Patient frailty: the elephant in the operating room



For most surgeons and many anaesthetists, patient frailty is currently the ‘elephant in the (operating) room’: it is easy to spot, but is often ignored. In this paper, we discuss different approaches to the measurement of frailty and review the evidence regarding the effect of frailty on peri-operative outcomes. We explore the limitations of ‘eyeballing’ patients to quantify frailty, and consider why the frail older patient, challenged by seemingly minor insults in the postoperative period, may suffer falls or delirium. Frailty represents a state of increased vulnerability to stressors, and older inpatients are exposed to multiple stressors in the peri-operative setting. Quantifying frailty is likely to increase the precision of peri-operative risk assessment. The Frailty Index derived from Comprehensive Geriatric Assessment is a simple and robust way to quantify frailty, but is yet to be systematically investigated in the pre-operative setting. Furthermore, the optimal care for frail patients and the reversibility of frailty with prehabilitation are fertile areas for future research.

(While I can't define it)…I know it when I see it. — Justice Potter Stewart, US Supreme Court

Clinicians usually recognise a frail patient, who is often, but not always, elderly. However, at best, we usually quantify frailty in a largely subjective ‘end-of-the-bed’ or ‘eyeball’ way [1]. Furthermore, although studies have examined patient factors such as age and ASA physical status and their association with peri-operative complications and mortality [2-4], until very recently, few have examined frailty [5]. Frailty increases vulnerability to stressors, which, in the peri-operative setting, can include the procedure(s), complications and medical care administered.

The aims of this paper are to review the emerging science around defining and quantifying frailty, and to examine its associated peri-operative risks. Current evidence suggests that quantifying frailty improves the precision of risk assessment [5]. However, we think that, for most surgeons and many anaesthetists, frailty and its associated risks represent an ‘elephant in the (operating) room’: it is easy to spot, but is often ignored.

What is frailty and who is frail?

‘Frailty’ has become the focus of considerable research interest among geriatricians and gerontologists. A PubMed search of the terms ‘frail’ + ‘elderly’ yielded 20 publications before 1980, but 4300 since 2003. Indeed, a journal entirely devoted to the investigation of frailty (The Journal of Frailty and Aging) was established in 2012. Medically, the term ‘frail’ identifies patients with a diminished capacity to compensate effectively for external stressors, who are at greater risk of adverse outcomes, including prolonged hospital stay, institutionalisation, worsening disability and death [6, 7].

The pathophysiology of frailty has not yet been established, but there is a growing body of evidence linking frailty with inflammation [8]. Although this association does seem consistent across different frailty definitions [9], further studies are needed to establish the exact nature of the relationship. Inflammation may be causal (in which case, anti-inflammatory strategies would be desirable), compensatory in response to viral antigens or subclinical disease (for which anti-inflammatory strategies would be undesirable) or an epi-phenomenon, merely a marker of another primary pathophysiological process, such as excessive oxidative stress (where anti-inflammatory strategies would be irrelevant). More recently, frailty has been most strongly associated with a combination of immunological and physiological impairments rather than a single biomarker. In the Women's Health and Ageing Study in North America, analysis of several variables found that the likelihood of frailty increased non-linearly in relation to a number of abnormalities, including anaemia and inflammation, and derangements of insulin-like growth factor-1, dehydroepiandrosterone sulfate, haemoglobin A1c, micronutrients, adiposity and fine motor speed [10]. Large, epidemiological studies of community-dwelling older people have linked frailty to several extrinsic factors, notably smoking [11, 12], low levels of physical activity [13] and financial deprivation [14, 15]. Frailty is therefore unlikely to have a linear deterministic chain of causation, but rather exemplifies a disorder that is the stochastic (partly random) outcome of multiple subcellular events that occur secondary to the effect of lifestyle and environmental factors combined with genetic susceptibility [16].

The prevalence of frailty increases with chronological age up to ~95 years of age, when all individuals are frail and therefore vulnerable to adverse surgical outcomes [17]. It is pragmatic, rather than ageist, to take date of birth into account when considering the risks of surgical intervention. Geriatricians think of older people in terms of young old (65–75 years), old old (75–85 years) and oldest old (aged over 85 years) because of the increasing vulnerability across these age groups. Anecdotally, many surgeons rue their decision to offer surgery to the oldest old when adverse outcomes result, such as failure to wean from ventilation. Gestalt, or ‘eyeball’, assessments (‘but he was so good for 87!’) fail to account for the ubiquity of physiological decline in the oldest old.

