Apolipoprotein B, Fibrinogen, HDL Cholesterol, and Apolipoprotein(a) Phenotypes Predict Coronary...

Apolipoprotein B, Fibrinogen, HDL Cholesterol, and

Apolipoprotein(a) Phenotypes Predict Coronary Artery

Disease in Hemodialysis Patients

MICHAEL KOCH,* BERND KUTKUHN,* EVI TRENKWALDER,t DIETER BACH,*

BERND GRABENSEE,* HANS DIEPLINGER,t and FLORIAN KRONENBERGt*Department of Internal Medicine, Division of Nephrologv and Rheumatology. Heinrich Heine University,

D#{252}sseldorf Germany; and tlnstitute of Medical Biology and Human Genetics, University of Innsbruck,

Innsbruck, Austria.

Abstract. Patients with end-stage renal disease have a markedly

elevated risk for coronary artery disease (CAD). Lipids and

most lipoproteins, however, seem to be not predictive for CAD

in these patients. Although there is clear evidence that lipopro-

tein(a) [Lp(a)] is significantly elevated in these patients, no

study with a sufficiently large group of hemodiaiysis patients

has investigated the relationship between CAD and Lp(a), as

well as the genetically determined apolipoprotein(a) [apo(a)]

phenotype. This cross-sectional study determines the preva-

lence of CAD in relation to the cardiovascular risk profile in an

unselected population of 607 hemodialysis patients, of which

33% were diabetic patients. Twenty-six percent (n = I 58) of

all patients suffered from CAD as diagnosed by a definitive

myocardial infarction (ii = 102) and/or at least one stenosis

>50% of a coronary artery (ii = 143). In univariate analysis,

several classic risk factors, including the concentration of

lipids, bipoproteins, apolipoproteins. and fibrinogen. correlated

with CAD. Lp(a) in patients with CAD showed only a ten-

dency to higher levels, without reaching significance, corn-

pared with patients without CAD (26.6 ± 30.8 rng/dl versus

22. 1 ± 30.4 mg/db, P = 0. 1 0). The frequency of low molecular

weight apo(a) isoforms, however, was significantly greater in

the group with CAD (34.8% versus 23.6%, P < 0.01 ). Step-

wise logistic regression analysis found seven variables associ-

ated with CAD: apobipoprotein B. the low molecular weight

apo(a) phenotype, male sex, age, fibrinogen, diabetes mellitus,

and HDL cholesterol. The association of these variables with

CAD differed depending on age. These results indicate that,

besides classic risk factors such as age, sex, and diabetes

mellitus, additional factors of the lipoprotein and fibrinobytic

system contribute to the high prevalence of CAD in hemodi-

alysis patients. (J Am Soc Nephrol 8: 1889-1898, 1997)

Patients with end-stage renal disease have a well documented

high risk for atherosclerotic complications, and cardiovascular

disease is the major cause of death in patients on renal replace-

ment therapy (1-4). Depending on comorbidity, cardiac events

such as myocardial infarction are approximately five to 50

times higher than in the general population (5). Although very

common in these patients (4,6-8), it is still a matter of dis-

cussion whether and how much lipoprotein and fibrinobytic

disorders contribute to the excessive cardiovascular morbidity

and mortality of these patients (9).

Lipoprotein(a) [Lp(a)] is a cholesterol-rich lipoprotein con-

sisting of an LDL attached to a large and highly polymorphic

glycoprotein named apolipoprotein(a) [apo(a)]. Lp(a) concen-

trations are highly heritable and mainly determined by the size

polymorphism of apo(a) (10). The right-skewed distribution,

together with the more than 1000-fold interindividual range of

Received February 13, 1997. Accepted May 19, 1997.Correspondence to Dr. Florian Kronenberg, Institute of Medical Biology and

Human Genetics, Sch#{246}pfstrasse 4 1. A-6020 Innsbruck, Austria.

1046-6673/0801 2- 1889$03.00/0

Journal of the American Society of Nephrology

Copyright iD 1997 by the American Society of Nephrology

Lp(a) (10), requires a high number of cases and controls to

detect differences of clinical relevance. Apo(a) phenotyping

can improve the reliability of such findings ( 1 1 , 1 2). From

numerous epidemiological studies. it seems that Lp(a) is a risk

factor and predictor of premature atherosclerosis ( I 3- 1 6). In

patients with end-stage renal disease, Lp(a) was found to be

significantly elevated (reviewed in reference 1 1 ). The obser-

vation of arteriovenous differences of Lp(a) between the aorta

and the renal vein, with lower concentrations in the renal vein,

suggests that the kidney is involved in the catabolism of Lp(a)

( 1 7). In contrast to the general population, the apo(a) pheno-

type seems to be more predictive of atherosclerosis in this

patient group than Lp(a) concentrations ( I 8).

Fibrinogen. a further risk factor for atherosclerosis in the

general population (19), was found to be elevated in patients

with chronic renal failure (20-22). Studies in small patient

groups with renal disease and coronary artery disease (CAD),

however, produced inconsistent results (2 1-23).

The aim of our study was to investigate whether elevated

Lp(a) concentrations and/or apo(a) phenotypes. as well as

fibrinogen levels and other lipids or lipoproteins, are predictive

ofCAD in hemodialysis patients. Therefore, we determined the

lipid profile, Lp(a), apo(a) phenotypes, lipoproteins. apoli-

I 890 Journal of the American Society of Nephrology

poproteins A-I and B, fibrinogen, and CAD status in a large

and unselected group of 607 hemodialysis patients, including

diabetic and nondiabetic patients.

Materials and MethodsPatients

During 1993 and 1994, we investigated the prevalence of cardio-

vascular events and the cardiovascular risk profile in 607 unselected

hemodialysis patients of seven dialysis centers in the Rhineland,

Germany. We decided to include nondiabetic and diabetic patients,

because we also wanted to investigate the relative importance of

diabetes mellitus as risk factor for CAD in this high-risk patient

population. Informed consent was obtained from all patients. The

study included 3 18 men and 289 women with a mean age of 58 ± 14

yr (range. 19 to 87 yr) and a mean duration of hemodialysis treatment

of 44 ± 5 1 IT1() (median. 27; range. I to 297 nb). Patients were treated

with bicarbonate-buffered dialysate three times weekly with 3 to 6 h

per session. Patients were considered diabetic ifthey were treated with

insulin or oral hypoglycemic agents. and/or fasting blood glucose was

higher than 140 mg/dl. Diabetes classification in the 200 diabetic

patients (42 of type I and 158 of type II) was done according to the

criteria of the National Diabetes Data Group (24). Patients were

classified as hypertensive if either the BP measured at the beginning

of dialysis treatment was greater than I 60/95 mmHg or patients were

on antihypertensive treatment. Smoking history was recorded by

interview. Body mass index was calculated with the body weight at

the end of dialysis treatment.

CAD was established by medical record documentation of defini-

tive myocardial infarction with the corresponding changes in electro-

cardiogram (significant Q waves) and/or a coronary angiography.

Selective coronary angiography was performed in patients who re-

ported angina pectoris and/or other clinical hints (changes of the ST

segment) suggesting coronary insufficiency. Multiple projections of

the left and right coronary artery were assessed, and the presence of

heniodynamically relevant stenosis (stenosis > 50% of the luminal

diameter) was classified as CAD.

