ww.sciencedirect.com

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3

Available online at w

ScienceDirect

Clinical benefits of dexmedetomidine versus propofol in adultintensive care unit patients: a meta-analysis of randomizedclinical trials

journal homepage: www.JournalofSurgicalResearch.com

Zhi-Qiu Xia, MD,a,1 Shu-Qin Chen, PhD, MD,b,1 Xi Yao, MD,a Chuan-Bo Xie, MD,c

Shi-Hong Wen, MD,a and Ke-Xuan Liu, PhD, MDa,*aDepartment of Anesthesiology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, ChinabDepartment of Gynecology and Obstetrics, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, ChinacDepartment of Biostatistics and Epidemiology, School of Public Health, Sun Yat-sen University, Guangzhou, China

a r t i c l e i n f o

Article history:

Received 12 May 2013

Received in revised form

24 June 2013

Accepted 26 June 2013

Available online 24 July 2013

Keywords:

Intensive care unit

Sedation

Outcomes

Dexmedetomidine

Propofol

* Corresponding author. Department of AnesRoad, Guangzhou, China 510080. Tel.: þ86 2

E-mail addresses: liukexuan705@163.com1 These authors equally contributed to thi

0022-4804/$ e see front matter ª 2013 Elsevhttp://dx.doi.org/10.1016/j.jss.2013.06.062

a b s t r a c t

Background: This meta-analysis was performed to assess the influence of dexmedetomidine

and propofol for adult intensive care unit (ICU) sedation, with respect to patient outcomes

and adverse events.

Materials and methods: A systematic review was conducted of all randomized controlled

trials exploring the clinical benefits of dexmedetomidine versus propofol for sedation in

adult intensive care patients. The primary outcomes of this study were length of ICU

stay, duration of mechanical ventilation, and risk of ICU mortality. Secondary outcomes

included risk of delirium, hypotension, bradycardia and hypertension.

Results: Ten randomized controlled trials, involving 1202 patients, were included. Dexme-

detomidine significantly reduced the length of ICU stay by <1 d (five studies, 655 patients;

mean difference, �0.81 d; 95% confidence interval [CI], �1.48 to �0.15) and the incidence of

delirium (three studies, 658 patients; relative risk [RR], 0.40; 95% CI, 0.22e0.74) in

comparison with propofol, whereas there was no difference in the duration of mechanical

ventilation (five studies, 895 patients; mean difference, 0.53 h; 95% CI �2.66 to 3.72) or ICU

mortality (five studies, 267 patients; RR, 0.83; 95% CI, 0.32e2.12) between these two drugs.

Dexmedetomidine was associated with an increased risk of hypertension (three studies,

846 patients; RR, 1.56; 95% CI, 1.11e2.20) compared with propofol. Other adverse event rates

were similar between dexmedetomidine and propofol groups.

Conclusions: For ICU patient sedation, dexmedetomidine may offer advantages over pro-

pofol in terms of decrease in the length of ICU stay and the risk of delirium. However,

transient hypertension may occur when dexmedetomidine is administered with a loading

dose or at high infusion rates.

ª 2013 Elsevier Inc. All rights reserved.

thesiology, The First Affiliated Hospital, Sun Yat-sen University, No 58, Zhongshan Second0 87755766x8273; fax: þ86 20 37637600., liukexuan807@yahoo.com.cn (K.-X. Liu).s work.ier Inc. All rights reserved.

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3834

1. Introduction confined to RCTs within human adults, without any language

To maintain comfort and safety is essential for the treat-

ment of patients in the intensive care unit (ICU). Optimal

sedation regimen reduces anxiety, facilitates daily ICU

procedure, improves tolerance of mechanical ventilation, and

substantially shortens the length of ICU stay and decrea-

sesmorbidity andmortality [1e3]. The critical state of patients

in ICU requires an ideal sedative agent to have an efficacious

and safe profile. Benzodiazepines (e.g., midazolam and lora-

zepam), the gamma-aminobutyric acid (GABA) agonists, pre-

viously most commonly used sedatives for ICU patients, can

cause respiratory depression, delay awakening, and increase

the risk of delirium if administered continuously or exten-

sively [4,5].

Propofol, nowadays, has become a preferred sedative in ICU

because it offers advantages over benzodiazepines in terms

of lack of accumulation, quick onset, easy adjustment, and fast

recovery after discontinuation. It has sedative and hypnotic

effects that mediate the GABA receptor but has no analge-

sic action [3,6]. Adverse effects associated with propofol

included pain on injection, hypotension, bradycardia, respi-

ratory depression, and hypertriglyceridemia. Propofol infusion

syndrome is a rare but life-threatening adverse event and

remains a concern.

One relatively new agent, dexmedetomidine, presents an

alternative to the GABA agonists for ICU sedation therapy. It

exerts sedative property by the stimulation of a-2 receptors in

the locus coeruleus and thus blunts the central nervous

system excitation. In addition to sedation, dexmedetomidine

reduces concurrent analgesic requirements while maintain-

ing patient arousability without compromising respiratory

drive [7,8]. It reduced the incidence of delirium, shortened the

duration of mechanical ventilation, and lowered the total ICU

cost compared with benzodiazepines [9e11].

Dexmedetomidine was introduced for sedation in the ICU

setting approximately 10 y ago; there have been numerous

studies in which dexmedetomidine and propofol were

compared with respect to their clinical outcomes. The relative

benefits and harm between these two drugs, however, remain

controversial. We, therefore, did a meta-analysis of random-

ized controlled trials (RCTs) to clarify whether dexmedeto-

midine could be associated with improved outcomes for adult

ICU sedation in comparison with propofol.

2. Materials and methods

2.1. Search strategy

We performed an extensive search of MEDLINE (1950 to April

05, 2012), Embase (1966 to April 05, 2012), the Cochrane Library

(issue 1, 2012), and ClinicalTrials.gov (US Institutes of Health).

