DEXMEDETOMIDINA DELIRIO

Anaesthesia 2018
doi:10.1111/anae.14472
Review Article
The effect of dexmedetomidine on delirium and agitation in
patients in intensive care: systematic review and metaanalysis with trial sequential analysis
K. T. Ng,1 C. J. Shubash2 and J. S. Chong2
1 Medical Officer (Doctor), Department of Anaesthesiology, Faculty of Medicine, University of Malaya, Jalan University,
Kuala Lumpur, Malaysia
2 Medical Student, International Medical University, Bukit Jalil, Kuala Lumpur, Malaysia
Summary
Delirium is common in intensive care patients. Dexmedetomidine is increasingly used for sedation in this
setting, but its effect on delirium remains unclear. The primary aim of this review was to examine whether
dexmedetomidine reduces the incidence of delirium and agitation in intensive care patients. We sought
randomised clinical trials in MEDLINE, EMBASE, PubMed and CENTRAL from their inception until June 2018.
Observational studies, case reports, case series and non-systematic reviews were excluded. Twenty-five trials
including 3240 patients were eligible for inclusion in the data synthesis. In the patients who received
dexmedetomidine (eight trials, 1425 patients), delirium was reduced, odds ratio (95%CI) 0.36 (0.26–0.51),
p < 0.001 and high quality of evidence. The use of dexmedetomidine was associated with a reduced incidence
of agitation, OR (95%CI) 0.34 (0.20–0.59), p < 0.001, moderate quality of evidence. Patients who were randomly
assigned to dexmedetomidine had a significantly higher incidence of bradycardia, OR (95%CI) 2.18 (1.46–
3.24), p < 0.001, moderate quality of evidence; and hypotension, OR (95%CI) 1.89 (1.48–2.41), p < 0.001, high
quality of evidence. We found no evidence of an effect on mortality, OR (95%CI) 0.86 (0.66–1.10), p = 0.23,
moderate quality of evidence. The trial sequential analyses for the incidence of delirium, bradycardia and
hypotension was conclusive but not for the incidence of agitation and mortality. In summary, this meta-analysis
suggests that dexmedetomidine reduces the incidence of delirium and agitation in intensive care patients. The
general quality of evidence ranged from moderate to high.
.................................................................................................................................................................
Correspondence to: K. T. Ng
katingng1@gmail.com
Accepted: 10 September 2018
Keywords: anxiety; delirium; dexmedetomidine; intensive care units; mortality
Introduction
effective sedative agent which does not compromise the
Delirium is common in patients in the intensive care unit
respiratory system would be an advantage [1, 9].
(ICU). It ranges from 20% in non-intubated ICU patients to as
There has been considerable interest in using
high as 83% in those undergoing mechanical ventilation [1,
dexmedetomidine to reduce the incidence of delirium in
2]. It is an independent predictor of mortality and
ICU
patients.
Unlike
benzodiazepines
and
opioids,
prolonged duration of ICU stay, which in turn is associated
dexmedetomidine is a highly selective and potent a2-
with
cognitive
adrenergic agonist [8]. It has a potential role as a sedative
significant
morbidity
and
long-term
impairment after ICU discharge [3–7]. Many sedative agents
agent as it offers the prospect of inducing a calm yet easily
such as opiates, benzodiazepines and anticholinergics
rousable state in ICU patients without causing respiratory
predispose to delirium in these patients [8]. Thus, a safe and
depression [10, 11]. In addition, it has anxiolytic, analgesic
© 2018 Association of Anaesthetists
1
Anaesthesia 2018
Ng et al. | Dexmedetomidine in intensive care patients
and anti-inflammatory properties, and may modulate the
dates until June 2018. All RCTs that investigated
stress response to illness [12, 13]. Dexmedetomidine has
dexmedetomidine vs. placebo in the ICU were sought for
also been shown to have organ-protective effects
this review, regardless of measured outcomes. Trial
(neuroprotection, cardioprotection and renoprotection) in
registers (ClinicalTrials.gov, World Health Organisation
non-cardiac and cardiac surgery [14, 15]. Favourable clinical
International Clinical Trials Registry Platform, International
outcomes of dexmedetomidine in ICU patients are
Standard Randomised Controlled Trial Number Registry)
supported by several systematic reviews and meta-analyses,
were searched to identify unpublished and ongoing
and include a lower incidence of delirium, and shorter
studies. Inclusion criteria were: adults (≥ 18 years old);
durations of ICU stay and mechanical ventilation [11, 16–
medical or surgical ICUs; receiving dexmedetomidine or
21]. However, most reviews were performed to compare
placebo; and RCTs only. Observational studies, case-
dexmedetomidine with other sedatives (benzodiazepines
controlled studies, case reports, case series, non-systematic
and opioids); such sedatives may actually increase the
reviews and trials published as abstracts were excluded.
incidence of delirium and thus introduce significant
Studies comparing dexmedetomidine with other sedative
response bias [16–18, 20]. In recent years, several multi-
agents
centre randomised controlled trials (RCT) comparing
propofol) were also excluded.
(opiates,
benzodiazepines,
midazolam
and
dexmedetomidine with placebo/control were published
The search strategy and terms used are provided in the
with conflicting findings, and a meta-analysis is due [22–24].
