Gabapentin prophylaxis for postoperative nauseaand vomiting in abdominal surgeries: a quantitative analysisof evidence from randomized controlled clinical trialsS. Achuthan1, I. Singh1*, S. B. Varthya1, A. Srinivasan1, A. Chakrabarti1 and D. Hota2

1 Department of Pharmacology, Department of Pharmacology, Postgraduate Institute of Medical Education and Research (PGIMER),Chandigarh 160012, India2 Department of Pharmacology, AIIMS, Bhubaneshwar, India

* Corresponding author. E-mail: inderhazari@gmail.com

Editor’s key points

† The authors examined theevidence base for theeffectiveness ofgabapentin in preventingpostoperative nausea andvomiting (PONV).

† They found support for theuse of preoperativegabapentin in PONVprophylaxis, especially inabdominal surgery.

† Interestingly, the benefitsof gabapentin appearedreduced in the presenceof the use of propofol.

Introduction. Postoperative nausea and vomiting (PONV) is frequently encountered in thesurgical recovery room. Abdominal surgery is one important risk factor for increasedincidence of PONV. Gabapentin, an anticonvulsant with known postoperative analgesicproperties, has shown some activity against PONV. Results from clinical trials evaluating theanti-emetic efficacy of gabapentin are conflicting. The present meta-analysis wasperformed to examine this issue.

Methods. Seventeen randomized placebo-controlled trials reporting PONV with preoperativegabapentin administration in patients undergoing abdominal surgery were included foranalysis. Outcomes evaluated were nausea, vomiting, composite PONV and the use ofrescue anti-emetic medication in the postoperative period.

Results. The pooled relative risk (RR), estimated using the random effects model of themetafor package for R, was 0.76 (95% CI 0.58–0.98) for nausea, 0.62 (0.45–0.85) forvomiting, 0.71 (0.39–1.28) for data represented as composite PONV (possibly biased by asingle study, as observed in the sensitivity analysis), and 0.6 (0.41–0.89) for rescue anti-emetic use. There was a significant RR reduction for nausea and vomiting when propofolwas not used as induction and/or maintenance for anaesthesia. In the abdominalhysterectomy subgroup, there was a significant RR reduction for vomiting but not for nausea.

Discussion. The present analysis provides evidence supporting preoperative gabapentin as apharmacotherapy for prevention of PONV in patients undergoing abdominal surgeries.Future studies comparing preoperative gabapentin with 5HT3 antagonists are needed toprecisely define its role in PONV.

Keywords: gabapentin; hysterectomy; postoperative nausea and vomiting

Accepted for publication: 19 September 2014

Postoperative nausea and vomiting (PONV) in the surgical re-covery room is a common occurrence. Despite being self-limiting, it is associated with substantial distress to patientsin the early postoperative period. Its prevention and/or treat-ment significantly improves patient satisfaction and qualityof life.1 – 3 PONV prevention can lead to a substantial reductionin health care costs by reducing the duration of the post-anaesthesia care unit (PACU) stay.4 5 It is estimated thateach episode of emesis delays discharge from the PACU byapproximately 20 min.6 Furthermore, serious complications,including retinal detachment, aspiration, wound dehiscence,oesophageal rupture, and subcutaneous emphysema, relatedto PONV can be prevented. The general incidence of PONVranges from 25% to 30%,7 and in some subsets of high-riskpatients it can be as high as 80%,8 even after preventive

pharmacotherapy. The pharmacotherapies used in the preven-tion/management of PONV include 5-HT3, dopaminergic, hista-minic, and NK1 antagonists. Nevertheless, the need for cheaperand more effective therapies cannot be disputed.

Gabapentin, an anticonvulsant, has been shown to be ef-fective as a non-opioid ancillary agent for postoperative painmanagement.9 The safety profile, minimal drug interactions,pharmacokinetics (good oral bioavailability and renal elimin-ation), and relatively flat dose–response relationship withrespect to safety and efficacy favour its clinical use.10 11 Inter-estingly, in the recent past, gabapentin’s anti-nausea effectswere noted when there was a marked improvement in thechemotherapy-induced nausea in breast cancer patients12

and subsequently was observed to possess activity againstPONV. The exact mechanism for gabapentin’s efficacy in

& The Author 2015. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.For Permissions, please email: journals.permissions@oup.com

British Journal of Anaesthesia 114 (4): 588–97 (2015)Advance Access publication 8 January 2015 . doi:10.1093/bja/aeu449

0107/04/2015

Mdl. User

PONV is not known. While results of a few studies evaluatingthe efficacy of preoperative gabapentin as a prophylaxis werepromising, the results of others did not reach statistical signifi-cance and have been equivocal. Hence the present analysiswas performed to define the role of gabapentin as a preventivetherapy for PONV.

