- Email: [email protected]
Review of issues relating to iron and infection
Review of Issues Relating to Iron and Infection Steven Fishbane, MD ● Use of erythropoietin (EPO) therapy and iron supplementation has improved the management of anemia in patients with end-stage renal disease (ESRD). As more patients receive supplemental iron, however, concerns are being raised about a potential link between iron and infection. There is biologic plausibility for this link, since iron is a growth factor for bacteria and certain host defense mechanisms are iron-sensitive. Animal models show that injection of iron leads to increased susceptibility to bacterial infection. In some studies, patients with high serum ferritin levels have reduced neutrophil function. However, these studies did not determine whether serum ferritin levels were elevated because of increased iron stores or because of infection. If infection is present, it might cause both the elevated serum ferritin levels and the neutrophil dysfunction. Several clinical studies have found an association between high serum ferritin levels and increased infectious risk. In studies that control for important covariates such as use of catheters and previous infections, the infectious risk associated with iron administration or elevated serum ferritin levels is reduced or eliminated. Collectively, these studies suggest that our current understanding of the relationship between iron and infection is incomplete and further studies are needed. There is no reason to alter current iron treatment strategies based on this literature. 娀 1999 by the National Kidney Foundation, Inc. INDEX WORDS: Kidney failure, chronic; hemodialysis; infection; iron deficiency; erythropoietin.
A
NEMIA MANAGEMENT in end-stage renal disease (ESRD) has come a long way in recent years. A decade ago, recombinant human erythropoietin (EPO) was introduced into clinical use. Before that, many patients had iron overload caused by repeated blood transfusions. By 1993, before effective iron management strategies were in place, the majority of hemodialysis patients receiving EPO therapy were in need of additional iron supplementation.1 More than 50% of patients were iron-deficient, with transferrin saturations (TSATs) below 20%.2 We have learned that EPO must be used in conjunction with supplemental iron therapy to optimize anemia management.3 As a result, we practice better iron management today, and fewer ESRD patients are iron-deficient. This is a timely point for examining the relationship between iron and infection, since we now have more extensive experience with iron supplementation. We must begin by making clear that the accumulated evidence for the efficacy of intravenous (IV) iron is clear and incontrovertible. The literature addressing safety concerns with IV iron, however, is much thinner. This article is an attempt to synthesize some of the safety literature on the link between iron and infection and address what it means to us as clinicians. A PLAUSIBLE BIOLOGIC ROLE
Before examining the body of literature linking iron to clinical infection, one should determine if there is at least biologic plausibility for
an association. Indeed, there is firm ground here— iron affects both bacterial virulence and host defense mechanisms. Iron is an important growth factor for bacteria.4,5 Acquisition of iron by bacteria is a prerequisite for clinically important infections. Several mechanisms allow bacteria to acquire iron (Table 1). Some bacteria acquire iron directly from transferrin through a membrane-bound transferrin receptor. More commonly, bacteria secrete siderophores, iron chelators that compete with host proteins for iron binding.4 It is interesting to note that deferoxamine, a drug that is given to relieve iron overload, is actually a siderophore.5 Thus, it carries iron in a loose way that allows certain bacteria to acquire iron from it. Indeed, deferoxamine is a known growth factor and virulence-enhancing factor for Yersinia enterocolitica,5 which is an occasional pathogen in patients with hemochromatosis. Although Y enterocolitica does not produce siderophores, it does have a cell-bound transport system for iron.6 On the host defense end, several host mechanisms may be affected by iron. White blood cells (WBCs) need iron to perform phagocytosis, to From the Division of Nephrology, Winthrop-University Hospital, Mineola, NY. Address reprint requests to Steven Fishbane, MD, Associate Director, Division of Nephrology, Winthrop-University Hospital, 222 Station Plaza North, Suite 510, Mineola, NY. E-mail: [email protected]
娀 1999 by the National Kidney Foundation, Inc. 0272-6386/99/3404-2008$3.00/0
American Journal of Kidney Diseases, Vol 34, No 4, Suppl 2 (October), 1999: pp S47-S52
S47
S48
STEVEN FISHBANE
Table 1. Main Mechanisms of Bacterial Acquisition of Iron From Transferrin
Mechanism
Description
Secretion of siderophores
Bacteria secrete siderophores (phenolate/hydroxamate iron chelators) that complete with transferrin for iron and bring it back to the cell where it is brought in by receptor-mediated endocytosis Bacteria acquire iron from transferrin through a bacterial-bound transferrin receptor
Nonsiderophore mechanism
Examples of Organisms Using the Method
Escherichia coli; Klebsiella, Salmonella, and Pseudomonas
Staphylococcus aureus,16 Haemophilus influenzae
NOTE. Data reviewed in Williams and Griffiths, 1992.4
generate reactive oxygen species, and to kill bacteria,7,8 but too much or too little iron can impair WBC function. While the importance of antibody and complement for bacterial killing is well known, what might be surprising is that unsaturated iron-binding proteins are also important for this process. In the test tube, potential pathogens grow freely if binding proteins are oversaturated.9 In addition, several studies have shown that a higher TSAT is associated with less efficient bacterial killing.10-13 STUDIES OF IRON AND BACTERIAL VIRULENCE
Several studies have shown a link between iron and infection. Bullen was one of the pioneers in this area. In a study conducted in the late 1960s, Bullen and Rogers14 injected mice with an inoculum (between 1.5 ⫻ 105 and 2.9 ⫻ 106 bacteria) of Pasteurella septica. When lysed red blood cells (RBCs) or purified hemoglobin were added as a source of iron, the passive immunity to P septica was reduced and the death rate increased by 9- to 10-fold. These investigators suggested that the effect occurred because hemoglobin provided extra iron to support bacterial replication. Several other animal studies were
