Professional Documents
Culture Documents
2006 Pain Gabapentinmetaanalysis
2006 Pain Gabapentinmetaanalysis
net/publication/257551222
CITATIONS READS
211 183
3 authors, including:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Kok-Yuen Ho on 05 November 2019.
Received 7 February 2006; received in revised form 13 May 2006; accepted 12 June 2006
Abstract
The objective of this systematic review was to evaluate the efficacy and tolerability of perioperative gabapentin administration for
the control of acute postoperative pain. We searched Medline (1966–2006), the Cochrane Library (2006), Scopus, CINAHL and
bibliographies from clinical trials and review articles. We included randomized controlled trials (RCTs) comparing gabapentin with
inactive controls in surgical patients. Sixteen valid RCTs were included. Weighted mean difference (WMD) for postoperative pain
intensity (0–100 mm visual analogue scale) was 16.55 mm at 6 h and 10.87 mm at 24 h for treatment with a single preoperative
dose of gabapentin 1200 mg. Cumulative opioid consumption at 24 h was also significantly decreased with gabapentin (WMD,
27.90 mg). When gabapentin was administered at doses less than 1200 mg, pain intensity was also lower at 6 h (WMD,
22.43 mm) and 24 h (WMD, 13.18 mm). Cumulative 24 h opioid consumption was also lower (WMD, 7.25 mg). Gabapentin
was associated with an increased risk of sedation (Peto OR 3.86; 95% CI 2.50–5.94) but less opioid-related side effects such as vom-
iting (Peto OR 0.58; 95% CI 0.39–0.86) and pruritus (Peto OR 0.27; 95% CI 0.10–0.74). In conclusion, gabapentin has an analgesic
and opioid-sparing effect in acute postoperative pain management when used in conjunction with opioids.
2006 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
2. Methods
*
Corresponding author. Tel.: +1 919 681 4660; fax: +1 919 681
7901. We followed the QUOROM guidelines for reporting meta-
E-mail address: hokokyuen@yahoo.com.sg (K.-Y. Ho). analyses (Moher, 1999).
0304-3959/$32.00 2006 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.pain.2006.06.018
92 K.-Y. Ho et al. / Pain 126 (2006) 91–101
2.1. Search The following adverse effects were noted for analysis: nau-
sea, vomiting, sedation and dizziness. Side effects reported as
Published reports of randomized controlled trials (RCTs) drowsiness and somnolence were grouped under sedation.
that investigated perioperative gabapentin administration for Reports of lightheadedness and vertigo were analyzed together
postoperative pain management were searched in Medline with dizziness.
(1966–2006), The Cochrane Central Register of Controlled Pain intensity measured on the visual analogue scale (VAS,
Trials (2006), Scopus and CINAHL. Free text and MeSH 0–100 mm, 0 = no pain and 100 = maximum pain), was ana-
terms ‘gabapentin’, ‘pain’, ‘analgesia’, ‘analgesic’, ‘postopera- lyzed quantitatively. Pain verbal rating scale (VRS) reported
tive’ and ‘surgery’ were used for searching. Search was per- on a 0–10 scale was converted to 0–100. The reporting of pain
formed without language restriction but limited to RCTs in scores in the randomized trials was highly disparate. Pain
humans. The last electronic search was in January 2006. Addi- scores were documented at different time intervals and some
tional studies from the bibliographies of reviews or reports of these studies ended before 24 h from the end of surgery.
were also identified. Authors of original reports were contacted To facilitate pooling of data, only pain scores in the immediate
for additional information if needed. (within 6 h from end of surgery) and late (24 h from end of sur-
gery) postoperative period were analyzed. Pain scores that
2.2. Selection criteria were reported at 2 or 4 h postsurgery were grouped under
the immediate postoperative period and analyzed together if
Only randomized, placebo-controlled trials that reported data at the 6 h time point were not reported. Clinical trials that
on relevant pain outcomes such as pain scores, time to first were less than 24 h in duration would not have 24 h pain scores
analgesic request and postoperative cumulative opioid con- for data analysis. All postoperative opioid consumption was
sumption were included. Reviews and abstracts were not con- converted to morphine equivalents (Macintyre and Ready,
sidered. Studies incorporating a local anesthetic technique or 2001).
nerve block as part of the anesthetic regimen were excluded.
