Chapter 9: Hematology in Older Persons

The world’s population is aging at an unprecedented rate and the implications for health care delivery are profound.^1,2,3,4 Over the next several decades, the percentage of the population older than age 65 years will nearly double.⁵ In anticipation, an increased research effort is being made to better understand the basic biology of aging and the mechanisms whereby individuals become susceptible to disease.^6,7

A central dogma of gerontology is that aging is not a disease. Yet, intrinsic biologic aging is the major risk factor for virtually all major diseases of developed societies, including cancer, diabetes, atherosclerotic cardiovascular and cerebrovascular disease, diabetes, neurodegenerative diseases (e.g., Alzheimer and Parkinson), osteoporosis and infection. Examination of the mechanisms by which biologic aging contributes to the pathogenesis of these diseases has now become recognized in the mainstream of scientific inquiry over a broad range of disciplines.⁷ However, from a biogerontologist’s perspective, there remain certain features of aging that transcend the more discipline-focused investigations.

One of these common features of aging is heterogeneity. For example, for any measureable variable, the range of values among normal older individuals is much wider than the range of normal among younger individuals.^8,9 This variation is particularly relevant for hematologists consulted to examine older, but otherwise healthy, individuals with laboratory values outside the “normal range.”

Although functional declines that accompany normal aging have been well characterized,^10,11 in general these are not of sufficient magnitude to account for symptoms or be mistaken for disease. For example, that kidney function declines with age is well recognized,^12,13 and, in fact, has proven to be a useful biologic marker of aging. Yet, clinical consequences of this change in renal function, in the absence of a disease or the exposure to an exogenous nephrotoxic agent, do not occur commonly. Similarly, the marrow changes with age. Marrow stem cells increase in number and proliferative capacity, yet the in vitro proliferative potential of progenitor cells is less.^14,15,16 Although clinically significant cytopenias do not occur in the absence of disease, mild to moderate anemia that has not been fully characterized occurs with increasing frequency, especially in the frail elderly.^17,18,19,20 Furthermore, in frail individuals even a mild reduction in hemoglobin level is associated with untoward clinical outcomes.^19,21,22

Certain immune functions decrease with age,^23,24,25 but these may be of only marginal clinical significance. For example, whether the laboratory-observed declines in immune function contribute to a heightened susceptibility to infection is a subject of debate, but data support an association of age-associated qualitative change in lymphocyte function and susceptibility to reactivation of tuberculosis^26,27 or herpes zoster^28,29 and diminished response to influenza vaccine.^30,31,32,33 There is compelling evidence that a profound decline in immunity is causally related to certain malignancies,³⁴ but there remains debate over whether the more modest decline associated with normal aging is sufficient to account for the observed increased rate of cancer in the elderly.^35,36 In fact, there is some evidence that cancer incidence is even lower in frail elderly,³⁷ a population known to be functionally immunodeficient.³⁸ Similarly, development of autoantibodies appear with increasing frequency with advancing age, but these are considered markers of an acquired humoral immune dysregulation, and with regard to autoantibody, not of clinical importance.³⁹ In contrast, essential monoclonal gammopathy, a clonal B lymphocyte expansion that increases in frequency with age and that stabilizes at a clone size that does not impair normal immunoglobulin synthesis or inhibit hematopoiesis, does have a probability of undergoing clonal evolution to a B-cell malignancy, such as myeloma, lymphoma, or monoclonal light-chain amyloidosis at a rate of 1 percent per year (Chap. 106).⁴⁰

THEORIES OF AGING

Providing a rational, unifying explanation for the aging process has been the subject of a great number of theoretical expositions. Yet, no single proposal suffices to account for the complexities observed (Table 9–1).

Genetic Effects

That genetic controls are involved seems obvious when one considers that lifespan is highly species-specific. For example, mice generally live approximately 30 months and humans approximately 90 years. However, the aging phenomenon is not necessarily a direct consequence of primary DNA sequence. For example, mice and bats have 0.25 percent difference in their primary DNA sequence, but bats live for 25 years, 10 times longer than mice. Thus, regulation of gene expression seems likely to be the major source of species longevity differences.

Progeria Syndromes

Gerontologists have long been intrigued by the concept of accelerated aging and by examining those rare individuals who are so affected. From work with invertebrate models a number of genes have been identified that associate with longevity. Yet, the identification and functional analysis of analogous genes in humans remains elusive. With regard to genetic examples of accelerated aging, two syndromes have been well characterized: Hutchison-Guilford syndrome (early-onset progeria) and Werner syndrome (adult-onset progeria).^41,42 Although neither these nor other progeria syndromes manifest a complete phenotype of advanced age, the identification of the genes responsible for these particular syndromes is beginning to pay dividends by providing clues to the molecular mechanisms involved in the aging process. For example, Werner syndrome is now defined by mutations in a single gene on chromosome 8 that encodes a protein containing a helicase-like domain.^43,44 The activity of the Werner protein helps to maintain telomere structure and homology-dependent recombination.⁴⁵ Similarly, a mutation in the lamin A (LMNA) gene localized to chromosome 1 has been causally related to the Hutchison-Guilford syndrome.⁴⁶ The product of the mutated LMNA gene (termed progerin) accumulates, producing a variety of nuclear distortions of which telomere dysfunction and associated replicative senescence are notable.⁴⁷

Examination of aging in yeast has also been informative with regard to the genetic controls of aging. These single-cell organisms follow the replicative limits of mammalian cells and it has been observed that “life span” is related to silencing large chromosomal regions. Mutations in these silencing genes lead to increased longevity.⁴⁸

