by Amit Golding, MD
Postdoctoral Fellow, Division of Rheumatology
Johns Hopkins University School of Medicine.
Release Date: June 29, 2006
Expiration Date: June 29, 2008
Dr. Golding has no significant financial interest or relationships to disclose.
Today we’re going to discuss the anemia of inflammation — starting with its mechanisms, then looking at a fairly new genetic link between inflammation and anemia, and including how we can apply all of this to our patients.
Case Example 1
We will call our first patient Mr. F. He was a 78-year-old man who had been quite fit and active, but over the past two to three months he developed progressively worsening stiffness in his shoulders, in particular, as well as in his hands and hips. He also noticed somewhat diminished energy. He had been prescribed some NSAIDs, which gave him some relief, but he really did not feel back to the way he had been before this process began.
On exam, the only thing that we could find was some pain with stiffness on external rotation of his shoulders, but he was able to reach up, behind his head and behind his back, without difficulty. The first-year fellow (which was me, in this case) concluded that he had osteoarthritis of the shoulders, which for whatever reason had gotten a little bit worse, and that he would benefit from physical therapy. However, the precepting fellowship director said, and I think this is a fair paraphrasing: “Think again. An older man with a symmetric proximal joint stiffness and fatigue, slowly progressing over two months, is compelling for polymyalgia rheumatica, even if there are subtle findings on exam. Let us start him on a fairly low dose of prednisone (10 mg), stop the NSAIDs, and see how he does. Also get labs, just to support our diagnosis.” Mr. F called back in two days. He was feeling pretty well, though not terribly better, but after four days of on daily prednisone at 10 mg, he really felt a lot better. The precepting attending wanted to make sure: “Did he just feel a little bit better or a whole lot better?” At a two-week follow-up, Mr. F remarked that he “felt a tremendous difference.” The attending asked the fellow if, in reviewing the lab work, he had noticed that Mr. F.’s hematocrit had gone down 6 points from baseline before starting on prednisone. Again, the fellow had kind of brushed this off because it was still in the normal range. The attending retorted that a “6-point drop is still a 6-point drop,” and it would be interesting to see if it had gone back up with treatment. In fact, just being on two weeks of treatment, it had already started heading back up in the right direction.
That’s an introduction to the importance of anemia of inflammation and of the sometimes subtle effects of inflammation on hematocrit. The term that most of us are familiar with still is “anemia of chronic disease,” which is, to some extent, a diagnosis of exclusion because we always want to think in the back of our minds, “Is there an occult or not-so-occult bleeding, hemolysis, nutritional defect, or marrow placement by malignancies?” Assuming that those are ruled out, we then think of anemia of chronic disease in the setting of chronic infection, inflammatory disease, or neoplastic disorder. If you sum up the basic findings of laboratory testing, it is the combination of low serum iron but adequate iron stores with total body iron being normal.
Some argue that anemia of chronic disease is actually not a good phrase because there are a lot of chronic diseases — such as hypertension, diabetes, emphysema — that are not accompanied by anemia. Additionally, there are plenty of situations where people who do have chronic inflammatory diseases do not have anemia of chronic disease; rather, they have anemia due to other causes. The phrase “anemia of chronic disease” is not a helpful construct, and, at least in the hematology field, they have switched to “anemia of inflammation.” I would argue that we should do the same, primarily to clarify that the underlying mechanism is inflammatory.
Rheumatoid arthritis (RA) was one of the initial chronic diseases for which anemia of chronic inflammation was described. As seen in patients with RA, anemia of inflammation manifests in a number of ways. For one thing, there is increased erythropoietin production in RA patients, though it is not quite as robust as you see in non-RA controls with the same degree of anemia. Similarly, the response to erythropoietin, when given exogenous erythropoietin, is also less robust in an RA patient than a control. Finally, you have diminished serum iron in the setting of total body iron stores being normal or, in fact, elevated to a large extent in the form of ferratin in macrophages in the reticuloendothelial system. It has been observed that many inflammatory cytokines (such as IL1, interferon gamma, IL6, etc.), when given in mouse models, will produce a pattern of anemia very similar to what is seen in anemia of inflammation. It has also been observed that in RA patientsif you treat the disease, you often see a recovery of the anemia as well, indicating the same underlying process.
