Embed Size (px)
Citation preview
●
iefmD(va(�acasiL©
Im
SiTtaetcrwfmaap
tdmppIaodt
A
Ischemia-Modified Albumin Predicts Mortality in ESRD
Rajan Sharma, MD, David C. Gaze, PhD, Denis Pellerin, MD, Rajnikant L. Mehta, MSc,Helen Gregson, BSc, Christopher P. Streather, FRCP, Paul O. Collinson, FRCPath,
and Stephen J.D. Brecker, MD
Background: The primary study aim is to determine whether ischemia-modified albumin (IMA) levels predict mortalityn patients with end-stage renal disease (ESRD). The secondary aim is to determine characteristics of patients withlevated IMA levels. Methods: A prospective observational study of 114 renal transplantation candidates was per-ormed. All underwent coronary angiography and dobutamine stress echocardiography. The primary end point is totalortality. Results: During a follow-up period of 2.25 � 0.71 years, there were 18 deaths; 10 were cardiac related.iabetes, severe coronary artery disease, positive dobutamine stress echocardiography result, cardiac troponin T
cTnT) level, IMA level, left ventricular (LV) end-systolic diameter, LV ejection fraction, left atrial size, and mitral peakelocity of early filling (E)/early diastolic velocity (Ea) ratio all predicted mortality. The receiver operating characteristicrea under the curve for mortality prediction was similar for IMA and cTnT levels. An IMA level of 95 KU/L or greatern � 46) predicted mortality with a sensitivity of 76% and specificity of 74%. cTnT level of 0.06 ng/mL or greater (>0.06g/L; n � 51) predicted mortality with a sensitivity of 75% and specificity of 72%. Thirty-eight patients (33%) had both IMAnd cTnT levels elevated. With multivariate analysis, a positive dobutamine stress echocardiography result (P � 0.003),ombined elevated IMA and cTnT levels (P � 0.005), and E/Ea ratio (P � 0.009) were independent prognostic factors. IMAnd cTnT levels alone were not independent predictors of mortality. Patients with an elevated IMA level had aignificantly larger LV size, decreased LV systolic function, and greater E/Ea ratio compared with those without anncreased level. Conclusion: IMA level predicts mortality in patients with ESRD. Patients with elevated levels have largerV size, decreased systolic function, and greater estimated LV filling pressures. Am J Kidney Dis 47:493-502.2006 by the National Kidney Foundation, Inc.
NDEX WORDS: Ischemia-modified albumin; end-stage renal disease (ESRD); renal transplantation candidate;
ortality; risk stratification.wustIh
P
ca
lHS
N
o
mtU
EVERAL BIOMARKERS have been identi-fied that predict cardiac and total mortality
n patients with end-stage renal disease (ESRD).hese include cardiac troponin,1,2 natriuretic pep-
ides,3,4 high-sensitivity C-reactive protein,5,6 andsymmetrical dimethylarginine.7 It has not beenstablished whether targeted intervention to pa-ients based on biomarker findings improves out-ome. Nevertheless, given that cardiac diseaseemains the leading cause of mortality in patientsith ESRD,8 the identification of new prognostic
actors in these patients is important. Ischemia-odified albumin (IMA) is a novel biomarker of
cute ischemia with high sensitivity, but moder-te specificity.9,10 It has not been studied inatients with ESRD.In this single-center, prospective, observa-
ional study of 114 renal transplantation candi-ates, we hypothesized that baseline IMA levelsay predict both cardiac disease and mortality in
atients with ESRD. During a 2-year follow-uperiod, the primary end point was total mortality.MA level was compared with other markers ofdverse outcome in patients with ESRD. Theptimal IMA level that predicted mortality wasetermined. We then investigated cardiac struc-
ural and functional characteristics of patientsmerican Journal of Kidney Diseases, Vol 47, No 3 (March), 2006
ith elevated baseline IMA levels. All patientsnderwent coronary angiography, dobutaminetress echocardiography (DSE), and determina-ion of levels of baseline biochemical markers.nformed consent was obtained, and the studyad full ethical approval.
METHODS
opulationBetween January 2002 and December 2003, a total of 114
onsecutive patients referred for renal transplantation evalu-tion at St George’s Hospital, London, UK, were studied
From the Departments of Cardiology, Chemical Patho-ogy, and Renal Medicine, St George’s Hospital; The Heartospital, London; and the Department of Medical Statistics,outhampton General Hospital, Southampton, UK.Received October 24, 2005; accepted in revised form
ovember 29, 2005.Originally published online as doi:10.1053/j.ajkd.2005.11.026
n January 24, 2006.Support: None. Potential conflict of interest: None.Address reprint requests to Rajan Sharma, MD, Depart-
ent of Cardiology, E Level East Wing, Mailpoint 46, Southamp-on General Hospital, Tremona Rd, Southampton SO16 6YD,K. E-mail: [email protected]© 2006 by the National Kidney Foundation, Inc.0272-6386/06/4703-0016$32.00/0
doi:10.1053/j.ajkd.2005.11.026: pp 493-502 493
pytnEuw
T
wpmveeivdsmcc
Daptwvodww
D
adiaoBc1(aT(optos
epmnc
mdswscwTa
S
oau
C
tenlc�cnctn
I
cmAWto1up
E
mtb(
C
rmsgm((
SHARMA ET AL494
rospectively. Exclusion criteria were age younger than 18ears, severe aortic stenosis, unstable angina, and inabilityo consent. Clinicians reporting echocardiographic and coro-ary angiographic data were blinded to biochemical data.chocardiography and coronary angiography results weresed to determine suitability for renal transplantation andhether revascularization was required.
