Resistance to Erythropoiesis Stimulating Agent (ESA) Treatment

Definition of Hyporesponsiveness/Resistance to Erythropoiesis Stimulating Agents

The identification of patients with low responsiveness to ESAs is important for clinical and economic reasons. However, no standardized definition has been produced over the last three decades. In this respect, it should be emphasized that the concept of the optimal Hb to target with ESAs has progressively evolved over the years, together with the knowledge of possible safety risks with ESA use at high doses. For this reason, the understanding and definition of ESA hyporesponsiveness have evolved as well. What was accepted 15 years ago as the maximal ESA dose is not necessarily accepted today. However, the general concept has remained unchanged: a patient is hyporesponsive to ESAs when his dose requirements to either achieve or maintain a desired Hb level are higher than expected according to CKD stage, clinical characteristics, and geographical area.

Looking back, in the 1990s or early 2000s, available international guidelines on anemia treatment proposed their own definition.

According to the Revised European Best Practice Guidelines (EBPG), which were published in 2004, ESA hyporesponsiveness was defined as a failure to attain the Hb target concentration while receiving more than 300 IU/kg/week (> 20,000 IU/week) of epoetin or 1.5 μg/kg of darbepoetin alfa (> 100 μg/week) or has a continued need for such high dosages to maintain the target. This is > 2.5 times the average ESA dose, so most (> 90%) iron-replete patients given ESAs would be expected to respond to a lower dose. At that time, methoxy polyethylene glycol-epoetin beta was not available in the market yet and thus was not included in the definition.

In 2012, Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines defined ESA hyporesponsiveness as:

• 1)

  A significant increase in the ESA dose requirement to maintain a certain Hb level or a significant decrease in Hb level at a constant ESA dose.

• 2)

  A failure to increase the Hb level to greater than 11 g/dL despite an ESA dose equivalent to epoetin greater than 500 IU/kg/wk.

ESA hyporesponsiveness can be observed even at treatment start. Alternatively, higher dose needs can be seen after months or years of optimal response. This evolving clinical scenario was taken into consideration by the Kidney Disease: Improving Global Outcomes (KDIGO) guidelines in 2012. They classified ESA hyporesponsiveness as “initial” (if there is no increase in Hb concentration from baseline after the first month of ESA treatment on appropriate weight-based dosing) or “subsequent” (if there is a need for two increases in ESA doses up to 50% beyond the dose at which the patient had been stable in order to maintain a stable Hb level). Since a < 2% increase in the Hb concentration is likely to be within the variability range of Hb values in individual patients, this value was set as “no increase.” Differing from previous guidelines, the concept of a maximal dose to define hyporesponsiveness or resistance has disappeared. This has a number of implications from the practical point of view. Indeed, the definition of hyporesponsiveness was developed following the evidence coming from the Trial to Reduce cardiovascular Events with Aranesp Therapy (TREAT) study. This was a prospective, double-blind, randomized controlled trial of darbepoetin alfa versus placebo for the treatment of anemia in 4038 participants with type 2 diabetes mellitus, CKD not on dialysis, and anemia (Hb < 11.0 g/dL). The initial dose of darbepoetin alfa in the active treatment arm was 0.75 μg/kg, with doses adjusted monthly according to a computer-based algorithm to target Hb at 13 g/dL. The administration frequency was every 2 weeks. If lower initial dosages than those used in TREAT are chosen, the diagnosis of hyporesponsiveness must take this into account, and repeated escalations in ESA dose should be allowed to reach double the weight-based dose used in TREAT (i.e., two 50% dose escalations) starting from the TREAT initial dose.

Another point to be considered is that the TREAT study was performed in a population of patients with diabetes (and thus possibly with higher ESA dose needs) with CKD, not on dialysis. Interestingly, a post hoc analysis of TREAT showed that Hb declined precipitously in the year prior to the development of end-stage kidney disease (ESKD) (on average 1.15 g/dL lower in those who developed ESKD versus those who did not) with simultaneous increase in the mean darbepoetin dose (15% higher than those not reaching ESKD). Considering that hemodialysis (HD) patients have, on average, higher dose needs than patients not on dialysis, a higher setpoint has to be possibly considered.

