Question 10 from the second paper of 2021 was the first time frailty had appeared in the CICM Part II exam. The specific focus had been on the assessment of frailty, the limitation of assessment methods, and the use of frailty in the management of critically ill patients. These are obviously not the only possible questions on this topic.
Definition of frailty
Falvey & Ferrante (2019) give a solid one-liner to define frailty:
"a state of decreased physiologic reserve that heightens vulnerability to acute stressors"
Even though they were supposed to be quoting McDermid et al (2011), who actually defined it as
"a multidimensional syndrome characterized by the loss of physical and cognitive reserve that predisposes to the accumulation of deficits and increased vulnerability to adverse events"
If that does not seem wordy enough, De Biasio et al (2020) took it a few steps further:
"The frailty syndrome is ... a network of interrelated perturbations involving the age-related accumulation of molecular, cellular, and tissue damage that leads to multisystem dysregulation, functional decline, and disproportionately poor response to physiologic stress"
So is it a physiological state, or is it a multidimensional syndrome, or is it a ... network of interrelated perturbations? After becoming frustrated with the literature on this subject after only ten minutes of Googling, one would have to conclude that none of the definitions are entirely satisfactory. In case one would like to instead contemplate what a satisfactory definition should look like, Rockwood (2005) produced a rather comprehensive list of criteria by which to judge a definition of frailty. Since then there have been a multitude published definitions and so much disagreement that, when a consensus conference was convened to consolidate them (Rodríguez-Mañas et al, 2013), the group of thirty-one experts somehow failed to agree on even the fundamental basics. In short, the exam-ready reader is advised to pick a short pithy statement and memorise it for their exams, secure in the knowledge that all of them have equally little validity.
Assessment of frailty
The local Government Organ for Clinical Awesomeness has a great frailty page that brings together a few frailty screening tools, and from which the following extensively borrows. In case they lose their funding after the next state election and their website goes down, this paper by Panhwar et al (2019) would probably be a more resilient link. In summary, the assessment is multidimensional, and can consist of:
- Global subjective assessment by the intensivist, according to clinical findings, which are typically:
- Clinical features of malnutrition
- Evidence of decreased energy expenditure
- Features of multiple health deficits or comorbidities
- Self-reported historical features, eg. fatigue, mood, social participation
- Proxy functional assessment, eg. use of mobility equipment, reliance on services, hospital admissions, frequency of falls
- Functional disability assessment, eg. hand grip, gait, balance, cognitive assessment
- Biometric data, eg. weight loss, muscle mass
- Objective biological findings eg. , echo findings, chronic organ system dysfunction
The use of a validated frailty scoring system is recommended, as these tend to bring together all of these elements, and distil the complex and ineffable network of interrelated perturbations into a single numerical score. It seems as if around the years 2000-2001 people somehow suddenly realised that frailty assessment is an untapped gold mine of potential publications, and frailty assessment scales sprouted like mushrooms after the rain. Examples of such scales, listed in chronological order, include:
- Fried phenotype (Fried et al, 2001) which consists of five domains (weakness and exhaustion, grip strength, daily physical activity, weight loss and gait velocity). It's especially well suited to bioengineering research, as everything is fairly quantitative. Unfortunately, its applicability in the ICU is limited, as we are usually unable to walk the patients, weigh them accurately, or even get them to squeeze a grip dynamometer.
- Groningen frailty indicator (Steverink, 2001) which is a self-reported fifteen-item questionnaire made up of yes-or-no questions. Cognitive, physical, psychological and social function is tested; a score of 4 or more identifies you as "frail", and a maximum score corresponds to a fully dependent nursing home level of existence.
- The Frailty Index Comprehensive Geriatric Assessment (Jones et al, 2004), which uses a ten-item system and which is distinguished from the other scales by only being able to detect how frail you are, rather than being able to discriminate frail from non-frail.
- Rockwood scale (Rockwood et al, 2005), which is a seven-point scale (or nine-point, if you include some optional extras) developed by the same guy who (in the same year) published a thing about the necessary elements of a successful frailty definition. This also has the advantage of having been the most commonly used scale for ICU-based frailty research, according to a literature review by Pugh et al (2018). This scale is also occasionally referred to as the CFS (Clinical Frailty Scale), and takes the form of several functionally anchored descriptors which help the assessing clinician categorise the patient's level of functional reserve. It has the advantage (and the disadvantage!) of being almost entirely subjective, which means most of the data can be derived from relatives and other proxy sources.
- Reported Edmonton Frail Scale (Rolfson et al, 2006) is made up of nine components (cognition, functional performance, general health status, functional independence, social support, medication use, nutrition, mood, and continence). It is scored using a questionnaire, and is therefore also quite subjective, but seems to have high validity (insofar as the scores correlate with the learned opinion of an assessing geriatrician).
- Tilburg frailty indicator (Gobbens et al, 2014), which focuses on three main dimensions: physical, psychological and social. Unfortunately, they don't seem to relate to each other, which means one can score a high Tilburg score and still be relatively robust (i.e. you'd gain frail points by living alone and being depressed or incontinent, but you could totally kill the physical assessment by having an iron grip and steady gait).
There are about a thousand more (eg. see this review by De Vries et al, 2011), but the reader will have already got the point around Rockwood. The question is, are any of these any better than others? That's harder to assess than one might think, considering that we haven't agreed on any objectively measurable definition of frailty. The most that can be said of them is that they all have fairly similar construct validity, i.e. whatever the thing is that they are measuring, it seems to be associated with some genuine patient-centred outcome changes, eg. increased mortality.
