Frailty is an increasing problem in an ageing surgical population.
It is strongly correlated with adverse surgical outcomes, and may be a better predictor of complications and mortality than age itself.
Until recently, frailty has been difficult to define and measure robustly in the clinical setting.
Definitions of frailty have recently been developed, along with tools for measuring the severity of the frailty syndrome in clinical practice.
The aetiology of frailty syndrome is currently contested, but is likely to be multifactorial.
Clinical interventions have been shown to treat and reverse components of the frailty syndrome.
Multicomponent interventions targeting the frailty syndrome (e.g. exercise therapy, prehabilitation, nutritional optimisation and medical interventions) are likely to improve surgical outcomes.
Many experienced clinicians know frailty when they see it. For centuries, frailty has been described in older people, but it is only recently that knowledge of this clinical condition has evolved and allowed us to define, measure and study frailty as a clinical entity in its own right (Fried et al., 2001; Rockwood et al., 2005).
It has long been recognised that older patients undergoing surgery have poorer outcomes than their younger counterparts. However, it is increasingly clear that frailty may be more important than chronological age in determining outcomes from surgery. Biological age and chronological age are not always equivalent, and the rate at which individuals progress through the ageing process is heterogeneous. Professional experience tells us that an 85-year-old marathon runner is likely to tolerate surgery better than a frail 70-year-old. However, there is a vast spectrum where the presence, absence or severity of frailty is less clear.
In this chapter we will review the role of frailty in predicting the risks of surgery. We will also provide a brief overview of existing tools and metrics used to identify and measure frailty. Finally, we will overview the existing evidence base for modification and reversal of the frailty syndrome.
It Is Increasingly Common
Degenerative and neoplastic conditions are strongly correlated with age, and are therefore anticipated to increase in frequency. Older people increasingly expect access to gold-standard treatment, which in many cases necessitates surgery. It is therefore no surprise that in recent years the average age of the surgical patient has increased. More than 50 per cent of the operations now performed in the United States are on patients aged over 65 (Kim et al., 2014). By 2039, UK population projections anticipate a 50 per cent increase in the population of people aged 85 years and over (ONS, 2015). Ten per cent of community-dwelling persons aged over 65 can be considered frail; in community-dwelling persons aged over 80, this increases to up to 50 per cent (BGS, 2014). Surgical disease itself predisposes to frailty, and therefore the prevalence of frailty in older surgical patients is typically 30–40 per cent (Chen et al., 2015). It is often higher in selected emergency surgical populations.
Frailty Is an Important Predictor of Surgical Outcomes
It is well established that older patients suffer increased rates of complications, mortality and functional decline than their younger counterparts as a result of surgery. (Partridge et al., 2012; Robinson et al., 2013). These probably relate to a range of patient and healthcare system factors, which act with synergistic effect. Co-morbidity, and specifically multi-morbidity, is recognised to accumulate with advancing age. Undoubtedly, this is an important contributory factor. However, it is increasingly recognised that frailty is a strong, independent risk factor for adverse surgical outcomes. It may even be better than age alone in predicting adverse outcomes (Joseph et al., 2014).
Compared to the robust, frail surgical patients have elevated post-operative mortality and significantly increased risks of post-operative complications, inpatient length of hospital stay and complex discharge to an institution other than their home (Kim et al., 2014).
Frailty screening in the preoperative setting is becoming increasingly mainstream.
1 Risk Stratification
It is evident that surgical risk is not only associated with specific co-morbidity, and that age in itself is not necessarily a strong independent risk factor. The presence or absence of frailty is an important factor for predicting longer-term outcomes from surgery, including post-operative cognitive disorders, functional decline and loss of independence (Rowe et al., 2014). In an evolving medico-legal climate, managing patient and family expectations from surgery is essential to allow fully informed consent.
2 Medical Optimisation and Prehabilitation
Frailty is a complex entity with many component parts. Screening for this syndrome reveals many factors which may be amenable to optimisation. Evidence is accumulating that this may reduce the risks of surgery in frail older adults (O’Doherty et al., 2013; Gillis et al., 2014). As the evidence further evolves, it seems likely that clinical outcomes will be improved by selection of higher-risk patients and enrolling them into preoperative optimisation programmes.
However, not all older patients are frail, and interventions for frailty are likely to require considerable resources. To address frailty in the preoperative setting, we need effective strategies to identify and target vulnerable patients for specific interventions.
How Can We Define Frailty?
