This systematic review and meta-analysis assessed the efficacy of dexamethasone in reducing postoperative nausea, vomiting, pruritus, and enhancing postoperative analgesia in patients receiving neuraxial anesthesia with neuraxial morphine. The analysis included 8 randomized controlled trials with a total of 768 patients, 473 of which received dexamethasone and 295 received placebo. The results showed that dexamethasone was effective at reducing postoperative nausea, vomiting, and use of rescue antiemetics compared to placebo. Dexamethasone also reduced 24-hour pain scores and use of rescue analgesics. However, dexamethasone was not effective for reducing neuraxial morphine-induced pruritus.

1. Society for Obstetric Anesthesia and Perinatology
Section Editor: Cynthia A. Wong
Dexamethasone for the Prophylaxis of Postoperative
Nausea and Vomiting Associated with Neuraxial
Morphine Administration: A Systematic Review
and Meta-Analysis
Terrence K. Allen, MBBS, FRCA, Cheryl A. Jones, MD, DVM, and Ashraf S. Habib, MBBCh, MSc, MHS, FRCA
BACKGROUND: We performed a systematic review to assess the efficacy of dexamethasone in
reducing postoperative nausea, vomiting (PONV), pruritus, and enhancing postoperative analge-
sia in patients receiving neuraxial anesthesia with neuraxial morphine.
METHODS: We searched Medline (1966 –2011), the Cochrane Central Register of Controlled
Trials, EMBASE, and Web of Science for all randomized controlled trials comparing dexametha-
sone with placebo for the prevention of PONV and/or pruritus in patients receiving neuraxial
morphine as part of a neuraxial anesthetic technique. Data were extracted independently by the
authors on the incidence of PONV, pruritus, pain scores at 4 and 24 hours, and use of rescue
antiemetics, antipruritics, and analgesics.
RESULTS: Eight randomized controlled trials (4 cesarean deliveries, 4 total abdominal hyster-
ectomies) were included. From these trials, 768 patients were analyzed with 473 receiving
dexamethasone and 295 receiving placebo. The doses of dexamethasone investigated ranged
from 2.5 to 10 mg. Dexamethasone reduced the incidence of postoperative nausea (relative risk,
RR [95% confidence interval, CI] ϭ 0.57 [0.45, 0.72]), vomiting (RR [95% CI] ϭ 0.56 [0.43,
0.72]), and the use of rescue antiemetic therapy (RR [95% CI] ϭ 0.47 [0.36, 0.61]) compared
with placebo. There was no evidence of dose responsiveness with respect to its antiemetic
effect. Dexamethasone also reduced 24-hour visual analog pain scores (measured on an
11-point scale [0 –10]) (mean difference [95% CI] ϭ Ϫ0.30 [Ϫ0.46, Ϫ0.13]) and the use of
rescue analgesics (RR [95% CI] ϭ 0.72 [0.52, 0.98]). Dexamethasone did not reduce the
incidence of pruritus (RR [95% CI] ϭ 0.98 [0.84, 1.15]). Examination of the funnel plots and
Egger’s test revealed evidence of publication bias in the primary outcomes.
CONCLUSION: Dexamethasone is an effective antiemetic for patients receiving neuraxial
morphine for cesarean delivery and abdominal hysterectomy. In addition, the doses used for
antiemetic prophylaxis enhanced postoperative analgesia compared with placebo. However,
dexamethasone was not effective for the prophylaxis against neuraxial morphine–induced
pruritus. (Anesth Analg 2012;114:813–22)
D examethasone is an effective antiemetic when ad-
ministered prophylactically in patients receiving
general anesthesia.1– 4 However, its role as an anti-
emetic in patients receiving neuraxial anesthesia is not well
local anesthetics provides effective long-lasting postopera-
tive analgesia.5 However, neuraxial morphine has been
associated with a frequent incidence of postoperative nau-
sea, vomiting (PONV) and pruritus.6
described. Neuraxial anesthesia is frequently used for We performed this systematic review and meta-analysis
cesarean delivery and other lower abdominal and lower primarily to assess the efficacy of dexamethasone prophy-
limb procedures. The addition of neuraxial morphine to laxis in reducing the overall incidence of postoperative
nausea (PON) and postoperative vomiting (POV). Second-
arily, we also planned to assess its role in preventing
From the Department of Anesthesiology, Duke University Medical Center, pruritus and enhancing postoperative analgesia in patients
Durham, North Carolina. receiving neuraxial morphine as a part of a neuraxial
Accepted for publication December 9, 2011. anesthetic technique.
Supported solely by funding from the Duke University Medical Center’s
Department of Anesthesiology, Women’s Anesthesia and Critical Care
Division. METHODS
The authors declare no conflicts of interest. We followed the recommendations of the PRISMA
This report was previously presented, in part, at the American Society of statement.7
Anesthesia’s Annual Meeting in New Orleans, LA, October 2009.
Reprints will not be available from the authors. Eligibility Criteria
Address correspondence to Terrence K. Allen, MBBS, FRCA, Duke Univer- We performed a literature search of all published reports of
sity Medical System, Box 3094, Durham, NC 27710. Address e-mail to
terrence.allen@duke.edu. randomized controlled trials that compared a single dose of
Copyright © 2012 International Anesthesia Research Society IV dexamethasone with placebo primarily for the preven-
DOI: 10.1213/ANE.0b013e318247f628 tion of PONV in patients who received a single dose of
April 2012 • Volume 114 • Number 4 www.anesthesia-analgesia.org 813

2. Dexamethasone and Neuraxial Morphine–Induced PONV
neuraxial preservative-free morphine as an adjunct to a mean difference (MD) with 95% CI. If the 95% CI included
neuraxial anesthetic technique. In addition, we collected a value of 0, it was assumed that there was no difference
data on the incidence of pruritus and/or need for rescue between the groups. Mean and standard deviations were
antipruritic treatment, as well as pain scores and/or the estimated from studies reporting outcomes as median and
need for rescue analgesics. The primary anesthetic tech- range as recommended by Hozo et al.9 A fixed effects
nique had to be neuraxial anesthesia. model was used by default. Significant heterogeneity was
defined as I2 Ͼ 50%, and in such cases, the reason for
Literature Search and Study Selection heterogeneity was explored. Forest plots were used to
We searched Medline (1966 –2011), the Cochrane Central graphically represent and evaluate treatment effects. The
Register of Controlled Trials, EMBASE, and Web of Science number needed to treat was calculated to estimate the
using the terms “dexamethasone,“ “postoperative,” “nausea,” overall clinical impact of any statistically significant inter-
“vomiting,” “emesis,” “morphine,” “epidural,” “extradural,” ventions for dichotomous data for the primary outcomes.