MethodsRandomized placebo-controlled clinical trials comparing pre-operative gabapentin administration were identified for evalu-ation for inclusion in the analyses. The criteria for inclusionwere (i) patients undergoing abdominal surgery (open or lap-aroscopic) under general anaesthesia, (ii) preoperative admin-istration of gabapentin irrespective of dose and timing of thedose with respect to surgery, and (iii) trials reporting nausea,vomiting, postoperative nausea and vomiting, or a proportionof patients requiring rescue anti-emetic medication irrespect-ive of the objective of evaluation were included in the finalanalysis. Trials evaluating postoperative dosing alone or in add-ition to preoperative gabapentin were excluded from the finalanalysis. A literature search using the terms ‘gabapentin’,‘postoperative nausea and vomiting’, ‘abdominal surgery’,and ‘postoperative pain’ was performed in Medline, Embaseand the Cochrane library. Additional studies were identifiedthrough cross-references and meta-analyses conducted forevaluation of gabapentin in postoperative pain. All humanstudies published in full-text or abstract forms were initiallyscreened for inclusion without applying any language restric-tions. Titles and abstracts from the electronic search werereviewed. After initial review of the abstracts, the relevantstudies were identified and a detailed evaluation of the fulltext was done. All data with regard to authorship, year of pub-lication, study design, study population (i.e. type of abdominalsurgery), baseline patient characteristics, gabapentin dose,type of surgery, type and duration of anaesthesia (when datawere not available, the duration of surgery was taken as theduration of anaesthesia), number of patients, and relevantoutcomes were extracted from the selected studies. The meth-odological quality of the included studies was assessed usingthe Downs and Black score.13 The whole process was under-taken independently by two authors and all the conflictswere resolved by discussion among all authors.

Outcomes

The outcomes evaluatedwere postoperative nausea, vomiting,composite of nausea and vomiting (PONV, as some studieshad reported a composite outcome of nausea and vomiting),and the proportion of patients requiring rescue anti-emeticmedication.

Statistical analyses

Statistical analyses were performed using the metaforpackage for R.14 For each individual study the relative risk(RR) was calculated using the reported events of relevantoutcomes. Pooled RR for individual outcomes was estimatedusing the random effects model of the metafor package.

Heterogeneity was assessed based on the calculated I2 (theproportion of total variability explained by heterogeneity), esti-mated using the restricted maximum likelihood–basedmethod. A moderator analysis was performed using themixed effects modelling approach (incorporating weightedleast squares regression for the moderator in the randomeffects model) to assess the contribution of study level covari-ates to the overall heterogeneity and their impact on the effectsize. Age, percentage of females in each study, postoperativeopioid consumption, and duration of anaesthesia were thecontinuous variables, while surgery type and anaestheticagent used, dose of gabapentin, and timing of the dose werethe categorical variables used as covariates in the mixedeffects model. The effect of postoperative opioid consumptionwas evaluated after computing the standardized mean differ-ence between the placebo and gabapentin groups. To assessthe publication bias we plotted the log relative risk vs the stand-ard error of the individual studies. The symmetry of the plotwas assessed using Egger’s test. The trim and fill method wasused to estimate the number of studies missing from themeta-analysis due to the suppression of the most extremeresults on one side of the funnel plot. A Galbraith plot wasalso plotted to assess the same. A sensitivity analysis wasconducted using leave-one-out analysis. Briefly, in this proced-ure individual studies were excluded from the model and theinfluence of the exclusion of that study on the model para-meters was assessed.

ResultsSeventeen studies were included in the final analysis. The lit-erature search and study selection is represented in Fig. 1.The characteristics of the included studies are shown inTable 1. The included studies differed with respect to the typeof surgery and the gabapentin dose evaluated. The selectedstudies contained a total of 1605 patients, with 810 in thegabapentin group and 795 in the placebo group. Eight studieswere conducted in patients undergoing hysterectomy, threestudies each for open and laparoscopic cholecystectomypatients, and one study each with donor nephrectomy, majorbowel surgery, and laparoscopic-assisted reproductive surgerypatients. Gabapentin was administered either 1 or 2 h priorto surgery in all the studies. In two studies an additional doseof gabapentin was administered the night before surgery.The data primarily collected from the studies were the studydesign; gabapentin dose; number of nausea, vomiting, andPONV events; number of rescue anti-emetic events; and alsothe induction and maintenance agents used for anaesthesia.

Outcomes

Postoperative nausea

Ten studies comprising a total of 632 patients (324 in the gaba-pentin group and 308 in the placebo group) were included forpooling the RR for postoperative nausea. The pooled RR was0.76 (95% CI 0.58–0.98). I2 was 5.8% and the test for hetero-geneity was not significant (P¼0.36) (Fig. 2).

Gabapentin and postoperative nausea and vomiting BJA

589

Postoperative vomiting

Ten studies comprising a total of 632 patients (324 in the gaba-pentin group and 308 in the placebo group) were included forpooling the RR for postoperative vomiting. The pooled RR was0.62 (95% CI 0.45–0.85). I2 was 0% and the test for heterogen-eity was not significant (P¼0.5) (Fig. 3).

PONV

Seven studies comprising a total of 973 patients (487 each inthe gabapentin and placebo groups) were included forpooling the RR for postoperative nausea. The pooled RR was0.71 (95% CI 0.39–1.28). I2 was 94% and the test for hetero-geneity was significant (P,0.0001) (Fig. 4).

Postoperative use of anti-emetic rescue medication

Seven studies comprising a total of 578 patients (289 each inthe gabapentin and placebo groups) were included forpooling the RR for postoperative nausea. The pooled RR was0.6 (95% CI 0.41–0.89). I2 was 70% and the test for heterogen-eity was significant (P¼0.0045) (Fig. 5).