Supporting Information (Table S2). Articles not written in
To date, the efficacy and safety of dexmedetomidine for the
the English language were included if the journal provided
reduction in delirium among ICU patients remain unclear in
an English-translated version. The bibliographies of
the literature.
included papers and relevant systematic reviews were
The primary aim of this systematic review and meta-
hand-searched for additional papers. Experts and authors
analysis with trial sequential analysis was to determine
of papers identified in the search strategy were contacted at
whether the use of dexmedetomidine reduces the
least twice if data were missing.
incidence of delirium and agitation in ICU patients. We
Titles and abstracts were independently screened
also aimed to examine the effect of dexmedetomidine
against eligibility criteria by two authors (CS and JC). If both
on mortality and the duration of hospital stay, and
authors (CS and JC) were confident that a study was
explore its adverse effects, namely bradycardia and
unsuitable, based on the title and abstract, this study was
hypotension.
excluded. Any disagreements or different opinions at this
stage were resolved by the third author (KN) via a group
Methods
discussion. Selection of the final included articles was based
This meta-analysis was conducted and reported according
on consensus between all the three reviewers (CS, JC and
to the `Preferred Reporting Items for Systematic Review and
KN).
Meta-analysis’ (PRISMA) statement 2015 [25]. The review
All the included RCTs were assessed for risk of bias
protocol was pre-registered on PROSPERO in July 2017.
by two authors (JC and CS) independently using the
However, the protocol was updated on 12 April 2018 to
Cochrane Collaboration Risk of Bias Assessment Tool
change some of the original primary outcomes (length of
(https://handbook.cochrane.org). If their opinion was split,
ICU/hospital stay and duration of ventilation) to secondary
the article was discussed with the third author (KN) until
outcomes as those outcomes were influenced by many
consensus was achieved. In addition to the measures of
confounding factors. The review process was only started
outcome, the following fields were extracted: citation;
after approval of the updated protocol. The research
country; population; setting; study designs; total number of
questions were formulated using a population, intervention,
patients; dosage of dexmedetomidine; and type of ICU
comparison
see
(surgical or medical). When the reported values were
Supporting Information (Table S1). Primary outcomes were
and
outcomes
(PICO)
approach;
presented as median (IQR), they were converted to mean
incidence of ICU delirium and agitation. Secondary
(SD) [26].
outcomes were: all-cause mortality; length of hospital and
The GRADE assessments of the evidence and summary
ICU stay; time to extubation; duration of ventilation; and the
of findings were independently performed by two authors
adverse
(KN
effects
of
dexmedetomidine
(incidence
of
bradycardia and hypotension).
and
CS)
using
the
GRADEpro/GDT
software
(https://gradepro.org/). Based on the Cochrane Handbook,
The databases Ovid Medline, Embase, PubMed and
we downgraded a starting rating of ‘high quality’ evidence
CENTRAL were systematically searched from their start
of RCT based on the five criteria (risk of bias, inconsistency,
2
© 2018 Association of Anaesthetists
Ng et al. | Dexmedetomidine in intensive care patients
Anaesthesia 2018
indirectness, imprecision and publication bias) by one level
authors reached a consensus after the discussion with the
for serious concern or by two levels for very serious
third author (KN) to exclude any studies about the use of
concerns [27]. Any disagreements were resolved by the
regional anaesthesia with dexmedetomidine. During the
third author (JC).
full-text screening, the third author (KN) was involved in the
Statistical analyses were undertaken using Review
discussion to include Yang [34] as the recovery area is the
Manager version 5.3 (The Cochrane Collaboration,
ICU for the authors’ hospital. After applying inclusion and
Copenhagen, Denmark). Analysis of funnel plots for primary
exclusion criteria, 25 RCTs with a total of 3240 participants
and secondary outcomes (incidence of delirium, agitation,
were included in this meta-analysis. The study design and
mortality, bradycardia and hypotension) was conducted to
clinical characteristics of all included RCTs are illustrated in
assess the risk of publication bias [28–30]. The I2 test was
the Supporting Information Table S3. Details of the
used to assess the heterogeneity of studies; values of less
excluded studies are outlined in the Supporting Information
than 40%, 40–60% and more than 60% were used to
(Table S4). Searching the clinical trials registries identified
determine low, moderate and substantial heterogeneity,
eight relevant ongoing studies; see Supporting Information
respectively [27, 31]. A two-sided p value of < 0.05 was
(Table S5).
considered to denote the statistical significance of
Fourteen studies [35–48] were single-centre RCTs,
heterogeneity. Findings were reported as odds ratios (OR)
whereas 11 studies [22–24, 34, 49–55] were multi-centre
or mean difference (MD) with 95%CI. If no substantial
RCTs. Most studies examined the prophylactic effect of
heterogeneity was noted, a fixed-effect model analysis
dexmedetomidine on the reduction of delirium in ICU
(Mantel–Haenszel method) was used to pool estimates. If
patients, except two trials [35, 49] which investigated the
evidence of substantial heterogeneity (I2 > 60%) was
therapeutic effect of dexmedetomidine. Twenty studies [22,
observed, a random-effects model analysis (DerSimonian–
23, 34, 36–48, 51, 53–55] were conducted in surgical ICU,
Laird method) was used. To investigate the presence of
three [35, 50, 52] were medical ICU and two [24, 49] were
small study effects on the measured outcomes with
based on a combination of medical and surgical ICU
substantial heterogeneity (I2 > 60%), both fixed and
patients. The dexmedetomidine regimen varied between
random effects models were compared. In addition,
studies, with most adopting a goal-directed approach to
sensitivity analyses were performed by analysing only
titrate the dosage of dexmedetomidine to meet a targeted
studies at a low risk of bias and by sequentially removing
sedation level. In terms of risk of bias assessment, most of
each study (from the most recent trials) and re-analysing the
the included RCTs were at low risk (overall), with the
remaining dataset of those outcomes with substantial
exception of six trials [38–40, 44, 50, 54] rated as high risk of
heterogeneity. For measured outcomes with zero events in
bias (overall) due to inadequate sequence generation and
either arm, we adhered to the guidance of the Cochrane
allocation concealment (Table 1). The third author (KN) was
Handbook (16.9.3) by using an OR-based method as it
involved in the discussion with regard to the potential
excludes those reporting bias whether or not they are
conflict of interest (funding) in all the included RCTs. The
published [32].
