Received: 9 July 2019 | Revised: 20 September 2019 | Accepted: 30 September 2019

DOI: 10.1111/pan.13752

RESEARCH REPORT

The effect of ketamine on emergence agitation in children:

A systematic review and meta‐analysis

Ka Ting Ng1 | Deep Sarode2 | Yuen Sin Lai1 | Wan Yi Teoh3 | Chew Yin Wang1

1
Department of Anaesthesiology, University
of Malaya, Kuala Lumpur, Malaysia Abstract
2
Department of Anaesthesiology, University Background: Ketamine is believed to reduce the incidence of emergence agitation in
of Glasgow, Glasgow, Scotland
children undergoing surgery or procedure. However, recent randomized controlled
3
Department of Anaesthesiology, University
of Liverpool, Liverpool, UK
trials reported conflicting findings.
Aims: To investigate the effect of ketamine on emergence agitation in children.
Correspondence
Dr Ka Ting Ng, Department of
Methods: Databases of MEDLINE, EMBASE, and CENTRAL were systematically
Anaesthesiology, Faculty of Medicine, searched from their start date until February 2019. Randomized controlled trials
University of Malaya, Jalan Universiti, Kuala
Lumpur 50603, Malaysia.
comparing intravenous ketamine and placebo in children were sought. The primary
Email: [email protected] outcome was the incidence of emergence agitation. Secondary outcomes included

Section Editor: Brian Anderson

postoperative pain score, duration of discharge time, and the adverse effects associ-
ated with the use of ketamine, namely postoperative nausea and vomiting, desatura-
tion, and laryngospasm.
Results: Thirteen studies (1125 patients) were included in the quantitative meta‐
analysis. The incidence of emergence agitation was 14.7% in the ketamine group and
33.3% in the placebo group. Children receiving ketamine had a lower incidence of
emergence agitation, with an odds ratio being 0.23 (95% confidence interval: 0.11
to 0.46), certainty of evidence: low. In comparison with the placebo, ketamine group
achieved a lower postoperative pain score (odds ratio: −2.42, 95% confidence inter-
val: −4.23 to −0.62, certainty of evidence: very low) and lower pediatric anesthe-
sia emergence delirium scale at 5 minutes after operation (odds ratio: −3.99, 95%
confidence interval: −5.03 to −2.95; certainty of evidence: moderate). However, no
evidence was observed in terms of incidence of postoperative nausea and vomiting,
desaturation, and laryngospasm.
Conclusion: In this meta‐analysis of 13 randomized controlled trials, high degree of
heterogeneity and low certainty of evidence limit the recommendations of ketamine
for the prevention of emergence agitation in children undergoing surgery or imaging
procedures. However, the use of ketamine is well‐tolerated without any notable ad-
verse effects across all the included trials.
PROSPERO registration: CRD42019131865.

KEYWORDS
emergence agitation, emergence delirium, ketamine, meta‐analysis, pain, systematic review

Pediatric Anesthesia. 2019;00:1–10. wileyonlinelibrary.com/journal/pan

© 2019 John Wiley & Sons Ltd | 1
2 | NG et al.