Moderator analysis

Continuous moderators, such as study level, differences inage, percentage of females, standardized mean difference inopioid consumption, and the duration of anaesthesia, did notinfluence the outcome measures in the univariate moderatoranalysis for nausea, vomiting, PONV, and rescue anti-emeticuse. The results of the categorical moderators are displayedin Supplementary Figs. 6 and 7. There was a significant reduc-tion in vomiting when gabapentin was used in abdominalhysterectomies. The RR reduction was not statistically signifi-cant in the cholecystectomy, nephrectomy, bowel surgeryand laparoscopic-assisted reproductive techniques (LART)subgroups. The type of the surgery was not associated with areduction in RR when nausea was analysed. There was also asignificant reduction in the RR of both nausea and vomitingwhen propofol was not used as an induction or maintenanceagent.

Sensitivity analysis and publication bias

The influence diagnostics and leave-one-out analysis of thestudies was conducted to assess whether the effect size of

Records identified through pubmeddatabase searching

(n=175)

Records identified through other sources (Crossreferences, EMBASE, Cochrane library)

(n=56)

Records after duplicates removed(n=196)

Articles assessed for eligibility(n=42)

Records excluded(n=154)

Studies included in qualitativesynthesis

(n=17)

Studies included in quantitativesynthesis (meta-analysis)

Outcomes (n)Nausea (10)Vomiting (10)

PONV (7)Rescue antiemetic use (6)

Excluded-2511- Abdominal surgery but not fitting

the criteria8- Orthopedic surgery

5- Other surgeries1- No placebo control

Fig 1 Flow diagram of the search strategy and study selection.

BJA Achuthan et al.

590

Table 1 Details of the included studies. AH, abdominal hysterectomy; DN, donor nephrectomy; H, halothane; I, isoflurane; LART, laparoscopic-assisted reproductive techniques; LC, laparoscopiccholecystectomy; MBS, major bowel surgery; NR, not reported; OC, open cholecystectomy; P, propofol; PONV, postoperative nausea and vomiting; S, sevoflurane; T, thiopentone. *Downs and Blackscore. †1, 1 h prior to surgery; 2, 2 h prior to surgery.

Study Ajoriet al.16

Durmuset al.17

Frouzanfardet al.18

Khademiet al.19

Mohammadiet al.20

Pandeyet al.21

Pandeyet al.22

Pandeyet al.23

Senet al.24

Srivastavaet al.25

Takmazet al.26

Turanet al.27

Ghafariet al.28

Ghaiet al.29

Behdadet al.30

Siddiquiet al.31

Bashiret al.32

Year 2012 2007 2013 2010 2008 2004 2005 2006 2009 2010 2007 2004 2009 2011 2012 2014 2009

Surgery AH AH AH OC LART LC DN LC AH OC OC AH AH AH AH MBS LC

Randomization Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Blinded Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Quality score* 21 22 21 22 19 19 20 19 23 23 19 21 22 22 21 21 22

Gabapentindose (mg)

600 1200 1200 600 300 300 600 600 1200 1200 900/1200

1200 300+300 900 100+300 600 600

Timing of dose† 1 1 2 2 1 2 2 2 1 2 1 2 1 1–2 2 1 2

Sample size (n)

Placebo 69 25 25 43 35 153 20 125 20 60 15 25 33 30 31 36 50

Gabapentin 69 25 25 44 35 153 20 125 20 60 15/15 25 33 30 30 36 50

Mean age,years(placebo)

48.3 48 44.6 52.18 31.9 43.49 41.5 41.8 46 44.7 43.5 50.4 48.16 43 44.55 37.2 NR

Opioid consumption, mean (SD)

Gabapentin 1.4 (0.6) 42.9(10.09)

1.2 (2.8) 28.33(12.9)

NR 221.16(52.39)

563.3(252.8)

221.2(92.4)

31 (12) 253.9(44.8)

165.1(11.1)

270.4(144.4)

15.78(1.15)

54.40(15.57)

NR NR NR

Placebo 2.5 (0.7) 66.45(10.93)

5.2 (2.8) 35.1(15.1)

NR 355.86(42.04)

924.7(417.5)

505.9(82)

48 (17) 375.8(83.5)

456(35.6)

419.6(83.6)

26.94(2.28)

105(23.6)

NR NR NR

Nausea, n (%)

Gabapentin 8 (11.6) 7 (28.0) NR NR 2 (5.7) NR 4 (20.0) NR 9 (45.0) NR 2 (13.3) 5(20.0)

5 (15.2) NR 5 (16.1) 17 (47.2) NR

Placebo 19 (27.5) 9 (36.0) NR NR 9 (25.7) NR 3 (15.0) NR 8 (40.0) NR 9 (60.0) 7(28.0)

7 (21.2) NR 5 (16.7) 17 (47.2) NR

Vomiting, n (%)

Gabapentin 5 (7.2) 3 (12.0) NR NR 0 (0.0) NR 1 (5.0) NR 8 (40.0) NR 1 (6.7) 6(24.0)