conducted that showed similar results. However, how these animal studies, which involved large amounts of bacteria and iron, correlate with what we see clinically is not clear. Mofenson et al15 reported a case in which iron overdose appeared to cause acute Y enterocolitica septicemia. Fifteen-month-old twins were chronic carriers of Y enterocolitica. The male twin accidentally ingested 30 ferrous sulfate tablets and was treated with deferoxamine. While his sister suffered only from diarrhea, the boy developed septicemia within 48 hours. The key factor leading to these very different clinical outcomes appears to be the boy’s ingestion of iron and treatment with deferoxamine. An important study was reported in 1998 by Modun et al, who investigated the means by which staphylococci acquire iron during infection.16 Staphylococci express a 42-kd protein on the cell wall that is a transferrin-binding protein. The investigators found that 30% of healthy participants and 70% of healthy continuous ambulatory peritoneal dialysis patients had antibodies against this protein. In 2 patients who experienced staphylococcal infections, antibodies to the transferrin-binding protein appeared concurrently with the onset of peritonitis. In vitro, the purified antibody blocked the binding of transferrin to whole staphylococcal cells; in other words, it was an active neutralizing agent. This study is of interest for the renal community because staphylococcus is an important pathogen in dialysis patients, while many of the other pathogens discussed in iron studies are not as relevant. However, the clinical relevance of the expression of the transferrin receptor or neutralizing antibodies is unknown. EFFECT OF IRON STATUS ON WHITE BLOOD CELL FUNCTION
The literature on the effects of iron on WBC function is also thin; this article will review 4 of 9 studies identified. Many of these studies are retrospective and cross-sectional. Cantinieaux et al17 examined neutrophil defense against Yersinia. The study involved neutrophils sampled from 3 groups of patients: 9 iron-overloaded hemodialysis patients (history of transfusion with 25 units of packed RBCs and serum ferritin levels ⬎ 800 ng/mL), 9 patients with no evidence of iron overload, and 10 healthy control patients. In
REVIEW OF IRON AND INFECTION
vitro, phagocytosis was significantly decreased in neutrophils from the iron-overloaded patients; the addition of normal serum for opsonization did not improve the function. In the presence of serum, actual killing of Yersinia was mildly decreased in neutrophils from the iron-overloaded patients compared with that in the other groups. Examination of these results reveals some of the more important issues confounding studies investigating the relationship between iron and infection. For example, while Cantinieaux’s hemodialysis patients may in fact have been ironoverloaded, they may have had leukocyte dysfunction because of the effect of transfusion of blood itself.18 In addition, as is discussed more extensively in Dr Cavill’s article in this supplement, inflammation and infection can cause serum ferritin levels to be elevated independently of iron status.3,19 Therefore, one cannot be sure whether a high serum ferritin level is a marker of infection or a marker of iron overload. If the patient is infected, the infection, rather than iron overload, might be the cause of the WBC dysfunction. Flament et al compared superoxide anion production in 21 hemodialysis patients whose serum ferritin levels were greater than 1,000 ng/mL with that in 19 patients whose serum ferritin levels were less than 1,000 ng/mL. Superoxide production was 33% lower in opsonized zymosan- and dialysis-stimulated patients with high serum ferritin levels (median, 3,770 ng/mL) than in those with lower values (median, 73 ng/mL).20 Here again, the cause of the high serum ferritin is not clear. In addition, the clinical significance of a 33% reduction in superoxide production is not known. Patruta et al conducted a study of neutrophil function that had an interesting categorization schedule. The study included hemodialysis patients with high serum ferritin levels, as well as hematology patients with hemosiderosis. Patient groups were subdivided based on serum ferritin levels and TSAT values (serum ferritin level ⬍ 60 ng/mL, between 100 and 350 ng/mL, and ⬎ 650 ng/mL but TSAT ⬍ 20%).21 An additional group of healthy patients was included as a control. Neutrophils from patients with serum ferritin levels between 100 ng/mL and 350 ng/mL exhibited mild inhibition of phagocytosis but significant inhibition of intracellular killing of
S49
bacteria compared with normal controls. Neutrophil function was also somewhat decreased in the interesting group of hemodialysis patients with serum ferritin levels of greater than 650 ng/mL and TSATs less than 20%. The case for excess iron causing the functional neutrophil deficiency is particularly strong in this study because the effect was seen in both ESRD patients and in hemochromatosis patients. But what about the effect of iron deficiency on neutrophil function? In a study by Cui et al, neutrophils from 20 iron-deficient patients had decreased stimulated reactive oxygen species generation compared with controls, although the basal release was not significantly different.22 This study illustrates that iron deficiency might lead to impaired host defense mechanisms, which can cause an increased infection rate. IRON AND INFECTIOUS RISK
The literature on the relationship between iron and clinical infectious risk is also thin. Seifert et al studied infectious risk in a group of hemodialysis patients who were treated with deferoxamine, some for aluminum overload, others for iron overload.23 Deferoxamine treatment for aluminum overload did not increase the risk for infection. However, deferoxamine treatment for iron overload led to a 3-fold increase in infections. It is likely that the deferoxamine made iron available for bacterial utilization. In a cross-sectional, retrospective part of the study, Seifert et al studied infectious risk in 184 hemodialysis patients who were never treated with deferoxamine.23 In patients with serum ferritin levels of 10 to 330 ng/mL, there were 0.18 infections per year. In patients with serum ferritin values between 331 and 1,000 ng/mL, there were 0.34 infections per year, and in those with serum ferritin levels between 1,001 and 2,000 ng/mL, there were 0.58 infections per year. These were significant differences, and the boundary of risk for serum ferritin levels appeared to be at some level greater than 330 ng/mL. The investigators concluded that treatment with deferoxamine itself does not cause septicemia or bacterial infection, but iron overload may predispose patients treated with the deferoxamine to bacterial infection. Tielemans et al performed a retrospective study in 61 hemodialysis patients with varying serum ferritin levels. The purpose was to relate iron
S50