Utility of frailty assessment for ICU management
What is the relevance of this term to the management of the ICU patient? Let's say you applied some sort of scale, and assessed the patient as being frail on the basis of your findings. What now? "We contend that the concept and measurement of frailty may have clinical, psychosocial and economic relevance to critical care medicine", opined McDermid et al in their 2011 editorial. However, their paper mainly focused on prognostic implications in terms of the use of ICU resources. Apart from "limitations of therapy", few novel suggestions are made by other authors.
Still, this can be cobbled together into an answer to the question, "what use is frailty to the intensivist".
- Frail patients have poorer ICU outcomes.
- The expectations of society on intensive care services is changing; more elderly and frail patients are expected to access intensive care services in the future.
- The cost of critical care is increasing, and the cost of caring for functionally dependent or chronically ill patients is already high; to say nothing of the unmeasurable cost of decreased quality of life for disabled survivors
- Thus, screening for frailty is probably most useful in the setting of appropriate goals of care, taking some of the subjectivity out of the otherwise emotionally charged process of decisionmaking.
- Frailty, in certain populations, may be more predictive of poor outcomes than physiology-based acute illness scores. For example, Le Maquet et al (2014) found that SAPS II and SOFA scores were fairly similar between frail and non-frail patients.
- To make the diagnosis of frailty early in the course of critical illness can result in increased clinician attention to management strategies designed to maximise the preservation of function and the prevention of further enfeeblement, such as:
- Nutrition (high caloric and high protein supplementation)
- Micronutrient and vitamin replacement
- Early extubation
- Early mobility and physiotherapy
- Delirium screening
- Referral to services such as social work, psychology, occupational therapy and specialist geriatrician services
Prognostic relevance of frailty scores
So, if you;'re going to make decisions about limitations of therapy and goals of care on the basis of frailty assessments, you should probably have some contemporary data to back your decisions. Darvall et al (2019), in a retrospective Australian study that used the ANZICS database, found that the in-hospital mortality for frail patients admitted to the ICU was higher (17.6% v 8.2%) when compared to non-frail patients of a similar age. Basically the same authors repeated a prospective version of the same study and got basically the same mortality rates (16% vs 5%), with in-hospital mortality of up to 39% for the "severely frail" category, and only 2% for patients regarded as "fit" on the basis of the CFS.
Limitations of frailty assessment
Limitation of making an assessment of frailty in the ICU are numerous:
- Clinical features of frailty may be exaggerated by critical illness. Disease processes produce disability even in healthy people, and weight loss can develop over a relatively short period in originally well-nourished individuals. The assessment of frailty at the time of ICU admission can be clouded by this.
- Frailty can develop during the ICU stay. To borrow a turn of phrase from McDermid et al (2011), "deficits associated with frailty, which typically take years to accumulate in the outpatient geriatric population, rapidly develop in a large proportion of critically ill patients independent of age and illness severity". After enough time in the ICU, there are substantial similarities between frail patients and critically ill patients, such as the inability to walk and poor upper body strength.
- There may be insufficient information. Unrecognised comorbidities or unknown functional capacity would obviously bedevil the evaluation of frailty. When an Unknown Male (Age Unknown) is found unconscious at the foot of some stairs, one would not be able to make an accurate assessment.
The standardised instruments designed to measure frailty also have several disadvantages in the ICU
- They are painful to use. The frailty index by Minitsky et al (2001) has seventy items. Nobody has time for this in the ICU. It does not matter how robust the measure, if nobody has the resources to apply it properly.
- They may require special training. For an extreme example, not everybody has operated a hand grip ergometer. In a more general sense, the ratings produced by uncalibrated users of a subjective scale may not have much validity.
- They were developed for community populations. These are outpatients who can be walked under the watchful gaze of the physiotherapist, interrogated about their social life by the geriatrician, instructed to handle squeezy ergometers, and so on. In short, the application of these frailty scoring instruments in the ICU is fraught with inaccuracy, as one ends up borrowing second-hand information from carers and family. This also limits the selection of scales - everything with an objective measurement such as grip strength is unsuitable, so one resorts to using scales like the CFS which are entirely subjective and dependent on accurately answered questions.
- They are unreliable even with intended use cases, let alone in the ICU. When De Vries et al (2011) tried to compare a few of them, they came to the conclusion that they probably all have some construct validity (eg. they predict poor health outcomes), but that their reliability is generally poor, or basically just not reported upon. For example, of the twenty-one instruments they studied, only two had some good published data on inter-rater agreement. In short, when you the intensivist determine that a patient is frail, there's no guarantee that the geriatrician or cardiologist will concur, even if you both used the same rating scale.
- They might lead to unjust triage. Subjective or proxy assessments of increased risk of poor outcome might lead to decisions about limiting therapeutic options, which would turn into self-fulfilling prophecies. This would be most apparent in scenarios of scarcity or systemic stress, such as some kind of pandemic for example. Care might be misallocated or withheld unfairly.
- They might not be especially predictive. Not all studies find a mortality disadvantage associated with frailty. In particular, Charles et al (2011) and Fisher et al (2015) did not find much of a mortality difference between frail and non-frail ICU patients.