Editor’s Choice - Frailty and the management of patients with acute cardiovascular disease: A position paper from the Acute Cardiovascular Care Association

Abstract

Frailty is increasingly seen among patients with acute cardiovascular disease. A combination of an ageing population, improved disease survival, treatable long-term conditions as well as a greater recognition of the syndrome has accelerated the prevalence of frailty in the modern world. Yet, this has not been matched by an expansion of research. National and international bodies have identified acute cardiovascular disease in the frail as a priority area for care and an entity that requires careful clinical decisions, but there remains a paucity of guidance on treatment efficacy and safety, and how to manage this complex group. This position paper from the Acute Cardiovascular Care Association presents the latest evidence about frailty and the management of frail patients with acute cardiovascular disease, and suggests avenues for future research.

What is frailty?

Frailty is increasingly seen among patients with acute cardiovascular disease.¹ A combination of an ageing population, improved disease survival, treatable long-term conditions as well as a greater recognition of the syndrome has accelerated the prevalence of frailty in the modern world.^2,3 National and international bodies have identified acute cardiovascular disease in the frail as a priority area for care and an entity that requires careful clinical decisions,⁴ but there remains a paucity of guidance on treatment efficacy and safety, and how to manage this complex group.⁵ Frailty is a syndrome characterised by reduced biological reserve.⁶ It is commoner, but not synonymous, with increasing age and results from the impairment of multiple physiological systems leading to vulnerability to minor stressors.⁷ Increasing attention to the recognition and implications of frailty reflects an emerging evidence base which, typically, defines frailty in two ways – a phenotype model and an accumulated deficit model.^8,9 The former focuses on the identification of a population in whom weakness is prominent, and where intervention could defer or prevent disability. In contrast, the latter reflects a population for whom disability and comorbidity, individually or in combination, produce a cumulative deficit. However defined, individuals with frailty are more vulnerable to deterioration in the face of concurrent illness, and have worse clinical outcomes.⁷

Despite an increasing awareness of the importance of frailty, no consensus exists as to the best method of evaluation.¹⁰ Although collectively frailty, often conceived of as a functional deficit, carries an adverse impact on outcome in observational analysis,¹¹ no large-scale prospective interventional evaluation has been reported in either hospital or community settings. Where studies have been conducted with in-patients, they are frequently observational and use a wide variety of scoring tools of uncertain validity.^12–14 Importantly the majority of the literature on measuring frailty focuses on patients in the non-acute setting.¹⁵ This reflects the confounding impact of acute illness and the uncertainty of recovery in the early post-acute phase.

There are two key circumstances when pre-existing frailty impacts on cardiovascular care. Firstly, in a situation where a treatment or intervention is thought necessary for optimal outcome, an evaluation of baseline frailty might guide anticipatory care during and after the intervention. Secondly, when the outcome is less certain, then management decisions may be contingent upon baseline status. This may be the case in the management of acute presentations or when considering the likely benefit of elective interventional procedures.¹⁶ The most appropriate evaluation of baseline frailty, and the potential benefit this might yield remain uncertain.^5,17 The increasing awareness of the impact of frailty on healthcare resources has prompted the European Union to issue a ‘call for action’ to accurately define frailty.⁴

Frailty in context

Frailty may account for the distinction between biological and chronological ageing. At any chronological point, a diminished physiological reserve renders the individual vulnerable to stressors.⁷ Frailty acts as a marker of ‘biological age’, and is more closely associated with prognosis than chronological age.¹⁸ Whilst this vulnerability is typically taken to be physical and manifest (such as slow gait speed, tiredness and weakness), cognitive and nutritional deficits are also common.⁷ Although accumulated co-morbid illnesses do not per se constitute frailty, their aggregated impact (and the associated therapeutic burden) is often manifest through reduced reserve and increased vulnerability.^19,20

This heterogeneity, not only of contributory factors, but also of the interplay of their manifestations, has three important consequences. Firstly, despite the appeal of a single quantitative measurement of frailty sufficient to guide clinical decisions, no inclusive index of frailty has been defined.¹¹ Secondly, the context of the evaluation is important: a measure of frailty in a falls-prevention programme may not be useful in an evaluation of recovery after acute myocardial infarction.²¹ Thirdly, simple measures used in the correct context may be just as useful as more complicated indices.^22,23 All three factors are key to considering the impact of frailty upon clinical outcomes.

Definitions of frailty

Several models of frailty exist, all of which rely upon overlapping domains and are based upon surveys of community dwelling individuals.

Phenotype models rely upon functional measures of physical performance such as gait speed (typically over 5–10 m) and grip strength (measured with a dynamometer). Of these, the Fried Scale is the most widely researched.²⁴ This scale requires three or more of the following five components for the diagnosis of physical frailty: (a) exhaustion; (b) unintentional weight loss; (c) low physical activity; (d) impaired grip strength; and (e) slowness. One commonly reported assessment included in other phenotype models is the ‘timed get up and go test’.²⁵

The accumulated deficit model represents an alternative approach, which includes measurement of disabilities, symptoms, diseases and laboratory measurements. One of the first to be systematically evaluated was the Frailty Index, a cumulative count of multiple clinical deficits derived from the Canadian Study of Health and Ageing Study, a community-based epidemiological evaluation of dementia.²⁶ The Rockwood Scale (also known as the Clinical Frailty Scale (CFS)) is a simplified global clinical measure of frailty validated against the Canadian Frailty Index. The CFS, which relies upon a brief assessment of the level of support needed for ordinary functional activity, was developed to simplify the time consuming multiple assessment process required by the Frailty Index.²⁷ A cumulative deficit approach has also recently been used to derive and validate an electronic Frailty Index (eFI) which uses electronic health records to automatically calculate a patient’s frailty score and has now been adopted into an automated evaluation of the older people in UK general practice surgeries.²⁸ The eFI shows good discriminatory performance for death, hospitalisation and nursing home admission at up to three years.²⁸

