Sci. Aging Knowl. Environ., 3 August 2005
Vol. 2005, Issue 31, p. pe24
[DOI: 10.1126/sageke.2005.31.pe24]


From Bedside to Bench: Research Agenda for Frailty

Linda P. Fried, Evan C. Hadley, Jeremy D. Walston, Anne B. Newman, Jack M. Guralnik, Stephanie Studenski, Tamara B. Harris, William B. Ershler, and Luigi Ferrucci

The authors are at the Center on Aging and Health and the Division of Geriatric Medicine and Gerontology at the Johns Hopkins University School of Medicine, Baltimore, MD 21224, USA (L.P.F. and J.D.W.), the National Institute on Aging, Bethesda, MD 20892, USA (E.C.H.), the Division of Geriatric Medicine at the University of Pittsburgh, Pittsburgh, PA 15213, USA (A.N. and S.S.), the Laboratory of Epidemiology, Demography, and Biometry at the National Institute on Aging, Bethesda, MD 20814, USA (J.M.G. and T.B.H.), the Institute for Advanced Studies in Aging and Geriatric Medicine, Washington, DC 20007, USA (W.B.E.), and the Clinical Research Branch of the National Institute on Aging, Baltimore, MD 21225, USA (L.F.). E-mail: (L.P.F.)

Key Words: frailty • inflammatory cytokines • stressors • loss of complexity • loss of redundancy • systems biology

Introduction: Conceptual Approaches to Frailty

A working conference titled "Research Agenda for Frailty in Older Adults: Towards a better understanding of physiology and etiology" was held in Baltimore in January 2004. It was sponsored by the American Geriatrics Society and funded by the National Institute on Aging. This Perspective summarizes key observations about frailty discussed at the meeting and proposes an agenda for future research in this area.

One reason for the appeal of the term "frailty" to researchers in geriatrics may be that it is thought to encompass a variety of important phenotypes that are not fully captured by current definitions of disease or disability. "Frail" has a variety of clinically relevant vernacular meanings, e.g., easily broken or destroyed, likely to fail or die quickly, unusually susceptible to disease or other infirmity, lacking normal strength or force, weak, tenuous, thin, and slight. There is a wide range of common frailty-associated phenotypes in geriatrics to which such terms are applied, including muscle weakness, bone fragility, very low body mass index, susceptibility to falling, vulnerability to trauma, vulnerability to infection, high risk for delirium, blood pressure instability, severely diminished physical capabilities, and others. Research on such traits is ongoing to determine their causes, including the potential role of aging-related changes, and possible common underlying processes that may contribute to multiple traits. Whether these different frailty-associated conditions are markers, correlates, or components of a central frailty process are subject to discussion.

As discussed in a recent editorial (1), differing concepts and definitions of frailty that are in use include a variety of domains such as physical characteristics and function, cognitive function, other psychological characteristics, and psychosocial factors. Researchers have applied differing conceptual approaches regarding these phenomena and methods to understand their causes. However, some investigators think that there is evolving consensus that frailty is a definable clinical state that may well involve definable etiology. Understanding this etiology, as well as having clarity on the phenotype(s), would lay the basis for defining opportunities for prevention and for treatment.

Frailty as a clinical state

A central definition of frailty in geriatric medicine is that it is a clinical state of vulnerability to stressors (2). Some researchers postulate that such a state of frailty results from aging-associated declines in resiliency and physiologic reserves and a progressive decline in the ability to maintain a stable homeostasis. They believe that, in its later stages, this state is clinically apparent and has a constellation of characteristic representations, including loss of muscle mass, weakness, weight loss, low exercise tolerance and energy, and low activity (2-6). This constellation is hypothesized by a number of scientists to result from cumulative alterations in multiple physiologic and molecular systems, acting in synergy with biological changes of aging. In addition, the hypothesized physiologic decline in adaptive capacity is proposed to result in the inability to tolerate such stresses as surgery, infection, or injury.

Many geriatricians identify frail older adults as the subset of older persons with just such increased vulnerability when stressed, often--but not exclusively--manifested by the presentations above. Although historically the concept of frailty has often been interchanged with comorbidity, disability, and extreme old age, some researchers believe that frailty is a distinct clinical entity and that there is increasing evidence for its intrinsic vulnerability being due to physiological processes identified at the organ system, cellular, and, potentially, molecular levels (2, 7, 8) (see Walston Perspective).

Some investigators have built on this conceptual framework about the etiology of frailty (discussed in more detail below) in defining a clinical phenotype (2, 3) for which considerable evidence has now been obtained on its clinical and pathophysiologic correlates and its predictive validity (9-12). This phenotype is based on the hypothesis that the clinical manifestations of frailty are related in a cycle (Fig. 1) and that, when clinically important, frailty is related to the presence of a critical mass of these manifestations in the cycle. Thus, the frailty phenotype was operationalized in a definition requiring the presence of three or more of the following five criteria, theorized to be related in a cycle: weight loss, weakness, sense of exhaustion, slow walking speed, and low physical activity. The presence of three or more criteria, representing a critical mass and hypothetically consistent with a syndrome, predicted the adverse outcomes clinically thought to result from frailty: disability, falls, hospitalization, and death (3). A prefrail stage, in which one or two criteria are present, identified a subset at high risk of progressing to frailty as defined above and at intermediate risk of the same adverse health outcomes. It is theorized that one criterion may constitute a risk factor but does not represent frailty itself, because by this conceptualization frailty is multisystemic. It was proposed by some participants at the conference that this finding may support the theory of frailty as an aggregate chronic, progressive process with a latent phase. Although it was not the goal of the conference to consider specific definitions or phenotypes of frailty, some conference participants identified this phenotype as a reasonable starting point, offering a physiologic rationale for etiologic investigation and against which alternative conceptualizations of frailty could be compared.

View larger version (9K):
[in this window]
[in a new window]
Fig. 1. Hypothesized cycle of frailty (1). Core components of clinical presentation of frailty are highlighted. All components are known to be related to each other in at least a pairwise manner with strong evidence, with the exception of the early evidence that suggests dysregulation of the balance between energy expended in physical activity with energy taken in through diet in a subset of older adults, potentially frail (2). VO2 max, maximum oxygen consumption.