“spinal,” “intrathecal,” “neuraxial,” “an(a)esthesia,” “analge- To assess whether there was a dose-response relationship
sia” and/or “pruritus,” and their combinations in all fields for the antiemetic efficacy of dexamethasone, we per-
without language restriction up to February 2011. The bibli- formed a meta-regression in which the outcome variable
ographies of retrieved trials were used to identify other was the log (RR) for PON or POV and the predictor
relevant articles. variable was dexamethasone dose. A fixed effect regression
The methodological quality of the included studies was model was used. Publication bias was assessed using
assessed using the risk of bias tool recommended by the funnel plots for the primary outcomes and the regression
Cochrane Collaboration.8 Two of the authors scored the test described by Egger et al.10 When outcomes were not
included studies independently (TKA and CAJ). Any dis- consistently reported and quantitative analysis was inap-
crepancies were resolved by discussion with the third propriate, they were reviewed qualitatively.
author (ASH). Analyses were performed using Review Manager
Software (Review Manager [Revman] Version 5.0. Co-
Data Collection and Presentation penhagen: The Nordic Cochrane Centre, The Cochrane
One author (TKA) extracted the data. A data collection Collaboration, 2008), and Comprehensive Meta-Analysis
sheet was used to collect the following data: (i) authors, (ii) software (Version 2.2.050).
year of publication, (iii) primary country of origin, (iv) type
of surgery, (v) anesthetic technique, (vi) dose, timing, and RESULTS
route of administration of neuraxial morphine, (vii) dose We initially identified 191 publications but subsequently
and time of administration of dexamethasone, (viii) out- excluded 178 articles (Fig. 1). Of the remaining 13 potentially
come measures including the incidence of PONV, pruritus, relevant articles, 8 were eventually analyzed.11–18 Of the
pain intensity, and the use of rescue antiemetics, antipru- excluded studies, no neuraxial morphine was administered in
ritics, and analgesics postoperatively, and (ix) side effects. 2 studies,19,20 intrathecal neostigmine was administered in 1
Two authors (CAJ and ASH) independently checked the study,21 intrathecal meperidine was administered in an-
extracted data for accuracy. When studies included mul- other,22 and there was no placebo group in the remaining
tiple group comparisons, only data from the dexametha- study.23 Seven of the studies were performed in Taiwan,
sone and placebo groups were extracted. For studies not Republic of China11–16,18 and the remaining study in the
reporting an overall incidence of PONV but instead report- United Kingdom.17 Authors of 2 publications were con-
ing PON and POV at different time intervals, we recorded tacted and provided missing data.14,17 The details of the
the highest incidence for both outcomes. Pain intensity was included trials are shown in Table 1. The methodological
recorded at 4 hours (early) and 24 hours (late) or the closest quality of the included studies is summarized in the risk of
reported time intervals in the included studies. Pain scores bias table (Table 2). There were no discrepancies between
were assumed to be at rest unless stated otherwise. We also the authors in the assessment of the methodological quality
collected the highest pain scores reported in each study of the included studies. The 8 randomized controlled trials
regardless of the reported time interval. Pain scores reported included 1014 patients; 246 of these patients received
as a 0- to 100-mm visual analog scale were converted to an another antiemetic (tropisetron,12 droperidol,11,15,18 cycliz-
11-point scale (0 ϭ no pain to 10 ϭ worst pain). We also ine,17 and metoclopramide16) either alone or in combina-
performed a subgroup analysis of different dexamethasone tion with dexamethasone. These patients were not included
doses, surgical procedures, and route of administration of in the analysis. Of the remaining 768 patients, 473 received
neuraxial morphine for the relevant outcomes when reported IV dexamethasone and 295 received a placebo. Neuraxial
by 2 or more studies. Authors were contacted to provide morphine was administered for abdominal hysterectomies
additional information or clarification when necessary. in 4 studies12,14,16,18 and cesarean deliveries in the remain-
ing 4 studies.11,13,15,17 Epidural anesthesia was the primary
Meta-Analysis anesthetic technique in 6 studies,12–16,18 combined spinal-
Dichotomous data were extracted and summarized using epidural anesthesia in 1 study,17 and spinal anesthesia in
relative risks (RRs) with 95% confidence intervals (CIs). If the remaining study.11 The dose of epidural morphine was
the 95% CI included a value of 1, it was assumed that there 3 mg12–16,18 and the dose of intrathecal morphine was 0.2
was no statistically significant difference between dexa- mg.11,17 The doses and time of administration of dexameth-
methasone and placebo. Continuous data were extracted as asone in each individual trial are listed in Table 1. Six stud-
mean and standard deviations and summarized using ies investigated a single dose of dexamethasone11,12,14 –17
814 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

3. Figure 1. Flow chart of screened, excluded, and
included trials. The included trials investigated
dexamethasone as an antiemetic in patients receiv-
ing neuraxial morphine who also had neuraxial
anesthesia as the primary anesthetic technique.
