Chapter 2: Use of Iron to Treat Anaemia in CKD
For adult CKD patients with anaemia not on iron or ESA therapy, we suggest a trial of IV iron (or in CKD ND patients alternatively a 1–3 month trial of oral iron therapy) if (2C):
- an increase in Hb concentration without starting ESA treatment is desired
- and TSATis ≤30% and ferritin is ≤500 ng/mL (≤500 μg/L)
It is well known that iron therapy is an important step in the treatment of anaemia in CKD patients, as both absolute and functional iron deficiencies are common. In the TREAT study, the patients in the control group maintained a relatively high mean Hb level during the follow-up despite the fact that they received minimal darbepoetin alfa doses [median 0, interquartile range (IQR) 0–5 μg/month] as a rescue therapy. This was partially influenced by the fact that many of these patients were not fully iron-replete [median transferrin saturation (TSAT) 23%, IQR 18–29%] and thus received a course of oral or IV iron. These results have certainly contributed to re-evaluate the role of iron therapy not only in patients who are iron deficient but also in those with apparently adequate iron stores (as defined by serum biomarkers).
Thus, we agree with this recommendation about the possibility first to perform a trial of IV iron (or oral iron therapy in the ND-CKD population when tolerated) in anaemic CKD patients if an increase in Hb levels is desired. This would also be helpful in reducing the need for blood transfusions. In ND-CKD patients with mild to moderate anaemia, oral iron should be used as first-line therapy for a minimum of 3 months in the absence of known gastrointestinal intolerance to preserve the veins of the arm for possible future dialysis access (AV fistula). Conversely, IV iron is the first choice in this population in the presence of severe anaemia or when oral iron is ineffective.
We also agree with the statement that 'For any individual patient the optimal balance of Hb level, ESA dose, and iron dose at which clinical benefit is maximized and potential risk is minimized is not known'. Indeed, peripheral-iron blood indices of iron storage transport and handling have limited utility in identifying depletion of bone marrow iron stores.
However, in our opinion, the proposed limits of serum ferritin and TSAT, which the new KDIGO guidelines suggest should help drive the decision on whether or not administer iron therapy in patients not receiving ESA, are too wide and are not adoptable for a number of reasons.
First, no clear distinction is made between absolute and functional iron deficiency when giving the strength of the recommendation whether to start iron therapy. For patients with absolute iron deficiency (serum ferritin <100 ng/mL and TSAT <20%), the indication for iron therapy should be stronger since the likelihood of obtaining an increase in Hb level following iron therapy is much higher. Conversely, it is true that even in patients with adequate bone marrow iron stores, sometimes, it is possible to obtain an increase in Hb levels following iron therapy. However, this quantitative effect is lower in patients who are not iron deficient. Stancu et al. showed that, following the administration of 1000 mg of IV iron to 100 patients with ND-CKD, an erythropoietic response was obtained in 63% of those who had iron-deplete bone marrow but only in 30% in those who were iron-replete. The chances of a positive response increased by 7% for each 1% decrease in TSAT. In this European, mostly Caucasian population, the median serum ferritin and TSAT values were much lower (176 ng/mL and 23%, respectively) than the upper threshold up to which iron therapy could be prescribed according to KDIGO recommendations (i.e. serum ferritin ≤500 ng/mL and TSAT ≤30%). Of note, ferritin values in the Stancu's paper were much <500 ng/mL despite the fact that two-thirds of the patients were chronically inflamed (C-reactive protein level >10 mg/L).
Spinowitz et al. studied another cohort of 304 ND-CKD patients who were given two 510-mg doses of ferumoxytol IV or 200 mg of elemental oral iron daily for 21 days in a 3:1 ratio. Among patients who were not receiving ESAs (n = 188), Hb increased by 0.62 ± 1.02 g/dL with ferumoxytol and by 0.13 ± 0.93 g/dL with oral iron. In this population, mean baseline serum ferritin and TSAT were ~145 ng/mL and 10%, respectively.
We agree with the KDIGO group that the available evidence on this topic is inadequate and scanty. However, we believe that before deciding whether or not to give a course of iron therapy, the physician should know that this evidence has been obtained only in the ND-CKD population and that in the available studies, mean/median ferritin levels were well below the now proposed upper limit of 500 ng/mL. These considerations clearly suggest the need to differentiate in this recommendation between non-dialysis and dialysis patients, and not lump these together in an overarching recommendation. Given the paucity of evidence, we have no information about safety when prescribing long-term iron therapy at higher ferritin levels than those previously recommended in the ND-CKD population.