Laborator’t’ Measurements

Blood was drawn in a fasting state before the beginning of dialysis

treatment. In patients with a myocardial infarction, blood was with-

drawn at the earliest 3 mo after the event to avoid the effect of acute

phase. After low-speed centrifugation, samples were frozen and kept

at -80#{176}Cbefore analysis (25). Lipids. lipoproteins. and apolipopro-

teins were measured in fresh serum and fibrinogen was measured in

fresh plasma in one laboratory, using commercially available kits

from Boehringer Mannheim (Mannheim, Germany): total cholesterol

with the CHOD-PAP method, HDL cholesterol after precipitation

with phosphotungstic acidIMgCl2, LDL cholesterol after precipitation

with heparin (26), triglycerides with the GPO-PAP method. and

apolipoprotein A-I and apolipoprotein B by nephelometry. Fibrinogen

was measured according to a modified method described by Clauss

(27). Plasma HbAlc was determined with DIAMAT-analysis system

(Bio-Rad, Richmond, CA) by ion exchange chromatography.

Lp(a) quantification was performed as described in detail (25) with

a double-antibody. enzyme-linked immunosorbent assay, using an

affinity-purified polycbonal apo(a) antibody for coating and the horse-

radish peroxidase-conjugated monoclonal antibody 1A2 for detection.

This anti-apo(a) antibody recognizes the epitope motif YYPN in

kringle IV (K-IV) type 2 (28). An Lp(a)-positive serum from Immuno

(Vienna, Austria) with the same apo(a) isoforms served as standard

throughout the study. Each sample was analyzed in duplicate, and

intra- and interassay coefficients of variation were 2.7 and 6%, re-

spectively. The reproducibility of our assay was monitored by four

control plasma samples used in each run. It was shown previously that

Lp(a) measurements with this assay give stable results in samples

stored for up to 2 yr (25). Apo(a) phenotyping was performed by

sodium dodecyl sulfate (SDS)-agarose gel electrophoresis under re-

ducing conditions, as outlined previously (29), with the following

modifications. In a first run, we applied 50 ng of Lp(a) to the gel. In

cases in which we detected only one apo(a) isoform, we repeated the

sample, applying up to 2000 ng. which often resulted in the detection

of a second apo(a) isoform that was weak in relation to the other one

detected in the previous run. By this procedure, it was possible to

detect two apo(a) isoforms in 79% of the patients under investigation.

In only two patients (0.3%), we were not able to detect any apo(a)isoform.

Statistical Analyses

Statistical analyses were performed with Statistical Analysis Sys-

tem (version 6. 1 1) and Statistical Package for the Social Sciences for

Windows 6. 1 .3. Univariate comparisons of continuous variables be-

tween groups were done by unpaired t test or by the nonparametric

Wilcoxon rank sum test in the case of non-normally distributed

variables [Lp(a), triglycerides. and time on dialysis]. Dichotomized

variables were compared using Pearson’s ,y� test. This test was also

applied to compare the prevalence of CAD between different catego-

ries of apolipoprotein B and fibrinogen levels, as well as K-IV repeats.

Variables contributing to diabetic status or CAD status were iden-

tified by stepwise logistic regression analysis. Lp(a) and triglyceride

concentrations and time on dialysis were included in the model after

logarithmic transformation. By removing single variables, we deter-

mined the most parsimonious model. Odds ratios (OR) as an approx-

imation of the relative risk of having CAD were calculated from the

regression coefficients, and adjustment was made for other variables.

To allow an easier interpretation of the OR of continuous variables,

we give the OR for clinically relevant increments as stated for each

variable. However, the calculation was not based on a “categoriza-

tion,” but on the continuous value of the variable divided by the step

of increment (e.g. . values of apolipoprotein B divided by 20).

To examine a possible effect modification, we built interactionterms of diabetes mellitus with age. fibrinogen, and triglyceride con-

centrations by calculating the product of the code for diabetic status

with age. fibrinogen, and triglycerides, respectively. These interaction

terms were used in logistic regression analysis.

To investigate whether differences for each variable between pa-tients with and without CAD in the total patient group depended on

the diabetic status, we introduced an interaction term of these two

variables. For continuous variables, we used 2-way ANOVA with

diabetic and CAD status and the interaction term as independent

variables. For dichotomized variables, multiple logistic regression

analysis was used with the same three variables as independent

variables.

Because of the high number of detectable apo(a) isoforms (>30),

many phenotypes were only represented in bow numbers. To account

for this problem, we decided a priori to combine apo(a) isoforms in

steps of three K-IV repeats according to the molecular weight of thesmaller apo(a) isoforms to get sufficient sample sizes in each category

(30). Because patients with 1 1 to 16 or >34 K-IV repeats were

represented relatively rarely. we built one group by combining 1 1 to

19 and one by combining >31 K-IV repeats. Furthermore, we divided

apo(a) phenotypes in two subgroups according to the molecular

weight of the smaller apo(a) isoforms, as done in previous works

CAD in Hemodialysis Patients 1891

(12,18,31-33). The low molecular weight group (LMW) included all

patients with at least one apo(a) isoform with I 1 to 22 K-IV repeats

(29); the high molecular weight (HMW) group comprised all patients

who had only isoforms with >22 K-IV repeats. If two apo(a) isoforms

were detectable, we generally used only the smaller apo(a) isoform for

categorization for the following reasons. First, phenotyping was per-

formed by SDS-agarose gel electrophoresis. Although we could detect

two apo(a) isoforms in 79% of the patients with this method. the

remaining patients cannot simply be classified as homozygous for the

observed isoform. From DNA phenotyping by pulsed-field gel elec-

trophoresis, only about 6% of the patients are known to be homozy-

gous (34,35). Therefore, more than two-thirds of the patients who

showed only one band in SDS-agarose gel electrophoresis are genet-

ically heterozygous. This discrepancy may be explained either by the

resolution power of SDS-agarose gel ebectrophoresis [e.g. . two iso-

forms differing only by one K-IV repeat are often seen as one apo(a)

band] or by the detection limit of immunoblotting in case of very low

expression of an isoform. Therefore, to classify a patient as homozy-

gous by SDS-agarose gel electrophoresis would result in a misclassi-

fication in most of the cases. Second, more than 80% of the patients

with two apo(a) bands expressed the smaller apo(a) band with more

intensity. Third, the results obtained were similar when we based our

calculations only on the predominantly expressed apo(a) isoform of

each patient.

ResultsWe investigated the prevalence of cardiovascular events in

relation to the cardiovascular risk profile of 607 unselected

hemodialysis patients. This group included 200 patients with

(33%) and 407 patients without diabetes mellitus. In univariate

analysis, diabetic patients had significantly higher triglyceride

and fibrinogen concentrations, a higher body mass index, and

were more often hypertensive. They were less likely to be

smokers and were on dialysis treatment for a shorter time.

Furthermore, the diabetic patients suffered more often from

CAD than the nondiabetic patients (35% versus 22%. P <

0.001) (Table 1). In logistic regression analysis, the differences

between diabetic and nondiabetic patients remained significant

for triglyceride concentration, body mass index, time on dial-

ysis, and frequency of hypertension and CAD (data not

shown).

