NEPHROLOGY

2004;

9

, 130–141

et al.

Correspondence: Dr Sophia Zoungas, Department of Vascular Sci-ences and Medicine, Dandenong Hospital, David St, Dandenong 3175,Victoria, Australia. Email: sophia.zoungas@med.monash.edu.au

Accepted for publication 15 January 2004.

Original Article

Atherosclerosis and folic acid supplementation trial in chronic renal failure: Baseline results

SOPHIA ZOUNGAS,

1

PAULINE BRANLEY,

2

PETER G KERR,

3

SONYA RISTEVSKI,

2

CHRISTINE MUSKE,

2

LISA DEMOS,

2

ROBERT C ATKINS,

3

GAVIN BECKER,

4

MARGARET FRAENKEL,

5

BRIAN G HUTCHISON,

6

ROBERT WALKER,

7

JOHN J MCNEIL

2

and BARRY P MCGRATH

1

1

Department of Vascular Sciences and Medicine, Monash University, Dandenong Hospital, Dandenong, Victoria, Australia,

2

Department of Epidemiology and Preventative Medicine, Monash University, Alfred Hospital, Prahran, Victoria, Australia,

3

Department of Nephrology, Monash Medical Centre, Clayton, Victoria, Australia,

4

Department of Nephrology, The Royal Melbourne Hospital, Parkville, Victoria, Australia,

5

Department of Nephrology, Austin & Repatriation Medical Centre, Heidelberg, Victoria, Australia,

6

Department of Nephrology, Sir Charles Gairdner Hospital, Perth, Western Australia, Australia and

7

Department of Nephrology, Dunedin Hospital, Dunedin, New Zealand

SUMMARY:Background:

Atherosclerosis and Folic Acid Supplementation Trial (ASFAST) is a randomized pla-cebo controlled trial assessing whether high-dose folic acid can reduce cardiovascular events and ath-erosclerosis progression in patients with chronic renal failure (CRF). Here we report the baselineresults and compare indices of arterial structure (carotid intima-medial thickness (IMT)) and function(systemic arterial compliance (SAC)), pressure augmentation index (AI

x

) and pulse wave velocity(PWV a-f and PWV f-d)) to age- and sex-matched controls.

Methods:

Three hundred and fifteen subjects with CRF (serum creatinine

≥

0.40 mmol/L) aged 24–79 years (mean

±

SD: 56.6

±

13.6 years) and 213 healthy controls (58.2

±

10.2 years) were studied.Fasting blood samples were assayed for lipids (both groups), total homocysteine (tHcy), red cell folate,cobalamin and fibrinogen (CRF group). Ultrasound B mode measurements were used to determinemean carotid IMT and applanation tonometry techniques to determine SAC, AI

x

, PWV (a-f), PWV(f-d) and central pressures.

Results:

Ninety-six per cent of the CRF group had at least one of: hypertension, hypercholester-olaemia, diabetes or smoking; 35% had established cardiovascular disease. The mean IMT wasgreater in CRF patients than in controls (0.86

±

0.19

vs

0.68

±

0.11 mm,

P

<

0.001). The SAC was sig-nificantly lower, and PWV (a-f) and AI

x

significantly higher. The tHcy was increased in 97% of the CRFgroup (27.3

±

2.9

m

mol/L (normal

<

13)). Total homocysteine did not correlate with IMT or any othermeasure of arterial function. However, those in the upper quantile of tHcy (

≥

25

m

mol/L) did havehigher PWV (a-f) and lower SAC than those in the lower quantile.

Conclusions:

Compared to normals, patients with CRF exhibited a 10–15-year shift to the right inage-related increases in carotid IMT and PWV (a-f), and significantly increased central pressure aug-mentation. This 5-year study is examining the impact of high-dose folic acid therapy on cardiovascularend-points, IMT progression and arterial function in CRF.

KEY WORDS:

atherosclerosis, chronic renal failure, folic acid.

INTRODUCTION

Chronic renal failure (CRF) is a risk factor for cardiovas-cular disease, and is similar in impact to diabetes andsmoking.

1

More than half of all dialysis deaths are car-diovascular in nature.

2–4

This is presumed to be caused by

ASFAST baseline results

131

rapidly progressive atherosclerosis, although the evi-dence for this is not well established.

The high incidence of cardiovascular disease in CRFhas been ascribed to an excess of conventional risk fac-tors including hypertension, dyslipidaemia and diabetes.Hypertension is common but other factors including leftventricular hypertrophy, insulin resistance, hyperpar-athyroidism, disturbed autonomic function, anaemia,increased lipid peroxidation and reduced myocardialreserve may all contribute. Recently, a role for homocys-teine accumulation has also been implicated.

5–8

Homocysteine is a non-essential, sulfur-containingamino acid produced during the normal metabolism ofmethionine, an essential, sulfur-containing amino acidfound in abundant supply in the usual high-proteinWestern diet. Once formed it circulates in plasma largelybound to albumin, but 30% remains unbound of whichonly a small amount is in the reduced form. All circu-lating forms are measured by the total plasmahomocyst(e)ine automated immunoassay.

Homocysteine is metabolized by one of two metabolicpathways: remethylation and trans-sulfuration. Whenmethionine levels are low, homocysteine is remethylatedback to methionine by the enzyme methionine synthase.In this reaction, N

5

-methyl-tetrahydrofolate is themethyl donor and cobalamin is an essential cofactor. Adeficiency of either of these two substances can result inthe accumulation of homocysteine. When methioninelevels are abundant or cysteine synthesis is required,homocysteine is condensed with serine to form cys-tathionine using the vitamin B

6

-dependent enzyme, cys-tathionine

b

-synthase. Cystathionine is then hydrolysedto form cysteine. Mutations of the cystathionine

b

-synthase gene result in the rare congenital disorder,homocystinuria.

In the general population, mild hyperhomocystein-aemia is associated with occlusive vascular disease,including ischaemic heart disease, stroke and carotidvessel disease.

9–11

Others have linked folic acid status tolevels of homocysteine and cardiovascular disease.

6

InCRF, homocysteine levels are elevated in approxi-mately 85% of patients,

5

and are higher in those withcardiovascular events as compared to those withoutcardiovascular events.

7,8

In this population, homocys-teine accumulation is thought to be related to both analtered metabolism in the liver and renal tubules,

6

andreduced renal excretion caused by declining glomeru-lar filtration.

5

Folic acid administration has been investigated inrenal failure patients for its potential to reduce extremehomocysteine levels. Treatment of patients in dosesranging from 500

m

g to 60 mg daily, has consistently ledto reductions in homocysteine levels by 30–50%, with noextra effect seen with doses greater than 15 mg/day.

6,12–14

The effect of folic acid therapy on cardiovascular mor-bidity and mortality has yet to be reported by largeclinical trials.

The role of cobalamin and pyridoxine in loweringextreme homocysteine levels in renal failure patients isless clear. Deficiencies in CRF have been associated withelevated homocysteine levels, but prospective studieshave failed to show any consistent additional effect. Onestudy examined 15 mg of folic acid daily in addition topyridoxine and cobalamin and compared this to a con-trol group on 1 mg folic acid daily and smaller doses ofpyridoxine and cobalamin.

