Anemia in Chronic Kidney Disease

Abstract

Epoetinalfa, the first erythropoiesis-stimulating agent (ESA), was approved in 1989 for the treatment of anemia in patients with chronic kidney disease (CKD). By 1991, in dialysis, there were no longer patients requiring regular transfusions for severe anemia (eg, hemoglobin concentration, <7g/dL [toconvert togramsper liter,multiplyby 10]), anddialysis center–based transfusions haddecreased bymore than 65%.1 For many reasons, including the establishment of financial incentives that rewarded overuse, aggressive treatment of anemia in dialysis-requiring patients with increasingly higher doses of ESA became the norm, and mean hemoglobin concentrations in patients receivingdialysis increased from10.5 g/dL in 1995 to 12.0 g/dL in 2005.1,2 Anemia during CKD stages 4 and 5ND (not receiving dialysis) is less frequent and less severe than in patients receiving dialysis. In 1 observational study, mean hemoglobin concentrationwas 12.0 g/dL in stage 4 and 10.9 g/dL in stage 5ND. Consequently, patients with CKD do not require transfusions except during acutemedical illness, atwhichpoint their prior degree of anemia is relatively unimportant. In this issueof JAMA InternalMedicine,Winkelmayer and colleagues3 examineddata onMedicare recipients 67 years or older who started dialysis to capture information on anemia management practices in the 2 years prior to initiation of dialysis. In 2010, nondialysis ESA administration in physician’s offices accounted for $701 million in Medicare Part B spending. Erythropoiesis-stimulating agents increase hemoglobin levels andmay reduce the need for transfusions, but trials in patients with CKD who either were or were not receiving dialysis show that ESAs increase the risk of death, thrombovascularandcardiovascularevents, cancer recurrence,andcancerrelated death and have not been proven to improve quality of life or fatigue.4-7 Howdid these drugs becomeblockbusters? I place the observational data of Winkelmayer et al in the context of contemporaneous trial results and guidelines addressing anemia management in CKD. Paralleling growth observed in patients receiving dialysis, Winkelmayer et al found marked growth in ESA use between 1995and2007, from3.2%to40.8% inpatientswithCKD who were not receiving dialysis.3 In adjusted analyses, patients in 2007 were 11.5 times more likely to receive an ESA, and treatment started 7months earlier relative to dialysis initiation than in 1995.3 Growth in ESA use was slower between 1995and2000butaccelerated in2001,correspondingwithpublication of the first Kidney Disease Outcomes Quality Initiative (KDOQI)guidelinesaddressinganemia in thenon–dialysisrequiring CKD population. Amgen, the maker of ESAs, is the “foundingandprincipal sponsorof theKDOQIguidelines.”8(pA6) Both the 1997 KDOQI guidelines, which focused on dialysis, and the 2001 guidelines recommended targeting hemoglobin to 11 to 12 g/dL, higher than the then-approvedUSFood and Drug Administration (FDA) target of 10 to 12 g/dL. The guidelines outlined 6 strands of largely associational evidence linking hemoglobin concentrations below 10 g/dLwith worse outcomes. The message from guidelines, ESA manufacturers,dialysis chains, andanemiaexperts to the renal communitywas that anemiawasharmful and thatmaintaininghemoglobin concentrations above 11 g/dL was beneficial.2 The KDOQI anemia guidelineswere fundedby themakers ofESAs, and many anemia experts, including me, were paid by ESA makers to expound on the value of treating anemia in CKD. By 2001, the only completed large outcome trial of ESAs was the Normal Hematocrit Trial. Terminated in 1996 by the data safetymonitoring committeebecause “...the resultswere nearing the statistical boundary of a higher mortality rate in the normal hematocrit group”4(p589) and published in 1998, it had randomized dialysis-requiring patients to an ESA-driven hemoglobin target of 9 to 11 g/dL or a normal range of 13 to 15 g/dL.4 In the published report, the group with the higher hemoglobin target had a statistically insignificant higher risk of the primary end point (death or nonfatal myocardial infarction [MI], 202 events [183 deaths, 19 MIs] vs 164 events [150 deaths, 14 MIs]; risk ratio [RR], 1.3 [95% CI, 0.9-1.9]) but had reduced transfusions and improved physical functioning as measured on the SF-36 quality-of-life instrument.4 The framers of the 2001 guidelines believed that these dangers conflicted with data from observational and small intervention trials and noted that the higher risk of death or nonfatal MI “was not statistically significant at the time the study was terminated.”9(pS191)However, thetrial reportwas less thanforthcoming as it had adjusted theCIs for repeated reviews anddid not report theunadjustedCIs.4,5 In2007, theFDAreported that the high-target arm had significantly higher unadjusted risk for death (RR, 1.27 [95%CI, 1.04-1.54]) and death plusMI (RR, 1.28 [95% CI, 1.06-1.56]).4,5 In 2012, it was also revealed that, contrary to the 1998 report, targetingnormalhemoglobin concentrations had led to no improvement in any component of quality of life.5 Knowledge of the serious risk of ESAswith no benefits in 1998maywellhavealtered the trajectoryof thedata of Winkelmayer et al. Although Amgen and the authors believed that the trialwas flawedanddidnot showdefinitiveevidence of harm,withholding the unadjusted statistics and the real quality-of-life results prevented physicians and guideline groups from reaching their own conclusions. Winkelmayer et al showthatmeanhemoglobinconcentrationpeaked in2006,useofESAsreachedaplateau in2007, then bothgraduallydeclined through2010.3 In2006, theKDOQIhad publishednewanemiaguidelines, stating that all patientswith Related article page 699 Research Original Investigation Trends in Predialysis Anemia Care