Anemia in Chronic Obstructive Pulmonary Disease: an insight intoits prevalence and pathophysiology
Afroditi K. Boutou, Nicholas S. Hopkinson, Michael I. Polkey
NIHR Respiratory Biomedical Research Unit at Royal Brompton and Harefield NHS Foundation Trust and Imperial College, London, UK
Short title: Anemia in Chronic Obstructive Pulmonary Disease
Corresponding author
Afroditi K Boutou
Stratigou Sarafi 9-13, 55132, Thessaloniki, Greece
Email:
Telephone number: 00306946611433
Keywords:
Anemia;
Chronic Obstructive Pulmonary Disease;
Epidemiology;
Systemic Inflammation;
Aetiology
Abstract
Chronic Obstructive Pulmonary Disease (COPD) is a major health problem, with increasing morbidity and mortality. There is a growingliterature regarding the extra-pulmonary manifestations of COPD, which can have a significant impact on symptomburdenand disease progression. Anemia is one of the more recently identified co-morbidities, with a prevalence that varies between 4.9% to 38% depending on patient characteristics and the diagnostic criteria used. Systemic inflammation seems to be an important factor for its establishment and repeatedbursts of inflammatory mediators during COPD exacerbations could further inhibit erythropoiesis. However, renal impairment, malnutrition, low testosterone levels, growth hormone level abnormalities, oxygen supplementation, theophylline treatment, inhibition of angiotensin converting enzyme and aging itself are additional factors that could be associated with the development of anemia. This review evaluates the published literature onthe prevalence and significance of anemia in COPD. Moreover, it attempts to elucidate the reasons for the high variability reported and investigates the complex pathophysiology underlying the development of anemia in these patients.
The prevalence of anemia in COPD
There is a huge variation of anemia prevalence in the published literature, most likely reflecting the different cohorts which have been studied. A systematic search in the electronic database of PubMed using the search terms Chronic Obstructive Pulmonary Disease (COPD) and an(a)emia, h(a)ematocrit, h(a)emoglobin, iron deficiency or red blood cells, identified 24 studies which were conducted in humans and published as full-text articles in English between 2005 and 2013. These studies reported the prevalence of anemia in COPD, using either percentages or absolute patient numbers and explored its potential association with the disease or its impact on several disease outcomes (Table 1). In these reports anemia frequency varied widely from 4.9% to 38%.(1)(2)In contrastto expectations, polycythemia was less common than anemia,its prevalence ranging from6-18.1%.(3)(4)(5), perhaps reflecting more widespread use of domiciliary oxygen and other forms of respiratory support.
Several reasons can be proposedfor this discrepancy in anemia prevalence, and the varyingcharacteristics of the populationsstudiedisone of them, as anemia prevalence has been investigated in stable COPD outpatients,(6)(4)(7)(8)(9)(10)(11)(12)hospitalized patients foran acute exacerbation of COPD (AECOPD),(13)(14)(15)(15)(17)intubated patients in the intensive care unit(18), COPD patients from the general population(10)(11)and COPD patients using long-term oxygen treatment (LTOT) or non-invasive ventilation (NIV).(3)(5)(19)These patients not only presented with a different COPD severity, but have a dissimilar health statusoverall, together with varying burden of concomitant disorders that could cause anemia,so results are not easily comparable.
COPD diagnosis, according to ERS/ATS guidelines, should be based on specific spirometric criteria.(20) Nevertheless, this is not the case for every published study which has investigated anemia prevalence in COPD patients. Although most authors have used the post-bronchodilator Forced Expiratory Volume in 1 second/Forced Vital Capacity (FEV1/FVC)<0.7,(4)(7)(8)(9)(14)(2)(21) several have applied the ICD 9/10 codes(13)(22)(23)(24) or identified COPD patients from existing databases without describing the diagnostic criteria in detail.(19)Thus it is possible that in these studies patients with respiratory symptoms of airflow obstruction without fulfillingspirometric criteria for COPD were misclassified, creating more variation regarding anemia prevalence.
Furthermore, anemia definition has been variable in published studies. According to current World Health Organization, anemia in the general population is defined by hemoglobin (Hb) levels <13 g/dL in male and <12 g/dL in female(25)and these thresholds have been used by several authors in order to identify anemic COPD patients.(5)(8)(11)(15)(26)However, the use of a Hb <12 g/dL threshold to define anemia in postmenopausal women is currently under debate,(27) so the 13g/dL threshold for both male and female has also been applied.(4)(7)(2)(19)Controversy is even more evident regarding COPD patients admitted in the ICU, since there is as yet no accepted definition of abnormal hemoglobin values in the critically ill.(18) A final problem with many of the studies is that the prevalence of anemia was often not measured in an age and sex matchedpopulation with similar comorbidity burden (control population);(3)(9)(14)(26) thus it is difficult to establish whether the prevalence of anemia is increased in COPD.