summary of findings/GRADE assessment of quality of
Using the trial sequential analysis viewer version 0.9.5.5
Beta (Copenhagen Trial Unit, 2016), trial sequential analysis
evidence and PRISMA checklist are shown in Table 2 and in
the Supporting Information Table S6, respectively.
was performed on the primary and secondary outcomes to
Our two primary outcomes were author-defined
prevent the risk of random error and multiplicity
incidence of ICU delirium and agitation based on different
phenomenon due to repeated significance testing in meta-
screening tools (Intensive Care Delirium Screening Checklist
analyses [33]. The required meta-analysis information size
(ICDSC),
and adjusted significance thresholds were calculated based
Confusion Assessment Method for Intensive Care Unit
on a two-sided sequential analysis-adjusted fixed effects
(CAM-ICU), Riker Sedation–Agitation Scale). For all-cause
model with 5% risk of type-1 error and power of 80%.
mortality, the longest follow-up data ICU, in-hospital, 30-day
Richmond
Agitation–Sedation
Scale
(RASS),
stay were selected for data analysis. Authors’ definitions of
Results
bradycardia included different threshold of low heart rate
The results of the literature search and study selection
(< 60 beats.min 1, < 45 beats.min 1, < 55 beats.min 1) or
process are outlined in the PRISMA flow chart (Fig. 1). The
a decrease of more than 20–30% from the baseline heart
titles and abstracts of 8251 non-duplicate articles were
rate. Definitions of hypotension included different threshold
screened, of which 58 articles were retrieved for full-text
of low systolic blood pressure (< 95 mmHg, < 90 mmHg,
review. In the stage of titles and abstracts screening, all the
< 80 mmHg) or a decrease of more than 20–30% from the
© 2018 Association of Anaesthetists
3
Anaesthesia 2018
Ng et al. | Dexmedetomidine in intensive care patients
MEDLINE
Cochrane database of systematic reviews
EMBASE
PUBMED
633 citation(s)
677 citation(s)
6,007 citation(s)
934 citation(s)
8,251 non-duplicate
989 duplicates removed
citations screened
7,262 articles for title
and abstract screening
inclusion/exclusion
criteria applied
7,204 articles excluded
after title/Abstract screen
58 articles
retrieved
inclusion/exclusion
criteria applied
33 articles excluded
after full text screen
25 articles
included
Figure 1 PRISMA flow diagram.
baseline systolic blood pressure. Studies which did not
based on 5% risk of type-1 error (two-sided), power 80%,
report on any of these primary and secondary outcomes
low bias-based relative risk reduction of 59.31% and
were excluded at this stage.
incidence in control arm of 18.5% with a model variance-
We combined eight trials [22–24, 34, 37, 41, 43, 52]
based heterogeneity correction (Fig. 3). The cumulative
(1425 patients) in the analysis of delirium. The incidence was
z-curve (blue line) crossed the boundary of required
7.6% in the dexmedetomidine group and 18.5% in the
information size and trial sequential analysis monitoring
placebo group. Patients who had dexmedetomidine had a
boundary for favouring dexmedetomidine. Thus, this
significant reduction in the incidence of delirium, OR (95%
pooled analysis is conclusive that dexmedetomidine
CI) 0.36 (0.26–0.51), p < 0.001, I = 0% high quality of
reduces incidence of delirium in ICU patients.
2
evidence and conclusive trial sequential analysis. Statistical
Dexmedetomidine reduced the incidence of agitation
heterogeneity was assessed as low in the pooled effect
based on the combined data of four RCTs [34,43,48,55], (a
(Fig. 2). The trial sequential analysis of a diversity-adjusted
total of 718 patients, OR (95%CI) 0.34 (0.20–0.59),
required information size for delirium was 296 patients,
p < 0.001, I2 = 42%; moderate quality of evidence and an
4
© 2018 Association of Anaesthetists
Low
High
Low
High
High
Low
Low
Low
High
Low
Low
Low
Low
Low
High
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Venn et al. [51]
Triltsch et al. [54]
Martin et al. [55]
Wahlander et al. [40]
Akin et al. [39]
Menda et al. [46]
Leino et al. [53]
Khalil et al. [47]
Abdel-Meguid et al. [38]
Devlin et al. [52]
Priye et al. [37]
Bielka et al. [35]
Yang et al. [34]
Balkanay et al. [43]
Guo et al. [44]
Reade et al. [49]
Wu et al. [41]
Cho et al. [42]
Su et al. [22]
Cheng et al. [45]
Zhao et al. [48]
Kawazoe et al. [50]
Li et al. [23]
Skrobik et al. [24]
Bielka et al. [36]
Sequence
generation
Study
Domain
© 2018 Association of Anaesthetists
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
High
Low
Low
Low
Low
Low
High
Low
Low
Low
High
High
Low
High
Low
Allocation
concealment
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
High (Open-label)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Low (Double-blind)
Blinding of participants
and personnel
Table 1 Summary of the risk of bias assessment for the included studies.
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Blinding of outcome
assessment
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Incomplete
outcome data
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Selective outcome
reporting
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Low
Other sources
of bias
Low
Low
Low
High
Low
Low
Low
Low
Low
Low
High
Low
Low
Low
Low
Low
High
Low
Low
Low
High
High
Low
High
Low
Overall
Ng et al. | Dexmedetomidine in intensive care patients
Anaesthesia 2018
5
6
Not serious b
Not serious b
Not serious b
Not serious
Not serious
Risk of bias
Not serious
Not serious
Not serious
Not serious
Not serious
Inconsistency
Not serious
Not serious
Serious c
Not serious
Not serious
Indirectness
Not serious
Not serious
Not serious
Not serious
Not serious
Imprecision
None
Publication
bias strongly
suspected a
None
Publication
bias strongly
suspected a
None
Publication
bias
⨁⨁⨁⨁
HIGH
⨁⨁⨁◯
MODERATE
⨁⨁⨁◯
MODERATE
⨁⨁⨁◯
MODERATE
⨁⨁⨁⨁
HIGH
Overall
certainty of
evidence
b
Funnel plot suggests publication bias.