2.5. Meta-analyses
2.3. Validity assessment
Both dichotomous and continuous data were extracted. Con-
One author (K.Y.H.) screened the abstracts of all retrieved tinuous data were analyzed as weighted mean differences
reports andexcluded thosethat did notmeet the inclusion criteria. (WMD) with 95% confidence intervals (CI). When mean values
All included reports were then independently read by two review- and standard deviations were not reported, the authors of the
ers (K.Y.H. and A.H.) who assessed the validity of the studies studies were contacted. If they did not reply and the data were
using the modified Oxford Scale (Box 1) (Pasquina et al., 2003; presented graphically, data were extracted from the graphs. If
Elia and Tramer, 2005). The minimum score of an included trial this was not possible, the data were not considered. Dichoto-
was 1 and the maximum was 7. Discrepancies were resolved by mous data on adverse effects were summarized using relative risk
discussion or by consulting the third reviewer (T.J.G.). (RR) with 95% CI. For rare outcomes, Peto odds ratios (OR)
Box 1: Modified Oxford Scale that deal better with zero cells were computed. A random effects
model was used by default. If the statistical test for heterogeneity
Validity score (0–7) was negative, a fixed effects model was utilized. If the 95% CI
Randomization included 1, it was assumed that there was no statistically signif-
0 None icant difference between gabapentin and control. Analyses were
1 Mentioned performed using Review Manager Software (version 4.2.8,
2 Described and adequate Cochrane Collaboration). Data were graphically plotted using
forest plots to evaluate treatment effects. For statistically signif-
Concealment of allocation
icant differences in the incidence of adverse effects between treat-
0 None
ment and control groups, Number-Needed-to-Treat (NNT) or
1 Yes
Number-Needed-to-Harm (NNH) was calculated to estimate
Double blinding the clinical impact of the beneficial or harmful effect of the
0 None intervention.
1 Mentioned
2 Described and adequate
Flow of patients 3. Results
0 None
1 Described but incomplete Twenty publications that described the use of gaba-
2 Described and adequate pentin in postoperative pain management were identi-
fied between 2002 and 2006 (Fig. 1). All reports were
published in English. Four were excluded; two were
2.4. Data abstraction
not clinical trials (Dahl et al., 2004; Wiffen et al.,
A data abstraction form was created and the following data 2005), one was only published in abstract form (Gregg
were collected: (i) type of surgery; (ii) number of patients; (iii) et al., 2001), and one was excluded because eutectic mix-
gabapentin dosage and regimen; (iv) study design and dura- ture of local anesthetics cream and a nerve block with
tion; (v) analgesia outcome measures; and (vi) adverse effects. ropivacaine were used in addition to general anesthesia
K.-Y. Ho et al. / Pain 126 (2006) 91–101 93
3.1.2. Twenty-four hour cumulative opioid consumption (Turan et al., 2004c). Meta-analysis of the remaining
Three studies reported on opioid consumption at 24 h two studies showed that gabapentin produced a sta-
(Turan et al., 2004a; Menigaux et al., 2005; Pandey tistically significant delay in time to first request for
et al., 2005a). Combined data showed that the WMD analgesia (WMD 7.42 min; 95% CI 0.49–14.34)
of 27.9 mg (95% CI 31.52 to 24.29) was in favor (Fig. 5).
of gabapentin (Fig. 4).