With this conceptualization, frailty is thought to be present along a continuum of severity, from an end-stage condition that presages death to modest alterations in homeostatic reserves that are not clinically apparent in the steady state but that have the potential for deterioration when confronted with stressors such as acute illness or injury. Older persons with advanced or end-stage frailty, as conceptualized above, appear to be at imminent risk for multiple negative health outcomes, including death (2, 3, 13, 14). From this perspective of frailty occurring along a continuum of severity, in many individuals frailty may occur prematurely and/or be detected at early stages, when treatment or prevention may have positive effects.

Three patient histories were presented at the conference to exemplify the spectrum described above:

(i) A 75-year-old man was admitted to the hospital for surgery for benign prostatic hypertrophy. Before admission, he had a history of ischemic cardiomyopathy, stable congestive heart failure, and osteoarthritis of the knees (see Loeser Perspective and Shakoor Case Study); he also lifted weights and exercised regularly. He tolerated surgery well. The following day, he walked while pushing his IV pole, took a sedative for sleep, and then was discharged home after an uneventful hospital course.

(ii) Another 75-year-old man was also admitted to the hospital for surgery for benign prostatic hypertrophy. He also had a history of congestive heart failure and knee osteoarthritis. He also tolerated surgery well. The following day, he fell while walking to the bathroom while pushing his IV pole. He was bruised and received pain medication for this. He subsequently became confused while on the pain medication; this was persistent over the next few days. As a result, he was kept in bed, where he became progressively weaker. He also became incontinent, perhaps as a result of both being confined to bed and his confusion. He ate little. He was ultimately discharged to a nursing home for rehabilitation.

(iii) An 85-year-old man was brought to the Emergency Room after being found by a neighbor on the floor of his home, where he had been lying for 5 hours after stumbling and falling (without loss of consciousness) and being unable to get up. He had a history of a fall with a hip fracture 3 years before; he also had a history of osteoarthritis of the hips and hands. In the previous year, he had lost 15 pounds in weight, his appetite was fair, and he had had increasing weakness and fatigue. His wife had died 2 years before, and he was still grieving, although not clinically depressed. He lived alone, and family and friends checked on him and brought him food. On examination, this patient was cachectic, with muscle wasting and diffuse weakness. He was unable to walk or rise from his chair independently. He was admitted to the hospital, where no specific acute change could be identified that explained his fall. He underwent rehabilitation, with a very slow course; after 2 weeks, he could only walk 40 feet with a walker. He remained unable to care for himself, and it was considered unsafe for him to live alone. For these reasons, he was transferred to an assisted-living facility--still hoping to return home eventually.

These three patients represent a spectrum of community-dwelling older adults who each experienced a stressor--surgery or a fall--and then manifested differential capacity for recovery, which strongly affected both the clinical course and the outcome. The latter two patients, who demonstrated the most vulnerability to stressors, manifested characteristics that many investigators would consider to be associated with frailty: low muscle mass (sarcopenia; see Hepple Perspective), weakness, fatigue, weight loss, low activity, and/or undernutrition, and risk for outcomes of frailty including falls, delirium, disability, and loss of independence (2-5). Individuals with these characteristics are also at high risk of mortality (3). It is notable that the degree of vulnerability of frail individuals cannot be captured solely by the list of their diseases (3) or by their clinical appearance prior to a stressor but is probably best captured by consideration of dynamic changes and response to challenging tests.

Frailty: One or multiple phenotypes?

Alternative conceptual approaches were presented at the conference regarding frailty, with the premises (i) that the general usage of "frailty" subsumes a very great diversity of physiologic vulnerabilities, weaknesses, instabilities, and limitations (which span a considerably more diverse range than those discussed in the preceding paragraphs) and (ii) that knowledge about the degree to which these phenotypes share common causes or effects is quite limited. From this perspective, the relative utility of a definition of "frailty" as a single entity, versus definitions of individual vulnerabilities, weaknesses, instabilities, and limitations, versus various clusters of such traits remains unclear, pending further data and analyses.

Learning the contributory factors to each of these diverse vulnerabilities, weaknesses, instabilities, and limitations could lead to better ways to prevent or ameliorate them. These traits do not always occur together, but rather individually and in various combinations. We do not yet know the degree to which any set of such traits included in any broad definition of frailty as a "clinical state of vulnerability" actually share common contributory factors or effects. From this perspective, it is risky to base broad definitions of frailty on combinations of traits specified a priori without empiric data on the degree to which they share causes or consequences (even with biologically plausible hypotheses as to the degree to which they do so), because a given factor could have differing, or even opposing, effects on different vulnerabilities, weaknesses, instabilities, or limitations included within any combined definition. To the extent this occurs, the use of such combined definitions could actually confuse, rather than clarify, interpretations of data from studies on causal factors and effects of frailty.

However, these limitations in our knowledge about frailty do not rule out the possibility that broader definitions of frailty may be useful, and even crucial, to identify persons at high risk and to understand the causes and consequences of the variety of traits subsumed under the general idea of frailty. It is quite possible that one or more fundamental etiologies or pathophysiologies contribute to multiple vulnerabilities, weaknesses, instabilities, and limitations or that a single frailty will ultimately be found. Research to identify commonalities is important, because it could identify pathways responsible for multiple adverse effects, which could serve as the target for testing prevention or treatment interventions with multiple benefits.

From this perspective, research on factors that contribute to individual vulnerabilities, weaknesses, instabilities, and limitations not only is valuable in itself but also may aid in identifying clusters of such traits in which all components of the cluster share common contributory factors. Such clusters could be tested as syndromic definitions of one or more "frailties" that encompass the set of related clinical phenomena in the cluster. The basic concept is straightforward and is illustrated in Fig. 2. Analogous approaches using more sophisticated analytic techniques are available. In contrast, strategies based on defining frailty a priori as a specified cluster of traits on theoretical bases, before such analyses are done, could potentially limit identification of contributory factors or effects because they may include groups of traits less likely to share contributory factors or fail to include others that do share contributory factors with those in the cluster. There is great need for research that, by combining these different fields of science, can generate hypotheses about clusters that are clinically relevant and will help understanding as well. Ultimately, understanding will be needed to determine how individual risk factors or clusters synergize in creating an aggregate syndrome differentiated from its individual components.