13

In this study, the high-dosefolic acid was more effective at lowering homocysteine,however, the effects of pyridoxine and cobalamin couldnot be differentiated. Others have studied pyridoxinesupplementation

12,15

and intramuscular cobalamininjection

16

and also shown no consistent effect.The Atherosclerosis and Folic Acid Supplementation

Trial (ASFAST) is an Australian/New Zealand ran-domized, double blind, placebo-controlled trial of high-dose folic acid therapy in patients with CRF examiningthe hypothesis that cardiovascular disease in CRF isrelated to hyperhomocysteinaemia and can be reducedby lowering homocysteine with folic acid therapy. Thestudy is powered to detect a significant differencebetween active and placebo-treated groups in cardiovas-cular events (a composite end-point of mortality andmorbidity from ischaemic heart disease and stroke). TheASFAST will also determine whether atheroma progres-sion, as measured by change in carotid artery intima-medial thickness (a well-established surrogate marker ofcardiovascular risk

17

), can be slowed by folic acid. Addi-tional secondary end- points are differences betweenthe groups in pulse wave velocity (PWV) (a-f), pressureaugmentation index (AI

x

), systemic arterial compliance(SAC) and homocysteine levels. Measurements ofPWV

18

and arterial wave reflection (AI

x

)

19

are predictiveof cardiovascular mortality in patients with chronic renalfailure.

This report gives the baseline results from ASFAST.The measurement of carotid intima-medial thickness(IMT) and arterial function in the CRF group have beencompared with those measured in a group of healthy,age- and sex-matched normotensive control subjects.

METHODS

Atherosclerosis and Folic Acid Supplementation Trial

The primary objective of ASFAST is to determine whether treatmentwith folic acid will reduce cardiovascular events in patients with CRF.The primary end-point being a composite of cardiovascular death,myocardial infarction and stroke. Secondary end-points will include achange in rate of progression of carotid artery intima-medial thickness,change in indices of arterial function, namely SAC, AI

x

and PWV anda difference between the treatment groups in terms of homocysteinelevel.

Approval by the Ethics Review Committees of Monash Universityand all participating hospitals was received and the study is being con-ducted in accordance with the National Health and Medical ResearchCouncil of Australia Guidelines on Human Experimentation.

132

S Zoungas

et al.

Protocol

Recruitment for ASFAST commenced on the 30 June 1998 and wasfinished on 31 December 2000. At closure, a total of 315 subjects wereentered from five dialysis units (four in Australia and one in NewZealand). Subjects were included on the basis that they were above18 years of age, had chronic renal failure of any cause, serum creati-nine of 0.40 mmol/L or greater (calculated creatinine clearance

<

25 mL/min) and awaiting commencement of dialysis or on continu-ous ambulatory peritoneal dialysis, intermittent peritoneal dialysis orhaemodialysis. Subjects were excluded on the basis of failure to obtainvalid informed consent, inability to comply with the study protocol,planned early living-related transplantation, the presence of a lifethreatening disease such as cancer, state of high cell turnover such asinflammatory bowel disease, ongoing treatment with methotrexate,phenytoin or trimethoprim-sulfamethoxazole, recent return to dialysisafter transplantation and still on immunosuppression, cobalamindeficiency without replacement, and previous bilateral carotid arterysurgery or carotid artery stenosis greater than 75% on screening.

Pre-randomization

Baseline consultation

This first consultation involved obtaining a medical history (comorbiddiseases and smoking status) and physical examination results includ-ing standardized blood pressure measurement (average of two readings,using an automated ‘Dinamap’ cuff; CRITIKON Compact TS, Johnson& Johnson, North Ryde, NSW, Australia); height, weight and waist-to-hip ratio measurements; a fasting blood sample for examination of lipidprofile, lipoprotein (a), homocysteine, zinc, cobalamin and red cell andserum folate levels; baseline carotid duplex scanning and arterial func-tion measurements. Each subject also completed a questionnaire thatassessed cardiovascular risk based on the National Heart Foundation ofAustralia Risk Factor Prevalence study.

20

History of prior cardiovasculardisease, defined as prior experience of angina, myocardial infarction orischaemic stroke, was then corroborated with a review of participants’hospital records. At study entry, those on folic acid or multivitaminscontaining folic acid were asked to withdraw this treatment. This wasjustified in view of the lack of randomised trials to that time showingany survival benefit from the use of folic acid supplementation. Sub-jects were then randomly assigned to treatment with folic acid 15 mgdaily or with an identical placebo tablet.

Post-randomization

Study consultations

Ongoing study consultations are occurring at yearly intervals to studycompletion or 5 years. Each study visit includes documentation ofchanges in therapy or medication in the case report form and repeatfasting blood sample collection for analysis of lipid profile, folate andcobalamin status and total plasma homocyst(e)ine (tHcy).

Every other 6 months, serum folic acid and cobalamin are also mea-sured as part of the routine follow up of the patients.

Follow up

All patients were contacted by telephone and at least on one occasionin person, at 1 and 4 weeks after randomization to treatment groups.

Subsequently, all are being contacted at 3 monthly intervals. Theobjective is to detect adverse events and to estimate patient compli-ance. If adverse events are found, a medical visit is organized with thetreating renal physician.

Safety

A safety committee is monitoring all adverse events in order to makedecisions regarding breaking the code or terminating the study, shouldthis be required.

Red cell folate levels below 510 nmoles/L on 6-monthly testing(reference interval: 510–1310 nmol/L) mandate folic acid supple-mentation at the smallest dose required to normalize the level.Such patients will continue on with the study according to theprotocol.

Compliance

Unused study medication is being returned to the hospital pharmacy toenable poor compliers to be identified. Patients with compliance tomedication usage estimated at less than 80% in the first month are con-tacted on a monthly basis, and every effort is made to obtain follow-updata. If harmful or distressing side-effects thought to be caused by folicacid are experienced, the medication is stopped and the patient is with-drawn from the study.

End-points

An end-point committee is adjudicating on all deaths and cardiovas-cular events occurring during the study.

Power calculation

The study sample size calculation was based on the rate of IMT pro-gression, and a power calculation was performed for the combined end-point of cardiovascular event or death caused by cardiovascular causesat 5 years, based on data obtained from the Australian and NewZealand Dialysis and Transplant Registry.

3

In the ‘MIDAS’ study,

21

the rate of IMT progression was 0.02 mmover 3 years with a standard deviation of 0.023. We expected a fasterrate in CRF and hence based the sample size calculation on a 0.03-mmIMT progression rate over 3 years, with a conservative standarddeviation (SD) estimate of 0.03. Estimating a 33% slower rate of pro-gression in the folic acid group, at a significance level (alpha) of 0.05(two-sided) with 80% power to detect a difference, the number of sub-jects required in each treatment arm was 143, that is, 286 subjects to befollowed over 3 years.

Based on an estimated cardiovascular event/death rate in the pla-cebo group of 1.5 events per 10 person-years of follow up, a mean fol-low-up duration of 3 years per participant (858 person-years of followup in total) and a two-sided significance level (alpha) of 0.05, thestudy of 286 participants split equally between groups has approxi-mately 91% power to detect a hazard ratio (relative risk) of 0.5, thatis, a reduction to a rate of 0.75 cardiovascular events per 10 person-years in the folic acid group. This calculation implies that the numberof cardiovascular events required in the study is 100.

22

Annual followup is now underway to provide the period of 858 person-years ofobservation.

ASFAST baseline results

133

Subjects

All 315 subjects recruited for ASFAST were included in the baselineassessment. Subjects with CRF were aged 24–84 years (mean

±

SD:56.6

±

13.6 years); 213 males and 102 females (Table 1). At entry, 190(60%) were on haemodialysis, 72 (23%) were on peritoneal dialysis and53 (17%) were predialysis. For those established on dialysis, the time ondialysis ranged from 1 to 392 months (mean: 3.3

±

4.9 years). For thosepredialysis, all were awaiting commencement of dialysis in the ensuing6 months. Main causes of CRF included glomerulonephritis (37%),diabetes mellitus (18%), polycystic kidney disease (11%), hypertension(10%) and reflux nephropathy (9%) (Table 2).

