndndnd flflflowowow oooffffffseseset t t erererroror rsrsrs wwwasasas rrrede uced191919 byyy




7

frame over one cardiac cycle as previously described.9, 11 Regurgitant fraction (regurgitant

volume/forward volume x 100%) was also calculated.

Echocardiography

Clinical echocardiograms were acquired a mean of 22.9 ±81.5 days from the baseline CMR scan,

according to standard protocols.21 Assessment of the grade of AR on echocardiography was

based on multiple semi-quantitative and qualitative two-dimensional imaging parameters, as

suggested in the American Society of Echocardiography guidelines,16 with senior advice sought

in difficult cases. The echocardiograms were not performed specifically for the research study

however, and did not include the quantification of LV volumes or aortic regurgitation as current

guidelines recommend. Because of this limitation, LV end-diastolic and end-systolic diameters

and the semi-quantitative echocardiographic grading were not included in the predictive analysis

for comparison with CMR parameters.

Data assessment and statistical analysis

Receiver operating characteristic (ROC) analysis was used to determine the ability of the various

parameters to discriminate patients who would develop symptoms or other indications for

surgery during follow up, from those that remained asymptomatic. Differences in ROC area were

compared using the method of DeLong et al.22 Cox proportional hazards and multiple logistic

regression analyses were applied to any parameters with reasonable discriminatory ability (area

under the curve (AUC) on ROC analysis >0.70) to determine if any of these were independent

predictors. Cox proportional hazard analysis was performed in a binary fashion, comparing

groups above & below the optimal threshold identified on ROC analysis. Multiple logistic

regression analysis was performed using continuous variables, with subsequent binary analysis

for independent variables, again based on the thresholds identified from ROC analysis. Kaplan-

guidelines recommend. Because of this limitation, LV end-diastolic and end-syssttotoliliicc dididiammameteteterers s

and the semi-quantitative echocardiographic grading were not included in the predictive analysis

foor r cococommpmparararisisisooon wwwititith h CMR parameters.

DDattata assessmmenent aanandd d ststtatatatiisisttiticcacal l anananalalysysis

ReRececec iviviverer oopepeperaratitingngg cchhaararaactcterrisisi tititiccc (R(R(ROCOCC) ) anananaaalysysysisis wwwaas usussededed tooo dededetetermrmrmininne ththe e ababbililititty y y oof tttheee vvaraariooousu

parameters to o dididiscsccriririmimiminananatee pppatata ieeentntnts s whwhwhooo wowowoulululd d d dedeevevevelolol pp sysysympmpmptototomsmm ooor r otototheheher r r inindidid cacacatititiononons s for




8

Meier survival curves are better for time-dependent events, and these were generated for any

independent parameters to illustrate their association with the progression to symptoms/surgery.

For group comparisons of CMR parameters, including the surgical group, one-way analysis of

variance (ANOVA) was used, with Bonferroni post-hoc analysis, after confirming normal

distributions of the variables using the Kolmogorov-Smirnov test. All analyses were performed

with SPSS version 17.0 (SPSS Inc., Chicago, USA) with the exception of the ROC and Cox

regression analyses which were performed with MedCalc version 9.3.1 (MedCalc Software,

Mariakerke, Belgium). Values shown are means standard deviation and a p-value of <0.05 was

considered the threshold for statistical significance.

Results

118 asymptomatic patients were considered for inclusion in the study, who had at least moderate

AR on echocardiography. Five were excluded because aortic valve surgery occurred within two

months of the CMR scan, leaving 113 patients, who were followed for up to 9 years (mean 2.6

±2.1 years). Thirty nine patients (35%) underwent aortic valve replacement during the follow-up

period, having developed symptoms (n=19) or other established echocardiographic indications

for surgery (excessive LV dilation [EDD >7.5cm or ESD >5.5cm], n=17); or reduced LV

function [echocardiographic ejection fraction <50%], n=3). These were designated the

‘crossover’ group. The mean time from CMR scan to the decision on surgery in this group was

21 months (median: 11 months), with 90% of events occurring within three years. Eight patients

underwent surgery primarily for aortic dilation and remained in the ‘conservative’ group but

censored at the time of surgery; mean regurgitant fraction in this group was 19% (range 5-30%).

One patient underwent aortic valve replacement surgery without conventional established

Results

111888 asasasymymptptptomommataticicic ppatients were considered for inininccllusion in the sstututudyy,, wwhwho had at least moderate

ARAR oon echocaardrdiioi grgrapapphyhyy. FiFiFivevee wwererereee eexcxccludeeeddd beccacauuse e e aaoaortrticic vvalallvee suurrrgeeery y oocccucurrrrrededd wwwitthihhin n twtwwo

momontntn hshshs ooofff thththee CMCMMRR R scscscanann,,, lel avavavining gg 11111 33 papapatititienenentsss,, , wwhwhooo wwewererere fffololo lololowwewed dd fofof rr r upupp tto o o 9 9 9 yeyeyeararars s (m(m(meaeaean n 222.66 6

±2.1 years).. TTThihihirtrtr y y y ninin nenene ppatatatieieientntn s s s (3(3( 5%5%5%) ) ) ununundedederwrwrwenenent t aoaoaortr icicc vvvalalalveveve rrrepee lalalacececememementntnt ddurururinining g g thththe e follow-uppp




9

indications3 and was also retained in the conservative group. This patient was asymptomatic and

had surgery for LV ‘dilation’, but the echocardiographic end-diastolic diameter was 6.5cm; his

regurgitant fraction on CMR was 19%, and EDV 222ml.

Association with events

The ability of CMR parameters to identify patients at baseline who would develop indications for

surgery is shown in table 1 (ROC analyses). Quantitative measures of AR showed excellent

discriminatory power, with aortic regurgitant volume having an area under the curve (AUC) of

0.96 (p<0.0001), and regurgitant fraction an AUC of 0.93 (p<0.0001), with no statistical

difference between the two. CMR LV volumetric indices also showed good discriminatory

ability, though slightly lower than regurgitation quantification, with an AUC of 0.88 for LVEDV

and AUC 0.78 for LVESV (both p 0.01 vs. regurgitant volume). On multivariate analyses (table

2), only regurgitant fraction, regurgitant volume and LVEDV remained as independent

predictors. Binary analyses (comparing groups above/below the threshold identified on ROC

analysis) showed higher hazard ratios for regurgitant volume and fraction than LVEDV. The

differences in AUC and hazard ratios were small however, with some overlap of the confidence

limits, the latter likely due to the binary nature of the analyses and moderate sample size. In

general, regurgitant volume and fraction showed very similar discriminatory power; regurgitant

fraction has the modest advantage of being a body size-independent variable.