The following search strategy was used: “dexmedetomidine”

AND “propofol” AND (“intensive care” OR “critical care” OR

“critical illness” OR “critically ill” OR “sepsis” OR “trauma”

OR “wounds and injuries” OR “shock” OR “postoperative” OR

“surgical procedures”). Reference lists of relevant articleswere

also reviewed for any additional studies. Our search was

restrictions.

2.2. Study selection

2.2.1. Inclusion criteriaRCTs that met each of the following items were included:

(1) the setting was an adult ICU; (2) the study compared dex-

medetomidine with propofol for sedative therapy; and (3) the

primary or secondary outcomes included length of ICU stay,

duration of mechanical ventilation, ICU mortality, delirium,

hypotension, bradycardia, and hypertension.

2.2.2. Exclusion criteriaWe excluded studies if they (1) used dexmedetomidine or

propofol for anesthesia intraoperatively rather than continued

its use for sedation in the ICU for >6 h, (2) included patients

with neurologic impairment, and (3) did not report the specific

results comparing dexmedetomidine with propofol.

2.3. Data extraction and risk of bias assessment

Thiswas performedby two authors (Z.Q.X and S.Q.C)whowere

masked to the journal and article titles and the study authors.

Any disagreement in opinion was resolved by consensus with

all investigators. The following data were extracted from the

trials: publication year, methodology, patient population,

number of patients, intervention, and outcomes.

The primary outcomes were length of ICU stay (defined as

the time from admission to discharge from ICU), duration of

mechanical ventilation, and ICU mortality. The secondary

outcomes included risks of delirium, hypotension, brady-

cardia, and hypertension.

The Cochrane Collaboration’s tool was applied for assess-

ing the risk of bias in each identified study [12].

2.4. Statistical analysis

Length of ICU stay data were recorded as mean (standard

deviation [SD]) in days. Duration of mechanical ventilation

data were recorded as mean (SD) in hours. For continuous

outcomes (length of ICU stay and duration of mechanical

ventilation), mean differences (MDs) with 95% confidence

intervals (CIs) were calculated. For categorical outcomes

(incidence of delirium, hypotension, bradycardia, hyperten-

sion, and ICU mortality), relative risks (RR) with 95% CIs were

calculated. Using the formula provided by Hozo et al. [13], we

estimated the mean and variance of the trials in which only

median, range, and size were reported. Using the formula

provided in chapter 7 of the Cochrane Handbook for Systematic

Reviews of interventions, we estimated the mean and variance

of the trials in which only median, interquartile range (IQR),

and size were reported [14]. According to the statistical

suggestion explained in chapter 18 of the Meta-analysis by

Sterne et al. [15], we discarded the studies when there were no

events in neither dexmedetomidine nor propofol group.

To explore the effect of long-time use of dexmedetomidine

on the length of ICU stay when compared with propofol, we

performed a sensitivity analysis excluding studies with

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3 835

a mean length of ICU stay <3 d. All analyses were based on

a random-effects model. We explored heterogeneity between

studies using a chi-squared test, and a P value of <0.10 was

indicative of significant heterogeneity. Meta-regression was

undertaken to assess the potential effect of publication year,

patient age, study size, and infusion pattern of dexmedeto-

midine (using loading dose or not) when heterogeneity exis-

ted. Publication bias was assessed by funnel plot using

hypotension as an end point. Begg’s tests were also per-

formed, and a P value of <0.05 was regarded to indicate

potential publication bias. Data management and statistical

analyses were done with STATA, version 12.0, (StataCorp,

College Station, TX), and a P value of <0.05 was considered as

significant in this meta-analysis.

3. Results

After excluding duplicates, 42 unique citations were identified

(Fig. 1). Of the total, 29 articles were excluded, of which nine-

teen were reviews [16e34], four were case reports [35e38], two

were retrospective studies [39,40], two compared dexmedeto-

midine with haloperidol or saline [41,42], one was not

a randomized trial [43], and one included brain-injured

patients [44]. Thirteen studies were full-text read for further

evaluation, of which ten trials, totaling 1202 patients, met the

Fig. 1 e Flowchart of th

inclusion criteria [45e54]. The other three were excluded

because of failure to apply randomization or no relevant

outcomes being reported [55e57]. The evaluated trials included

data from 1202 patients and were published between January

2001 and April 2012. The details of the identified studies were

described in Table 1. Risk assessment was listed in Table 2.

3.1. Primary outcomes

3.1.1. Length of ICU stayFive trials reported the effect of dexmedetomidine on the

length of ICU stay when compared with propofol

[45,46,48,51,54]. Three reported mean (SD) stay [45,46,51], one

reported median (range) [54], and one reported median (IQR)

[48]. Among the five included trials, one had a mean length of

ICU stay of<3 d, whereas other trials hadmean lengths of ICU

stay between 1 and 2 wk [46].

The combined data showed that the use of dexmedeto-

midine for sedation in ICU significantly reduced the length of

ICU stay when compared with propofol (MD, �0.81 d; 95% CI,

�1.48 to �0.15; P ¼ 0.017), without statistical evidence of

heterogeneity among the studies (c2 ¼ 3.80, P ¼ 0.434). We

recorded no publication bias with Begg’s test (P ¼ 0.806;

Fig. 2A).

After excluding one trial with an obvious shorter length of

ICU stay (<3 d), a secondary analysis was performed [46]. The

e literature search.

Table 1 e Main characteristics of RCTs in the meta-analysis.