Twenty-six percent (ii = 158) of all patients showed CAD

diagnosed by a definitive myocardial infarction (ii = 102) or at

least one stenosis >50% of a coronary artery (n = 143), or

both. In univariate analysis (Table 2), the group of patients

with CAD was significantly older and included more men and

diabetic patients. Apobipoprotein B, fibrinogen, total and LDL

Table 1. Univariate analysis between patients with and without diabetes meblitusa

Characteristic

Patient Group

p Value. .Nondiabetic

. .Diabetic

(Ii = 407) (‘7 =

Age (yr) 57 ± 15 61 ± 13 0.01

Male sex (%) 52 54 NS

Years of diabetes mellitus 17.7 ± 9.3

Months of dialysisc 53 ± 56 26 ± 34 0.001

Body mass index (kg/rn2) 22.9 ± 3.5 24.9 ± 4.5 0.0()l

Current smoker (%) 26 1 8 0.05

Hypertension (%) 48 72 0.001

Systolic BP (mmHg) 134 ± 19 141 ± 20 0.00 1

Diastolic BP (mmHg) 76 ± 10 77 ± 10 NS

CAD(%) 22 35 0.001

Total cholesterol (mg/dl) 2 19 ± 57 229 ± 63 NS

HDL cholesterol (mg/dl) 40.0 ± 16.0 37.8 ± 17.6 NS

LDL cholesterol (mg/db) 1 38 ± 46 1 40 ± 5 1 NS

Triglycerides (mg/dl; mean ± SD)C median 2 1 7 ± 1 56 300 ± 209 0.0() 1

180 229

Apolipoprotein A-I (mg/dl) 122 ± 28 1 19 ± 30 NS

Apobipoprotein B (mg/dl) 131 ± 40 137 ± 46 NS

Lp(a) (mg/dl; mean ± SD)’ median 2 1 .8 ± 28.6 26.3 ± 34.0 0.2

9.3 12.1

LMW apo(a) phenotypes (%)(� 27.3 25.0 NS

Fibrinogen (mg/dl) 386 ± I 13 420 ± 131 0.01

a CAD, coronary artery disease; Lp(a), lipoprotein(a): LMW, low molecular weight.b Including 42 patients with type I and I 58 with type II diabetes mellitus.C Wilcoxon rank sum test was used due to non-normal distribution of the variable; other continuous variables were compared by

unpaired I test.d Low molecular weight apo(a) phenotype.

1892 Journal of the American Society of Nephrology

Table 2. Univariate analysis of patients with and without CAD independent of diabetic status

Characteristic

Patient Group

‘) ValueWithout CAD With CAD(n - 449) (?i = 158)

Age (yr) 57 ± 15 63 ± 1 1 0.00 1

Male sex (%) 48 63 0.001

Diabetes melbitus (%) 29 44 0.001

Months of dialysisa 45 ± 52 43 ± 48 NS

Body mass index (kg/m2) 23.6 ± 4.0 23.5 ± 3.9 NS

Current smoker (%) 23 24 NS

Hypertension (%) 56 55 NS

Systolic BP (mmHg) 136 ± 19 136 ± 20 NS


Total cholesterol (mg/db) 219 ± 58 232 ± 61 0.05

HDL cholesterol (mg/db) 40.4 ± 16.9 36.1 ± 15.2 0.01

LDL cholesterol (mg/dl) 135 ± 47 149 ± 50 0.01

Triglycerides (mg/dl: mean ± SD)�’ median 233 ± 171 274 ± 197 0.001

194 205 0.001

Apobipoprotein A-I (mg/db) 123 ± 29 1 16 ± 27 0.05

Apolipoprotein B (mg/dl) 128 ± 39 146 ± 47 0.001

Lp(a) (mg/dl: mean ± SD)a median 22. 1 ± 30.4 26.6 ± 30.8 0.1

9.3 12.6 0.1

LMW apo(a) phenotypes (%)�‘ 23.6 34.8 0.01

Fibrinogen (mg/dl) 382 ± ii2 440 ± 132 0.001

a Wilcoxon rank sum test was used due to non-normal distribution of the variable; other continuous variables were compared by

unpaired t test.h Low molecular weight apo(a) phenotype.

cholesterol, and triglycerides were significantly higher,

whereas HDL cholesterol and apolipoprotein A-I were lower.

The prevalence of CAD increased markedly with increasing

apolipoprotein B and fibrinogen concentrations (P < 0.001)

(Figure 1 ). Lp(a) in patients with CAD showed only a tendency

toward higher levels without reaching statistical significance

(26.6 ± 30.8 versus 22. 1 ± 30.4 mg/dl, P 0. 1). However, the

frequency of LMW apo(a) isoforms was significantly higher in

patients with CAD compared to those without CAD (34.8%

versu.s 23.6%, P < 0.01) (Table 2). The prevalence of CAD

was particularly elevated in 105 patients with 20 to 22 K-IV

repeats (Figure 1 ). Surprisingly. in patients with the lowest

number of K-IV repeats (1 1 to 19 repeats). the prevalence of

CAD was similar to patients with HMW apo(a) phenotypes.

We therefore compared those two groups with the lowest (1 1

to 19 repeats, ii = 56) and the highest (>31 repeats, ii = 118)

number of K-tV repeats that had a similar prevalence of CAD.

The former one had, with the exception of Lp(a) concentra-

tions, a more favorable risk profile, with significantly higher

apobipoprotein A-I (128 ± 25 versus 117 ± 27 mg/dl, P =

0.01 ) and HDL cholesterol (46 ± 14 versus 38 ± 16 mg/dl.

P = 0.001), as well as lower triglyceride concentrations

( I 69 ± 82 versus 254 ± 2 1 7 mg/dl, P < 0.005) and a trend

toward a lower frequency of smokers ( 17% versus 28%) and

men (46% versus 57%).

Because one-third of the patients were diabetic, we per-

formed the univariate analysis between patients with and with-

out CAD separately for diabetic and nondiabetic patients (Ta-

bles 3 and 4). In both groups, patients with CAD were older,

were more often men, had higher apolipoprotein B and fibrin-

ogen concentrations, and had a higher frequency of LMW

apo(a) isoforms. The differences for total, LDL, and HDL

cholesterol and for triglycerides reached statistical significance

only in nondiabetic patients. In diabetic patients with CAD,

apolipoprotein A-I and the frequency of hypertension was

lower compared to those without CAD (Table 3).

To investigate whether differences between patients with

and without CAD in the total patient group depend simply on

the diabetic status, we introduced an interaction term of these

two variables. Two-way ANOVA for continuous variables and

multiple logistic regression analysis for dichotomized variables

found that this was not the case for the variables in Table 2.

By stepwise mubtivariate logistic regression analysis. we

investigated which variables were independently associated

with the prevalence of CAD. Because this analysis necessitates

evaluating a large group of patients depending on the number

of variables included (36), we decided to include in a first run

all patients independent of their diabetic status to increase the

power of the analysis. Furthermore, this procedure should

confirm the independence of diabetes mellitus as a risk factor

for CAD in a large group of unselected hemodialysis patients.