1 | I NTRO D U C TI O N
What is already known about this topic
Emergence agitation is a temporary state of dissociation, disorienta-
tion, and irritability during the early stage of children's recovery from • Emergence agitation in children after surgery and procedure
anesthesia.1,2 The incidence of emergence agitation varies between is distressing to parents, caregivers, and nurses, which may
3
18% and 80%, 4-6
depending on different types of emergence agita- impose risk of self‐injury such as accidental removal of cath-
tion scoring system, the choices of anesthetic agents, the types of eters or peripheral line and falling out of bed.
2,7
surgery, age, and preoperative anxiety level. Although this is a re-
What new information this study adds
versible and self‐limiting condition, it comprises a substantial risk of
self‐injury and contributes significant distress to both caregivers and • This meta‐analysis demonstrates the potential role of
family members.7 ketamine to prevent emergence agitation in children un-
Ketamine is a noncompetitive N‐methyl‐D‐aspartate receptor dergoing surgery or procedure.
antagonist. It manifests the effects of anesthesia and analgesia in a • Its use is well‐tolerated without any notable adverse ef-
dose‐dependent relationship.7,8 Several meta‐analyses of observa- fects, namely incidence of postoperative nausea and
tional studies have demonstrated an association between delirium vomiting, desaturation, and laryngospasm.
and pro‐inflammatory cytokines, such as tumor necrosis factor‐α, in-
terleukin‐1, interleukin‐6, and interleukin‐8, neuronal injury marker,
and C‐reactive protein.9-11 It is believed that ketamine has the prop-
erties of anti‐inflammation where patients who received ketamine MEDLINE (OvidSP), EMBASE (OvidSP), and Cochrane Controlled
were associated with lower serum inflammatory markers, namely Register of Trials were searched from their inceptions until February
C‐reactive protein and interleukin‐6.12 With the effect of anti‐in- 2019. Trial registries, namely the World Health Organization
flammation and antinociception of ketamine, several randomized International Clinical Trials Registry Platform and the clinicaltrial.
clinical trials have demonstrated the promising effects of ketamine gov.my, were searched for any ongoing studies. The search strategy
in reducing the incidence of postoperative delirium in both adults is outlined in Table S1. All randomized controlled trials (RCTs) exam-
and children.13-19 Ketamine is a low‐cost drug with wide therapeutic ining intravenous ketamine versus placebo/usual treatment only in
20
window, and as such likely to be cost‐effective if it is associated children (<18 years old) were included. No restrictions were imposed
with a lower rate of emergence agitation in children. However, some on publication language and length of follow‐up period. A manual
adverse effects associated with the use of ketamine, namely postop- search of references from the primary or relevant review articles
erative nausea and vomiting, incidence of laryngospasm, and desat- was performed. Study designs of observational studies, case series,
21,22
uration, were reported. To date, there have been no systematic and case reports were excluded in this review.
reviews and meta‐analyses examining the use of ketamine for the Primary outcome of this review was the incidence of emer-
prevention of emergence agitation in children. While the preven- gence agitation during the stay in PACU. In studies where Paediatric
tive effect of ketamine on emergence agitation seems desirable, it is Anaesthesia Emergence Delirium (PAED) ≥10 and ≥15 were available,
timely warranted for a robust systematic review and meta‐analysis a PAED score of ≥10 was used to define the incidence of emergence
to summarize the evidence use of ketamine in children. agitation as it has the highest diagnostic sensitivity and specific-
We speculated that intravenous ketamine reduced the inci- ity. 25-27 For studies that measured the PAED scores at different time
dence of emergence agitation in children. The primary objective of intervals, all the data were extracted and only the outcomes of more
this systematic review was to determine the effect of ketamine on than two studies were pooled as secondary outcomes. Other sec-
emergence agitation in children undergoing surgery or procedures. ondary outcomes included postoperative pain score at the arrival
Secondary objectives were to evaluate the effect of ketamine on in PACU using the Modified Children's Hospital of Eastern Ontario
postoperative pain score, duration of discharge from postanaesthe- Pain Scale (mCHEOP), duration of discharge time from PACU with
sia care unit (PACU), and the adverse effects associated with the use the Aldrete score of ≥9, incidence of postoperative nausea and vom-
of ketamine, namely postoperative nausea and vomiting, desatura- iting, desaturation, and laryngospasm.
tion, and laryngospasm. Prior to the initiation of article screening, two authors (YL and DS)
were briefed on study inclusion and exclusion criteria by the main au-
thor (KN). Both authors (YL and DS) independently screened through
2 | M ATE R I A L S A N D M E TH O DS the titles/abstracts. Articles coded as “no” by both authors (YL and DS)
were excluded. In articles which coded as “yes” by both authors (YL
Our review adhered to the Cochrane Handbook for Systematic and DS), full‐text articles were retrieved and screened independently
23
Reviews and Interventions and the Preferred Reporting Items by both authors (YL and DS). For articles coded as “maybe” or any dis-
for Systematic Reviews and Meta‐analyses. 24 Prior to the literature crepancy on the coding of screened articles, it was resolved by con-
search, our review protocol was registered and published in a public sulting with a third author (KN). The final selection of all the included
database, PROSPERO (CRD42019131865). studies was discussed and agreed upon by all the authors.
NG et al. | 3

Data extraction form was piloted by the main author (KN). studies used ketamine as a comparator, except for two studies15,29
Data extraction was performed by two authors (YL and DS) inde- comparing ketamine‐propofol with the control group only re-
pendently using a standardized online data extraction form. All the ceiving propofol. The majority of studies14-16,32-35 administered
extracted data were checked by a third author (KN). In addition to 0.25 mg kg−1 of ketamine, with three studies administering19,32,36,37
the reported outcomes, the following fields, namely author, year, 0.5 mg kg−1 and four studies giving17,29,34,36 1 mg kg−1. All the stud-
study design, route, timing and dosage of ketamine, age, country, ies gave intravenous ketamine as a bolus dose, except one study17
and type of surgery, were extracted. All the included RCTs were which administered a bolus dose of ketamine followed by ketamine
assessed for risk of bias by two authors (YL and DS) independently, infusion throughout the surgery. For the purpose of data analysis,
using the Cochrane Collaboration Risk of Bias Assessment Tool. 23 the measured outcomes for different dosages of ketamine were
Any conflicts were resolved via a discussion with a third author combined in three studies.32,34,36 Ten RCTs anesthetized children
(KN). with sevoflurane, one36 desflurane, one18 isoflurane and one29 with
Review Manager Version 5.3 was used to generate the results a combination of sevoflurane and propofol. In terms of assessment
of forest plots for statistical meta‐analyses. 28 P < .05 (two‐tail) was tools for emergence agitation, six RCTs14,17,29,33,34,37 used PAED,
considered statistically significant. All findings were described as five15,18,19,32,36 utilized Aono's Four‐point Scale, and two16,35 used
odds ratios (OR) and mean difference (MD) with 95% confidence the Emergence Agitation Scale.
interval (CI) for binary and continuous outcomes, respectively. I‐ The overall risk of bias assessment was low for five stud-
square (I2) test was performed to examine the degree of hetero- ies17,19,29,33,34 and unclear for eight studies (Table S4).14-16,18,32,35-37
geneity in all the measured outcomes. I2 value of <40%, 40%‐60%, The highest risk of bias across all included studies was the domain
and >60% was categorized as low, moderate, and substantial, re- of sequence generation, followed by the allocation concealment
spectively. Fixed‐effect model was used to pool the estimates for and the blinding of outcome assessment. The summary of findings
all the measured outcomes. If substantial heterogeneity (I2 > 60%) and the PRISMA checklist are outlined in Table 2 and Table S5,
was detected, a random‐effect model was used. A subgroup anal- respectively.
ysis was performed on the primary outcome by stratifying all the The meta‐analytical findings of primary and secondary out-
included studies into different types of emergence agitation as- comes are tabulated in Table 3. Based on the combined data of
sessment tools. To further investigate the substantial heteroge- 11 RCTs (959 children), the incidence of emergence agitation was
neity of the incidence of emergence agitation, sensitivity analyses 14.7% in the ketamine group and 33.3% in the placebo group.
were conducted based on low risk of bias studies only and studies Ketamine reduced the incidence of emergence agitation, with
that diagnosed emergence agitation with PAED score only as it an OR being 0.23 (ρ < .0001; 95% CI: 0.11 to 0.46) (Figure 2).
was the only validated score and tool for emergence delirium. We However, statistical heterogeneity was substantial across the
also performed another sensitivity analysis by excluding two stud- included studies (I2 = 67%). The certainty of evidence for emer-
ies15,29 which compared propofol‐ketamine (ketofol) and propofol gence agitation was low due to the potential risk of bias, incon-
in children. sistency, and imprecision. All the included studies were of small
The assessment of evidence and summary of findings were con- sample size and underpowered where type 1 error may exist. To
ducted independently by two authors (YL and DS) using an online investigate for substantial heterogeneity, our subgroup analysis
software (https​://grade​pro.org).30 The level of evidence was as- indicated that different types of emergence agitation tools intro-
sessed based on risk of bias, inconsistency, indirectness, impreci- duced bias to the overall pooled estimate. Sensitivity analyses
sion, and publication bias. Any conflicts were resolved by consulting based on studies of low risk of bias and studies with PAED tool
a third author (KN). detected a change in the direction and magnitude of the inci-
dence of emergence agitation, indicating that our overall finding
may be skewed by studies with high risk of bias or studies with
3 | R E S U LT S other types of emergence agitation tools. Sensitivity analysis by
removing studies15,29 comparing ketofol versus propofol showed
The PRISMA diagram summarized the flow of study selection in no changes in the direction and magnitude of statistical estimate
Figure 1. Our search yielded 299 non‐duplicate articles for title/ effect (OR: 0.16, 95% CI: 0.11 to 0.24; ρ < .00001, I2 = 43%; par-
abstract screening. Applying the inclusion and exclusion criteria, ticipants = 568, studies = 9).
twenty‐two articles were selected for full‐text review. Of these, Four studies (217 children) measured the incidence of emergence
thirteen studies (1125 patients) were included for quantitative delirium using PAED score at 5 minutes after operation. Ketamine
meta‐analysis. Nine studies were excluded for the reasons detailed users were associated with a lower PAED score as compared to the
in Table S2. Searching clinical trial registries identified one relevant placebo group (ρ < .00001; OR: −3.99, 95% CI: −5.03 to −2.95; cer-
31
ongoing study (NCT02828566), which was estimated to conclude tainty of evidence = moderate). No substantial heterogeneity was
in February 2019 (Table S3). found across the studies.
The study characteristics of all included RCTs are illustrated in In comparison with the placebo group, children receiving in-
Table 1. All the included RCTs were single‐centered. Of all, eleven travenous ketamine had a lower postoperative pain score at the
4 | NG et al.

