Effect of dialyzer geometry during hemodialysis with cuprophane membranes

Kidney International, Vol. 42 (1992), pp. 442—447

Effect of dialyzer geometry during hemodialysis with cuprophane membranes J.E. TAYLOR, M. MCLAREN, R.A. MACTIER, I.S. HENDERSON, W.K. STEWART, and J.J.F. BELCH Renal Unit and Department of Medicine, Ninewells Hospital and Medical School, Dundee, Scotland, United Kingdom

Effect of dialyzer geometry during hemodialysis with cuprophane membranes. The effect of dialyzer geometry, both flat plate (FP) and hollow fiber (HF), on platelet and granulocyte activation during dialysis with cuprophane membranes was studied in 12 patients. A subset of six patients was restudied alter correction of their anemia with recombinant human erythropoietin (EPO). Granulocyte count and aggregation in vitro fell significantly (P <0.01) at 20 minutes of dialysis, followed by a gradual return towards pre-dialysis values at 240 minutes, Malondi-

aldehyde (MDA), a product of free radical reactions generated by activated granulocytes, increased significantly during dialysis [predialysis MDA (median, range): 8.4 (5.8 to 11.6) nmol/ml, 240 minutes MDA: 9.7 (6.6 to 12.5) nmol/mI, P <0.01 Wilcoxon test). This increase,

however, was not affected by dialyzer geometry or EPO therapy. Neither type of dialyzer was associated with significant platelet loss at the end of dialysis. Whole blood platelet aggregation in vitro (spontaneous and collagen-induced) decreased significantly, (P < 0.01) during dialysis, the fall in spontaneous aggregation being significantly less following EPO therapy [spontaneous aggregation 240 minutes; preEPO: 34 (13 to 52) %; post-EPO 50: (16 to 76) %, P < 0.01)1. The ratio of the platelet release proteins /3-thromboglobulin and platelet factor 4 increased significantly during dialysis, indicating platelet activation in vivo, although there was no effect of dialyzer geometry or EPO. Factor VIII von Willebrand Factor antigen, a putative marker of endothelial damage, was raised pre-dialysis, and increased further during dialysis,

irrespective of dialyzer geometry or EPO. In conclusion, dialyzer geometry had no significant effect on granulocyte and platelet counts and activity during hemodialysis with cuprophane membranes. Erythropoietin was associated with higher spontaneous platelet aggregation during dialysis, which could be relevant to the development of extracorporeal thrombosis.

Hollow fiber (HF) or fiat plate (FP) dialyzers are currently utilized for almost all hemodialysis treatments. Although there have been few prospective controlled studies of these different dialyzer designs, it seems that neither type of dialyzer offers definite advantages over the other, such as: (a) similar cost; (b) all biomaterials currently used in membrane manufacture can be incorporated into both forms of dialyzers; (c) dialyzers of equivalent performance as measured by solute clearance and ultrafiltration coefficient are available in each form of dialyzer; and (d) equal residual volumes. Hollow fiber dialyzers. however, have lower blood priming volumes, and have been shown Received for publication November 21, 1991 and in revised form March 16, 1992 Accepted for publication March 16, 1992

© 1992 by the International Society of Nephrology

by some [1], but not others [2], to induce less thrombocytopenia during dialysis. Red cells, white cells and platelets have profound effects on

microcirculatory flow, and their behavior following passage through a dialyzer may be important. Hemodialysis is associated with the activation of white blood cells, in particular granulocytes [3], and platelets [4]. Activated granulocytes produce reactive oxygen species (free radicals) which are prothrombotic. They may also effect ischemic damage by their physical ability to occlude the microcirculation due to endothehal adherence, and the formation of aggregates [5]. Platelets have a central role in thrombus formation, and there is some evidence for greater platelet activation during dialysis with flat plate designs [1]. The increase in red cell number due to the use of recombinant human erythropoietin (EPO) is associated with

enhanced platelet aggregability [6, 7], which could further augment any dialysis-associated coagulation changes.