Thus, if there are certain genes that regulate normal aging, or at least are associated with the development of an aged phenotype, it stands to reason that acquired mutations of those genes might influence the rate of aging. Over the years several theories have been proposed that relate to this supposition. In general, they hypothesize a random or stochastic accumulation of damage, either to DNA or protein that leads eventually to dysfunctional cells, cell death and subsequent organ dysfunction, and ultimately death. Prominent among these is the somatic mutation theory,⁴⁹ which predicts that genetic damage from background radiation, for example, accumulates and produces mutations that ultimately result in functional decline. A variety of refinements have been suggested to this theory, invoking the importance of mutational interactions,⁵⁰ transposable elements,⁵¹ and changes in DNA methylation status.⁵²

A related hypothesis is Burnet’s intrinsic mutagenesis theory,⁵³ which proposes that spontaneous or endogenous mutations occur at different rates in different species and that this accounts for the variability observed in life span. Closely related to this notion is the DNA repair theory.⁵⁴ Initially, there was great excitement about this idea as it was found that long-lived animals had demonstrably more active DNA repair mechanisms than shorter-lived species.⁵⁴ However, longitudinal studies within a species have not revealed a consistent decline in repair mechanisms with age. This, of course, does not rule out the possibility that repair of certain specific and critical DNA lesions is altered with advancing age. We now understand that there are multiple DNA repair mechanisms, including base excision repair, transcription-coupled repair, and even mitochondrial DNA repair mechanisms. Disorders involving one or a subset of repair mechanisms could lead to accumulation of DNA damage and dysfunction.

In yet another intrinsic/stochastic model, Orgel⁵⁵ proposed the error catastrophe theory in which he suggested that random errors in protein synthesis occur and when the proteins involved are those responsible for DNA or RNA synthesis, there is resultant DNA damage and the consequences thereof to daughter cells. Although this model has appeal, there has been no reported evidence for impaired or inaccurate protein synthesis machinery with advancing age. However, a candidate protein that may eventually be shown to be so affected is telomerase. This critical enzyme is necessary for maintaining telomere length and cell replicative potential. In vitro cellular senescence is associated with diminished telomerase activity,⁵⁶ but whether this relates to aging of the organism as a whole remains controversial.⁵⁷

Posttranslational Effects

Evidence that exogenous factors are involved in the acquisition of age-associated damage to DNA and protein is derived from a number of observations, many of which are circumstantial or correlative, but nonetheless provocative. It now appears that the accumulation of abnormal protein within senescent cells, as predicted by the error catastrophe theory, actually reflects posttranslational events, such as oxidation or glycation, and resultant crosslinking. There is theoretic appeal to the concept that key proteins, such as collagen or other extracellular matrix proteins, and DNA become dysfunctional with age as a consequence of the impairment produced by these crosslinks.^58,59,60

Glycation

One mechanism producing crosslinks is called glycation, the nonenzymatic reaction of glucose with the amino groups of proteins. Presumably, glycation would occur more readily in the presence of higher serum levels of glucose, and, thus, this theory fits well with the observed, age-associated dysregulation of glucose metabolism and prevalent hyperglycemia in geriatric populations. Of course, these findings also point out the theory’s deficiency as a unifying mechanism, as there is no question that individuals with well-maintained glucose levels throughout their life span will still be subject to the acquired changes typical of aging.

Free Radical Hypothesis

Another mechanism held responsible for crosslinking is the damage produced by free radicals, which forms the basis of the free radical hypothesis initially promoted by Harman.^61,62 This theory offers that aging is the result of DNA and protein damage (e.g., mutagenesis or crosslinking) by atoms or molecules that contain unpaired electrons (free radicals). These highly reactive species are produced as byproducts of a variety of metabolic processes and are normally inhibited by intrinsic cellular antioxidant defense mechanisms. Nitrate-based free radicals are also generated by in vivo processes and another set of nitrogen free-radical scavenging mechanisms are in place. If free radical generation increases with age, or the defense mechanisms that scavenge free radicals (e.g., glutathione) or repair free radical damage decline, the accumulated free radical damage may account for altered DNA and protein function. Evidence to support this widely held notion is incomplete. It is known that free radical generation in mammals correlates inversely with longevity⁶³ and, similarly, the level of free radical inhibiting enzymes, such as superoxide dismutase were higher in those species with longer life spans.⁶³ However, efforts at enhancing antioxidant mechanisms with dietary vitamin E have resulted in only a modest enhancement of median survival in mice and no effect on maximum life span.^64,65,66

Much attention has been focused on mitochondrial function in the context of free radical damage because the bulk of oxidative metabolism and the production of reactive oxygen species occur in these organelles. Although mitochondrial DNA codes for antioxidant enzymes in addition to enzymes involved in energy production, it is currently believed that energy production declines with age as a result of mitochondrial DNA damage by those reactive products. Indeed, mitochondrial damage increases with age in experimental models,^67,68,69 and the shortened survival of knockout mice deficient in mitochondrial antioxidant enzymes has supported the potential importance of this mechanism.⁷⁰

The most compelling data to date in support of the free radical hypothesis come from experiments in which transgenic Drosophila producing enhanced levels of superoxide dismutase and catalase had a maximum survival 33 percent greater than controls.²⁶ Furthermore, it is known that flies produce high levels of free radicals associated with their impressive metabolic requirements, and that survival is enhanced dramatically when the ability to fly is experimentally hindered.⁶⁵ However, the generalizability of these findings has been questioned. It has been noted that transgenic mice overexpressing free radical scavenging enzymes have produced very modest effects on life span.⁷¹ Thus, the conclusion that augmentation of free radical scavenging mechanisms increases longevity in mammalian species is not established.