Study Population Intervention Cocurrenttreatment

Number of patients Outcomes usedin the meta-

analysis

Sedationlevel

Dexmedetomidine Propofol Dexmedetomidine Propofol

Jakob

et al. [48]

Patients (�18 y) requiring

invasive mechanical

ventilation and light

to moderate sedation

IV 0.2e1.4 mg/kg/h IV 0.3e4 mg/kg/h Fentanyl boli for

pain relief

n ¼ 251

Surgical ¼ 92

Nonsurgical ¼ 159

n ¼ 247

Surgical ¼ 77

Nonsurgical ¼170

Duration of

mechanical

ventilation,

delirium,

hypotension,

bradycardia,

and

hypertension

RASS: �3 to 0

Tasdogan

et al. [54]

Patients (�18 y) requiring

sedation and ventilation

after ileus surgery and

met at least two of the

criteria of sepsis

IV 1 mg/kg over 10 min loading,

followed by 0.2e2.5 mg/kg/h

over 24 h

IV 1 mg/kg over

15 min loading,

followed

by 1e3 mg/kg/h

over 24 h

IV alfentanil for

pain relief and

paracetamol

for hyperthermia

n ¼ 20

Surgical ¼ 20

Nonsurgical ¼ 0

n ¼ 20

Surgical ¼ 20

Nonsurgical ¼ 0

Length of ICU

stay, duration of

mechanical

ventilation, ICU

mortality, and

hypotension

RSS: <2

Memis

et al. [51]

Patients (�18 y) with

septic shock

IV 1 mg/kg over 10 min loading,

followed by 0.2e2.5 mg/kg/h

over 24 h

IV 1 mg/kg over

15 min loading,

followed by 1e3

mg/kg/h over 24 h

IV alfentanil for

pain relief

n ¼ 20

Surgical ¼ 0

Nonsurgical ¼ 20

n ¼ 20

Surgical ¼ 0

Nonsurgical ¼ 20

Length of ICU

stay, and ICU

mortality

RSS: <2

Maldonado

et al. [46]

Patients (�18 y) after

cardiac valve operations

IV 0.4 mg/kg loading,

followed by 0.2e0.7 mg/kg/h

IV 20e50 mg/kg/min Fentanyl, ketorolac,

hydrocodone, and

oxycodone for pain

relief and IV

haloperidol and

IV lorazepam for

agitation

n ¼ 40

Surgical ¼ 40

Nonsurgical ¼ 0

n ¼ 38

Surgical ¼ 38

Nonsurgical ¼ 0

Length of ICU

stay, duration

of mechanical

ventilation, ICU

mortality, and

delirium

RSS of 3 during

intubation

and RSS of 2

after extubation

Kaneko [53] Patients (�18 y) requiring

sedation after carotid

endarterectomy (CEA)

IV 0.2e0.7 mg/kg/h IV 1e3 mg/kg/h Diclofenac sodium

suppositories for

pain relief

n ¼ 33

Surgical ¼ 33

Nonsurgical ¼ 0

n ¼ 33

Surgical ¼ 33

Nonsurgical ¼ 0

Hypotension and

hypertension

RSS: 3e4

Memis

et al. [45]

Patients (�18 y) with

critical illness

IV 2.5 mg/kg/h over 10 min,

followed by 0.2 mg/kg/h

over 5 h

IV 2 mg/kg/h Not mentioned n ¼ 12

Surgical ¼ 1

Nonsurgical ¼ 11

n ¼ 12

Surgical ¼ 0

Nonsurgical ¼ 12

Length of ICU

stay,

hypotension,

and bradycardia

Not mentioned

Corbett

et al. [47]

Patients (�18 y) requiring

mechanical ventilation

of <24 h after

nonemergent

coronary artery bypass

graft surgery (CABG)

IV 1 mg/kg over 15 min loading,

followed by 0.4 mg/kg/h

IV 5e75 mg/kg/min Midazolam for

breakthrough anxiety,

morphine for pain

relief, and meperidine

for postoperative

shivering

n ¼ 43

Surgical ¼ 43

Nonsurgical ¼ 0

n ¼ 46

Surgical ¼ 46

Nonsurgical ¼ 0

ICU mortality,

delirium,

hypotension,

and bradycardia

RSS: 5 for the

first 2 h

postoperatively

RSS: 3e4 during

intubation

Elbaradie

et al. [49]

Patients (�18 y) requiring

sedation and ventilation

of >6 h after major

thoracic, abdominal,

or pelvic cancer surgeries

IV 2.5 mg/kg/h over 10 min

loading, followed by

0.2e0.5 mg/kg/h

IV 1 mg/kg loading,

followed by 0.5e1

mg/kg/h

IV fentanyl for

pain relief

n ¼ 30

Surgical ¼ 30

Nonsurgical ¼ 0

n ¼ 30

Surgical ¼ 30

Nonsurgical ¼ 0

Hypotension

and bradycardia

RSS: 2e5

journal

of

surgic

al

research

185

(2013)833e843

836

Herr etal.[52]

Patients

(�18y)after

CABG

IV1mg/kgover20min

loading,

followedby0.2e0.7

mg/kg/h

,

and

pro

pofolco

uld

begivenifmore

heavilyse

dationwasrequired

whereasth

einfu

sionrate

of

dexmedetomidinewasalready

atth

emaxim

um

of0.7

mg/kg/h

Acc

ord

ingto

each

investigator’s

usu

alpractice

Morp

hineand

nonstero

idalanti-

inflammatory

dru

gs

forpain

relief

n¼

148

Surgical¼

148

Nonsu

rgical¼

0

n¼

147

Surgical¼

147

Nonsu

rgical¼

0

Hypotension,

bradyca

rdia,

and

hypertension

RSS:�3

during

intu

bation

Venn

etal.[50]

Patients

(�18y)requiring

8hpostoperative

sedationandventilation

afterco

mplexmajor

abdominalorpelvic

surgery

IV2.5

mg/kg/h

over10min

loading,followedby

0.2e2.5

mg/kg/h

IV1e3mg/kg/h

,

abolusdose

of1mg/kgwas

given

initiallyifrequires

IValfentanilfor

pain

relief,atracu

rium

formusc

lerelaxation,

andparace

tamolfor

hypertherm

iaif

nece

ssary

n¼

10

Surgical¼

10

Nonsu

rgical¼

0

n¼

10

Surgical¼

10

Nonsu

rgical¼

0

ICU

mortality,

hypotension,

andbradyca

rdia

RSS:>2

RASS¼

RichmondAgitationSedationSca

le;RSS¼

Ramsa

ySedationSco

re.