RCTs ϭ randomized controlled trials.
whereas 2 dose ranging studies compared 3 doses with a There was no evidence of heterogeneity observed with this
placebo group.13,18 Dexamethasone was used at a dose of outcome (I2 [95% CI] ϭ 32% [0%, 70%]). A fixed effect
2.5 mg in 2 studies,13,18 5 mg in 4 studies,12,13,16,18 8 mg in meta-regression demonstrated no evidence of dose-
4 studies,11,14,15,17 and 10 mg in 2 studies13,18 (Tables 1 and responsiveness for PON. The slope (95% CI) was 0.01160
3). Dexamethasone was administered before neuraxial mor- (Ϫ0.09684, 0.12003) log RR per milligram of dexamethasone
phine in 6 studies12–16,18 and after in 2 studies.11,17 In 7 of (P ϭ 0.83). In a subgroup analysis by type of surgery,
the 8 studies,11–16,18 dexamethasone was administered as dexamethasone significantly reduced the incidence of PON
an IV bolus, and in the remaining study,17 the drug was compared with placebo in patients having abdominal hys-
administered as an infusion. Visual examination of the terectomy (15% vs 32%, RR [95% CI] ϭ 0.46 [0.31, 0.68]) and
funnel plots for both PON (Fig. 2) and POV (not shown) cesarean delivery (23% vs 41%, RR [95% CI] ϭ 0.68 [0.51,
revealed evidence of publication bias. Egger’s test was also 0.90]). In a subgroup analysis by route of administration of
statistically significant for both PON (intercept [95% CI] ϭ neuraxial morphine, dexamethasone significantly reduced
Ϫ2.84 [Ϫ4.26, Ϫ1.42], P ϭ 0.002) and POV (intercept [95% PON compared with placebo in patients receiving epidural
CI] ϭ Ϫ2.80 [Ϫ3.87, Ϫ1.73], P Ͻ 0.001). morphine (14% vs 31%, RR [95% CI] ϭ 0.45 [0.32, 0.62]). In
these studies, dexamethasone was also administered before
Postoperative Nausea and Vomiting epidural morphine. In the studies in which intrathecal
Postoperative Nausea morphine was administered, there was no difference in
All 8 studies reported on the incidence of PON PON between the groups (52% vs 58%, RR [95% CI] ϭ 0.89
(Table 3).11–18 Overall, when all doses of dexamethasone for [0.64, 1.22]).
all included studies were combined, there was a significant
reduction in the incidence of PON compared with placebo Postoperative Vomiting
(19% vs 36%, RR [95% CI] ϭ 0.57 [0.45, 0.72]). Doses of 5, 8, POV was investigated in all 8 studies (Table 3).11–18 Overall,
and 10 mg significantly reduced the incidence of PON when all doses of dexamethasone were combined, there was
compared with placebo. However, the lowest dose of 2.5 a significant reduction in the incidence of POV compared with
mg did not significantly reduce the incidence of PON. placebo (15% vs 30%, RR [95% CI]) ϭ 0.56 [0.43, 0.72]). Doses
April 2012 • Volume 114 • Number 4 www.anesthesia-analgesia.org 815

4. 816
Table 1. Characteristics of Included Trials
Time of Timing of dexamethasone Dose and route Postoperative Postoperative Postoperative
First author, year administration administration in relation Type of Anesthetic of neuraxial rescue rescue rescue
of publication Study groups (n) of study drug to neuraxial morphine surgery technique morphine antiemetic antipruritic analgesic Outcomes analyzed
Wang, 199914 Dexamethasone 8 mg End of surgery Before TAH Epidural 3 mg epidural Metoclopramide Diphenhydramine Diclofenac 75 PON, POV, use of rescue antiemetics,
(38), saline (36) 10 mg IV 20 mg IM mg IM VAS pain scores,a use of rescue
analgesia, incidence of pruritus,
use of rescue antipruritica
Tzeng, 200015 Dexamethasone 8 mg End of surgery Before Cesarean Epidural 3 mg epidural Metoclopramide Diphenhydramine Diclofenac 75 PON, POV, use of rescue antiemetics,
(38), droperidol delivery 10 mg IV 20 mg IM mg IM use of rescue analgesia, incidence
1.25 mg (38), of pruritus, incidence of side
saline (37) effects
Ho, 200118 Dexamethasone End of surgery Before TAH Epidural 3 mg epidural Ondansetron 4 Diclofenac 75 PON, POV, use of rescue antiemetic,
2.5 mg (43), mg IV mg IM VAS pain scores, use of rescue
www.anesthesia-analgesia.org
dexamethasone antiemetic, incidence of side
5 mg (42), effects
dexamethasone 10
mg (44), droperidol
1.25 mg (42),
saline (43)
Dexamethasone and Neuraxial Morphine–Induced PONV
Wang, 200113 Dexamethasone 2.5 End of surgery Before Cesarean Epidural 3 mg epidural Ondansetron 4 Diphenhydramine Tenoxicam 20 PON, POV, use of rescue antiemetics,
mg (44), delivery mg IV 20 mg IM mg IV VAS pain scores, use of rescue
dexamethasone 5 analgesics, incidence of pruritus,
mg (44), use of rescue antipruritics,
dexamethasone 10 incidence of side effects
mg (43), saline
(44)
Tzeng, 200216 Dexamethasone 5 mg End of surgery Before TAH Epidural 3 mg epidural Ondansetron 4 Diclofenac 75 PON, POV, use of rescue antiemetics,
(38), mg IV mg IM VAS pain scores, use of rescue
metoclopramide 10 analgesia, incidence of side
mg (39), saline effects
(38)
Wang, 200212 Dexamethasone 5 mg End of surgery Before TAH Epidural 3 mg epidural Droperidol 1.25 Diphenhydramine Diclofenac 75 PON, POV, use of rescue antiemetics,
(38), tropisetron 5 mg IV 20 mg IV mg IM VAS pain scores, use of rescue
mg (38), saline analgesia, incidence of pruritus,
(37) use of rescue antipruritics,
incidence of side effects
Nortcliffe, 200317 Dexamethasone 8 mg In recovery After Cesarean CSE 0.2 mg intrathecal Prochlorperazine Chlorpheniramine Acetaminophen PON, POV, use of rescue antiemetics,
(30), cyclizine 50 room delivery 12.5 mg IM 4 mg orally 500 mg/ VAS pain scores, use of rescue
mg (30), saline codeine 30 analgesia, incidence of pruritus,a
(30) mg orally use of rescue antipruritics,
incidence of side effects
Wu, 200711 Dexamethasone 8 mg Before skin After Cesarean Spinal 0.2 mg intrathecal Ondansetron 4 Diphenhydramine Diclofenac 75 PON, POV, use of rescue antiemetics,
(30), droperidol incision delivery mg 20 mg IV mg IM VAS pain scores, use of rescue
1.25 mg (30), analgesics, incidence of pruritus,
dexamethasone 8 incidence of side effects
mg ϩ droperidol
1.25 mg (30),
saline (30)
TAH ϭ total abdominal hysterectomy; CSE ϭ combined spinal-epidural; PON ϭ postoperative nausea; POV ϭ postoperative vomiting; VAS ϭ visual analog scale.
a
Additional data obtained from authors.
