●

nsgmtt5nas0aaT©

Ir

Aipsttdtsb

o

U

M

o

FP

ssu

4

Arterial Stiffness Assessed by Digital Volume Pulse Correlates WithComorbidity in Patients With ESRD

Daniel Sollinger, MD, Markus G. Mohaupt, MD, Albina Wilhelm, MD, Dominik Uehlinger, MD,Felix J. Frey, MD, and Ute Eisenberger, MD

Background: Digital volume pulse (DVP), a noninvasive method for indirect assessment of arterial stiffness, wasot tested previously in patients with end-stage renal disease (ESRD). Therefore, we compared the DVP-derivedtiffness index (SIDVP) with aortic pulse wave velocity (PWV) determined by means of Doppler ultrasonography in 2roups of patients with ESRD and analyzed the correlation between SIDVP and comorbidity. Methods: Photoplethys-ography was performed on the index finger of the dominant hand or the hand from the nonfistula arm in 49 renal

ransplant (TX) recipients and 48 hemodialysis (HD) patients. Pulse curves were analyzed with computer assis-ance. Comorbidity was assessed by using an established index. Results: The intrasubject variability of SIDVP was.7%. SIDVP and aortic PWV values correlated significantly (r � 0.66; P � 0.001) in patients with ESRD. SIDVP couldot be assessed reliably in 25% and 6% of HD patients and TX recipients, respectively. Multivariate regressionnalyses showed that SIDVP increased with age in both HD patients and TX recipients (r � 0.61; P < 0.001) and withystolic blood pressure (r � 0.53; P < 0.025), mean arterial pressure (r � 0.47; P < 0.05), and pulse pressure (r �.52; P � 0.02) in TX recipients. Severity of comorbid status was associated highly with individual residuals ofge-adjusted SIDVP in HD patients and TX recipients (P < 0.001). Conclusion: DVP allows the measurement ofrterial stiffness in most, but not all, patients with ESRD. SIDVP values correlate with comorbidity in HD patients andX recipients. Am J Kidney Dis 48:456-463.2006 by the National Kidney Foundation, Inc.

NDEX WORDS: Arterial stiffness; digital volume pulse; photoplethysmography; pulse wave velocity; end-stage

enal disease (ESRD); comorbidity.

pcttsfltlruvs

avpfhacfvcommf

RTERIAL STIFFNESS is a major indepen-dent risk factor for cardiovascular morbid-

ty and overall mortality in the general and renalopulations.1-4 Several methods have been con-idered for the assessment of arterial stiffness inhe past. To date, pulse wave velocity (PWV) ishe most widely used noninvasive method toetermine arterial stiffness.1 PWV, measured be-ween the carotid and femoral artery, reflects thetiffness of the aorta and increases with age andlood pressure (BP).2,5

Recently, Millasseau et al6 described an indexf stiffness based on analysis of digital volume

From the Department of Nephrology and Hypertension,niversity Hospital of Bern, Switzerland.Received November 17, 2005; accepted in revised formay 8, 2006.Originally published online as doi:10.1053/j.ajkd.2006.05.014

n July 13, 2006.Support: Research support provided by Swiss National

oundation for Scientific Research, grant no. 31-102 153.otential conflicts of interest: None.Address reprint requests to Ute Eisenberger, MD, Univer-

ity Hospital Bern, Department of Nephrology and Hyperten-ion, Freiburgstr 10, CH-3010 Bern, Switzerland. E-mail:te.eisenberger@insel.ch© 2006 by the National Kidney Foundation, Inc.0272-6386/06/4803-0012$32.00/0

ldoi:10.1053/j.ajkd.2006.05.014

American Journal of Kidn56

ulse (DVP; SIDVP). The contour of the DVP isharacterized by 2 distinct waves, an early sys-olic peak caused by direct antegrade pressureransmission along the arterial tree followed by aecond peak caused by the pressure wave re-ected at sites of impedance mismatch, mainly in

he lower limb. Increased arterial stiffness ofarge conduit vessels leads to earlier return of theeflected wave, resulting in elevated SIDVP val-es. SIDVP is based upon the height of the indi-idual divided by the time between the earlyystolic peak and inflection point (Fig 1).

