ACD is extremely common and, overall, is probably more common than any anemia syndrome other than blood loss with consequent iron deficiency. Cash and Sears evaluated all the anemic individuals admitted to the medical service of a busy municipal hospital during two 2-month periods in 1985 and 1986.⁹ After patients with active bleeding, hemolysis, or known hematologic malignancy were excluded, 52% of anemic patients met laboratory criteria for ACD.⁹ The syndrome is also observed in 27% of outpatients with rheumatoid arthritis¹⁰ and in 58% of new admissions to inpatient rheumatology units.¹¹ However, it should be remembered that 40% of patients in the series reported by Cash and Sears lacked one of the traditional “ACD-associated disorders.”⁹ Approximately one-third of this latter group had renal insufficiency, in which pathophysiologic mechanisms implicated in ACD are active.¹² Clinical disorders commonly associated with ACD are listed in Table 41.1.

Because this type of anemia occurs in association with so many diseases, the clinical manifestations necessarily vary widely. Usually, the signs and symptoms of the underlying disorder overshadow those of the anemia, but on rare occasions, developing anemia provides the first evidence of the primary condition. This situation may be observed particularly in difficult-to-diagnose clinical syndromes, such as temporal arteritis.¹³

The concentration of free protoporphyrin in the erythrocytes (FEP) tends to be elevated in patients with ACD.¹⁴, ⁵⁵, ⁵⁶ However, FEP increases more slowly in ACD than it does in iron deficiency, and it does not become clearly abnormal until significant anemia has developed.

A variety of other biochemical changes often are detected in patients with chronic diseases. Many of these changes reflect alteration of levels of particular plasma proteins, often called acute-phase reactants.⁵⁷, ⁵⁸, ⁵⁹, ⁶⁰, ⁶¹ The concentrations of certain plasma proteins, such as fibrinogen, ceruloplasmin, haptoglobin, C-reactive protein, orosomucoid, C3, and amyloid A protein, increase⁶², ⁶³, ⁶⁴, whereas the concentrations of albumin and transferrin characteristically decrease.⁶⁵ The increase in ceruloplasmin accounts for the increase in serum copper levels often noted in association with chronic diseases.², ¹⁴ An elevated fibrinogen level is probably the most important factor in the increased sedimentation rate.

Patients with chronic illness develop accelerated protein catabolism and negative nitrogen balance associated with muscle proteolysis.⁶⁴, ⁶⁵ Over time, this phenomenon can result in muscle wasting, increased urea excretion, weight loss, and growth impairment in children. However, protein catabolism generates amino acids that can be used by the patient as alternative energy sources or to supply substrates for biosynthetic processes related to host response. In this context, it is perhaps relevant that elevated serum levels of tumor necrosis factor (TNF), a cytokine implicated in the pathogenesis of ACD, are noted in patients with significant malnutrition.⁶⁶, ⁶⁷

Erythrocyte survival is modestly but significantly reduced in patients with ACD. In two studies comparing red cell survival in anemic patients with rheumatoid arthritis to red cell survival in normal individuals, the mean red cell survivals noted were 81 days versus 98 days, and 90 days versus 114 days, respectively.⁶⁸, ⁶⁹ In a similar study comparing red cell survivals in 10 anemic patients with a variety of chronic inflammatory states to 10 normal individuals, the mean values observed were 80 days versus 88 days, respectively.⁶⁸ The usual manifestations of increased blood destruction, such as increases in serum bilirubin values and urobilinogen excretion, are not typically observed.², ⁷⁰

