Anemia in the Elderly

Abstract

As the population ages, increasing attention has become focused on the prevalence of anemia in elderly individuals. Anemia occurs in more than 10% of individuals who are older than the age of 65 years, and it increases to more than 50% in individuals who are older than the age of 80 years. Although the anemia is typically mild and unlikely to result in symptoms, it is uniformly associated with increased morbidity and mortality as assessed in large cohort studies. Anemia is an independent predictor of these adverse outcomes both in healthy community-dwelling subjects and in patients with significant co-morbidities. Efforts to understand the pathophysiology of anemia in this population, especially the one third of patients with “unexplained” anemia, have focused on the potential contributions of inflammatory pathways, erythropoietin resistance, and changes in hematopoietic stem cells to the age-dependent decrease in red cell mass. We would argue that these pathways are closely interrelated and combine to lead to anemia in aging individuals. This brief review summarizes the current understanding of this entity and our studies aimed at further delineating its pathophysiology.

INTRODUCTION

Several large cohort studies of healthy community-dwelling individuals have shown that anemia is an independent predictor of morbidity and mortality in elderly patients (1–3). The basis for this negative impact on health and survival remains obscure. The anemia tends to be mild, and it is therefore unlikely to contribute to negative health outcomes on its own; it seems more likely that it is a biomarker of other biological pathways that affect the health and survival of the elderly population.

Study of the Third National Health and Nutrition Examination Survey (NHANES III) cohort has shown that the prevalence of anemia in men (hemoglobin, Hb <13 g/dL) and women (Hb < 12g/dL) who are older than 65 years was 11% and 10%, respectively. Prevalence was higher in non-Hispanic blacks when compared with non-Hispanic whites. One third of anemic individuals had a nutritional basis for anemia (mostly iron deficiency), whereas the other two thirds were divided between “anemia of chronic inflammation” and “unexplained anemia.”(4) A more recent study examined anemia prevalence and incidence in a longitudinal manner and showed that whereas the prevalence of anemia is 5% to 7% at 65 years, it increases rapidly with age, exceeding 40% in individuals who are older than 80 years. Again, anemia is consistently mild, with the prevalence of severe anemia (Hb<8 g/dL) shown to be less than 0.5% (5).

It is unknown why such a large fraction of elderly individuals are anemic, but there are several potential clues to the pathogenesis of anemia in this population. This discussion will be focused around the following questions:

1. What is the role of hepcidin-induced anemia of inflammation in anemia in the elderly?

2. What is the role of erythropoietin (EPO) resistance and/or EPO insufficiency in anemia in the elderly?

3. What is the role of an age-related pro-inflammatory state in the development of anemia in the elderly?

4. What is the basis for the increased morbidity and mortality associated with anemia in the elderly?

HEPCIDIN AND THE ANEMIA OF INFLAMMATION

The anemia of inflammation (AI) has long been recognized as a complication of systemic inflammatory diseases such as rheumatoid arthritis. It is characterized by the presence of decreased iron incorporation into developing erythrocytes despite the presence of normal or increased iron stores in the bone marrow. Biochemically, it is marked by decreased serum iron levels with normal or decreased iron binding capacity and normal or increased ferritin levels (6). Although the level of ferritin also increases as an acute phase reactant in the presence of acute inflammation, a serum ferritin level of more than 200 mg/dL serves to confirm that the anemia is not related to iron deficiency (7). Our understanding of the mechanism of this anemia has been transformed by the discovery, isolation, and characterization of hepcidin as the primary factor influencing this iron-deficient hematopoiesis in the presence of adequate iron stores (8).

Hepcidin is an antimicrobial peptide synthesized by the liver that serves as a key regulator of iron metabolism. By targeting the major iron transporter, ferroportin, for degradation, it inhibits intestinal iron absorption and also blocks iron release from macrophages. Transgenic mice overexpressing hepcidin die of severe anemia (9), and hepdicin-null mice die of iron overload (10). Furthermore, hepcidin is potently induced by the pro-inflammatory cytokine interleukin-6 (IL-6), and the link between inflammation and anemia is therefore postulated to be mediated by IL-6-induced hepcidin elevation and consequent changes in iron availability for developing erythrocytes (11).