Regurgitant fraction >33% had high sensitivity (85%) and specificity (92%) for

identifying patients who progressed to symptoms and surgery (Figure 2). A single threshold

value may not provide all information however, and the data showed further useful thresholds:

all patients with a regurgitant fraction >51% (n=8) progressed to surgery (100% positive

predictive value), while all but two patients with a regurgitant fraction >43% (n=20) progressed

ability, though slightly lower than regurgitation quantification, with an AUC of 00.0.88888 ffforor LLLVEVEVEDDV

and AUC 0.78 for LVESV (both p 0.01 vs. regurgitant volume). On multivariate analyses (table

2)), , onononlylyly rregegegururrgitataanntnt fraction, regurgitant volumeee aaanndd LVEDV remmmaia needd d aaas independent

predddici tors. Binanarrry ananaalylyysesesesss (c(c(coomompapaparirir nngg ggrouupups abbbovve/bebebelolowww ththee ththhreeshshshololdd ididdeenentitifififiededed oonnn ROROROCC C

annalalalysysysisisi )) ) shshhowoowededd hhhigighhherr r hahazaaardrdd rrratatatioioios s fofoforrr rerereguguurrrgigitttannnt vvvolollumumumee anana ddd frfrfracaca ttitiononn tthhahan n n LVLVVEEDEDVV.V. TThehee

differences inin AAAUCUCUC aandndnd hazazazararard rararatitiososo wwwererreee smsmsmaaallll hhhowowwevvvererer,, wiwiwiththth ssommmeee ovovoverererlalal pp ofofof ttthehehe cconfidence




10

to surgery (90% positive predictive value). At the other end of the scale, no patients with a

regurgitant fraction <26% (n=45) progressed to surgery (100% negative predictive value).

Survival curves are however better for assessing the effect of time on events (to account for the

fact that some events require adequate follow up to occur). There was significant separation of

the groups over time, with survival without surgery at the median time point (2.0 years) of 95%

for patients with regurgitant fraction 33% compared to 33% for patients with regurgitant

fraction >33% (p<0.0001 by logrank test). The data were also analysed using the highest

regurgitant fraction during follow up (there were 20 patients with serial CMR scans). This may

allow for increasing values over time and may also be closer to clinical practice (waiting until a

threshold is reached). Using the highest regurgitant fraction showed similar discriminatory power

on ROC analysis (AUC 0.93), and similar separation of survival curves at the median time point

of 1.9 years: surgery-free survival was 93% and 34% for regurgitant fractions 33% and >33%

respectively (Figure 3). In patients with highest regurgitant fraction 33%, longer term surgery-

free survival at 8 years showed a slight increase compared to analysis using the first recorded

regurgitant fraction (91% and 83% respectively), indicating the few patients that developed

higher degrees of regurgitation over time and were moved to the crossover group. All patients

with a regurgitant fraction >33% eventually had surgery over 8-9 years of follow up, but subject

numbers were small in the later years. The average time to surgery when using the highest

regurgitant fraction was slightly reduced, as would be expected (mean: 2.4 years, median: 1.9

years).

The association of aortic regurgitant fraction with outcome remained robust in sub-group

analyses. There was no significant difference between Oxford and the other participating centres;

table 3, p=0.59 by logrank test on Kaplan-Meier survival analysis (Figure 3c). Comparing

hreshold is reached). Using the highest regurgitant fraction showed similar disccriririmimim naaatototoryryry pppowowe

on ROC analysis (AUC 0.93), and similar separation of survival curves at the median time point

off 111.9.9.9 yyyeaeaarsrsrs:: ssurgrggeereryy-free survival was 93% and d 34343 %% for regurgittanaa t frfrracacactit ons 33% and >33%

eespppece tively (FiFiiggugurere 333).. IIn nn pappatititienentstst wwwiitthh higghheeest rregeggurgigigittatanntnt fffraraccttiooon n 33333 %,%, llloonongegeer rr tetetermrmm sssururggegerrry-

frreeeee sssurururvivivavaval l l atat 888 yyyeaearrss ssshohoweweed dd aaa slslsligiighthtt iincncncrerereasasse e e ccocompmpmpararededed tto o anana aaalysysysisisi uuussiingng tthehehe ffiririrststst reeecoorordedeedd

egurgitant frfracacactitit ononon (((91919 % %% ananand dd 83833%%% rrrese pepepectctctivivivelelely)y)y),,, ininndidid caaatititingngng ttthehehe fefef w w w papapatititienenentststs thahahat t t dededevevev loped




11

centres using free-breathing CMR flow sequences (Oxford and Leeds) with the other two centres

using breath-hold sequences also showed no difference in the association of regurgitant fraction

with outcome (p=0.84 by logrank test on Kaplan-Meier survival analysis). Restricting the

analyses to only those patients that developed LV dilation or dysfunction as an indication for

surgery (excluding patients for whom symptoms developed) again showed a similar association

with outcome to the whole group. The area under the curve on ROC analysis was 0.91, and

survival without surgery to 2.0 years (the median time point) was 97% for those with regurgitant

fractions 33%, and 39% for those with regurgitant fraction >33% (p<0.0001 by logrank test).

The association of LVEDV with outcome appeared slightly lower than regurgitation

quantification, though the differences were slight and confidence limits overlap. Combining LV

end-diastolic volume with regurgitant fraction provided further improvement on either parameter

alone however (Figure 3d). The combination may thus provide the most robust discrimination,

especially given that both parameters are measured in one CMR examination. LV ejection

fraction was not able to predict events (AUC 0.55; p=0.43). CMR LV mass showed some

predictive power (AUC 0.74; p<0.0001), but this parameter is closely related to LVEDV, and the

similar mass:volume ratios in all groups (table 4) suggests that LVEDV is likely to be a

significant confounding factor.