Because of the simplicity of its determination,ssessment of SIDVP might be a valuable nonin-asive alternative to detect arterial stiffness inatients with end-stage renal disease (ESRD). Soar, the utility of this index was tested only inealthy individuals, patients with hypertension,nd diabetes.7,8 In patients with ESRD, factorsausing arterial stiffness are different in partrom those in other populations investigated pre-iously because chronic fluid overload, arterialalcifications, microinflammation, and activationf the renin-angiotensin-aldosterone system areore pronounced. As a corollary, observationsade in other populations are not directly trans-

erable to patients with ESRD.Therefore, we studied SIDVP in an ESRD popu-

ation and compared these results with values

ey Diseases, Vol 48, No 3 (September), 2006: pp 456-463

oftmww

S

ctatc

(lipnpaTw

pattvaewtg

licHa

nh4ss

D

eamcoeiatAtfi(f

MDdcsp(pacbpepscab

baS

iamamTti

f

DIGITAL VOLUME PULSE IN ESRD 457

btained by measuring PWV, the standard methodor assessment of arterial stiffness. To determinehe biological significance of arterial stiffnesseasured by using SIDVP, the values obtainedere analyzed with respect to their correlationith comorbidity.9

METHODS

ubjectsA cross-sectional study was performed in 140 subjects,

onsisting of 97 patients with ESRD and 43 healthy volun-eers in our institution. All subjects were without signs ofcute infection or arrhythmia. The study was approved byhe internal review board of the institution, and informedonsent was obtained from all participants.

Of 97 patients with ESRD, 49 were kidney transplantTX) recipients with stable graft function and follow-up of ateast 3 months, administered either a double (calcineurinnhibitor plus prednisone) or triple (calcineurin inhibitor,rednisone, plus azathioprine or mycophenolic acid) immu-osuppressive regimen. In addition, 48 hemodialysis (HD)atients participated in the study. Dialysis was performed onn outpatient basis thrice weekly for more than 3 months.wo patients had previously received a kidney allograft, butere back on dialysis therapy because of TX failure.Comorbidity was scored by using the comorbidity index

roposed by Khan et al.9 This index, derived from patientge and a number of comorbidities, appropriately expresseshe prognostic impact of comorbidity on mortality in pa-ients with ESRD. For this purpose, the presence of cardio-ascular disease (history of myocardial infarction, coronaryrtery disease, stroke, or peripheral arterial occlusive dis-ase), cardiovascular risk factors, or other comorbiditiesere reported as indicated in patients’ files. Diabetes melli-

us was defined as a fasting glucose level of 126 mg/dL or

Fig 1. Calculation of SIDVP by using the DVP wave-orm.

reater (�7.0 mmol/L) and/or a nonfasting serum glucose t

evel of 200 mg/dL or greater (�11.1 mmol/L) and/or use ofnsulin or oral antidiabetic medication. Hypertension wasonsidered to be present if systolic BP (SBP) was 140 mmg or greater, diastolic BP (DBP) was 90 mm Hg or greater,

nd/or at least 1 antihypertensive drug was used.A control group of 43 subjects (27 women; 62.8%) with

o history of renal impairment was recruited from theospital medical staff and local community. Mean age was6 � 14 years, 7 subjects (16%) had a history of hyperten-ion; 4 subjects (9%) had a history of diabetes, and 6ubjects (14%) were active smokers.

VP and PWV RecordingA Pulse Trace PCA (Micro Medical Ltd, Rochester, UK)

quipped with an infrared light-emitting diode (940 nm) andphoto diode receiver was used to record DVP photoplethys-ographically. A finger clip containing an emitter and re-

eiver was applied to the index finger of the dominant armr, in the presence of an arteriovenous fistula, the contralat-ral arm. Determination of SIDVP was described previouslyn detail.7 In brief, the contour of the DVP was recorded andnalyzed. The systolic peak and inflection point were ob-ained by analyzing the first derivative of DVP waveforms.

single waveform was obtained by averaging DVP con-ours during a period of 10 seconds. The time between therst systolic peak and the inflection point in the waveforms�TDVP) was determined. SIDVP was calculated by using theollowing equation: body height/�TDVP (Fig 1).

