The Center for Medicare & Medicaid Services retired its Quality Incentive Program’s anemia performance measures for Hb levels < 10 g/dl for 2013 and beyond, triggering a great deal of discussion among patients, physicians, nurses, dialysis providers, and others about the potential risks to patients of this decision. This change was prompted by, among other things, findings of several clinical trials conducted in patients with CKD who were not on dialysis (and who were not maintained with Hb levels < 10 g/dl) and recent U.S Food and Drug Administration labeling changes for erythropoiesis-stimulating agents.

In discussing this decision, Dr. Patrick Conway, CMS chief medical officer and director of the agency’s Office of Clinical Standards & Quality, stated that “CMS believes that new concerns about the safety of ESAs for dialysis patients strongly argue for providers to work more closely with their patients to develop anemia management strategies that respond to the patient’s unique medical issues, rather than adopting a one-size fits all approach to care,” he said. “This patient-centered approach should result in better treatment outcomes. We plan to monitor hemoglobin levels by facility and to transparently share this information with consumers.”  Dr. Conway went on to say, “… anemia management and therapy should be determined by the patient’s physician in light of the patient’s individual needs and in consultation with the patient.”

Is elimination of an Hb “floor” really consistent with a patient-centered approach to care? It depends. It depends on who is providing the patient care and calling the shots when it comes to anemia management, and to what extent patients can and do participate in determining their optimal Hb level.

Who is regulating Hb range?

There are huge differences in two of Dr. Conway’s comments about elimination of an Hb lower limit of 10 g/dl; one refers to “providers” and one refers to “physicians.” These are not always the same. The management of anemia in dialysis patients has from the beginning of ESA availability been intertwined with regulatory and economic matters. Until recently there were financial incentives to use more ESA than might have been really necessary. This was likely exacerbated by how clinical practice guideline recommendations were interpreted and put into practice as well as the use of dialysis facility “quality” metrics that monitor the percent of patients with Hb levels above and below certain limits, incentivizing facilities, nurses and physicians to manage anemia to a metric rather than to the needs of the patient. With the advent of bundling, there is now a financial disincentive to ESA use and a new need to ensure that patients are adequately protected from any tendency to underutilize ESAs. This is where the distinction between “provider” and “physician” becomes critical. I will return to this.

If a floor, what is a fair measurement?

If one were to make a strong case for an Hb floor, it might be reasonably argued that 10 g/dl is not the right floor; but it is also hard to make a strong case for any other specific number. I would suggest though, with relatively few exceptions, that the vast majority of hemodialysis patients should probably have Hb levels maintained above 9.5 g/dl or so most of the time, particularly if this can be achieved with a reasonable ESA dose and avoidance of excessive iron administration. Of course, defining “reasonable” and “excessive” is not easy and any attempt to do so will surely meet with debate and disagreement.

Why consider an Hb floor of 9.5 g/dl? If we go back to the Canadian Erythropoietin Study Group randomized controlled trial of hemodialysis patients, we find some evidence in support of a 9.5 g/dl lower limit.1In this study, patients with a mean baseline Hb of 7.0 g/dl were randomized to placebo or epoetin alfa with a goal Hb of either 9.5-11.0 g/dl or > 11.0 g/dl. After only two months, the transfusion rate had fallen to under 3% in the epoetin-treated patients compared to 58% in the placebo group. At six months there were significant quality of life and physical function benefits in the epoetin-treated patients compared to the placebo group (but no difference between the high and low Hb target epoetin-treated patients).  Unfortunately, there is little else in the way of high quality evidence from randomized controlled trials in patients on dialysis; most such trials, at least recently, have been conducted in patients with CKD who are not on dialysis. These and other studies in dialysis patients have examined higher Hb levels rather than attempting to determine whether there is an optimal lower Hb limit. To what extent it is clinically appropriate to extrapolate the findings from studies in patients not on dialysis to patients on dialysis is not known.

Recently, treatment of anemia has appropriately begun to shift away from being focused on achieving a specific Hb level in all patients regardless of need to avoidance of transfusions and improvement in quality of life on a more individualized basis. Debate rages on regarding the extent to which ESAs and higher Hb levels actually enhance dialysis patient quality of life. If we are really able and willing to provide truly patient-centered, individualized anemia management for our hemodialysis patients we need to be able to make clear patient-specific assessments of the potential risks of iron, ESAs, transfusion and no treatment, as well as goals of therapy. The latter would ideally include objective measures of benefits of therapy other than determination of just Hb level and other laboratory parameters. This could include a validated and easily administered quality of life assessment tool and some sort of assessments of physical function; something simple might suffice—walking a block or up a flight of stairs, for instance.

While an Hb floor of 9.5 g/dl might be very reasonable for the “typical“ hemodialysis patient, this is probably too high a floor for some patients, i.e. a higher Hb than they really need, and too low a floor for others. The former group of patients might include those with severe debility who are bed-bound, have dementia and/or limited functional capacity, limited life-expectancy, and are likely to be quite ESA hyporesponsive. For these patients anemia management might be focused only on avoidance (or more correctly minimization) of transfusions. It might even be best to transfuse such patients when necessary for treatment of symptomatic anemia; decisions regarding the Hb level that triggers a transfusion should take into account recent studies indicating that a Hb level of 7 or 8 g/dl may be an appropriate transfusion threshold.2

Of greater concern are patients in whom benefits of an Hb level of 10-12 g/dl or even higher might well far exceed the risks. These are patients not really included in any of our currently available studies—young, healthy, active, awaiting transplant, no or few comorbidities, and likely to respond briskly to low ESA doses. This is a patient population that is virtually unstudied—and yet the group that probably suffers the most from severe anemia. For these patients we shouldn’t be thinking about a Hb floor so much as Hb optimization…we are unfortunately at a loss as to how to best identify such patients and monitor their response to treatment. But we all know that they are out there.

Who should be determining EPO dose?

Returning to the “provider” vs. “physician” distinction I made earlier––if by “provider” it is meant that anemia management is delegated to an anemia management nurse using a dialysis facility-developed anemia treatment algorithm––a Hb floor quality metric is probably necessary despite its limitations. If physicians, however, actively and consistently identify specific goals of anemia management for each patient, justified and documented in the medical record, and tailor anemia management to the individual patient without consideration of the financial ups or downs of such a patient-centered approach, a floor becomes less necessary.

For now, I think we need one. Recent evidence is that ESA doses have fallen, Hb levels < 10 g/dl have become more common, ferritin levels are skyrocketing, and transfusion rates are increasing. It has been suggested that the transfusion rate among hemodialysis patients increases steeply as Hb levels fall below the 9-10 g/dl range. The impact of all of this on meaningful clinical outcomes such as hospitalization and mortality remains to be seen but it seems that we should err on the side of being a bit overprotective of our patients rather than underprotective.


1. Association between recombinant human erythropoietin and quality of life and exercise capacity of patients receiving haemodialysis. Canadian Erythropoietin Study Group. BMJ 300: 573-578, 1990

2. Carson JL, Carless PA, Hebert PC.Outcomes using lower vs higher hemoglobin thresholds for red blood cell transfusion. JAMA 309: 83-84, 2013