In general, in order to fulfill the definition of ESA hyporesponsiveness, the patient has to be iron replete. Even if this could appear simple at first sight, looking more in depth, it is not necessarily easy to rule out this point. Indeed, many patients who are hyporesponsive to ESAs are inflamed and experience functional iron deficiency (i.e., high ferritin levels but low transferrin saturation [TSAT]).

The response to ESAs is made by the relationship between the desired Hb and ESA needs; this relationship is complex, nonlinear, and dynamic. In order to mathematically represent this concept, the ESA resistance index (ERI) has been proposed. This is the ratio between average weekly EPO dose (IU) (or other ESA molecules) per kilogram of body weight divided by Hb (g/dL). Considering the nonlinear and dynamic kinetics of Hb response to ESAs and that Hb has its own variability, the use of average values over at least a 3-month period is suggested. Indeed, in a moving 3-month window, ERI varies over time as Hb increases in response to the constant ESA dose. ERI is generally used for research purposes, and it is not commonly used in everyday clinical practice. However, some authors proposed its use as a useful tool to guide dosing; when elevated, ERI could trigger an evaluation for remediable factors causing hyporesponsiveness, even when Hb goals have been reached. However, ERI has the limitation of not taking into account iron stores or availability.

Epidemiology of Hyporesponsiveness to Erythropoiesis Stimulating Agents

The true incidence of ESA hyporesponsiveness is still a matter of study, and probably, it differs among countries (in the United States, it is likely to be higher than in Europe and even higher again than in Japan). Also, the lack of a universally accepted definition influences the high variability of estimates.

It is well known that, on average, ESA dose requirements differ among countries, reflecting a different burden of comorbidities (higher in the United States), possible differences in age, gender, race, ethnicity, and body mass index, dialysis vintage, treatment practices, and reimbursement policies. These differences are clearly shown by the Dialysis Outcomes and Practice Patterns Study (DOPPS). This is a prospective cohort study of HD practices based on the collection of observational longitudinal data for a random sample of patients from dialysis facilities in a representative and random sample of units in 20 countries. Patients from Japan, France, and Italy had been on dialysis significantly longer. The body mass index was significantly lower in Japan and significantly higher in the United States. More patients in Canada and the United States had diabetes. ESA doses and the percentage of patients with a weekly ESA dose > 35,000 IU/week were highest in the United States (18.1%). In European countries, this percentage ranged from only 1.8% in Germany to 12.7% in Sweden. In Japan, no patients received these high treatment ESA doses. These figures are partially influenced by reimbursement policies that were in place at the time of data collection in the different countries. Indeed, ESA reimbursement was often limited to a maximum dose applying not necessarily to all the ESA molecules.

The prevalence of ESA hyporesponsiveness in HD patients has been reported between 7.3% and 17.6%. However, a small cross-sectional study of 550 Italian HD patients reported a much lower percentage (only 2.4% when using the EBPG definition of ESA hyporesponsiveness) due to iron deficiency in many cases. Conversely, the prevalence of hyporesponsiveness to ESAs was much greater in a cohort from a large dialysis organization (DaVita) in the United States (43% of the patients received more than 18,000 IU/week of epoetin). However, these estimates did not discriminate between chronic and acute (and therefore reversible) ESA hyporesponsiveness. Indeed, taking this difference into consideration and using three definitions of hyporesponsive dialysis patients (total ESA dose, ESA dose per kg of body weight, or ESA dose per Hb level), Gilbertson et al. showed a prevalence of 4.5%, and 15% of patients with chronic and acute ESA hyporesponsiveness, respectively. According to this estimate, 50% of total ESA costs were spent on the 15% of patients requiring the highest dosage.