Inone of the first studies in the field, John et al studied 101 stable severe COPD outpatients; the prevalence of anemia was 13%.(9) Althoughpatients with disorders that could be accompanied bylow Hb levels, such as heart failure, gastrointestinal bleeding and malignancy were excluded, renal function was not investigated. Results of two more studies conducted in COPD hospital outpatients of various severities were comparable, with anemia prevalence ranging between 15%-17.1%.(7)(4)The only study which used both clinical and laboratory criteria to exclude all potential causes of anemia, apart from anemia of chronic disease, estimatedthe prevalence of the latter to be 10.2% among stable hospital outpatients.(8)Interestingly,the prevalence of anemia in selected populations with respiratory failure was not very different from that seen in hospital outpatients. Chambellan et al reported that 12.6% male and 8.2% female of the total 2,542 outpatients receiving LTOT were anemic(3); these results were similar to the ones of Dal Negro et al(19) and of Kollert et al,(5) who reported an anemia prevalence of 11.3% and 14.9% among outpatients under long-term oxygen treatment or domiciliary non-invasive ventilation, correspondingly. Conversely, two large population-based studies estimated a lower prevalence of anemia among COPD patients, rangingbetween 7.3-7.5%(10)(11); the inclusion of COPD patients with less severe disease burden compared to the ones with respiratory failure or under a hospital follow-up, is probably the main reason for this discrepancy.
As expected, anemia is even more frequent among patients hospitalized for AECOPD or other causes. In a retrospective study of a series of patients with various disorders who were discharged from hospital, anemia prevalence among COPD patients was 23.1%, comparable to the one among individuals with heart failure.(23)In a longitudinal study by Almagro et al anemia prevalence among hospitalized patients for AECOPD was 19.3%, while other authors have reported even higher frequencies, ranging from 26 to 33%.(15)(28) Hospital-acquired anemia (29) is a unique entity affecting patients with various disorders who are admitted in hospital; nevertheless, during AECOPD the burst of systemic inflammation is a factor further inhibiting erythropoiesis, as described below in detail. In contrast to previous data, Nowinskiet al(1) and Barbra et al(13) reported a much lower frequency for anemia among patients with AECOPD (4.9% and 9.8%,respectively); however, the different methodology used to define the parameters of interest in studies might have caused this discrepancy (Table 1).
Much of the information regarding the frequency of anemia came from two large retrospective studies using healthcare databases. Shorr et al studied a population of 2,404 COPD patients and identified that 788 (33%) of them were anemic.(24) The study used the WHO definition of anemia and was conducted in a large patient population; however, patients with chronic kidney disease were not excluded, while there is no information whether hemoglobin was measured during stable state or hospitalization. Halpern et al indicated that out of the 132,424 COPD patients whowere included in the US Medicare Claims Database, 27,932 COPD (21%) were anemic.(22) Although it is not known whether anemia was identified on an inpatient or outpatient basis, the lower percentages of anemia are probably due to the exclusion of patients with renal insufficiency, along with other causes of anemia. Although these two studies do not offer further information regarding the specific anemia prevalence in different COPD populations, they indicate that in a general COPD population of various severity and several comorbidities, anemia is a commoncomplication.
In summary,although anemia occurs frequently in COPD patient, its prevalence varies widely in the published literature. The baseline characteristics of the study population, the various comorbidities present and the different methodology adopted to define the measures of interestare the main causes of this discrepancy, which make the results of published studies difficult to compare.
Pathogenesis of anemia in COPD
Although the presence of anemia has been repeatedly reported, studies that have investigated the specific causes of anemia in COPD are scarce and manypotential etiological mechanisms, which are not mutually exclusive, exist.(Figure 1)Nevertheless, COPD has been increasingly recognized as a disorder with important systemic manifestations,(30) sothe development of “anemia of chronic disease” (ACD) among COPD patients could be expected.
