1672 Volume 3 ‘ Number 10 ‘ 1993

Effect of Recombinant Human Erythropoietin on PlateletProduction in Dialysis Patients1

Charles J. Kaupke,2 George C. Butler, and N.D. Vaziri

C.J. Kaupke, GO. Butler, ND. Vaziri, Division of Ne-phrology. Department of Medicine, University of Cali-fornia, Irvine, Irvine, CA, for the Epoetin Beta U.S. Mul-ticenter Trials

(J. Am. Soc. Nephrol. 1993; 3:1672-1679)

ABSTRACTTwo hundred forty-four anemic hemodialysis patientswere randomized into recombinant erythropoietinand placebo-treated groups during a 12-wk double-blind phase, followed by a 24-wk open-label period.Mean platelet count rose from the baseline value of242 x 109/L to 264 x 109/L on day 5 of epoetin

therapy (P < 0.001, paired I test). Mean plateletcount peaked at 290 x 109/L on day 40 and re-mained at a significantly elevated level below thepeak thereafter. The peak platelet count did notexceed the normal range in a majority of cases.

Platelet count was unaffected by placebo. Patients

without an erythropoietic response during the firstfew weeks of therapy exhibited a rise in platelet

count comparable to that in patients with a satisfac-tory erythropoiesis. Patients with low initial serum fer-

ritin concentrations had baseline platelet countscomparable to those with normal or high ferritin val-ues and showed a similar rise in platelet count during

therapy. As a group, patients with baseline plateletcounts above 400 x 109/L showed no rise in plateletcount, whereas those with normal or reduced plate-let counts showed a marked thrombopoietic re-

sponse to epoetin. Erythropoietin therapy did not

significantly alter the incidence of blood accessthrombosis when compared with placebo treatment.

Key Words: Dialysis, recombinant erythropoietin. platelets

C hronic renal failure is associated with impairedproduction of erythropoietin. the major erythro-

I Received October 5, 1992. Accepted January 15, 1993.2 Correspondence to Dr. C.J. Kaupke, Division of Nephroiogy, Department of

Medicine, University of California, irvine Medlcai Center, 101 The City Drive,Orange, CA 92668.

1046-6673/0310-1672503.00/0Journal of the American Society of NephroiogyCopyr$9ht C 1993 by the American Society of Nephroiogy

poietic growth factor. In most cases, the resultingerythropoietin deficiency leads to hypoproliferativeanemia, roughly correlating in severity with the de-

gree of renal insufficiency. Before the advent of re-combinant human erythropoietin (epoetin), a greatmajority of patients with ESRD exhibited moderateto severe anemia, with blood transfusion required in

approximately 25% of cases. However, widespreaduse of epoetin has greatly improved anemia-associ-

ated morbidity and markedly lowered the need forblood transfusion in the ESRD population (1-3).

Although the primary effect of epoetin is stimula-tion of erythropoiesis, in vitro studies have suggested

an effect on platelet production as well (4-7). Inaddition, in vivo animal experiments and human

trials of epoetin have raised the possibility of en-hanced platelet production and function (3.8-10).

The use of epoetin has been suspected to increasethe risk of thrombosis, particularly of the vascular

access and hemodiabysis circuit ( 1 1 . 1 2). Accordingly,a possible effect of epoetin on platelet production and

function is not only of physiologic interest but ofclinical significance as well. This article reports theeffect of epoetin therapy on platelet production and

the rate of vascular access thrombosis in chronichemodialysis patients studied during two identical,multicenter, phase III clinical trials.

METHODS

Patients

Two hundred forty-four chronic hemodialysis pa-tients (138 men and 106 women; mean age, 50.9 yr)were studied at 20 different sites. Patients under-went in-center hemodiabysis treatment three timesweekly.

To be included in the study. patients met the fob-

lowing criteria: baseline hemoglobin concentrationless than 85 g/L, clinical stability on hemodialysis

for at least 3 months, and age of older than 1 8 yr.Women were postmenopausal or treated with oralcontraceptives. Exclusion criteria included: poorlycontrolled hypertension, a history of myocardial in-

farction within 1 yr. a history of cerebral infarctionor seizure, significant hepatic or neoplastic disease,

sickle cell anemia, presence of antibodies to HIV.evidence of hemolysis or blood loss other thanthrough the dialyzer. evidence of aluminum toxicityor otherwise unexplained microcytosis. substanceabuse within 5 yr. and recent therapy with androgens

� � Kaupke et al

Journal of the American Society of Nephrology 1673

or IU�t��1V� � P�i11�ilt� with dI�b�t�mellitus were Included If metabolic control was sat-

isfactory.