Clinical Use of Erythropoiesis Stimulating Agents in Hyporesponsive Patients

All ESAs are effective in correcting renal anemia and increasing Hb levels. However, ESAs differ in amino acid sequence, carbohydrate content, charge, and molecular weight. These characteristics influence their half-life and biological activity and, thus, their clinical use in hyporesponsive patients. Indeed, the higher the molecular weight and the carbohydrate content, the longer the half-life and the lower the affinity to the EPO receptor. Epoetin alfa and beta behave similarly with a half-life of nearly 8 hours when administered intravenously and of 24 hours for the subcutaneous route. Darbepoetin alfa has a half-life of 25 and 48 hours when given intravenously and subcutaneously, respectively. Methoxy polyethylene glycol-epoetin beta has a half-life of nearly 130 hours, which is independent of the administration route.

According to pharmacokinetics/pharmacodynamics characteristics, less frequent dosing of ESAs is more effective for the molecules with longer half-life (i.e., darbepoetin alfa or methoxy polyethylene glycol-epoetin beta) or in patients not on dialysis. However, the labels of all the different ESA molecules allow delayed administration frequencies. However, patients with high dose requirements are less likely to benefit from less frequent administration of short-acting ESAs. Indeed, several studies have shown that this strategy is effective without the need for extradosing, mainly in patients with low dose requirements. In hyporesponsive patients, administering a short-acting ESA at a low frequency implies very high serum peaks, potentially exposing them to ESA-related adverse events. It also implies extra costs. For this reason, it is reasonable to suggest in hyporesponsive patients that administration frequencies should respect as much as possible the original recommendations based on the pharmacokinetics of every single agent.

It has been suggested that at high doses, long-acting ESAs may have a more favorable conversion factor from epoetin alfa and beta, possibly reflecting different kinetics in the activation of the EPO receptor. However, this has not received definitive confirmation from the scientific community.

Recently, a large cohort study from the Japanese Registry of Dialysis compared the mortality risk associated with the use of short-acting versus long-acting ESAs. According to crude data analysis, the long-acting ESAs were associated with a 20% higher risk of all-cause death than short-acting ESAs. Of interest for this chapter is the fact that the difference in risk was higher among patients receiving high ESA doses and those who had high ERI. Long-acting ESA use was also associated with an increased rate of death from cardiovascular diseases, infection, and malignancies. Darbepoetin alfa users had the highest rate of all-cause death, death from cardiac diseases, and malignancies. Despite several strengths, this study also has several limitations. In particular, despite the large sample size and similar baseline characteristics between long- and short-acting ESA users, potential prescription biases at patient and facility level and residual confounders cannot be ruled out. A large amount of missing data on ESA type, major covariates, and TSAT may have influenced patient selection and adjusted analyses. Interestingly, the study showed that the outcomes of ESA users might be different for achieved Hb levels. Indeed, for achieved Hb levels < 11 g/dL, long-acting ESA users had a slightly higher risk of all-cause death compared to those treated with short-acting agents; no difference was found for higher achieved Hb values. Of note, in Japan, there is a maximum reimbursable dose for short-acting ESAs of 9000 units per week (determined from prescribing information on the ESA label). This may account for a possible prescription bias toward a higher use of long-acting ESAs in hyporesponsive patients.

In the same period, a large, randomized study aimed at comparing postapproval safety of methoxy polyethylene glycol-epoetin beta in comparison to reference ESAs did not confirm a different safety profile among ESA molecules with respect to the risk of major adverse cardiovascular events or all-cause mortality.

Another concept to be kept in mind is that dose requirements of long-acting ESAs are similar either subcutaneously or intravenously. On the contrary, short-acting ESAs have, on average, a 30% extra dose requirement when administered intravenously. This difference is important, considering that now, virtually all HD patients receive ESAs intravenously.

ESA dose requirements are hard to predict in individual patients. However, patients with comorbidities, diabetes, cardiovascular disease, and inflammation are more likely to have high dose needs. The role of body weight and obesity is still controversial, with some studies showing lower per-kilo dose requirements in obese patients. According to an analysis of the national database of the DaVita dialysis organization on 38,328 prevalent HD patients, low iron stores, hyperparathyroidism, and high-turnover bone disease, together with lower estimated dietary protein intake and signs of malnutrition, were associated with a greater risk of ESA hyporesponsiveness.