PWV was recorded using a Pulse Trace PWV (Microedical Ltd) with a 4-MHz continuous wave directionaloppler pencil probe. In brief, pulse trace measures PWV byetermining the time lag between the R wave of the electro-ardiogram and the arrival of the arterial pulse at 2 differentites. After application of electrocardiogram electrodes, theencil probe was placed over (1) the right carotid artery and2) the right or left femoral artery. Carotid positioning of theencil probe was not influenced by the position of anrteriovenous fistula. The femoral position of the probe washosen randomly, regardless of the position of the allograft,ecause femoral PWV values obtained by positioning theencil probe contralateral and ipsilateral to the graft werequal in 5 TX recipients (results not shown). At each site, theosition was optimized to achieve a good quality signal andtable pulse waveform. About 10 systolic peaks were re-orded at both sites to obtain mean time lag at the carotidrtery and time lag at the femoral artery. The distanceetween the 2 detection sites was recorded.Brachial BP was measured while the subject was recum-

ent by using a mercury sphygmomanometer, and phases Ind V of Korotkoff sounds were considered to representBP and DBP, respectively.All subjects were examined at ambient room temperature

n a supine position after at least 5 minutes of rest. To ensurehemodynamically stable condition, BP and heart rateeasurements were performed before and after each DVP

nd PWV determination. All patients examined by bothethods were investigated subsequently at the same time.he sequence of methods was randomly assigned. No atten-

ion was given to the long (3 days) or short (2 days) dialysisnterval in HD patients; all measurements were performed in

he middle of the dialysis session while the patients were

hmo

S

escwvtm1wlpG

C

t

TfeoiprdwnH(b

R

1om0

NABMMDCSSSSPK

S

MHP

MH

mm

SOLLINGER ET AL458

emodynamically stable, assessed by using BP measure-ents. Individual results of 3 to 4 consistent measurements

f SIDVP and PWV were averaged.

tatistical AnalysisCorrelations between PWV, DVP, age, and BP were

xamined by using univariate and multiple regression analy-is. Chi-square or Fisher exact test were applied to analyzeategorical variables, whereas 1-way analysis of varianceas used for comparison of normally distributed continuousariables. Receiver operating characteristic curves were es-ablished to test the association of DVP and PWV measure-ents with a comorbid risk score (Khan Index) by plotting� specificity versus sensitivity. Unless indicated other-ise, all results are presented as mean � SD. P of 0.05 or

ess is considered significant. All statistical analyses wereerformed using Systat 10.0 software (Systat SoftwarembH, Erkrath, Germany).

RESULTS

haracteristics of Patients With ESRD

Characteristics of both ESRD study popula-ions at the time of inclusion are listed in Table 1.

Table 1. Clinical and Demographic Cha

Renal TX Reci

o. of patients (men/women) 49 (28/21ge (y) 50 � 14ody mass index (kg/m2) 25 � 4onths on dialysis —onths since transplantation 59 � 66iabetes (%) 33oronary artery disease (%) 18moking (%) 27erum creatinine (mg/dL) 1.7 � 0.erum calcium (mg/dL) 9.8 � 0.erum phosphate (mg/dL) 3.2 � 0.arathyroid hormone (pg/mL) 97 � 62han scoreLow risk 22 (45)Medium risk 14 (29)High risk 13 (26)

IDVP (m/s) 9.0 � 2.Men/women (m/s) 10.1 � 2.2/7.6AP† (mm Hg) 100 � 6eart rate† (beats/min) 68 � 13WV (m/s) 10.3 � 2.Men/women (m/s) 10.6 � 2.5/9.9AP‡ (mm Hg) 99 � 5eart rate‡ (beats/min) 69 � 14

NOTE: Values expressed as mean � SD or number (pg/dL to �mol/L, multiply by 88.4; serum calcium in mg/dmol/L, multiply by 0.3229.*Cross-tabulation showed a greater presence of high-risk†SI measurements.

