Considerations in Iron Measures:
What Is an Appropriate Serum Ferritin Level?
A CME Symposium at the Renal Research Institute
7th International Conference on Dialysis: Advances in ESRD 2005
Introductory remarks: What is an appropriate serum ferritin level?
In this section we're going to take a look at what is an appropriate serum ferritin level? And I would hope, by the end of this talk, you'll find this talk provocative. It's intended to be. And it's also intended to question a lot of presumptions, assumptions, and information that we as nephrologists have been fed over the last several years.
Current serum ferritin upper limits
To begin with, let's take a look at the present ferritin limits as recommended by two expert panels. The K/DOQI criteria are obviously the US data, and they recommend a serum ferritin cut off of 800 ng/mL for CKD patients, which initially was the value set for ESRD or dialysis patients. This was opinion based, and there's really a marvelous story about the committee arguing over whether it should be 1,000 ng/mL or not, and the number 1,000 sounded kind of high, and there was no reason not to pick 1,000, but 1,000 sounded high, so they decided they would go with 800 because that sounded like a lower number. The European Best Practices Panel recommends targeting serum ferritin. That is, treating patients to get their serum ferritin to 500 ng/mL, and not giving iron when the ferritin is greater than 800 ng/mL.
Controversy of the ferritin upper limit
But we have to take a step back and ask, what's been happening with ferritin, and what are the implications for the future? In 1993, in dialysis patients, the mean serum ferritin was 154 ng/mL. And eight years later, the mean serum ferritin is 526. And serum ferritin, certainly, in the last few years, been maintaining itself in the mid 500s. If IV iron administration is the driving force behind this change in ferritin, and in large part it is, then we have to recognize that there has been no increase in mortality over those years. In fact, there's been a small decrease, not an appreciable amount, but a small decrease. There's been no increase in infections or unusual infections over those years. And there have been no reports of an iron overload syndrome or an epidemic of iron overload disease, such as liver failure or heart failure, in our patients. We do know that IV iron use reduces EPO administration. And it also decreases EPO income in the present system that we operate in. And also discussing the use of IV iron and the risks of IV iron deflect our discussion of the risks of EPO. And especially the higher doses of EPO, which, in nonrenal areas, is becoming a very hot topic.
Serum ferritin and inflammation
In an ideal world, serum ferritin would reflect the iron stores in our patients. We would measure the ferritin and we'd say, here are their iron stores. But our patients are not ideal by any means. And in addition to reflecting their iron stores, the serum ferritin level also reflects their inflammatory state, and in this study by Kalantar-Zadeh, serum ferritin values, with mean values of 800 to 2,000 ng/mL correlated with higher C-reactive protein levels, a marker for inflammation.
Serum ferritin and nutrition
In addition, higher ferritin stores reflect patient nutritional status; so maybe we can end the discussion with the thought that serum ferritin is an almost valueless test, as you'll see.
Is the serum ferritin cutoff warranted?
Why should we stop iron when the ferritin is high? Well, it would make sense to stop iron administration as the ferritin gets high if it's not safe. If there's a point at which we recognize that the safety of the drug and the safety of iron stores in the patients is no longer present. We certainly shouldn't give them iron if the ferritin's high, and we don't have reasonable evidence that it's beneficial, that it's not making them less cold, that it's not improving their hemoglobin. And thirdly, we shouldn't do it if it's not cost effective. And I suppose if the EPO price falls by 90%, one day iron might not be cost effective. But right now, it pretty much is.
Iron and infection: USRDS data
Let's take a look at the data that questions and says that iron is dangerous. So, is it safe to give iron? One of the main drivers of the data to say that iron was unsafe was Allan Collins at the University of Minnesota. Perhaps part of Allan's 15 minutes of fame was pushing the concept that higher doses of IV iron were associated with worse outcomes. And in this case, it was infection-related deaths. And this was based on the USRDS data, and also correlated with Medicare billing. The higher the bars, the higher the relative risk. But a few things you ought to notice. First, the relative risk doesn't even go up to .2 increased risk of death, and second, we don't really see a particularly good dose response curve in that this is the highest bar, when patients were getting the second lowest dose of IV iron on a regular basis. And in fact, the patients who got much more, these two bars, are actually lower. So there wasn't a perfect correlation. All of these have the flaw of using a database, the USRDS and Medicare billing, that are very suspect in terms of accurately identifying the amount of administered iron.