MEDLINE Cochrane database of systematic reviews EMBASE

33 Citation(s) 88 Citation(s) 254 Citation(s)

Non-Duplicate
76 Duplicates removed
375 citations screened

299 articles for title and

abstract screening

Inclusion/exclusion 277 Articles excluded after

criteria applied title/abstract screen

22 Articles
retrieved

Inclusion/exclusion 9 Articles excluded

criteria applied after full text screen

13 Articles
included

FIGURE 1 PRISMA flow diagram

arrival in PACU (ρ = .009; OR: −2.42, 95% CI: −4.23 to −0.62; chil- Our meta‐analysis demonstrated no evidence on the adverse
dren = 321; studies = 5, certainty of evidence: very low). However, effects associated with the use of ketamine, namely the inci-
substantial heterogeneity was detected across all the included dence of postoperative nausea and vomiting (ρ = .31; OR: 1.24,
RCTs (I2 = 95%). 95% CI: 0.81 to 1.90; children = 928; studies = 9, certainty of ev-
Based on the combined data of eight RCTs (748 children), idence: moderate), desaturation (ρ = .74; OR: 0.92, 95% CI: 0.55
there was no evidence that ketamine reduced the discharge time to 1.52; children = 571; studies = 4, certainty of evidence: high),
from PACU where the Aldrete score ≥ 9 (ρ = .57; MD: 0.66, 95% and laryngospasm (ρ = .96; OR: 0.95, 95% CI: 0.13 to 7.00; chil-
CI: −1.60 to 2.93; certainty of evidence: moderate). The degree dren = 120; studies = 2, certainty of evidence: high). Statistical
of heterogeneity was found to be substantial across studies heterogeneity was found to be low for all the three measured
(I2 = 92%). outcomes.
TA B L E 1 Clinical characteristics of included studies
NG et al.

% of EA
Age Dosage of Regime of Timing of (ketamine/
Author Year Design (mean ± SD)‐ year Comparator ketamine ketamine ketamine Control Anesthesia Type of EA scale placebo) Type of surgery Setting Country n
35 −1
Dalens 2006 Single‐ K: 36.8 ± 17.9 mo Ketamine 0.25 mg kg Bolus At the Saline Sevoflurane Emergence 0%/ 10.7% Elective MRI MRI room Canada 61
center C: 29.1 ± 20.2 mo end of Agitation
RCT procedure Scale ≥ 4
−1
Abu‐ 2007 Single‐ K: 5.3 ± 0.9 Ketamine 0.25 mg kg Bolus Before the Saline Sevoflurane PAED ≥ 15 16.6%/ Dental repair with Operation Canada 80
Shahwan33 center C: 5.4 ± 0.8 end of the 34.2% no extraction theater
RCT surgery

Lee32 2010 Single‐ K0.25:5.0 ± 0.4 Ketamine K0.25:0.25 Bolus Before the Saline Sevoflurane Aono's Four‐ 20%80% Adenotonsillectomy Operation Korea 90
center K0.5:5.0 ± 0.4 mg kg−1; end of the point Scale > 2 theater
RCT C: 4.8 ± 0 K0.5:0.5 surgery
mg kg−1