Case Example 2
Another example, also a patient of mine, take the case of a 56-year-old man, CCP positive RA. He started on methotrexate for a flare of RA, with synovitis in his hands. His symptoms resolved after being on methotrexate for 4 months. Looking back at his baseline, he had a normal range hemoglobin/hematocrit, which went down significantly during his flare, and as his symptoms improved with methotrexate treatment, his hemoglobin/hematocrit improved as well. An abstract presentation by Wolfe et al. from an American College of Rheumatology meeting in 20051, which involved more than 2,000 RA patients and close to 4,000 controls with non-inflammatory rheumatic disorders, looked at the prevalence and incidence of anemia. The researchers looked at both acute anemia and more chronic anemia, finding that up to a third of patients in the RA group had mild, intermediate, or severe chronic anemia. In addition, they found that lower hemoglobin was often correlated with clinical activity and other markers, as measured by the physician, laboratory testing, and patients’ assessments.. They also found that, in terms of statistical significance, C-reactive protein (CRP) was the strongest predictor of anemia – in other words, that the underlying inflammation was causing the anemia.
They also looked specifically at the association between hemoglobin and creatinine clearance because worsening renal insufficiently often accompanies low creatinine and decreased production of erythropoietin. They also found a similar trend whereby the lower the creatinine clearance, the lower the hemoglobin, presumably because of the same mechanism. However, that drop occurred at much lower diminutions in creatinine clearance than in the general public, indicating that there is a process causing the anemia that is totally separate from chronic renal insufficiency.
We need to briefly consider iron homeostasis in the body. Very fundamentally, iron is essential for the basic mechanism of energy metabolism, oxygen transport of hemoglobin, electron transport chain, oxidative phosphorylation, etc. However, it can also be very dangerous because of its potential to generate free radicals that can cause oxidative damage. We have many examples — hemochromatosis is one of them — of how iron overload can essentially be very toxic to the body. Therefore, it is very tightly regulated and, on a given day, only a few milligrams of iron are needed to be absorbed from the gastrointestinal tract for maintenance of normal erythropoiesis. However, the majority of total body iron (from 400 to 1,000 mg stored in the body) is mostly cycled in between the bone marrow and the reticuloendothelial system; very little is actually absorbed on a daily basis and even that is very tightly regulated.
Some of the regular values you should keep in mind include:
- Serum iron, usually in the range of 65-170 microgram/dL.
- Total iron-binding capacity, 250-450 mg/dL.
- Serum transferrin, a component of total iron-binding capacity, 200-400 mg/dL. This is the most important iron-binding protein; it shuttles among different sites, dropping off iron, for example, at the bone marrow, picking it up at the gut or elsewhere.
- Percent saturation, which indicates the amount of saturation of the binding capacity with iron, 20-55%
- Ferritin, 10-300. This is the storage form of iron, in macrophages for example.
- Transferrin receptor and soluble transferring receptor, which reflect how starved cells, especially bone marrow cells, are for iron; they’re expected to be elevated when iron stores are depleted, as in iron deficiency but not in anemia of inflammation alone.
Case Example 3
Now, let’s consider another patient. He is a 20-year-old gentleman who had significant shortness of breath, very severe gas transfer defect on his pulmonary function test (PFT) due to interstitial lung disease, active myositis, and arthritis. He turned out to be OJ-1 synthetase antibody positive. He had been on oral daily Cytoxan plus prednisone for 6 months, so his PFTs had improved, his creatinine kinase had normalized and his arthritis had resolved. Again, I was focusing on inflammatory markers: erythrocyte sedimentation rate (ESR), CRP, iron studies, and hemoglobin/hematocrit. We can see that prior to treatment he had anemia, elevated inflammatory markers, and his iron studies showed that he had low serum iron but also low total iron-binding capacity (TIBC) and low transferrin (a component of TIBC), borderline low percent saturation, but a markedly elevated ferritin. That goes along exactly with our definition of anemia of chronic inflammation with elevated total body iron stores but having low serum iron and anemia, which then improves over time with treatment of the underlying disease.
Case Example 4
The next example is a patient with gout. He is an 82-year-old man, with severe tophaceous gout and a polyarticular flare of gout. It was interesting to note, again, that he had significant anemia while at the same time elevated ferretin, indicating total body iron source was elevated, low serum iron, low TIBC, borderline low saturation, and improvement in his anemia as the disease was treated.
Case Example 5
The next patient was a woman with a very impressive bilateral scleritis, unilateral sensorineural hearing loss, and vestibular dysfunction. She responded very impressively to high-dose steroids, with almost complete resolution of the scleritis in two weeks. I did not have the follow-up iron studies, but again the initial findings of elevated inflammatory markers, low serum iron, low TIBC, in this case low saturation as well, but elevated ferritin — again, telling you that the total body iron stores were up.