There is little evidence of a compensatory erythropoietic response to this reduction in red cell survival. The reticulocyte count usually is normal or decreased, and little or no erythroid hyperplasia of the marrow is observed. The pathogenetic significance of these findings is discussed in the following section. Kinetic data indicate that anemia develops because the bone marrow fails to increase red cell production sufficiently to compensate for a mild decrease in the lifespan of the red cells.², ⁸, ⁶⁸ Ferrokinetic studies involving patients with chronic infections,⁷¹, ⁷² rheumatoid arthritis,⁷³ and various malignant diseases²¹, ²², ⁷⁴ reveal that the rate of disappearance of iron from the plasma is rapid, the plasma iron transport rate is normal or slightly increased, the uptake of iron into erythrocytes and the amount of iron turning over through red cells daily are normal or increased, and the fraction of red cells renewed daily is increased.² When techniques that allow the division of marrow iron turnover (a measure of total erythropoiesis) into red cell iron turnover and ineffective iron turnover were used, marrow iron turnover was normal in patients with ACD.⁶⁸, ⁶⁹, ⁷⁵, ⁷⁶ Ineffective iron turnover was also normal, or even less than normal, indicating a lack of ineffective erythropoiesis as traditionally defined.⁶⁸ In iron-deficient subjects, ineffective iron turnover is increased, perhaps because of greater stimulation of the marrow.

Efforts to clarify the pathogenesis of ACD have focused on three principal abnormalities: shortened erythrocyte survival, impaired marrow response, and disturbance in iron metabolism. The modest shortening of the erythrocyte survival creates an increased demand for red cell production on the marrow. Normally, the marrow could easily accommodate this demand, but in the setting of ACD, the marrow is unable to respond fully because of a combination of a blunted erythropoietin (Epo) response, an inadequate progenitor response to Epo, and limited iron availability (Fig. 41.2).

ACD is one manifestation of the systemic response to immunologic or inflammatory stress, which results in the production of various cytokines.³, ⁷⁷ The ability to trigger this cytokine response appears to be the common pathogenetic factor shared by the various conditions associated with this anemia syndrome.⁷⁸ The central role of these molecules suggests that ACD may be best understood as a cytokine-mediated process.⁶ The cytokines most often implicated in the pathogenesis of ACD are TNF,⁷⁹, ⁸⁰, ⁸¹ IL-1,⁸², ⁸³ IL-6,⁸⁴ and the interferons,⁸⁵, ⁸⁶, ⁸⁷ concentrations of which have been reported to be increased in the serum or plasma of patients with disorders associated with ACD.⁸⁵, ⁸⁷, ⁸⁸, ⁸⁹, ⁹⁰, ⁹¹ Therapeutic administration of TNF or interferon may induce anemia.⁹², ⁹³ The role of IL-6 is complex. IL-6 administration itself does not suppress erythropoiesis but rather is associated with increased plasma volume; therefore there is a dilutional component to any observed anemia.¹⁷ The association of IL-6 with hepatocyte activation may thus explain the dilutional component of the anemia observed in liver disease.⁹⁴, ⁹⁵, ⁹⁶ However, IL-6 is also a potent inducer of hepcidin, proposed as a major mediator of the iron abnormalities of ACD.⁵

While inflammation-induced cytokine activation is clearly the initial event in the pathogenesis of ACD, the liver-produced antimicrobial peptide hepcidin appears to be the most important mediator through which cytokines exert their effects on the pathogenetic mechanisms of ACD.⁵, ⁹⁷, ⁹⁸ Hepcidin is an acute-phase-reacting peptide, largely regulated by IL-6 but also by cytokine pathways not linked to IL-6.⁵, ⁹⁹, ¹⁰⁰ Patients with hepatic adenomas secreting hepcidin exhibit a hypoferremic anemia that resolves with resection of the adenoma.⁹⁷ Urinary hepcidin excretion is strongly correlated with serum ferritin concentration and is markedly elevated in patients with anemia of inflammation (defined as anemia
with a characteristic clinical setting and an elevated serum ferritin) compared to iron-deficient patients.⁵ Hepcidin promotes macrophage iron retention by causing internalization of the iron export protein ferroportin.¹⁰¹

Of the pathogenetic mechanisms shown in Figure 41.2, hepcidin clearly drives the abnormalities in iron metabolism. However, it may be linked to other elements as well. Under conditions of limited Epo availability, hepcidin is associated with impaired erythroid colony formation in vitro.¹⁰² Increased circulating hepcidin appears to be linked to the degree of resistance to recombinant (rh) Epo therapy in dialysis patients.¹⁰³ Hepcidin and Epo production appear to be regulated in an inverse relationship by hypoxiainducible factor,¹⁰⁴ which may potentially provide some linkage between hepcidin and the relative Epo deficiency of ACD. This latter connection remains to be demonstrated.

