Non‐antiepileptic drugs for trigeminal neuralgia

Abstract

Background

Trigeminal neuralgia was defined by the International Association for the Study of Pain as a sudden, usually unilateral, severe, brief, stabbing recurrent pain in the distribution of one or more branches of the fifth cranial nerve. Standard treatment is with anti‐epileptic drugs. Non‐antiepileptic drugs have been used in the management of trigeminal neuralgia since the 1970s. This is an update of a review first published in 2006 and previously updated in 2011.

Objectives

To systematically review the efficacy and tolerability of non‐antiepileptic drugs for trigeminal neuralgia.

Search methods

On 20 May 2013, for this updated review, we searched the Cochrane Neuromuscular Disease Group Specialized Register, CENTRAL (2013, Issue 4), MEDLINE (January 1966 to May 2013), EMBASE (January 1980 to May 2013), LILACS (January 1982 to May 2013) and the Chinese Biomedical Retrieval System (1978 to May 2013). We searched clinical trials registries for ongoing trials.

Selection criteria

We included double‐blind, randomised controlled trials in which the active drug was used either alone or in combination with other non‐antiepileptic drugs for at least two weeks.

Data collection and analysis

Two authors decided which trials fitted the inclusion criteria and independently graded risk of bias. We assessed the quality of the evidence according to the GRADE criteria for this update.

Main results

In this 2013 update, we updated the searches, but identified only two new ongoing studies. The review includes four trials involving 139 participants. The primary outcome measure in each was pain relief. Three trials compared one of the oral non‐antiepileptic drugs tizanidine, tocainide or pimozide with carbamazepine. The quality of evidence for all outcomes for which data were available was low. In a trial of tizanidine involving 12 participants (one dropped out due to unrelated disease), one of five participants treated with tizanidine and four of six treated with carbamazepine improved (risk ratio (RR) 0.30, 95% confidence interval (CI) 0.05 to 1.89). Few side effects were noted with tizanidine. For pimozide, there was evidence of greater efficacy than carbamazepine at six weeks. Up to 83% of participants reported adverse effects but these did not lead to withdrawal; the report did not provide comparable data for carbamazepine. Limited data meant that we could not assess the effects of tocainide; however, data from non‐randomised studies (not included in this review) indicate that serious haematological adverse events can occur. A trial involving 47 participants compared 0.5% proparacaine hydrochloride eyedrops with placebo but did not show any significant benefits, again according to low‐quality evidence. The report did not mention adverse events. The proparacaine trial was at low risk of bias; the other trials were at unclear risk of bias overall.

Authors' conclusions

There is low‐quality evidence that the effect of tizanidine is not significantly different than that of carbamazepine in treating trigeminal neuralgia. Pimozide is more effective than carbamazepine, although the evidence is of low quality and the data did not allow comparison of adverse event rates. There is also low‐quality evidence that 0.5% proparacaine hydrochloride eye drops have no benefit over placebo. Limitations in the data for tocainide prevent any conclusions being drawn. There is insufficient evidence from randomised controlled trials to show significant benefit from non‐antiepileptic drugs in trigeminal neuralgia. More research is needed.

Plain language summary

Drugs other than those used for epilepsy for treating trigeminal neuralgia

Review question

The purpose of this review was to assess the effects of non‐antiepileptic drugs in trigeminal neuralgia.

Background

Trigeminal neuralgia is a condition that affects the trigeminal nerve, the nerve which provides sensation in the skin of the face. The condition causes a sudden, severe, stabbing facial pain near the nose, lips, cheek, eye or ear. The incidence of trigeminal neuralgia is three to five new cases per 100,000 people each year. Standard treatment is with drugs that are used to treat epilepsy (antiepileptic drugs) but non‐antiepileptic drugs, such as baclofen and tocainide, have been used to treat trigeminal neuralgia since the 1970s.

Study characteristics

After a wide search for studies, we identified four trials involving 139 participants in total that met our criteria for inclusion in the review. Three randomised controlled trials compared the three different non‐antiepileptic drugs tizanidine, tocainide and pimozide with carbamazepine, which is the standard drug treatment. No new trials were identified for the update of this review in 2013.

Results and quality of the evidence

Tizanidine did not produce significantly more benefit than carbamazepine according to low‐quality evidence. The reporting of the tocainide trial did not allow us to assess whether the drug helped the pain of trigeminal neuralgia, but studies that were not part of this review suggest that this treatment can have serious harmful effects on the blood. Side effects of pimozide were very common but there was low‐quality evidence that it was more effective than carbamazepine. In a fourth trial there was low‐quality evidence that proparacaine hydrochloride eye drops did not show any significant benefit.

There is insufficient evidence from randomised controlled trials to show significant benefit from non‐antiepileptic drugs in trigeminal neuralgia. Further well‐designed randomised controlled trials are needed to establish whether non‐antiepileptic drugs are beneficial in trigeminal neuralgia.

The searches for the review are up to date to May 2013.

Summary of findings

Summary of findings for the main comparison. Tizanidine versus carbamazepine for trigeminal neuralgia.

Tizanidine versus carbamazepine for trigeminal neuralgia
Patient or population: patients with trigeminal neuralgia Settings: unclear, Norway Intervention: tizanidine versus carbamazepine
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Carbamazepine	Tizanidine
Immediate improvement in pain relief Number of participants improved (good/very good efficacy) on a self assessed categorical 'overall efficacy score' (0 = no efficacy to 4 = very good) Follow‐up: 21 days (end of treatment)	667 per 1000	200 per 1000 (33 to 1000)	RR 0.3 (0.05 to 1.89)	11 (1 study)	⊕⊕⊝⊝ low1,2	‐
50% pain reduction 12 weeks after the start of treatment Not measured	See comment	See comment	Not estimable	‐	See comment	Only trial of this intervention was of short (21 days) duration
Quality of life assessment 12 weeks after the start of treatment Not measured	See comment	See comment	Not estimable	‐	See comment	Only trial of this intervention was of short (21 days) duration
Adverse effects (not reported)	See comment	See comment	Not estimable	11 (1 study)	See comment	Adverse event data collected but no details provided. "Tizanidine was generally better tolerated than carbamazepine, and few side effects were noted"2
*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% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RR: risk ratio
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.