The safety of administering IV iron therapy to ESA-naïve haemodialysis patients with high serum ferritin levels has not been established. For instance, increased circulating ferritin levels have been found associated with an impaired immune response of monocytes, possibly increasing infection risk. We feel that it is likely that the magnitude of the obtained increase in Hb level after IV iron therapy would be rather small in haemodialysis patients; so, the start of ESA treatment would be probably more effective (also considering that evidence-based concerns related to ESA use in the dialysis population seem less concerning than in the ND-CKD population). Therefore, a course of IV iron therapy despite ferritin values >300 ng/mL should be considered in those haemodialysis patients in whom ESA therapy may be contraindicated or considered risky (see KDIGO section about ESA initiation). Conversely, in ESA-naïve anaemic patients who have adequate iron stores, the concomitant start of ESA and iron therapy may be appropriate to prevent iron deficiency due to increased erythropoiesis stimulation.
Moreover, it is clear that an operative interval is lacking. After reading this recommendation, the physician remains confused because the upper limit for prescribing iron therapy coincides with the proposed limit not to be exceeded in treatment. This implies the risk of reaching very high serum ferritin levels when IV iron therapy is started at the upper edge of the interval (especially with single high doses given for long periods).
Finally, whether or not to treat CKD patients with Hb values >12 g/dL and absolute iron deficiency remains an open issue. At present, it seems wise to suggest that we do treat these patients (in the absence of clear risks of targeting towards higher Hb values) but as we do so being careful to avoid intentionally exceeding an Hb value of 13 g/dL. If these patients are receiving ESA treatment, this additional iron therapy should be at least temporarily halted (see following sections about ESA initiation and ESA maintenance therapy).
For adult CKD patients on ESA therapy who are not receiving iron supplementation, we suggest a trial of IV iron (or in CKD ND patients alternatively a 1–3 month trial of oral iron therapy) if (2C) an increase in Hb concentration or a decrease in ESA dose is desired and TSAT is ≤30% and ferritin is ≤500 ng/mL (≤500μg/L).
In CKD patients receiving ESA therapy, iron stores may be nearly normal, but they may be insufficient for the increased erythropoiesis which typically follows bone marrow ESA stimulation. In this context, iron therapy reduces significantly ESA doses requirements. This is of particular importance given concern related to ESA use especially at high doses (even if this association is limited by the bias that patients receiving higher ESA doses are usually those with more comorbidities). The fact that iron deficiency (absolute or relative) is a major cause of ESA hyporesponsiveness in CKD patients suggests that there is still room to augment iron therapy in many CKD patients. Unfortunately, TSAT and ferritin have limitations and low power to diagnose functional iron deficiency and predict response to IV iron. Other markers such as hypochromic red cells or reticulocyte Hb may add some information when available. According to recent findings of 120 dialysis facilities of the Dialysis Outcomes and Practice Patterns Study (DOPPS) Practice Monitor in the USA, following the change in the ESA label by the Food and Drug Administration (FDA) in June 2010, from August 2010 to August 2011, the percentage of dialysis patients receiving IV iron went from 57 to 71%. This went together with a significant decline in ESA dosing and a slight decrease in median Hb levels.
However, a larger use of iron therapy caused a substantial increase in ferritin levels. Indeed, in these patients the median ferritin level increased from 556 to 650 ng/mL with 34% of the patients exceeding the value of 800 ng/mL. Conversely, the percentage of patients with TSAT ≥50% remained around 10%. Interestingly, every 100 mg of IV iron raised TSAT by 0.43% in those subjects having TSAT values <30% but only by 0.10% in those having higher TSAT levels; Hb values remained unchanged. This may suggest that targeting to TSAT levels ≥30% with IV iron therapy does not improve erythropoiesis and exposes patients to the risk of iron overload.
In Europe, dialysis patients have lower median ferritin levels than those in the USA. According to the data of the UK Renal Registry, in 2009 in haemodialysis patients, the median ferritin value was of 417 ng/mL (IQR 270–598). This value was similar in 2010 [444 ng/mL (IQR 299–635)]. However, no major changes in guideline recommendations about anaemia management took place in Europe in this period (thus, it is too early to observe ferritin changes in the European population).
The safety of persistently very high ferritin levels is still unknown. In 453 men with non-dialysis CKD, a trend towards higher mortality was observed in patients with a serum ferritin level >250 ng/mL. However, the study was not adequately powered to properly analyse survival data. In dialysis patients, high serum ferritin has been associated with increased mortality as well. In a cohort of 58 058 prevalent haemodialysis patients in the USA, both all-cause and cardiovascular mortality had increasing rates across increasing ferritin levels, whereas the opposite (inverse) association was observed for TSAT increments. Serum ferritin levels between 200 and 1200 ng/mL and iron saturation ratio between 30 and 50% were associated with the lowest all-cause and cardiovascular death risks. However, association studies are biased by the fact that serum ferritin is also a marker of inflammation. Indeed, in unadjusted, time-varying model, serum ferritin >800 ng/mL during each quarter was associated with increased death rate.
