The Journal of Pain, Vol -, No - (-), 2010: pp 1-9

Available online at www.sciencedirect.com

Critical Review
Systematic Review of the Comparative Effectiveness
of Antiepileptic Drugs for Fibromyalgia
Anne Chamberlin Siler, Hallie Gardner, Keenan Yanit, Tera Cushman,
and Marian McDonagh
Oregon Health & Science University, School of Medicine, Portland Oregon.

Abstract: Fibromyalgia is a difficult-to-treat chronic pain syndrome that affects 2% of the US

population. Pregabalin is an antiepileptic recently FDA approved for fibromyalgia treatment. Other
antiepileptics have been suggested for treatment. This systematic review examines the relative
benefits and harms of antiepileptic drugs in the treatment of fibromyalgia. A literature search was
conducted and 8 studies matched criteria (7 studies of pregabalin, 1 of gabapentin). Both drugs
reduced mean pain scores more than placebo at a modest rate (pregabalin, 38% to 50%; gabapentin,
51%). In a 6-month trial of pregabalin responders, 32% continued to have response at 6 months, with
a mean time to loss of response of 34 days. Compared to placebo, the drugs had similarly high rates
of adverse events and withdrawals. Without a head-to-head trial it is not possible to conclude if 1
antiepileptic is more effective or harmful than the other, although limited evidence suggests poten-
tial differences. Future studies must directly compare the drugs, include a more broadly defined
population, examine long term benefits and harms, and include cointerventions. We conclude that
pregabalin and gabapentin are modestly effective for the treatment of fibromyalgia but that their
long-term safety and efficacy remain unknown.
Perspective: This systematic review evaluates the benefits and harms of using the antiepileptic
drugs gabapentin and pregabalin for the treatment of fibromyalgia. Conclusions from this paper
can help clinicians to more effectively treat the pain associated with fibromyalgia.
ª 2010 by the American Pain Society
Key words: Review, antiepileptic, fibromyalgia, gabapentin, pregabalin, anticonvulsant.

F
ibromyalgia is a chronic syndrome characterized by
widespread musculoskeletal pain.11 It is defined by the
American College of Rheumatology as a combination
of pain for a minimum of 3 months and tenderness in 11
of 18 specific sites on the body.29 Other symptoms can in-
clude fatigue, anxiety, depression, sleep disturbances, bowel
dysfunction, and joint swelling.30 This syndrome is also asso-
ciated with increased work absenteeism, disability, and
decreased quality of life.17 Fibromyalgia affects 2% of the
U.S. population, predominantly women,6 although its
prevalence is likely underestimated and was more recently
Received March 11, 2010; Revised August 31, 2010; Accepted September
23, 2010.
Address reprint requests to Marian McDonagh, PharmD, Associate
Professor, Department of Medical Informatics and Clinical Epidemiology
3181 S.W. Sam Jackson Park Road, Mail Code BICC, Portland, OR 97239-
3098. E-mail: mcdonagh@ohsu.edu
1526-5900/$36.00
ª 2010 by the American Pain Society
doi:10.1016/j.jpain.2010.09.007
estimated at 3.3% in Canada.19 Prevalence is highest in
those aged 60 to 79 years, so the overall prevalence in the
US will continue to grow as our population ages.30 Although
the exact etiology is unknown, growing evidence suggests
that fibromyalgia may be a syndrome of dysfunctional
pain processing in the central nervous system. This abnormal
pain processing may be associated with multiple pathways,
including central pain sensitization and alterations in neu-
rotransmitters.4,25 Treatment of fibromyalgia is usually
multifactorial and concentrated on symptomatic relief.
Pharmacological interventions frequently include
nonsteroidal anti-inflammatory drugs (NSAIDS), opioids,
sedatives, muscle relaxants, and analgesics, but relief is of-
ten elusive.8,13,27 While nonpharmacological interventions
are frequently used, recent research has focused on finding
more effective pharmacological treatments for the chronic
pain associated with this disorder.16
Interest in antiepileptic drugs for treatment of fibro-
myalgia has increased in recent years. Antiepileptic drugs,