Neuroendocrine Theory

From a different perspective, very good evidence implicates a nonrandom, perhaps genetically regulated endogenous mechanism involved in aging. For example, the neuroendocrine theory suggests that the decrements in neuronal and associated hormonal function are central to aging. It has been suggested that age-associated decline of hypothalamic–pituitary–adrenal axis function results in a physiologic cascade leading, ultimately, to the “frail” phenotype. This hypothesis is appealing because it is well established that this neuroendocrine axis regulates much of development and also the involution of ovarian and testicular function. Furthermore, age-associated declines in growth hormone and related factors,⁷² dehydroepiandrosterone,⁷³ and secondary sex steroids⁷⁴ are implicated in age-associated impairments, including a reduction in lean body mass and bone density. Furthermore, pharmacologic reconstitution using these or related hormones has met with some success at reversing age-associated functional decline.^75,76

Immunologic Theory

Similarly, it has been argued that involution of the thymus gland and subsequent decline in immune function discussed below is a key regulator of aging.⁷⁷ The argument is based upon the observation that the decline in immune function occurs in all mammalian species, but occurs later in those with longer survival.⁷⁸ Furthermore, dietary restriction is associated with maintained thymic mass and measurable immune function as well as prolonged survival, suggesting an association of a decline in immunity with primary aging processes.

The possibility is highlighted by the observation that differences in maximum survival of different mouse strains has been associated with specific alleles in the major histocompatibility complex, which, in turn, code for immunologic determinants.⁷⁹ This hypothesis, although not without its appeal, is not widely accepted as a major explanation for aging. Perhaps this relates to the fact that biologic aging is a universal phenomenon and certain features are held in common, even in organisms with primitive or no immune function. (The same could also be said for the neuroendocrine theory.) It is obvious that the immune system is of great importance in minimizing the chance of early death, particularly from infectious diseases. However, immunologic reconstitution of middle aged or old animals has not been shown to prolong survival.⁸⁰

LIFE SPAN: MEDIAN AND MAXIMUM SURVIVAL

From the perspective of those who study aging, an important distinction is made between median (life expectancy) and maximum life span. Over the past century, a dramatic increase in median survival has been mostly attributable to modern sanitation and refrigeration, as well as public health measures, including vaccination and antibiotics.⁸¹ Early deaths have been diminished and more individuals are reaching old age. In the United States today, expected survival from birth is approximately 80 years.⁸² Median survival is what concerns public health officials and healthcare providers. In contrast, maximum survival is the focus of those gerontologists interested in the biology of aging and longevity.

The oldest human being alive today is approximately 120 years old. It is intriguing that the oldest age limit has remained stable, unchanged by the public health initiatives mentioned above. In the laboratory, limits on age have been established for a variety of species. Drosophila, free of predators, disease or fly swatters, can live 30 days, whereas C57BL/6 mice in a laboratory environment and allowed to eat a healthy diet ad libitum, may survive 40 months. Unlike health-related interventions in humans, certain experimental interventions in lower species are associated with a prolongation of maximum survival. In Drosophila, for example, transgenic offspring producing extra copies of the free radical scavenging enzymes superoxide dismutase and catalase survive approximately 33 percent longer than controls.⁸³ In nutritional intervention studies involving lower species, controlled restriction of dietary intake (dietary restriction) has become a common experimental paradigm exploited in the investigation of primary processes of aging and maximum survival.^84,85

Dietary Restriction

Dietary restriction typically involves a reduction of 30 to 40 percent in caloric intake with careful attention to the provision of adequate amounts of essential nutrients. It is associated with both a delay in the acquisition of age-related diseases (including cancer) and a reduction in the rate of achieving certain established biomarkers of aging (i.e., a retardation in primary aging). The critical questions remain: What is the mechanism of the effect of dietary restriction, and will it be applicable to higher species? With regard to the latter, there are now comprehensive and interactive studies within the United States in which dietary restriction is being examined in nonhuman primates^86,87 and human studies are also underway.^88,89 Although it appears that the calorie restricted monkeys in these studies are assuming a more youthful phenotype in a variety of physiologic measures,^86,90,91 it remains too early to predict whether maximum survival will be affected.

CELLULAR SENESCENCE AND ORGANISMAL AGING

After a finite number of divisions, normal somatic cells invariably enter a state of irreversibly arrested growth, a process termed replicative senescence.⁹² In fact, it has been proposed that escape from the regulators of senescence is what oncologists term malignant transformation. However, the role of replicative senescence as an explanation of organismal aging remains the subject of vigorous debate (for review, see references 93 and 94). The controversy relates, in part, to the fact that certain organisms (e.g., Drosophila, Cunninghamella elegans) undergo an aging process, yet all of their adult cells are postreplicative.

