Dialysis Facility Ownership and Epoetin Dosing in Hemodialysis Patients: A US Physician Perspective

How Does This Study Compare With Prior Studies: Why Was the Target Recommended By the FDA Not Followed? 

One argument was that the FDA was wrong, and outcomes would be better if the target was higher. The 1997 National Kidney Foundation Dialysis Outcomes Quality Initiative (DOQI) clinical practice guidelines9 reflected this belief by recommending that the hemoglobin target be 11 to 12 g/dL (110-120 g/L), rather than 10 to 12 g/dL (100-120 g/L) as suggested by the FDA. However, at the time that these guidelines were published, there were no large RCT results supporting the belief that a target higher than the FDA’s target improved outcomes (Table 1). The first RCT to test this hypothesis, published in 1998, was stopped early when the data safety monitoring board, observing a trend toward increased mortality in the group targeting hemoglobin 14 g/dL (140 g/L) versus 10 g/dL (100 g/L), considered it futile to continue.5 Despite this mortality finding, it remains argued that quality of life is better at higher than recommended hemoglobin levels. However, the reporting of quality of life in these studies has been seriously flawed by a lack of masking, selective reporting of components that were different between groups (without statistical adjustment for multiple comparisons), and the use of different quality of life instruments. It is no surprise that quality of life results have varied between studies.

Another argument for not following FDA recommendations was that fluctuations in hemoglobin levels made it impossible to maintain hemoglobin in a narrow target range, such as the targets suggested by the FDA or even DOQI.10, 11 Implicit in this argument was the assumption that it was better to err too high rather than too low. However, there are no data from RCTs to support this hypothesis either.

The most plausible argument for not following FDA recommendations is the financial benefits inherent with targeting higher hemoglobin levels. The best evidence that the administration of ESAs to hemodialysis patients has been influenced by financial motives comes from several analyses of registry data comparing ESA use in for-profit and not-for-profit dialysis units. In 2006, Thamer et al reported that the most frequent use of subcutaneous ESAs, known to require lower doses and reduce cost, was in dialysis units not affiliated with for-profit chains.12 Similarly, Collins et al reported that the administration of ESAs to patients whose hemoglobin levels exceeded recommendations was more common in for-profit chain facilities than in not-for-profit facilities.13 In the 2006 US Renal Data System Annual Data Report,14 there were marked differences between dialysis providers in the proportion of patients with hemoglobin of at least 12 g/dL (120 g/L). The nonprofit provider, Dialysis Clinic, Inc, had only 20% of the 2004 prevalent population with hemoglobin greater than or equal to 12 g/dL (120 g/L), while a major for-profit provider, DaVita, had over 60% with hemoglobin of at least 12 g/dL (120 g/L). Most recently, an analysis of registry data led Thamer et al to likewise conclude that large for-profit chains targeted higher hemoglobin levels than not-for-profit dialysis facilities.15

The pattern from observational studies is clear. More ESA is used in for-profit facilities than in not-for-profit facilities. Of course, there may be nonfinancial explanations that could explain these results. However, careful statistical adjustment for differences in patient populations failed to provide a cogent, alternative explanation. This, combined with the fact that for-profit dialysis facilities (like other corporations) must show an increase in quarterly earnings to survive, provides a compelling argument that the quest to increase earnings has influenced the administration of ESAs to hemodialysis patients.

What Should Physicians Do? 

Recent data from RCTs in nondialysis patients have shown that targeting hemoglobin levels higher than those recommended by the FDA can have adverse consequences.16, 17, 18 There is little reason to believe that this would not also be the case for hemodialysis patients. Responding to these and other recent RCT results, the FDA issued an alert in November 2006. The FDA recommended that physicians should “[w]ithhold the dose of ESA if the hemoglobin exceeds 12 g/dL (120 g/L) or rises by 1 g/dL (10 g/L) in any 2 week period.”19 In the absence of RCT data to the contrary, perhaps we should heed the advice of the FDA.

In an editorial accompanying the study by Thamer et al, Daniel Coyne suggested that physicians should maintain hemoglobin levels between 10.5 (105 g/L) and 11.5 g/dL (115 g/L), rarely initiate an ESA if hemoglobin is above 10 g/dL (100 g/L), change the ESA dose by approximately 25% monthly when the hemoglobin falls outside this range, and either reduce by 50% or stop the ESA when the hemoglobin level exceeds 12.5 g/dL (125 g/L).20 This is very similar to the recommended protocol in our dialysis unit, although our protocol strictly follows the FDA recommendation targeting hemoglobin 10 to 12 g/dL (100-120 g/L), and withholding the ESA when the hemoglobin exceeds 12 g/dL (120 g/L). Of course, treating iron deficiency and other barriers to the effective use of ESAs should be included in the management paradigm.

There is a need for additional RCTs in hemodialysis patients. Trials should address not only outcomes, but how to achieve those outcomes in the most cost effective manner possible. For example, how long should ESAs be withheld when the target is exceeded? When should the hemoglobin be retested? How should iron replacement best be integrated with ESA dosing? There is also a need for the CMS to revisit its reimbursement policy for ESAs. In 2003, Congress asked the CMS to submit a report on how a bundled payment system for end-stage renal disease could be designed and to carry out a demonstration project. This has not yet occurred. Finally, ethics in industry need not be an oxymoron. Dialysis providers and the pharmaceutical industry should both consider how they could improve not only next quarter’s earnings, but also long-term patient outcomes. The incidence rate for end-stage renal disease has been slowing in the United States, and the best way for the industry to increase earnings in the future may be to increase the longevity of hemodialysis patients. This is a goal that we can all profit from.