Types of studies

We included studies if they were randomised controlled trials (RCTs) with at least 10 participants per treatment group and double‐blind (participant and observers) assessment of participant‐reported outcomes, following two weeks of treatment or longer, although the emphasis of the review is on studies of eight weeks or longer. Full journal publication was required, with the exception of extended abstracts of otherwise unpublished clinical trials (for example detailed information from PDFs of posters that typically include all important details of methodology used and results obtained). We did not include short abstracts (usually meeting reports with inadequate or no reporting of data). We excluded studies of experimental pain, case reports, and clinical observations.

Migraine and headache studies previously included in an earlier version of this review were excluded (Wiffen 2000). This subject is being dealt with in greater depth by the Cochrane Pain, Palliative Care and Supportive Care Review Group.

Types of participants

We included adult participants aged 18 years and above. Participants could have one or more of a wide range of chronic neuropathic pain conditions including (but not limited to):

• painful diabetic neuropathy (PDN);

• postherpetic neuralgia (PHN);

• trigeminal neuralgia;

• phantom limb pain;

• postoperative or traumatic neuropathic pain;

• complex regional pain syndrome (CRPS) Type II;

• cancer‐related neuropathy;

• HIV‐neuropathy;

• spinal cord injury;

or

• fibromyalgia;

• complex regional pain syndrome (CRPS) Type I.

We also included studies of participants with more than one type of neuropathic pain. We analysed results according to the primary condition.

Types of interventions

Carbamazepine in any dose, by any route, administered for the relief of neuropathic pain or fibromyalgia, and compared to placebo, no intervention or any other active comparator. We did not include studies using carbamazepine to treat pain resulting from the use of other drugs.

Types of outcome measures

We anticipated that studies would use a variety of outcome measures, with the majority of studies using standard subjective scales (numerical rating scale (NRS) or visual analogue scale (VAS)) for pain intensity or pain relief, or both. We were particularly interested in Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) definitions for moderate and substantial benefit in chronic pain studies (Dworkin 2008). These are defined as at least 30% pain relief over baseline (moderate), at least 50% pain relief over baseline (substantial), much or very much improved on Patient Global Impression of Change (PGIC) (moderate), and very much improved on PGIC (substantial). These outcomes concentrate on dichotomous outcomes where pain responses do not follow a normal (Gaussian) distribution. People with chronic pain desire high levels of pain relief, ideally more than 50%, and with pain intensity not worse than mild (O'Brien 2010).

We include a Summary of findings table as set out in the Cochrane Pain, Palliative and Supportive Care Group author guide (AUREF 2012). The Summary of findings table includes outcomes of at least 30% and at least 50% pain intensity reduction, PGIC, at least one adverse event, adverse event withdrawals, serious adverse events and death.

Electronic searches

For the earlier review, RCTs of antiepileptics in acute, chronic or cancer pain were identified by searching MEDLINE (originally via Silver Platter, then Ovid) from 1966 to June 2010, EMBASE 1994 to Dec 2009, SIGLE 1980 to July 1999, and CENTRAL (Issue 4, 2010).

When the review was split in to individual drugs, this search strategy was narrowed to include carbamazepine only. Appendix 3 has the search strategies for CENTRAL, MEDLINE and EMBASE.

For this update we searched:

• Cochrane Central Register of Controlled Trials (CENTRAL, 2014 Issue 1 in The Cochrane Library);

• MEDLINE (via Ovid) (January 2010 to 7 February 2014);

• EMBASE (via Ovid) (January 2010 to 7 February 2014);

• PhRMA clinical study results database (clinicaltrials.gov) to 7 February 2014;

• WHO International Clinical Trials Registry Platform (apps.who.int/trialsearch) to 7 February 2014

Searching other resources

Additional studies were identified from the reference list of the retrieved papers. In the first version, a letter was sent to the first author of a report for further information on their published report (method of randomisation, double blinding, outcome measures and dropouts) and to ask if they knew of any other studies which met our inclusion criteria, either undertaken by them or by other investigators. In addition, 41 medical journals were hand searched, chosen from the 50 with the highest number of reports in MEDLINE, and nine specialist journals which were either not on that list or were not indexed (Jadad 1994). The search process included volumes published between 1950 and 1990. No further hand searching has been undertaken as the key journals are now being searched by the Cochrane Collaboration.

In an earlier version of this review data were requested from 19 authors but only one (Leijon 1989) was able to supply information relevant to this review. In the two updates, no further attempt was made to contact authors.

Selection of studies

Two review authors independently read the titles and abstracts of all studies identified by the search, and the full text of all potentially relevant studies. Agreement on eligibility was reached by discussion. We did not anonymise the studies in any way before assessment.

Data extraction and management

Two review authors extracted data using a standard data extraction form, and agreed data before entry into RevMan (RevMan 2012) or any other analysis method. Data extracted included information about the pain condition and number of participants treated, drug and dosing regimen, study design, study duration and follow up, analgesic outcome measures and results, withdrawals and adverse events (participants experiencing any adverse event, or a serious adverse event).

Assessment of risk of bias in included studies

We independently scored each study for quality using a three‐item scale (Jadad 1996) and agreed a 'consensus' score for each study. Scores of two and below have been associated with greater estimates of efficacy than studies of higher quality (Khan 1996). Quality scores were not used to weight the results in any way.

We used the 'Risk of bias' tool to assess the likely impact on the strength of the evidence of various study characteristics relating to methodological quality (randomisation, allocation concealment, blinding, freedom from selective reporting), study validity (duration, outcome reporting, and handling of missing data), and size (Moore 2010a).

