Nephrol Dial Transplant (2000) 15: 10141021Nephrology [623435]

Nephrol Dial Transplant (2000) 15: 1014–1021Nephrology
Dialysis
TransplantationOriginal Article
Arterial sti ffening and vascular calcifications in end-stage renal disease
Alain P. Gue ´rin, Ge ´rard M. London, Sylvain J. Marchais and Fabien Metivier
Service d’He ´modialyse, Ho ˆpital F.H. Manhe `s, Fleury-Me ´rogis, France
Abstract Introduction
Background. Epidemiological studies have identified
aortic sti ffness as an independent predictor of cardio-Epidemiological and clinical studies have shown thatvascular mortality in end-stage renal disease (ESRD)damage of large arteries is a major contributory factor
patients. In these patients, aortic pulse wave velocity to the high cardiovascular morbidity and mortality of
(PWV ) was associated with mediacalcosis, but theend-stage renal disease (ESRD) patients [1,2]. Theinfluence of arterial calcifications on the viscoelasticadverse e ffects of macrovascular disease are attribut-properties of large arteries was not well characterized.able to two principal mechanisms: (i) the presence ofThe purpose of the present study was to analyse theocclusive lesions, principally atherosclerotic plaques,influence of arterial calcifications on arterial sti ffnessresponsible for ischaemic lesions and /or infarctionin stable haemodialysed patients.downstream from the lesion and (ii) sti ffening ofMethods. We studied 120 stable ESRD patients onarterial walls associated with arterial dilatation andhaemodialysis. All patients underwent B-mode ultra-hypertrophy [3]. The increased sti ffness causes ansonography of common carotid artery (CCA), aorta,increase of systolic blood pressure (BP) and pulseand femoral arteries to determine CCA distensibility,pressure and a decrease of diastolic BP, thereby causingthe elastic incremental modulus (Einc), and the pres-increased left ventricular (LV ) afterload and alteredence of vascular calcifications. All patients underwentcoronary perfusion [4]. The principal outcomes ofmeasurement of aortic PWV and echocardiogram. Thethese changes are LV hypertrophy, aggravation ofpresence of calcifications was analysed semiquantit-coronary ischaemia, and increased fatigue of arterialatively as a score (0 to 4) according to the number ofwall tissues [3–5]. ESRD patients have sti ffer arteriesarterial sites with calcifications.than age- and BP-matched non-uraemic subjects [3,6].Results. Our observations indicate that arterial andFurthermore, in these patients, arterial hypertrophy isaortic sti ffness is significantly influenced by the pres-accompanied by increased incremental elastic modulusence and extent of arterial calcifications. The extent of(Einc) and increased pulse-wave velocity (PWV )arterial calcifications is in part responsible for increased[3,6,7]. Recent epidemiological studies on ESRDleft ventricular afterload, and is inversely correlatedpatients have identified aortic PWV and commonwith stroke volume. The influence of calcifications iscarotid artery (CCA) Einc as strong independent pre-independent of the role of ageing and blood pressure.dictors of their cardiovascular morbidity and mortalityArterial calcifications density increases with age, dura-[8,9]. While the observation that the Einc is elevatedtion of haemodialysis, the fibrinogen level, and thein ESRD patients strongly favours altered intrinsicprescribed dose of calcium-based phosphate binders.elastic properties or major architectural abnormalities,Conclusions. The results of this study showed that thethe pathogenic factors contributing to arterial sti ffeningpresence of vascular calcifications in ESRD patientsin ESRD are less obvious. Calcification of elasticwas associated with increased sti ffness of large capacity,lamellae and increased calcium contents are observedelastic-type arteries, like the aorta and CCA. Thein the arteries of uraemic patients [10,11]. In haemodia-extent of arterial calcifications increased with the uselysed patients, aortic PWV was found to be associatedof calcium-based phosphate-binders.with mediacalcosis of the aorta and an elevated
calcium×phosphate product (Ca ×P) [6,12]. Never-
Keywords: arteries; calcifications; calcium carbonate; theless, the influence of arterial calcifications and /or
haemodialysis; viscoelasticity factors associated with calcifications on the viscoelastic
properties of large conduit arteries in ESRD has not
yet been well characterized. The purposes of the presentstudy were to analyse the influence of arterial calcifica-
tions on large conduit artery sti ffness, and to identifyCorrespondence and o ffprint requests to 😀 rG e ´rard M. London,factors associated with arterial calcifications and Hoˆpital F. H. Manhe `s, 8 Grande Rue, F-91712 Fleury-Me ´rogis,
Cedex, France. stiffening.
