Research

Preliminary Communication | CARING FOR THE CRITICALLY ILL PATIENT

Effect of Dexmedetomidine Added to Standard Care

on Ventilator-Free Time in Patients With Agitated Delirium
A Randomized Clinical Trial
Michael C. Reade, DPhil, FCICM; Glenn M. Eastwood, RN, PhD; Rinaldo Bellomo, MD, FCICM; Michael Bailey, PhD; Andrew Bersten, MD, FCICM;
Benjamin Cheung, MBBS, FCICM; Andrew Davies, MBBS, FCICM; Anthony Delaney, PhD, FCICM; Angaj Ghosh, MBBS, FCICM;
Frank van Haren, PhD, FCICM; Nerina Harley, MD, FCICM; David Knight, MBBS, FCICM; Shay McGuiness, MBChB, FCICM;
John Mulder, MBChB, FCICM; Steve O’Donoghue, MBChB, FCICM; Nicholas Simpson, MBBS, FCICM; Paul Young, MBChB, FCICM;
for the DahLIA Investigators and the Australian and New Zealand Intensive Care Society Clinical Trials Group

Editorial page 1455

IMPORTANCE Effective therapy has not been established for patients with agitated delirium Supplemental content at
receiving mechanical ventilation. jama.com

OBJECTIVE To determine the effectiveness of dexmedetomidine when added to standard

care in patients with agitated delirium receiving mechanical ventilation.

DESIGN, SETTING, AND PARTICIPANTS The Dexmedetomidine to Lessen ICU Agitation

(DahLIA) study was a double-blind, placebo-controlled, parallel-group randomized clinical
trial involving 74 adult patients in whom extubation was considered inappropriate because of
the severity of agitation and delirium. The study was conducted at 15 intensive care units in
Australia and New Zealand from May 2011 until December 2013. Patients with advanced
dementia or traumatic brain injury were excluded.

INTERVENTIONS Bedside nursing staff administered dexmedetomidine (or placebo) initially

at a rate of 0.5 μg/kg/h and then titrated to rates between 0 and 1.5 μg/kg/h to achieve
physician-prescribed sedation goals. The study drug or placebo was continued until no longer
required or up to 7 days. All other care was at the discretion of the treating physician.

MAIN OUTCOMES AND MEASURES Ventilator-free hours in the 7 days following randomization.
There were 21 reported secondary outcomes that were defined a priori.

RESULTS Of the 74 randomized patients (median age, 57 years; 18 [24%] women), 2 withdrew
consent later and 1 was found to have been randomized incorrectly, leaving 39 patients in the
dexmedetomidine group and 32 patients in the placebo group for analysis. Dexmedetomidine
increased ventilator-free hours at 7 days compared with placebo (median, 144.8 hours vs 127.5
hours, respectively; median difference between groups, 17.0 hours [95% CI, 4.0 to 33.2 hours];
P = .01). Among the 21 a priori secondary outcomes, none were significantly worse with
dexmedetomidine, and several showed statistically significant benefit, including reduced time
to extubation (median, 21.9 hours vs 44.3 hours with placebo; median difference between
groups, 19.5 hours [95% CI, 5.3 to 31.1 hours]; P < .001) and accelerated resolution of delirium
(median, 23.3 hours vs 40.0 hours; median difference between groups, 16.0 hours [95% CI, 3.0
to 28.0 hours]; P = .01). Using hierarchical Cox modeling to adjust for imbalanced baseline
characteristics, allocation to dexmedetomidine was significantly associated with earlier
extubation (hazard ratio, 0.47 [95% CI, 0.27-0.82]; P = .007).
Author Affiliations: Author
CONCLUSIONS AND RELEVANCE Among patients with agitated delirium receiving mechanical affiliations are listed at the end of this
ventilation in the intensive care unit, the addition of dexmedetomidine to standard care article.
compared with standard care alone (placebo) resulted in more ventilator-free hours at 7 days. Group Information: The DahLIA
Investigators are listed at the end of
The findings support the use of dexmedetomidine in patients such as these.
this article.
Corresponding Author: Michael C.
TRIAL REGISTRATION clinicaltrials.gov Identifier: NCT01151865 Reade, DPhil, FCICM, University of
Queensland, Health Sciences Bldg,
JAMA. 2016;315(14):1460-1468. doi:10.1001/jama.2016.2707 Herston, Queensland, Australia 4029
Published online March 15, 2016. (m.reade@uq.edu.au).

1460 (Reprinted) jama.com

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Associação de Medicina Intensi User on 04/16/2016

 Dexmedetomidine Plus Standard Care in Patients With Agitated Delirium Preliminary Communication Research