Comparison with the surgical group

Data from the surgical group are shown in table 4. This showed similar mean aortic regurgitation

and LV volumetric indices to the crossover group and both were significantly larger than in the

conservative group. Ejection fraction was lower in the surgical group (mean 57.1% versus 62.9%

and 63.6% for the crossover and conservative groups respectively; p<0.01 for both comparisons),

perhaps reflecting a more advanced stage of the disease. The higher proportion of bicuspid

quantification, though the differences were slight and confidence limits overlap.. CoCoCombmbbininininining g g LVLV

end-diastolic volume with regurgitant fraction provided further improvement on either parameter

allononne e hohohoweweweveveer ((FiFiFiggugure 3d). The combination mmmayaya tthus provide ththhe e momooststst robust discrimination,

espeeecic ally givvenen thhhat t bobob ththth pppaararaamametettererrs s araree meeeaassureeded in ononone e CMCMCMRR exxamaminii aatatioioon.n. LLV VV ejejejecectit oonon

frracacctitiiononon wwasasas nnotot abblblee ttto pppreredidiictct eeeveveventnnts s (A(AAUCUCUC 000.5.5. 555; ppp==00.4.443)3)3).. CMCMCMR R R LVLVLV mmasasa s s shshshowowowededed somomome e

predictive ppowowwerere (((AUAUAUCCC 0.747474;;; p<<<0.0.0 0000000111))),, bububut t thththisiss ppparararamammetete ererer iiis s s clclclosoo elelely y y rererelalalateteted d tototo LLLVEVEVEDVD , and thhee




12

valves in the crossover group (0.55, compared to 0.29 in the conservative group, p=0.003 by Chi-

squared analysis), might be explained by the slightly higher mean regurgitant fraction (mean

33.4% vs. 25.8% for tricuspid valves; p=0.004).

Discussion

The association of aortic regurgitation quantitation with outcome

Our data demonstrate the potential value of quantifying aortic regurgitation with CMR, which

showed a significant association with the future need for surgery, including patients who

developed asymptomatic LV dilation or dysfunction. Patients already destined for surgery (the

‘surgical’ group) also had measures of aortic regurgitation that were not significantly different

from the ‘crossover’ group, suggesting that a similar threshold of regurgitation had been reached

in the surgical group before symptoms occurred. These CMR parameters might thus be

predictors of future events, but this requires testing in a future prospective study.

Comparison with LV volumetric indices

Quantifying the regurgitation showed a slightly better association with events than CMR-derived

LV indices, despite highly accurate measurements of LV volumes and function by CMR.

LVEDV still had good discriminatory power however, and was an independent predictor on

multivariate analyses. The combination of LVEDV with regurgitant fraction provided a slight

enhancement over aortic regurgitation alone, and LV volumes and function are important in the

overall assessment of the patient. Given these factors, and that both are readily available from a

standard CMR scan, the combination of CMR quantification of aortic regurgitation and LV

volumes could be a valuable component of the work-up in patients with aortic regurgitation. The

slightly stronger association of outcome with aortic regurgitation indices compared to LV

surgical’ group) also had measures of aortic regurgitation that were not significacaa tntlylyl dddifififfefeferererentnt

from the ‘crossover’ group, suggesting that a similar threshold of regurgitation had been reached

nn tthehehe sssururgigigicacacal grgrrouououpp before symptoms occurredd. ThThhese CMR pararaameeteteersrsrs might thus be

predddici tors of fufututuurree eevvvenntnts,s,s bbbututut tthihiiss rerer qququiires teeestinnngg in aaa ffututuuurree pprrorospspecctit vveve ssstututudydy..

CoCoompmpmparara isissononon wwititthh LVLVV vvvololummmetetetririricc c iininddidicececesss

Quantifying g ththhe e e rereegugugurgrgrgitititatioioion n n shhhowowowedede aaa sslililighghghtltltly y y bebebettttt erere aaassssssococociaiai tititiononon wwwititth h h evevevenenentsts ttthahahan n n CMCMC R-derivedddd




13

volumes, and the ability of regurgitant fraction to identify patients who would develop excess

LV dilation or dysfunction as indications for surgery, suggests that increases in regurgitation

may occur before LV dilation. This would be logical given that regurgitation is the physiological

stimulus for LV dilation in this patient group, though is not conclusively proven with our data.

LV mass showed reasonable discriminatory ability in identifying patients likely to

progress to surgery (AUC 0.74). It is however closely related to LV volume, and was not an

independent predictor on multivariate analysis. Other studies have not shown any predictive

power of wall thickness23 and the LV mass to volume ratios were similar for all three subject

groups in our study, suggesting that there is no excess increase in mass over that required for the

chamber volume increase, and that the apparent association of LV mass with outcome is likely to

be confounded by its close link to LV volume.

Comparison with echocardiography and other CMR studies

Echocardiographic techniques for quantifying aortic regurgitation also exist,24, 25 although there

are potential inaccuracies from the calculations and assumptions involved. Quantitative AR

assessment by echocardiography is primarily used to aid the grading of regurgitation severity,

and may have improved grading from the semi-quantitative assessment used to identify patients

in our study. It has also shown an association with the need for surgery,26 though in that paper

43% of patients with moderate AR on echocardiography progressed to surgery, suggesting a

lower ability of quantitative echocardiographic AR grading to identify patients at risk of events.

Previous studies comparing CMR with echocardiographic grading27, 28 suggest only a moderate

correlation, with significant overlap of CMR-quantified AR values across the echocardiographic

grades, particularly between moderate and severe grades. It is of interest that in the Gabriel

paper27 the AR groups defined as ‘truly severe’ had regurgitant volumes & fractions

chamber volume increase, and that the apparent association of LV mass with ouutctccommmee isisis lllikikikeleely y to

be confounded by its close link to LV volume.

CoCompmpmpaararisisisonono wititthh h eechocardiography and othererer CCCMR studies

EEchhohocardiograpaphhih ccc tetechchc nininiquqqueseses ffororr qqquauau ntntiifyingngng aorrrtiicc reeeggugurgrggittatatioioonn alalsoo exixiststt,24,24,24, 25255 aaaltltthohohougugugh hh thththererre

arre e e popopotetet ntntiaiaal l l ininacacccucuuraracccieeses ffrooom m m thththe ee ccacalclcculululatatatiioionsnsns aandndnd asssssumumumppptioioonsnsns invnvnvololvveved.d. QQQuauauantntiititaatativvvee AARAR

assessment bby y y ecece hohohocacac rdrdrdioi grgrgrapapaphyhyhy iiiss prpp imimmarararililly y y usususedede totoo aaididd tthehehe gggrararadidd ngngng ooof f f rereregugugurgrggitititatatatioioon n n ses verity,




14

predominantly above the thresholds we identified in our study. Other CMR studies of AR

quantitation11, 28, 29 also showed significant overlap between quantitative values and AR grades

by other techniques, but none have examined the potential value of AR quantitation in clinical

management.

Comparison with previous studies of outcome in aortic regurgitation

Bonow et al’s 1991 study23 had similar methodology to our own, in a similar sized group of

patients, and examined the prediction of clinical events in initially asymptomatic patients who

underwent echocardiography. Both this and the 1995 Tornos study4 showed a predictive ability

for end-diastolic diameter (>7.0cm) and end-systolic diameter (>4.0cm but especially >5.0cm).