Recombinant Human Erythropoietin

Recombinant human erythropoietin (epoetin beta;

Marogen#{174}; Chugai-Upjohn, Inc. . Rosemont. IL) was

provided by Chugai Pharmaceutical Co. , Ltd. Epoetin

beta has 1 65 amino acid residues and a molecular

size of approximately 30 kd, with 40% carbohydrate.Endogenous human erythropoietin, epoetin beta, and

epoetin alfa have identical biologic properties. iden-

tical amino acid sequences, and similar carbohydrate

structures. Epoetin beta is a lyophilized powder thatwas reconstituted with saline on the day of use andwas injected iv at the end of each dialysis session.

Study Design

The study protocol was approved by the Institu-

tional Review Board of each of the participating cen-ters, and all patients gave informed consent before

participating in the study.The study period consisted of two parts. The first

part was a 1 2-wk placebo-controlled. double-blindperiod. Patients were randomized at a 2: 1 ratio toreceive either epoetin. 100 lU/kg, or vehicle alonethree times per week. After 6 wk, the dose of epoetinwas increased to 1 50 lU/kg if the hemoglobin con-

centration was below 95 g/L and had not increasedby at least 20 g/L over baseline. The dose was reducedto 50 lU/kg for hemoglobin values greater than 125g/L and temporarily withheld when hemoglobin con-centration was greater than 1 35 gIL.

The second part was a 24-wk open-babel period.

Patients who completed the double-blind portion ofthe study could elect to enter the 24-wk open-babelperiod. Epoetin was administered to all patients atan initial dose of 50 lU/kg for 6 wk, regardless of

their hemoglobin levels at the end of the double-blindperiod.

Thus, patients who had been randomized to receiveplacebo during the initial 1 2 wk constituted a groupfor whom this open-label period was the first expo-sure to epoetin, at a dose of 50 lU/kg. In contrast,

patients randomized to receive epoetin at the begin-

ning of the double-blind study had received an initialdose of 1 00 lU/kg. allowing a comparison of theeffects of two different dosages and placebo.

After the first 6 wk of the open-babel period, dos-ages were adjusted at 4-wk intervals as follows: thedrug was continued at the same dose if hemoglobinconcentration was between 95 and 1 25 gIL or if ithad increased by at beast 20 g/L since the beginning

of the open-babel period; the dose was reduced by50% for hemoglobin concentrations above 1 25 g/Land temporarily withheld for values exceeding 1 35g/

L. if th� t1�fflO�!1OtJ1fleono�ntration r�nia!ne� �e1ow95 g/L and had not Increased by at least 20 gILcompared with the concentration at the beginning ofthe open-label period, the dose was increased by 50lU/kg.

Ferrous sulfate, 325 mg thrice daily, was admin-istered to all patients whose baseline serum ferritinwas below 300 ng/mL. Intravenous iron dextran and

blood transfusion were administered as needed. Anti-

hypertensive and other medications were institutedor continued as appropriate.

Hemoglobin. hematocrit. and platelet counts were

determined in predialysis arterial blood samples be-

fore the first dose of epoetin and were repeated atdayss, 12,26,40,54,82, 110, 138, 166, and250of

the study.Patients were carefully monitored for dialysis ar-

teriovenous access thrombosis during the studyperiod. In patients experiencing more than one epi-sode, the thromboses were considered to be separate

and distinct events only if they were separated by at

beast 4 days.

STATISTICAL ANALYSIS

Data for hematologic parameters were analyzed byt test, paired t test, and logistic or linear regression

analysis as appropriate. The rate of hemodiabysisaccess thrombosis, comparing the placebo and treat-ment groups and the multiple study sites, was com-

pared by logistic regression, the Mantel-Haenszeltest, and the Wilcoxon rank sum test. Survival analy-sis was used to compare differences between treat-ment and placebo groups and between the study

sites, with the occurrence of access thrombosis de-fined as the survival endpoint.