© 2000 European Renal Association–European Dialysis and Transplant Association

Vascular calci fications and enhanced arterial sti ffness 1015
[13]. The intima –media cross-sectional area (IMCSA) Subjects and methods
was calculated as p(CCA diameter /2+IMT)2−p(CCA
diameter/2)2, and the wall /lumen ratio as 2IMT /CCA dia-Patientsmeter [3]. The repeatability coe fficient of the IMT measure-
ment was ±60mm [3]. A localized echostructure encroachingOne hundred and twenty stable, non-diabetic, ESRD patientsinto the vessel lumen was considered to be plaque if theon haemodialysis for at least 12 months (89 ±66 months;CCA IMT was >50% thicker than neighbouring sites [14].range: 12 –304 months) were included. Based on a clinicalMeasurements of CCA diameter and CCA IMT were alwayswork-up and complementary paraclinical investigations (ech-performed in plaque-free arterial segments in CCA oppositeocardiography, ECG, echo-Doppler examination) patientsto the site of AV shunts. The presence of plaques wasor controls with acute myocardial infarction, haemo-determined in both CCAs.dynamically signi ficant valvular heart disease, cerebral vascu-Diastolic internal aortic diameter was measured ultrasono-lar disease, CCA stenosis or heart failure were excluded.graphically (Sonel 300, Compagnie Ge ´ne´rale de Radiologie,Patients were dialysed with synthetic membranes (AN69 orSaint-Cloud, France) using 3.5-MHz transducers. The meas-polysulphone) and a bicarbonate dialysate with 1.25 orurements were made blindly by two observers at the level of1.5 mmol /l of calcium according to serum calcium –phosphateaortic root diameter and the aortic bifurcation diameter.equilibrium and the necessity to use 1,25-vitamin D for theInterobserver reproducibility was ±1 mm, their values werecontrol of parathyroid hormone (PTH ) activity. The dura-then averaged. Aortic PWV was determined as carotid –tion of dialysis was individually tailored (4 –6 h thrice weekly)femoral PWV using the foot-to-foot method [6,15].to control body fluids and blood chemistries, and to achieveTranscutaneous Doppler flow velocity recordings were car-aK t/V>1.2 (1.46 ±0.13). Seventy-two patients were regu-ried out simultaneously at the base of the neck over thelarly receiving recombinant human erythropoietin. Forty-CCA and the femoral arteries in the groin with a SEGAeight patients were taking an antihypertensive drug. PatientsM842 8-MHz Doppler unit (Socie ´te d’Electronique Ge ´ne´raleregularly took iron and vitamin supplements. Calcium car-et Applique ´e, Paris, France) and a Gould 8188 recorder. Thebonate (CaCO3) was used to maintain predialysis phosphata-time delay (t) was measured between the feet of the flowemia<2.0 mmol /l. Investigations were performed in thewaves recorded at these di fferent points. The distance trav-morning before the first weekly haemodialysis session. Eachelled by the pulse wave was measured over the body surfacesubject gave informed written consent to participate in theas the distance between the two recording sites minus thatstudy, which was approved by our institutional review board.from the suprasternal notch to the carotid (D). PWV was
calculated as PWV =D/t. The reproducibility of the measure-
ment has been published previously [16,17]. Arterial geometry and function
CCA distensibility was determined from changes of CCA
diameter during systole and simultaneously measured CCA Brachial BP was measured, after 15 min of recumbency with
pulse pressure ( DP). CCA pulse pressure was recorded non- a mercury sphygmomanometer with a cu ffadapted to arm
invasively by applanation tonometry with a pencil-type probe circumference. The appearance of Korotko ffsounds was
incorporating a high- fidelity Millar strain gauge transducer taken as the systolic BP and their disappearance (phase V )
(SPT-301, Millar Instruments, Houston, Texas), and sphyg- as diastolic BP. CCA diameter and wall motion were meas-
mocardiograph (Sphygmocor PWV, Sydney, Australia). The ured by high-resolution B-mode ultrasonography (Scanner
CCA pressure wave was calibrated assuming that brachial 350, PIE Medical, Maastricht, The Netherlands) with a
and carotid diastolic and mean BPs were equal. Mean BP 7.5-MHz transducer, and echotracking system (Wall-Track
on the CCA pressure wave was computed from the area of System, Maastricht, The Netherlands) enabling assessment
the CCA pressure wave in the corresponding heart period, of arterial wall displacement during the cardiac cycle. A
and set equal to brachial mean BP. CCA pressure amplitude complete detailed description of this system has been pub-
was then computed from diastolic BP and the position of lished previously [3]. The radiofrequency signal over six
mean BP on the directly recorded CCA pressure wave. A cardiac cycles is digitized and stored in a large memory bank.
detailed description of this system has been published previ- Two sample volumes, selected under cursor control, are
ously [4,18]. CCA distensibility was determined according positioned on the anterior and posterior walls. The vessel
to following formulae: 2[(Ds −Dd)/Dd]/DP (kPa −1.10−3). walls are continuously tracked by sample volumes according
The repeatability coe fficient of the measurement was to phase and the displacement of the arterial walls is obtained
±1 kPa−1.10−3for CCA distensibility [4]. While distensibil- by autocorrelation processing of the Doppler signal. The
ity provides information about the ‘elasticity ’of the artery accuracy of the system is ±30mm for CCA diastolic diameter
as a hollow structure, the Einc gives information on the (Dd) and <± 1mm for stroke change in CCA diameter
intrinsic properties of the arterial wall biomaterial. Einc was (Ds–Dd, where Ds is systolic diameter). The repeatability
calculated as 3(1 +LCSA/IMCSA) /CCA distensibility [4]. coefficient of the measurements was ±0.273 mm for CCA
diameter, and ±0.025 mm for Ds –Dd. Measurements were
made on the right CCA, 2 cm beneath the bifurcation. CCA-lumen cross-sectional area (LCSA) was calculated as p(CCA
Echocardiography
diameter) 2/4. CCA intima –media thickness (IMT) was meas-
ured on the far wall at the same level as the diameter All subjects underwent echocardiography with a Hewlett-
Packard Sonos 100 device equipped with a 2.25-MHz probe. measurements with computer-assisted acquisition, pro-
cessing, and storage. The computing equipment was linked LV measurements were made according to the recommenda-
tions of the American Society of Echocardiography [19]. to an 80386 /16 MHz processor and an imaging card provid-
ing real-time digitizing of the video analogue signal from the Measurements were performed blindly by two physicians.