These findings are in keeping with our data which showed an association with outcome for end-

diastolic and end-systolic volumes, though the addition of CMR measurements of LV volume

and quantification of AR adds to these existing studies with more modern imaging techniques.

Other studies have examined the prediction of outcome post-surgery in patients with aortic

regurgitation,5, 6, 30 or in a mixed asymptomatic and pre-surgical group of patents,31 and

confirmed that symptoms, reduced ejection fraction and excess LV dilation are associated with

worse long term outcome. These studies helped inform the current guidelines for surgery in

aortic regurgitation,3 but also highlight the value in identifying patients prior to symptoms or

significant LV dilation or dysfunction, as this study aims to do.

Clinical utility

The ability to identify patients prior to symptoms or excess LV dilation/dysfunction would be

clinically important. These patients might be considered for early surgery, and at the very least

could be followed more closely. Our sample size was modest however, and to support a change

in clinical practice, particularly where cardiac surgery is concerned, requires better

These findings are in keeping with our data which showed an association with oouuutcocoomememe fffororor eeendnd-

diastolic and end-systolic volumes, though the addition of CMR measurements of LV volume

annd d d quququaanantititififificaationonon oof AR adds to these existing stststududdies with moree mmodddererernn imaging techniques.

OOOthheher studies hahaveveve eexxaamimiminenened d d thththe e prprpredede iicttiion offf outtctcomome e e popossts --ssururgegeeryy iiinnn ppapatitiienenentsts wwwititth h h aoaortrttiicic

eegugugurgrgrgititi atatioioon,nn,5, 5, 6,6, 300 oor r inin aa mmixixixededed aaasysysympmpptototomamamatiiicc c aandndnd pprrere-s-ssurururgigigicacacal l grgrgrouououpp p offf ppatatatenenntsts,,,3331 andndd

confirmed thhatatat sssymymymptptp omomoms, rrredededucuccededed eeejejej ctctctioioon n n frfrracacactitiiononn aaandndd eeexcxcxcesesess s s LVLVLV dddilililatatatioioion n n ararre ee asasassososocicic ated with




15

demonstration of patient benefit in a randomised trial comparing early surgery with surgery

based on conventional indications. Quantitative CMR indices may provide the appropriate tool

for identifying suitable patients for such a trial. Conversely, patients with lower quantities of

aortic regurgitation and LV end-diastolic volume might be reassured of the good medium term

prognosis, and may require less frequent follow-up, aiding the efficient use of healthcare

resources.

Limitations

The moderate sample size and relatively small number of events limit the strength of our

conclusions, though follow up was for a reasonable period of time (mean 2.6 years, and up to 9

years).

The lack of blinding to the CMR data in three of the centres may also have biased results.

There are however no current CMR criteria/thresholds for recommending surgery, and we

attempted to minimise any bias where possible, and confirmed that there were no significant

differences in the association with the progression to surgery between centres. It is possible

however that some bias remains, particularly given the subjective nature of symptom assessment.

The CMR sequence for flow measurement also differed between centres, as did the analysis

software, but the associations with outcome were no different between sub-groups, which

suggests the results may be generalizable for both types of sequence and different vendor

software.

There remain a limited number of contra-indications to MRI, including most pacemakers

and other implanted metallic devices, and a few patients are unsuitable for CMR. Prosthetic heart

valves are not a contra-indication, however.

years).

The lack of blinding to the CMR data in three of the centres may also have biased results

Thhererere e e aararee hohohowwweveveverrr nno current CMR criteria/threeeshshshooolds for recommmemm ndddininingg surgery, and we

attteempm ted to mminininimmmisiseee anananyy bibibiaasas wwheheherere ppoossiiblblee, andndnd conononfifirmrmmeeded tthahat t ththheeeree weweereree nno o sisiigngng ifii iciccanant tt

didifffffferererenenencecesss ininin tthehee aaassssooociiaatitionnn wwwititithhh thththe e prprprogogogrreressssssioioi nnn ttoto sssururrgegegerryry bebeb twtwtweeeeeen n cecentntrerees.s. IIt t iisis pppossssiiblblb e e

however thatt sssomomomee e bibib asasa remmmaiaiainsss, , papapartrtr icccululu arararlylyly gggiiiveveen n n thththe sususubjbjbjecece tititivevev nnnatatatururureee ofofof ssymymymptptptomomom assessmenttt




16

Conclusions

Quantification of aortic regurgitation with CMR showed significant associations with outcome,

particularly when combined with CMR-derived LV volume. The study was of moderate size

however and not all clinicians were blinded to the CMR results. These CMR parameters might

prove useful for identifying suitable patients for early aortic valve replacement, but a clinical

trial is recommended to confirm this and determine clinical benefit.

Acknowledgements: The authors would like to thank Jacqueline Birks for expert statistical

advice.

Funding Sources: The research study was funded by a project grant from the Garfield-Weston

Trust, London, UK (PMS/MMS – 02/03-620). The work was also supported by the Oxford

National Institute for Health Research (NIHR) Biomedical Research Centre programme, and the

NIHR Cardiovascular Biomedical Research Unit of the Royal Brompton and Harefield NHS

Foundation Trust and Imperial College London. SN acknowledges support from the Oxford

British Heart Foundation Centre of Research Excellence.

Conflict of Interest Disclosures: None.

References:

1. Iung B, Baron G, Butchart EG, Delahaye F, Gohlke-Barwolf C, Levang OW, Tornos P, Vanoverschelde JL, Vermeer F, Boersma E, Ravaud P, Vahanian A. A prospective survey of patients with valvular heart disease in europe: The euro heart survey on valvular heart disease. Eur Heart J. 2003;24:1231-1243. 2. Singh JP, Evans JC, Levy D, Larson MG, Freed LA, Fuller DL, Lehman B, Benjamin EJ. Prevalence and clinical determinants of mitral, tricuspid, and aortic regurgitation (the framingham heart study). Am J Cardiol. 1999;83:897-902. 3. Bonow RO, Carabello BA, Kanu C, de Leon AC, Jr., Faxon DP, Freed MD, Gaasch WH, Lytle BW, Nishimura RA, O'Gara PT, O'Rourke RA, Otto CM, Shah PM, Shanewise JS, Smith SC, Jr., Jacobs AK, Adams CD, Anderson JL, Antman EM, Faxon DP, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Lytle BW, Nishimura R, Page RL, Riegel B. Acc/aha 2006 guidelines for

Funding Sources: The research study was funded by a project grant from the GGaaarfifiieleld-d-d-WeWeWestststonon

Trust, London, UK (PMS/MMS – 02/03-620). The work was also supported by the Oxford

National Institute for Health Research (NIHR) Biomedical Research CeCentre programme, and the

NINIIHRHRHR CCarrdididiooovasascuculalar BiB omomeddicical RResseaarcr h UnU itit offf the RoRoyay l BrBrromoo ptptonon andd HHara efieeldld NNHS

FFouunundation Trurussst aaandndd Impmpmpeereriaiaialll CoCoolllllegege ee LLLonnddooon. SNSNSN ackckcknonowlwlwlededgggess susuppppoorort t frfrfromom ttthehehe OOOxfxffoorord dd

BrBrrititisisi h h Hearrtt t FoFoF ununundadattiionn CCeenentrtrtre e ofoff RRReeseeaearcrchhh EExExceellllenenncecee..