status to risk for infection. In this study, the infection risk began to increase at serum ferritin levels greater than 500 ng/mL. Twenty-one patients who were treated with deferoxamine for iron overload experienced a decreased incidence of infection from 1 per 19 patient-months to 1 per 112 patient-months (P ⬍ 0.005, chi-squared test).24 In a study conducted in 158 hemodialysis patients with transfusional overload, Boelaert et al studied episodes of bacteremia as they relate to serum ferritin levels in hemodialysis patients. They found that the relative risk of infection tripled when the serum ferritin level was greater than 1,000 ng/mL.25 In interpreting these studies, one should keep several points in mind. These studies were performed in the pre-EPO era, at a time when anemia was usually managed through blood transfusions, which can cause both iron overload and immunosuppression. Iron overload was more common then; in EPO-treated patients, iron is shifted into the RBC compartment, so storage iron overload is less abundant. Also, the need for transfusions itself may have been a problem: a recent study by He´bert et al showed that transfusion is an independent risk factor for mortality in some critically ill patients.26 Several studies conducted by Hoen et al are reviewed more extensively in his article in this supplement. The first of these studies was the study reported by Kessler et al, which identified risk factors for bacterial infection in hemodialysis patients.27 In this study, significantly more patients with bacterial infection had serum ferritin levels higher than 1,000 ng/mL than did nonbacteremic patients. A second study was conducted to further probe risk factors for bacterial infection in stable hemodialysis patients. Upon multivariate analysis, 3 parameters were found to be significant independent risk factors for bacterial infection: (1) history of bacterial infection, (2) type of vascular access (catheter v arteriovenous fistula), and (3) serum ferritin levels greater than 500 ng/mL.28 The most recent study from this group is the most valuable contribution to the literature on iron and infection. It is a multicenter study of large size, of prospective design, and rich in clinical data. From 19 centers, 988 adult hemodialysis patients were monitored for 6 months.29 Multivariate analysis identified 4 risk factors for
STEVEN FISHBANE
bacteremia: (1) type of vascular access (catheter v fistula), (2) history of bacteremia (ⱖ 2 v no previous episode), (3) immunosuppressive therapy (current v none), and (4) anemia based on corpuscular hemoglobin (per 1-g/dL increment).29 The serum ferritin level and iron treatment did not stand up as significant risk factors in the multivariate analysis of this study. The powerful design of this study allows us to conclude that it is unlikely that either iron overload or iron treatment per se causes any significant increase in infectious risk in hemodialysis patients. While studies from Dr Hoen’s group were based on clinical data, Collins et al used claimsbased data for their analysis of the relationship between iron and infection. This study used the number of vials of iron given rather than serum ferritin levels or iron dose. In their 1997 abstract, data analyzed from 33,120 patients demonstrated a 35% increased risk of infectious death with frequent, low-dose IV iron therapy compared with other dosing methods.30 They also found a trend toward reduced mortality with low-frequency, high-dose iron compared with no iron (personal communication, July 1997, A. Collins). Several important issues arise from this study. Is there a causal relationship between IV iron and infectious risk? This is unlikely given the better-designed Hoen study showing no relationship between iron and infection. If there is a link between iron and infection, is it due to frequent low-dose therapy? Is IV iron a surrogate for another factor that may be a causal factor for infection? In a follow-up study that addressed some of these issues, Collins et al increased the number of observations to 309,219 and examined some of the clinical factors that the previous study did not consider as covariates. They excluded patients with dialysis catheters or previous admissions for sepsis. The study included multivariate analysis of 12 categories of factors affecting outcome over a 6-month entry period, and the IV iron claims were based on frequency and number of vials administered. They found that the risk for mortality was slightly increased in patients who received more than 17 vials of IV iron administered over 3 to 4 or 5 to 6 months. The infectious risk for all other frequency and dosing patterns was not different from that in patients
REVIEW OF IRON AND INFECTION
receiving no IV iron.31 This finding must be interpreted in context. If 17 vials of iron were administered over 4 months, it could amount to as much as 5.1 g iron per year; most patients do not require or receive that much iron.3,32 The Collins 1998 study controlled for 2 important covariates (catheters and sepsis admissions) that were not controlled for in the 1997 study. With this adjustment, relative risk for mortality associated with iron therapy decreased from a maximum of 1.5933 to a maximum of 1.2. A similar trend was seen in the series of studies reported by Dr Hoen’s group. As more variables were controlled for, the contribution of iron to infectious risk was eliminated. One wonders whether the finding and inclusion of additional covariates could lower the infectious risk associated with IV iron to 1 or even lower, possibly indicating a protective effect of IV iron on infectious risk. SUMMARY/CONCLUSIONS
A review of the literature suggests that a link between iron and infection is biologically plausible. While studies confirm that iron is important for bacterial virulence/growth and that lack of iron and excess iron may impair WBC function, studies that attempt to link iron and clinical infectious risk are not as clear-cut. Part of the difficulty lies in the relatively faulty nature of using the serum ferritin level as a marker of iron overload. A clear boundary of increased infectious risk for serum ferritin levels has not been established. Given all of the above considerations, the attributable infectious risk for iron, if any, is probably small. Better studies that factor in potentially confounding variables are needed to more carefully dissect out the contribution of iron to infectious risk in ESRD patients. The best study to date, by Hoen et al (1998),29 showed no role of iron in causing infections. As clinicians, we must weigh these risks against the more established benefits of improved anemia management in ESRD patients, which include decreased mortality34-36 and improved quality of life.37-39 Further study will provide more complete information with which to approach this clinical issue.
