Delirium—Beyond the CAM-ICU*
Richard R. Riker, MD
Gilles L. Fraser, PharmD
Department of Critical Care
Maine Medical Center
Portland, ME
T
wenty-five years ago, Hansen-Flaschen et al (1) deliv-
ered several hard truths to the critical care community
in an editorial, “Beyond the Ramsay Scale,” identifying
limitations in this scale and our broader approach to ICU seda-
tion, and proposing targets for improvement. Since then, tre-
mendous advances have produced validated sedation scales and
new medications, clarified the downside of benzodiazepines
and the importance of medication selection, and changed our
sedation target to a lighter level for most ICU patients. This
progress has helped reduce time on mechanical ventilation and
in the ICU and decrease interventions such as tracheostomy
and neurodiagnostic testing (2, 3).
In a parallel process, ICU delirium has been studied using
two recommended scales, the Confusion Assessment Method
for the ICU (CAM-ICU) and Intensive Care Delirium Screen-
ing Checklist (ICDSC), providing greater understanding of
the adverse outcomes associated with it (3). Despite this in-
sight, efforts to reduce delirium and its associated clinical
impact have often yielded negative results (4–8). Although
delirium has sometimes been reduced by select interventions,
the “improved” results still leave 48% (9) and 54% (10) of
ICU patients experiencing delirium. Multiple pharmacologic
approaches to reduce delirium have failed to show improve-
ments (4–7), and a reduced delirium burden was not associ-
ated with reduced adverse outcomes (6).
Why have our efforts to reduce delirium been so disap-
pointing? If we are to advance further, we must follow the
footsteps by Hansen-Flaschen et al (1), identifying flaws in
our current efforts, and investigating different approaches
which might yield improvements. In this issue of Critical Care
Medicine, Reznick et al (11) study a challenging problem—
accurately assessing delirium in patients with aphasia and
other neurologic deficits. Rather than ignoring these deficits
based on established screening recommendations (12), these
investigators broke new ground, incorporating additional
components to define inattention. In addition, they expanded
criteria defining unable to assess (UTA) beyond a Richmond
*See also p. 111.
Key Words: Confusion Assessment Method for the Intensive Care
Unit; delirium; Intensive Care Delirium Screening Checklist; intracerebral
hemorrhage; sensitivity
The authors have disclosed that they do not have any potential conflicts
of interest.
Copyright © 2019 by the Society of Critical Care Medicine and Wolters
Kluwer Health, Inc. All Rights Reserved.
DOI: 10.1097/CCM.0000000000004056
134 www.ccmjournal.org January 2020 • Volume 48 • Number 1
Copyright © 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Agitation-Sedation Scale (RASS) less than –3, including neu-
rologic barriers to following commands when stroke location
was consistent. Among 60 patients with intracerebral hemor-
rhage, daily assessment with the CAM-ICU, ICDSC, and ex-
pert evaluation based on the Diagnostic and Statistical Manual
of Mental Disorders, fifth edition (DSM-5) (considered the gold
standard) identified delirium in 57% of patients.
Comparing these screening tools was complex since both
were assessed two ways. The best results were obtained using
the ICDSC as an ordinal scale from 0 to 8, slightly less accurate
when using it as a binary tool (delirium yes or no). By com-
parison, the CAM-ICU showed reduced accuracy, worst using
the judicious approach which differed from its design. Similar
limitations of the CAM-ICU when assessing neurologic ICU
patients were previously reported by van Eijk et al (13).
Why was the ICDSC more accurate in the study by Reznick
et al (11)? Rather than defining inattention only with the vigi-
lance A test, it allows incorporation of other factors similar to
the nonverbal approach by Reznick et al (11). The two CAM-
ICU components not linked to RASS (inattention and disor-
ganized thinking) were present in 36% and 18% of assessments
and were frequently rated UTA (32% and 43% of assessments).
Inattention was much less commonly UTA when nonverbal
assessments were included (dropping from 32% to 11%). The
authors identify several limitations in their article, including
potential bias from a single investigator assessing delirium
using all three methods (CAM-ICU, ICDSC, and DSM-5
criteria).
An underemphasized result identified that 12% of patients
were UTA due to RASS scores less than –3, but half of them
could follow commands (67%) or demonstrate visual atten-
tion and tracking (33%). It is unclear what RASS score reli-
ably identifies an assessable patient (14, 15), and the data by
Reznick et al (11) suggest that assessing more than level of se-
dation may increase diagnostic accuracy. As Reznick et al (11)
bravely challenged existing dogma (11, 16), we must find other
“commandments” guiding our approach to ICU delirium that
might be targets for potential change and improvement. Here
are some to consider.