It was thought that measurement of serum hepcidin would therefore provide an excellent screening test for the definitive diagnosis of AI, which has previously been based on screening serum iron studies and clinical parameters. It was further thought that mild increases in hepcidin levels might indicate subtle inflammatory changes that might allow a diagnosis of AI in some patients with unexplained anemia for whom iron studies were not definitive. Unfortunately, although the link between hepcidin and inflammatory anemia is undisputed, hepcidin measurements have not been helpful in delineating subtypes of anemia in practice. A study of hepcidin levels in the InCHIANTI Study showed the expected finding that iron deficiency was associated with low hepcidin levels. However, in patients who have low serum iron levels without iron deficiency who were diagnosed with AI, hepcidin levels were, surprisingly, also low. This subset of patients had increased IL-6 and C - reactive protein (CRP) levels, and low serum iron levels, as predicted; however, hepcidin levels were not correspondingly elevated. The authors speculated that the serum hepcidin may not need to remain elevated to maintain the abnormal iron metabolism leading to anemia, or that perhaps the level measured at a single time point may not adequately reflect iron kinetics in the setting of inflammatory anemia (12). Regardless, it suggests that hepcidin may not be a good biomarker for diagnosis of AI, nor is it likely to shed light on the pathogenesis of unexplained anemia.

EPO RESISTANCE AND INSUFFICIENCY IN ANEMIA IN THE ELDERLY

Aging appears to be associated with a decreased sensitivity of the hematopoietic stem cell to EPO, as evidenced by increasing EPO levels with age (13). Study of the Baltimore Longitudinal Study on Aging documented that EPO levels increase with age in healthy, non-anemic individuals (13, 14). The slope of the increase was greater in those without diabetes, renal insufficiency, or hypertension, and a decreased slope of increase is associated with anemia. These data suggest that anemia occurs as a failure of a normal compensatory increase in EPO levels required to maintain the hemoglobin level in aging individuals. It has been proposed that this reflects secondary co-morbidities leading to decrement in renal function and EPO production. Two possible mechanisms have been proposed to explain why EPO demand increases with age. It is well-known that pro-inflammatory cytokine expression increases with age, and that the downstream inflammatory pathways induce EPO resistance (15–17). On the other hand, in vitro colony assays have shown that there is also a cell-intrinsic change in hematopoietic stem cells with age. Hematopoietic stem cells from elderly patients show decreased erythroid colony-forming activity and a shift toward myeloid maturation (18). This has been further suggested to contribute to the predisposition to myeloid malignancy in the elderly population.

INFLAMMATION AND ANEMIA IN THE ELDERLY

As already noted, it is well-established that levels of inflammatory cytokines and acute phase proteins increase with age; however, it is controversial whether this increase is characteristic of a healthy elderly population or whether it reflects co-morbidities that accompany the aging process. In the InCHIANTI study of more than 1300 healthy community-dwelling elderly individuals, it was documented that IL-6, IL-1 receptor antagonist, IL-18, CRP, and fibrinogen levels were all increased in individuals older than 65 years of age, although the independent effect of age on these parameters was small after adjusting for cardiovascular risk and other co-morbidities (16). In other studies, anemia and inflammation have both been shown to contribute to the development of frailty, a syndrome characterized by weight loss, generalized weakness, impaired mobility, and poor balance (1, 2, 19, 20).

We hypothesize that non-nutritional anemia in the elderly develops from an interaction between an increased inflammatory milieu and age-related co-morbidities. Furthermore, we postulate that two major inflammatory pathways contribute to the pathophysiology of age-associated anemia. First, inflammatory signalling through IL-6 leads to iron-limited hematopoiesis with increased hepcidin, poor iron incorporation into developing erythrocytes, and classic anemia of inflammation. Second, tumor necrosis factor α (TNFα)–mediated pathways lead to increased EPO resistance of the hematopoietic stem cell, increasing EPO demand to maintain erythropoiesis, and eventually outstripping the ability of the aging kidney to maintain an adequate EPO level to sustain a normal haemoglobin (21). Individuals in whom the latter pathway predominates do not have the characteristic iron studies of patients with anemia of inflammation; and this group may represent a subset of patients with “unexplained” anemia.

Vitamin D in Anemia in the Elderly

We investigated the potential role of vitamin D in modulating inflammatory pathways related to anemia. Review of the NHANES III cohort revealed that vitamin D deficiency was very common and was associated with anemia. The odds ratio for anemia in the setting of vitamin D deficiency was 1.6, and was independent of age, sex, and race/ethnicity. Interestingly, among those with anemia, vitamin D deficiency was most prevalent in those with anemia of inflammation, with an odds ratio of 1.85 (22).