In this randomized crossover study, we investigated the effect of dialyzer geometry (FP vs. HF) on granulocyte and platelet aggregation in vitro in a group of 12 stable hemodialysis

patients. Six of these patients, who were anemic, were restudied after correction of their anemia with EPO. In this smaller subset of patients, indirect tests of free radical activity [plasma malondialdehyde (MDA) and thiol], the platelet release proteins [/3-thromboglobulin (BTG) and platelet factor 4 (PF4)], and Factor VIII von Willebrand factor antigen (FVIIIvWFAg), a marker of endothelial damage [8], were also measured. Methods

Patients Twelve patients, median age 57.5 (range 20 to 75) years, seven males, five females, were studied. Their underlying causes of renal failure were: polycystic kidneys (3), chronic glomerulonephritis (3), obstructive uropathy (2), unknown (2), renovascular disease (1), and hypertensive nephropathy (1). All had been receiving hemodialysis for at least 12 months, and had reliable vascular access via a Cimino fistula. No patient was receiving medication known to influence blood coagulation, and none had chronic infection or inflammation. They were randomly assigned to be dialyzed with either a flat plate [Gambro Lundia (surface area) 4 N(l.0 m2), 5 N(1.1 m2) or 6 N(l.6 m2)] or hollow fiber [Gambro Alwall GFE11(1.1 m2) or

442

443

Taylor et at: Dialyzer geometry during dialysis

(b) Collagen: Addition of 10 pi of a 0.6 jsg!ml collagen GFE18(1.8 m2), Gambro AB, Lund, Sweden] cuprophane dialyzer for a period of four weeks, following which the alternative solution (Semmelweis, Italy) followed by four mindialyzer of approximately comparable surface area was used for utes of incubation in a shaking water bath at 37°C a further four week period. All dialyzers were ethylene oxide (c) ADP: Addition of 10 1.d of a 1 tM ADP solution sterilized, and were used on a single occasion only. There were (Sigma, UK) followed by one minute of incubation slight differences in the non-cuprophane materials in the two in a shaking water bath at 37°C, dialyzer types: FP, pressure plates (styreneacrylonitrile), side Platelet aggregation was determined from the percentage fall plates (acrylonitrilebutadienestyrene), connecting plates (polypropylene); and HF, potting material (polyurethane) housing in platelet count following each aggregation procedure. and caps (polycarbonate). Duration of dialysis, blood and The following tests were only performed in the six patients dialysate flow rates, dialysis fluid [8 patients bicarbonate (ace- who were additionally treated with EPO.

tate concentration 3 mEq/liter), 4 patients acetate], target weight, ultrafiltration and heparin dose were kept constant for Malondialdehyde (MDA) each patient throughout the study. Blood was taken at the following time points during a dialysis MDA is an oxidative product of free radical reactions and at the end of the four week period on a particular dialyzer type: would therefore be expected to increase following granulocyte

(a) arterial line, pre-dialyzer, at commencement of dialysis

(b) venous line post-dialyzer, at commencement of

activation and the release of toxic oxygen species. Plasma malondialdehyde was measured using Aust's spectrophotomet-

nc assay [10] on 2.5 ml blood anticoagulated with lithium

(c) after 20 minutes of dialysis (venous line)

heparin (10 lU/mi). The plasma was mixed with thiobarbituric acid reagent (Sigma), and heated in a water bath at 100°C for 30 minutes. The optical density of the supernatant was measured at 532 nm, and was converted to MDA concentration using an

(d) after 240 minutes of dialysis (venous line)

extinction coefficient of 1.50 x l0 mol' cm (normal range:

dialysis

The following tests were performed in all 12 patients on both types of dialyzer, and repeated in six patients (on both dialyzers) following EPO therapy.

