Acute event‐related anaemia

Acute events in medicine often lead to anaemia. Many patients become anaemic not only after surgery but also during sepsis and during their stay in an Intensive Care Unit (ICU). At first sight, this seems the result of blood loss including frequent blood drawing and invasive procedures. But apart from hypoferraemia, the acute event-related anaemia (AERA) bears no resemblance to iron deficiency anaemia. AERA, with high ferritin and low-to-normal transferrin and serum transferrin receptor levels, shares characteristics with the anaemia of chronic disease (ACD) found in patients with infectious diseases, inflammatory disorders such as rheumatoid arthritis and with malignancies (Lee, 1983; Means & Krantz, 1992; Means, 1995; Spivak, 2000). In ACD, inflammatory stimuli influence erythropoiesis and iron metabolism leading to a state of functional iron deficiency: hypoferraemia in the presence of normal iron stores (Lee, 1983; Means & Krantz, 1992; Means, 1995; Weiss et al, 1995). Although different types of acute events will show different accents, temporary/transient inflammation is a common pathophysiological factor contributing to AERA. This review focuses on the possible mechanisms of AERA with the emphasis on erythropoiesis and iron metabolism and their relationship with inflammation. After surgery, blood loss seems to be the most obvious and determining factor influencing post-operative haemoglobin (Hb) concentration. However, surgical procedures without or with a limited amount of blood loss also induce a decrease in Hb level, which can not simply be explained by dilution (van Iperen et al, 1998). An average of 40 d is required to regenerate 500 ml of blood lost, whereas the loss of 3500 ml of blood is replaced in about 100 d (Coleman et al, 1953). Six weeks after total hip surgery, Hb levels have not returned to their preoperative values (Biesma et al, 1995). Furthermore, the average blood loss of 1 l, measured during total hip replacement should lead to a decrease in Hb concentration of 2·5 g/dl, yet decreases of 3 g/dl are found (Biesma et al, 1995; van Iperen et al, 1998). Although internal, un-measurable blood loss cannot totally be excluded other mechanisms are more likely to be involved. In critical illness, like sepsis, anaemia is also common and multifactorial. Blood loss may originate from surgery, insertion of catheters and diagnostic procedures like frequent phlebotomies. Gastrointestinal bleeding can be a source of blood loss and may be caused by gastric tube lesions, ulcers and stress-induced mucosa ulceration, which is seen especially in patients requiring mechanical ventilation (Smoller et al, 1989; Dale & Pruett, 1993; Cook et al, 1994; Tryba, 1994). Coagulation disorders may aggravate blood loss and include disorders in platelet function, decreased platelet count, (hyper)fibrinolysis, loss of plasma coagulation factors and endothelium-related coagulation disorders (Bone, 1992; Boldt et al, 1994; Boldt, 1996). Apart from blood loss several other mechanisms may underlie AERA either by affecting red cell survival or by decreasing their production. In cardiac surgery, using cardiopulmonary bypass or in valve replacement, haemolysis influences the erythrocyte's lifespan (Kusserow & Clapp, 1966; Roeser et al, 1968; Tabak et al, 1981). Apart from mechanical stress, some procedures induce oxidative stress, especially in case of ischaemia-reperfusion injury and tissue damage (McCord, 1985). Oxygen radicals are able to enhance lipid peroxidation, a process that can be considered a sign of ageing when it takes place in the red cell membrane, contributing to the removal of those cells from the circulation (Bartosz, 1990; Lutz, 1990; Gutteridge, 1995). Finally, surgery as well as sepsis has a well-known inflammatory effect with elevated plasma levels of cytokines like interleukin 1 (IL-1) and IL-6 (Cruickshank et al, 1990; Baigrie et al, 1992; van Deuren et al, 1998). This inflammatory state may have a negative influence on erythrocyte survival as reported in patients with rheumatoid arthritis (Dinant & de Maat, 1978; Salvarani et al, 1991). Despite these potential hazards, a recent study on erythrocyte survival in total hip surgery showed the same half life of 51Cr-labelled cells before and after the operation (van Iperen et al, 2000a). This indicates that in most circumstances red cells have a normal survival unless there is ongoing contact with foreign material. Another mechanism involved in AERA is a blunted erythropoietin response. Production of red cells in the bone marrow is stimulated by the hormone erythropoietin (EPO), which acts on erythroid progenitor cells, the erythroid burst forming unit (BFU-E) and the erythroid colony-forming unit (CFU-E) (Spivak, 1986; Krantz, 1991). This 34 kDa glycoprotein is produced mainly in the peritubular cells of the cortex and outer medulla of the kidney. The primary stimulus for EPO production is hypoxia, often caused by a reduced red cell mass. Haemoglobin concentration has to fall below 10·5 g/dl before EPO levels start to rise significantly (Kickler & Spivak, 1988). In the case of surgery, even