ANESTHESIA & ANALGESIA

5. Table 2. Risk of Bias
Adequate sequence Allocation Incomplete data Free of selective
Included studies generation concealment Blinding addressed reporting
Wang, 199914 Yes No Yes Yes Yes
Tzeng, 200015 Yes No Yes Yes Yes
Ho, 200118 Unclear No Yes Yes Yes
Wang, 200113 Unclear No Yes No Yes
Tzeng, 200216 Yes No Yes Yes Yes
Wang, 200212 Yes No Yes Yes Yes
Nortcliffe, 200317 Yes Yes Yes Yes Yes
Wu, 200711 Yes No Yes Yes Yes
Each study risk of bias was assessed as Yes (low risk of bias), No (high risk of bias), or Unclear for each question-based entry.
Table 3. The Effect of Dexamethasone on the Incidence of Postoperative Nausea, Vomiting, Pruritus, and
the Use of Rescue Antiemetics and Antipruritics
Risk of treatment group Risk of control group RR (95% CI) NNT (95% CI)
Incidence of postoperative nausea
Dexamethasone 2.5 mg13,18 19/87 (22%) 26/87 (30%) 0.73 (0.44, 1.22) —
Dexamethasone 5 mg12,13,16,18 21/163 (13%) 50/162 (31%) 0.42 (0.26, 0.66) 6 (4, 11)
Dexamethasone 8 mg11,14,15,17 39/136 (29%) 57/133 (43%) 0.68 (0.50, 0.92) 8 (4, 35)
Dexamethasone 10 mg13,18 10/87 (12%) 26/87 (30%) 0.38 (0.20, 0.75) 5 (3, 15)
All studies combined11–18 89/473 (19%) 107/295 (36%) 0.57 (0.45, 0.72) 6 (4, 9)
Incidence of postoperative vomiting
Dexamethasone 2.5 mg13,18 14/87 (16%) 20/87 (23%) 0.70 (0.38, 1.29) —
Dexamethasone 5 mg12,13,16,18 14/163 (9%) 39/162 (24%) 0.36 (0.20, 0.63) 7 (4, 13)
Dexamethasone 8 mg11,14,15,17 35/136 (26%) 50/133 (38%) 0.69 (0.50, 0.96) 9 (4, 121)
Dexamethasone 10 mg13,18 6/87 (7%) 20/87 (23%) 0.30 (0.13, 0.71) 7 (4, 17)
All studies combined11–18 69/473 (15%) 89/295 (30%) 0.56 (0.43, 0.72) 7 (5, 11)
Use of rescue antiemetics
Dexamethasone 2.5 mg13,18 19/87 (22%) 28/87 (32%) 0.68 (0.41, 1.12) —
Dexamethasone 5 mg12,13,16,18 20/163 (12%) 56/162 (35%) 0.35 (0.22, 0.56) 5 (3, 8)
Dexamethasone 8 mg11,14,15,17 26/136 (19%) 46/133 (35%) 0.56 (0.38, 0.82) 6 (4, 15)
Dexamethasone 10 mg13,18 9/87 (10%) 28/87 (32%) 0.32 (0.16, 0.64) 5 (3, 10)
All studies combined11–18 74/473 (16%) 102/295 (35%) 0.47 (0.32, 0.69) 6 (4, 8)
Incidence of pruritus
All studies combined11–15,17 158/306 (52%) 120/214 (56%) 0.98 (0.84, 1.15) —
Use of rescue antipruritic
All studies combined12–14,17 29/238 (12%) 26/147 (18%) 0.71 (0.43, 1.18) —
RR ϭ relative risk; CI ϭ confidence interval; NNT ϭ number needed to treat.
Figure 2. A funnel plot of all included studies
investigating dexamethasone for the prevention of
postoperative nausea.
of 5, 8 and 10 mg significantly reduced the incidence of POV surgery, dexamethasone significantly reduced the inci-
when compared with placebo (Table 1). There was no evi- dence of POV compared with placebo in patients having
dence of heterogeneity among the trials (I2 [95% CI] ϭ 44% total abdominal hysterectomy (9% vs 25%, RR [95% CI] ϭ
[0%, 75%]). A fixed effect meta-regression demonstrated no 0.36 [0.22, 0.61]) and cesarean delivery (20% vs 36%, RR
evidence of a dose-response relationship. The slope (95% CI) [95% CI] ϭ 0.70 [0.52, 0.96]). In a subgroup analysis by
was 0.1127 (Ϫ0.0244, 0.2498) log RR per milligram of dexa- route of administration of neuraxial morphine, dexa-
methasone (P ϭ 0.11). In a subgroup analysis by the type of methasone significantly reduced the incidence of POV in
April 2012 • Volume 114 • Number 4 www.anesthesia-analgesia.org 817

6. Dexamethasone and Neuraxial Morphine–Induced PONV
Figure 3. Forest plots showing the effect of dexamethasone on (A) early pain scores (4 hours), (B) late pain scores (24 hours) (measured on
an 11-point scale [0 –10]), and (C) use of rescue analgesics. M-H ϭ Mantel-Haenszel; IV ϭ inverse variance.
those patients receiving epidural morphine compared subgroup analysis by route of administration of neuraxial
with placebo (9% vs 24%, RR [95% CI] ϭ 0.38 [0.25, 0.57]). morphine, dexamethasone significantly reduced the use of
However, in the studies in which intrathecal morphine postoperative rescue antiemetic therapy in those patients
was administered, there was no difference in POV be- receiving epidural morphine compared with placebo
tween the groups (50% vs 55%, RR [95% CI ϭ 0.91 [0.65, (12.8% vs 32.8%, RR [95% CI] ϭ 0.37 [0.26, 0.52]). A similar
1.28]). effect was not observed in patients receiving intrathecal
morphine (35% vs 42%, RR [95% CI ϭ 0.84 [0.56, 1.26]).
