Should Hemoglobin be Normalized in Patientswith Chronic Kidney Disease?

Lesley Stevens,* Caroline Stigant,² and Adeera Levin**Department of Medicine, Nephrology Division, University of British Columbia, Vancouver, British Columbia,and ²Department of Medicine, Nephrology Division, University of Toronto, Toronto, Ontario, Canada

ABSTRACT

In the last decade the nephrology community has learnedmuchabout the impact of anemia on patients with kidney disease.Therapy of anemia can correct many of the symptoms whichseriously compromise patient function. Despite the obviousbene®ts, controversy continues regarding the optimal targethemoglobin concentration both in patients prior to dialysis andin dialysis populations. In this editorial we review the clinicaldata that contribute to this controversy and the physiologicconcepts underlying the treatment of anemia. Furthermore, wediscuss the need to individualize hemoglobin targets for speci®cpatient populations and the importance of early identi®cationand treatment of anemia in patients with kidney disease. Theeconomic impact of normalizing hemoglobin with the use oferythropoietin and intravenous or oral iron has a�ected clinicalpractice over the last decade. Current guidelines published by

KidneyDiseaseOutcomes andQuality Initiative (KDOQI), theEuropean Working Group on Anemia Management, and theCanadian Society of Nephrology all recommend target hemo-globin concentrations and thresholds for initiation of therapyand also suggest the need for reevaluation of current targets inlight of new evidence. This editorial supports those guidelinesand challenges the reader to critically evaluate current practicein the context of the accumulating data and the physiologicprinciples discussed herein. The therapy of anemia in patientswith chronic kidney disease (CKD) is becoming increasinglysophisticated and is an essential component of care in patientswith CKD. However, the e�ects of therapy will be mostimpressive when accompanied by the optimal care of allhemodynamic and metabolic abnormalities that are associatedwith CKD.

Treatment of anemia with recombinant human ery-thropoietin has resulted in signi®cant symptomaticimprovement in patients with chronic kidney disease(CKD). Over the past decade, numerous studies havedemonstrated multiple bene®ts of this treatment inclu-ding improved quality of life, brain and cognitivefunction, exercise capacity and cardiac function, andreductions in hospitalization and mortality. Despitethese obvious bene®ts, controversy continues regardingthe optimal target hemoglobin concentration both inpatients prior to dialysis and in dialysis populations. Inthis editorial we will brie¯y review the clinical data thatcontribute to this controversy, as well as some of thephysiologic concepts underlying the treatment ofanemia. Furthermore, our recommendations are basedon the need to individualize hemoglobin targets forspeci®c patient populations and the importance of earlyidenti®cation and treatment of anemia in patients withkidney disease.The economic impact of normalizing hemoglobinwith

the use of erythropoietin and intravenous or oral ironhasa�ected clinical practice over the last decade (1). Some ofthe controversy about higher hemoglobinmay be relatedto cost rather than physiology or appropriate therapy.Current guidelines published by the Kidney Disease

Outcomes and Quality Initiative (KDOQI), the Euro-peanWorking Group on AnemiaManagement, and theCanadian Society of Nephrology all recommend targethemoglobin concentrations and thresholds for initiationof therapy, and also suggest the need for reevaluation ofcurrent targets in light of new evidence. While recom-mendations di�ering from those published by theseesteemedgroups are not o�eredhere,wedo challenge thereader to critically evaluate currentpractice in the contextof the accumulating data and the physiologic principlesdiscussed herein.

Higher Hemoglobin Targets: The Evidence

Historical Perspective

The original clinical studies using erythropoietin weredesigned to double hemoglobin levels in dialysis patientsfrom a baseline of 45±60 g/L to approximately 100 g/L.This targetwas chosenas a reasonablephysiologic target,below which symptoms were thought to be exacerbated.These initial studies predate the current well-establishedlink betweenhemoglobin andmorbidity andmortality indialysis patients (1±4). The initial well-designed Cana-dian (5) randomized asymptomatic, young, nondiabeticdialysis patients to di�erent target hemoglobin concen-trations. The investigators were unable to demonstratesigni®cant improvements in these patients above a meanhemoglobin concentration of 102 g/L (versus a meanhemoglobin of 117 g/L) with respect to quality of life orexercise capacity. Furthermore, given the initial reportsof hypertension and seizures, it was felt that the dialysis

Address correspondence to: Adeera Levin, MD, FRCPC,Chronic Kidney Disease Care Clinic, St. Paul's Hospital,University of British Columbia, Vancouver, British Columbia,Canada, or e-mail: alevin@providencehealth.bc.c.