Seven variables correlated with CAD in the final model: apo-

__80

0

0

I� 20>ci)

Rn


Apoli poprotein B p<0.OOl

0<=60 61-

80

81-

1 00101-

1 20121-

140141- 161- 181-

160 180 200

201- >220

220 mg/dL

n = 12 34 85 115 126 98 62 32 24 19

Fibrinogen p<0.O01

<=250 250-299

300-349

350-399

400-449

450- 500- 550-499 549 599

600- >650649 mg/dL

n = 52 62 123 122 73 71 43 22 13 26

.-. 80

0

0

� 40

ce 20>�i)

0�

e�_� 60

K-IV repeats pczO.005

0

40

20

‘ I I I I1 1 -1 9

20-22

23-25

26-28

29-31

>31K-lV repeats

n= 56 105 99 126 103 118

Figure 1. Prevalence of coronary artery disease (CAD) in relation toapolipoprotein B and fibrinogen concentrations and the number of

kringle-IV (K-IV) repeats of apobipoprotein(a). n is the number of

samples in each group.

lipoprotein B, the LMW apo(a) phenotype, male sex, age,

fibrinogen level, diabetes mellitus, and HDL cholesterol (Table

5). This final model classified 76% of the patients correctly for

having CAD. We compared univariate OR of these variables

with conditional OR of the final multivariate model and found

similar values, with the exception of HDL cholesterol (data not

shown), which indicates that HDL cholesterol is not indepen-

dent from the other variables.

Because age and levels of fibrinogen and triglyceride were

associated with both CAD and diabetic status in univariate

analysis, we introduced interaction terms for the diabetic status

and these three variables to detect a possible effect modifica-

tion. Repeating the logistic regression analysis, we found no

effect by the introduction of the diabetes-fibrinogen and dia-

betes-triglyceride interaction terms, but did find a significant

effect by the introduction of the diabetes-age interaction term

(x2 5.62, P < 0.05). We therefore decided to perform theanalysis separately for patients bel4’ and above the mean age

of 58 yr. This cutoff is justified by the high rate of patients with

diabetes mellitus type II who had a mean age of 65 yr. Because

0

CAD rarely occurred in the 44 patients younger than 35 yr of

age (2.3%, or one case), we excluded these patients in the

following step to avoid a “dilution” of the effects. The patient

group aged 35 to 58 yr included 237 patients. of whom 46

(19.4%) showed CAD. Logistic regression analysis found that

fibrinogen, the LMW apo(a) phenotype. and HDL cholesterol

were significant predictors for the prevalence of CAD (Table

5). This final model classified 83% of the patients correctly for

having CAD. These variables can be considered independent

because univariate OR did not markedly differ from the con-

ditional OR of the multivariate model. Although the frequency

of CAD in this age group tended to be higher in diabetic than

nondiabetic patients (26% versus 17%), diabetes mellitus

failed to be a significant marker for CAD in multivariate

analysis (P = 0.19). The older patient group consisted of 326

patients and showed a markedly higher prevalence of CAD

compared with the younger group (34.0% versus’ 19.4%, P <

0.001 ). Regression analysis found that apolipoprotein B. male

sex, and diabetes mellitus were associated with CAD, and this

model classified 68% of the patients correctly. All three van-

abbes were independently associated with CAD. The LMW

apo(a) phenotype showed a borderline association (P < 0. 1);

fibninogen failed to be a marker in this age group.

DiscussionWe describe three major findings in this investigation of a

large group of unselected hemodiabysis patients with a wide

range in age. ( 1 ) CAD occurs frequently in hemodialysis

patients, with an even higher prevalence in diabetic patients.

(2) CAD is strongly associated with age. male sex, diabetes

mellitus, LMW apo(a) isoforms, high levels of apolipoprotein

B and fibrinogen. and low bevels of HDL cholesterol. (3) The

association of these variables with CAD differs between age

groups.

Patients with end-stage renal disease have an extremely high

risk for atherosclerotic complications (1-5). A 3- to 5-yr fol-

low-up of 6742 end-stage renal disease patients revealed a

relative risk of death from myocardial infarction five times

higher than in the general population (2). A prospective study

of 305 patients with end-stage renal disease described a car-

diovascular mortality 10 times higher and for diabetic patients

more than 40 times higher than for the general population (3).

In agreement with these studies, we found a prevalence of

CAD in an unselected group of diabetic and nondiabetic he-

modialysis patients of 35 and 22%, respectively. This preva-

lence is probably markedly higher in reality because only 30

and 22%, respectively, of these patients underwent a coronary

angiography and because a study in diabetic patients showed a

high frequency of symptomless but angiographically verified

CAD (37). On the other hand, approximately 50% of dialysis

patients with clinical angina pectoris have no significant cor-

onary artery occlusion (38). Therefore. we minimized false

positive results by performing a coronary angiography in pa-

tients with symptoms of angina pectoris. We are aware that

designation of CAD would be more reliable if performed by

coronary angiography in the whole patient group independent

of whether patients have symptoms. whether they are younger


Table 3. Univaniate analysis of diabetic patients with and without CAD

Characteristic

Patient Group

‘#{176}ValueWithout CAD With CAD

(‘I = 131) (0 = 69)

Age (yr) 59 ± 13 64 ± 12 0.05

Male sex (%) 50 62 0.1

Months of dialysis” 26 ± 36 25 ± 3 1 NS

Body mass index (kg/m2) 25.0 ± 4.6 24.7 ± 4.2 NSCurrent smoker (%) 16 21 NS

Hypertension (%) 77 62 0.05Systolic BP (mmHg) 141 ± 19 140 ± 21 NS


Total cholesterol (mg/db) 227 ± 60 233 ± 70 NS

HDL cholesterol (mg/dl) 39.0 ± 17.7 35.6 ± 17.3 NSLDL cholesterol (mg/dl) 137 ± 49 146 ± 54 NS

Triglycerides (mg/dl: mean ± SD)�’ median 293 ± 197 314 ± 230 NS

217 245 NS

Apolipoprotein A-I (mg/dl) 123 ± 32 1 13 ± 26 0.05

Apobipoprotein B (mg/dl) 132 ± 40 148 ± 56 0.05Lp(a) (mg/dl: mean :�: SD)� median 25.3 ± 34.9 28.1 ± 32.4 NS

12.5 10.9 NS

LMW apo(a) phenotypes (%)b 20.6 33.3 0.05

Fibninogen (mg/dl) 400 ± 123 457 ± 139 0.01

.‘ Wilcoxon rank sum test was used due to non-normal distribution of the variable; other continuous variables were compared by

unpaired I test.h Low molecular weight apo(a) phenotype.

or older. and whether they are men or women. Because this

was not possible in our study, we decided to use very strict

criteria, i.e. , definitive myocardial infarction or coronary an-

giography. or both, in patients with clinical symptoms sugges-

tive of CAD. We found for diabetic patients a similar preva-

lence of CAD (35%) in a recent study that included 105

consecutive diabetic patients of the same geographic area un-

dergoing coronary angiography regardless of clinical symp-

toms. Of these patients. 36% presented with hemodynamically

significant CAD, and an additional 1 1% showed coronary

atherosclerosis without hemodynamically relevant stenosis

(39).