4 (12.1) NR 5 (16.1) 5 (13.9) NR

Placebo 16 (23.2) 6 (24.0) NR NR 4 (11.4) NR 2 (10.0) NR 8 (40.0) NR 7 (46.7) 9(36.0)

9 (27.3) NR 7 (23.3) 6 (16.7) NR

PONV, n (%)

Gabapentin NR NR 7 (28) 16 (36.4) NR 38(24.8)

NR 46(36.8)

NR 15 (25.0) NR NR NR 21 (70.0) NR NR 20 (40)

Placebo NR NR 24 (96.0) 28 (65.1) NR 8 (5.2) NR 75(60.0)

NR 31 (51.7) NR NR NR 29 (96.7) NR NR 31 (62)

Rescue medication use, n (%)

Gabapentin 7 (10.1) NR NR NR 0 (0.0) NR NR NR 8 (40.0) 9 (15.0) 2 (13.3) 11 NR NR NR NR 17 (34)

Placebo 19 (27.5) NR NR NR 4 (11.4) NR NR NR 7 (35.0) 23 (38.3) 11 (73.3) 16 NR NR NR NR 27 (54)

Induction agent

T T T T P P P P P P T P T P P P P

Maintenance anaesthetic agent

I S I I P P P P S H S S I S p S P

Gabapen

tinan

dpostoperative

nau

seaan

dvom

iting

BJA591

gabapentin on the four main parameters, i.e. nausea, vomiting,PONVand rescue anti-emetic use, was influenced by individualstudies. The pooled RR for PONV was not significant [0.71 (95%CI 0.39–1.28)], but when the study by Pandey and colleagues21

was excluded, the RR was significant [0.59 (95% CI 0.5–0.7)]and the I2 value for heterogeneity decreased from 94 to27.2% (Supplementary Table 2). The Galbraith plot showedno evidence of publication bias (Supplementary Figure 8). TheEgger’s test for asymmetry of the funnel plot was non-significant (P¼0.27). The trim and fill method estimated thenumber of missing studies in the funnel plot as one.

DiscussionThe quantitative analysis of the collected data showedthat preoperative administration of gabapentin in patients

undergoing abdominal surgery was associated with a lower in-cidence of nausea and vomiting. We observed a reduction of24% for nausea, 38% for vomiting, no significant reduction inthe composite PONV (possibly biased by a single study), anda 40% reduction in the rescue anti-emetic use in patientswho received preoperative gabapentin. An earlier analysis byAlayed and colleagues15 also showed that there was a reduc-tion in the incidence of postoperative nausea and vomiting.However, the analysis was focussed on only abdominal hyster-ectomy patients (11 studies). Also, we consider that combiningthe composite PONV data in the analysis of both nauseaand vomiting outcomes as performed in that analysis couldpossibly result in biased estimates. In the present analysiswe included studies on all abdominal surgeries16 – 32 thatevaluated only preoperative gabapentin (even postoperativegabapentin use was included in the earlier analysis). We

Pooled Relative Risk for Nausea

Favors Gabapentin Favors Placebo

PlaceboGabapentin

Studies Events EventsSample SampleGabapentin Weights RR

(n) (N) Dose (mg) (n) (N) (95% Cl)

Mohamaddi et al

Behdad et al

Ajori et al

Pandey et al*

Ghafari et al

Siddiqui et al

Takmaz et al

Durmus et al

Sen et al

Takmaz et al

Turan et al

RE Model

2

5

8

4

5

17

7

7

9

2

5

0.01

9

5

19

3

7

17

9

9

8

9

0.22 (0.05, 0.96)

0.97 (0.31, 3.01)

0.42 (0.20, 0.90)

1.33 (0.34, 5.21)

0.71 (0.25, 2.02)

1.00 (0.61, 1.63)

0.78 (0.39, 1.54)

0.78 (0.34, 1.76)

1.12 (0.55, 2.32)

0.22 (0.06, 0.86)

7

35

30

69

20

33

36

15

25

20

15

25 0.71 (0.26, 1.95)

3.19%

5.22%

11.28%

3.64%

6.14%

24.61%

13.68%

9.74%

12.24%

3.69%

6.58%

0.76 (0.58, 0.98)100.00%

35

31

69

20

33

36

15

25

20

15

25

300

400

600

600

600

600

900

1200

1200

1200

1200

0.10 0.50 4.00

Relative risk (RR)

Fig 2 Pooled analysis of postoperative nausea.

BJA Achuthan et al.

592

additionally performed subgroup analysis to determine thefactors that could influence the effect of gabapentin on post-operative nausea and vomiting (which was not performed inthe earlier analysis).