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3 837

pooled result of the remaining four trials showed that the

use of dexmedetomidine for sedation in ICU did no reduce

the length of ICU stay compared with that of propofol

(MD, �0.09 d; 95% CI, �1.34 to �1.17; P ¼ 0.894), without

statistical evidence of heterogeneity among the studies

(c2 ¼ 1.98, P ¼ 0.576). We recorded no publication bias with

Begg’s test (P ¼ 1.000).

3.1.2. Duration of mechanical ventilationFive trials evaluated the effect of dexmedetomidine on

the duration of mechanical ventilation compared with that

of propofol [46e48,52,54], of which three reported mean (SD)

time duration [46e48], one reported median (range) [54],

and one reported median (IQR) [52]. The combined data sug-

gested that the use of dexmedetomidine for sedation in ICU

did not appear to reduce the duration of mechanical venti-

lation compared with that of propofol (MD, 0.53 h; 95% CI,

�2.66 to 3.72; P ¼ 0.744). However, significant heterogeneity

was noted among the studies (c2 ¼ 13.69, P ¼ 0.008). The

Begg’s test revealed no evidence of publication bias (P¼ 1.000;

Fig. 2B).

Meta-regression showed that for dexmedetomidine

therapy, compared with propofol, none of the publication

year, patient age, study size, and infusion pattern of dexme-

detomidine predicted the duration of mechanical ventilation

(P � 0.641) on univariate analysis.

3.1.3. ICU mortalityAll-cause morality during ICU period was available in five

trials [46,47,50,51,54]. No difference in ICU mortality was

found between patients receiving dexmedetomidine and

those receiving propofol (RR, 0.83; 95% CI, 0.32e2.12; P¼ 0.695),

without statistical evidence of heterogeneity among the

studies (c2 ¼ 2.42, P ¼ 0.658). The Begg’s test revealed no

evidence of publication bias (P ¼ 0.462; Fig. 2C).

3.2. Secondary outcomes

3.2.1. DeliriumThe incidences of delirium were determined in three studies

[46e48]. Delirium rates were significantly reduced with dex-

medetomidine compared with those with propofol (RR, 0.40;

95% CI, 0.22e0.74; P ¼ 0.003), without statistical evidence of

heterogeneity among the studies (c2 ¼ 2.21, P ¼ 0.332). The

Begg’s test revealed no evidence of publication bias (P ¼ 1.000;

Fig. 3).

3.2.2. HypotensionNine trials included hypotension as an outcome of interest

[45,47e54], of which three were excluded because no hemo-

dynamic emergency occurred in neither dexmedetomidine

nor propofol group [45,49,50]. Aggregation of the remaining six

studies showed that the use of dexmedetomidine for sedation

in ICUwas not associatedwith significant reduction of the risk

of hypotension compared with that of propofol (RR, 1.12; 95%

CI, 0.86e1.47; P ¼ 0.402), without statistically significant

heterogeneity among the studies (c2 ¼ 7.68, P ¼ 0.174). The

Begg’s test revealed no evidence of publication bias (P ¼ 1.000;

Fig. 4A).

Table

2e

Riskofbiasofin

cludedtrials.

Stu

dy

Randomization?

Alloca

tionco

nce

alm

ent?

Patients

blinded?

Personnelandoutcom

eass

essors

blinded

Withdrawal

Selectivereporting?

Jakobet

al.[48]

Yes,

byace

ntralinteractive

voice-resp

onse

system

Unknown,notsp

ecifica

llystated

Yes

Yes

2/500

No,alloutcomesreported

Tasd

oganet

al.[54]

Yes,

byaco

mputer-generatedtable

Yes,

byse

aledenvelopes

No

No

0No,alloutcomesreported

Memis

etal.[51]

Yes,

byaco

mputer-generatedtable

Yes,

byse

aledenvelopes

No

No

0No,alloutcomesreported

Maldonadoet

al.[46]

Yes,

byrandom

drawing

Unknown,notsp

ecifica

llystated

No

No

0No,alloutcomesreported

Kaneko[53]

Yes,

notsp

ecifica

llydesc

ribedth

e

meth

odofrandomization

Unknown,notsp

ecifica

llystated

No

No

0No,alloutcomesreported

Memis

etal.[45]

Yes,

byaco

mputer-steered

perm

uted-block

design

Unknown,notsp

ecifica

llystated

Yes

Yes

0No,alloutcomesreported

Corb

ett

etal.[47]

Yes,

byarandom-n

umbertable

Unknown,notsp

ecifica

llystated

No

No

0No,alloutcomesreported

Elbaradie

etal.[49]

Yes,

usingatoss

Unknown,notsp

ecifica

llystated

Yes

No

0No,alloutcomesreported

Herr

etal.[52]

Yes,

perform

edbyth

estatistician

Yes,

byse

aledenvelopes

No

No

0No,alloutcomesreported

Vennet

al.[50]

Yes,

notsp

ecifica

llydesc

ribed

themeth

odofrandomization

Yes,

byse

aledenvelopes

No

No

0No,alloutcomesreported

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3838

3.2.3. BradycardiaFive trials included bradycardia as an outcome of interest

[45,48e50,52], of which three were excluded because no

hemodynamic emergency occurred in neither dexmedetomi-

dine nor propofol group [45,49,50]. Aggregation of the

remaining two studies showed that the use of dexmedeto-

midine for sedation in ICU was not associated with significant

reduction of risk of bradycardia compared with that of pro-

pofol (RR, 1.36; 95% CI, 0.85e2.18; P ¼ 0.203), without statisti-

cally significant heterogeneity among the studies (c2 ¼ 0.58,

P ¼ 0.446). The Begg’s test revealed no evidence of publication

bias (P ¼ 1.000; Fig. 4B).