Conflict of InInIntetetererereststst DDDisissclclosososuuureses::: NoNoN nenene.




17

the management of patients with valvular heart disease: A report of the american college of cardiology/american heart association task force on practice guidelines (writing committee to revise the 1998 guidelines for the management of patients with valvular heart disease): Developed in collaboration with the society of cardiovascular anesthesiologists: Endorsed by the society for cardiovascular angiography and interventions and the society of thoracic surgeons. Circulation. 2006;114:e84-231. 4. Tornos MP, Olona M, Permanyer-Miralda G, Herrejon MP, Camprecios M, Evangelista A, Garcia del Castillo H, Candell J, Soler-Soler J. Clinical outcome of severe asymptomatic chronic aortic regurgitation: A long-term prospective follow-up study. Am Heart J. 1995;130:333-339. 5. Hwang MH, Hammermeister KE, Oprian C, Henderson W, Bousvaros G, Wong M, Miller DC, Folland E, Sethi G. Preoperative identification of patients likely to have left ventricular dysfunction after aortic valve replacement. Participants in the veterans administration cooperative study on valvular heart disease. Circulation. 1989;80:I65-76. 6. Turina J, Milincic J, Seifert B, Turina M. Valve replacement in chronic aortic regurgitation. True predictors of survival after extended follow-up. Circulation. 1998;98:II100-106; discussion II106-107. 7. Chaliki HP, Mohty D, Avierinos JF, Scott CG, Schaff HV, Tajik AJ, Enriquez-Sarano M. Outcomes after aortic valve replacement in patients with severe aortic regurgitation and markedly reduced left ventricular function. Circulation. 2002;106:2687-2693. 8. Enriquez-Sarano M, Tajik AJ. Clinical practice. Aortic regurgitation. N Engl J Med. 2004;351:1539-1546. 9. Dulce MC, Mostbeck GH, O'Sullivan M, Cheitlin M, Caputo GR, Higgins CB. Severity of aortic regurgitation: Interstudy reproducibility of measurements with velocity-encoded cine mr imaging. Radiology. 1992;185:235-240. 10. Honda N, Machida K, Hashimoto M, Mamiya T, Takahashi T, Kamano T, Kashimada A, Inoue Y, Tanaka S, Yoshimoto N. Aortic regurgitation: Quantitation with mr imaging velocity mapping. Radiology. 1993;186:189-194. 11. Sondergaard L, Lindvig K, Hildebrandt P, Thomsen C, Stahlberg F, Joen T, Henriksen O. Quantification of aortic regurgitation by magnetic resonance velocity mapping. Am Heart J. 1993;125:1081-1090. 12. Bellenger NG, Burgess MI, Ray SG, Lahiri A, Coats AJ, Cleland JG, Pennell DJ. Comparison of left ventricular ejection fraction and volumes in heart failure by echocardiography, radionuclide ventriculography and cardiovascular magnetic resonance; are they interchangeable? Eur Heart J. 2000;21:1387-1396. 13. Hudsmith LE, Petersen SE, Francis JM, Robson MD, Neubauer S. Normal human left and right ventricular and left atrial dimensions using steady state free precession magnetic resonance

True predictors of survival after extended follow-up. Circulation. 1998;98:II100-101006;6;6 ddisiscucussssioi nI106-107.

7. Chaliki HP, Mohty D, Avierinos JF, Scott CG, Schaff HV, Tajik AJ, Enriquez-Sarano M.Outcomes after aortic valve replacement in patients with severe aortic regug rgitation and mamarkrkrkedededllyly rrredededuucededed llefe t ventricular function. Circccuululaattion. 2002;10666:2: 6887-7-7-22693.

88.. EEEnrn iquez-Saararaanooo MMM, TaTaTajijijik k k AJAAJ.. ClClClininii acaal prrracccticeee. AoAorrrtitiic c rreeggugurgrggittatatioionnn. NN N EnEnEngglgl JJJ MMMededed... 2020004044;351:153539-1151546.

9. DDululce MMC,C MMosstbtbece k GHG , O'O'SuSulll ivann MM, ChCheieitlin MM, CaCapuputot GGR,R, HHiggiginsns CCB.B. Seveveritity y ofo aortic regurgigig tatatatitit ononon::: InInnteteersstututudydyd rrreeeprprp ododo ucucucibibbilillititi yy y ofofof mmmeaeae suuurererememementntntsss wwititith h h vevevelololocicicityty-e-eencncncodododedee cine mr mmagaginingg RaRadidiolologogyy 19199292;1;18585:2:23535 2-24040