We also found in vitro evidence that vitamin D may also impinge on the TNF-signaling pathway. TNF has been shown by several investigators to decrease colony-forming activities of hematopoietic stem cells in vitro. We have previously shown that resveratrol, postulated to serve as an anti-inflammatory, can partially reverse the TNF-induced decrease in colony-forming activity through modulation of NFκB signalling pathways. We have shown a similar impact of vitamin D in restoring colony-forming capacity to cultured CD34+ hematopoietic progenitor cells (23). We are currently assessing the potential clinical significance of these results by assessing the impact of vitamin D on anemia in the elderly as an ancillary study to a large cohort examining the impact of vitamin D and omega-3 fatty acids, separately and together, on cardiovascular and cancer outcomes (VITamin D and OmegA-3 TriaL (VITAL) Study). By assessing the impact of vitamin D on anemia subtypes, we hope to tease out the contributions of vitamin D to pathways governing unexplained anemia and anemia of inflammation in a healthy community-dwelling population of elderly individuals.

Anemia of Aging and the Anemia of HIV

Many chronic illnesses, including cardiovascular disease, chronic obstructive pulmonary disease, and cancer, appear to develop at an earlier age in the HIV population, suggesting that HIV may be a model for premature aging. Whether this is going to be modified by the availability of earlier effective therapy remains an open question, but many of these complications appear in successfully treated patients.

The impact of anemia on the prognosis of HIV is reminiscent of that of anemia of the elderly. Although frequently only mild, anemia is a potent predictor of survival from HIV, even in the era after the development of combined anti-retroviral therapy (cART). This has been studied in the Veterans Administration Cohort Study (VACS), a large cohort study focused on the natural history of HIV in aging veterans. Based on this population, Justice et al have developed a “VACS index,” which predicts the outcome of HIV-related disease, focusing on the contribution of age, CD4 counts, viral load, anemia, and inflammatory markers on survival (24).

We studied anemia in the VACS, taking advantage of stored specimens on 3213 veterans with and 3161 veterans without HIV. The median age of the patients is 50 years, and they have a projected follow-up of 8 years. We examined the prevalence of subtypes of anemia, and correlated anemia with 10 functional inflammatory cytokine polymorphisms, including IL-6, IL-1β, TNFα, leptin, and leptin receptor. The data showed that anemia was the strongest predictor of survival in the VACS index, eclipsing CD4 count and viral load as an independent predictor of mortality. Furthermore, persistent anemia despite successfully treated disease with an undetectable viral load continued to predict for poor survival. A similar strong negative predictive value for survival in association with anemia has been observed in several other large cohort studies.

We also established a link between a functional polymorphism in the leptin receptor and anemia in the HIV-positive population. Because the prevalence of anemia in the HIV-negative population was considerably lower, we were unable to establish the same link in HIV-negative individuals (25). However, we did interrogate the cohort from the Jupiter study, and preliminary results did not confirm that the polymorphism was linked to anemia in that population (unpublished data). How this polymorphism has an impact on anemia in the HIV population remains uncertain. However, because leptin is produced by adipocytes and increases local TNF production in the marrow (which is more than 50% adipose tissue in patients who are older than 50 years), it is attractive to postulate that leptin responsiveness may positively regulate the impact of inflammation, at least in the setting of HIV. This hypothesis remains to be tested.

THE MORBIDITY AND MORTALITY OF ANEMIA IN THE ELDERLY

As previously noted, anemia in the elderly is usually mild and unlikely to be of direct clinical significance. Hence, it remains unexplained why anemia of any degree has such a strong correlation with morbidity and mortality in aging populations. Although there are as yet no data to explain this, we hypothesize that non-nutritional anemia in the elderly is all related to inflammatory signaling, either as the anemia of inflammation mediated through IL-6 pathways or as iron-independent anemia mediated through TNF and NFκB. We are currently investigating whether other interventions in large cohort studies aimed at determining the impact of changes in the inflammatory milieu on cardiovascular and cancer outcomes will have an impact on anemia in the elderly. Specifically, we are investigating whether, if interventions have an impact on outcomes, they are correlated with changes in the prevalence of anemia. We are studying this in two settings: 1) as described above, we are determining whether either vitamin D or omega-3 fatty acids have an impact on anemia as an ancillary investigation to the larger trial determining their impact on cardiovascular disease and cancer; and 2) we are proposing an ancillary study of the prevalence and incidence of anemia in a Cardiovascular Inflammation Reversal Trial (CIRT). The parent study is investigating the efficacy of low-dose methotrexate in the secondary prevention of cardiovascular events. The two studies should provide complementary information because the first is a study of a healthy aging population, whereas the latter is a study of patients with diabetes and metabolic syndrome who have already had a cardiovascular event. The results of these two investigations should elucidate whether anemia is responsive to inflammatory modulation in patients with and without significant co-morbid disease. Furthermore, correlation with the primary study outcomes will suggest whether response of anemia is a useful biomarker for predicting decreased negative outcomes in the elderly population.