T
he incidence of delirium in critically ill patients is high.1,2 Adult patients (aged ≥18 years) were eligible for the study
Delirium is associated with increased mortality 3 if, in the opinion of their treating physician, they continued
and decreased long-term cognitive function.4,5 Agi- to require mechanical ventilation only because their degree of
tated delirium is particularly problematic in patients receiv- agitation was so severe as to make lessening their sedation and
ing mechanical ventilation because it increases the risk extubation unsafe. These criteria were quantified objectively
of self-extubation and removal of other essential medical by requiring that the patient should meet all of these addi-
devices.6 Identification of an agent that shortens the dura- tional criteria during the 4 hours prior to randomization:
tion of established delirium would be an important therapeu- (1) need for mechanical restraint, antipsychotic or sedative
tic advance. medication, or both restraint and medication; (2) have Con-
Dexmedetomidine, a sedative α2-agonist, is theoretically fusion Assessment Method for the ICU (CAM-ICU)8 results that
an attractive treatment for patients with agitated delirium in indicated presence of delirium; and (3) have a Motor Activity
the intensive care unit (ICU) because unlike other sedatives, Assessment Scale (MAAS) score9 of 5 or greater, confirming psy-
it induces a calm yet rousable state with preserved respira- chomotor agitation.
tory drive, thereby allowing it to be continued after extubation.7 Patients were excluded if they (1) were pregnant or breast-
However, to our knowledge, no trial has compared dexme- feeding, (2) had dementia that required professional nursing
detomidine with placebo for the treatment of patients receiv- care, (3) had a head injury as the cause of their altered mental
ing mechanical ventilation who would be candidates for ex- state, (4) were already receiving dexmedetomidine or clonidine
tubation based on respiratory, cardiovascular, and metabolic for sedation, (5) had been enrolled previously in the study,
criteria but who remain intubated because of severe agitated or (6) there was a known contraindication to haloperidol or
delirium. Accordingly, we tested the hypothesis that dexme- α2-agonists.
detomidine, when added to all other elements of standard care, Patients were randomized, stratified by site and age (<55
would result in shorter duration of delirium and earlier extu- years and ≥55 years), in concealed permuted blocks of 2 to 6
bation in such patients. by a computer-generated algorithm accessed via Internet con-
nection to the Australian and New Zealand Research Centre at
Monash University. Unblinded pharmacists or nurses neither
caring for the patient nor involved in the trial prepared the
Methods study drug in identically labeled syringes. Patients in the dex-
The Dexmedetomidine to Lessen ICU Agitation (DahLIA) study medetomidine group were started with a dose of 0.5 μg/kg/h.
was a double-blind, parallel-group, placebo-controlled mul- There was a clinician-directed option of a bolus of 1.0 μg/kg
ticenter randomized trial in which intubated ICU patients were over 20 minutes.10
allocated randomly 1:1 to receive dexmedetomidine or saline Patients randomized to placebo received an identically la-
as a treatment for agitated delirium. No other aspect of pa- beled infusion of saline at an equivalent rate. Study medication
tient care was constrained, with the exception that clonidine was titrated by the bedside nurse between 0 and 1.5 μg/kg/h to
was prohibited due to its potential interaction with dexme- achieve a Richmond Agitation-Sedation Scale score11 of 0 or to
detomidine. achieve physician-prescribed goals. After 48 hours of study drug
The trial was conducted between May 9, 2011, and Decem- infusion, the treating physician could prescribe open-label dex-
ber 23, 2013, in the ICUs of 15 hospitals in Australia and medetomidine and the study drug infusion would be stopped.
New Zealand, 14 of which are mixed medical-surgical units More than 7 days of infusion of study drug was considered treat-
(7 tertiary academic and 7 metropolitan) and 1 of which ad- ment failure; at that point, the study drug was stopped and open-
mits primarily postoperative cardiac surgical patients. The trial label dexmedetomidine could be commenced.
protocol, which contains the statistical analysis plan, ap- Extubation timing (or the decision to insert a tracheos-
pears in Supplement 1. The trial protocol was approved by the tomy) was determined by senior ICU physicians, taking into
Austin Hospital human research ethics committee and, where account the assessments of bedside nurses. This decision
required, by individual hospital ethics committees. was not part of the protocol, but instead was tailored to indi-
Consent was sought from the person responsible for the vidual patient circumstances, with a physician constantly
patient. In some jurisdictions, if this person could not be con- present at each ICU. Physicians and nurses treating study pa-
tacted, the patient could be enrolled in anticipation of retro- tients, and the study staff at each site, remained blinded to
spective consent. In other jurisdictions, eligible patients were group allocation.
enrolled when the treating clinician considered participation The primary outcome was the number of ventilator-free
to be in the patient’s best interest; however, patients were not hours (the number of hours alive and free from requiring in-
included if relatives indicated that the patient would not wish vasive mechanical ventilation up until day 7 postrandomiza-
to participate. tion) during the incident ICU admission. Patients who had tra-
Once the patient had recovered sufficiently, all had the op- cheostomies inserted were considered to be free of sedation
portunity to provide fully informed consent to the use of data (and so likely to have been extubated had a tracheostomy not
and ongoing study participation. Either the patient or person been inserted) after a 4-hour period in which they received no
responsible could withdraw consent at any stage. A data and sedatives or opioid analgesics.
safety monitoring committee reviewed all adverse effects. The 21 reported secondary outcomes that were defined
There was no interim analysis. a priori included time to extubation (in the case of tracheos-

jama.com (Reprinted) JAMA April 12, 2016 Volume 315, Number 14 1461

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Associação de Medicina Intensi User on 04/16/2016

 Research Preliminary Communication Dexmedetomidine Plus Standard Care in Patients With Agitated Delirium

Figure 1. Patient Flow Diagram of the DahLIA Trial

Approximately 21 500 admissions of

intubated patients at 15 intensive care
units in 2 countries a

74 Patients randomized
The number of patients not
randomized due to not meeting the
41 Randomized to receive 33 Randomized to receive placebo inclusion criteria or having met 1 of
dexmedetomidine 33 Received placebo as randomized the exclusion criteria was not
40 Received dexmedetomidine recorded. DahLIA indicates
as randomized
1 Study treatment discontinued Dexmedetomidine to Lessen
(met an exclusion criterion and ICU Agitation.
inappropriately randomized) a
Estimated from total admission
numbers to each intensive care unit
1 Withdrew consent to use data 1 Withdrew consent to use data along with data from the Australian
and New Zealand Intensive Care
Society Centre for Outcomes and
39 Included in primary intent-to- 32 Included in primary intent-to- Resource Evaluation14 on the
treat analysis treat analysis
proportion of these admissions that
were intubated.

tomy, with liberation from sedation and mechanical ventila- All statistical analyses were performed using Stata version
tion defined in the same manner as used for the primary out- 11.2 (StataCorp) or SAS version 9.4 (SAS Institute Inc) with a
come), time taken to achieve a satisfactory sedation score 2-sided P value of less than .05 considered significant. No adjust-
(Richmond Agitation-Sedation Scale score of −2 to 1), time taken ment was made for multiple comparisons, and so the secondary
to achieve a satisfactory agitation score (MAAS score of 2 to outcomes presented (although all prespecified) should be con-
4), proportion of study time with a satisfactory MAAS score, sidered exploratory, yielding hypothesis-generating findings.
period until the nurse caring for the patient thought it was time Based on a pilot study with a mean control estimate of 108
to extubate, time to the first CAM-ICU results that indicated ventilator-free hours (SD, 46 ventilator-free hours),12 a sample
absence of delirium, time spent having CAM-ICU results that size of 96 patients was estimated to provide 80% power to de-
indicated presence of delirium, the requirement for sedative tect a 20-hour difference (ie, half the effect size observed in a
and antipsychotic medications, the proportion who under- pilot study) using a 2-tailed hypothesis at an α level of .05.
went tracheostomy, requirement for reintubation, daily Sepsis- These calculations include an inflation rate of 15% to account
related Organ Failure Assessment score, and lengths of stay in for the possibility that ventilation-free days would not be nor-
the ICU and hospital. Adverse events were recorded both pro- mally distributed.13
spectively and by review of each clinical chart. However, the sponsoring pharmaceutical company (Ho-
Modified intention-to-treat analyses were performed. spira Australia) decided against extending funding and pro-
Modification was permitted to account for postrandomiza- vision of study drug beyond a date that had been earlier agreed.
tion circumstances that prevented use of data from certain pa- Consequently, the trial was terminated prematurely in Decem-
tients. Because there were no missing data for the primary out- ber 2013 after 74 patients had been randomized. At no stage
come and less than 5% missing for all secondary outcomes, no did the pharmaceutical company have access to the study data,
data imputation was performed. Due to nonnormality, all con- and no data analysis by the study investigators had occurred
tinuous outcomes were compared using Mann-Whitney tests prior to this decision.
with location shifts between treatment groups calculated using To account for the possibility that early termination may
the Hodges-Lehmann estimate and reported using distribution- exaggerate the effect size, additional analyses were per-
free 95% confidence limits. formed post hoc to assess the likelihood of a null finding had
The sensitivity analysis accounting for multiple sites was the study been completed as originally planned. These analy-
performed using the van Elteren statistic. Categorical out- ses were performed using 10 000 simulations based first on
comes were compared using χ2 or Fisher exact tests and re- the assumption that nonenrolled patients came from the origi-
ported as differences in proportion (95% confidence inter- nal projected population and then second based on the as-
val). Time-to-event data were compared using log-rank tests sumption that the nonenrolled patients came from the ob-
and presented as Kaplan-Meier curves. To account for any ef- served population.
fect of site and for baseline imbalances, a Cox proportional haz-
ards regression model was used with patients nested within
site, and site treated as a random effect with the following co-
variates included in the model: Acute Physiology and Chronic
Results
Health Evaluation II diagnosis, duration of intubation, and elec- From May 2011 until December 2013, we randomized 74 pa-
tive status. Proportionality assumptions were determined using tients (Figure 1). However, 1 patient allocated to dexmedeto-
log survival plots. midine had been randomized in error, and 1 patient in each