Iron and infection: European data
In fact, in 1998, Hoen et al, in a study of just under 900 patients, did a prospective multicenter study in Europe and found no association between serum ferritin and the risk of getting bacteremic. And there was no association of cumulative iron dose administered to the patients and the risk of developing bacteremia. So why did patients get bacteremic in this prospective study? If you had a catheter in, you were much more likely to get bacteremic. If you had a previous bacteremic episode, if you were on immunosuppressives, you were at increased risk. But the cumulative iron dose, in particular, did not correlate with the risk of increased infection. But because serum ferritin is an acute phase reactant, we run into this problem where ferritins do rise in infected patients, and it's easy to say, "Oh, maybe they got too much iron and this is what led to the infection". But this prospective study did not support such a notion.
Iron and cardiovascular risk
When we look at cardiovascular disease in a soon to be published review article, Fishbane, Nissenson and Kalantar-Zadeh make a few points. First, in a study of 2,000 patients there was really no association between higher serum ferritins and increased cardiovascular risk. And if you look at the ultimate iron overloaded patient, those with hemochromatosis, when you have matched controls, iron-overloaded patients actually have a lower incidence of coronary artery disease than patients with normal iron stores. While there have been some conflicting studies, overall, the risk of a cardiovascular event due to high iron stores is probably not a factor in considering what should be the upper limit, in the opinion of these authors.
IV iron dose and outcomes: Time-dependent analysis
And when we go back to one study that looked at a analysis of risk of IV iron administration and increased mortality from cardiovascular infectious deaths, we have the data by Harv Feldman and others at the University of Pennsylvania. In this particular study they used data from the Fresenius database where they could really understand how much iron patients received over a period of time. More importantly, they looked at a time varying and weighted model, because if you think of your sick patients, they get anemic. And when you read our protocols, one of the things we do with our anemic patients is, we give them more EPO, and we give them more iron. So both of those are otherwise going to segregate with sicker patients and ultimately higher mortality.So by doing multivariable analyses and sophisticated epidemiologic adjustments, which I know nothing about, by the way, so don't dare ask a question at the end ... he finds that when you take the Collins-like data (from their USRDS analysis), which gives you an increasing risk with increasing iron administration, and correct for time-dependent covariates, there is no risk of increased risk of death with administering more and more iron to our patients. That doesn't mean we should do it. It simply means that the concern about safety is not supported by the present data that we have.
Benefits of IV iron therapy
What about beneficial? Is IV iron therapy not beneficial? We certainly know that IV iron administration to dialysis patients and CKD patients improves erythropoiesis. It also corrects all those non-hematologic effects that Rajiv talked about in the first talk. Without adequate iron stores, EPO is relatively ineffective.
Effectiveness of IV iron in randomized trials
That point is demonstrated in this study, which is a classic dialysis study.This is the MacDougall study. Brand new dialysis patients. Very anemic. They get a set dose of EPO, and they get either IV iron, oral iron or no iron at all. And only the patients getting EPO plus IV iron normalized their hemoglobin into the target range over the 16 weeks of the study. And oral iron was no better than no iron. Four separate randomized trials showing that.
Independent erythropoietic effect of IV iron
But what we also sometimes don't appreciate is that when you give maintenance iron to patients, you are really having an independent erythropoietic effect on the patients. You're stimulating red cell production. To demonstrate that, I'll show this study by Fudin and others, published in 1998. In this particular study, there are all brand new dialysis patients. 39 hemo patients. None of them received any EPO during this entire study. And this was not a US study, so if you have ethical objections to the design of this study, don't talk to me about it. All the patients were iron deficient at baseline by bone marrow aspiration. They had zero iron on staining. And of interest, the average ferritin was about 280 in these patients and ranged from about 80 to over 400. And that again demonstrates, how can you have no iron present and you have ferritins of 400 in these patients? They then got IV iron, oral iron or no iron at all. And just getting IV iron on a maintenance basis, the mean hemoglobin went from about eight to ten at one year, and almost 11 by 26 months. And oral iron, again, was no better than no iron at all. So we're actually stimulating an erythropoiesis. And therefore, also, reducing the need for epoetin.