ransthoracic EchocardiographyThe General Electric Vingmed System 7 (Horten, Nor-
ay) was used. For those on dialysis therapy, studies wereerformed 16 to 24 hours postdialysis, when patients wereost likely to be closest to their euvolemic state.11,12 Left
entricular (LV) end-diastolic diameter, LV end-systolic diam-ter, and interventricular and LV posterior wall thickness atnd-diastole were measured from parasternal M-mode record-ngs of the left ventricle, with the cursor at the tips of the mitralalve leaflets. LV end-systolic and end-diastolic volumes wereetermined by using a modified biplane Simpson rule and thetandard formula applied to give LV ejection fraction. Measure-ents were averaged over 3 cardiac cycles. LV mass was
alculated according to Devereux and Reichek.13 This wasorrected for body surface area to give LV mass index.
Transmitral inflow was recorded by using pulsed-waveoppler recordings at the mitral valve leaflet tips in the
pical 4-chamber view. Peak velocity of early filling (E),eak velocity of atrial filling (A), E/A ratio, and E decelera-ion time were measured. Flow propagation velocity (Vp)as determined from color M-mode in apical 4-chamberiew. From pulsed-wave real-time tissue Doppler imagesbtained in the 4-chamber view, peak systolic and earlyiastolic (Ea) velocities were measured. The sample volumeas placed at the lateral mitral annulus. LV filling pressureas estimated from E/Ea and E/Vp ratios.14
obutamine Stress EchocardiographyImages were acquired in parasternal long- and short-axis
nd apical 2-, 3-, and 4-chamber views at baseline anduring stepwise infusion of dobutamine, administered accord-ng to a protocol based on 3-minute stages of 5, 10, 20, 30,nd 40 �g/kg/min. Atropine was administered up to a totalf 1.0 mg intravenously if target heart rate was not achieved.lood pressure and 12-lead electrocardiograms were re-orded at each infusion stage. Baseline, low-dose (heart rate,0 to 15 beats/min greater than baseline), peak, and recovery10 minutes after drug infusion was terminated) stage im-ges were stored and analyzed in digital quad screen format.he test was stopped if: (1) target heart rate was achieved
[220 � age] � 0.85), (2) ST depression greater than 2 mmccurred, (3) significant tachyarrhythmia (sustained su-raventricular tachycardia or � 3-beat run of ventricularachycardia) occurred, (4) symptomatic severe hypotensionccurred, or (5) blood pressure exceeded 240 mm Hgystolic or 140 mm Hg diastolic.
All images were reported off-line by 2 experienced observ-rs blinded to the rest of the study. Qualitative analysis waserformed with the left ventricle divided into a 17-segmentodel. Regional wall motion was described as hyperkinetic,
ormal, hypokinetic, akinetic, and dyskinetic. Results were
lassified as a normal response with an overall increase in wall ootion or abnormal response. An abnormal response wasescribed by the occurrence under stress of hypokinesia, akine-ia, or dyskinesia in 1 or more resting normal segments and/ororsening of wall motion in 1 or more resting hypokinetic
egments.15 Resting and peak wall motion index scores werealculated. The level of agreement between the 2 sonographersas � � 0.82. Consensus was obtained in discordant cases.here were 6 discordant cases, and agreement was reachedfter discussion between the 2 observers.
ample AcquisitionWhole-blood venous samples were collected at the time
f DSE. Samples were allowed to clot and were centrifugedt 3,000 rpm for 10 minutes. Serum was decanted and frozenntil analysis.
ardiac Troponin T AssayCardiac troponin T (cTnT) was measured by using the
hird-generation TROP T STAT assay on an Elecsys 2010lectrochemiluminescent immunoassay system (Roche Diag-ostics, Lewes, East Sussex, UK).16 The assay detectionimit was less than 0.01 ng/mL (�0.01 �g/L). Total assayoefficients of variation were 5.5% at 0.32 ng/mL (0.32g/L) and 5.4% at 6.0 ng/mL (6.0 �g/L). The correspondingoncentration with a 10% coefficient of variation was 0.03g/mL (0.03 �g/L). The receiver operating characteristicurve medical decision cutoff value for myocardial infarc-ion is 0.05 ng/mL (0.05 �g/L).17 The 99th percentile is 0.01g/mL (0.01 �g/L).
schemia-Modified AlbuminSerum IMA was measured by means of the albumin
obalt-binding (ACB) test on a Cobas MIRA PLUS instru-ent (ABX Diagnostics, London, UK). The first-generationCB test has been validated and described previously.18,19
e used a second-generation ACB test, and in our labora-ory, assay coefficients of variation were 5.1% in the rangef 56.67 to 66.57 KU/L and 3.1% in the range of 147.17 to58.03 KU/L. According to the manufacturer, the IMApper limit of normal is 85 KU/L, determined from the 95th
ercentile of 283 apparently healthy individuals.
stimated Glomerular Filtration RateIn patients with residual urine production (n � 70), esti-ated glomerular filtration rate (GFR) was calculated by using
he Cockcroft-Gault equation: GFR � (140 � age) � leanody weight (kg)/72 � stable serum creatinine (�mol/L) �0.85 for females).
oronary AngiographyCoronary angiograms were interpreted blindly by 2 expe-
ienced observers and consensus was obtained for disagree-ent. Level of agreement was � � 0.85. The stenosis
everity of each epicardial artery was assessed visually andraded as follows: normal, mild (�50% luminal narrowing),oderate (50% to 70% luminal narrowing), and significant
�70% luminal narrowing). Severe coronary artery diseaseCAD) is defined as luminal stenosis greater than 70% in 1
r more epicardial artery.