3.2.4. HypertensionThe incidences of hypertension were examined in three

studies [48,52,53]. The analysis demonstrated that dexmede-

tomidine significantly increased the risk of hypertension

comparedwith propofol (RR, 1.56; 95% CI, 1.11e2.20; P¼ 0.010).

No statistically significant heterogeneity among the studies

(c2 ¼ 2.42, P ¼ 0.298) was found. The Begg’s test revealed no

evidence of publication bias (P ¼ 0.296; Fig. 4C). Table 3 listed

the detailed results of this meta-analysis.

Using hypotension as an end point, the funnel plot did not

suggest the presence of publication bias (Fig. 5).

4. Discussion

This meta-analysis showed that dexmedetomidine signifi-

cantly reduced the length of ICU stay and the risk of delirium

compared with propofol, whereas there was no difference in

the duration of mechanical ventilation or ICU mortality

between the two drugs. Moreover, the pooled analysis sug-

gested an increased risk of hypertension after use of dexme-

detomidine in comparison with that of propofol.

Distinct from GABA agonists, dexmedetomidine produces

its pharmacologic effect through a unique binding with a-2

receptor, and this may explain the improved clinical

outcomes that we detected. Dexmedetomidine could reduce

the requirements of opioids and thus decrease the risk of

oversedation associated with it. Complications of over-

sedation may be short term and long term, including

ventilator-associated pneumonia because of delayed libera-

tion from mechanical ventilation, cognitive impairment,

adverse events associated with inadvertent drug overdose,

and even post-traumatic stress disorder [25]. In addition,

dexmedetomidine produces a state called “cooperative seda-

tion,” which allows patients to interact with health care

providers. The better arousability and orientation, which is

recommended by 2002 Society of Critical Care Medicine pain

and sedation guideline, is of vital importance because it

allows patients to express their discomfort and perform

spontaneous breathing trials as well as aids in routine

assessment for the prevalence of delirium [3]. Furthermore,

dexmedetomidine does not have respiratory depressive

effects that may prolong extubation and ICU discharge.

However, our sensitivity analysis suggested that when it came

to patients with longer lengths of ICU stay, the beneficial

effect of dexmedetomidine in reducing the length of ICU stay

Table 3 e Pooled analysis of dexmedetomidine versus propofol for sedation in ICU patients.

Outcomes Studies(patients)

Heterogeneity test Results Hypothesistest

Begg’s test

c2 P I2 (%) Z P z P

For continuous variables MD (95% CI)

Length of ICU stay, d 5 (655) 3.80 0.434 0.0 �0.81 (�1.48 to �0.15) 2.40 0.017 0.24 0.806

Duration of mechanical ventilation, h 5 (895) 13.69 0.008 70.8 0.53 (�2.66 to 3.72) 0.33 0.744 �0.24 1.000

For binary variables RR (95% CI)

ICU mortality 5 (267) 2.42 0.658 0.0 0.83 (0.32e2.12) 0.39 0.695 0.73 0.462

Delirium 3 (658) 2.21 0.332 9.4 0.40 (0.22e0.74) 2.96 0.003 0.00 1.000

Hypotension 6 (1015) 7.68 0.174 34.9 1.12 (0.86e1.47) 0.84 0.402 0.00 1.000

Bradycardia 2 (788) 0.58 0.446 0.0 1.36 (0.85e2.18) 1.27 0.203 0.00 1.000

Hypertension 3 (846) 2.42 0.298 17.4 1.56 (1.11e2.20) 2.56 0.010 1.04 0.296

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3 839

did not exist, possibly because more opioids and other medi-

cations were used and ventilator-related diseases developed

during the ICU stay that complicated the situation.

In contrast with propofol, dexmedetomidine significantly

reduced the risk of delirium. A recently published meta-

Fig. 2 e Effects of dexmedetomidine versus propofol on (A) lengt

in hours, and (C) ICU mortality.

analysis also presented similar results for the effect of dex-

medetomidine on delirium, in which dexmedetomidine was

compared with a placebo or an alternative sedative agent in

elective cardiac surgery [58]. The development of delirium is

associated with prolonged ventilator dependence and

h of ICU stay in days, (B) duration of mechanical ventilation

Fig. 3 e Effects of dexmedetomidine versus propofol on the risk of delirium.

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3840

hospitalization and increased mortality and health care cost

[59e61]. The favorable role of dexmedetomidine in prevention

and treatment of delirium is more likely because of its

high affinity to a-2 receptor rather than GABA receptor.

The evidence suggests that stimulation of GABA receptor

increases levels of deliriogenic neurotransmitters, which may

predispose ICU patients to delirium [9,10,62]. Besides, the

analgesia-sparing property of dexmedetomidine lessens the

Fig. 4 e Effects of dexmedetomidine versus propofol on the risk

requirement of additional opioids, exposure to which may

also cause neurocognitive disorders, and thus further mini-

mizes the prevalence of delirium. Our analysis showed that

compared with propofol, dexmedetomidine decreased the

length of ICU stay without reduction in the duration of

ventilation, which could be because of better cognitive

performance of patients in the dexmedetomidine group.

Specific observations of cognitive changes, including cognitive

s of (A) hypotension, (B) bradycardia, and (C) hypertension.

Fig. 5 e Funnel plot.