18

imaging. J Cardiovasc Magn Reson. 2005;7:775-782. 14. Lorenz CH, Walker ES, Morgan VL, Klein SS, Graham TP, Jr. Normal human right and left ventricular mass, systolic function, and gender differences by cine magnetic resonance imaging. J Cardiovasc Magn Reson. 1999;1:7-21. 15. Myerson SG, Bellenger NG, Pennell DJ. Assessment of left ventricular mass by cardiovascular magnetic resonance. Hypertension. 2002;39:750-755. 16. Zoghbi WA, Enriquez-Sarano M, Foster E, Grayburn PA, Kraft CD, Levine RA, Nihoyannopoulos P, Otto CM, Quinones MA, Rakowski H, Stewart WJ, Waggoner A, Weissman NJ. Recommendations for evaluation of the severity of native valvular regurgitation with two-dimensional and doppler echocardiography. J Am Soc Echocardiogr. 2003;16:777-802. 17. Chatzimavroudis GP, Walker PG, Oshinski JN, Franch RH, Pettigrew RI, Yoganathan AP. Slice location dependence of aortic regurgitation measurements with mr phase velocity mapping. Magn Reson Med. 1997;37:545-551. 18. Myerson SG. Heart valve disease: Investigation by cardiovascular magnetic resonance. JCardiovasc Magn Reson. 2012;14:7. 19. Kilner PJ, Gatehouse PD, Firmin DN. Flow measurement by magnetic resonance: A unique asset worth optimising. J Cardiovasc Magn Reson. 2007;9:723-728. 20. Sakuma H, Kawada N, Kubo H, Nishide Y, Takano K, Kato N, Takeda K. Effect of breath holding on blood flow measurement using fast velocity encoded cine mri. Magn Reson Med. 2001;45:346-348. 21. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MS, Stewart WJ. Recommendations for chamber quantification: A report from the american society of echocardiography's guidelines and standards committee and the chamber quantification writing group, developed in conjunction with the european association of echocardiography, a branch of the european society of cardiology. J Am Soc Echocardiogr. 2005;18:1440-1463. 22. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics. 1988;44:837-845. 23. Bonow RO, Lakatos E, Maron BJ, Epstein SE. Serial long-term assessment of the natural history of asymptomatic patients with chronic aortic regurgitation and normal left ventricular systolic function. Circulation. 1991;84:1625-1635. 24. Enriquez-Sarano M, Seward JB, Bailey KR, Tajik AJ. Effective regurgitant orifice area: A noninvasive doppler development of an old hemodynamic concept. J Am Coll Cardiol. 1994;23:443-451.

18. Myerson SG. Heart valve disease: Investigation by cardiovascular magnetic rrresesonono anana cecece. JJJCardiovasc Magn Reson. 2012;14:7.

19. Kilner PJ,J, GGatehouse PD, Firmin DN. Flow measurement by magnetic resonance: A uniqueassseseettt wwowortrtrth h h opoptimimimisising. J Cardiovasc Magn Resosoonn. 2007;9:723-772822 .

2200. SSakuma HH, KaKaKawawadadad NNN, KuKuKubbobo HHH,, NNisshhideee YYY, TTakkkannoo KK,, KKaKatoto NNN, TaTaT kekeedada KKK.. EfEfffefefectctc ooof brbrbreaeaathhh hohoholdddini g on bblolooddd ffflowww mmmeasasururrememennt t t uussininng ffafasst veellooccitity y y enenencccodddeddd cciine mmmrrii.. MaMaMagngn RRResssononn MMMedd. 200001010 ;4;4;45:5:343446-6-6-3434888.

21. Lang RM,M,M, BBBieieeriririg gg M,M,M DDDeveveverere eueueuxx x RBRBRB, , FlFlFlacacchshshskakakampmpmpfff FFA,A,A FFFososo teteter r r E,E,, PPPelele lililikkkkkka aa PAPAA,, PiPiPicacacardr MH, ffRoRomamann MJMJ SeSewawardrd JJ ShShananewewisisee JSJS SoSololommonon SSDD SSpepencncerer KKTT SSututtotonn MSMS StStewewarartt WJWJ




19

25. Tribouilloy CM, Enriquez-Sarano M, Fett SL, Bailey KR, Seward JB, Tajik AJ. Application of the proximal flow convergence method to calculate the effective regurgitant orifice area in aortic regurgitation. J Am Coll Cardiol. 1998;32:1032-1039. 26. Detaint D, Messika-Zeitoun D, Maalouf J, Tribouilloy C, Mahoney DW, Tajik AJ, Enriquez-Sarano M. Quantitative echocardiographic determinants of clinical outcome in asymptomatic patients with aortic regurgitation: A prospective study. JACC Cardiovasc Imaging. 2008;1:1-11. 27. Gabriel RS, Renapurkar R, Bolen MA, Verhaert D, Leiber M, Flamm SD, Griffin BP, Desai MY. Comparison of severity of aortic regurgitation by cardiovascular magnetic resonance versus transthoracic echocardiography. Am J Cardiol. 2011;108:1014-1020. 28. Kutty S, Whitehead KK, Natarajan S, Harris MA, Wernovsky G, Fogel MA. Qualitative echocardiographic assessment of aortic valve regurgitation with quantitative cardiac magnetic resonance: A comparative study. Pediatr Cardiol. 2009;30:971-977. 29. Ambrosi P, Faugere G, Desfossez L, Habib G, Bory M, Luccioni R, Bernard P. Assessment of aortic regurgitation severity by magnetic resonance imaging of the thoracic aorta. Eur Heart J. 1995;16:406-409. 30. Klodas E, Enriquez-Sarano M, Tajik AJ, Mullany CJ, Bailey KR, Seward JB. Optimizing timing of surgical correction in patients with severe aortic regurgitation: Role of symptoms. J Am Coll Cardiol. 1997;30:746-752. 31. Dujardin KS, Enriquez-Sarano M, Schaff HV, Bailey KR, Seward JB, Tajik AJ. Mortality and morbidity of aortic regurgitation in clinical practice. A long-term follow-up study. Circulation. 1999;99:1851-1857.

of aortic regurgitation severity by magnetic resonance imaging of the thoracic aortrta.a.a. EuEur r HeHearart J1995;16:406-409.

30. Klodas E, Enriquez-Sarano M, Tajik AJ, Mullany CJ, Bailey KR, Seward JB. Optimizing imingg of surggical correction in patients with severe aortic regurgitation: Role of symptoms. J Am

CoCollllll CCCaarardididiololol.. 199979797;3;30:746-752.

331. DDujardin KKS,S,S EEEnrnriqiqi uueuez-z-SaSaSarraranonoo MMM, SScchaffff HHVV,, BBBaillleyeyey KKKRR,R, SSewewwararddd JBBB, TaTaTajjijik k AJAJAJ.. MMMortrtrtalalititi yy anannd d mom rbiditity yy offf aaaortiicc rreguguurgrggiti atatioon n n iinn cccliniiicaaal prraractcticicce.e.e. AAA looong--teeerm fffolllloow-w-w-upup sttutuddydy.. CiCiircrcrculululatata ioionnn.. 119199999;9;999:9:18185151-11858585777.




20

Table 1. Receiver operating characteristic (ROC) data. Comparison of the ability of each CMR parameter to identify the initially asymptomatic patients who would develop indications for surgery, using receiver operating characteristic (ROC) analysis.