Use of Postoperative Rescue Antiemetic Treatment
All 8 studies investigated the use of postoperative rescue
antiemetic therapy (Table 3).11–18 Ondansetron 4 mg was
Postoperative Pain
the rescue antiemetic used in 4 studies,11,13,16,18 metoclo-
Postoperative Pain Intensity
pramide 10 mg in 2 studies,14,15 droperidol 1.25 mg in 1
Seven studies reported on postoperative pain intensity in
study,12 and prochlorperazine 12.5 mg in the remaining
patients receiving dexamethasone for PONV prophy-
study.17 Overall, when all doses of dexamethasone were
laxis.11–14,16 –18 Six of these studies reported early pain
combined, there was a significant reduction in the use of
scores at or around 4 hours (Fig. 3A).11–14,16,18 Five studies
postoperative rescue antiemetic therapy when compared
with placebo (16% vs 35%, RR [95% CI] ϭ 0.47 [0.36, 0.61]). reported late postoperative pain scores at 24 hours (Fig.
There was no evidence of significant heterogeneity (I2 ϭ 3B).11–13,16,18 One study reported highest and lowest pain
47%). Dexamethasone at doses of 5, 8, and 10 mg also scores only, without specifying a time period.17 There were
significantly reduced the need for a rescue antiemetic no differences in early pain scores in patients receiving
therapy (Table 2). In a subgroup analysis by the type of dexamethasone compared with those receiving placebo.
surgery, dexamethasone reduced the need for postopera- Overall, when all doses were combined, dexamethasone
tive rescue antiemetic therapy in patients undergoing total reduced postoperative pain scores at 24 hours when com-
abdominal hysterectomy (13% vs 31%, RR [95% CI] ϭ 0.38 pared with the placebo group (MD [95% CI] ϭ Ϫ0.30
[0.25, 0.59]) and cesarean delivery (18% vs 38%, RR [95% [Ϫ0.46, Ϫ0.13]). There was evidence of significant hetero-
CI] ϭ 0.53 [0.29, 0.98]) compared with placebo. In a geneity among the included trials (I2 ϭ 72%). Removing the
818 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

7. only study11 in which intrathecal morphine was adminis- not consistently reported quantitatively and so were re-
tered abolished this heterogeneity (I2 ϭ 0). When combin- viewed qualitatively. Five studies investigated delayed
ing the remaining studies, there was still a reduction in wound healing as a possible side effect during the patient’s
pain scores (MD [95% CI] ϭ Ϫ0.20 [Ϫ0.37, Ϫ0.02]). Doses of postoperative stay.11,13,15,16,18 There were no reported cases
dexamethasone of 5 mg (MD [95% CI] ϭ Ϫ0.19 [Ϫ0.37, in either the treatment or placebo groups. Similarly, there
Ϫ0.01]) and 8 mg (MD [95% CI] ϭ Ϫ1.00 [Ϫ1.45, Ϫ0.55]) were 5 studies investigating wound infection,11,13,15,16,18
significantly reduced pain scores at 24 hours compared with no reported cases in either the treatment or interven-
with placebo. No conclusion could be drawn for the tion group. Other less relevant side effects investigated
comparison of the highest reported pain scores between depended on the antiemetics being investigated in conjunc-
patients receiving dexamethasone and those in the placebo tion with dexamethasone. Three studies comparing the
group because of the wide CIs of the pooled results (MD antiemetic efficacy of droperidol alone or in combination
[95% CI] ϭ Ϫ1.87 [Ϫ4.16, 0.41]). with dexamethasone reported no difference in the inci-
dence of restlessness between the dexamethasone and
Use of Postoperative Rescue Analgesics placebo groups.11,15,18 In 2 of these studies, there was no
All 8 studies reported on the use of rescue analgesics in reported incidence of this side effect in any patient.11,15
patients receiving dexamethasone for PONV prophy- Two studies assessed and reported postoperative seda-
laxis.11–18 In 6 of the studies, diclofenac 75 mg IM was used tion.11,17 One study compared droperidol alone or in com-
as the rescue analgesic.11,12,14 –16,18 IV tenoxicam 20 mg was bination with dexamethasone11 whereas the other used
administered in 1 study13 and oral acetaminophen 500 cyclizine.17 Both studies used a 4-point sedation scale and
mg/codeine 30 mg in the other.17 Overall, dexamethasone reported no difference in sedation scores over the study
reduced the use of postoperative rescue analgesics com- period between the dexamethasone and placebo groups.
pared with placebo (27% vs 38%, RR [95% CI] ϭ 0.76 [0.62, One study investigating the antiemetic efficacy of tropise-
0.93]) (Fig. 3C). In subgroup analysis by dexamethasone tron reported an incidence of headaches of 5% in both the
dose, only dexamethasone at a dose of 5 mg significantly dexamethasone and placebo groups.12 Another study com-
reduced the need for rescue analgesia when compared with paring dexamethasone with metoclopramide reported no
placebo (27% vs 38%, RR [95% CI] ϭ 0.72 [0.52, 0.98]). There cases of extrapyramidal side effects in the dexamethasone
was no evidence of heterogeneity among the trials (I2 ϭ 0). and placebo groups.16
In a subgroup analysis by type of surgery, dexamethasone
compared with placebo reduced the need for postoperative
rescue analgesia in patients undergoing total abdominal DISCUSSION
hysterectomy (28% vs 40%, RR [95% CI] ϭ 0.74 [0.56, 0.99]) The results of this systematic review indicate that a single
but not after cesarean delivery. In a subgroup analysis by dose of IV dexamethasone is an effective antiemetic in
route of administration, dexamethasone was effective in patients receiving neuraxial morphine as part of a neuraxial
significantly reducing the use of postoperative rescue an- anesthetic technique. In addition, dexamethasone pro-
algesia after epidural morphine administration (26% vs duced a small but statistically significant reduction in
37%, RR [95% CI] ϭ 0.74 [0.59, 0.95]). However, in patients 24-hour postoperative pain scores and reduced the need for
receiving intrathecal morphine, no conclusion could be rescue analgesics when compared with placebo. Dexameth-
made regarding the use of postoperative rescue analgesics asone, however, did not reduce the incidence of neuraxial
between the treatment and placebo groups (35% vs 43%, RR morphine–induced pruritus when compared with placebo.