Two review authors independently assessed risk of bias for each study, using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and adapted from those used by the Cochrane Pregnancy and Childbirth Group, with any disagreements resolved by discussion. We assessed the following for each study.

• Random sequence generation (checking for possible selection bias). We assessed the method used to generate the allocation sequence as: low risk of bias (any truly random process, e.g. random number table; computer random number generator); unclear risk of bias (method used to generate sequence not clearly stated). We excluded studies using a non‐random process (e.g. odd or even date of birth; hospital or clinic record number).

• Allocation concealment (checking for possible selection bias). The method used to conceal allocation to interventions before assignment determines whether intervention allocation could have been foreseen in advance of or during recruitment, or changed after assignment. We assessed the methods as: low risk of bias (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes); unclear risk of bias (method not clearly stated). We excluded studies that did not conceal allocation (e.g. open list).

• Blinding of outcome assessment (checking for possible detection bias). We assessed the methods used to blind study participants and outcome assessors from knowledge of which intervention a participant received. We assessed the methods as: low risk of bias (study states that it was blinded and describes the method used to achieve blinding, e.g. identical tablets; matched in appearance and smell); unclear risk of bias (study states that it was blinded but does not provide an adequate description of how this was achieved). We excluded studies that were not double‐blind.

• Incomplete outcome data (checking for possible attrition bias due to the amount, nature, and handling of incomplete outcome data). We assessed the methods used to deal with incomplete data as: low risk of bias (< 10% of participants did not complete the study and/or used ‘baseline observation carried forward’ analysis); unclear risk of bias (used 'last observation carried forward' analysis); high risk of bias (used 'completer' analysis).

• Size of study (checking for possible biases confounded by small size). We assessed studies as being at low risk of bias (≥ 200 participants per treatment arm); unclear risk of bias (50 to 199 participants per treatment arm); high risk of bias (< 50 participants per treatment arm).

Measures of treatment effect

Relative risk (or risk ratio, RR) was used to establish statistical difference. Numbers needed to treat (NNT) and pooled percentages were used as absolute measures of benefit or harm.

The following terms are used to describe adverse outcomes in terms of harm or prevention of harm:

• When significantly fewer adverse outcomes occurred with carbamazepine than with control (placebo or active) we use the term the number needed to treat to prevent one event (NNTp).

• When significantly more adverse outcomes occurred with carbamazepine compared with control (placebo or active) we use the term the number needed to harm or cause one event (NNH).

Unit of analysis issues

The control treatment arm would be split between active treatment arms in a single study if the active treatment arms were not combined for analysis.

Dealing with missing data

We planned to use intention‐to‐treat (ITT) analysis wherever possible. The ITT population consisted of participants who were randomised, took the assigned study medication and provided at least one post‐baseline assessment. Missing participants were assigned zero improvement (baseline observation carried forward, BOCF) where this could be done. We were aware that imputation methods might be problematical and examined trial reports for information about them.

Assessment of heterogeneity

We planned to deal with clinical heterogeneity by combining studies that examined similar conditions. We assessed statistical heterogeneity visually (L'Abbe 1987) and with the use of the I² statistic.

Assessment of reporting biases

The aim of this review was to use dichotomous data of known utility (Moore 2009b). The review did not depend on what authors of the original studies chose to report or not report, though clearly there were difficulties with studies failing to report any dichotomous results. Continuous data, which probably poorly reflect efficacy and utility, were extracted and used only when useful for illustrative purposes.

We undertook no statistical assessment of publication bias.

We looked for evidence of possible enrichment, either complete or partial, in enrolment of participants into the studies. Enrichment typically means including participants known to respond to a therapy, and excluding those known not to respond, or to suffer unacceptable adverse effects, though for gabapentin no significant effects have been shown from partial enrichment (Straube 2008). Enriched enrolment randomised withdrawal studies, known to produce higher estimates of efficacy, would not be pooled (McQuay 2008).

Data synthesis

We analysed data for each painful condition in three tiers, according to outcome and freedom from known sources of bias.

• The first tier uses data meeting current best standards, where studies report the outcome of at least 50% pain intensity reduction over baseline (or its equivalent), without the use of last observation carried forward (LOCF) or other imputation method for dropouts, report an intention‐to‐treat (ITT) analysis, last eight or more weeks, have a parallel‐group design, and have at least 200 participants (preferably at least 400) in the comparison (Moore 2010a; Moore 2012b). These top‐tier results are reported first.

• The second tier uses data from at least 200 participants but where one or more of the above conditions is not met (for example reporting at least 30% pain intensity reduction, using LOCF or a completer analysis, or lasting four to eight weeks).

• The third tier of evidence relates to data from fewer than 200 participants, or where there are expected to be significant problems because, for example, of very short duration studies of less than four weeks, where there is major heterogeneity between studies, or where there are shortcomings in allocation concealment, attrition, or incomplete outcome data. For this third tier of evidence, no data synthesis is reasonable, and may be misleading, but an indication of beneficial effects might be possible

Subgroup analysis and investigation of heterogeneity

We planned for all analyses to be according to individual painful conditions, because placebo response rates with the same outcome can vary between conditions, as can the drug‐specific effects (Moore 2009a). We also planned subgroup analysis for dose of carbamazepine and duration of study.

Sensitivity analysis

We planned no sensitivity analyses because the evidence base was known to be too small to allow reliable analysis.