Pathogenesis of anemia of chronic disease
ACD is an immune-driven disorder;(31) it accompanies several diseases which are characterized by sub-acute or chronic immune activation, such as malignancies, systemic autoimmune disorders and inflammatory diseases.(32)(33)(34)(35)For this reasonACD has also beencharacterized as “anemia of inflammation”.(31)ACDcan be classified as anemia due to reduced erythropoiesis andis usually a mild to moderate normochromic, normocytic anemia, though less frequently, it can have a hypochromic microcyticpattern.(36)The pathophysiological background of ACD is immunological; cytokines and cells of the reticuloendothelial system induce changes in:a) iron homeostasis, b) proliferation and differentiation of progenitor erythroid cells and c) production of erythropoietin which all contribute toits establishment.(31)
a) disorders of iron homeostasis
One of the most characteristic features of ACD is the development of disorders in iron homeostasis, with enhanced uptake and retention of iron within the cells of the reticuloendothelial system. This leads to a diversion of iron from the circulation, resulting inreduced intake of iron from erythroid progenitor cells and, thus, to restricted erythropoiesis.(31)
Iron homeostasis involves several mechanisms. Dietary iron is transported, asferrous iron, across the apical surface of the intestinal epithelium cell membrane by means of a transmembrane protein, the divalent metal transporter 1 (DMT1), via a coupling-proton mechanism.(37)(38)After iron has entered the cells, it can either be stored in cytoplasmic storage, ferritin, or be exported to the plasma through protein-carriers of the basolateral membrane,(39) the most significant of which is ferroprotein.(40)Exported iron is then bound toplasma transferrin, which is the primary form by which iron is transported in blood and delivered to various cells.(39)Different cells use iron in different ways; however, erythrocytes, hepatocytes and reticuloendothelial macrophages are the most important.(41)This complex iron homeostasis is regulated by several molecules, with hepcidin -a 25 amino acid protein secreted by liver- being the most important.(42)Studies in both mice and humans have indicated that hepcidin is involved in the mechanisms of response tohypoxia and anemia; in these conditions hepcidin levels decrease, its inhibitory effect on ironis diminished, and more iron is made available from diet and the iron storage of macrophages for erythropoiesis. The opposite happens during infection or systemic inflammation; hepcidin synthesis is markedly induced and, thus, iron availability decreases.(43)
Chronic inflammation, as seen in ACD, disrupts iron homeostasis in multiple ways. Tumor Necrosis Factor-1α (TNF-a) and Interleukin(IL)-1, increase the synthesis of ferritin by liver cells and macrophages by inducing its transcription and translation.(44)IL-6 distorts iron metabolism,(45) since it modulates ferritin translation, expression of transferrin mRNA and, possibly, expression of DMT1.(32)TNF-a and Interferon-γ (INF-γ) induce the production of DMT1 and block the release of iron from macrophages by down-regulating ferroprotein expression. (31) IL-10 can also impair iron homeostasis by inducing ferritin expression and increasing the acquisition of iron by macrophages.(46)Finally, lipopolysaccharide and IL-6 are major stimulants of hepcidin synthesis, leading to hypoferremia within hours of inflammatory stimulant in animal models.(43)
b) disorders of proliferation and differentiation of erythroid progenitor cells(35)
Cytokine-mediated stem cell proliferation arrest or induction of apoptosis, as well as an interaction with erythropoietin or other major factors that promote erythropoiesis are the potential underlying mechanisms for these disorders.(32)Several cytokines, exert inhibitory effects on erythroid progenitor cells; IL-1a inhibits erythropoiesis in vivo in mice and in vitro in humans,(47) TNF-a and INF-γ inhibit erythroid colony formation in uremic sera,(48) while there is an inverse relationship between INF-γconcentration, reticulocytes count and hemoglobin (Hb) concentration.(49).Inflammatory cytokines can also exert a direct toxic effect on progenitor cells, promoting the formation of unstable free radicals such as nitric oxide or peroxide from macrophages.(31)(50)Moreover, erythropoiesis is further inhibited by the limited availability of iron to erythroid progenitor cells.(32)
c) resistance to the action of erythropoietin
Erythropoietin (EPO) is a protein hormone which promotes erythroid cell proliferation; the most potent known stimulus for EPO production is tissue hypoxia.(51) Most patients with ACD have disproportionately low erythropoietin levels for the severity of anemia present.(52)(53)In vitro studies indicate that IL-1 and TNF-α inhibit hypoxia-induced EPO production in a dose-dependent manner,(54) while in anemic individuals high levels of IL-1, IL-6 and TNF-α areassociated with low levels of EPO.(53)Cytokine overproduction is the most probable cause for hyporesponsiveness to EPO treatment in anemic individuals without iron deficiency.(55)Abnormal iron metabolism also contributes to EPO resistance; iron not only becomes unavailable for erythroid progenitor cells,(31) but its depletionmay also impair erythropoietin transgene expression.(56)
Systemic inflammation: the link between COPD and ACD?