Jeong36 2012 Single‐ K1.0:5.0 ± 0.4 Ketamine K0.5:0.5 Bolus Before the Saline Desflurane Aono's Four‐ 25%/ 85% Ophthalmic surgery Operation Korea 60
center K0.5:5.0 ± 0.4 mg kg−1; end of the point Scale > 2 theater
RCT C: 4.8 ± 0.4 K1.0:1.0 surgery
mg kg−1

Abdelhalim19 2013 Single‐ K: 5.1 ± 1.6 Ketamine 0.5 mg kg−1 Bolus Before the Saline Sevoflurane Aono's Four‐ 15%/ 42.5% Tonsillectomy ± ad- Operation Saudi 80
center C: 4.8 ± 1.9 end of the point Scale > 2 enoidectomy theater Arabia
RCT surgery

Chen17 2013 Single‐ K: 4.2 ± 1.2 Ketamine 1 mg kg−1 Bolus + Before the Saline Sevoflurane PAED ≥ 10 29.6%/ Strabismus surgery Operation China 51
center C: 4.3 ± 1.1 infusion end of the 70.8% theater
RCT surgery

Eghbal18 2013 Single‐ K: 9.1 ± 3.3 Ketamine 0.25 mg kg−1 Bolus Before the Saline Isoflurane Aono's Four‐ 30.3%/ Adenotonsillectomy Operation Iran 66
center C: 8.2 ± 3.1 end of the point Scale > 2 90.9% theater
RCT surgery
14 −1
Ozcan 2014 Single‐ K: 4.0 ± 1.6 Ketamine 0.25 mg kg Bolus Before the Saline Sevoflurane PAED ≥ 10 30%/ 55% Inguinal hernia Operation Turkey 40
center C: 4.8 ± 1.3 end of the repair, theater
RCT surgery circumcision, or
orchidopexy

Rashad16 2014 Single‐ K: 26.9 ± 9.1 mo Ketamine 0.25 mg kg−1 Bolus Before the Saline Sevoflurane Emergence 20%/ 40% hypospadias repair Operation Egypt 40
center C: 24.6 ± 10.2 mo end of the Agitation theater
RCT surgery Scale ≥ 4
15 −1
Rizk 2014 Single‐ K: 4.3 ± 1.5 Ketamine‐ 0.25 mg kg Bolus Before the Usual Sevoflurane/ Aono's Four‐ 16.7%/ Tonsillectomy ± ad- Operation Egypt 60
center C: 4.1 ± 1.3 propofol end of the treatment propofol point Scale > 2 23.3% enoidectomy theater
RCT surgery (propofol)
34
Moawad 2015 Single‐ K0.25:4.3 ± 1.4 Ketamine K0.25:0.25 Bolus Before the Saline Sevoflurane PAED ≥ 10 ‐ Elective MRI MRI room Egypt 120
center K1.0:4.2 ± 1.5 mg kg−1; end of the
RCT C: 4.6 ± 1.4 K1.0:1.0 surgery
mg kg−1
|
5

(Continues)
6 | NG et al.

4 | DISCUSSIONS

331
46

Abbreviations: C, control; EA, emergence agitation; k, ketamine; K0.25, ketamine 0.25mg kg−1; K0.5, ketamine 0.50 mg kg−1; K1.0, ketamine 1 mg kg−1; MRI, magnetic resonance imaging; n, sample size;
n
Our review demonstrated that ketamine reduced the incidence of

Switzerland
emergence agitation in children undergoing surgery or magnetic

Country

Turkey
resonance imaging (MRI) scan. The postoperative pain score was
lower in the ketamine users. This review observed no evidence of

Operation

MRI room
theater
Setting ketamine in the duration of discharge time from PACU and no ad-
verse effects were associated with the use of ketamine, namely the
incidence of postoperative nausea and vomiting, desaturation, and
laryngospasm. The general quality of evidence ranged from very low
Type of surgery

bronchoscopy

Elective MRI to high, mainly due to the potential risk of bias, inconsistency, impre-
Fiberoptic

cision, publication bias, and dose‐response gradient.

Emergence agitation and emergence delirium are used inter-
changeably in the literature.7 In our review, the incidence of emer-
4.3%/ 1.8%
(ketamine/

gence agitation was 14.7% in the ketamine group and 33.3% in

placebo)
% of EA

the placebo group. Several reviews have supported this promising

effect of ketamine for the prevention of emergence agitation in
‐

children.7,26,38 However, our findings need to be interpreted with

Type of EA scale

cautions as several important confounding factors were identified,

PAED ≥ 15

PAED ≥ 10

which introduced variances to our pooled estimates. Different emer-

gence agitation scoring tools were used across studies, with differ-
ent degrees of sensitivity and specificity. 25,27 In comparison with
Sevoflurane/
Sevoflurane
Anesthesia

other assessment scores (Watcha Scale, Cravero Scale, Emergence

propofol

Agitation Scale, Aono's Four‐point Scale), the PAED score is known

as the most reliable, validated, and comprehensive scoring system to
treatment
(propofol)

measure incidence of emergence delirium because it incorporates

Control

cognitive and agitation assessment criteria. 25-27,39 All the included

Saline

Usual

RCTs utilized different inhalational anesthetic agents (isoflurane,

sevoflurane, and desflurane) for the induction of anesthesia. Among
procedure

procedure
Before the
Regime of Timing of
ketamine

all the volatile anesthetic agents, sevoflurane was reported to be

end of
At the

associated with a higher incidence of emergence agitation than iso-

flurane and desflurane. However, a Cochrane review found no evi-
ketamine

dence in the incidence of emergence agitation between sevoflurane,

Bolus

Bolus

isoflurane, and desflurane.40

Of all the included studies, two studies compared ketofol and
0.5 mg kg−1
Dosage of