[95% CI] ϭ 0.81 [0.55, 1.18]). Dexamethasone has a well-established role as a prophy-
lactic antiemetic in children and adults receiving general
Pruritus anesthesia.1– 4,24 However, PON and POV are also frequent
Six studies reported the incidence of pruritus associated adverse effects after neuraxial morphine administration.
with neuraxial morphine in patients receiving dexametha- Dahl et al.6 reported that the number needed to harm for
sone or placebo.11–15,17 Overall, dexamethasone did not PON and POV in patients receiving intrathecal morphine
significantly reduce the incidence of pruritus when com- for cesarean delivery under spinal anesthesia was 6.3 and
pared with placebo (Table 3). 10.1, respectively. The mechanism for the antiemetic effect
Four studies reported the use of rescue antipruritic of dexamethasone remains unknown. The ability of dexa-
medication after the administration of neuraxial mor- methasone to deplete ␥-aminobutyric acid stores, reduce
phine.11,12,14,17 IM diphenhydramine 20 mg was the rescue the blood-brain barrier’s permeability to emetic toxins,
antipruritic in 3 studies12–14 and oral chlorpheniramine 4 inhibit brainstem enkephalin release, central prostaglandin
mg was the rescue drug in the remaining study.17 Overall, synthesis, and serotonin synthesis and release are among
no conclusion could be made regarding the impact of the proposed mechanisms.25,26 Its long duration of action
dexamethasone on the use of rescue antipruritic medication makes it an ideal drug for prophylaxis in patients receiving
when compared with placebo because of the wide CIs of long-acting neuraxial opioids.1 In this patient population,
pooled results (Table 3). There was no evidence of hetero- dexamethasone reduced the overall incidence of PON. The
geneity for the trials included for both outcomes. lowest effective dose for the prevention of PON and POV
was 5 mg. The lowest effective dose of dexamethasone for
Side Effects PONV prophylaxis in patients receiving neuraxial mor-
Seven of the 8 studies investigated and reported possible phine has not been established but at least one study found
side effects of dexamethasone.11–13,15–18 Side effects were doses as low as 2.5 mg to be an effective antiemetic in adult
April 2012 • Volume 114 • Number 4 www.anesthesia-analgesia.org 819

8. Dexamethasone and Neuraxial Morphine–Induced PONV
patients receiving general anesthesia.27 Published consen- Pruritus is a common adverse effect of neuraxial mor-
sus guidelines for the management of PONV recommend phine but currently there seem to be no consistently
doses of 4 to 5 mg for antiemetic prophylaxis.24 effective therapies.36 The lack of an antipruritic effect of
Interestingly, dexamethasone was an effective anti- dexamethasone is disappointing in light of its antiemetic
emetic in patients receiving epidural morphine but ineffec- and antiinflammatory properties. However, dexametha-
tive in patients receiving intrathecal morphine. However, sone administration has been associated with perineal
based on the small sample size of the pooled results in the pruritus, a poorly understood phenomenon that is associ-
intrathecal morphine subgroup analysis, the power to ated with bolus administration and more frequently re-
detect a difference may have been inadequate (type II ported in females.37,38
error). We pooled both epidural and intrathecal adminis- This meta-analysis has several limitations. We identified
tration of morphine for this meta-analysis based on studies evidence of publication bias. Of the 8 trials included in this
that reported no difference in the incidence of PONV when review, 6 were able to demonstrate some degree of anti-
comparing comparable doses using both routes.28,29 In this emetic efficacy of single-dose dexamethasone. The causes
review, the overall incidence of PON and POV in the of publication bias have been highlighted in previous
control groups in both trials in which intrathecal morphine publications and its presence may make the validity of the
was administered was 58% and 55%, respectively, which is findings of this review questionable.3,10 However, tests of
much higher than the corresponding incidence of 31% and publication bias are unreliable in meta-analyses based
24%, respectively, in the trials investigating epidural mor- exclusively on multiple small trials and therefore the re-
phine. It is possible that the mechanism of PONV in sults of these tests should be interpreted with caution.10
patients receiving intrathecal and epidural opioids may not Another limitation is that 7 of the 8 studies were performed
be similar. In addition, the pharmacokinetics of both routes in Taiwan, Republic of China. In fact, the studies seemed to
of administration differs significantly, potentially leading have been performed at the same institution and the study
to more rapid rostral spread and higher intrathecal concen- population also only included female patients. The results
trations after direct intrathecal administration compared may therefore not be applicable to the male population or
with administration by the epidural route.5,30,31 The intra- those of different ethnicity. With respect to the study
thecal morphine dose may also have had a role. Doses of design, in the majority of the studies, dexamethasone was
intrathecal morphine as high as 0.2 mg are associated with administered at the end of surgery or postoperatively. This
an increase in the incidence of PONV, compared with lower is in contradiction to evidence suggesting a superior anti-
doses, without providing any additional analgesic effi- emetic efficacy of dexamethasone when administered at the
cacy.32,33 The timing of administration of dexamethasone in induction of anesthesia rather than at the end of surgery in
relation to the administration of neuraxial morphine may patients receiving general anesthesia.24,34 The optimal tim-
also have accounted for the observed difference. Dexameth- ing of administration of dexamethasone in patients receiv-
asone was administered before morphine in the epidural ing neuraxial anesthesia has not been determined but is
studies but after its administration in the intrathecal stud- likely similar to general anesthesia given the slow onset of
ies. With the onset of action of the antiemetic effect of action of dexamethasone. Pain was a secondary end point
dexamethasone estimated to be 2 hours and the possibility in the included trials and details about postoperative
of a more rapid onset of emetic symptoms after intrathecal analgesia were mainly reported as number of patients
morphine administration, it may have been more prudent requiring rescue and not consistently as doses of rescue
to administer the prophylactic drug before initiation of drugs given. Overall, side effects were poorly reported in
spinal anesthesia.31,34 the included trials. Adverse effects such as wound infection
Dexamethasone reduced pain scores and need for rescue and delayed wound healing are of interest but were inad-
analgesics in this systematic review. None of the trials equately reported. Increased risk of bleeding after a single
individually reported any reduction in postoperative pain dose of dexamethasone in pediatric tonsillectomies has also
intensity or need for rescue analgesia between the treat- been reported, but this risk was not explored in any of the
ment and placebo groups. However, analgesic outcomes studies.39 Future studies should accurately report side
were not a primary end point in any of the studies and the effects so that the overall safety of dexamethasone can be
studies were probably not powered to detect this end point. established.