S51
REFERENCES 1. Eschbach JW, Adamson JW: Iron overload in renal failure patients: Changes since the introduction of erythropoietin therapy. Kidney Int 69:S35-S43, 1999 (suppl) 2. Young EW, Bloembergen WE, Woods JD, Emmert G, Port FK, Wolfe RA, Jones CA, Held PJ: Iron use among erythropoietin-treated U.S. hemodialysis patients. J Am Soc Nephrol 7:1469A, 1996 (abstr) 3. National Kidney Foundation: NKF-DOQI clinical practice guidelines for the treatment of anemia of chronic renal failure. Am J Kidney Dis 30:S192-S237, 1997 (suppl) 4. Williams P, Griffiths E: Bacterial transferrin receptors— Structure, function and contribution to virulence. Med Microbiol Immunol 181:301-322, 1992 5. Gallant T, Freedman MH, Vellend H, Francombe WH: Yersinia sepsis in patients with iron overload treated with deferoxamine. N Engl J Med 314:1643, 1986 (letter) 6. Perry RD, Brubaker RR: Accumulation of iron by yersiniae. J Bacteriol 137:1290-1298, 1979 7. Walter T, Arredondo S, Are´valo M, Stekel A: Effect of iron therapy on phagocytosis and bactericidal activity in neutrophils of iron-deficient infants. Am J Clin Nutr 44:877882, 1986 8. Murakawa H, Bland CE, Willis WT, Dallman PR: Iron deficiency and neutrophil function: Different rates of correction of the depressions in oxidative burst and myeloperoxidase activity after iron treatment. Blood 69:1464-1468, 1987 9. Goldoni P, Visca P, Pastoris MC, Valenti P, Orsi N: Growth of Legionella spp. under conditions of iron restriction. J Med Microbiol 34:113-118, 1991 10. Struelens MJ, Mondal G, Roberts M, Williams PH: Role of bacterial and host factors in the pathogenesis of Shigella septicemia. Eur J Clin Microbiol Infect Dis 9:337344, 1990 11. Towns ML, Bennett B, Check IJ, Hunter RL: An improved method for determining the microbial inhibitory activity of serum and its application to the study of patients with leukemia. Am J Clin Pathol 92:192-198, 1989 12. Ward CG, Hammond JS, Bullen JJ: Effect of iron compounds on antibacterial function of human polymorphs and plasma. Infect Immun 51:723-730, 1986 13. Hunter RL, Bennett B, Towns M, Vogler WR: Transferrin in disease II: Defects in the regulation of transferrin saturation with iron contribute to susceptibility to infection. Am J Clin Pathol 81:748-753, 1984 14. Bullen JJ, Rogers HJ: Effect of haemoglobin on experimental infections with Pasteurella septica and Escherichia coli. Nature 217:86, 1968 15. Mofenson HC, Caraccio TR, Sharieff N: Iron sepsis: Yersinia enterocolitica septicemia possibly caused by an overdose of iron. N Engl J Med 316:1092-1093, 1987 16. Modun BJ, Cockayne A, Finch R, Williams P: The Staphylococcus aureus and Staphylococcus epidermidis transferrin-binding proteins are expressed in vivo during infection. Microbiology 144:1005-1012, 1998 17. Cantinieaux J, Boelaert J, Hariga C, Fondu P: Impaired neutrophil defense against Yersinia enterocolitica in patients with iron overload who are undergoing dialysis. J Lab Clin Med 111:524-528, 1988 18. Bordin JO, Heddle NM, Blajchman MA: Biologic
S52
effects of leukocytes present in transfused cellular blood products. Blood 84:1703-1721, 1994 19. Harrison PM, Arosio P: The ferritins: Molecular properties, iron storage function and cellular regulation. Biochim Biophys Acta 1275:161-203, 1996 20. Flament J, Goldman M, Waterlot Y, Dupont E, Wybran J, Vanherweghem J-L: Impairment of phagocyte oxidative metabolism in hemodialyzed patients with iron overload. Clin Nephrol 25:227-230, 1986 21. Patruta SI, Edlinger R, Sunder-Plassman G, Ho¨rl WH: Neutrophil impairment associated with iron therapy in hemodialysis patients with functional iron deficiency. J Am Soc Nephrol 9:655-663, 1998 22. Cui Y, Zhang ZN, Li RS: Influence of latent iron deficiency on generation of oxygen species in polymorphonuclear leucocytes. Chin Med J 103:642-646, 1990 23. Seifert A, Von Herrath D, Schaefer K: Iron overload, but not treatment with desferrioxamine favours the development of septicemia in patients on maintenance hemodialysis. Q J Med 65:1015-1024, 1987 24. Tielemans CL, Lenclud CM, Wens R, Collart FE, Dratwa M: Critical role of iron overload in the increased susceptibility of haemodialysis patients to bacterial infections. Beneficial effects of desferrioxamine. Nephrol Dial Transplant 4:883-887, 1989 25. Boelaert JR, Daneels RF, Schurgers ML, Matthys EG, Gordts BZ, van Landuyt HW: Iron overload in haemodialysis patients increases the risk of bacteraemia: A prospective study. Nephrol Dial Transplant 5:130-134, 1990 26. He´bert PC, Wells G, Blajchman MA, Marshall J, Martin C, Pagliarello G, Tweeddale M, Schweitzer I, Yetisir E: A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med 340:409-417, 1999 27. Kessler M, Hoen B, Mayeux D, Hestin D, Fontenaille C: Bacteremia in patients on chronic hemodialysis. Nephron 64:95-100, 1993 28. Hoen B, Kessler M., Hestin D, Mayeux D: Risk factors for bacterial infections in chronic haemodialysis
STEVEN FISHBANE
adult patients: A multicentre prospective survey. Nephrol Dial Transplant 10:377-381, 1995 29. Hoen B, Paul-Dauphin A, Hestin D, Kessler M: EPIBACDIAL: A multicenter prospective study of risk factors for bacteremia in chronic hemodialysis patients. J Am Soc Nephrol 9:869-876, 1998 30. Collins A, Ebben J, Ma J: Frequent IV iron dosing is associated with higher infectious deaths. J Am Soc Nephrol 8:190A, 1997 (abstr) 31. Collins A, Ebben J, Ma J, Xia H: I.V. iron dosing patterns and mortality. J Am Soc Nephrol 9:205A, 1998 (abstr) 32. Fishbane S, Maesaka JK: Iron management in endstage renal disease. Am J Kidney Dis 29:319-333, 1997 33. Collins AJ: IV iron and mortality issues in ESRD patients. Slide 28, on file, Amgen, Thousand Oaks, CA, 1997 34. Collins A, Ma J, Ebben J: Patients survival is associated with hematocrit (HCT) level. J Am Soc Nephrol 8:190A, 1997 (abstr) 35. Madore F, Bridges K, Brugnara C, Lew NL, Lowrie EG, Lazarus JM, Owen WF: A population study of the interplay between iron, nutrition and inflammation in erythropoiesis in hemodialysis patients. J Am Soc Nephrol 7:1456A, 1996 (abstr) 36. Collins A, Xia H, Ebben J, Ma J: Change in hematocrit and risk of mortality. J Am Soc Nephrol 9:204A, 1998 (abstr) 37. Nissenson AR: Optimal hematocrit in patients on dialysis therapy. Am J Kidney Dis 32:S142-S146, 1998 (suppl) 38. Canadian Erythropoietin Study Group: Association between recombinant human erythropoietin and quality of life and exercise capacity of patients receiving haemodialysis. BMJ 300:573-578, 1990 39. Eschbach JW, Glenny R, Robertson T, Gurhrie M, Rader B, Evans R, Chandler W, Davidson R, Easterling T, Denney J, Schneider G: Normalizing the hematocrit (HCT) in hemodialysis patients (HDP) with EPO improves quality of life (Q/L) and is safe. J Am Soc Nephrol 4:425A, 1993 (abstr)