Serum ferritin and responsiveness of anemia to IV iron
So why is it that certain groups are saying that, what we really need to do is not give iron when the ferritin is greater than 500 ng/mL, that when the ferritin reaches 500, there is a lack of efficacy of IV iron. And probably the best demonstration of that argument is set forth in a JASN supplement which just came out in December, 2004 by George Aronoff, and this data has also been quoted several times by Dr. David Van Wyck. They showed the treatment success rate percentages shown as a function of their serum ferritin values.
So when you look at this, first you have serum ferritin values around 300 ng/mL, and four different studies were analyzed. And there are various markers that were reputed to be taken out of that data to make this graph. If you look at this and you take this ferritin of 300, and you've got these two marks here, representing the Lin and Fishbane studies; and in one the anemia response rate is about 15% and in the other the response rate (of anemia to IV iron) is at 10%. Now, I think any human being looking at this chart would say that that means, that if you have 100 people with a ferritin of 300, and you gave them IV iron, then only 10 or 15% would have an increase of a hemoglobin of one over eight weeks, or a five point increase in hematocrit. Or a 10% decrease in EPO over four weeks. That's how I would read this, and when you get to a serum ferritin of 500 ng/mL, he (Aronoff) says that the Fishbane study shows that nobody above 500 ng/mL responded to IV iron.
Serum ferritin and responsiveness of anemia to IV iron: Lin study
Let's take a look at those studies. This is the Lin study. First, there were only 32 patients in this study, and 75% of them responded to IV iron by having their hemoglobin go up by at least one gram with administration of iron. The conclusion of the study investigators was, that there were no significant differences in the basal TSAT or ferritin between the people whose Hgb responded to IV iron vs. those who didn't. In addition, there were no iron indices with adequate or acceptable sensitivity and specificity to guide you when to give iron or when not to give iron. And that's why these investigators concluded that the way to decide if a patient was going to respond to IV iron was to give the patient a dose of IV iron and see if their hemoglobin went up.
Lin study (continued)
When you look at the actual data in this study, here is the sensitivity data, and this is the percent of patients who responded to the IV iron, and there were 26 of them. And, between ferritins of 300 and 500 ng/mL, there were four patients who got IV iron and who responded with a hemoglobin increase of at least 1.0 g/dL. In fact, there was only one patient who didn't respond to IV iron who had a serum ferritin between 300 and 500 ng/mL. So when he took five patients between 300 and 500 and gave them iron, four of the five had their hemoglobin go up by 1.0 gram or more. By my count, that's not zero to 15% response rate; that's an 80% response rate.
What about the Fishbane study? Dr. Fishbane is world renowned as an expert in anemia, and he has done really seminal studies in this area. And in this particular study there were 47 patients total. As in the Lin study, there was not a single patient who had a serum ferritin above 500 ng/mL. Presence of anemia was not required as an entry criterion; you could have a hemoglobin of 12.5 g/dL and get into this study. And Dr. Fishbane concluded that there was no statistically significant difference in baseline ferritin or transferrin saturation between responders to IV iron vs. the nonresponders. Although he did suggest, and I have to talk to him about this ... that a ferritin of less than 300 ng/mL would be a good guide for iron management. In fact, his data shows no support for such a recommendation, because this is his data.
Fishbane study (continued)
And when you look at the patients (in the Fishbane study) with ferritins between 300 and 500 ng/mL, there were three patients who responded to IV iron with a 5% increase in hematocrit or a 10% drop in EPO requirements, and there were three patients who didn't respond. That's a 50% response rate. So when you go back to this summary slide, what this really is, is a sensitivity curve, and actually, it should be going from 100% sensitivity for identifying all the patients in this population who are iron deficient. But this says nothing about using ferritin as a way to identify your iron deficient patients.
Summary of studies supporting a serum ferritn cutoff of 500 ng/mL
In fact, four studies of differing design have been quoted to claim that there's no benefit of giving IV iron when the serum ferritin is high. Two studies didn't even include patients with serum ferritins above 500 ng/mL. Those are the two I showed you, and you can't figure out from reading the papers what happened to those individuals.