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3 841

state after extubation, duration, and degree of delirium,

should be reported in further studies. Previous studies showed

that patients treated with dexmedetomidine experienced less

delirium compared with those treated with midazolam or

lorazepam [9,10]. As such, dexmedetomidine may be a prom-

ising agent because it could offer sedationwithout necessarily

compromising cognitive function.

Although bradycardia and hypotension are common

adverse effects of dexmedetomidine, our study showed that

the risks of these two events were comparable with those of

propofol. In all the studies that we evaluated, no withdrawal

effects including rebound hypertension or tachycardia were

observed on the discontinuation of dexmedetomidine infu-

sion. It is worthy of note, based on our analysis, that the use of

dexmedetomidine seemed to increase the risk of hyperten-

sion compared with that of propofol. The risk of hypertension

was determined by three studies, of which one used both

a loading dose of 1.0 mg/kg over 20min followed by continuous

infusion of 0.2e0.7 mg/kg/h, one used a relatively high main-

tenance dose of 0.2e1.4 mg/kg/h, and one used a maintenance

dose of 0.2e0.7 mg/kg/h. Transient hypertension could be

observed after a loading dose or high infusion rate because of

the activation of peripheral a-2b receptors, leading to vaso-

constriction that overwhelms the competing vasodilatory

effect of a-2a receptors [63e65]. Many clinicians chose to avoid

loading infusions of dexmedetomidine, reporting satisfactory

sedative effects, and hemodynamic stability could be attained

[65e67]. For health providers, the observed hypertension after

a loading dose or high infusion rate of dexmedetomidine

should be concerned. Further studies should focus more on

details of hemodynamic changes, such as the time line and

degree of hypertension. All cointerventions, including use of

analgesics and neuromuscular blockers, and preexisting

diseases of patients should also be reported.

Among the included RCTs, one trial estimated the average

total cost of postoperative therapy for patients receiving

dexmedetomidine and propofol, suggesting that dexmedeto-

midine was associated with a nonsignificant decrease of the

cost (dexmedetomidine, $7025; propofol, $9875; P ¼ 0.12) [46].

Future pharmacoeconomic assessment of dexmedetomidine

and propofol need to be conducted to evaluate their effec-

tiveness, safety, and expense.

Several limitations in this meta-analysis warrant discus-

sion. First of all, we integrated the results of all relevant indi-

vidual RCTs; however, our conclusion was still based on

a relative small number of trials. Although for most outcomes,

the tests of heterogeneity were not significant, the clinical

profiles of the patients included were not same, which may

make this study underpowered to detect the unrevealed but

statistically important difference between these two sedatives.

Second, different defined sedation goals and protocols were

used among the trials. Some trials allowed a wide range of

relatively deep sedation, such as Ramsay score of 2e5, 3e5,

2e6, or 3e6, whereas some trials required Ramsay score 1e3 or

Richmond Agitation Sedation Scale of �3 to 0. This could lead

to heterogeneity in the doses and the amount of these two

drugs and then influence the outcomes. Another limitation is

that the extensive exclusion criteria inmost of the trials, which

includedhemodynamic instability, neurologic disease, renal or

hepatic insufficiency, suspected pregnancy, grossly obese, and

spinal or epidural anesthesia, limited the applicability of the

results to the general critically ill patient population.

There have been several other meta-analyses of dexme-

detomidine in the literature, which compared dexmedetomi-

dine with a placebo or an alternative sedative agent or had

different inclusion criteria with our meta-analysis and also

included non-randomized studies [58,68]. Our study specifi-

cally compared two common sedatives, dexmedetomidine

and propofol, with respect to their clinical outcomes of ICU

patients. In summary, dexmedetomidine may have a prom-

ising role in ICU sedation for improved outcomes. Compared

with propofol, dexmedetomidine shortened the length of ICU

stay and decreased the risk of delirium. Consider avoiding

a loading dose or high infusion rates of dexmedetomidine

because transient hypertension may occur. Larger, high-

quality randomized trails of dexmedetomidine are needed,

with a focus on duration of mechanical ventilation, long-term

mortality, and pharmacoeconomic analysis.

Acknowledgment

This research was supported by a grant from the Natio-

nal Natural Science Foundation of China (number: 81171847,

to K.X.L).

r e f e r e n c e s

[1] Mazzeo AJ. Sedation for the mechanically ventilated patient.Crit Care Clin 1995;11:937.

[2] Tung A, Rosenthal M. Patients requiring sedation. Crit CareClin 1995;11:791.

[3] Jacobi J, Fraser GL, Coursin DB, et al. Clinical practiceguidelines for the sustained use of sedatives and analgesicsin the critically ill adult. Crit Care Med 2002;30:119.

[4] Laine GA, Hossain SM, Solis RT, et al. Polyethylene glycolnephrotoxicity secondary to prolonged high-doseintravenous lorazepam. Ann Pharmacother 1995;29:1110.

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3842

[5] McCollam JS, O’Neil MG, Norcross ED, et al. Continuousinfusions of lorazepam, midazolam, and propofol forsedation of the critically ill surgery trauma patient:a prospective, randomized comparison. Crit Care Med 1999;27:2454.

[6] Trapani G, Altomare C, Liso G, et al. Propofol in anesthesia.Mechanism of action, structure-activity relationships, anddrug delivery. Curr Med Chem 2000;7:249.

[7] Bhana N, Goa KL, McClellan KJ. Dexmedetomidine. Drugs2000;59:263.

[8] Venn RM, Karol MD, Grounds RM. Pharmacokinetics ofdexmedetomidine infusions for sedation of postoperativepatients requiring intensive care. Br J Anaesth 2002;88:669.

[9] Pandharipande PP, Pun BT, Herr DL, et al. Effect of sedationwith dexmedetomidine vs lorazepam on acute braindysfunction in mechanically ventilated patients: the MENDSrandomized controlled trial. JAMA: J Am Med Assoc 2007;298:2644.