1462 JAMA April 12, 2016 Volume 315, Number 14 (Reprinted) jama.com

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Associação de Medicina Intensi User on 04/16/2016

 Dexmedetomidine Plus Standard Care in Patients With Agitated Delirium Preliminary Communication Research

Table 1. Baseline Patient Characteristicsa

Dexmedetomidine Placebo
(n = 39) (n = 32)
Tertiary ICU with >18 beds 17 (43.6) 15 (46.9)
Age, median (IQR), y 58 (47-65) 56.5 (46-69.5)
Male sex 28 (71.8) 25 (78.1)
Weight, median (IQR), kg 83 (72-100) 85 (78-105)
Living at home 39 (100) 31 (96.9)
APACHE II score immediately prior to randomization,
median (IQR)
Acute physiology 11 (8-16) 11.5 (8.5-16.5)
Total 14 (10-22) 14 (11-20)
APACHE II comorbidity score ≥2 10 (25.6) 6 (18.8)
APACHE II diagnostic category
Nonoperative 17 (43.6) 12 (37.5)
Respiratory 5 (29.4) 5 (41.7)
Cardiovascular 4 (23.5) 0
Neurological 4 (23.5) 2 (16.7)
Other 4 (23.5) 5 (41.7)
Operative 22 (56.4) 20 (62.5)
Multiple trauma 4 (18.2) 2 (10.0)
Cardiovascular 10 (45.5) 15 (75.0)
Respiratory 1 (4.5) 0
Neurosurgery including neurotrauma 0 1 (5.0)
Gastrointestinal 5 (22.7) 1 (5.0)
Other 2 (9.1) 1 (5.0)
Emergency ICU admission 29 (74.4) 18 (56.3)
During the 24 h prior to randomization
Mechanical restraint 13 (33.3) 11 (34.4)
Use of pharmacotherapy (n = 38) (n = 32)
Midazolam 4 (10.5) 5 (15.6)
Propofol 38 (100.0) 29 (90.6)
Morphine 9 (23.7) 9 (28.1) Abbreviations: APACHE II, Acute
Physiology and Chronic Health
Fentanyl 14 (36.8) 11 (34.4)
Evaluation II; ICU, intensive care unit;
Antipsychotic (haloperidol, olanzapine, risperidone, 9 (23.7) 6 (18.8) IQR, interquartile range.
or quetiapine)
a
Data are expressed as No. (%)
Duration of intubation prior to enrollment, median (IQR), h 63 (26-96) 43.5 (23-72)
unless otherwise indicated.

group withdrew consent, leaving data from 39 patients in the fore, open-label dexmedetomidine was commenced, and the
dexmedetomidine group and 32 in the placebo group for analy- placebo study drug was stopped. Five patients (1 allocated to
sis. The baseline characteristics of study participants appear dexmedetomidine and 4 to placebo) received open-label
in Table 1. dexmedetomidine after administration of the study drug for
Almost all patients were sedated with propofol. Approxi- 7 days.
mately one-third of the patients required mechanical re- Compared with the dexmedetomidine group, patients al-
straint immediately prior to randomization and 20% re- located to the placebo group received a significantly greater
ceived an antipsychotic drug. The median Acute Physiology rate and volume of study drug on days 1 and 2 (Table 2). More
and Chronic Health Evaluation II score immediately prior to patients in the placebo group received antipsychotic medica-
randomization was low, reflecting the inclusion criterion that tions (haloperidol, risperidone, olanzapine, or quetiapine) on
these patients should be ready for extubation except for hav- any day than was true for the dexmedetomidine group (65.6%
ing agitated delirium. vs 36.8%, respectively; mean difference between groups,
One male patient randomized to receive dexmedetomi- −28.8% [95% CI, −51.3% to −6.3%; P = .02). There were no sig-
dine did not receive any study drug because his physician nificant differences in the use of individual antipsychotic drugs
decided the delirium had resolved almost immediately (eTable 1 in Supplement 2). On several days, significantly lower
after randomization. While blinded to study drug allocation, quantities of intercurrent sedatives (propofol and mid-
the treating physician of 1 patient randomized to placebo azolam) and opioids (morphine and fentanyl) were used in the
decided, after 48 hours of administering the placebo study dexmedetomidine group compared with the placebo group
drug, optimal treatment would be dexmedetomidine. There- (eTable 2).

jama.com (Reprinted) JAMA April 12, 2016 Volume 315, Number 14 1463

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Associação de Medicina Intensi User on 04/16/2016

 Research Preliminary Communication Dexmedetomidine Plus Standard Care in Patients With Agitated Delirium