The rate of survival of cells from patients with arthritis, when transfused into normal subjects, is normal, and the survival of red cells from normal individuals in the circulation of patients with arthritis is less than the normal rate.², ²³ Therefore, shortened red cell survival in patients with chronic inflammatory disorders is attributed to an extracorpuscular mechanism.³⁰ IL-1 levels and shortened red cell survival are correlated in anemic patients with rheumatoid arthritis,¹⁰⁵ and mice that become anemic after exposure to TNF in vivo also exhibit a shortened red cell survival.¹⁰⁶ Neocytolysis, a selective hemolysis of newly formed erythrocytes associated with Epo deficiency,¹⁰⁷, ¹⁰⁸ has been proposed as a mechanism for shortened red blood cell survival in ACD. Peroxynitrite, derived from the reaction of the cytokine second messenger nitric oxide and superoxide, may contribute to red cell rigidity and thus to decreased survival.¹⁰⁹

Normal bone marrow, capable of a six- to eightfold increase in the red cell production rate, should easily compensate for such a modest reduction in erythrocyte survival. Its failure to do so in ACD suggests that impaired production capacity is of fundamental importance in the pathogenesis of this condition. The possible defects in erythropoiesis fall into three categories: inappropriately low Epo secretion, diminished marrow response to Epo, and ironlimited erythropoiesis.

An inverse relationship between serum or plasma Epo levels and hemoglobin normally exists: As the hemoglobin decreases, the Epo level rises.¹¹⁰ A similar inverse relationship between hemoglobin and Epo level exists in anemic individuals with rheumatoid arthritis, cancer, and human immunodeficiency virus infection¹¹¹, ¹¹², ¹¹³, ¹¹⁴; however, for any given anemic patient in these disease categories, the Epo level was lower than that found in equally anemic individuals with iron deficiency, indicating that the Epo response to anemia was blunted. This impaired Epo response is cytokine-mediated. IL-1, TNF-α, and transforming growth factorβ inhibit production of Epo by various hepatoma cell lines and by isolated perfused rat kidneys in vitro.¹¹⁵, ¹¹⁶ This effect occurs at the messenger RNA level.¹¹⁵ It has been proposed that this inadequate Epo secretion is adaptive; that is, it reflects reduced tissue oxygen use so that normal oxygenation is maintained despite reduced hemoglobin levels. One colorful description of this process is that the “hematologic thermostat has been turned down a bit”,⁸ perhaps to allow diversion of substrates normally used in erythropoiesis to more immediately critical activities. Changes that ordinarily signify erythrocyte adaptation to tissue hypoxia, such as increased erythrocyte 2,3-diphosphoglycerate (2,3-DPG) levels, are observed in ACD.¹¹⁷ Hemoglobin oxygen affinity is also slightly decreased in these patients but overlaps the range observed in normal subjects.¹¹⁷

Although the Epo levels of patients with ACD are lower than those observed in equally anemic iron-deficient individuals, they are still higher than are observed in normal individuals who are not anemic. This implies that inhibition of Epo production cannot entirely account for the impaired erythropoiesis associated with ACD, and that the erythroid progenitors themselves exhibit an abnormal response to Epo. Studies of anemic cancer patients support this concept.¹¹⁸, ¹¹⁹

TNF, IL-1, and interferon have all been reported to inhibit erythropoiesis in vivo and in vitro.⁹⁰, ⁹², ¹²⁰, ¹²¹, ¹²², ¹²³, ¹²⁴, ¹²⁵, ¹²⁶, ¹²⁷, ¹²⁸ TNF inhibits in vitro colony formation by human erythroid colony-forming units (CFU-E) indirectly, through interferon-β and possibly other soluble factors released from marrow accessory cells in response to TNF.¹²⁹, ¹³⁰ The inhibitory effect of rhIL-1 on human CFU-E colony formation was also indirect and dependent on marrow accessory cells; however, the responsible accessory cells in this case were T lymphocytes, and the mediator was interferon-γ, which also exhibits a direct inhibitory effect on human CFU-E.¹³¹ Although interferon-α and interferon-β share a common receptor, interferon-α does not exert a direct inhibitory effect on human CFU-E colony formation. Instead, its effect is mediated through an unidentified soluble factor released by T lymphocytes.¹³² The inhibitory effects of interferon-α and interferon-β exhibit synergy with interferon-γ.¹³² An important point to consider in reviewing the various models for cytokine effects on erythropoiesis is that none of the systems described operates in isolation in vivo. Thus, TNF induces IL-1 production by macrophages, IL-1 induces interferon-γ production by T lymphocytes, and interferon-γ can exhibit positive or negative feedback on production of IL-1 and TNF.