The standard Comprehensive Geriatric Assessment (CGA), which may include over 40 items and requires formal clinical assessment, captures not only accumulated clinical and functional deficits and nutrition, but also social factors, and cognitive evaluation. The CGA, which has been thoroughly validated,²⁹ may best be seen as a model for the management of disability (be it cognitive, nutritional, functional or social) in the in-patient setting. However, the CGA is time-consuming, and dependent on environment and capability. Several tools have, therefore, been derived from the CGA, including the Edmonton Frailty Scale (EFS),³⁰ Multidimensional Prognostic Index (MPI)³¹ and CGA score.³² However, such abbreviated versions of the CGA developed as frailty screening tools are less well validated.

The presence of frailty is defined by thresholds. Thus, for the Rockwood Scale (or CFS) moderate frailty is deemed to be present when there is a degree of dependence on others for ‘high order’ activities of daily living (such as finances, transportation, heavy housework, medications), with lesser dependence stratified as levels of vulnerability.⁹ In an accumulated deficit model the number of deficits is expressed as a continuum of the total assessed. Frailty in the Fried model is present when three or more of the five criteria are met.⁸

Recent UK National Institute for Health and Social Care Excellence (NICE) guidance discourage the use of frailty performance tools in individuals who are acutely unwell.³³ This means that, ideally, all patients with potential frailty should be assessed in the community, and their index or score made available to clinicians in the event of an acute admission.

Prevalence and impact of frailty in cardiovascular disease

Frailty is more common in older people. Depending upon the definition used and the location studied, the prevalence in community-dwelling populations is reported to be between 4–59% with a weighted prevalence of frailty of ~10% for ‘physical frailty’ and ~14% for a broader frailty phenotype.³⁴ Frailty appears to be more prevalent in women and in people of lower educational status and income. Geographic variation also exists, with a higher prevalence in Southern than Northern Europe, and in African Americans.^35,36 The prevalence of frailty in nursing homes is higher than in community dwellers.³⁷

Frailty is reported to be more common among people with cardiovascular disease. The relationship may be bidirectional with frail subjects more likely to have cardiovascular disease, and those with subclinical cardiovascular disease more likely to have functional impairment.³⁸ The prevalence of frailty in cardiovascular disease reflects the rise seen in general populations – with 20% of patients aged over 65 years receiving percutaneous intervention (PCI) and 27% of patients aged over 70 years receiving coronary angiography reported as having frailty.^39,40 Rates of frailty in older patients, for example those with heart failure (HF) and receiving trans-catheter aortic valve replacement (TAVI), are even higher.^41,42

There are extensive data supporting an association between increasing frailty and an increased mortality risk. The Canadian Study of Health and Aging reported an adjusted odds ratio for death at five years of 4.82 (95% confidence interval (CI): 3.74–6.21) among mildly frail people and 7.34 (95% CI 4.73–11.38) among severely frail people. Frailty was the most important predictor of death and institutionalization with an odds ratio of 7.28 (95% CI 5.01–10.58) among mildly frail people and 8.64 (95% CI 4.92–15.17) among severely frail people.⁴³ In the Women’s Health and Aging Study, frail women had a six-fold higher risk of death and a more than 10-fold higher risk of disability and nursing home entry.⁴⁴

For people with cardiovascular disease, the Cardiovascular Health Study found a similar predictive association between frailty and pre-frailty status and falls, reduced independent function, and incident hospitalization and death over three or seven years of follow-up.²⁴ An association between frailty and poorer clinical outcomes including death has also been shown for patients aged over 65 years receiving percutaneous coronary intervention (PCI).³⁸ The addition of frailty, assessed using the Fried criteria, to conventional risk scores improved the prediction of mortality in the TaRgeted platelet Inhibition to cLarify the Optimal strateGy to medicallY manage Acute Coronary Syndromes (TRILOGY) study of patients with acute coronary syndrome (ACS).⁴⁵ Frailty is also strongly associated with worse outcomes for patients with heart failure, and following transcatheter aortic valve implantation (TAVI) or cardiac surgery.^46–50 Moreover, a recent systematic review and meta-analysis found that frailty assessment in an already vulnerable TAVI population identifies individuals at even greater risk of poor outcomes, and the use of objective frailty tools may help to inform patient selection.⁵¹ Indeed, the prognostic information afforded by measures of frailty may be greater than that predicted by conventional risk models in a wide range of cardiovascular conditions.^41,49,52,53

Implications of frailty for the treatment of cardiovascular disease

In medical care, the optimal outcome of any treatment is dependent upon the inherent safety of the treatment and the likelihood of benefit in the person to whom it is administered. The presence of frailty is clearly associated with poorer outcomes and must, therefore, be taken into account when assessing the balance of benefit. There is a case for a tailored approach to the identification of components of established frailty in order to allow an optimal strategy for care, such as the anticipation of post-procedural cognitive problems following a necessary interventional strategy.

An evaluation of frailty may also facilitate decisions regarding the value of any contingent intervention – that is, where functional improvement and prognosis depend upon baseline status. Therefore, in patients where the benefit of an intervention accrues over time, the identification of increasing physical and cognitive frailty could suggest a more supportive approach. Examples would include decision-making around primary prevention implantable cardioverter-defibrillator (ICD) insertion or TAVI. In such contingent scenarios, the management strategy for a frail individual must take account of the reduced likelihood of realised benefit, but also the possibility of iatrogenic curtailment of limited remaining independence and quality of life.