High serum ferritins and low transferrin saturations
So to say that there's no benefit of giving IV iron when the serum ferritin is above 500 ng/mL, I think, is a gross distortion of these results. None of these studies examined actually have looked at the patients whom we're most interested in - our anemic patients who have high ferritins and low transferrin saturations. And we really need to do studies in such patients; in fact, Watson is funding a study to look at just that population. So when we get to beneficial, there's clearly benefits to IV iron, but a serum ferritin cut-off of 500 ng/mL is not supported by the data that's presently being cited.
So what about cost effectiveness? We as nephrologists need to recognize that this is a multibillion dollar business. This looks at the top ten recombinant DNA protein products between 2001 and 2003 from Nature Biotech. Number one, two and nine on the list are Procrit (epoetin), EPO and Aranesp (darbepoetin). The present increase in growth of this is about 12% to 13% per year, and in 2003, this represented a world wide sales of $8 billion. The entire US IV iron market is estimated to be $300 million, with the world wide market estimated to be $500 million. So if you lower the serum ferritin cutoff 800 to 500 ng/mL to guide IV iron therapy, it doesn't matter what you as the individual will do, there will be selected patients who would have benefitted from iron who won't get it, and they'll get an increased EPO dose. And this cost figure (for erythropoetin-type drugs) will continue to rise unabated.
Maintenance IV iron: Costs and outcomes
Maintenance IV iron studies have been done and this dosing strategy has been shown to be virtually universally efficacious. This is a study by Chang which looked at 149 patients treated for an entire year with 50 to 100 milligrams of IV iron per week. Hematocrit increased over the year by 3.6% and EPO use went down by 25% with a cost savings for the patients, but not necessarily a cost savings for the dialysis unit because of our present reimbursement mix. Too often providers make profits from drugs (to treat anemia) and this drives, I think, imprudent decisions.
Maintenance IV iron and EPO requirements
Other clinical studies have shown significant improvements in EPO usage when you give maintenance iron, and here's one of them. Duration anywhere from four months to 12 months. Various drugs. Iron dextrans, iron saccharate. Ferric gluconate. Huge reductions in EPO with maintaining patients on maintenance iron.
So when we go back to the list of reasons to have a 500 ng/mL serum ferritin cut-off - Not beneficial - this clearly hasn't been shown in the patients that we want to look at, and certainly not based on the data. Not safe - this is not supported by the present data that we have. And not cost effective - the studies clearly show that regular IV iron administration is cost effective. At the present time, we are in, and hopefully are moving out, of a management model that does not have our providers respond to cost efficiencies. Physicians have to recognize that there are economic pressures that influence the debate over iron use, EPO dose and hemoglobin targets.
Reticulocyte hemoglobin content
Where do we go if ferritin has almost no value in identifying our iron deficient patients? One alternative is reticulocyte hemoglobin content, or CHR. This measures the amount of hemoglobin in the reticulocytes that have left the bone marrow in the last 24 to 48 hours. If you are iron deficient today, tomorrow, your reticulocytes will have decreased amounts of hemoglobin. And if you're profoundly anemic from iron deficiency but I repair it, within 48 hours, the reticulocyte hemoglobin will normalize. So this gives you an immediate snapshot of whether the patient is iron deficient today. It's performed on autoanalyzers, and is usually available by simply writing an order, and in most of the major dialysis chains, a physician is able to add the CHr on to the regular patient CBC (complete blood count) that they send off. This has been of some value.
CHr in assessment of iron status: Tessitore study
Tesitore et al, in a paper published in the NDT, looked at 125 patients. He gave all 125 a gram of iron, or 992 milligrams. Fifty-one of the patients responded, so he labeled those as iron deficient. And when he went back to the baseline parameters that he measured in these patients, the responders could be identified by looking at the percent hypochromic red cells, which I'll get back to, and second best was CHr (reticulocyte hemoglobin content). Markers that were not particularly effective included the transferrin saturation, and, in particular, the serum ferritin level. He used a cut-off for CHr of 29. If the value was below 29, patients were very likely to respond to IV iron. If CHr was above 29, patients were unlikely to respond.