DVP

‡PWV measurements.

X recipients were slightly younger and lessrequently had a history of coronary artery dis-ase than patients on dialysis therapy. Analysisf comorbidity score as proposed by Khan et al9

ndicated a more severe overall disease state inatients on dialysis therapy compared with TXecipients. Mean PWV was greater in patients onialysis therapy than in renal TX recipients,hereas the corresponding SIDVP values wereot significantly different (Table 1). In 12 of 48D patients (25%) and 3 of 49 TX recipients

6.1%), SIDVP could not be assessed reliablyecause no inflection point was detectable.

eproducibility

To assess the intrasubject variability of SIDVP,9 patients with ESRD were investigated on 2ccasions. Mean SIDVP values were 8.9 � 0.3/s in kidney graft recipients (n � 8) and 9.4 �

.7 m/s in HD patients (n � 11). Within-subject

stics of TX Recipients and HD Patients

HD Patients P

48 (33/15)60 � 14 0.00125 � 5 0.854 � 52

—31 0.936 0.0319 0.4—

9.5 � 0.6 0.015.6 � 2.2 0.001304 � 248 0.001

9 (19) 0.002*10 (21) 0.002*29 (60) 0.002*10.3 � 2.8 0.1

10.7 � 2.7/9.5 � 2.692 � 14 0.0273 � 15 0.4

11.4 � 4.0 0.0311.9 � 4.2/11.1 � 3.7

93 � 13 0.0367 � 14 0.9

) unless noted otherwise. To convert serum creatinine inmol/L, multiply by 0.2495; serum phosphate in mg/dL to

rbid factors in HD patients (P � 0.002).

racteri

pients

)

648

4� 1.8

9� 3.3

ercentL to m

como

cpf

CFW

e(sc0sa

A

smma00E0im0i00

oormpS0ansigmi

A

ai

btmt

DIGITAL VOLUME PULSE IN ESRD 459

oefficients of variation were 3.7% and 7.2% foratients with ESRD and TX recipients and 5.7%or all 19 patients, respectively.

omparison of Arterial Stiffness Values Derivedrom DVP and PWV Measurements in Patientsith ESRD

Because PWV, but not SIDVP, values were differ-nt between the 2 groups of patients investigatedTable 1), analysis of results of the 49 patientstudied with both methods was performed. SIDVP

orrelated significantly with aortic PWV (r �.66; P � 0.001). The Bland-Altman plot (Fig 2)howed acceptable consistency between SIDVP

nd PWV (mean difference, �1.1 � 2.59 m/s).

ge Dependency of SIDVP

The relationship between age and SIDVP ishown in Fig 3. The slope (0.114 versus 0.117/s/y) and intercept on the y-axis (3.5 versus 3.4/s) of significant linear correlations were virtu-

lly identical in TX recipients (r � 0.65; P �.0001) and dialysis patients (r � 0.59; P �.0001; Fig 3A and B). In all patients withSRD, SIDVP increased with age (r � 0.61; P �.0001). A similar age dependency was observedn 43 controls with no history of renal impair-ent (slope, 0.099 m/s/y; intercept, 4.9 m/s; r �

.62; P � 0.0001). Correspondingly, aortic PWVncreased with age in renal TX recipients (r �.69; P � 0.0001) and dialysis patients (r �

Fig 2. Bland-Altman plot showing the differenceetween SIDVP minus PWV measurements as a func-ion of the mean of both measurements. Dashed lines,ean difference � SD; ‘, TX recipients; �, HD pa-

ients.