Granulocyte aggregation in whole blood This was measured using the method of Fisher et a! [9], on 5 ml blood anticoagulated with sodium heparin (10 lU/mi). The sample was stirred (150 rpm) at 37°C for three minutes, following which the synthetic aggregating agent N-formyl-methionylleucyl-phenyl-alanine (fMLP, Sigma, UK) was added, and the sample stirred for a further 1.5 minutes. Aliquots were removed from the sample pre- and post-aggregation, and fixed in 0.2% formalin in isotonic diluent (Isoton II, Coulter Electronics, UK). An erythrocyte lysing agent (Zapoglobin, Coulter Electronics) was added to remove the red cells, following which the formalin-fixed white cells were processed through a Coulter

5.0 to 8.0 nmol/ml).

Plasma thiol Plasma thiol, mainly albumin thiol, can donate electrons to neutralize free radicals, and is a useful measure of the degree of oxidation of the plasma [11]. This was measured using Eliman's method [12] on 2.5 ml blood anticoagulated with lithium heparin

(10 lU/mi). Ellman's reagent (Sigma) and 0.1 M di-sodium hydrogen orthophosphate buffer (pH 7.6) were added to the sample, and the mixture left at room temperature for five minutes. Optical density was measured at 440 nm to give an estimate of the thiol concentration of the sample (normal range:

400 to 600 mol/liter). /3-thromboglobulin (BTG) and platelet factor 4 (PF4) BTG and PF4 are two platelet-specific proteins released from

counter (Coulter ZN, Coulter Electronics) equipped with a platelets when they aggregate, and their measure can be used as particle volume analyzer (Pulse, Width and Height Analyser, markers of platelet activation. Both proteins are measured, and Bio-engineering Unit, University of Strathclyde). This apparatus transferred particle volume and frequency information to a microcomputer where it was displayed as a curve with several peaks representing lymphocytes, single granulocytes, and granulocyte aggregates. Granulocyte aggregation was calculated from the percentage fall in the area under the single granulocyte curve post-aggregation.

Platelet aggregation in whole blood This was measured on 5 ml blood anticoagulated with 3.8% trisodium citrate in the proportion 9:1. The sample was separated into 1 ml aliquots which were tested for spontaneous, collagen-induced and ADP-induced aggregation as follows: (a) Spontaneous: Rotation (30 rpm) for six minutes at 37°C

an increase in the ratio of BTG to PF4 is indicative of platelet activation in vivo [13]. BTG and PF4 were determined using commercial enzyme immunoassay kits (Elisa BTG, Boehringer Mannheim GmbH Diagnostica (Enzygnost Platelet Factor 4; Behring, Germany). Blood (4.5 ml) was collected into tubes containing citrate, adenosine, theophylline and dipyridamole, cooled on ice for 15 minutes, and then centrifuged at 2000 g for 30 minutes at 4°C. The middle layer of supernatant was removed for testing. Test plasma was incubated in tubes coated with antibodies to BTG or PF4. The bound BTG/PF4 was then allowed to couple with anti-BTG/PF4 bound to peroxidase. This subsequently reacted with hydrogen peroxide and the chro-

mogen ortho-phenylenediamine to produce a color change detected spectrophotometrically at 492 nm which was proportional to the BTG/PF4 concentration of the sample. [Normal ranges: BTG 10 to 40 lU/mi; PF4 2.5 (1.4 to 6.1) pg/liter].

444

Taylor et al: Dialyzer geometry during dialysis 12

.

A

11

100

10

A . a090

9

I

.' a

8

•

7 6

..

5 4

3 2 1

0

80-

Mo

.

£

+&8. it ..

•

.

.'•

'I •

A

a

27o —a 60.j,Aa a50'

•'! S

.* $a

Pre Post Post

'

40-

"30H 0 20

.

.*

.

4

.4

a

S

f1 .

a

lisa

La •5

W

Post

1st 20mm 240 mm

r-4

Pre Post

Post

Post

1st 20mm 240 mm

Fig. 1. Change in granulocyte count and aggregation during dialysis with flat plate (•) and hollow fiber (A) cuprophane membranes in 12 hemodialysis patients. (Horizontal lines depict median values.)