¹Quality of the evidence downgraded for imprecision. This trial had relatively few patients and thus the measure of effect has wide CIs.
 ²Further downgrading for risk of bias and potential indirectness. Three participants taking tizanidine discontinued taking the optimal dose: two because of lack of efficacy, one due to unrelated disease. Other potential risks of bias were from unclear randomisation, allocation concealment and blinding procedures. Investigators used a non‐validated outcome scale.

Summary of findings 2. Tocainide versus carbamazepine for trigeminal neuralgia.

Tocainide versus carbamazepine for trigeminal neuralgia
Patient or population: patients with trigeminal neuralgia Settings: unclear, Sweden Intervention: tocainide versus carbamazepine
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Carbamazepine	Tocainide
Pain relief (frequency and intensity) from the baseline 0 to 10‐point scale Follow‐up: mean 2 weeks	See comment	See comment	Not estimable	24 (1 study)	See comment	A carry‐over effect may exist and the report does not provide data from the first phase, so we are unable to analyse data for this outcome1,2
50% pain reduction 12 weeks after the start of treatment Not measured	See comment	See comment	Not estimable	‐	See comment	Only trial of this intervention was of short (2 weeks) duration
Quality of life assessment 12 weeks after the start of treatment Not measured	See comment	See comment	Not estimable	‐	See comment	Only trial of this intervention was of short (2 weeks) duration
Adverse effects	See comment	See comment	Not estimable	24 (1 study)	See comment	3 participants reported adverse effects of tocainide3
*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% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval
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.

¹Four participants in this trial could not tolerate being off the drug and so they had no baseline score.
 ²This is a cross‐over study and it may have carry‐over effects.
 ³Tocainide produced nausea, distal paraesthesias and skin rash in three participants, but the report provides no information about adverse effects of carbamazepine.

Summary of findings 3. Proparacaine hydrochloride 0.5% eye drops versus placebo for trigeminal neuralgia.

Proparacaine hydrochloride 0.5% eye drops versus placebo for trigeminal neuralgia
Patient or population: patients with trigeminal neuralgia Settings: inpatients, the United States Intervention: proparacaine hydrochloride 0.5% versus placebo
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Proparacaine hydrochloride 0.5%
Immediate improvement in pain relief Number of participants improved (≥ 2 level decrease in pain on 10‐point verbal rating scale, or pain gone or rare) Follow‐up: 30 days	227 per 1000	241 per 1000 (84 to 680)	RR 1.06 (0.37 to 2.99)	47 (1 study)	⊕⊕⊝⊝ low1,2	‐
50% pain reduction 12 weeks after the start of treatment Not measured	See comment	See comment	Not estimable	‐	See comment	Only trial of this intervention was of short (30 days) duration
Quality of life assessment 12 weeks after the start of treatment Not measured	See comment	See comment	Not estimable	‐	See comment	Only trial of this intervention was of short (30 days) duration
Adverse effects (not reported)	See comment	See comment	Not estimable	47 (1 study)	See comment	‐
*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% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RR: risk ratio
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.

¹Two participants in the control group required medication.
 ²This trial had relatively few patients and thus the RR has wide CIs.

Summary of findings 4. Pimozide versus carbamazepine for trigeminal neuralgia.

Pimozide versus carbamazepine for trigeminal neuralgia
Patient or population: patients with trigeminal neuralgia Settings: outpatients, Venezuela Intervention: pimozide versus carbamazepine
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Carbamazepine	Pimozide
Pain relief (TN score) Follow‐up: 8 weeks (outcome measured at 6 weeks)	The mean improvement in pain at 6 weeks in the carbamazepine group was 49.7%	The mean improvement in TN score at 6 weeks in the pimozide group was 28.7% greater (20.88% to 36.52% greater)	Not estimable	96 (1 study)	⊕⊕⊝⊝ low1,2	No definition is given for 'improvement' in this paper and we have not therefore presented the number of participants improved
50% pain reduction 12 weeks after the start of treatment Not measured	See comment	See comment	Not estimable	96 (1 study)	See comment	Cross‐over study, each phase 8 weeks
Quality of life assessment 12 weeks after the start of treatment Not measured	See comment	See comment	Not estimable	96 (1 study)	See comment	Cross‐over study, each phase 8 weeks
Adverse effects The number of participants with adverse effects Follow‐up: 24 weeks	See comment	See comment	Not estimable	96 (1 study)	See comment	Study report does not provide the number of participants with adverse events for carbamazepine
*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% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; RR: risk ratio
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.

¹This is a cross‐over study and it may have carry‐over effects. Results are from first period only. Risk of selective reporting bias present: first period TN score available at six weeks, when improvement greatest, but not at end of treatment (eight weeks) (other than in graphical format). Second period TN scores available in graphical format only. Insufficient information to assess selection bias and fully analyse blinding.
 ²No baseline characteristics provided.

Background

Trigeminal neuralgia was defined by the International Association for the Study of Pain as a sudden, usually unilateral, severe, brief, stabbing recurrent pain in the distribution of one or more branches of the fifth cranial nerve (Merskey 1994). The International Headache Society (Anonymous 2004) set the following criteria for trigeminal neuralgia.

