Pathogenesis and clinical implications of HIV-related anemia in 2013

Anemia is among the most common hematologic abnormalities in patients with HIV and has been associated with disease progression and poor clinical outcomes.¹  Study of the unique pathophysiology of anemia in HIV has been recognized as an important step toward improving therapeutic options and disease management. Recent emerging studies have also begun to suggest that there may be a connection between the underlying mechanisms driving the development of anemia in the context of HIV infection and the pathophysiology of anemia in the elderly. A growing body of data has established that HIV infection is associated with increased expression of proinflammatory cytokines,²  including many of the same biomarkers previously associated with inflammation and anemia of aging.³  Ongoing studies of the development and progression of anemia in the HIV⁺ population will help characterize whether this common clinical presentation reflects an acceleration of the systemic inflammation already well known to be associated with the biology of aging. Because effective antiretroviral therapy (ART) has dramatically increased the life expectancy of patients with HIV, improved understanding of the ways in which HIV infection alters physiologic parameters will be an important component of developing targeted therapies to affect morbidity and mortality in this population.

Anemia in HIV⁺ patients is a well-documented phenomenon. An earlier systemic review of the literature identified 31 distinct studies reporting either the incidence or prevalence of anemia in the HIV⁺ population,¹  and the connection between anemia and HIV continues to be an active area of research investigation.⁴  Anemia has been recognized as a clinical problem in HIV⁺ patients both before and after the advent of combination ART (cART), suggesting that there may be a fundamental physiologic change in the regulation of erythropoiesis resulting from viral infection. However, when evaluating the large body of literature on this topic, it is important to note the various definitions used to clinically diagnose a patient with anemia. The World Health Organization (WHO) defines anemia as a hemoglobin concentration of < 13 g/dL for men and < 12 g/dL for nonpregnant women.⁵  Although this definition is widely used in the literature, many of the largest cohort studies on HIV are done in Africa or Asia, where a more stringent hemoglobin cutoff of < 10 g/dL is often used for clinical stratification. However, even while acknowledging differences in definitions, there is no doubt that anemia is a significant clinical challenge for the HIV⁺ population, because somewhere between 1.3% and 95% of patients in published cohorts meet diagnostic criteria.¹ 

A closer review of epidemiological data within this population reveals several noteworthy trends. First, it appears that the degree of anemia is correlated with HIV/AIDS disease progression in both retrospective and prospective analyses.¹  Several of these studies suggested that the anemia in some advanced AIDS patients results from therapy-related toxicity,^1,6  but adverse drug effects alone are not sufficient to account for the consistent association between anemia and HIV stage. Furthermore, several studies in a range of patient cohorts demonstrated that anemia has prognostic significance for morbidity and mortality measures in HIV⁺ patients,^7,8  supporting the possibility that the virus itself may play an independent role in driving dysfunctional erythropoiesis. Studies evaluating anemia in patients undergoing cART compared with patients not on active treatment extend this hypothesis, because cART appears to correct the anemia associated with HIV effectively in a significant number of patients in both adult and pediatric populations.^9,10  Conversely, a recent study demonstrated that ongoing anemia is statistically associated as a marker of treatment failure in patients newly initiated on cART who fail to achieve appropriate viral suppression.¹¹ 

A wealth of older literature has established anemia as a prognostic marker for mortality in HIV⁺ patients.^6,8,12  Emphasizing the centrality of the role played by anemia in this patient cohort, anemia has been associated with shortened median survival in HIV⁺ patients irrespective of their initial CD4⁺ count.¹²  Risk factor stratification demonstrates that when a range of potentially relevant demographic and clinical variables are considered, CD4 count and hemoglobin are the 2 factors independently associated with mortality.⁸  More recently, several studies in Europe,¹³  Africa,⁷  and Asia^14,15  have evaluated the association between anemia and outcomes specifically in patients initiating cART. Once again, despite the widely divergent socioeconomic factors that characterize HIV worldwide, the degree of clinical anemia remains an independent risk factor for predicting mortality from HIV even upon initiation of therapy. These findings identify anemia as an important variable in monitoring HIV disease progression and management, particularly in resource-limited settings. These data also begin to suggest that understanding the mechanisms by which HIV causes anemia may have significant prognostic and therapeutic significance.

There are several factors believed to contribute to the pathophysiology of anemia observed in HIV⁺ patients. First, many of the opportunistic infections or malignancies to which HIV⁺ patients are susceptible can lead to anemia. This is particularly problematic in the developing world, where endemic infections such as malaria or hookworm can lead to substantial anemia even in those without concomitant HIV infection. Similarly, HIV infection rates are also particularly high in regions of the world with a high prevalence of thalassemia or sickle cell trait. Furthermore, as with anemia in the HIV⁻ population, micronutrient deficiencies can play a role in contributing to either microcytic or macrocytic anemia in HIV⁺ patients. Finally, as mentioned previously, many of the most common drugs included in standard cART also have a negative effect on hematopoiesis and can thus contribute to anemia in HIV⁺ patients undergoing therapy.