General principles of management

Introduction

Inpatients who are frail present an increasingly common challenge for physicians and represent a high-risk subgroup. The prevalence of frailty is now greater than 40% of those aged over 75 years admitted to a cardiology unit with an acute cardiovascular disease.⁵⁴ Given that most trials on which guidelines are based do not include the frail, there are limited data on which to inform decisions.⁵ In addition, the potential reversibility of frailty in its early stages and the predictive value for adverse outcomes at older ages⁵⁵ would suggest that a coordinated approach across primary and secondary healthcare is essential.

Comorbidity and polypharmacy in frail patients with acute cardiovascular diseases

Co-morbidities increase the risk of frailty, an outcome of which can be disability.⁵⁶ Co-morbidities are frequently seen in frail patients with acute cardiovascular disease, and contribute to poor outcomes and an increased length of stay. In one study almost 40% of hospitalised patients with HF had five or more co-morbidities, and this group accounted for 81% of the total hospital days experienced by all HF patients.⁵⁷ Moreover, hospitalization becomes more likely as the number of comorbidities increases. Chronic obstructive pulmonary disease (COPD), renal failure, diabetes mellitus, and depression are independently associated with a higher risk of mortality and preventable hospitalization.⁵⁷

Older patients with comorbidity are at particular risk of adverse drug reactions from polypharmacy. In particular, the use of five or more drugs per day is associated with increased drug interactions.⁵⁸ Polypharmacy is also associated with the use of drugs which are not clinically indicated.⁵⁹ During hospitalization, adverse drug reactions in older people are more frequent in those using at least one inappropriate agent than in those not taking such drugs.⁶⁰ Drug-induced adverse events may also trigger the prescription of further drugs in a ‘domino effect’, for example proton pump inhibitors to counteract indigestion from anti-platelet agents. Knowledge of the screening tool of older people’s prescriptions (STOPP) and screening tool to alert right treatment (START)⁶¹ and Beers criteria,⁶² two lists including drugs of particular risk to be avoided in older people, may help to improve the quality of prescription in frail patients.

Older people with acute cardiovascular disease and frailty or co-morbidity are under-represented in randomised clinical trials, which has resulted in a paucity of evidence in these patients.⁵ This has rendered many aspects of their care largely empirical, based on results obtained in trials with younger patients, even if it is known that older people respond differently. Frailty and co-morbidity often limit the implementation and tolerability of conventional cardiovascular therapies. For example, diuretic therapy in a frail patient with heart failure is more likely to lead to urinary incontinence, electrolyte imbalance, progression of renal dysfunction, delirium and falls. Use of vasodilators in a frail person can more frequently lead to orthostatic hypotension due to co-existent arteriosclerosis and autonomic dysfunction.⁴² Thus, the ‘dose objective’ should not be that demonstrated to be effective in clinical guidelines, but the one that keeps the patient active and without symptoms of dizziness.

The general principle of ‘start low and go slow’ is of even greater importance in frail patients. Antithrombotic and anticoagulant treatments in ACS are another example of drugs of greater risk to frail patients. The choice of agents with better risk/benefit ratio (e.g. avoiding prescription of prasugrel over the age of 75 years) is recommended in these patients,⁶³ while adjustment of doses is crucial to minimise the higher risk of bleeding, particularly with parenteral drugs and oral anticoagulants.⁶⁴

Evidence-based medicine has evolved through research on the management of single diseases. Less attention, however, has been paid to the investigation of the optimal management of coexisting chronic and acute diseases. These may be of equal importance and have competing management strategies, for example acute heart failure in patients with renal failure. On the whole, clinical guidelines tend to focus on treating the primary disease and a more patient-centred approach is needed to manage cardiovascular disease in the context of multiple chronic conditions, geriatric syndromes and reduced functional status.^3,65 Cardiovascular disease may not be the main protagonist, but a component in the treatment of a complex patient.

Management of frail patients with acute cardiovascular disease and delirium

Delirium is an acute disturbance of consciousness and cognition and can be considered as an ‘acute brain failure’ resulting from noxious insults or stressful situations. Delirium may indicate a loss of cognitive reserve, and is a frequent complication during hospitalization of frail older people.⁶⁶ It is under-recognised, and associated with poor outcomes.⁶⁷ The incidence in older patients hospitalised for acute cardiovascular disease may exceed 17%,⁶⁸ and is higher in those admitted to cardiac intensive care units.^69,70

Patients with delirium have a longer hospital stay, higher mortality (51.4% during the first year) and higher rates of functional decline.^68,69 At three months, 40% of patients who suffered delirium have global cognition scores 1.5 standard deviations below the mean of the population, and similar to scores of patients with moderate traumatic brain injury.⁷¹ The assessment of delirium during admission is important, as are measures to prevent it occurring. Simple screening tools, such as the Confusion Assessment Method or the intensive care version CAM-ICU for critically ill patients with ventilation support can be used.⁷²

Delirium prevention during hospitalization for acute disease is challenging, but different multicomponent non-pharmacological interventions have demonstrated benefits,^73,74 even in intensive care units.⁷⁵ All patients at risk should be assessed within 24 h of admission, and interventions should be tailored to their particular needs.⁷⁶

Pharmacological treatment such as haloperidol or other anti-psychotics are not effective in the prevention of delirium, and are not recommended.^66,77 If sedation is required, then the use of lighter levels of sedation with short-acting sedative agents is preferable. There is some evidence to recommend the use of dexmedetomidine, a non-benzodiazepine sedative agent with low associated levels of delirium, in the coronary or cardiac intensive care unit.⁷⁸