Case Example 6
The next gentleman had a somewhat unusual periodic fever of unknown origin and oligoarthritis; I’m not sure we knew exactly what he had. He was admitted with fevers, in part attributed to pneumonia. Iron studies showed that serum iron was low normal, TIBC low normal, saturation normal, but, again, the strikingly elevated ferratin and the anemia.
A study published in 2000 by Mirzayan et al2 looked at multiple factors correlating biomarkers lab results with disease activity. One of the more significant findings was the degree of association between anemia and the SLEDAI score of disease activity This also held up with predicting SLEDAI score inaudible a year after the initial anemia was found.
In another study by Voulgarelis et al3, there was evidence of a blunted erythropoietin response to anemia in lupus patients. This points to one of the observations in RA, which is that, in lupus patients (similar to RA patients), those who have anemia of chronic disease or anemia of inflammation do not have a significant increase in erythropoietin production, despite becoming anemic. However, those who have acute hemolytic anemia, iron-deficiency anemia, in fact do increase their erythropoietin production with lowering of the hemoglobin. This again indicates that one of the things chronic inflammation affects is production of erythropoietin. That mechanism is not fully understood.
To change gears a bit, and address the missing link between inflammation and anemia, we need to talk about hepcidin. It is a fairly recent discovery, a very serendipitous one, that involved two groups in 2001 — the Ganz lab and the Nicholas lab. The Ganz lab4 discovered what they called LEAP: Liver Expressed Antimicrobial Peptide. They were not looking for anything having to do with inflammation per se, or regulation of iron. They were trying to screen for cationic peptides that would have antimicrobial properties in urine, and they found this peptide. At about the same time completely by chance, the Nicholas lab was trying to knock out a glucose metabolism gene, USF2. Instead they unintentionally knocked out the mouse hepcidin gene,the mouse with multi-organ system iron overload. What helped them figure out that they knocked out the wrong gene was that another group knocked out USF2 and did not get this phenotype at all. There are severe forms of juvenile hemochromatosis that are, in fact, due to mutations in the human gene for hepcidin, similar to was seen in the hepcidin knock-out mouse.
They made incredible leaps after both of these serendipitous discoveries. The same group that had unintentionally knocked out the hepcidin gene in mice then went on to transgenically over-express it in mice, and that led to the opposite phenotype of severe anemia.
One of the mechanisms by which this works is modulating iron transport across the cell membrane, mediated by a transporter called ferroprotein. Without hepcidin, the ferroprotein would continuously transfer iron from inside the cell to the adjacentt vessel. For example, if you have a cell in the crypt of the small intestine, there is going to be iron input from the luminal side. Then there are tight junctions, so that iron transport into the blood vessel side can be regulated. Hepcidin, the small peptide, binds to the ferroprotein, causing it to be internalized and degraded so that it can no longer transport iron from the gut into the blood stream.
The same iron membrane transporter, ferroprotein, is also present in macrophages, one of the body’s storage areas for iron. In the absence of significant hepcidin secretion from the liver, again, you will have constitutive transport of iron from the gut into the bloodstream and out of macrophages at some equilibrium. With the secretion of significant amounts of hepcidin, down-regulation the transporter, occurs and will again stop transport of iron into the bloodstream as well as release of iron from macrophages5. You can imagine that if you knock out hepcidin, you will always have a significant amount of absorption of iron, even if the body does not need it anymore. You will have constitutive release of iron into the bloodstream from macrophage sources as well. This is what is seen in hemochromatosis.
At this point, there are two ways of measuring hepcidin. There is a urine assay and there is a serum assay of the pro-peptide. But they are not available in mass quantity; only a few labs are currently doing this test.
Other points to remember:
- In iron-deficiency anemia you do not get hepcidin elevation because you want to try to absorb from the gut and from the body’s stores such as macrophages as much as possible.
- With anemia of inflammation, levels of hepcidin in general are much higher. The hepcidin level correlates well with the level of ferritin, which points to the fact that hepcidin prevents release of iron from macrophages.
- The exact mechanisms of how hepcidin is turned on have not been worked out. There are a number of upstream promoter binding sites, such as NF Kappa B, for example, in the mouse gene, but this has yet to be worked out.
- The process is not just about iron homeostasis. Two other key areas related to anemia of inflammation are either inappropriate production of erythropoietin or inability of the bone marrow to respond to erythropoietin.