[10] Riker RR, Shehabi Y, Bokesch PM, et al. Dexmedetomidinevs midazolam for sedation of critically ill patients:a randomized trial. JAMA: J Am Med Assoc 2009;301:489.

[11] Dasta JF, Kane-Gill SL, Pencina M, et al. A cost-minimizationanalysis of dexmedetomidine compared with midazolam forlong-term sedation in the intensive care unit. Crit Care Med2010;38:497.

[12] Higgins JPT, Green S, eds. Cochrane handbook for systematicreviews of interventions. Version 5.1.0 [updated March 2011].The Cochrane Collaboration, 2011. Available from www.cochrane-handbook.org.

[13] Hozo SP, Djulbegovic B, Hozo I. Estimating the mean andvariance from the median, range, and the size of a sample.BMC Med Res Methodol 2005;5:13.

[14] Higgins JP, Deeks JJ. Selecting studies and collecting data. In:Higgins J, Green S (Ed.). Cochrane handbook for systematicreviews of interventions. Version 5.0.1 [updated September2008]. The Cochrane Collaboration, 2008. Available fromwww.cochrane-handbook.org.

[15] Sterne JAC, Bradburn MJ, Egger M. Meta-analysis inStata. In: Egger M, Smith GD, Altman DG, editors. Systematicreviews in health care: meta-analysis in context. London:BMJ Publishing Group; 2008. p. 357.

[16] Riker RR, Fraser GL. Altering intensive care sedationparadigms to improve patient outcomes. Anesthesiol Clin2011;29:663.

[17] Monk TG, Price CC. Postoperative cognitive disorders. CurrOpin Crit Care 2011;17:376.

[18] Bracco D, Donatelli F. Volatile agents for ICU sedation?Intensive Care Med 2011;37:895.

[19] Dexmedetomidine (Precedex) for ICU sedation. Med LettDrugs Ther 2011;53:41.

[20] Nseir S, Makris D, Mathieu D, et al. Intensive Care Unit-acquired infection as a side effect of sedation. Crit Care 2010;14:R30.

[21] Arnold HM, Hollands JM, Skrupky LP, et al. Optimizingsustained use of sedation in mechanically ventilatedpatients: focus on safety. Curr Drug Saf 2010;5:6.

[22] Riker RR, Fraser GL. Altering intensive care sedationparadigms to improve patient outcomes. Crit Care Clin 2009;25:527.

[23] Arcangeli A, D’Alo C, Gaspari R. Dexmedetomidine use ingeneral anaesthesia. Curr Drug Targets 2009;10:687.

[24] Keegan MT. Sedation in the neurologic intensive care unit.Curr Treat Options Neurol 2008;10:111.

[25] Devlin JW. The pharmacology of oversedationin mechanically ventilated adults. Curr Opin Crit Care 2008;14:403.

[26] Carollo DS, Nossaman BD, Ramadhyani U.Dexmedetomidine: a review of clinical applications.Curr Opin Anaesthesiol 2008;21:457.

[27] Szumita PM, Baroletti SA, Anger KE, et al. Sedation andanalgesia in the intensive care unit: evaluating the role ofdexmedetomidine. Am J Health Syst Pharm 2007;64:37.

[28] Zapantis A, Leung S. Tolerance and withdrawal issues withsedation. Crit Care Nurs Clin North Am 2005;17:211.

[29] Tung A. New anesthesia techniques. Thorac Surg Clin 2005;15:27.

[30] Riker RR, Fraser GL. Adverse events associated withsedatives, analgesics, and other drugs that provide patientcomfort in the intensive care unit. Pharmacotherapy 2005;25:8S.

[31] Taiji K. [Dexmedetomidine hydrochloride (Precedex), a newsedative in intensive care, its pharmacologicalcharacteristics and clinical study result]. Nihon YakurigakuZasshi 2004;124:171.

[32] Nasraway SA Jr. Use of sedative medications in the intensivecare unit. Semin Respir Crit Care Med 2001;22:165.

[33] Coursin DB, Maccioli GA. Dexmedetomidine. Curr Opin CritCare 2001;7:221.

[34] Angelini G, Ketzler JT, Coursin DB. Use of propofol and othernonbenzodiazepine sedatives in the intensive care unit. CritCare Clin 2001;17:863.

[35] Ohashi Y, Ohta N, Hirao O, et al. Analgesic effect ofdexmedetomidine in a patient with herpetic stomatitis afterliving-donor lung transplantation. J Anesth 2008;22:297.

[36] Matsumoto K, Yanagita W, Miki S. [Ultra fast-trackanesthesia with operating room extubation in a patientundergoing OPCABG by combination of propofol-remifentanil-dexmedetomidine]. Masui 2008;57:206.

[37] Santos MC, Vinagre RC. [Dexmedetomidine forneurocognitive testing in awake craniotomy: case report].Rev Bras Anestesiol 2006;56:402.

[38] Braz LG, Camacho Navarro LH, Braz JR, et al. [Clonidineas adjuvant therapy for alcohol withdrawal syndromein intensive care unit: case report]. Rev Bras Anestesiol 2003;53:802.

[39] Reichert MG, Jones WA, Royster RL, et al. Effect ofa dexmedetomidine substitution during a nationwidepropofol shortage in patients undergoing coronary arterybypass graft surgery. Pharmacotherapy 2011;31:673.

[40] Devabhakthuni S, Pajoumand M, Williams C, et al. Evaluationof dexmedetomidine: safety and clinical outcomes incritically ill trauma patients. J Trauma 2011;71:1164.

[41] Reade MC, O’Sullivan K, Bates S, et al. Dexmedetomidinevs. haloperidol in delirious, agitated, intubated patients:a randomised open-label trial. Crit Care 2009;13:R75.