Previous studies have shown that the most reliable inde-pendent predictors of PONV were female gender, history ofPONV or motion sickness, non-smoker, younger age, durationof anaesthesia with volatile anaesthetics, and postoperativeopioids.33 34 From the extracted data we were able to analysewhether age, female gender, duration of anaesthesia andpostoperative opioid use influence the effect of gabapentinon postoperative nausea and vomiting. None of these factorshad a significant influence on the effect size of gabapentin’sreduction in nausea and vomiting. The factors that seemedto influence the efficacy of gabapentin in PONV were the typeof surgery and the type of anaesthetic agent used for induction

and maintenance. The statistically significant reduction ofpostoperative vomiting observed in the present analysis inthe subgroup of patients undergoing abdominal hysterectomyconcurs with the earlier findings in the same patient popula-tion. However, the imprecise estimates found in the cholecyst-ectomy group do not rule out the anti-emetic efficacy ofpreoperative gabapentin. The same could explain the statistic-ally non-significant effects in the individual dosage groups.Also, the subgroups had significant heterogeneity, possiblydue to the small number of studies in the individual subgroups.Moreover, the dose–response (efficacy) relationship curve ofgabapentin is likely to be flat, as observed in epilepsy,10 satur-able absorption being a plausible explanation. Conversely, theadverse events (commonly somnolence and dizziness) maypossibly have a relationship with dose.11 Although it is difficult,based on the present analysis, to recommend a specific dose,

Pooled Relative Risk for Vomiting

Favors Gabapentin Favors Placebo

PlaceboGabapentin

Studies Events EventsSample SampleGabapentin Weights RR

(n) (N) Dose (mg) (n) (N) (95% Cl)

Mohamaddi et al

Behdad et al

Ajori et al

Pandey et al*

Ghafari et al

Siddiqui et al

Takmaz et al

Durmus et al

Sen et al

Takmaz et al

Turan et al

RE Model

0

5

5

1

4

5

6

3

8

1

6

0.01

4

7

16

2

9

6

7

6

8

7

0.11 (0.01, 1.99)

0.69 (0.25, 1.94)

0.31 (0.12, 0.81)

0.50 (0.05, 5.08)

0.44 (0.15, 1.30)

0.83 (0.28, 2.49)

0.86 (0.38, 1.95)

0.50 (0.14, 1.78)

1.00 (0.47, 2.14)

0.14 (0.02, 1.02)

9

35

30

69

20

33

36

15

25

20

15

25 0.67 (0.28, 1.59)

1.27%

9.91%

11.77%

1.96%

9.15%

8.84%

15.60%

6.55%

18.33%

2.72%

13.90%

0.62 (0.45, 0.85)100.00%

35

31

69

20

33

36

15

25

20

15

25

300

400

600

600

600

600

900

1200

1200

1200

1200

0.10 0.50 4.00

Relative risk (RR)

Fig 3 Pooled analysis of postoperative vomiting.

Gabapentin and postoperative nausea and vomiting BJA

593

a preoperative dose of 300 or 600 mg can be justified forevaluation in future studies, providing evidence regarding thebenefit–risk profile of gabapentin in this clinical setting.

Another interesting finding not previously reported wasthe differential efficacy of gabapentin with respect to the useof propofol, either as a maintenance or induction agent.In the pooled analysis of subgroup of studies using propofol,the reduction in the nausea and vomiting events with gaba-pentin was not statistically significant, but was significantwhen propofol was not used. The anti-emetic properties ofpropofol35 may explain the differential effects of gabapentin,consequently negating the effects of the latter. This suggestsa complex interaction between gabapentin and propofol, al-though synergistic effects between propofol and gabapentincan be expected due to different pharmacological properties,highlighting the need for additional studies. The anti-emeticeffects of propofol are short-lived (up to 6 h),35 while gabapentinpossibly provides long-lasting effects (possibly .24 h), as

it has been shown to be effective in delayed chemotherapy-induced nausea and vomiting12 and has postoperative anal-gesia even after 24 h36. This precludes making an evaluationon any temporal relationship, as the included studies re-ported the events occurring only in the first 24 h postoperativeperiod.

The speculated mechanisms of anti-emetic properties ofgabapentin include decreased tachykinin neurotransmis-sion,12 37 – 39 decreased calcium influx in area postrema,40 41

reduced inflammation at the surgical trauma site,42 as in-flammation could lead to postoperative ileus and subsequentlyto PONV,43 and reduced anxiety,44 apart from the direct effect ofdecreased opioid consumption (seen with preoperative gabapen-tin use) on nausea and vomiting. Even though the postoperativeopioid consumption is significantly lower with preoperative gaba-pentin as compared with the placebo group (observed in the pre-vious analysis), we did not observe any influence on the effectsize of gabapentin on postoperative nausea and vomiting in the

Pooled Relative Risk for PONV

Favors Gabapentin Favors Placebo

PlaceboGabapentin

Studies Events EventsSample SampleGabapentin Weights RR

(n) (N) Dose (mg) (n) (N) (95% Cl)

Pandey et al

Ghai et al

Khademi et al

Pandey et al**

Shrivastava et al

Bashir et al

Frouzanfard et al

RE Model

38

21

16

46

15

20

7

0.01

8

29

28

75

31

31

24

4.75 (2.29, 9.84)

0.72 (0.57, 0.92)

0.56 (0.36, 0.87)

0.61 (0.47, 0.80)

0.48 (0.29, 0.80)

0.65 (0.43, 0.97)

0.29 (0.15, 0.55)

153

30

43

125

60

50

25

12.71%

15.37%

14.46%

15.27%

14.16%

14.69%

13.34%

0.70 (0.39, 1.28)100.00%

153

30

44

125

60

50

25

300

300

600

600

600

600

1200

0.10 0.50 4.00

Relative risk (RR)

Fig 4 Pooled analysis of composite postoperative nausea and vomiting.