At present, there is an agreement to avoid the use of excessively high ESA doses. How this recommendation should be translated into clinical practice is less clear. In 2010, the European Renal Best Practice (ERBP) position statement gave generic suggestions to balance increased cardiovascular risk with the possible benefits of anemia correction obtained with high ESA doses. In particular, dose escalation should be avoided in those patients who do not respond to treatment or in whom it is obvious that worsening of anemia is linked to nonrenal factors. Risks and benefits should also be carefully evaluated in patients with cancer in whom cure is an anticipated outcome.

In 2015, National Institute for Health and Care Excellence (NICE) guidelines confirmed the recommendations that were given in 2006, focusing on a balance between the pros and cons of a trial of anemia management to be discussed among the clinician, the person with anemia of CKD, and their families and caregivers if applicable. Consideration to not administer ESAs may be appropriate in the presence of comorbidities or when the prognosis is likely to negate the benefits of correcting the anemia. As an alternative, a trial with ESAs could be tried in the presence of uncertainty over the benefit of correcting with ESAs. When appropriate, ESA withdrawal should be considered.

KDIGO guidelines on anemia therapy, which were issued in 2012, also reported a number of recommendations about avoiding excessive ESA dose escalation in those who are either hyporesponsive at treatment start or those who become later on. Unfortunately, these recommendations are all issued starting from a low degree of evidence and thus are either not graded or are graded 2D at maximum. In particular, it is suggested to avoid repeated escalations in ESA dose beyond double the initial weight-based dose (initial hyporesponsiveness, 2D) or beyond double the dose at which the patient had been stable (acquired hyporesponsiveness, 2D). If a patient remains hyporesponsive despite correcting treatable causes, KDIGO suggests individualizing further anemia therapy after having balanced its pros and cons. Among possible options, physicians can decide to keep the patient at a lower Hb level, to continue a high ESA dose, and/or give blood transfusions when needed.

Hyporesponsiveness and Outcomes

Hyporesponsiveness to ESAs has been clearly associated with adverse outcomes in HD or nondialysis CKD patients. The gathering of this knowledge comes from either observational studies or secondary analyses of randomized clinical trials. According to a retrospective cohort study of 128,598 HD patients, there is a dose-dependent positive association between weekly epoetin-α doses ≥ 18,000 U/week and mortality risk. Similar findings were shown by a cohort study with data obtained from Clinical Research Center registry for ESKD in Korea. In this population of 1594 prevalent HD patients, those with high-dose ESAs and low Hb levels had a significantly higher risk of all-cause mortality in comparison with those with a better response to ESA therapy.

Similarly, using data from the U.S. Renal Data System, Zhang et al. found that epoetin requirements predict mortality in a large cohort of prevalent HD patients ( n = 94,569). In particular, for every hematocrit cohort studied, patients administered higher doses of epoetin had significantly lower hematocrit values and greater mortality rates; a significant nonlinear relationship between increased epoetin dose and mortality was also found regardless of hematocrit, with the steepest increase in relative risk for death in the 25th higher percentile.

The RISCAVID (“RISchio CArdiovascolare nei pazienti afferenti all'Area Vasta In Dialisi”) was an observational study, which was performed in Tuscany, Italy, between 2004 and 2007. Among a population of 651 prevalent HD patients receiving ESAs, quartiles of ERI correlated with all-cause mortality and fatal/nonfatal cardiovascular (CV) events, with those in the highest quartile having the highest risk.