The fact that diabetic patients have a higher prevalence of

CAD than nondiabetic patients (35% versus 22%, P < 0.001)

may (besides the diabetes mellitus per se) be partially ex-

plained by the higher frequency of hypertension, higher body

mass index, and triglyceride serum concentrations. The time on

dialysis treatment, however, was markedly shorter in diabetic

patients (26 versus 53 mo, P < 0.001 ), which was possibly

caused by a decreased survival time of dialyzed diabetic pa-

tients (40) compared with nondiabetic patients. In agreement

with an earlier study (12), we observed no significant differ-

ences in Lp(a) concentrations or apo(a) phenotype distribution,

as well as total, HDL, and LDL cholesterol between diabetic

and nondiabetic hemodialysis patients. Only triglyceride con-

centrations were higher in diabetic patients (Table I).

Besides the three classic risk factors for atherosclerosis (age,

male sex, and diabetes mellitus), serum concentrations of apo-

lipoprotein B, fibrinogen, and HDL cholesterol, as well as the

frequency of LMW apo(a) phenotype, were significantly asso-

ciated with CAD in the entire patient group (Table 5). Fibrin-

ogen is described as a risk factor for atherosclerosis in the

general population ( 1 9). and elevated levels have been reported

in patients with chronic renal failure (20-22). Peritoneal dial-

ysis patients with CAD were found to have higher fibrinogen

levels than those without CAD (22). However, a study of 71

end-stage renal disease patients treated by hemodialysis or

peritoneab dialysis found cardiovascular disease in 23 patients

that was not associated with higher fibrinogen levels in mul-

tivaniate analysis (21 ). Another report on 24 patients with

vascular disease described fibrinogen levels similar to those in

26 patients without such complications (23). These contrasting

results may be explained by the small patient groups investi-

gated (21-23). In the study presented here, fibrinogen was

significantly higher in the entire group of CAD patients ( 1 58 of

607 patients) in univariate and multivariate analysis. As with

apolipoprotein B, this association is supported by a continu-

ously increasing prevalence of CAD with increasing levels of

fibninogen (Figure 1). It is unclear, however, why fibrinogen

was not predictive of CAD in the older patient group (Table 5).

In the general population, a prospective study in 2 1 ,520 men

found apolipoprotein B most strongly and independently asso-

ciated with the risk for ischemic heart disease. This association

was stronger than for other lipids and apolipoproteins (41).


Table 4. Univariate analysis of nondiabetic patients with and without CAD

Characteristic

Patient Group

�#{176}ValueWithout CAD With CAD(n = 318) (ii = 89)

Age (yr) 56 ± 16 62 ± 10 0.001

Male sex (%) 48 65 0.01

Months of dialysisa 53 ± 56 57 ± 55 NS

Body mass index (kg/rn2) 23.0 ± 3.5 22.5 ± 3.3 NS

Current smoker (%) 26 26 NS

Hypertension (%) 47 49 NS

Systolic BP (mmHg) I 34 ± 1 9 1 33 ± 1 8 NS


Total cholesterol (mg/dl) 216 ± 58 231 ± 54 0.05

HDL cholesterol (mg/dl) 41.0 ± 16.5 36.5 ± 13.5 0.05

LDL cholesterol (mg/dl) 1 35 ± 46 1 5 1 ± 47 0.01

Triglycerides (mg/dl; mean ± SD)� median 209 ± 153

175

244 ± 164

197

0.06

0.06

Apolipoprotein A-I (mg/dl) 123 ± 28 1 19 ± 27 NS

Apolipoprotein B (mg/dl) 1 27 ± 39 145 ± 40 0.001

Lp(a) (mg/dl; mean ± SD)� median 20.8 ± 28.3

8.6

25.4 ± 29.5

14.4

0.1

0.1

LMW apo(a) phenotypes (%)“ 24.8 36.0 0.05

Fibrinogen (mg/dl) 374 ± 107 427 ± 125 0.001

a Wilcoxon rank sum test was used due to non-normal distribution of the variable; other continuous variables were compared by

unpaired t test.b Low molecular weight apo(a) phenotype.

When we separated our patients into two age groups (35 to 58

yr versus >58 yr). apolipoprotein B was found to correlate

with the risk for CAD only in the older group. This may be due

to the fact that diabetes was more frequent in the older group

compared with the younger group (38.7% versus 27.8%, P <

0.01) and to the reported observation that apolipoprotein B is

higher in diabetic patients (42). Total and LDL cholesterol did

not correlate with CAD by multivariate analysis. These find-

ings are in accordance with those of Attman and Alaupovic,

who suggested that the atherogenic potential of dyslipidemia in

renal disease depends more on apolipoprotein than on lipid

abnormalities and therefore may not always be recognized by

measurement of lipids alone (9,43). Previous studies described

elevated triglyceride and decreased HDL cholesterol levels in

hemodialysis patients with CAD (8,44). Our study, which

included a markedly higher number of patients, observed these

changes in univariate analysis. In multivariate analysis, HDL

cholesterol was associated with CAD, especially in younger

patients (Table 5); triglycerides, however, clearly failed to be

predictive.

In the general population, numerous studies found lipids and

lipoproteins related to CAD. High total and LDL cholesterol

and low HDL cholesterol values, as well as in some studies

high triglyceride concentrations, have been demonstrated to be

associated with an increased risk for CAD. Apolipoproteins

such as apolipoprotein A-I or B were related to CAD in several

of these studies (reviewed in reference 45). This raises the

question, Why does a parameter such as LDL cholesterol,

which is a potent predictor in the general population. fail to be

predictive in hemodialysis patients? Besides other reasons, we

must bear in mind that the measurement of parameters in the

hemodialysis phase provides a single static view of exposure to

risk factors at a given moment, ignoring the changeable expo-

sure to risk factors during the development of renal disease.

These patients often have experienced a long phase of hyper-

cholesterolemia during the development of renal disease and

are now normo- or hypocholesterobemic. Therefore, the pre-

dictive value of this actually measured parameter might be

diminished.

There is clear evidence from numerous studies that Lp(a) is

elevated in hemodialysis patients (reviewed in reference 1 1).

This is of special importance because most, but not all, of the

prospective and numerous retrospective studies in the general

population have described an independent association between

high Lp(a) concentrations and CAD or other atherosclerotic

complications (reviewed in references 45 and 46). Only a few

studies with a sufficient number of patients have investigated

this association in renal disease patients. Cressman et al. fob-

lowed 1 29 hemodialysis patients prospectively for 48 mo (47).