Results are presented as mean ± SD unless other-wise noted. A P value less than 0.05 was considered

significant.

RESULTS

Effect on Platelets

Platelet count immediately before epoetin admin-

istration was 242 x 1 09/L ± 1 03 x 1 09/L (range 67 x

1 09 to 837 x 1 09/L; N 244). Thrombocytopenia(platelet count < 1 50 X 1 09/L) was found in 33 pa-tients ( 1 3%), and thrombocytosis (platelet count>400 x 109/L) was found in 8 patients (3%). Nosignificant difference was present in baseline plate-

bet counts between the group treated with 1 00 lU/kgand those treated with 50 lU/kg as the initial dose.

Mean platelet count in patients receiving epoetinrose from the baseline value to 264 x 1 09/L at day 5

of therapy (P < 0.001 ; paired t test), reaching a peakvalue of 290 x 1 09/L (a 1 9.8% rise above baseline) atday 40 (Figure 1). Subsequently. it remained signifi-

�* � ±�

� �- EPOETIN OROUP

���CEBOGROUP�

Pecombinant Erythropoietin and Platelets

1674 Volume 3 ‘ Number 10 ‘ 1993

.- C

L�H0

rI�ED-9DSE PERIOD DOSE-ADJUSTMENT PERIOD

20 �o 40 50 6o

DAYS

70 80

Figure 1. Platelet counts in the group receiving epoetin compared with the group receiving placebo. Data are mean ± SE.‘ P< 0.05 between groups; t P< 0.05 versus baseline value.

cantly elevated at a level below the peak. For eachindividual patient, the peak platelet count occurredwithin the first 54 days of therapy in the majority of

cases. The mean of the individual peak values

achieved at any point during the course of epoetintherapy was 332 x 1 09/L, representing an averageincrease of 43% compared with the baseline value.Mean platelet count remained unchanged in patientsreceiving placebo (Figure 1).

Baseline platelet counts in patients with initialserum ferritin concentrations below 100 �gIL (N =

55) were not different from those of the remainder ofthe study group. Baseline platelet counts did not

correlate with serum ferritin concentrations for thestudy group as a whole. Moreover, changes in plateletcount did not correlate with changes in serum ferritinconcentration during epoetin therapy.

Serum ferritin concentration decreased signifi-cantly during epoetin therapy in patients with Initialvalues above 500 �zg/L (1 .883 ± 1 .768 Mg/L at base-

line to 998 ± 1 ,4 1 2 Mg/L at day 25 1 : P < 0.05) butremained largely unchanged In those with low base-line values receiving iron supplementation.

To examine the relationship between changes inerythrocyte and platelet production during the firstfew weeks of therapy. patients with complete retic-

ubocyte and platelet data during the first 26 dayswere divided Into two subgroups. The first subgroup(N = 47) consisted of patients who showed no initial

erythropoietic response to epoetin (I.e. , an un-changed or decreased reticulocyte count at day 5

compared with the baseline value). The secondsubgroup (N = 1 66) consisted of those exhibiting arise in reticulocyte count. A significant increase inplatelet count was noted in both subgroups at day 5.A similar comparison was made between 1 26 pa-tients exhibiting a hematocrit rise of at least 0.01(mean hematocrit increased from 0.21 to 0.25; P <

0.05) and 90 patients who showed no rise in hema-tocrit during the first 1 2 days of epoetin therapy(mean hematocrit, 0.24 versus 0.23; P = not signifi-

cant (NSJ). These two subgroups showed comparablebaseline platelet counts (239 x 109/L versus 241 x109/L; P = NS) and nearly identical rises in platelet

count with epoetin administration (Figure 2). No cor-relation was found between changes in hematocritand platelet count during the initial 6 wk of therapy.

Patients with initial platelet count below 1 50 x109/L (N = 33) showed a significant rise in meanplatelet count from 125 x 109/L at baseline to 191 x1 09/L by day 26 of epoetin therapy (P < 0.0 1 ). Plateletcount remained significantly above the baselinevalue throughout the study in this subgroup. In con-trast, patients with initial platelet counts greaterthan 350 x 109/L (mean 434 x 109/L; N 26) showedvirtually no change in platelet count, whereas thosewith baseline values between the two extremesshowed an Intermediate rise. The relative change inplatelet count during epoetin therapy correlated neg-

atively with the baseline platelet count (r = -0.26;

P = .001 comparing baseline to day 27 values) Fig-ure 3).

