Frailty in the critically ill: a novel concept

The concept of frailty has been defined as a multidimensional syndrome characterized by the loss of physical and cognitive reserve that predisposes to the accumulation of deficits and increased vulnerability to adverse events. Frailty is strongly correlated with age, and overlaps with and extends aspects of a patient's disability status (that is, functional limitation) and/or burden of comorbid disease. The frail phenotype has more specifically been characterized by adverse changes to a patient's mobility, muscle mass, nutritional status, strength and endurance. We contend that, in selected circumstances, the critically ill patient may be analogous to the frail geriatric patient. The prevalence of frailty amongst critically ill patients is currently unknown; however, it is probably increasing, based on data showing that the utilization of intensive care unit (ICU) resources by older people is rising. Owing to the theoretical similarities in frailty between geriatric and critically ill patients, this concept may have clinical relevance and may be predictive of outcomes, along with showing important interaction with several factors including illness severity, comorbid disease, and the social and structural environment. We believe studies of frailty in critically ill patients are needed to evaluate how it correlates with outcomes such as survival and quality of life, and how it relates to resource utilization, such as length of mechanical ventilation, ICU stay and duration of hospitalization. We hypothesize that the objective measurement of frailty may provide additional support and reinforcement to clinicians confronted with end-of-life decisions on the appropriateness of ICU support and/or withholding of life-sustaining therapies.

adverse events [1]. Frailty overlaps and extends beyond disability (functional limitation) and comorbidity (coexistence of two diseases), and acknowledges that patients can be disabled and/or have comorbidities without being frail, and vice versa.
Ageing is a complex interplay between genetics and environment that begins during embryonic and fetal develop ment. Th e mechanisms underlying the ageing pro cess are only beginning to be clarifi ed. In recognizing that ageing is a very heterogeneous process, many clinicians seek a method to quantify physiologic age rather than simply chronologic age. Since frailty has been shown to closely correlate with the ageing process [2], it has been suggested that the syndrome of frailty may be just such a measure.
One hypothesis whereby ageing is associated with and may predispose to development of frailty relates to the concept of infl ammaging: the dynamic interplay between the protective proinfl ammatory response to invading microorganisms and the similarly protective com pensatory anti-infl ammatory system, which defends against uncontrolled infl ammation. Genetic polymorph isms in the proinfl ammatory and anti-infl ammatory responses have been proposed as one potential mechanism to explain some of the individual variability in the rate of ageing, and may partly explain the poor discriminatory power of age alone to predict outcome [3]. An excessively strong proinfl ammatory response that may be protective during the reproductive years may become maladaptive later in life [4]. By exhausting the compensatory antiinfl ammatory system, the proinfl ammatory response results in unintended damage to the host organism and predisposes to a vicious cycle of decreasing muscle mass, malnutrition and reduced energy expenditure. Th is cycle eventually culminates in the inability to maintain homeostasis and an 'avalanche-like destruction of the organism' [2,4]. One expres sion of this unbridled infl ammation may be the syndrome of frailty, a state in which physiologic defi cits accumulate that individually may be reversible but collectively often represent an insurmountable burden of disease and consequently vulnerability to adverse outcomes [5] (Figure 1).

Measuring and quantifying frailty
Th e syndromic nature of frailty presents challenges in creating an eff ective defi nition of the state. As previously