Moving away from ‘End-of-the-Bed’ Assessment
Surgeons and anaesthetists have always subconsciously assessed frailty, but traditionally this has been by means of an underlying gut instinct honed over years of experience. Experienced clinicians may well be able to identify frailty accurately on this basis, but the end-of-the-bed approach is highly subjective. While it may be easy for even inexperienced doctors to differentiate between profoundly frail and robust individuals, frailty is a spectrum and this is reflected in considerable heterogeneity in frailty diagnosis between assessors.
Ultimately, frailty can be considered a syndrome of decreased physiological reserve across a range of organ systems, rendering the frail individual vulnerable to physiological decompensation when subjected to external stressors (Figure 14.1).
Figure 14.1 Impact of physiological reserve on response to external stressors.
The Frailty Phenotype
However, a universal definition and diagnostic criteria for frailty has not yet been unanimously agreed. The first model to be proposed was the frailty phenotype, as described by Fried and colleagues (2001) and derived from the Cardiovascular Health Study. This model considers frailty a binary entity defined by the presence of three or more of the following characteristics:
a. Self-reported exhaustion (on questionnaire correlating effort and activity)
b. Weakness of grip strength
c. Unintentional weight loss (>5% body weight in the preceding year)
d. Slowness of gait speed (>6 seconds to walk 15 feet)
e. Low levels of physical activity (weekly energy expenditure in fifth quintile: males <383Kcal; females <270Kcal).
Patients displaying three or more of these parameters were observed to be at increased risk of frailty-associated outcomes including declining mobility, falls, dependence in activities of daily living, hospital admission and death (reviewed Partridge et al., 2012). It has been suggested that these problems may be born of physiological insufficiency characterised by deconditioning, sarcopenia, balance disturbances, osteopenia and nutritional deficits (Topinkova et al., 2008). Many of these parameters are simple to identify; by having a measureable variable, they provide information that can be monitored chronologically.
However, there are limitations to such a simple model of frailty. Criticism has been raised that the phenotype excludes important characteristics, such as cognitive dysfunction and physical function. These intuitively seem to be closely associated with our understanding of frailty as a syndrome of reduced physiological (and especially neurophysiological) reserve. Secondly, specific medical conditions, especially neurodegenerative diseases such as Parkinson’s or stroke, are characterised by gait impairment, weakness and low levels of activity. It is easy to see how a patient we would not otherwise consider physiologically frail can be abruptly rendered frail according to phenotypic criteria.
The Frailty Index
A response to these concerns can be seen in the evolution of the frailty index described by Rockwood: the ‘Cumulative Deficit Model’ (2005). This model of frailty considers there to be a wide and diverse range of parameters (70 in total) including a range of medical and psychosocial factors. The extent to which these have been accumulated contribute to overall frailty in an individual. The precise extent to which an individual has accumulated these various deficits allows calculation of a frailty index ranging from 0–1. An index of 0.25 was epidemiologically associated with adverse frailty outcomes. It can therefore be considered as a cut-off defining frailty, though risk is closely correlated with overall score. Index models incorporate a wide variety of contributory factors and therefore their breadth can prove a useful screening tool. However, some versions of the frailty index have included up to 92 possible deficits. While this may be feasible in the research setting, concerns exist as to the practical application of such an index. Subsequent refinements have been made to include only the most strongly prognostic parameters (BGS, 2014; Jones et al., 2005).
How to Measure Frailty in Clinical Practice?
In clinical practice, detection of frailty using the frailty phenotype or index is cumbersome. This has resulted in the development of modified scales and tools designed for clinical practice. The literature also reports on the use of multiple metrics that are closely correlated with frailty (e.g. grip strength) which have been used as proxy, surrogate measures for detecting frailty. Strictly, these, however, are not synonymous with frailty itself.
Which Tool to Use?
There is currently insufficient evidence to unequivocally determine which clinical tool is best to identify frailty in the preoperative setting. However, several tools are emerging which bear promise. The selection of an optimal preoperative frailty tool depends on a number of factors. Certain tools have been shown to more accurately reflect the adverse prognosis of frail persons in a population, though not all have been studied in surgical populations. There is also variation depending on the nature of the planned surgery, and whether the information is used purely for risk assessment, or as an adjunct towards a personalised prehabilitation programme.
Recent developments have revealed the potential role for sarcopenia measurement as a predictor of adverse surgical outcome. Computed tomography (CT) combined with analytic morphomics software can determine skeletal muscle mass. Data have linked CT-derived sarcopenic parameters with surgical outcomes, including survival, complications and cost in patients undergoing abdominal, transplant and vascular surgery (Friedman et al., 2015). This is an appealing strategy in patients who need to undergo CT imaging for staging or surgical planning. The cost implication and radiation exposure do make it less appealing in other settings.
