Frailty is a complex clinical syndrome with multiple causes and contributing factors in which there is increased vulnerability when exposed to a minor stressor and increased risk for adverse outcomes, such as disability, hospitalization and/or mortality.4,5 This is often manifested by maladaptive response to stressors, leading to a vicious cycle toward functional decline and other serious adverse health outcomes. It is characterized by diminished strength, endurance and physiological reserve across the neuromuscular, metabolic and immune systems.6 It is important to note that old age itself does not define frailty, because some patients remain vigorous despite advanced age, while others can have functional decline in the absent of apparent stress factors or failure to rebound following hospitalization or illness.5 Thus, it is important to note the difference between biological age and chronological age.7

The prevalence of frailty ranges from 4% to 17%, and is higher among women (almost double that in men), increasing significantly in patients older than 80 years of age.8,9 Pre-frailty (which describes patients at risk for frailty who fulfill some, but not all, criteria for frailty) has been also been the subject of various studies, which show a prevalence around 28-44%.9 Several factors are thought to contribute to the development of frailty, including poor nutrition, reduced exercise tolerance, aging, chronic inflammation and immunological decline.7 Thus, it can potentially be prevented or treated with specific modalities, such as exercise, protein-calorie and vitamin D supplementation, and reduction of polypharmacy.6 However, a common feature of frail persons is mild cognitive impairment, which hampers the application of therapeutic regimens, especially exercise programs.

Frailty is a multifactorial condition. The literature shows that certain changes in physiological systems are associated with an increased risk for frailty, including a proinflammatory state and elevated markers of blood clotting,10 sarcopenia,7 anemia,11 anabolic hormone deficiencies,12,13 insulin resistance,14 significant immune system alterations,15 and oxidative stress.16

The aging of populations is increasing the number of frail patients with CVD. Thus, identifying frailty has important implications for clinical care. Frailty worsens prognosis in patients with CVD (Table 1) and of patients undergoing cardiac surgery and other cardiovascular interventions, and can reduce the net benefits of some cardiac interventions because of competing risks.17,18 Frailty also increases the risk of cardiovascular and non-cardiovascular mortality and the need for rehabilitation and institutional care.2,19,20 In the large TRILOGY ACS trial, which included 4671 patients older than 65 years with ACS, 25% were considered pre-frail and 5% frail (according to the Fried score).21 Frail patients were more likely to suffer stroke or cardiovascular death after adjusting for the Global Registry of Acute Coronary Events (GRACE) score. Frailty is also a strong independent predictor of mortality in heart failure patients. One study found a population-attributable risk associated with frailty for emergency department visits of 35% and for hospitalizations of 19% among patients with heart failure.22 Kang et al. found that frailty was strongly and independently associated with short-term outcomes for elderly patients with ACS.23 In a study by Ekerstad et al. frailty was strongly and independently associated with in-hospital mortality, one-month mortality, prolonged hospital care, and the primary composite outcome in patients with non-ST-segment elevation myocardial infarction.24 Ricci et al. also found that frail and pre-frail older individuals accounted for a substantial proportion of those with more cardiovascular risk factors, especially diabetes, highlighting the need for preventive strategies in order to avoid the co-occurrence of CVD and frailty.25

Similarly, frailty is associated with higher mortality and morbidity and greater need for health care in patients with valvular disease undergoing cardiac surgery.20,26

The number of elderly patients undergoing cardiac surgery is increasing. Frailty screening may be useful to identify patients with increased risk of adverse outcomes. Sundermann et al. found that patients who died within one year had a median Comprehensive Assessment of Frailty (CAF) score of 16 [5;33] compared to 11 [3;33] in one-year survivors (p=0.001), proving the prognostic value of frailty in cardiac surgery.20 Afilalo et al. also demonstrated the association between frailty and mortality or major morbidity after coronary artery bypass grafting (CABG) and/or valve surgery (odds ratio [OR] 2.63; 95% confidence interval [CI] 1.17-5.90).28 In a study by Jung et al., frailty was associated with a 3- to 8-fold increase in risk of postoperative delirium. According to these authors, ‘frail’ and ‘fit’ may be considered two ends of a continuum, and the risk of postoperative delirium grows as one becomes increasingly frail.29 There has also been interest in understanding whether preoperative frailty is associated with worse outcomes after implantation of a left ventricular assist device (LVAD) as destination therapy. A study by Dunlay et al. using a deficit index to assess frailty found that patients who were intermediate frail (adjusted HR 1.70, 95% CI 0.71-4.31) and frail (HR 3.08, 95% CI 1.40-7.48) were at increased risk for death (p=0.004 for trend). The mean number of days alive out of hospital the first year after LVAD implantation was higher for patients who were not frail.30 Schoenenberger et al. studied elderly patients undergoing transcatheter aortic valve implantation (TAVI) and showed that all the components of their geriatric baseline examination helped predict functional decline after intervention (OR: 3.31; 95% CI 1.21-9.03).26 Stortecky et al. found that frailty was associated with increased all-cause mortality (OR: 3.68; 95% CI 1.21-11.19), and with increased major adverse cardiovascular and cerebrovascular events (MACCE) (OR: 4.89; 95% CI 1.64-14.60) one year after TAVI (OR: 3.68; 95% CI 1.21-11.19).31

