Frailty and sarcopenia: findings from epidemiological studies and consensus reports

Frailty

There has been wide agreement among experts in the field that frailty is a distinct clinical entity, with a recent consensus statement defining frailty as⁽ Reference Morley, Vellas and van Kan ^{9 )}:

‘…a medical syndrome with multiple causes and contributors that is characterised by diminished strength, endurance and reduced physiologic function that increases an individual's vulnerability for developing increased dependency and/or death.’

There is also wide agreement that frailty is distinct from disability⁽ Reference Morley, Vellas and van Kan ^{9 ,} Reference Rodríguez-Mañas, Féart and Mann ^{10 )} and co-morbidity⁽ Reference Morley, Vellas and van Kan ^{9 ,} Reference Fried, Ferrucci and Darer ^{11 )}, although all may co-exist. In general, the commonly used operational definitions of frailty lie on a spectrum between two different conceptual approaches to frailty measurement: summation of health deficits to create a frailty index (FI) and measurement of a physical frailty phenotype (PFP). These approaches are summarised below.

In brief, the FI characterises frailty as an accumulation of deficits across multiple body systems, in line with the general concept of frailty as a multi-system disorder⁽ Reference Mitnitski, Mogilner and Rockwood ^{6 ,} Reference Rockwood and Mitnitski ^{12 )}. Any number of health deficits from 30 to 70 can be included, with each deficit carrying an equal weight. Deficits can be symptoms, signs, disabilities, diseases or even laboratory abnormalities and can cover all aspects of health and wellbeing, although deficits should increase in prevalence with age, not saturate too early and be associated with adverse outcomes⁽ Reference Howlett, Rockwood and Mitnitski ^{13 ,} Reference Searle, Mitnitski and Gahbauer ^{14 )}. Frailty is then quantified according to the proportion of deficits present and, although designed to be used as a continuous scale, an index value of about 0·20–0·25 (regardless of age) is usually accepted as the threshold above which frailty is present⁽ Reference Searle, Mitnitski and Gahbauer ^{14 )}. The deficit approach to frailty measurement was pioneered by Kenneth Rockwood and Arnold Mitnitski in the Canadian Study of Health and Ageing⁽ Reference Mitnitski, Graham and Mogilner ^{16 )} but has since been applied in other cohorts⁽ Reference Armstrong, Mitnitski and Launer ^{15 ,} Reference Goggins, Woo and Sham ^{17 –} Reference Peña, Theou and Wallace ^{19 )}. The components of one FI are exemplified in Fig. 1.

Fig. 1. (Colour online) Components of the frailty index operationalised in the Honolulu-Asia Aging Study.⁽ Reference Armstrong, Mitnitski and Launer ^{15 )} SBP, systolic blood pressure; DBP, diastolic blood pressure; PD, Parkinson's disease.

In contrast the PFP characterises frailty as the presence of a constellation of attributes: weakness, slow walking speed, unintentional weight loss, exhaustion and low physical activity⁽ Reference Fried, Tangen and Walston ^{7 )}. Frailty is present when three or more of these characteristics are present and those with just one or two characteristics are termed pre-frail. The PFP was initially operationalised by Linda Fried et al. using the infrastructure of the Cardiovascular Health Study, after considering consensus clinical opinion on the most salient hallmarks of frailty in patients (Fig. 2). Other frailty measurement tools such as the FRAIL scale⁽ Reference Morley, Malmstrom and Miller ^{20 )} and the Gérontopôle Frailty Screening Tool⁽ Reference Vellas, Balardy and Gillette-Guyonnet ^{21 )} are also derived from the concept of the PFP.

Fig. 2. (Colour online) The Cardiovascular Health Study physical frailty phenotype⁽ Reference Fried, Tangen and Walston ^{7 )} and other related frailty measurement tools (the FRAIL scale⁽ Reference Morley, Malmstrom and Miller ^{20 )} and the Gérontopôle Frailty Screening Tool⁽ Reference Vellas, Balardy and Gillette-Guyonnet ^{21 )}).

(Colour online) The PFP is based on the theory of a vicious cycle of frailty, linking reduced physical activity, chronic undernutrition and loss of muscle mass to reduced resting metabolic rate, reduced strength and low mobility⁽ Reference Fried, Watson, Hazzard, Blass, Ettinger, Halter and Ouslader ^{22 )}. This cycle has remained the most plausible biological explanation for the mechanisms underpinning the frailty syndrome and has provided a standard framework upon which aetiological investigations have been based.