A. Paroxysmal attacks of pain lasting from a fraction of a second to two minutes, affecting one or more divisions of the trigeminal nerve and fulfilling criteria B and C.
 B. Pain that has at least one of the following characteristics:

1. intense, sharp, superficial or stabbing;

2. precipitated from trigger areas or by trigger factors.

C. Attacks are stereotyped in the individual patient.
 D. There is no clinically evident neurological deficit.
 E. Not attributed to another disorder.
 
 The diagnosis of trigeminal neuralgia is based on clinical findings and there have not been any case controlled studies to validate diagnostic methods. A careful clinical history and examination can distinguish between trigeminal neuralgia and other disorders that cause facial pain (Zakrzewska 2002a).

Trigeminal neuralgia has an annual incidence of three to five per 100,000 people. It is more common in women than men (age‐adjusted ratio: 1.74:1) and is most widespread in people aged 50 to 69 years. Attacks are commonly on the right side (Katusic 1990; Yoshimasu 1972). More recent data from primary care practices suggest that the incidence may be around 12.7 per 100,000 (Koopman 2009).

The aetiology is not clear. Compression of the trigeminal nerve root, at or near the dorsal root entry zone, by a blood vessel is a major causative or contributing factor (Devor 2002). Compression is not always found and it is then necessary to look for other causes. In a few cases trigeminal neuralgia is due to multiple sclerosis. Other rare causes include infiltration of the nerve root, trigeminal ganglion or nerve by a tumour or amyloid, and small infarcts or angiomas in the pons or medulla (Cheng 1993). Once all of these possibilities have been excluded, there remain a small proportion of patients in whom the aetiology is undetermined (Nurmikko 2001). Obermann has shown that atypical type trigeminal neuralgia may be the result of central changes (Obermann 2007). The pain of idiopathic trigeminal neuralgia is indistinguishable from that caused by a demonstrable structural lesion other than vascular compression.

Spontaneous remission of trigeminal neuralgia is common but the disorder is often progressive. Remission may last for months or even years but as the attacks become more frequent, the patient may develop persistent pain between episodes. Attacks may come in clusters and can completely disrupt activities of daily living if left untreated (Katusic 1990; Zakrzewska 2002b).

Antiepileptic drugs have been used in pain management since the 1960s. The clinical impression is that they are useful for neuropathic pain, especially when the pain is lancinating or burning (Jacox 1994). Based on two placebo‐controlled studies of carbamazepine in trigeminal neuralgia, a Cochrane systematic review deduced a number needed to treat to benefit of 1.9 (95% confidence interval (CI) 1.4 to 2.8) (Wiffen 2011). In one randomised controlled trial (RCT), lamotrigine was effective as an add‐on therapy (Zakrzewska 1997). Little evidence supports a beneficial role of other antiepileptic drugs such as clonazepam, phenytoin and sodium valproate (Sindrup 2002). Gabapentin has been evaluated in a small RCT (Lemos 2008) and pregabalin in a cohort study (Obermann 2008). The long‐term effects of carbamazepine have been evaluated in only one study, which showed either loss of effect or poor tolerability in half the participants after 10 years (Taylor 1981). In a long‐term cohort study of oxcarbazepine, pain increased with progression of disease severity and the drug became less effective in long‐term pain control (Zakrzewska 2002b).

Non‐antiepileptic drugs have been used to treat trigeminal neuralgia since the 1970s. Conventional analgesics such as aspirin, codeine and non‐steroidal anti‐inflammatory drugs are not used since experience suggests that they are not helpful. There has been a reported RCT on the use of dextromethorphan, an antitussive (cough suppressant) drug with analgesic effects, in patients with mixed facial neuralgias (Gilron 2000). Other drugs that are used include baclofen, pimozide, racemic ketamine, proparacaine, tocainide, misoprostol and capsaicin. Baclofen is a gamma‐aminobutyric acid receptor agonist that has the least serious adverse effects (Steardo 1984). Pimozide is a neuroleptic that has shown significant relief of pain when compared with carbamazepine in a double‐blind, cross‐over trial. Racemic ketamine is an anaesthetic which showed some benefit when treating acute pain but was ineffective for pain lasting more than five years (Mathisen 1995). Topical ophthalmic anaesthetics, such as proparacaine, have been reported to relieve trigeminal neuralgia pain in some patients but not in a randomised trial (Kondziolka 1994). Tocainide is a drug used to prevent cardiac arrhythmias. It has been withdrawn from the market because of bone marrow suppression (Rawson 1998) but was tried in trigeminal neuralgia. Misoprostol is a long‐acting prostaglandin E1 analogue which, in an open‐label study, relieved trigeminal neuralgia in six of seven participants with definite multiple sclerosis who had failed to respond to conventional pharmacologic therapy (Reder 1995). Use of capsaicin, the pungent component of red pepper, has been reported in two case studies (Fusco 1992). There have been previous systematic reviews of all drugs for trigeminal neuralgia (Sindrup 1999; Zakrzewska 2009). The review in Clinical Evidence (Zakrzewska 2009) is updated every 18 months. International guidelines which used systematic review methodology have also been published (Cruccu 2008; Gronseth 2008). The aim of this review is to systematically review all double‐blind RCTs of non‐antiepileptic drugs for trigeminal neuralgia. Antiepileptic drugs for trigeminal neuralgia have been included in other Cochrane reviews (Moore 2009; Wiffen 2010a; Wiffen 2010b; Wiffen 2011).

Objectives

To systematically review the efficacy and tolerability of non‐antiepileptic drugs for trigeminal neuralgia.

Methods

Criteria for considering studies for this review

Types of studies

We searched for all double‐blinded, randomised controlled trials (RCTs) which lasted for at least two weeks and involved non‐antiepileptic drugs in the treatment of trigeminal neuralgia.