S
dpccvpclldcmttca(
F
itm
B
121tldrwfpeic�c(epr
F
0ct
1dyb
AS
E
SSCcISE
D
DPC
D
HISP
PS
nmm0
ISCHEMIA-MODIFIED ALBUMIN IN ESRD 495
tatistical AnalysisContinuous variables are expressed as mean � 1 SD, and
ifferences between groups were determined by using un-aired t-test. Categorical variables were compared by usinghi-square analysis or Fisher exact test. Pearson correlationoefficient was calculated for the correlation of continuousariables with IMA level. The optimal IMA cutoff value toredict mortality was determined from receiver operatingharacteristic analysis. The area under the curve was calcu-ated and compared with cTnT level. Stepwise multivariateogistic regression analysis was used to determine indepen-ent predictors of mortality. Long-term survival related toTnT and IMA levels was analyzed in a Kaplan-Meierodel. Log-rank test was used to evaluate differences be-
ween Kaplan-Meier curves. All statistical tests were 2ailed, with P less than 0.05 to indicate statistical signifi-ance. The level of agreement between 2 observers wasnalyzed by using � statistics. The SPSS statistics packageSPSS Inc, version 12; Chicago, IL) was used.
ollow-UpFollow-up of clinical status was obtained by review of
npatient and outpatient medical records and telephone callso the transplantation unit. The primary end point is all-causeortality.
RESULTS
aseline Characteristics
Of 117 renal transplantation candidates screened,patient was excluded for severe aortic stenosis,patients refused to consent for the study. Thus,
14 patients were enrolled. Baseline characteris-ics are listed in Table 1. Diabetes (40%) was theeading cause of renal failure and 55% were onialysis therapy. Eighty-one patients (71%) hadesidual urine production. Mean estimated GFRas 16 � 9 mL/min (0.27 � 0.15 mL/s). Thirty-
our patients (30%) had severe CAD and 23atients (20%) had impaired LV function (LVjection fraction � 50%), but none had severempairment (LV ejection fraction � 30%). Medi-ation use included aspirin in 51 patients (45%),-blockers in 40 patients (35%), angiotensin-onverting enzyme inhibitors in 53 patients46%), statins in 58 patients (51%), erythropoi-tin in 60 patients (53%), and diuretics in 49atients (44%). Medication was not adjusted inelation to DSE or coronary angiogram results.
ollow-UpMean follow-up was 2.25 � 0.71 years (range,
.19 to 3.27 years). Thirty-eight patients re-eived a renal transplant, and 7 patients were
aken off the transplant list. There were 18 deaths, t0 of which were cardiac related. Mean time toeath was 0.97 � 0.45 years (range, 0.19 to 3.15ears). Eight patients underwent coronary arteryypass surgery, and 11 patients underwent percu-
Table 1. Baseline Characteristics
ge (y) 52 � 12 (44-76)exMen 76 (67)Women 38 (33)
thnicityWhite 51 (45)Asian 27 (24)Afro-Carribean 33 (29)Chinese 2Other 1
erum creatinine (mg/dL) 6.9 � 3.1 (2.9-15.0)erum hemoglobin (g/dL) 10.2 � 1.6 (8-14)holesterol (mg/dL) 189 � 70 (54-580)TnT (ng/mL) 0.08 � 0.05 (0-1.08)MA (KU/L) 88 � 29 (39-164)erum albumin (g/dL) 3.6 � 0.7 (2.1-4.4)SRD causeDiabetes 46 (40)Glomerulonephritis 24 (21)Hypertension 14 (12)Adult polycystic kidney
disease 13 (11)Obstructive uropathy 10 (9)Other 7 (5)ialysis modalityPredialysis 48 (42)Hemodialysis 43 (38)Peritoneal dialysis 23 (20)ialysis time (mo) 2.73 � 2.19 (0-11)revious transplant 13 (11)ardiac symptoms 52 (46)NYHA class 1 35 (31)NYHA class 2 17 (15)NYHA classes 3 and 4 0 (0)iabetes 46 (40)Type 1 20Type 2 26ypertension 105 (92)
ncreased cholesterol 59 (52)moker 14 (12)reexisting cardiac diseaseIschemic heart disease 12 (11)Heart failure 10 (8)
ositive DSE result 35 (31)evere CAD 34 (30)
NOTE. Values expressed as mean � 1 SD (range) orumber (percent). To convert creatinine in mg/dL to �mol/L,ultiply by 88.4; hemoglobin and albumin in g/dL to g/L,ultiply by 10; cholesterol in mg/dL to mmol/L, multiply by.02586; cTnT in ng/mL to �g/L, multiply by 1.Abbreviation: NYHA, New York Heart Association.
aneous intervention. The decision for revascular-
ig
Ca
v
odBgwafi
NAMFSHIDPC
PPPDDBSDCHCCPAcILLLLLLLLEEPS
�0m
SHARMA ET AL496
zation was driven by DSE and coronary angio-ram results.