Despite the role of such nonspecific factors in potentiating anemia in many HIV patients, several lines of evidence suggest that the pathophysiology of anemia in HIV may include direct effects of the virus itself. First, studies from the pediatric literature indicate that, even in populations in which micronutrient deficiency is expected to be highly prevalent, the HIV⁺ cohort is less likely to be affected directly compared with the general population. HIV⁺ children in Malawi were less likely to be iron deficient than HIV⁻ children,¹⁶  In a cohort of Thai children, the majority of HIV⁺ children with anemia were not iron deficient.¹⁷  If HIV itself did not have pathogenic effects on the development of anemia, then the relative role of other risk factors such as nutritional deficiencies would likely be more extensive. In addition, HIV infection may also contribute to aberrant immune activation that exacerbates other etiologies of anemia. A recent report demonstrates that molecular mimicry between erythropoietin (EPO) and the HIV-1 p17 protein can lead to circulating auto-antibodies against endogenous EPO in some HIV⁺ patients, blunting the normal physiologic cytokine response to anemia.¹⁸ 

Several studies have also demonstrated a direct effect of HIV on hematopoietic progenitor cells and EPO responsiveness. HIV-2 infection of BM progenitor cells in the in vitro setting has been shown previously to inhibit erythyropoiesis directly at the BFU-E and CFU-E stage of differentiation.¹⁹  More recently, in a pediatric cohort in Malawi, BM analysis demonstrated that HIV⁺ children with anemia have a decreased number of both CD34⁺ progenitor cells and primitive erythroid progenitors compared with HIV⁻ children with anemia.²⁰  In an evaluation of EPO signaling in the setting of HIV infection, a small cohort of HIV⁺ patients was monitored before and after cART initiation.²¹  Compared with uninfected controls, serum levels of soluble transferrin receptor and EPO were decreased.²¹  Furthermore, after ex vivo differentiation of peripheral blood CD34⁺ cells, an increase was noted in the early BFU-E stage of erythropoiesis in untreated HIV⁺ patients compared with uninfected controls.²¹  The addition of exogenous EPO to the growth media for these studies suggests that the progenitor cells from the HIV⁺ patients were intrinsically refractory to the growth effects of EPO, resulting in accumulation of colonies at a relatively undifferentiated stage of erythroid development. However, after initiation of cART therapy, circulating levels of EPO and transferrin levels increased in patient serum, and the number of BFU-E colonies normalized compared with uninfected controls in ex vivo studies.²¹  Although this was a small patient cohort, these findings suggest that HIV may have a direct suppressive effect on the cytokine cascade responsible for normal erythropoiesis. The hematopoietic stem cell may even be a reservoir for latent viral infection, although this remains controversial. Several studies have suggested that CD34⁺ cells are resistant to HIV infection,^22-24  but others have identified detectable virus in the CD34⁺ hematopoietic progenitor pool in patients with well-controlled disease, including undetectable serum viral loads.²⁵  A recent study extended these findings to the CD133⁺ cohort within the BM of patients on cART.²⁶ 

Studies evaluating hematopoiesis in patients undergoing HIV treatment also support a direct role for the virus in inducing anemia. In animal studies of HIV infection using glycolipid administration to activate natural killer cells, decreased HIV-1 viral replication was accompanied by restoration of normal hematopoiesis in treated mice.²⁷  At the molecular level, BM from HIV⁺ patients receiving cART but without adequate response—so-called immunologic nonresponders—demonstrated reduced clonogenic ability and a decrease in the number of the more multipotent primitive progenitor cells compared with those with response to therapy.²⁸  Actively replicating virus thus appears to suppress the normal physiologic response to anemia and the complex network of regulatory cytokines that maintain normal hematopoiesis. Intriguingly, the cytokine profile produced by the BM cells from patients without immunologic response to cART was notable for decreased IL-2 and increased TNF-α and IL-7, consistent with a proinflammatory milieu.²⁸ 

The emerging role of an active proinflammatory state in contributing to anemia in patients with HIV suggests a potential connection to the pathophysiology of anemia in the aging population, because this is also a process associated with a shift toward a proinflammatory state. An individual HIV⁺ patient may present with any type of anemia and may well have obvious risk factors such as micronutrient deficiency or use of specific cART medications, but anemia in the HIV⁺ population is most often characterized by a low reticulocyte count, normochromic and normocytic erythrocytes, normal iron stores, and impaired erythropoietin response. These descriptive features are also classically associated with anemia of inflammation (previously known as anemia of chronic disease), a state that is particularly well described in the aging population.²⁹  cART therapy has led to a growing number of aging adults with HIV, thereby providing a growing database of patients in which to examine the relationships between HIV infection, aging, and inflammation.