.61; P � 0.0001). a

Because arterial stiffness is influenced notnly by the age of an individual, but also byther hemodynamic parameters, we tested theelationship between SIDVP and SBP, DBP,ean arterial blood pressure (MAP), or pulse

ressure. By means of univariate analysis,IDVP correlated positively with SBP (r �.53; P � 0.03), MAP (r � 0.47; P � 0.05),nd pulse pressure (r � 0.52; P � 0.02), butot with DBP in the TX group. Multiple regres-ion analysis confirmed that SIDVP correlatedndependently with age and MAP in the TXroup. Conversely, none of the mentioned he-odynamic parameters correlated with SIDVP

n the HD group (data not shown).

ssessment of Risk Factors

To analyze whether individual SIDVP values a marker of arterial stiffness is an age-ndependent predictor of comorbidity in pa-

Fig 3. SIDVP as a function of age in (A) TX recipientsnd (B) HD patients.

tiberasjc

e4moHc(ab

w(0ScSTbwtutarijnPS

aaaeFv

TS

C

K

SOLLINGER ET AL460

ients with ESRD, patients were subdividednto groups. For this purpose, the differenceetween measured SIDVP and a theoreticallyxpected stiffness index derived from the cor-elation of SIDVP as a function of age (Fig 3And B) was calculated for every patient. Threeubgroups comprising an equal number of sub-ects were formed according to the residualhange in SIDVP, ie, the difference between the

Fig 4. Residuals (�SIDVP) of SIDVP as a function ofge in TX recipients and HD patients. Individual dotsre the difference (residuals) between expected andctual SIDVP. The zero line represents the theoreticallyxpected age-dependent value (SIexpect) derived fromig 3 A and B. Dashed lines are given by the cutoffalues of � 0.7 m/s, as rationalized in Results.

Table 2. Distribution of Sex, Cardiovascular Risk FAge-Correcte

Group 1SIDVP � Expected SI

otal patients 33 (100)exMen 15 (45)*Women 18 (55)*ardiovascular risk factorsCoronary artery disease 5 (15)†Diabetes 8 (24)Smoking 5 (15)Hypertension 30 (91)

han scoreLow risk 18 (55)‡Medium risk 7 (21)‡High risk 8 (24)‡

NOTE. Values expressed as number (percent).*P � 0.002, group 1 versus group 3.†P � 0.03, group 1 versus group 3.

‡Cross tabulations showed a significant (P � 0.001) deviance fr

xpected stiffness index and actual SIDVP (Fig). Analysis showed that a cutoff value of �0.7/s (dashed lines in Fig 4) led to the formation

f 3 groups of similar size in TX recipients andD patients. In addition, a fourth group was

omposed of patients with no detectable SIDVP

12 HD patients, 3 TX recipients). Informationbout age, sex, cardiovascular risk, and comor-idity of these 4 groups is listed in Table 2.A significantly greater percentage of men than

omen had SIDVP values greater than averageP � 0.015) or a not measurable SIDVP (P �.075). Subjects with a greater than expectedIDVP more frequently had (42%) a history oforonary artery disease than patients with anIDVP value less than average (17%; P � 0.031;able 2). Severity of comorbid status as definedy the Khan et al9 was associated significantlyith SIDVP (Table 2; Fig 5). Fifty-five percent of

he group with lower-than-expected SIDVP val-es had low comorbidity risk, whereas 58% ofhe group with SIDVP values greater than thege-corrected average had high comorbidityisk (P � 0.002). Patients with no detectablenflection point were older than all other sub-ects (68 � 11 versus 54 � 15 years; P � 0.005;

� 15 versus 81), had significantly greaterWV values than patients with a measurableIDVP (13.3 � 3.8 versus 10.4 � 3.4 m/s; P �

, and Comorbidity Status According to Groups ofP Residuals

p 2xpected

Group 3SIDVP � Expected

Group 4SIDVP Not Measurable

00) 24 (100) 15 (100)

0) 19 (79)* 11 (73)0) 5 (21)* 4 (27)

6) 11 (46)† 7 (47)2) 7 (29) 7 (47)2) 8 (33) 1 (7)8) 23 (96) 14 (93)

2)‡ 4 (17)‡ 1 (7)‡0)‡ 6 (25)‡ 1 (7)‡8)‡ 14 (58)‡ 13 (86)‡

actorsd SIDV

Grou

DVP as E

25 (1

15 (610 (4

4 (18 (38 (3

22 (8

8 (310 (47 (2

om expected values.