What is clear is that the loss of proliferative capacity of human cells in culture is intrinsic to the cells and not dependent on environmental factors or even culture conditions.⁹² Unless transformation occurs, cells age with each successive division. The number of divisions turns out to be more important than the actual amount of time passed. Thus, cells held in a quiescent state for months, when allowed back into a proliferative environment, will continue approximately the same number of divisions as those that were allowed to proliferate without a quiescent period.⁹⁵ The question remains whether this in vitro phenomenon is relevant to animal aging.⁹⁵ Although when various species are compared, replicative potential is directly and significantly related to life span,⁹⁶ within an organism there is great variability in proliferative capacity from tissue to tissue and organ to organ. As such, age-associated changes in the marrow or gut might relate to replicative senescence, whereas in muscle or brain other processes most certainly are involved. But, added to this heterogeneity within an individual, is that fact that certain commonly employed models of aging (e.g., Drosophila, C. elegans) undergo an aging process despite a long held belief that all of their adult cells were post replicative.^94,97,98 However, this notion has been countered by the demonstration of multipotent intestinal stem cells within the midgut near the intestinal basement membrane of Drosophila.⁹⁹ Unlike intestinal stem cells in vertebrates that interact with stromal cells within a niche, analogous cells within Drosophila reside on the surface of the basement membrane and interact directly with daughter cells. Nevertheless, the presence of such cells has rekindled an interest in Drosophila as a model for stem cell biology, cancer, and whole-animal aging.^100,101

CELLULAR SENESCENCE AND CANCER

A feature of cellular senescence is diminished proliferative capacity. In fact, it is now understood that genes considered tumor suppressors (e.g., p53, RB [retinoblastoma gene]) prevent cancer by inducing programmed cell death (apoptosis), particularly in cells at risk for neoplastic transformation. Alternatively, they can prevent potential cancer cells from proliferating by inducing permanent withdrawal from the cell cycle (cellular senescence). Although little is known about how cells choose between apoptotic and senescence responses, both are crucial for suppressing cancer^102,103 and both are highly relevant to functional decline and longevity.¹⁰⁴

There are a multitude of oncogenic stimuli that may result in either cancerous transformation or cellular senescence.^105,106 Epigenetic changes within chromatin, histones, or nucleic acids may be caused by pharmacologic agents or altered expression of proteins.^107,108,109 Such changes can alter the expression of protooncogenes or tumor-suppressor genes and are a frequent occurrence among malignant tumors. Thus the senescence response aborts the uncontrolled proliferative response an assortment of potentially oncogenic stimuli.

Although diverse stimuli can induce a senescence response, they appear to converge on one or both of the two pathways that establish and maintain the senescence growth arrest. These pathways are governed by the gatekeeper tumor-suppressor proteins p53 and pRB.^104,110,111 Furthermore, the senescence response to dysfunctional telomeres requires the integrity of the p53 pathway.^112,113 Overexpression of the RAS gene may also trigger a p53-dependent damage response by producing high levels of reactive oxygen species.^113,114,115 However, oncogenic RAS can also induce p16, an activator of the pRB pathways, which provides a second barrier to the proliferation of potentially oncogenic cells. There is an emerging consensus that senescence occurs through one pathway or the other, with the p53 pathway mediating senescence primarily as a result of telomere dysfunction and DNA damage and p16/pRB pathway–mediating senescence primarily as a result of oncogenes, chromatin disruption, and various stresses.

A more speculative, but potentially important consequence of cellular senescence may be its impact on stem cells.¹¹⁶ Embryonic stem cells, whether human or rodent, express a high level of telomerase and thus are considered resistant to replicative senescence.^117,118 However, mammalian adult stem cells or progenitor cells do not proliferate indefinitely.^{119,120,121,122} The ability of stem cells to undergo senescence and apoptosis is likely to be an important mechanism for preventing cancer.^123,124

Aging is a universal phenomenon that affects all normal cells, tissues, organ systems, and organisms. Accordingly, the marrow undergoes changes with age. Age-related hematologic changes are reflected by a decline in marrow cellularity, an increased risk of clonal myeloid neoplams¹²⁵ and anemia,^{17,126,127,128,129,130} and a decline in adaptive immunity.^{131,132,133,134}

MARROW: ANATOMIC CHANGES

The percentage of marrow space occupied by the hematopoietic tissue declines from 90 percent at birth to a level of approximately 50 percent at age 30 years and 30 percent at age 70 years.^135,136 A similar change occurs in the thymus, where involution begins at an earlier age and is reflected anatomically by a reduction in lymphoid mass with an increase in fat¹³⁷ and functionally by a steady decrease in the production of naïve T cells.^80,138 Fat infiltration into the marrow and thymus results in a diminished volume of hematopoietic tissue.

Although age-related change in the marrow is well described, the exact mechanisms that regulate these changes remains speculative.¹³⁹ For example, it remains unclear whether the age-associated expansion of marrow fat is a cause or an effect of aging and whether the changes seen in marrow and histologically similar changes within the thymus are intrinsically related. Because of the intricate association of hematologic and immune functions and these common histologic patterns of change with age, both changes in blood and innate immunity are discussed below in the sections on Blood Cell Changes with Age and Aging and Immunity.

MARROW: STEM CELLS

The ontogeny of hematopoietic stem cells is the focus of much attention.^140,141 In fetal development the manufacture of blood cells occurs in several organs, but after birth this function is subsumed by the marrow.^{142,143,144,145,146,147,148} The process of embryonic and fetal hematopoiesis is described in Chap. 7. Hematopoietic cells appear in the medullary cavities of bone around 14 weeks of gestation,¹⁴⁹ and by birth the marrow is the primary site of hematopoiesis.

