Stroke Medicine Is Frailty Medicine: Clinical and Research Priorities for Frailty in Stroke

TQ: Frailty is interesting because it is one of those terms we all use, often loosely, but what has emerged over the last decade or so is a much more precise scientific understanding of what we mean by “frailty.” There are lots of frailty theories, but they all have a common thread: frailty is about systems not coping with a stressor, and this tends to be age-related. Applying that abstract theory to stroke, we can think of stroke as being an insult that a robust older adult is able to weather, but in frailty, the outcomes from that stroke insult are exaggerated much more than would be expected for a person of that age.

Although age is a very important part of frailty, age and frailty are not synonymous. The other concept that often gets conflated with frailty is disability. Again, there is a lot of overlap of disability and frailty, but they are they are different constructs.

I guess the second part of any definition of frailty is quite important: why are we defining this thing and why are we interested in it? I would say that the reason I became interested in frailty is because it is so common. There have been a number of epidemiological studies looking at the prevalence of frailty in different settings. A lot of the literature comes from the acute stroke setting, and it is fairly consistent that about one in four people admitted with an acute stroke are living with frailty before their stroke, so this is not a niche condition by any means. I would argue that stroke medicine is frailty medicine.

JM: To the points that Terry raised, we have a population that is growing older with comorbidity, and this presents challenges. Stroke has a significant economic and societal burden, and the ability to improve prognosis by consideration of frailty on the individual’s disease trajectory, and to begin to think about how treatment may be attenuated or perhaps focused on the right individuals, is an important issue. So this is a problem that is not going to go away.

LB: I think something that struck me from the work that I have done looking at frailty in transient ischaemic attack (TIA) or minor stroke is that the stressor event does not have to always be that significant. Even in quite minor events like TIA and minor stroke, we see that frailty has an impact on outcomes. In our TIA clinic, we found frailty had a significant impact on mortality after even a TIA, so I think this is a really important concept to consider for how stroke survivors move through the care pathway depending on their level of frailty.

NE: Terry’s mentioned one in four individuals presenting with acute stroke have frailty, but actually when you measure those who are on their way to becoming frail – those with pre-frailty – the burden increases dramatically, with almost three-quarters of individuals having some degree of pre-frailty or frailty. I expect that figure will rise with the ageing population. So, we have this demographic time bomb on our hands where frailty is going to play a large role in what we will see coming in acutely to the stroke unit.

TQ: Another reason I was attracted to looking at frailty in stroke is that stroke survivors have told us this is important to them. In the UK, a research priority setting exercise – led by The Stroke Association and NIHR – identified one of the research priorities to be an understanding of how frailty can modify the stroke journey. So, that was coming from people with lived experience of stroke and that is very compelling to me that this is something that is important to stroke survivors, so it is a place where we need to be doing robust research and looking at clinical pathways.

LB: There is a plethora of different ways of assessing frailty. They range from doing a face-to-face clinical assessment with the patient, with probably the best known being the Rockwood Clinical Frailty Scale. Frailty can also be characterised in terms of phenotypes, and probably the best known of these is the Fried phenotype, which has constructs such as weight loss, feeling exhausted all the time, reduced walking speed, and poor hand grip strength. Those are things that geriatricians or stroke physicians see commonly in their patients. There are also electronic ways of measuring frailty: you can use a cumulative deficit model based on the total number of comorbidities that a person has in their health records and that can be in primary care or in the secondary care healthcare records system, to determine a frailty score. Those tend to be used in the research setting and also maybe in primary care to try to identify or characterise people who might be at risk of deterioration or where you might want to start thinking about advance care planning.

I woul’d say the most common assessment, and certainly one that is become embedded in our hospital, is the Rockwood Clinical Frailty Scale. I think people prefer the face-to-face assessment where you are able to get a real feel for that person; what they are able to do for themselves and how they function in everyday life. It can also be used as a common language between specialities when communicating frailty.

Probably the last thing that I want to mention with regards to frailty assessment is the difference to disability. I am sure that everyone is very familiar with the modified Rankin Scale (mRS) used as a major trial outcome in stroke studies, and there has been some interest in looking at whether there is an overlap between the mRS and frailty. Sometimes they are considered to be synonymous, but frailty really should be viewed as a different concept and we cannot just replace it with the mRS. I think we do need to be measuring frailty independently and separately to the mRS. It will be nice to see that coming through in future studies and trials in stroke.

NE: I agree: I have been struck by the number of times where I have seen somebody that has disability but not frailty, and vice-versa. A number of studies have shown a variable relationship between frailty and mRS: somewhere between a correlation of 0.5 and 0.8, so they capture something very different.

