Pragmatic &
aspirational
approaches to postoperative
delirium & neurocognitive
disorders
Liz Whitlock, MD, MSc
Assistant Professor
University of California, San Francisco
Department of Anesthesia & Perioperative Care
MSQC/ASPIRE Meeting
8 April 2022
No financial relationships to disclose
Current & past funding:
National Institute on Aging (NIH)
National Center for Advancing Translational Sciences (NIH)
Foundation for Anesthesia Education and Research
Learning objectives
• Understand recommendations from
international medical groups about risk
stratification and prevention of delirium in
older adults
• Describe components of a pragmatic
delirium risk stratification and mitigation
quality improvement pathway
• Assess the need for more comprehensive
programs and/or the incorporation of long-
term cognitive outcomes into local
perioperative pathways for older adults
Which is the real postoperative course?
78yo ASA3 woman s/p femur ORIF
• 143/85, HR 87, SpO2 95% on RA
• Ambulated with PT yesterday
• Ready for discharge to rehab
Mom broke her femur
• I’ve spent 3 days in the hospital
trying to help her understand
what happened
• Is there anything we could have
done to keep her out of a
nursing home?
• 6 months later: still cognitively
unable to live independently
“[my mother]
wasn’t there at all,
not at all. It’s a
shell.”
She was doing OK
before, so it must
have been the
surgery/anesthesia.
Day J & Higgins I Qual Health Res 2015 Dec;25(12):1700-18
N Engl J Med 2019; 380:408-409
Which is the real postoperative course?
78yo ASA3 woman s/p femur ORIF
• 143/85, HR 87, SpO2 95% on RA
• Ambulated with PT yesterday
• Ready for discharge to rehab
Mom broke her femur
• I’ve spent 3 days in the hospital
trying to help her understand
what happened
• Is there anything we could have
done to keep her out of a
nursing home?
• 6 months later: still cognitively
unable to live independently
The nurse told me she’s
delirious, and the
anesthesiologist warned me
this could happen. It’s scary
but I hope it will improve as
she gets better.
The geriatrician who cleaned up
Mom’s med list said it sounds like she
had moderate dementia before all
this happened. We talked about her
end-of-life wishes, and we know
more about what to expect.
CMAJ 2017 189 (14) E542
Centering the patient in
perioperative care
For older adults, this means:
Function
=
Independence
Surgical
outcomes
Blocking
other bad
stuff
Talking
about stuff
you can’t
prevent
Considering
alternatives
Oh, you mean like…
Delirium and neurocognitive
disorders?
7
Perioperative Neurocognitive Disorders
•Neurocognitive disorder (mild vs. major)
Prior to surgery
•Emergence excitation OR delirium
Emergence
•Delirium OR
•Delayed neurocognitive recovery
Immediately postop
to 30 days
•Postoperative neurocognitive disorder (mild vs. major)
Expected recovery (30 days)
to 12 months
•Neurocognitive disorder (mild vs. major)
•Postoperative neurocognitive disorder (if dx made prior to 12 months)
Beyond 12 months
Evered L, et al. Anesth Analg. 2018. 127(5):1189.
J Clin Anesth. 2022 Feb;76:110574.
Unrecognized cognitive
impairment:
Elective noncardiac: 37%
Urgent/emergent: 50%
Timing of diagnosis:
How common is delirium?
Inouye et al, Lancet 2013; Zenilman, JAMA 2017
Medical patients:
Prevalence: 18-35%
Incidence: 11-14%
Surgical patients:
Incidence: 5-52%
ICU patients:
Prevalence + Incidence: 80-85%
Ann Surg. 2018 July. 268(1):93.
Great data, no surprises:
• Patients with higher risk of delirium were…
• Older (mean 80.7 years)
• Higher ASA Class (Class IV-V 24.4% vs 8%)
• More likely to undergo emergency surgery
• More likely to have preop cognitive impairment
• 4 strongest predictors of postop delirium:
• Preoperative Cognitive Impairment (OR 2.9)
• Surgery-specific Risk (OR 2.4)
• ASA Class (ASA3: OR 1.5, ASA4-5: OR 2.1)
• Age (70-74 OR 1.4, 75-79 OR 1.9, 80+ OR 2.7)
Aren’t your eyeballs already
predicting something here?
What are the consequences of delirium?