Factor VIII von Willebrand factor antigen (FVIIIvWFAg)

0.05, HF) at 240 minutes (Fig. 1). With both types of dialyzer, granulocyte aggregation fell in parallel with the fall in granuloblood flow. FVIIIvWFAg was released from damaged endothe- cyte count to approximately 25% of pre-dialysis levels at 20 hum and was measured to investigate the effects on the endo- minutes. There was, however, no significant difference between thelium of blood that had passed through a dialyzer. A com- the two dialyzer types at any other time point during dialysis mercial enzyme immunoassay was used (Dako Ltd, Denmark). (Fig. 1). EPO treatment had no significant effect on granulocyte Citrated plasma was added to a microwell coated with anti- count or aggregation during dialysis. In patients dialyzed with FP dialyzers there was a slight, but FVIIIvWFAg. This was then further incubated with a peroxidase conjugated anti-FVIIIvWFAg antibody, and reacted with significant (P < 0.05), fall in platelet count 20 minutes into ortho-phenylenediamine to produce a color change at 492 nm, dialysis [FP pre-dialysis: 151 (78 to 253) x 109/hiter; 20 mm: 135 proportional to the amount of FVIIIvWFAg in the sample. The (71—244) x l09/liter}. By the end of dialysis (240 mm) platelet results were expressed as a percentage of values obtained from counts had returned to levels which were not significantly different from those pre-dialysis. A fall in platelet count at 20 pooled normal plasma. minutes was not observed when patients were dialyzed with HF Statistical analysis dialyzers, however, with this dialyzer type, platelet counts Non-parametric statistical methods were used, and results tended to overshoot at the end of dialysis (P < 0.05 vs. expressed as medians and ranges. Wilcoxon rank sum tests for pre-dialysis) to levels significantly higher (P <0.02) than those matched-paired data were used to compare pre-dialysis values observed with FP dialyzers. [HF pre-dialysis: 170(106 to 262) x with different time points during dialysis for a single dialyzer 109/liter; HF 240 mm: 195 (155 to 290) x 109/liter; FP pretype, and to compare the two dialyzer types at the same time dialysis: 151 (78 to 253) x 109/liter; FP 240 mm: 153 (100 to 249) point during dialysis. In the six patients additionally tested after x 109/liter]. There was no significant difference between the treatment with EPO, analysis of variance (Kruskal Wallis Test), two dialyzer types at any other time points during dialysis (Fig. was used to determine the significance of dialyzer geometry and 2), and in six patients there was no additional effect of EPO on EPO therapy. Changes in parameters during the course of a platelet count. Spontaneous and collagen-induced platelet agdialysis were correlated with the change in plasma albumin gregation fell significantly during the course of dialysis irrespecusing Spearman's rank correlation in order to determine the tive of dialyzer geometry (Fig. 2), but there was no significant change in aggregation induced by ADP. In the six patients effect of hemoconcentration. treated with EPO, spontaneous platelet aggregation was signif-

An intact functioning endothelium is required for normal

Results

icantly higher (P < 0.01) at 20 and 240 minutes of dialysis There was no significant difference in the patients' pre- compared to pre-EPO [20 mm; pre-EPO: 37 (19 to 82)%; dialysis plasma potassium, urea, phosphate, creatinine and post-EPO: 58 (40 to 91)%, 240 mm; pre-EPO: 34 (13 to 52)%, albumin on the two different dialyzer types with or without EPO therapy. Hematocrit, however, increased significantly in the six patients who were treated with EPO regardless of dialyzer type [pre-EPO: 22.0 (17.5 to 30.8) %, post-EPO: 31.4 (27.2 to 40.6)

%, P < 0.01).

post-EPO: 50 (16 to 76)%J. These findings were not influenced by dialyzer geometry.

In the six patients studied before and after EPO treatment,

MDA levels were higher pre-dialysis than in normal non-uremic

subjects, and increased further during the course of dialysis

In all patients, granulocyte counts fell significantly to approx- (Table 1), whereas plasma thiol was lower than normal preimately 25% of pre-dialysis values 20 minutes into dialysis. This dialysis, but increased during dialysis. This increase, however, was followed by a return to just above pre-dialysis values (P < was significantly correlated with the change in plasma albumin

445

Taylor et a!: Dialyzer geomer,y during dialysis 300

A A

Si

250'-

:.