JM: I wanted to add that there is a bit of tension in the idea as to whether we should be assessing frailty from an individual domain perspective – for example, thinking about physical, cognitive, or brain structural change – or whether it is the total picture of combining these individual components. I think that as we move forward, we may get a better perspective on operationalising these tools.

TQ: I just want to come back to Jatinder’s point because I think it is fundamental: I hear a lot of people who dismiss the frailty concept say due to the many different ways of measuring it, and I want to just challenge that. I think in any area of medicine, you have different ways of assessing things, and you assess things according to the purpose of the assessment and the resources you have, and that assessment can move from repurposing data that is already there, through to a screening or a triage assessment, through to the clinical formulation, through to the very detailed phenotype needed for research. That is standard, and the same exists in frailty, so these different methods for assessing frailty I see as being complementary, rather than conflicting. I think what I would encourage clinical services to do is to choose the one that works best for their service and use it.

TQ: I think you cannot have a panel talking about stroke – a brain disease – and not consider brain frailty. Frailty is a multisystem syndrome, and in the same way that it can affect physical systems, it can affect cognitive systems. Now, we have a definition of cognitive frailty that says cognitive frailty is the mix of physical frailty and cognitive decline that is not yet sufficient to have a dementia diagnosis. I think that is helpful to an extent, but I think what is more helpful is to go back to basics and that original definition of frailty, which is that the system does not respond well to a stressor. You can think of the brain as a system, and if there is a stressor like stroke, it may not respond well to that, and that may manifest as further cognitive decline or other cognitive symptoms like delirium. All of those to my mind are manifestations of cognitive frailty.

TQ: One of the theories of frailty is of cumulative deficits: insults to the system that have accumulated over the lifespan and each one makes that system more vulnerable to the next insult that comes along. In the brain, we can see radiological evidence of various historic insults: you may see silent stroke, old traumatic brain injury, cerebral small vessel disease, and atrophy. Each of these makes the brain more vulnerable, and when they are present cumulatively, that vulnerability increases. There is no magic to brain frailty, you are just using the scan that you order as part of routine stroke care and looking beyond the acute stroke lesion to evaluate what other evidence is there that suggests this brain may not respond so well to the stroke.

LB: Ideally, you want to be assessing frailty from before the hospital admission, usually more than 2 weeks. If you scored most older people in the hospital bed on hospital admission then they would come out as highly frail, so it is not reflective of them at their baseline prior to hospitalisation. This can lead to inappropriate decision-making and that is something that we have sometimes seen in clinical services. Obviously if that person has been declining for some time, it can be challenging to decide at what time point you should measure frailty: is that decline a result of an underlying cognitive impairment that means that it is going to continue to progress, or is it something that is more of an acute reversible illness where you might be inappropriately frailty scoring somebody if you do it too close to the hospital admission? So it is really important to consider the timing of your assessment.

TQ: I am a pragmatist. If I am working with a research dataset and it has sufficient data to allow a frailty index to be calculated, it just seems obvious to me that you would want to look at frailty as part of the baseline descriptors. Every clinical service is extremely busy, so we cannot ask clinical teams to start doing research-level frailty assessments. We need something that is quick, we need something that can be done by any member of the team, and we need something that has training available but does not require lots of training. In our service, we have tended to use the Clinical Frailty Scale that Lucy spoke about earlier.

JM: Just to add, I think standardising the recording of frailty assessments within your institution also enables a culture of understanding how that tool functions, enables, and empowers other members of the multidisciplinary team to be able to perform those assessments. Importantly, if you are collecting it from the outset, you can track how patients are re-admitted, and you can look at the trajectory over a period of time. The difficulty is that you have got interfaces of care, such as moving from inpatient care towards rehabilitation hospitals and beyond, so how do we ensure that frailty recording continues? I think that there needs to be some implementation work around how that gets carried forward and tracked. Clearly, data are very powerful if it is collected through the entire stroke pathway.

NE: Fortunately, we work in a field where there is a lot of data generated across the multidisciplinary team, so I very often do use a frailty index in the research setting, and the advantage of it is that it captures the data from across biochemical results, functional results, and comorbidities. However, it takes time. I suppose a challenge for us in the field is the number of different frailty indices, and the slightly different ways of approaching them, can make it very difficult to compare one study to the next. Should we be working towards a standardised frailty index for use in stroke that means that we can then compare populations and facilitate meta-analysis across a number of different studies?