There is a huge amount of evidence in the published literature regarding the presence of systemic inflammatory responses among patients with COPD. Serum levels of C-reactive protein (CRP), TNF-α, IL-6, IL-8 and fibrinogen (57)(58)(59)(60)(61)are only a few of the inflammatory markers that have been found to be significantly increased among patients with stable disease compared to healthy controls. Cytokine levels have already been associated with disease burden, as established by severity of obstruction, BODE index, free fat mass, body mass index and exercise capacity,(57)(58)(62)(63) and with several systemic COPD manifestations and comorbidities, such as muscle cachexia, pulmonary hypertension, heart disease and depression.(64)(65)(66)(67)(68)
Currently, two studies have evaluated the potential association between inflammatory mediators and anemia in stable COPD patients. John et al studied a population of 101 COPD patients, 13 ofwhomwere anemic, and foundthat CRP and IL-6 levels were significantly elevated inanemic COPD patients compared to controls, while CRP was also significantly higher in anemic compared to non-anemic COPD patients. EPO concentration was also higher in anemic individuals compared to both non-anemic patients and healthy controls.(9) In another case-control study where ACD in COPD patients was clinically and laboratory defined, the concentration of all studied cytokines (that is TNF-α, IL-6, IL-10 and INF-γ) was higher in the group of anemic compared to non-anemic COPD patients; However, the between group differences were statistically significant only for INF-γ and IL-10. Likewise, EPO levels were also higher in anemic individuals, indicating the presence of EPO resistance due to systemic inflammation.(69)
Systemic inflammation and ACD: The role of exacerbations
One of the most important complications in the course of COPD is acute exacerbation (AECOPD), during which a further burst of inflammatory mediators occurs. Sputum or serum levels of CRP, TNF-α, IL-6, IL-8, IL-1β, IL-10, fibrinogenand total cell counts are significantly increased, compared to stable patients or controls,(59)(70)(71)(72) and this increase often persists after the improvement of lung function.(70)
Two studies have studied the potential association between systemic inflammation and EPO levels during an AECOPD, but results are conflicting. Markoulaki et al(73) used measurements at three time points in a selected cohort of 93 COPD patients who presented with AECOPD. Haemoglobin levels were significantly decreased and EPO levels were significantly increased during the acute phase compared to resolution and steady phases; EPO and Hb were negatively correlated during the acute phase and positively correlated during resolution and stable phases. Moreover, IL-6 levels were negatively correlated with Hb and positively correlated with EPO, indicating the presence of EPO resistance during the acute phase of AECOPD.
In a previous report Sala et al(74)identified lower levels of EPO among exacerbated COPD patients, compared to stable COPD patients, non-COPD smokers and healthy controls. In COPD patients EPO levels correlated with CRP and circulating neutrophils, while in a small (n=8) subgroup of COPD patients who were studied both at AECOPD and stable phase, EPO levels significantly increased, when acute phase resolved. These conflicting results indicate that more studies are needed to reveal the complex pathophysiology underlying EPO regulation during AECOPD, especially now that a distinct COPD phenotype, “the frequent exacerbator” with increased airway and systemic inflammation and a high prevalence of extrapulmonary comorbidities has beenproposed.(75)
Oxygen supplementation
As described above, tissue hypoxia is the most potent trigger for EPO production which results inincreased erythropoiesis.(51) Thus, one could hypothesize that the treatment of hypoxia in COPD patients would result to the reduction of EPO production and inhibition of erythroid progenitor cells proliferation, leading to anemia. However, results from human studies regarding both the EPO response to hypoxia and the impact of oxygen treatment on EPO concentration are conflicting and sometimes paradoxical.
Guidetet al studied 21 COPD patients with severe hypoxemia to find that EPO levels were not significantly different between polycythemic and non-polycythemic groups. The absence of adaptive polycythemia in the presence of severe hypoxia was not associated with a quantitative deficit of EPO, nor to a lack of sensitivity of progenitor cells to its action.(76) In a case-control study of 32 patients and 34 matched non-smokers healthy subjects, Tsantes et al found that although erythrocytocis and macrocytosis, which are both induced by hypoxia, occur more often among hypoxemic COPD subjects, they are not a consistent feature of hypoxia in COPD.(77)Moreover, the severity of hypoxemia could be of some importance, apart from its presence; Fitzpatrick et al(78) concluded, after studying 8 COPD and 9 healthy subjects, that, mild nocturnal oxygen desaturation is not associated with elevated EPO levels, whereas daytime hypoxaemia accompanied by severe nocturnal desaturation is associated with increased serum EPO levels both by day and by night. After comparing a cohort of 40 COPD patients with 40 healthy subjects, Casale et al indicated that normal circardian rhythm of circulating serum EPO levels is lost in COPD patients and mean daily levels of EPO are significantly higher, suggesting that daytime hypoxemia and severe nocturnal desaturation might be the cause of this abnormality.(79)