Lp(a) levels in the 26 patients with atherosclerotic events were

significantly higher than in those without events. A heteroge-

neous study group, including hemodialysis, peritoneal. and

renal transplant patients, reported symptomatic arterial disease

in 92 of 325 patients (48). The symptomatic patients had


Table 5. Stepwise logistic regression analysis to predict CAD status in hemodialysis patients�’

Variable Coefficient SEM (9S�RCI) P Value

Total group (n = 607)

constant -4.271 0.977 19.10

apolipoprotein B (increment 20 mg/dl) 0.176 0.052 12.61 1.19 (1.08 to 1.32) <0.001

LMW apo(a) phenotyp&’ 0.746 0.230 10.51 2.1 1 (1.34 to 3.31) 0.001

sex ( 1 = female, 2 = male) 0.737 0.228 10.44 2.09 (1 .34 to 3.27) 0.001

age (increment 5 yr) 0.131 0.042 9.94 1.14 (1.05 to 1.24) <0.005

fibrinogen (increment 50 mg/dl) 0.1 16 0.044 7.13 1.12 (1.03 to 1.22) <0.01

diabetes mellitus (0 = no, 1 = yes) 0.521 0.215 5.84 1.68 (1.10 to 2.57) <0.02

HDL cholesterol (increment 10 mg/dl) -0.015 0.007 3.98 0.99 (0.97 to 0.999) <0.05

Age 35 to 58 yr (ii = 237)

constant -0.296 1.166 0.06

fibrinogen (increment 50 mg/dl) 0.267 0.077 1 1.96 1.31 (1.12 to 1.52) <0.001

LMW apo(a) phenotyp&’ I .237 0.434 8. 1 3 3.45 (1 .47 to 8.07) <0.005

HDL cholesterol (increment 10 mg/dl) -0.292 0.133 4.79 0.75 (0.58 to 0.97) <0.05

Age >58 yr (ii = 326)

constant -2.770 0.794 12.17

sex (I = female, 2 = male) 0.864 0.260 1 1.07 2.37 (1.43 to 3.95) <0.001

apolipoprotein B (increment 20 mg/dl) 0.183 0.057 10.27 1.20 (1.07 to 1.34) <0.005

diabetes mellitus (0 = no, I = yes) 0.532 0.250 4.54 1.70 (1.04 to 2.78) <0.05

LMW apo(a) phenotype” 0.442 0.264 2.80 1.56 (0.93 to 2.61) 0.09

a Analysis was performed for the total group and separately for the age groups 35 to 58 yr and >58 yr (58 yr corresponds to the mean

age of the total patient group). OR, odds ratio; CI, confidence interval. Other abbreviations as in Table 1.h Apo(a) phenotype: I = high molecular weight; 2 = low molecular weight.

significantly higher Lp(a) concentrations and a higher fre-

quency of LMW apo(a) phenotypes (48). Prospective studies in

the general population demonstrated that patients who devel-

oped CAD during the observation period had 30 to 50% higher

Lp(a) concentrations than those remaining free of CAD (46).

Recent statistical simulations with Lp(a) concentrations dem-

onstrated that at least 100 patients per group in case control

studies are necessary to detect such a difference in an unam-

biguous manner (11,12). Although we investigated 607 pa-

tients, we did not find a significant elevation of Lp(a) in

hemodialysis patients with CAD. Instead of Lp(a), the fre-

quency of LMW apo(a) phenotypes was significantly higher in

patients with CAD compared to those without CAD (34.8%

versus 23.6%, P < 0.01). This result is in accordance with a

recent study (1 8), which found that the LMW apo(a) phenotype

was more predictive of the prevalence and degree of carotid

atherosclerosis than were Lp(a) plasma concentrations in a

group of 167 hemodialysis patients. This finding was explained

by the apo(a) phenotype-specific elevation of Lp(a) found in

hemodialysis patients ( 12,33): because Lp(a) in hemodialysis

patients increases only in patients with HMW apo(a) pheno-

types, the Lp(a) levels of these patients come closer to those of

LMW apo(a) phenotypes. It has therefore been suggested that

hemodialysis patients can no longer be discriminated for high

and low risk by means of Lp(a) concentrations alone. In

contrast, the genetically determined apo(a) phenotype that de-

termines predisease Lp(a) concentrations also reflects the

former “baseline” atherosclerosis risk before development of

end-stage renal disease (18).

The proposed model is in agreement with observations in

our patients in whom the apo(a) phenotype was more predic-

tive of CAD status than Lp(a) concentrations. Furthermore, it

concurs with a study in peritoneal dialysis patients with CAD,

in which Lp(a) was only slightly elevated but the frequency of

LMW apo(a) phenotypes was twice as high as in patients

without CAD (49). Renal disease seems to be the only condi-

tion in which the apo(a) phenotype is a better predictor for

CAD (49) or atherosclerotic complication ( 1 8) than the Lp(a)

concentration. Nonrenal disease patients with CAD show an

increased frequency of LMW apo(a) phenotypes (3 1,50). In

multivariate analysis, however, the Lp(a) concentration ex-

eludes the apo(a) phenotype from the model due to the strong

negative correlation between the molecular weight of apo(a)

isoforms and Lp(a) plasma concentrations (10). In patients

with renal disease, this relation is weakened by the elevation of

Lp(a) in HMW apo(a) phenotypes. Therefore, the apo(a) phe-

notype, which also reflects the former atherosclerosis risk, is

more predictive of CAD than Lp(a) levels.

When we divided the patients in six groups depending on the

number of K-IV repeats of apo(a), we observed that the fre-

quency of CAD was especially elevated in those 105 patients

with 20 to 22 K-IV repeats (Figure 1). Comparing patients with

the lowest number of K-IV repeats (1 1-19 K-IV) to patients

with the highest number of repeats revealed, surprisingly, a


similar prevalence of CAD, which seems to be a contradiction.

However, patients with 1 1 to 19 K-IV repeats had [with the

exception of Lp(a) concentrations] a more favorable athero-

sclerosis risk profile. Because this is a case control study

without prospective observation, we cannot exclude a survival

bias: A higher number of patients with 1 1 to 19 K-IV repeats

and a less favorable atherosclerosis risk profile might have

passed away at the time of examination and the surviving

patients compensated for their disadvantage of the apo(a) size

polymorphism by a more favorable risk profile. This hypoth-

esis should, however, be confirmed by future large prospective

studies.

Similar to a previous study (1 8), the apo(a) phenotype was

not predictive in the older age group of our patients. The OR

for patients with an LMW apo(a) phenotype was significantly

elevated in the total patient group (OR, 2. 1 1 ; P = 0.001 ) and

in the younger patient group (OR, 3.45; P < 0.005), but did not

reach significance in the older group (OR, 1 .56; P = 0.09)

(Table 5). This agrees with observations from the general

population that Lp(a) is a risk factor for premature CAD

( 1 3- 15), but less of a risk factor in patients >55 yr old( IS).

In summary, CAD risk in hemodialysis patients may be

explained in part by classic risk factors such as age, male sex,

and diabetes mellitus. In contrast to most lipids and lipopro-

teins, apolipoprotein B, HDL cholesterol, the LMW apo(a)

phenotype, and fibrinogen are predictive of CAD in various

age groups. This study extends previous findings in a smaller

hemodialysis patient group investigating carotid atherosclero-

sis (1 8) to CAD and demonstrates that the apo(a) phenotype is

of reliable predictive value for CAD in patients with end-stage

renal disease.