If a patient does become distressed, verbal and then non-verbal de-escalation techniques should be used. Pharmacological intervention should be reserved for patients with severe agitation who risk interruption of essential medical treatment (e.g. intravenous infusions), who risk injury to themselves or other, and for those with distressing psychotic symptoms. Antipsychotic drugs are used in this context to control symptoms, and may be started at low doses and titrated until the effect is achieved. NICE guidance recommends the use of haloperidol first-line, although this should be avoided in patients with Parkinson’s disease or Lewy body dementia.⁷⁶ A baseline electrocardiogram (ECG) is recommended, as haloperidol can prolong the QT interval. If anti-psychotics are contraindicated, lorazepam is an alternative for short-term use (up to a maximum of 48 h). The patient should be monitored for sedative and hypotensive effects. Based on data reported from randomised trials, the use of anti-psychotics is not associated with changes in delirium duration, severity, length of stay or mortality.⁷⁷

Acute coronary syndromes

Introduction

ACS is a major stressor, and the presence of frailty or pre-frailty is an independent risk factor for mortality.⁷⁹ This may in part be due to an underuse of guideline-recommended therapies in the frail,^80,81 but it is also recognised that frail patients are under-represented in the trials that underpin the guidelines.⁵ Frail individuals with an ACS are less likely to be admitted to intensive cardiac units,^14,82,83 which may be appropriate in some cases. They have longer lengths of hospital stay,^14,79,82 and are more likely to be readmitted with recurrent cardiovascular events.^14,83,84

The evidence base in frailty and ACS

As with other acute presentations, ACS presents a challenging situation for evaluating frailty, and the population-based frailty scores are not ideal in this situation.¹⁵ However, the Fried score and the EFS have been used in the setting of acute myocardial infarction, with a reported frailty prevalence of 27–34%.^79,85,86 Simpler indices using gait speed or blood biomarkers have also been investigated.^87,88 Regardless of the index used, frailty is a strong predictor of mortality in ACS after adjusting for prognostic factors including age. Likewise, frailty added significant prognostic information over the Global Registry of Acute Coronary Events (GRACE) score in this group of patients.^86,89

ST-elevation myocardial infarction (STEMI)

There is a paucity of data concerning frail patients with ST-elevation myocardial infarction (STEMI). The evidence that exists mainly relates to the very elderly, and does not take into consideration whether or not they are frail. Small-sized clinical trials demonstrate the superiority of primary PCI in comparison to fibrinolytic therapy in older people, but without specific reference to the frail.⁹⁰ One recent trial (the TRatamiento del Infarto Agudo de miocardio eN Ancianos (TRIANA) study) pooled its data with those from two previous trials, which corroborated evidence of the benefit of primary PCI in patients over the age of 75 years.⁹¹ A pooled meta-analysis of individual patient data from 22 randomised trials showed that the benefit of primary PCI over fibrinolysis was independent of age,⁹² and yet the subset of patients aged 75 years and older had a worse outcome after primary PCI.^93–96 As expected, this worse prognosis was more evident at long-term follow-up, and classical cardiac risk factors played a less influential role than they did in young patients.^97,98 This, perhaps, implies that co-morbidity and frailty are of relevance. Slow gait speed measured at discharge was associated with a worse long-term outcome after primary PCI.⁸⁷ Other factors that contribute to the poorer prognosis of older people treated with primary PCI are a longer ischaemic time, lower success rate of the procedure, higher frequency of Killip class>1 HF, more multi-vessel disease and higher frequency of bleeding complications.^95,96,99,100 Nevertheless, the immediate peri-procedural complication rate in the very elderly is acceptable, suggesting that primary PCI should not be routinely denied to frail patients.

The benefit of fibrinolytic treatment in those over 80 years of age is questionable in view of bleeding risk and haemorrhagic stroke.¹⁰¹ A sub-analysis of the STrategic Reperfusion Early After Myocardial infarction (STREAM) trial suggests that using half-dose tenecteplase reduces haemorrhagic stroke risk without compromising reperfusion efficacy.¹⁰² This approach may be a reasonable alternative when primary PCI is not available.

Non-ST-elevation myocardial infarction (non-STEMI)

No clinical trials have been specifically designed to investigate the management of frail patients in non-STEMI: however, there are relevant data available. Comorbidities are a strong predictor of one-year mortality following non-STEMI.¹⁰³ Likewise, in a series of non-STEMI cases, frailty predicted both 30-day and mid-term mortality, as well as cardiovascular events. This was independent of co-morbidities and other risk factors.^14,82,89 In another study, which analysed a wide range of geriatric conditions (physical disability, instrumental disability, cognitive impairment and comorbidities), frailty captured most of the prognostic information derived.⁸⁶

Although there are no data about the best management of the frail patient with non-STEMI, the guidelines recommend a routine invasive strategy.¹⁰⁴ Still, invasive management is underused in the frail and among co-morbid patients.^79,82,103 Furthermore, people with frailty who receive coronary angiography have lower revascularization rates compared with their less frail counterparts.^79,82 This may be, in part, a reflection of the more complex coronary artery disease identified in those with frailty, who were more likely to have left main stem disease, three-vessel disease or proximal disease, which may not be amenable to PCI.⁴⁰ Conversely, it is also possible that lower rates reflect an aversion to the perceived risk of invasive management in frail patients, whereby those with some potential to gain benefit may have been deemed inappropriate for coronary intervention.¹⁰⁵ After PCI or cardiac surgery, frailty and comorbidity are associated with adverse long-term outcomes.^40,106 Co-morbid patients, however, could benefit the most from in-hospital revascularization in non-STEMI.^107,108

Although frail patients have usually been excluded from randomised clinical trials, there are some data regarding older people. Savonitto et al. conducted the first randomised trial comparing invasive and conservative strategies in older people with non-STEMI.¹⁰⁹ No significant differences were observed. This is probably because of the low statistical power resulting from the small sample size. The ‘After Eighty’ randomised trial included patients aged over 80 years with non-STEMI and demonstrated a benefit from an invasive strategy in terms of the composite endpoint of death or cardiovascular events at 1.5 years of follow-up.¹¹⁰ In the conservative arm of this study, no patient underwent cardiac catheterization under any circumstance. Recently, a randomised trial of older people with non-STEMI was performed which focused on patient comorbidities.¹¹¹ Although the small sample size limited the conclusions, the invasive strategy reduced the probability of death or ischaemic events at three months but this effect disappeared at the 2.5-year follow-up.