Both characterisations of frailty have face validity. We would expect older adults with more health deficits or older adults who have slowed up, become weaker, less active and more fatigued to be more vulnerable. Additionally, regardless of the definition used, frailty increases with advancing age and female sex providing construct validity⁽ Reference Fried, Tangen and Walston ^{7 ,} Reference Armstrong, Mitnitski and Launer ^{15 ,} Reference Goggins, Woo and Sham ^{17 )}. For example, frailty was present in 2·1 % of 65–69-year olds compared with 20·1 % of 80–84-year olds in a Spanish population and 7·7 % of men were frail compared with 9·8 % of women⁽ Reference Garcia-Garcia, Gutierrez Avila and Alfaro-Acha ^{23 )}. Most prevalence estimates of frailty are based on the phenotypic definition of frailty and range from 4·0 to 27·3 % in community-based populations of older adults (⩾65 years old)⁽ Reference Fried, Tangen and Walston ^{7 ,} Reference Garcia-Garcia, Gutierrez Avila and Alfaro-Acha ^{23 –} Reference Santos-Eggimann, Cuénoud and Spagnoli ^{26 )}.

Using both constructs, frailty has also been shown to predict the negative health outcomes we associate with vulnerable older people such as disability, institutionalisation, hospitalisation, falls and death⁽ Reference Fried, Tangen and Walston ^{7 ,} Reference Romero-Ortuno and Kenny ^{18 ,} Reference Ensrud, Ewing and Taylor ^{27 –} Reference Bandeen-Roche, Xue and Ferrucci ^{29 )}. Although the FI arguably predicts these outcomes with increased precision compared with the PFP⁽ Reference Rockwood, Andrew and Mitnitski ^{30 )}, the PFP has gained the most favour in epidemiological studies⁽ Reference Syddall, Roberts and Evandrou ^{25 ,} Reference Sirola, Pitkala and Tilvis ^{31 ,} Reference Gill, Gahbauer and Allore ^{32 )} because it allows frailty to be easily distinguished from co-morbidity and disability⁽ Reference Fried, Ferrucci and Darer ^{11 )} facilitating exploration of its determinants and consequences⁽ Reference Gale, Cooper and Deary ^{33 ,} Reference Baylis, Bartlett and Syddall ^{34 )}. In contrast, the FI often contains co-morbidity and disability in its construct making it difficult to disentangle associations (Fig. 1). Thus, the FI has been predominantly used when there is a need to use readily available or retrospectively collected data, e.g. in studies of healthcare utilisation for health-service planning⁽ Reference Albaba, Cha and Takahashi ^{35 )}.

Sarcopenia

Sarcopenia was first described by Rosenberg as the age-related loss of skeletal muscle mass⁽ Reference Rosenberg ^{8 ,} Reference Rosenberg ^{36 )}. It can be distinguished from cachexia by the more moderate degree of muscle wasting observed and the absence of either associated adipose tissue wasting and/or a high inflammatory state⁽ Reference Morley ^{37 )}. Rosenberg's first observations concluded⁽ Reference Rosenberg ^{8 )}:

‘…there is probably no decline in structure and function more dramatic than the decline in lean body mass or muscle mass over the decades of life.’

Although early operational definitions were based on low muscle mass alone⁽ Reference Baumgartner, Koehler and Gallagher ^{38 )}, research over the past few decades has emphasised the strong predictive relationships between measures of muscle quality i.e. strength and/or physical performance, and health outcomes. In particular, measures of physical capability such as grip strength, usual walking speed, timed chair stands performance and standing balance have been the focus of a wealth of research interest⁽ Reference Guralnik, Ferrucci and Simonsick ^{39 –} Reference Guralnik, Simonsick and Ferrucci ^{42 )}. Thus, more recent proposals for definitions of sarcopenia recommend including some measure of muscle quality in addition to muscle mass⁽ Reference Fielding, Vellas and Evans ^{43 –} Reference Studenski, Peters and Alley ^{46 )} and these definitions are summarised in Fig. 3.