TQ: Standardisation is always powerful; it removes variation and allows you to compare things, so I am always in favour of standardisation. However, I have seen some real crimes against frailty committed under the auspices of a frailty index with people putting in values that really are not related to older age and are not related to system decline and hence should not be in a frailty index. I would point out that Ken Rockwood and his team have made a guide with best practice on how you create a frailty index, and I think if we are going to move to standardisation, we should really be looking to emulate that best practice.

TQ: An argument that I hear frequently is that by assessing frailty we may be guilty of a therapeutic nihilism; that we may use frailty as a tool to deny treatment to older adults. Again, I want to challenge that because although we have clinical practice guidelines, we are not automatons. We make clinical decisions on every patient we see, and we make those decisions in a patient-centred way. We consider how they are going to react to the interventions that we are going to give. What I see frailty doing is just formalising and standardising that assessment. Going back to some of the points that Lucy made before, it gives us a common language so that other people in the team can understand that assessment. It also allows us to quantify things because I do not see frailty as being binary present or absent; it exists across a continuum. Doing the assessment allows us to give that gradation and understand various levels of frailty that might exist.

JM: I definitely agree with that point. I think one of the challenges we have got is whether we are going to operationalise frailty, and we also need to answer some of the questions from critics. For example, how confident can you be that frailty models link to the lack of physiological reserve, and therefore, how resilient is that individual likely to be in the face of certain treatments? We know that where disease is often severe and has a high morbidity and mortality, for example, intracerebral haemorrhage (ICH), this nihilism exists because we believe that we are unlikely to secure a good outcome. However, simple interventions like bundled care in ICH have led to a revolution whereby mortality can be reduced. Understanding how frailty is a modifier in these instances is absolutely crucial. I think what we have not yet got is population-level studies around frailty and thinking about them on a national scale if we are going to implement and standardise the use of a particular tool. I think we have got a little way to go still to convince some people that these tools are robust, but increasingly they are the best way to think about some of the terminology that gets banded around in terms of disability, dependency, multimorbidity, and frailty. These are quantitative measures that take away some of that subjectivity and that can surely only be good.

LB: I wanted to highlight rehabilitation – and I think this links nicely with discussions in general geriatric medicine outside of stroke care – particularly in relation to rehabilitation goals and labelling people as “not having rehab potential.” Whether frailty moderates and influences decisions about who goes to rehab, how much rehab they receive, and how intensive it is, is not something that has been looked at either in clinical services or from a research perspective. I think we need to do a lot more on how we integrate frailty not only into our assessments and scoring, but how it frames our treatment discussions and decisions with patients as well.

NE: I would echo this and the importance of avoiding “treatment nihilism.” There have been a number of studies that show individuals with frailty may not benefit from some standard interventions from which individuals without frailty do benefit. However, I think there is a risk of missing an important point: does that indicate that somebody with frailty is not going to benefit from any intervention, or does it suggest we should be looking at how to adapt these interventions for when stroke and frailty co-exist? A number of specialities have started doing this – cardiology being an example – where rehabilitation may target both the heart but also frailty more generally. I think we should be looking at the same in stroke. Maybe we are just failing to adapt our treatments and interventions to this different population, rather than treating everybody in a uniform fashion.

My other consideration is about hyperacute treatment. There is increasing evidence indicating that frailty has a treatment-modifying effect on reperfusion therapies. However, frailty is not a binary entity and represents a spectrum. I do worry that this overlooks some individuals who have intermediate frailty – but who have not necessarily reached frank frailty yet – where they may continue to have some benefit from these acute reperfusion therapies, albeit attenuated. However, individuals with frailty are often not being enrolled into studies or being considered for treatment, so I think we need better evidence to be able to come to more robust decision-making for individuals with frailty. That is hopefully an area that we can work towards in the future.

JM: This has relevance for clinical trials and end points. In order to think about how well some of the treatments we deliver affect outcomes, we have got to think about frailty being recorded throughout these trials. A good example is how we utilise the mRS in clinical practice to make some decisions around mechanical thrombectomy, but actually, if those trials had included the Clinical Frailty Scale as one potential tool to enable decisions, would more patients who may benefit from thrombectomy have received one (given its potential as a prognostication tool for case selection, particularly when individuals may have disability but not frailty)? So I think it is interesting to think about how some frailty tools may replace the mRS in the future, particularly in hyperacute therapy decision-making. I would love to see future clinical trials being empowered to consider frailty as part of the inclusion and exclusion criteria, and also carried forward all the way through the trial.