View larger version (5K):
[in this window]
[in a new window]
Fig. 2. Schematic of findings from hypothetical study or studies of four specific traits (vulnerabilities, weaknesses, instabilities, and/or limitations) and four factors potentially contributing to them. Xs refer to positive findings of association of specific factors with specific traits. Findings indicate that traits A and C have in common etiologic factors 1 and 2 (and not 3 or 4), and that traits B and D have in common factors 3 and 4 (and not 1 or 2), suggesting two possible clusters of "frailties" with distinct and different pathophysiologies or etiologies.

Regardless of whether the conceptual viewpoint is of one frailty, or of multiple frailty-associated vulnerabilities, weaknesses, instabilities, and limitations, or of a variety of clusters of such traits, understanding the etiology and pathophysiology of the condition(s) of interest is essential to establish criteria for diagnosis, including screening and early recognition, and to identify potentially effective preventive and treatment strategies and optimal intervention point(s) over the course of the development of frailty in old age. Preventive or treatment interventions could be effective at a number of points before the end stage.

Etiology and Pathophysiology of Frailty

A range of theories has been developed as to what biology may underlie the clinical presentation of frailty. As described above, one recent physiologically based hypothesis seeking to explain the basis for the observed co-occurrence of clinical manifestations presents a theory of a cycle of frailty (Fig. 1) (2, 3). Building on clinical consensus that skeletal muscle-related declines in strength, exercise tolerance, energy, and physical activity are components of the presentation of frailty, along with undernutrition and weight loss (2-6, 13, 14, 15), and building further on much evidence in the literature for pairwise associations between these, it was hypothesized that these can be organized into a cycle of dysregulated energetics (2, 3). This cycle could, hypothetically, be activated by many different age-related conditions or physiologic alterations. Once triggered, it was hypothesized that this vicious cycle could become self-sustaining and an independent contributor to health outcomes, and even dominate the clinical picture independent of its initial cause. This theory can offer one possible basis for conceptualizing the clinical presentation as an aggregate entity, possibly a clinical syndrome--and a starting point for seeking underlying etiologies as well as a basis for comparison against other hypotheses.

Research to date to characterize the etiology of frailty has focused heavily on determining changes in individual physiologic systems associated with the clinical presentation. Evidence was presented by a number of conference participants that demonstrated, in cross-sectional studies, significant relations between frailty, by a range of definitions, and multiple physiologic systems that are known to contribute to homeodynamic regulation. Presenters reviewed evidence that frailty is associated with loss of muscle mass (sarcopenia) and with activated inflammatory systems and that it is characterized by increased serum levels of the inflammatory cytokine interleukin-6 (IL-6), C-reactive protein, total white blood cells, monocytes, and neutrophils (10-12, 16, 17). In addition, studies were presented that reported decreased concentrations of hormones important in muscle mass maintenance and muscle strength, including the steroid dehydroepiandrosterone sulfate, insulin-like growth factor 1, and testosterone, and their potential interactions with inflammatory factors in the pathophysiology of frailty (16-20). Evidence was also reported for increased clotting activity in frail older individuals, characterized by increased levels of factor VIII, fibrinogen, and D-dimers (derived from cross-linked fibrin blood clots), perhaps induced by activated inflammatory pathways (10). Other physiologic systems that may be implicated in the development of age-related frailty, because they show important age-related changes, include hematopoiesis (see Fuller Perspective), neurological control of movement, and cognition. Loss of optimal feedback mechanisms in such systems may be particularly important (21). Specificity to frailty in old age remains to be examined for some of these systems.

The concept that the hallmark of frailty is a dysregulated response to stressors suggests the need to assess human stress responses for insights into the pathway to frailty. Human stress responses involve the central nervous system (CNS), with both hormonal and neural responses in the hypothalamic-pituitary and sympathetic axes (involving glucocorticoids, epinephrine, norepinephrine, and the neurotransmitter dopamine), the autonomic nervous system, and several other physiologic systems. Etiologically, one organizing hypothesis, articulated at the conference by Hannelore Ehrenreich, is that frailty may arise from a decreased ability of the brain to mount appropriate responses to stressors. Such altered brain function could contribute to current observations of altered hormonal axes, metabolic pathways, slowed motor processing and speed, and decreases in physical activity and/or undernutrition, i.e., components of the physical frailty phenotype. Interestingly, hormones, stress response systems, and the autonomic nervous system all affect the production of inflammatory cytokines such as IL-6. Given the role of IL-6 and other inflammatory cytokines in sarcopenia and the resulting clinical presentations of frailty, this raises the possibility that inflammatory cytokines are a final common pathway of multiple system declines toward frailty.

Dysregulation of any or all of these physiologic systems (i.e., CNS, sympathetic nervous system, hypothalamic-pituitary axis, inflammatory, and endocrine) that contribute to the ability to adapt could contribute, hypothetically, to frailty. Beyond those above, a number of emerging biologic hypotheses as to potential etiologic mechanisms may explain either the occurrence of these multisystem changes or their impact.

One particular focus at the conference was on the early evidence for loss of complexity across multiple systems with aging (22). It was pointed out that systems may be dysregulated with aging and that loss of complexity may result in responses that are inadequate or exaggerated or prolonged where no longer needed. Responses to stressors (22), as well as increased randomness and variability (23-25), may be constituents of the vulnerability at the core of frailty. Because frailty may result from a loss of the redundancy important in resilient and adaptable biologic systems that are in reciprocal communication (see Gavrilov Review), the dynamics of aging systems, and then their interactions, need to be characterized in a nonlinear, dynamic manner (22). Aging, and the vulnerability of frailty in particular, may be associated with loss of complexity of both the fractal architecture of anatomic structure and the dynamics of physiologic processes of these same systems; this has been shown in neurons, bone, kidney, and subepidermis and in heart rate variability, postural control, blood pressure regulation, and gait dynamics (see Lipsitz Perspective for further discussion of loss of complexity, fractals, and frailty). The decline in gait fractal index in individuals who fall often (23-25) provides further support for this theory, because frailty is predictive of frequent falls. Building on chaos theory may provide a basis for understanding the complex, irregular, and seemingly unpredictable dynamics of physiologic systems, which consistently produce fractal patterns.

Each of the alterations above could lead to impairments in multiple physiological systems. Multisystem decrements may also result from defects in signaling networks that connect multiple physiological systems and allow their reciprocal interactions; examples are the cytokine/endocrine and inflammation/clotting system interactions (26, 27). Alterations in such communication systems are often not detectable in the absence of external stressors such as infection, injury, or organ system-based illness. In addition, it is apparent that each of these physiological communication systems has the profound ability to influence other key communication systems. Participants noted that future dynamic studies that test these communication systems will likely produce high yields toward the understanding and treatment of the vulnerability associated with frailty.

Both evidence and theory were presented at the conference to suggest that alterations in multiple physiologic systems as well as an associated, reduced ability for reciprocal compensation, rather than the impairment of one specific physiological system, are key to the development of frailty. At a physiologic level, the aggregate loss of redundancy resulting from decrements in multiple systems could hypothetically constitute what is clinically thought of as loss of reserves with aging and could result in the vulnerability of the frail person to stressors and to illness. The implications of this clinically are substantial: If the multiplicity of systems affected is the issue in the etiology of frailty, then repletion of one system could, in some cases, be ineffective therapy in itself or, on the other hand, could restore a sufficient degree of redundancy to produce clinical benefits.

There is much theory, based on clinical observation, that frail older adults have diminished physiologic reserves and compromised ability to compensate for stressors and maintain homeostasis under dynamic conditions. It has been hypothesized by many investigators that the ultimate characteristic of frailty at a physiologic level is this aggregate diminution in loss of reserves and vulnerability to stressors. This could, theoretically, result from the aggregate dysregulations of the multiple systems that, when well integrated, comprise a robust and resilient organism, as suggested in Fig. 3. This loss of reserves may be possible to observe under static or resting circumstances or may only be recognizable under stressed conditions, using markers within single systems (delayed, abnormal, or abnormally regulated stress response) or across systems.

View larger version (19K):
[in this window]
[in a new window]
Fig. 3. Systems biology of frailty. Examples of dynamics within and across multiple levels (clinical presentation, physiologic dysregulation, cellular function, and molecular and genetic characteristics), which are theorized to result in the aggregate loss of organismal resilience and ability to compensate for stressors that is frailty. This multilevel network emphasizes the complexity of an intact system and visually suggests the loss of complexity in response to challenges for the organism as a whole likely to result from dysregulation of multiple systems. DHEA-S, dehydroepiandrosterone sulfate; GH, growth hormone; ANS, autonomic nervous system.

Thus, some conference attendees identified the ultimate phenotype of frailty as that of physiologic vulnerability, likely apparent in response to stressors but potentially not in the steady state. Latent vulnerability, when unmasked by the presence of stressors, may be manifested as an inadequate homeodynamic response as well as a resulting decline in function and an increase in falls and mortality. The clinical presentation identified with being frail likely occurs at the threshold point in this latent vulnerability. There are, as yet, no known measures that can identify those with such vulnerability before the development of the clinical phenotype--particularly without applying a stressor to the system. Development of definitions that include one or more test results or other indicators of vulnerability are necessary to delineate frail groups more sharply and will help identify central etiologies. This is a critical area, and a pressing next stage, for frailty research.

Potential Roles of Age-Related Molecular Changes and Genetic Variation in Frailty

This conference provided a valuable opportunity for interactions between clinical investigators and investigators engaged in studying the basic biology and genetics of aging. With the multiple physiologic systems involved in the clinical syndrome of frailty, a search for ultimate molecular or genetic causes has increasing validity. Several potential candidates were discussed.

Age-related alterations in mitochondrial function in many tissue types were considered as a possible causal mechanism of frailty. Such changes could include declines in cellular energy production, increased generation of free radical oxygen species, and effects on cellular redox potential within the cell (28) (see Nicholls Perspective, Dugan Perspective, and "The Two Faces of Oxygen"). These factors, in turn, could alter gene transcription and cause DNA damage, which can affect cell survival, replicative capacity, and function. Factors affecting replicative capacity, which also include processes such as telomere shortening, can lead to the appearance of senescent cell types that are resistant to apoptosis and incapable of division (29-33). These senescent cells, because of the altered nature of their gene expression patterns, are also functionally different and may be more likely to secrete inflammatory cytokines, which could disrupt the normal morphology of tissues (29) (see also "Faustian Bargain" and "Led Astray"). Senescent cells that are functionally different, combined with chronic loss of cells from apoptosis, could drive the decline in tissue integrity and lead to loss of, or altered function across, multiple physiologic systems that may underlie the physiologic vulnerability in frailty. Ultimately, it is likely that these proposed biologic mechanisms are intimately linked with frailty and its clinical expression. Research into these areas was recommended. Further, some conference participants recommended the study of the molecular declines in cells in specific communication systems known to influence vulnerability, including the immune, hematopoietic, and central and sympathetic nervous systems.

Hundreds, if not thousands, of genes are involved in signal transduction pathways important in inflammation, skeletal muscle maintenance, endocrine systems, energy metabolism, and protein production. Individual polymorphic variants in single genes, or many polymorphic variants of modest effect in several signal transduction pathways, may cumulatively contribute to molecular and physiologic change and ultimately lead to the vulnerability of frailty. Genetic and molecular research agendas to help identify common molecular pathways that bridge across physiologic systems were suggested as potential high-yield studies.

In summary, frailty research could benefit from the application of advances in the basic biology of aging and from advances in the genetics of complex diseases and syndromes. In addition, targeting physiologic pathways relevant to frailty by basic science investigators is likely to facilitate meaningful translational research efforts.

Frailty as a Potential Exemplar of Systems Biology

Through the conference discussions, it became apparent that the clinical presentation of frailty was likely a cumulative outcome of several levels of interrelationships: (i) between physiologic systems; (ii) between different aspects of cellular and molecular changes, including their age-related changes, that could influence each other; (iii) between these specific cellular and molecular functions and their influence on physiologic systems and, beyond that, on clinical manifestations; and (iv) potentially, between clinical manifestations, such as physical activity or undernutrition, and each other and with physiologic and biologic levels. Thus, the rich web of interconnectedness within and across clinical, physiologic, and biologic levels of organismal functioning may, hypothetically, contribute to biological stability or, if depleted, to the development of the vulnerability and clinical presentation of frailty; this is portrayed in Fig. 3. This multilevel network exemplified in Fig. 3 emphasizes the complexity of an intact system and visually suggests the loss of complexity for the organism as a whole that could result from dysregulation of multiple systems and then be manifested in its response to challenges.