Measurement of frailty in clinical practice


A variety of measurement tools identify frailty as a clinical syndrome or phenotype, that is, a set of signs and symptoms that tend to occur together, characterising a specific medical condition. Several algorithms derived from clinical judgment have been validated against adverse outcomes, including physical inactivity and weight loss [18], the FRAIL scale of fatigue, resistance, ambulation, illness, loss of weight [19] and the Easy Prognostic Score (age, sex, daily drug use, sensory deficits, physical inactivity, calf circumference, independent activities of daily living, gait and balance, pessimism about one's health) [20]. Fried et al. defined the most well-known and widely used phenotype [21], which identifies frailty as the presence of three or more out of five criteria – weight loss, exhaustion, weak grip strength, slow walking speed and low physical activity. These criteria were initially categorised by quintile, such that those with the lowest 20% of physical activity, for example, fulfilled that criterion, an approach that correlates well with population-independent, literature-derived cut-offs [22]. Fried's model, however, is a predominantly ‘physical’ conceptualisation, identifying frailty as a wasting process, with the degenerative loss of skeletal muscle mass (sarcopenia) as a key pathophysiological feature, and omits disorders of cognition and mood.

Ideally, a peri-operative frailty score should be quick and easy to calculate when administered by non-geriatricians, and highly predictive of complications, mortality and level of care after hospital discharge. Although phenotypic frailty measures (observable morphology, biochemistry and physiology) have been derived and investigated in many thousands of community-dwelling older people, their inpatient clinical utility is unclear [23]. Although the Fried criteria are the most widely used currently, many of the most vulnerable older inpatients are unable to mobilise for gait speed estimation, hence cannot be stratified accordingly [24].

Gait speed and exercise testing

Measures of frailty that depend on physical performance-based tests have limited applicability in the peri-operative setting. Nevertheless, slow gait speed has been used in isolation as a marker of frailty, is the strongest predictor among Fried's criteria of chronic disability and institutionalisation, and is the only significant predictor of injury from falls [25]. In community-dwelling older people, those who walk more slowly are at greater risk of adverse outcomes, including disability, cognitive impairment and mortality [26].

Slow gait speed is likely to be a reflection of cerebrovascular and chronic neuronal disease, being associated with white matter hyperdensities and with grey matter changes in the medial temporal area that are linked with memory [27]. Associated smaller volume of the prefrontal area (a region particularly vulnerable to hypertensive damage) contributes to slow walking through impaired information processing [28]. These features provide a mechanistic understanding of the association between impaired executive functioning in the Walking While Talking Test and later development of frailty [29].

Although some patients cannot physically perform the tests, gait speed measured over 5 m has been found to be an independent predictor of mortality and major morbidity in older patients undergoing cardiac surgery, and is increasingly used to stratify risk in these cohorts [30].

Aerobic capacity

Although a relationship between gait speed and frailty is clearly established, less is known about the association between frailty and aerobic capacity. Maximal aerobic capacity declines by approximately 10% per decade aged 20 onwards, secondary to reductions in maximal heart rate, ejection fraction and loss of cardiac muscle mass. However, up to 50% of this decline may be due to deconditioning rather than ageing [31]. Cardiopulmonary exercise testing (CPET) is a well-established stratification tool, particularly for complex surgery, such as heart transplantation [32]. The validity of CPET in other settings, such as vascular surgery, has been questioned [33], and it is not routinely available in many hospitals [34]. Furthermore, only the fittest of the oldest old are able to undertake any form of CPET – 55% of men and 72% of women in this age group report severe limitations with even basic activities of daily living and/or are unable to climb a single flight of stairs [35].

Subjective opinion

The clinician's subjective opinion can provide an alternative to phenotypic and performance-based assessments of frailty, and is consistent with the aphorism that frailty is difficult to define, but easy to identify. Several scales have been developed and validated against adverse outcomes, such as institutionalisation and death [36]. Experienced geriatricians show good inter-rater agreement in the estimation of patients’ biological age [37] and these subjective measures have strong construct validity. The Clinical Frailty Scale [36] can predict outcomes from severe burns in older people [38]. However, inter-rater agreement, and therefore utility, decreases when frailty scales are used by clinicians not trained in geriatric medicine, particularly as an accurate assessment of functional status underpins many frailty scales. Furthermore, risk is context-dependent. Geriatricians may judge that a patient does not have the physiological reserve to benefit from a procedure such as coronary artery bypass graft, but may advocate for the same patient to receive elective hip replacement, and even the most frail may benefit from cataract surgery, or from inpatient rehabilitation. As the ‘frailty threshold’ is decision-specific, subjective scales may be too simplistic to inform decision-making about whether or not a patient should undergo surgery.