0

z000

Li

0�

II

280

260

240

220

200 I � I I I I I Effect on Erythrocytes

0.5

0.0

-0.5 -

0 200 400 600

BASELINE PLATELET COUNT (x109/L)

800 1000

‘�-�‘�‘ Kaupkeetal

Journal of the American Society of Nephrology 1675

Initial dosage of epoetin was 50 or 1 00 lU/kg, thrice

weekly. Periodic dosage adjustments were made ac-

cording to the protocol. Of the 244 patients initially

enrolled in the double-blind study, 185 completed

both the 1 2-wk double-blind and 24-wk open-label

300

0 5 10 15 20 25

DAYS OF EPOETIN THERAPY

Figure 2. Mean platelet countsforthe first 26 days of epoetin

therapy. Two subgroups are shown: group I, patients withno increase in hematocrit between days 0 and 12 of ther-apy (N= 90); group II, patients with at least a 0.01 increasein hematocrit between days 0 and 12 (N = 126). The T barsrepresent SE. P = NS for all points shown.

1.0

periods, at which time the mean dose was 91 lU/kg

thrice weekly (range, 6 to 300 lU/kg) (Figure 4). At

the conclusion of the study, 1 1 (6%) of 185 patients

were receiving doses greater than 250 lU/kg, whereas

42 (23%) of 185 were receiving less than 50 lU/kg.

Eighteen patients (8%) failed either to achieve a he-

mogbobin concentration above 95 g/L or to experience

a hemoglobin rise exceeding 20 g/L by day 82 of

epoetin therapy.

At the time of the first dosage adjustment (day 40),

1 8 (29%) of the 62 patients who first received epoetinat a dose of 50 lU/kg thrice weekly in the open-label

phase of the trial required an upward dosage adjust-ment, whereas 44 (7 1 %) of 62 required no change inepoetin dose; none required a dose reduction. In thedouble-blind phase, 3 1 (20%) of 1 56 patients receiv-ing epoetin. 100 lU/kg thrice weekly, required up-ward adjustment. 1 23 (79%) remained on the same

dose, and 2 ( 1 %) required downward adjustment. Asexpected. the group receiving placebo failed to exhibit

a discernible erythropoietic response to either theinitial dose or subsequently increased doses of thevehicle.

A significant increase in mean reticubocyte countwas observed by day 5 of epoetin therapy at bothinitial dosages (P < 0.00 1 ). Reticubocyte counts in the1 00-lU/kg group increased steadily from 1 .7 ± 1.3%to a peak of 4.2 ± 4.7% at day 26 and remainedsignificantly above the baseline thereafter. Thegroup receiving 50 lU/kg as initial dosage followed asimilar course, exhibiting a rise from the baseline

Figure 3. Relative change in platelet count from baseline to day 27 of epoetin beta therapy versus baseline platelet count.The linear regression line is shown, with 95% confidence intervals. Relative change in platelet count = 0.47-0.00 14 (baselineplatelet count). P< 0.001; r= -0.26.

40 r

30

20

10

0

EPOETIN DOSE (lu/KG 11W)

0-25 25-49 50-99 100-149 150-199 200-249 250+

Figure 4. Distribution of epoetin dosages at the conclusion of the study. The drug was administered iv, three times weekly.with dosages adjusted to maintain hemoglobin concentration between 95 and 125 g/L.

4-�

F iFS EDSE PER CD 55SF ADJUSTMENT PERIOD

.- . . . 1 � . I I 1

.. . (\ : �I �O 40 50 50 70 80

Figure 5. Mean hematocrit values during the first 82 days of epoetin administration. Three groups are shown: patients

receiving initial dosages of 50 lU/kg and 100 lU/kg of recombinant erythropoietin and the group receiving placebo. ‘ P <

0.05, for each of the epoetin-treated groups compared with placebo. # P < 0.05, comparing the 50 and 100 lU/kg groupswith each other. Data shown are mean ± SE.