Examinations were undertaken in a quiet, air-conditioned roomwhile subjects were in the supine position, after a 10 min rest period.On the morning of study, subjects were advised to take all usual ther-

apy, which included antihypertensive agents in 210 subjects. For thoseon haemodialysis, measurements were performed on a non-dialysisday.

Assessment of artery wall thickness (common carotid intima-medial thickness (IMT)) was performed in all subjects. All subjectsattending two of the five study centres (

n

= 204) also underwentmeasurement of arterial function (SAC, AI

x

, PWV). The age andsex distribution of this group was similar to that of the entirecohort.

Two hundred and thirteen healthy control subjects, all normo-tensive, non-smokers were recruited from the general community andstudied over the same time period. These subjects were aged 23–76 years (mean

±

SD: 58.2

±

10.2 years) and consisted of 142 malesand 71 females. The controls were age- and sex-matched with theCRF patients to ensure a similar distribution in each 10-year agegroup. One hundred and sixty of these healthy controls also under-went a measurement of arterial function (SAC, AI

x

, PWV). Lipidlowering or antihypertensive therapy was not taken by any controlsubject.

Measurement of artery wall thickness

Imaging studies of the common carotid artery were performed by usinga high-resolution ultrasound machine (Diasonics DRF-400, Milwau-kee, WI, USA) and a hand-held transducer with 7.5 MHz central fre-quency (7.5-SPC mechanic sector transducer; Diasonics). A region1.0 cm proximal to the origin of the bulb of both the right and leftcommon carotid artery was identified by using B-mode ultrasonogra-phy. The transducer was manipulated such that the near wall of thecarotid artery was parallel to the transducer footprint and the lumenmaximized in the longitudinal plane. Three images of each B-mode,taken at three different angles (anterior, lateral, medial) were recordedand then digitized and saved electronically. Stored images were thenanalysed to determine artery wall intima-medial thickness (IMT) byusing a customized software program.

23

The mean of 60 IMT measure-ments from the far wall of both common carotid arteries was used inthe final data analysis to give mean IMT values. Maximum values forIMT were also obtained in each of the three planes of both commoncarotid arteries and averaged to give the mean, maximum IMT valuesfor each patient.

The same operator performed all image readings. The repeatabilityof this measurement was assessed in a small subgroup (

n

= 20) of studyparticipants who returned to the study centre on two separate occasions2 weeks apart without having changed any therapy. The correlation

Table 1

Demographic and clinical characteristics of athero-sclerosis and folic acid supplementation trial patients versuscontrols

CRF Control

P

Age (years)Male 57.1

±

13.4 59.2

±

9.9 NSFemale 55.4

±

14.2 56.4

±

10.6 NS

P

NS NSMale/Female 213/102 142/71 NSSystolic BP(mmHg)

Male 141

±

24 126

±

10

<

0.001Female 134

±

25 120

±

10

<

0.001

P

0.039

<

0.001Diastolic BP(mmHg)

Male 78

±

12 76

±

7 0.04Female 73

±

12 72

±

6 NS

P

0.001

<

0.001Pulse pressure(mmHg)

Male 62

±

18 50

±

8

<

0.001Female 61

± 20 49 ± 7 <0.001P NS NS

Heart rate (bpm)Male 70 ± 12 62 ± 9 <0.001Female 71 ± 14 62 ± 9 <0.001P NS NS

Height (m)Male 1.72 ± 0.08 1.75 ± 0.07 <0.001Female 1.57 ± 0.07 1.63 ± 0.07 <0.001P <0.001 <0.001

Weight (kg)Male 78 ± 15 79 ± 12 NSFemale 66 ± 14 70 ± 13 NSP <0.001 <0.001

BMI (kg/m2)Male 27 ± 4 26 ± 3 0.026Female 27 ± 6 26 ± 5 NSP NS NS

Results are presented as mean ± SD. BMI, body mass index; BP,blood pressure; CRF, chronic renal failure. P values compare CRF sub-jects with controls and males with females. NS, not significant atP > 0.05.

Table 2 Comparison of primary renal disease of Atherosclero-sis and Folic Acid Supplementation Trial (ASFAST) patientsversus the Australian and New Zealand Dialysis and TransplantRegistry (ANZDATA) population

ASFAST ANZDATA (1995)

Glomerulonephritis 37 35Diabetic nephropathy 18 20Polycystic kidney disease 11 8Hypertension 10 8Reflux nephropathy 9 5Analgesic nephropathy 1 7Miscellaneous 10 11Uncertain diagnosis 4 6

Values are a percentage of the totals (ASFAST n = 315; ANZDATAAustralian population n = 1358).

134 S Zoungas et al.

coefficient for IMT between the two visits was 0.89 and the intra-observer coefficient of variation 10.43%.

Measurement of arterial compliance and stiffness

The total systemic arterial compliance (SAC) was estimated byusing the ‘area method’, which requires measurement of volumetricblood flow and associated driving pressure to derive an estimatedcompliance over the total systemic arterial tree. A hand-held Dop-pler flow velocimeter (Multidoplex MD1, Huntleigh Technology,Cardiff, UK) placed on the suprasternal notch at the base of theneck was used to estimate arterial blood flow. Aortic driving pres-sure was estimated by applanation tonometry of the right commoncarotid artery using a non-invasive pressure transducer (MillarMikro-tip; Millar Instruments Houston, Houston, TX, USA). Dop-pler flow and pressure traces were recorded continually for 1 minand analysed over 10 cardiac cycles using purpose written software(Microsoft visual C++, Department of Vascular Sciences and Medi-cine, Monash University, Dandenong, VIC, Australia). Blood pres-sures, derived from carotid waveforms, were calibrated againstdiastolic and mean brachial artery blood pressure measurements fromthe left arm. Systemic arterial compliance was then calculatedaccording to the formula:

SAC = Ad/R(Pes - Pd)

where R is the total peripheral resistance calculated as mean arterialblood pressure/mean blood flow; Ad is the area under the diastolic por-tion of the pulse pressure contour; Pes is the end-systolic aortic bloodpressure; and Pd is the end-diastolic arterial blood pressure.

The same carotid blood pressure waveforms used for the determi-nation of SAC were also analysed to identify the ‘shoulder’ and ‘peak’of the waves. The carotid artery augmentation index (AIx) was thencalculated as the ratio of the pressure difference between the shoulderof the wave and peak systolic pressure (DP) and the pulse pressure (PP)according to the formula:

AIx = (DP/PP) ¥ 100.

The augmentation point (Pi) is reliably identified mathematically asthe first zero-crossing from positive to negative, after the beginning ofsystole, of the third derivative of the pressure waveform. A value for AIx

is defined as positive or negative depending on whether Pi occurs beforeor after the peak pressure.

Pulse wave velocity (PWV) measurements were obtained forboth the aorto-femoral (PWV (a-f)) and femoral-dorsalis pedis arte-rial segments (PWV (f-d)). Continuous pulse pressure wave signalswere recorded with two tonometers (Millar Mikro-tip, SPT-301;Houston, TX, USA) positioned at both the base of the right com-mon carotid artery and over the femoral artery, or over the femoralartery and the ipsilateral dorsalis pedis artery. Distances from thecarotid sampling sites to the manubrium sternum, manubrium ster-num to femoral artery, and femoral artery to dorsalis pedis were mea-sured as straight lines between these points on the body surface witha tape measure. The low point of the pulse wave was defined as thepoint of intersection of the end of diastole and beginning of systole,identified from the waveform analysis as the maximum of the firstderivative at the pressure signal. The mean transit time (Dt) betweenthe feet of simultaneously recorded waves was determined from 10consecutive cardiac cycles. The PWV was calculated from the dis-tance between measurement points and the measured time delay (Dt)as follows:

PWV = D/Dt(m/s)

where D is distance in metres and Dt is the time interval in seconds.