1 mg kg−1
ketamine

propofol for the prevention of emergence agitation in children.15,29

Several meta‐analyses have demonstrated that children anesthetized
with propofol had a lower incidence of emergence agitation than
Comparator

those administered sevoflurane.4,41,42 Thus, it is unclear whether the

Ketamine‐

RCT, randomized controlled trial; SD, standard deviation.

propofol
Ketamine

prevention of emergence agitation comes from either propofol, ket-

amine, or the additive effect of both propofol and ketamine. Despite
(mean ± SD)‐ year

all the included studies showing positive effect of ketamine on emer-

gence agitation, Schmitz and colleagues reported a contradictory find-
K: 4.3 ± 1.3
C: 4.0 ± 1.6

C: 3.7 ± 0.8
K: 4.1 ± 0.8

ing where ketamine increased the incidence of emergence agitation

in children, who anesthetized with either total intravenous anesthesia
Age

(propofol) or inhalation anesthetic agent (sevoflurane).29 There was a

center

center
Design

Single‐

Single‐
(Continued)

hypothesis asserting that inhalational anesthetic agents may alter the

RCT

RCT

balance between neuronal synapse excitation and inhibition, resulting

2018
2016
Year

in a higher incidence of emergence agitation than total intravenous

anesthesia.43,44 Thus, the inclusion of both total intravenous anesthe-
TA B L E 1

sia and inhalational anesthetic agents may introduce significant bias to

Schmitz29
Ozturk37
Author

the finding, making the interpretation of the benefits of ketamine dif-

ficult. In a sensitivity analysis by removing the two studies15,29 which
NG et al. | 7

TA B L E 2 Summary of findings’ table

Anticipated absolute effects

№ of participants Certainty of the Relative effect Risk difference with

Outcomes (studies) follow‐up evidence (GRADE) (95% CI) Risk with placebo ketamine

Incidence of emergence 959 (11 RCTs) LOWa,b,c OR 0.23 (0.11 333 per 1000 230 fewer per 1000
agitation to 0.46) (281 fewer to 146
fewer)
PAED score at 5 min 217 (4 RCTs) MODERATEc,d ‐ MD 3.99 lower (5.03
postoperatively lower to 2.95 lower)
Pain score at PACU 321 (5 RCTs) VERY LOWa,b,c,d ‐ MD 2.42 lower (4.23
lower to 0.62 lower)
Discharge time (time to 748 (8 RCTs) MODERATEb,d ‐ MD 0.66 higher (1.6
Aldrete score ≥ 9) lower to 2.93 higher)
Incidence of nausea/ 928 (9 RCTs) MODERATEc,d OR 1.24 (0.81 96 per 1000 20 more per 1000 (17
vomiting to 1.90) fewer to 72 more)
Incidence of desaturation 571 (4 RCTs) HIGHc OR 0.92 (0.55 121 per 1000 9 fewer per 1000 (51
to 1.52) fewer to 52 more)
Incidence of 120 (2 RCTs) HIGHc OR 0.95 (0.13 to 34 per 1000 2 fewer per 1000 (30
laryngospasm 7.00) fewer to 166 more)

Note: The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect
of the intervention (and its 95% CI). GRADE Working Group grades of evidence: High certainty: We are very confident that the true effect lies close
to that of the estimate of the effect; Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to
the estimate of the effect, but there is a possibility that it is substantially different; Low certainty: Our confidence in the effect estimate is limited:
The true effect may be substantially different from the estimate of the effect; Very low certainty: We have very little confidence in the effect esti-
mate: The true effect is likely to be substantially different from the estimate of effect.
Abbreviations: CI, confidence interval; OR, odds ratio; MD, mean difference.
a
Majority of the included studies were unclear risk of bias.
b
Substantial heterogeneity.
c
Total number of event <300 for binary outcome and <400 for continuous outcome.
d
Funnel plot suggests of publication bias.

included both inhalational and total intravenous anesthetic agents, the A Cochrane review has demonstrated that a single low dose
incidence of emergence agitation remained lower with the ketamine (<0.5 mg kg−1) to moderate dose (≥0.5 mg kg−1) of ketamine prior to
group and a smaller confidence interval and lower degree of hetero- the end of surgery as an analgesic adjunct reduces the postopera-
geneity were observed. In addition, all the included studies were not tive pain score.47 Ketamine manifests the effect of antinociception
adequately powered for the outcome of emergence agitation, with the via the inhibition of pain signals by blocking N‐methyl‐D‐aspar-
majority of the included studies were of high risk of bias due to inad- tate receptors.13,48 Our review finding was consistent with the
equate blinding and randomization. Thus, future adequately powered Cochrane review that the ketamine group reported lower post-
RCTs addressing the aforementioned confounding factors must seek operative pain score at the arrival in PACU. However, this finding
to provide clearer answers. needs to be cautiously interpreted due to the significant degree of
Although pain is recognized as one of the most important con- heterogeneity and the subjective assessment of pain among chil-
tributing factors to emergence agitation, the phenomenon of emer- dren. Pain measurement is challenging and complicated at different
gence agitation was also seen in children underwent non‐painful stages of emotional and cognitive development.7 It is sometimes
29,34,35
procedure such as MRI scan. Apart from pain, preoperative difficult to differentiate pain and emergence agitation in a crying
anxiety and rapid emergence following sedation in an unfamiliar child.46 Pain assessment is mainly based on observations by physi-
environment are among the risk factors of emergence agitation in cians or nurses, which could introduce significant observer bias in
45
children. Chen and team reported that children who randomized the reporting of pain. In our review, some confounding factors such
to benzodiazepine (midazolam) had a lower incidence of emergence as the dosage of ketamine ranging from 0.25 mg kg−1 to 1 mg kg−1
46
agitation as compared to the ketamine group. However, Ozcan and and different types of surgery may affect the analgesic effect of
colleagues failed to support the use of midazolam as no significant ketamine and severity of pain, which may influence the occurrence
differences were observed in the incidence of emergence agitation of emergence agitation. Thus, future adequately powered RCT on
between the midazolam, ketamine, and placebo groups.14 The find- non‐painful procedure (MRI) is warranted to eliminate the compo-
ing has to be interpreted with cautions as it was underpowered to nent of pain in order to evaluate the effect of ketamine on emer-
detect the incidence of emergence agitation.14 gence agitation.
8 | NG et al.

TA B L E 3 Meta‐analytic findings of primary and secondary outcomes

Outcomes Trials n I2 (%) FEM/REM MD/OR (95% CI) ρ

1 Emergence agitation 11 959 67 REM 0.23 (0.11 to 0.46) <.0001

1.1 Emergence agitation—subgroup analysis by different
Emergence Agitation Scales
PAED 4 502 65 REM 0.47 (0.17 to 1.32) .15
Aono's Four‐point Scale 4 356 68 REM 0.12 (0.05 to 0.33) <.0001
Emergence Agitation Scale 2 101 0 REM 0.30 (0.08 to 1.08) .07
Test for subgroup differences: Chi2 = 8.69, df = 2
(P = .01), I2 = 77.0%
1.2 Emergence agitation—sensitivity analysis (removal of 9 568 43 FEM 0.16 (0.11 to 0.24) <.00001
study with propofol‐ketamine as comparator)
Emergence agitation—sensitivity analysis (low risk of 4 542 68 REM 0.42 (0.15 to 1.19) .10
bias)
Emergence agitation—sensitivity analysis (PAED only) 4 502 65 REM 0.47 (0.17 to 1.32) .15
2 PAED score at 5 min postoperatively 4 217 9 FEM −3.99 (−5.03 to −2.95) <.0000
3 Pain score at PACU (mCHEOP) 5 321 95 REM −2.42 (−4.23 to −0.62) .009
4 Discharge time with the Aldrete score ≥ 9 (min) 8 748 92 REM 0.66 (−1.60 to 2.93) .57
5 Incidence of nausea and vomiting 9 928 0 FEM 1.24 (0.81 to 1.90) .31
6 Incidence of desaturation 4 571 0 FEM 0.92 (0.55 to 1.52) .74
7 Incidence of laryngospasm 2 120 0 FEM 0.95 (0.13 to 7.00) .96

Abbreviations: mCHEOP, Modified Children's Hospital of Eastern Ontario Pain Scale; PACU, postanaesthesia care unit; PAED, Paediatric Anaesthesia
Emergence Delirium.

FIGURE 2 Forest plot of incidence of emergence agitation

In this review, there was no evidence of ketamine use in short- Given the limited studies on the incidence of desaturation and laryn-
ening the duration of recovery time to achieve the Aldrete score of gospasm, our findings need to be interpreted with a caveat due to
≥9. Different dosages and regimes (bolus only versus bolus followed underpowered sample sizes and potential type II error.
by infusion) of ketamine were administered across different RCTs, One of the main limitations in this review was none of the
which could affect the recovery time of children after surgery or pro- included RCTs were adequately powered for the incidence of
cedures. Although the clinical effect of ketamine is dose‐dependent, emergence agitation. Varied primary outcome and lack of stan-
the dose‐response effect of ketamine and incidence of emergence dardization on the assessment tool for emergence agitation may
agitation were not explored in this review due to inadequate data introduce variances to our findings. Other limitations included
for pool analysis. Our review reported no evidence on the adverse different patients’ characteristics, types of surgical procedures,
effects associated with ketamine use, namely the incidence of post- age of patients, types of anesthetic agents, dosages and regimes
operative nausea and vomiting, desaturation, and laryngospasm. of ketamine administered, and other concomitant medications.
NG et al. | 9