Inter- and intraobserver reproducibilities have been reported echo-recording (processing corresponding to 256 levels of
grey). The IMT was automatically determined from changes previously [16,17]. LV mass (LVM ) was calculated as
1.05×[(PWT+IVST+LVEDD) 3−(LVEDD) 3]−13.6 [20], of density on the section perpendicular to the vessel wall
with speci fic software (Eurequa, TSA, Meudon, France) where PWT is the posterior wall thickness, IVST is the

A. P. Gue ´rinet al. 1016
interventricular septal thickness, and LVEDD is the LV method [22] and measured only once on the day of haemo-
dynamic evaluation. Smoking habits, prescriptions for calci- end-diastolic diameter. LV mean wall thickness (LVMWT)
was calculated as (IVST +PWT )/2. The fractional triol, and the dose of CaCO3expressed in grams of elemental
calcium prescribed to each patient were recorded from the shortening of the LV was calculated as [(LVEDD −
LVESD)/LVEDD] ×100, where LVESD is the LV end- patients ’files.
systolic diameter. Stroke volume was calculated as the aorticvalve annular cross-sectional area multiplied by the LV
Statistical analysisoutflow velocity integral, and cardiac output as stroke volume
multiplied by heart rate. Total peripheral resistance wasData are expressed as means ±SD. Patients were classi fiedcalculated as mean BP ×80/cardiac index and expressed ininto five groups according their calci fication scores (0 –4).dynes.s.cm −5.m2. LV diastolic filling was evaluated fromAnalysis of variance (ANOVA) was used for comparisonspulsed Doppler studies obtained from the apical 4-chamberof the di fferent groups. Di fferences in frequency were deter-view of the heart. The sample volume was positioned in themined by x2analysis. Single and multiple regression analysesinflow area just at the tip of the mitral lea flets. Maximalwere conducted using the least-squares method and wereearly diastolic flow velocity (E) and maximal late atrial flowperformed on the entire population. Gender (1, male; 2,velocity (A) were measured and their ratio (E /A) calculated.female), was used as the dummy variable. Statistical analysis
was performed using NCSS 6.0. software (J. I. Hintze,
Arterial calcifications: calcification score Kaysville, Utah, USA). Repeatability and reproducibility ofthe methods were de fined as recommended by the British
The presence of arterial calci fications was evaluated ultra- Standards Institution [23]. A Pvalue<0.05 was considered
sonographically in the CCA, the abdominal aorta, the ili- signi ficant.
ofemoral axis, and the legs. Ultrasound examinations were
performed by the sonographer at the time of the determina-tion of arterial geometry. The protocol involved scanning of
Results
the near and the far walls of CCA(s) in a 4-cm segmentpreceding the carotid bifurcation. A 10-cm segment of
Patient characteristicsabdominal aorta above its bifurcation was scanned, and thefemoral arteries were examined distal to the inguinal ligament
Among the BP and cardiac characteristics of the five
proximal to the site of the division of super ficial and deepcalci fication score groups, only diastolic BP, strokefemoral arteries. Arteries were scanned longitudinally andvolume, LV fractional shortening and the E /A ratiotransversely to determine the presence of plaques. A localizeddiffered signi ficantly (Table 1). Diastolic BP was lowerechostructure encroaching into the vessel lumen was consid-in patients with the highest calci fication score andered to be plaque when the arterial wall was >50% thickernegatively correlated with the score ( r=− 0.331, than neighbouring sites. Highly echogenic plaques producing
P<0.001). Signi ficant decrease in stroke volume, LV bright white echoes with shadowing were considered to be
fractional shortening, and E /A ratio of transmitral calci fications [21]. Assessment of the presence of calci fica-
tions was complemented with posteroanterior and lateral velocities were observed in subjects with higher calci-
fine-detail radiographs of the abdomen and pelvis. Arterial fication score. When clinical parameters and blood
calci fications of the femorotibial arterial axis were evaluated chemistries were considered as a function of the calci-
by soft-tissue native radiographs. Arterial calci fications infication score (Table 2), age, smoking habits, the dura-each arterial region were quanti fied qualitatively as absenttion of dialysis, lower plasma albumin, increases of(0) or present (1). The final overall score was obtained byCRP and fibrinogen, were signi ficantly associated withthe addition of calci fications from all studied zones. Thehigher calci fication scores. The other variables evalu-final score ranged from 0 (absence of calcium deposits), to 4ated did not appear to be a ffected except PTH levels(calci fications present in all arterial segments examined). The
which was negatively correlated with calci fication score calci fication score was independently checked by two obser-
(P=0.047) and age ( P<0.0001). Twenty-eight patients vers. The reproducibility was absolute for patients with the
score from 2 to 4. Five patients classi fied as score 0 by one presented one or more hypercalcaemic episodes (Ca
of the observer were classi fied as score 1 by the second >2.6 mmol /l ), with a signi ficant trend towards more
observer. The discordance was related to calci fications of incidences for patients with high calci fication scores
iliofemoral axis which were absent on ultrasonography and(P=0.034). Also prescribed dose of CaCO3increasedvisible only after magni fication on radiographs. Thesesigni ficantly from group 0 to group 4. All the para-patients were finaly classi fied as having score 1.meters used to de fine arterial structures and function
were signi ficantly in fluenced by increasing calci fication
Blood chemistry scores (Table 3). CCA diameter, IMT, aortic root and
bifurcation diameters increased progressively with the
Predialysis blood chemistries determined twice monthlyscores, as did the aortic PWV and CCA Einc. Inincluded serum creatinine, urea, calcium, phosphataemia,parallel the CCA distensibility decreased.sodium, potassium, bicarbonates, and haemoglobin. Serum
albumin, blood lipids, plasma fibrinogen, C-reactive protein
(CRP), and PTH were measured every 3 months. The values Correlation study
considered in the present study are the averages of all above-The correlation matrix (Table 4) shows the interrela-mentioned measurements over the 1 to 4-year period preced-
tionships between age, blood chemistries, smoking, ing the study. The occurrence of at least one hypercalcaemic
and duration of dialysis, and calci fication scores. Note episode (Ca >2.60 mmol /l ) was carefully monitored. Total
plasma homocysteine was determined by fluorimetric HPLC the age-dependency of many studied parameters and

Vascular calci fications and enhanced arterial sti ffness 1017
Table 1. Blood pressure and cardiac parameters according to the calci fication score (0 to 4)
Parameters 0 1 2 3 4
(n=40) ( n=15) ( n=12) ( n=22) ( n=31) ANOVA
Systolic BP (mmHg) 145 ±22 150 ±28 149 ±39 151 ±32 151 ±34 NS
Diastolic BP (mmHg) 88 ±13 85 ±11 78 ±14 77 ±12 77 ±15 0.001
Mean BP (mmHg) 108 ±15 110 ±15 103 ±21 101 ±17 106 ±17 NS
Pulse pressure (mmHg) 57 ±14 64 ±22 70 ±22 82 ±28 82 ±27 0.001
Heart rate (beats /min) 71 ±11 66 ±11 66 ±10 74 ±13 68 ±12 NS
Stroke volume (ml ) 68 ±13 74 ±14 66 ±24 60 ±27 55 ±19 0.01
TPR (dynes.s.cm −5.m2) 2430 ±562 2668 ±912 2524 ±582 2167 ±565 2406 ±695 NS
LV shortening (%) 36 ±6±63 3 ±83 8 ±83 3 ±73 2 ±5 0.01
E/A (ratio) 1.06 ±0.33 1.03 ±0.29 1.08 ±0.50 0.80 ±0.22 0.88 ±0.33 0.001
LV mass (g) 256 ±94 288 ±123 245 ±51 313 ±105 295 ±100 0.07
Values are means ±SD. BP, blood pressure; TPR, total peripheral resistance; LV, left ventricle.