0b

utsspcshSd

nicvc

dmtpSa

sititawlaasvtdpelPy

EvniScfi

Grgget

cwtu

DIGITAL VOLUME PULSE IN ESRD 461

.05), and predominantly showed a high comor-id risk score (Table 2).Receiver operating characteristic curves were

sed to compare SIDVP and PWV for their abilityo indicate a high- or low-risk group of comorbidcore as defined by Khan et al.9 Sensitivity andpecificity of SIDVP tended to be greater com-ared with PWV to identify the group with highomorbid risk (P � 0.08; Fig 6). Sensitivity andpecificity for the low-risk group were equallyigh with both methods (area under the curve forIDVP and PWV, 0.80 and 0.75, respectively;ata not shown).

DISCUSSION

We report for the first time that arterial stiff-ess, assessed by using contour analysis of DVP,s applicable in patients with ESRD. The studylearly shows that elevated age-adjusted SIDVP

alues are associated significantly with overallomorbidity in patients with ESRD.

Arterial stiffness can be measured by usingifferent methods.1 Aortic PWV is the indirecteasurement of large-arterial stiffness most of-

en used, a method applied in many patientopulations, including patients with ESRD.1-4

IDVP derived by contour analysis of DVP is an

Fig 5. Comorbid status and groups of residuals.roup 1, less than �0.7 m/s (11 HD patients, 11 TX

ecipients); group 2, �0.7 m/s or less and �0.7 m/s orreater (10 HD patients, 15 TX recipients); group 3,reater than �0.7 m/s (15 HD patients, 17 TX recipi-nts); and group 4, SIDVP not measurable (12 HD pa-ients, 3 TX recipients).

lternative simple method to analyze arterialcd

tiffness. SIDVP and aortic PWV do not providedentical information. Aortic PWV, measured overhe carotid-femoral region, determines the veloc-ty of pressure transmission between both detec-ion sites and is related inversely to aortal compli-nce by using the Bramwell-Hill equation,10

hereas SIDVP is a composite parameter modu-ated by elastic properties of the large centralrteries and reflective properties of the peripheralrterial bed. In addition to the velocity of pres-ure wave reflection along large central conduitessels, the DVP waveform is affected by charac-eristics of ventricular ejection and the exactistribution of aortic and more peripheral sites ofressure wave reflection. 11 Despite these differ-nces, Millasseau et al6 found a significant corre-ation (r � 0.65; P � 0.001) between SIDVP andWV in 87 healthy individuals aged 21 to 68ears.In the present investigation of patients with

SRD, the correlation between PWV and SIDVP

alues was highly significant. However, it wasot possible to determine a reliable SIDVP valuen 15 of 97 patients. Patients with no detectableIDVP were significantly older and had signifi-antly greater aortic PWV values than patientsor whom SIDVP was detected successfully. Thus,t is possible that PWV and hence arterial stiff-

Fig 6. Receiver operating characteristic curvesomparing SIDVP and PWV measurements for the groupith high comorbid risk. Comparison of areas under

he curves tended to be different (P � 0.08). Areasnder the curve for SIDVP and PWV were 0.76 (95%

onfidence interval, 0.66 to 0.87) and 0.66 (95% confi-ence interval, 0.52 to 0.80), respectively.

ntr

EtcPsdictfptrawdR1i

sP6Hmaapfatoace

ompng

cwnve

Ppaccfacm

wpmapirdcdtcvhp

tgvaaondt

s4

Me

ce

nc1

Sc

SOLLINGER ET AL462

ess have progressed so far in these patients thathe method of contour analysis of the DVP haseached its natural technical limits.