AcknowledgmentsThis study was supported by grants from the “Austrian National-

bank” to Dr. Kronenberg (Project 5553) and from the Austrian “Fonds

zur Forderung der Wissenschaftlichen Forschung” to Dr. Dieplinger(P-10090 and P-12358). We appreciate the helpful critical contribu-

tion of Dr. G. Utermann (Institute of Medical Biology and Human

Genetics, University of Innsbruck, Innsbruck. Austria) and the statis-tical advice of Dr. G. Seeber (Department of Statistics, University of

Innsbruck, Innsbruck, Austria) and Dr. R. Willers (Department of

Biostatistics, University of D#{252}sseldorf, D#{252}sseldorf, Germany). For the

excellent collaboration, we thank Dr. G. van Endert and Dr. W.Kleophas (Renal Unit Karlstrasse, D#{252}sseldorf),Dr. J. Erkenbrecht and

Dr. B. Hartmann (Diakoniewerk Kaiserswerth, D#{252}sseldorf), Dr. W.Herms (Evangelisches Krankenhaus, D#{252}sseldorf),Dr. W. Hust (RenalUnit Adolf-Flecken-Straf3e, Neuss, Germany), Dr. L. Jennessen and

Dr. W. Roperts (Renal Unit Eickener H#{246}h.Monchengladbach, Ger-

many), Dr. W. Kulas and Dr. R. Trapp (Renal Unit Gartenstraj3e,Mettmann, Germany) and Dr. W. Kreusser (Marienhospital, Duis-burg, Germany).

References1. Held PJ, Port FK. Webb RL, Wolfe RA, Bloembergen WE,

Turenne MN, Hobzman E, Ojo AO, Young EW, Mauger EA:Excerpts from United States Renal Data System 1995 AnnualData Report. Am J Kidney Dis 26: Sl-Sl86, 1995

2. Raine AEG, MacMahon SH, Selwood NH, Wing AJ, Brunner

FP: Mortality from myocardial infarction in patients on renal

replacement therapy in the U.K. [Abstractb. Nephrol Dial Traits-

p/alit 6: 902, 1991

3. Brown JH, Hunt LP, Vites NP. Short CD, Gokal R. Mallick NP:

Comparative mortality from cardiovascular disease in patientswith chronic renal failure. Nephrol Dial Transplant 9: 1 136-

1142. 1994

4. Kronenberg F. Dieplinger H. Konig P. Utermann G: Lipoprotein

metabolism in renal replacement therapy: A review. !sr J Med

Sci 32: 371-389. 1996

5. Brunner FP. Selwood NH, on behalf of the EDTA Registration

Committee: Profile of patients on RRT in Europe and death rates

due to major causes of death groups. Kidney lot 42lSuppl 381:

54-515. 1992

6. Avram MM, Goldwasser P. Burrell DE. Antignani A. Fein PA.Mittman N: The uremic dyslipidemia: A cross-sectional and

longitudinal study. Am J Kidney Dis 20: 324-335. 1992

7. Attman P-O, Alaupovic P: Lipid abnormalities in chronic renal

insufficiency. Kidney ml 39[Suppl 31]: S16-S23. 1991

8. Goldberg AP. Harter HR. Patsch W, Schechtman KB. Province

M, Weerts C. Kuisk I, McCrate MM, Schonfeld G: Racialdifferences in plasma high-density lipoproteins in patients re-

ceiving hemodialysis: A possible mechanism for accelerated

atherosclerosis in white men. N EngI J Med 308: 124�-12S2,

1983

9. Ritz E: Why are lipids not predictive of cardiovascular death in

the dialysis patient? Miner Electrolyte Metab 22: 9 - 1 2. 199610. Utermann G: The mysteries of lipoprotein(a). Science 246: 904-

910. 1989

1 1. Kronenberg F, Utermann G. Dieplinger H: Lipoprotein(a) in

renal disease. Am J Kidney Dis 27: 1-25, 1996

12. Kronenberg F, Konig P. Neyer U. Auinger M. Pribasnig A. Lang

U. Reitinger I. Pinter G. Utermann G. Dieplinger H: Multicenter

study of lipoprotein(a) and apolipoprotein(a) phenotypes in pa-

tients with end-stage renal disease treated by hemodialysis or

continuous ambulatory peritoneal dialysis. J A,,: Soc Nep/irol 6:

110-120. 1995

13. Sandkamp M, Funke H, Schulte H, K#{246}hlerE, Assmann G:

Lipoprotein(a) is an independent risk factor for myocardial in-

farction at a young age. C/in C/tern 36: 20-23, 1990

14. Dumngton PN, Ishola M, Hunt L, Arrol 5, Bhatnagar D: Apo-

lipoproteins (a). Al and B and parental history in men with early

onset ischaemic heart disease. Lancet i: 1070-1073, 1988

15. Sunayama 5, Daida H, Mokuno H. Miyano H. Yokoi H. Lee YJ.

Sakurai H, Yamaguchi H: Lack of increased coronary athero-sclerotic risk due to elevated lipoprotein(a) in women > or 55years of age. Circulation 94: 1263-1 268. 1996

16. Assmann G, Schubte H, Von Eckardstein A: Hypertniglyceride-mia and elevated lipoprotein(a) are risk factors for major coro-

nary events in middle-aged men. Am J C’ardiol 77: 1 179-1 184,

I 996

I 7. Kronenberg F, Trenkwalder E. Lingenhel A. Fniedrich G. Lhotta

K, Schober M, Moes N, Konig P. Utermann G. Dieplinger H:Renovascular arteniovenous differences in Lp(a) plasma concen-

trations suggest removal of Lp(a) from the renal circulation.

J Lipid Res 38: 1755-1763, 1997

18. Kronenberg F, Kathrein H, Konig P. Neyer U. Sturm W. Lhotta

K. Gr#{246}chenig E. Utermann G. Dieplinger H: Apolipoprotein(a)

phenotypes predict the risk for carotid atherosclerosis in patients

with end-stage renal disease. Arterioscler Thromb I 4: 140S-

1411, 1994

19. Wilhelmsen L, Sv#{228}rdsuddK, Korsan-Bengtsen K. Larsson BO.


Welin L, Tibblin G: Fibninogen as a risk factor for stroke and

myocardial infarction. N Engl J Med 31 1: 501-505, 1984

20. Haaber AB, Eidemak I, Jensen T, Feldt-Rasmussen B,

Strandgaard S: Vascular endothelial cell function and cardiovas-

cular risk factors in patients with chronic renal failure. J Am Soc

Nep/zrol5: 1581-1584, 1995

2 1. Bostom AG, Shemin D, Lapane KL. Sutherland P, Nadeau MR.

Wilson PWF. Yoburn D, Bausserman L. Tofler G, Jacques PF,

Selhub I, Rosenberg IH: Hyperhomocysteinemia. hyperfibrino-

genemia. and lipoprotein (a) excess in maintenance dialysis

patients: A matched case-control study. Atherosclerosis 125:

91-101, 1996

22. Bartens W. Nauck M. Schollmeyer P. Wanner C: Elevated Ii-

poprotein(a) and fibrinogen serum levels increase the cardiovas-

cular risk in continuous ambulatory peritoneal dialysis patients.

Peril Dial In! 16: 27-33. 1996

23. Bachmann J, Tepel M. Raidt H. Riezler R, Graefe U, Langer K,

Zidek W: Hyperhomocysteinemia and the risk for vascular dis-

ease in hemodialysis patients. J Am Soc Nephrol 6: 121-125,

I 995

24. National Diabetes Data Group: Classification and diagnosis of

diabetes mellitus and other categories of glucose intolerance.

