Use of Epoetin in Chronic Renal Failure

In this issue of JAMA, Thamer and colleagues¹ examine the use of epoetin in US dialysis facilities. This report comes at an opportune time given the recent US Food and Drug Administration (FDA) advisory warning that epoetin and darbepoetin (erythropoiesis-stimulating agents [ESAs]) result in “an increased number of deaths and of non-fatal heart attacks, strokes, heart failure, and blood clots when ESAs were adjusted to maintain . . . hemoglobin more than 12 g/dL.”² The FDA recommends using just enough ESA to maintain the lowest hemoglobin level necessary to avoid the need for transfusions and ensure hemoglobin level does not exceed 12 g/dL.² The data from Thamer and colleagues¹ suggest higher epoetin use was the goal, and targeting higher hemoglobin levels was the means to that goal.

Thamer et al¹ report that for-profit facilities administer about 3300 U/wk more epoetin per patient than nonprofit facilities. The 2004 reimbursement rate of $10 per 1000 U of epoetin corresponds to $1700/y higher expenses per patient. Additionally, for-profit facilities increased the epoetin dose much more when hemoglobin levels were less than 11 g/dL (hemoglobin × 3 ≈ hematocrit), and even increased epoetin use significantly when hemoglobin levels were within the recommended target range of 11 to 12 g/dL.

Among the 4 largest for-profit dialysis chain facilities, all administered significantly more epoetin than the largest nonprofit chain. The heaviest administrator of epoetin, chain 2 facilities, administered 5845 U/wk more epoetin for patients with hematocrit levels of 33% or higher, corresponding to an additional $3000/y per patient, and an astounding 31 915 U/wk more epoetin for patients with hematocrit levels of less than 33%.

These discrepancies help to explain the explosion in the cost of epoetin therapy for dialysis patients. Between 1998 and 2004, annual Medicare payments for ESAs (epoetin and darbepoetin) in patients with end-stage renal disease increased 115% from $840 million to $1.8 billion while patient volume increased just 33%.³ Amgen is the sole manufacturer of both epoetin (brand name Epogen) and darbepoetin to US dialysis facilities. Johnson & Johnson is the distributor of epoetin (brand name Procrit) in all US nondialysis markets, such as oncology and chronic kidney disease.

As of 2004, 309 000 patients were receiving long-term hemodialysis in the United States,³ mostly in dialysis facilities now concentrated into a few large for-profit dialysis chains owned by large corporations. The distribution of hemoglobin levels among patients receiving dialysis in for-profit chains strongly suggests the goal apparently was to maximize epoetin use by increasing hemoglobin levels to higher than 12 g/dL.

Targeting a hemoglobin level of higher than 12 g/dL with epoetin exceeded the FDA recommended target range of 10 to 12 g/dL,⁴ and the then-applicable clinical practice recommendations, which suggested a target hemoglobin range of 11 to 12 g/dL.⁵ Randomized controlled trials (RCTs) of epoetin comparing target hemoglobin levels above and below 12 g/dL have shown that target levels of higher than 12 g/dL increased cardiovascular events and mortality with limited or no evidence of benefit.⁶ Also, the higher cost of anemia management at for-profit dialysis chains is borne by Medicare and commercial insurers, not the dialysis corporations.

Observational studies have noted that higher hemoglobin levels correlate with lower hospitalization rates, fewer cardiovascular events, and better survival in dialysis patients. A report by Regidor et al⁷ based on data from the dialysis chain operated by DaVita Inc showed a lower mortality rate among hemodialysis patients who maintained hemoglobin levels between 12 and 13 g/dL.

However, these observational analyses are seriously confounded by the interrelationship of anemia and epoetin resistance with illness in patients with chronic kidney disease. The sicker the patient (and the longer the patient is sick), the longer anemia is present and the greater the cumulative epoetin dose that must be administered to achieve the target hemoglobin level. The sickest patients rarely achieve target hemoglobin levels and also are the most likely patients to die. Healthier patients achieve target hemoglobin levels more easily and persistently with less epoetin^{8 - 9} and may exceed the target hemoglobin levels routinely, especially when that level is the goal of the dialysis chain. Therefore, the time-averaged hemoglobin level reflects both the targeted hemoglobin level and the health of the patient.