AUC Threshold p Sens (%) Spec (%)

Regurgitant fraction (%) 0.93 (0.87-0.97) > 33 <0.0001 85 92

Regurgitant volume (ml) 0.96 (0.90-0.99) > 42 <0.0001 92 85

Regurgitant volume index (ml/m2) 0.95 (0.89-0.98) > 23 <0.0001 82 92

LVEDV (ml) 0.88 (0.80-0.93) > 246 <0.0001 87 77

LVEDV index (ml/m2) 0.86 (0.79-0.92) > 129 <0.0001 85 82

LVESV (ml) 0.78 (0.70-0.86) > 88 <0.0001 77 70

LVESV index (ml/m2) 0.77 (0.68-0.84) > 45 <0.0001 74 72

Ejection fraction (%) 0.55 (0.45-0.65) < 59 0.43 38 77

LV mass (g) 0.74 (0.64-0.81) > 187 <0.0001 72 68

LV mass index (g/m2) 0.73 (0.63-0.81) > 90 <0.0001 74 64 AUC = area under the curve; LVEDV = left ventricular end-diastolic volume; LVESV = left ventricular end-systolic volume; p = p value for ROC curve; Sens = sensitivity; Spec = specificity; threshold = value for each parameter which best identified the ‘crossover’ group.

Ejection fraction (%) 0.55 (0.45-0.65) < 59 0.43 383838 777777

LV mass (g) 0.74 (0.64-0.81) > 187 <0.0001 72722 6688

LV mass index (gg/m2) 0.73 (0.63-0.81) > 90 <0.0001 74 64 AUAUCCC === aareaa uuundndnder tthehehe ccurve; LVEDV = left ventricular endndnd-dddiastolic volume;; LLLVEESVSVV == left ventricular end-systolicvovoolulummme; p = p vavavaluee fofoorr ROROCCC cucuurvrve;e; SSenene ss == sesensnsititivivitity;y;y; SSpeeec == spepecicifificicityt ; ththhrereshsholold d == vav luue e fofoorr eaeachch ppararamama etetterer wwhiccchh best identififieieied d ttthe e ‘c‘c‘ rooossssssovovovererer’ grgroououpp.p.




21

Table 2. Cox proportional hazard regression and multiple logistic regression analyses for the variables with significant discriminatory ability on ROC analysis (AUC>0.70).

Cox proportional hazard regression (binary analysis)

Multiple logistic regression

Univariate analysis Multivariate analysis Continuous variables Binary analysis Hazard ratio

(B-exponent) p-value Hazard ratio

(B-exponent) p-value Odds ratio per unit

increase Odds ratio

Regurgitant fraction (%) 16.0 (6.7-38.3) <0.0001 7.4 (3.0-18.6) <0.0001 1.23 (1.11-1.35) 28.9 (8.0-104.1)

Regurgitant volume (ml) 13.2 (3.8-45.8) <0.0001 13.2 (3.8-45.8) 0.0001 1.14 (1.08-1.20) 31.4 (7.6-129.4)

LVEDV (ml) 16.3 (5.8-45.9) <0.0001 6.1 (2.0-19.1) 0.002 1.02 (1.00-1.03) 6.8 (1.8-25.8) LVESV (ml) 7.0 (3.2-15.3) <0.0001 - - -

LV mass (g) 3.2 (1.6-6.5) 0.0007 - - - Analysis was binary for the Cox regression (comparing groups above/below the optimal threshold identified from ROC analysis); and continuous for multiple logistic regression (per unit increase), with subsequent binary analysis. Only absolute rather than indexed values were used, as these showed marginally better discriminatory power on ROC analysis and including both would result in significant confounding of the closely related variables. On multivariate analysis, values are shown for the variables with significant independent predictive value. 95% confidence limits are shown in brackets for all results.

urgitant fraction (%) 16.0 (6.7-38.3) <0.0001 7.4 (3.0-18.6) <0.0001 1.23 (1.11-1.35)) 2228.8.8 9 99 (((8.8.8.0-0-0 101 4

urgitant volume (ml) 13.2 (3.8-45.8) <0.0001 13.2 (3.8-45.8) 0.0001 1.14 (1.08-1.20)) 331.1.1 4 4 4 (((7.7.6-6-6 121212999

EDV (ml) 16.3 (5.8-45.9) <0.0001 6.1 (2.0-19.1) 0.002 1.02 (1.00-1.03) 6.8 (1.8-25.8ESV ((ml) 7.0 (3.2-15.3) <0.0001 - - -

mmasasssss (((g)g)g) 3.3 2 (1.6-6.5) 0.0007 -- - - ssisiss wasasa bbininary fofor r thththe Coox x reregrgrese siionon ((cocompm arrining g grgrououpsps aaboboveve/b/bele owow tthehe optpp mmi al thrhreseshoholdld identntififfieiei d d d frf omm RROCOC anaalyysisis)s); ; aandnd cconntitinunuouous for r mum ltipcc r ggegreression (per uuninit t ininncrcrc aaea ese),),, wwiti hh h sus sbseqeqe ueueuentntnt bbbinini rarary yy anananalal sysy isis. OnOnO lyly aabsb llolute e rathherere ttthahahan nn ninindedexexexed d d vaaalululuesese wweree usedd, , , aaass thththesese ee shshs owwowededed mmmarara gigig nan lllllly y bet

mim anatotory power on ROROC C ana alaa yssisis aa ddnd iincncluludidd ngngn bbotothh h wowow ulu d dd eresult in nn isigng if cicanant t co fnfnfouououndnd nining ofoo tthehe c ooloselyl rrelee tateded vvaraa aiaiablblees.s OOOn mumum ltltivi rarariaiatee aaa analylysisaaare ssshown for the variablblb es ss wi hth s ggignificacantnn iindn eppendedentn pppredid ctiveee aavalue. 9599 % %% confnfididdenenencecece l mmimits araa e shs own n in bbrackcc tets s for a llll r sesulu ssts. ff




22

Table 3. Proportion of asymptomatic patients developing indications for surgery over time, according to CMR regurgitant fraction, and stratified by CMR centre.

Proportion developing indications for surgery

CMRCentre n

Highest CMR regurgitant fraction 33%

Highest CMR regurgitant fraction >33%

Kaplan-Meierp-value for difference

by logrank test Oxford 39 0.04 (n=24) 0.80 (n=15) Other centers 74 0.08 (n=48) 0.81 (n=26)

0.59

Table 4. CMR parameters by group. Comparison of CMR parameters between the three groups of patients with aortic regurgitation.