omparison Between Survivorsnd Nonsurvivors
Differences between survivors and nonsurvi-ors are listed in Table 2. Nonsurvivors were
Table 2. Comparison of
Parameter Survivors
o. of patients 96 (88)ge (y) 50 � 13ale 66 (61)emale 30 (31)moker 10 (11)ypertension 90 (92)
ncreased cholesterol 50 (44)iabetes 31 (33)ositive family history 10 (11)ardiac symptoms 44 (46)NYHA class 1 30 (31)NYHA class 2 14 (15)
ast history ischemic heart disease 7 (8)ast history heart failure 6 (7)revious renal transplant 9 (11)ialysis 56 (58)ialysis time (mo) 2.58 � 1.26ody mass index (kg/m2) 25 � 16ystolic blood pressure (mm Hg) 145 � 18iastolic blood pressure (mm Hg) 76 � 11holesterol (mg/dL) 189 � 48emoglobin (g/dL) 11.2 � 1.4reatinine (mg/dL) 6.6 � 3.1alcium (mg/dL) 9.4 � 1.0hosphate (mg/dL) 5.4 � 1.2lbumin (g/dL) 3.7 � 1.5TnT (ng/mL) 0.07 � 0.04MA (KU/L) 83 � 26V end-systolic diameter (cm) 2.8 � 0.8V end-systolic volume (cm)3 34 � 22V end-diastolic diameter (cm) 4.7 � 0.9V end-diastolic volume (cm)3 105 � 43V fractional shortening (%) 39 � 10V ejection fraction (%) 68 � 13V mass index (g/m2) 132 � 18A size (cm) 3.8 � 0.6/Ea ratio 11.2 � 4.2/Vp ratio 2.02 � 0.72ositive DSE result 25 (26)evere CAD 24 (23)
NOTE. Values expressed as number (percent) or mean �mol/L, multiply by 88.4; hemoglobin and albumin in g/dL t.02586; cTnT in ng/mL to �g/L, multiply by 1; calcium imol/L, multiply by 0.3229.*Comparison of survivors and nonsurvivors.
†Comparison of survivors and mortality from cardiac causes.lder and a significantly greater proportion hadiabetes, severe CAD, and a positive DSE result.aseline cTnT and IMA levels were significantlyreater in nonsurvivors. LV cavity dimensionsere larger, systolic function was lower, left
trial (LA) size was larger, and estimated LVlling pressures were greater in nonsurvivors.
vors and Nonsurvivors
survivors P* Cardiac Mortality P†
8 (16) 10 (9)� 12 0.05 57 � 4 0.05
0 (55) 0.69 5 (50) 0.918 (44) 0.61 4 (40) 0.744 (22) 0.57 2 (20) 0.785 (83) 0.88 8 (80) 0.929 (50) 0.71 6 (60) 0.375 (83) �0.001 9 (90) �0.0015 (27) 0.35 3 (30) 0.418 (44) 0.63 5 (50) 0.515 (28) 0.52 2 (20) 0.443 (17) 0.51 1 (10) 0.815 (28) 0.05 4 (40) 0.034 (23) 0.07 3 (30) 0.034 (22) 0.11 2 (20) 0.630 (55) 0.74 6 (60) 0.32
� 1.66 0.26 2.83 � 0.26 0.37� 10 0.89 23 � 11 0.88� 10 0.66 150 � 18 0.38� 9 0.49 82 � 11 0.79� 45 0.32 175 � 43 0.37� 1.3 0.68 11.5 � 1.4 0.52� 2.7 0.35 6.9 � 2.5 0.44� 0.9 0.74 9.7 � 0.5 0.22� 1.1 0.86 5.2 � 1.1 0.85� 1.9 0.37 3.3 � 1.4 0.32� 0.08 0.02 0.14 � 0.07 0.03� 25 0.02 101 � 38 0.03� 0.9 0.008 3.7 � 0.6 0.001� 22 0.004 66 � 27 0.002� 1.2 0.007 5.6 � 0.6 0.001� 46 0.01 144 � 74 0.009� 10 0.005 29 � 78 0.004� 14 0.006 51 � 16 0.001� 11 0.05 158 � 18 0.05� 0.7 0.02 4.3 � 0.6 0.03� 8.2 0.009 18.6 � 9.1 0.008� 0.78 0.04 2.61 � 0.72 0.03
0 (55) 0.01 8 (80) 0.0050 (55) 0.02 9 (80) 0.007
unless noted otherwise. To convert creatinine in mg/dL tomultiply by 10; cholesterol in mg/dL to mmol/L, multiply byL to mmol/L, multiply by 0.2495; phosphate in mg/dL to
Survi
Non
1591
1
1
12.85
2414883
17411.77.39.55.23.4
0.161023.764
5.71472752
1544.4
19.12.56
11
1 SDo g/L,n mg/d
Ss
CP
ippsomaKpanps2cue
P
fscera[ta19tct
CB
pI
A T in n
ISCHEMIA-MODIFIED ALBUMIN IN ESRD 497
ex and the proportion on dialysis therapy wereimilar in the 2 groups.
omparison of IMA and cTnT Levels toredict Mortality
Figure 1 shows receiver operating character-stic analysis for IMA and cTnT levels toredict mortality. Area under the curve com-arison showed both markers had high, butimilar, predictive accuracy (P � 0.83). Theptimal biomarker concentrations to predictortality were determined from receiver oper-
ting characteristic curves. IMA level of 95U/L or greater, seen in 46 patients (40%),redicted mortality with a sensitivity of 76%nd specificity of 74%. cTnT level of 0.06g/mL or greater (�0.06 �g/L), seen in 51atients (45%), predicted mortality with a sen-itivity of 75% and specificity of 72%. Figureshows Kaplan-Meier survival curves for both
TnT and IMA according to these cutoff val-es. Thirty-eight patients (33%) had both el-
Diagonal segments are produced by ties.
1 - Specificity
1.00.75.50.250.00
Se
nsi
tivity
1.00
.75
.50
.25
0.00
IMA
AUC = 0.76 (95% CI 0.59, 0.94)
P = 0.05
Optimal IMA concentration to predict mortality i
95 KU/L (sensitivity 76%, specificity 74%)
Fig 1. Receiver operating characteristic curves for bbbreviation: AUC, area under the curve. To covert cTn
vated IMA and cTnT levels. T
redictors of Mortality
Table 3 shows logistic regression analysisor predictors of mortality. In univariate analy-is, diabetes, severe CAD, positive DSE result,TnT level, IMA level, LV end-systolic diam-ter, LV ejection fraction, LA size, and E/Eaatio all predicted mortality. In multivariatenalysis, a positive DSE result (odds ratioOR], 8.11; 95% confidence interval [CI], 5.12o 9.67; P � 0.003), combined elevated IMAnd cTnT levels (OR, 7.12; 95% CI, 4.14 to0.12; P � 0.005), and E/Ea ratio (OR, 6.21;5% CI, 4.95 to 8.11; P � 0.009) were foundo be independent prognostic factors. IMA andTnT levels alone were not independent predic-ors of mortality.
haracteristics of Patients With Elevatedaseline IMA Levels
Cumulative mortalities during the follow-uperiod for patients with and without increasedMA levels were 79% and 96%, respectively.
Diagonal segments are produced by ties.
1 - Specificity
1.00.75.50.250.00
Se
nsi
tivity
1.00
.75
.50
.25
0.00
cTnT
UC = 0.82 (95% CI 0.64, 0.99)
= 0.02
ptimal cTnT concentration to predict mortality is
06 ng/mL (sensitivity 75%, specificity 72%)
e IMA and cTnT concentrations that predict mortality.g/mL to �g/L, multiply by 1.
s
A
P
O
0.
aselin
able 4 lists differences in patients with and
wgsngltpilct(Pean
iitrm
neiwgcKsPhsittiTadqenets
S
m
SHARMA ET AL498
ithout baseline IMA levels of 95 KU/L orreater. Those with elevated IMA levels hadignificantly greater LA and LV cavity size, sig-ificantly decreased LV systolic function, andreater estimated LV filling pressures. Age, cTnTevel, medication, and proportions on dialysisherapy, with diabetes, with severe CAD, with aositive DSE result, and undergoing revascular-zation were similar in the 2 groups. When corre-ation coefficients were calculated for IMA withontinuous variables, there was a significant posi-ive correlation with LV end-systolic diameterr � 0.39; P � 0.04) and E/Ea ratio (r � 0.41;
� 0.05) and a negative correlation with LVjection fraction (r � �0.44; P � 0.02) andlbumin level (r � �0.38; P � 0.04). There waso correlation with age or cTnT level.
DISCUSSION
This is the first study to investigate IMA levelsn patients with ESRD. We found that IMA levels a marker of mortality in a group of renalransplantation candidates. The predictive accu-acy was similar to that of cTnT level. By using
IMA (p = 0.02)Survival Functions
3.53.02.52.01.51.0.50.0
1.1
1.0
.9
.8
.7
.6
Follow Up Time (Years)
urvival
IMA < 95 KU/L
IMA > 95KU/L
Fig 2. Kaplan-Meier survival curves according to IMultiply by 1.
ultivariate analysis, cTnT and IMA levels were s
ot independent predictors of mortality. How-ver, combined elevated IMA and cTnT levelsndependently predicted mortality when testedith clinical, echocardiographic, and angio-raphic parameters. We determined the optimalutoff value for IMA to predict prognosis as 95U/L. This is greater than the IMA cutoff value
uggested for ischemia diagnosis (85 KU/L).atients with IMA levels of 95 KU/L or greaterad larger LA and LV size, more decreased LVystolic function, and greater estimated LV fill-ng pressures than those with IMA levels lesshan 95 KU/L. Age and proportions on dialysisherapy and with diabetes, severe CAD, andnducible ischemia were similar in the 2 groups.here was a negative correlation between IMA andlbumin levels. Additional studies are required toetermine whether baseline albumin level affectsuantitative IMA level. In this study, albumin lev-ls were not significantly different in survivors andonsurvivors and patients with and without el-vated IMA levels. Therefore, we do not believehe observed differences in IMA levels betweenurvivors and nonsurvivors can be explained by
cTnT (p = 0.001)
cTnT < 0.06 µg/LcTnT > 0.06 µg/L
3.53.02.52.01.51.0.50.0
cTnT < 0.06 ng/mL
cTnT > 0.06 ng/mL
Follow Up Time (Years)
vival
TnT concentrations. To covert cTnT in ng/mL to �g/L,
1.1
1.0
.9
.8
.7
Sur
A and c
ignificant differences in baseline albumin levels.
ttdphftdtr
oobkstIitcodcb
ecphw
rtplbiafmmadb3iaNhsceaapcnrhsttptlEbtt
plmhs
U
M
ISCHEMIA-MODIFIED ALBUMIN IN ESRD 499
The patients studied, namely renal transplanta-ion candidates, represented a lower risk grouphan previously studied; patients with ESRD onialysis therapy. This reflects current referralractice for renal transplantation. Only a thirdad severe CAD, 81% had normal LV systolicunction, none had an LV ejection fraction lesshan 30%, and a large proportion were not onialysis therapy. However, the use of biomarkerso assess risk in patients with ESRD is highlyelevant for patients referred for transplantation.