BJA Achuthan et al.

594

moderator analysis. This shows that mechanisms other than thedecreased consumption of opioids are possibly responsible forgabapentin’s anti-emetic property.

In the sensitivity analysis, the study by Pandey and collea-gues21 had a significant impact on the results of the PONVend-point. The pooled RR estimates showed that the effect ofgabapentin on the incidence of PONV was not significant [RR0.71 (95% CI 20.39–1.28)]. Exclusion of this study resulted ina statistically significant finding [RR 0.59 (95% CI 20.5–0.7)].This study by Pandey and colleagues21 showed that preope-rative gabapentin (300 mg) use caused an increased incidenceof PONV (38/153 compared with 8/153 in the placebo group).However, in a study by the same authors in 2006,23 it wasshown that the preoperative use of gabapentin (600 mg)indeed had a beneficial effect on the incidence of PONV (46/125 compared with 75/125 in the placebo group). Both thestudies were done in laparoscopic cholecystectomy patients.

The reasons for this disparity are unknown; we believethat this difference could be due to the fact that PONV wasnot the primary endpoint in the first study, whereas it wasin the second. This could have led to differences in studyconduct and analysis. Similarly, concerns about the reliabilityof the data not formally collected as the primary endpointwere raised in another recent review.45

In the present analysis the observed reduction in post-operative nausea and vomiting with preoperative gabapentinsupport our recommendation of including gabapentin inthe armamentarium of pharmacotherapies for PONV in pati-ents undergoing abdominal surgery. The current treatmentguidelines recommend 5-HT3, dopaminergic, and histaminicantagonists as prophylaxis and treatment for PONV.34 Basedon the additional analgesic properties, safety profile, andcost, gabapentin might be usefully included in the current listof PONV prophylactic agents.

Pooled Relative Risk for Rescue Antiemetic Use

Favors Gabapentin Favors Placebo

PlaceboGabapentin

Studies Events EventsSample SampleGabapentin Weights RR

(n) (N) Dose (mg) (n) (N) (95% Cl)

Mohamaddi et al

Ajori et al

Shrivastava et al

Bashir et al

Takmaz et al

Sen et al

Takmaz et al

Turan et al

RE Model

0

7

9

17

12

8

2

11

0.01

4

19

23

27

11

7

11

16

0.11 (0.01, 1.99)

0.37 (0.17, 0.82)

0.39 (0.20, 0.77)

0.63 (0.40, 1.00)

1.09 (0.73, 1.62)

1.14 (0.51, 2.55)

0.18 (0.05, 0.68)

0.69 (0.40, 1.17)

35

69

60

50

15

20

15

25

1.69%

11.97%

13.89%

18.07%

19.40%

11.92%

6.34%

16.72%

0.60 (0.41, 0.89)100.00%

35

69

60

50

15

20

15

25

300

600

600

600

900

1200

1200

1200

0.10 0.50 4.00

Relative risk (RR)

Fig 5 Pooled analysis of postoperative use of rescue medication.

Gabapentin and postoperative nausea and vomiting BJA

595

ConclusionThe present analysis overcomes the limitations of earlier ana-lyses and the additional moderator and sensitivity analysesconducted provide strong evidence favouring preoperativegabapentin as a pharmacotherapy for prevention of PONV inpatients undergoing abdominal surgeries. The use of propofolin the perioperative period could possibly be an importantfactor influencing the decision to use gabapentin as a prevent-ive agent. Nevertheless, the analgesic properties and a possiblelonger duration of action favour preoperative gabapentinuse irrespective of the anaesthetic agent. Future studies com-paring the efficacy of gabapentin with the more commonlyused 5HT3 antagonists are needed to precisely define therole of preoperative gabapentin in PONV.

Supplementary materialSupplementary material is available at British Journal ofAnaesthesia online.

Authors’ contributionsS.A., I.S., and S.B. were involved in the data extraction. S.A., I.S.,A.S., D.H., and A.C. were involved in the data analysis and thedrafting and editing of the manuscript. D.H. and A.C. reviewedand approved the final manuscript.

Declaration of interestNone declared.

References1 Wagner DS, Yap JM, Bradley KM, Voepel-Lewis T. Assessing parents

preferences for the avoidance of undesirable anesthesia sideeffects in their children undergoing surgical procedures. PaediatrAnaesth 2007; 17: 1035–42

2 Macario A, Weinger M, Carney S, Kim A. Which clinical anesthesiaoutcomes are important to avoid? The perspective of patients.Anesth Analg 1999; 89: 652–8

3 Kerger H, Turan A, Kredel M, et al. Patients’ willingness to pay foranti-emetic treatment. Acta Anaesthesiol Scand 2007; 51: 38–43

4 Habib AS, Chen YT, Taguchi A, Hu XH, Gan TJ. Postoperative nauseaand vomiting following inpatient surgeries in a teaching hospital: aretrospective database analysis. Curr Med Res Opin 2006; 22:1093–9

5 Hill RP, Lubarsky DA, Phillips-Bute B, et al. Cost-effectiveness ofprophylactic antiemetic therapy with ondansetron, droperidol, orplacebo. Anesthesiology 2000; 92: 958–67