Determination of blood pressures

Brachial blood pressure recordings were taken at 5-min intervalsthroughout the imaging period by using a Dinamap device (CRI-TIKON Compact TS, Johnson & Johnson, North Ryde, NSW, Aus-tralia). The first blood pressure recording was excluded and allsubsequent measurements were averaged. Central blood pressure wasmeasured by applanation tonometry of the right common carotid arteryusing a non-invasive pressure transducer (Millar Mikro-tip; MillarInstruments Houston). Carotid artery pressure waveforms obtainedsimultaneously with the brachial artery Dinamap pressure recordingswere analysed to obtain central pulse pressure (PPcen) measurements.

Measurement of total plasma homocyst(e)ine

Venous blood taken after an 8-h fast was analysed for total plasmahomocyst(e)ine (tHcy) by using an ABBOTT Fluorescence Polariza-tion immunoassay (Abbott Laboratories, Abbott Park, IL, USA) on anABBOTT IMx Analyser. This assay is comparable to the high perfor-mance liquid chromatography (HPLC) method in its precision of mea-surement of both normal and high tHcy levels.24 The between runcoefficient of variation for this assay was 4.3% at a level of 12.5 mmol/L.

Data analysis

Statistical analysis was performed by using Microsoft Excel 97(Microsoft Corp., Redmond, WA, USA) and SPSS version 10 for Win-dows program (SPSS Inc., Chicago, IL, USA). Data are reported asmean ± standard deviation with P < 0.05 considered to be statisticallysignificant. The Student’s unpaired t-test and Chi-squared test wereused to determine the significance of the differences between groups forcontinuous and categorical variables, respectively. Relationshipsbetween vascular indices were examined by using univariate Pearsoncorrelation. Multiple regression analysis was used to examine determi-nants of IMT, SAC, AIx and PWV by including the variables associatedwith these that had a P-value <0.10 in the univariate analysis.

RESULTS

Demographic and clinical characteristics

Table 1 shows the demographic and clinical characteris-tics in CRF patients and normal controls. The groupswere similar in age, sex distribution, weight and bodymass index.

Of the CRF patients, 35% reported a history ofcardiovascular disease (angina (32%), acute myocardialinfarction (11%) and/or cerebrovascular disease (9%)).Ninety per cent reported hypertension, 23% diabetesmellitus and 44% hypercholesterolaemia. Two or more ofthese risk factors were observed in 50% of the CRFpatients. Fifteen, 67 and 25% of patients had drug treat-ment for angina, blood pressure and cholesterol, respec-tively. The commonest class of antihypertensive drugused was calcium antagonists (39%), followed by bblockers (26%), angiotensin-converting enzyme inhibi-tors (21%) and diuretics (21%). Twenty-one per cent ofpredialysis subjects, and 46 and 22% of haemodialysis

ASFAST baseline results 135

Table 4 Biochemical data of Atherosclerosis and Folic AcidSupplementation Trial (ASFAST) patients versus controls

CRF Control P

Total cholesterol(mmol/L)

Male 5.0 ± 1.2 5.8 ± 0.9 <0.001Female 5.7 ± 1.2 5.8 ± 1.1 NSP <0.001 NS

Triglycerides(mmol/L)

Male 2.0 ± 1.2 1.30 ± 0.7 <0.001Female 2.2 ± 1.4 1.0 ± 0.6 <0.001P NS 0.02

HDL Cholesterol(mmol/L)

Male 1.0 ± 0.4 1.3 ± 0.4 <0.001Female 1.3 ± 0.4 1.7 ± 0.5 <0.001P <0.001 <0.001

LDL Cholesterol(mmol/L)

Male 3.2 ± 1.0 3.9 ± 0.9 <0.001Female 3.4 ± 0.9 3.7 ± 1.0 NSP 0.044 NS

LDL : HDL ratioMale 3.4 ± 1.5 3.2 ± 1.3 NSFemale 2.8 ± 1.0 2.3 ± 1.1 0.023P <0.001 <0.001

Lipoprotein a(mg/L)

Male 525 ± 580 260 ± 305 <0.001Female 644 ± 681 436 ± 500 0.032P NS 0.021

Total homocysteine(umol/L;RR = 3–13 mmol/L)

Male 27.6 ± 13.8Female 26.8 ± 11.0P NS

Red cell folate(nmol/L;RR = 510–1310)

Male 1519 ± 991Female 1389 ± 905P NS

Cobalamin(pmol/L;RR = 155–670 pmol/L)

Male 446 ± 233Female 404 ± 203P NS

Fibrinogen(g/L;RR = 1.5–4.0 g/L)

Male 3.9 ± 1.4Female 3.6 ± 1.3P NS

Results are presented as mean ± SD. CRF, chronic renal failure;HDL, high density lipoprotein; LDL, low density lipoprotein; RR, ref-erence range. P-values compare CRF subjects with controls and maleswith females. NS, not significant at P > 0.05.

and peritoneal dialysis-dependent subjects, respectively,took folic acid prior to study commencement (mediandose 5 mg, range 1–35 mg weekly).

With regards to smoking, 38% were life-long non-smokers, 46% were former smokers and 16% were cur-rent smokers. In the control group, 25% were formersmokers and none were current smokers. Of those CRFpatients that smoked, the average length of use was 29pack years (total cigarettes smoked per day divided by 20;range 0.1–160 pack years). On the basis of body massindex, 21% of the CRF patients were considered over-weight (BMI ≥ 30) as compared to 13% of the controls.

In Table 2, the distribution of renal disease diagnosticgroups in the ASFAST study group is compared to thedistribution seen in the Australian and New ZealandDialysis Registry.3 These were very similar. Table 3reports the distribution of cardiovascular risk factors,which again is similar to that reported by the Australianand New Zealand Dialysis Registry.3

Blood biochemistry

Biochemical data is given in Table 4. Total homocysteinewas increased in 97% of the CRF population((normal < 13 mmol/L), mean ± SD: 27.3 ± 12.9 mmol/L;median 24.8 mmol/L; range 4.5–134.4 mmol/L) andcorrelated inversely with red cell folate (r = -0.30,P < 0.001) and cobalamin (r = -0.15, P = 0.009) levels.Total homocysteine was similarly elevated in those onhaemodialysis as compared with those on peritonealdialysis (mean ± SD, 27.2 ± 11.4 vs 26.1 ± 11.0 mmol/L,respectively). Folate deficiency, defined as a red cellfolate less than 510 mmol/L, was identified in 6.4% ofthe population. For those CRF patients taking folic acidat entry to the study, total homocysteine was lower thanin those not taking supplementation (24.0 ± 8.4 vs29.2 ± 14.6 mmol/L, P < 0.001). A significantly greaterproportion of patients on haemodialysis were taking folic

Table 3 Prevalence of cardiovascular risk factors in Athero-sclerosis and Folic Acid Supplementation Trial (ASFAST)patients versus the Australian and New Zealand Dialysis andTransplant Registry (ANZDATA) population

ASFAST(n = 315)

ANZDATA (1995)(n = 1358)

Former smokers 46 50Current smokers 16 12Non-smokers 38 38Diabetes 23 24Ischaemic heart disease 33 36Hypertension 90 82Hypercholesterolaemia 44 NA

Values are a percentage of the total (ASFAST n = 315; ANZDATAAustralian population n = 1358). NA, not available.