THE PUZZLE OF SEDATION
Moderate and deep levels of sedation are known to influence
delirium assessments (12), and several studies suggest that
even levels deemed “assessable” by the CAM-ICU (RASS –3)
may affect delirium assessment (14, 15, 17). During sedation
interruption trials, CAM-ICU positive results dropped from
97% to 31% with the improved arousal (12). It is clear that
many patients being diagnosed with delirium may simply be
too sedated to follow commands. The clinical importance of
separating sedation effects from delirium was identified in a
study screening delirium before and after sedation interrup-
tion (17). Patients receiving sedation while being assessed were

10.5 (5.3–21) times more likely to be CAM-ICU positive than
after sedation was interrupted. The patients who were CAM-
ICU positive only while receiving sedation had outcomes sim-
ilar to those with no delirium, and much better than those with
delirium that persisted after sedation interruption. A separate
study reported that 21% (95% CI, 12–33%) of patients screen-
ing positive with the CAM-ICU while receiving sedation were
negative when sedation was interrupted; clinical outcomes for
these patients were also similar to patients without delirium
(18).
Some may ask what is gained by more restrictive definitions
of delirium. If patients we diagnose with “delirium” are actu-
ally just “sedated,” best treatment may be reducing that seda-
tion. Adding haloperidol or ziprasidone should not improve
outcomes for patients whose ability to follow commands is im-
paired purely by sedation. To reduce such confounding, some
studies considered patients with RASS of –3 unassessable,
requiring a score of –2 or higher (4, 10). Others have proposed
an even higher RASS threshold to reduce sedation confounding
of delirium assessments (14, 15). An alternative to using RASS
levels may be to reduce sedation until patients can follow three
out of four simple commands (open eyes in response to voice,
use eyes to track clinicians, stick out tongue, squeeze hands) as
per the original sedation interruption study (2). Reznick et al
(11) found that some patients with lower RASS scores could
follow commands, and also performed a multicomponent
assessment of attention as outlined in their methods.
LUMPING ETIOLOGIES TOGETHER
Delirium is a not a diagnosis with uniform pathophysiology; it
is a syndrome associated with differing etiologies. It is oversim-
plistic to think that one treatment will benefit all patients who
screen positive for delirium, whether related to withdrawal
from alcohol or other substances, medication toxicity, hepatic
or renal encephalopathy, sepsis, hyponatremia, or other causes.
Searching for and treating predisposing factors for delirium is
always appropriate, but lumping them all together may not be.
This may explain why intervention trials enrolling heteroge-
nous causes for delirium have failed to show benefit.
COMBINING HYPERACTIVE AND
HYPOACTIVE DELIRIUM
Perhaps an even bigger problem affecting many delirium stud-
ies is lumping together hypoactive and hyperactive delirium,
and treating them with a common protocol. Such approaches
have consistently resulted in no improvements (4, 5, 7). In
comparison, studies treating only hyperactive delirium have
shown benefits with quetiapine and dexmedetomidine (19,
20). The recent Modifying the Impact of ICU-Associated Neu-
rological Dysfunction-USA trial compared placebo, haloper-
idol, and ziprasidone to treat ICU patients testing positive for
delirium, but 89% were hypoactive at enrollment, and at least
75% had 1 day or less of hyperactive delirium during the study
(5). Applying their conclusions that haloperidol or ziprasidone
had no effect on the duration of delirium must be limited to
Critical Care Medicine www.ccmjournal.org 135
Copyright © 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Editorials
hypoactive patients, and we should not consider this negative
finding when selecting treatment for agitated patients.
EVER-NEVER COHORTING
Almost all studies of ICU delirium assessed patients once or
twice a day, then grouped them into cohorts of “Ever” delirium
(if at least one assessment was positive) versus “Never” de-
lirium (if ALL assessments were negative). With this approach,
a patient with one positive and 11 negative delirium assess-
ments would join another patient with all 12 positive assess-
ments in the “Ever” group. Svenningsen et al (14) showed that
20% of patients had more than 10 positive CAM-ICU assess-
ments, 45% had fewer than four, and 18% had only one, yet
these patients were all grouped together. It is not certain that
this approach is worse compared to attempts to stratify based
on the burden or severity of delirium, but we are likely missing
important differences between patients by lumping them this
way, potentially contributing to negative study results.
APPROPRIATE OUTCOMES
The best treatment goals for ICU delirium trials have not been
defined. Should we treat all patients with delirium, even those
who remain calm, in an effort to “clear” their delirium, or should
we only treat patients if they are agitated or distressed, with
comfort and safety as the goal? If we believe that delirium is in-
jurious and causes the associated adverse outcomes, we should
try to eliminate it, but this belief remains unproven (6). Alter-
natively, if we believe delirium is a manifestation of underlying
pathophysiology that causes both delirium and the adverse out-
comes, it would be reasonable to treat delirium only if associ-
ated with undesirable or unsafe symptoms. Studies designed to
reduce delirium prevalence or duration as the primary outcome
might require treating all delirium, but a growing list of nega-
tive trials should compel us to reconsider this approach.
Reznick et al (11) have challenged the diagnostic approach
to ICU delirium that has been written in stone for almost
20 years. Although improvements and progress have been
made, the growing number of negative trials, the persisting
high prevalence of delirium, and the continuing uncertainty
whether the link between delirium and adverse outcomes
represents causality or only an association mandates we make
changes in our approach. The focus by Reznick et al (11) on a
select group of patients and the diagnostic challenges associ-
ated with them is an important step forward. We need more
research incorporating similar “outside the box” changes to
facilitate progress.