1676 Volume 3 . Number 10 ‘ 1993

Recombinant Erythropoietin and Platelets

0.:3

P

0

0

:3

84

51)H-

z50

F-

0�

�o,-’ie�, �: .�L: �

��LACEBC�

value of 1 .8 ± 1 .7% to a peak reticubocyte count of3.3 ± 4.1%.

With epoetin administration, hemoglobin and hem-

atocrit rose to values that were significantly greaterthan those found at baseline and in comparison to

those of the placebo-treated group at day 1 2 (Figure

5). The observed rise was greater in patients given1 00 lU/kg than that found in patients treated with

50 lU/kg. Consequently, hemoglobin and hematocritvalues in the 1 00-lU/kg group were significantlygreater than those in the 50-lu/kg group at days 26

to 82.

- .. -�-. ... -. Kaupkeetal

Journal of the American Society of Nephrology 1677

Reticubocyte count, hemoglobin, and hematocrit

values did not change in patients receiving placebo.

Effect on the Rate of Venous AccessThrombosis

During the 83-day double-blind period, 2 1 epoetin-

treated patients (1 2%) experienced 3 1 episodes ofgraft thrombosis (0.58 events/patient year). In theplacebo-treated group, six patients (8%) exhibited

single episodes of access thrombosis (0.34 events/patient year) during the observation period (P = NScomparing epoetin-treated and placebo groups). Therate of access thrombosis observed during the open-

label period (0.57 events/patient year) was compa-

rable to that observed in the epoetin-treated patientsduring the double-blind period. Access survival, de-fined as time to first thrombotic event, did not differ

significantly between the epoetin-treated and pla-cebo-treated groups. Stratification of the data bytreatment site demonstrated no significant intersitedifferences in the rate of thrombosis or access sur-vivab time. Moreover, access thrombosis was not cor-related with baseline or subsequent platelet count orhematocrit during the study period.

DISCUSSION

Erythropoietin, a 30-kd gbycoprotein, is the pri-

mary hormonal regulator of erythropoiesis. Erythro-

poietin promotes proliferation and differentiation oferythrocyte progenitor cells, leading to productionand release of mature reticulocytes. Concurrently,

erythropoietin stimulates transferrin receptor

expression and hemoglobin synthesis (13).A number of observations suggest that, in addition

to its erythrocytopoietic property. erythropoietin mayalso influence platelet production. This study exam-med the effect of epoetin on platelet count, erythro-poiesis, and incidence of vascular access thrombosisin a barge group of anemic patients maintained on

hemodialysis. As expected, a significant increase inreticubocyte count was observed within the first week

of therapy in a great majority of patients, preceding

a detectable rise in hemoglobin and hematocrit val-ues. In addition, our patients exhibited a rise in plate-let count that reached statistical significance in 1

wk. On the average. platelet count rose by nearby20% and remained significantly elevated at a level

below the peak for the duration of the study. Theobserved rise was confined to the normal limits in

the large majority of cases. These results are sugges-

tive of a persistent stimubatory effect of epoetin onplatelet production.

A significant rise in platelet count was foundwhether or not reticubocytosis had occurred during

the first week of epoetin administration. Likewise,platelet count remained significantly above the base-

line with epoetin administration during the ensuing

3 wk whether a rise in hematocrit or hemoglobin

concentration had occurred or not. Moreover,

changes in platelet count did not correlate with the

resultant changes in hematocrit and hemoglobin con-

centration. It thus appears that the effect of epoetin

on platelet production is largely independent of the

associated erythropoietic response. Although hema-

tocrit rose somewhat more rapidly with the higherinitial dose of epoetin (1 00 lU/kg) than with the bower

initial dose (50 lU/kg). the rate of rise in platelet countwas similar with both dosages. This observation sug-

gests that the maximum thrombopoietic effect maybe achieved at an iv dose of 50 lU/kg or bess.

Patients with initial platelet counts below the nor-

mab range showed a greater rise, reaching normalvalues within a few weeks of therapy in most cases,as compared with those with initial thrombocytosiswho, as a group. showed no significant increase in

platelet count.