136 S Zoungas et al.

acid than patients on peritoneal dialysis (46 vs 22%,respectively; P < 0.001).

Dyslipidaemia, defined as total cholesterol greaterthan 6.0 mmol/L, HDL cholesterol less than 1.0 mmol/Land/or triglycerides greater than 3.0 mmol/L, or treat-ment with lipid lowering agents was identified in 36% ofthe renal subjects. Those peritoneal dialysis had highertotal cholesterol, LDL cholesterol, triglycerides and lipo-protein (a) levels than those predialysis or on haemodi-alysis. When compared with controls not on any lipidlowering therapy, CRF patients had lower totalcholesterol and LDL cholesterol but also lower HDL cho-lesterol and higher triglyceride and lipoprotein (a) levels.

Blood pressure measurements

Blood pressure results are shown in Tables 1 and 5.Comparing CRF patients with controls, systolic blood

pressure, diastolic blood pressure and pulse pressure, allwere significantly higher in the CRF patients. In theCRF population, central systolic and pulse pressures werehighly correlated with corresponding supine brachialblood pressures (systolic BP (r = 0.94, P < 0.001); pulsepressure (r = 0.90, P < 0.001)).

Arterial indices

As shown in Table 5, the mean carotid wall IMT wassignificantly greater in CRF patients than controls(0.86 ± 0.19 vs 0.68 ± 0.11 mm, P < 0.001). The mean,maximum IMT was 1.07 ± 0.25 mm. In the CRF group,pulse wave velocity (a-f) and pressure augmentationindex were significantly higher (heart rate-adjustedPWV (a-f): 10.6 ± 3.5 vs 9.2 ± 1.8 m/s, P < 0.001; heartrate-adjusted AIx: 23.3 ± 11.1 vs 14.3 ± 9.9%, P < 0.001)and systemic arterial compliance was significantly lowerthan that for controls (0.44 ± 0.22 vs 0.59 ± 0.21 units,P < 0.001).

At any age, both the IMT and PWV (a-f) were greaterin CRF patients than in controls (Figs 1,2). The differ-

Table 5 Vascular indices of Atherosclerosis and Folic AcidSupplementation Trial patients versus controls

CRF Control P

Carotid IMT (mm)Male 0.86 ± 0.18 0.68 ± 0.12 <0.001Female 0.84 ± 0.20 0.67 ± 0.09 <0.001P NS NS

Systemic arterialcompliance (units)

Male 0.46 ± 0.24 0.60 ± 0.21 <0.001Female 0.40 ± 0.18 0.59 ± 0.22 <0.001P 0.033 NS

PWV (a-f) (m/s)Male 11.0 ± 3.7 9.4 ± 1.8 <0.001Female 10.2 ± 3.1 8.6 ± 1.7 <0.001P NS 0.004

PWV (f-d) (m/s)Male 10.5 ± 1.8 10.8 ± 1.5 NSFemale 10.2 ± 2.0 10.2 ± 1.6 NSP NS 0.013

Augmentationindex (%)

Male 21.4 ± 11.6 12.0 ± 9.3 <0.001Female 25.2 ± 11.5 20.4 ± 9.9 0.014P 0.028 <0.001

Central SBP(mmHg)

Male 126 ± 23 117 ± 12 <0.001Female 124 ± 23 110 ± 11 <0.001P NS < 0.001

Central PP(mmHg)

Male 48 ± 18 41 ± 10 <0.001Female 48 ± 17 39 ± 9 <0.001P NS NS

Results are presented as mean ± SD. CRF, chronic renal failure; IMT,intima-medial thickness; PP, pulse pressure; PWV, pulse wave velocity;SBP, systolic blood pressure. P values compare CRF subjects withcontrols and males with females. NS, not significant at P > 0.05.

Fig. 1 Carotid artery intima-medial thickness (IMT) by agegroup in Atherosclerosis and Folic Acid Supplementation Trial(ASFAST) patients versus controls. Data are means for eachage group. P < 0.05 for comparison at each age group. (�)Control, (�) chronic renal failure (CRF).

<40 40–49 50–59 60–69 >700.00

0.02

0.04

0.06

0.08

0.10

Car

otid

IM

T (

cm)

Age (years)

Fig. 2 Pulse wave velocity (PWV (a-f)) by age group in Ath-erosclerosis and Folic Acid Supplementation Trial (ASFAST)patients versus controls. Data are means for each age group.P < 0.05 for comparison at each age group. (�) Control, (�)chronic renal failure (CRF).

<40 40–49 50–59 60–69 >70Age (years)

0

5

10

15

PWV

af

(m/s

ec)

ASFAST baseline results 137

ences between the groups were equivalent to a 10–15-year shift to the right in age-related increases in IMT andPWV in the CRF group.

After adjustment for age, the presence of diabetes or apast history of cardiovascular disease was associated withgreater PWV(a-f) but not with IMT, SAC or centralpulse pressure.

Comparison by cause of renal failure is shown inTable 6. The diagnosis of polycystic kidney disease wasassociated with lower IMT and PWV(a-f), while thediagnosis of diabetic end-stage renal disease was associ-ated with greater IMT and PWV (a-f).

Dialysis, per se, and current mode of dialysis did notappear to influence any of the vascular parameters; IMT,PWV, AIx and SAC were similar in those established ondialysis as compared to those not on dialysis, and similarto those on haemodialysis as compared with peritonealdialysis (Fig. 3). However, the duration of dialysis didshow a significant positive correlation with AIx (r = 0.28;P < 0.001).

As shown in Table 7, all arterial indices were related toage except AIx (IMT vs age (r = 0.44, P < 0.001); PWV(a-f) vs age (r = 0.46, P < 0.001); PWV (f-d) vs age(r = 0.17, P = 0.032); SAC vs age (r = -0.22, P = 0.015)).

Fig. 3 Indices of arterial structure andfunction by mode of dialysis. Data areage-adjusted means ± SD (except forpressure augmentation index whereunadjusted data is given). HD, haemodi-alysis; PD, peritoneal dialysis. (a)Carotid intima-medial thickness (IMT)by mode of dialysis; (b) systemic arterialstructure (SAC) by mode of dialysis; (c)pressure augmentation index (AIx) bymode of dialysis; (d) pulse wave velocity(PWV) by mode of dialysis.

0.00

0.20

0.40

0.60

0.80

1.00

1.

(a)

20

Age

-adj

uste

d m

ean

IMT

(cm

)

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.

(b)

80

Age

-adj

uste

d S

AC

(m

L/m

mH

g)

0

5

10

15

20

25

30

35

40

(c)

AIx

(%

)

Predialysis HD PD0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

(d)

Age

-adj

uste

d PW

V (a

f) (

m/s

ec)

Predialysis HD PD

Table 6 Vascular indices by cause of renal failure

All other CRF (n = 222) DM (n = 57) PCKD (n = 36) P

Age (years) 56.3 ± 14.3 57.9 ± 11.0 56.1 ± 12.6 NSCarotid IMT (mm)

Mean 0.86 ± 0.20 (217) 0.89 ± 0.16 (54) 0.80 ± 0.13 (35) NS (0.086)Mean, maximum 1.08 ± 0.27 (217) 1.10 ± 0.20 (54) 0.99 ± 0.19 (35) NS (0.082)

Systemic arterial compliance (units) 0.44 ± 0.22 (139) 0.41 ± 0.22 (41) 0.50 ± 0.28 (23) NSPWV (a-f) (m/s) 9.97 ± 2.94 (138) 13.74 ± 4.32 (42) 9.73 ± 1.48 (22) <0.001PWV (f-d) (m/s) 10.26 ± 1.82 (123) 10.59 ± 1.90 (31) 10.52 ± 1.93 (21) NSAugmentation index (%) 24.6 ± 11.6 (138) 16.4 ± 10.0 (41) 22.5 ± 11.4 (23) <0.001Heart rate (bpm) 70 ± 13 72 ± 12 68 ± 10 NSCentral PP (mmHg) 46 ± 16 (139) 59 ± 20 (41) 44 ± 13 (23) <0.001

Results are presented as mean ± SD. Data in parentheses indicates number of subjects. CRF, chronic renal failure; DM, diabetes mellitus;IMT, intima-medial thickness; PCKD, polycystic kidney disease; PP, pulse pressure; PWV, pulse wave velocity. P value for between group differences(one-way ANOVA). NS, not significant at P > 0.05.

138 S Zoungas et al.

By using multiple regression analysis, the main deter-minants of IMT were age and PWV (a-f) (r2 = 0.146,P < 0.001). For systemic arterial compliance, the maindeterminants were age, mean arterial pressure, heightand red cell folate (r2 = 0.388, P < 0.001). For centralpressure augmentation, the main determinants wereheart rate, duration of dialysis, mean arterial pressure andheight (r2 = 0.286, P < 0.001). For aortofemoral pulse-wave velocity, the main determinants were age, meanarterial pressure and fibrinogen (r2 = 0.405, P < 0.001).

After adjustment for age, there was no significant dif-ference between the sexes in terms of IMT and PWV,however, SAC was significantly lower in women withCRF than in men (0.39 ± 0.17 vs 0.46 ± 0.24 units,P = 0.012). The AIx was also significantly greater inwomen with CRF than in men (25.2 ± 11.5 vs21.4 ± 11.6%, P = 0.028). This was explained by theheight difference between the genders as the differenceswere no longer apparent after adjustment for height.

By using univariate analysis, no correlation was foundbetween homocysteine and IMT or any other measure ofartery function in CRF. A comparison of patients in theupper quantile (homocysteine levels ≥ 25 mmol/L) tothose in the lower quantile of homocysteine (homocys-teine levels < 25 mmol/L) found SAC to be significantlylower (0.40 ± 0.19 vs 0.47 ± 0.25 units, P = 0.021) andPWV (a-f) to be higher (11.2 ± 3.4 vs 10.2 ± 3.6 m/s,P = 0.062).

DISCUSSION

Patients with CRF have a high prevalence of cardiovas-cular risk factors and this is shown in the ASFAST studygroup. At baseline, 35% had had a prior cardiovascularevent and 96% had at least one of the following risk fac-tors: hypertension (90%), hypercholesterolaemia (44%),diabetes mellitus (23%) or were smokers (46% former,16% current). With regard to cardiovascular risk factors,the CRF group was similar to the total group of patientswith CRF on the Australian and New Zealand Dialysisand Transplant Registry.3 This Registry has recentlyreported dialysis-dependent mortality rates of 15.7 deathsper 100 patient years at risk, of which 46% were accounted

for by cardiac causes.25 This represents 12% of patientstreated at any time during the preceding year. In contrast,mortality rates reported for renal transplant recipients aremuch lower (3.2 deaths per 100 patient years at risk). Ifwe were to assume similar mortality rates for this CRFpopulation, then we would forecast 19 cardiovasculardeaths in the first year, 16 in the second year and 15 in thethird year of follow up. Considering that the ASFASTpopulation is highly representative of the wider Austra-lian CRF population, we believe that the present study isadequately powered to meet its projected accumulationof 100 primary end-points by 5 years.

Changes in arterial structure and function in CRF

Compared to an age- and sex-matched control group ofnormotensive non-smokers, the CRF patients exhibiteda 10–15-year shift to the right in the age-relatedincreases in carotid IMT and PWV (a-f). Both of theseindices have been shown to be predictive markers offuture cardiovascular disease.17,18 In addition, the CRFgroup had increased mean values for heart rate, centralpulse pressure (PP) and AIx, and reduced SAC comparedwith controls. There is compelling evidence thatincreased PP is associated with increased cardiovascularrisk in the general population.26–28 The increases in cen-tral PP, AIx and reduced SAC in the CRF group are con-sistent with increased arterial stiffness and arterial wavereflection, as has been reported previously.29–32

The present study has permitted an evaluation of thekey determinants of carotid IMT and arterial functionindices in CRF. After age adjustment, PWV (a-f) wasgreater in diabetics and in patients with a history of priorcardiovascular events but lower in patients with polycys-tic renal disease, and was the only vascular marker toshow such a discrimination. Our results support theoutcome studies of Blacher et al.,18 which identifiedpulse-wave velocity as a valuable vascular marker for car-diovascular outcome prediction.

Interestingly, red cell folate was a significant determi-nant of systemic arterial compliance. This may or maynot be because of its relationship to homocysteine, inview of the recent finding that folate improved endothe-

Table 7 Interrelationships of age, intima-medial thickness (IMT) and indices of arterial function in chronic renal failure (CRF)

Age IMT SAC AIx PWV (a-f) PWV (f-d)

Age 1.00 0.44** -0.22** 0.02 (NS) 0.46** 0.17*IMT 1.00 -0.12 (NS) 0.07 (NS) 0.21** -0.14 (NS)SAC 1.00 -0.31** -0.40** -0.04 (NS)AIx 1.00 -0.09 (NS) -0.04 (NS)PWV (a-f) 1.00 0.04 (NS)PWV (f-d) 1.00

Results are presented as correlation coefficients. AIx, pressure augmentation index; IMT, carotid intima-medial thickness; SAC, systemic arterialcompliance. PWV, pulse wave velocity. **P < 0.001; *P < 0.05; NS, not significant.

ASFAST baseline results 139

lial function by increasing nitric oxide bioavailability intype II diabetic patients.33 Whether this is the caseremains to be corroborated by further studies.

Height and heart rate were important influences onAIx, as has been reported in normal subjects.34–37 Arterialwave reflection was also correlated with the duration ofdialysis. None of the vascular indices were influenced bythe mode of dialysis.

Homocysteine in chronic renal failure

Plasma homocysteine levels were increased two to three-fold in patients with CRF, even in patients alreadyreceiving small doses of folic acid therapy. This confirmsthe results of previous studies.6,38,39 Plasma homocysteinewas inversely correlated with red cell folate and serumB12 levels.

A reduction of plasma homocysteine levels in CRFhas been consistently achieved by using high-dose folicacid therapy.12,13,40 Homocysteine levels are lowered butoften still above the normal range. It has been suggestedthat this may be related to the mode of dialysis5 or unrec-ognized vitamin B6 deficiency, which is often seen inCRF patients.41 In this large cohort of CRF patients,there were no significant differences in mean totalhomocysteine, red cell folate or cobalamin levelsbetween the haemodialysis and peritoneal dialysisgroups.

In healthy subjects, modestly elevated homocysteinelevels have been related to carotid IMT measure-ments.42,43 In the ARIC study, subjects with intima-medial thickening had significantly higher homocysteinelevels than those without.44 In CRF, higher carotid IMTmeasurements and homocysteine levels have been as-sociated with the methylenetetrahydrate reductase(MTHFR) gene polymorphism, an enzyme critical inproviding the methyl group donor (methyltetrahydro-folate) for the remethylation of homocysteine to S-adenosyl methionine (TT vs CC genotype, 0.93 ± 0.07vs 0.79 ± 0.13 mm).45 In the present study, there was nocorrelation between homocysteine levels and carotidIMT. However, those in the upper quantile of plasmahomocysteine levels (≥25 mmol/L) had higher meanPWV (a-f) and significantly lower mean SAC than thosein the lower quantile.