The effect estimates are unchanged when conducting a sensitivity analysis with only low risk of bias studies.
c
Some patients were admitted to intensive care units after major surgery.
a
2568
(18 RCTs)
Incidence of
hypotension
2501
(17 RCTs)
Incidence of
bradycardia
2426
(15 RCTs)
Mortality
718
(4 RCTs)
Incidence of
agitation
1425 (8 RCTs)
Incidence of
delirium
No. of
participants
(studies)
follow-up
GRADE assessment
Dexmedetomidine vs. placebo in ICU
Table 2 Summary of findings table with GRADE assessment quality of evidence.
115/1255
(9.2%)
36/1230
(2.9%)
176/1202
(14.6%)
53/341
(15.5%)
129/698
(18.5%)
Placebo
218/1313
(16.6%)
80/1271
(6.3%)
158/1224
(12.9%)
26/377
(6.9%)
55/727
(7.6%)
Dexmedetomidine
Study event rates (%)
Summary of findings
1.89
(1.48–2.41)
2.18
(1.46–3.24)
0.86
(0.66–1.10)
0.34
(0.20–0.59)
0.36
(0.26–0.51)
Odds
ratio
(95%CI)
92 per
1000
29 per
1000
146 per
1000
155 per
1000
185 per
1000
Risk with
placebo
absolute
68 more
per 1000
(38 more to
104 more)
32 more
per 1000
(13 more
to 60 more)
18 fewer
per 1000
(45 fewer to
12 more)
97 fewer
per 1000
(120 fewer
to 57 fewer)
109 fewer
per 1000
(129 fewer
to 81 fewer)
Risk
difference
with dexmedetomidine
Anticipated
effects
Anaesthesia 2018
Ng et al. | Dexmedetomidine in intensive care patients
© 2018 Association of Anaesthetists
Ng et al. | Dexmedetomidine in intensive care patients
Anaesthesia 2018
Figure 2 Forest plot of the incidence of delirium. MH, Mantel-Haenszel
Figure 3 Trial sequential analysis of the incidence of delirium.
inconclusive trial sequential analysis). The incidence of
relative risk reduction of 57.58%. The trial sequential
agitation in patients was 6.8% in the dexmedetomidine
analysis diversity-adjusted required information size for
group and 15.5% in the placebo group. Statistical
agitation would be 741 randomly allocated participants.
heterogeneity was moderate across the studies (Fig. 4). At
Although the cumulative z-curve (blue line) crossed the trial
this stage of review, with 718 patients, only 96.9% of the
sequential analysis monitoring for benefit (red line), it is still
required information size is available to detect or reject a
inconclusive that dexmedetomidine reduces agitation in
© 2018 Association of Anaesthetists
7
Anaesthesia 2018
Ng et al. | Dexmedetomidine in intensive care patients
ICU patients as it did not achieve the required information
p < 0.001, I2 = 28%, moderate quality of evidence and
size (Fig. 5).
conclusive trial sequential analysis (Fig. 6, Fig. S7); and
Based on data from 15 RCTs [22–24, 35, 36, 40–42, 47,
hypotension (OR 95%CI 1.86 (1.23–2.80), p = 0.003,
49–52, 54, 55], with a total of 2426 ICU patients, we found no
I2 = 47%, high quality of evidence and conclusive trial
evidence of a decrease in mortality in the dexmedetomidine
sequential analysis) (Fig. 7, Fig. S8).
group compared with placebo OR (95%CI) 0.86 (0.66–1.10),
Sensitivity analyses were performed on the secondary
p = 0.23, I2 = 0%; moderate quality of evidence and
outcomes length of hospital/ICU stay and time to extubation
inconclusive trial sequential analysis), see Supporting
and the effect estimate remained unchanged. The
Information (Fig. S1). Currently, with 2426 patients, only
sequential removal of the most recent trials and re-analysing
8.3% of the required information size to detect or reject a
the remaining dataset detected major changes in the
relative risk reduction of 9.19% is actually available at this
direction and magnitude of statistical findings for length of
stage. The required information size for a conclusive meta-
hospital/ICU stay, time to extubation and duration of
analysis would be 29,134 patients; see Supporting
mechanical ventilation, which indicated the potential
Information (Fig. S2).
evidence of between-study heterogeneity among all the
In comparison with placebo, patients who were
included trials. The funnel plots for delirium, mortality and
randomly assigned to receive dexmedetomidine spent less
hypotension did not reveal important asymmetry in the
time in hospital, with a mean difference (95%CI) of
Supporting
0.87 day ( 1.42 to
Information
(Figs. S9,
S10
and
S11,
0.32), p = 0.002, I2 = 99%); and in
respectively). However, the funnel plots for agitation and
0.08), p < 0.001,
bradycardia were asymmetrical, indicating potential risk of
I2 = 98%). However, the statistical heterogeneity for both
publication bias; see Supporting Information (Figs. S12 and
measured outcomes was very high; see Supporting
S13).
ICU, MD (95%CI)
0.11 day ( 0.14 to
Information (Figs. S3 and S4).
Eight trials [22, 36, 38, 47, 49, 53–55] with 1090 patients
Discussion
indicated a shorter time to extubation in patients who
Our
received dexmedetomidine, MD (95%CI) 1.74 h ( 2.64 to
dexmedetomidine reduces the incidence of delirium and
0.84), p < 0.001, I2 = 100%). The test of heterogeneity was
agitation in ICU patients when compared with placebo. It
high in the pooled effect; see Supporting Information
also shortens the duration of ICU and hospital stay,
(Fig. S5).
duration of mechanical ventilation and time to tracheal
Six RCTs [23, 24, 46, 50–52] including 747 patients
meta-analysis
extubation.