Seminars in DialysisÐVol 15, No 1 (January±February) 2002pp. 8±13

8

patient had adapted to lower hemoglobin levels, and thusthat attainment of a lower hemoglobin target would bereasonable. No cardiac outcomes were measured in thisstudy.As experience with erythropoietin was obtained, and

more analyses of existing databases and studies wereperformed, the possibility of improvements in numerousaspects of patients' functioning by increasing hemoglo-bin was hotly debated. Furthermore, the link betweencardiac disease, mortality, and anemia became wellestablished, and the goal of reduction in mortalitythrough treatment of anemia became the target ofpractice.Several recent trials examine the e�ect of normaliza-

tion of hemoglobin with erythropoietin therapy in boththe dialysis and predialysis populations (5±12). The®ndings of these studies are extremely favorable regard-ing improvement in exercise capacity, quality of life, andcognitive function with treatment of anemia. Regardingthese endpoints, these trials generally support higherhemoglobin levels as anappropriate therapeutic strategy.However, the results of the Normal Hematocrit Trial (6)have fueled the debate as to whether we can in factrecommend hemoglobin normalization in the entireCKD cohort.

Clinical Trials Assessing HemoglobinNormalization and Cardiovascular Disease

in Dialysis Patients

The Normal Hematocrit Trial was a randomized,prospective, open-label trial of hemodialysis patientswith symptomatic ischemic heart disease or congestiveheart failure. The study population had been on dialysisfor approximately 3 years and had a mean hematocrit(Hct) of 30.5% at the start of the study. A total of 1233hemodialysis patients were randomized to either anormal Hct (42%) or conventional Hct (30%). Theprimary end point was length of time to death or ®rstnonfatal myocardial infarction. The study was stoppedearly because, according to intention to treat analysis,patients in the higher Hct target group had a 7% excessmortality. Yet subsequent regression analysis demon-strated that mortality decreased with increasing hemo-globin concentration in both groups. The results of thisregression analysis are, however, misleading due to asurvival bias and the concept of competing risksÐthosewhodied earlydid sobefore they couldachieve thehigherHct. Thus this post hoc analysis is not proof that higherhemoglobin is good, only that if one lives long enoughone can achieve a higher hemoglobin.Factors that confound the interpretation of the

Normal Hematocrit Trial include the fact that thenormal hematocrit group had a decline inKt/V (whereastheconventionalHctgrouphadahigherKt/V) comparedto baseline, increased vascular access thrombosis, andincreased intravenous iron dextran administration. It ispossible that these three factors contributed to pooreroutcomes: vascular access thromboses led to a reductionin Kt/V that ultimately resulted in poorer dialysis and amore in¯ammatory state.Theneed formore iron therapy

during in¯ammatory activity may have reduced theability to utilize the iron, leading to tissue deposition andsubsequent damage. Furthermore, intravenous ironadministration itself may result in an increased risk ofinfection and subsequent death (13±15). Nonetheless,this well-conducted randomized, controlled trial invol-ving more than 1200 patients cannot be dismissed, andthus the normalization of hemoglobin in patients withsymptomatic cardiac disease is not supported.The next question is if patients with less advanced

cardiac disease bene®t from normalization of hemoglo-bin. This is addressed by Foley et al. (8) in the CanadianStudy of Normalization of Hemoglobin, a randomized,prospective, open-label study of patients with asympto-matic cardiomyopathy. In this study a total of 146hemodialysis patients were randomized to target hemo-globin concentrations of 110 g/Lor 135 g/L. This cohorthad been on dialysis for a mean of 4.5 years and had amean hemoglobin concentration of 100 g/L at the startof the study. Patients had established left ventricularhypertrophy (LVH) at study entry, either eccentric orconcentric enlargement, andwere randomized accordingto their left ventricular geometry. The study demonstra-ted a trend toward a decrease in the progression of leftventricular dilation in patients with normal chamber sizeat the beginning of the study, but no positive impact ofhemoglobin normalization with respect to left ventriclesize in those with already hypertrophied hearts. Inaddition, this study was able to demonstrate an im-provement in quality of life in patients assigned to thehigher hemoglobin group. This is consistent with the®ndings of multiple other studies, all with small samplesize, that also demonstrate improved quality of life withtreatment of anemia.Together, the Besarab et al. and Foley et al. studies