IMA is a marker of ischemia. The N-terminusf albumin becomes damaged under conditionsf ischemia, resulting in IMA. IMA is unable toind metals at the N-terminus. In the ACB test, anown amount of cobalt is added to the serumample. Unbound cobalt is measured spectropho-ometrically and is proportional to the amount ofMA in the sample. Serum IMA concentrationsncrease acutely after percutaneous coronary in-ervention9 and in patients with spontaneous is-hemia.20,21 Its exact role in the risk stratificationf patients with acute chest pain has yet to beefined. The recommended cutoff value for is-hemia diagnosis is 85 KU/L. Sensitivity is high,
Table 3. Univariate and Multivariate LogisticRegression Analysis to Determine Predictors
of Mortality
OR for Prediction ofMortality (95% CI) P
nivariate predictorsAge 0.07 (0.01-0.18) 0.79Sex 0.25 (0.12-0.51) 0.61Dialysis 3.34 (2.12-5.58) 0.07Diabetes 6.24 (3.32-8.18) 0.01cTnT 7.14 (5.71-10.22) 0.004IMA 6.91 (3.11-7.19) 0.04LV end-systolic
diameter (cm) 5.68 (2.21-7.22) 0.02LV end-diastolic
diameter (cm) 2.99 (1.63-5.11) 0.08LV ejection fraction (%) 6.99 (3.21-8.75) 0.008LA size (cm) 5.32 (2.14-7.34) 0.02E/Ea ratio 7.12 (3.44-8.12) 0.002LV mass index 0.08 (0.02-1.21) 0.77Positive DSE result 9.16 (4.45-11.95) 0.001Severe CAD 6.91 (2.18-12.11) 0.009Elevated IMA and cTnT 7.81 (2.16-12.11) 0.002ultivariate predictorsPositive DSE result 8.11 (5.12-9.67) 0.003Elevated IMA and cTnT 7.12 (4.14-10.12) 0.005E/Ea ratio 6.21 (4.95-8.11) 0.009
ut specificity is low. Serum concentrations are e
levated in patients with skeletal muscle is-hemia and should be interpreted with caution inatients with peripheral vascular disease.22 IMAas not been investigated previously in patientsith ESRD.Several biomarkers have been proposed for
isk stratification in patients with ESRD. Al-hough their ultimate role in the management ofatients with renal failure has yet to be estab-ished, they identify a high-risk group that mayenefit from therapeutic intervention. cTnT levels elevated in a proportion of patients with ESRDnd is associated with worse survival.1,2 Reasonsor this have yet to be determined. Potentialechanisms include CAD,23 microinfarctions,yocarditis, or arrhythmia. Natriuretic peptide
nd high-sensitivity C-reactive protein levels pre-ict poorer outcome in patients with ESRD.3-5 Arain natriuretic peptide cutoff concentration of90 ng/L has been shown to predict mortalityn hemodialysis patients.3 Apple et al6 proposed
multibiomarker model incorporating cTnT,-terminal pro-B-type natriuetic peptide, andigh-sensitivity C-reactive protein levels for risktratification in patients with ESRD. IncreasedTnT and high-sensitivity C-reactive protein lev-ls independently predicted death in 399 hemodi-lysis patients during a 2-year period. Tertilenalysis for N-terminal pro-B-type natriureticeptide also showed a prognostic value at a highutoff value. Levels of symmetrical dimethylargi-ine, a nitric oxidase inhibitor linked to atheroscle-osis and endothelial dysfunction, are increased inemodialysis patients and predict mortality.7 Thistudy suggests that IMA is an additional biomarkero predict outcome in patients with ESRD. Al-hough IMA and cTnT levels alone were not inde-endent predictors of mortality, this study showshat a combination of elevated IMA and cTnTevels independently predicts outcome, along with/Ea ratio and a positive DSE result. Such a com-ined biochemical marker model may be used inhe future to evaluate cardiac risk in potential renalransplantation candidates.