Even
has
though
demonstrated
dexmedetomidine
that
was
showed that dexmedetomidine reduced the duration of
associated with a higher incidence of bradycardia and
mechanical ventilation, MD (95%CI)
0.20 days ( 0.36 to
hypotension, no significant effects on mortality were
0.05), p = 0.01, I2 = 93%); see Supporting Information
detected. The general quality of evidence ranged from
(Fig. S6).
Seventeen studies (2501 patients) [22–24, 34–36, 39–
41, 43, 45, 46, 48, 50, 52, 54, 55] and 18 studies (2568
moderate to high, although we strongly suspected
publication bias and serious indirectness in some of the
measured outcomes.
patients) [22–24, 34–36, 39–41, 43–46, 48, 51, 52, 54, 55]
To the best of our knowledge, this is the first meta-
reported the incidence of bradycardia and hypotension,
analysis to compare dexmedetomidine with placebo in ICU
respectively. Patients who received dexmedetomidine were
patients. Many meta-analyses have demonstrated a
more likely to be bradycardic, OR (95%CI) 2.18 (1.46–3.24),
significant reduction in the incidence of delirium in ICU
Figure 4 Forest plot of the incidence of agitation. MH, Mantel-Haenszel
8
© 2018 Association of Anaesthetists
Ng et al. | Dexmedetomidine in intensive care patients
Anaesthesia 2018
Figure 5 Trial sequential analysis of the incidence of agitation.
Figure 6 Forest plot of incidence of bradycardia. MH, Mantel-Haenszel
© 2018 Association of Anaesthetists
9
Anaesthesia 2018
Ng et al. | Dexmedetomidine in intensive care patients
Figure 7 Forest plot of incidence of hypotension. MH, Mantel-Haenszel
patients who received dexmedetomidine instead of
the included RCTs may have introduced variances to our
benzodiazepines or opiates [16–18, 20]. It is believed that
findings.
the modulation of gamma-aminobutyric acid (GABA) type-
Patients who received dexmedetomidine had a
A receptors by both benzodiazepines and opioids induces
shorter duration of ventilation, time to extubation and
delirium in critically ill patients [7, 22]. However, it is unclear
length of ICU/hospital stay, which corresponds with
whether these ICU patients gained delirium-sparing
previous meta-analyses [16–21]. This may be due to
benefits from the prophylactic effect of dexmedetomidine
dexmedetomidine reducing the incidence of delirium
or by avoiding those delirium-inducing sedatives and
and its complications. However, our review does not
analgesics. In this review, we have addressed this
provide evidence of a causal relationship between the
distinction.
use
The incidence of delirium varies widely from 16.1% to
of
dexmedetomidine
and
delirium.
Many
confounding factors, such as: different populations of
83.3% in ICU patients, especially among those undergoing
ICUs
mechanical ventilation (60–80%) [56, 57]. The variable
dexmedetomidine (goal-directed vs. fixed rate); and total
incidence
baseline
dose of dexmedetomidine, may have influenced the
characteristics; types of ICU (medical, surgical or both);
duration of ICU and hospital stay in all the included
illness severity; screening tools for diagnosis of delirium;
RCTs. Given the high degree of heterogeneity across
and concomitant medications [56, 58]. Several international
studies for all the aforementioned outcomes, their
studies have suggested a strong association between
pooled estimate effects need to be interpreted with
delirium and short-term/long-term mortality [1, 58–61]. In
caution before making any generalisation on the clinical
keeping with previous observational studies [62, 63], the
benefits of dexmedetomidine to all ICU patients. The
incidence of delirium in our review was 18.5% in the control
nature and clinical characteristics of medical and surgical
group and 7.6% in the dexmedetomidine group, the
ICU patients varied and different centres have their own
difference being statistically significant. Nevertheless, the
criteria for ICU admission. Future trials focusing on the
standard of care and choices of concomitant medications
use
(benzodiazepine or anticholinergic) for critically ill patients
conditions may provide better information with a lesser
varied among different centres. For instance, Su et al.
degree of heterogeneity.
is
due
to:
different
patient
(surgical
of
vs.
medical);
dexmedetomidine
in
more
regimens
of
homogeneous
limited the use of anticholinergics and benzodiazepines,
Our review confirmed that dexmedetomidine was
and their primary sedative was propofol, all of which may
associated with a higher incidence of bradycardia and
have decreased the control group’s incidence of ICU
hypotension, but no significant differences in mortality
delirium [22]. Different delirium assessment tools used in
were demonstrated. The definition of hypotension and
10
© 2018 Association of Anaesthetists
Ng et al. | Dexmedetomidine in intensive care patients
Anaesthesia 2018
bradycardia used in each study was not standardised
References
and it may either over- or under-estimate the adverse
1. Ely EW, Inouye SK, Bernard GR, et al. Delirium in mechanically
ventilated patients: validity and reliability of the confusion
assessment method for the intensive care units (CAM-ICU).
Journal of the American Medical Association 2001; 286: 2703–
10.
2. Van Rompaey B, Schuurmans MJ, Shortridge-Baggett LM,
Truijen S, Elseviers MBL. A comparison of the CAM-ICU and the
NEECHAM Confusion Scale in intensive care delirium
assessment: an observational study in non-intubated patients.
Critical Care Medicine 2008; 12: R16.
3. Thomason JW, Shintani A, Peterson JF, et al. Intensive care unit
delirium is an independent predictor of longer hospital stay: a
prospective analysis of 260 non-ventilated patients. Critical
Care 2005; 9: R375–R381.