Types of participants

We included all causes of trigeminal neuralgia, both idiopathic and symptomatic, where the diagnostic criteria were clearly defined.

Types of interventions

We included all non‐antiepileptic drugs used to achieve analgesia. Non‐antiepileptic drugs include: baclofen, pimozide, racemic ketamine, proparacaine, dextromethorphan, tocainide or any other single non‐antiepileptic drug in any dose or combination. Comparison interventions were placebo, no treatment or carbamazepine.

The hypotheses below were to be tested.

1. Therapy A* is more effective in controlling trigeminal neuralgia than placebo treatment.

2. Therapy A* is more effective in controlling trigeminal neuralgia than no treatment.

3. Therapy A* is more effective in controlling trigeminal neuralgia than carbamazepine.

4. Therapy B** is more effective in controlling trigeminal neuralgia than placebo treatment.

5. Therapy B** is more effective in controlling trigeminal neuralgia than no treatment.

6. Therapy B** is more effective in controlling trigeminal neuralgia than therapy A.

7. Therapy B** is more effective in controlling trigeminal neuralgia than carbamazepine.

*Therapy A = baclofen, pimozide, racemic ketamine, proparacaine, tocainide or any other single non‐antiepileptic drug.
 **Therapy B = combination of any two drugs in therapy A.

Types of outcome measures

Primary outcomes

The primary outcome was immediate improvement in pain relief evaluated as decreased pain intensity or trigeminal neuralgia score, that is number of attacks per day and their intensity.

Secondary outcomes

Secondary outcomes were:

1. improvement in pain intensity or trigeminal neuralgia score at least 12 weeks after the start of treatment;

2. the number of participants with 50% pain reduction at least 12 weeks after the start of treatment;

3. quality of life measured with a validated scale, such as the Short Form 36 Health Survey questionnaire (SF‐36) (Ware 1992) at least 12 weeks after the start of treatment; and

4. adverse effects that were attributable to treatment by two weeks from the start of treatment.

Outcomes for inclusion in a 'Summary of findings' table

We included a 'Summary of findings' table to illustrate the findings of the review and our assessment of the quality of evidence for key outcomes:

• immediate improvement in pain relief;

• improvement in pain intensity or trigeminal neuralgia score at least 12 weeks after the start of treatment;

• the number of participants with 50% pain reduction at least 12 weeks after the start of treatment;

• quality of life measured with a validated scale, such as the Short Form 36 Health Survey questionnaire (SF‐36) (Ware 1992) at least 12 weeks after the start of treatment; and

• adverse effects.

We graded the quality of the evidence from included RCTs as high, moderate, low or very low based on the GRADE criteria. These start from a grading of high for RCTs and reasons for downgrading are: study limitations, consistency of effect, imprecision, indirectness and publication bias. Also, evidence from downgraded RCTs can be upgraded for a large effect size, when all plausible confounding would tend to underestimate the size of the effect or when there is a dose‐response gradient (GRADE working group 2004; Schünemann 2011a; Schünemann 2011b).

Search methods for identification of studies

Electronic searches

We searched the Cochrane Neuromuscular Disease Group Specialized Register (20 May 2013), CENTRAL (2013, Issue 4 in The Cochrane Library), MEDLINE (January 1966 to May 2013), EMBASE (January 1980 to May 2013), LILACS (January 1982 to May 2013), the Chinese Biomedical Retrieval System (1978 to May 2013), ClinicalTrials.gov (http://www.clinicaltrials.gov/) (May 2013) and the World Health Organization International Clinical Trials Registry Platform (ICTRP) (http://www.who.int/ictrp/en/). There were no restrictions on the basis of language. We reviewed the bibliographies of the randomised trials found and contacted the authors and known experts in the field to identify additional published or unpublished data. For the detailed search strategies see Appendix 1 (MEDLINE), Appendix 2 (EMBASE), Appendix 3 (CENTRAL), Appendix 4 (LILACS), Appendix 5 (Chinese Biomedical Retrieval System), Appendix 6 (ClinicalTrials.gov) and Appendix 7 (ICTRP).

Searching other resources

In this 2013 update, we searched the following 10 journals in electronic database of Chinese Biomedical Retrieval System rather than by handsearching:

• Chinese Journal of Neurology;

• Chinese Journal of Neuroscience;

• Journal of Apoplexy and Nervous Diseases;

• Journal of Brain and Nervous Diseases;

• Journal of Clinical Neurology;

• Chinese Journal of Nervous and Mental Diseases;

• Chinese Journal of Clinical Neurosciences;

• Chinese Journal of Pain Medicine;

• Chinese Journal of New Drugs and Clinical Remedies;

• Chinese Journal of Neuromedicine.

We also searched the retrieved articles for any other potentially missing studies and a personal reference database collected over 20 years (JZ).

Data collection and analysis

Selection of studies

Two review authors (JZ and NC) scrutinised the titles and abstracts identified from the updated searches. We obtained the full text of any study of possible relevance for independent assessment. The authors decided which trials fitted the inclusion criteria and graded their risk of bias. The authors resolved disagreements about inclusion by discussion. Two review authors (LH and JZ) arbitrated if there was a failure in resolving disagreement. If necessary and if possible we contacted the trial authors for clarification.

Data extraction and management

We collected information from each study about diagnostic criteria, comparability of groups at baseline, pain condition and number of patients studied, non‐antiepileptic drug type and dose regimen, study design, study duration and follow‐up, length of washout, analgesic outcome measures and results, withdrawals and adverse effects (minor and major).

Two review authors (JZ and MY) independently extracted and checked data on participants, methods, interventions, outcomes and results using a data extraction form. We obtained missing data from the authors whenever possible. One author (MY) entered data into the Cochrane software Review Manager, currently RevMan 5.2 (RevMan 2012). A second author (MZ) checked the data entry and a third author (JZ) did another independent assessment.