Studies suggest lower rates of angotensin converting enzyme inhibitor (ACEi), angiotensin receptor blocker (ARB) and statin prescribing following ACS in frail patients.^87,89 Prescribing decisions are especially complex for those with frailty, who are at increased risk of medication-related side effects.¹¹² This may be a particular problem for frail ACS patients where guideline-mandated treatment includes drugs with a high risk of bleeding or hypotensive complications. Careful clinical judgment is required to weigh up the risk: benefit ratio, based on likely duration of treatment.⁶⁴ Some patients with advanced frailty are entering the terminal phase of life and a decision to withhold a medication that cannot provide long-term benefit, but might increase polypharmacy and the risk of side effects, may be considered appropriate.

To date only one prospective study has used a validated tool to assess frailty in the context of ACS.⁸⁹ Within this study, prevalence of frailty was notably low at 4.7%, compared with the median prevalence estimate across studies of 38.1%. In this trial, cardiovascular and all-cause mortality rates for people with frailty were lower than results reported in the observational studies. This may be because the trial participants did not include people with advanced frailty who were in their terminal phase of life, which would be consistent with the relatively low study prevalence of frailty. The study did show that frail patients benefited more with respect to outcomes if they received prasugrel compared to clopidogrel, however, the prasugrel group had higher rates of bleeding complications. The bleeding complications may in part be due to a lower use of proton pump inhibitors in the frail group, although these drugs also have the potential for side effects such as hyponatraemia.¹¹³

There is a relationship between age and bleeding risk in patients on dual antiplatelet therapy,¹¹⁴ but the same has not been demonstrated for the frail.⁸⁹ On the whole, physicians tend to avoid the new more potent antiplatelet agents in older or frail patients based on the risk of bleeding. Yet, risk estimation by clinicians is imperfect.¹¹⁵ The ‘POPular Age study’ will compare clopidogrel versus more potent antiplatelet agents (ticagrelor or prasugrel) in patients aged over 75 years old in order to provide guidance on this issue.¹¹⁶

Future research

Lower levels of provision of ACS care may be the result of a lack of evidence from randomised controlled trials to guide ACS management in the context of frailty (see Table 1). The under-representation of older people with frailty in randomised studies of ACS interventions risks excluding those at greatest risk of adverse outcomes following ACS and limits the generalizability of trial findings. Prospective data on the outcome of patients with ACS and frailty are required. Frailty or measures of frailty should be included as baseline characteristics in all prospective ACS registries.

Future trials of invasive and pharmacological ACS management should include methods to select and stratify participants with frailty, as an especially high-risk group. Robust randomised evidence may help guide appropriate decision-making when considering invasive and pharmacological management for older people with frailty. Further research is needed in the frail with cardiogenic shock and/or after cardiopulmonary resuscitation. These patients have the highest mortality and might have the largest benefit from early revascularization therapy.

Heart failure

Introduction

Frailty is present in 50–70% of older people hospitalised with acute heart failure (AHF).^117–120 It is associated with in-hospital,¹²¹ short- and long-term adverse outcomes,^46,117,119 and so an understanding of the interaction between frailty and AHF is critical for acute cardiovascular physicians.

Management of heart failure in the frail patient

The recognition of frailty may help to detect older people at risk of poor outcomes (i.e. disability, death, hospitalization or institutionalization)^122,123 and it is, therefore, important to identify those patients that may benefit from early intervention and close monitoring during and after hospitalisation. The optimal tool and time to diagnose frailty in AHF patients have not yet been established.¹¹ Most of the current evidence regarding AHF is based on the previously discussed frailty phenotype (Fried Scale)²⁴ or from some of its domains (i.e. physical performance)¹⁷ such as in the Short Physical Performance Battery (SPPB).⁴

Frailty should be considered at the time of stratifying AHF older patients in the emergency department setting¹²⁴ as well as in those undergoing invasive procedures^125,126 such as TAVI,^127,128 cardiac surgery,¹²⁹ cardiac resynchronization therapy¹³⁰ and post-PCI,⁴⁰ all of which may be required in frail patients with HF. However, the assessment of frailty in the acute phase of hospital admission undoubtedly provides the greatest challenge. Taking into account the acute phase of the HF condition, self-reported assessments may be more appropriate at hospital admission, while more objective performance measures would be more suitable at hospital discharge.

The evidence base in frailty and heart failure

Table 2 summarises the main studies published regarding the prognosis of older patients with frailty in AHF. The Fried criteria predict 30-day functional decline, one-year all-cause mortality and readmission in non-dependent patients aged 70 years and older admitted to hospital with AHF.¹¹⁹ With respect to physical performance tests, a total SPPB score ≤4 applied at hospital admission is an independent predictor of the length of hospital admission,¹²¹ one-year mortality and rehospitalization^46,117 in older patients with AHF.