Abnormal platelet counts have been described inpatients with iron storage or availability disorders( 1 4. 1 5). However, in our patients. the size of tissueiron stores (as indicated by serum ferritin concentra-

tion) at the beginning of the study did not correlatewith the subsequent changes in platelet count during

epoetin administration. Moreover, platelet count in asubgroup of 55 patients with iron depletion was not

significantly different from that seen in their iron-

replete counterparts. Hence. the associated rise inplatelet production is not mediated by epoetin-in-duced iron depletion and appears to represent a direct

effect on megakaryocytopoiesis instead.

The regulation of megakaryocytopoiesis is complex

and incompletely understood, apparently involving avariety of humoral factors. Among these are interleu-kins-la,-3,-4, and -5, granulocyte macrophage-col-ony-stimubating factor, transforming growth factor

beta, and thrombopoietin (4,5. 1 6). Thrombopoietin.whose structure has not yet been elucidated, is the

major growth factor for platelet progenitors. Whereasthrombopoietin activity has been shown to increase

in thrombocytopenia (1 6). the nature of the feedback

mechanism remains unknown.

Several clinical and experimental observationshave suggested a role for erythropoietin in the regu-

lation of platelet production. In vitro experimentswith various preparations of erythropoietin havedemonstrated increased megakaryocyte colony for-

mation in several assay systems with plasma clot,agar, and serum-free culture media (4-7,17-20). Re-

cent studies with recombinant erythropoietin have

shown stimulation of megakaryocyte colony growth

and promotion of the differentiation of murine mega-karyocytes in serum-free liquid cultures (6,2 1 ). In

several studies, recombinant erythropoietin alone

stimulated megakaryocytopoiesis (6.9). whereas in

Recombinant Erythropoietin and Platelets

1678 Volume 3 . Number 10 . 1993

others, an effect was seen only in combination with

other cytokines such as interleukin-la, interleukin-

3, or granubocyte macrophage-colony-stimulating

factor (5,22). In contrast, a few studies found eryth-ropoietin to have no apparent effect (23.24).

In vivo studies have yielded similar results. The

administration of epoetin to rats resulted in a dose-dependent increase in both reticubocyte and plateletcounts, with an effect on megakaryocytes observed

as early as 24 h after the initial dose. Higher doseswere required to significantly raise platelet count ascompared with those required to produce reticulocy-

tosis. Peak values for both reticubocytes and platelets

were observed after 4 to 6 days of epoetin adminis-

tration, with platelet count rising by 20% and sub-sequently returning to baseline bevels despite contin-

ued epoetin administration (8).Clinical trials in anemic ESRD patients have dem-

onstrated the efficacy of epoetin in raising hemoglo-bin and hematocrit, but the reported effects on plate-lets have been inconsistent. Although some studies

have reported no change in platelet count (1), othershave demonstrated rises in the order of 20% or more

(3.25,26).The mechanism by which erythropoietin alters

platelet production is unknown. However, a direct

effect on platelet precursors is suggested by the dem-onstration that megakaryocytes express erythropoi-

etin receptors and that epoetin binds to the high-

affinity receptors on murine megakaryocytes. The

apparent receptor density is similar to that of pro-normoblasts (6. 500 receptors/cell) and increases

with advancing megakaryocyte differentiation. Ma-ture murine platelets. however, are devoid of eryth-ropoietin receptors (27). Thrombopoietin. the major

regulator of terminal megakaryocyte differentiation,

is believed to act through as-yet-unidentified cell

membrane receptors. Whether erythropoietin and

thrombopoietin share a common receptor is uncer-

tam. Likewise, the interaction, if any. between eryth-ropoietin and thrombopoietin remains unclear. How-

ever, the observation that epoetin promoted a marked

rise in platelet count in patients with low plateletcount and prcsumabby high thrombopoietin. but hadno effect in those with thrombocytosis. is of interest.

These observations may suggest that the effect ofepoetin on thrombopoiesis may be influenced by thepresence of thrombopoietin. Although the molecular

structure of thrombopoietin has not yet been fullyelucidated, some investigators have postulated struc-tural similarities between erythropoietin and throm-

bopoietin (9,16).In conclusion, this study demonstrates that, in ad-

dition to promoting erythropoiesis. epoetin enhancesplatelet production. The associated rise in platelet

count is evident within the first week of epoetin

therapy. Platelet count generally increases by as

much as 20% and remains elevated throughout the

course of therapy. The epoetin-induced thrombo-poiesis is independent of the associated erythropoi-

etic response and ferritin values but is inverselyrebated to the baseline platelet concentration. Finally.

epoetin administration does not appear to signifi-cantly increase the rate of hemodialysis accessthrombosis.