Table 2. Clinical characteristics and blood chemistries according to calci fication score (0 to 4)
Parameters 01234 ANOVA
Age (years) 41.4 ±13.9 48.4 ±14.2 53.9 ±18.4 65.7 ±9.6 63.2 ±11.7 0.001
M/F ratio 1.4 1.5 1.3 1.45 1.3 NS
Smoking (pack years) 4.8 ±9.3 11 ±15.4 5 ±8.8 17.3 ±25 17.1 ±24.4 0.01
BMI (kg /m2) 22.3 ±4 24.6 ±3.8 23.2 ±2.5 23 ±3.4 24.6 ±3.9 NS
Dialysis (months) 52 ±58 1 ±67 85 ±52 101 ±104 107 ±86 0.001
Cholesterol (mmol /l ) 4.8 ±1.04 5.17 ±0.60 4.85 ±1.25 5.30 ±1.12 5.05 ±1.26 NS
HDL (mmol /l ) 1.01 ±0.36 1.16 ±0.30 1.01 ±0.19 1.24 ±0.27 1.09 ±0.59 NS
Triglycerides (mmol /l ) 1.81 ±0.94 2.07 ±1.22 2.02 ±0.70 1.70 ±0.50 2.03 ±1.18 NS
Albumin (g /l ) 41.9 ±2.2 40 ±2.2 40.6 ±2.27 39.2 ±2.4 38 ±2.4 0.001
Fibrinogen (g /l ) 3.98 ±0.77 4.19 ±0.89 4.26 ±1.32 4.98 ±0.74 5.04 ±1.0 0.001
CRP (mg /l ) 5.2 ±3.6 6.9 ±3.6 6.4 ±4.8 11.8 ±8.3 12.8 ±11.9 0.001
Homocysteine (mmol /l ) 36.1 ±12.7 38.7 ±20.1 32.4 ±12.3 40.5 ±12.6 31.3 ±8.3 NS
Haematocrit (%) 32 ±5.3 29 ±5.1 31 ±33 1 ±3.6 32 ±4.6 NS
Calcium (mmol /l ) 2.30 ±0.12 2.31 ±0.18 2.31 ±0.12 2.35 ±0.09 2.36 ±0.12 0.07
Phosphataemia (mmol /l ) 1.94 ±0.43 1.97 ±0.50 2.0 ±0.42 1.74 ±0.40 1.97 ±0.46 NS
Ca×P (product) (mmol 2/l2) 4.40 ±0.80 4.54 ±1.21 4.57 ±0.77 4.14 ±0.83 4.64 ±1.6 NS
PTH (pg /ml ) 360 ±278 409 ±332 477 ±388 221 ±148 237 ±285 0.047
CaCO3(g/day) 1.35 ±1.10 1.35 ±0.74 1.50 ±0.81 1.84 ±0.94 2.18 ±0.93 0.001
1,25 (OH)D3 ( mg/day) 0.21 ±0.27 0.28 ±0.18 0.30 ±0.49 0.14 ±0.20 0.21 ±0.32 NS
% Hypercalcaemia 8 10 18 36 42 0.034
BMI, body mass index; CRP, C-reactive protein; PTH, parathyroid hormone; HDL, high-density lipoprotein; % hypercalcaemia, % of
patients having presented at least one episode of calcaemia >2.60 mmol /l.