6 Carroll NV, Miederhoff PA, Cox FM, Hirsch JD. Costs incurred by out-patient surgical centers in managing postoperative nausea andvomiting. J Clin Anesth 1994; 6: 364–9

7 Franck M, Radtke FM, Apfel CC, et al. Documentation of post-operative nausea and vomiting in routine clinical practice. J IntMed Res 2010; 38: 1034–41

8 Apfel CC, Laara E, Koivuranta M, Greim CA, Roewer N. A simplifiedrisk score for predicting postoperative nausea and vomiting: con-clusions from cross-validations between two centers. Anesthesi-ology 1999; 91: 693–700

9 Chang CY, Challa CK, Shah J, Eloy JD. Gabapentin in acute post-operative pain management. Biomed Res Int 2014; 2014: 631756

10 Bockbrader HN, Wesche D, Miller R, Chapel S, Janiczek N, Burger P. Acomparison of the pharmacokinetics and pharmacodynamics ofpregabalin and gabapentin. Clin Pharmacokinet 2010; 49: 661–9

11 VanMeter SA, Kavanagh ST, Warren S, Barrett RW. Dose response ofgabapentin enacarbil versus placebo in subjects with moderate-to-severe primary restless legs syndrome: an integrated analysis ofthree 12-week studies. CNS Drugs 2012; 26: 773–80

12 Guttuso T Jr, Roscoe J, Griggs J. Effect of gabapentin on nauseainduced by chemotherapy in patients with breast cancer. Lancet2003; 361: 1703–5

13 Downs SH, Black N. The feasibility of creating a checklist for the as-sessment of the methodological quality both of randomised andnon-randomised studies of health care interventions. J EpidemiolCommunity Health 1998; 52: 377–84

14 Wolfgang V. Conducting meta-analyses in R with the metaforpackage. J Stat Software 2010; 36: 1–48

15 Alayed N, Alghanaim N, Tan X, Tulandi T. Preemptive use of gaba-pentin in abdominal hysterectomy: a systematic review andmeta-analysis. Obstet Gynecol 2014; 123: 1221–9

16 Ajori L, Nazari L, Mazloomfard MM, Amiri Z. Effects of gabapentin onpostoperative pain, nausea and vomiting after abdominal hyster-ectomy: a double blind randomized clinical trial. Arch GynecolObstet 2012; 285: 677–82

17 Durmus M, Kadir But A, Saricicek V, Ilksen Toprak H, Ozcan Ersoy M.The post-operative analgesic effects of a combination of gabapen-tin and paracetamol in patients undergoing abdominal hysterec-tomy: a randomized clinical trial. Acta Anaesthesiol Scand 2007;51: 299–304

18 Frouzanfard F, Fazel MR, Abolhasani A, Fakharian E, Mousavi G,Moravveji A. Effects of gabapentin on pain and opioid consumptionafter abdominal hysterectomy. Pain Res Manag 2013; 18: 94–6

19 Khademi S, GhaffarpasandF, HeiranHR, Asefi A. Effects of preopera-tive gabapentin on postoperative nausea and vomiting after opencholecystectomy: a prospective randomized double-blind placebo-controlled study. Med Princ Pract 2010; 19: 57–60

20 Mohammadi SS, Seyedi M. Effects of gabapentin on early post-operative pain, nausea and vomiting in laparoscopic surgery forassisted reproductive technologies. Pak J Biol Sci 2008; 11: 1878–80

21 Pandey CK, Priye S, Singh S, Singh U, Singh RB, Singh PK. Preemptiveuse of gabapentin significantly decreases postoperative pain andrescue analgesic requirements in laparoscopic cholecystectomy.Can J Anaesth 2004; 51: 358–63

22 Pandey CK, Singhal V, Kumar M, et al. Gabapentin provides effectivepostoperative analgesia whether administered pre-emptively orpost-incision. Can J Anaesth 2005; 52: 827–31

23 Pandey CK, Priye S, Ambesh SP, Singh S, Singh U, Singh PK. Prophy-lactic gabapentin for prevention of postoperative nausea andvomiting in patients undergoing laparoscopic cholecystectomy: arandomized, double-blind, placebo-controlled study. J PostgradMed 2006; 52: 97–100

24 Sen H, Sizlan A, Yanarates O, et al. A comparison of gabapentin andketamine in acute and chronic pain after hysterectomy. AnesthAnalg 2009; 109: 1645–50

25 Srivastava U, Kumar A, Saxena S, Mishra AR, Saraswat N, Mishra S.Effect of preoperative gabapentin on postoperative pain and tra-madol consumption after minilap open cholecystectomy: a rando-mized double-blind, placebo-controlled trial. Eur J Anaesthesiol2010; 27: 331–5

26 Takmaz SA, Kaymak C, Pehlivan BS, Dikmen B. [Effect of preo-perative 900 and 1200 mg single oral dose of gabapentin on post-operative pain relief and tramadol consumption in opencholecystectomy surgery]. Agri 2007; 19: 32–8

BJA Achuthan et al.