ACKNOWLEDGMENTS

The principal coinvestigators participating in this study were: Paul

A. Abraham, MD, Hennepin County Medical Center. Minneapolis.

MN; Sergio R. Acchiardo. MD. University of Tennessee, Memphis.

TN; Howard Alfred, MD, Worcester Memorial Hospital. Worcester.

MA; William Bennett, MD, Oregon Health Sciences University. Port-

land. OR; Anatole Besarab, MD, Thomas Jefferson Medical College.

Philadelphia. PA; Walter Flamenbaum, MD, Health and Sciences

Research, Inc. , Englewood. NJ; Theodore S. Herman. MD, Buffalo,

NY; Abdullah Iravani, MD, University of Florida, College of Medicine.

Gainesville. FL; John Josselson. MD. University of Maryland Medical

Center, Baltimore, MD; Benjamin Levy. MD, National Medical Re-

search Corporation. Hartford, CT; E. Paul MacCarthy. MD. Dialysis

Clinic Incorporated. Cincinnati, OH; Douglas Maxwell. MD, Indiana

University Hospitals. Indianapolis. IN; Kenneth Miller. MD, Lutheran

General Hospital. Park Ridge. IL; Marshall R. Sack. MD. The Austin

Diagnostic Clinic. Austin, TX: Richard Solomon. MD, Westchester

Medical Center, Valhalla, NY; Bruce S. Spinowitz. MD, Booth Me-

mortal Medical Center, Flushing. NY; Arkady Synhavsky. MD. Kid-

ney Disease and Critical Care Associates. Minneapolis. MN; ND.

Vazirt, MD. University of California Irvine. Irvine, CA; P. Viasses.

PharmD, Thomas Jefferson Medical College. Philadelphia. PA: Beat

von Albertini, MD. The George Washington University Medical Cen-

ter, Washington, DC; Michael Walczyk. MD. Northwest Renal Clinic.

Portland, OR; Duane G. Wombolt, MD. Clinical Research Association

of Tidewater, Norfolk, VA; John J. Whalen. MD. Besselaar Associ-

ates, Princeton Forrestal Center, Princeton, NJ; and Barry C. Wood.

MD. Kansas City. MO. The authors thank Dr. Howard Tucker and

James Ashurst for their assistance with statistical analysis of the

data.

REFERENCES

1 . Eschbach JW, Egrie JC, Downing MR. BrowneJK, Adamson JW: Correction of the anemia ofend-stage renal disease with recombinant hu-man erythropoietin: results of a combined phaseI and II clinical trial. N Engl J Med 1987;316:73-78.

2. Erslev AJ: Erythropoietin. N Engb J Med 1991;324:1339-1344.

3. Eschbach JW. Abclulhadi MH. Browne JK, etat.: Recombinant human erythropoletin in ane-mic patients with end-stage renal disease: Re-suits of a phase III multicenter clinical trial. AnnIntern Med 1989:111:992-1000.

4. Bruno E, Briddell R, Hoffman R: Effect of re-combinant and purified hematopoietic growthfactors on human megakaryocyte colony forma-tion. Exp Hematol 1988;16:371-377.

5. Teramura M, Mizoguchi H: The effect of cyto-kines on the pboidy of megakaryocytes. mt J CellCloning 1 990;8:245-252.

6. Ishibashi T, Koziol JA, Burstein SA: Humanrecombinant erythropoietin promotes differen-tiation of murine megakaryocytes in vitro. J ClinInvest 1 987;79:286-289.

7. Clark DA, Dessypris EN: Effects of recombinant

Kaupke et al

Journal of the American Society of Nephrology 1679

erythropoietin on murine megakaryocyte colonyformation in vitro. J Lab Clin Med 1 986; 108:423-429.

8. Berridge MV, Fraser JK, Carter JM, Lin F-K:Effects of recombinant human erythropoietin onmegakaryocytes and on platelet production inthe rat. Blood 1988;72:970-977.