A
NEMIA MANAGEMENT in end-stage renal disease (ESRD) has come a long way in recent years. A decade ago, recombinant human erythropoietin (EPO) was introduced into clinical use. Before that, many patients had iron overload caused by repeated blood transfusions. By 1993, before effective iron management strategies were in place, the majority of hemodialysis patients receiving EPO therapy were in need of additional iron supplementation.1 More than 50% of patients were iron-deficient, with transferrin saturations (TSATs) below 20%.2 We have learned that EPO must be used in conjunction with supplemental iron therapy to optimize anemia management.3 As a result, we practice better iron management today, and fewer ESRD patients are iron-deficient. This is a timely point for examining the relationship between iron and infection, since we now have more extensive experience with iron supplementation. We must begin by making clear that the accumulated evidence for the efficacy of intravenous (IV) iron is clear and incontrovertible. The literature addressing safety concerns with IV iron, however, is much thinner. This article is an attempt to synthesize some of the safety literature on the link between iron and infection and address what it means to us as clinicians. A PLAUSIBLE BIOLOGIC ROLE
Before examining the body of literature linking iron to clinical infection, one should determine if there is at least biologic plausibility for
an association. Indeed, there is firm ground here— iron affects both bacterial virulence and host defense mechanisms. Iron is an important growth factor for bacteria.4,5 Acquisition of iron by bacteria is a prerequisite for clinically important infections. Several mechanisms allow bacteria to acquire iron (Table 1). Some bacteria acquire iron directly from transferrin through a membrane-bound transferrin receptor. More commonly, bacteria secrete siderophores, iron chelators that compete with host proteins for iron binding.4 It is interesting to note that deferoxamine, a drug that is given to relieve iron overload, is actually a siderophore.5 Thus, it carries iron in a loose way that allows certain bacteria to acquire iron from it. Indeed, deferoxamine is a known growth factor and virulence-enhancing factor for Yersinia enterocolitica,5 which is an occasional pathogen in patients with hemochromatosis. Although Y enterocolitica does not produce siderophores, it does have a cell-bound transport system for iron.6 On the host defense end, several host mechanisms may be affected by iron. White blood cells (WBCs) need iron to perform phagocytosis, to From the Division of Nephrology, Winthrop-University Hospital, Mineola, NY. Address reprint requests to Steven Fishbane, MD, Associate Director, Division of Nephrology, Winthrop-University Hospital, 222 Station Plaza North, Suite 510, Mineola, NY. E-mail: [email protected]
娀 1999 by the National Kidney Foundation, Inc. 0272-6386/99/3404-2008$3.00/0
American Journal of Kidney Diseases, Vol 34, No 4, Suppl 2 (October), 1999: pp S47-S52
S47
S48
STEVEN FISHBANE
Table 1. Main Mechanisms of Bacterial Acquisition of Iron From Transferrin
Mechanism
Description
Secretion of siderophores
Bacteria secrete siderophores (phenolate/hydroxamate iron chelators) that complete with transferrin for iron and bring it back to the cell where it is brought in by receptor-mediated endocytosis Bacteria acquire iron from transferrin through a bacterial-bound transferrin receptor
Nonsiderophore mechanism
Examples of Organisms Using the Method
Escherichia coli; Klebsiella, Salmonella, and Pseudomonas
Staphylococcus aureus,16 Haemophilus influenzae
NOTE. Data reviewed in Williams and Griffiths, 1992.4
generate reactive oxygen species, and to kill bacteria,7,8 but too much or too little iron can impair WBC function. While the importance of antibody and complement for bacterial killing is well known, what might be surprising is that unsaturated iron-binding proteins are also important for this process. In the test tube, potential pathogens grow freely if binding proteins are oversaturated.9 In addition, several studies have shown that a higher TSAT is associated with less efficient bacterial killing.10-13 STUDIES OF IRON AND BACTERIAL VIRULENCE
Several studies have shown a link between iron and infection. Bullen was one of the pioneers in this area. In a study conducted in the late 1960s, Bullen and Rogers14 injected mice with an inoculum (between 1.5 ⫻ 105 and 2.9 ⫻ 106 bacteria) of Pasteurella septica. When lysed red blood cells (RBCs) or purified hemoglobin were added as a source of iron, the passive immunity to P septica was reduced and the death rate increased by 9- to 10-fold. These investigators suggested that the effect occurred because hemoglobin provided extra iron to support bacterial replication. Several other animal studies were
conducted that showed similar results. However, how these animal studies, which involved large amounts of bacteria and iron, correlate with what we see clinically is not clear. Mofenson et al15 reported a case in which iron overdose appeared to cause acute Y enterocolitica septicemia. Fifteen-month-old twins were chronic carriers of Y enterocolitica. The male twin accidentally ingested 30 ferrous sulfate tablets and was treated with deferoxamine. While his sister suffered only from diarrhea, the boy developed septicemia within 48 hours. The key factor leading to these very different clinical outcomes appears to be the boy’s ingestion of iron and treatment with deferoxamine. An important study was reported in 1998 by Modun et al, who investigated the means by which staphylococci acquire iron during infection.16 Staphylococci express a 42-kd protein on the cell wall that is a transferrin-binding protein. The investigators found that 30% of healthy participants and 70% of healthy continuous ambulatory peritoneal dialysis patients had antibodies against this protein. In 2 patients who experienced staphylococcal infections, antibodies to the transferrin-binding protein appeared concurrently with the onset of peritonitis. In vitro, the purified antibody blocked the binding of transferrin to whole staphylococcal cells; in other words, it was an active neutralizing agent. This study is of interest for the renal community because staphylococcus is an important pathogen in dialysis patients, while many of the other pathogens discussed in iron studies are not as relevant. However, the clinical relevance of the expression of the transferrin receptor or neutralizing antibodies is unknown. EFFECT OF IRON STATUS ON WHITE BLOOD CELL FUNCTION