Table 3. Arterial structure and function according to calci fication score (0 to 4)
Parameters 0 1 2 3 4 ANOVA
CCA diameter (mm) 6 ±0.82 6.3 ±0.89 6.3 ±0.70 6.67 ±0.97 6.73 ±0.80 0.001
CCA IMT ( mm) 700 ±95 784 ±156 800 ±102 849 ±80 830 ±81 0.001
Aorootdiameter (mm) 26.4 ±4 26.3 ±4 27.8 ±3.9 27.5 ±4.1 29.7 ±4.5 0.01
Aobifdiameter (mm) 16 ±2.4 16.7 ±4.4 16.7 ±1.5 17.8 ±3.3 18.1 ±2.6 0.01
CCA distens. (kPa −1.10−3) 22.6 ±9.5 17.5 ±7.8 18.7 ±11.9 13.5 ±6.3 11.5 ±6.3 0.001
CCA Einc (kPa.10 3) 0.47 ±0.24 0.61 ±0.31 0.62 ±0.48 0.76 ±0.37 1.01 ±0.60 0.001
Aortic PWV (cm /s) 914 ±180 946 ±141 1040 ±268 1270 ±384 1302 ±317 0.001
CCA, common carotid artery; Ao, aortic; Aobif, aortic bifurcation; CCA distens, CCA distensibility; Einc, incremental elastic modulus;
PWV, pulse wave velocity.
the positive correlations between fibrinogen, CRP, and was negatively correlated with age ( P<0.0001), blood
pressure ( P<0.0001), calci fication score ( P<0.0001), smoking, and the negative correlations between albu-
min and these parameters. Multiple regression analyses serum albumin ( P=0.001), CRP ( P<0.0001), fibrino-
gen ( P<0.0001), and smoking ( P<0.0001). In were performed to adjust the roles of di fferent patho-
genic factors to age and other confounding variables univariate analysis, the aortic PWV and CCA Einc
were positively correlated with all above-mentioned (Table 5). In univariate analysis the CCA distensibility

A. P. Gue ´rinet al. 1018
Table 4. Correlation matrix between calci fication score (Ca score) and variables signi ficantly di fferent between the groups
Age Albumin Smoking PTH CRP Fibrinogen CaCO3Dialysis Ca score
Age 1.00
Albumin −0.483** 1.00
Smoking 0.285* −0.262* 1.00
PTH −0.413** 0.147 −0.193§ 1.00
CRP 0.338** −0.435** 0.336** −0.185§ 1.00
Fibrinogen 0.443** −0.480** 0.361** −0.162 0.610** 1.00
CaCO30.128 −0.186§ 0.095 −0.053 0.255* 0.200 § 1.00
Dialysis −0.246* −0.054 −0.150 0.165 0.026 0.010 0.324** 1.00
Ca score 0.578** −0.434** 0.285* −0.204§ 0.374** 0.423** 0.396** 0.260* 1.00
§P<0.05; * P<0.01; ** P<0.001. PTH, parathyroid hormone; CRP, C-reactive protein; CaCO3, calcium carbonate therapy; dialysis,
duration of dialysis in months.
Table 6. Multiple regression analysis between arterial calci fication Table 5. Multiple regression analysis between indexes of arterial
stiffness and signi ficantly associated variables score and signi ficantly correlated variables
Dependent variable Dependent variable
Independent Sequential Partial Independent Sequential Partial
variable b-coefft-value Pr 2 r2 variable b-coefft-value Pr 2 r2
CCA distensibility Calci fication score
Age (years) 0.057 6.53 <0.0001 0.352 0.360 Age (years) −0.333−8.71<0.0001 0.411 0.404
Mean BP −0.262−8.83<0.0001 0.636 0.411 Fibrinogen (g /l ) 0.268 2.21 0.0292 0.400 0.042
CaCO3(g/day) 0.296 2.78 0.0061 0.483 0.067 (mmHg)
Ca score (0 –4)−1.155−3.00 0.0034 0.671 0.078 Dialysis (months) 0.007 5.03 <0.0001 0.580 0.188
F ratio=36.67 adjusted r2=0.565 F ratio=73.5 adjusted r2=0.663
(P<0.001) ( P<0.0001)
CCA Einc
Age (years) 0.013 5.74 <0.0001 0.278 0.227CaCO3expressed in grams of elemental calcium; dialysis, durationMean BPof haemodialysis in months.(mmHg) 0.014 7.63 <0.0001 0.503 0.342
Ca score (0 –4) 0.08 3.10 0.0024 0.543 0.080
F ratio=44.5 adjusted r2=0.533 Discussion
(P<0.001)
Aortic PWVThe arterial system of ESRD patients is characterized Age (years) 10.22 6.56 <0.0001 0.385 0.273
Mean BP by dilatation and intima –media hypertrophy of elastic-
(mmHg) 6.52 5.40 <0.0001 0.494 0.203 type, capacitative arteries, such as the aorta or the
Ca score (0 –4) 52.85 3.41 0.0009 0.540 0.092CCA [3,7]. This remodelling is associated with arterialF ratio=43.1 adjusted r2=0.528stiffening [3,6,7]. As the arteries become sti ffer, the(P<0.001)PWV increases and is responsible for an early return
of wave re flections from the periphery to the ascendingCCA, common carotid artery; BP, blood pressure; Einc, elasticaorta during systole, causing an abnormal rise of aortic incremental modulus; PWV, pulse-wave velocity; Ca, calci fication
score. systolic BP and a decrease of diastolic BP [3,4]. This
abnormal pressure pattern places an additional load
on the LV, leading to LV hypertrophy and alteredcoronary perfusion [3,4,24,25]. Recent studies have parameters. As shown in Table 4, many of these factors
are correlated with age and are mutually interrelated. shown that arterial sti ffness is a major predictor of all-
cause and cardiovascular mortality in haemodialysis Therefore after adjustment for all these interrela-
tionships, arterial sti ffening, whether expressed as CCA patients [8,9].
Arterial sti ffness increases with age and hypertension distensibility, aortic PWV, or CCA Einc was signi fic-
antly associated with three factors, i.e. age, mean BP, [3,15]. In ESRD patients, the arterial sti ffness is
increased in comparison to age- and BP-matched non- and the calci fication score. After adjustment for all
confounding factors, four factors were independently uraemic subjects [3,6,12,26]. This modi fication a ffects
elastic- and muscular-type arteries independently of associated with the calci fication score: the age, duration
of dialysis, dose of CaCO3prescribed, and plasma the presence of atherosclerotic plaques [7]. The increase
of arterial sti ffness is certainly multifactorial, but the fibrinogen concentration. The negative correlation
between PTH and calci fication score was not signi ficant exact mechanisms are not clear. In ESRD patients, the
arterial sti ffening is due to alterations of arterial wall after adjustment for age. Together, these four para-
meters accounted for 56.5% of the variance of the materials as characterized by increased Einc [3,7],
and/or major architectural abnormalities like those calci fication score (Table 6).