Fig. 3. (Colour online) Sarcopenia is defined by international working groups as the presence of low muscle mass with low muscle strength and/or low physical performance. International Academy on Nutrition and Aging: International working group on sarcopenia⁽ Reference Fielding, Vellas and Evans ^{43 )}; EWGSOP, European working group on sarcopenia in older people⁽ Reference Cruz-Jentoft, Baeyens and Bauer ^{44 )}; AWGS, Asian working group for sarcopenia⁽ Reference Chen, Liu and Woo ^{45 )}; FNIH, Foundation for the National Institutes of Health Sarcopenia project⁽ Reference Studenski, Peters and Alley ^{46 )}. *Measured by dual X-ray absorptiometry.

These definitions are broadly comparable, with all including a combination of low muscle function with low muscle mass. The main differences occur in the detail, with different cut-points suggested in each definition for the different parameters. This is partly due to variation in normative ranges between populations, particularly with respect to muscle strength and muscle mass⁽ Reference Chen, Liu and Woo ^{45 ,} Reference Yoshimura, Oka and Muraki ^{47 –} Reference Kamarul, Ahmad and Loh ^{50 )}. However, there is also ongoing debate about how to define valid cut-points. For example, should low muscle mass be identified using a cut-point 2·5 sd below a young adult population, as low bone density was defined in the context of osteoporosis? Alternatively, others suggest that cut-points should be identified by threshold values beyond which the risk of adverse outcomes significantly increases⁽ Reference Cesari, Kritchevsky and Penninx ^{51 ,} Reference Lauretani, Russo and Bandinelli ^{52 )}.

Sarcopenia according to the European definition has been identified in 13·8 % of men and 12·4 % of women (mean age 75 years) participating in a Japanese study (using a definition of low muscle mass 2 sd below a young Japanese cohort mean)⁽ Reference Akune, Muraki and Oka ^{53 )}. Additionally, sarcopenia was identified in 4·6 and 7·9 % of men and women participating in a UK cohort study (mean age 67 years; low muscle mass defined as the lowest sex-specific tertile of lean mass)⁽ Reference Patel, Syddall and Jameson ^{54 )} and in 10·8–14·9 and 7·8–16·6 % of older men and women in Taiwan (mean age 73 years), depending on the method used to define low muscle mass⁽ Reference Lee, Liu and Peng ^{55 )}. Regardless of the definition used, prevalence increases with age but women do not always have a higher prevalence than men⁽ Reference Akune, Muraki and Oka ^{53 –} Reference Lee, Liu and Peng ^{55 )}. Early evidence suggests that sarcopenia defined by the European definition is associated with health outcomes including self-reported health, disability and mortality⁽ Reference Patel, Syddall and Jameson ^{54 ,} Reference Landi, Cruz-Jentoft and Liperoti ^{56 ,} Reference Alexandre and Duarte ^{57 )}. However, theses definitions are all relatively new and have been little scrutinised. A surprisingly low prevalence of sarcopenia (0·9 %) was reported when using the European definition in Finnish older women⁽ Reference Patil, Uusi-Rasi and Pasanen ^{58 )}. Additionally, studies comparing the different operational definitions suggest that they only exhibit mild–moderate positive per cent agreement, although negative per cent agreement is high⁽ Reference Lee, Liu and Peng ^{55 ,} Reference Dam, Peters and Fragala ^{59 )}.

Relationship between sarcopenia and frailty

The aetiology of sarcopenia is unclear but it is unlikely to be attributable to a single cause. Evidence suggests that loss of motor units as a result of motor axonal degeneration, dysregulation of cell-signalling pathways, persistent low-grade inflammation (‘inflammaging’), low habitual physical activity and endocrine dysfunction all contribute to the pathophysiology of sarcopenia⁽ Reference Matthews, Huang and Sun ^{60 )}. Indeed, the likely significant role of motor neuron degeneration in the pathophysiology of sarcopenia has led some investigators to re-characterise it as a primary neurogenic disease, influenced by a multitude of systemic factors, rather than a primary disease of muscle⁽ Reference Kwan ^{61 )}.