emphasize the importance of de®ning the study groupand of de®ning important outcomes. While reduction inmortality is our ultimate goal in the treatment of dialysispatients, improvements in quality of life, exercisecapacity, and other aspects of functioning are alsoimportant goals. It may be that in patients at the end oftheir kidney and cardiac ``life cycle,'' improving longevityby attaining higher hemoglobin is not possible. Otherpopulations, however, may bene®t from aggressiveanemia therapy.

Clinical Trials Assessing Exercise Capacity,Cognitive Function, and Quality of Life

Several recent trials examine the e�ect of normaliza-tion of hemoglobin with erythropoietin therapy onexercise capacity, cognitive function, and quality of life(11,16±18). These studies generally involve a smallnumber of patients receiving hemodialysis. Speci®coutcomes include exercise capacity as measured by workoutput and oxygen consumption; brain and cognitivefunction as measured with evoked potentials, EEG, andsleep; and quality of life as measured with standardizedquestionnaires. Despite the improvements in physicaland cognitive function and quality of life noted in thesestudies, none of these parameters improved to the level ofnondialysis controls.

NORMALIZED HEMOGLOBIN IN CKD 9

These ®ndings remind us, as both clinicians andresearchers, that the impact of uremia on the health andfunction of patients with CKD is sustained beyond thechanges in any speci®c laboratory parameter.

Physiology of Anemia

Anemia in the Non-CKD Population

In all people with anemia, maintenance of adequatetissueoxygenation is achievedbybothnonhemodynamicand hemodynamic adaptations. Nonhemodynamicadaptations include increases in erythropoietin produc-tion and increases in the intraerythrocytic concentrationsof 2,3-diphosphoglycerate (2,3-DPG) (19). Hemody-namic adaptations begin to occur when hemoglobinconcentrations decline to less than 100 g/L or, notably,in nonrest conditions at hemoglobin concentrationsbetween 100 and 140 g/L. Brie¯y, such adaptationsinclude an increase in cardiac preload and a decrease insystemic vascular resistance leading to decreased after-load, both of which contribute to the high cardiac outputstate. Although the cardiovascular responses are appro-priate, long-term activation leads to left ventricularremodeling, including initial dilation due to the increasein preload,with subsequent hypertrophy in an attempt todecrease the highwall tension of the dilated left ventricle.In the non-kidney disease population, these changes arefrequently reversible.

Anemia in the CKD Population

In patients with CKD, the response to anemia isdiminished and altered compared to that of the generalpopulation. Although erythropoietin production doesrise as the hemoglobin falls, the rise is not as high asexpected for the degree of anemia. The relative failure oferythropoietin production is linked to loss of kidneymass, whereas erythropoietin resistance is likely secon-dary to hyperparathyroidism or the in¯ammatory stateassociated with uremia. Preliminary evidence in supportof the multifactorial nature of erythropoietin de®ciency(absolute and relative) comes from observations thatpatients on daily or nocturnal hemodialysis requiresigni®cantly less erythropoietin than their counterpartson conventional hemodialysis. Furthermore, some dailyor nocturnal hemodialysis patients do not require anyerythropoietin therapy to maintain normal hemoglobinlevels (20).Anemia is a signi®cant contributor to the development

of LVH in patients with CKD. LVH has been docu-mented in approximately 80%of patients with end-stagerenal disease (ESRD) and in nearly 40% of those withmild to moderate renal insu�ciency (21,22). It is notknown what level of hemoglobin triggers hemodynamiccompensatory changes in CKD, but in a multicenterstudy of patients with early kidney disease, it was thedecrease in hemoglobin rather than the absolute levelwhichwas identi®edas the importantdeterminant for thedevelopment of LVH. In fact, mean hemoglobin levelswere at target values for dialysis patients (121 g/L) and