3.1.4. Adverse effects
3.1.3. Time to first request for rescue analgesic Data from six studies showed that single preoperative
Three studies reported data on the time to first dose of 1200 mg gabapentin were associated with a low-
rescue analgesic. Out of these three studies, only er risk of vomiting (Peto OR 0.42; 95% CI 0.24–0.76)
two studies had suitable data available. One study (Rorarius et al., 2004; Turan et al., 2004a,b,c; Tuncer
measured time to first request for analgesic from et al., 2005; Pandey et al., 2005a). NNT was 8. Patients
the instance the study drug (gabapentin or placebo) in the gabapentin group were at less risk of urinary
was administered and not from the end of surgery retention too (Turan et al., 2004a,b). NNT was 7. There
Fig. 3. Meta-analysis: VAS of pain intensity (0–100 mm) at 6 and 24 h in patients receiving 1200 mg gabapentin preoperatively. VAS, visual
analogue scale; WMD, weighted mean difference; CI, confidence interval.
96 K.-Y. Ho et al. / Pain 126 (2006) 91–101
Fig. 4. Meta-analysis: 24 h morphine consumption (mg) in patients receiving 1200 mg gabapentin preoperatively. WMD, weighted mean difference;
CI, confidence interval.
3.2.3. Time to first request for rescue analgesic 3.3.1. Pain intensity
None of the five studies measured time to first analge- Only two out of the five trials had data suitable for
sia as an outcome. meta-analysis (Fassoulaki et al., 2002; Gilron et al.,
Fig. 5. Meta-analysis: time to first request for rescue analgesic (min) in patients receiving 1200 mg gabapentin preoperatively. WMD, weighted mean
difference; CI, confidence interval.
K.-Y. Ho et al. / Pain 126 (2006) 91–101 97
Fig. 6. Meta-analysis: VAS of pain intensity (0–100 mm) in patients receiving single dose of gabapentin <1200 mg preoperatively. WMD, weighted
mean difference; CI, confidence interval.
2005). Combined data did not show any difference studies (Fassoulaki et al., 2002; Dierking et al., 2004;
between gabapentin and control groups at both 6 and Gilron et al., 2005; Radhakrishnan et al., 2005; Turan
24 h after surgery. et al., 2006). The gabapentin group was also associated
with a lower incidence of pruritus (Peto OR 0.21; 95%
3.3.2. Twenty-four hour cumulative opioid consumption CI 0.05–0.87) (Gilron et al., 2005; Radhakrishnan
Only one study measured 24 h morphine consump- et al., 2005; Turan et al., 2006). NNT was 13. Peto
tion as an outcome (Turan et al., 2006). It showed a OR was 0.12 (95% CI 0.02–0.9) for a lower incidence
24% reduction in total patient-controlled analgesia mor- of constipation in the gabapentin group based on one
phine usage in the gabapentin group compared with the study (Turan et al., 2006). NNT was 6. There was no dif-
control group. ference between gabapentin and control groups for the
other adverse effects.
3.3.3. Time to first request for rescue analgesic
Only one study presented these data as an outcome 3.4. Adverse effects
and reported no difference between the gabapentin and
control groups (Fassoulaki et al., 2002). In the subgroup analyses of adverse effects, the
number of clinical trials analyzed for each adverse
3.3.4. Adverse effects effect was small because adverse effects were variably
A statistically significant lower incidence of nausea reported. Therefore we pooled data on adverse effects
was observed in the patients receiving multiple doses from all studies to perform a meta-analysis (Fig. 8).
of gabapentin perioperatively (Peto OR 0.54; 95% CI All 16 studies reported nausea as an adverse outcome.
0.31–0.95). NNT was 9. Data were obtained from five Three studies did not report the incidence of vomiting
Fig. 7. Meta-analysis: 24 h cumulative morphine consumption (mg) in patients receiving single dose of gabapentin <1200 mg preoperatively. WMD,
weighted mean difference; CI, confidence interval.
98 K.-Y. Ho et al. / Pain 126 (2006) 91–101
Fig. 8. Meta-analysis: Adverse effects (nausea, vomiting, sedation, dizziness, pruritus, urinary retention, constipation and respiratory depression).
OR, odds ratio; WMD, weighted mean difference; CI, confidence interval.
K.-Y. Ho et al. / Pain 126 (2006) 91–101 99
(Dirks et al., 2002; Gilron et al., 2005; Menigaux pronounced in the early postoperative period. However,
et al., 2005). One study reported nausea and vomiting this reduction was still significant at 24 h and was
collectively as a single adverse event and these data associated with a significant reduction in opioid con-
were not included in the meta-analysis (Pandey sumption. Gabapentin given as multiple doses perioper-
et al., 2004a). Quantitative analysis showed that atively did not demonstrate lower VAS pain scores
patients who received gabapentin experienced less compared to the control group. However, this result
vomiting (Peto OR 0.58; 95% CI 0.39–0.86) compared should be interpreted with caution, as there were only
to controls with a NNT of 11. On the other hand, the two studies with appropriate data for meta-analysis in
incidence of sedation was higher in the gabapentin this subgroup.
group (Peto OR 3.86, 95% CI 2.50–5.94) with a While this analysis demonstrated a statistically signif-
NNH of 8. There was no statistically significant differ- icant prolongation to the time of first request for rescue
ence between the groups in the incidence of nausea analgesics, the prolongation was only by about 7 min
(Peto OR 0.72; 95% CI 0.51–1.01) or dizziness (Peto and therefore is of no clinical importance. However, this
OR 1.34; 95% CI 0.86–2.10). outcome was only reported in four of the 16 randomized
Six trials reported on the incidence of pruritus controlled trials.
(Turan et al., 2004a,b, 2006; Gilron et al., 2005; Pain after surgery is predominantly nociceptive in
Radhakrishnan et al., 2005; Pandey et al., 2005b). nature. However, prolonged central sensitization mani-
Combined data showed that gabapentin was associat- festing as hyperalgesia does occur after surgical trauma
ed with less pruritus compared with control (Peto OR (Field et al., 1997). Gabapentin has been shown to pre-
0.27; 95% CI 0.10–0.74) with a NNT of 15. Four tri- vent central sensitization by reducing hyperexcitability
als reported urinary retention as an adverse effect of secondary nociceptive neurons in the dorsal horn.
(Turan et al., 2004a,b, 2006; Radhakrishnan et al., This may be related to suppression of noxious stimu-
2005). Meta-analysis showed no statistically significant lus-induced excitatory amino acid release in the spinal
difference in the incidence of urinary retention cord (Feng et al., 2003). Little is known about how
between gabapentin and the control group (Peto OR gabapentin modulates postoperative pain in the pres-
0.51; 95% CI 0.23–1.14). Only three trials reported ence of opioids. Gabapentin and morphine may be syn-
on the incidence of constipation and combined data ergistic due to their separate action on the peripheral
again demonstrated no statistically significant differ- and central nervous system (Matthews and Dickenson,
ence between gabapentin and the control group (Peto 2002). Gabapentin may also decrease postoperative
OR 0.34; 95% CI 0.11–1.11) (Turan et al., 2004a,b, morphine requirement through preventing the develop-
2006). Combined data from two studies (Pandey ment of opioid tolerance (Gilron et al., 2003).
et al., 2004a, 2005a) did not show any difference in The incidence of postoperative vomiting and pruritus
the incidence of respiratory depression between gaba- was significantly lower in the gabapentin group. This
pentin and the control group (Peto OR 1.07; 95% may be explained by a reduction in opioid consumption
CI 0.21–5.49). and the associated decrease in opioid-related adverse
effects. There was also a trend towards a lower incidence
4. Discussion of nausea, urinary retention and constipation in the
gabapentin group, but the difference did not achieve sta-
4.1. Efficacy and adverse effects tistical significance. One study however reported a sig-
nificantly greater incidence of postoperative nausea
This systematic review demonstrated that preopera- and vomiting in patients undergoing laparoscopic sur-
tive gabapentin administration was useful for postoper- gery who had received gabapentin (Pandey et al.,
ative pain management. A single preoperative dose of 2004a). This study was not included in our systematic
gabapentin, 1200 mg or less, effectively reduced pain review since the authors did not report the incidence
intensity and opioid consumption for the first 24 h after of nausea and vomiting separately but reported the inci-
surgery. In the subgroup that received a single 1200 mg dence of both outcomes collectively. The incidence of
of gabapentin preoperatively, the time to first request sedation was significantly higher in the gabapentin
for rescue analgesia was also prolonged. However, mul- group. There was also a trend for more dizziness in
tiple dosing with gabapentin preoperatively and contin- patients receiving gabapentin. This was in line with the
ued postoperatively did not appear to reduce VAS pain side effect profile of gabapentin and might limit the use-
scores. fulness of gabapentin in ambulatory settings. Whether
This systematic review therefore demonstrates a this effect is clinically significant and whether it leads
potential role for preemptive gabapentin as an adjunct to prolongation of recovery time and postanesthesia
to standard postoperative pain management. In the care unit stay needs further evaluation. No serious
groups receiving a single dose of gabapentin preopera- adverse effects were reported with the perioperative
tively, the reduction in pain scores appeared to be more administration of gabapentin.
100 K.-Y. Ho et al. / Pain 126 (2006) 91–101
Elia N, Tramer MR. Ketamine and postoperative pain – a quantitative Pandey CK, Priye S, Singh S, Singh U, Singh RB, Singh PK.
systematic review of randomised trials. Pain 2005;113:61–70. Preemptive use of gabapentin significantly decreases postoperative
Fassoulaki A, Patris K, Sarantopoulos C, Hogan Q. The analgesic pain and rescue analgesic requirements in laparoscopic cholecys-
effect of gabapentin and mexiletine after breast surgery for cancer. tectomy. Can J Anaesth 2004a;51:358–63.
Anesth Analg 2002;95:985–91. Pandey CK, Sahay S, Gupta D, Ambesh SP, Singh RB, Raza M, et al.
Fassoulaki A, Triga A, Melemeni A, Sarantopoulos C. Multimodal Preemptive gabapentin decreases postoperative pain after lumbar
analgesia with gabapentin and local anesthetics prevents acute and discoidectomy. Can J Anaesth 2004b;51:986–9.
chronic pain after breast surgery for cancer. Anesth Analg Pandey CK, Singhal V, Kumar M, Lakra A, Ranjan R, Pal R, et al.
2005;101:1427–32. Gabapentin provides effective postoperative analgesia whether admin-
Feng Y, Cui M, Willis WD. Gabapentin markedly reduces acetic istered pre-emptively or post-incision. Can J Anaesth 2005b;52:827–31.
acid-induced visceral nociception. Anesthesiology 2003;98: Pasquina P, Tramer MR, Walder B. Prophylactic respiratory physiother-
729–33. apy after cardiac surgery: systematic review. BMJ 2003;327:1379.
Field MJ, Holloman EF, McCleary S, Hughes J, Singh L. Evaluation Radhakrishnan M, Bithal PK, Chaturvedi A. Effect of preemptive
of gabapentin and S-(+)-3-isobutylgaba in rat model of postoper- gabapentin on postoperative pain relief and morphine consump-
ative pain. J Pharmacol Exp Ther 1997;282:1242–6. tion following lumbar laminectomy and discectomy: a randomized,
Gilron I, Biederman J, Jhamandas K, Hong M. Gabapentin blocks double-blinded, placebo-controlled study. J Neurosurg Anesthesiol
and reverses antinociceptive morphine tolerance in the rat paw- 2005;17:125–8.