A A

A

200 -

•A

150 -

I

__

•A

St

"As a 55 A

.

A

50 -

Pre Post Post 1st 20mm C

0 a)

0) a)

0) 0) a)

4a) a) a)

a a)

90

C a)

0) a)

0

0

- 'SI —

!

S

80 - S 70 - — :— •• ,A 60 50 - S * ••

•

40 30 20 10

0) 7Q I

S

'a

-

•g

''

Pre Post

A

.. •A .

S

A

S

St •5 'A A

A

50 .6 A I Li U) 401 Si S As 0a) c 201I

••

a)

a)j a0 Post

S

0

$a

4-

£

701 i

r

Post 1st 20 mm

•

Ta

SC 2

20 I

10] 0 Post 240 mm

i. •AA

SS

'

I

C)

''

!•

A

501

•0

Post 240 mm

a

A6o1: •i0 .A5

• '40] :1 A

;A

t

Pre Post Post 1st 20 mm

9if•

AO)8Oj a) i

As

.•A .& SA ,a

A

S. A

A

101

240 mm

•A S.

..C 100

a

0-

A A

a)

S

100 -

C)

0C

i

80

0) 4•:-Aa)601

E o

0-

100

90-

a) Ai A $a)

£

.

A

•

A

Si

100 -

0

• S

A

A

'

Pre Post Post 1st 20 mm

Fig. 2. Change in platelet count and aggregation during dialysis with flat plate (•)

''

and hollow fiber (A) cuprophane membranes

Post

in 12 hemodialysis patients. (Horizontal lines depict median values.)

240 mm

Table 1. Change in malondialdehyde, thiol, /3-thromboglobulin, platelet factor 4 and factor VIII von Willebrand factor antigen during hemodialysis in six patients Post- 1st

Malondialdehyde nmol/ml Thiol u.mol/liter j3-thromboglobulin lU/mi

Platelet factor 4 p.g/liter

Factor VIII von Willebrand factor antigen % normal pooi

Post-240

Pre-HD

HD

Post-20 mm

8.40

8.55

(5.8—11.6)

(2.3—11.5)

9. 15a (7.1—12.3)

314.0

335.5

(188—444)

(74—477)

145.5

156.5

353•5a (294-531) 160.9a

(85.6—172.8)

(71.7—179.0)

(92.8—198.6)

37.1

37.1

37.2

(9.2—140.0)

(19.2—237.5)

(14.3—241.6) 219.Oa (100—400)

199.5

211.5

(94—390)

(75—411)

mm

97Ø (6.6—12.5)

4535

(351—747)

168.8a (98.6—203.6)

21.4 (3.4-223.5) 262.5a (102—545)

Values are expressed as medians and ranges. Normal ranges are: malondialdehyde (N = 23), 6.3 (5.4—8.7) nmol/ml; thiol (N = 23), 445 (358—501) smo1/l; /3-thromboglobulin, 10—40 lU/mI; platelet factor 4, 1.4-6.1 pg/liter. a P < 0.01, Wilcoxon versus pre-dialysis

due to hemoconcentration (r = 0.671, P < 0.01). The platelet release proteins, BTG and PF4, were elevated by comparison with normals pre-dialysis. BTG significantly increased during the course of dialysis, whereas PF4 decreased, but not significantly (Table 1). The ratio of BTG to PF4 increased during

dialysis from 3.7 (1.1 to 17.1) pre-dialysis to 7.3 (0.8 to 44.1) at

240 minutes (P < 0.01). FVIIIvWFAg was also increased pre-dialysis and rose significantly during dialysis (Table 1). Changes in all these parameters were not influenced by dialyzer

geometry or EPO therapy, and only plasma thiol correlated

446

Taylor et al: Dialyzer geometry during dialysis

significantly with the change in albumin due to hemoconcentration. Discussion

This study has demonstrated further evidence for the activation of granulocytes and platelets during the course of dialysis with cuprophane membranes. These findings, however, were not greatly influenced by dialyzer geometry or erythropoietin treatment.

increased pre-dialysis; too large to be removed by dialysis; and its structure is not changed during dialysis [181. In our study, the change in FVIIIvWFAg could not be explained by hemo-

concentration effects alone. It is possible that free radicals generated during hemodialysis are harmful to the patients' vaseulature, and that this damage is reflected by the rise in FVIIIvWFAg. Certainly, in patients who receive renal transplants, cessation of dialysis is associated with a reversal of uremic vasculopathy [19].