Assessment of risk of bias in included studies

We assessed the risk of bias of each trial. The assessment took into account the security of randomisation, allocation concealment, blinding of participants and personnel and blinding of outcome assessors, completeness of outcome data, selective reporting and any other potential bias, as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Two review authors assessed these items independently and subdivided trials.

Measures of treatment effect

We calculated a treatment effect with the Cochrane statistical package RevMan. We expressed results as risk ratios (RRs) or risk differences (RDs) with 95% confidence intervals (CIs) for dichotomous outcomes and mean differences (MDs) with 95% CIs for continuous outcomes. According to the method of assessing risk of bias in cross‐over trials (Higgins 2011), we evaluated the following items for cross‐over trials: (i) whether the cross‐over design was suitable; (ii) whether there was a carry‐over effect; (iii) whether the first period data were available; (iv) incorrect analysis; and (v) comparability of results with those from parallel‐group trials. We planned to analyse all outcomes under consideration.

Dealing with missing data

We analysed data on an intention‐to‐treat basis. Therefore, we included all participants with available data in the analysis in the group to which they were allocated, regardless of whether or not they received the allocated intervention. If in the original reports participants were not analysed in the group to which they were randomised, and there was sufficient information in the trial report, we attempted to restore them to the correct group. We obtained missing data from the study authors whenever possible.

Assessment of reporting biases

As there were insufficient studies, we did not use a funnel plot to investigate the possibility of publication bias but used a descriptive analysis to evaluate the possible reporting bias.

Data synthesis

As the non‐antiepileptic and control group interventions tested in the included studies were different, we did not use RevMan to perform meta‐analyses of the studies. Instead, we described the results of each trial as reported by the authors.

Results

Description of studies

Results of the search

In this 2013 update, we performed electronic searches and retrieved 71 references from the Cochrane Neuromuscular Disease Group Specialized Register, 185 from CENTRAL, 1571 from MEDLINE, 314 from EMBASE, 17 from LILACS, 156 from the Chinese Biomedical Retrieval System, 54 from Clinicaltrials.gov and 24 from WHO ICTRP. We have identified two new ongoing studies since the original review in 2005. In total 29 reports potentially met our inclusion criteria.

We included four trials with 139 participants (Kondziolka 1994; Lechin 1989; Lindström 1987; Vilming 1986) (see Characteristics of included studies). We excluded 23 studies and there were two ongoing studies.

Included studies

The review includes four trials. Three studies compared an oral non‐antiepileptic drug with carbamazepine; the drugs used were tizanidine (Vilming 1986), tocainide (Lindström 1987), and pimozide (Lechin 1989). The duration of treatments was generally short: two weeks (Lindström 1987), 21 days (Vilming 1986) and eight weeks per treatment period in Lechin 1989. The trials took place in Norway (Vilming 1986), Sweden (Lindström 1987), and Venezuela (Lechin 1989).

The fourth study, Kondziolka 1994 (47 participants), was conducted in the US and compared a single application of 0.5% proparacaine hydrochloride eye drops with buffered saline placebo.

Two studies had a parallel‐group design (Kondziolka 1994; Vilming 1986), the other two were cross‐over trials.

Outcome measures in Vilming 1986 (12 participants) were VAS pain and overall efficacy of medication (no efficacy, slight, moderate, good, very good) daily. Lindström 1987 (12 participants) used a zero to 10‐point scale to measure the frequency and severity (duration and intensity) of attacks; Lechin 1989 (68 participants) used a TN symptom score (1 to 100); and Kondziolka 1994 reported a zero to 10‐point verbal pain rating scale and a descriptive scale (unchanged, moderately better, markedly better, or worse) to grade outcome.

None of the study reports stated the investigators' conflicts of interest and only Lindström 1987 and Lechin 1989 provided information on funding, which came from an institute or foundation.

Excluded studies

We excluded 23 studies (Carasso 1979; Fromm 1984; Fromm 1987; Fromm 1993; Gazerani 2009; Gilron 2000; Jiang 2007; Jiang 2008; Kanai 2006a; Kanai 2006b; Kanai 2006c; Laurinda 2008; Liu 1999; Luo 1995; Parekh 1989; Shao 1999; Shao 2011; Silver 2007; Stajcic 1990; Wang 2007; Xu 1998; Yan 2011; Zhao 2011) as: they were not RCTs (Kanai 2006b; Liu 1999; Luo 1995; Xu 1998; Shao 2011; Yan 2011; Zhao 2011), not double‐blind (Carasso 1979; Wang 2007), were of a combination of treatments rather than a single non‐antiepileptic drug (Laurinda 2008; Shao 1999; Silver 2007), were repeated (Jiang 2007), the participants enrolled did not correspond with the protocol (Jiang 2008) or their characteristics were not described (Gilron 2000), the duration of treatment was less than two weeks (Fromm 1984; Fromm 1987; Fromm 1993; Parekh 1989), loss to follow‐up was more than 20% (Stajcic 1990) or the study objective did not correspond to the protocol (Kanai 2006a; Kanai 2006c). (See Characteristics of excluded studies).

Risk of bias in included studies

The 'Risk of bias' summary (Figure 1) shows the review authors' judgements about each 'Risk of bias' item for the included studies. One study had an overall low risk of bias (Kondziolka 1994). For the other three the risk of bias was generally unclear and in particular there was not enough information to assess the adequacy of randomisation and blinding.

1.