Diabetes 28: 1039-1057, 1979

25. Kronenberg F. Lobentanz E-M, Konig P. Utermann G,

Dieplinger H: Effect of sample storage on the measurement of

Iipoprotein(a), apolipoproteins B and A-IV, total and high-den-

sity lipoprotein cholesterol and triglycerides. J Lipid Res 35:

1318-1328. 1994

26. Wieland H. Seidel D: A simple specific method for precipitation

of low density lipoproteins. J Lipid Res 24: 904-909, 1983

27. Clauss A: Gerinnungsphysiologische Schnellmethode zur

Bestimmung des Fibrinogens. Acta Haematol 17: 237-246, 1957

28. Dieplinger H. Gruber G. Krasznai K, Reschauer 5, Seidel C.

Burns G, Muller H-i, Cs#{225}sz#{225}rA, Vogel W, Robenek H, Uter-

mann G: Kringle 4 of human apolipoprotein[a] shares a linear

antigenic site with human catalase. J Lipid Res 36: 813-822.

I 995

29. Kraft HG. Lingenhel A, Bader G, Kostner GM, Utermann G: The

relative electrophoretic mobility of apo(a) isoforms depends on

the gel system: Proposal of a nomenclature for apo(a) pheno-

types. Atherosclerosis 125: 53-61, 1996

30. Budowle B, Giusti AM, Waye iS. Baechtel FS, Fourney RM,

Adams DE, Presley LA, Deadman HA, Monson KL: Fixed-bin

analysis for statistical evaluation of continuous distributions of

allelic data from VNTR loci, for use in forensic comparisons.

Am J Hum Genet 48: 841-855, 1991

3 1 . Sandholzer C. Saha N, Kark ID. Rees A, iaross W, Dieplinger H,

Hoppichler F, Boerwinkle E, Utermann G: Apo(a) isoforms

predict risk for coronary heart disease: A study in six popula-

tions. Arteriosc/er Thromb 12: 1214-1226, 1992

32. Molgaard i. Klausen IC. Lassvik C, Fa�rgeman 0, Gerdes LU,

Olsson AG: Significant association between low-molecular-

weight apolipoprotein(a) isoforms and intermittent claudication.

Arterioscler Thromb 12: 895-901 , 1992

33. Dieplinger H, Lackner C, Kronenberg F, Sandholzer C. Lhotta K,

Hoppichler F, Graf H, Konig P: Elevated plasma concentrations

of lipoprotein(a) in patients with end-stage renal disease are not

related to the size polymorphism of apolipoprotein(a). J C/in

hues! 91: 397-401. 1993

34. Boerwinkle E. Leffert CC, Lin I, Lackner C, Chiesa G. Hobbs

HH: Apolipoprotein(a) gene accounts for greater than 90% of the

variation in plasma lipoprotein(a) concentrations. J Clin Invest

90: 52-60, 1992

35. Kraft HG, K#{246}chl5, Menzel Hi, Sandholzer C. Utermann G: The

apolipoprotein(a) gene: A transcribed hypervariable locus con-

trolling plasma lipoprotein(a) concentration. Hum Genet 90:

220-230, 1992

36. Nunnally IC: Psychometric Theory, 2nd Ed., New York,

McGraw-Hill, 1978.

37. Braun WE, Phillips DG, Vidt AC, Novick AC, Nakamoto S,

Popowniak KL. Paganini E, Magnusson M, Pohl M, Steinmuller

DR. Protiva D, Buszta C: Coronary artery disease in 100 diabet-

ics with end-stage renal failure. Transplant Proc 16: 603-607,

1984

38. Rostand SG, Kirk KA, Rutsky EA: Dialysis-associated ischemic

heart disease: Insights from coronary angiography. Kidney hit

25: 653-659. 1984

39. Koch M, Gradaus F, Schoebel F-C, Leschke M, Grabensee B:

Relevance of conventional cardiovascular risk factors for the

prediction of coronary artery disease in diabetic patients on renal

replacement therapy. Nephrol Dial Transplant 12: 1187-1191,

I 997

40. Koch M, Thomas B, Tschope W, Ritz E: Survival and predictors

of death in dialysed diabetic patients. Diabeto/ogia 36: 1 1 13-

1117, 1993

41. Wald NJ, Law M, Watt HC, Wu TS, Bailey A, iohnson AM,

Craig WY, Ledue TB, Haddow JE: Apolipoproteins and isch-

aemic heart disease: Implications for screening. Lancet 343:

75-79. 1994

42. Attman P-O, Nyberg G. William-Olsson T, Knight-Gibson C,

Alaupovic P: Dyslipoproteinemia in diabetic renal failure. Kid-

nevlnt42: 1381-1389, 1992

43. Attman P-O. Alaupovic P: Lipid and apolipoprotein profiles of

uremic dyslipoproteinemia: Relation to renal function and dial-

ysis. Nephron 57: 401-410, 1991

44. Hahn R, Oette K, Mondorf H, Finke K, Sieberth HG: Analysis of

cardiovascular risk factors in chronic hemodialysis patients with

special attention to the hyperlipoproteinemias. Atherosclerosis

48: 279-288, 1983

45. Rader Di, Hoeg JM, Brewer HB ir: Quantitation of plasma

apolipoproteins in the primary and secondary prevention of cor-

onary artery disease. Ann Intern Med 120: 1012-1025, 1994

46. Kronenberg F, Steinmetz A, Kostner GM, Dieplinger H: Li-

poprotein(a) in health and disease. Crit Rev C/in Lab Sci 33:

495-543, 1996

47. Cressman MD, Heyka Ri, Paganini EP, O’Neil I, Skibinski CI,

Hoff HF: Lipoprotein(a) is an independent risk factor for cardio-

vascular disease in hemodialysis patients. Circulation 86: 475-

482, 1992

48. Webb AT, Reaveley DA, O’Donnell M, O’Connor B, Seed M,

Brown EA: Lipids and lipoprotein(a) as risk factors for vasculardisease in patients on renal replacement therapy. Nephrol Dial

Transplant 10: 354-357. 1995

49. Wanner C, Bartens W, Walz G, Nauck M, Schollmeyer P:

Protein loss and genetic polymorphism of apolipoprotein(a)

modulate serum lipoprotein(a) in CAPD patients. Nephrol Dial

Transplant 10: 75-81, 1995

50. Kraft HG, Lingenhel A, K#{246}chl5, Hoppichler F, Kronenberg F,

Abe A, MUhlberger V. Sch#{246}nitzerD, Utermann G: Apolipopro-

tein(a) Kringle IV repeat number predicts risk for coronary heart

disease. Arterioscler Thromb Vasc Biol 16: 7 13-7 19, 1996

Apolipoprotein B, Fibrinogen, HDL Cholesterol, and Apolipoprotein(a) Phenotypes Predict Coronary...

Documents

Transcript of Apolipoprotein B, Fibrinogen, HDL Cholesterol, and Apolipoprotein(a) Phenotypes Predict Coronary...