Vascular calci fications and enhanced arterial sti ffness 1019
seen in experimental uraemia and in the arteries of or age-independently correlated with arterial sti ffness
indexes. Arterial calcium contents are increased and uraemic patients, namely fibroelastic intimal thicken-
ing, calci fication of elastic lamellae, increased extracel- arterial calci fications are frequently observed in ESRD
patients [10,11]. Cardiovascular calci fications are risk lular matrix with more collagen and relatively less
elastic fibre content [10,11]. While the published data factors for cardiovascular events and, in ESRD
patients have been linked to cardiac dysfunction and agree that smoking is a factor associated with arterial
wall hypertrophy in ESRD patients [4,27], no consist- aortic stenosis [34,35]. Multiple risk factors have been
implicated in the development of arterial calci fications. ent and/or constant associations could be established
between arterial remodelling and common vascular The prevalence of arterial calci fications and calcium
content increase with age both in the general popula- risk factors. An association between arterial remodel-
ling and/or functional alterations and lipid abnormalit- tion and in ESRD patients [6,10,11]. In ESRD
patients, the calcium content is higher than in age-and ies in ESRD patients were found inconstantly. While
London et al . [6 ] and Saito et al . [12] reported an sex-matched controls [10,11]. In agreement with data
reported in the literature [10,11], our findings showed inverse relationship between aortic PWV and HDL
cholesterol, and Shoji et al . [28] reported a positive an association between the calci fication score and the
duration of haemodialysis. Secondary hyperparathy- relationship between aortic PWV and IDL cholesterol,
in other studies no correlations were found between roidism, hyperphosphataemia, elevated Ca ×P prod-
ucts, and increased vitamin D concentrations are serum lipids and arterial changes [3,10,21]. In the
present study, we were also unable to establish any frequently evoked as the principal causes associated
with vascular remodelling and /or arterial calci fications correlation between alterations of arterial structure
and function, and blood lipid alterations. [36,37]. Kawagishi et al. [27] reported that CCA IMT
(but not the presence of calci fications) was associated The present results demonstrate that plasma fib-
rinogen level was independently and signi ficantly asso- with hyperphosphataemia, while the femoral artery
IMT was associated with increased serum PTH. In ciated with the extent of arterial calci fications. In
addition, fibrinogen concentration was signi ficantly contrast, Savage et al. [21] did not find any relationship
between the presence of carotid plaques and PTH and positively correlated with age, smoking habits and
CRP levels, and negatively correlated with serum albu- levels, and found a negative link between femoral
plaque and phosphataemia or Ca ×P products. In their min level. Multiple stepwise regression analysis indi-
cated that fibrinogen was the only biochemical factor study using electron beam computed tomography,
Arad et al . [38] did not find the serum concentrations signi ficantly associated with the calci fication score.
Fibrinogen is elevated in ESRD patients [29] and could of calcium, 1,25-vitamin D, and PTH to be associated
with the presence of arterial calci fications. In the be associated with endothelial dysfunction and risk
factors like smoking, or could merely re flect an present study the Ca ×P product was lower than
the critical value for calcium precipitation (i.e. inflammatory response as shown by its close associ-
ation with the CRP level. Several studies have shown 5.65 mmol 2/l2), serum calcium was not increased, and
phosphataemia was kept at a reasonable level. that acute-phase reactions, characterized by increased
fibrinogen and /or CRP are prominent risk factors for Moreover, neither phosphataemia nor the Ca ×P prod-
uct was associated with the calci fication score or arter- cardiovascular events in the general population as well
as in ESRD patients [30,31]. The link between plasma ial geometry and sti ffness indexes. Patients with higher
calci fication scores had lower PTH levels, and the only fibrinogen and the presence of arterial plaques and
calci fications was also con firmed by Levenson et al. in variable that was independently associated with the
score of vascular calci fications was the amount of the general population [32].
The factors most frequently quoted in association CaCO3(elemental calcium) prescribed as a phosphate
binder. One of the adverse e ffects of calcium- with arterial sti ffening in ESRD patients are altered
calcium and phosphate metabolism and parathyroid based phosphate binders is hypercalcaemia, which
may in turn result in arterial calci fications. Hyper- activity. In haemodialysed patients, aortic PWV was
found to be associated with the presence of aortic phosphataemia is frequently observed in patients with
low bone metabolic activity and those with adynamic calci fications and increased Ca ×P products [6,12].