The overlap between the mechanisms linking the anemia of aging and the anemia of HIV begins with aberrant cytokine expression. It has been established previously that a host of markers associated with inflammation, including IL-1, IL-6, acute phase proteins, and TNF-α, are increased in the aging population, sometimes irrespective of underlying health status.²⁹  In the aging population, elevation in these cytokines has been linked with multiple comorbidities, including anemia. IL-6 may play a particularly significant role in potentiating anemia of inflammation through its role in regulating hepcidin, an acute phase reactant that is a critical regulator of iron stores and iron-limited erythropoiesis and has been linked with anemia of inflammation.²⁹  In a striking overlap with the aging population, proinflammatory cytokines and altered iron metabolism have also been linked to HIV infection. Several independent studies have identified an association between inflammatory biomarkers and morbidity and mortality in HIV⁺ patients, specifically soluble CD14 (a marker of monocyte activation), D-dimer, and IL-6.^2,30  In addition, hepcidin regulation is altered in HIV infection, suggesting an overlap with the mechanisms driving anemia of inflammation. In vitro T-cell cultures have shown that hepcidin induces HIV-1 transcription.³¹  In a cross-sectional study of 200 HIV⁺ women in Rwanda, 6% of the population was noted to have a mutation in the iron transporter ferroportin, rendering it resistant to the negative regulatory effects of hepcidin.³²  The women with the ferroportin mutation had significantly increased rates of some opportunistic infections. These findings provide a possible mechanistic link for the previously discussed hypothesis that HIV infection contributes directly to anemia and thus poor outcomes in HIV⁺ patients. The alterations in iron metabolism and resulting anemia in the context of viral infection continue to be an active area of research investigation.³³ 

Studies evaluating the inflammatory axis in the context of obesity and adipokines secreted by adipose tissue have also demonstrated a noteworthy connection between the aging population and the HIV⁺ population. Leptin, a protein associated with body mass and energy metabolism, is known to induce hepcidin, thus linking inflammation and iron metabolism.²⁹  However, decreased leptin expression is also correlated with impaired response to EPO in the elderly, suggesting that metabolic dysregulation at the molecular level may be associated with anemia in this population.³⁴  Similarly, in the HIV⁺ population, leptin is now emerging as a regulator of anemia. A recent analysis of a large cohort of more than 2000 HIV⁺ patients with matched controls revealed that a specific polymorphism in the leptin gene promoter was associated with the development of anemia in HIV⁺ but not HIV⁻ subjects.³⁵  Body habitus and inflammation have also been linked in patients undergoing cART. In one study, incremental increases in body mass index were associated with increasing levels of C reactive protein and TNF-α, whereas increased expression of IL-6 and migration inhibitory factor were specifically correlated with patients considered to be obese by BMI measurements.³⁶ 

Finally, there is evidence that EPO signaling can be impaired both during aging and in HIV infection, a process that may be accelerated by the inflammation also associated with both processes. EPO levels rise with age, suggesting that, over time, hematopoietic progenitors become less responsive to EPO.²⁹  The role of a proinflammatory milieu in impairing progenitor cell EPO responsiveness or mediating increased EPO expression remains to be determined, but it is noteworthy that in aging patients with comorbidities associated with increased inflammation, the rate of EPO elevation is diminished,³⁷  suggesting that inflammation may play a role in blunting the normal physiologic compensation for aging. As discussed above, HIV infection can have direct effects on decreasing the function of hematopoietic progenitors and specifically their response to EPO.

The Veterans Aging Cohort Study (VACS) has become a particularly useful clinical tool in investigating the overlapping contributions of aging and inflammation to the anemia associated with HIV. The VACS is a well-described patient cohort of 1302 HIV⁺ patients on cART with age/race/site-matched HIV⁻ controls on whom blood and DNA specimens have been banked.³⁸  This patient cohort has been used to develop the VACS index, an evaluation of 7 variables (age, CD4 count, HIV-1 status, hemoglobin, FIB-4 index, hepatitis C infection, and eGFR) used to help predict mortality.³⁹  The VACS index is more predictive of mortality than the Restricted Index, which only considers age, CD4 count, and HIV-1 status.⁴⁰  This improved predictive value based on measures of clinical end points associated with inflammation (anemia, liver injury, kidney injury) supports a central role for inflammation in disease progression caused by HIV.³⁸  In the case of anemia, what is particularly striking is that when analyzing the factors that make up the VACS index, anemia was not only correlated with other markers of inflammation, but was also found to be an independent prognostic factor for mortality.⁴¹  When serum levels of sCD14, IL-6, and D-dimer were evaluated in the VACS cohort and included in mortality predictions, D-dimer and sCD14 further improved predictive accuracy. These findings suggest that IL-6 likely plays a role in contributing to the other parameters already assessed by the VACS index and further support the centrality of inflammatory pathways in mediating adverse outcomes in HIV.³⁸ 