Patient Harm
• Mortality
• Discharge to institution
• Accelerated cognitive decline
• Functional decline
• Falls
Increased Cost
• Length of stay
• Readmissions
• Staff time
Gleason et al, JAMA Surg 2015
Berian et al, Ann Surg 2018
1
2
What are the consequences of delirium?
Gleason et al, JAMA Surg 2015
Berian et al, Ann Surg 2018
Probably
same patients
Mortality
Institutional
discharge
Unsure
Readmissions
Cognitive decline
Functional decline
Probably caused by
delirium
Falls
Length of stay
Cognitive decline
Inouye et al, Alz Dement 12:766-775, 2016
Accelerated cognitive decline after delirium
• Delirious patients…
• Started off lower
• Declined faster
• Causality not yet
established… but it is still
important
“[My father would] just look at me
and there was nothing in his eyes,
nothing, no recognition, just dead
patches in his eyes.”
Delirium often has more than one victim
• Disconnect between clinical and
observed scenario
• Isolation
• Psychological consequences (PTSD)
“[my mother]
wasn’t there at all,
not at all. It’s a
shell.”
Day J & Higgins I Qual Health Res 2015 Dec;25(12):1700-18
Svenningsen H et al J Clin Nurs 2016 Oct;25(19-20):2807-15
“The color purple was especially frightening. It had a
hold on me and surrounded me. I tried to run, but it
was there all the time, it followed me and swallowed
me. The ceiling was right over my nose and the walls
cracked like a spider web. I lost the sense of who I
was, where I was – and the worst thing was that they all
just stood there and laughed.”
15
What we should be doing
Everybody’s got an opinion
Curtis et al, Curr Opin Anesthesiol 2020 Oct 35(5): 668-73
Whoa!
So, obviously, we should all be doing cognitive
screens.
• No consensus on how
https://www.dal.ca/sites/gmr/our-tools/clinical-frailty-scale.html
Whitlock EL. Anesth Analg 2021. 133(5):1090-1093, from McIsaac et al, Anesth Analg 2020; 131:263-272.
So, are you doing cognitive screening?
Or the other stuff?
Absolutely not.
Expensive
Whose problem is it?
Which test?
How do we talk about it?
Clearly we are!
It’s the right thing to do
We use it to improve care
We picked a test
We’re muddling through
UCSF is
here-ish.
19
What we can be doing: talking!!
Here’s your next patient, in preop, with her caregiver son
• You can tell she’s at high risk (elderly, ASA 3, you
suspect cognitive impairment, and she’s getting
a fem-fem bypass)
• Instead of ignoring it, you…
• Tell her and her son she’s at high risk
• Explain that delirium is scary but temporary
• Counsel them about what to look for
• Explain what you’ll do to try to minimize her risk
• Describe what can be done after admission to
prevent or treat
UCSF started doing this proactively (on
medicine floors)
http://delirium.ucsf.edu
Now we’re doing this
perioperatively, for all patients,
through a massive QI effort.
Postop delirium at UCSF
What we had
Universal hospital delirium
screening (NuDESC)
A poorly-performing risk
stratification instrument
(AWOL)
An electronic medical record
& programming expertise
What we did
Use routinely-collected data
to develop a delirium risk
instrument for periop
Program it (largely
automated) into the EMR
Develop a pathway for high-
risk surgical patients
How things
changed
Decreases in Beers List
meds in older adults
(3% relative decrease per
month)
Decreases in
anticholinergics given to
high-risk patients
(20% relative decrease)
Donovan et al, Anesth Analg. 2020 Dec;131(6):1911-1922.
Whitlock et al, Anesth Analg. 2020 Dec;131(6):1901-1910.
AWOL-S and how it’s used
Bishara et al, BMC Anesthesiol. 2022 Jan 3;22(1):8
Whitlock et al, Anesth Analg. 2020 Dec;131(6):1901-1910.
Donovan et al, Anesth Analg. 2020 Dec;131(6):1911-1922.
A machine learning
model performs better
but isn’t clinically
implementable (yet)
But remember how we’re only here?
UCSF is
here-ish.
• Our risk stratification tool isn’t amazingly predictive
• Misses 25% of patients who will be delirious
• Overcalls delirium risk in 40% of those “at high risk”
• It only works if you do it
• And what are you going to do, anyway?