Using an IMA cutoff value of 95 KU/L, 46atients with ESRD (40%) had elevated baselineevels. This represents the optimal IMA level forortality prediction in this study. These patients
ad significantly increased LA and LV cavityize, decreased systolic function, and greater
stimated LV filling pressures. The mechanism
ftTGicdsrrscciiwdhaohtw
sEgcsgbpwvpaaptowcgrs
c
SHARMA ET AL500
or this remains unclear. There was no associa-ion with severe CAD or a positive DSE result.he proportions on dialysis therapy and meanFRs were similar in those with and without
ncreased IMA levels. This suggests that in-reased IMA levels are genuine markers of car-iac damage in patients with renal failure and notimply a false-positive result related to impairedenal excretion. Although high IMA levels mayeflect low-grade ischemia, cTnT level was notignificantly increased in this group. This indi-ates that increased levels do not represent signifi-ant myocardial necrosis. ACB test results arenfluenced by ischemia, with or without irrevers-ble myocardial necrosis.19 IMA level previouslyas studied primarily for the detection of myocar-ial ischemia. However, increased IMA levelsave been reported because of gastrointestinalnd skeletal muscle ischemia.24,25 Modificationf albumin during ischemia may be caused byypoxia, free radical damage, or acidosis. Allhese processes may be occurring in patients
Table 4. Characteristics of Patients With and Witho
IM
Age (y) 5DialysisCreatinine (mg/dL)GFR (mL/min)DiabetesPast history heart failurePast history ischemic heart diseaseNYHA classPositive DSE resultSevere CADHemoglobin (g/dL) 1cTnT (ng/mL) 0Albumin (g/dL)LV end-systolic diameter (cm)LV end-diastolic diameter (cm)LV fractional shortening (%) 3LV end-systolic volume (mL)LV end-diastolic volume (mL) 1LV ejection fraction (%)Peak systolic velocity (m/s) 0E/Ea ratio 1E/Vp ratioLA size (cm)LV mass index (g/m2) 15
NOTE. To convert creatinine in mg/dL to �mol/L, multiplTnT in ng/mL to �g/L, multiply by 1; GFR in mL/min to mL/Abbreviation: NYHA, New York Heart Association.
ith renal disease as a result of tissue oxidative s
tress. Increased IMA levels in patients withSRD therefore may reflect skeletal muscle,astrointestinal, and renal ischemia, as well asardiac ischemia. However, cTnT level is a veryensitive marker for myocardial necrosis and noteneralized ischemia in other organs. This maye a reason that IMA level adds incrementalrognostic information to cTnT level in patientsith ESRD. Additional studies are required toerify this hypothesis. The decreased survival inatients with increased IMA levels is explained,t least in part, by the increased LV cavity sizend decreased systolic function because both areredictors of mortality in patients on dialysisherapy26,27 and after renal transplantation.28 Thebserved greater LA size and E/Ea ratio in thoseith elevated IMA levels, suggestive of in-
reased LV filling pressure, is associated withreater mortality in nonuremic patients.29 E/Eaatio is an independent prognostic factor in thistudy.
Main study limitations are the small sample
Elevated Baseline IMA Level of 95 U/mL or Greater
KU/L)
IMA � 95 KU/L(n � 68) P
3.8 49.6 � 11.2 0.68) 40 (58) 0.84.9 6.5 � 1.8 0.582 19 � 10 0.39) 25 (37) 0.42) 5 (7) 0.88) 7 (10) 0.87.5 1.6 � 0.6 0.16) 20 (29) 0.39) 20 (29) 0.33.8 10.8 � 1.1 0.32.05 0.07 � 0.04 0.11.8 3.7 � 1.5 0.22.9 2.6 � 0.8 0.04.9 4.5 � 1.0 0.021.2 39.6 � 11.7 0.054 28 � 16 0.031 94 � 26 0.052 69 � 15 0.05.02 0.09 � 0.02 0.02.1 11.9 � 4.3 0.05.7 1.8 � 0.6 0.07.4 3.9 � 2.4 0.041.6 159.1 � 54.3 0.62
.4; hemoglobin and albumin in g/dL to g/L, multiply by 10;iply by 0.01667.
ut an
A � 95(n � 46
2.7 � 126 (576.7 � 217 � 121 (455 (115 (13
1.9 � 015 (3314 (311.4 � 1.09 � 03.4 � 13.1 � 05.1 � 05.1 � 146 � 221 � 460 � 2.07 � 06.5 � 82.2 � 04.3 � 22.3 � 3
y by 88s, mult
ize and short follow-up. Larger scale studies are
rilt
p9Lfscga
td
Pd2
ii2
tA
Cm4
Mnca
cis
D
Jmt
mg1
cp
K
dh
mt
CDpt
t
ngS
eCst
taC
s(d1
hmC
EC
aend
tcu2
ric1
adC
Dd
ISCHEMIA-MODIFIED ALBUMIN IN ESRD 501
equired to validate our findings and show anyncremental prognostic information that IMAevel provides over existing risk-stratificationools in patients with ESRD.
In conclusion, IMA level predicts mortality inatients with ESRD. Using a cutoff IMA value of5 KU/L, patients with elevated levels had largerA and LV cavity size, decreased LV systolic
unction, and greater estimated LV filling pres-ures than those in whom levels were not in-reased. These findings suggest that IMA is aenuine biomarker of mortality and cardiac dam-ge in patients with ESRD.
