Guidelines 3.4 and 3.5 – Treatment of Anaemia with/without ESA therapy – Target Haemoglobin Range

We recommend that people with non-dialysis dependent CKD (NDD-CKD) or those receiving dialysis, who are on ESA therapy, should achieve Hb between:

• 100 and 120 g/L in adults, young people and children aged 2 years and older. (1B)
• 95 and 115 g/L in children younger than 2 years of age (reflecting the lower normal range in that age). (2B)

Guideline 3.5 – Treatment of Anaemia without ESA Therapy – Target Haemoglobin Range:

We suggest that this Hb target range applies exclusively to people with CKD receiving ESAs and is not intended to apply to the treatment of iron deficiency in people receiving iron therapy without the use of ESAs. (2B)

Rationale

In determining target Hb guidelines it is important to assess potential benefits (in terms of possible improved survival, improvement in health-related quality of life (HRQoL) and avoidance of transfusion requirement and hospitalisation) vs. potential harms (increased mortality, increased risk of vascular events).

Although several studies have shown that higher Hb targets could be associated with improvements in both physical and mental health domains (1), the HRQoL benefits of higher Hb targets diminish over time (1). In addition, there is no apparent Hb threshold above which there is definitively a quality-of-life improvement in the higher Hb treatment arms. Besarab et al (2) reported a study of normalisation of haemoglobin in 1233 prevalent people with receiving HD with high cardiovascular risk on haemodialysis. Normalisation of haemoglobin showed no benefit in risk reduction but did show an improvement in quality of life. The treatment arm showed a trend towards increased risk of death or cardiac event (183 deaths and 19 myocardial infarcts, producing 202 primary events, compared to 164 events (150 deaths, 14 myocardial infarcts) and vascular access (39% versus 29%) and the trial was terminated before completion on the grounds that the study was unlikely to show benefit from normalisation.

Two trials evaluated the effect of ESA on people not yet receiving dialysis – CHOIR (3) and CREATE (4). outcome of the CHOIR trial showed no benefit of higher Hb outcome in people with NDD-CKD (eGFR 15-50ml/min/1.73m2) randomised to Hb of 113 g/L vs. 135 g/L. Higher outcome target Hb had an increased risk (using composite endpoints of death, myocardial infarction, or hospitalisation for congestive cardiac failure) and no incremental improvement in quality of life (3). The limitation of this trial is that, compared with the group assigned to the lower Hb treatment target, the higher Hb target group showed at baseline a statistically greater proportion of patients with a history of hypertension and coronary artery bypass graft. A report posted by the trial sponsor (5) indicates that patients assigned to the higher Hb treatment arm also had a significantly greater severity of congestive heart failure (CHF) at baseline. The results of a multivariate analysis, included in this report, indicate that after adjustment for baseline conditions (CHF by National Health and Nutrition Examination Survey CHF score, atrial fibrillation/flutter, serum albumin level, reticulocyte count, and age), the relationship between treatment assignment and primary composite outcome events is no longer statistically significant (HR, 1.24; 95% confidence interval [CI], 0.95 to 1.62; p = 0.11 compared with the unadjusted HR of 1.34; 95% CI: 1.03 to 1.74; p = 0.03) (6). A secondary analysis of the CHOIR trial suggested that higher doses of epoetin α, rather than target Hb per se, were associated with an increased risk of death, myocardial infarction, congestive heart failure or stroke compared with lower epoetin doses, and with poorer outcomes (6). Another secondary analysis of the CHOIR trial found that, among patients with diabetes mellitus, the percentage of people reaching the primary end point of death, myocardial infarction, congestive heart failure or stroke within 3 years was similar in the high and low haemoglobin arms of the trial (24.8% versus 24.7%, respectively; p = 0.249). By contrast, among patients without diabetes mellitus at baseline, 36.4% of participants randomised to the higher Hb target had reached the primary end point after 3 years compared with 24% of those randomised to the lower haemoglobin target (HR 1.70; 95% CI: 1.03–2.81; p= 0.04). Individuals without diabetes mellitus randomised to the higher haemoglobin target had a significantly greater risk of reaching the primary end point after 3 years than individuals with diabetes mellitus randomised to the lower haemoglobin target (7).