Table 2. Quantification of Study Drug Use and Administration of Intercurrent and Subsequent Medications
Dexmedetomidine Placebo Difference Between
(n = 39) (n = 32) Groups (95% CI) P Value
Bolus at start of study drug infusion, No./total No. of 2/37 (5.4) 2/32 (6.3) −0.8 (−12.0 to 10.3) >.99
observations (%)
Time until peak infusion rate of study drug reached, 8.3 (5.0 to 17.0) 8.3 (4.0 to 15.3) 0 (−3.0 to 2.0) .68
median (IQR), h
Total duration of study drug infusion, median (IQR), h 23.5 (19.5 to 35.0) 35.0 (24.8 to 71.5) −10.0 (−262.8 to −2.8) .004
Study drug continued after extubation, No. (%) 4 (10.3) 4 (12.5) −2.2 (−17.1 to 12.7) >.99
Day 1
Study drug rate, median (IQR), mL/ha 12.8 (8.3 to 22.2) 25.4 (21.3 to 30.4) −12.2 (−16.2 to −7.7) <.001
Total volume of study drug administered, 9.8 (5.5 to 18.4) 25.2 (20.5 to 29.5) −13.7 (−17.6 to −8.7) <.001
median (IQR), mL/hb
Day 2 (n = 13) (n = 18)
Study drug rate, median (IQR), mL/ha 11.0 (4.6 to 20.3) 25.9 (20.4 to 32.0) −14.1 (−22.1 to −5.1) .004
Total volume of study drug administered, 13 (5.7 to 21.6) 25.9 (19.4 to 32.8) −12.0 (−21.2 to −2.9) .01
median (IQR), mL/hb
Day 3 (n = 8) (n = 13)
Study drug rate, median (IQR), mL/ha 17.9 (7.0 to 27.2) 26.0 (20.0 to 30.2) −7.6 (−19.8 to 6.2) .26
Total volume of study drug administered, 17.9 (6.9 to 27.3) 27.4 (18.7 to 30.6) −8.6 (−19.6 to 5.2) .19
median (IQR), mL/hb
Day 4 (n = 3) (n = 6)
Study drug rate, median (IQR), mL/ha 22.2 (18.7 to 32.3) 21.9 (9.6 to 35.6) 3.6 (−22.5 to 23.3) .90
Total volume of study drug administered, 22.2 (17.5 to 32.3) 20.3 (9.0 to 35.6) 4.6 (−23.7 to 31.5) .90
median (IQR), mL/hb
Day 5 (n = 3) (n = 4)
Study drug rate, median (IQR), mL/ha 17.8 (14.4 to 19.2) 30.2 (15.1 to 38.4) −13.4 (−26.8 to 13.7) .38
Total volume of study drug administered, 17.8 (14.4 to 19.2) 30.2 (15.1 to 38.4) −13.3 (−26.8 to 13.7) .38
median (IQR), mL/hb
Day 6 (n = 2) (n = 2)
Study drug rate, median (IQR), mL/ha 18.4 (13.6 to 23.3) 38.4 (35.6 to 41.2) −20.0 (−27.6 to 12.4) .25
Total volume of study drug administered, 18.4 (13.6 to 23.3) 22.6 (4.0 to 41.2) −4.2 (−27.6 to 19.3) >.99
median (IQR), mL/hb
Day 7 (n = 1) (n = 2)
Study drug rate, median (IQR), mL/ha 9.8 (9.8 to 9.8) 38.4 (35.6 to 41.2) −28.6 (−31.4 to 25.8) .54
Total volume of study drug administered, 9.8 (9.8 to 9.8) 38.4 (35.6 to 41.2) −28.6 (−31.4 to 25.8) .54
median (IQR), mL/hb
Received open-label dexmedetomidine, No./total No. of
observations (%)
Between 48 h and 7 d of receiving the study drug 0/39 1/32 (3.1) −3.1 (−9.2 to 2.9)c .45
After 7 d of receiving the study drug 2/39 (5.1) 3/32 (9.4) −4.2 (−16.5 to 8.0)c .65
Received any antipsychotic on any day, No./total No. of 14/38 (36.8) 21/32 (65.6) −28.8 (−51.3 to −6.3)c .02
observations (%)
Study days in which any antipsychotic medication was 26.3 (21.3 to 39.0) 40.0 (20.6 to 49.0) −11.4 (−20.6 to 3.3) .08
administered, median (95% CI), %d
Study days requiring vasopressor for blood pressure support, 0 (0 to 35.5) 16.7 (3.0 to 20.7) −9.5 (−16.7 to 10.1) .27
median (95% CI), %d
d
Abbreviation: IQR, interquartile range. Indicates the days following randomization until discharge from the intensive
a
In patients receiving the study drug. care unit or day 7, whichever was sooner. Each daily percentage was calculated
b
using the number of patients who received medication that day divided by the
In patients remaining in the study.
number of patients remaining in the study on that day.
c
Data are expressed as percentages.

During the 7 days after randomization, propofol use was CI, −41.1% to −2.0%]; P = .03), and received a significantly
common in both groups (71.8% of the dexmedetomidine group higher median dosage of fentanyl (1543 μg [IQR, 335 to
and 87.5% of the placebo group). The median dosage was sig- 6629 μg] vs 310 μg [IQR, 200 to 680 μg], respectively; me-
nificantly higher in the placebo group (5390 mg [interquar- dian difference between groups, 609 μg [95% CI, 50 to
tile range {IQR}, 1880 to 10 803 mg]) compared with the dex- 2225 μg]; P = .03) (eTable 3 in Supplement 2).
medetomidine group (980 mg [IQR, 280 to 3050 mg]) (median Patients randomized to dexmedetomidine had signifi-
difference between groups, −3094.5 mg [95% CI, −5852 to cantly more ventilator-free hours at 7 days (median, 144.8
−940 mg]; P < .001). Patients in the placebo group were more hours vs 127.5 hours in the placebo group; median difference
likely to receive morphine (34.4% vs 12.8% of the dexmedeto- between groups, 17.0 hours [95% CI, 4.0-33.2 hours]; P = .01;
midine group; mean difference between groups, −21.6% [95% van Elteren site-adjusted P = .04) (Table 3). There were no

1464 JAMA April 12, 2016 Volume 315, Number 14 (Reprinted) jama.com

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Associação de Medicina Intensi User on 04/16/2016

 Dexmedetomidine Plus Standard Care in Patients With Agitated Delirium Preliminary Communication Research