40% of patients demonstrated evidence of LVH despitehemoglobin levels of more than 110 g/L (22).Since uremia is a hypermetabolic state (23), we

hypothesize that hemodynamic changes in response toanemia may occur at higher hemoglobin concentrationsthan in the non-CKDcohort. It is certainly plausible thatthe heart, in a hypermetabolic milieu, is more susceptibleto the hemodynamic e�ects of anemia. Supporting thishypothesis is animal evidence that hearts of uremic ratsare more vulnerable to ischemic damage due to rapiddegradation of energy-rich nucleotides and diminishedexpression of insulin-sensitive glucose transporter(24,25). Furthermore, in contrast to the general popula-tion, there are other potential contributors to thedevelopment of cardiac disease in patients with CKD.Hypertension, volume expansion, diabetes, hyperpara-thyroidism, and uremia, in addition to anemia, allcontribute to the high prevalence of LVH observed inCKD (26). Moreover, the presence of these comorbidconditions contributes to intermyocardiocystic ®brosis,calcium deposition, increased left ventricle sti�ness, andarteriosclerosis that is commonly observed in the heartsof patients with CKD. It is the chronic exposure to thesemaladaptive processes that likely prevent the reversibilityof LVH in uremic individuals.We are also beginning to understand some of the

e�ects of anemia on muscle and brain function. Prelim-inary evidence suggests that anemia may contribute toabnormal sodium-potassium (Na-K) ATPase activity,which is in part responsible for disordered potassiumregulation that may be responsible for the early fatiguenoted in patients with kidney failure. Similarly, musclesof uremic rats demonstrate impaired Na-K ATPaseactivity (27). Finally, erythropoietin therapy, independ-ent of hemoglobin level, is antiapoptotic to neurons (28)and is known to decrease ischemic injury after traumaand stroke (29).In summary, there is ample clinical and physiologic

evidence to support aggressive therapy of anemia inpatients with CKD. The only study with a negativeoutcome was the Besarab study; thus we must carefullyconsider patient subpopulations and comorbidity whenfaced with therapeutic decisions.

Primary Prevention of Anemia

The di�erent trials that have assessed cardiovascularoutcomes with normalization of hemoglobin with ery-thropoietin have studied patients at various stages andseverity of cardiac disease. Besarab et al. (6) studiedsymptomatic patients with advanced cardiac disease,whereas Foley et al. (8) studied those with early butasymptomatic cardiac disease. Levin et al. (22) thenexamined predialysis patients and determined that thedecline in hemoglobin, rather than the actual hemoglo-bin level, is the major determinant of left ventriculargrowth. It may be that correction of anemia in the earlystages of kidney disease allows for improvements incardiac status, whereas anemia correction after pro-longed exposure to the uremic milieu, and thus after thedevelopment of irreversible cardiac structural changes, is

Stevens et al.10

not e�cacious in improving cardiac function. Indeed,several small uncontrolled studies in patients withmoderate kidney impairment demonstrate regression ofleft ventricular mass with the treatment of anemia witherythropoietin therapy (9,30,31).As in cardiology, with a shift in emphasis from

secondary to primary prevention, so there is in nephrol-ogy an increasing awareness that it may be moreimportant to prevent the cardiovascular consequencesof anemia than to attempt to ameliorate them onceestablished. There are at least three ongoing trials thatattempt to determine the e�ect of anemia prevention, ortreatment at early phases, versus correction of prolongedanemia in patients prior to dialysis who do not have anyevidence of advanced cardiovascular disease.

Adverse Effects of Higher Hemoglobin

The main concerns regarding higher hemoglobinconcentrations include vascular access thrombosis,hypertension, and seizures. It is di�cult to separate thee�ects of higher hemoglobin from those of highererythropoietin doses. However, the fact that patientswith naturally high hemoglobin concentrations, such asthose with polycythemia rubra vera, posttransplanterythrocytosis, or polycythemia secondary to polycystickidneydisease, donot appear to have symptomsuntil thehemoglobin exceeds 170 g/L supports the hypothesisthat some of the adverse events may be related toerythropoietin itself, or to the administration of the drugin nonphysiologic manners. Furthermore, the cost ofachieving a higher target hemoglobin is signi®cant due toincreased requirements of erythropoietin and iron.Studies describing an increase in vascular access