Several self-reported scales (e.g. Tilburg Frailty Indicator (TFI)) or those derived from comprehensive geriatric assessments (e.g. MPI, CGA score and EFS) have been shown to have a high predictive value of adverse short-term outcomes in AHF.^32,131–134 The MPI predicts the 30-day mortality in patients aged 65 years and older hospitalised with AHF.¹³¹ The CGA score estimatesthe in-hospital and two year mortality in older patients hospitalised for AHF.^32,132 The TFI is a self-reported questionnaire that addresses determinants and components (with physical, psychological and social domains) of frailty, and a study has demonstrated that social components of the frailty syndrome adversely affect the ability of self-care in older patients with AHF.¹³³

Future research

There is a paucity of evidence about optimal care of the frail older patient with AHF. This is because clinical trials have excluded this population and fail to consider other outcomes such as functional and cognitive deterioration, institutionalization, out-of-hospital days alive or quality of life. It is also necessary to unify how, when and where frailty should be diagnosed in older patients attending hospital with AHF. Notably, frailty is a potentially reversible state and, therefore, research should also focus on better understanding the pathophysiology of this syndrome with the aim of identifying new biomarkers and treatments. As such, a necessary next step is to develop high-performing risk models that predict and discriminate short- and long-term prognosis and quality of life. Early holistic intervention, monitoring and care transition recommendations in frail older patients with AHF should be defined, integrating patient preferences into the decision-making process.⁵

Atrial fibrillation and other cardiac arrhythmias

Introduction

Atrial fibrillation (AF) is the commonest sustained cardiac arrhythmia and, like frailty, is more prevalent with increasing age. Over 10% of the population aged over 80 years have AF.¹³⁶ Much of the published literature on frailty and cardiac arrhythmias focuses on AF, particularly in relation to stroke prevention (Table 3).

Oral anti-coagulation for stroke prevention

The relative benefit from the use of oral anticoagulants (OACs) for reducing stroke and mortality in the older person with AF is similar to that seen in younger subjects.¹³⁷ As with management of any AF patient, stroke and bleeding risks should be assessed, and re-assessed at appropriate review visits.¹³⁸ Stroke prevention in AF has been comprehensively addressed in many recent reviews.^139,140

For older people with AF, frailty assessment may influence OAC prescribing for stroke prevention, since frailty is associated with an increased risk of falls. Falls are often cited as a reason to stop OACs, but surprisingly they are not an independent predictor of bleeding on OAC.¹⁴¹ The AF patients who have sustained falls represent a high risk group for adverse events, although the net clinical benefit of stroke prevention is likely to outweigh the risk of serious bleeding.¹⁴² Other barriers to OAC prescription include history of bleeding, impaired cognitive status, increasing co-morbidities, polypharmacy and patient adherence. All of these common features seen in the AF population have been associated with increased frailty.¹⁴³

Assessment of frailty in patients with atrial fibrillation

Formal frailty assessment as part of a multidisciplinary team approach may help clinicians identify patients who are at increased risk of adverse outcomes from AF and/or OAC therapy.¹⁴³ A physician’s subjective judgement of frailty may lead to under-prescription of an OAC. The eFI may be used to identify older AF people with mild, moderate and severe frailty, with robust predictive validity for outcomes of mortality, hospitalization and nursing home admission.²⁸ Even so, a patient’s risk does not remain static and regular re-assessments should be part of the holistic approach to management.

In a study by Perera et al., frailty was the strongest predictor of the OAC therapy prescribed on discharge from a tertiary referral hospital, with frail patients with AF being less likely to be prescribed warfarin.¹⁴⁴ Importantly, they found that frailty identified a particularly high-risk patient group, as after six months frail participants were significantly more likely to die or to have an embolic stroke. They also appeared to have higher rates of major/severe haemorrhage, although this finding did not reach statistical significance.

A recent Canadian study found that the decision to prescribe OACs to octogenarians with AF was multifactorial and influenced by a complex interplay of thromboembolic risk, bleeding risk and frailty status.¹⁴⁵ Unsurprisingly, rates of OAC use were clearly diminished among patients who were severely frail with advanced disability. In a Dutch study by Tulner et al., age appeared to be the main determinant of OAC non-prescription among frail geriatric outpatients with AF.¹⁴⁶ After four years follow-up, the cumulative rate of bleeding causing discontinuation of OAC therapy was not much higher than in other usual-care studies. A population-based study in the United Kingdom found that higher risk older patients with AF were less likely to receive OACs, and were instead prescribed aspirin or no stroke-prophylactic treatment.¹⁴⁷ However, in older people aspirin is not safer than well-managed vitamin K antagonists (VKAs) or a non-VKA oral anticoagulant (NOAC) with regard to serious bleeding.^148–150 Nonetheless, patient values and preferences are important considerations regarding OAC use.¹⁵¹ Patients are desperate to avoid strokes, whilst physicians tend to focus more on bleeding concerns. In a study by Lahaye et al., patients were prepared to sustain four major bleeds to avoid one stroke.¹⁵² Patient values and preferences have been the subject of a recent consensus document from the European Heart Rhythm Association.¹⁵³

In AF patients, other practical considerations pertain to the use of rate or rhythm control strategies.¹⁵⁴ Frail older people at risk of falls, pre-syncope or postural symptoms should be assessed for underlying conduction abnormalities, including sinus node disease, which may be remedied by pacing. Excessive bradycardia due to drugs, such as rate-limiting calcium-channel blockers, beta-blockers and/or cardiac glycosides (e.g. digoxin) when given in combination may exacerbate symptoms, and regular medication reviews should be undertaken. The risk of drug interactions increases with reduced renal reserve as seen in frail older people and with polypharmacy.¹¹² Ultimately, decisions on rate or rhythm control should be patient-centred and symptom-directed, consistent with published guidelines.^155,156

Other arrhythmias

The general principles of a multidisciplinary, holistic approach to management apply to the assessment and treatment of the frail patient with other cardiac arrhythmias. With an ever increasing range of drugs and technologies available for treating arrhythmias, realistic expectations in the frail patient are needed in order to carefully balance quality and quantity of life.