Table 3. Primary and Secondary Study Outcomes

Dexmedetomidine Placebo Difference Between
(n = 39) (n = 32) Groups (95% CI) P Value
Primary Outcome
Time ventilator-free during the first 7 d after randomization, 144.8 (114.0 to 156.0) 127.5 (92.0 to 142.8) 17.0 (4.0 to 33.2) .01
median (IQR), h
Secondary Outcomes
Time taken to achieve a satisfactory sedation score, 1 (1 to 1) 1 (1 to 1) 0 (0 to 0) .90
median (IQR), da
Time until bedside nurse thought patient was ready 19.1 (16.7 to 25.8)b 40.5 (21.1 to 90.7)c −21.1 (−34.5 to −6.0) <.001
for extubation (not tracheostomy), median (IQR), h
Time to extubation (not tracheostomy), median (IQR), h 21.9 (18.3 to 27.7) 44.3 (25.3 to 94.2) −19.5 (−31.1 to −5.3) <.001
Underwent tracheostomy, No. (%) 7 (17.9) 2 (6.3) 11.7 (−3.0 to 26.4) .14
Time to tracheostomy, median (IQR), h 41.9 (20.2 to 101.8)d 71.1 (70.3 to 71.9)e −29.2 (−71.9 to 95.2) .88
Required intubation for a second time, No. (%) 0 1 (3.2) −3.1 (−9.2 to 2.9) .45
Extubated, No. (%)
<24 h after randomization 8 (20.5) 2 (6.3) 14.3 (−0.9 to 29.5) .09
<48 h after randomization 30 (76.9) 19 (59.4) 17.5 (−4.0 to 39.1) .13
On the same calendar day as randomization 6 (15.4) 2 (6.3) 9.1 (−5.0 to 23.2) .23
On the first or second calendar day after randomization 22 (56.4) 14 (43.8) 12.6 (−10.6 to 35.8) .28
Confusion Assessment Method for the ICU
Time to first results indicating absence of delirium, 23.3 (13.0 to 54.0) 40.0 (25.3 to 76.0) −16.0 (−28.0 to −3.0) .01
median (IQR), h
Time spent with results indicating presence of delirium, 36 (24 to 66) 62 (42.5 to 106.75) −24 (−41 to −6) .009
median (IQR), h
Proportion of days postrandomization spent with results 47 (30 to 76) 62 (46 to 86) −10 (−30 to 0) .05
indicating presence of delirium, median (IQR), %
Time spent with at least 1 assessment indicating presence 1 (1 to 3) 3 (1 to 4) −1 (−2 to 0) .02
of delirium postrandomization, median (IQR), d
Required mechanical restraint on any day, No. (%) 10 (26.3)f 15 (46.9)b −20.6 (−42.8 to 1.7) .07
Proportion of study days in which mechanical restraint 20.0 25.0 −6.8 (−15.6 to 6.1) .34
was required, median, %
Proportion of study days spent lightly sedated, median, % 82.4 74.1 7.8 (−6.6 to 20.2) .37
ICU length of stay, median (IQR), d
Postrandomization 2.9 (2.1 to 4.9) 4.1 (3.0 to 7.9) −1.0 (−2.1 to 0.1) .09
Overall 5.9 (3.7 to 10.2) 7.5 (4.7 to 11.7) −1 (−3 to 1) .29
Hospital length of stay, median (IQR), d
Postrandomization 8.5 (6.2 to 13.6) 9.5 (6.5 to 13.5) 0 (−3 to 3) .96
Overall 14.0 (10.0 to 20.0) 12.5 (9.0 to 21.0) 1 (−3 to 5) .61
Location of death, No. (%)
ICU 1 (2.6) 0 2.6 (−2.4 to 7.5) >.99
Hospital 2 (5.1) 0 5.1 (−1.8 to 12.1) .50
Discharged to rehabilitation rather than home 5 (13.2)f 3 (9.7)c 3.5 (−11.5 to 18.4) .65
or other acute hospital, No. (%)
Adverse event, No. (%)
Related to bradycardia 2 (5.3)f 0g 5.3 (−1.8 to 12.4) .50
Related to agitation 1 (2.6)f 2 (6.7)g −4.0 (−14.3 to 6.2) .58
d
Abbreviations: ICU, intensive care unit; IQR, interquartile range. There were 7 patients with recorded data.
a e
Assessed as Richmond Agitation-Sedation Scale score of −2 to 1. There were 2 patients with recorded data.
b f
There were 32 patients with recorded data. There were 38 patients with recorded data.
c g
There were 31 patients with recorded data. There were 30 patients with recorded data.

differences between groups in the proportion of patients In the time to event analysis, dexmedetomidine was as-
who required tracheostomy (Table 3). A sensitivity analysis sociated with earlier extubation (hazard ratio [HR], 0.58 [95%
examining only patients who did not receive a tracheostomy CI, 0.36-0.95]; P = .03) (Figure 2). In the hierarchical Cox pro-
showed the same qualitative difference in ventilator-free portional hazards regression model adjusting for baseline char-
hours (median, 147.3 hours [IQR, 131-158 hours] in the dex- acteristics (eTable 4 in Supplement 2), allocation to dexme-
medetomidine group vs 128 hours [IQR, 92-143 hours] in the detomidine remained associated with earlier extubation (HR,
placebo group; median difference between groups, 19.5 0.47 [95% CI, 0.27-0.82]; P = .007). In a sensitivity analysis that
hours [95% CI, 6.5-36.0 hours]; P = .002). used right censoring for patients with tracheostomy at the time

jama.com (Reprinted) JAMA April 12, 2016 Volume 315, Number 14 1465

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Associação de Medicina Intensi User on 04/16/2016

 Research Preliminary Communication Dexmedetomidine Plus Standard Care in Patients With Agitated Delirium

tion was provided to educate the nurses regarding collection

Figure 2. Kaplan-Meier Analysis of the Proportion of Patients Remaining
Intubated During the First 7 Days of the Study
of the MAAS score than the CAM-ICU. This absence of data was
missed during interim monitoring, preventing remediation of
1.00
Hazard ratio, 0.58 (95% CI, 0.36-0.95);
the problem during the conduct of the trial.
Proportion Remaining Intubated