Looking at randomized clinical trials, already back in the nineties, Besarab et al. showed that the patients who did not achieve a given Hb target have a higher risk of death. In this randomized clinical trial, 1233 HD patients with clinical evidence of congestive heart failure or ischemic heart disease were assigned to either a lower (30%) or higher (42%) hematocrit during therapy with epoetin alfa. In both the randomization groups, the mortality rate in each group at various hematocrit values, calculated as the average of all values until death, loss to follow-up, or study end, increased at lower hematocrit values. Using a Cox multivariate analysis, a 30% decrease in the risk of death or myocardial infarction was shown per 10-point increase in average hematocrit. Unexpectedly, a higher epoetin dose was not associated with increased mortality.

More recently, the secondary analysis of a recent, randomized postmarketing study aimed at comparing postapproval safety of methoxy polyethylene glycol-epoetin beta in comparison to reference ESAs showed that the patients with a 3-month average Hb of < 10 g/dL (i.e., those who did not achieve the Hb target of the study despite ESA therapy) had a threefold higher risk of experiencing a primary endpoint compared with the reference category (10–11 g/dL).

It is still matter of debate whether high ESA doses may cause damage or hyporesponsiveness to ESAs is just a marker of increased comorbidities.

It is well known that, in addition to erythropoiesis stimulation, ESAs have a number of pleiotropic effects in several tissues where the EPO receptor is expressed. Some of them may be beneficial, but others could be harmful, especially at high ESA doses. This would suggest a possible negative effect of excessive ESA doses per se.

However, as shown by a secondary analysis of the TREAT study, the predictive role of ESA response on patient outcome seems to be present already at treatment start. Indeed, the subgroup of patients who demonstrated an Hb change of < 2% after 1 month following a fixed starting dose of darbepoetin alfa (i.e., those having an initial poor response) had a higher risk of reaching the cardiovascular composite endpoint than the better Hb responders (who had an event rate similar to the placebo group). Even if this does not exclude a negative effect of exposure to higher ESA doses, it may suggest that preexisting inflammation or comorbidities are of more importance. Accordingly, these patients were more likely to have previous cardiovascular disease or receive treatment for heart failure, have higher C-reactive protein (CRP) levels, and be overweight. Interestingly, ESA hyporesponsiveness has been linked to insulin resistance, which is more likely to occur in overweight patients and is a well-known cardiovascular risk factor.

Some years ago, Gillespie et al. performed a case-crossover study on a cohort of 6645 European CKD HD patients and evaluated the reversibility of ESA hyporesponsiveness and the factors associated with the transition to this state. ESA responsiveness was defined on the basis of the median dose threshold. Transition to hyporesponsiveness was associated with hospitalization, vascular access changes, or worsening inflammation, with these factors accounting for over two-thirds of transitions. Clear causes or risk factors of hyporesponsiveness could be identified in 68% of the patients. In the remaining ones, hyporesponsiveness remained unexplained. The fact that the identified factors were independent risk factors for poorer outcomes is a further confirmation that ESA hyporesponsiveness is a risk marker rather than a risk factor.

Causes of Hyporesponsiveness and Their Management

The pathogenesis of the anemia related to ESA resistance can be attributable mainly to three mechanisms: iron-restricted erythropoiesis , inflammation, and bone marrow suppression. The relationship of one with the other is complex, but very often, they are closely interwoven. Bearing this concept in mind, many of the risk factors that have been described as causes of ESA hyporesponsiveness can be ascribed to these three mechanisms ( Box 38.1 ). In particular, the development of iron deficiency or the occurrence of an inflammatory state is very common. In this regard, intercurrent infections are the most frequent reason why patients become temporarily (and reversibly) ESA-resistant.

Occult blood loss, severe hyperparathyroidism, malnutrition, and inadequate dialysis are also important. Angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor antagonists may play a role in some instances. Compliance should also be checked in patients self-administering an ESA.

Some causes of ESA resistance are little evident; their identification requires an in-depth diagnostic workout. As suggested by KDIGO guidelines, patients with either initial or acquired ESA hyporesponsiveness should be evaluated and then treated for specific causes of poor ESA response (Not Graded).

Unfortunately, as shown by Gillespie et al. in an observational study, one-third of the patients have no identifiable cause.