COVID-19: Use of the Clinical Frailty Scale for Critical Care Decisions.

To the Editor: Amid the coronavirus disease 2019 (COVID‐19) pandemic crisis, doctors are now faced with tough ethical decisions in determining who would benefit most from intensive care and ventilator support. COVID‐19 has the potential to cause serious life‐threatening complications in both the young and old. However, it is especially detrimental among older adults with multimorbidities.2, 3

Given the scarcity of resources faced by many areas, critical care triage becomes ethically complex and challenging. This dilemma has led to countries across the globe rapidly preparing triage guidelines to aid doctors in making the difficult decision on the appropriateness and potential benefits of admitting patients, especially older adults, to critical care. However, Le Couteur and his coauthors highlighted concerns regarding allocating critical care interventions based solely on age cutoffs and that a frailty‐based approach, prioritizing on an individual’s functional abilities, multimorbidities, prognosis, and treatment preferences, may prove more valuable.

The National Institute for Health and Care Excellence (NICE) recently published COVID‐19 rapid guidelines for critical care in adults. It recommends the use of the Clinical Frailty Scale (CFS) in patients aged 65 years and older, and states that decisions about admission to intensive care units (ICUs) should be made on the basis of the potential for medical benefit. We refer to NICE guidelines and wish to share our concerns about the use of a CFS score of 5 as the pivot point to define a “more frail” state where uncertainty exists regarding the likely benefit of ICU care.

We recently published data on acutely hospitalized older adults (mean age = 89.4 ± 4.6 years; n = 210) where we compared patients with a CFS score of 1 to 5 (non‐frail to mildly frail) against those with a score of 6 to 8 (moderate to severely frail). Our data revealed that being in the latter group was independently predictive of mortality and institutionalization following hospitalization. Using the same data set, we separated our cohort into five groups (CFS scores 1‐3, 4, 5, 6, and 7‐8) and compared them against the risk of mortality and institutionalization during hospitalization and 6 and 12 months after enrollment. We performed the Fisher exact test for a comparison between the groups and incident mortality and institutionalization. We then performed logistic regression analysis, adjusted for age, sex, and severity of illness, to investigate the independent association between increasing frailty (using a CFS score of 5 as reference point) and the adverse outcomes of interest.

We found mortality rate during hospitalization to be higher in CFS 6 (4.7%) and CFS 7 to 8 (8.6%) patients with no mortality recorded among CFS 1 to 3, CFS 4, and CFS 5 patients (Table 1). At 12‐month follow‐up, the mortality rate in CFS 5 patients was low (8.3%) in comparison with CFS 6 (29.2%) and CFS 7 to 8 (60.0%) (P < .001). This was similarly observed for institutionalization. Logistic regression analysis revealed that increasing frailty, when compared with CFS 5, was independently predictive of mortality (CFS 6: odds ratio (OR) = 4.52; 95% confidence interval [CI] = 1.61‐12.73; P = .004; CFS 7‐8: OR = 17.99; 95% CI = 5.36‐60.34; P < .001) and institutionalization and/or mortality (CFS 6: OR = 5.11; 95% CI = 1.97‐13.27; P = .001; CFS 7‐8: OR = 15.87; 95% CI = 5.04‐49.95; P < .001) at 12 months after hospitalization.

Table 1

Comparison of Short‐ and Long‐Term Adverse Health Outcomes between Clinical Frailty Scale Groups

CFS groups	Mortality			Institutionalization and/or mortality
	Initial hospitalization (n = 210)	6 mo (n = 210)	12 mo (n = 210)	Initial hospitalization (n = 210)	6 mo (n = 206) b	12 mo (n = 206) b
Univariate analysis (%)
CFS 1‐3	0/4 (.0)	0/4 (.0)	0/4 (.0)	0/4 (.0)	0/4 (.0)	0/4 (.0)
CFS 4	0/5 (.0)	0/5 (.0)	0/5 (.0)	0/5 (.0)	0/5 (.0)	1/5 (20.0)
CFS 5	0/60 (.0)	3/60 (5.0)	5/60 (8.3)	0/60 (.0)	4/57 (7.0)	6/57 (10.5)
CFS 6	5/106 (4.7)	23/106 (21.7)	31/106 (29.2)	10/106 (9.4)	29/105 (27.6)	39/105 (37.1)
CFS 7‐8	3/35 (8.6)	16/35 (45.7)	21/35 (60.0)	3/35 (8.6)	18/35 (51.4)	22/35 (62.9)
P value a	.27	<.001	<.001	.15	<.001	<.001
CFS groups	Mortality			Institutionalization and/or mortality
	Initial hospitalization c	6 mo (n = 201) d	12 mo (n = 201) d	Initial hospitalization c	6 mo (n = 197) e	12 mo (n = 197) e
		Adjusted OR (95% CI)	Adjusted OR (95% CI)		Adjusted OR (95% CI)	Adjusted OR (95% CI)
Multivariate analysis f
CFS 5		Ref. g	Ref. g		Ref. g	Ref. g
CFS 6		5.11 (1.44‐18.06) h	4.52 (1.61‐12.73) h		5.00 (1.64‐15.23) h	5.11 (1.97‐13.27) h
CFS 7‐8		15.69 (3.98‐61.87) i	17.99 (5.36‐60.34) i		14.11 (4.07‐48.94) i	15.87 (5.04‐49.95) i

Abbreviations: CFS, Clinical Frailty Scale; CI, confidence interval; OR, odds ratio.

Fisher exact test performed.

Four participants excluded from 6‐ and 12‐month analysis because they were lost to follow‐up.

No mortality or institutionalization among CFS 5 subjects. Hence unable to use CFS 5 as reference point.

Total n = 201 (excluded nine who had CFS <5 as all did not have incident mortality).

eTotal :n = 197 (excluded nine who had CFS <5 and 4 who were lost to follow‐up).

Logistic regression performed adjusted for age, sex, and severity of illness.

CFS 5 used as reference for multivariate analysis.

P < .05.

P < .001.

Even as we make our case for a thoughtful consideration of the CFS, we wish to acknowledge that the lower 12‐month mortality and institutionalization rate in our oldest‐old population with CFS 5 scores comes with the caveats of good geriatric care supported by a multidisciplinary team and a healthcare system that was not overwhelmed. We also acknowledge that clinical outcomes may well be dissimilar in older adults hospitalized with COVID‐19. Nonetheless, our data in an indirect way demonstrate what is possible with good geriatrics care, and even during our battle with COVID‐19, unplanned admissions for illnesses unrelated to COVID‐19 still need to be managed adequately.

We note that previous studies reported poorer ICU outcomes in frail older adults that include hospital and long‐term mortality, and reduced likelihood to be discharged home.8, 9 However, what was noteworthy in the study by Darvall and his coauthors, as highlighted in an editorial by Mudge, was that the outcomes for critically ill frail older patients were reasonably good with 91% surviving ICU admission, 82% surviving to hospital discharge, and fewer than 5% discharged to new nursing home care. In many ways, our data support findings from Darvall et al and corroborate the concept of frailty: mildly frail older adults may still have enough intrinsic capacity to withstand the stressors of hospitalization and make a good recovery.

Thus although there is a pressing need for clear critical care guidance during this pandemic, frailty assessment should not lead to automatic disqualification of less frail individuals (eg, with a CFS 5 score) from the potential benefits of ICU care, especially if critical care capacity remains sufficient. In conclusion, we wish to emphasize that frailty identification should not simply result in a “label” but rather impact management in a meaningful context‐appropriate way that is used to make care rational and not to ration care.