REFERENCES
1. Hansen-Flaschen J, Cowen J, Polomano RC: Beyond the Ramsay
scale: Need for a validated measure of sedating drug efficacy in the
intensive care unit. Crit Care Med 1994; 22:732–733
2. Kress JP, Pohlman AS, O’Connor MF, et al: Daily interruption of sed-
ative infusions in critically ill patients undergoing mechanical ventila-
tion. N Engl J Med 2000; 342:1471–1477
3. Barr J, Fraser GL, Puntillo K, et al; American College of Critical Care
Medicine: Clinical practice guidelines for the management of pain,
Editorials
agitation, and delirium in adult patients in the intensive care unit. Crit
Care Med 2013; 41:263–306
4. Page VJ, Ely EW, Gates S, et al: Effect of intravenous haloperidol on
the duration of delirium and coma in critically ill patients (Hope-ICU):
A randomised, double-blind, placebo-controlled trial. Lancet Respir
Med 2013; 1:515–523
5. Girard TD, Exline MC, Carson SS, et al; MIND-USA Investigators:
Haloperidol and ziprasidone for treatment of delirium in critical illness.
N Engl J Med 2018; 379:2506–2516
6. Al-Qadheeb NS, Balk EM, Fraser GL, et al: Randomized ICU trials
do not demonstrate an association between interventions that reduce
delirium duration and short-term mortality: A systematic review and
meta-analysis. Crit Care Med 2014; 42:1442–1454
7. Girard TD, Pandharipande PP, Carson SS, et al; MIND Trial
Investigators: Feasibility, efficacy, and safety of antipsychotics for in-
tensive care unit delirium: The MIND randomized, placebo-controlled
trial. Crit Care Med 2010; 38:428–437
8. Bigatello LM, Amirfarzan H, Haghighi AK, et al: Effects of routine mon-
itoring of delirium in a surgical/trauma intensive care unit. J Trauma
Acute Care Surg 2013; 74:876–883
9. Balas MC, Vasilevskis EE, Olsen KM, et al: Effectiveness and safety
of the awakening and breathing coordination, delirium monitoring/
management, and early exercise/mobility bundle. Crit Care Med
2014; 42:1024–1036
10. Riker RR, Shehabi Y, Bokesch PM, et al; SEDCOM (Safety and
Efficacy of Dexmedetomidine Compared With Midazolam) Study
Group: Dexmedetomidine vs midazolam for sedation of critically ill
patients: A randomized trial. JAMA 2009; 301:489–499
11. Reznick ME, Drake J, Margolis SA, et al: Deconstructing Poststroke
Delirium in a Prospective Cohort of Patients With Intracerebral
Hemorrhage Crit Care Med 2019; 47:111–118
136 www.ccmjournal.org January 2020 • Volume 48 • Number 1
Copyright © 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
12. Devlin JW, Skrobik Y, Gélinas C, et al: Clinical practice guidelines for
the prevention and management of pain, agitation/sedation, delirium,
immobility, and sleep disruption in adult patients in the ICU. Crit Care
Med 2018; 46:e825–e873
13. van Eijk MM, van den Boogaard M, van Marum RJ, et al: Routine use
of the confusion assessment method for the intensive care unit: A
multicenter study. Am J Respir Crit Care Med 2011; 184:340–344
14. Svenningsen H, Egerod I, Videbech P, et al: Fluctuations in sedation
levels may contribute to delirium in ICU patients. Acta Anaesthesiol
Scand 2013; 57:288–293
15. Haenggi M, Blum S, Brechbuehl R, et al: Effect of sedation level
on the prevalence of delirium when assessed with CAM-ICU and
ICDSC. Intensive Care Med 2013; 39:2171–2179
16. Ioannidis JP: Scientific inbreeding and same-team replication: Type D
personality as an example. J Psychosom Res 2012; 73:408–410
17. Patel SB, Poston JT, Pohlman A, et al: Rapidly reversible, sedation-
related delirium versus persistent delirium in the intensive care unit.
Am J Respir Crit Care Med 2014; 189:658–665
18. Kenes MT, Stollings JL, Wang L, et al: Persistence of delirium after ces-
sation of sedatives and analgesics and impact on clinical outcomes in
critically ill patients. Pharmacotherapy 2017; 37:1357–1365
19. Devlin JW, Roberts RJ, Fong JJ, et al: Efficacy and safety of quetiapine
in critically ill patients with delirium: A prospective, multicenter, ran-
domized, double-blind, placebo-controlled pilot study. Crit Care Med
2010; 38:419–427
20. Reade MC, Eastwood GM, Bellomo R, et al; DahLIA Investigators;
Australian and New Zealand Intensive Care Society Clinical Trials
Group: Effect of dexmedetomidine added to standard care on venti-
lator-free time in patients with agitated delirium: A randomized clinical
trial. JAMA 2016; 315:1460–1468