4. Salluh JI, Soares M, Teles JM, et al. Delirium epidemiology in
critical care (DECCA): an international study. Critical Care 2010;
14: R210.
5. Van Rompaey B, Elseviers MM, Schuurmans MJ, et al. Risk
factors for delirium in intensive care patients: a prospective
cohort study. Critical Care 2009; 13: R77.
6. Choi JG. Delirium in the intensive care unit. Korean Journal of
Anesthesiology 2013; 65: 195–202.
7. Shelton KT, Qu J, Bilotta F, et al. Minimizing ICU neurological
dysfunction with dexmedetomidine- induced sleep (MINDDS):
protocol for arm, placebo-controlled trial. British Medical
Journal Open 2018; 8: 1–8.
8. Ibrahim K, Mccarthy CP, Mccarthy KJ, et al. Delirium in the
cardiac intensive care unit. Journal of the American Heart
Association 2018: 7: 1–12.
9. Cavallazzi R, Saad MMP. Delirium in the ICU: an overview.
Annals of Intensive Care 2012; 2: 49.
10. Iirola T, Aantaa R, Laitio R, et al. Pharmacokinetics of prolonged
infusion of high-dose dexmedetomidine in critically ill patients.
Critical Care 2011; 15: R257.
11. Pasin L, Greco T, Feltracco P, et al. Dexmedetomidine as a
sedative agent in critically ill patients: a meta-analysis of
randomised controlled trials. PLoS ONE 2013; 8: e82913.
12. Guo TZ, Jiang JY, Buttermann AE, Maze M. Dexmedetomidine
injection into locus ceruleus produces antinociception.
Anesthesiology 1996; 84: 873–81.
13. Aantaa R, Kanto J, Scheinin M, Kallio A, Scheinin H.
Dexmedetomidine, an alpha 2-adrenoceptor agonist reduces
anesthetic requirements for patients undergoing minor
gynecologic surgery. Anesthesiology 1990; 73: 230–5.
14. Biccard BM, Goga S, Buers JD. Dexmedetomidine and cardiac
protection for non-cardiac surgery: a meta-analysis of
randomised controlled trials. Anaesthesia 2008; 63: 4–14.
15. Wijeysundera DN, Naik JS, Beattie WS. Alpha-2 adrenergic
agonists
to
prevent
perioperative
cardiovascular
complications: a meta-analysis. American Journal of Medicine
2003; 114: 742–52.
16. Xia ZQ, Chen SQ, Yao X, Xie CB, Wen SHLK. Clinical benefits of
dexmedetomidine versus propofol in adult intensive care unit
patients: a meta-analysis of randomised clinical trials. Journal
of Surgical Research 2013; 185: 833–43.
17. Constantin JM, Momon A, Mantz J, et al. Efficacy and safety of
sedation with dexmedetomidine in critical care patients: a
meta-analysis of randomized controlled trials. Anaesthesia
Critical Care and Pain Medicine 2016; 35: 7–15.
18. Pasin L, Landoni G, Nardelli P, et al. Dexmedetomidine
reduces the risk of delirium, agitation and confusion in critically
ill patients: a meta-analysis of randomized controlled trials.
Journal of Cardiothoracic and Vascular Anesthesia 2014; 28:
1459–66.
19. Chen K, Lu Z, Xin YC, Cai Y, Chen Y, Pan SM. Alpha-2 agonists
for long-term sedation during mechanical ventilation in
effects of dexmedetomidine in ICU patients. Our current
meta-analysis was underpowered to detect a statistically
significant
reduction
in
mortality
as
the
required
information size of 29,134 subjects for mortality was not
achieved. In the ICU setting, it is not recommended that
a
loading
dose
of
dexmedetomidine
should
be
administered for sedation as many ICU patients are
haemodynamically unstable. [64] However, five RCTs [39,
40,
47,
51,
54]
gave
a
loading
dose
of
dexmedetomidine within the range of 0.5–6 lg.kg
1
before continuing with the infusion of dexmedetomidine
at a rate of 0.1–0.7 lg.kg
1
.h
1
. The different dosages
of dexmedetomidine used in RCTs may potentially affect
the
interpretation
dependent
of
adverse
findings
our
effects
of
on
the
dose-
dexmedetomidine
(bradycardia and hypotension).
Limitations of this study are common to other metaanalyses, namely: differences in population characteristics;
small
patient
numbers;
dexmedetomidine;
use
dosage
of
and
regime
concomitant
drugs;
of
and
variations in primary outcomes. In some studies, other
sedative agents (benzodiazepine and
opioid) were
administrated at the discretion of attending physicians or
intensivists. The combination of different types of sedatives
(propofol, fentanyl, midazolam, morphine, lorazepam,
pethidine, sufentanil) was used across all the included
RCTs, which may introduce variability to our findings. Thus,
it is very difficult to perform a sub-group analysis based on
the concomitant use of sedative agents due to the
combinations of different types of sedative agents in each
trial. Future trials with standardised protocols to minimise
the use of benzodiazepines and opioids could improve
inconsistency. Unfortunately, we were not able to control
the aforementioned confounding factors at this review
level. The exclusion of trials published as abstract or
presented at conferences may potentially introduce
publication bias to our findings.
In this meta-analysis of 25 RCTs, dexmedetomidine
demonstrated significant benefits in reducing the incidence
of delirium and agitation in ICU. Although it was associated
with a significantly higher incidence of bradycardia and
hypotension, no effects on mortality were noted.
Acknowledgements
This
review
was
registered
in
PROSPERO
(CRD42017072304). We thank Dr Y. Skrobik for sharing the
full text for the data analysis in this review. No external
funding and no competing interests declared.
© 2018 Association of Anaesthetists
11
Anaesthesia 2018
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
12
critically ill patients. Cochrane Database of Systematic Reviews
2015; 1: CD010269.