lower Hb levels do not seem to trigger an impressive plasma EPO peak (Levine et al, 1991; Clemens & Spivak, 1994; van Iperen et al, 1998). Clemens & Spivak (1994) demonstrated in a variety of surgical patients a blunted EPO response in both women and men at the first post-operative day with normalization of the response from d 5 onwards. Inflammatory mediators, such as IL-1β, present after a surgical procedure, are thought to be responsible for inhibition of EPO gene transcription and, thus, the restricted EPO response to anaemia (Johnson et al, 1989, 1990; Faquin et al, 1992). Elevated levels of IL-1β and tumour necrosis factor-α (TNF-α) are also found in critically ill patients and might explain the severely blunted EPO response in these patients (Corwin et al, 1997; Rogiers et al, 1997; Gabriel et al, 1998; van Iperen et al, 2000b). This applies not only to septic patients but also to patients without sepsis. However, during the first days after sepsis or septic shock, increased EPO levels have been found together with high levels of IL-6 (Abel et al, 1996; Krafte-Jacobs & Bock, 1996). This high EPO level was irrespective of Hb concentration and resembled that of acute-phase proteins; the clinical significance of this short EPO peak remains unknown. Because EPO is primarily produced in the kidney one might expect the lowest EPO levels in critically ill patients with renal failure. However, all critically ill patients show extremely low values with only a relatively preserved EPO response in patients without sepsis and without renal failure (Rogiers et al, 1997). Acute renal failure (ARF), often seen in the ICU, leads to equally low EPO levels as in chronic renal failure (Nielsen & Thaysen, 1990). In those patients, EPO levels even remain depressed during the recovery phase of ARF. The low EPO activity after the acute event is also reflected by non-elevated serum transferrin receptor levels (Biesma et al, 1995; van Iperen et al, 1998, 2000b). Serum transferrin receptors have a close relationship to transferrin receptor expression on erythroblasts and the rate of erythropoiesis (Huebers et al, 1990; Ferguson et al, 1992). Proliferation of erythroid progenitor cells can be directly impaired in AERA as a consequence of the same inhibitory cytokines that suppress EPO production (Johnson et al, 1990; Means & Krantz, 1992). This reduced ability to proliferate in response to the EPO hormone can be overcome by sufficiently high doses of rHu EPO (Johnson et al, 1990). Indeed, the application of recombinant human (rHu) EPO in transfusion-dependent surgery shows that regimens with rHu EPO given subcutaneously during the perioperative period are effective in increasing Hb levels and reducing the need for homologous blood transfusions (Levine et al, 1989a,b; Kyo et al, 1992; Canadian orthopedic perioperative erythropoietin study group, 1993; Green et al, 1996; Faris et al, 1996). A dose of 600 IU/kg in patients with multiple organ failure induces a rise in reticulocyte count despite high levels of cytokines (Gabriel et al, 1998). Van Iperen et al (2000b) showed that five injections of 300 U/kg s.c., given to critically ill patients in an ICU, were able to establish a more than threefold increase in reticulocytes; the effect gradually disappeared within about 10 d after the last injection. Apart from a rise of reticulocytes, an increase in serum transferrin receptors was also seen, indicating that the bone marrow is able to respond to exogenous rHu EPO when given with intravenous iron supplementation. However, in the short observation period of this study, no significant differences were observed in Hb level and transfusion need between EPO- and non-EPO-treated patients, probably because of the many confounding factors in the ICU situation. A randomized controlled trial reported a 50% reduction in red cell transfusions as a result of the administration of EPO to critically ill patients (Corwin et al, 1999). Iron is essential for the number of Hb molecules produced, and striking changes in iron metabolism are observed in AERA with the characteristics of functional iron deficiency: hypoferraemia in the presence of adequate iron stores. The drop in plasma iron, which occurs within 6–12 h after an incision, has been reported after various types of procedures such as abdominal, orthopaedic and cardiac surgery (Feldthusen et al, 1953; Fitzsimmons & Ballantyne, 1983; Paterson, 1986; Fraser et al, 1989; Biesma et al, 1995; van Iperen et al, 1998). Low serum iron is also found in critically ill patients and even in non-septic patients, indicating that a moderate, but not overwhelming, release of cytokines seems sufficient for this effect (van Iperen et al, 2000b). Recently, we described a significant decrease in serum iron even after minor surgery with hardly any blood loss such as correction of hallux valgus and arthroscopy of the knee (van Iperen et al, 1998). It is likely that the inflammatory effect of the procedure induces the changes in iron metabolism. The short peak of IL-1β in plasma, which can be noted within 2 h of surgery could be responsible for the early hypoferraemia (Baigrie et al, 