A 2005 paper by Ripley et al6 that focused on interleukin 6 (IL6) in lupus looked at the correlation between IL6 levels and various end organ manifestations of lupus — connective tissue manifestations, mucous membranes, central nervous system, musculoskeletal, cardiovascular, vascular, renal. Compared with inactive patients with lupus, there was no significant difference in IL6 levels. However, specifically for anemia associated with active lupus, there was a significant increase in IL6 level. Another observation was that as the hemoglobin level goes down, you also have increased IL6 levels. There’s a significant difference in IL6 levels between lupus patients with anemia present versus anemia absent.
A number of studies have shown that hepcidin is activated very rapidly during acute inflammation and also that it is downstream of a number of inflammatory cytokines, especially IL-6. This was shown very elegantly in a study by Nemeth et al7 showing up-regulation of hepcidin mRNA in human hepatocytes treated with IL-6.
How can we apply the discovery of hepcidin and better understanding of anemia of chronic inflammation to our patients? A key point is how we decide whether somebody not only has an inflammatory condition causing anemia, but possibly coincident iron deficiency in particular due to occult bleed. A review by Weiss and Goodnough7 showed in anemia of chronic disease, serum iron is reduced, and the total iron-binding capacity, or the transferrin component of that, is reduced.
Ferritin, however, is normal to increased; the total body iron stores are elevated. They referred to a ratio of the soluble transferrin receptor to the log of a ferritin because the soluble transferrin is shed from developing cells in the bone marrow. They will highly up-regulate the serum transferrin receptor, especially when they are starved for iron. When elevated, this indicates the total body iron sources are probably low. If you have pure iron deficiency anemia without anemia of inflammation, serum iron is reduced but total iron-binding capacity is increased, and the serum soluble transferrin receptor is increased. If you have both conditions, anemia of inflammation together with iron-deficiency anemia, the main difference is in the total body iron stores. We would expect ferritin to be quite elevated. But if it is normal, low normal, or low, that indicates in addition to anemia of inflammation, there is also probably occult blood loss or hemolysis.
Again, this is an algorithm8 starting with anemia.
If you see the transferrin saturation at less than 16%, you can go to the ferretin first. If that is less than 30, there is no question that there is an iron-deficiency anemia. If it is greater than 100, then there is anemia of inflammation. If there are both, you would be in a gray range and would have to rely on other tests.
So, let’s consider another patient: an 84-year-old woman with limited scleroderma. Her disease had not been particularly active. She does not have significant lung disease or pulmonary hypertension, and to her knowledge did not have a significant history of anemia. When we saw her, we noticed her hemoglobin/hematocrit were significantly low. The key thing was that her ferritin was less than 30 which automatically means that we have to consider occult gastrointestinal loss. We have a number of therapeutic options. Obviously, treating the underlying disease often improves the anemia. Transfusions may be appropriate if hematocrit drops particularly low or if there is co-morbidity, such as cardiomyopathy. Iron supplementation is somewhat controversial; it is usually first-line only if you are convinced the patient has true iron deficiency from blood loss in addition to chronic inflammation. However, I should say that many, especially in the renal field, favor the prescription of erythropoietin, which also requires giving people iron at the same time. Standard practice is if you are going to be giving erythropoietin, even if the patient does not have true iron deficiency, in order for there to be a robust response to the erythropoietin, you usually include iron supplementation.
The big question involves the hepcidin molecule, which is very clearly down-stream of inflammatory markers and probably plays an important role in anemia of inflammation. So, is hepcidin elevated in autoimmune disease? I think that I have shown you enough evidence to suggest that it probably is. But, in searching PubMed, I did not find a single publication where anybody had directly measured hepcidin in any of the rheumatic diseases. If I were going to start with one, it would be adult-onset Still’s disease. As an example, a 33-year-old man presented with arthritis, myositis, rash, and fevers. He did very well with high-dose steroids that were slowly tapered. His original ferritin, commonly very elevated in adult-onset Still’s disease, was 19,000. His hemoglobin/hematocrit was low, in a pattern consistent with anemia of inflammation, and not with iron-deficiency anemia. With treatment of the underlying inflammatory disorder, the anemia was corrected. I would guess that hepcidin is very strongly elevated in patients with adult-onset Still’s disease, but, as mentioned, this has yet to be studied.
1. F Wolfe, K Michaud, national Data Bank for Rheumatic Diseases. Prevalence and Characteristics of Anemia in Patients with Rheumatoid Arthritis (RA). Arthritis & Rheumatism 52, no. 9 suppl., 2005.