log-rank P = . 03 Adverse events (bradycardia requiring interruption of

or Sedated With Tracheostomy

0.80
study drug, hypotension requiring vasopressor support, and
agitation requiring temporarily increased sedation) were rare
0.60
and not different between study groups. Two of 39 patients
0.40 in the dexmedetomidine group and 2 of 32 patients in the pla-
Placebo
cebo group received a bolus of dexmedetomidine. None of
0.20 these patients were among those who experienced a brady-
Dexmedetomidine cardia-related adverse event. A patient with known cardio-
0
0 20 40 60 80 100 120 140 160 myopathy developed ventricular tachycardia 8 hours after ces-
Hours After Randomization sation of the study drug; however, the data and safety
No. at risk monitoring committee ruled that this was not related to the
Dexmedetomidine 39 10 4 2
Placebo 32 13 6 2
study, and the protocol continued without modification.
There were no statistically significant differences be-
tween the groups in Sepsis-related Organ Failure Assessment
of the procedure, there was no difference in qualitative out- score on any study day. As expected in this comparatively re-
come (HR, 0.39 [95% CI, 0.21-0.71]; P = .002). covered cohort of critically ill patients, ICU and hospital mor-
The median time to extubation was 21.9 hours for the dex- tality were low and not different between groups. Only 1 pa-
medetomidine group vs 44.3 hours for the placebo group (me- tient required reintubation, which occurred 2 hours following
dian difference between groups, 19.5 hours [95% CI, 5.3 to 31.1 elective extubation. No patient self-extubated.
hours]; P < .001; van Elteren site-adjusted P = .004). Bedside
nurses thought their patients were ready to extubate signifi-
cantly earlier (P < .001) if they were receiving dexmedetomi-
dine (median, 19.1 hours [IQR, 16.7 to 25.8 hours]) than if they
Discussion
were receiving placebo (median, 40.5 hours [IQR, 21.1 to 90.7 In this double-blind placebo-controlled randomized trial in-
hours]) (median difference between groups, −21.1 hours [95% volving patients with agitated delirium receiving mechanical
CI, −6.0 to −34.5 hours]). The median ICU length of stay was ventilation, who were primarily receiving propofol-based se-
2.9 days (IQR, 2.1 to 4.9 days) with dexmedetomidine vs 4.1 dation and antipsychotic medications determined by their
days (IQR, 3.0 to 7.9 days) with placebo (P = .09). treating physicians, dexmedetomidine increased the num-
An additional post hoc simulation analysis calculating the ber of ventilator-free hours during the 7 days following ran-
probability of finding no difference in the median duration of domization. Compared with placebo, dexmedetomidine has-
ventilator-free hours during the first 7 days after randomiza- tened the resolution of delirium and extubation in patients by
tion if the trial included the planned number of patients (using approximately 1 day. Adverse events were rare and not differ-
both the original design effects and the observed effects) found ent between the groups.
a chance of less than 7% of producing a null result (P > .05) (cal- The results of this study are consistent with earlier large
culated using either approach). This lack of likely qualitative randomized clinical trials comparing dexmedetomidine with
difference occurred because the observed treatment effect in benzodiazepines or propofol as a sedative, which found dex-
this study was very similar to that projected. medetomidine was associated with less delirium in the ICU15,16
Allocation to dexmedetomidine was associated with sev- and reduced time to extubation.17 However, these were trials
eral improved indices of delirium (Table 3). With dexmedeto- of dexmedetomidine as a sedative rather than as a treatment
midine, delirium resolved more rapidly (median, 23.3 hours for delirium. Suggestion of a therapeutic effect of dexmedeto-
vs 40.0 hours in the placebo group; median difference be- midine in established delirium was present in the Safety and
tween groups, 16.0 hours [95% CI, 3.0-28.0 hours]; P = .01; van Efficacy of Dexmedetomidine Compared With Midazolam
Elteren site-adjusted P = .04). Compared with the patients who trial.16 In the post hoc analysis of the 60% of patients who had
received placebo, the patients who received dexmedetomi- CAM-ICU results indicating the presence of delirium at the time
dine had delirium for a lower proportion of their ICU stay, and of randomization, there was a reduction in the prevalence of
had a median of 2 additional delirium-free days during their delirium from 95.5% to 68.7%.
ICU stay. There was no between-group difference in the time Dexmedetomidine had a propofol- and fentanyl-sparing
taken to achieve a satisfactory sedation score. effect on day 1. It is possible that deliriogenic sedatives were
The time taken to achieve a satisfactory MAAS score (or replaced with alternatives less prone to cause delirium.
proportion of time spent with a satisfactory MAAS score) can- However, propofol and opioids are probably less delirio-
not be reported because almost no patients had a MAAS score genic than benzodiazepines, 18 and nonbenzodiazepine
recorded after that which was assessed by the study research alternatives were the overwhelming choice for intercurrent
coordinator at the time of study entry. The protocol required care. Therefore, a direct antidelirium effect of dexmedeto-
bedside nurses to collect MAAS scores; however, less atten- midine remains possible. How an α2-agonist might exert this

1466 JAMA April 12, 2016 Volume 315, Number 14 (Reprinted) jama.com

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Associação de Medicina Intensi User on 04/16/2016

 Dexmedetomidine Plus Standard Care in Patients With Agitated Delirium Preliminary Communication Research