Other psychological influences, namely parental anxiety, chil- 7. Mason KP. Paediatric emergence delirium: a comprehensive review
dren's coping ability, and preexisting maladaptive behavior, may and interpretation of the literature. Br J Anaesth. 2017;118(3):335‐343.
8. Gao M, Rejaei D, Liu H. Ketamine use in current clinical practice.
also affect the incidence of emergence agitation. The exclusion
Acta Pharmacol Sin. 2016;37:865‐872.
of trials published as abstracts or presented at conferences may 9. Liu X, Yu Y, Zhu S. Inflammatory markers in postoperative delirium
introduce publication bias to this review. (POD) and cognitive dysfunction (POCD): a meta‐analysis of obser-
In this meta‐analysis of 13 RCTs, high degree of heterogeneity vational studies. PLoS ONE. 2018;13(4):e0195659.
10. Androsova G, Krause R, Winterer G, Schneider R. Biomarkers of
and potential type I error limit the recommendation on the use of
postoperative delirium and cognitive dysfunction. Front Aging
ketamine to prevent emergence agitation in children undergoing sur- Neurosci. 2015;7:112.
gery or imaging procedures. However, ketamine is well‐tolerated in 11. Ayob F, Lam E, Ho G, Chung F, El‐Beheiry H, Wong J. Pre‐opera-
all the included studies without any notable adverse effects, namely tive biomarkers and imaging tests as predictors of post‐operative
delirium in non‐cardiac surgical patients: a systematic review. BMC
postoperative nausea and vomiting, incidence of desaturation, and
Anesthesiol. 2019;19(1):25.
laryngospasm. Future powered RCTs are warranted to increase the 12. Dale O, Somogyi AA, Li Y, Sullivan T, Shavit Y. Does intraopera-
certainty of evidence on the use of ketamine in children by overcom- tive ketamine attenuate inflammatory reactivity following sur-
ing the aforementioned confounding factors. gery? A systematic review and meta‐analysis. Anesth Analg.
2012;115(4):934‐943.
13. Hudetz JA, Patterson KM, Iqbal Z, et al. Ketamine attenuates
AC K N OW L E D G M E N T delirium after cardiac surgery with cardiopulmonary bypass. J
Cardiothorac Vasc Anesth. 2009;23:651‐657.
This review was registered in PROSPERO (CRD42019131865). 14. Ozcan A, Kaya AG, Ozcan N, et al. Effects of ketamine and midaz-
olam on emergence agitation after sevoflurane anaesthesia in chil-
dren receiving caudal block: a randomized trial. Rev Bras Anestesiol.
C O N FL I C T O F I N T E R E S T 2014;64(6):377‐381.
15. Rizk SN, Samir EM. Use of ketofol to control emergence agitation
The authors report no conflict of interest. in children undergoing adenotonsillectomy. Egyptian J Anaesth.
2014;30:13‐19.
16. Rashad MM, Soud D. The effect of different drugs on sevoflurane
AU T H O R S ’ C O N T R I B U T I O N S
emergence agitation in pediatric patients undergoing hypospadia
KN, DS, and YL involved in protocol/project management, data repair surgery. Egypt J Anaesth. 2014;30:123‐127.
17. Chen JY, Jia JE, Liu TJ, Qin MJ, Li WX. Comparison of the effects of
collection, or management, analyzed the data analysis, and wrote/
dexmedetomidine, ketamine and placebo on emergence agitation
edited the manuscript. WT and CW involved in protocol/project after strabismus surgery in children. Can J Anesth. 2013;60:385‐392.
management, analyzed the data, and wrote/edited the manuscript. 18. Eghbal MH, Taregh S, Amin A, Sahmeddini MA. Ketamine im-
proves postoperative pain and emergence agitation following ade-
notonsillectomy in children: a randomized clinical trial. Middle East J
ORCID Anaesthesiol. 2013;22(2):155‐160.
19. Abdelhalim AA, Alarfaj AM. The effect of ketamine versus fentanyl
Ka Ting Ng https://orcid.org/0000-0001-7825-2628 on the incidence of emergence agitation after sevoflurane anesthe-
sia in pediatric patients undergoing tonsillectomy with or without
Wan Yi Teoh https://orcid.org/0000-0001-9791-659X
adenoidectomy. Saudi J Anaesth. 2013;7(4):392‐398.
20. Radvansky BA, Puri S, Sifonios AN, Eloy JD, Vanny L. Ketamine‐ a nar-
rative review of its uses in medicine. Am J Ther. 2016;23:1414‐1426.
REFERENCES 21. Kye YC, Rhee JE, Kim K, et al. Clinical effects of adjunctive atro-
pine during ketamine sedation in pediatric emergency patients. Am
1. Yamashita M. Postanaesthetic excitation and agitation. Pediatr J Emerg Med. 2012;30:1981‐1985.
Anesth. 2003;13:641. 22. Song JW, Shim JK, Song Y, Yang SY, Park SJ, Kwak YL. Effect of
2. Dahmani S, Stany I, Brasher C, et al. Pharmacological preven- ketamine as an adjunct to intravenous patient‐controlled analgesia,
tion of sevoflurane‐ and desflurane‐related emergence agitation in patients at high risk of postoperative nausea and vomiting under-
in children: a meta‐analysis of published studies. Br J Anaesth. going lumbar spinal surgery. Br J Anaesth. 2013;111:630‐635.
2010;104(2):216‐223. 23. Higgins J, Green S (editors). Cochrane Handbook for Systematic
3. Voepel‐Lewis T, Malviya S, Tait AR. A prospective cohort study Reviews of Interventions Version 5.1.0 [updated March 2011]. The
of emergence agitation in the pediatric postanesthesia care unit. Cochrane Collaboration. https​://train​ing.cochr​ane.org/handbook.
Anesth Analg. 2003;96:1625‐1630. Published 2011. Accessed March 31, 2019.
4. Chandler JR, Myers D, Mehta D, et al. Emergence delirium in chil- 24. Shamseer L, Moher D, Clarke M, et al. Preferred reporting items for
dren: a randomized trial to compare total intravenous anesthesia systematic review and meta‐analysis protocols (PRISMA‐P) 2015.
with propofol and remifentanil to inhalational sevoflurane anesthe- BMJ. 2015;349:g4647.
sia. Pediatr Anesth. 2013;23(4):309‐315. 25. Bong CL, Ng AS. Evaluation of emergence delirium in Asian children
5. Locatelli BG, Ingelmo PM, Emre S, et al. Emergence delirium in using the Pediatric Anesthesia Emergence Delirium Scale. Pediatr
children: a comparison of sevoflurane and desflurane anesthesia Anesth. 2009;19:593‐600.
using the Paediatric Anesthesia Emergence Delirium scale. Pediatr 26. Wong D, Bailey CR. Emergence delirium in children. Anaesthesia.
Anesth. 2013;23(4):301‐308. 2015;70:375‐392.
6. John M, Berger TM. Paediatric anaesthesia and inhalation agents. 27. Janseen NJ, Tan EY, Staal M, et al. On the utility of diagnostic in-
Best Pract Res Clin Anaesthesiol. 2005;19:501‐522. struments for pediatric delirium in critical illness: an evaluation of
10 | NG et al.