Arterial stiffness is increased in patients withSRD compared with individuals with no his-

ory of renal impairment.12,13 This finding is notonfined to patients with ESRD on HD therapy.WV also is increased in patients with hyperten-ion with mild to moderate renal dysfunction,efined by decreased creatinine clearance.14 Thisnverse relationship between arterial stiffness andreatinine clearance is stronger in younger pa-ients, presumably because such confoundingactors as hypertension and diabetes are lessresent in this cohort.14,15 The relationship be-ween arterial stiffness and renal function waseported to be bidirectional. Greater PWV valuesre associated with decreased renal function,2

hereas impaired renal function is an indepen-ent predictor of arterial stiffness acceleration.16

eported PWV values are in the range of 7.2 to4.6 m/s, depending on age and severity of renalmpairment, in the study population.2,14-16

Most studies analyzing arterial stiffness in theetting of renal transplantation reported lowerWV values in kidney graft recipients (range,.6 to 11.1 m/s) than in patients on maintenanceD treatment (range, 7.2 to 12.8 m/s).17,18 Thisost likely is caused in part by differences in

ge, which is usually younger in graft recipients,nd selection bias because younger and healthieratients tend to undergo transplantation morerequently. Interestingly, 2 publications reporteddecrease in PWV values after successful renal

ransplantation.18,19 In line with previous data,ur kidney graft recipients had lower mean PWVnd SIDVP values than HD patients, although thisomparison is limited by a significant age differ-nce in both groups.

Despite these limitations of contour analysisf DVP at high PWV values, SIDVP measure-ents correlated significantly with age in our

atients with ESRD. Thus, age dependency defi-itely is a hallmark of arterial stiffness in theeneral20-22 and renal populations.13,23

Arterial stiffness is a prognostic parameter ofardiovascular and overall mortality in patientsith renal disease.2-13 To evaluate arterial stiff-ess as an individual risk factor, absolute SIDVP

alues have to be corrected for age. Blacher

t al23 showed that transformation of absolute s

WV values into an age-dependent PWV indexrovided information with high predictive powernd greater discriminatory prospective signifi-ance for mortality than conventional cardiovas-ular risk factors. Interestingly, our cutoff valuesor residuals of age-adjusted SIDVP matched thege-dependent PWV increments defined by Bla-her et al,23 suggesting similar utility of bothethods.The comorbidity score proposed by Khan et al9

as shown to be predictive for mortality in aopulation of 1,407 patients on renal replace-ent therapy. Our data clearly show that age-

djusted SIDVP is associated with comorbidity inatients with ESRD. This is particularly interest-ng because the association of such traditionalisk factors as serum cholesterol level with inci-ence or severity of cardiovascular disease isontroversial and possibly confounded by nontra-itional factors in the ESRD population.24,25 Inhe present investigation, correlation of SIDVP toomorbidity is based on a cross-sectional obser-ation. Therefore, the prognostic value of SIDVP

as to be confirmed in a prospective trial inatients with ESRD.In conclusion, results presented here indicate

hat SIDVP obtained by means of photoplethysmo-raphic analysis of DVP is a reliable marker ofascular aging, and elevated values are associ-ted with a state of increased coronary and over-ll comorbidity in patients with ESRD. The usef this method is limited in patients on mainte-ance HD treatment because of an inability toetect correct waveforms in 25% of tested pa-ients.

REFERENCES1. Covic A, Gusbeth-Tatomir P, Goldsmith DJ: Arterial

tiffness in renal patients: An update. Am J Kidney Dis5:965-977, 20052. Blacher J, Guerin AP, Pannier B, Marchais SJ, SafarE, London GM: Impact of aortic stiffness on survival in

nd-stage renal disease. Circulation 99:2434-2439, 19993. London GM, Blacher J, Pannier B, Guerin AP, Mar-

hais SJ, Safar ME: Arterial wave reflections and survival innd-stage renal failure. Hypertension 38:434-438, 2001

4. Laurent S, Boutouyrie P, Asmar R, et al: Aortic stiff-ess is an independent predictor of all-cause and cardiovas-ular mortality in hypertensive patients. Hypertension 37:236-1241, 20015. Smulyan H, Siddiqui DS, Carlson RJ, London GM,

afar ME: Clinical utility of aortic pulses and pressuresalculated from applanated radial-artery pulses. Hyperten-

ion 42:150-155, 2003

Dn1

sw9

Tst

ic

p

Hbg6

NaA

Ms1

ce

ic

mno

Rp

hcd

f6

pt

Oac

opr

Sm1

ps

i

DIGITAL VOLUME PULSE IN ESRD 463

6. Millasseau SC, Kelly RP, Ritter JM, Chowienczyk PJ:etermination of age-related increases in large artery stiff-ess by digital pulse contour analysis. Clin Sci (Lond)03:371-377, 20027. Millasseau SC, Guigui FG, Kelly RP, et al: Noninva-

ive assessment of the digital volume pulse. Comparisonith the peripheral pressure pulse. Hypertension 36:952-56, 20008. Millasseau SC, Kelly RP, Ritter JM, Chowienczyk PJ:

he vascular impact of aging and vasoactive drugs: Compari-on of two digital volume pulse measurements. Am J Hyper-ens 16:467-472, 2003