In similarity with sarcopenia, the aetiology of frailty is also likely to be multi-factorial⁽ Reference Fried, Tangen and Walston ^{7 )} and it is possible that both frailty and sarcopenia are the final common pathway of many pathological processes. In addition, frailty (certainly physical frailty) also shares with sarcopenia the appearance of skeletal muscle decline as a key feature. Therefore, both conditions share low physical capability as a common attribute⁽ Reference Cesari, Landi and Vellas ^{62 )} and almost all definitions of both sarcopenia and frailty include low physical function as a component, either measured by self-report or using objective measures such as usual walking speed (Fig. 4).

Fig. 4. (Colour online) Low physical function is common to both frailty and sarcopenia (adapted from⁽ Reference Cesari, Landi and Vellas ^{62 )}).

Furthermore, weakness has been identified as the most common first manifestation of the PFP⁽ Reference Xue, Bandeen-Roche and Varadhan ^{63 )} and low mobility has been associated with organism fragility (e.g. premature mortality) in animal models, emphasising the fundamental importance of mobility for survival⁽ Reference Fisher ^{64 )}. Indeed, Schrack et al. provided evidence that the decline in walking speed with increasing age reflects the need to conserve energy to support essential metabolic functions such as homeostasis, which become less efficient and increase their metabolic cost as we age⁽ Reference Schrack, Simonsick and Ferrucci ^{65 )}. Therefore, physical function is not just a marker of musculoskeletal health but encapsulates (or is an epiphenomenon of) the health of the whole organism⁽ Reference Kwan ^{61 )} and it is not surprising that measures of low physical function such as low grip strength and slow walking speed have been established as important independent predictors of mortality and morbidity in their own right⁽ Reference Studenski, Perera and Patel ^{40 ,} Reference Cooper, Kuh and Hardy ^{41 ,} Reference Cooper, Kuh and Cooper ^{66 )}. These measures have also been recommended as biomarkers of the healthy ageing phenotype by the National Institutes for Health, being included in the National Institutes for Health toolbox⁽ Reference Reuben, Magasi and McCreath ^{67 )}. Indeed some have suggested that markers of low function such as low grip strength could be used as single clinical markers of frailty⁽ Reference Syddall, Cooper and Martin ^{68 )} and usual walking speed has been postulated as the sixth vital sign of health due to its association with a wide range of health states, e.g. cognitive function, mood, motivation, musculoskeletal health and cardiovascular fitness⁽ Reference Fritz and Lusardi ^{69 )}.

The present consensus opinion still holds the view that frailty is broader than just low function and sarcopenia alone⁽ Reference Morley, Vellas and van Kan ^{9 )} and it is unclear to what extent sarcopenia and frailty overlap as clinical syndromes⁽ Reference Cesari, Landi and Vellas ^{62 )}. However, it is generally agreed that low physical function is a feature common to both conditions and a range of simple and validated measures of physical function are available, which could be used in the clinical setting to identify those at risk of these conditions⁽ Reference Cesari, Landi and Vellas ^{62 )}.

Evidence that sarcopenia and frailty are potentially reversible conditions

In a longitudinal study of community-based older persons (⩾70 years old), who were non-disabled at baseline, frailty was measured at 18 month intervals over 54 months⁽ Reference Gill, Gahbauer and Allore ^{32 )}. Over the course of the study, 57·6 % of participants had at least one transition between non-frail, pre-frail and frail states. Although transitions to greater frailty states were more commonly observed (up to 43 % of transitions at any time interval), a significant number of transitions occurred as participants became less frail (up to 23 % of transitions at any time interval). Thus, frailty is a dynamic process with evidence of reversibility and it should be possible to design interventions to ameliorate or improve frailty.

To date the most evidence has accrued to support exercise interventions for both frailty and sarcopenia. In particular, interventions that are delivered at least three times per week, include resistance exercise training and become progressively more challenging may be effective. For example, progressive resistance exercise training has been shown to improve physical performance in many studies of older adults⁽ Reference Liu and Latham ^{70 ,} Reference Peterson, Rhea and Sen ^{71 )} and also to reduce the common clinical manifestations of frailty and sarcopenia, e.g. falls⁽ Reference Gillespie, Robertson and Gillespie ^{72 )}.

Nutritional interventions have also been considered but the evidence is less consistent. In particular for all the considered nutritional interventions, e.g. protein, vitamin D and antioxidant supplementation, there is a disparity between observational and experimental evidence. For example, nutritional intake declines during older age and reductions in protein intake may reduce muscle protein synthesis, both through reduced substrate availability and reduced anabolic stimulation (leucine, an amino acid, stimulates muscle protein synthesis)⁽ Reference Drummond and Rasmussen ^{73 )}. In support of this hypothesis, observational evidence from longitudinal cohort studies has shown that those with the lowest protein intake have the highest rates of muscle mass decline⁽ Reference Houston, Nicklas and Ding ^{74 )}. However, protein supplementation studies have failed to consistently demonstrate benefit⁽ Reference Milne, Potter and Vivanti ^{75 )} although investigation of the role of protein supplementation as part of a multifactorial intervention is ongoing⁽ Reference Romera, Orfila and Segura ^{76 )}. Additionally, although vitamin D receptors are found on skeletal muscle cells and myopathy is a feature of vitamin D deficient diseases, low serum vitamin D is not always associated with low physical function⁽ Reference Annweiler, Beauchet and Berrut ^{77 –} Reference Visser, Deeg and Lips ^{81 )} in observational studies and supplementation studies also show mixed results⁽ Reference Gillespie, Robertson and Gillespie ^{72 ,} Reference Stockton, Mengersen and Paratz ^{82 ,} Reference Bischoff-Ferrari, Dawson-Hughes and Staehelin ^{83 )}.

However, the recognition of sarcopenia and frailty as important medical syndromes has fuelled interest in the development of effective interventions and it is likely that this will be an area of change over the coming few years. In particular, sarcopenia research is stimulating new drug discovery and several novel pharmaceutical interventions are being explored. These both consider new roles for existing drugs, e.g. angiotensin-converting enzyme inhibitors and the development of novel pharmaceutical agents, e.g. myostatin inhibitors⁽ Reference Laosa, Alonso and Castro ^{84 )}.

Sarcopenia and frailty in clinical practice

The prevalence of sarcopenia and frailty is higher in patient populations than community-based cohorts. For example, 40 % of the older emergency medical admissions (mean age 83 years) were identified as frail according to the Cardiovascular Health Study phenotype definition in a Belgium study⁽ Reference Joosten, Demuynck and Detroyer ^{85 )} and a recent prospective study in the UK identified 28 % of older patients (mean age 77 years) attending three acute surgical admission units as frail⁽ Reference Hewitt, Moug and Middleton ^{86 )}. In the outpatient setting, 26 % of urology, surgical oncology and general surgery patients in the USA (mean age 62 years) were deemed frail or pre-frail⁽ Reference Revenig, Canter and Taylor ^{87 )} and 37·0 % of men and 29·3 % of women (mean age 64 years) attending a dialysis unit in Korea had sarcopenia (defined as low muscle mass and strength)⁽ Reference Kim, Choi and Choi ^{88 )}. Presently many studies of prevalence in the clinical setting do not use internationally recognised definitions to identify cases and some caution must be used when comparing the results between clinical and epidemiological studies. For example, in studies of sarcopenia in surgical patients there has been a tendency to identify cases by the presence of low muscle mass alone⁽ Reference Sheetz and Waits ^{89 ,} Reference Jones, Doleman and Scott ^{90 )}. This tends to produce higher estimates of sarcopenia prevalence than definitions, including both muscle mass and muscle quality parameters.

However, it is likely that sarcopenia and frailty will be more prevalent in patient populations compared with community-based cohorts. Therefore, given the known associations of sarcopenia and frailty with negative health outcomes, it is not surprising that the estimated healthcare cost of these conditions is high. For example, in 2000 $18·5 billion dollars of spending on healthcare in the USA were attributed to sarcopenia⁽ Reference Janssen, Shepard and Katzmarzyk ^{91 )}. Similarly, one study calculated the absolute costs associated with elective surgical procedures in frail patients to be three times higher ($76 363 (sd 48 495) per patient) than non-frail patients ($27 731 (sd 15 693) per patient)⁽ Reference Robinson, Wu and Stiegmann ^{92 )}.

These high estimates, of prevalence and cost, have led to calls for these conditions to be considered more routinely in clinical practice^{( 93 )}. Case-finding has been suggested in the acute medical setting^{( 94 )}, in the care of older adults being considered for aggressive oncological treatments⁽ Reference Handforth, Clegg and Young ^{95 )}, in elective and acute surgery⁽ Reference Anaya, Johanning and Spector ^{96 ,} Reference Partridge, Harari and Dhesi ^{97 )} and in primary care^{( 93 )}. In the primary care setting, there is evidence that frailty may be important when considering treatment of traditional disease risk factors, such as hypertension, which may behave differently in frail patients⁽ Reference Muller, Smulders and de Leeuw ^{98 ,} Reference Odden, Peralta and Haan ^{99 )}. Additionally, there is scope for early intervention to prevent or delay late-life disability since both frailty and sarcopenia often precede disability but are potentially reversible conditions. In contrast, disability is difficult to reverse once it has occurred⁽ Reference Fried, Ferrucci and Darer ^{11 )}. For example, frailty could be used to identify those older people presenting to medical services who might benefit the most from comprehensive geriatric assessment (CGA). CGA is a multi-dimensional and inter-disciplinary review of the medical, functional and psycho-social needs of an older person in order to formulate a personalised plan for treatment and long-term follow-up. It has been shown to improve the likelihood of older people being alive and in their own home 1 year following an emergency hospital admission⁽ Reference Ellis, Whitehead and Robinson ^{100 )} but present clinical resources are not sufficient to provide CGA for all older people. Therefore, frailty could help to promote equity of access to CGA services.

With respect to medical practice in secondary care, one particular area that has recently received a great deal of interest is frailty in the older surgical patient. Presently a growing number of prospective studies evaluating the associations of frailty with surgical outcomes have been published (Supplementary Material, Table S1).

It is clear from these studies that frailty adds a new dimension to the surgical assessment. Preliminary evidence suggests that frailty measurement can aid risk prediction and outperform traditional anaesthetic or surgical risk scores, in terms of prediction of post-operative complications, longer in-patient stay, discharge to institutional care and mortality⁽ Reference Revenig, Canter and Taylor ^{87 )}. Therefore, adding frailty to the assessment of older surgical patients could aid decision making and improve patients’ (and their families’) understanding of the operative risks, in order to make a more informed choice regarding treatment options⁽ Reference Partridge, Harari and Dhesi ^{97 )}. Secondly, if frailty (and/or sarcopenia) was identified, optimisation interventions to reduce or reverse frailty could result in improved outcomes. For example, CGA has also been associated with improved outcomes when applied to the management of older surgical patients⁽ Reference Harari, Hopper and Dhesi ^{101 )} and the pre-operative identification and treatment of anaemia, a common condition associated with frailty, has been found to be beneficial in older patients presenting for elective orthopaedic surgery⁽ Reference Goodnough, Maniatis and Earnshaw ^{102 )}. Additionally, the concept of ‘pre-habilitation’ has been suggested⁽ Reference Gillis, Li and Lee ^{103 )}. This would involve the design of multifaceted interventions to improve the fitness and nutritional health of older patients prior to surgery. However, to date few trials have evaluated exercise or nutritional interventions peri-operatively and this is an area in urgent need of further research.

Challenges to clinical translation

Presently neither sarcopenia nor frailty are recognised with International Classification of Diseases codes and debate over the exact definitions and diagnostic criteria of both sarcopenia and frailty has been a major limiting factor with respect to clinical appetite to incorporate these conditions into practice⁽ Reference Cooper, Dere and Evans ^{104 )}. As demonstrated, the lack of consistency when defining sarcopenia or frailty even extends to within sub-specialty areas of the research field. For example, a striking observation from Table S1 (Supplementary Material) is the variety of tools used to measure frailty and the different ways even the same measure is operationalised. For example, the PFP definition proposed by Fried is sometimes operationalised incorrectly as a continuous score (from 0 to 5) rather than a three-level category (frail (0), pre-frail (1–2) and robust (⩾3))⁽ Reference Leung, Tsai and Sands ^{105 )}. Additionally, while consensus meetings often agree on the concepts of frailty and sarcopenia they often fail to achieve agreement on diagnostic criteria⁽ Reference Rodríguez-Mañas, Féart and Mann ^{10 )}, including methods to identify low muscle mass⁽ Reference Studenski, Peters and Alley ^{46 )} and the choice of appropriate cut-points. A particular problem has been the use of cohort-specific cut-points in research studies, e.g. low muscle mass defined as the ‘lowest tertile’ of muscle mass in the cohort under investigation. While use of different percentile groups to explore correlations and trends within cohorts can provide useful observations, it makes findings of different studies hard to compare and replication of results more difficult. This will be especially true when sarcopenia and frailty are the subjects from interventional studies. For example, in absolute terms the lowest ‘tertile’ cut-point of one cohort may be very different from another. Thus, if the same intervention was applied in two study populations and the results differed, then this difference may be attributed to the inherent differences between the cases included in each study. Doubt would be cast that the same people, and thus the same underlying condition, were identified by each interventional study. This problem was exemplified in a report from the Leiden Longevity Study which considered seven proposed operational definitions of sarcopenia and found that only one individual in their cohort of 654 older men and women was sarcopenic by all definitions⁽ Reference Bijlsma, Meskers and Ling ^{106 )}.

Additionally, the operational definitions that have worked well in epidemiological studies do not always function as smoothly in the clinical context. For example, although a FI can be derived from standard clinical assessments such as the CGA ⁽ Reference Jones, Song and Mitnitski ^{107 )}, it is not feasible to conduct a CGA on every older person presenting to healthcare services and one of the main indications for identifying frailty in clinical practice would be as an indication for CGA. Even the PFP has proved difficult to execute in busy clinical settings, since it requires measurement of attributes which are not part of the routine examination and which require additional equipment, e.g. dynamometer to measure grip strength. To this end several other frailty scales have been developed in recent years, which may be more feasible, e.g. the FRAIL scale (exemplified in Fig. 3) or the SHARE-FI75+⁽ Reference Romero-Ortuno and Soraghan ^{108 )}. Rockwood et al. also developed a Clinical Frailty Scale based on clinical judgement that is simple to use and performs well in comparison with the multiple deficits based FI⁽ Reference Rockwood, Song and MacKnight ^{109 )}.

With respect to sarcopenia, the main area of the present debate resides in the necessity to measure muscle mass in clinical practice. Compared with measures of strength or performance, relationships between muscle mass and health outcomes are weaker⁽ Reference Lauretani, Russo and Bandinelli ^{52 ,} Reference Janssen ^{110 ,} Reference Newman, Kupelian and Visser ^{111 )}. Additionally, low muscle mass does not always correlate well with low strength or performance⁽ Reference Patil, Uusi-Rasi and Pasanen ^{58 )} and it is not clear how best to measure muscle mass in clinical practice or whether it would be feasible⁽ Reference Keevil, Hayat and Dalzell ^{112 )}. Dual X-ray absorptiometry is widely regarded as the safest and most accurate measure for clinical practice but equipment is bulky, with limited availability in many health-care settings. Additionally, bioelectrical impedance analysis, while more portable and potentially feasible in primary as well as secondary care, is not clearly superior to simple anthropometric methods of body composition assessment⁽ Reference Elia ^{113 )}, which are themselves not endorsed due to concerns over accuracy⁽ Reference Cruz-Jentoft, Baeyens and Bauer ^{44 )}.

The challenges of measuring muscle mass in the clinical setting have led to proposals for function-based sarcopenia screening tools. For example, SARC-F measures risk of sarcopenia based upon responses to questions pertaining to lifting or carrying, rising from a chair, assistance with walking, stair climbing and falls and has been validated for use⁽ Reference Malmstrom and Morley ^{114 –} Reference Morley and Malmstrom ^{116 )}.

Thus, in the clinical setting we may need to adapt what has been used in research and take a pragmatic approach⁽ Reference Martin and Brighton ^{117 )}. For example, we could take advantage of the commonality of low physical function to both frailty and sarcopenia in order to utilise simple tests, such as usual walking speed, to identify those at risk of both conditions. Whether further assessment would then be warranted, e.g. with tests to identify low muscle mass, would depend on the further development of interventions. At present the exercise interventions with the best evidence base work equally well in frail patients, sarcopenic patients or simply those patients identified to have low physical function. Therefore, until more specific treatment is available, more specific identification and differentiation of sarcopenia and frailty in the clinical setting will not be justified in terms of cost or resource allocation. However, the pragmatic assessment of frailty and sarcopenia in clinical practice and research will be justified as long as it adds value to help explain the observed heterogeneity in risk among older people.