The literature on the effects of iron on WBC function is also thin; this article will review 4 of 9 studies identified. Many of these studies are retrospective and cross-sectional. Cantinieaux et al17 examined neutrophil defense against Yersinia. The study involved neutrophils sampled from 3 groups of patients: 9 iron-overloaded hemodialysis patients (history of transfusion with 25 units of packed RBCs and serum ferritin levels ⬎ 800 ng/mL), 9 patients with no evidence of iron overload, and 10 healthy control patients. In
REVIEW OF IRON AND INFECTION
vitro, phagocytosis was significantly decreased in neutrophils from the iron-overloaded patients; the addition of normal serum for opsonization did not improve the function. In the presence of serum, actual killing of Yersinia was mildly decreased in neutrophils from the iron-overloaded patients compared with that in the other groups. Examination of these results reveals some of the more important issues confounding studies investigating the relationship between iron and infection. For example, while Cantinieaux’s hemodialysis patients may in fact have been ironoverloaded, they may have had leukocyte dysfunction because of the effect of transfusion of blood itself.18 In addition, as is discussed more extensively in Dr Cavill’s article in this supplement, inflammation and infection can cause serum ferritin levels to be elevated independently of iron status.3,19 Therefore, one cannot be sure whether a high serum ferritin level is a marker of infection or a marker of iron overload. If the patient is infected, the infection, rather than iron overload, might be the cause of the WBC dysfunction. Flament et al compared superoxide anion production in 21 hemodialysis patients whose serum ferritin levels were greater than 1,000 ng/mL with that in 19 patients whose serum ferritin levels were less than 1,000 ng/mL. Superoxide production was 33% lower in opsonized zymosan- and dialysis-stimulated patients with high serum ferritin levels (median, 3,770 ng/mL) than in those with lower values (median, 73 ng/mL).20 Here again, the cause of the high serum ferritin is not clear. In addition, the clinical significance of a 33% reduction in superoxide production is not known. Patruta et al conducted a study of neutrophil function that had an interesting categorization schedule. The study included hemodialysis patients with high serum ferritin levels, as well as hematology patients with hemosiderosis. Patient groups were subdivided based on serum ferritin levels and TSAT values (serum ferritin level ⬍ 60 ng/mL, between 100 and 350 ng/mL, and ⬎ 650 ng/mL but TSAT ⬍ 20%).21 An additional group of healthy patients was included as a control. Neutrophils from patients with serum ferritin levels between 100 ng/mL and 350 ng/mL exhibited mild inhibition of phagocytosis but significant inhibition of intracellular killing of
S49
bacteria compared with normal controls. Neutrophil function was also somewhat decreased in the interesting group of hemodialysis patients with serum ferritin levels of greater than 650 ng/mL and TSATs less than 20%. The case for excess iron causing the functional neutrophil deficiency is particularly strong in this study because the effect was seen in both ESRD patients and in hemochromatosis patients. But what about the effect of iron deficiency on neutrophil function? In a study by Cui et al, neutrophils from 20 iron-deficient patients had decreased stimulated reactive oxygen species generation compared with controls, although the basal release was not significantly different.22 This study illustrates that iron deficiency might lead to impaired host defense mechanisms, which can cause an increased infection rate. IRON AND INFECTIOUS RISK
The literature on the relationship between iron and clinical infectious risk is also thin. Seifert et al studied infectious risk in a group of hemodialysis patients who were treated with deferoxamine, some for aluminum overload, others for iron overload.23 Deferoxamine treatment for aluminum overload did not increase the risk for infection. However, deferoxamine treatment for iron overload led to a 3-fold increase in infections. It is likely that the deferoxamine made iron available for bacterial utilization. In a cross-sectional, retrospective part of the study, Seifert et al studied infectious risk in 184 hemodialysis patients who were never treated with deferoxamine.23 In patients with serum ferritin levels of 10 to 330 ng/mL, there were 0.18 infections per year. In patients with serum ferritin values between 331 and 1,000 ng/mL, there were 0.34 infections per year, and in those with serum ferritin levels between 1,001 and 2,000 ng/mL, there were 0.58 infections per year. These were significant differences, and the boundary of risk for serum ferritin levels appeared to be at some level greater than 330 ng/mL. The investigators concluded that treatment with deferoxamine itself does not cause septicemia or bacterial infection, but iron overload may predispose patients treated with the deferoxamine to bacterial infection. Tielemans et al performed a retrospective study in 61 hemodialysis patients with varying serum ferritin levels. The purpose was to relate iron
S50