Conservative Crossover Surgical p-value Number in group 74 39 45 Age (years) 50.8 16.8 45.7 18.7 55.6 16.5* 0.04 Height (cm) 173.4 10.0 176.6 9.0 173.8 8.1 0.24 Proportion of male subjects 0.69 0.92† 0.91† 0.002 Weight (kg) 79.1 16.4 83.0 13.0 83.2 18.4 0.30 Body surface area (m2) 1.94 0.22 2.00 0.18 1.98 0.25 0.31 Bicuspid valve frequency 0.29 0.55† 0.24 0.006 Systolic BP (mmHg) 132.9 19.3 134.2 16.0 135.1 21.4 0.82 Diastolic BP (mmHg) 68.8 15.0 59.4 13.5 63.5 16.3 0.17 Regurgitant volume (ml) 27.5 15.5 74.7 28.5†† 80.5 38.7†† <0.0001 Regurgitant volume index (ml/m2) 14.1 7.6 37.4 14.5 †† 41.4 21.0†† <0.0001 Regurgitant fraction (%) 21.8 9.8 42.0 9.5 †† 45.6 11.0†† <0.0001 LVEDV (ml) 209.5 54.7 301.1 61.6 †† 315.8 91.1†† <0.0001 LVEDV index (ml/m2) 108.0 25.3 151.6 31.6 †† 160.6 48.6†† <0.0001 LVESV (ml) 78.7 35.5 113.0 33.0 †† 138.4 59.4††* <0.0001 LVESV index (ml/m2) 40.5 17.8 57.1 17.3† 70.3 31.8††* <0.0001 LV Ejection fraction (%) 63.6 8.7 62.9 6.4 57.1 10.9†* <0.0001 Echo LVEDD (cm) 5.9 0.6 6.6 0.7 †† 6.6 0.8†† <0.0001 Echo LVESD (cm) 3.7 0.6 4.1 0.6† 4.4 0.8†† <0.0001 LV mass (g) 173.2 73.1 232.2 80.1†† 277.4 77.8††* <0.0001 LV mass index (g/m2) 88.5 34.8 117.1 40.0† 143.7 41.1††* <0.0001 LV mass/LVEDV ratio (g/ml) 0.83 0.26 0.78 0.25 0.91 0.24 0.69

Abbreviations same as for table 1. Values are means ±standard deviation; p-values shown for one-way ANOVA with the exception of bicuspid valve frequency and proportion of male subjects which were by chi-squared analysis; bold figures indicate p<0.05. Individual group comparisons using Bonferroni correction are also shown: †p<0.01, †p<0.001 and ††p<0.0001 vs. the conservative group; *p<0.05 surgical vs. crossover groups.

Height (cm) 173.4 10.0 176.6 9.0 173.8 8.11 00.0.242424 Proportion of male subjects 0.69 0.92† 0.91† 000.000020202 Weight (kg) 79.1 16.4 83.0 13.0 83.2 18.4 0.30 Bodydy suru facece areea a ((mm2) 1.94 0.22 2.00 0.18 11.9.98 0.25 0.31 BiBiBicuuuspspid vvalalalvvve ffrereququencyy 0.29 0.55† 0.24 0.006 SySyststolic BPP (m(mmHmHmHg)g) 13131 2.22 99 9 1911 ..33 133434.2.2.2 16161 .0.00 113355.1 22121.4.4.4 00.0.828282 DDiD aasastolic BPP ((mmmmHHHg) 6868.8.8 155.000 59.9.9.444 1333.5 633.3.55 166.3 0.1717 ReReegugugurgrgrgitititananant t t vovolululumemee (((mmml) ) 2727.5.5. 11515.555 77744.4.7 7 28282 .5.5.5†††† 8880.55 5 383838 7.7.7††††† <<0<0.0.00000011ReReRegugugurgrgrgitititanananttt vovovolululumememe iiindndndexexex (((mlmlml/m/m/m22))) 141414 11.1 777.666 373737 44.4 141414 55.5 †† 444111.444 212121 00.0†† <0<0<0 00.0000000111ReRegugurgrgititanant t frfrf acactitit onon (((%)%)%) 2121.88.8 9.9.88 44242.0.0 9.9.9 55 ††††† 4445.5.6 6 11111.0.0††††† <00.0.0000011




23

Figure Legends:

Figure 1. CMR flow measurement in aortic regurgitation. Top: still frame from steady-state free

precession cine showing left ventricular outflow tract view in diastole with the aortic

regurgitation jet (arrowed) and the slice location for through-plane flow measurement (parallel

lines). Middle: example through-plane flow images; left – anatomical (magnitude) image in

systole, middle – flow (phase) image in systole showing forward flow in white, right – flow

(phase) image in diastole showing regurgitant flow in black. Bottom: resulting flow-time curve

showing volume of regurgitation. LV = left ventricle, LA = left atrium, Ao = aorta.

Figure 2. Discriminatory ability of aortic regurgitant fraction. Top: receiver operating

characteristic (ROC) curve for the ability of aortic regurgitant fraction to identify asymptomatic

patients who would develop symptoms or other indications for surgery. Bottom: dot plot

showing regurgitant fraction in the conservative and crossover groups, with the optimal threshold

of 33% shown.

Figure 3. Surgery-free survival by aortic regurgitant fraction. Kaplan-Meier graphs for survival

without surgery in 113 asymptomatic subjects with at least moderate aortic regurgitation initially

treated conservatively and followed for up to 9 years. a) Stratified by the highest aortic

regurgitant fraction measured by CMR during follow up 33% (n=74) and >33% (n=39), with

the time of the scan with highest regurgitant fraction used as the baseline. b) Similar graph

stratified by CMR-derived LV end-diastolic volume 246ml (n=60) and >246ml (n=53). c) Same

graph as in a), stratified by both highest CMR regurgitant fraction and CMR centre. d) Similar

graph stratified by both highest CMR regurgitant fraction and LV end-diastolic volume.

Figure 2. Discriminatory ability of aortic regurgitant fractif on. Top: receiver operating

chararacacacteteteriririststs icicc (ROROOC)CC curve for the ability of aorrtitiiccc rerr gurgitant fractionn tttooo identify asymptomatic rrr

ppaatiienents whoo wwoououlddl dddeveve elelelopopop sssymymymptptptomomomss ooror ooothththeer r innnddiicaatititiononons s s fofofor r susus rgrgrgerereryy.y. BBBotttototom:m:m: dddototo ppplololot t t

hhhowowowining g regugugurgrgr itttannnt t frfrraccctionnn inn n ththe e cocconnsnserervavaatiiveve aanandd ccrcroosossosooveeer ggrgroooupspp ,, wwiwiththh theheh oopptptimimmaaal tthrhreeshhhold

ofofof 3333%3%3% ssshohohownwnwn.









Clinical Outcome in Aortic Regurgitation with ... - .pdfAortic regurgitation (AR) remains an important cardiac condition,1, 2 although substantial chronic regurgitation can be tolerated

Documents