Unlike the commonly held notion that stem cell compartments diminish either in number or function with age ultimately resulting in an inability to meet homeostatic demands, age-related hematopoietic stem cell (HSC) changes appear to be an exception, at least for murine species in which this question has been most directly addressed.^150,151 Early work demonstrated that marrow serially transplanted could reconstitute hematopoietic function for an estimated 15 to 20 life spans.¹⁵² Furthermore, the capacity for old marrow to reconstitute proved superior to that of young.¹⁵³ Subsequently, a number of investigators using a variety of techniques have concluded that HSC frequency in old mice and humans is approximately 2 to 10 times greater than in the young.^{16,150,151,154,155,156} Some evidence suggests that the intrinsic function of HSCs changes somewhat with age, most notably in a shift in lineage potential from lymphoid to myeloid development. This may contribute to an observed relative increase in neutrophils and decrease in lymphocytes in the blood of older people.¹⁵⁷

There is an intrinsic change in HSCs with age, most notably resulting in a shift in lineage potential from lymphoid to myeloid development. This may contribute to a relative increase in neutrophils and decrease in lymphocytes in the blood of older persons.¹⁵⁷ As HSCs age, they accumulate genotypic (mutational) and phenotypic alterations. Indeed, human stem-progenitor cells from healthy volunteers were found to accumulate 13 exonic (private) mutations per year of age.^157a Current opinion is that such changes are responsible for the development of immune senescence and that such changes are responsible for the development of immune senescence, as well as the increased occurrence of age-associated diseases such as myelodysplasia and leukemia. Thus, the process of “immunosenescence,” as it affects the innate and adaptive immune system, may result from HSC aging. For example, an age-related decrease in the provision of B-cell precursors may be the result of HSC aging.¹⁴⁸

MARROW DURING ADULT LIFE

The most apparent change seen in the marrow with aging is decreased cellularity (Fig. 9–1).¹³⁵ Under normal circumstances, the marrow is the only site of hematopoiesis. Foci of extramedullary hematopoiesis may occur in the liver, spleen, or lymph nodes in pathologic states, but they are not of functional consequence. Until puberty the entire skeleton remains hematopoietically active, but by age 18 years only the vertebrae, ribs, sternum, skull, pelvis, proximal epiphyseal regions of humerus and femur remain active sites of blood production, with other medullary sites infiltrated with fatty tissue. By age 40 years, the marrow in sternum, ribs, pelvis and vertebrae is composed of equal amounts of hematopoietic tissue and fat and cellularity declines gradually thereafter. By age 65 years, marrow cellularity is estimated to be approximately 30 percent,^135,136 with a corresponding increase in marrow fat. Age-associated imbalanced bone remodeling and osteoporosis results in decreased trabecular bone which itself may contribute to diminished hematopoiesis.¹⁵⁸ The presence of fat correlates with the occurrence and severity of osteoporosis, both of which are evident with aging.¹⁵⁹ Several age-related qualitative changes have been identified in hematopoietic cells, including skewed X-chromosome inactivation, telomere shortening,^160,161,162 accumulation of mitochondrial DNA mutations,^163,164 and micronuclei formation,¹⁶⁵ any of which could result in cellular dysfunction. Furthermore, growth hormone production declines with age, and this, too, is linked with deposition of fat within the marrow.¹⁶⁶ Administration of growth hormone to old rats reduces marrow fat and increases hematopoietic tissue.¹⁶⁷

BLOOD CELL CHANGES WITH AGE

Red Cells

Anemia is a significant health problem in the elderly because of a high prevalence and significant associated morbidity, including reduced quality of life, clinical depression, falls, functional impairment, slower walking speed, reduced grip strength, loss of mobility, worsening comorbidities, and mortality.^168,169

In older men and women, anemia defined using the World Health Organization (WHO) criteria of hemoglobin levels less than 13 g/dL for men and 12 g/dL for women¹⁷⁰ is associated with an increase in mortality.^{171,172,173,174,175,176} It has been pointed out that the WHO criteria do not take into account inherent ethnic variations, particularly with respect to Americans of African descent who have lower levels of hemoglobin without significant adverse outcomes.^177,178 In a study that analyzed 1018 Americans of African descent and 1583 Americans of European descent adults aged 71 to 82 years, anemia defined by the WHO criteria was associated with increased mortality in those of European descent but not those of African descent.^177,178 The reasons for these ethnic differences are undefined. However, the difference is one of degree. In general, the impact of anemia on functional status and mortality in Americans of African descent becomes apparent at hemoglobin levels approximately 1 g/dL lower than in whites. The issue of establishing criteria for the diagnosis of anemia is relevant in the context of age, as well. Older women, for example, have better physical performance and function at hemoglobin values between 13 and 15 g/dL than at between 12.0 and 12.9 g/dL,¹⁷⁹ suggesting perhaps that the cut off level of 12 g/dL is too low. Nevertheless, the WHO definition remains the standard used in most current epidemiologic surveys and many clinical laboratories.

Prevalence of Anemia

In the third National Health and Nutrition Examination Survey (NHANES III) database, a nationally representative sample of community-dwelling persons and determined age- and sex-specific prevalence rates of anemia in the total U.S. population,¹²⁷ of those individuals older than age 65 years approximately 11 percent were anemic by WHO criteria (see Table 9–1). The prevalence of anemia was lowest (1.5%) among males between 17 and 49 years of age and highest (26.1%) in males older than 85 years. Among those 65 years and older, the prevalence rate was notably higher in Americans of African descent as compared to Americans of European descent and Americans of Hispanic descent. Prevalence rates of anemia in the elderly vary in community-dwelling and institutionalized populations. Also, anemia is more common among frail elderly. In the nursing home, for example, anemia prevalence approaches 50 percent or higher.^{126,180,181,182}

Unexplained Anemia

Hematologists are usually successful in uncovering the cause of anemia in young adults. However, in older populations a specific explanation cannot be defined by routine evaluation in approximately one-third of anemic patients (Table 9–2).^20,183,184 Typically, this anemia is mild (hemoglobin concentration in the 10–12 g/dL range), normocytic, and hypoproliferative (low reticulocyte index). It has been postulated that the cause relates to a number of factors including declining testosterone level,¹⁸⁵ occult inflammation,¹⁸⁶ impaired renal function with inappropriately low serum erythropoietin,¹⁸⁷ or incipient myelodysplasia.¹⁸⁸ Likely, unexplained anemia represents an amalgam of these and perhaps other factors, such as shortened red cell survival, refractoriness of the erythroid precursors to erythropoietin stimulation, and/or the presence of as yet undiagnosed illness.