status to risk for infection. In this study, the infection risk began to increase at serum ferritin levels greater than 500 ng/mL. Twenty-one patients who were treated with deferoxamine for iron overload experienced a decreased incidence of infection from 1 per 19 patient-months to 1 per 112 patient-months (P ⬍ 0.005, chi-squared test).24 In a study conducted in 158 hemodialysis patients with transfusional overload, Boelaert et al studied episodes of bacteremia as they relate to serum ferritin levels in hemodialysis patients. They found that the relative risk of infection tripled when the serum ferritin level was greater than 1,000 ng/mL.25 In interpreting these studies, one should keep several points in mind. These studies were performed in the pre-EPO era, at a time when anemia was usually managed through blood transfusions, which can cause both iron overload and immunosuppression. Iron overload was more common then; in EPO-treated patients, iron is shifted into the RBC compartment, so storage iron overload is less abundant. Also, the need for transfusions itself may have been a problem: a recent study by He´bert et al showed that transfusion is an independent risk factor for mortality in some critically ill patients.26 Several studies conducted by Hoen et al are reviewed more extensively in his article in this supplement. The first of these studies was the study reported by Kessler et al, which identified risk factors for bacterial infection in hemodialysis patients.27 In this study, significantly more patients with bacterial infection had serum ferritin levels higher than 1,000 ng/mL than did nonbacteremic patients. A second study was conducted to further probe risk factors for bacterial infection in stable hemodialysis patients. Upon multivariate analysis, 3 parameters were found to be significant independent risk factors for bacterial infection: (1) history of bacterial infection, (2) type of vascular access (catheter v arteriovenous fistula), and (3) serum ferritin levels greater than 500 ng/mL.28 The most recent study from this group is the most valuable contribution to the literature on iron and infection. It is a multicenter study of large size, of prospective design, and rich in clinical data. From 19 centers, 988 adult hemodialysis patients were monitored for 6 months.29 Multivariate analysis identified 4 risk factors for
STEVEN FISHBANE
bacteremia: (1) type of vascular access (catheter v fistula), (2) history of bacteremia (ⱖ 2 v no previous episode), (3) immunosuppressive therapy (current v none), and (4) anemia based on corpuscular hemoglobin (per 1-g/dL increment).29 The serum ferritin level and iron treatment did not stand up as significant risk factors in the multivariate analysis of this study. The powerful design of this study allows us to conclude that it is unlikely that either iron overload or iron treatment per se causes any significant increase in infectious risk in hemodialysis patients. While studies from Dr Hoen’s group were based on clinical data, Collins et al used claimsbased data for their analysis of the relationship between iron and infection. This study used the number of vials of iron given rather than serum ferritin levels or iron dose. In their 1997 abstract, data analyzed from 33,120 patients demonstrated a 35% increased risk of infectious death with frequent, low-dose IV iron therapy compared with other dosing methods.30 They also found a trend toward reduced mortality with low-frequency, high-dose iron compared with no iron (personal communication, July 1997, A. Collins). Several important issues arise from this study. Is there a causal relationship between IV iron and infectious risk? This is unlikely given the better-designed Hoen study showing no relationship between iron and infection. If there is a link between iron and infection, is it due to frequent low-dose therapy? Is IV iron a surrogate for another factor that may be a causal factor for infection? In a follow-up study that addressed some of these issues, Collins et al increased the number of observations to 309,219 and examined some of the clinical factors that the previous study did not consider as covariates. They excluded patients with dialysis catheters or previous admissions for sepsis. The study included multivariate analysis of 12 categories of factors affecting outcome over a 6-month entry period, and the IV iron claims were based on frequency and number of vials administered. They found that the risk for mortality was slightly increased in patients who received more than 17 vials of IV iron administered over 3 to 4 or 5 to 6 months. The infectious risk for all other frequency and dosing patterns was not different from that in patients
REVIEW OF IRON AND INFECTION
receiving no IV iron.31 This finding must be interpreted in context. If 17 vials of iron were administered over 4 months, it could amount to as much as 5.1 g iron per year; most patients do not require or receive that much iron.3,32 The Collins 1998 study controlled for 2 important covariates (catheters and sepsis admissions) that were not controlled for in the 1997 study. With this adjustment, relative risk for mortality associated with iron therapy decreased from a maximum of 1.5933 to a maximum of 1.2. A similar trend was seen in the series of studies reported by Dr Hoen’s group. As more variables were controlled for, the contribution of iron to infectious risk was eliminated. One wonders whether the finding and inclusion of additional covariates could lower the infectious risk associated with IV iron to 1 or even lower, possibly indicating a protective effect of IV iron on infectious risk. SUMMARY/CONCLUSIONS
A review of the literature suggests that a link between iron and infection is biologically plausible. While studies confirm that iron is important for bacterial virulence/growth and that lack of iron and excess iron may impair WBC function, studies that attempt to link iron and clinical infectious risk are not as clear-cut. Part of the difficulty lies in the relatively faulty nature of using the serum ferritin level as a marker of iron overload. A clear boundary of increased infectious risk for serum ferritin levels has not been established. Given all of the above considerations, the attributable infectious risk for iron, if any, is probably small. Better studies that factor in potentially confounding variables are needed to more carefully dissect out the contribution of iron to infectious risk in ESRD patients. The best study to date, by Hoen et al (1998),29 showed no role of iron in causing infections. As clinicians, we must weigh these risks against the more established benefits of improved anemia management in ESRD patients, which include decreased mortality34-36 and improved quality of life.37-39 Further study will provide more complete information with which to approach this clinical issue.