‘Eyeballing’ patients to estimate frailty status is fraught with other potential pitfalls. Frailty is often viewed as synonymous with cachexia and sarcopenia, yet identifying only those with a low body mass index (BMI) as vulnerable would omit many high-risk patients. In the English Longitudinal Study of Ageing, the association between BMI and frailty showed a U-shaped curve, with the lowest prevalence of frailty in those with BMI 25–29.9 kg.m−2 [39]. Similarly, assigning frailty status according to medical diagnoses or prescribed medication count may be misleading. Frailty is related to, but distinct from, co-morbidity, as there is considerable variation in how older people tolerate disease, even when illness severity is taken into account [40]. Likewise, polypharmacy does not necessarily indicate frailty, although frailty has been linked to excessive polypharmacy in some community dwellers [41]. By comparison, patients with advanced dementia are often prescribed relatively fewer medications [42], but are at higher risk for important complications, including delirium [43], and failure to consider patients with dementia as frail and vulnerable compromises clinical care.

Deficit accumulation

Rather than assessing frailty as a global phenotype, another way to quantify frailty is to consider accumulated deficits, that is, viewing frailty as a multidimensional risk state and assessing the quantity, rather than the quality, of health problems accumulated throughout patients’ lives [44]. The more ‘deficits’ accumulated, the higher the likelihood of frailty [45, 46]. The Frailty Index model uses well-defined methodology to create a proportional index of deficits; for example, someone with 6 deficits out of 40 has a Frailty Index of 0.15 [47]. This approach has been extensively investigated in community-dwelling populations throughout the developed and developing world. Among the 4721 participants of the Cardiovascular Health Study, for example, the risk of adverse outcomes was defined more precisely by deficit indices than by the frailty phenotype [48]. In Chinese studies, involving 2000 and 13 700 older adults, respectively, Frailty Indices predicted mortality risk [49] and type of death (bedridden > 30 days <, with/without suffering) [50]. More recently, in almost 30 000 older Europeans, a Frailty Index was found to be a stronger predictor of mortality than chronological age [51].

For institutionalised or very ill populations, frailty levels are high, but no longer accumulate with age [52]. The frailest older people are comparable to a system with no redundancy: they can no longer accumulate deficits (i.e. become more frail), and the next insult will cause failure of the system (i.e. death). This resonates with the redundancy exhaustion hypothesis of the reliability theory of ageing, in which the body fails when its exhausted physiological reserves are challenged [53], and explains why operating electively on someone who seems to be ‘very frail but stable’ still carries a high risk of adverse outcomes. Conceptualising a frail older person as a complex system on the threshold of failure also provides an understanding of geriatric syndromes. When a complex system fails, it fails with ‘higher order functions’ first. In humans, higher order functions, which require a coordinated, integrated and precise interaction between many components, include upright bipedal ambulation, and advanced cognitive processes, such as divided attention. This explains why a seemingly minor insult, such as a urinary tract infection, can present with falls or delirium in the frail elderly patient [54].

Peri-operative implications

Researchers have only recently begun to recognise the importance of frailty as a prognostic indicator of postoperative outcome and discharge destination. Measuring frailty might provide more accurate risk stratification and facilitate better care provision, be it interventional or otherwise, for older patients. The National Surgical Quality Improvement Program (NSQIP) of the American College of Surgeons collects comprehensive data on large numbers of patients in more than 450 hospitals throughout the USA. In comparing the Canadian Study of Health and Aging Frailty Index with the NSQIP data set, Velanovich et al. identified 11 pre-operative variables collected by each [55] and translated these into a modified (deficit accumulation) Frailty Index (Table 1). They found that 30-day morbidity and mortality among almost a million patients who underwent surgery between 2005 and 2009 correlated with the modified Frailty Index across all surgical specialties. This effect was more pronounced when the relative surgical risk was low, the converse of which the researchers confirmed in a second study of patients undergoing very complex surgery, in whom complications and mortality were less well correlated with frailty [56]. Perhaps of greater clinical relevance, when the researchers analysed data from 58 000 patients undergoing open and laparoscopic colectomy, they found that the proportion of patients experiencing complications requiring intensive care admission correlated with increases in the same modified 11-point Frailty Index [57]; laparoscopic colectomy was associated with lower complication rates among frail patients.