In keeping with the findings of Daugirdas et al [14], we In conclusion, dialyzer geometry (flat plate and hollow fiber) demonstrated a comparable fall in granulocyte count during does not influence granulocyte or platelet number and activadialysis with FP and HF dialyzers. In addition, concomitant tion during dialysis with cuprophane membranes. Patients with the fall in granulocyte count, we also observed a fall in treated with EPO may be more prone to extracorporeal thromgranulocyte aggregation in vitro. We believe this to represent bosis due to the effect of the increased red cell number on removal of a more active subpopulation of granulocytes from the circulation due to activation and aggregation in vivo with subsequent deposition in the pulmonary vasculature, The remaining cells were therefore more resistant to aggregation in vitro. Malondialdehyde levels were increased pre-dialysis and increased further during the course of dialysis. These observations have also been demonstrated by Miguel et al [15], and provide indirect evidence for enhanced free radical activity in hemodialysis patients, with further generation of reactive oxygen species during the course of dialysis due to granulocyte activation. We also found reduced plasma thiol levels predialysis, indicating lowered plasma antioxidant capacity. Although plasma thiol might have been expected to decrease during dialysis, due to the generation of free radicals from granulocytes, we in fact observed an increase. This correlated with the change in albumin concentration and was therefore most likely due to the effects of hemoconcentration. Changes in plasma MDA and thiol were not affected by dialyzer geometry or EPO therapy. Neither type of dialyzer was associated with a significant loss of platelets at the end of dialysis. Platelet aggregation (spontaneous and collagen-induced) decreased during the course of dialysis, reflecting the development of a population of "exhausted" platelets which had undergone the release reaction upon initial exposure to cuprophane, during the earlier stages of dialysis. In keeping with this, there was an increase in the ratio of BTG to PF4 during dialysis confirming platelet activation in vivo. Levels of BTG and PF4 were also significantly elevated in the pre-dialysis samples. Although clearance of BTG is impaired in renal failure [16], the elevated pre-dialysis PF4 would suggest a degree of non-dialysis-related platelet activation in these patients. Changes in platelet aggregation were not affected by dialyzer geometry, but were affected by EPO, such that higher spontaneous aggregation was observed by 20 minutes and at the end of dialysis in EPO treated patients compared to pretreatment. The increased red cell count produced by EPO has previously been shown by our group to enhance spontaneous platelet aggregation, possibly by a combination of physical and chemical effects [6], and may contribute to the increased tendency to extracorporeal thrombosis observed with this ther-

platelet function.

apy [17].

13. KAPLAN KL, OWEN J: Plasma level of /3-thromboglobulin and

The significance of the increase in FVIIIvWFAg during dialysis is unclear. Measurement of FVIIIvWFAg has been shown to be a marker of endothelial damage in other patients with chronic vascular disease [8]. In hemodialysis patients it is:

Acknowledgment We thank Gambro Limited, Sweden, for supporting this research. Reprint requests to Dr. J.E. Taylor, Renal Unit, Ninewells Hospital, Dundee, DDI 9SY, Scotland, United Kingdom.