Liu X, Xie GH, Zhang K, et al. Dexmedetomidine vs propofol
sedation reduces delirium in patients after cardiac surgery: a
meta-analysis with trial sequential analysis of randomised
controlled trials. Journal of Critical Care 2017; 38: 190–6.
Fraser GL, Devlin JW, Worb CP, et al. Benzodiazepine versus
nonbenzodiazepine-based
sedation
for
mechanically
ventilated, critically ill adults: a systematic review and metaanalysis of randomized trials. Critical Care Medicine 2013; 41
(Suppl 1): S30–8.
Su X, Meng ZT, Wu XH, et al. Dexmedetomidine for prevention
of delirium in elderly patients after non-cardiac surgery: a
randomised, double-blind, placebo-controlled trial. Lancet
2016; 388: 1893–902.
Li X, Yang J, Nie XL, et al. Impact of dexmedetomidine on the
incidence of delirium in elderly patients after cardiac surgery: a
randomized controlled trial. PLoS ONE 2017; 12: 1–15.
Skrobik T, Duprey MS, Hill NSDJ. Low-dose nocturnal
dexmedetomidine prevents ICU delirium: a randomized,
placebo-controlled trial. American Journal of Respiratory and
Critical Care Medicine 2018; 197: 1147–56.
Shamseer L, Moher D, Clarke M, et al. Preferred reporting
items for systematic review and meta-analysis protocols
(PRISMA-P) 2015. British Medical Journal 2015; 349: g4647.
Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and
variance from the median, range, and the size of a sample. BMC
Medical Research Methodology 2005; 5: 1–10.
Ryan R, Hill S. How to GRADE the quality of the evidence.
Cochrane Consumers and Communication Group. 2016.
http://cccrg.cochrane.org/author-resources (accessed 27/11/
2017).
Sterne JAC, Sutton AJ, Ioannidis JPA, et al. Recommendations
for examining and interpreting funnel plot asymmetry in metaanalyses of randomised controlled trials. British Medical
Journal 2011; 343: d4002.
Ng KT, Yap J. Continuous infusion versus intermittent bolus
injection of furosemide in acute decompensated heart failure: a
meta-analysis of randomised controlled trials. Anaesthesia
2017; 73: 238–47.
Smith AF, Carlisle JC. Reviews, systematic reviews and
Anaesthesia. Anaesthesia 2015; 70: 644–70.
Choi SW, Lam DMH. Heterogeneity in meta-analyses.
Comparing apples and oranges? Anaesthesia 2017; 72: 532–4.
Cochrane. 16.9.3 Studies with no events. Cochrane handbook
for systematic reviews of interventions. 2017. http://handbook5-1.cochrane.org/(accessed 14/11/2017).
Thorlund K, Engstrom J, Wetterslev J, et al. User manual for trial
sequential analysis (TSA). Copenhagen Trial Unit. 2017:1–114.
http://www.ctu.dk/tools-and-links/trial-sequential-analysis.a
spx (accessed 2/01/2018).
Yang XY, Li ZZ, Gao CY, Liu RC. Effect of dexmedetomidine on
preventing agitation and delirium after microvascular free flap
surgery: a randomized, double-blind, control study. Journal of
Oral Maxillofacial Surgery 2015; 73: 1065–72.
Bielka K, Kuchyn LGF. Addition of dexmedetomidine to
benzodiazepines for patients with alcohol withdrawal
syndrome in the intensive care unit: a randomized controlled
study. Annals of Intensive Care 2015; 5: 1–7.
Bielka K, Kuchyn L, Babych V, Martycshenko K, Innozemtsev O.
Dexmedetomidine infusion as an analgesic adjuvant during
laparoscopic c holecystectomy: a randomized controlled study.
BMC Anesthesiology 2018; 18: 1–6.
Priye S, Jagannath S, Singh D, Shivaprokash SRD.
Dexmedetomidine as an adjunct in postoperative analgesia
following cardiac surgery: a randomized, double-blind study.
Saudi Journal of Anaesthesia 2015; 9: 353–8.
Ng et al. | Dexmedetomidine in intensive care patients
38. Abdel-meguid MS. Dexmedetomidine as anesthetic adjunct
for fast tracking and pain control in off-pump coronary artery
bypass. Saudi Journal of Anaesthesia 2013; 7: 6–8.
39. Akin S, Aribogan AAG. Dexmedetomidine as an adjunct to
epidural analgesia after abdominal surgery in elderly intensive
care patients: a prospective, double-blind, clinical trial. Current
Therapeutic Research 2008; 69: 16–28.
40. Wahlander
S,
Frumento
RJ,
Wagener
G,
et al.
Dexmedetomidine as an adjunct to epidural analgesia after
thoracic surgery. Journal of Cardiothoracic and Vascular
Anesthesia 2005; 19: 630–5.
41. Wu XH, Cui F, Zhang C, et al. Low-dose dexmedetomidine
improves sleep quality pattern in elderly patients after
noncardiac surgery in the intensive care unit. Critical Care
Medicine 2016; 125: 979–91.
42. Cho JS, Shim JK, Soh S, Kim MK, Kwak YL. Perioperative
dexmedetomidine reduces the incidence and severity of acute
kidney injury following valvular heart surgery. Kidney
International 2016; 89: 693–700.
43. Balkanay OO, Goksedef D, Omeroglu Sn IG. The dose-related
effects of Dexmedetomidine on renal functions and serum
neutrophil gelatinase-associated lipocalin values after coronary
artery
bypass
grafting:
a
randomized.
Interactive
CardioVascular and Thoracic Surgery 2015; 20: 209–14.
44. Guo Y, Sun L, Zhang JF, Li QF, Jiang HJW. Preventive effects of
low-dose dexmedetomidine on postoperative cognitive
function and recovery quality in elderly oral cancer patients.