596

27 Turan A, Karamanlioglu B, Memis D, Usar P, Pamukcu Z, Ture M. Theanalgesic effects of gabapentin after total abdominal hysterec-tomy. Anesth Analg 2004; 98: 1370–3

28 Ghafari MH, Akrami M, Nouralishahi B, Sadegh A. Preoperative gaba-pentin or clonidine decreases postoperative pain and morphineconsumption after abdominal hysterectomy. Res J Biol Sci 2009;4: 458–63

29 Ghai A, Gupta M, Hooda S, Singla D, Wadhera R. A randomized con-trolledtrial tocomparepregabalinwithgabapentinforpostoperativepain in abdominal hysterectomy. Saudi J Anaesth 2011; 5: 252–7

30 Behdad S, Ayatollahi V, Bafghi AT, Tezerjani MD, Abrishamkar M.Effect of gabapentin on postoperative pain and operation compli-cations: a randomized placebo controlled trial. West Indian Med J2012; 61: 128–33

31 Siddiqui NT, Fischer H, Guerina L, Friedman Z. Effect of a pre-operative gabapentin on postoperative analgesia in patients withinflammatory bowel disease following major bowel surgery: arandomized, placebo-controlled trial. Pain Pract 2014; 14: 132–9

32 Bashir F, Mohd K, Qazi S, Hashia AM. A randomized, double blind,placebo controlled study evaluating preventive role of gabapentinfor PONV in patients undergoing laparoscopic cholecystectomy.JK Sci J Med Educ Res 2009; 11: 190–3

33 Apfel CC, Heidrich FM, Jukar-Rao S, et al. Evidence-based analysis ofrisk factors for postoperative nausea and vomiting. Br J Anaesth2012; 109: 742–53

34 Gan TJ, Diemunsch P, Habib AS, et al. Consensus guidelines for themanagement of postoperative nausea and vomiting. AnesthAnalg 2014; 118: 85–113

35 Tramer M, Moore A, McQuay H. Propofol anaesthesia and post-operative nausea and vomiting: quantitative systematic review ofrandomized controlled studies. Br J Anaesth 1997; 78: 247–55

36 Sen H, Sizlan A, Yanarates O, et al. The effects of gabapentin onacute and chronic pain after inguinal herniorrhaphy. Eur J Anaes-thesiol 2009; 26: 772–6

37 Guttuso T Jr, Kurlan R, McDermott MP, Kieburtz K. Gabapentin’seffects on hot flashes in postmenopausal women: a randomizedcontrolled trial. Obstet Gynecol 2003; 101: 337–45

38 Maneuf YP, Hughes J, McKnight AT. Gabapentin inhibits the sub-stance P-facilitated K+-evoked release of [3H]glutamate from ratcaudial trigeminal nucleus slices. Pain 2001; 93: 191–6

39 Fehrenbacher JC, Taylor CP, Vasko MR. Pregabalin and gabapentinreduce release of substance P and CGRP from rat spinal tissuesonly after inflammation or activation of protein kinase C. Pain2003; 105: 133–41

40 Felix R, Gurnett CA, De Waard M, Campbell KP. Dissection of func-tional domains of the voltage-dependent Ca2+ channel a2d

subunit. J Neurosci 1997; 17: 6884–91

41 Stahl SM. Mechanism of action of alpha2delta ligands: voltage sen-sitive calcium channel (VSCC) modulators. J Clin Psychiatry 2004;65: 1033–4

42 Dias JM, de Brito TV, de Aguiar Magalhaes D, et al. Gabapentin, asynthetic analogue of gamma aminobutyric acid, reverses system-ic acute inflammation and oxidative stress in mice. Inflammation2014; 37: 1826–36

43 Rychter J, Clave P. Intestinal inflammation in postoperative ileus:pathogenesis and therapeutic targets. Gut 2013; 62: 1534–5

44 Clarke H, Kirkham KR, Orser BA, et al. Gabapentin reduces preopera-tive anxiety and pain catastrophizing in highly anxious patientsprior to major surgery: a blinded randomized placebo-controlledtrial. Can J Anaesth 2013; 60: 432–43

45 Guttuso T Jr. Gabapentin’s anti-nausea and anti-emetic effects: areview. Exp Brain Res 2014; 232: 2535–9

Handling editor: J. G. Hardman

Gabapentin and postoperative nausea and vomiting BJA

597

Annotations

Gabapentin prophylaxis for postoperative nausea andvomiting in abdominal surgeries: a quantitative analysis of

evidence from randomized controlled clinical trialsAchuthan, S.; Singh, I.; Varthya, S. B.; Srinivasan, a.; Chakrabarti, a.; Hota, D.

01 Mendeley User Page no. 1

7/4/2015 9:45

Les NVPO sont un événement fréquent en post-anesthésie puisqu'ils touchent environ un tiers despatients. Les différents scores et prophylaxies utilisées bien que souvent efficaces ne closent pas lechapitre de leur prévention. La gabapentine, antivonvusilvant, a montré par ailleurs son effetanalgésique en post-opératoire.Plus récemment, la gabapentine a montré un effet anti-émétiquelorsqu'elle était administrée en prévention dans la chimiothérapie du cancer du sein.Cette étude estune méta-analyse des essais randomisés de la gabapentine en prévention des NVPO. Elle conclut àson efficacité, efficacité d'autant plus marquée que le propofol n'est pas utilisé comme agentd'induction et/ou d'entretien.