LB: Building on some of the things said above, we know that frailty is extremely prevalent in stroke and TIA populations, and there is a consideration that the presentation with stroke/TIA is a reflection of the frailty of the individual as well. The fact that they are presenting because they are vulnerable – and the stressor has unmasked that vulnerability – may frame discussions and decisions around secondary prevention and further investigations, and what may or may not be appropriate for individuals depending on their preferences.

JM: I recall some studies around carotid revascularisation that identified frailty was independently associated with post-procedure mortality and stroke. So I think it is important that we understand that there are some large-scale datasets demonstrating some interventions may not be beneficial for some populations. However, let’s also remember that a lot of historical secondary prevention trials did not include individuals of older age.

There is also the concept that those with the greatest risk sometimes have the greatest benefit. I think going back to the concern around nihilism, perhaps we do deny individuals certain treatments because we feel that their lifespan may be 5–10 years; therefore, there is uncertainty about what we are trying to modify, but arguably, there is actually a large potential benefit in that first 12 months or 2 years that we need to think about. Dual antiplatelet therapy is another example where sometimes we can be a bit nihilistic, but actually, there are fairly convincing data that it is beneficial in the right populations.

TQ: I think what is coming across from the responses from all of us is there is some equipoise about how we should use frailty: there is an uncertainty as to what we should actually do with it and I think that is where we need research. That is where we need data, because currently, we have our opinions, and opinion should not be what we base our clinical care on. It should be based on data from robust clinical research and I think it is quite exciting that there are trials happening now that are starting to look at these questions. When I think about secondary prevention in older adults I think about in both ways: I think it is about starting things and stopping things because a lot of physicians will – in a well-intentioned way – stop every medicine when someone reaches a particular age threshold, and actually, we do not know if that is the right thing to do or not.

JM: On that topic Terry, do you think that there has also some work to be done on the language we use to frame this for patient populations? I think patients benefit from understanding that they may be on a trajectory towards something, and how we utilise some of the frailty tools to describe where the patient is at that particular point in time is helpful. There is that well-noted diagram that was included in one of the review articles published by some of those on the call today that explains the impact that frailty has on treatment trajectories. I sometimes feel that if you are frail and you had a stroke and you do manage to leave hospital, it is important to frame that actually your reserve, should you have a recurrent stroke in the future, is extremely limited and therefore the risk of long-term dependency and/or death is certainly there. I do not know how well we have those conversations.

TQ: Language is so important and I think there is a public relations issue with frailty: from some of the work that we have done, older adults never perceive themselves as frail. They see the importance of frailty, they see that other people are frail, but they do not want to be frail themselves. So we just need to think do we use the “F word” or do we frame it in a different way? I am not sure. Again, I think there is research to be done there.

LB: I think it is really important to consider this in the context of the patient’s priorities. Although frailty in itself perhaps should not preclude a particular treatment, the patient’s priorities may change as their frailty progresses, and what they want out of life and their care may change as well. So I think it is really important to use frailty within the context of what is important for that patient. I think statins are often the ones that are viewed negatively in older people, and geriatricians have been seen to stop them, but I think it is important to consider the level of polypharmacy that patient has, are they struggling with taking all of those medications each day, what are their symptoms, and which medication is going to benefit symptoms rather than perhaps secondary prevention. I think there is an awful lot of other factors that go into consideration and we should be careful that we are not just using cutoff frailty scores or ages to make those decisions and that we are actually framing it in the context of the patient and what is important to them as well.

NE: I think that ability to tailor information, to be able to incorporate the individual’s views of what is important to them, is key. When we speak to stroke survivors across a range of different frailty levels, I agree there is often a perception by individuals that they are not frail, but they will often refer to themselves as “getting older.” However, very often, a lot of what they describe are actually characteristics of frailty. When it comes to hyperacute decision-making, the number of people that we have spoken to that report that they are getting older and perhaps less robust, but even if they can still get some benefit – be it attenuated – they would not want people withholding it from them simply because they are that bit older and frailer, even if there is an increased risk of complications. I think being able to have that data to hand – which I do not feel we have at the moment – is going to be key to being able to have those more informed discussions and being able to reach a shared decision and a better more informed decision at the time.