Studying renal transplant recipients, Barenbrock et al. bone disease, and these patients are frequently pre-
scribed higher doses of calcium-based phosphate [33] observed a relationship between high PTH levels
and decreased CCA distensibility. In the present study, binders to maintain acceptable phosphate levels. On
the other hand, patients with low bone activity and /or we observed a negative association between calci fica-
tion score and PTH levels, but this correlation was adynamic bone disease are more susceptible to develop
iatrogenic hypercalcaemia. In the present study, the related to the negative correlation between patients ’
age and PTH level. The present study showed that the patients with higher calci fication scores and lower PTH
concentrations (and probably lower metabolic bone presence and extent of arterial calci fications per se is
associated with arterial sti ffening, independently of age activity) received higher doses of CaCO3to control
phosphataemia and more frequently experienced hyp- and BP (Table 5). The association of calci fications
with arterial sti ffening was independent of factors such ercalcaemic episodes, even if they were transient. PTH
levels decreased with age (Table 4) and this decline as serum calcium, phosphataemia, Ca ×P product, or
PTH concentration, all these factors not being directly was coupled with more arterial calci fications. This

A. P. Gue ´rinet al. 1020
4. Nichols WW, O ’Rourke MF. Vascular impedance. In: picture, associating lower bone activity with arterial
McDonald ’s Blood Flow in Arteries: Theoretical, Experimentalcalci fication, looks like that seen in osteoporosis of theand Clinical Principles , 3rd edn. Edward Arnold, London, 1991;elderly, where the loss of bone calcium and osteopeniapp 330 –342is associated with ectopic calci fications in the arteries 5. Watanabe H, Ohtsuka S, Karihana M, Sugishita Y. Coronary
circulation in dogs with an experimental decrease in aortic [39,40]. Osteopenia and reduced bone mass have been
compliance. J Am Coll Cardiol 1993; 21: 1497 –1506 found to be associated with stroke, carotid athero-6. London GM, Marchais SJ, Safar ME et al . Aortic and largesclerosis, and aortic calci fications [41,42]. Growingartery compliance in end-stage renal failure. Kidney Int 1990;evidence supports the hypothesis that vascular calci-37: 137 –142
fications develop by mechanisms similar to bone forma- 7. Mourad JJ, Girerd X, Boutouyrie P, Laurent S, Safar ME,
London GM. Increased sti ffness of radial artery wall material tion and that bone-associated proteins known as Gla-
in end-stage renal disease. Hypertension 1997; 30: 1425 –1430containing proteins, such as osteocalcin and matrix8. Blacher J, Gue ´rin AP, Pannier B, Marchais SJ, Safar ME,Gla protein, have been localized to advanced athero-London GM. Impact of aortic sti ffness on survival in end-stagesclerotic lesions. In addition to these molecules, osteo- renal disease. Circulation 1999; 99: 2434 –2439
9. Blacher J, Pannier B, Gue ´rin AP, Marchais SJ, Safar ME, pontin has been also found at sites of arterial
London GM. Carotid arterial sti ffness as a predictor of cardio- calci fications [43,44].
vascular and all-cause mortality in end-stage renal disease.The principal limitation of the present study is theHypertension 1998; 32: 570 –574semiquantitative evaluation of vascular calci fication.10. Ibels LS, Alfrey AL, Hu ffer WE, Craswell PW, Anderson JT,Using the score to evaluate arterial calci fication is Weil R. Arterial calci fication and pathology in uremic patients
undergoing dialysis. Am J Med 1979; 66: 790 –796 rather crude and is likely to underestimate the true
11. Vincenti F, Amend WJ, Abele J, Feduska NJ, Salvatierra O Jr. calcium load. It also does not allow a quantitativeThe role of hypertension in hemodialysis-associated atheroscler-assessment of the calcium concentration in the arteries.osis. Am J Med 1980; 68: 363 –369More sensitive methods (helical or electron-beam com- 12. Saito Y, Shirai K, Uchino J et al.Effect of nifedipine administra-
puted tomography) should be used in the future to tion on pulse wave velocity (PWV ) of chronic hemodialysis
patients: 2-year trial. Cardiovasc Drug Ther 1990; 4: 987 –990 quantify vascular calci fication and to assess the time-
13. Bonithon-Kopp C, Ducimetie `re P, Touboul JP et al . Plasmarelated alterations of vascular calcium content.converting-enzyme activity and carotid wall thickening.However, using a simple score is reproducible andCirculation 1994; 89: 952 –954speci fic, and the method is inexpensive and readily 14. Roman MJ, Saba PS, Pini R et al. Parallel cardiac and vascular
adaptation in hypertension. Circulation 1992; 86: 1909 –1918 available. The second limitation is represented by the
15. Avolio AP, Chen S, Wang R, Zhang C, Li M, O ’Rourke MF. observational nature of the relationship between the
Effects of aging on changing arterial compliance and left ventri-calci fications-associated risk of calcium-based phos-cular load in a northern Chinese urban community. Circulationphate binders. To clarify more precisely the eventual1983; 68: 50 –58
long-term risk associated with the use of these phos- 16. London GM, Marchais SJ, Gue ´rin AP et al . Salt and water
retention and calcium blockade in uremia. Circulation 1990; phate binders, prospective controlled studies compar-
82: 105 –113ing these drugs with alternative ones should be17. London GM, Pannier B, Gue ´rin AP, Marchais SJ, Safar ME,instigated.Cuche JL. Cardiac hypertrophy, aortic compliance, peripheralIn conclusion, the present study shows that the resistance, and wave re flection in end-stage renal disease:
presence of vascular calci fications in ESRD patients is Comparative e ffects of ACE inhibition and calcium channel
blockade. Circulation 1994; 90: 2786 –2796 associated with increased sti ffness of large, capacitative,
18. Kelly R, Hayward C, Ganis J, Daley J, Avolio A, O ’Rourke M.elastic-type arteries like the aorta and CCA. The extentNoninvasive registration of arterial pressure pulse waveformof arterial calci fication increases with age and theusing high- fidelity applanation tonometry. J Vasc Med Biolduration of dialysis. The other factors are increased 1989; 1: 142 –149
19. Sahn DJ, De Maria A, Kisslo J, Weyman A. Recommendations fibrinogen levels, and the use of calcium-based phos-
regarding quantitation in M-mode echocardiographic measure- phate binders. The use of calcium binders is widely
ments. Circulation 1978; 58: 1072 –1083advocated in ESRD patients to replace aluminium20. Devereux RB, Reichek N. Echocardiographic determination ofsalts, but it seems that development of non-calciumleft ventricular mass in man: Anatomic validation of the method.binders would be preferable, especially for older Circulation 1977; 55: 613 –618
