Journal Club

Frailty

Clinical Question

Over a stratified range of ages, how does frailty affect the chance of death in ICU?

Frailty Definition

“multidimensional syndrome indicating an increased vulnerability to stressors…”

“…a reflection of a biological rather than chronological age…”

“…dimension of health that is not meaurable using current risk scoring systems…”

Clinical Frailty Score(CFS): 5+

CFS Limitations

[Rockwood 2020 - Using the Clinical Frailty Scale in Allocating Scarce Health Care Resources]

” The key points are that the Clinical Frailty Scale…
is not widely validated in younger people
or those with stable single-system disabilities,
and it requires clinical judgement”

Frailty In Younger Populations

[Spiers 2020 - Measuring frailty in younger populations: a rapid review of evidence]

“There was no evidence about the impact of measuring frailty in younger populations on patient outcomes and care.”

Paper: Methods

• Prospective Observational Trial (Cross Sectional)
• Single Centre Tertiary ICU:
  • Cardiothoracics ❌
• 18 Months

Methods - Measurement

• Clinical Frailty Scale
  • Frail (5+) vs Not Frail (4-)

Outcomes

• Primary - 180 day all cause mortality
• Secondary - ICU Complications
• Secondary - ICU Length Of Stay

Stats

• Mann-Whitney U Test
• Chi-Square Test
• Anova test - comparison of multiple groups
• Multivariable cox regression analyses

Cox Regression Analyses

• Multivariable cox regression analyses:
  • adjusted for the following as independent factors:
    • comorbidities
    • saps3
    • decision to withhold or withdraw therapy

Results

Demographics

Frailty Prevalence: 41%

Outcomes - Dead At 180 Days

Outcomes - ICU Complications…?

Outcomes - ICU Length Of Stay - Median Hours

Reported Findings - Length of Stay

• Longer length of stay
  • (But 9-22 hours longer)
  • (No split of elective vs not)

Reported Findings - Mortality

• Frailty = Increased Hazard of 180 day mortality
• In all age groups
• (Not just older cohorts)

Reported Findings - Multivariable

“Multivariable analysis with illness severity, comorbidities, decisions to withhold or withdraw therapy, and frailty demonstrated that frailty was independently associated with 180-day mortality among patients aged 50–64 but not in the other age groups”

Summary

1. Frailty is common in ICU: ~27-40%
2. Frail patients are more likely to die in all age groups
3. Length of stay is longer for frail patients:
   • But maybe not meaningfully
4. I’m not sure about their adjusting for confounders…

SAPS. Severity and Comorbidities

Confounds

What did they do in the paper?

• Stratification - Splitting into seperate age groups
• Multivariable Adjustment:
  • Illness Severity
  • Comorbidities
  • Decisions to Withhold or Withdraw
  • Clinical Frailty Score

Strategies

1. Fork
2. Pipe
3. Collider
4. Descendant

Fork

Comorbidities is common cause of Frailty and Mortality
You wanted to know if there was a relationship between Frailty and Mortality
You need to stratify for your common cause (put comorbidities in model)

Pipe

Total effect of Frailty on Mortality goes through Illness Severity and Decisions
If you adjust for them, you block the indirect effect

Collider

None in this model
X and Y have no shared cause
But they do both influence Z
If you stratify by Z, you make it look like X and Y are associated.

Descendant - ?Another Pipe

Here there's a pipe of X Z Y again,
but also Z influences A
if you stratify for A, you're basically stratifying for Z

• Comorbidities (Fork)
• Age (Fork)
• Illness Severity (Pipe)
• Decision to withhold (Pipe)