thrombosis (VAT) have either involved small patientnumbers or have evaluated VAT as a secondaryendpoint. Nevertheless, the results of these studies havebeen con¯icting. Besarab et al. (6) found a higherincidence of VAT of both native artery ®stula andsynthetic grafts in the normal Hct group than in the lowHct group that reached statistical signi®cance. Churchillet al. (32) described an increased incidence of graftthrombosis but not ®stula thrombosis. Other investiga-tors have found no increase in either ®stula or graftthrombosis (33,34). Although it is generally believed thatgraft thrombosis increaseswith erythropoietin therapy, alarger prospective analysis with normalization of hemat-ocrit with erythropoietin therapy and VAT as a primaryendpoint is required to fully address this issue.Studies have generally suggested that 30% of patients

treated with erythropoietin become hypertensive orrequire an increase in the dose or number of antihyper-tensivemedications tomaintain adequate blood pressurecontrol. Erythropoietin has also been shown to activateendothelin in a subset of patients (35), and to increaseangiotensin II expression and stimulate vascular smoothmuscle cell proliferation (36). Furthermore, throughmultiple mechanisms, including the increase in hyper-tension, theremay be an increase in diastolic dysfunctionwith erythropoietin therapy. Furthermore, one reportdescribes that, independent of hemoglobin, a decrease in

bleeding time, serum plasminogen, and antithrombin IIIlevels occurred during therapy with erythropoietin (37).It is possible that some people are more susceptible tothese vascular e�ects of erythropoietin, and thus thatthese patientsmayhave anassociated cardiovascular riskwith the use of erythropoietin, particularly in high doses.Furthermore, as higher doses are required to raise, ratherthan to maintain, hemoglobin concentration, it may bethat e�ects are lessened at the lower doses required forprimary prevention.Seizures have rarely been associated with erythropoie-

tin use.However, it has been observed that seizures occurmost commonly when large erythropoietin doses havecaused a rapid increase in hemoglobin. Therefore the useof lowerdoses toprevent adecrease inhemoglobinwouldobviate the concern regarding seizures.

Conclusion

The controversy of the appropriate hemoglobin targetis related to two key points: the previously held view ofthe CKD population as a homogeneous group ofpatients, and the cost of normalization of hemoglobinwith erythropoietin in the context of a perceived paucityof data on long-term bene®ts and, in addition, someconcerns about adverse e�ects.With respect to the ®rst point, it is now clear that

patients need to be classi®ed according to the level ofkidney function, dialysis duration, and the presence ofcardiovascular disease. There are data to support thenormalization of hemoglobin in patients with asympto-matic cardiac disease, or no cardiac disease, if oneassumes that endpoints to justify such therapy includeimproved quality of life, exercise capacity, and cognitivefunction. Randomized controlled data suggesting that ahigher hemoglobin confers a survival advantage arescarce, though several large observational cohort studiesdo suggest this.Given that there are di�erent cardiovascular diseases

in patients with kidney disease, the need to categorizepatients with respect to cardiovascular disease prior toestablishing targets is also clear. Those on dialysis withsevere cardiac symptoms are not likely to bene®t fromhigher hemoglobin levels. However, patients not ondialysis, with clearances well below 50 ml/min (30) andsevere congestive heart failure (CHF), may bene®t fromhemoglobin levels in the 120±130 g/L range.As treatments for ESRD patients change (such as the

implementation of daily or nocturnal dialysis programs),the de®nitions of hemoglobin targets may need to beadjusted according toglomerular ®ltration rate (actual orachieved). Of interest, reports to date suggest highhemoglobin levels in these patients, with reduction orcessation of supplemental erythropoietin therapy.Would we suggest phlebotomizing these individuals?The costs of therapy with current available agents in

the absence of clear survival data bene®ts have alsofueled the controversy. As clinicians and patient advo-cates, are we prepared to sacri®ce the improvement inquality of life, cognitive function, and exercise capacity inpatients who do not have contraindications to higher

NORMALIZED HEMOGLOBIN IN CKD 11

hemoglobin levels on the basis that no long-term survivalbene®t has been demonstrated yet? As well, the adversee�ects of erythropoietin and iron therapy need to beseparated from the adverse e�ects of higher hemoglobin,so that we do not get confused between treatment ofanemia and the treatment options for anemia. Thisdichotomy may help in separating the cost issues forsome.What can we say with the current available evidence?