9. McDonald TP, Cottrell MB, Chit RE, Cullen WC,Lin F-K: High doses of recombinant erythropoi-etin stimulate platelet production in mice. ExpHematol 1987;15:719-721.

10. Moia M, Mannucci PM, Vizzotto L, Casati S.Cattaneo M, Ponticelli C: Improvement in thehemostatic defect of uraemia after treatmentwith recombinant human erythropoietin. Lancet1987;2:1227-1229.

1 1 . Shinaberger JH, Miller JH, Gardner PW: Eryth-ropoietin alert: Risks of high hematocrit hemo-dialysis. Trans Am Soc Artif Intern Organs1988;34:179-184.

12. Eschbach JW, Adamson JW: Recombinant hu-man erythropoietin: Implications for nephrobogy.Am J Kidney Dis 1 988; 1:203-209.

1 3. Nimer SD: Molecular biology and hematology oferythropoietin. In: Nissenson AR. moderator. Re-combinant human erythropoietin and renal ane-mia: Molecular biology, clinical efficacy, andnervous system effects. Ann Intern Med 1 99 1;114:402-407.

1 4. Baldwin JG, Sellers GA: Profound thrombocy-tosis in a patient with iron-back anemia. SouthMed J 1989;82:662-663.

1 5. Soff GA, Levin J: Thrombocytopenia associatedwith repletion of iron in iron-deficiency anemia.Am J Med Sci 1988;295:35-39.

1 6. Vainchenker W, Kieffer N: Human megakary-ocytopoiesis: in vitro regulation and character-ization of megakaryocyte precursor cells by dif-ferentiation markers. Blood Rev 1988;2:102-107.

1 7. Williams N, Jackson H, Iscove NN, Dukes PP:The role of erythropoietin. thrombopoietic stim-ulating factor, and myeboid colony-stimulatingfactors on murine megakaryocyte colony for-mation. Exp Hematol 1984; 12:734-737.

18. Mizoguchi H, Fujiwara Y, Sasaki R, Chiba H.The effect of interbeukin 3 and erythropoietin onmurine megakaryocyte colony formation. In:Levine RF, Williams N, Levin J, Evatt BL, eds.Megakaryocyte Development and Function. NewYoric: Liss; 1986:111-113.

19. McLeod DL, Shreeve MM, Axelrad AA: Induc-tion of megakaryocyte colonies with platelet for-mation in vitro. Nature (Lond) 1 976;26 1:492-495.

20. Dukes PP. Egrie JC, Strickland TW, BrowneJK, Lm F-K: Megakaryocyte colony stimulatingactivity of recombinant human and monkeyerythropoietin. In: Levine RF. Williams N. LevinJ, Evatt BL. eds. Megakaryocyte Developmentand Function. New York: Liss; 1986:105-109.

2 1 . Dessypris EN, Gleaton JH, Armstrong OL: Ef-fect of human recombinant erythropoietin onhuman megakaryocyte colony formation in vitro.Br J Hematol 1987:65:265-269.

22. Bruno E, Miller ME, Hoffman R: Interacting cy-tokines regulate in vitro human megakaryocy-topoiesis. Blood 1989:73:671-677.

23. Grossi A, Vannuchi AM, Rafanelli D, FerriniPR: Recombinant human erythropoietin has bit-tle influence on megakaryocytopoiesis in mice.BrJ Hematol 1989;71:4653-4658.

24. Kamatsu N, Suda T, Moroi M, et at.: Growthand differentiation of a human megakaryoblas-tic cell line, CMK. Blood 1989:74:42-48.

25. Schaefer RM, Kuerner B, Zech M, Denninger G,Borneff C, Heidland A: Treatment of the anemiaof hemodiabysis patients with recombinant hu-man erythropoietin. mt J Artif Org 1 988; 1 1:249-254.

26. Faulds D, Sorkin EM: Epoetin (recombinant hu-man erythropoietin): A review of its pharmaco-dynamic and pharmacokinetic properties andtherapeutic potential in anaemia and the stim-ulation of erythropoiesis. Drugs 1989:38:863-899.

27. Fraser JK, Tan AS, Lm F-K, Berridge MV:Expression of specific high-affinity binding sitesfor erythropoietin on rat and mouse megakary-ocytes. Exp Hematol 1 989; 17:10-16.