Avoid preoperative midazolam?
Avoid postop meperidine?
Intraoperative EEG monitoring?
Multicomponent interventions?
Talk people out of surgery?
Seriously: how about
we just talk about it,
until we can do more?
(the secret subtheme of my talk)
Perfect is the enemy of
good enough.
Pragmatic &
aspirational
approaches to postoperative
delirium & neurocognitive
disorders
Long
-term cognitive
decline
How
likely?
How
bad?
Patient-
centered
risk
discussion
Death or disablement
from cardiac disease
How
likely?
How
bad?
Sternal wound
infection
How
likely?
How
bad?
The current state: “POCD” in 10-15%
Put it
together
Anecdote
Popular
news
articles
Medical
literature
“I have a 15% chance of permanent,
functionally impactful cognitive decline
after surgery: the Post said so!”
A pragmatic perspective on POCD (PND)
• 15% risk of WHAT, exactly??
• Is major surgery/anesthesia systematically harmful to older adults’ long-term
cognition?
• What characteristics are associated with clinically meaningful late cognitive
decline?
Options for coronary revascularization
Severe
CAD
CABG
On-pump Off-pump
PCI
Whitlock et al. JAMA. 2021 May 18;325(19):1955-1964.
AVOIDS:
Sternotomy (trauma)
Chronic pain
General anesthetic
CPB exposure
Mechanical ventilation
ICU stay
Postoperative delirium
Pre-procedure biennial
assessments
Post-procedure biennial
assessments
Time (years): -4 -2 0 2 4
Participant
undergoes
CABG or PCI
Memory score;
dementia probability
Memory score;
dementia probability
Memory
Time (years)
PCI?
CABG?
CABG
PCI
PCI
CABG
PCI
9.6%
CABG
10.5%
CABG
PCI
PCI
Trad
CABG
OPCAB
CABG
PCI
Criticism: Averages obscure meaningful differences
Cognition
-5 0 5 10
Time (years)
Cognition
-5 0 5 10
Cognition
-5 0 5 10
Moving from averages to individuals
Cognition
-5 0 5 10
Time (years)
Cognition
-5 0 5 10
Cognition
-5 0 5 10
(And we knew
about this: it’s the
difference between
“anecdote-POCD”
and “research-
POCD”)
5% PCI
6 % CABG
Major
PND
25y of
cognitive
aging
How
likely?
How
bad?
Prediction model for PND:
Older ages (~20% ⬆ per year)
Frailty (doubles risk)
Overweight/obese (~30-40% ⬇)
Whitlock et al. Unpublished data not for distribution
• Using this (unpublished!) model:
• 25% risk of major PND if you screen high-risk
• 4% risk of major PND if you screen low-risk
• But… it only flags half the major PND people as high-risk
• Limitations: Validated? Clinical use? PND definition?
• What do we need to do to do this better??
With good data, there is so much
more we could do
Outcomes risk prediction was revolutionized by
data collection practices. So will NCDs be!
77yo
CABG
Diabetic
Declining slower
Can’t exercise
0 2w 6w 3m 6m 1y 2y 5y
85yo
Ovarian cancer resection
Mild cognitive impairment
Declining faster
Return to previous function?
Independent
IADL difficulty
ADL difficulty
Current
“Postoperative delirium is
important at my institution.”
Feasible
“Postoperative delirium is
important to anesthesiologists.”
Focus on
delirium
Aspirational
“Postoperative neurocognitive
disorders are important to all who
provide perioperative care for
older adults.”
Focus on
cognitive
recovery
DATA
MPOG, NACOR
DATA None / local
Medicare, NSQIP, STS
Specialty
initiatives
Local
initiatives
Multidisciplinary
initiatives
Generaliz-
able best-
practices
study data
Limited best-
practices for OR
and critical care
Local quality
improvement
Universal buy-in
Specialty-level buy-in
What I can leave you with:
• Everything is hard right now.
• Doing something is better than nothing – particularly
when best evidence is for patient-centered care
• What can you, your department, your institution, or
your specialty do to move things forward?
(Honestly, and from a caring place: I think this is a way to help
with the burnout and moral injury we feel seeing older adults &
families blindsided by NCDs!
Who's at
risk
Name
it
Talk
about
it
(Prevent
it?)
“Avoid hypotension and
hypoxia.”
Love,Cardiology