21. Savage T, Clarke AL, Giles M, Tomson CRV, Raine AEG. patients with lower bone metabolic activity.
Calci fied plaque is more common in the carotid and femoral
arteries of dialysis patients without clinical vascular disease. Acknowledgements. This work was supported by GEPIR (Groupe
Nephrol Dial Transplant 1998; 13: 204 –212 d’Etude de Pathophysiologie de l ’Insuffisance Renale) and UMIF
22. Fortin LJ, Genest J. Measurement of homocyst(e)ine in the (Union des Mutuelles de l ’Ile-De-France).
prediction of atherosclerosis. Clin Biochem 1995; 28: 155 –162
23. British Standards Institution. Precision of test methods. 1: Guide
for the determination of reproducibility for a standard test References
method. British Standards Institution, London, 1979: BS 5497,part 1 1. Raine AEG, Margreiter R, Brunner FP et al . Report on
24. Marchais SJ, Gue ´rin AP, Pannier BM, Levy BI, Safar ME, management of renal failure in Europe, XXII, 1991. Nephrol
London GM. Wave re flections and cardiac hypertrophy in Dial Transplant 1992; 7[Suppl. 2]: 7 –35
chronic uremia. In fluence of body size. Hypertension 1993; 2. Lindner A, Charra B, Sherrard DJ, Scribner BH. Accelerated
22: 876 –883 atherosclerosis in prolonged maintenance hemodialysis. N Engl
25. London GM, Gue ´rin AP, Pannier B, Marchais SJ, Benetos A, JM e d 1974; 290: 697 –701
Safar ME. Increased systolic pressure in chronic uremia: Role 3. London GM, Gue ´rin AP, Marchais SJ et al. Cardiac and arterial
of arterial wave re flections. Hypertension 1992; 20: 10 –19 interactions in end-stage renal disease. Kidney Int 1996; 50:
600–608 26. Barenbrock M, Spieker C, Laske J et al . Studies of the vessel

Vascular calci fications and enhanced arterial sti ffness 1021
wall properties in hemodialysis patients. Kidney Int 1994; 45: 35. Raine AEG. Acquired aortic stenosis in dialysis patients.
Nephron 1994; 68: 159 –168 1397 –400
36. Goldsmith DJ, Covic A, Sambrook PA, Ackrill P. Vascular 27. Kawagishi T, Nishizawa Y, Konishi T et al . High-resolution
calci fication in long-term haemodialysis patients in a single unit: B-mode ultrasonography in evaluation of atherosclerosis in
A retrospective analysis. Nephron 1997; 77: 37 –43uremia. Kidney Int 1995; 48: 820 –82637. Nakagawa K. A study of aortic calci fication in uremia. Jpn28. Shoji T, Nishizawa Y, Kawagishi T et al . Intermediate-densityJ Nephrol 1997; 39: 135 –143lipoprotein as an independent factor for aortic atherosclerosis38. Arad Y, Spadaro LA, Roth M et al . Serum concentrations ofin hemodialysis patients. J Am Soc Nephrol 1998; 9: 1277 –1284calcium, 1,25 vitamin D and parathyroid hormone are not29. Vaziri ND, Gonzales EC, Wang J, Said S. Blood coagulation,correlated with coronary calci fications. An electron beam com-fibrinolytic, and inhibitory proteins in end-stage renal disease:puted tomography study. Coron Artery Dis 1998; 9: 513 –518Effect of hemodialysis. Am J Kidney Dis 1994; 28: 828 –83539. Vogt MT, San Valentin R, Forrest KY, Nevitt MC, Cauley JA.
30. Ernst E. Fibrinogen. Its emerging role as a cardiovascular risk Bone mineral density and aortic calci fication: The study of
factor. Angiology 1994; 45: 87 –93 osteoporotic fractures. J Am Geriatr Soc 1997; 45: 140 –145
31. Zimmermann J, Herrlinger S, Pruy A, Metzger T, Wanner C. 40. von der Recke P, Hansen MA, Hassager C. The association
between low bone mass at the menopause and cardiovascular Inflammation enhances cardiovascular risk and mortality in
mortality. Am J Med 1999; 106: 273 –278 hemodialysis patients. Kidney Int 1999; 55: 648 –658
41. Browner WS, Pressman AR, Nevitt MC et al . Association 32. Levenson J, Giral P, Megnien JL, Gariepy J, Plainfosse MC,
between low bone density and stroke in elderly women. TheSimon A. Fibrinogen and its relations to subclinical extracoron-study of osteoporotic fractures. Stroke 1993; 24: 940 –946ary and coronary atherosclerosis in hypercholesterolemic men.42. Uyama O, Yoshimoto Y, Yamamoto Y, Kawai A. Bone changesArterioscl Thromb Vasc Biol 1997; 17: 45 –50and atherosclerosis in postmenopausal women. Stroke 1997; 45:33. Barenbrock M, Hausberg M, Kosch M, Kisters K, Hoeks APG,1730 –1732Rahn K-H. E ffect of hyperparathyroidism on arterial distensibil-43. Shanahan CM, Cary NR, Metcalfe JC, Weissberg PL. Highity in renal transplant recipients. Kidney Int 1998; 54: 210 –215expression of genes for calci fication-regulating proteins in human
34. Detrano RC, Wong ND, Tang W et al . Prognostic signi ficance atherosclerotic plaques. J Clin Invest 1994; 93: 2393 –2402
of cardiac cine fluoroscopy for coronary calci fic deposits in 44. Wada T, McKee MD, Steitz S, Giachelli CM. Calci fication of
asymptomatic high-risk subjects. J Am Coll Cardiol 1994; 24: vascular smooth muscle cell cultures: Inhibition by osteopontin.
Circ Res 1999; 84: 166 –178 354–358
Received for publication: 20.8.99
Accepted in revised form: 5.3.00