Abstract
The concept of frailty has been defi ned as a multidimensional syndrome characterized by the loss of physical and cognitive reserve that predisposes to the accumulation of defi cits and increased vulnerability to adverse events. Frailty is strongly correlated with age, and overlaps with and extends aspects of a patient's disability status (that is, functional limitation) and/ or burden of comorbid disease. The frail phenotype has more specifi cally been characterized by adverse changes to a patient's mobility, muscle mass, nutritional status, strength and endurance. We contend that, in selected circumstances, the critically ill patient may be analogous to the frail geriatric patient. The prevalence of frailty amongst critically ill patients is currently unknown; however, it is probably increasing, based on data showing that the utilization of intensive care unit (ICU) resources by older people is rising. Owing to the theoretical similarities in frailty between geriatric and critically ill patients, this concept may have clinical relevance and may be predictive of outcomes, along with showing important interaction with several factors including illness severity, comorbid disease, and the social and structural environment. We believe studies of frailty in critically ill patients are needed to evaluate how it correlates with outcomes such as survival and quality of life, and how it relates to resource utilization, such as length of mechanical ventilation, ICU stay and duration of hospitalization. We hypothesize that the objective measurement of frailty may provide additional support and reinforcement to clinicians confronted with endof-life decisions on the appropriateness of ICU support and/or withholding of life-sustaining therapies.  , frailty is strongly correlated with chronological age [6], but it is not an inevitable part of ageing [1]. Furthermore, the prevalence of frailty within closely aligned age strata, even in the very old person, is variable [7]. Consequently, a number of descriptive tools have been developed to defi ne and quantify frailty.
One of the most widely adopted tools is the operational defi nition described by Fried and colleagues [2] (Table 1). Th e Frailty Index, a detailed 70-item inventory of clinical defi cits, is also broadly used in studies of frailty [8]. A more generic, less detailed but no less clinically valid impression of patient frailty has also been developed by Rockwood and colleagues [1]. In 2,305 patients aged 65 years or older participating in the second stage of the Canadian Study on Health and Aging, Rockwood and colleagues developed and validated a judgment-based seven-point Clinical Frailty Scale (CFS) to measure frailty [1] (Table 2). In their study, the CFS was highly correlated with the Frailty Index. Participants with higher CFS scores were older, more often female, and more likely to have cognitive impairment and impaired mobility. By multi-variable analysis, each one-point increase in the CFS translated into signifi cantly higher hazards of death (hazard ratio = 1.30) and entry into an institutional facility (hazard ratio = 1.46). Each of these tools appears to perform similarly well in identifying older patients at risk for adverse outcomes, but to date have not been evaluated in other populations [1,2,9,10].

How is frailty relevant to critical care?
Th e prevalence of frailty in the older demographic may be as high as 43% [1,11]. Based on evidence showing that utilization of intensive care unit (ICU) resources by older people is rising, the prevalence of pre-existing frailty in patients admitted to the ICU is probably also increasing [3].
Th e relevance of frailty, however, is not limited to admission demographics. Whether due to chronic disease depleting the reserve or acute disease over whelming the reserve, the critically ill patient is vulnerable to adverse clinical outcomes, as evidenced by the number and severity of unexpected deteriorations in clinical status requiring increases in the degree of life support, without which the critically ill patient would die. Additionally, defi cits associated with frailty, which typically take years to accumulate in the outpatient geriatric population, rapidly develop in a large proportion of critically ill patients independent of age and illness severity. Th ese features include muscle wasting, clinically signifi cant weakness and poor functional status following discharge from the ICU [12,13]. A recent editorial underscored the potential importance of infl ammation in the development of acquired muscle weakness in the critically ill patient [14]. Additional pathophysiologic mechanisms proposed for these fi ndings have included immobilization, suboptimal nutritional supplementation and ineff ective substrate utilization -all of which may be further compounded by medications such as neuromuscular blockers and corticosteroids [15]. In fact, functional dependence after critical illness is correlated with two of the phenotypic features of characterizing frailty: inability to walk and poor upper extremity strength [13].
Since critically ill patients of all ages may share many of the features seen in frail geriatric patients, we contend that the concept and measurement of frailty may have clinical, psychosocial and economic relevance to critical care medicine. Accordingly, we hypothesize that the objective evaluation of frailty in critical illness may comple ment and/or contribute important prognostic information in the clinical care of patients.