1
2 The Journal of Pain
also known as anticonvulsant drugs, are a diverse group of
medications initially developed for the treatment of sei-
zures. Some of the drugs in this group have also been
used to effectively treat neuropathic pain.14,18
Neuropathic pain shares some similarities to fibromyalgia
in etiology and symptoms experienced. Pregabalin
(Lyrica), a classic antiepileptic drug, was approved by
the US Federal Drug Administration for use to treat
fibromyalgia in 2007. Pregabalin and gabapentin are
structurally similar to the inhibitory neurotransmitter
gamma-aminobutyric acid (GABA). Both anticonvulsants
bind with high affinity to the alpha-2-delta subunit of
voltage-gated calcium channels in the CNS. In vitro, prega-
balin and gabapentin reduce the calcium-dependent
release of several neurotransmitters, including glutamate,
norepinephrine, calcitonin gene-related peptide, and
substance P. Although the exact mechanism of action of
these drugs is not known, it is thought that this action
may be involved in their analgesic and anticonvulsant
effects.21 It is not yet clear how pregabalin compares to
other antiepileptic drugs or other treatments in terms of
overall effectiveness when considering both benefits and
harms. Antiepileptics are a diverse group of drugs with
wide variation in mechanisms of action, adverse event
profiles, and potential efficacy in treating fibromyalgia.
It will be important for clinicians to understand the avail-
able evidence regarding their use in this difficult-to-treat
disease.
This systematic review analyzes the clinical evidence on
comparative benefits and harms of antiepileptic drugs
used to treat fibromyalgia. We posed the following
critical questions: 1) Can antiepileptic drugs effectively
treat fibromyalgia and is one drug more effective than
the others? 2) What are the adverse effects of antiepilep-
tic use and is one drug more harmful than the others?
3) Are there differences in effectiveness of antiepileptics
across subpopulations of patients with the disease, eg,
patients with comorbidities?
Methods
Inclusion Criteria
Studies eligible for this review had inclusion criteria of
adults ($18 years) who had been diagnosed with fibro-
myalgia according to American College of Rheumatology
standards.29 We included studies of any antiepileptic
drug compared to placebo or another antiepileptic using
a randomized controlled trial or observational design.
Outcomes included were the impact on pain (the propor-
tion of patients with response, change in pain score from
baseline, functional status, or sleep quality) and adverse
events (discontinuation from study due to adverse
events, and the type and incidence of adverse events).
Literature Search
To identify relevant studies, we conducted literature
searches in Ovid Medline (1996-2009, including in-
process and nonindexed citations), the Database of
Abstracts of Reviews of Effects, and the Cochrane Registry
of Controlled Trials. We used the search terms fibromyal-
Antiepileptic Drugs for Fibromyalgia
gia/or fibromyalg$ and the following antiepileptic drugs:
carbamazepine, valproic acid, ethotoin, gabapentin,
lamotrigine, levetiracetam, methsuximide, oxcarbaze-
pine, phenobarbital, phenytoin, primidone, pregabalin,
tiagabine hydrochloride, topiramate, divalproex sodium,
and zonisamide. We used Medical Subject Headings as
search terms when available or key words when appropri-
ate. Searches were limited to human and English language.
The timeframe of the searches was limited to 1996-2009.
Additionally, manual searches of the reference lists of in-
cluded studies, reviews, and documents on the FDA Center
for Drug Evaluation Research web site were conducted.
Study Selection
Two reviewers independently screened each title and
abstract for potential full-text review, applying the
aforementioned inclusion criteria. Disagreements were
resolved through discussion and consensus. After retriev-
ing the full-text of these studies, each was again indepen-
dently screened for eligibility by 2 reviewers with
disagreements resolved through consensus to arrive at
the final set of included studies.
Data Abstraction
Two reviewers abstracted the following data from each
included study: study design, eligibility criteria, interven-
tion (drug, dosage, and length of treatment), other
permitted concurrent medications and interventions,
methods of outcome assessment, study demographics,
sample size, loss to follow-up, reported adverse events,
withdrawals due to adverse events, and results. We
recorded intention-to-treat (ITT) results when available.
Abstracted data were checked by a second reviewer.
Quality Assessment
Internal validity of included studies was graded ‘‘good’’,
‘‘fair’’, or ‘‘poor’’ based on the US Preventive Services Task
Force (USPSTF) criteria.15 Elements of internal validity
assessment for trials include methods of randomization
and allocation concealment, the presence of balanced