This conference, in discussing the multiple systems associated with frailty, the dysregulation of individual systems, and the complex interactions across systems, lent itself to considering a systems biology conceptualization of frailty. These theories are allied with the concept that these physiologic systems, along with molecular and cellular systems, are integrated into complex networks, with feedback and feed-forward loops, signaling pathways, and genetic switches. When intact, this complex, integrated web of systems offers redundancy and the ability to compensate effectively for stressors. Theoretically, it was posited that frailty could be the result of the deterioration of the intact system, with resulting homeostatic vulnerability due to aggregate losses of redundancy and network complexity. Fig. 3 exemplifies the normal, rich web of interrelated healthy systems and suggests the implications, in terms of the ability to maintain homeostasis, if such a network were depleted.

Of note, evidence was reviewed at the conference that indicates that components of the clinical presentation of frailty can feed back on other levels; for example, as indicated in Fig. 3, physical activity can affect physiological factors and it was hypothesized during the conference to affect molecular factors as well. Intriguingly, this suggests the potential for physical activity interventions to be used as treatments of potential physiologic triggers for frailty. Additional studies may allow the elucidation of direct connections between physical activity and molecular triggers, which in turn may enable more targeted approaches in the treatment of frailty.

The above suggests that frailty may present a new theoretical model for systems biology. Kitano has proposed (34) that the structure and dynamics of cellular and organismal function must be examined to understand biology at the system level, rather than the characteristics of isolated parts of a cell or organism. Frailty research would suggest that the properties of systems, such as robustness in the face of stressors, need to be viewed at multiple levels beyond (but including) cellular and organismal function, including communication networks essential to the functionality of multiple physiologic systems and the relationship of these system properties to the clinical presentation. Components of both clinical presentation and physiologic dysregulation may feed back to affect both cellular and organismal function and molecular characteristics. This is consistent with the theory of Csete and Doyle (35) as to the layers of feedback regulation that underlie the complexity and robustness of organisms. These observations suggest new directions for organizing research on the biology of frailty at the level of systems biology.

Future Research Development and Potential Interventions in Frailty: From Bench to Bedside and Back

Frailty research began with clinical observation of a subgroup of older patients who had a vulnerability to stressors and high risk for adverse health outcomes and who had a recognizable pattern of clinical characteristics, including loss of muscle mass, weight loss, little energy, weakness, and diminished activity levels. To date, there has been substantial progress in the development of phenotypic definitions that offer an approximation of the clinical presentation and early insight into some of the multiple physiologic mechanisms that may contribute to frailty. Data from the biology of aging literature provides additional evidence to support hypotheses regarding molecular and/or genetic etiologies of frailty. This frailty conference offered an integrated perspective suggesting that alterations at every biologic level may contribute, likely in the aggregate, to frailty.

In addition to theories about the multilevel nature of frailty (Fig. 3), this conference's interdisciplinary scientific interactions could be summarized as being consistent, conceptually, with the stages of development of research on frailty and the ongoing translational needs, as summarized in Fig. 4. Although it is a simplified model, it suggests the flow of translation between scientific approaches that is ultimately required to determine the etiology and therapies, as appropriate, for this clinical problem. Ultimately, it is likely that there are effective preventive approaches or treatments that can be developed without fully understanding pathophysiology and that interventions targeted to different levels may have different types of efficacy and different points in the progressive development of frailty at which they may be effective. Intervention studies can provide information on mechanisms as well as efficacy. Ultimately, the knowledge of both mechanisms and effective interventions at different levels and points in the natural history of frailty will offer the opportunity to identify the critical points for most effective, and parsimonious, interventions.

View larger version (11K):
[in this window]
[in a new window]
Fig. 4. The progression of research on frailty has followed the natural history of investigation driven by a clinical question.


This conference was important in bringing together scientists with widely varying expertise and areas of research focus, facilitating identification of the next generation of data needed to understand the etiology of frailty in aging, as well as highlighting the multidisciplinarity essential to its investigation. Recommendations for future research made by one or more conference participants included:

(1) Formal agreement on standardized, preliminary criteria for a clinical phenotype of frailty, recognizing that this conference was not organized to accomplish this. Other approaches could then be compared to this.

(2) Determination of the contributions to frailty from other clinical domains not currently included in broad definitions of frailty that may enhance predictive value and potential for etiologic insights from these definitions. One example is cognitive impairment or subclinical central and peripheral neurological degeneration, which may be etiologic factors, components, or correlates of frailty. Some attendees suggested that an important part of a future research agenda would be the enhancement of understanding of the role of cognition and neurodegeneration in frailty and consideration of markers of neurological impairment as part of future definitions of frailty.

(3) Identification of contributors to individual vulnerabilities, weaknesses, instabilities, and/or limitations, identification of clusters of such traits that share common contributing factors, and determination of the predictive value of phenotypes defined by such clusters. One approach would be to identify combinations that occurred more commonly than would be expected by chance.

(4) Research that would aid in assessing the utility of different phenotypes relating to frailty, including (i) determination of whether frailty comprises a clinical syndrome; (ii) evaluation of whether there is one frailty with multiple risk factors and/or a common ultimate cause, or multiple frailty phenotypes or frailty-associated conditions with different constellations of risk factors, etiologies, and/or natural histories; and (iii) assessment of whether there are more precise clusters of signs and symptoms that identify frail older adults than are currently recognized.

(5) Development of methods to identify physiologic vulnerabilities of frailty, such as poor responses to stressors. Potential approaches could include stimulation tests or simple tests that identify those who are frail, regardless of the presence (or absence) of the clinical presentation in the stable state.

(6) Identification of subclinical components and mechanisms of frailty and relations between the molecular, cellular, and physiologic levels.

(7) Development of animal and cellular models explicitly designed for the investigation of the etiology or treatments of frailty and delineation of premature or preventable frailty.

(8) Application of large population-based studies and other clinical studies to the evaluation of contributory factors to frailty, including multisystem factors, and its outcomes, and characterization of the natural history of frailty.