Table 1. National Surgical Quality Improvement Program (NSQIP) items used to derive a modified Frailty Index and corresponding deficit in the Canadian Study of Health and Ageing (CSHA) [55].
History of diabetes mellitusDiabetes mellitus requiring insulin or oral hypoglycaemic therapy

Changes in everyday activity

Problems with getting dressed

Problems with bathing

Problems with carrying out personal grooming

Problems with cooking

Problems with going out alone

Functional health status before surgery partially or totally dependent
History of lung or respiratory problems

History of severe chronic obstructive pulmonary disease

Current pneumonia

Congestive heart failureCongestive heart failure within 30 days before surgery
Myocardial infarctionHistory of myocardial infarction within 6 months before surgery
Cardiac problems

Previous cardiac surgery

Angina within 1 month before surgery

Previous percutaneous coronary intervention

HypertensionHypertension requiring medication

Clouding or delirium

History relevant to cognitive impairment or loss

Family history relevant to cognitive impairment

Acute confusion or impaired sensorium
Cerebrovascular problemsHistory of transient ischaemic attack
History of strokeCerebrovascular accident or stroke with neurological deficit
Decreased peripheral pulses

History of revascularisation or amputation for peripheral vascular disease

Rest pain or gangrene

These findings are supported by other papers, which have used a variety of frailty scales, but have found that increasing frailty correlates with poorer outcomes after surgery. Age and frailty, assessed using the Edmonton Frailty Scale, were both found to be independent factors for complications and institutionalisation after surgery in Canada [58]. Frailty (Fried's criteria), but not ASA ≥ 3, among 84 older, medically optimised Singaporeans was associated with postoperative complications [59]. Frailty (Fried's criteria) was associated with increasing complication rates (adjusted odds ratio 2.5, 95% CI 1.1–5.8), length of stay and institutionalisation among 594 patients in Baltimore [60]. Including additional proxy markers of chronic disease and/or poor nutritional status, such as plasma albumin concentration, may increase the predictive value of pre-operative frailty assessment [61].

A frailty index from comprehensive geriatric assessment

Comprehensive Geriatric Assessment (CGA) is used in geriatric medicine to capture relevant information about the health and function of an older person, facilitating accurate diagnosis, holistic management and effective communication within the multidisciplinary team. The CGA collects data on function, co-morbidities and cognitive/psychological status; all this can be coded as deficits so as to derive a Frailty Index score (the FI-CGA) [62]. In community-dwelling participants of the Canadian Study of Health and Ageing, the FI-CGA was a valid and reliable means of quantifying health status, stratifying patients’ risk of institutionalisation and death [62]. The FI-CGA is feasible for all patients, even those unable to attempt performance-based tests. In 178 consecutive patients with low-trauma hip fracture, a FI-CGA (Table 2) documented 3–5 days postoperatively was a significant predictor of mortality, length of hospital stay and discharge home [63] (Table 3). Although derivation of a FI-CGA from 55 deficits may seem cumbersome, the information should be collected by a treating clinician as part of routine practice. In the hip fracture study [63], the use of a tablet-based spreadsheet, which included prompts for each deficit, automatically calculated the final Frailty Index and took less than 10 minutes to complete, highlights that new technologies may facilitate frailty quantification. Whether the FI-CGA can translate to pre-operative assessment should be the focus of further study.

Table 2. Deficits used to derive a frailty index from Comprehensive Geriatric Assessment in patients with hip fracture [63].
Domain (maximum deficits)Deficit score
Motivation (1)Low = 1, Usual/High = 0
Self-rated health (1)

Excellent = 0

Good = 0.25

Fair = 0.5

Poor = 1

Can't say = 1

Cognition (5)

Dementia = 1, Mild cognitive impairment = 0.5

Delirium = 1

Agitation = 1

Delusions or hallucinations = 1

Abbreviated mental test score ≤ 6 = 1

Abbreviated mental test score > 6 = 0

Emotional state (4)1 point for each of Anxiety, Bereavement, Depression, Fatigue
Sleep (2)