References 1. LYNCH RE, BOSL RH, STREIFEL AJ, EBBEN JP, EHLERS SM, KJELLSTRAND CM: Dialysis thrombocytopenia: Parallel plate vs. hollow fiber dialysers. Trans Am So Artif Intern Organs 24:704— 707, 1978 2. Doccl D, TURd F, DEL VECCHIO C, BILANCIONI R, CENclorri L, PRETOLANI E: Hemodialysis-associated platelet loss: Study of the relative contribution of dialyzer membrane composition and geometry. mt J Artif Organs 7:337—340, 1984 3. CRADDOCK PR, FEHR J, DALMASSO AP, BRIGHAM K, JACOB HS: Hemodialysis leukopenia. Pulmonary vascular leukostasis resulting from complement activation by dialyzer cellophane membranes. J Clin Invest 59:879—888, 1977 4. SULTAN Y, LONDON GM, GOLDFARB B, TOULON P, MARCHAIS SJ:

Activation of platelets, coagulation and fibrinolysis in patients on long-term hemodialysis: Influence of cuprophan and polyacrylonitrite membranes. Nephrol Dial Transplant 5:362—368, 1990

5. BELCH JJF: The role of the white blood cell in arterial disease. Blood Coagulation and Fibrinolysis 1:183—192, 1990 6. TAYLOR JE, HENDERSON IS, STEWART WK, BELCH JJF: Erythro-

poietin and spontaneous platelet aggregation in haemodialysis patients. Lancet 338:1361—1362, 1991 7. VAN GEET C, HAUGLUSTAINE D, VERRESEN L, VANRUSSELT M,

VERMYLEN J: Haemostatic effects of recombinant human erythro-

poietin in chronic haemodialysis patients. Thromb Haemostas 64:7—10, 1990

8. BELCH JJF, ZOMA AA, RICHARDS IM, MCLAUGHLIN K, FORBES

CD, STURROCK RD: Vascular damage and factor-Vill-related antigen in the rheumatic diseases. Rheumatol mt 7:107—111, 1987 9. FISHER TC, BELCH JJF, BARBENEL JC, FISHER AC: In vitro human whole blood granulocyte aggregation. Cliii Sci 76:183—187, 1989 10. AUST S: Lipid peroxidation, in Handbook of Methods for Oxygen Free Radical Research, edited by GREEN WALD RA, Florida CRC Press, 1988, pp 203—207 11. BELCH JJF, CHOPRA M, HUTCI-IINSON 5, LORIMER R, STURROCK

RD, FORBES CD, SMITH E: Free radical pathology in chronic arterial disease. Free Rad Biol Med 6:375—378, 1989

12. ELLMAN GL: Tissue sulfiiydryl groups. Arch Biochem Biophys 82:70—77, 1969

platelet factor 4 as indices of platelet activation in vivo. Blood 57:199—202, 1981

14. DAUGIRDAS JT, POTEMPA LD, DINH N, GANDHI VC, IVANOVICH

PT, ING TS: Plate, coil, and hollow-fiber cuprammonium cellulose dialyzers: discrepancy between incidence of anaphylactic reactions

Taylor et a!: Dialyzer geometry during dialysis

447

and degree of complement activation. Artif Organs 11:140—143,

of one year's treatment in a European multicentre study of 150

1987 15. MIGUEL A, MIGUEL A, LINARES M, PEREZ A, MOLL R, SANCHIS J,

haeniodialysis-dependent patients. Nephrol Dial Transplant 4:979— 987, 1989 18. GRALNICK HR, MCKEOWN LP, WILLIAMS SB, SHAFER BC,

ESCOBEDO JM, MIGUEL-BOIUA JM: Evidence of an increased

susceptibility to lipid peroxidation in red blood cells of chronic renal failure patients. Nephron 50:64—65, 1988 16. DEPPERMANN D, ANDRASSY K, SEELIG H, RITZ E, POST D: Beta-thromboglobulin is elevated in renal failure without thrombosis. ThrombRes 17:63—69, 1980 17. SUNDAL E, KAESER U: Correction of anaemia of chronic renal failure with recombinant human erythropoietin: Safety and efficacy

PIERCE L: Plasma and platelet von Willebrand factor defects in uraemia. Am J Med 85:806—810, 1988 19. GILCHREST BA, ROWE JW, MIHM MC: Clinical and histological

skin changes in chronic renal failure: Evidence for a dialysisresistant transplant responsive microangiopathy. Lancet ii: 127 1— 1275, 1980