Epidemiological evidence is strong for a central linkbetween hyperhomocysteinaemia and premature cardio-vascular disease and death.46–51 There is also a strongassociation between hyperhomocysteinaemia and venousthrombosis.52,53 To date, randomized controlled trials ofhomocysteine-lowering therapy in non-renal popula-tions have reported improvement in abnormal exerciseelectrocardiographic testing54 and decreased rates ofrestenosis55 and lower adverse event rates after percuta-neous coronary intervention.56 These studies are, how-ever, limited by their small numbers and short periods of

follow up. No trials have reported cardiovascular out-come with homocysteine lowering in CRF populations.Using folic acid supplementation for cardiovascularprotection in CRF patients is currently widespread butnot evidence-based. The ASFAST, with its largenumbers and 5-year period of observation, is adequatelypowered to answer the question of cardiovascular out-come in patients with CRF after high-dose folic acidsupplementation.

CONCLUSION

The present study shows that patients with CRF enteredinto the ASFAST study are representative of the popu-lation on the Australian and New Zealand Dialysis reg-istry, which has a high prevalence of cardiovascular riskfactors. Compared to an age- and sex-matched group ofnormotensive control subjects, ASFAST participantshave a 10–15-year increase in age-related changes incarotid IMT and PWV (a-f), a reduction in SAC and asignificant increase in central wave reflection. Vascularchanges were more marked in patients with diabetes andless in patients with polycystic kidney disease comparedwith other causes of CRF. The ASFAST is adequatelypowered to determine whether lowering homocysteinewith folic acid therapy in CRF patients will improve car-diovascular outcome.

STUDY ORGANIZATION

The study steering committee consisting of Professors BPMcGrath, JJ McNeil, RC Atkins, PG Kerr, G Becker,and R Walker, and Drs S Zoungas, P Branley, M Fraen-kel, and BG Hutchison is overviewing the overallprogress and direction of the study.

ACKNOWLEDGEMENTS

The study is supported by a grant-in-aid from theNational Health and Medical Research Council ofAustralia and National Heart Foundation of NewZealand.

Thanks to all the patients and staff of the participat-ing renal units for their ongoing support and enthusiasm:Monash Medical Centre, Clayton, Victoria, Australia;Royal Melbourne Hospital, Parkville, Victoria, Austra-lia; Austin and Repatriation Medical Centre, Heidel-berg, Victoria, Australia; Sir Charles Gardiner Hospital,Perth, Australia; and Dunedin Hospital, Dunedin, NewZealand.

Thanks also to Dr Rory Wolfe from the Departmentof Epidemiology and Preventative Medicine, MonashUniversity for his statistical support in the derivation ofthe study power.

140 S Zoungas et al.

REFERENCES

1. Ritz E. Atherosclerosis and cardiac death: Are they related to dial-ysis procedure and biocompatibility? Nephrol. Dial. Transplant.1994; 9 (Suppl. 2): 165–72.

2. Lindner ACB, Sherrard DJ, Scribner BH. Accelerated atheroscle-rosis in prolonged maintenance hemodialysis. N. Engl. J. Med.1974; 290: 697–701.

3. Disney APS (ed). ANZDATA. Registry report 1996. Adelaide:Australia and New Zealand Dialysis and Transplant Registry,1996.

4. Wheeler DC. Cardiovascular disease in patients with chronicrenal failure. Lancet 1996; 348: 1673–4.

5. Robinson K, Gupta A, Dennis V et al. Hyperhomocysteinemiaconfers an independent increased risk of atherosclerosis in end-stage renal disease and is closely linked to plasma folate andpyridoxine concentrations. Circulation 1996; 94: 2743–8.

6. Dennis VW, Robinson K. Homocysteinemia and vascular diseasein end-stage renal disease. Kidney Int. 1996; 50 (Suppl. 57): 11–17.

7. Jungers P, Massy ZA, Khoa TN et al. Incidence and risk factors ofatherosclerotic cardiovascular accidents in predialysis chronicrenal failure patients: A prospective study. Nephrol. Dial. Trans-plant. 1997; 12: 2597–602.

8. Moustapha A, Naso A, Nahlawi M et al. Prospective study ofhyperhomocysteinemia as an adverse cardiovascular risk factor inend-stage renal disease. Circulation 1998; 97: 138–41.

9. Stampfer MJ, Malinow MR, Willett WC et al. A prospective studyof plasma homocyst(e)ine and risk of myocardial infarction in USphysicians. JAMA 1992; 268: 877–81.

10. Verhoef P, Hennekens CH, Malinow MR, Kok FJ, Willett WC,Stampfer MJ. A prospective study of plasma homocyst(e)ine andrisk of ischemic stroke. Stroke 1994; 25: 1924–30.

11. Petri M, Roubenoff R, Dallal GE et al. Plasma homocysteine as arisk factor for atherothrombotic events in systemic lupus erythe-matosus. Lancet 1996; 348: 1120–4.

12. Arnadottir M, Brattstrom L, Simonsen O et al. The effect of high-dose pyridoxine and folic acid supplementation on serum lipid andplasma homocysteine concentrations in dialysis patients. Clin.Neph. 1993; 40: 236–40.

13. Bostom AG, Shemin D, Lapane KL et al. High dose B-vitamintreatment of hyperhomocysteinemia in dialysis patients. KidneyInt. 1996; 49: 147–52.

14. Wilcken DEL, Dudman NPB, Tyrell PA, Robertson MA. Folicacid lowers elevated plasma homocysteine in chronic renal insuf-ficiency possible implications for prevention of vascular disease.Metabolism 1988; 37: 697–701.

15. Chauveau P, Chadefax B, Coude M et al. Long-term folic acid (butnot pyridoxine) supplementation lowers elevated plasmahomocysteine level in chronic renal failure. Miner. ElectrolyteMetab. 1996; 22: 106–9.

16. Polkinghorne K, Zoungas S, Branley P et al. A randomised placebocontrolled trial of intramuscular vitamin B12 for the treatment ofhyperhomocysteinaemia in dialysis patients. Int. Med. J. 2003; 33:489–94.

17. O’Leary DH, Polak JF, Kronmal RA, Manolio TA, Burke GL,Wolfson SK. Carotid-artery intima and media thickness as a riskfactor for myocardial infarction and stroke in older adults. N. Engl.J. Med. 1999; 340: 14–22.

18. Blacher J, Guerin AP, Pannier B, Marchais SJ, Safar ME, LondonGM. Impact of aortic stiffness on survival in end-stage renaldisease. Circulation 1999; 99: 2434–9.

19. London GM, Blacher J, Pannier B et al. Arterial wave reflectionsand survival in end-stage renal failure. Hypertension 2001; 38:434–8.

20. Hodge RL. Risk factors in Australians: National Heart Founda-tion’s Risk Factor Prevalence Study, 1980. Aust. N. Z. J. Med.1984; 14: 395–9.

21. Borhani NO, Mecuri M, Borhani PA et al. Final outcome resultsof the multicenter israpidine diuretic atherosclerosis study(MIDAS). JAMA 1996; 276: 785–91.

22. Freedman L. Tables of the number of patients required in clinicaltrials using the logrank test. Stat. Med. 1982; 1: 121–9.

23. Gamble G, Zorn J, Sanders G et al. Estimation of arterial stiffness,compliance and distensibility from M mode ultrasound measure-ments of the common carotid artery. Stroke 1994; 25: 11–16.