'Risk of bias' summary: review authors' judgements about each methodological quality item for each included study. Red = high risk of bias; yellow = unclear risk of bias; green = low risk of bias

Allocation

Four trials (Kondziolka 1994; Lechin 1989; Lindström 1987; Vilming 1986) stated that they were randomised. Only one trial (Kondziolka 1994) described an adequate method of randomisation; the participants were randomly assigned by the tossing of a coin. This trial also compared the characteristics of the two groups, which showed that the randomisation had been effective. The report states that "The code was not broken until the study was completed". The other three trials did not describe their method of 'random allocation' or allocation concealment.

Blinding

Most trial reports described the method of blinding only in broad terms such as 'double‐blind', which made it impossible to know exactly who was blinded (Higgins 2011). All four trials (Kondziolka 1994; Lechin 1989; Lindström 1987; Vilming 1986) claimed that they were double‐blind; one trial (Kondziolka 1994) described this in detail. Lechin 1989 stated that trial and placebo medications were identical.

Incomplete outcome data

Two trials reported that there were dropouts or withdrawals (Lechin 1989; Vilming 1986) and they reported the reason but did not use intention‐to‐treat analysis to analyse the results.

Selective reporting

We considered the risk of selective reporting bias introduced by the trial investigators to be high in two trials (Lechin 1989; Vilming 1986). In Vilming 1986, the trial authors did not report visual analogue scale (VAS) measurements, one of the prespecified outcomes, in a suitable way. Lechin 1989 reported data primarily in graphs. The first period data for reduction in trigeminal neuralgia score were reported in the text after six weeks of treatment, which the trial authors describe as the point of "maximal improvement" for both pimozide and carbamazepine. This outcome is therefore at high risk of bias owing to selective reporting. We considered the risk of reporting bias in Lindström 1987 and Kondziolka 1994 to be low.

Other potential sources of bias

Only one trial (Kondziolka 1994) claimed that the treatment groups were the same with regard to gender, age, duration of disease and pain scores. The others did not compare the baseline characteristics of the groups in detail. Many of the trials lacked clear diagnostic criteria and referred to typical trigeminal neuralgia but did not define it and they did not report whether participants had idiopathic trigeminal neuralgia. One cross‐over trial (Lindström 1987) did not report a washout period, so there is a high risk of a carry‐over effect in this study. Vilming 1986 did not report whether there was a washout period prior to the start of the trial.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

As the non‐antiepileptic and control group interventions tested in the included studies were different, we did not combine the results. Instead, we describe the results of each trial as reported by the authors.

Three trials reported the proportion of participants with a good initial effect evaluated by pain intensity or trigeminal neuralgia scores two weeks after treatment.

Tizanidine versus carbamazepine

This comparison was reported in Vilming 1986, a very small parallel‐group trial (12 participants, with six in each group).

The duration of treatment was 21 days. Three participants taking tizanidine discontinued taking the optimal dose, one because of unrelated disease and two as a result of intolerable pain. The report provides results for 11 participants.

Primary outcome: immediate improvement in pain relief

A visual analogue scale (VAS) was used to evaluate pain but the report does not include the results. The investigators reported overall efficacy assessed by participants and investigators on a four‐point scale (no efficacy, slight, moderate, good or very good) but provide no other details of the measure. At the end of treatment, one of five participants improved with tizanidine and four of six with carbamazepine (RR 0.30, 95% CI 0.05 to 1.89) (Analysis 1.1). The difference was not statistically significant.

1.1. Analysis.

Comparison 1 Tizanidine versus carbamazepine, Outcome 1 Number improved (reported by participant on global efficacy scale).

Secondary outcomes

The duration of treatment was 21 days, which was too short for the measurement of our secondary efficacy outcomes at 12 weeks.

Side effects were not specifically reported, other than, "Tizanidine was generally better tolerated than carbamazepine, and few side effects were noted. Laboratory tests did not show any deviations that could be related to the test medication."

Tocainide versus carbamazepine

One double‐blind, cross‐over RCT (Lindström 1987) in 12 participants compared tocainide with carbamazepine.

The duration of each phase was two weeks. However, the authors did not report the washout period or data from the end of the first arm in this trial. Since there is a high risk of carry‐over effects, we have not reported the quantitative data from this trial.

For adverse effects, the authors reported that tocainide produced nausea, distal paraesthesias and skin rash which led to withdrawal from the study, but non‐randomised studies showed an unacceptable rate of serious haematological side effects and deaths (see Discussion). The Lindstrom study did not report whether the adverse effects of tocainide were dose‐dependent.

Proparacaine hydrochloride 0.5% versus placebo

In a parallel‐group RCT (Kondziolka 1994), investigators assigned 47 participants to receive a single dose of proparacaine hydrochloride 0.5% or placebo eye drops. Each participant was treated only once.

Primary outcome: immediate improvement in pain relief

A verbal pain rating scale (VRS) was used to assess the effect. There was no significant reduction of pain after 3, 10 or 30 days. At 30 days, six of 25 participants improved (i.e. had a 2 or more level decrease in pain on 10‐point scale, or pain was gone or rare) with proparacaine hydrochloride 0.5%, and five of 22 improved with placebo (RR 1.06, 95% CI 0.37 to 2.99) (Analysis 2.1). Eleven participants in the treatment group and 14 participants in the control group underwent surgery during follow‐up. Two participants in the control group required medication.

2.1. Analysis.

Comparison 2 Proparacaine hydrochloride 0.5% versus placebo, Outcome 1 Number improved (reported by participant on verbal pain rating scale).

Secondary outcomes

The study duration was too short to report secondary efficacy outcomes at 12 weeks. For adverse events, the authors state "We observed no morbidity related to placement of either proparacaine or saline eyedrops."

Pimozide versus carbamazepine

One double‐blind, cross‐over RCT (Lechin 1989) compared pimozide and carbamazepine in 68 people but nine were not admitted to the treatment phase. Eleven participants did not contribute data to the statistical analysis: 10 because of protocol deviations and one dropped out. The trial lasted 24 weeks: four weeks of placebo washout, eight weeks of treatment following random assignment to carbamazepine or pimozide, four weeks of abrupt withdrawal with placebo substitution and eight weeks of cross‐over treatment.