the Pediatric Anesthesia Emergence Delirium Scale, the Delirium 40. Costi D, Cyna AM, Ahmed S, et al. Effects of sevoflurane versus
Rating Scale, the Delirium Rating Scale‐revised R‐98. Intensive Care other general anaesthesia on emergence agitation in children.
Med. 2011;37:1331‐1337. Cochrane Database Syst Rev. 2014;9:CD007084.
28. The Cochrane Collaboration. Review Manager (RevMan) [Computer 41. Kanaya A, Kuratani N, Satoh D, Kurosawa S. Lower incidence of
program]. Version 5.3. Copenhagen, Denmark: The Nordic Cochrane emergence agitation in children after propofol anesthesia com-
Centre; 2014. pared with sevoflurane: a meta‐analysis of randomized controlled
29. Schmitz A, Weiss M, Kellenberger C, et al. Sedation for magnetic trials. J Anesthesia. 2014;28:4‐11.
resonance imaging using propofol with or without ketamine at in- 42. Anouad MT, Yazbeck‐Karam VG, Nasr VG, El‐Khatib MF, Kanazi
duction in pediatrics‐ a prospective randomized double‐blinded GE, Bleik JH. A single dose of propofol at the end of surgery for
study. Pediatr Anesth. 2017;28(3):264‐274. the prevention of emergence agitation in children undergoing stra-
30. McMaster University and Evidence Prime Icn. GRADEpro|GDT. bismus surgery during sevoflurane anaesthesia. Anesthesiology.
2015. https​://grade​pro.org/. Accessed June 25, 2019. 2007;107:733‐738.
31. Poonai N, Langford C.Intranasal Ketamine for Procedural Sedation 43. Yli‐Hankala A, Vakkuri A, Sarkela M, Lindgren L, Korttila K, Jantti V.
(INK). clinicaltrials.gov. https​://clini​c altr​ials.gov/ct2/show/NCT02​ Epileptiform electroencephalogram during mask induction of anes-
828566. Published 2017. Accessed May 19, 2019. thesia with sevoflurane. Anesthesiology. 1999;91:1596‐1603.
32. Lee YS, Kim WY, Choi JS, Son JH, Kim JH, Park YC. The effect of 44. Cohen IT, Finkel JC, Hannallah RS, Humer KA, Patel KM. Rapid
ketamine on the incidence of emergence agitation in children un- emergence does not explain agitation following sevoflurane anaes-
dergoing tonsillectomy and adenoidectomy under sevoflurane gen- thesia in infants and children: a comparison with propofol. Pediatr
eral anesthesia. Korean J Anesthesiol. 2010;58(5):440‐445. Anesth. 2003;13:63‐67.
33. Abu‐Shahwan I, Chowdary K. Ketamine is effective in decreasing 45. Ozcengiz D, Gunes Y, Ozmete O. Oral melatonin, dexmedetomi-
the incidence of emergence agitation in children undergoing den- dine, and midazolam for prevention of postoperative agitation in
tal repair under sevoflurane general anesthesia. Pediatr Anesth. children. J Anesth. 2011;25(2):184‐188.
2007;17:846‐850. 46. Chen J, Li W, Hu X, Wang D. Emergence agitation after cataract
34. Moawad HS, El‐Diasty T. Efficacy of ketamine in prevention of ag- surgery in children: a comparison of midazolam, propofol and ket-
itation in children undergoing magnetic resonance imaging under amine. Pediatr Anesth. 2010;20(9):873‐879.
face mask sevoflurane: a randomized trial. Egyptian J Anaesth. 47. Chaparro L, Smith SA, Moore R, Wiffen PJ, Gilron I. Pharmacotherapy
2015;31(2):121‐125. for the prevention of chronic pain after surgery in adults. Cochrane
35. Dalens BJ, Pinard AM, Letourneau DR, Albert NT, Truchon RJ. Database Syst Rev. 2013;7:CD008307.
Prevention of emergence agitation after sevoflurane anesthesia 48. Hayhurst CJ, Farrin E, Hughes CG. The effect of ketamine on de-
for pediatric cerebral magnetic resonance imaging by small doses lirium and opioid‐induced hyperalgesia in the Intensive Care Unit.
of ketamine or nalbuphine administered just before discontinuing Anaesth Crit Care Pain Med. 2018;37:525‐527.
anesthesia. Anesth Analg. 2006;102:1056‐1061.
36. Jeong WJ, Kim WY, Moon MG, et al. The effect of ketamine on the
separation anxiety and emergence agitation in children undergo-
ing brief ophthalmic surgery under desflurane general anesthesia. S U P P O R T I N G I N FO R M AT I O N
Korean J Anesthesiol. 2012;63(3):203‐208.
37. Ozturk T, Acikel A, Yilmaz O, Topcu I, Cevikkalp E, Yuksel H. Effects Additional supporting information may be found online in the
of low‐dose propofol vs ketamine on emergence cough in children Supporting Information section at the end of the article.
undergoing flexible bronchoscopy with sevoflurane‐remifentanil
anesthesia: a randomized, double‐blinded, placebo‐controlled trial.
J Clin Anesth. 2016;35:90‐95.
How to cite this article: Ng KT, Sarode D, Lai YS, Teoh WY,
38. Reduque LL, Verghese ST. Paediatric emergence delirium. Anaesth
Crit Care Pain Med. 2013;13:39‐41. Wang CY. The effect of ketamine on emergence agitation in
39. Lee CA. Paediatric emergence delirium: an approach to di- children: A systematic review and meta‐analysis. Pediatr
agnosis and management in the postanaesthesia care unit. Anesth. 2019;00:1–10. https​://doi.org/10.1111/pan.13752​
J Perioper Crit Intensiv Care Nurs. 2017;3(140).https​://doi.
org/10.4172/2471-9870.10000140