(9) Identifying genetic, cellular, physiologic, psychological, or sociobehavioral factors with pleiotropic effects on multiple vulnerabilities, weaknesses, instabilities, and/or limitations, e.g., studies of effects of individual hormones or cytokines and their interactions.

(10) Determination of whether there are clusters of risk factors, such as physiologic or molecular systems that are dysregulated, that most strongly identify those at risk for frailty or specific frail phenotypes, and evaluation of such information for insights into potential targets for preventive or therapeutic strategies.

(11) Development of innovative analytical techniques critical to understanding the altered dynamics and important interactions that underlie the vulnerability of frailty. These are likely to include computational biology, especially nonlinear methods, and development of the ability to analyze multisystem and multilevel interactions in the dynamic state.

(12) Development of large collaborative networks where populations and resources can be shared within and across institutions for the study of frailty, to enhance the range of methods and research strategies focused on frailty. Novel interactive approaches will be needed for ongoing integration of discoveries at the cellular and molecular level with those at the system and multisystem level and with clinical outcomes.

For complementary, and in some cases expanded, discussion of the background relating to the recommendations, see (36).

Ultimately, research on frailty will need to tackle the unanswered question of where the biology of frailty and the biology of aging overlap and where, if anyplace, they differ. If frailty exemplifies systems biology, then the study of frailty will serve the understanding of systems biology broadly, as well as informing core issues in aging-related research. Unanswered, but hypothesized at the conference, are theories that frailty may be a continuum of the expression of aging effects on multisystem regulation and resulting resilience, and that the clinical manifestations are its most extreme or aggregate expression. The current hypothesis is that early manifestations can be prevented or delayed and that the development of effective prevention and treatment is warranted. These questions remain to be answered.