What are the prognostic implications of frailty?
Frailty is recognized as a major determinant of mortality, hospitalization, institutionalization and functional outcome in geriatric patients, and outperforms chronological  age [6,16,17]. In fact, frailty may represent a surrogate for many of the diffi cult-to-measure aspects of a patient's prehospital health state. To date, however, no study has prospectively evaluated the prevalence or associated outcomes of frailty in critically ill patients.
Traditionally, prognostication in critical illness has relied heavily upon measures of acute physiologic derange ments present at or within 24 hours of ICU admission -that is, Acute Physiology and Chronic Health Evaluation II [18], Sequential Organ Failure Assessment [19], Simplifi ed Acute Physiology Score II [20] -and has modeled illness severity to the estimate probability of survival [21][22][23]. Th ese scoring systems incorporate a limited assessment of sociodemographic characteristics (that is, age, social support, education and comorbidity) and do not integrate any signifi cant measures of prehospital functional status, scope or severity of comorbid illness, disability or frailty. Addressing these limitations is particularly important when considering long-term outcomes (that is, 6 or 12 months) following critical illness. Th e increasing recognition of poor intermediate and long-term outcomes of critical illness -including not only survival, but also functional status, institutionalization and quality of life -coupled with the huge fi nancial cost of critical care therapy means that better tools to predict those patients who will benefi t most from critical care treatment are urgently needed [24].
Th ere is increasing evidence to suggest that physiologic reserve may be an important determinant of clinical outcome in critically ill patients and that baseline functional status and the burden of pre-existing comorbid illness have prognostic value [18,[25][26][27][28]. We currently have no method, however, to estimate this physiologic reserve or capacity to heal in critically ill patients. Moreover, when considering the increasing age and complexity of patients admitted to the ICU, advances in life-support technology and changing societal expectations for recovery, accurate prognostication in the ICU becomes very emotionally charged and challenging [11].
A prospective multi-center study of 980 survivors of critical illness found recently that pre-existing comorbid disease was the strongest predictor of post-ICU quality of life [28]. Additionally, simple measures of burden of preexisting disease and global function, such as residency in a nursing home facility, have also been shown to correlate with mortality [29]. Frailty explicitly captures this sort of functional dependence in an easily comprehensible and more descriptive fashion, but has not yet been evaluated in the critically ill patient. Owing to the potential similarities in frailty between geriatric patients and critically ill patients, the concept of frailty in critical illness may have clinical relevance, be independently predictive of outcomes and show interaction with several factors, including illness severity, comorbid illness, and the social and structural environment. We contend that studies of frailty in critically ill patients are needed to evaluate how frailty correlates with clinical outcomes such as survival and quality of life, but also how frailty correlates with resource utilization, such as lengths of mechanical ventilation, ICU stay and duration of hospitalization. We also believe that if frailty is proven to have clinical and prognostic relevance, its objective measurement may provide additional support and People who have no active disease symptoms but are less fi t than those of category 1. Often, they exercise or are very active occasionally (that is, seasonally).

3
Managing well People whose medical problems are well controlled, but are not regularly active beyond routinely walking.

Vulnerable
While not dependent on other for daily help, symptoms often limit activities. A common complaint is being slowed up, and/or being tired during the day.

5
Mildly frail These people often have more evident slowing, and need help in high-order independent activities of daily living (fi nances, transportation, heavy housework, medications). Typically, mild frailty progressively impairs shopping and walking outside alone, meal preparation and housework. reinforcement to critical care clinicians engaged in goals of care planning and/or end-of-life decisions [11].

How might we quantify frailty in the critically ill patient?
Time constraints necessitate a tool that is simple to under stand and easy to administer rapidly, if the tool is to be used clinically on admission to the ICU. Furthermore, the required information must be obtainable from friends and family, as the patient is often unable to participate in the assessment in an active way. Th ese make the commonly-used operational defi nition of Fried and colleagues diffi cult to apply. Th e Frailty Index is also diffi cult to incorporate into a busy critical care practice, as the degree of detail required makes the tool cumbersome and time consuming to use. On the other hand, the CFS is readily available at the bedside and is easier to understand and use than other frailty assessment tools. Consequently, the CFS may be the optimal tool for use on admission to the ICU. Furthermore, while the CFS is judgment based and has some subjectivity, it captures a spectrum of information that transcends several aspects of a patient's premorbid health state. Th is fl exibility is also likely to be advan tageous and has been validated in other clinical settings [1,30]. Th e CFS has now been adapted and validated for administration by health research coordinators and by telephone interview, making it practically useful in the critical care setting where obtaining collateral history from family members and friends is an integral part of the informationgathering process [31].
Recognizing that frailty is not a static state, evaluating the patient for frailty in the recent past and quantifi cation of developing frailty during hospitalization in the ICU may also add considerable predictive power to the assessment. In addition to quantifying admission frailty, obtaining a historical point estimate of frailty by retrospectively administering the CFS through friends and family regarding function in the recent past could create an estimate of premorbid health trajectory. Additionally, by tracking features related to Fried's operational defi nition of frailty during the ICU stay (such as weight, nitrogen balance, adequacy of caloric supplementation, walking distance, upper extremity strength and selfreported exhaustion), one may be able to create an objective measure of healing that to date is limited to serial administration of admission prognostic scores [14]. Such an assessment that includes both premorbid trajectory and response to critical care intervention would be very useful both for physicians with respect to individualization of prognosis and for families during end-of-life discussions, by providing objective, easily comprehensible critical care benchmarks for response to treatment.