Serum Erythropoietin and Aging

Data on erythropoietin levels in nonanemic older persons are inconsistent. Some suggest that nonanemic older persons have higher erythropoietin levels compared to younger adults,^189,190,191 but other studies fail to confirm these findings.^10,11,12 One longitudinal analysis clearly demonstrates that serum erythropoietin levels rose gradually in healthy individuals who maintained normal hemoglobin levels but the rise was not observed in those who were to develop diabetes or hypertension during the evaluation period.¹⁹² An explanation for the rise in serum erythropoietin with age is not established, but in theory, it could be the result of age-associated shortened red cell survival or reduced sensitivity of erythroid progenitor cells to the erythropoietin signal. Studies in older subjects are ongoing to define the basis for the increasing need for erythropoietin to maintain normal levels of red cells.

Leukocytes

Although no significant change is seen in the blood leukocyte count with normal aging,^193,194 among those who acquire features of frailty, an increased neutrophil count may be observed.^157,195 Furthermore, several qualitative neutrophil defects have been described. For example, a decreased respiratory burst response to soluble signals,¹⁹³ defective phagocytosis,¹⁹⁴ and impaired neutrophil migration to sites of stress¹⁹⁶ have been described in accordance with advanced age. Although the exact cause for these functional changes has not been clarified, it may be associated with an age-related alteration in actin cytoskeleton and receptor expression in leukocytes.¹⁹⁷ A mild decrease in the blood lymphocyte count is first noticeable in the fourth decade with a gradually progressive decrease thereafter throughout the remainder of the lifespan.¹⁹⁸ Qualitative alterations in T-lymphocyte function in the elderly have also been demonstrated.¹⁹⁹

Platelets

At present, knowledge about the influence of age on platelet counts has been limited to cross-sectional data derived from selected populations. From those data, no or very limited changes in platelet number are noted with age.^{200,201,202,203} To date, a longitudinal data set describing alterations in platelet number with advancing age has not been produced nor are there conclusive studies describing age-associated changes in platelet function.

AGING AND COAGULATION

Coagulation Factors and Aging

A number of proteins critical to clot formation and fibrinolysis change in characteristic ways with advancing age.^204,205,206 Plasma concentrations of factor VII coagulant activity and antigen,^{204,205,206,207,208} and factor VIIIC,^191,206,209 as well as von Willebrand factor,^191,209 fibrinogen,^{191,206,208,210} fibrinopeptide A,^191,206,207 and tissue plasminogen activator antigen^{191,211,212,213} increase with age. In healthy centenarians, levels of activated factor VII, activation peptides of prothrombin, factors IX and X, and thrombin–antithrombin complex concentration were increased, which are signs of higher-than-expected coagulation enzyme activity.²⁰⁶ Age-associated increases in levels of protein C occur in both sexes. Aging is also associated with increasing levels of free protein S.²⁰⁴ In contrast, antithrombin tends to decrease with age in males and increases with age in females following menopause.²¹⁴ Higher D-dimer and plasmin–antiplasmin complexes indicate an accompanying increase in fibrinolytic activity.^206,215 In contrast, plasma tissue-plasminogen activator inhibitor levels increase with increasing age, as do levels of thrombin-activatable fibrinolysis inhibitor in women²¹⁶ and its proenzyme form, procarboxypeptidase U, in both sexes.²¹⁷ These latter findings are suggestive of a possible age-dependent compromise in fibrinolytic activity.²¹⁸ Thus, procoagulant and, in some studies, fibrinolytic activities appear to be increased in older subjects by both in vitro^206,219,220 and in vivo studies,^190,221 but the changes in fibrinolytic activity are inconsistent. Older patients may show an exaggerated anticoagulant response to warfarin.²²²

Aging as a Prothrombotic State

Activation of the coagulation system and increase in procoagulant markers are associated with the pathogenesis of atherosclerosis.^223,224 However, procoagulant markers, most notably D-dimer,²²⁵ fibrinogen, and factor VIII,²²⁶ also increase with advancing age, and may, in fact, correlate better with aging than with cardiovascular disease.^223,224 In a study examining 1729 participants age 70 years and older in the Established Populations for the Epidemiological Study of the Elderly (EPESE) cohort, increasing age was associated with high D-dimer levels. For example, 23 percent of the participants age 90 to 99 years had high D-dimer levels (>600 mcg/L) compared to 13 percent in the 80- to 89-year-old age group and 7 percent in the 70- to 79-year-old age group.²¹⁵ Investigators measured fibrinogen concentrations in healthy subjects ages 19 to 96 years and found levels to be significantly higher in participants older than age 60 years when compared to younger subjects.²²⁷ Healthy individuals across the life span had fibrinogen levels increased by 25 mg/dL per decade of life, and levels as high as 320 mg/dL were found in more than 80 percent of people older than 65 years of age.²²⁸ Other markers of activated coagulation, such as plasminogen activating inhibitor-I (PAI-1) and factor VIII also increased with age.^209,229,230 Thus, it is now apparent that aging is associated with markers of activated coagulation. In this context, it is notable that the incidence of venous thrombosis and pulmonary emboli increases dramatically in geriatric populations.^231,232 Bleeding complications from anticoagulation therapy are also increased in older patients. No interventional study has identified an at-risk population of normal-age subjects without prior thrombosis in whom prophylactic anticoagulation is of value.