August 3, 2005
  1. K. Rockwood, Frailty and its definition: A worthy challenge. J. Am. Geriatr. Soc. 53, 1069-1070 (2005). [CrossRef][Medline]
  2. L. P. Fried, J. Walston, in Principles of Geriatric Medicine and Gerontology, W. R. Hazzard, J. P. Blass, W. H. Ettinger Jr., J. B. Halter, J. Ouslander, Eds. (McGraw-Hill, New York, ed. 5, 2003), chap. 116, pp. 1487-1502.
  3. L. P. Fried, C. M. Tangen, J. Walston, A. B. Newman, C. H. Hirsch, J. S. Gottdiener, T. Seeman, R. Tracy, W. J. Kop, G. Burke et al. Frailty in older adults: Evidence for a phenotype. J. Gerontol. A Biol. Sci. Med. Sci. 56, M146-M156 (2001). [Abstract/Free Full Text]
  4. L. Ferrucci, J. M. Guralnik, S. Studenski, L. P. Fried, G. B. Cutler, J. D. Walston, Designing randomized controlled trials aimed at preventing or delaying functional decline and disability in frail, older persons: A consensus report. J. Am. Geriatr. Soc. 52, 625-634 (2004). [CrossRef][Medline]
  5. S. Studenski, R. P. Hayes, R. O. Leibowitz, R. Bode, L. Lavery, J. Walston, P. Duncan, S. Perera, Clinical global impression of change in physical frailty: Development of a measure based on clinical judgement. J. Am. Geriatr. Soc. 52, 1560-1566 (2004). [CrossRef][Medline]
  6. A. Chin, M. J. Paw, J. M. Dekker, E. J. Feskens, E. G. Schouten, D. Kromhout, How to select a frail elderly population? A comparison of three working definitions. J. Clin. Epidemiol. 52, 1015-1021 (1999). [CrossRef][Medline]
  7. L. P. Fried, L. Ferrucci, J. Darer, J. D. Williamson, G. Anderson, Untangling the concepts of disability, frailty, and comorbidity: Implications for improved targeting and care. J. Gerontol. A Biol. Sci. Med. Sci. 59, M255-M263 (2004). [Abstract/Free Full Text]
  8. J. Walston, Frailty--the search for underlying causes. Sci. Aging Knowledge Environ. 2004 (4), pe4 (2004). [Abstract/Free Full Text]
  9. A. B. Newman, J. S. Gottdiener, M. A. McBurnie, C. H. Hirsch, W. J. Kop, R. Tracy, J. D. Walston, L. P. Fried, for the Cardiovascular Health Study Research Group. Associations of subclinical cardiovascular disease with frailty. J. Gerontol. A Biol. Sci. Med. Sci. 56, M158-M166 (2001). [Abstract/Free Full Text]
  10. J. Walston, M. A. McBurnie, A. Newman, R. Tracy, W. J. Kop, C. H. Hirsch, J. Gottdiener, L. P. Fried, Frailty and activation of the inflammation and coagulation systems with and without clinical morbidities: Results from the Cardiovascular Health Study. Arch. Intern. Med. 162, 2333-2341 (2002).[CrossRef][Medline]
  11. S. Leng, P. Chaves, K. Koenig, J. Walston, Serum interleukin-6 and hemoglobin as physiological correlates in the geriatric syndrome of frailty: A pilot study. J. Am. Geriatr. Soc. 50, 1268-1271 (2002).[CrossRef][Medline]
  12. S. Leng, Q.-L. Xue, Y. Huang, R. Semba, P. Chaves, K. Bandeen-Roche, L. Fried, J. Walston, Total and differential white blood cell counts and their associations with circulating interleukin-6 levels in community-dwelling older women J. Gerontol. A Biol. Sci. Med. Sci. 60, 195-199 (2005).[Abstract/Free Full Text]
  13. B. Berkman, L. W. Foster, E. Campion, Failure to thrive: Paradigm for the frail elderly. Gerontologist 29, 654-659 (1989). [Abstract/Free Full Text]
  14. R. B. Verdery, Failure to thrive in older people. J. Am. Geriatr. Soc. 44, 465-466 (1996). [Medline]
  15. S. B. Roberts, P. Fuss, M. B. Heyman, W. J. Evans, R. Tsay, H. Rasmussen, M. Fiatarone, J. Cortiella, G. E. Dallal, V. R. Young, Control of food intake in older men. JAMA 272, 1601-1606 (1994). [CrossRef][Medline]
  16. R. Roubenoff, H. Parise, H. A. Payette, L. W. Abad, R. D'Agostino, P. F. Jacques, P. W. Wilson, C. A. Dinarello, T. B. Harris, Cytokines, insulin-like growth factor 1, sarcopenia, and mortality in very old community-dwelling men and women: the Framingham Heart Study. Am. J. Med. 115, 429-435 (2003). [CrossRef][Medline]
  17. R. Roubenoff, Catabolism of aging: Is it an inflammatory process? Curr. Opin. Clin. Nutr. Metab. Care 6, 295-299 (2003). [CrossRef][Medline]
  18. S. X. Leng, A. R. Cappola, R. E. Andersen, M. R. Blackman, K. Koenig, M. Blair, J. D. Walston, Serum levels of insulin-like growth factor-I (IGF-I) and dehydroepiandrosterone sulfate (DHEA-S), and their relationships with serum interleukin-6, in the geriatric syndrome of frailty. Aging Clin. Exp. Res. 16, 153-157 (2004).[CrossRef][Medline]
  19. J. V. Hennessey, J. A. Chromiak, S. DellaVentura, S. E. Reinert, J. Puhl, D. P. Kiel, C. J. Rosen, H. Vandenburgh, D. B. MacLean, Growth hormone administration and exercise effects on muscle fiber type and diameter in moderately frail older people. J. Am. Geriatr. Soc. 49, 852-858 (2001).[CrossRef][Medline]
  20. A. R. Cappola, Q. L. Xue, L. Ferrucci, J. M. Guralnik, S. Volpato, L. P. Fried, Insulin-like growth factor I and interleukin-6 contribute synergistically to disability and mortality in older women. J. Clin. Endocrinol. Metab. 88, 2019-2025 (2003).[CrossRef][Medline]
  21. L. Ferrucci, C. Cavazzini, A. Corsi, B. Bartali, C. R. Russo, R. F. Lauretani et al. Biomarkers of frailty in older persons. J. Endocrinol. Invest. 25 Suppl 10, 10-15 (2002).[Medline]
  22. L. A. Lipsitz, Dynamics of stability: The physiologic basis of functional health and frailty. J. Gerontol. A Biol. Sci. Med. Sci. 57, B115-B125 (2002).[Abstract/Free Full Text]
  23. J. M. Hausdorff, Y. Ashkenazy, C. K. Peng, P. C. Ivanov, H. E. Stanley, A. L. Goldberger, When human walking becomes random walking: Fractal analysis and modeling of gait rhythm fluctuations. Physica A 302, 138-147 (2001).[CrossRef][Medline]
  24. J. M. Hausdorff, Gait variability: Methods, modeling and meaning. J. Neuroengineering Rehabil. 2, 19 (2005). [CrossRef][Medline]
  25. T. Herman, N. Giladi, T. Gurevich, J. M. Hausdorff, Gait instability and fractal dynamics of older adults with a "cautious" gait: Why do certain older adults walk fearfully? Gait Posture 21, 178-185 (2005). [CrossRef][Medline]
  26. J. P. de Magalhaes, O. Toussaint, GenAge: A genomic and proteomic network map of human ageing.�FEBS Lett. 571,�243-247 (2004). [CrossRef][Medline]
  27. J. B. Strosznajder, H. Jesko, A. Zambrzycka, A. Eckert, M. Chalimoniuk, Age-related alteration of activity and gene expression of endothelial nitric oxide synthase in different parts of the brain in rats. Neurosci. Lett. 370, 175-179 (2004). [CrossRef][Medline]
  28. G. A. de Erausquin, K. Hyrc, D. A. Dorsey, D. Mamah, M. Dokucu, D. H. Masco, T. Walton, K. Dikranian, M. Soriano, J. M. Garcia Verdugo et al. Nuclear translocation of nuclear transcription factor-kappa B by alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors leads to transcription of p53 and cell death in dopaminergic neurons. Mol. Pharmacol. 63, 784-790 (2003). [Abstract/Free Full Text]
  29. J. Campisi, Cellular senescence and apoptosis: How cellular responses might influence aging phenotypes. Exp. Gerontol. 38, 5-11 (2003).[CrossRef][Medline]
  30. S. H. Kim, P. Kaminker, J. Campisi, Telomeres, aging and cancer: In search of a happy ending. Oncogene 21, 503-511 (2002).[CrossRef][Medline]
  31. R. M. Cawthon, K. R. Smith, E. O'Brien, A. Sivatchenko, R. A. Kerber, Association between telomere length in blood and mortality in people aged 60 years or older. Lancet 361, 393-395 (2003).[CrossRef][Medline]
  32. J. S. Kim-Han, S. A. Reichert, K. L. Quick, L. L. Dugan, BMCP1: A mitochondrial uncoupling protein in neurons which regulates mitochondrial function and oxidant production. J. Neurochem. 79, 658-668 (2001).[CrossRef][Medline]
  33. N. C. Souza-Pinto, V. A. Bohr, The mitochondrial theory of aging: Involvement of mitochondrial DNA damage and repair. Int. Rev. Neurobiol. 53, 519-534 (2002).[Medline]
  34. H. Kitano, Systems biology: A brief overview. Science 295, 1662-1664 (2002).[Abstract/Free Full Text]
  35. M. E. Csete, J. C. Doyle, Reverse engineering of biological complexity. Science 295, 1664-1669 (2002). [Abstract/Free Full Text]
  36. J. Walston, E. Hadley, L. Ferrucci, J. Guralnik, A. Newman, A. Studenski, W. Ershler, T. Harris, L. P. Fried, Research agenda for frailty in older adults: Towards a better understanding of physiology and etiology. Summary from the American Geriatrics Society/National Institute on Aging research conference on frailty in older adults. J. Am. Geriatr. Soc. (in press).
  37. The authors thank Jill Epstein and Julie Pestana and the American Geriatrics Society for their expert support in the organization of this conference. This work was supported by grants from the National Institute on Aging: Conference Series: Bedside to Bench; U13AG022361 and Pathogenesis of Physical Disability in Aging Women; R37-AG19905-03.
  38. The following scientists participated in this conference: Mark Bach, Merck & Co., Inc., Rahway, NJ, USA; Ludovico Balducci, H. Lee Moffitt Cancer Center Research Institute, Tampa, FL, USA; Brock Beamer, Johns Hopkins University School of Medicine, Baltimore, MD, USA; Howard Bergman, Canadian Initiative on Frailty and Aging, Montreal, Canada; Caroline Blaum, University of Michigan Medical School, Ann Arbor, MI, USA; Vilhelm Bohr, National Institute on Aging, Baltimore, MD, USA; Cynthia Brown, University of Alabama at Birmingham, Birmingham, AL, USA; Judith Campisi, Lawrence Berkeley National Laboratory, Berkeley, CA, USA; Anne Cappola, University of Pennsylvania, Philadelphia, PA, USA; Christy Carter, Wake Forest University School of Medicine, Winston-Salem, NC, USA; Chiara Cavazzini, National Public Health System, Rome, Italy; Richard Cawthon, University of Utah, Salt Lake City, UT, USA; Aravinda Chakravarti, Johns Hopkins University, Baltimore, MD, USA; Laura Dugan, Washington University, St. Louis, MO, USA; Hannelore Ehrenreich, Max Planck Institute for Experimental Medicine, Goettingen, Germany; William Ershler, Institute for Advanced Studies in Aging and Geriatric Medicine, Washington, DC, USA; William Evans, University of Arkansas, Little Rock, AR, USA; Neal Fedarko, Johns Hopkins Bayview Medical Center, Baltimore, MD, USA; Luigi Ferrucci, National Institute on Aging, Baltimore, MD, USA; Linda Fried, Johns Hopkins Medical Institutions, Baltimore, MD, USA; Tom Gill, Yale University School of Medicine, New Haven, CT, USA; Brett Goodpaster, University of Pittsburgh, Pittsburgh, PA, USA; Jack Guralnik, National Institute on Aging, Bethesda, MD, USA; Evan Hadley, National Institute on Aging, Bethesda, MD, USA; Mark Hallet, National Institute of Neurological Disorders and Stroke, Bethesda, MD, USA; Tamara Harris, National Institute on Aging, Bethesda, MD, USA; Jeffrey Hausdorff, Beth Israel Deaconess Medical Center, Boston, MA, USA; Richard Hodes, National Institute on Aging, Bethesda, MD, USA; Nancy Jenny, College of Medicine, University of Vermont, Colchester, VT, USA; Fran Kaiser, University of Texas Southwestern and St. Louis University, St. Louis, MO, USA; Doug Kiel, Harvard Medical School, Boston, MA, USA; James Kirkland, Boston University School of Medicine, Boston, MA, USA; Lewis Lipsitz, Beth Israel Deaconess Medical Center, Boston, MA, USA; Heather McDonald, SAGE KE, Washington, DC, USA; John Morley, Journals of Gerontology, St. Louis, MO, USA; Anne Newman, University of Pittsburgh, Pittsburgh, PA, USA; Marco Pahor, Wake Forest University School of Medicine, Winston-Salem, NC, USA; Mukaila Raji, University of�Texas Geriatrics and Memory Loss Clinics, Galveston, TX, USA; Ronenn Roubenoff, Tufts University, Boston, MA, USA; Judy Salerno, National Institute on Aging, Bethesda, MD, USA; Jeffrey Silverstein, Mount Sinai School of Medicine, West Nyack, NY, USA; Stephanie Studenski, University of Pittsburgh, Pittsburgh, PA, USA; George Taffet, Baylor College of Medicine, Houston, TX, USA; Mary Tinetti, Yale University, New Haven, CT, USA; Russell Tracy, College of Medicine, University of Vermont, Colchester, VT, USA; Bruce Troen, University of Miami School of Medicine, Miami, FL, USA; Ravi Vardahan, Johns Hopkins University School of Medicine, Baltimore, MD, USA; Jeremy Walston, Johns Hopkins University School of Medicine, Baltimore, MD, USA; Michi Yukawa, Harborview Medical Center, Seattle, WA, USA.
Citation: L. P. Fried, E. C. Hadley, J. D. Walston, A. B. Newman, J. M. Guralnik, S. Studenski, T. B. Harris, W. B. Ershler, L. Ferrucci, From Bedside to Bench: Research Agenda for Frailty. Sci. Aging Knowl. Environ. 2005 (31), pe24 (2005).