effect remains speculative.19 It is possible that the analgesic The study also has some limitations. The planned 96 pa-
effects of dexmedetomidine might have lessened both agi- tients were not recruited. Unplanned early termination of clini-
tation and delirium. cal trials can exaggerate effect size. However, the probability
Our study is the same size as the largest previous trial20 of finding a qualitatively different result in the primary out-
of therapy for patients with agitated delirium determined by come had the trial recruited to the target sample size was less
the CAM-ICU; that study compared (nonblinded) dexmedeto- than 7%. The relatively small sample size led to several chance
midine with midazolam in 72 patients who had been intu- imbalances in baseline characteristics, notably the duration of
bated after undergoing elective cardiac surgery. Patients in the ventilation before randomization. However, when adjusted for
dexmedetomidine group were extubated earlier (46.6 hours this imbalance, dexmedetomidine remained associated with
vs 58.3 hours in the midazolam group, P < .001), but the be- earlier extubation. Although there was a difference in the pri-
tween-group comparisons of agitated delirium were not re- mary outcome and several congruent secondary outcomes, the
ported. This study showed the superiority of dexmedetomi- study was underpowered to detect differences in important
dine over midazolam as a sedative for patients with agitated end points including ICU length of stay.
delirium who had undergone intubation. However, given con- Many patients (n = 21 500) were screened in the effort to
sensus recommendations against benzodiazepines in these recruit only 74 patients; therefore, the results may not be gen-
circumstances,18 this is less relevant than the question ad- eralizable to patients earlier in the course of their critical ill-
dressed by our study. ness, with other forms of delirium, or not intubated. Only pa-
To our knowledge, the only other published trial target- tients who could be extubated (were it not for agitated delirium)
ing patients with agitated delirium was a single-center pilot were recruited. Although we cannot say if the results apply to
study of 20 patients, which compared nonblinded infusions patients with agitated delirium in the ICU earlier in their ill-
of dexmedetomidine or haloperidol. Dexmedetomidine short- nesses, the dexmedetomidine sedative trials provide reassur-
ened time to extubation (from 42.5 hours to 19.2 hours, P = .02) ing evidence of safety and a suggestion of efficacy in this
and ICU length of stay (from 6.5 days to 1.5 days, P = .004). In patient group.
our larger trial, only 30.4% of patients received haloperidol and We cannot comment on whether dexmedetomidine might
a higher percentage received atypical antipsychotics. be effective in patients with traumatic brain injury or demen-
The only previous delirium pharmacotherapy placebo- tia. However, there is no evidence that dexmedetomidine
controlled trial (in which only 30.6% had a Riker Sedation- would harm such patients. Resolution of delirium was one of
Agitation Scale21 score of ≥5 at the time of enrollment, sug- the most important end points, but identification of delirium
gesting agitated delirium)22 had a similar design to that of our in critically ill patients is problematic, as previously argued.23
study. Patients (not all intubated at the time of enrollment) were Delirium was defined using the CAM-ICU, which has been the
randomized to quetiapine or placebo, with all other elements subject of criticism for its false-positive results in the context
of care as directed by the physician. Both groups received as- of recently discontinued sedation. 24,25 Nonetheless, the
needed haloperidol. Delirium resolved faster with quetiap- CAM-ICU is recommended by consensus guidelines18 and our
ine but duration of mechanical ventilation and ICU length of study was blinded.
stay were similar to placebo. Even though clinicians were blinded to study drug allo-
This study has several strengths. First, it used a double- cation, dexmedetomidine often causes bradycardia,10 which
blind, multicenter, randomized, permuted block, placebo- might have suggested study drug allocation. However, in agi-
controlled design. Second, the study had objective enroll- tated patients receiving strong doses of sedating drugs to avoid
ment criteria. Third, the primary end point was patient centered self-injury, changes in heart rate are common.
and with likely financial cost-benefit implications. Fourth, the
protocol replicated current practice by having bedside nurses
independently titrating the study drug to either a physician-
prescribed sedation goal or a default goal of light sedation.
Conclusions
Fifth, all other therapies were consistent with current consen- Among patients with agitated delirium receiving mechanical
sus recommendations.18 Sixth, the development of lack of phy- ventilation in the ICU, the addition of dexmedetomidine to
sician equipoise was accommodated by permitting open- standard care compared with standard care alone (placebo) re-
label dexmedetomidine after patients had received the study sulted in more ventilator-free hours at 7 days. The findings sup-
drug for 48 hours, but this had no effect on the results. port the use of dexmedetomidine in patients such as these.

ARTICLE INFORMATION Australian and New Zealand Intensive Care Australia (van Haren); Royal Melbourne Hospital,
Published Online: March 15, 2016. Research Centre, School of Public Health and Melbourne, Australia (Harley); Christchurch
doi:10.1001/jama.2016.2707. Preventive Medicine, Monash University, Hospital, Christchurch, New Zealand (Knight);
Melbourne, Australia (Bailey); Flinders Medical Auckland City Hospital, Auckland, New Zealand
Author Affiliations: Burns, Trauma, and Critical Centre, Adelaide, Australia (Bersten); Toowoomba (McGuiness); Western Hospital, Melbourne,
Care Research Centre, University of Queensland Hospital, Toowoomba, Australia (Cheung); Australia (Mulder); Royal Brisbane and Women’s
and Joint Health Command, Australian Defence Peninsula Health, Melbourne, Australia (Davies); Hospital, Brisbane, Australia (O’Donoghue);
Force, Brisbane, Australia (Reade); Austin Hospital, Royal North Shore Hospital of Sydney, Sydney, Geelong Hospital, Geelong, Australia (Simpson);
Melbourne, Australia (Eastwood); School of Australia (Delaney); Northern Hospital, Melbourne, Wellington Hospital, Wellington, New Zealand
Medicine, University of Melbourne and Austin Australia (Ghosh); Canberra Hospital, Canberra,
Hospital, Melbourne, Australia (Bellomo);

jama.com (Reprinted) JAMA April 12, 2016 Volume 315, Number 14 1467

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Associação de Medicina Intensi User on 04/16/2016

 Research Preliminary Communication Dexmedetomidine Plus Standard Care in Patients With Agitated Delirium