The reduction in 24-hour pain scores highlights the slow Further research is needed to determine the role of
onset and long duration of action of dexamethasone and a dexamethasone as an antiemetic in patients receiving intra-
possible synergistic action between neuraxial morphine thecal morphine after neuraxial anesthesia. With the popu-
and dexamethasone. The administration of a single dose of larity of intrathecal morphine for postcesarean analgesia,
dexamethasone has been associated with a reduction in the efficacy of dexamethasone as an antiemetic, particularly
tissue inflammatory mediators, including bradykinin, pros- in this at-risk population, needs further investigation. Stud-
taglandins, and other nociception-promoting neuropep- ies investigating dexamethasone in combination with other
tides; this may contribute to its analgesic properties.35 pharmacological and nonpharmacological antiemetic
Dexamethasone has been shown to significantly reduce therapy are also needed. Dexamethasone’s role as an anal-
postoperative peritoneal inflammation and abdominal pain gesic has been extensively investigated in patients receiving
after colectomy.35 However, although the overall reduction general anesthesia but its role as an analgesic adjunct in
in late pain scores was statistically significant, it was small patients receiving neuraxial anesthesia remains less clear.
and may not be clinically relevant, especially in patient Studies investigating the analgesic effect of dexamethasone
populations in which the pain scores were low. in this patient population as a primary outcome are needed.
820 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

9. In conclusion this systematic review presents relatively 7. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred report-
strong evidence that a single IV dose of dexamethasone 5 to ing items for systematic reviews and meta-analyses: the
PRISMA statement. J Clin Epidemiol 2009;62:1006 –12
10 mg was an effective antiemetic for women receiving 8. Higgins JPT, Green S, eds. Cochrane Handbook for System-
neuraxial morphine for cesarean delivery or abdominal atic Reviews of Interventions. Version 5.0.2 (updated Sep-
hysterectomy. Although this review suggests that dexa- tember 2009). The Cochrane Collaboration, 2008. Available
methasone may not be effective in patients receiving intra- at: www.cochrane-handbook.com; accessed March 30, 2011
9. Hozo S, Djulbegovic B, Hozo I. Estimating the mean and
thecal morphine, this should be interpreted with caution
variance from the median, range, and the size of a sample.
because of the small number of patients receiving intrathe- BMC Med Res Methodol 2005;5:13
cal morphine that were included in the analysis. There is 10. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis
also some evidence that the doses used for antiemetic detected by a simple, graphical test. BMJ 1997; 315:629–34
prophylaxis also improved postoperative analgesia, al- 11. Wu JI, Lo Y, Chia YY, Liu K, Fong WP, Yang LC, Tan PH.
Prevention of postoperative nausea and vomiting after intra-
though the reduction in pain scores was small and may not thecal morphine for cesarean section: a randomized compari-
be clinically significant. There was no evidence that dexa- son of dexamethasone, droperidol, and a combination. Int J
methasone was an effective antipruritic. Further studies are Obstet Anesth 2007;16:122–7
needed to investigate this drug’s safety profile and its 12. Wang JJ, Tzeng JI, Ho ST, Chen JY, Chu CC, So EC. The
prophylactic effect of tropisetron on epidural morphine-related
efficacy in patients receiving intrathecal morphine. In light
nausea and vomiting: a comparison of dexamethasone with
of possible publication bias, these findings need to be saline. Anesth Analg 2002;94:749 –53
interpreted with caution. 13. Wang JJ, Ho ST, Wong CS, Tzeng JI, Liu HS, Ger LP. Dexa-
methasone prophylaxis of nausea and vomiting after epidural
morphine for post-cesarean analgesia. Can J Anaesth 2001;
DISCLOSURES 48:185–90
Name: Terrence K. Allen, MBBS, FRCA. 14. Wang JJ, Ho ST, Liu YH, Ho CM, Liu K, Chia YY. Dexameth-
Contribution: This author helped design the study, analyze asone decreases epidural morphine-related nausea and vomit-
the data, and write the manuscript. ing. Anesth Analg 1999;89:117–20
15. Tzeng JI, Wang JJ, Ho ST, Tang CS, Liu YC, Lee SC. Dexameth-
Attestation: Terrence K. Allen has seen the original study data, asone for prophylaxis of nausea and vomiting after epidural
reviewed the analysis of the data, approved the final manu- morphine for post-cesarean section analgesia: comparison of
script, and is the author responsible for archiving the study droperidol and saline. Br J Anaesth 2000;85:865– 8
files. 16. Tzeng JI, Hsing CH, Chu CC, Chen YH, Wang JJ. Low-dose
Name: Cheryl A. Jones, MD, DVM. dexamethasone reduces nausea and vomiting after epidural
Contribution: This author helped conduct the study and write morphine: a comparison of metoclopramide with saline. J Clin
Anesth 2002;14:19 –23
the manuscript. 17. Nortcliffe SA, Shah J, Buggy DJ. Prevention of postoperative
Attestation: Cheryl A. Jones has seen the original study data, nausea and vomiting after spinal morphine for caesarean
reviewed the analysis of the data, and approved the final section: comparison of cyclizine, dexamethasone and placebo.
manuscript. Br J Anaesth 2003;90:665–70
Name: Ashraf S. Habib, MBBCh, MSc, MHS, FRCA. 18. Ho ST, Wang JJ, Tzeng JI, Liu HS, Ger LP, Liaw WJ. Dexameth-
Contribution: This author helped design the study, conduct asone for preventing nausea and vomiting associated with
epidural morphine: a dose-ranging study. Anesth Analg
the study, and write the manuscript. 2001;92:745– 8
Attestation: Ashraf S. Habib has seen the original study data, 19. Clarke H, Pereira S, Kennedy D, Andrion J, Mitsakakis N,
reviewed the analysis of the data, and approved the final Gollish J, Katz J, Kay J. Adding gabapentin to a multimodal
manuscript. regimen does not reduce acute pain, opioid consumption or
This manuscript was handled by: Cynthia A. Wong, MD. chronic pain after total hip arthroplasty. Acta Anaesthesiol
Scand 2009;53:1073– 83
20. Mathiesen O, Jacobsen LS, Holm HE, Randall S, Adamiec-
REFERENCES Malmstroem L, Graungaard BK, Holst PE, Hilsted KL, Dahl JB.