Primary outcome: immediate improvement in pain relief

A trigeminal neuralgia score ranging from 0 to 100 was used to assess the effect. However, the authors did not state how they defined 'improvement', other than stating that all of the 48 participants on pimozide and 27 out of 48 participants on carbamazepine "showed improvement". They do give a percentage improvement score for each group after six weeks' treatment (when the authors state that there was "maximal improvement"); they also report that "similar differences were obtained" at weeks seven and eight, week eight being the end of the first treatment period. Week seven and eight results are provided in graph form in the study report as are second period results. In the first period, the reduction in total trigeminal neuralgia score was 78.4% with pimozide and 49.7% with carbamazepine, six weeks after starting treatment (MD 28.70%, 95% CI 20.88% to 36.52%) (see Analysis 3.1).

3.1. Analysis.

Comparison 3 Pimozide versus carbamazepine, Outcome 1 Improvement (reduction in total TN score).

Secondary outcomes

The study duration was too short (eight‐week treatment phases) to report secondary efficacy outcomes at 12 weeks.

Adverse effects were recorded in 40 (83.3%) participants with pimozide. These were physical and mental retardation, hand tremors, memory impairment, involuntary movements during sleep and slight Parkinson's disease manifestations. In the carbamazepine period, 22 serious adverse effects occurred (one abnormal full blood count, one abnormal liver function, 18 lethargy, one inappropriate secretion of vasopressin and one erythematous exanthem that disappeared after the trial. However, the report does not provide the total number of participants with adverse effects in each group. Since some participants might have had more than one adverse effect with carbamazepine, we could not calculate the incidence of adverse effects between the two drugs. For pimozide, the adverse effects could be attenuated by reducing the dose, so may be dose‐dependent.

No studies were alike, therefore we did not perform a meta‐analysis.

Discussion

We did not identify any new randomised controlled trials (RCTs) for this update. In a previous update, we re‐evaluated the nine trials in the original review in the light of changed inclusion criteria. Since most drugs used in trigeminal neuralgia need to be escalated slowly in dose, and a steady state reached at the maximum dose, we considered it important that a trial should report the results from at least two weeks at the maximum steady state dose. Trigeminal neuralgia is a chronic condition and drugs need to provide pain relief for long periods. Our inclusion criteria therefore stated that the trial had last at least two weeks using the maximum steady dose. After reassessment of the trials, four clinical trials investigating non‐antiepileptic drugs in the treatment of trigeminal neuralgia, with a total of 139 participants, met the inclusion criteria. The drugs tested were tizanidine, tocainide, proparacaine hydrochloride (0.5% eye drops) and pimozide. We were unable to pool the results because of heterogeneity between the studies. The results of these trials can be summarised as follows.

1. There was no significant difference in effect between carbamazepine and tizanidine measured using the selected outcomes. Larger and better studies are required to draw any conclusions (Vilming 1986) (Table 1).

2. The efficacy of tocainide was not analysed because of the high risk of a carry‐over effect and no separate first phase data were presented that would have enabled us to analyse these separately (Table 2). In non‐randomised studies, tocainide produced serious haematological side effects limiting its use (Rawson 1998).

3. Proparacaine hydrochloride 0.5% instilled into the eyes was simple and safe but no significant effect was observed when compared with placebo (Kondziolka 1994) (Table 3). In this trial three participants had first division (V1) pain only and a further four had V1 as well as other division involvement. The rest of the participants did not have V1 involvement, therefore they were looking for a systemic effect of the drug. The authors did not report the results separately for the V1 participants but presumably if the drug had worked better in that group they would have reported it. We therefore assume that the trial showed that the drug neither works centrally nor peripherally.

4. The effect of pimozide was evaluated in one double‐blind, cross‐over RCT (Table 4). Adverse effects were common and the authors suggested it should not be chosen as a first‐line treatment option (Lechin 1989).

These conclusions are limited by methodological considerations such as the method of randomisation, quality of double‐blinding, allocation concealment, study duration and selection of outcome measures. All of the included trials were either inadequately reported or had methodological flaws. The quality problems were as follows:

1. the method of randomisation and allocation concealment, diagnostic criteria and comparability of the baseline groups were not reported;

2. the washout period was not reported in a cross‐over trial;

3. the blinding methods were unclear;

4. dropouts were high and not reported in detail and intention‐to‐treat analysis was not carried out;

5. the follow‐up time was short;

6. the selection of assessment tools was not consistent and adverse effects were not reported in some trials;

7. the participants recruited in some trials were not typical and did not meet the diagnostic criteria for trigeminal neuralgia.

At present we cannot draw any definitive conclusions about the effect and reliability of any non‐antiepileptic drug in trigeminal neuralgia.

All these trials were carried out before more rigorous methods were introduced for the conduct of RCTs. No study included a power calculation and the primary and secondary outcome measures were not clearly stated. One of the studies (Kondziolka 1994) had a low risk of bias; the other three had an unclear risk of bias.The Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) has provided some guidelines on the use of outcome measures which need to be considered when designing future trials in this field (Dworkin 2005). Trials in patients with trigeminal neuralgia are difficult to conduct due to the lack of knowledge about the natural history of the condition. Patients can go into a natural period of complete pain remission, as is reported to have potentially occurred in one participant in the tizanidine study (Vilming 1986). The following outcomes need to be measured: the frequency of attacks of pain, the duration of each attack and the severity of attack. Evoked or spontaneous pain could also be measured as there is a suggestion that this can be altered differently (Beydoun 2002). A 50% pain reduction is considered a good outcome in all chronic pain studies of other conditions but a higher target is possible in patients with trigeminal neuralgia. Quality of life assessment is important because patients may be unable to eat or socialise and if rendered pain‐free with drugs they may be unable to walk due to ataxia, diplopia and drowsiness. In the current update of the review we revised our outcomes to include 50% pain reduction and quality of life assessment.