24. Tcherkas YV, Denisenko AD. Simultaneous determination of sev-eral amino acids, including homocysteine, cysteine and glutamicacid, in human plasma by isocratic reversed-phase high-perfor-mance liquid chromatography with fluorimetric detection. J.Chromatogr. 2001; 913: 309–13.

25. Russ GR (ed). ANZDATA. Registry report 2001. Adelaide: Aus-tralia and New Zealand Dialysis and Transplant Registry, 2001.

26. Darne B, Girerd X, Safar M, Cambien F, Guize L. Pulsatile versussteady component of blood pressure: A cross-sectional analysis anda prospective analysis on cardiovascular mortality. Hypertension1989; 13: 392–400.

27. Benetos A, Safar M, Rudnichi A et al. Pulse pressure: A predictorof long-term cardiovascular mortality in a French male population.Hypertension 1997; 30: 1410–15.

28. Domanski MJ, Mitchell GF, Norman JE, Exner DV, Pitt B, PfefferMA. Independent prognostic information provided by sphyg-momanometrically determined pulse pressure and mean arterialpressure in patients with left ventricular dysfunction. J. Am. Coll.Cardiol. 1999; 33: 951–8.

29. London GM, Marchais SJ, Safar ME et al. Aortic and large arterycompliance in end-stage renal failure. Kidney Int. 1990; 37: 137–42.

30. London G, Guerin A, Pannier B, Marchais S, Benetos A, Safar M.Increased systolic pressure in chronic uremia. Role of arterial wavereflections. Hypertension 1992; 20: 10–19.

31. Barenbrock M, Spieker C, Laske V et al. Studies of the vessel wallproperties in hemodialysis patients. Kidney Int. 1994; 45: 1397–400.

32. Blacher J, London GM, Safar ME, Mourad JJ. Influence of age andend-stage renal disease on the stiffness of carotid wall material inhypertension. J. Hypertens. 1999; 17: 237–44.

33. Van Etten RW, De Koning EJ, Verhaar MC et al. Impaired NO-dependent vasodilatation in pateints with Type II (non insulin-dependent) diabetes mellitus is restored by acute administration offolate. Diabetologia 2002; 45: 1004–10.

34. London GM, Guerin AP, Pannier B, Marchais SJ, Stimpel M.Influence of sex on arterial hemodynamics and blood pressure.Role of body height. Hypertension 1995; 26: 514–19.

35. Hayward CS, Kelly RP. Gender-related differences in the centralarterial pressure waveform. J. Am. Coll. Cardiol. 1997; 30: 1863–71.

36. Smulyan H, Marchais SJ, Pannier B, Guerin AP, Safar ME,London GM. Influence of body height on pulsatile arterial hemo-dynamic data. J. Am. Coll Cardiol. 1998; 31: 1103–9.

37. Cameron JD, McGrath BP, Dart AM. Use of radial artery appla-nation tonometry and a generalized transfer function to determineaortic pressure augmentation in subjects with treated hyperten-sion. J. Am. Coll. Cardiol. 1998; 32: 1214–20.

ASFAST baseline results 141

38. Suliman M, Qureshi A, Barany P et al. Hyperhomocysteinemia,nutritional status and cardiovascular disease in hemodialyispatients. Kidney Int. 2000; 57: 1727–35.

39. Tremblay R, Bonnardeaux A, Geadah D et al. Hyperhomocys-teinemia in hemodialysis patients: Effects of 12-month supple-mentation with hypersoluble vitamins. Kidney Int. 2000; 58:851–8.

40. Perna AF, Ingrosso D, De Santo NG, Galletti P, Brunone M,Zappia V. Metabolic consequences of folate-induced reduction ofhyperhomocysteinemia in uremia. J. Am. Soc. Nephrol. 1997; 8:1899–905.

41. Lindner A, Bankson DD, Stehman-Breen C, Mahuren JD, CoburnSP. Vitamin B6 metabolism and homocysteine in end-stage renaldisease and chronic renal insufficiency. Am. J. Kidney Dis. 2002;39: 134–45.

42. McQuillan B, Beilby J, Nidorf M et al. Hyperhomocysteinemia butnot the C677T mutation of methylenetetrahydrofolate reductaseis an independent risk determinant of carotid wall thickening: thePerth carotid ultrasound disease assessment study (CUDAS).Circulation 1999; 99: 2383–8.

43. Willinek WA, Ludwig M, Lennarz M, Holler T, Stumpe KO.High-normal serum homocysteine concentrations are associ-ated with an increased risk of early atherosclerotic carotidartery wall lesions in healthy subjects. J. Hypertens. 2000; 18:425–30.

44. Malinow MR, Nieto FJ, Szklo M, Chambless LE, Bond G. Carotidartery intimal-medial wall thickening and plasma homocyst(e)inein asymptomatic adults. The Atherosclerosis Risk in CommunitiesStudy. Circulation 1993; 87: 1107–13.

45. Lim PS, Hung WR, Wei YH. Polymorphism in methylenetetrahy-drofolate reductase gene: Its impact on plasma homocysteinelevels and carotid atherosclerosis in ESRD patients receivinghemodialysis. Nephron 2001; 87: 249–56.

46. Perry IJ, Refsum HM, Morris RW et al. Prospective study of serumtotal homocysteine concentration and risk of stroke in middle-aged British men. Lancet 1995; 346: 1395–8.

47. Graham IM, Daly LE, Refsum HM et al. Plasma homocysteine as arisk factor for vascular disease. The European Concerted ActionProject. JAMA 1997; 277: 1775–81.

48. Bots ML, Launer LJ, Lindemans J et al. Homocysteine and short-term risk of myocardial infarction and stroke in the elderly: TheRotterdam Study. Arch. Intern. Med. 1999; 159: 38–44.

49. Nygard O, Nordrehaug JE, Refsum H, Ueland PM, Farstad M,Vollset SE. Plasma homocysteine levels and mortality in patientswith coronary artery disease. N. Engl. J. Med. 1997; 337: 230–6.

50. Taylor L, Moneta G, Sexton G et al. Prospective blinded study ofthe relationship between plasma homocysteine and progression ofsymptomatic peripheral arterial disease. J. Vas. Surg. 1999; 29: 8–19.

51. Wald DS, Law M, Morris JK. Homocysteine and cardiovasculardisease: Evidence on causality from a meta-analysis. BMJ 2002;325: 1202–9.

52. Fermo I, Vigano’D’Angelo S, Paroni R, Mazzola G, Calori G,D’Angelo A. Prevalence of moderate hyperhomocysteinemia inpatients with early-onset venous and arterial occlusive disease.Ann. Intern. Med. 1995; 123: 747–53.

53. den Heijer M, Brouwer IA, Bos GM et al. Vitamin supplementa-tion reduces blood homocysteine levels: A controlled trial inpatients with venous thrombosis and healthy volunteers. Arterio-scler. Thromb. Vasc. Biol. 1998; 18: 356–61.

54. Vermeulen EG, Stehouwer CD, Twisk JW et al. Effect of homo-cysteine-lowering treatment with folic acid plus vitamin B6 onprogression of subclinical atherosclerosis: A randomised, placebo-controlled trial. Lancet 2000; 355: 517–22.

55. Schnyder G, Roffi M, Pin R et al. Decreased rate of coronary res-tenosis after lowering of plasma homocysteine levels. N. Engl. J.Med. 2001; 345: 1593–600.

56. Schnyder G, Roffi M, Flammer Y, Pin R, Hess OM. Effect ofhomocysteine-lowering therapy with folic acid, vitamin B (12),and vitamin B (6) on clinical outcome after percutaneous coro-nary intervention: The Swiss Heart study: A randomizedcontrolled trial. JAMA 2002; 288: 973–9.