Mittman also did a similar study. He used a less than ideal readout. He gave everybody IV iron, 500 milligrams, and looked at a one percent increase in reticulocyte index and found that a CHr of less than 28 identified patients who were going to respond to the IV iron. This was using an older machine (to measure CHr), and I'll talk about at the end what CHr limits you probably ought to use.
Chuang looked at administering a large dose of iron split over a six month period, and looked at changes in hemoglobin and found a CHr of 28 as sensitive and specific in terms of predicting responsiveness of anemia to IV iron. In fact, 78% of the patients who responded IV iron had CHr values that were below 28. And of the patients who didn't respond to IV iron, 87% had a CHr value above 28.
A CHr test is easy to perform. You justify use of the CHr test because the ferritin and TSAT are confounding in your particular patients. It's accurate and it's inexpensive. The CHr is an ideal test in the US, because it's not affected by shipping off the blood to a central laboratory, and it can be measured two days later. It's actually widely available in commercial dialysis units. Most commercial dialysis chains and Spectra, which does a lot of private dialysis chain patient laboratory measurements, has the CHr test available as a simple add on, and you can order it. The optimal cut-off value for the CHr has not been clearly identified, and in general, I would say that if you measure it in a patient and it's less than 30, I would consider giving them iron. Probably a gram of iron as a cumulative dose over a couple weeks. If the CHr is above 30, it's very unlikely that such a patient is going to respond to IV iron, and the higher the CHr value is above 30, the less likely a response to IV iron is. So in my high ferritin, low TSAT patients, this is a handy test to check to decide on whether I'm going to give them iron.
Percentage of hypochromic RBCs
The other test which is, at least in studies that have looked at it, turns out to be better than CHr, is the percentage of hypochromic red blood cells (on a complete blood count). If you've been iron deficient for the last month, your percent red cells that are hypochromic is increased. So this test is not looking at whether you're iron deficient today, it's really looking at whether you've been iron deficient for the last one or two months. But the problem with this test is that it's a test of concentration (of hemoglobin in red cells). So if the sample sits around, the red cells swell, and you get a higher and higher percentage of hypochromic red cells identified on this test. So the test loses sensitivity when you have to ship it from St. Louis to California, like we do for our labs. If you have this measured in your local lab, if you have a patient on dialysis and you get a local CBC (complete blood count) and they can measure percent hypochromic red cells, then this is a very sensitive test for identifying iron deficiency. And in general, most centers have used cut-offs of five or ten percent. If it's above five or ten percent, they're definitely iron-deficient, and you can consider giving them a course of IV iron.
Reassessing the serum ferritin cutoff: Conclusions
So to conclude, serum ferritin alone has a low sensitivity and specificity for determining both iron overload and iron depletion, or responsiveness of anemia to IV iron therapy. The primary goal is to improve erythropoiesis in our patients, and not some rigid adherence to iron measures so that our data might look good for the state monitors who come in. The setting of a low ferritin ceiling may actually prevent anemia correction. It may prevent us from treating the nonanemic iron deficiency side effects that Dr. Agarwal talked about. And it will almost certainly lead to higher EPO usage and the risks associated with higher EPO doses. So I'll stop there, and I'll be glad to answer any questions you might have. Thank you for your attention.
- IV. NKF-K/DOQI Clinical Practice Guidelines for Anemia of Chronic Kidney Disease: update 2000. Am J Kidney Dis. 2001 Jan;37(1 Suppl 1):S182-238.
- Horl WH, Jacobs C, Macdougall IC, Valderrabano F, Parrondo I, Carveth BG, Thompson K. European best practice guidelines 17-18: adverse effects. Nephrol Dial Transplant. 2000;15 Suppl 4:51-6.
- Kalantar-Zadeh K, Rodriguez RA, Humphreys MH. Association between serum ferritin and measures of inflammation, nutrition and iron in haemodialysis patients. Nephrol Dial Transplant. 2004 Jan;19(1):141-9.
- Collins AJ, Hao W, Xia H, Ebben JP, Everson SE, Constantini EG, Ma JZ. Mortality risks of peritoneal dialysis and hemodialysis. Am J Kidney Dis. 1999 Dec;34(6):1065-74.
- Hoen B. Iron and infection: clinical experience. Am J Kidney Dis. 1999 Oct;34(4 Suppl 2):S30-4.