In a recent study of 171 patients undergoing pacemaker implantation for sinus node dysfunction or atrio-ventricular block, a frailty syndrome was diagnosed in 25%, and pre-frailty in 37% of patients.¹⁵⁷ This may have an impact on patient’s ability to attend clinic for follow-up, and may increase the risk of erosion if there is associated poor nutrition. An ICD is a well-established option for patients at risk of significant ventricular arrhythmias and sudden cardiac death. However, 20% of older people die within one year of an episode of life-threatening ventricular arrhythmia, mostly from non-arrhythmic causes.¹⁵⁸ Certainly, the incidence of sudden death as a percentage of all-cause mortality decreases with age. A small Swedish cohort (n=76) reported that in octogenarians who are due for an elective ICD unit replacement, high one-year mortality rates suggested that the ICD was unlikely to offer a survival benefit.¹⁵⁹ They concluded that careful consideration should be given to the overall clinical status, comorbidities and frailty of patients in this situation.

The impact of frailty on critical care

Critical care use is increasing. There is a greater proportion of in-hospital beds devoted to critical care, and an estimated one in five patients now die in the intensive care unit setting.¹⁶⁰ Frailty, however, is a relatively new concept for critical care, although it is increasingly recognised as an important consideration when determining the appropriateness of intensive care admission. Assessment of frailty can now be used to help predict critical care outcomes, and more recently to address some of the pathological processes involved in critical illness and its complications.

Introduction

With an ageing population, patients aged over 80 years comprise 10–20% of critical care admissions, and this is increasing by 6% per annum across the world.¹⁶¹ Mortality in critically ill older people requiring intensive care remains high, in particular those with medical or emergency admissions, with an intensive care mortality of 38–64%, in-hospital mortality of 45–55% and one-year mortality of 80–89% in this patient group. The increasing demand for critical care usage, as well as poor outcomes in certain patient populations has led to greater consideration of the actual benefit that intensive care admission might provide to an individual. Decisions not to admit to the intensive care unit on the basis of perceived lack of benefit are frequent, however, predicting individual patient outcomes remains one of the major challenges. Increased age is associated with a worse mortality in critical care, resulting in clinician bias with respect to critical care admission. In one study, only 40% of patients assessed in the emergency department over 80 years who met criteria for admission were referred, and only 50% of those referred were subsequently admitted to the intensive care unit. The denial of intensive care admission based upon age is controversial, with some proposing that older people may actually benefit most from critical care admission.¹⁶² Indeed, the Triage decision-making for the Elderly in European Intensive Care Units (ELDICUS) consensus stated that ‘age should never be the sole determining factor in triage decisions’, and additionally that ‘physiological status is more important than chronological age in decisions’,¹⁶³ which is also our view. A number of studies have shown frailty to be an independent risk factor for intensive care, hospital and long-term mortality, as shown in Table 4.

Frailty is not limited to older people. The prevalence and impact of frailty has been evaluated in a number of studies involving cohorts of younger patients, including lung transplant candidates (median age 59 years) and those with end-stage liver or kidney disease (mean age 55 years).¹³ Recognition of frailty in younger, critically ill patient populations is increasing. In one study of patients admitted to six intensive care units (50–65 years, mean age 58.5±4 years), frailty was relatively common on admission (>25%, 34% classified as vulnerable) when compared with the age-matched general population (10–15%). In this younger patient group, there were no differences between frail/non-frail in terms of critical-care interventions, indications for admission and treatment intensity. However, frail patients had a high burden of co-morbid disease including HF. Further, one year mortality was higher, quality of life significantly worse, and re-hospitalization more common in those young patients with pre-admission frailty.¹³ These findings contribute to the increasing evidence regarding the burden of critical-illness survival on patients and their carers, including impaired physical function, neurocognitive and neuropsychological sequelae, increased healthcare costs and reduced quality of life.

Frailty and prognostication

A number of scoring systems are used to predict prognosis following critical-care admission. However, these were primarily designed to compare mortality between intensive care units, rather than predict the likely outcome for any one individual. These scores rely predominantly upon measures of acute physiological derangement at the time of admission or within the first 24 hours. They include limited demographic data, with no measures of prehospital functional status, disability or frailty. Indeed, in one French study two such commonly used scores (Simplified Acute Physiology Score (SAPS II) and Sequential Organ Failure Assessment (SOFA) scores) were unable to differentiate between frail and non-frail patients.¹⁶⁴ Increasing recognition of the burden that critical illness imposes upon the patient and their family, has led to the search for tools to predict who will benefit the most from critical-care treatment¹⁶⁵ beyond merely surviving to discharge, but also in terms of longer-term functional status and quality of life. It seems increasingly likely that including an assessment of frailty may enhance such decision-making tools.