A more simplistic approach has been to use basic physiological parameters that can be measured in an outpatient clinic. Examples include gait speed, timed-up-and-go and grip strength, all of which have been shown to correlate well with overall frailty scores. However, they have limitations in specific surgical settings (e.g. orthopaedics) where the primary pathology is likely to selectively affect these surrogates and fail to select out the truly frail individual.
The Edmonton Frail Scale (EFS) has been shown to correlate with post-operative mortality and complications in a mixed surgical population (Dasgupta et al., 2009). The scale is derived from an abbreviated frailty index and screens for frailty across a number of domains that have been shown to be key components of frailty. It has been shown to be quick and simple to complete, and requires little specialist training. It has good inter-rater reliability, and has been studied in surgical cohorts, making it one of relatively few composite tools clinically validated in this setting (Table 14.1).
Data now exist which directly correlate EFS score with surgical outcomes. In surgical patients, EFS scores of >8 were associated with increased mortality, complications and discharge to an institution other than home. In vascular surgical patients, a score of >6.5 predicted length of stay of >12 days, increased rates of post-operative infection, medical complications following surgery or poor functional outcomes (Dasgupta et al., 2009; Partridge et al., 2015).
In the United States, the Johns Hopkins Frailty Assessment Tool has been studied in surgical cohorts. This tool can be completed in approximately 10 minutes, but requires access to a handheld dynamometer, which may limit its application in a wider clinical setting. Patients are scored on a scale of five parameters, each of which carries an equally weighted score leading to a maximum of 5:
Weight loss (10lbs in previous year)
Poor strength (as determined by a handheld dynamometer)
Low levels of physical activity
Slow gait speed
A patient is deemed frail with a score of 4 or more. Intermediately frail patients score 2–3. When applied to planned surgical admissions, patients deemed frail using this scale were at greater risk of post-surgical complications. Comparing frail to non-frail patients following major surgery; inpatient stay was 3.5 days longer, post-operative complications more than doubled and institutionalised care at the point of discharge was increased 14-fold (Makary et al., 2010).
Which tool to use in clinical practice will depend on the specific surgical setting and may evolve as new evidence emerges. A frailty assessment tool that offers multiple targets for optimisation is an advantage, and our current clinical practice finds the EFS to be most suitable for our needs at present. Future screening for frailty may be achieved through coded population primary care data, though this is not currently available (Clegg et al., 2013).
Why Do People Get Frail?
Before considering how we can treat the frailty syndrome, it is important to consider why older people become frail.
The underlying pathophysiology resulting in frailty may be a dysregulation of the ageing process resulting in accelerated ageing and the onset of a pathological state of homeostatic failure. This loss of physiological reserve leads to a failure to mount an adequate response to stressors. Frail older persons present atypically in response to relatively minor physiological stressors. The Geriatric Giants are a cluster of geriatric presentations (instability, immobility, falls, incontinence and delirium) which reflect the breakdown in complex neurophysiological function in frail older adults. Such breakdowns typically occur secondary to a physiological stressor. Medical students have long recognised that low-grade infection, such as UTI, can have this effect, but so too the inflammatory response of surgery and drugs of anaesthesia.
However, it remains unclear which processes lead to loss of physiological reserve and the ensuing state of frailty. Rowe and colleagues (2014) stated that ‘apoptosis, senescence, repair, inflammation and immune activation have all been implicated as pathways responsible for this decline.’ Other theories of ageing and frailty relate to endocrine observations that have been made in frail persons, especially those with sarcopenia. These associations include low IGF-1, low adrenal DHEA-S, high afternoon cortisol levels and low 25(OH) D. From a metabolic perspective, ageing mitochondria and alterations in the renin-angiotensin-aldosterone axis have been implicated. Other contributory factors to frailty include loss of autonomic regulation and alterations within the ageing immune system such as increased inflammatory cytokine levels and altered clotting factors (reviewed in Walston et al., 2016). This is an active field of current research, and many of these observed associations with frailty have been employed as targets for specific frailty interventions (Table 14.2).
|Low sex-hormone binding globulin||Elevated CRP|
|Low adrenal DHEA-S||Elevated Il-6|
|Low vitamin D||Elevated WCC (and monocytes)|
|High cortisol||Elevated TNF-α|
Legend: CRP= c reactive protein, DHEA= dihydroepiandrosterone, IL=interleukin, WCC=white cell count, TNF=tumour necrosis factor
How Can Frailty Be Treated?
Longitudinal studies have observed that not all frail persons progress towards more advanced frailty and death: a small minority spontaneously becomes less frail with time. Research suggests that frailty may be reversible and that specific interventions may be able to enhance this process. A clear evidence base to address frailty in its entirety does not yet exist, though there is established evidence for treatment of components of the frailty syndrome. Until wider research has addressed this important question, interventions for frailty are best directed at individual components or outcomes of frailty. Common targets for intervention are discussed in what follows.