JM: I think we are at an exciting juncture where we have got pockets of ongoing research into different aspects of frailty. I think we have also got some focused work within our communities looking at neuroimaging and trying to determine markers of brain frailty. What I would love to see, and I think we need more work on, is marrying the brain phenotypes – which, in the era of advanced imaging, we have a much greater understanding of – with systemic frailty, and there are some fantastic studies from colleagues up in Scotland that have really tried to look at this in detail, both in trial and non-trial populations. I would also like to see a great deal more research looking at the subtypes. As Lucy mentioned earlier, there has been some work in milder stroke subtypes and also in ICH, which is fantastic to see. Perhaps we also need some more work in chronic stroke populations months to years after stroke to see the trajectories of mid-range/higher frailty scores – is it really all entirely pessimistic and poor outcomes?

NE: To me, mechanistic research underpinning frailty is an important area for us to develop. We are getting a better understanding of the associations between frailty and outcomes, but we do not necessarily understand how and why frailty influences outcomes. There are some very interesting data that looks at frailty being a systemic inflammatory state and the effect it can have on the brain as a consequence. Other roles as well – of frailty being related to sarcopenia in the rehab setting, about blood flow and cerebral autoregulation – all these multisystem effects by which frailty can influence outcomes in stroke affected by frailty. Having a better understanding of how frailty impacts stroke would be advantageous, but even more exciting is that understanding may allow us to intervene in some fashion, either before or around the time of stroke. Marrying the mechanisms to the observational data that we have seen would be really interesting.

LB: I would like to see more about how we use and embed frailty assessments into clinical services and looking at how that does some of those things we have been discussing about: how it influences treatment and decision-making, how it frames the way we have conversations with patients, how it affects who goes for rehab, and how well they rehab. Perhaps we should also think on a bigger scale about how we can organise services and streamline care for patients, and how we can factor frailty into our services to make sure that our services are frailty-friendly. Also, that we are considering the holistic aspects of patient care, and not just the presentation or single-organ pathology. I think frailty as a concept could be really useful in how we do that, but certainly, I think at the moment it is being applied quite heterogeneously in different clinical settings and services. I think looking at how we can better utilise and embed that to support both us and our patients in decision-making and care would be extremely useful.

TQ: I just want to echo Lucy’s point because I think she is exactly right: we have got a real implementation gap, and this is unusual for stroke because stroke medicine is traditionally a very evidence-based speciality. However, with frailty, stroke is lagging behind other disciplines and I think we need to learn from other specialities and we need to look at how we implement that in clinical care.

The other area that I am particularly interested in is really the flip side of frailty: in the same way that a person of a certain age can have a poor outcome than you might expect following an insult, you also have people who do really well and they probably have this concept called “resilience.” It is moving from a deficit-based approach to an asset-based approach is how I conceptualise it. Where my research team is moving at the moment is starting to look at resilience and what predicts this and what are the mechanisms that underpin resilience.

NE: Thank you. I think that is an excellent note on which to end. I think the field is exciting: it is expanding and is incredibly important. I thank all of you for the discussions about these areas and topics, not only for highlighting what we have achieved so far, but also where we are going in the field.

For those interested in further reading on the topics discussed in this interview, the authors have prepared a list of relevant reference articles below [1‒13].

Dr. Jatinder S. Minhas was a member of the journal’s Editorial Board at the time of submission. Dr. Nicholas Evans, Dr. Lucy C. Beishon, and Professor Terence J. Quinn have no conflicts of interest to declare.

N.R.E. is supported by a Stroke Association Senior Clinical Lectureship (SA-SCL-MED-22\100006) and the National Institute for Health and Care Research (NIHR) Cambridge Biomedical Research Centre (NIHR203312). J.S.M. is supported by a Stroke Association Senior Clinical Lectureship (SA SCLM23\100003) and a UKRI Future Leaders Fellowship (MR/Y016807/1). L.C.B. is a clinical lecturer funded by the NIHR. J.S.M. and L.C.B. are supported by the NIHR Leicester Biomedical Research Centre (BRC). The views expressed are those of the author(s) and not necessarily those of the Stroke Association, UKRI, NIHR, or the Department of Health and Social Care.

N.R.E., J.S.M., L.C.B., and T.J.Q. orchestrated the collection and prepared this manuscript jointly.

Dr. Nicholas R. Evans

Department of Clinical Neurosciences, University of Cambridge, Cambridge, UK

[email protected]

Dr. Jatinder S. Minhas

NIHR Leicester Biomedical Research Centre, Department of Cardiovascular Sciences, University of Leicester, Robert Kilpatrick Clinical Sciences Building, Leicester, UK

Dr. Lucy C. Beishon

NIHR Leicester Biomedical Research Centre, Department of Cardiovascular Sciences, University of Leicester, Robert Kilpatrick Clinical Sciences Building, Leicester, UK

Prof. Terence J. Quinn

Institute of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow, UK