Coagulation and Functional Decline

In the EPESE study, increases in D-dimer and interleukin (IL)-6 were related to increases in both morbidity and mortality.²¹⁵ In fact, the correlation for adverse outcomes was stronger with D-dimer than IL-6.²³³ In this, and other studies,^234,235,236 D-dimer and other markers of activated coagulation were associated with limitation in a wide variety of functional domains, including independent activities of daily living (IADL), lower-extremity function, and performance on cognitive testing. The age-associated changes in coagulation markers occur earlier than other aging biomarkers, and hence it has been argued that they could be early predictors of those elderly at increased risk for functional decline.²³⁷

This age-associated prominence of coagulation factors has also been reproduced in animals. For example, when stress associated with physical restraint was compared in aged versus young C57BL/6J mice, significantly increased expression of PAI-1 mRNA was noted in almost all the tissues in older mice.²³⁸ Similar results were seen for expression of tissue factor mRNA in aged mice.²³⁸ In both these experimental models an increase in microthrombi was noted with clots distributed through multiple organ systems in the older mice.

In humans, both the presence of depression and/or psychological stress are associated with increased coagulation^239,240,241 and decreased fibrinolytic activity.²⁴² In elderly subjects without cardiovascular disease, physical exhaustion, a characteristic frequently used to distinguish frail from nonfrail individuals, was associated with significant increases in both inflammatory and coagulation factors as assessed by fibrinogen, C-reactive protein, and white cell levels.²³⁹ Frail and prefrail subjects from the Cardiovascular Health Study had significantly higher levels of fibrinogen, factor VIII, and D-dimer levels as compared to the nonfrail group. The association with frailty persisted even after adjusting for the presence of cardiovascular disease and diabetes.²⁴³ Frailty is also associated with increased risk of venous thromboembolism when compared to nonfrail individuals of the same age, especially in association with increased factor VIII levels.²⁴⁴

AGING AND IMMUNITY

Whether related to primary processes of aging or not, the thymus gland undergoes a very characteristic pattern of involution beginning well in advance of other phenotypic changes attributed to aging (see Figs. 9–1 and 9–2; Chap. 6).²⁴⁵ Among the consequences are a decreased generation of naïve T cells.²⁴⁶ Despite this, the total lymphocyte count does not decline greatly because peripheral T cells are capable of expanding to fill the T-cell niche in the absence of generation of new T cells. However, when they do so, the repertoire for antigen recognition becomes less comprehensive. Thymic involution may result from the aging T-cell progenitor population,²⁴⁷ from the defects in rearrangement of T-cell receptor β genes,^248,249 from loss of self-peptide expressing thymic epithelium,²⁵⁰ and/or from the loss of thymic trophic cytokines.²⁵¹ Thymic epithelial cells produce a variety of colony-stimulating factors and hematopoietic cytokines, such as IL-1, IL-3, IL-6, IL-7, transforming growth factor-β, oncostatin M, and leukemia inhibitory factor,^252,253,254 which influence the complex process of T-cell production. It is proposed that thymic atrophy and decreased thymopoiesis is an active process and mediated by the upregulation of thymosuppressive cytokines (leukemia inhibitory factor, IL-6, and oncostatin M), which results in the altered peripheral lymphatic tissue T-lymphocyte function with aging.²⁵⁵

There is a notable shift in the overall blood T-cell population toward lymphocytes with memory T-cell markers,²⁵⁶ and many of these are thought to have attained replicative senescence.²⁵⁷ With the decreasing numbers of naïve T cells in the peripheral lymphatic tissue and increasing memory T cells reaching senescence, elderly persons have difficulties responding to old and new antigens and demonstrate impaired reactions to vaccinations. The decreased efficacy of vaccines may also be a result of alterations in antigen presentation with age. Within the T-helper cell fraction there is a shift to the T-helper (Th) type 2 subset and away from Th1,²⁵⁸ thereby influencing cytokine production and overall immune response.

In addition to the anatomic changes within marrow and thymus (see Figs. 9–1 and 9–2), similar age-associated morphologic changes within the paracortical and medullary zones of secondary lymphoid tissues (spleen, and lymph nodes) occur, including a decline in the paracortical and medullary zones and increased deposition of fat within the germinal centers.^259,260 It remains unclear to what extent these changes contribute to the overall change in immune function with age.

A wide range of lymphocyte functional changes have been described in the context of aging; however, cataloguing these would be beyond the scope of this chapter. Such changes are detailed in several excellent reviews.^{24,261,262,263,264} Briefly stated, there is a shift in the T-cell population toward memory T cells,²⁵⁶ which attain replicative senescence in response to repeated antigen exposures.²⁵⁷ With the relative and absolute decrease in numbers of naïve T cells in the peripheral lymphatic tissue and the accumulation of functionally diminished memory senescent T cells, primary and secondary immune responses are reduced in elderly persons.