Pre-frailty, which is a potentially reversible state, also appears to have some prognostic value. The findings of Sergi et al. suggest that pre-frailty is independently associated with a higher risk of older adults developing CVD. Among the physical domains of pre-frailty, low gait speed seems to be the best predictor of future CVD.32

Assessment of frailty may lead to patients being reclassified to different clinical risk categories, suggesting it signals risk not captured by currently used risk assessment scores.2,32

An ideal frailty screening tool should (1) be able to accurately identify frailty; (2) predict the response of frail patients to potential therapies; and (3) be simple and easy to apply and have low cost.33 Simple and rapid screening tests have been developed and validated to enable the objective recognition of frail persons. They differ mainly in the nature and number of deficits they measure, in line with two contrasting conceptual models: the frailty phenotype, or physical frailty, and the frailty index or deficit accumulation.34 The phenotype concept considers frailty as a syndrome, consisting of a small number of highly specific deficits in health, such as unintentional weight loss, exhaustion, slowness, low physical activity and impaired grip strength. Further specific health deficits, such as cognitive deficits, have been proposed as part of a frailty phenotype scale. By contrast, frailty indices are based on the concept of cumulative deficit, assessing frailty through a larger number of unspecified age-associated health deficits (usually at least 30). The most commonly used are the FRAIL Questionnaire screening tool,35–38 the Cardiovascular Health Study Frailty Screening Scale (Fried criteria),9 the Clinical Frailty Scale (CFS),39 frailty indices,40,41 and the Edmonton Frail Scale42 (Table 2).

The FRAIL Questionnaire screening tool considers deficits accumulated in five domains, forming its acronym: Fatigue (self-reported), Resistance, Ambulation (slow walking speed), Illnesses, and Loss of weight (5% or more in the past year). The five domains are weighted equally. Individuals with two deficits are considered pre-frail, and those with three or more deficits are classified as frail.

The Fried criteria for frailty (also known as the Cardiovascular Health Study Frailty Screening Scale, the Physical Frailty Phenotype and the Hopkins Frailty Phenotype) was first developed in the Cardiovascular Health Study.9 It assesses physical characteristics or phenotype, which include five domains: unintentional weight loss (4.5 kg or more in the last year), exhaustion (self-reported), low physical activity, weakness (low grip strength), and walking speed.9,43 Low physical activity is assessed through the frequency of moderate intensity activities, such as gardening or household chores. Similarly to the previous scale, individuals with two deficits are considered pre-frail, and those with three or more deficits are classified as frail.29

The CFS is a global clinical assessment of frailty based on physical function and level of independence with activities of daily living proposed by Rockwood et al.39 Each point on its scale has a visual chart and a written description of frailty to assist the classification process. Scoring is based on clinical judgment and ranges from 1 (very fit) to 9 (terminally ill).39

Frailty indices are based on the deficit accumulation approach to measuring frailty, and are commonly used tools to assess frailty in order to estimate the related risk for adverse health outcomes, such as mortality.44 A frailty index is based on the concept that frailty is a consequence of interacting physical, psychological, and social factors. As deficits accumulate, people become increasingly vulnerable to adverse outcomes. The subject answers 20 or more questions related to medical and functional issues. The tool can be adapted to information available in the medical record and does not require a patient interview or exam to assess frailty. The 70 items of the original version are not to be considered a fixed set of variables (Table 3). It has been reported that estimates of risk are stronger when a minimum of 50 items are considered, but shorter versions (as few as 20 items) have also been studied.39–41 Rockwood and Mitnitski44 proposed a deficit accumulation-based frailty index using a comprehensive geriatric assessment (FI-CGA). This involves the accumulation of 30 or more comorbidities, symptoms, diseases, disabilities and other health deficits and is expressed as a ratio calculated as the number of deficits in an individual divided by the number of total deficits measured; the greater the number of deficits, the higher the score. The comprehensive geriatric assessment (CGA) includes medical, nutritional, functional and psychological assessments by a multidimensional team. The FI-CGA was initially developed as a ten-domain index with 14 CGA components and was later expanded to include 52 CGA components.33

The Edmonton Frail Scale (EFS) was developed to be practical and usable in the community setting or at the bedside. It is scored out of 17 and contains the following components: cognition, general health status, self-reported health, functional independence, social support, nutrition, mood, continence, and functional performance. The component scores are summed and the following cut-offs are used to classify frailty severity: not frail (0-5), apparently vulnerable (6-7), mildly frail (8-9), moderately frail (10-11) and severely frail (12-17).

The scales described above are those most commonly used to assess frailty. However, other frailty scales are available, as listed below.

The Groningen Frailty Indicator (GFI) considers 15 dichotomous self-reported deficits in four domains: physical, cognitive, social, and psychological.45

The Tilburg Frailty Indicator contains 15 self-reported items in physical, psychological, and social domains.46

The Gérontopôle Frailty Screening Tool (GFST) comprises two steps: an initial questionnaire (containing six components: living alone, involuntary weight loss, fatigability, mobility, memory complaints and slow gait speed) followed by the clinician's judgment of frailty status.47

PRISMA-7 contains seven self-reported components: older than 85 years; male; health problems which limit activities; health problems requiring staying at home; support of another person needed; social support; and use of a cane or walker or wheelchair. Frailty is defined by a score of 3 or more.48

The Multidimensional Prognostic Instrument (MPI) is a multidimensional prognostic tool used for hospitalized older patients. It includes eight CGA components: ADL, instrumental ADL, risk of developing pressure sores, comorbidity, medication number, nutritional status, cognitive status, and living status.49

Some frailty assessment tools have been designed to be used specifically in the setting of CVD (Table 4). Some of the tools described above are also used for this purpose.23,24,50

Kang et al. used the CFS, which was useful in assessment of elderly patients with ACS, predicting all-cause mortality, unscheduled return visit, and in-hospital and recurrent major adverse cardiovascular events.23

Ekerstad et al. used frailty as measured by the CFS to assess short-term outcomes for elderly patients with non-ST-segment elevation myocardial infarction, and showed that the combined use of frailty and other comorbidity tools (such as the coronary artery disease-specific index) may constitute a novel risk prediction concept in regard to cardiovascular patients with complex needs.24

Uchmanowicz et al. investigated the correlation of a scale for assessing frailty – the Tilburg Frailty Indicator and its mental and physical domains – with other screening tools commonly used for CGA in patients with ACS. Significant correlations were demonstrated between the values of the TFI and other scales.50

Boxer et al. also found that the six-minute walk and the five-item Cardiovascular Health Study were independently predictive of mortality in older adults with heart failure, with hazard ratio (HR) 0.82 (95% CI 0.72-0.94) and 1.64 (95% CI 1.19-2.26), respectively, and both could be useful as a measure of frailty.17

In a prospective observational study by Jung et al. in elective cardiac surgery patients, frailty was defined using the seven-item Cardiovascular Health Study score, the Short Physical Performance Battery (SPPB) and a 35-item frailty index. They found that the addition of frailty improved the ability of the EuroSCORE II to predict postoperative delirium, pointing to opportunities for improved prevention and management.29

The CAF is a tool created by Sundermann et al.20 to assess the prognosis of elderly patients before cardiac surgical interventions and accurately predicts mortality. It comprises grip strength, walking speed, balance, and ability to pick up a pen from the floor, rise from a chair three times and put on and remove a jacket, thus combining characteristics of the CHS criteria9 of patient phenotype, physical performance, and laboratory results. According to the authors, a combination of the CAF and traditional scoring systems may facilitate more accurate risk scoring in elderly high-risk patients scheduled for conventional cardiac surgery or transcatheter aortic valve replacement.27 The CAF was preoperatively applied to 400 patients aged ≥74 years admitted to a cardiac surgical department between September 2008 and January 2010. For 213 of these patients one-year follow-up was assessed by telephone interview until April 2010. One hundred and ten male and 103 female patients were included. Twenty-five percent underwent isolated coronary revascularization, 35% isolated valve procedures and 26% underwent combined procedures. One-year mortality was 12.2%. Patients who died within one year had a median frailty score of 16 [5;33] compared to 11 [3;33] in one-year survivors (p=0.001).20 Sundermann et al. showed that the CAF score facilitates prediction of mid-term outcome of high-risk elderly patients and the modified CAF score showed a promising ability to predict one-year mortality in patients undergoing cardiac surgery.20,27

The study by Dunlay et al. assessed the association between preoperative frailty and worse outcomes after implantation of an LVAD. Patients undergoing LVAD implantation as destination therapy at the Mayo Clinic, Rochester, MN between February 2007 and June 2012 were included in this study. Frailty was assessed using a deficit index (including 31 impairments, disabilities and comorbidities) and defined as the proportion of deficits present. Patients were then divided based on tertiles of the deficit index (>0.32=frail, 0.23 to 0.32=intermediate frail, <0.23=not frail). The authors concluded that frailty before destination LVAD implantation, as assessed by their deficit index, is associated with increased risk of death and may represent a significant patient selection consideration.30

Some tools have also been applied to patients undergoing TAVI. Stortecky et al. assessed the Multidimensional Geriatric Assessment (MGA) as a predictor of mortality and MACCE after TAVI. This prospective cohort comprised 100 consecutive patients aged ≥70 years undergoing TAVI. Global risk scores (Society of Thoracic Surgeons [STS] score, EuroSCORE) and MGA-based scores (cognition, nutrition, mobility, activities of daily living [ADL], and frailty index) were assessed as predictors of all-cause mortality and MACCE 30 days and one year after TAVI. This study provides evidence that risk prediction can be improved by adding MGA-based information to global risk scores.31

Schoenenberger et al. used the EuroSCORE, the STS score, and a geriatric baseline examination (based on assessment of cognition, mobility, nutrition, instrumental and basic activities of daily living) to predict functional decline in elderly patients undergoing TAVI. Overall predictive performance was best for the geriatric baseline examination and low for the EuroSCORE and STS score. In univariate analysis, all components of the geriatric baseline examination helped predict functional decline. The authors concluded that the geriatric baseline examination, but not established risk scores, was predictive of functional decline.26

In the 2012 study by Afilalo et al., a total of 152 patients were enrolled in a prospective, multicenter cohort study of elderly patients (>70 years) undergoing CABG and/or valve surgery in the US and Canada. Four different frailty scales, three disability scales, and five cardiac surgery risk scores were measured in all patients. The primary outcome was the STS composite endpoint of in-hospital postoperative mortality or major morbidity. The four frailty scales examined in this study are described in Table 4. The authors concluded that clinicians should use an integrative approach combining frailty, disability, and risk scores to better characterize elderly patients referred for cardiac surgery and identify those that are at increased risk.28

The same author, in a recent study (2017),51 compared the incremental predictive value of seven different frailty scales to predict poor outcomes following TAVI and valve surgery: the Fried criteria (described above),9 Fried+(the Fried criteria plus cognition assessed by the Mini Mental State Exam [MMSE] and mood assessed by the Short-form Geriatric Depression Scale), the Rockwood CFS,39 the SPPB (three physical tests, with each scored 0 to 4 for a composite score of 0 to 12: gait speed, time to stand five times from a seated position without using arms and the ability to stand 10 s with the feet in tandem or side-by-side positions),52 the Bern Scale (six items for a composite score of 0 to 7: gait speed, mobility, cognition, nutrition and disability in activities of daily living and instrumental activities),26,31 the Columbia Scale (four items, with each scored 0 to 3 for a composite score of 0 to 12: gait speed, grip strength, serum albumin and disability),53 and the Essential Frailty Toolset (EFT) (four items for a composite score of 0 to 5: time to stand five times from a seated position without using arms (1 point if ≥15 s, 2 points if unable to complete), cognition (1 point if MMSE <24), hemoglobin (1 point if <13 g/dl in men or <12 g/dl in women), and serum albumin (1 point if <3.5 g/dl).51 Frailty as measured by the EFT was the strongest predictor of death at one year (p<0.001) and of worsening disability at 1 year (adjusted OR:2.13; 95% CI:1.57 to 2.87) and death at 30 days (adjusted OR: 3.29; 95% CI: 1.73 to 6.26).