Poor or disrupted = 1

Daytime drowsiness = 1

Communication (3)
SpeechImpaired = 1
Hearing (with/without aid)Impaired = 1
Vision (+/− glasses)Impaired = 1
Strength (4)
HemiparesisWeak arm = 1, Weak leg = 1
Grip strengthWeak = 1 (on non-hemiplegic side)
Proximal muscle strengthWeak = 1 (on non-hemiplegic side)
Nutrition (3)
WeightUnderweight = 1, Obese = 1
 Weight changeLoss = 1, Significant gain = 1
 AppetitePoor = 1, Fair = 0.5
Continence (2)Bowels incontinent = 1, Bladder incontinent/catheter = 1
Medical history (15)1 point for each of Hypertension, Asthma/COPD, TIA/Stroke, Angina/MI, Congestive cardiac failure, Diabetes, Cancer, Alcohol excess, Pressure sore, Hip fracture, Osteoarthritis/rheumatoid arthritis, Osteoporosis, Parkinson's disease, 1st ‘other medical problem’, 2nd ‘other medical problem’
Prescribing (5)
No. of medications/24 h0–4 = 0, 5–9 = 1, 10–14 = 2, 15–19 = 3, 20–24 = 4, 25+ = 5
Mobility (5)
TransfersDependent = 1, Assistance = 0.5
WalkingDependent = 1, Assistance = 0.5
MovementSlow = 1
Sitting balanceImpaired = 1
Falls3 or more in last 6 months = 1
Activities of daily living (3)
FeedingDependent = 1, Assistance = 0.5
WashingDependent = 1, Assistance = 0.5
DressingDependent = 1, Assistance = 0.5
Instrumental activities of daily living (2)
Could have managed own medications pre-admissionDependent = 1, Assistance = 0.5
Could have managed own finances pre-admissionDependent = 1, Assistance = 0.5
Total score for patient[e.g. 20]
Frailty index[e.g. 0.36]
Table 3. Frailty Index from Comprehensive Geriatric Assessment in patients with hip fracture [63]. Values are mean (SD) or proportion.
OutcomeLow FI ≤ 0.25Intermediate FI 0.25–0.4High FI > 0.4
Age; years74 (12)82 (10)86 (9)
Length of stay; days21 (17)36 (23)68 (39)
Discharged home within 30 day80%41% 6%
Inpatient mortality05.2%28.1%

Can frailty be reversed?

‘Prehabilitation’ describes the relatively new concept of exercise training before undergoing surgery. Although there are no proven strategies to reverse frailty, exercise (like optimal nutrition and better education) is of particular therapeutic interest because it involves a complex intervention that may modify deficit accumulation in many systems. Exercise is safe and feasible, even for very frail older adults, and has beneficial effects on multiple domains, increasing muscle strength, improving balance and postponing cognitive decline [64]. However, despite improving fitness, the impact on surgical outcome after prehabilitation programmes involving continuous or interval training remains ill-defined [65]. Whether frail older patients would benefit from longer periods of prehabilitation that combine tailored exercise programmes and nutritional support requires further investigation.


Although ‘frailty’ has long been recognised, clinicians are able to quantify the health status of older patients with increasing precision. However, no method of quantifying frailty, particularly in the peri-operative period, has yet been implemented into routine clinical practice. There are currently two broad approaches: phenotype and accumulated deficit. Slightly complicating this area are different scales with similar names. Of those currently available, the Frailty Index using accumulated deficits [6, 63] approach seems to hold the highest promise, as its assessments are independent of physical performance tests, and better differentiate among the most frail. Accurate peri-operative frailty assessment appears to provide useful additional information about care provision for the frail elderly, based on prognostic stratification of postoperative morbidity and mortality, and discharge destination. Peri-operative research around frailty is evolving, and much work still needs to be carried out for refining and validating frailty scores, optimising care for frail patients and prehabilitating frail patients with exercise training and nutritional support.

As peri-operative physicians, anaesthetists must learn to recognise that, for many older people, ‘frailty’ defines their most feared aspects of the ageing process – wasting, decrepitude, dependency and decline. The Oxford English Dictionary defines frailty rather negatively both as ‘the condition of being weak and delicate’ and as ‘weakness in character or morals’ [66]. In caring for frail patients with respect and dignity, anaesthetists must seek to reassure both patients and their families that frailty in the medical sense, far from indicating a waning moral compass or fading personhood, indicates a useful measure of how high-quality care might best be provided to elderly, vulnerable patients according to their individual needs.

Competing interests

No external funding and no competing interests declared.