9. Khan IH, Campbell MK, Cantarovich D, et al: Compar-ng outcomes in renal replacement therapy: How should weorrect for case mix? Am J Kidney Dis 31:473-478, 1998

10. Bramwell JC, Hill AV: Velocity of transmission of theulse and elasticity of arteries. Lancet 199:891-892, 192211. van der Heijden-Spek JJ, Staessen JA, Fagard RH,

oeks AP, Boudier HA, van Bortel LM: Effect of age onrachial artery wall properties differs from the aorta and isender dependent: A population study. Hypertension 35:37-642, 200012. Blacher J, Demuth K, Guerin AP, Safar ME, Moatti

, London GM: Influence of biochemical alterations onrterial stiffness in patients with end-stage renal disease.rterioscler Thromb Vasc Biol 18:535-541, 199813. Guerin AP, Blacher J, Pannier B, Marchais SJ, SafarE, London GM: Impact of aortic stiffness attenuation on

urvival of patients in end-stage renal failure. Circulation03:987-992, 200114. Mourad JJ, Pannier B, Blacher J, et al: Creatinine

learance, pulse wave velocity, carotid compliance andssential hypertension. Kidney Int 59:1834-1841, 2001

15. Wang MC, Tsai WC, Chen JY, Huang JJ: Stepwisencrease in arterial stiffness corresponding with the stages of

hronic kidney disease. Am J Kidney Dis 45:494-501, 2005 o

16. Benetos A, Adamopoulos C, Bureau JM, et al: Deter-inants of accelerated progression of arterial stiffness in

ormotensive subjects and in treated hypertensive subjectsver a 6-year period. Circulation 105:1202-1207, 200217. Bahous SA, Stephan A, Barakat W, Blacher J, Asmar

, Safar ME: Aortic pulse wave velocity in renal transplantatients. Kidney Int 66:1486-1492, 200418. Covic A, Goldsmith DJ, Gusbeth-Tatomir P, Bu-

aescu I, Covic M: Successful renal transplantation de-reases aortic stiffness and increases vascular reactivity inialysis patients. Transplantation 76:1573-1577, 200319. Zoungas S, Kerr PG, Chadban S, et al: Arterial

unction after successful renal transplantation. Kidney Int5:1882-1889, 200420. Dillon JB, Hertzman AB: The form of the volume

ulse in the finger pad in health, atherosclerosis, and hyper-ension. Am Heart J 21:172-190, 1941

21. Avolio AP, Chen SG, Wang RP, Zhang CL, Li MF,’Rourke MF: Effects of aging on changing arterial compli-

nce and left ventricular load in a northern Chinese urbanommunity. Circulation 68:50-58, 1983

22. Avolio AP, Deng FQ, Li WQ, et al: Effects of agingn arterial distensibility in populations with high and lowrevalence of hypertension: Comparison between urban andural communities in China. Circulation 71:202-210, 1985

23. Blacher J, Safar ME, Guerin AP, Pannier B, MarchaisJ, London GM: Aortic pulse wave velocity index andortality in end-stage renal disease. Kidney Int 63:1852-

860, 200324. Wanner C, Krane V, Marz W, et al: Atorvastatin in

atients with type 2 diabetes mellitus undergoing hemodialy-is. N Engl J Med 353:238-248, 2005

25. Kielstein JT, Zoccali C: Asymmetric dimethylargin-ne: A cardiovascular risk factor and a uremic toxin coming

f age? Am J Kidney Dis 46:186-202, 2005