In patients prior to dialysis, hemoglobin levels shouldbe monitored and a decline in hemoglobin should beprevented or at least minimized. Currently there areno ®rm data supporting a recommendation thathemoglobin levels be normalized. However, maintain-ing hemoglobin levels between 125 and 140 g/L isreasonable and supported by data. The impact ofanemia in the context of other variables that accom-pany the uremic state and progressive decline of kidneyfunction is not known and requires further study.Outcome measures for this group should includequality of life, exercise tolerance, cognitive function,and cardiovascular health.In dialysis patients with chronic anemia, target

hemoglobin concentrations should be in keeping withcurrent published guidelines. Such targets should bemodi®ed in speci®c patient groups such as those withsymptomatic ischemic heart disease or those withpremorbid conditions in which polycythemia is a feature(thereby potentially necessitating higher target hemoglo-bin levels), including polycystic kidney disease or chronicobstructive pulmonary disease.Of importance, we need to encourage completion of

ongoing clinical trials, to establish new trials, and toclearly articulate the generalizability of clinical studies tothe populations thatwe treat. It is our opinion thatmuchof the current controversy about higher hemoglobin hasbeen generated through improper articulation of theissues, the lack of clear identi®cation of the keysubpopulations in dialysis and nondialysis patients, andthe confusion of therapeutic strategies and costs versusphysiologic rationale. Furthermore, as clinicians, scien-tists, and patient advocates, it is important that weidentify meaningful endpoints on which we should basetherapeutic decisions.Randomized trials should be designed to address these

issues. One such trial in patients with a glomerular®ltration rate (GFR) of less than 50 ml/min is random-izing patients to maintenance of normal hemoglobinversus ``resuscitation'' of hemoglobin to levels of 135 g/Lonce it has dropped to less than 110 g/L. Future trialsshould also address the question of vascular accessthrombosis, and the issue of targeting hemoglobinaccording tomarkers of in¯ammation such asC-reactiveprotein.The therapy of anemia in patients with CKD is

becoming increasingly sophisticated. The results of theongoing studies are eagerly awaited. We emphasizeanemia management as an essential aspect in the care ofpatients withCKD, but believe that the e�ects of therapywill be more impressive in the context of care of all of thehemodynamic and metabolic abnormalities that accom-pany CKD.

References

1. Collins AJ, Ma JZ, Xia A, Ebben J: Trends in anemia treatment with ery-thropoietinusageandpatientoutcomes.AmJKidneyDis 32(6 suppl4):S133±S141, 1998

2. Foley RN, Parfrey PS, Harnett JD, Kent GM,Murray DC, Barre PE: Theimpact of anemia on cardiomyopathy, morbidity andmortality in end-stagerenal disease. Am J Kidney Dis 28:53±61, 1996

3. Harnett JD, Kent GM, Foley RN, Parfrey PS: Cardiac function andhematocrit level. Am J Kidney Dis 25(4 suppl 1):S3±S7, 1995

4. Madore F, Lowrie E, Brugnara C, LewN, LazarusM, BridgesK, OwenW:Anemia in hemodialysis patients: variables a�ecting this outcome predictor.J Am Soc Nephrol 8:1921±1929, 1997

5. Canadian Erythropoietin Study Group: Association between recombinanthuman erythropoietin and quality of life and exercise capacity of patientsreceiving hemodialysis. BMJ 300:573±578, 1990

6. Besarab A, Bolton WK, Browne JK, Egrie JC, Nissenson AR, OkamotoDM, Schwab SJ, Goodkin DA: The e�ects of normal as compared with lowhematocrit values in patients with cardiac disease who are receivinghemodialysis and epoetin. N Engl J Med 339:584±590, 1998

7. Evans RW, Radar B,ManninenDL, CooperativeMulticenter EPOClinicalTrial Group: The quality of life of hemodialysis recipients treated withrecombinant human erythropoietin. JAMA 263:825±830, 1990

8. FoleyRN, Parfrey PS,Morgan J, Barre PE, Campbell P, Cartier P, CoyleD,FineA,HandaP,KingmaI,LauCY,LevinA,MendelssohnD,MuirheadN,Murphy B, Plante RK, Posen G, Wells GA: E�ect of hemoglobin levels inhemodialysis patients with asymptomatic cardiomyopathy. Kidney Int58:1325±1335, 2000

9. HayashiT, SuzukiA, Shoji T, TogawaM,OkadaN,TsubakiharaY, ImaiE,Hori M: Cardiovascular e�ects of normalizing the hematocrit level duringerythropoietin therapy in predialysis patients with chronic renal failure.AmJKidney Dis 35:250±256, 2000

10. Marsh JT, Brown WS, Wolcott D, Carr CR, Harper R, Schweitzer SV,NissensonAR:RHuEPOtreatment improvesbrainandcognitive functionofanemic dialysis patients. Kidney Int 39:155±163, 1991

11. McMahon LP, Mason K, Skinner SL, Burge CM, Grigg LE, Becker GJ:E�ects of hemoglobin normalization on quality of life and cardiovascularparameters in end-stage renal failure.Nephrol Dial Transplant 15:1425±1430,2000

12. Nisensson AR, Pickett JL, Theberge DC, BrownWS, Schweitzer SU: Brainfunction is better in hemodialysis patients when hematocrit is normalizedwith erythropoietin [abstract]. J Am Soc Nephrol 7:1459, 1996

13. Collins A, Ebben J, Ma J: Frequent IV iron dosing is associated with higherinfectious deaths [abstract]. J Am Soc Nephrol 8:190A, 1997

14. HoenB,KesslerM,HestinD,MayeuxD:Risk factors forbacterial infectionsin chronic haemodialysis adult patients: a multicentre prospective survey.Nephrol Dial Transplant 10:377±381, 1995

15. TielemansCL, LencludCM,WensR, Collart FE,DratwaM:Critical role ofiron overload in the increased susceptibility of haemodialysis patients tobacterial infections: bene®cial e�ects of desferrioxamine. Nephrol DialTransplant 4:883±887, 1989

16. Benz RL, Pressman MR, Hovick ET, Peterson DD: A preliminary study ofthe e�ects of correction of anemia with recombinant human erythropoietintherapy on sleep, sleep disorders, and daytime sleepiness in hemodialysispatients (the SLEEPO study). Am J Kidney Dis 34:1089±1095, 1999

17. Nissenson AR: Optimal hematocrit in patients on dialysis therapy. Am JKidney Dis 32(suppl 4):S142±S146, 1998

18. Stivelman JC: Bene®ts of anaemia treatment on cognitive function.NephrolDial Transplant 15(suppl 3):29±35, 2000

19. Horina JH, Schwaberger G, Brussee H, Sauseng-Fellegger G, Holzer H,Krejs GJ: Increased red cell 2,3-diphosphoglycerate levels in haemodialysispatients treated with erythropoietin. Nephrol Dial Transplant 8:1219±1222,1993

20. O'Sullivan DA, McCarthy JT, Kumar R, Williams AW: Improved bio-chemical variables, nutrient intake, and hormonal factors in slow nocturnalhemodialysis: a pilot study.Mayo Clin Proc 73:1123±1124, 1998

21. Levin A, Singer J, Thompson CR, Ross H, Lewis M: Prevalent left ventric-ular hypertrophy in the predialysis population: identifying opportunities forintervention. Am J Kidney Dis 27:347±354, 1996

22. Levin A, Thompson CR, Ethier J, Carlisle EJ, Tobe S, Mendelssohn D,Burgess E, Jindal K, Barrett B, Singer J, Djurdjev O: Left ventricular massindex increase in early renal disease: impact of decline in hemoglobin. Am JKidney Dis 34:125±134, 1999

23. Franch HA, Mitch WE: Catabolism in uremia: the impact of metabolicacidosis. J Am Soc Nephrol 9(suppl 12):S78±S81, 1998

24. Raine AEG, Seymour AML, Robert AFC, Radda GK, Ledingham JGG:Impairment of cardiac function and energetics in experimental renal failure.J Clin Invest 92:2934±2940, 1993

25. Matthais S, Honack C, Rosen P, Zebe H, Ritz E: Glucose uptake andexpression of glucose transporters in the heart of uremic rats [abstract]. J AmSoc Nephrol 6:1023, 1995

26. Metivier F, Marchais SJ, Guerein AP, Pannier B, London GM: Pathophy-siology of anemia: focus on the heart and blood vessels. Nephrol DialTransplant 15(suppl 3):14±18, 2000

Stevens et al.12

27. Bonilla S, Goercke IA, Bozzo S, Alvo M, Michea L, Marusic ET: E�ect ofchronic renal failure on Na,K-ATPase alpha 1 and alpha 2 mRNA tran-scription in rat skeletal muscle. J Clin Invest 29:39±43, 1991

28. Digicaylioglu M, Lipton SA: Erythropoietin-mediated neuroprotectioninvolves cross-talk between Jak2andNF-kappaB signaling cascades.Nature412:641±647, 2001

29. BrinesML,Ghezzi P, Keenan S, deAgnelloD, Lanerolle NC,Cerami C, ItriLM, Cerami A: Erythropoietin crosses the blood-brain barrier to protectagainst experimental brain injury. ProcNatl Acad Sci USA 97:10526±10531,2000

30. SilverbergDS,WexlerD,ShepsD,BlumM,KerenG,BaruchR,SchwartzD,Yachnin T, Steinbruch S, Shapira I, Laniado S, Iaina A: The e�ect ofcorrection of mild anemia in severe, resistant congestive heart failure usingsubcutaneous erythropoietin and intravenous iron: a randomized controlledstudy. J Am Coll Cardiol 37:1775±1780, 2001

31. Portoles J, Torralbo A, Martin P, Rodrigo J, Herrero JA, Barrientos A:Cardiovascular e�ects of recombinant human erythropoietin in predialysispatients. Am J Kidney Dis 29:541±548, 1997

32. Churchill DN, Muirhead N, Goldstein M, Posen G, Fay W, Beecroft ML,Gorman J, Taylor DW: Probability of thrombosis of vascular access amonghemodialysis patients treatedwith recombinant human erythropoietin. J AmSoc Nephrol 4:1809±1810, 1994

33. MartinoMA,VogelKM,O'Brien SP,KersteinMD:Erythropoietin therapyimproves graft patency with no increased incidence of thrombosis orthrombophlebitis. J Am Coll Surg 187:616±619, 1998

34. Besarab A, Medina F, Musial E, Picarello N, Michael H: Recombinanthuman erythropoietin does not increase clotting in vascular accesses.ASAIOTrans 36:M749±M753, 1990

35. Stefanidis I, Mertens PR, Wurth P, Bach R, Makropoulos W, Mann H,HeintzB: In¯uence of recombinant human erythropoietin therapy onplasmaendothelin-1 levels during hemodialysis. Int J Artif Organs 24:367±373, 2001

36. Vaziri ND: Cardiovascular e�ects of erythropoietin and anemia correction.Curr Opin Nephrol Hypertens 10:633±637, 2001

37. Tsao CJ, Kao RH, Cheng TY, Huang CC, Chang SL, Lee FN. The e�ect ofrecombinant human erythropoietin on hemostatic status in chronic uremicpatients. Int J Hematol 55:197±203, 1992

REYNARD THE FOX AND THE KING'S URINE

Uroscopy, the art of diagnosis based on inspection of the urine, was widely practiced fromthe twelfth through the seventeenth centuries. The image of the physician-uroscopist wascelebrated in art and literature. Logan Clendening, a clinican and historian of medicnefrom the University of Kansas, wrote about the fable of Reynard the Fox (1). A popularanimal legend originating in monastic story-telling traditions, the adventures of Reynardwere set down by Heinrich der GlichesaÈ re in the form of an epic poem in the late twelfthcentury. One story involves an ailing lion who consults Reynard, the physican. Noble, thelion, recites his symptoms: a great pain in the head, blurry vision, a foul taste, diffuse bodyaches, and dif®culdty breathing. The pain, says the lion, has left him shivering and unableto speak further. Reynard considers:

So, said Fox, you shall be cured,When the third day has passed, bring me a urinal,And we shall see therein the ill. . . .The urinal was brought, Noble is sitting upright,And urined more than half.And Fox said, Just so,Took it and went into the sunlight;Lifted high the urinal, looks at it externally,Turns it over and over again To see whether it will alter. . . .Sire, he says, By God's help I see you have ague.Yet I have the portion which kills it.Sir King, I shall have to administer it to you. . . .A little longer and I should have been too late. . . .The sickness of your body I will cure.The quartane fever will go off, which takes away your breath so much.

And so it happened.

1. Clandening L: The history of urinalysis. Medical Leaves 3:25±32, 1940.

NORMALIZED HEMOGLOBIN IN CKD 13