treatment arms at baseline, ITT analysis, consideration of
confounders in analysis, and overall and differential loss
to follow-up. For each identified study, 2 reviewers inde-
pendently rated quality with any disagreements resolved
through discussion and consensus. According to the
USPSTF methods, good studies meet all criteria; poor
studies fail to meet combinations of criteria thought to
constitute a fatal flaw, and the remainder are fair quality.
Synthesis
We analyzed results qualitatively and quantitatively.
Our primary outcome measure for analyses was response
to treatment, defined as a reduction of pain greater than
or equal to 30% from baseline to end point using an
11-point pain scale. Secondary measures included dura-
tion of response measured in days, overall discontinua-
tion rates, and rates of specific adverse events. We
examined discontinuations from study due to any cause
and due to adverse events, and we assessed the rates of
Siler et al
the most commonly occurring adverse events. Where we
had at least 2 trials reporting the same outcome, we con-
ducted meta-analyses based on ITT populations using
a Mantel Haenszel random-effects model in MetaAnalyst
Beta 3.1.3. Where findings were statistically significant,
numbers needed to treat (NNT) or harm (NNH) were
calculated. Statistical heterogeneity was assessed using
the Q and the I2 statistics. The volume of evidence
available did not allow formal analysis of publication
bias.
Results
Our electronic literature searches identified 38
citations, and 2 additional studies were later found on
ClinicalTrials.gov (Fig 1). No additional studies were
identified from reference lists or documents on the
FDA web site. After dual review, we included 8 studies
involving 3,114 patients (Table 1): 2,964 in pregabalin
trials and 150 in a gabapentin trial. No study directly
compared 1 antiepileptic drug to another and no studies
of other antiepileptic drugs were found. Five random-
ized controlled trials compared pregabalin to pla-
cebo,3,9,10,22 (Pfizer Protocol #A0081100, ClinicalTrials.
gov, and ClinicalStudyResults.org 2009), 1 compared
gabapentin to placebo,2 an open-label extension study
of pregabalin alone evaluated longer-term safety (Pfizer
Protocol # A0081101, ClinicalTrials.gov 2009) and a non-
randomized, single-arm study examined the effects of
pregabalin added to quetiapine.5 Additionally, we
included a post hoc analysis of 1 of the pregabalin trials
Figure 1. Study Eligibility Flow Diagram.
The Journal of Pain 3
exploring the impact of anxiety or depression at baseline
on improvement in pain scores.1
The trials were conducted in populations that were
predominantly women aged 47 to 50 years (Table 1).
Six were short-term trials (8–14 weeks) with primary end-
points relating to pain response.2,3,5,9,22 (Pfizer Protocol
# A0081100, ClinicalTrials.gov, and ClinicalStudyResults.
org 2009) The longest study was 26 weeks in duration,
assessing the durability of the response to pregabalin
achieved during a run-in period of 6 weeks.10 During
the 6-week open-label run-in with dose escalation,
patients were removed from the study due either to
lack of response or lack of tolerability. To continue on
to the double-blind period, patients also had to meet 2
response criteria.
The study of gabapentin differed from the randomized,
controlled trials of pregabalin in that it was smaller (150
compared to 529–748), a lower proportion of patients
enrolled were female (68% compared to 91–100%), and
dosing was handled differently (Table 1).1 Gabapentin
was titrated to final dose over 5 weeks with a minimum
final dose (1,200 mg daily), while pregabalin was titrated
over 1 to 3 weeks, depending on the study. Baseline sever-
ity of pain in these studies is also presented in Table 1.
Based on an 11-point scale, patients in the gabapentin trial
had baseline mean pain scores approximately 1 point
lower (5.7–6.0) than patients in the short-term trials of
pregabalin (6.6–7.3). Patients in the 26-week pregabalin
study (at open-label baseline) and patients in the
quetiapine add-on study had higher baseline pain scores
(mean 7.8).1,10
4 The Journal of Pain Antiepileptic Drugs for Fibromyalgia
Table 1. Trials of Pregabalin and Gabapentin
DESIGN DAILY DURATION of ENROLLED MEAN % BASELINE PAIN
AUTHOR YEAR FUNDER INTERVENTION DOSE (mg) STUDY (WEEKS) (N) AGE FEMALE SCORE (0–10)
Arnold 2007 RCT Gabapentin 2,400* 12 150 48 68** 5.7–6.0
NIH
Crofford 2005 RCT Pregabalin 150, 300, 450 8 529 50 91 6.9–7.3
Pfizer
Mease 2008 RCT Pregabalin 300, 450, 600 13 748 49 94 7.0–7.2
Pfizer
Arnold 2008 RCT Pregabalin 300, 450, 600 14 745 50 95 6.6–6.7
Pfizer
Crofford 2008y RCT Pregabalin 300, 450, 600 26 566 49 93 7.8***
Pfizer
Pfizer Protocol # RCT Pregabalin 300, 450, 600 14 735 49 91 NR
A0081100 Pfizer
Pfizer Protocol # Open-label Pregabalin 150 – 600 12 357 48 90 NR
A0081101 extension
Pfizer
Calandre 2007 Non-random Pregabalin added 180 (mean) 12 19 47 100 7.8
NR to quetiapine

Abbreviation: NR, not reported.

*Doses titrated from 300 mg per day to 2,400 mg per day over 5 weeks with a minimum final dose of 1,200 mg a day.
yEnriched enrollment study.
**Computed from means reported for placebo and drug groups.
***Converted from visual analog scale results.

The internal validity of all the randomized controlled
trials was fair. There were no good-quality studies be-
cause the methods of blinding, allocation concealment,
and randomization were unclear. Considering the titra-
tion of doses in the active drug arms, lack of description
of blinding methods is important. Although loss to
follow up was generally high, ITT analysis was performed
and groups were similar at baseline. The gabapentin trial
was government funded and the pregabalin RCTs were
funded by the pharmaceutical manufacturer. The source
of funding for the quetiapine add-on study was not
stated.1
Can Antiepileptic Drugs Effectively Treat
Fibromyalgia and Is One Drug More
Effective Than the Others?
Compared to placebo, pregabalin and gabapentin were
superior in achieving short-term response (Table 2). The
highest rate of response was found with 1,800 mg per
day of gabapentin (51%), although this study was the
smallest of the 5 RCTs (N = 150). The range of response
with pregabalin was 26 to 50%. The pooled response
rate among patients receiving placebo was 28% (95% CI,
19 to 35%) For the single gabapentin study, the relative
risk of response compared to placebo was 1.7 (95% CI,
1.1 to 2.5) with a NNT of 5. The pooled relative risk for re-
sponse with pregabalin compared with placebo was 1.4
(95% CI, 1.3 to 1.6, Fig 2), with a NNTof 8. Statistical hetero-
geneity was not found (I2 = 0) in these analyses using a Der-
Simonian and Laird random effects model. Analysis of
response stratified by dose did not suggest a dose-
response effect (Fig 2), although the 450-mg doses
resulted in a slightly higher pooled response rate
compared to 300 mg or 600 mg per day (Table 2).
Durability of the response to pregabalin compared to
placebo was evaluated in a group of patients first enrolled
in a 6-week open-label run-in period with pregabalin.10 At
the end of 6 weeks, those who had responded to and
tolerated the drug were randomized to 26 weeks of

Table 2. Short-Term Response to Treatment With Pregabalin Or Gabapentin

DAILY DOSE (mg)
STUDY DURATION
AUTHOR YEAR DRUG N WEEKS 300 450 600 2400 PLACEBO
Crofford 2005 Pregabalin 529 8 38%* 48% – – 27%
Arnold 2008 Pregabalin 745 14 42% 50% 48% – 30%
Mease 2008 Pregabalin 748 13 43%* 43%* 44%* – 35%
Pfizer Protocol # Pregabalin 735 14 32% 33% 26% 19%**
A0081100
Arnold 2007 Gabapentin 150 12 – – – 51% 31%

Abbreviation: NR, not reported.

*Difference between drug and placebo not statistically significant.
**Patients who responded to placebo during 1-week run-in period (24.3%) were excluded from study.
Siler et al The Journal of Pain 5

Figure 2. Forest plot of meta-analysis comparing pooled effectiveness of pregabalin in reducing pain by 30% on an 11-point scale for
3 placebo controlled trials.

pregabalin (300, 450, or 600 mg) or placebo. This study What Are the Adverse Effects of
differed from those presented above because in order to Antiepileptic Use and Is One Drug More
continue into the randomized portion of the study,
Harmful Than the Others?
patients had to meet 2 criteria for response. First, patients
The short-term incidences of adverse events associated
had to have >50% reduction in pain based on visual analog
with gabapentin and pregabalin were reported in 4 fair-
scale (VAS) score from baseline. The short-term trials
quality trials.2,3,9,22 Dizziness, headache, somnolence,
required >30% reduction in pain to meet criteria. Second,
and edema were the most commonly reported adverse
a rating of ‘‘much improved’’ or ‘‘very much improved’’ on
events for both drugs (Table 3). Dizziness was the most
the patient-rated Global Impression of Change scale was
commonly reported adverse event with pregabalin
also required. Of the 1,050 patients enrolled in the
(38%), while headache (27%) was the most common
6-week portion of the trial, only 54% enrolled in the
adverse event with gabapentin. The confidence intervals
6-month trial. The number of days to loss of therapeutic
for headache with the 2 drugs do not overlap, suggest-
response was the primary outcome. Loss of therapeutic re-
ing a statistically significant difference in incidence
sponse was defined as either <30% reduction in pain VAS
between the drugs. We could not conduct an adjusted
score relative to baseline or worsening in the judgment of
indirect meta-analysis of these data because there is
the investigator. All pregabalin doses (300, 450, and 600
only 1 study of gabapentin.
mg daily) resulted in significantly longer time to loss of
While the 4 most common adverse events in Table 3
therapeutic response than placebo, with 300 mg daily
were common to both drugs, other adverse events varied
having the largest difference.
by drug (Table 4). Weight gain was the most common
The mean time to loss of response for pregabalin
other adverse event for pregabalin, with 11% of patients
patients was 34 days (95% CI, 21 to 48) compared to
reporting this effect. The amount of weight or the
7 days (95% Cl, 5 to 9) with placebo (Kaplan-Meier
proportion of patients with clinically important weight
P < .0001). At 6 months, the proportion of patients
gain was not reported in most studies. The difference
with loss of response was 32% in the pregabalin group
compared to placebo was statistically significant, with
and 61% in the placebo group (relative risk 1.9 (95% CI,
a NNH of 11. The number of patients reporting dry
1.5 to 2.3); NNT = 3). These analyses only represent the
patients who lost response and do not include patients
who withdrew from treatment due to adverse events or Pooled Proportion of Patients
Table 3.
other reasons. Because all patients had been stabilized Reporting Common Adverse Events
on pregabalin prior to the randomized portion of the
ADVERSE PREGABALIN GABAPENTIN PLACEBO
study, it is possible that unblinding occurred when pa- EVENT (%, 95% CI) (%, 95% CI) (%, 95% CI)
tients were switched to placebo after the run-in period.
Dizziness 38% (34 to 41%) 25% (16 to 37%) 10% (8 to 12.5%)
Analysis excluding patients who withdrew in the first 8
Headache 9% (6.5 to 13%) 27% (17 to 38%) 5% (4 to 8%)
days resulted in statistically significant findings of 74
Somnolence 21% (17 to 24%) 19% (11 to 29%) 14% (6 to 21%)
days to loss of response for pregabalin, compared Edema 7% (3 to 9%) 16% (9 to 26%) 3% (.07 to 8%)
with 15 days for placebo (Kaplan-Meier P < .0001).
6 The Journal of Pain Antiepileptic Drugs for Fibromyalgia
Table 4. Adverse Events Reported by Drug In the 6-month study of pregabalin among patients with re-
sponse during run-in, significantly more patients in the
% (95% CI) RELATIVE RISK (95% CI) drug groups withdrew due to adverse events compared
Pregabalin to the placebo group (relative risk 2.2, 95% CI, 1.2 to 3.9),
Weight gain 11% (7 to 15%) 4.5 (2.8 to 7.3) corresponding to a NNH of 12.10 Analysis of withdrawals
Dry Mouth 8% (5 to 12%) 5.0 (2.7 to 9.1) due to adverse events by dose level in this case shows an in-
Amblyopia 6% (2 to 9%) 6.4 (2.6 to 15.7) verse relationship, with decreasing risk as dose increases.
Euphoria 6% (5 to 7%) 4.9 (1.1 to 21.0)
However, with limited numbers of subjects in each group
Gabapentin
this analysis should be considered exploratory.
Nausea 21% (13 to 32%) 1.0 (0.5 to 1.6)
Sedation 24% (15 to 35%) 6.0 (2.0 to 18.6)
Insomnia 12% (96 to 22%) 1.5 (0.6 to 3.9)
Lightheadedness 15% (8 to 25%) 11.0 (2.0 to 65.6)
Are There Differences in Effectiveness of
Antiepileptics Across Subpopulations?
The effect of depression and anxiety at baseline on the
mouth, amblyopia, and euphoria with pregabalin was response to pregabalin was examined in a post hoc
also significantly higher compared to placebo (NNH of analysis of the results of a placebo-controlled trial.1 A
17, 19, and 21, respectively). With gabapentin, sedation path analysis (based on linear regression models) was
(24%) and light-headedness (15%) occurred significantly conducted using the Hospital Anxiety and Depression
more often than with placebo (NNH of 5 and 8, respec- Scale and pain-score diaries. Approximately 25% of the
tively). Consistent with these findings from short-term pain relief was associated significantly with the
trials, data from the 6-month trial of pregabalin indi- combined improvements in anxiety and depression,
cated that dizziness, somnolence, headache, and weight divided evenly between the 2. The remaining 75% of
gain were the most common adverse events in the pain relief was not associated with improvements in
pregabalin group. anxiety or depression. Additionally, neither the anxiety-
While discontinuation due to any reason was not by-treatment nor the depression-by-treatment interac-
significantly different between drug and placebo groups tions were significant alone.
in short-term trials, discontinuation due to adverse events In a small study of 19 women with residual pain who
was significantly greater in pregabalin groups compared were taking quetiapine, the addition of pregabalin was
to placebo at 300, 450, and 600 mg per day with increasing found to improve pain scores.5 The study was too small
relative risks with increasing dose (Fig 3). As can be seen in to indicate whether these findings were statistically
Fig 3, withdrawal due to adverse events with gabapentin significant. While these women are described as having
was similar but not statistically significantly different than had response to quetiapine, the baseline mean pain
placebo, likely due to the smaller sample size in this study. score (7.8 out of 11) was higher than those in the

Figure 3. Forest plot of meta-analysis comparing rates of withdrawal from study due to adverse events.
Siler et al
randomized trials (Table 1), and the mean dose of prega-
balin was lower (180 mg per day). Almost a third (31.6%)
withdrew during the 12-week trial and approximately
half of these withdrew due to adverse events, with
light-headedness, dizziness, and dry mouth being the
most common reasons. This adverse event panel is similar
to that of the RCTs. Finally, this study reported that the
patients who continued to take pregabalin were free
of adverse events at the end of the study.
Examination of the impact of other factors such as age,
race, gender, and socioeconomic status on the benefits
or harms associated with pregabalin or gabapentin was
not possible due to narrow inclusion criteria and limited
reporting in the studies.
Discussion
A comprehensive review of the medical literature
revealed that current evidence of the efficacy of antiepi-
leptics in treating fibromyalgia is limited to studies of
pregabalin and gabapentin compared to placebo. There
is no direct evidence comparing antiepileptics to each
other or to other treatments. The evidence for gabapentin
is limited to a single, small short-term trial (150 patients
total) while the evidence for pregabalin includes multiple
trials including almost 3,000 patients. Both were found to
be better than placebo in reducing pain in the short-term;
however, in absolute terms, the magnitude of benefit is
limited with half of patients (at most) experiencing
a 30% decrease in pain (while up to one-third have a simi-
lar response when taking placebo). The NNTs for both
drugs were not impressive; for every 5 patients treated
with gabapentin or 8 patients treated with pregabalin
(rather than placebo) for 8 to 14 weeks, 1 additional
patient will have a response of at least 30% improvement
in symptoms. In general, for treatment comparisons, NNTs
of 2 to 4 are considered favorable.26 Of pregabalin
responders, approximately one-third continued to have
response at 6 months. While this rate of continued
response was better than for those who switched to
placebo, it is still disappointing considering the long-
term duration of symptoms in most patients. Anxiety
and depression at baseline did not impact response to
pregabalin, although this evidence should be considered
preliminary. The pregabalin trials differed from the gaba-
pentin trial in dosage in that all of the pregabalin trials
used fixed dosing, the gold standard for evaluating dose
response, while the single gabapentin trial used flexible
dosing.20 However, fixed dosing does not reflect real-
world clinical practice and is not as useful in evaluating
outcomes in the longer term as it can result in high
dropout rates due the lack of flexibility in making dose
changes for adverse events or inadequate response.
The short-term adverse events, which included dizziness,
headache, somnolence, and edema, were similar for both
drugs and generally occurred at slightly higher rates than
placebo. There is insufficient evidence to conclude
whether 1 antiepileptic drug is more harmful than an-
other. Discontinuation from the studies due to adverse
events was greater in the drug groups, with a dose-
response relationship present for pregabalin in short-term
The Journal of Pain 7
trials, but the longer-term trial finding was contradictory.
While it appears that gabapentin is slightly more
efficacious, it may have a higher rate of headache than pre-
gabalin. Only a direct comparison can determine the rela-
tive benefits and harms of these drugs.
The main components in assessing the strength of
evidence are precision of estimates, directness and consis-
tency of evidence, risk of bias, and magnitude of effect. If
this body of evidence were subjected to the criteria evalu-
ating the strength of the overall body of evidence, it
would merit only low strength of evidence.24 While studies
met most criteria for internal validity, the study design,
statistical analyses, and conclusions of the pregabalin stud-
ies may have been subject to further bias. Within these
placebo-controlled trials, the precision of the resulting
pooled estimates (based on confidence intervals) is fairly
good. Current evidence for the effectiveness of antiepilep-
tics in the treatment of fibromyalgia is limited by lack of
direct comparison between drugs. The evidence is indirect
because there are no direct comparisons of the drugs and
because effectiveness outcomes (eg, time lost from work)
are not reported. Based on the indirect evidence
(placebo-controlled efficacy trials), the risk of bias is
moderate and the magnitude of effect is modest at 40 to
50% response short-term with a large dissipation of effect
over 6 months.
The applicability (external validity) of the studies was
limited by the short duration of all clinical trials and
the selective patient population. While symptoms of
fibromyalgia often begin in the 40s, the peak prevalence
of symptoms occurs among a somewhat older popula-
tion, with one estimate of 7.4% among 70 to 79 year
olds,30 and another more recent estimate of 7.9% among
55 to 64 year olds.19 Fibromyalgia may also occur second-
arily to comorbidity, a situation that has not been
studied with these drugs. Additionally, with the
exception of 1 small study, the populations studied
were exclusively treated with acetaminophen plus the
trial drug. This situation does not represent the typical
population with fibromyalgia who are often on
multidrug regimens and further limits the applicability
of these studies.
To make better treatment decisions for patients with
fibromyalgia, clinicians need more information.
Longer-term comparative trials would help provide addi-
tional data on the effectiveness of pregabalin and gaba-
pentin. Future research is needed to determine the
applicability and relative effectiveness of these drugs,
with equal focus on benefits and risks. Such trials would
need to be pragmatic, including patients with a wide va-
riety of baseline symptomatology and comorbidity. Clini-
cians need studies that examine these issues and other
real-world effectiveness outcomes such as quality of life
and impact on social or functional factors (ability to
work, coincident depression, and sleep) over the long
term. Recent research indicates that patients with fibro-
myalgia incur significant costs every year, including med-
ical costs, pharmacy costs, and indirect costs that include
work-loss estimates.28 White et al28 found that the yearly
costs for a patient with fibromyalgia were over $10,000
in 2005, with indirect costs accounting for 29% of the
8 The Journal of Pain
total. Currently there is a price differential between the
drugs, further underlining the importance of future
comparison studies to patients, clinicians, and policy-
makers alike. In a UK analysis of the cost effectiveness
of pregabalin, 300 mg and 450 mg doses were found to
be cost-effective compared to placebo for the treatment
of fibromyalgia, with cost per quality adjusted life years
estimated to be £23,166 and £22,533, respectively. The
model did not use data for the 600-mg dose and was lim-
ited to evaluating the impact in patients with severe fi-
bromyalgia.7 Finally, future studies must also allow for
and adjust for cointerventions because fibromyalgia is
commonly treated with multiple drugs. Sample sizes
adequate to accommodate these requirements will be
required.
Potential limitations of our work include the use only of
English language studies and the use of the American
College of Rheumatology criteria for fibromyalgia as an
inclusion criterion. There may have been additional
studies of antiepileptic drugs in the non-English language
literature, although there is debate about the value of
including such literature.12,23 Allowing other methods of
diagnosing fibromyalgia would have broadened the
scope of the review, but this may have also reduced the
clarity of the applicability of the results. The primary
outcome of interest for this review was overall response
at the end of study, selected because it is a clinically
meaningful measure. In the primary trials, sample sizes
were calculated based on mean change in pain
compared to baseline. Thus, there is a chance that there
was inadequate power to show a statistically significant
difference using the response rate as a primary outcome.
With the analysis of pregabalin, this is not likely to be
a problem because the pooled estimate has a much
larger sample size than the individual studies. For the
References
1. Arnold LM, Crofford LJ, Martin SA, Young JP,
Sharma U: The effect of anxiety and depression on im-
provements in pain in a randomized, controlled trial of
pregabalin for treatment of fibromyalgia. Pain Med 8:
633-638, 2007
2. Arnold LM, Goldenberg DL, Stanford SB, Lalonde JK,
Sandhu HS, Keck PE, Welge JA, Bishop F, Stanford KE,
Hess EV, Hudson JI: Gabapentin in the treatment of fibro-
myalgia: A randomized, double-blind, placebo-controlled,
multicenter trial. Arthritis Rheum 56:1336-1344, 2007
3. Arnold LM, Russell IJ, Diri EW, Duan WR, Young JP,
Sharma U, Martin SA, Barrett JA, Haig G: A 14-week, ran-
domized, double-blinded, placebo-controlled monotherapy
trial of pregabalin in patients with fibromyalgia. J Pain 9:
792-805, 2008
4. Buskila D, Buskila D: Developments in the scientific and
clinical understanding of fibromyalgia. Arthritis Res Ther
11:242, 2009
5. Calandre EP, Morillas-Arques P, Rodriguez-Lopez CM,
Rico-Villademoros F, Hidalgo J: Pregabalin augmentation
of quetiapine therapy in the treatment of fibromyalgia:
An open-label, prospective trial. Pharmacopsychiatry 40:
68-71, 2007
Antiepileptic Drugs for Fibromyalgia
results of the gabapentin study, this is still a limitation of
our approach. However, the results of the mean change
in pain score resulted in similar findings to the analysis of
response.
Results of this review indicate that pregabalin and
gabapentin can be used for the treatment of fibromyal-
gia with moderate benefits in the short-term. However,
clinicians must be cautious not to generalize these results
to other unstudied antiepileptic drugs and realize the
limitations of this evidence with regard to the popula-
tions included, durations studied, and the potentials
for important differences in adverse effect profiles
between the drugs. Direct comparison between these
drugs, or to other drugs used to treat fibromyalgia
(eg, nonsteroidal anti-inflammatory drugs, tricyclic anti-
depressants, selective serotonin reuptake inhibitors,
benzodiazepines, and muscle relaxants) will be the only
way to determine the relative benefits and harms of
these drugs. These studies should be longer term, and
should include functional and health outcomes in order
to truly assess their relative utility in treating this difficult
disease.
Acknowledgments
Primary authors (Anne Chamberlin Siler, Hallie Gardner,
and Keenan Yanit) contributed equally to this work. We
would like to thank Khloth Lim, BS, Anthony Eskander,
BS, and Mollie Lane, MPH, for assistance with topic
development and data collection and Susan Carson,
MPH, Nancy Lee, PharmD, and Kim Peterson, MS, for their
assistance in development of the systematic review
methods and mentoring the process. No financial support
was provided for this project.
6. Chakrabarty S, Zoorob R, Chakrabarty S, Zoorob R: Fibro-
myalgia. Am Fam Physician 76:247-254, 2007
7. Choy E, Richards S, Borwin K, Watson P, Lloyd A,
Sadosky A, Zlateva G: Cost effectiveness of pregabalin in
treatment of fibromyalgia from a UK perspective. Curr
Med Res Opin 26:965-975, 2010
8. Crofford LJ: Pharmaceutical treatment options for
fibromyalgia. Curr Rheumatol Rep 6:274-280, 2004
9. Crofford LJ, Rowbotham MC, Mease PJ, Russell IJ,
Dworkin RH, Corbin AE, Young JP, LaMoreaux LK,
Martin SA, Sharma U: Pregabalin -105 Study G: Pregabalin
for the treatment of fibromyalgia syndrome: Results of a ran-
domized, double-blind, placebo-controlled trial. Arthritis
Rheum 52:1264-1273, 2005
10. Crofford LJ, Mease PJ, Simpson SL, Young JP, Martin SA,
Haig GM, Sharma U: Fibromyalgia relapse evaluation and
efficacy for durability of meaningful relief (FREEDOM): A
6-month, double-blind, placebo-controlled trial with
pregabalin. Pain 136:419-431, 2008
11. Deary V, Chalder T, Sharpe M: The cognitive behavioural
model of medically unexplained symptoms: A theoretical
and empirical review. Clin Psychol Rev 27:781-797, 2007
12. Egger M, Zellweger-Zahner T, Schneider M, Junker C,
Lengeler C, Antes G: Language bias in randomised
Siler et al
controlled trials published in English and German. Lancet
350:326-329, 1997
13. Goldenberg DL: Pharmacological treatment of fibro-
myalgia and other chronic musculoskeletal pain. Baillieres
Best Pract Res Clin Rheumatol 21:499-511, 2007
14. Goodyear-Smith F, Halliwell J: Anticonvulsants for
neuropathic pain: Gaps in the evidence. Clin J Pain 25:
528-536, 2009
15. Harris RP, Helfand M, Woolf SH, Lohr KN, Mulrow CD,
Teutsch SM, Atkins D: Methods Work Group Third U.S.
Preventive Services Task Force: Current methods of the US
Preventive Services Task Force: A review of the process. Am
J Prev Med 20:21-35, 2001
16. Hassett AL, Gevirtz RN: Nonpharmacologic treatment for
fibromyalgia: Patient education, cognitive-behavioral therapy,
relaxation techniques, and complementary and alternative
medicine. Rheum Dis Clin North Am 35:393-407, 2009
17. Hudson JI, Goldenberg DL, Pope HG Jr., Keck PE Jr.,
Schlesinger L: Comorbidity of fibromyalgia with medical
and psychiatric disorders. Am J Med 92:363-367, 1992
18. Jensen TS, Madsen CS, Finnerup NB: Pharmacology and
treatment of neuropathic pains. Curr Opin Neurol 22:
467-474, 2009
19. Lawrence RC, Felson DT, Helmick CG, Arnold LM, Choi H,
Deyo RA, Gabriel S, Hirsch R, Hochberg MC, Hunder GG,
Jordan JM, Katz JN, Kremers HM, Wolfe F: Estimates of the
prevalence of arthritis and other rheumatic conditions in
the United States. Part II. Arthritis Rheum 58:26-35, 2008
20. Lipkovich I, Adams DH, Mallinckrodt C, Faries D,
Baron D, Houston JP: Evaluating dose response from flexible
dose clinical trials. BMC Psychiatry 8:1471-1480, 2008
21. McEvoy GK, ed: AHFS: Drug Information. Bethesda, MD,
American Society of Health-System Pharmacists: 1125-1126,
2008
22. Mease PJ, Russell IJ, Arnold LM, Florian H, Young JP,
Martin SA, Sharma U: A randomized, double-blind,
placebo-controlled, phase III trial of pregabalin in the treat-
The Journal of Pain 9
ment of patients with fibromyalgia. Journal Rheumatol 35:
502-514, 2008
23. Moher D, Pham B, Klassen TP, Schulz KF, Berlin JA,
Jadad AR, Liberati A: What contributions do languages
other than English make on the results of meta-analyses?
J Clin Epidemiol 53:964-972, 2000
24. Owens DK, Lohr KN, Atkins D, Treadwell JR, Reston JT,
Bass EB, Chang S, Helfand M: Grading the strength of
a body of evidence when comparing medical interven-
tions–Agency for Healthcare Research and Quality and the
Effective Health Care Program. J Clin Epidemiol 63:
513-523, 2010
25. Russell IJ, Larson AA: Neurophysiopathogenesis of fibro-
myalgia syndrome: A unified hypothesis. Rheum Dis Clin
North Am 35:421-435, 2009
26. Sackett DL, Straus SE, Richardson WS, Rosenberg W,
Haynes RB: Evidence-Based Medicine: How to Practice and
Teach EBM. JAMA 284:2382-2383, 2000
27. Sarzi-Puttini P, Torta R, Marinangeli F, Biasi G, Spath M,
Buskila D, Gracely RH, Giamberardino MA, Bazzichi L,
Cazzola M, Di Franco M, Stisi S, Salaffi F, Casale R,
Leardini G, Gorla R, Marsico A, Carignola R, Altomonte L,
Ceccherelli F, Cassisi G, Arioli G, Alciati A, Atzeni F, Italian
Fibromyalgia N: Fibromyalgia syndrome: The pharmacolog-
ical treatment options. Reumatismo 60(Suppl 1):50-58, 2008
28. White LA, Birnbaum HG, Kaltenboeck A, Tang J,
Mallett D, Robinson RL: Employees with fibromyalgia:
Medical comorbidity, healthcare costs, and work loss.
J Occup Environ Med 50:13-24, 2008
29. Wolfe F, Smythe HA, Yunus MB, Bennett RM,
Bombardier C, Goldenberg DL, Tugwell P, Campbell SM,
Abeles M, Clark P: The American College of Rheumatology
1990 Criteria for the Classification of Fibromyalgia. Report
of the Multicenter Criteria Committee. Arthritis Rheum 33:
160-172, 1990
30. Wolfe F, Ross K, Anderson J, Russell IJ, Hebert L: The
prevalence and characteristics of fibromyalgia in the
general population. Arthritis Rheum 38:19-28, 1995