Assessing frailty at the time of critical illness can be challenging due to the acuity of decision-making¹⁶⁶ and the need to obtain information from friends and family as, frequently, the patient is unable to participate actively in any assessment. It has been proposed that the CFS may be the most appropriate tool to use in the critical-care setting as it is flexible, relatively easy to use, and has been validated where only collateral history is available. It may predict outcomes more effectively than the commonly used intensive care unit illness scores alone.^164,167 It is well-recognised that the trajectory of symptoms prior to intensive care admission is a key determinant of outcome; not just in terms of mortality, but also independent functional status and quality of life.⁶ Thus, in addition to quantification of frailty on admission, obtaining a historical picture from relatives/friends with retrospective administration of the CFS has been proposed in order to provide an estimate of the patient’s premorbid health trajectory, and therefore their likely post-admission outcome. Further, the development of features measured in Fried’s operational definition of frailty during critical illness may provide additional predictive power. Evaluation of all three areas (premorbid trajectory, assessment on admission, and evaluation of measures of frailty during admission) when combined with existing illness severity scores potentially provide an in-depth objective measure of the status of the patient, and the likely outcomes following critical illness.

Frailty, physiological reserve and critical-care interventions

Selected patients of any age may benefit from screening for frailty, as well as early implementation of interventions aimed at minimising the impact of critical care on their physiological and psychological status. Although historically the main focus for improving outcomes from critical illness has been related to high-end interventions intended to support organ dysfunction and/or alter disease trajectory, the most successful interventions relate to minimising the physical deconditioning and psychological consequences of critical illness. These concepts are highly relevant in the frail patient population. Such interventions include implementation of care bundles such as attention to nutritional support, sedation interruption and early physiotherapy and mobilisation, using a multidisciplinary approach.⁶

The role of frailty in the critically ill is complex, and further confounded by the finding that critical illness may not only accelerate a patient’s premorbid frail state, but also lead to the development of many of the characteristics of frailty. In one study, frailty assessment immediately prior to hospital discharge following intensive care admission for respiratory failure demonstrated frail patients had a 41% six-month mortality, and that with each increased Fried phenotype domain, mortality increased three-fold.¹⁶⁸ The application of frailty scoring after intensive care admission may improve assessment of the physical, nutritional, cognitive and psychological disabilities present in survivors of critical illness. This in turn should promote interventions designed to minimise the longer-term adverse effects of intensive care.

Future research

Frailty as a measurable concept in critical illness is relatively new, and its true relevance is not yet known. Larger studies are needed to characterise cardiac patients most likely to benefit from frailty screening, and for development of more rigorous scores relevant for the critically ill. Additionally, the concept of the patient trajectory (before, during and after critical illness) demands further attention. The development of evidence-based interventions to enhance recovery, individualised for the patient is urgently required. The concept of frailty as an inflammatory process is particularly interesting when considering either the acceleration of frailty, or acquisition of some aspects of the frailty phenotype during critical illness, as this may provide an avenue for future research into the pathophysiology of both processes.¹⁶⁹ Finally, application of modified (or existing) frailty screening tools, particularly in our younger cardiac patients, will help to identify those at increased risk of critical illness. Early intervention may thereby avoid the need for an intensive care admission.

Conclusions

Which frailty model?

This review highlights the importance of the choice of frailty model used to evaluate any individual. It is evident that elements of the models overlap considerably, but it would also seem likely that different models lend themselves to different predictive uses. Presently, there are no objective measures of frailty that are sufficiently robust and validated to act as decisive arbiters in acute clinical decision making. The current effort directed toward the proliferation of frailty scores and initiatives might be better used defining the targeted utility of the available tools in specific contexts such as cardiovascular disease.

When to assess?

It is clear that there are problems with the assessment of frailty at the time of an acute admission. Many of the reported models require time and co-operation which is not always possible or practical in the acute setting. Of the currently available measurements of frailty acutely, the Rockwood Scale (or CFS) is probably the most useful.²⁷ Even so, given difficulties with acute frailty assessment, an obvious conclusion is that greater efforts should be made towards the identification, assessment, and management of frailty in the community. All data should then be made available on admission. Where elective cardiovascular procedures are undertaken (i.e. complex PCI, TAVI, cardiac surgery), evaluation of frailty using an appropriate tool should form part of the pre-assessment process.

Importance of frailty

Frailty may be a greater driver of clinical outcomes than conventional (‘medical’) measures of cardiovascular risk. Clinicians and, in particular, cardiologists are faced with an expanding range of possible interventions in aging and multi-morbid populations, with limited time to make decisions. The best outcome is what is best for the individual person being treated, and doing less may be better in some cases. To recognise frailty and the impact it may have on the individual alongside the benefit of any intervention is to start to understand the person independently of the labels clinicians provide, and is an essential component of optimal care.

Consensus statements

• Ideally all patients with potential frailty should be assessed in the community, and their index or score made available to clinicians in the event of an acute admission.

• Where elective cardiovascular procedures are undertaken, evaluation of frailty using an appropriate tool should form part of the pre-assessment process.

• For the frail, greater attention to pre-existing and post-procedure medication and co-morbidities is required. Knowledge of the Beers and ‘STOPP/START’ criteria may help to improve the quality of prescription in frail patients.

• The frail with cardiovascular disease represent a group at high risk of mortality, morbidity and medical complications. Decision making should involve a multidisciplinary team approach that takes into account patient values and preferences.

• For the frail with heart failure, the ‘dose objective’ should not be that demonstrated to be effective in clinical guidelines, but the one that keeps the patient active and free of restricting symptoms such as dyspnoea or dizziness.

• The frail with acute STEMI should not be routinely denied primary PCI. Equally, consideration should be given to limited intervention, where appropriate.

• There are no data about the best management of the frail patient with non-STEMI. Lower levels of provision of ACS care may be the result of a lack of evidence from randomised controlled trials.

• Frailty is common in older people with AF. Decisions regarding rate or rhythm control and the prescription of OACs should be patient-centred and symptom-directed, with careful consideration needed of overall clinical status, comorbidities and frailty.

• Physiological status is more important than chronological age when making decisions regarding acute cardiovascular care for the frail.

Conflict of interest

The authors declare that there is no conflict of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

References