The CREATE (4) trial reported that early correction of anaemia to normal Hb (130-150 g/L vs. 105-115 g/L) did not reduce risk of cardiovascular events. Indeed, the hazards ratio for primary endpoints of death from any cause or death from cardiovascular disease consistently (but not significantly) favoured the lower haemoglobin target group. The trend to increase in events appeared to occur after initiation of dialysis but there was no difference in endpoints after censoring of data from patients who started dialysis. Quality of life was significantly better in the higher Hb outcome group. Although GFR was not significantly different between the two groups, more people started kidney replacement therapy earlier in the higher Hb outcome group (p =0.03) with the difference apparent from 18 months. An important limitation of this trial is that the event rate was much lower than predicted; thus, the power to detect a difference in event rates was decreased (4). Other important limitation (s) of these trials is that important subgroups of participants enrolled in large trials, such as young adults, people returning to dialysis after failed kidney transplant, or people with chronic lung disease were not identified or assessed in any of these trials. Further analysis of outcome of high target Hb was performed by the KDOQI team (8). An Evidence Review Team analysed all data from randomised controlled trials of anaemia management in NDD-CKD, including, CHOIR, CREATE and other studies. Combining mortality outcomes from eight studies involving 3038 subjects with CKD (the CHOIR and CREATE trials contributed most of the weight to the analysis) revealed no difference between the higher and lower Hb target (7), but combining adverse cardiovascular events from six studies involving 2850 subjects showed an increased risk among the patients assigned to the higher Hb targets (a RR of 1.24, 95% CI: 1.02–1.51) (8), although it is worth noting that the CHOIR and CREATE trials also contributed most of the weight to the analysis. Among dialysis patients, combining mortality (four studies, 2391 subjects) or cardiovascular outcomes (three studies, 1975 subjects) showed no statistically significant difference between the higher and lower Hb level with The US Normal Haematocrit Trial (2) contributing most of the weight to the analysis.

In the TREAT trial (9), 4038 people with diabetes, chronic kidney disease not on dialysis, and anaemia, were randomly assigned in a 1:1 ratio to darbepoetin α, to achieve Hb level of approximately 130 g/L or to placebo, with rescue darbepoetin α when the haemoglobin level was less than 90 g/L. The primary end points were the composite outcomes of death or a cardiovascular event (nonfatal myocardial infarction, congestive heart failure, stroke, or hospitalisation for myocardial ischaemia) and of death or end-stage kidney disease. After a median follow up of 29 months, there was no difference between the two arms in the primary outcome of death, cardiovascular event or end stage kidney disease. Fatal or nonfatal stroke occurred in 101 participants assigned to darbepoetin α and 53 participants assigned to placebo (HR, 1.92; 95% CI: 1.38 to 2.68; p <0.001). The investigators concluded that for many involved in clinical decision making this risk of prescribing an ESA in this population will outweigh the potential benefits (9). Data from observational studies have, however, not shown increased hazard risk among people who achieved higher Hb. In one study, data from people receiving haemodialysis in the UK Renal Registry from 1999 to 2005 were analysed for the relative risk of death at different Hb concentrations. Hb concentrations above the reference range (100–110 g/L) consistently showed a 35% lower relative risk of death, while people with haemoglobin below 100 g/L had a 28% higher mortality. The greatest mortality was seen in people with haemoglobin <90 g/L (73% increased risk of death, although due to the small numbers, this was not statistically significant). On the other hand, the lowest death rate was seen in people with haemoglobin levels between 120 and 139 g/L (64% reduced mortality) (10).

The effect of cumulative ESA dose was also reported in another retrospective study (11). In this study, which looked at data from Medicare‘s end-stage kidney disease program between 1999 and 2007, different US dialysis centres annual anaemia management practice were characterised by estimating their typical use of ESAs and intravenous iron in people receiving haemodialysis within 4 hematocrit categories. Monthly mortality rates were assessed using Cox proportional hazards regression to correlate centre-level patterns of ESA and iron use with 1-year mortality risk in 269,717 people on incident haemodialysis. Monthly mortality rates were highest in people with haematocrit less than 30% (mortality, 2.1%) and lowest for those with haematocrit of 36% or higher (mortality, 0.7%). After adjustment for baseline case-mix differences, dialysis centres that used larger ESA doses in people with haematocrit less than 30% had lower mortality rates than centres that used smaller doses (highest vs. lowest dose group: HR, 0.94; 95% CI: 0.90-0.97). Centres that administered iron more frequently to people with haematocrit less than 33% also had lower mortality rates (highest vs. lowest quintile, HR, 0.95; 95% CI: 0.91-0.98). However, centres that used larger ESA doses in people with haematocrit between 33% and 35.9% had higher mortality rates (highest vs. lowest quintile, HR, 1.07; 95% CI: 1.03-1.12). More intensive use of both ESAs and iron was associated with increased mortality risk in people with haematocrit of 36% or higher (11).