1. Henzi I, Walder B, Tramer MR. Dexamethasone for the pre- Pregabalin and dexamethasone for postoperative pain control:
vention of postoperative nausea and vomiting: a quantitative a randomized controlled study in hip arthroplasty. Br J An-
systematic review. Anesth Analg 2000;90:186 –94 aesth 2008;101:535– 41
2. Apfel CC, Korttila K, Abdalla M, Kerger H, Turan A, Vedder I, 21. Tan PH, Liu K, Peng CH, Yang LC, Lin CR, Lu CY. The effect
Zernak C, Danner K, Jokela R, Pocock SJ, Trenkler S, Kredel M, of dexamethasone on postoperative pain and emesis after
Biedler A, Sessler DI, Roewer N; the IMPACT Investigators. A intrathecal neostigmine. Anesth Analg 2001;92:228 –32
factorial trial of six interventions for the prevention of postop- 22. Movafegh A, Soroush AR, Navi A, Sadeghi M, Esfehani F,
erative nausea and vomiting. N Engl J Med 2004;350:2441–51 Akbarian-Tefaghi N. The effect of intravenous administration
3. Bolton CM, Myles PS, Nolan T, Sterne JA. Prophylaxis of post- of dexamethasone on postoperative pain, nausea, and vomit-
operative vomiting in children undergoing tonsillectomy: a sys- ing after intrathecal injection of meperidine. Anesth Analg
tematic review and meta-analysis. Br J Anaesth 2006;97: 593– 604 2007;104:987–9
4. Karanicolas PJ, Smith SE, Kanbur BM, Davies E, Guyatt GH. 23. Szarvas S, Chellapuri RS, Harmon DC, Owens J, Murphy D,
The impact of prophylactic dexamethasone on nausea and Shorten GD. A comparison of dexamethasone, ondansetron
vomiting after laparoscopic cholecystectomy: a systematic re- and dexamethasone plus ondansetron as prophylactic anti-
view and meta-analysis. Ann Surg 2008;248:751– 62 emetic and antipruritic therapy in patients receiving intrathe-
5. Gadsden J, Hart S, Santos AC. Post-cesarean delivery analge- cal morphine for major orthopedic surgery. Anesth Analg
sia. Anesth Analg 2005;101:S62–9 2003;97:259 – 63
6. Dahl JB, Jeppesen IS, Jorgensen H, Wetterslev J, Moiniche S. 24. Gan TJ, Meyer TA, Apfel CC, Chung F, Davis PJ, Habib AS,
Intraoperative and postoperative analgesic efficacy and ad- Hooper VD, Kovac AL, Kranke P, Myles P, Philip BK, Samsa G,
verse effects of intrathecal opioids in patients undergoing Sessler DI, Temo J, Tramer MR, Vander Kolk C, Watcha M.
`
cesarean section with spinal anesthesia: a qualitative and Society for Ambulatory Anesthesia guidelines for the manage-
quantitative systematic review of randomized controlled trials. ment of postoperative nausea and vomiting. Anesth Analg
Anesthesiology 1999;91:1919 –27 2007;20:1615–28
April 2012 • Volume 114 • Number 4 www.anesthesia-analgesia.org 821

10. Dexamethasone and Neuraxial Morphine–Induced PONV
25. Glare PA, Dunwoodie D, Clark K, Ward A, Yates P, Ryan S, 34. Wang JJ, Ho ST, Tzeng JI, Tang CS. The effect of timing of
Hardy JR. Treatment of nausea and vomiting in terminally ill dexamethasone administration on its efficacy as a prophylactic
cancer patients. Drugs 2008;68:2575–90 antiemetic for postoperative nausea and vomiting. Anesth
26. Jordan K, Schmoll HJ, Aapro MS. Comparative activity of Analg 2000;91:136 –9
antiemetic drugs. Crit Rev Oncol Hematol 2007;61:162–75 35. Zargar-Shoshtari K, Sammour T, Kahokehr A, Connolly AB,
27. Liu K, Hsu CC, Chia YY. The effective dose of dexamethasone Hill AG. Randomized clinical trial of the effect of glucocorti-
for antiemesis after major gynecological surgery. Anesth Analg coids on peritoneal inflammation and postoperative recovery
1999;89:1316 – 8 after colectomy. Br J Surg 2009;96:1253– 61
28. Duale C, Frey C, Bolandard F, Barriere A, Schoeffler P. 36. George RB, Allen TK, Habib AS. Serotonin receptor antago-
Epidural versus intrathecal morphine for postoperative anal- nists for the prevention and treatment of pruritus, nausea, and
gesia after cesarean section. Br J Anaesth 2003;91:690 – 4 vomiting in women undergoing cesarean delivery with intra-
29. Sarvela J, Halonen P, Soikkeli A, Korttila K. A double-blinded, thecal morphine: a systematic review and meta-analysis.
randomized comparison of intrathecal and epidural morphine Anesth Analg 2009;109:174 – 82
for elective cesarean delivery. Anesth Analg 2002;95:436 – 40 37. Perron G, Dolbec P, Germain J, Bechard P. Perineal pruritus
´
30. Saldman LJ, Cousins MJ, Mather LE. Intrathecal and epidural after iv dexamethasone administration. Can J Anaesth
administration of opioids. Anesthesiology 1984;61:276 –310 2003;50:749 –50
31. Baraka A, Noueihid R, Hajj S. Intrathecal injection of morphine 38. Taleb N, Geahchan N, Ghosn M, Brihi E, Sacre P. Vulvar
for obstetric analgesia. Anesthesiology 1981;54:136 – 40 pruritus after high-dose dexamethasone. Eur J Cancer 1988;
32. Uchiyama A, Nakano S, Ueyama H, Nishimura M, Tashiro C. 24:495
Low dose intrathecal morphine and pain relief following 39. Czarnetzki C, Elia N, Lysakowski C, Dumont L, Landis BN,
caesarean section. Int J Obstet Anesth 1994;3:87–91 Giger R, Dulguerov P, Desmeules J, Tramer MR. Dexametha-
33. Milner AR, Bogod DG, Harwood RJ. Intrathecal administration sone and risk of nausea and vomiting and postoperative
of morphine for elective caesarean section. Anaesthesia 1996; bleeding after tonsillectomy in children: a randomized trial.
51:871–3 JAMA 2008;300:2621–30
822 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA