ORIGINAL ARTICLE

Efficacy and Safety of Gabapentin vs.

Carbamazepine in the Treatment of Trigeminal
Neuralgia: A Meta-Analysis

Min Yuan, MD*; Huang-yan Zhou, MD†; Zhi-long Xiao, MD*; Wei Wang, PhD*;
Xue-li Li, MD*; Shen-jian Chen, MD*; Xiao-ping Yin, PhD*; Li-jun Xu, BM*
*Department of Neurology, The Second Affiliated Hospital of Nanchang University, Nanchang,
Jiangxi; †Department of Immunology and Pathogenic Biology, The Basic Medical College of
Nanchang University, Nanchang, Jiangxi, China

& Abstract: To evaluate the safety and efficacy of gaba-
pentin in comparison with carbamazepine in the treatment of
trigeminal neuralgia, a meta-analysis of randomized con-
trolled trials was performed. Two reviewers independently
selected studies, assessed study quality, and extracted data.
Sixteen randomized controlled trials that included 1,331
patients were assessed. The meta-analysis showed that the
total effective rate of gabapentin therapy group was similar
with carbamazepine therapy group (OR = 1.600, 95% CI
1.185, 2.161, P = 0.002). While the effective rate of gabapen-
tin therapy for 4 weeks was higher than that of carba-
mazepine therapy (OR = 1.495, 95% CI 1.061, 2.107,
P = 0.022, heterogeneity: x2 = 7.12, P = 0.625, I2 = 0.0%),
the life satisfaction improvement is also better in the
gabapentin therapy group after a 4-week treatment
(SMD = 0.966, 95% CI 0.583, 1.348, P < 0.001). Furthermore,
our meta-analysis suggested that the adverse reaction rate of
gabapentin therapy group was significantly lower than that
of carbamazepine therapy group (OR = 0.312, 95% CI 0.240,
0.407, P < 0.001). In conclusion, present trials comparing
gabapentin with carbamazepine are all poor in terms of
Address correspondence and reprint requests to: Li-jun Xu, BM,
Department of Neurology, The Second Affiliated Hospital of Nanchang
University, No. 1, Minde Road, Nanchang 330006, Jiangxi, China. E-mail:
Xulijun20050901@sina.com (L.-J.Xu).
Submitted: April 12, 2015; Revision accepted: September 5, 2015
DOI. 10.1111/papr.12406
© 2016 World Institute of Pain, 1530-7085/16/$15.00
Pain Practice, Volume , Issue , 2016 –
methodological quality. Based on the available evidence, it is
not possible to draw conclusions regarding the efficacy and
side effects of gabapentin being superior to carba-
mazepine. &
Key Words: gabapentin, carbamazepine, trigeminal neu-
ralgia, meta-analysis, efficacy, safety
INTRODUCTION
Trigeminal neuralgia (TN) is a common type of neuro-
pathic pain, which is characteristically sudden, usually
unilateral, severe, brief, stabbing, and the recurrent
electric shock-like episodes of pain lasts from a few
seconds to < 2 minutes in the area of one or more
branches of the trigeminal nerve.1,2 The disease usually
has trigger points and is often induced in the process of
daily routines such as washing face, brushing teeth,
talking, and even shaving. The annual incidence is 5.9/
100,000 women and 3.4/100,000 men. The incidence
tends to be slightly higher among women at all ages, and
even increases with age.3,4
Currently, drug therapy is most commonly used in the
treatment of TN, and carbamazepine (CBZ) is recom-
mended as the first-line medical treatment according to
the American Academy of Neurology and the European
Federation of Neurological Societies,5 and it has
achieved a reduction in attacks in up to 88% of
2  YUAN ET AL.
patients.6,7 However, the therapeutic index of CBZ is
relatively narrow. Its efficacy is compromised by poor
tolerability of serious adverse effects such as vertigo,
nausea, and white blood cell reduction.8–10 Besides,
some studies show that CBZ has no significant efficacy
in partial patients.11
However, in recent years, gabapentin (GBP), a newer
anti-epileptic drug, is widely used in clinical treatment of
TN. Studies have demonstrated that GBP has broad
application prospects in chronic pain syndromes, espe-
cially in the neuropathic pain.12 Furthermore, GBP has
been the first-choice drug therapy for all types of
neuropathic chronic pain in several international pain
control centers. Its effects contain relatively low rate of
adverse reactions, lack of interaction with other drugs
acting on the nervous system, and evident perception of
its efficacy.9,13 And whenever CBZ fails to control TN,
GBP can be used as an alternative for reducing its
intensity.9,14,15 But in comparison with CBZ, its efficacy
and safety remain controversial. Accordingly, we con-
ducted a meta-analysis to critically evaluate all of the
currently available, randomized controlled trials that
compared GBP with CBZ in the treatment of TN.
METHODS
Search Strategy
The following electronic databases were searched
(to March 2015: Cochrane Central Register of Con-
trolled Trials (CENTRAL) [Issue 3 of 12, March 2015],
PubMed, MEDLINE, Chinese Biomedical Database
(CBM), and the Wanfang database. The first search
term was “trigeminal neuralgia” or “trigeminal neu-
ropathy;” the second search term was “gabapentin” or
“Neurontin;” and the third search term was “carba-
mazepine” or “tegretol.” These three terms were com-
bined for the electronic search, and corresponding
Chinese terms were also searched. In addition, all
relevant journals were manually searched. There were
no language restrictions.
Study Selection
Two reviewers independently identified potentially rel-
evant studies and evaluated each trial according to
predetermined eligibility criteria. Studies were selected if
meeting the following criteria: (1) articles reporting
randomized controlled trials; (2) articles including the
trials comparing GBP with CBZ in the treatment of TN;
(3) the diagnosis method of TN conforming with a
standard criterion such as the International Classifica-
tion of Headache Disorders (ICHD)16 and exclusion of
secondary TN by CT or MRI. (4) The included studies
having similar baseline characteristics in our analysis.
(5) The studies referring to at least one of the outcomes,
such as the visual analog acale (VAS) score, effective
rate, life satisfaction index B (LSI-B), or adverse reac-
tions. (6) Psychiatric diseases and/or other kinds of facial
pain being excluded, as well as studies using other
medications.
Data Extraction and Quality Assessment
All data were independently abstracted in duplicate
using a standard data collection form. When necessary,
the original authors were contacted for supplementary
information. The following data were extracted from
each study: first author’s last name, year of publication,
number of patients, age, number of male patients,
dosage regimen, treatment duration, follow-up period,
and outcome assessment. Risk of bias was assessed
according to the latest Cochrane classification in the
Cochrane Handbook for Systematic Reviews of Inter-
ventions (version 5.1.4). The criteria include the follow-
ing: random sequence generation, allocation
concealment, blinding of participants and personnel,
blinding of the outcome assessment, incomplete out-
come data, the selective reporting, and other bias. Each
criterion was indicated as low risk of bias, high risk of
bias, and unclear risk of bias. We assessed the evidence
quality of the outcomes as high, moderate, low, and very
low using the Grading of Recommendations Assess-
ment, Development and Evaluation (GRADE) system.
Disagreements were resolved by discussion between the
authors.
Data Synthesis
The meta-analysis was performed using STATA meta-
analysis software, version 12.0 (Stata Corporation, Col-
lege Station, TX, U.S.A.). For each study, the odds ratio
(OR) or standardized mean difference (SMD) was used for
the measurement data. The OR and SMD were presented
with 95% confidence intervals. The degree of heterogene-
ity among results was estimated by the chi-square test (The
P-value < 0.1 was considered significant). Whenever
significant heterogeneity (P-value < 0.10 or I2 score
> 50%) was achieved, the random effect model was used.
If no significant heterogeneity across studies was found, a
 Gabapentin vs. Carbamazepine in the Treatment of TN  3
fixed effect model was selected to pool the data. The
presence of publication bias was determined with Egger’s
test and Begg’s test. Sensitivity analyses and subgroups
were performed where appropriate.
RESULTS
Summaries of Included Trials
The literature search revealed 869 articles, among which
844 studies were excluded as duplicates, unrelated or
not clinical trials. The 25 relevant articles were selected
and reviewed by two independent authors. We further
excluded another 9 unrelated articles. Ultimately, 16
trials met our inclusion criteria (Figure 1), and the
general information was listed in Table 1. All trials were
originated from China, and all articles were published in
Chinese. Only six trials included patients diagnosed of
TN by the criterion of the International Classification of
Headache Disorders (ICHD), while other ten studies
used other local and foreign standards. Sixteen eligible
randomized controlled trials were identified, but defini-
tive randomization was found only in six studies,17–22
among which four studies used the method of random
number table for random distribution,17,18,20,21 and the
remaining two were randomly assigned according to
the date of treatment and the visiting sequence.19,22 The
other ten studies were found to be lacking definitive
randomization, with only one article describing blinding
method.17 The allocation concealment was not reported
in any of the trials, and the number of dropouts or
withdrawals was clearly reported in three trials.18,22,23
Only ten studies mentioned the follow-up time.18,22,23
Figure 1. Flow chart of studies selection. TN, trigeminal neural-
gia.
The baseline characteristics were similar between
groups in all trials. Three outcomes populate the
summary of findings for the main comparison: total
effective rate, life satisfaction Index B, and adverse
reactions rate. Table 2 shows the quality of the included
RCTs’ evidence of the outcomes as moderate and low
using the GRADE system.
Total Effective Rate
Fourteen randomized controlled trials compared the
total effective rate of GBP vs. CBZ in the treatment of
TN.17–20,22–31 In these studies, ten described the total
effective rate after 4 weeks,17,18,20,22–28 two after
10 weeks,29,31 and only one study described that after 3
and 7 weeks, respectively.19,30 A total of 588 patients
were randomized to the GBP therapy group, and 568
patients were randomized to the CBZ therapy group. A
low degree of heterogeneity between the trials was
observed (x2 = 13.08, P = 0.441, I2 = 0.6%), so we
conducted a meta-analysis using fixed effects models.
The meta-analysis showed that both therapies had
therapeutic effect on TN, and the total effective rate of
GBP therapy group was similar with that of CBZ
therapy group (OR = 1.600, 95% CI 1.185, 2.161,
P = 0.002) (Figure 2).
Risk of bias across studies: Egger’s regression test
suggested no significant bias (P = 0.983), and neither
did Begg’s test (P = 0.661) (figure not shown).
We also did a subgroup analysis on the total efficiency
according to the different treatment period. In subgroup
analyses, the results showed significant differences
between groups after a treatment duration of 4 weeks:
(OR = 1.495, 95% CI 1.061, 2.107, P = 0.022, hetero-
geneity: x2 = 7.12, P = 0.625, I2 = 0.0%). However,
we found similar efficacy between the two groups after a
treatment duration of 10 weeks: (OR = 0.723, 95% CI
0.236, 2.222, P = 0.572, heterogeneity: x2 = 0.00,
P = 0.996, I2 = 0.0%).
Life Satisfaction Index B (LSI-B)
The number of patients who experienced life satisfaction
improvement was provided in two studies.18,21 Both
studies demonstrated that the life satisfaction improve-
ment in the GBP therapy group is better than that in the
CBZ therapy group after a duration of 4 weeks. There
were 118 patients enrolled in the study: A total of 60
patients were randomized to the GBP therapy group,
and 58 patients were in the CBZ therapy group. No
 4 
YUAN ET AL.

Table 1. Characteristics of Included Studies

No. of Treatment Follow-Up

Participants Male Patients Duration Period Risk of Bias
Study, Year (GBP/CBZ) Age (year) (GBP/CBZ) Dosage Regimen (GBP) Dosage Regimen (CBZ) (week) (year) Outcome Assessment Assessment

Zhu, 2008 45/46 GBP: 61.03  8.3 23/22 ID: 300 mg qn ID: 100 mg qn 4 1.5 Vas score, effectiveness, ADR L-U-L-L-L-L-U
CBZ: 59.69  8.72 MD: 3,600 mg/day MD: 1,200 mg/day
Qu, 2013 25/23 GBP: 59.6  5.2 11/10 ID: 300 mg qn ID: 100 mg qn 4 NR Vas score, effectiveness, ADR, LSI-B L-U-U-U-H-L-U
CBZ: 60.7  4.1 MD: 2,400 mg/day MD: 1,200 mg/day
Huang, 2012 48/46 GBP: 56.13  3.75 22/23 ID: 300 mg qn ID: 100 mg bid 4 2 Vas score, effectiveness, ADR U-U-U-U-H-L-U
CBZ: 55.73  4.97 MD: 1,800 mg/day MD: 900 mg/day
ZHou, 2006 47/36 GBP: 18~81 NR ID: 100 mg tid ID: 200 mg tid 3 1 Vas score, effectiveness, ADR L-U-U-U-H-L-U
CBZ: 18~81 MD: 2,000 mg/day MD: 1,200 mg/day
Gu, 2010 34/34 GBP: 52.13  5.75 16/15 ID: 100 mg tid ID: 200 mg/day 10 NR Vas score, effectiveness, ADR U-U-U-U-H-L-U
CBZ: 53.57  5.14 MD: 2,400 mg/day ID: 900 mg/day
Lu, 2013 29/29 GBP: 52.11  5.78 13/12 ID: 100 mg tid ID: 100 mg tid 10 NR Vas score, effectiveness, ADR U-U-U-U-H-L-U
CBZ: 53.54  5.19 MD: 2,400 mg/day MD: 2,400 mg/day
Yang, 2008 78/80 GBP: 37~78 25/26 ID: 300 mg qn ID: 100 mg bid 4 1.5 Effectiveness, ADR L-U-U-U-U-L-U
CBZ: 36~82 MD: 1,800 mg/day MD: 800 mg/day
ZHao, 2012 51/51 GBP: 21~87 28/29 ID: 100–300 mg tid ID: 100 mg bid 4 1 ADR U-U-U-U-H-L-U
CBZ: 21~87 MD: 3,600 mg/day MD: 600 mg/day
Guo, 2011 20/20 GBP: 22~89 NR ID: 100 mg tid ID: 200 mg tid 4 5 Vas score, effectiveness U-U-U-U-H-L-U
CBZ: 22~89 MD: 1,800 mg/day MD: 1,200 mg/day
Huang, 2010 31/31 GBP: 56.6  12.8 10/9 ID: 100 mg tid ID: 100 mg bid 4 NR Vas score, effectiveness, ADR L-U-U-U-H-L-U
CBZ: 57.9  13.4 MD: 2,400 mg/day MD: 800 mg/day
Guo, 2010 19/19 GBP: 22~89 NR ID: 100 mg tid ID: 200 mg tid 4 5 Effectiveness, ADR U-U-U-U-H-L-U
CBZ: 22~89 MD: 1,800 mg/day MD: 1,200 mg/day
Wu, 2011 20/16 GBP: 58.7  6.2 12/9 ID: 300 mg qn ID: 200 mg qn 4 1.6 Effectiveness, ADR U-U-U-U-H-L-U
CBZ: 56.5  6.2 MD: 3,600 mg/day MD: 1,200 mg/day
Jiang, 2014 50/47 GBP: 41~75 22/20 ID: 300 mg qn ID: 100 mg qn 4 3.5 Vas score, effectiveness, ADR U-U-U-U-H-L-U
CBZ: 42~74 MD: 3,600 mg/day MD: 1,200 mg/day
ZHou, 2014 43/43 GBP: 42~69 20/18 ID: 300 mg qn ID: 100~200 mg qn 4 1 Vas score, effectiveness, ADR U-U-U-U-H-L-U
CBZ: 41~70 MD: 3,600 mg/day MD: 1,600 mg/day
Wang, 2013 100/100 GBP: 33~68 42/40 ID: 100 mg tid ID: 100 mg bid 7 NR Effectiveness, ADR U-U-U-U-H-L-U
CBZ: 35~71 MD: 2,400 mg/day MD: 1,200 mg/day
Liu, 2014 35/35 GBP: 45~65 16/15 ID: 300 mg qd ID: 100 mg qd 4 NR Vas score, ADR, LSI-B L-U-U-U-H-L-U
CBZ: 46~65 MD: 2,400 mg/day MD: 1,200 mg/day

Risk of bias was evaluated for 7 criteria in order: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other
bias. Each criterion was indicated as a low risk of bias (L), a high risk of bias (H), and an uncertain risk of bias (U).
GBP, gabapentin; CBZ, carbamazepine; ID, initial dosage; MD, maximal dosage; NR, not reported; qd, one time daily; qn, Once a night; bid, twice daily; tid, three times daily; ADR, Adverse reaction rate; LSI-B, life satisfaction index B.
 Gabapentin vs. Carbamazepine in the Treatment of TN  5

Table 2. Assessment of Evidence Quality by the GRADE Criteria

Gabapentin Compared to Carbamazepine for Trigeminal Neuralgia

Patient or Population: Patients with Trigeminal Neuralgia
Settings:
Intervention: Gabapentin
Comparison: Carbamazepine

Illustrative Comparative Risks* (95% CI) Quality of

Assumed Risk Corresponding Risk Relative Effect No of Participants the Evidence
Outcomes Carbamazepine Gabapentin (95% CI) (Studies) (GRADE) Comments

Total effective rate 769 per 1,000 842 per 1,000 OR 1.600 1,156 ⊕⊕ΟΟ
VAS (798 to 878) (1.185 to 2.161) (14 studies) low1,2
Follow-up: 1 to 5 years
Life Satisfaction Index B The mean life satisfaction The mean life satisfaction 118 ⊕⊕ΟΟ
Scale from: 0 to 22 index b in the control index b in the intervention (2 studies) low1,2
groups was groups was
11 points 0.966 standard deviations
higher
(0.583 to 1.348 higher)
Adverse reaction rate 479 per 1,000 223 per 1,000 OR 0.312 1,293 ⊕⊕⊕Ο
Follow-up: 1 to 5 years (181 to 272) (0.240 to 0.407) (15 studies) moderate1

*The basis for the assumed risk (e.g, the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the
assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI, confidence interval; OR, odds ratio; VAS, visual analog scale.
GRADE Working Group grades of evidence.
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
1
The evidence was downgraded one level for serious study limitations.
2
Downgraded for possible imprecision as the result of wide confidence intervals.

Figure 2. Forest plot of the total effective rate of gabapentin vs. carbamazepine for the treatment of trigeminal neuralgia.

heterogeneity between the trials was observed analysis showed significant differences between the two
(x2 = 0.00, P = 0.992, I2 = 0.0%). We conducted a groups in a pooled analysis (SMD = 0.966, 95% CI
meta-analysis using fixed effects models. The meta- 0.583, 1.348, P < 0.001) (Figure 3).
6  YUAN ET AL.

Figure 3. Forest plot of the life satisfaction index B of gabapentin vs. carbamazepine for the treatment of trigeminal neuralgia.

Adverse Reaction Rate the efficacy of GBP interventions were similar with CBZ
in the treatment of trigeminal neuralgia (TN); However,
Fifteen studies examined the incidence of adverse reac-
in 4 weeks, GBP interventions were significantly more
tions.17–32 Among participants allocated to receive GBP,
effective than CBZ. Furthermore, the life satisfaction
170 of 654 (26.0%) experienced some adverse events,
index was higher than before in both groups, and the
mainly manifested as vertigo, somnolence, fatigue and
index of GBP group was higher than that of CBZ group
dizziness. And 306 of 639 (47.9%) participants allo-
after treatment. Most importantly, the adverse reaction
cated to receive CBZ presented with adverse events,
of GBP group was obviously lower than that of CBZ
mainly manifested as vertigo, somnolence, nausea, and
group, which was consistent with the conclusions of
fatigue. No heterogeneity between the trials was
some case reports and large sample studies.9,14,17,33,34
observed (x2 = 3.14, P = 0.999, I2 = 0.0%), so we
Some studies also show that GBP is effective for TN, but
conducted a meta-analysis using fixed effects models
the efficacy was also similar to CBZ.23 Whether the
(Figure 4). The meta-analysis showed that the
efficacy of GBP for the treatment of TN is indeed similar
adverse reaction rate of GBP therapy group was signif-
with than that of CBZ and the onset time is faster or
icantly lower than that of CBZ therapy group
whether the included studies might have been inade-
(OR = 0.312, 95% CI 0.240, 0.407, P < 0.001). Mean-
quately designed and the effects of GBP might have been
while, we did meta-analysis on individual adverse events
overestimated, further studies are required to substan-
such as vertigo, nausea, fatigue, and somnolence, and
tiate these conclusions.
almost all of the studies showed significant differences
To our knowledge, the potential mechanisms that
between the two groups.
underlie TN remain elusive. The central theory indicates
Risk of bias across studies: Egger’s regression test
that TN is caused by a similar seizure activity in the
suggested significant bias (P = 0.014), and so did Begg’s
central pathway of the trigeminal nerve, which
test (P = 0.029) (figure not shown).
explained the continued attacks and expansibility of
TN. Some scholars believed that TN was sensory
DISCUSSION seizures, and its site of action was in the hypothalamic
cortical or trigeminal nucleus.35 Studies indicated that
In this study, we performed a meta-analysis of all
anti-epileptic drugs had been used in neuropathic pain
available randomized controlled trials on the compar-
since the 1960s, and the newer anti-epileptic drugs, such
ative total effective rate, life satisfaction index B, and
as GBP, show considerable promise in the management
adverse events of gabapentin (GBP) treatment vs.
of TN in recent years,9,36–39 and the study also suggested
carbamazepine (CBZ) treatment. The results found that
 Gabapentin vs. Carbamazepine in the Treatment of TN  7
Figure 4. Forest plot of the adverse reactions of gabapentin vs. carbamazepine for the treatment of trigeminal neuralgia.
that GBP can be effective as first- or second-line
treatment of TN, even in cases resistant to traditional
CBZ treatment.14 Previous studies have shown that
GBP’s primary pain-resistant mechanism is as follows:
(1) It is combined with NMDA receptors, thus inhibiting
the activity of NMDA receptors, and then plays against
pain effects.40 (2) Raise the level of the effect of GABA
receptors in the brain, increasing the synthesis of GABA,
reducing degradation of GABA, for GBP is a synthetic,
just like neurotransmitter (GABA) drugs that can cross
the blood–brain barrier to produce inhibitory effect of
GABA, then produces sedative and analgesic effects.41
(3) When GBP reaches appropriate concentrations, it
combines with subtypes of voltage-dependent calcium
channel protein (a2/d) and then regulates calcium
channels and neurotransmitter.42
This systematic review summarized the best available
data on the effectiveness, life satisfaction index, and
safety of GBP compared with CBZ in the management of
TN patients. To our knowledge, this is the first systematic
review on this topic. All the relevant literature was
collected, and VAS score was used to assess the
effectiveness of different stages. The total effective
number was defined as the number of obvious effective
patients plus the effective number. Although we demon-
strate that the VAS scores significantly decrease in both
groups after treatment, GBP group’s total effective rate is
similar with that of CBZ group, but the adverse reactions
are lower than that of CBZ group. Meanwhile, the
statistical heterogeneity is small in analyses. However,
our search identified sixteen systematic reviews pub-
lished only in Chinese language, and the pain measure
would be affected by its subjective nature and the strong
influence of social context, emotion, and other nonphys-
iological variables. As we all know, the quality of life is
not only the natural state of human existence, but also the
advantages and disadvantages of social conditions. Life
satisfaction is one of the indicators of life quality
assessment.43 This study used LSI-B to evaluate patients’
life satisfaction and the results showed that among
patients who took GBP, the LSI-B improvement was even
more apparent. This phenomenon would be linked to
dizziness, drowsiness, nausea, vomiting, and other side
effects, which are factors influencing life satisfaction and
quality in patients, and GBP treatment is related to lower
incidence of these adverse reactions.
Although we had made great efforts to seek out all
randomized controlled trials about GBP compared with
CBZ for the treatment of TN, we could only find eligible
studies published in the Chinese language and surprised
to see that fewer English literature was found on this
subject, at least no randomized controlled trials. There-
8  YUAN ET AL.
fore, further higher-quality studies on this subject in
future are needed to determine the differences in
therapeutic and safety response to GBP in comparison
with CBZ in the treatment of TN.
According to the GRADE system, there are five
factors that may contribute to a downgrade quality of
evidence, including risk of bias, inconsistency of
results, indirectness of evidence, imprecision of results,
and publication bias. The evidence of the studies
included in our meta-analysis was downgraded one
level for serious study limitations for the risk of bias
and for possible imprecision as the result of wide
confidence intervals which resulted in possible exag-
geration of the results. However, our studies had
consistent results, with no indirectness of evidence and
publication bias. Therefore, the quality of evidence
was generally rated as moderate and low by the
GRADE criteria.
There were also many limitations in our meta-
analysis. Firstly, a limited number of studies were
included, although we strove to find all related studies.
Only sixteen eligible randomized controlled trials were
found published in Chinese, and no English literature
that met the requirements was found. Secondly, in this
review, sixteen eligible randomized controlled trials
were identified, but definitive randomization was found
only in six studies, and the other ten studies were found
to be lack of definitive randomization, in addition that
only one article described blinding method. And
allocation concealment was not applied in any study.
All these factors described above may lead to bias in
selection, performance, and detection and may result in
false-positive findings. Thirdly, according to the risk
evaluation of deviations, correctness of bias may affect
the conclusions of the study. Furthermore, different
aspects of individual studies were included in our
analysis, for example, diseases of different inclusion
and exclusion criteria, and different drug doses. There
were six studies in our research with ICHD’s diagnostic
criterion, while the other ten studies used other local
and foreign standards. Although all of the standards
served the same purpose, different standards made it
difficult to compare the results of different studies and
would affect the results of our meta-analysis. In
addition, there were differences among the included
studies in terms of processing time, the severity of the
disease, and results evaluation of time point, as well as
methods for evaluating results. The above problems
could not be evaluated with sufficient studies and data,
and it may affect our results more or less. Therefore,
more studies concerning the above problems are
required in future and they will help to clarify the
specific problem more clearly on GBP vs. CBZ for
treatment of TN.
CONCLUSIONS
Based on the available evidence, it is not possible to
draw conclusions regarding the efficacy and side effects
of gabapentin being superior to carbamazepine.
REFERENCES
1. Tamburin S, Schweiger V, Magrinelli F, Brugnoli MP,
Zanette G, Polati E. Effect of 5% lidocaine medicated plaster
on pain intensity and paroxysms in classical trigeminal
neuralgia. Ann Pharmacother. 2014;48:1521–1524.
2. Carmichael A. Trigeminal Neuralgia: rapidly success-
ful treatment. Br Dent J. 2014;217:388.
3. Zakrzewska JM, Linskey ME. Trigeminal neuralgia.
BMJ. 2014;348:g474.
4. Katusic S, Beard CM, Bergstralh E, Kurland LT.
Incidence and clinical features of trigeminal neuralgia,
Rochester, Minnesota, 1945–1984. Ann Neurol. 1990;27:
89–95.
5. Di Stefano G, La Cesa S, Truini A, Cruccu G. Natural
history and outcome of 200 outpatients with classical trigem-
inal neuralgia treated with carbamazepine or oxcarbazepine in
a tertiary centre for neuropathic pain. J Headache Pain.
2014;15:34.
6. Cruccu G, Gronseth G, Alksne J, et al. AAN-EFNS
guidelines on trigeminal neuralgia management. Eur J Neurol.
2008;15:1013–1028.
7. Attal N, Cruccu G, Haanpaa M, et al. EFNS guidelines
on pharmacological treatment of neuropathic pain. Eur J
Neurol. 2006;13:1153–1169.
8. Flickinger JJ, Kim H, Kano H, et al. Do carba-
mazepine, gabapentin, or other anticonvulsants exert sufficient
radioprotective effects to alter responses from trigeminal
neuralgia radiosurgery? Int J Radiat Oncol Biol Phys.
2012;83:e501–e506.
9. Lemos L, Flores S, Oliveira P, Almeida A. Gabapentin
supplemented with ropivacain block of trigger points improves
pain control and quality of life in trigeminal neuralgia patients
when compared with gabapentin alone. Clin J Pain.
2008;24:64–75.
10. Sindrup SH, Jensen TS. Pharmacotherapy of trigeminal
neuralgia. Clin J Pain. 2002;18:22–27.
11. Taylor JC, Brauer S, Espir ML. Long-term treatment
of trigeminal neuralgia with carbamazepine. Postgrad Med
J. 1981;57:16–18.
12. Wiffen PJ, McQuay HJ, Edwards JE, Moore RA.
Gabapentin for acute and chronic pain. Cochrane Database
Syst Rev. 2005:D5452.
 Gabapentin vs. Carbamazepine in the Treatment of TN  9
13. Backonja MM. Anticonvulsants (antineuropathics) for
neuropathic pain syndromes. Clin J Pain. 2000;16:S67–S72.
14. Cheshire WJ. Defining the role for gabapentin in the
treatment of trigeminal neuralgia: a retrospective study. J Pain.
2002;3:137–142.
15. Solaro C, Messmer UM, Uccelli A, Leandri M,
Mancardi GL. Low-dose gabapentin combined with either
lamotrigine or carbamazepine can be useful therapies for
trigeminal neuralgia in multiple sclerosis. Eur Neurol.
2000;44:45–48.
16. Steiner TJ, Olesen J. Implementing the international
classification of headache disorders, 2nd edition (ICHD-II).
Handb Clin Neurol. 2010;97:147–159.
17. Zhu DP, Zhang YN, Xu PR. Randomized control trial
of gabapentin and carbamazepine in the treatment of primary
trigeminal neuralgia. West China Med J. 2008;23:790–792.
18. Qu PS, Wan HF, Wang Z, Yu JM, Tao F. Randomized
control trial of gabapentin and carbamazepine in the treatment
of recurrent trigeminal neuralgia. Chin Gen Pract.
2013;16:2444–2446.
19. Huang KW, Sun GY, Chen JX. Controlled trial of
gabapentin and carbamazepine in the treatment of primary
trigeminal neuralgia. Chin J Pract Nerv Dis. 2012;15:33–35.
20. Zhou B, Wu XY, Song XZ, Liu L. Clinical observation
of gabapentin in the treatment of primary trigeminal neuralgia.
J Clin Anesthesio. 2006;22:864–865.
21. Gu WX, Chen H, Ye XR, Zhu J, Chen XH. Contrast
study on treatment of primary trigeminal neuralgia with
gabapentin. Mod Diagn Treat. 2010;21:8–10.
22. Lu YQ. Clinical observation of gabapentin in the
treatment of primary trigeminal neuralgia. Mod Diagn Treat.
2013;24:1114–1115.
23. Yang QH, Wang YX, Mei D, Guo Q. Effect of
gabapentin on primary trigeminal neuralgia. Pain Clin J.
2008;4:268–270.
24. Zhao XL, Xu YG, Song WG, Fu ZJ. Clinical use of
antiepileptic drugs in the management of chronic pain. Chin J
Pain Med. 2012;18:73–77.
25. Guo YP, Kan HQ. Clinical observation of gabapentin
combined with nerve block in the treatment of trigeminal
neuralgia. Chin J Pain Med. 2011;17:380–381.
26. Huang LF, Shi JJ, Wu HF, Lin SQ, Dai SB. Carba-
mazepine combined Gabapentin in the treatment of primary
trigeminal neuralgia. Chin J Prim Med Pharm. 2010;17:3002–
3003.
27. Guo YP, Chen HQ, Xiao TQ, Zhu J, Yi L, Nie H.
Clinical observation of gabapentin combined with nerve block
in the treatment of trigeminal neuralgia. Neurology Medical
Association of Guizhou province, the eighth annual Confer-
ence. China’s Guizhou anshun. 2010: 2.
28. Wu X, Ning QM, Yan XF. Effects of gabapentin
on plasma substance P and b-endorphin of patients with
trigeminal neuralgia. Chin J Pain Med. 2011;17:663–665.
29. Jiang PP. Control study of gabapentin and carba-
mazepine in the treatment of primary trigeminal neuralgia.
Med Recap. 2014;20:2478–2479.
30. Zhou L. Comparison and research on the clinical effects
of Gabapentin and Carbamazepine in the treatment primary
trigeminal neuralgia. Chin J Pract Nerv Dis. 2014;17:59–61.
31. Wang S. Gabapentin in treatment of primary trigem-
inal neuralgia with clinical analysis of 100 cases. China Health
Care Nutr. 2013;23:7549–7550.
32. Liu MJ. Clinical comparison of gabapentin and carba-
mazepine in the treatment of recurrent trigeminal neuralgia.
Gansu Med J. 2014;33:759–760.
33. Sist T, Filadora V, Miner M, Lema M. Gabapentin for
idiopathic trigeminal neuralgia: report of two cases. Neurol-
ogy. 1997;48:1467.
34. Solaro C, Lunardi GL, Capello E, et al. An open-label
trial of gabapentin treatment of paroxysmal symptoms in
multiple sclerosis patients. Neurology. 1998;51:609–611.
35. Rusu MC. The spinal trigeminal nucleus–considera-
tions on the structure of the nucleus caudalis. Folia Morphol
(Warsz). 2004;63:325–328.
36. Prisco L, Ganau M, Bigotto F, Zornada F. Trigeminal
neuralgia: successful antiepileptic drug combination therapy in
three refractory cases. Drug Healthc Patient Saf. 2011;3:43–45.
37. Rozen TD. Antiepileptic drugs in the management of
cluster headache and trigeminal neuralgia. Headache. 2001;41
(Suppl 1):S25–S32.
38. Karlov VA, Savitskaia ON. [Comparative effectiveness
of antiepileptic preparations in the treatment of patients with
trigeminal neuralgia]. Zh Nevropatol Psikhiatr Im S S
Korsakova. 1980;80:530–535.
39. Bergouignan M. Anti-epileptic drugs in the treatment
of essential trigeminal neuralgia. Presse Med. 1970;78:1832–
1834.
40. Chen SR, Eisenach JC, McCaslin PP, Pan HL. Syner-
gistic effect between intrathecal non-NMDA antagonist and
gabapentin on allodynia induced by spinal nerve ligation in
rats. Anesthesiology. 2000;92:500–506.
41. Smith TE, Chong MS. Neuropathic pain. Hosp Med.
2000;61:760–766.
42. Jensen TS. Anticonvulsants in neuropathic pain:
rationale and clinical evidence. Eur J Pain. 2002;6(Suppl
A):61–68.
43. Hills D, Joyce CM. Workplace aggression in clinical
medical practice: associations with job satisfaction, life
satisfaction and self-rated health. Med J Aust. 2014;201:535–
540.