Exercise is well recognised to have multiple wider health benefits, including in frail persons, in whom both the onset and progression of frailty is delayed by exercise participation (Peterson et al., 2009). Gait speed is commonly used as a proxy measure for frailty, and improved speed has demonstrated increased survival. Resistance exercise is known to improve muscle mass, gait speed and exercise tolerance in frail patients residing in care homes. Moderate to intense exercise programmes can improve functional outcomes and prevent the accumulation of disability. The impact on exercise in obese frail patients has been studied and shown not only to improve parameters of physical fitness, but to reverse frailty (Villareal et al., 2006).
The role of exercise in the preoperative evaluation and optimisation of the frail older surgical patient is currently undergoing evaluation. The anaesthetic literature has long recognised the predictive role of anaerobic threshold and maximal oxygen consumption in surgical outcomes. Prehabilitation programmes can improve these parameters of cardiorespiratory fitness, and early data suggest this may additionally translate into reduced length of stay and reduced use of critical resources (O’Doherty et al., 2013). Data are currently limited, and uncertainty exists regarding the extent, type and quantity of exercise needed to improve outcomes among different surgical populations.
Most studies evaluating nutritional optimisation of the frail patient have used weight loss or grip strength as outcome measures. Though not entirely synonymous with frailty itself, these are key components of the frailty phenotype. Meta-analysis evidence supports oral nutritional supplementation combined with dietetics advice in frail older adults. In addition to weight gain, a small beneficial effect on all-cause mortality has also been observed in community populations.
Many ageing outcomes associated with frailty, such as increased hospitalisation and mortality, are closely linked with low protein intake, and these observations have resulted in revisions to recommended daily protein intake in frail elders. The precise recommended daily protein intake for older adults who may be subject to changes in protein metabolism with advanced ageing is a matter for debate.
In the perioperative setting, nutritional impairment is clearly associated with adverse surgical outcomes. Preoperative nutrition in at-risk individuals can reduce complications and length of stay (Schiesser et al., 2009). Guidelines now exist to support preoperative nutrition in impaired individuals, including in extreme circumstances the potential need to delay surgery for oral, enteric or parenteral feeding (Weimann et al., 2006).
Other avenues to optimise frailty through nutrition may focus on anaemia. This is well known to contribute to fatigue: another key component of the frailty phenotype. Consensus guidelines recommend replacement of deficient iron, folate and vitamin B12 before surgery, though the impact of this on frailty itself has not been studied. However, preoperative treatment of anaemia has been shown to reduce both morbidity and mortality (Goodnough et al., 2011).
Multiple associations have been made between various endocrine factors and frailty. Therapeutic treatment of these observed deficiencies has been attempted in some trials. An example includes testosterone supplementation for frail older men, which has been shown to successfully improve strength and muscle mass in this population. However, this has not been consistent and the adverse side effect profile prevents widespread use. Growth hormone and exogenous steroid precursors have also failed to show any advantageous effect (reviewed in Walston et al., 2016).
Vitamin D supplementation may play a future role in the wider treatment of frailty. This has a well-recognised role in muscle function and postural stability, and been shown in large-scale studies to reduce falls in older adults. Falls are commonly considered to be an indicator of frailty, and one of the Geriatric Giants.
Low-serum vitamin D is closely correlated with frailty, and replacement studies have been shown to improve proxy parameters of frailty such as sarcopenia. This may be expected to contribute to improvements in components of the frailty phenotype such as grip strength and gait speed, though direct trial evidence showing a reversal of frailty per se is not yet available.
Appetite stimulants such as megestriol acetate have been used in cancer patients and frail residents of care homes. Meta-analytical data for these specific populations have been shown to improve appetite, weight gain and quality of life, and they may therefore have a future role in the treatment of nutritional aspects of frailty. However they can be associated with fluid retention in congestive cardiac failure. Robust evidence to support their widespread use is not currently available, though in targeted patients, particularly surgical cancer patients with anorexia, their use may be justified (Berenstein et al., 2005).
Other pharmacological interventions to target frailty include the treatment of severe depression. Though this is not part of the frailty phenotype model, it is a component of the frailty index and it has independently been shown to impact surgical outcomes. Selective serotonin reuptake inhibitors (SSRIs) have been associated with increased risk of perioperative bleeding, and therefore where treatment for depression is being considered, an alternate class of antidepressant is recommended in the frail patient (Gartner et al., 2010).