Some authors, instead of validating existing frailty screening tools, used certain parameters to derive their own frailty score. Green et al. used a modification of the Fried frailty criteria that included gait speed, grip strength, serum albumin, and activities of daily living status to derive a frailty score among older adults with severe aortic stenosis who underwent TAVI. In this study frailty was not associated with increased periprocedural complications in patients selected as candidates to undergo TAVR, but was associated with increased one-year mortality after TAVR.53

We have described individual multiple frailty measurement scales. There have been various studies comparing the most commonly used screening tools, but agreement on which has the best ability to predict prognosis and all-cause mortality is lacking. While some studies found similar prognostic performance in some of these tools,54,55 other studies found significant differences.56 In 2013, a consensus conference identified some of these tools as allowing physicians to objectively recognize frail persons.6 However, they range from short, fast and crude screening tools to sophisticated and time-consuming scales. A source of concern is the fact that many frailty scales have been modified somewhat from their original and validated version, leading to significant differences in frailty classification.33

Since the frailty phenotype and frailty indices are based on different concepts, it is inappropriate to consider them as alternatives and/or interchangeable.57Table 5 describes the main characteristics of these two different instruments.

The FRAIL screening tool is clinically advantageous due to its simple nature and ability to be obtained from data already included in a CGA.2 It has been found to be predictive of mortality in specific populations, such as patients with CVD.43

The CHS scale (Fried criteria) is a widely used scale applied in multiple epidemiological studies, and has good predictive value for adverse clinical outcomes, including mortality. However, a major factor precluding its clinical application is the inclusion of measurements not routinely used for patient assessment (such as grip strength measured by a dynamometer). Another important limitation of this scale is that it does not include psychosocial components of frailty.9

The CFS has been validated as a predictor of adverse outcomes in hospitalized older people, such as all-cause, in-hospital mortality, one-month mortality and prolonged hospital care.23,24,33

Deficit accumulation-based frailty indices are well validated and are better at predicting adverse clinical events than other frailty measurements in both hospital and community settings. They have been applied to multiple datasets, but can be time-consuming to calculate. The FI-CGA is used as a clinical standard for frailty assessment and has been found to predict patient response in multiple fields, including cardiology.2

The EFS is a valid and reliable measurement tool in the hospital setting and, since it has only nine components, it is much simpler to extract from CGAs than the FI-CGA.42 In a community-based sample, even when administered by non-specialists with no formal training in geriatric care, the EFS compared favorably with the clinical assessment of geriatric specialists who completed a more comprehensive evaluation.42

Compared with other frailty assessment tools, the MPI appears to have greater ability to predict adverse outcomes. Nevertheless, additional research is needed to confirm these results.33,49

The GFI has moderate internal consistency and adequate discriminative ability, and shows good feasibility and reliability as a frailty measurement. Some authors have proposed that it should be used together with a frailty index as part of a two-step screening process.58

The TFI shows good reliability for identifying frailty in community-dwelling older people. However, although its physical components show good ability to predict adverse events, its social components appear to be weak predictors.59,60

The GFST, designed for early recognition of frailty in community-dwelling older people, appears to be a good frailty screening tool; however, it gives no specific guidance for clinicians on how to identify frailty and the clinician's judgment of frailty status is quite subjective.33

Although PRISMA-7 shows good accuracy in identifying frailty in community-dwelling older people, its ability as a screening tool is limited since it has a tendency to over-screen for frailty.59

Some of these tools have also been used in cardiovascular patients, such as the Fried frailty criteria,17,23,24,29 frailty indices,29,30 and the TFI.50

The geriatric baseline examination developed by Schoenenberger et al.,26 the multidimensional geriatric assessment used by Stortecky et al.,31 and the CAF,20,27 developed by Sundermann et al., are three recent frailty tools which have been shown to be useful in predicting mortality and assessing prognosis of elderly patients with CVD or before cardiac surgical interventions.

Some authors, such as Afilalo et al., prefer to use an integrative approach combining frailty, disability, and risk scores to better characterize elderly patients referred for cardiac surgery, which has proved to be useful for identifying those at increased risk.28

The same author recently showed the superiority of the EFT compared to other frailty scales when predicting poor outcomes following TAVI and valve surgery, since it is a relatively simple tool that is neither particularly burdensome nor time-consuming and at the same time captures multiple domains of frailty. The authors suggested it had the highest predictive value for death and worsening disability at one year and recommended its use in this setting.51

Assessment of frailty as a preoperative surgical risk factor has been shown to be useful and may increase the number of elderly patients considered eligible for surgical interventions, since it enables prior screening for risk and prediction of surgical success and safety.

When selecting a screening tool, it is also important to take into account the ecology of its application, since some frailty measurements are more suitable for use in population health studies as screening, whereas others are appropriate in the clinical setting for screening or diagnosis of frailty.42,55,57