REFERENCES
1. Dierkes J, Domrose U, Westphal S, et al: Cardiacroponin T predicts mortality in patients with end-stage renalisease. Circulation 102:1964-1969, 20002. Apple FS, Murakami MM, Pearce LA, Herzog CA:
redictive value of cardiac troponin I and T for subsequenteath in end-stage renal disease. Circulation 106:2941-2945,0023. Goto T, Takase H, Toriyama T, et al: Increased circulat-
ng levels of natriuretic peptides predict future cardiac eventn patients with chronic hemodialysis. Nephron 92:610-615,0024. Naganuma T, Sugimura K, Wada S, et al: The prognos-
ic role of brain natriuretic peptides in hemodialysis patients.m J Nephrol 22:437-444, 20025. Yeun JY, Levine RA, Mantadilok V, Kaysen GA:
-Reactive protein predicts all-cause and cardiovascularortality in hemodialysis patients. Am J Kidney Dis 35:469-
76, 20006. Apple FS, Murakami MM, Pearce LA, Herzog CA:ulti-biomarker risk stratification of N-terminal pro-B-type
atriuretic peptide, high-sensitivity C-reactive protein, andardiac troponin T and I in end-stage renal disease forll-cause death. Clin Chem 50:2279-2285, 2004
7. Zoccali C, Bode-Boger S, Mallamaci F, et al: Plasmaoncentration of asymmetrical dimethylarginine and mortal-ty in patients with end stage renal disease: A prospectivetudy. Lancet 358:2113-2117, 2001
8. US Renal Data System: Causes of death. Am J Kidneyis 32:S81-S88, 1998 (suppl 2)9. Sinha MK, Gaze DC, Tippins JR, Collinson PO, Kaski
C: Ischemia modified albumin is a sensitive marker ofyocardial ischemia after percutaneous coronary interven-
ion. Circulation 107:2403-2405, 200310. Pollack C, Peackock W, Summers R, et al: Ischemia-odified albumin is useful in risk stratification of emer-
ency department chest pain patients. Acad Emerg Med0:555-556, 200311. London GM, Fabiani F, Marchais SJ, et al: Uremic
ardiomyopathy: An inadequate left ventricular hypertro-hy. Kidney Int 31:973-980, 198712. Harnett JD, Murphy B, Collingwood P, Purchase L,
ent G, Parfrey PS: The reliability and validity of echocar- 1
iographic measurement of left ventricular mass index inemodialysis patients. Nephron 65:212-214, 199313. Devereux RB, Reichek N: Echocardiographic deter-ination of left ventricular mass in man. Anatomic valida-
ion of the method. Circulation 55:613-618, 197714. Ommen SR, Nishimura RA, Appleton CP, et al:
linical utility of Doppler echocardiography and tissueoppler imaging in the estimation of left ventricular fillingressures: A comparative simultaneous Doppler-catheteriza-ion study. Circulation 102:1788-1794, 2000
15. Armstrong WF: Stress echocardiography for detec-ion of coronary artery disease. Circulation 84:I43-I49, 1991
16. Hallermayer K, Klenner D, Vogel R: Use of recombi-ant human cardiac troponin T for standardization of thirdeneration troponin T methods. Scand J Clin Lab Investuppl 230:S128-S131, 199917. Collinson PO, Stubbs PJ, Kessler AC: Multicentre
valuation of the diagnostic value of cardiac troponin T,K-MB mass, and myoglobin for assessing patients with
uspected acute coronary syndromes in routine clinical prac-ice. Heart 89:280-286, 2003
18. Christenson RH, Duh SH, Sanhai WR, et al: Charac-eristics of an albumin cobalt binding test for assessment ofcute coronary syndrome patients: A multicenter study. Clinhem 47:464-470, 200119. Wu AH, Morris DL, Fletcher DR, Apple FS, Christen-
on RH, Painter PC: Analysis of the albumin cobalt bindingACB) test as an adjunct to cardiac troponin I for the earlyetection of acute myocardial infarction. Cardiovasc Toxicol:147-151, 200120. Bhagavan NV, Lai EM, Rios PA, et al: Evaluation of
uman serum albumin cobalt binding assay for the assess-ent of myocardial ischemia and myocardial infarction.lin Chem 49:581-585, 200321. Morrow DA, de Lemos JA, Sabatine MS, Antman
M: The search for a biomarker of cardiac ischemia. Clinhem 49:537-539, 200322. Roy D, Quiles J, Aldama G, et al: Ischemia modified
lbumin for the assessment of patients presenting to themergency department with acute chest pain but normal oron-diagnostic 12-lead electrocardiograms and negative car-iac troponin T. Int J Cardiol 97:297-301, 200423. deFilippi C, Wasserman S, Rosanio S, et al: Cardiac
roponin T and C-reactive protein for predicting prognosis,oronary atherosclerosis, and cardiomyopathy in patientsndergoing long-term hemodialysis. JAMA 290:353-359,00324. Apple FS, Quist HE, Otto AP, Mathews WE, Mu-
akami M: Release characteristics of cardiac biomarkers andschemia modified albumin as measured by the albuminobalt binding test after a marathon race. Clin Chem 48:1097-100, 2002
25. Roy D, Quiles J, Sharma R, et al: Ischemia modifiedlbumin concentrations in patients with peripheral vascularisease and exercise induced skeletal muscle ischaemia.lin Chem 50:1656-1660, 200426. Parfrey PS, Foley RN, Harnett JD, Kent GM, Murray
C, Barre PE: Outcome and risk factors for left ventricularisorders in chronic uraemia. Nephrol Dial Transplant 11:
277-1285, 1996
tsS
o
c1
da
SHARMA ET AL502
27. Zoccali C, Benedetto FA, Mallamaci F, et al: Prognos-ic value of echocardiographic indicators of left ventricularystolic function in asymptomatic dialysis patients. J Amoc Nephrol 15:1029-1037, 200428. McGregor E, Jardine AG, Murray LS, et al: Pre-
perative echocardiographic abnormalities and adverse out- C
ome following renal transplantation. Nephrol Dial Transplant3:1499-1505, 1998
29. Wang M, Yip GW, Wang AY, et al: Peak earlyiastolic mitral annulus velocity by tissue Doppler imagingdds independent and incremental prognostic value. J Am
oll Cardiol 41:820-826, 2003