The Effect and Relationship of Frailty Indices on Survival After Transcatheter Aortic Valve Replacement.
S. Kiani, A. Stebbins, V. H. Thourani, J. Forcillo, S. Vemulapalli, A. S. Kosinski, V. Babaliaros, D. Cohen, S. K. Kodali, A. J. Kirtane, et al. (2020)
JCIN 13, 219-231
   Abstract »    Full Text »    PDF »
Valvular Heart Disease in Patients >=80 Years of Age.
S. K. Kodali, P. Velagapudi, R. T. Hahn, D. Abbott, and M. B. Leon (2018)
JACC 71, 2058-2072
   PDF »
Impact of frailty on outcomes after discharge in older surgical patients: a prospective cohort study.
Y. Li, J. L. Pederson, T. A. Churchill, A. S. Wagg, J. M. Holroyd-Leduc, K. Alagiakrishnan, R. S. Padwal, and R. G. Khadaroo (2018)
CMAJ 190, E184-E190
   Abstract »    Full Text »    PDF »
Frail Elderly, the Ideal Patients for MitraClip.
S. V. Arnold (2017)
JCIN 10, 1930-1931
   Full Text »    PDF »
Considerations in Starting a Patient with Advanced Frailty on Dialysis: Complex Biology Meets Challenging Ethics.
M. Swidler (2013)
Clin. J. Am. Soc. Nephrol. 8, 1421-1428
   Abstract »    Full Text »    PDF »
The Impact of Frailty Status on Survival After Transcatheter Aortic Valve Replacement in Older Adults With Severe Aortic Stenosis: A Single-Center Experience.
P. Green, A. E. Woglom, P. Genereux, B. Daneault, J.-M. Paradis, S. Schnell, M. Hawkey, M. S. Maurer, A. J. Kirtane, S. Kodali, et al. (2012)
JCIN 5, 974-981
   Abstract »    Full Text »    PDF »

Science of Aging Knowledge Environment. ISSN 1539-6150