S51
REFERENCES 1. Eschbach JW, Adamson JW: Iron overload in renal failure patients: Changes since the introduction of erythropoietin therapy. Kidney Int 69:S35-S43, 1999 (suppl) 2. Young EW, Bloembergen WE, Woods JD, Emmert G, Port FK, Wolfe RA, Jones CA, Held PJ: Iron use among erythropoietin-treated U.S. hemodialysis patients. J Am Soc Nephrol 7:1469A, 1996 (abstr) 3. National Kidney Foundation: NKF-DOQI clinical practice guidelines for the treatment of anemia of chronic renal failure. Am J Kidney Dis 30:S192-S237, 1997 (suppl) 4. Williams P, Griffiths E: Bacterial transferrin receptors— Structure, function and contribution to virulence. Med Microbiol Immunol 181:301-322, 1992 5. Gallant T, Freedman MH, Vellend H, Francombe WH: Yersinia sepsis in patients with iron overload treated with deferoxamine. N Engl J Med 314:1643, 1986 (letter) 6. Perry RD, Brubaker RR: Accumulation of iron by yersiniae. J Bacteriol 137:1290-1298, 1979 7. Walter T, Arredondo S, Are´valo M, Stekel A: Effect of iron therapy on phagocytosis and bactericidal activity in neutrophils of iron-deficient infants. Am J Clin Nutr 44:877882, 1986 8. Murakawa H, Bland CE, Willis WT, Dallman PR: Iron deficiency and neutrophil function: Different rates of correction of the depressions in oxidative burst and myeloperoxidase activity after iron treatment. Blood 69:1464-1468, 1987 9. Goldoni P, Visca P, Pastoris MC, Valenti P, Orsi N: Growth of Legionella spp. under conditions of iron restriction. J Med Microbiol 34:113-118, 1991 10. Struelens MJ, Mondal G, Roberts M, Williams PH: Role of bacterial and host factors in the pathogenesis of Shigella septicemia. Eur J Clin Microbiol Infect Dis 9:337344, 1990 11. Towns ML, Bennett B, Check IJ, Hunter RL: An improved method for determining the microbial inhibitory activity of serum and its application to the study of patients with leukemia. Am J Clin Pathol 92:192-198, 1989 12. Ward CG, Hammond JS, Bullen JJ: Effect of iron compounds on antibacterial function of human polymorphs and plasma. Infect Immun 51:723-730, 1986 13. Hunter RL, Bennett B, Towns M, Vogler WR: Transferrin in disease II: Defects in the regulation of transferrin saturation with iron contribute to susceptibility to infection. Am J Clin Pathol 81:748-753, 1984 14. Bullen JJ, Rogers HJ: Effect of haemoglobin on experimental infections with Pasteurella septica and Escherichia coli. Nature 217:86, 1968 15. Mofenson HC, Caraccio TR, Sharieff N: Iron sepsis: Yersinia enterocolitica septicemia possibly caused by an overdose of iron. N Engl J Med 316:1092-1093, 1987 16. Modun BJ, Cockayne A, Finch R, Williams P: The Staphylococcus aureus and Staphylococcus epidermidis transferrin-binding proteins are expressed in vivo during infection. Microbiology 144:1005-1012, 1998 17. Cantinieaux J, Boelaert J, Hariga C, Fondu P: Impaired neutrophil defense against Yersinia enterocolitica in patients with iron overload who are undergoing dialysis. J Lab Clin Med 111:524-528, 1988 18. Bordin JO, Heddle NM, Blajchman MA: Biologic
S52
effects of leukocytes present in transfused cellular blood products. Blood 84:1703-1721, 1994 19. Harrison PM, Arosio P: The ferritins: Molecular properties, iron storage function and cellular regulation. Biochim Biophys Acta 1275:161-203, 1996 20. Flament J, Goldman M, Waterlot Y, Dupont E, Wybran J, Vanherweghem J-L: Impairment of phagocyte oxidative metabolism in hemodialyzed patients with iron overload. Clin Nephrol 25:227-230, 1986 21. Patruta SI, Edlinger R, Sunder-Plassman G, Ho¨rl WH: Neutrophil impairment associated with iron therapy in hemodialysis patients with functional iron deficiency. J Am Soc Nephrol 9:655-663, 1998 22. Cui Y, Zhang ZN, Li RS: Influence of latent iron deficiency on generation of oxygen species in polymorphonuclear leucocytes. Chin Med J 103:642-646, 1990 23. Seifert A, Von Herrath D, Schaefer K: Iron overload, but not treatment with desferrioxamine favours the development of septicemia in patients on maintenance hemodialysis. Q J Med 65:1015-1024, 1987 24. Tielemans CL, Lenclud CM, Wens R, Collart FE, Dratwa M: Critical role of iron overload in the increased susceptibility of haemodialysis patients to bacterial infections. Beneficial effects of desferrioxamine. Nephrol Dial Transplant 4:883-887, 1989 25. Boelaert JR, Daneels RF, Schurgers ML, Matthys EG, Gordts BZ, van Landuyt HW: Iron overload in haemodialysis patients increases the risk of bacteraemia: A prospective study. Nephrol Dial Transplant 5:130-134, 1990 26. He´bert PC, Wells G, Blajchman MA, Marshall J, Martin C, Pagliarello G, Tweeddale M, Schweitzer I, Yetisir E: A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med 340:409-417, 1999 27. Kessler M, Hoen B, Mayeux D, Hestin D, Fontenaille C: Bacteremia in patients on chronic hemodialysis. Nephron 64:95-100, 1993 28. Hoen B, Kessler M., Hestin D, Mayeux D: Risk factors for bacterial infections in chronic haemodialysis
STEVEN FISHBANE
adult patients: A multicentre prospective survey. Nephrol Dial Transplant 10:377-381, 1995 29. Hoen B, Paul-Dauphin A, Hestin D, Kessler M: EPIBACDIAL: A multicenter prospective study of risk factors for bacteremia in chronic hemodialysis patients. J Am Soc Nephrol 9:869-876, 1998 30. Collins A, Ebben J, Ma J: Frequent IV iron dosing is associated with higher infectious deaths. J Am Soc Nephrol 8:190A, 1997 (abstr) 31. Collins A, Ebben J, Ma J, Xia H: I.V. iron dosing patterns and mortality. J Am Soc Nephrol 9:205A, 1998 (abstr) 32. Fishbane S, Maesaka JK: Iron management in endstage renal disease. Am J Kidney Dis 29:319-333, 1997 33. Collins AJ: IV iron and mortality issues in ESRD patients. Slide 28, on file, Amgen, Thousand Oaks, CA, 1997 34. Collins A, Ma J, Ebben J: Patients survival is associated with hematocrit (HCT) level. J Am Soc Nephrol 8:190A, 1997 (abstr) 35. Madore F, Bridges K, Brugnara C, Lew NL, Lowrie EG, Lazarus JM, Owen WF: A population study of the interplay between iron, nutrition and inflammation in erythropoiesis in hemodialysis patients. J Am Soc Nephrol 7:1456A, 1996 (abstr) 36. Collins A, Xia H, Ebben J, Ma J: Change in hematocrit and risk of mortality. J Am Soc Nephrol 9:204A, 1998 (abstr) 37. Nissenson AR: Optimal hematocrit in patients on dialysis therapy. Am J Kidney Dis 32:S142-S146, 1998 (suppl) 38. Canadian Erythropoietin Study Group: Association between recombinant human erythropoietin and quality of life and exercise capacity of patients receiving haemodialysis. BMJ 300:573-578, 1990 39. Eschbach JW, Glenny R, Robertson T, Gurhrie M, Rader B, Evans R, Chandler W, Davidson R, Easterling T, Denney J, Schneider G: Normalizing the hematocrit (HCT) in hemodialysis patients (HDP) with EPO improves quality of life (Q/L) and is safe. J Am Soc Nephrol 4:425A, 1993 (abstr)