1992; van Deuren et al, 1998). This cytokine is able to increase ferritin H-subunit production and not l-subunits, which are translationally controlled by iron-responsive protein, with enhanced intracellular entrapment of iron (Rogers et al, 1990). However, serum ferritin levels only rise 2–3 d after surgery, whereas iron concentration decreases within 6–12 h (Rubin et al, 1984; Zdravkovic, 1987; Biesma et al, 1995), indicating that inflammation-induced ferritins are not immediately reflected in plasma ferritin levels. IL-1β and TNF-α induce hypoferraemia by increasing iron uptake into monocytes and macrophages in vitro and in mouse models (Gordeuk et al, 1988; Alvarez-Hernandez et al, 1989; Tsuji et al, 1991). TNF-α also hampers iron release from the mononuclear phagocytic system (Moldawer et al, 1989; Tsuji et al, 1991). TNF-α is unlikely to play a main role because it is usually not detectable after uncomplicated surgery, but it is found in sepsis and complicated cases. IL-6 is demonstrable within the first 48 h after surgery and in critically ill patients depending on the duration of the insult. It increases ferritin synthesis either by enhanced total protein synthesis or by redistribution of intracellular iron through the IRE/IRP (iron responsive elements/iron responsive protein) system (Rogers, 1996). This additional ferritin may act as a trap for iron (Konijn & Hershko, 1977). IL-6 also induces direct uptake of iron in hepatocytes (Kobune et al, 1994). Not only hypoferraemia and high levels of ferritin but also low plasma levels of the iron transporter transferrin, one of the negative acute-phase proteins, and a low transferrin saturation are demonstrated after surgery as well as after trauma and in critically ill patients (Fraser et al, 1989; Biesma et al, 1995; Walsh et al, 1996; van Iperen et al, 2000b). Data on the behaviour of the number of serum transferrin receptors in acute medical situations are scarce. Biesma et al (1995) found non-elevated serum transferrin receptor values shortly after total hip surgery with a gradual rise after 4–6 weeks, indicating that erythropoiesis is becoming more active several weeks after operation. In critically ill patients, they are low to normal despite the anaemia (van Iperen et al, 2000b). Iron absorption from the gastrointestinal tract decreases in cases of inflammation (Weber et al, 1988). The observation that oral iron supplementation during the first weeks after an operation has no measurable effects on serum iron and Hb concentration is in accordance with this (Zauber et al, 1992; Biesma et al, 1995). Ineffectiveness of oral iron treatment and impaired increase in erythropoiesis might also be caused by defective iron incorporation in the erythron. However, a recent study using [59Fe]-ferric citrate bound to transferrin given intravenously post operatively in patients receiving total hip replacement showed a very rapid iron utilization (van Iperen et al, 2001). Therefore, stimulation of iron uptake and impairment of iron release from the macrophage system and hepatocytes seems the most plausible mechanism. The question remains whether other inflammatory cytokines or proteins influence iron uptake and release by the mononuclear phagocytic system and hepatocytes. Although patients with acute coronary syndromes are not usually anaemic, patients with acute myocardial infarction show the same changes in iron metabolism as observed in critically ill patients (Nitter-Hauge, 1971; Birgegård et al, 1979; Fitzsimmons & Kaplan, 1980; Griffiths et al, 1985). Because patients admitted with unstable angina do not show such changes, it is unlikely that pain and stress are important mediators in developing functional iron deficiency (van Iperen et al, 2000c). The two coronary syndromes differ considerably in the extent of tissue damage and inflammatory response, making these factors more likely causes for the changes in iron metabolism (Tsuji et al, 1991). Acute event-related anaemia shows a high degree of analogy with the anaemia of chronic disease. Several mechanisms may be involved, as summarized in Fig 1, with a central role for inflammation. This induces functional iron deficiency, and blunts the responsiveness and action of erythropoietin. Pain and psychological stress are not sufficient to induce these changes while the bone marrow remains capable to incorporate iron and to react to erythropoietin. There are strong indications that the mononuclear phagocytic system plays a key role with enhanced iron entrapment and a temporary inhibition of iron release. Duration of anaemia is related to continuation of inflammation and suggests AERA to be an acute variant of ACD. A better understanding of the transient character of the mechanisms operative in AERA will contribute to a better and more appropriate treatment of patients in those circumstances. Schematic representation of the pathophysiological mechanisms underlying acute event-related anaemia (AERA). Acute events are, for example, surgical procedures, tissue damage, critical illness and sepsis. *IL-6 induces functional iron deficiency but does not impair EPO production nor erythropoiesis.