Post Hoc analysis of the PIVOTAL Trial of 2141 people receiving dialysis, examining the baseline data, and potential relationship to the primary outcome (all-cause mortality, myocardial infarction, stroke, and heart failure hospitalisation), and associations with key baseline characteristics and QoL was measured using EQ5D index, Visual Analogue Scale (VAS), and the KD-QoL (Physical Component Score (PCS) and Mental Component Score (MCS)). The mean baseline EQ5D index and VAS scores were 0.68 and 60.7; 33.7 (PCS) and 46.0 (MCS). Female sex, higher BMI, diabetes mellitus, history of myocardial infarction, stroke or heart failure were associated with significantly worse EQ5D index and VAS. CRP and transferrin saturation (TSAT) were associated with QoL. Hemoglobin was not an independent predictor of QoL. TSAT was an independent predictor of PCS. CRP was associated with most aspects of QoL. Impaired functional status was associated with mortality. Quality of life was impaired in people starting hemodialysis. CRP was a consistent independent predictor of the majority of QoL scores. TSAT ≤20% was associated with worse physical component scores of QoL (12).

Several systematic reviews have looked at ESA or ESA and iron use, comparing various Hb target ranges and their effect on quality of life. One review by Spinowitz et al. looked at people receiving dialysis being treated with ESAs and without ESAs and its impact on their health-related quality of life (HRQoL). This showed that there may be some increase in HRQoL in those people treated with ESAs (with or without iron) vs those without treatment. However, there was no evidence for higher Hb targets leading to improved reporting of HRQoL (13).

Guedes et al. found that, similarly, higher Hb targets (>115 g/L) may be associated with a small decrease in fatigue levels when compared to those people with CKD aiming for the recommended target of 100-115 g/L. However, there was no significant effect seen on physical function or HRQoL(14).

The findings of all the above studies have made it difficult to define a safe target Hb in CKD patients treated with ESA. As a result, target Hb in this patient group has been the subject of extensive debate in the literature:

• KDIGO suggests that for adults receiving dialysis, ESA therapy could be used to avoid having the Hb concentration fall below 90 g/L by starting ESA therapy when the haemoglobin is between 90– 100 g/L (15).
• The Anaemia Working Group of ERBP expressed its view that Hb values of 110-120 g/L should be generally sought in the CKD population without intentionally exceeding 130 g/L In low-risk patients (i.e., in younger patients with very few comorbidities). In those with ischaemic heart disease with worsening ischaemic symptoms associated with anaemia, or in those in whom a clear benefit on quality of life can be foreseen, the start of ESA therapy could be considered at higher Hb values but not exceeding 120 g/L. In high-risk patients, including those with asymptomatic ischaemic heart disease, treatment initiation with ESA should be started at Hb values between 90 and 100 g/L to maintain a Hb value ∼100 g/L during maintenance therapy (16).

NICE guidelines on managing anaemia in people with CKD suggest maintaining the “aspirational” Hb range between 100 and 120 g/L for adults (17). The rationale behind choosing a wide target Hb range (100-120 g/L) for this guideline is that when the target Hb level is narrow (i.e. 10 g/L), variability in achieved Hb levels around the target is high, the fraction of prevalent patients with achieved Hb levels within the target range is low and ESA dose titration is required frequently during maintenance therapy.

It is also suggested to consider accepting Hb levels below the lower limit of the target range, if the target range cannot be achieved despite high, escalating doses of ESAs. High ESA doses are defined as: over 175 IU/kg per week for people receiving haemodialysis, over 125 IU/kg per week for peritoneal dialysis patients and over 100IU/kg per week for pre-dialysis patients (17).

• The health economics of anaemia therapy using ESAs has been subject to a NICE systematic review (17) which concludes that treating to a target Hb 100-120 g/L is cost effective in HD patients (17). Table 1 summarises the mean Hb data for prevalent UK dialysis patients from the 24th (2020) UK Renal Registry Reports (18).
• The Medicines and Healthcare products Regulatory Agency (MHRA) guidance (2007) notes that using ESAs to achieve Hb levels greater than 120 g/L is associated with an increased risk of death and serious cardiovascular events in people with CKD. The MHRA advises that Hb levels greater than this should be avoided, and that patients should be monitored closely to ensure that the lowest approved dose of ESA is used to provide adequate control of the symptoms of anaemia. Use of ESAs to achieve Hb levels greater than 120 g/L is not consistent with UK marketing authorisations for ESAs. Informed consent should be obtained and documented (19).