After adjustments for demographic and laboratory information, Regidor et al⁷ noted that a time-averaged hemoglobin level of 12 to 12.9 g/dL was associated with approximately half the mortality rate compared with a hemoglobin level of 10 to 10.4 g/dL, suggesting a higher hemoglobin target would increase survival. Advocates for higher hemoglobin levels claim that these types of findings support targeting hemoglobin level higher than 12 g/dL, and at all times maintaining the hemoglobin level higher than 11 g/dL.¹⁰

However, increasing the mean hemoglobin level in the US dialysis population from 10 g/dL in 1993 to approximately 12 g/dL in 2004 certainly did not reduce mortality by 50%. The adjusted mortality rate for prevalent dialysis patients was 231 deaths (per 1000 patient-years at risk) in 1993 and 230 deaths (per 1000 patient-years at risk) in 2004.³ The profound effect predicted in the analysis by Regidor et al⁷ was not evident despite the marked improvements in health care for cardiovascular disease and infectious diseases, which are the major causes of death in the dialysis population.

Moreover, the beneficial survival effect predicted by observational studies also was not observed in RCTs of higher vs lower target hemoglobin levels in patients with chronic kidney disease. Indeed, it is possible that mortality rates in dialysis patients may not have decreased because higher hemoglobin levels, higher doses of an ESA, or both, increase cardiovascular events and death.

To date, no major RCT has shown a significant reduction in cardiovascular events or death by targeting hemoglobin levels higher than 12 g/dL.¹¹ The 2 largest RCTs in hemodialysis patients found equal or superior outcomes when target and mean achieved hemoglobin levels were less than 11.5 g/dL.^{12 - 13} The Normal Hematocrit Trial randomized 1233 hemodialysis patients with coronary artery disease or congestive heart failure (most dialysis patients have one or both of these disorders) to a target a hemoglobin level of 14 g/dL or 10 g/dL. In the 14-g/dL group, 32.7% of patients experienced myocardial infarction or died vs 26.7% in the 10-g/dL group (P = .07). This study was stopped early following an interim analysis because “ . . . differences in mortality between the groups were recognized as sufficient to make it very unlikely that continuation of the study would reveal a benefit for the normal hematocrit group and the results were nearing the statistical boundary of a higher mortality rate in the normal hematocrit group.”¹²

In a 96-week, double-blind RCT in 596 healthier (ie, recent initiation on dialysis, absence of symptomatic heart disease, and lack of left ventricular dilatation) hemodialysis patients, in which twice the epoetin dose was used to target hemoglobin level of 13.5 to 14.5 g/dL vs 9.5 to 11.5 g/dL, higher hemoglobin levels afforded no beneficial effect on left ventricular dilatation or left ventricular mass index, and resulted in a stroke rate of 4% vs 1% in the higher vs lower hemoglobin groups, respectively (P = .045).¹³

Two recent RCTs of hemoglobin level targets in patients with chronic kidney disease not receiving dialysis reported better outcomes when the target and achieved hemoglobin levels were less than 11.6 g/dL. In the Correction of Hemoglobin and Outcomes in Renal Insufficiency (CHOIR) study (N = 1432), patients randomized to a target hemoglobin level of 13.5 g/dL (and an achieved hemoglobin level of 12.6 g/dL) had a 34% higher rate of death and cardiovascular complications than those with a target hemoglobin level of 11.3 g/dL (17.5% vs 13.5%; P = .03).¹⁴ Quality of life data showed a similar improvement from baseline values in both groups. The authors stated that “there is no cost benefit for either patients or payers in this population, even before considering risk.”¹⁴ In the Cardiovascular Risk Reduction With Early Anaemia Treatment by Epoetin Beta (CREATE) study (N = 607), patients with chronic kidney disease randomized to a hemoglobin target level of 13.0 to 15.0 g/dL had a cardiovascular event rate of 19.2% vs 15.5% in the group with hemoglobin level targeted to 10.5 to 11.5 g/dL (P = .20).¹⁵ The higher target group also had significantly more patients who required initiation of dialysis (42% vs 37%; P = .03).¹⁵

A meta-analysis of the major RCTs in patients with chronic kidney disease and receiving dialysis indicates that targeting a hemoglobin level higher than 12 g/dL results in a 17% increased risk of death (P = .03) compared with target hemoglobin levels less than 12 g/dL.⁶ A review of 4 studies involving oncology patients reported that increasing the hemoglobin level beyond 12 g/dL is associated with a higher risk of death, poor disease outcome, or both.¹⁶