Treatment with recombinant human Epo (rhEpo) can correct ACD in many cases. Inhibition of in vitro erythroid colony formation by interferon-γ, but not interferon-α and interferon-β, can be corrected by exposure to high concentrations of Epo.¹³², ¹³³ Epo also suppresses hepcidin production.¹³⁴, ¹³⁵

It has been proposed that lack of iron for erythropoiesis contributes to the inadequate marrow response in ACD. Evidence of a functional iron deficiency in this syndrome includes erythrocyte microcytosis, increased FEP,¹⁴, ⁵⁶ reduced transferrin saturation,²⁴ and decreased marrow sideroblasts.²⁴

Ferrokinetic studies have proved confusing. It has been reported that erythroid iron turnover correlated with serum iron levels, permitting a conclusion that marrow proliferation was limited by availability of iron in ACD.¹¹⁸ Other studies found no correlation between marrow iron turnover and serum iron concentration.⁶⁸, ⁶⁹, ⁷⁶, ¹³⁶

The major contributor to hypoferremia in patients with ACD is probably a shift of iron from a transferrin-bound, available circulating state to an intracellular storage state. Iron absorption appears to be normal, but iron tends to remain in the mucosal cell and in hepatocytes.¹³⁷, ¹³⁸ Macrophages, the major site from which iron is obtained for erythropoiesis, also exhibit increased iron storage. This process is mediated by hepcidin, which downregulates the iron egress regulatory ferroportin, thus trapping iron intracellularly.¹⁰¹

The dominant factor in the iron abnormalities of ACD is clearly hepcidin, discussed above and at greater length in Chapter 23. A number of other processes of less overarching significance may contribute under particular circumstances to the iron anomalies observed. Apoferritin is normally synthesized in response to increased intracellular iron concentration.¹³⁹ It has been suggested that excess apoferritin is made in inflammatory and malignant conditions, and the surplus binds a larger-than-usual amount of iron entering the cell.¹⁴⁰, ¹⁴¹, ¹⁴² In effect, such a mechanism would divert iron from the rapid to the slow pathway of iron release. Rodents injected with recombinant TNF developed
a hypoferremic anemia associated with impaired storage iron release and incorporation into erythrocytes.¹⁰⁶, ¹⁴³ It has also been reported that IL-1 increases translation of ferritin messenger RNA and that this additional ferritin could act as a trap for iron that might otherwise be available for erythropoiesis.¹⁰⁵ Nitric oxide, which is a common mediator of cytokine effects, has similar effects on ferritin expression.¹⁴⁴, ¹⁴⁵ Lactoferrin is a transferrinlike protein found in neutrophil-specific granules.¹⁴⁶, ¹⁴⁷ It is released from the neutrophil during phagocytosis or stimulation by IL-1.¹⁴⁸ At low pH, lactoferrin binds iron more avidly than does transferrin.¹⁴⁹ Lactoferrin-bound iron is not immediately available for erythropoiesis¹⁵⁰; rather, it binds to specific receptors on macrophages (particularly in the liver and spleen) and is endocytosed and subsequently incorporated into ferritin.¹⁵¹, ¹⁵² Thus, lactoferrin transfers iron from its transferrin-bound, circulating state to a storage state, from which it cannot be rapidly mobilized.¹⁵³ One group of investigators found that reducing the available lactoferrin by inducing neutropenia in rats blunted the hypoferremic response to IL-1,¹⁵⁴ but another group found that administration of recombinant IL-1-induced hypoferremia in mice, even in the face of severe neutropenia.¹⁵⁵

In addition to decreased availability of iron, erythroid progenitors may also be unable to fully use the iron available to them. Erythroblasts from anemic patients with rheumatoid arthritis express fewer surface TfR than do erythroblasts from normal individuals. These TfR also exhibit lower binding affinity for transferrin.¹⁵⁶ Furthermore, acute-phase reactants, such as α₁-antitrypsin, impair transferrin binding to erythroblasts and also inhibit transferrin internalization.¹⁵⁷ rhEpo appears to induce a greater level of TfR expression on erythroid cells.¹⁵⁸

Much of the anemia commonly observed in patients with cancer can be attributed to the mechanisms involved in ACD; however, certain processes unique to malignancy may also contribute. Erythroid precursors may be displaced from marrow by metastatic tumor,¹⁵⁹ tumor-induced fibrosis,¹⁶⁰ or tumor-associated marrow necrosis.¹⁶¹ The treatment of cancer can also produce or exacerbate anemia by a variety of mechanisms, including impaired Epo production¹⁶² and cytotoxic effects of therapy on erythroid progenitors.¹⁶³

Chronic renal failure occurs during the final stages of several renal diseases. As a rule, the nature of the underlying disease bears little relation to the degree of anemia, although anemia may be less severe in patients with hypertensive renal disease²⁰⁵ and is considerably less severe in patients with polycystic disease.²⁰⁶, ²⁰⁷ In one series, the mean hematocrit in 12 subjects with polycystic disease was 0.297, compared with 0.212 in 24 subjects with other types of chronic renal failure.²⁰⁷ Apparently, the Epo-secreting function of the kidney is preserved in polycystic disease—even when the filtering function is lost.²⁰⁷, ²⁰⁸ Erythropoietic activity can be found in the cystic fluid and may arise from single interstitial cells juxtaposed to proximal tubular cysts.²⁰⁷

In most instances, the patient seeks medical attention because of symptoms related to the underlying renal disease, and anemia is an incidental finding. Occasionally, however, the renal symptoms are so subtle and so slowly progressive that the patient cites only symptoms of pallor, exertional dyspnea, or other signs of the cardiovascular adjustment to anemia. The severity of the anemia bears a rough relationship to the degree of renal insufficiency. Anemia is not routinely observed until the creatinine clearance falls to <45 ml/minute/1.73 m² body surface area, which corresponds roughly to a serum creatinine of 2.0 to 2.5 mg/ml in an average-sized adult.²⁰⁹ At creatinine clearance rates below that, a statistically significant correlation between creatinine clearance and hematocrit has been reported.²¹⁰, ²¹¹ However, the variation in the results of these studies is so great that one cannot reliably predict the hemoglobin level in an individual patient on the basis of renal function.

Anemia tends to become more severe as renal failure worsens, but in most patients the hematocrit ultimately stabilizes between 0.15 and 0.30.² Because regulation of body water and electrolyte balance is impaired in renal disease, the apparent degree of anemia may be exaggerated or minimized by alterations in plasma volume.

The erythrocytes usually are normocytic and normochromic, but slight macrocytosis is occasionally observed.²¹² The majority of red cells appear normal on blood smears. Occasionally, however, “burr” cells (Fig. 41.3) are observed along with some triangular, helmet-shaped, or fragmented cells. The reticulocyte count often is within normal limits,²¹³ but it may be moderately increased.²¹², ²¹⁴ In one study, the numbers of reticulocytes were normal when the blood urea nitrogen (BUN) value was <130 mg/dl; at higher BUN levels, however, their number often was increased.²¹⁵ The highest values (average, 6%) were observed with extreme azotemia (BUN, 300 to 350 mg/dl).

The leukocyte count typically is normal, but slight neutrophilic leukocytosis may be observed. In one series, the leukocyte count averaged 10.7 × 10⁹/L.²¹³ The platelet count is either normal or slightly increased,²¹³ but platelet function may be severely impaired, resulting in defective hemostasis (see Chapter 52).