Coincident with the age-related changes in lymphocyte function is the increase in levels of circulating proinflammatory cytokines, measureable in some, even in the absence of definable inflammatory disease. IL-6 is the prototype in this regard. In young adults, expression of IL-6 is tightly regulated and serum levels are usually unmeasurable or very low in the absence of inflammatory conditions. Animal studies reveal an increased production of IL-6^265,266 from mononuclear cells and lymphoid cells after stimulation with lipopolysaccharide or other mitogens. Similarly, in humans serum IL-6 levels increase significantly with age.^{267,268,269,270,271} Other inflammatory proteins, including tumor necrosis factor-α (TNF-α) and C-reactive protein are also seen at higher levels in the elderly.^272,273,274 Visceral adipose tissue from older mice express greater levels of both IL-6 and TNF-α mRNA than tissue from younger mice,²⁷⁵ and thus, some of the age-associated rise in IL-6 may be the consequence of those metabolic shifts mentioned above.

CLINICAL CONSEQUENCES

Marrow Aging

Although a number of measureable changes occur in the marrow, not the least of which is a reduction in cellularity, apparent compensatory stem cell changes allow the sustenance of normal or near-normal blood counts throughout the life span. It is notable from transplant experience that even when marrow is donated from a 65-year-old person to an human leukocyte antigen (HLA)-matched younger recipient, the donor marrow supports hematopoiesis for the life of the recipient, although allogeneic marrow from older donors has a greater chance of being associated with graft-versus-host disease.²⁷⁶

Unexplained Anemia

To the extent that marrow contains continuously repopulating cell lines, it is quite remarkable that changes attributable to aging alone (i.e., in the absence of disease) are quite subtle. Nonetheless “unexplained anemia” (UA) accounts for up to one-third of cases of anemia in older patients and its frequency increases with advancing age.²⁰

UA is usually mild, with hemoglobin levels approximately 1 g/dL lower than the WHO standard. The red cells are typically of normal size and examination of the blood film reveals no evidence for intravascular destruction or morphologic features suggestive of myelodysplasia. Although inflammatory cytokine levels may be elevated, the intensity of inflammation is insufficient to produce increased levels of hepcidin; thus, UA has a distinct pathogenesis from the anemia of chronic disease (Chap. 37). Because UA is typically mild, it is likely to be overlooked. In fact, in one population-based cohort that included elderly patients with even more significant anemia, the medical records of affected individuals did not mention anemia as a problem in 75 percent of the cases.¹⁷¹ However, there is now evidence that this casual acceptance of lower hemoglobin levels in older populations may not be advisable.¹⁶⁸ Not only can a decline in important functional measures be related to mild anemia,^{172,277,278,279,280} but longitudinal studies demonstrate increased mortality among individuals with even mild anemia.^174,179,281 Furthermore, a retrospective cohort study of the U.S. Veterans Administration National Surgical Quality Improvement database, indicated that of 310,311 subjects age 65 years and older who underwent noncardiac surgery, the 30-day mortality and cardiac event rates increased by 1.6 percent for each 1 percent decrease in hematocrit below the level of 39 percent.²⁷⁵ Thus, although in younger individuals mild anemia may be well tolerated, in many older individuals it is associated with important negative consequences. That stated, it remains to be established whether the correction of anemia for those with UA will result in improved quality of life, physical function, or survival.

For elderly patients with UA, the presence of macrocytosis; thrombocytopenia; neutropenia; splenomegaly; or unexplained constitutional findings of fever, chills, or weight loss; or symptoms of early satiety; or bone pain should prompt consideration of a marrow examination to rule out myelodysplasia or other diseases affecting marrow function, most of which occur with increasing frequency with advancing age.

Immune Senescence

The complex alterations in immune function with age have been described comprehensively in several reviews.^{261,262,263,264} The changes may explain an age-associated predisposition to certain infections (herpes zoster, tuberculosis reactivation) and perhaps a failure to mount a sufficient vaccine response (e.g., influenza hemagglutinin^{282,283,284,285,286}). The more profound immune deficiency commonly observed in older people most often reflects the debilitating effects of concurrent diseases, most of which occur more commonly with age, and side effects of the medicines used to manage those diseases.

Inflammation/Coagulation Dysregulation and Frailty

Presumably on the basis of chronic inflammatory stimuli, there is an age-associated activation of coagulation²³² and fibrinolytic²⁸⁷ pathways that favor thrombus formation. Fibrinogen levels are typically high with more than 80 percent of those age 65 years and older having levels above 320 mg/dL.²²⁷ Similarly an analysis of D-dimer levels in the EPESE, including 1727 community elderly, revealed an age-associated increase that correlated with declining overall physical function.²¹⁵ Furthermore, when combining D-dimer and IL-6 levels, those individuals who had elevations of both were at greatest risk for mortality over a 4-year interval.²³³ In the Cardiovascular Health Study, which included relatively healthy elderly, higher fibrinogen and factor VIII levels were associated with a greater risk for cardiovascular disease and mortality, even after adjustment for other cardiovascular risk factors.^232,288 Summarizing what has now become a robust literature, higher IL-6, TNF-α, D-dimer, and C-reactive protein are each associated with negative physiologic consequences, including reduced lower-extremity muscle mass and strength,^289,290 cognitive decline,²⁹¹ insulin resistance,²⁹² subclinical and clinical cardiovascular disease,^293,294 renal insufficiency,²⁹⁵ loss of bone mineral density,²⁹⁶ depression,²⁹⁷ anemia,¹⁸⁶ dementia,²³⁵ and mortality.²⁸⁹ As a result, a general consensus has emerged that activated inflammatory mediators are, at least in part, contributing to the physiology of aging, and to the extent that these pathways are dysregulated, important functional outcomes are impaired.