An unanswered question in the study by Thamer et al¹ is why would nephrologists prescribe so much more epoetin, and make such dubious dosing decisions, when caring for patients at certain for-profit chains compared with patients treated at nonprofit facilities. Most nephrologists do not own dialysis facilities or profit from increased ESA use by facilities or chains. However, some nephrologists may not even know they are making these prescribing decisions. Nephrologists frequently sign multipage standing orders for treatment of long-term dialysis patients that include an anemia protocol that may subtly increase epoetin and hemoglobin levels. Nephrologists often turn over management to anemia managers¹⁷ (dialysis chain employees), who have a vested interest in maintaining hemoglobin level higher than 11 g/dL, not necessarily between 11 and 12 g/dL. In addition, many nephrologists may have been convinced that it is in the interest of patients to always maintain hemoglobin levels higher than 11 g/dL, and that exceeding a hemoglobin level of 12 g/dL is acceptable (even desirable) based on observational data.¹⁰

Another important question is why for-profit dialysis chain facilities permit such management and whether they actually encourage or even orchestrate such management. Epoetin is profitable for dialysis facilities. The 2 largest dialysis chains in the United States reported in their 2006 annual reports filed with the Securities and Exchange Commission^{18 - 19} that 21% to 25% of their revenue was from epoetin. Epoetin is reimbursed by Medicare at 6% above the average sales price.^{18 - 19} In addition, contractual agreements with Amgen provide rebates to facilities and chains for growth in epoetin purchases and in patient outcomes,¹⁸ which could potentially include achieving a high percentage of patients at hemoglobin levels higher than 11 g/dL. Chief medical officers have clear conflicts of interest because they generally own stock in the company and directly or indirectly influence anemia management policies. Dialysis chains also may provide to their employees bonuses or other payments tied to certain employee-specific “outcome” performance, and generally tied to the financial performance of the corporation.^{18 ,20} The significant dependence of dialysis providers on epoetin for income, and the ease at which a higher hemoglobin target affords greater epoetin use, creates a tempting situation for all involved.

A third key question is whether anemia treatment guidelines for patients with renal failure should promote more prudent management. The most widely used renal anemia guidelines in the United States are those of the National Kidney Foundation's Kidney Disease Outcomes Quality Initiative (NKF-KDOQI). In 2005, the NKF, a nonprofit organization, received more than $19.7 million from various corporations and organizations, including Amgen Inc and Ortho Biotech Products LP (sellers of epoetin in nondialysis areas), which are both “platinum” corporate partners.²¹

According to the 2006 KDOQI anemia guideline,¹¹ Amgen is the “founding and principal sponsor of KDOQI,” and is the sole corporate sponsor acknowledged in the guideline publication. A chairman of the anemia work group is a “part-time employee of DaVita”²² and 11 of 16 work group members reportedly have conflicts of interest with the ESA producers.²³ Despite much recent public scrutiny and comment,^{23 - 27} the KDOQI leadership seem unable to recognize and decisively address these issues. Indeed, with RCTs showing superior to comparable outcomes with lower hemoglobin levels, the NKF-KDOQI anemia guidelines in May 2006 stated that it was an “evidence-based” fact that “in patients with chronic kidney disease, (hemoglobin) should be 11.0 g/dL or greater” and it should not be maintained at levels higher than 13 g/dL, effectively raising the target hemoglobin level from 11 to 12 g/dL.¹¹ These recommendations rest solely on improvements in quality of life in patients with higher hemoglobin levels, although a higher target hemoglobin level did not improve quality of life in patients in the 2 largest RCTs.^{12 ,14} The evidence-based claim that all patients with chronic kidney disease should have hemoglobin maintained at a level of 11 g/dL or higher has already been used to justify major dialysis chains' anemia management practices.²⁸

In light of the publication of the CHOIR and CREATE trials,^{14 - 15} the NKF's Anemia Workgroup is reevaluating the anemia guidelines and is expected to release revisions for public comment in April 2007.²⁹ But physicians caring for patients with chronic kidney disease and receiving dialysis should not wait for the NKF opinions, nor necessarily trust or follow them.

The FDA has stated that the goal of ESA therapy is to reduce transfusions, and the risks and benefits in choosing to use an ESA must be carefully weighed for each patient. Because the FDA is responsible for drug safety, it is important that clinicians adhere to its recommendations. Based on the available evidence from clinical trials, maintaining hemoglobin levels between 10.5 and 11.5 g/dL will reduce transfusion requirements in patients with chronic kidney disease who are receiving dialysis. This means rarely initiating an ESA if hemoglobin level is above 10 g/dL, and changing the ESA dose by approximately 25% monthly whenever the hemoglobin level is above or below this range in a patient treated with an ESA. The ESA dose should be reduced by approximately 50% or stopped when the hemoglobin level exceeds 12.5 g/dL.

This management strategy is good for patients, and would be far less expensive for society. Physicians need to challenge industries that appear to be using patients as profit centers based on bad science.