- Fishbane S, Kalantar-Zadeh K, Nissenson AR. Serum ferritin in chronic kidney disease: reconsidering the upper limit for iron treatment. Semin Dial. 2004 Sep-Oct;17(5):336-41.
- Feldman HI, Joffe M, Robinson B, Knauss J, Cizman B, Guo W, Franklin-Becker E, Faich G. Administration of parenteral iron and mortality among hemodialysis patients. J Am Soc Nephrol. 2004 Jun;15(6):1623-32.
- Macdougall IC, Tucker B, Thompson J, Tomson CR, Baker LR, Raine AE. A randomized controlled study of iron supplementation in patients treated with erythropoietin. Kidney Int. 1996 Nov;50(5):1694-9.
- Fudin R, Jaichenko J, Shostak A, Bennett M, Gotloib L. Correction of uremic iron deficiency anemia in hemodialyzed patients: a prospective study. Nephron. 1998;79(3):299-305.
- Aronoff GR. Safety of intravenous iron in clinical practice: implications for anemia management protocols. J Am Soc Nephrol. 2004 Dec;15 Suppl 2:S99-106.
- Lin JL, Chang MY, Tan DT, Leu ML. Short-term small-dose intravenous iron trial to detect functional iron deficiency in dialysis patients. Am J Nephrol. 2001 Mar-Apr;21(2):91-7.
- Fishbane S, Kowalski EA, Imbriano LJ, Maesaka JK. The evaluation of iron status in hemodialysis patients. J Am Soc Nephrol. 1996 Dec;7(12):2654-7.
- Mittman N, Sreedhara R, Mushnick R, Chattopadhyay J, Zelmanovic D, Vaseghi M, Avram MM. Reticulocyte hemoglobin content predicts functional iron deficiency in hemodialysis patients receiving rHuEPO. Am J Kidney Dis. 1997 Dec;30(6):912-22.
- Low CL, Bailie GR, Eisele G. Sensitivity and specificity of transferrin saturation and serum ferritin as markers of iron status after intravenous iron dextran in hemodialysis patients. Ren Fail. 1997 Nov;19(6):781-8.
- Chang CH, Chang CC, Chiang SS. Reduction in erythropoietin doses by the use of chronic intravenous iron supplementation in iron-replete hemodialysis patients. Clin Nephrol. 2002 Feb;57(2):136-41.
- Besarab A, Amin N, Ahsan M, Vogel SE, Zazuwa G, Frinak S, Zazra JJ, Anandan JV, Gupta A. Optimization of epoetin therapy with intravenous iron therapy in hemodialysis patients. J Am Soc Nephrol. 2000 Mar;11(3):530-8.
- Fishbane S, Frei GL, Maesaka J. Reduction in recombinant human erythropoietin doses by the use of chronic intravenous iron supplementation. Am J Kidney Dis. 1995 Jul;26(1):41-6.
- Senger JM, Weiss RJ. Hematologic and erythropoietin responses to iron dextran in the hemodialysis environment. ANNA J. 1996 Jun;23(3):319-23; discussion 324-5.
- Sunder-Plassmann G, Horl WH. Importance of iron supply for erythropoietin therapy. Nephrol Dial Transplant. 1995 Nov;10(11):2070-6.
- Taylor JE, Peat N, Porter C, Morgan AG. Regular low-dose intravenous iron therapy improves response to erythropoietin in haemodialysis patients. Nephrol Dial Transplant. 1996 Jun;11(6):1079-83.
- Tessitore N, Solero GP, Lippi G, Bassi A, Faccini GB, Bedogna V, Gammaro L, Brocco G, Restivo G, Bernich P, Lupo A, Maschio G. The role of iron status markers in predicting response to intravenous iron in haemodialysis patients on maintenance erythropoietin. Nephrol Dial Transplant. 2001 Jul;16(7):1416-23.
- Chuang CL, Liu RS, Wei YH, Huang TP, Tarng DC. Early prediction of response to intravenous iron supplementation by reticulocyte haemoglobin content and high-fluorescence reticulocyte count in haemodialysis patients. Nephrol Dial Transplant. 2003 Feb;18(2):370-7.