Moving forward with an evaluation of the role of anemia in HIV and connections with the biology of aging and inflammation, several outstanding questions remain and are likely to drive future research efforts. First, several large studies evaluating the epidemiology of anemia in the HIV⁺ population have noted that more significant anemia is often found in women and in those of black race.^12,35  A similar ethnic association has also been noted in studies of the aging population, specifically data reported from the Third National Health and Nutrition Examination Survey (NHANES III) study.⁴²  Whether race or gender-specific genetic differences contribute to differential regulation of inflammatory pathways and thus divergent outcomes in the HIV⁺ population remains to be determined. However, given disparities in HIV outcomes in women and minorities, the answer to this question is of significant clinical and social significance. The VACS cohort offers a particularly valuable clinical context in which to address one aspect of these questions, because it includes a significant number of non-Hispanic African-American patients. However, it is not surprising that the number of women enrolled in this study is small. Additional studies specifically enrolling women may be required to further address the overlap between anemia and inflammation in HIV⁺ patients.

Second, the role of genetic polymorphisms in modulating leptin and hepcidin signaling suggests that polymorphisms in other key mediators of inflammatory pathways may play a role in contributing to outcomes in patients with HIV. Some studies suggest that variation in the balance between pro- and anti-inflammatory cytokines in those exposed to HIV may play a role in determining rates of postexposure infection.⁴³  In addition, variation in inflammatory mediators has already been linked to outcomes in other diseases associated with chronic inflammation.⁴⁴  The correlation between anemia and morbidity and mortality in HIV⁺ patients raises the intriguing possibility that a more targeted genetic evaluation of the inflammatory cytokines linked to anemia may provide further insight into the mechanistic underpinnings of this process and suggest possible targets for therapy. Anemia itself may function as a biomarker for poor outcomes in HIV, but it remains unclear whether it is an end point or part of a more global disruption in the axis of inflammation.

Finally, although effective cART therapy can help to ameliorate anemia in HIV⁺ patients, as discussed above, it is not clear whether specifically targeting pathways of inflammation or aberrant erythropoiesis may also improve morbidity and mortality in HIV, particularly in the subset of aging HIV⁺ patients. To address this question, it will be important to evaluate in more detail not only the parameters of erythropoiesis and iron balance in the HIV⁺ population, but also the changes in these pathways resulting from cART therapy and occurring over time as patients age. It is not clear why some patients remain anemic despite cART therapy, and some data seem to suggest that continued anemia and inflammation after initiation of cART can be used to identify lack of immunologic response. However, more effective treatment of the anemia or inflammation caused by the direct effects of HIV infection may independently improve outcomes. The role of comorbid conditions in modulating outcomes in HIV is being increasingly recognized and may provide additional treatment options to improve quality of life in the HIV⁺ population. However, in the absence of data showing that treatment specifically targeting HIV-related anemia improves other outcomes, indications for such therapy should be similar to those in HIV⁻ patients: that is, addressing nutritional deficiencies and EPO insufficiency as indicated by laboratory data, with transfusion for severe, symptomatic anemia.

The etiology of anemia in HIV is complex and multifactorial. However, emerging evidence suggests that not only is anemia an important and independent prognostic indicator in HIV, but it may also reflect a common biologic end point associated with the increased inflammation and aberrant cytokine regulation seen in other clinical contexts such as aging. Increased expression of inflammatory markers and impaired EPO response are clearly associated with both the aging process and HIV infection, raising the intriguing possibility that at the molecular level, HIV infection may function to accelerate processes that would normally occur more slowly as a consequence of age. A better understanding of the mechanisms driving anemia in both the aging population and the HIV⁺ population will be critical in not only improving quality of life, but also in suggesting ways to reduce mortality in these vulnerable populations. Because cART therapy has allowed more and more HIV⁺ patients to age, the intersection of the biology of aging and the biology of HIV will remain an intriguing area of investigation in the field of hematology.

Conflict-of-interest disclosure: The authors declare no competing financial interests. Off-label drug use: None disclosed.

Nancy Berliner, Hematology, Mid-Campus 3, Brigham and Women's Hospital, Boston, MA 02115; Phone: 617-732-5840; Fax: 617-264-5215; e-mail: nberliner@partners.org.