pressure and tail-flick tests. Anesthesiology 2003;98:1288–92. Rorarius MG, Mennander S, Suominen P, Rintala S, Puura A,
Gilron I, Orr E, Tu D, O’Neill JP, Zamora JE, Bell AC. A placebo- Pirhonen R, et al. Gabapentin for the prevention of postoperative
controlled randomized clinical trial of perioperative administration pain after vaginal hysterectomy. Pain 2004;110:175–81.
of gabapentin, rofecoxib and their combination for spontaneous Rose MA, Kam PC. Gabapentin: pharmacology and its use in pain
and movement-evoked pain after abdominal hysterectomy. Pain management. Anaesthesia 2002;57:451–62.
2005;113:191–200. Rowbotham M, Harden N, Stacey B, Bernstein P, Magnus-Miller L.
Gregg AK, Francis S, Sharpe P, Rowbotham DJ. Analgesic effect of Gabapentin for the treatment of postherpetic neuralgia: a ran-
gabapentin premedication in laparoscopic cholecystectomy: a domized controlled trial. JAMA 1998;280:1837–42.
randomized double-blind placebo-controlled trial. Br J Anaesth Tuncer S, Bariskaner H, Reisli R, Sarkilar G, Cicekci F, Otelcioglu S.
2001;87:174 P. Effect of gabapentin on postoperative pain: a randomized, placebo-
Macintyre PE, Ready LB. Pharmacology of opioids. In: Acute pain controlled clinical study. The Pain Clinic 2005;17:95–9.
management – a practical guide. 2nd ed. W.B. Saunders; 2001. p. Turan A, Karamanlioglu B, Memis D, Hamamcioglu MK, Tukenmez
15–49. B, Pamukcu Z, et al. Analgesic effects of gabapentin after spinal
Matthews EA, Dickenson AH. A combination of gabapentin and surgery. Anesthesiology 2004a;100:935–8.
morphine mediates enhanced inhibitory effects on dorsal horn Turan A, Karamanlioglu B, Memis D, Usar P, Pamukcu Z, Ture M.
neuronal responses in a rat model of neuropathy. Anesthesiology The analgesic effects of gabapentin after total abdominal hyster-
2002;96:633–40. ectomy. Anesth Analg 2004b;98:1370–3.
Menigaux C, Adam F, Guignard B, Sessler DI, Chauvin M. Preoper- Turan A, Memis D, Karamanlioglu B, Yagiz R, Pamukcu Z, Yavuz E.
ative gabapentin decreases anxiety and improves early functional The analgesic effects of gabapentin in monitored anesthesia care for
recovery from knee surgery. Anesth Analg 2005;100:1394–9. ear–nose–throat surgery. Anesth Analg 2004c;99:375–8.
Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Turan A, White PF, Karamanlioglu B, Memis D, Tasdogan M,
Improving the quality of reports of meta-analyses of randomised Pamukcu Z, et al. Gabapentin: an alternative to the cyclooxygen-
controlled trials: the QUOROM statement. Quality of Reporting of ase-2 inhibitors for perioperative pain management. Anesth Analg
Meta-analyses. Lancet 1999;354:1896–900. 2006;102:175–81.
Pandey CK, Navkar DV, Giri PJ, Raza M, Behari S, Singh RB, et al. van de Vusse AC, Stomp-van den Berg SG, Kessels AH, Weber WE.
Evaluation of the optimal preemptive dose of gabapentin for Randomised controlled trial of gabapentin in complex regional
postoperative pain relief after lumbar diskectomy: a randomized, pain syndrome type 1 [ISRCTN84121379]. BMC Neurol 2004;4:13.
double-blind, placebo-controlled study. J Neurosurg Anesthesiol Wiffen PJ, McQuay HJ, Edwards JE, Moore RA. Gabapentin for acute
2005a;17:65–8. and chronic pain. Cochrane Database Syst Rev 2005:CD005452.