(Young); Medical Research Institute of John Durning, Robert Frengley, Alex Kazemi, Laura 11. Sessler CN, Gosnell MS, Grap MJ, et al.
New Zealand, Wellington (Young). Rust, Rima Song, and Anna Tilsley (Middlemore The Richmond Agitation-Sedation Scale. Am J
Author Contributions: Dr Reade had full access to Hospital, Auckland, New Zealand); Angaj Ghosh, Respir Crit Care Med. 2002;166(10):1338-1344.
all of the data in the study and takes responsibility Mary Park, and Yasmin Sungkar (Northern Hospital, 12. Reade MC, O’Sullivan K, Bates S, et al.
for the integrity of the data and the accuracy of the Melbourne, Australia); Benjamin Cheung, Indranil Dexmedetomidine vs haloperidol in delirious,
data analysis. Chatterjee, and Judy Smith (Toowoomba Hospital, agitated, intubated patients. Crit Care. 2009;13(3):
Study concept and design: Reade, Eastwood, Toowoomba, Australia); Rachel Dunlop, Steve R75.
Bellomo. O’Donoghue, Michael C. Reade, and Jason Roberts
(Royal Brisbane and Women’s Hospital, Brisbane, 13. Lehmann EL, D’Abrera HJM. Nonparametrics:
Acquisition, analysis, or interpretation of data: All Statistical Methods Based on Ranks. New York, NY:
authors. Australia); Anthony Delaney, Naomi Hammond,
Anne O’Connor, and Melissa Passer (Royal North Springer; 2006.
Drafting of the manuscript: Reade, Eastwood,
Bellomo, Young. Shore Hospital of Sydney, Sydney, Australia); 14. Hart GK. The ANZICS CORE. Crit Care Resusc.
Critical revision of the manuscript for important Deborah Barge, Nerina Harley, Andrea Jordan, and 2008;10(2):83-88.
intellectual content: All authors. Elizabeth Moore (Royal Melbourne Hospital, 15. Pandharipande PP, Pun BT, Herr DL, et al. Effect
Statistical analysis: Reade, Bailey. Parkville, Australia); Lynn Andrews, Dick Dinsdale, of sedation with dexmedetomidine vs lorazepam
Obtained funding: Reade, Bellomo, Young. Kristy Whitelaw, and Paul Young (Wellington on acute brain dysfunction in mechanically
Administrative, technical, or material support: All Hospital, Wellington, New Zealand); and Samantha ventilated patients. JAMA. 2007;298(22):2644-2653.
authors. Bates, Anna Tippett, Forbes McGain, and John
Mulder (Western Hospital, Melbourne, Australia). 16. Riker RR, Shehabi Y, Bokesch PM, et al.
Study supervision: Reade, Eastwood. Dexmedetomidine vs midazolam for sedation of
The data and safety monitoring committee
Conflict of Interest Disclosures: The authors have comprised Paul Myles, MD, FANZCA, Enjarn Lin, critically ill patients. JAMA. 2009;301(5):489-499.
completed and submitted the ICMJE Form for MBBS, FANZCA, and David Daly, MBBS, FANZCA 17. Jakob SM, Ruokonen E, Grounds RM, et al.
Disclosure of Potential Conflicts of Interest. (all 3 at the Alfred Hospital, Melbourne, Australia). Dexmedetomidine vs midazolam or propofol for
Dr Reade reported receiving single fee of A$1000 in The committee was not compensated for its work. sedation during prolonged mechanical ventilation.
2009 to contribute to a Hospira clinician advisory JAMA. 2012;307(11):1151-1160.
board preparing guidelines for the use of REFERENCES
dexmedetomidine. Dr Young reported receiving 18. Barr J, Fraser GL, Puntillo K, et al. Clinical
grants and personal fees from Baxter Healthcare 1. Bergeron N, Dubois MJ, Dumont M, et al. practice guidelines for the management of pain,
Corporation. No other disclosures were reported. Intensive Care Delirium Screening Checklist. agitation, and delirium in adult patients in the
Intensive Care Med. 2001;27(5):859-864. intensive care unit. Crit Care Med. 2013;41(1):263-
Funding/Support: This study was partly funded by 306.
Hospira Australia through an unrestricted grant of 2. Ely EW, Girard TD, Shintani AK, et al.
A$25000 plus free study drug supply. Individual Apolipoprotein E4 polymorphism as a genetic 19. Mo Y, Zimmermann AE. Role of
site funding was supplemented by grants from the predisposition to delirium in critically ill patients. dexmedetomidine for the prevention and
Wellington Hospital Research Office and the Austin Crit Care Med. 2007;35(1):112-117. treatment of delirium in intensive care unit
Hospital Intensive Care Specialists Trust Fund. 3. Ely EW, Shintani A, Truman B, et al. Delirium as a patients. Ann Pharmacother. 2013;47(6):869-876.

Role of the Funder/Sponsor: The funders had no predictor of mortality in mechanically ventilated 20. Yapici N, Coruh T, Kehlibar T, et al.
role in the design and conduct of the study; patients in the intensive care unit. JAMA. 2004;291 Dexmedetomidine in cardiac surgery patients who
collection, management, analysis, and (14):1753-1762. fail extubation and present with a delirium state.
interpretation of the data; preparation or approval 4. Pandharipande PP, Girard TD, Jackson JC, et al. Heart Surg Forum. 2011;14(2):E93-E98.
of the manuscript; or decision to submit the Long-term cognitive impairment after critical 21. Riker RR, Picard JT, Fraser GL. Prospective
manuscript for publication. illness. N Engl J Med. 2013;369(14):1306-1316. evaluation of the Sedation-Agitation Scale for adult
Group Information: The DahLIA Investigators were 5. van den Boogaard M, Schoonhoven L, Evers AW, critically ill patients. Crit Care Med. 1999;27(7):1325-
Vic Bennett, Jasmin Board, Andrew Davies, Vinodh et al. Delirium in critically ill patients. Crit Care Med. 1329.
Nanjayya, and Shirley Vallance (Alfred Hospital, 2012;40(1):112-118. 22. Devlin JW, Roberts RJ, Fong JJ, et al. Efficacy
Melbourne, Australia); Rinaldo Bellomo, Glenn M. 6. Dubois MJ, Bergeron N, Dumont M, et al. and safety of quetiapine in critically ill patients with
Eastwood, Leah Peck, and Helen Young (Austin Delirium in an intensive care unit. Intensive Care Med. delirium. Crit Care Med. 2010;38(2):419-427.
Hospital, Melbourne, Australia); Charlotte Firth, 2001;27(8):1297-1304. 23. Reade MC, Aitken LM. The problem of
Eileen Gilder, Mike Gillham, Lianne McCarthy, Shay definitions in measuring and managing ICU
McGuiness, and Rachael Parke (Auckland City 7. Hoy SM, Keating GM. Dexmedetomidine. Drugs.
2011;71(11):1481-1501. cognitive function. Crit Care Resusc. 2012;14(3):
Hospital, Auckland, New Zealand); Natalie Brook, 236-243.
Elisha Fulton, Amy Harney, Frank van Haren, Katie 8. Ely EW, Inouye SK, Bernard GR, et al. Delirium in
Milburn, and Helen Rogers (Canberra Hospital, mechanically ventilated patients. JAMA. 2001;286 24. Haenggi M, Blum S, Brechbuehl R, et al. Effect
Canberra, Australia); Seton Henderson, David (21):2703-2710. of sedation level on the prevalence of delirium
Knight, Jan Mehrtens, Patrick Seigne, and Duncan when assessed with CAM-ICU and ICDSC. Intensive
9. Devlin JW, Boleski G, Mlynarek M, et al. Motor Care Med. 2013;39(12):2171-2179.
Sugden (Christchurch Hospital, Christchurch, New Activity Assessment Scale. Crit Care Med. 1999;27
Zealand); Andrew Bersten, Alana St John, Kate (7):1271-1275. 25. Patel SB, Poston JT, Pohlman A, et al. Rapidly
Schwartz, and Amy Waters (Flinders Medical reversible, sedation-related delirium versus
Centre, Adelaide, Australia); Allison Bone, Tania 10. Therapeutic Goods Administration. persistent delirium in the intensive care unit. Am J
Elderkin, Melissa Fraser, Tania Salerno, and Nicholas Dexmedetomidine Hydrochloride (Precedex) Respir Crit Care Med. 2014;189(6):658-665.
Simpson (Geelong Hospital, Geelong, Australia); Product Information. 6.0 ed. Canberra, Australia:
Therapeutic Goods Administration; 2014.

1468 JAMA April 12, 2016 Volume 315, Number 14 (Reprinted) jama.com

Copyright 2016 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Associação de Medicina Intensi User on 04/16/2016