International Journal of Clinical and Experimental Medicine
2015; 8: 16183–90.
45. Cheng MH, Shi JL, Gao T, et al. The addition of
dexmedetomidine to analgesia for patients after abdominal
operations: a prospective randomized clinical trial. World
Journal of Surgery 2017; 41: 39–46.
€ Sayın M, T€
46. Menda F, K€
oner O,
ure H, _Imer P, Aykacß B.
Dexmedetomidine as an adjunct to anesthetic induction to
attenuate hemodynamic response to endotracheal intubation
in patients undergoing fast-track CABG. Annals of Cardiac
Anaesthesia 2010; 13: 16–21.
47. Khalil MAAAM. The impact of dexmedetomidine infusion in
sparing morphine consumption in off-pump coronary artery
bypass grafting. Seminars in Cardiothoracic and Vascular
Anesthesia 2013; 17: 66–71.
48. Zhao LH, Shi ZH, Chen GQ, et al. Use of dexmedetomidine for
prophylactic analgesia and sedation in patients with delayed
extubation after craniotomy: a randomized controlled trial.
Journal of Neurosurgical Anesthesiology 2017; 29: 132–9.
49. Reade MC, Eastwood GM, Bellomo R, et al. Effect of
dexmedetomidine added to standard care. Journal of the
American Medical Association 2016; 315: 1460–8.
50. Kawazoe Y, Miyamoto K, Morimoto T, et al. Effect of
dexmedetomidine on mortality and ventilator-free days in
patients requiring mechanical ventilation with sepsis a
randomized clinical trial. Journal of the American Medical
Association 2017; 317: 1321–8.
51. Venn RM, Bradshaw CJ, Spencer R, et al. Preliminary UK
experience of dexmedetomidine, a novel agent for
postoperative sedation in the intensive care unit. Anaesthesia
1999; 54: 1136–42.
52. Devlin JW, Al-Qadheeb NS, Chi A, et al. Efficacy and safety of
early dexmedetomidine during noninvasive ventilation for
patients. Chest 2014; 145: 1204–12.
53. Leino K, Hynynen M, Jalonen J, et al. Renal effects of
dexmedetomidine during coronary artery bypass surgery : a
randomized placebo-controlled study. BMC Anesthesiology
2011; 11: 1–10.
54. Triltsch AE, Welte M, Homeyer P, et al. Bispectral index-guided
sedation with dexmedetomidine in intensive care: a
© 2018 Association of Anaesthetists
Ng et al. | Dexmedetomidine in intensive care patients
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
prospective, randomized, double blind, placebo-controlled
phase II study. Critical Care Medicine 2002; 30: 1007–14.
Martin E, Ramsay G, Mantz J, Sum-Ping STJ. The role of the a 2 adrenoceptor agonist dexmedetomidine in postsurgical
sedation in the intensive care unit. Journal of Intensive Care
Medicine 2003; 18: 29–41.
King JGA. Delirium in intensive care. Anaesthesia Critical Care
and Pain Medicine 2009; 9: 144–7.
Kallenbach TF, Amado LA. Assessment of delirium in the
intensive care unit. Southern African Journal of Anaesthesia
and Analgesia 2017; 23: 57–63.
Lee A, Mu JL, Joynt GM, Chiu CH, Lai VKW, Gin T. Underwood
MJ. Risk prediction models for delirium in the intensive care
unit after cardiac surgery: a systematic review and independent
external validation. British Journal of Anaesthesia 2017; 118:
391–9.
McCusker J, Cole M, Abrahamowicz M, Primeau FBE. Delirium
predicts 12-month mortality. Journal of the American Medical
Association 2002; 162: 457–63.
Klein Klouwenberg PM, Zaal IJ, Spitoni C, et al. The attributable
mortality of delirium in critically ill patients: prospective cohort
study. British Medical Journal 2014; 349: g6652.
Gottesman RF, Grega MA, Bailey MM, et al. Delirium after
coronary artery bypass graft surgery and late mortality. Annals
of Neurology 2010; 67: 338–44.
Pauley E, Lishmanov A, Schumann S, et al. Delirium is a robust
predictor of morbidity and mortality among critically ill patients
treated in the cardiac intensive care unit. American Heart
Journal 2015; 170: 79–86.
Falsini G, Grotti S, Toccafondi G, et al. Long-term prognostic
value of delirium in elderly patients with acute cardiac diseases
admitted to two cardiac intensive care units: a prospective
study (DELIRIUM CORDIS). European Heart Journal: Acute
Cardiovascular Care 2017; 1: 204887261769523.
Scott-Warren VL, Sebastian J. Dexmedetomidine: its use in
intensive care medicine and anaesthesia. British Journal of
Anaesthesia 2016; 16: 242–6.
© 2018 Association of Anaesthetists
Anaesthesia 2018
Supporting Information
Additional supporting information may be found online in
the Supporting Information section at the end of the article.
Figure S1. Forest plot of mortality.
Figure S2. TSA of mortality.
Figure S3. Forest plot of length of hospital stay.
Figure S4. Forest plot of length of ICU stay
Figure S5. Forest plot of time to extubation.
Figure S6. Forest plot of duration of mechanical
ventilation.
Figure S7. TSA of incidence of bradycardia.
Figure S8. TSA of incidence of hypotension.
Figure S9. Funnel plot of delirium.
Figure S10. Funnel plot of mortality.
Figure S11. Funnel plot of hypotension.
Figure S12. Funnel plot of agitation.
Figure S13. Funnel plot of bradycardia.
Table S1. PICO table.
Table S2. Search strategy.
Table S3. Clinical characteristics of included studies.
Table S4. Characteristics of excluded studies.
Table S5. Clinical characteristics of ongoing studies.
Table S6. PRISMA checklist.
13