[42] Martin E, Ramsay G, Mantz J, et al. The role of the alpha2-adrenoceptor agonist dexmedetomidine in postsurgicalsedation in the intensive care unit. J Intensive Care Med 2003;18:29.

[43] Venn M, Newman J, Grounds M. A phase II study to evaluatethe efficacy of dexmedetomidine for sedation in the medicalintensive care unit. Intensive Care Med 2003;29:201.

[44] Mirski MA, Lewin JJ 3rd, Ledroux S, et al. Cognitiveimprovement during continuous sedation in critically ill,awake and responsive patients: the Acute Neurological ICUSedation Trial (ANIST). Intensive Care Med 2010;36:1505.

[45] Memis D, Dokmeci D, Karamanlioglu B, et al. A comparisonof the effect on gastric emptying of propofol ordexmedetomidine in critically ill patients: preliminary study.Eur J Anaesthesiol 2006;23:700.

[46] Maldonado JR, Wysong A, van der Starre PJ, et al.Dexmedetomidine and the reduction of postoperativedelirium after cardiac surgery. Psychosomatics 2009;50:206.

j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 5 ( 2 0 1 3 ) 8 3 3e8 4 3 843

[47] Corbett SM, Rebuck JA, Greene CM, et al. Dexmedetomidinedoes not improve patient satisfaction when compared withpropofol during mechanical ventilation. Crit Care Med 2005;33:940.

[48] Jakob SM, Ruokonen E, Grounds RM, et al. Dexmedetomidinevs midazolam or propofol for sedation during prolongedmechanical ventilation: two randomized controlled trials.JAMA: J Am Med Assoc 2012;307:1151.

[49] Elbaradie S, El Mahalawy FH, Solyman AH.Dexmedetomidine vs. propofol for short-term sedation ofpostoperative mechanically ventilated patients. J Egypt NatlCanc Inst 2004;16:153.

[50] Venn RM, Bryant A, Hall GM, et al. Effects ofdexmedetomidine on adrenocortical function, and thecardiovascular, endocrine and inflammatory responses inpost-operative patients needing sedation in the intensivecare unit. Br J Anaesth 2001;86:650.

[51] Memis D, Kargi M, Sut N. Effects of propofol anddexmedetomidine on indocyanine green eliminationassessed with LIMON to patients with early septic shock:a pilot study. J Crit Care 2009;24:603.

[52] Herr DL, Sum-Ping ST, England M. ICU sedation aftercoronary artery bypass graft surgery: dexmedetomidine-based versus propofol-based sedation regimens. JCardiothorac Vasc Anesth 2003;17:576.

[53] Kaneko T. [Postoperative management of carotidendarterectomy with dexmedetomidineea comparison withpropofol]. Masui 2008;57:696.

[54] Tasdogan M, Memis D, Sut N, et al. Results of a pilot study onthe effects of propofol and dexmedetomidine oninflammatory responses and intraabdominal pressure insevere sepsis. J Clin Anesth 2009;21:394.

[55] Anger KE, Szumita PM, Baroletti SA, et al. Evaluation ofdexmedetomidine versus propofol-based sedation therapy inmechanically ventilated cardiac surgery patientsat a tertiary academic medical center. Crit Pathw Cardiol2010;9:221.

[56] Aoki M, Nishimura Y, Baba H, et al. [Effects ofdexmedetomidine hydrochloride on postoperative sedationin cardiovascular surgery]. Kyobu Geka 2006;59:1181.

[57] Ruokonen E, Parviainen I, Jakob SM, et al. Dexmedetomidineversus propofol/midazolam for long-term sedationduring mechanical ventilation. Intensive Care Med 2009;35:282.

[58] Lin YY, He B, Chen J, et al. Can dexmedetomidine be a safeand efficacious sedative agent in post-cardiac surgerypatients? a meta-analysis. Crit Care 2012;16:R169.

[59] Ely EW, Shintani A, Truman B, et al. Delirium asa predictor of mortality in mechanically ventilated patientsin the intensive care unit. JAMA: J Am Med Assoc 2004;291:1753.

[60] Ely EW, Inouye SK, Bernard GR, et al. Delirium inmechanically ventilated patients: validity and reliability ofthe confusion assessment method for the intensive care unit(CAM-ICU). JAMA: I Am Med Assoc 2001;286:2703.

[61] Milbrandt EB, Deppen S, Harrison PL, et al. Costs associatedwith delirium in mechanically ventilated patients. Crit CareMed 2004;32:955.

[62] Kahraman S, Zup SL, McCarthy MM, et al. GABAergicmechanism of propofol toxicity in immature neurons. JNeurosurg Anesthesiology 2008;20:233.

[63] Bloor BC, Ward DS, Belleville JP, et al. Effects of intravenousdexmedetomidine in humans. II. Hemodynamic changes.Anesthesiology 1992;77:1134.

[64] Ebert TJ, Hall JE, Barney JA, et al. The effects of increasingplasma concentrations of dexmedetomidine in humans.Anesthesiology 2000;93:382.

[65] Dasta JF, Kane-Gill SL, Durtschi AJ. Comparingdexmedetomidine prescribing patterns and safety in thenaturalistic setting versus published data. AnnPharmacother 2004;38:1130.

[66] Shehabi Y, Ruettimann U, Adamson H, et al.Dexmedetomidine infusion for more than 24 hours incritically ill patients: sedative and cardiovascular effects.Intensive Care Med 2004;30:2188.

[67] Gerlach AT, Dasta JF. Dexmedetomidine: an updated review.Ann Pharmacother 2007;41:245.

[68] Tan JA, Ho KM. Use of dexmedetomidine as a sedativeand analgesic agent in critically ill adult patients:a meta-analysis. Intensive Care Med 2010;36:926.