Are there therapeutic implications for frailty?
Previous studies looking at multi-dimensional inter ventions to prevent adverse events in older patients have shown promise [32,33]; patient deterioration after the completion of the trial was common, however, and inferences may have been limited due to lack of an agreedupon defi nition for frailty. Recognition of the multifaceted nature of frailty has recently led to investigation of multi-dimensional home-based interven tions intended to interrupt the vicious cycle of frailty. In the ongoing British Frailty Intervention Trial, individualized nutritional, social, psychological and physical interventions targeted at frailty are being evaluated in a group of older adults who are considered frail by the operational defi nition proposed by Fried and colleagues [2,34]. Th ese interventions include nutritional intake analysis, home meal delivery and high-calorie/high-protein meal supplementation, day activity groups, psychiatric referral and home physiotherapy. It is hoped that this multi-faceted approach in a validated high-risk frail patient population will be eff ective.
Similarly, the importance of adequate nutritional support [35], the value of sedation interruption [36] coupled with early mobilization [37] and physiotherapy [38] to prevent physical deconditioning, and the psychological consequences of critical illness for both patients and their caregivers [39] are being increasingly recognized in the ICU setting. Since single interventions have historically had limited success in altering critical care outcomes with a few notable exceptions [40][41][42], a more eff ective approach may be to stratify critically ill patients based on frailty and intervene in a similarly-styled multidisciplinary way that targets multiple facets of the vicious cycle of frailty. It is con ceivable that the preexisting and/or newly developing frailty modifi es the potential attributable benefi t of timely and eff ective acute physiologic support in the critically ill patient. Accordingly, we hypothesize that a better character i zation of pre-existing frailty and its ongoing development may represent a novel method for risk identifi cation and stratifi cation for future clinical and therapeutic interventions in critical illness.

Conclusion
Frailty is common in geriatric populations and has shown clear association with risk of death and institution alization. Th e burden and potential modifying impact of frailty on the course and outcomes in critically ill patients is unknown.
Although not yet clearly established in the ICU population, we believe that frailty has clinical relevance and may predict both short-term and longterm outcomes. Th e validation of available frailty instruments, such as the CFS, in critical care settings would be an important fi rst step.
If the measures of frailty are proven to have compelling prognostic value, such evidence could then be used for risk identifi cation for novel therapeutic interventions or could potentially be integrated into clinical decisionmaking -not only at the bedside, but also at a health policy/societal level. Similarly, the inclusion of measures of frailty into cost-utility analyses would aid in identifying subgroups of ICU patients for whom the ICU would be least likely to preserve quality-adjusted survival and/or functional independence. Furthermore, therapeutic strategies in the ICU designed to minimize the development and consequences of frailty may have signifi cant benefi cial eff ects on utilization, cost and eff ectiveness of ICU support.
Until recently, the main thrust of critical care diagnosis and management has been on the acute processes leading to homeostatic imbalance. A paradigm that includes a better understanding of frailty may cause a fundamental shift of focus, with the diagnosis, treatment and prevention of frailty being considered equally as important as acute physiologic support for critical illness.

Competing interests
The authors declare that they have no competing interests.