The Dialysis Patients' Response to IV Iron with Elevated Ferritin (DRIVE) trial, found that IV iron was effective in increasing Hb levels and reducing ESA doses in patients with high ferritin (500–1200 ng/mL) and low transferrin saturation levels (TSAT ≤25%). However, the sample size was quite small and the overall follow-up (6 weeks + 6 weeks in the DRIVE II extension period) was adequate for testing acute iron toxicity but too short to provide information about safety and iron overload in the long term.
High-dose baseline iron therapy has been found to be associated with poor outcome in haemodialysis patients. However, this likely reflects an indication bias, because no statistically significant association was detected between mortality and any level of iron dosing.
According to an autopsy study of 36 haemodialysis patients published 30 years ago when erythropoietin was still not available, serum ferritin did not always correlate with bone-marrow iron stores but correlated well with the degree of hepato-splenic siderosis, probably because hepato-splenic stores failed to be mobilized to the bone marrow. This should be taken into account when administering IV iron, which by passes the intestinal mechanism for the regulation of iron absorption, especially in inflamed patients in whom inhibitory factors, such as hepcidin, decrease iron release from reticulo-endothelial and hepatocyte stores. The regulatory role of hepcidin may thus change the relationship between ferritin levels, iron stores and Hb levels. This is why a number of patients with high serum ferritin may have functional iron deficiency and have an increase in Hb levels following iron therapy. Compared with those days, this mechanism is likely to be amplified at any level of serum ferritin level. Indeed, the CKD population has substantially changed compared with the first haemodialysis patients, who were much younger and with fewer comorbidities compared with nowadays. After the introduction of ESA in clinical practice, following a much lower use of blood transfusion, clinically significant iron overload has become a rare event. However, it has been hypothesized that iron administration may exacerbate oxidative stress and increase the risk of infection, cardiovascular events and death well before causing signs of iron overload. At present, evidence coming from clinical trials testing IV iron molecules has not shown a significant increase in deaths, cardiovascular events or infections following IV iron use. However, these studies were not adequately sized to test mortality or hard end points. Long-term safety studies examining these practices are urgently required and long overdue.
Some years ago, a study was published which featured direct, non-invasive measurements of non-haeme hepatic iron content by magnetic resonance in 40 dialysis patients treated with IV iron. This study showed that two-third of the patients had signs of mild to severe iron overload despite the fact that only one-third of the patients had serum ferritin exceeding 500 ng/mL. According to the receiver operating characteristic analysis, the best specificity/sensitivity ratio to identify iron overload was obtained for ferritin >340 ng/mL. Recently, significant iron overload in the liver and spleen (assessed through T2 magnetic resonance) has been described in 19 of 21 haemodialysis patients with serum ferritin >1000 ng/mL and severe comorbidities who were treated with IV iron. Similarly, Rostoker et al. prospectively studied a cohort of 119 fit haemodialysis patients who were receiving iron and ESA therapy and measured their liver iron content by means of T1 and T2 magnetic resonance. Mild to severe hepatic iron overload was observed in 84% of the patients, 36% of whom had severe iron overload approaching that found in haemocromatosis. Liver iron content is significantly related to the cumulative iron dose received and can rapidly decrease following iron therapy discontinuation.
Despite the fact that excess iron in the liver is potentially harmful, the clinical consequences of high iron content estimated by magnetic resonance is not known.
As is the case for ESA-naïve patients, the indication to iron therapy should be stronger for absolute iron deficiency that increases ESA dose requirements inappropriately and perhaps also the risk of cardiovascular events.
It is true that we have no clear evidence indicating an upper limit for ferritin level that can be considered either safe or dangerous. However, the fact that the current KDIGO guideline on this topic indicates the possibility of administering a trial of iron therapy even in patients who already have high serum ferritin levels (~500 ng/mL) will certainly cause a significant rise in ferritin levels in the CKD population, especially in the haemodialysis setting, shifting to the right the frequency distribution curve. This will occur regardless of the presence or absence of signs of chronic inflammation. In our opinion, also considering that increasing TSAT above 30% does not substantially modify Hb levels, in the absence of clear evidence, prudence should prevail to limit over treatment, at least in European countries with no particular restriction to ESA treatment.
Following these considerations, we agree with the KDIGO recommendation 2.1.3 that in adult CKD patients treated with ESA a trial of iron therapy may be useful if an increase in Hb concentration or a decrease in ESA dose is desired. However, caution is needed in patients with already high ferritin levels, because the safety of treating these patients is still unknown. Conversely, we agree with the KDIGO group that this is an important topic that needs to be investigated by future research.