The results of our review are consistent with the previous studies described in the Background section.

Authors' conclusions

Implications for practice.

There is low‐quality evidence that the effect of tizanidine is not significantly different than that of carbamazepine in treating trigeminal neuralgia. Pimozide is more effective than carbamazepine, although the evidence is low‐quality and the data did not allow comparison of adverse event rates. There is also low‐quality evidence that 0.5% proparacaine hydrochloride eye drops have no benefit over placebo. Limitations in the data for tocainide prevent any conclusions being drawn.

Implications for research.

More research is needed to determine whether non‐antiepileptic drugs are beneficial in trigeminal neuralgia. Future trials should describe their methods adequately and use clear inclusion criteria, allocation concealment, randomisation, double‐blind assessment, intention‐to‐treat analysis and long‐term follow‐up. The length of washout period before interventions, and between periods in cross‐over trials, must be adequate. Outcome measures should include the frequency, duration and severity of attacks of pain, and quality of life. Adverse events and the number of and explanation for dropouts must be reported. Trials should be adequately powered.

What's new

Date	Event	Description
17 April 2013	New citation required but conclusions have not changed	Searches updated. A new author has joined the review update team: Jingjing Zhang
10 March 2013	New search has been performed	We added new references and updated the searches but identified no new trials. We revised some 'Risk of bias' assessments and removed some data from the results upon re‐examination of trial reports.

History

Protocol first published: Issue 1, 2003
 Review first published: Issue 3, 2006

Date	Event	Description
7 December 2010	New citation required but conclusions have not changed	Additional authors: Mi Yang, Ning Chen and Joanna M Zakrzewska. Bo Wu withdrew.
30 April 2010	New search has been performed	We updated the database searches in April 2010. No new randomised controlled trial was identified. We re‐evaluated the nine trials included, and five were excluded for reasons listed in the text. The overall review conclusions remain unchanged in the update.
2 July 2008	Amended	Converted to new review format.
24 April 2006	New citation required and conclusions have changed	Substantive amendment.

Appendices

Appendix 1. MEDLINE (Ovid SP) search strategy

Database: Ovid MEDLINE(R) <1946 to May Week 2 2013>
 Search strategy:
 ‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐
 1 randomized controlled trial.pt. (351508)
 2 controlled clinical trial.pt. (86242)
 3 randomized.ab. (253157)
 4 placebo.ab. (138957)
 5 drug therapy.fs. (1619317)
 6 randomly.ab. (181024)
 7 trial.ab. (261494)
 8 groups.ab. (1175082)
 9 or/1‐8 (3026869)
 10 exp animals/ not humans.sh. (3847813)
 11 9 not 10 (2571885)
 12 exp Trigeminal Nerve/ (13968)
 13 Facial Neuralgia/ (1113)
 14 exp Trigeminal Nerve Diseases/ (7128)
 15 (trigeminal or trigeminus).tw. (18945)
 16 tic douloureux.mp. (250)
 17 (cranial or face or facial or oro?facial or tri?facial or cranio?facial or maxillo?facial).tw. (231226)
 18 or/12‐17 (252495)
 19 neuralgia$.mp. (15849)
 20 18 and 19 (7506)
 21 trigeminal neuralgia/ (5383)
 22 20 or 21 (7506)
 23 11 and 22 (1591)
 24 remove duplicates from 23 (1571)
 25 randomized controlled trial.pt. (351508)
 26 controlled clinical trial.pt. (86242)
 27 randomized.ab. (253157)
 28 placebo.ab. (138957)
 29 drug therapy.fs. (1619317)
 30 randomly.ab. (181024)
 31 trial.ab. (261494)
 32 groups.ab. (1175082)
 33 or/25‐32 (3026869)
 34 exp animals/ not humans.sh. (3847813)
 35 33 not 34 (2571885)
 36 exp Trigeminal Nerve/ (13968)
 37 Facial Neuralgia/ (1113)
 38 exp Trigeminal Nerve Diseases/ (7128)
 39 (trigeminal or trigeminus).tw. (18945)
 40 tic douloureux.mp. (250)
 41 (cranial or face or facial or oro?facial or tri?facial or cranio?facial or maxillo?facial).tw. (231226)
 42 or/36‐41 (252495)
 43 neuralgia$.mp. (15849)
 44 42 and 43 (7506)
 45 trigeminal neuralgia/ (5383)
 46 44 or 45 (7506)
 47 35 and 46 (1591)
 48 remove duplicates from 47 (1571)
 49 randomized controlled trial.pt. (351508)
 50 controlled clinical trial.pt. (86242)
 51 randomized.ab. (253157)
 52 placebo.ab. (138957)
 53 drug therapy.fs. (1619317)
 54 randomly.ab. (181024)
 55 trial.ab. (261494)
 56 groups.ab. (1175082)
 57 or/49‐56 (3026869)
 58 exp animals/ not humans.sh. (3847813)
 59 57 not 58 (2571885)
 60 exp Trigeminal Nerve/ (13968)
 61 Facial Neuralgia/ (1113)
 62 exp Trigeminal Nerve Diseases/ (7128)
 63 (trigeminal or trigeminus).tw. (18945)
 64 tic douloureux.mp. (250)
 65 (cranial or face or facial or oro?facial or tri?facial or cranio?facial or maxillo?facial).tw. (231226)
 66 or/60‐65 (252495)
 67 neuralgia$.mp. (15849)
 68 66 and 67 (7506)
 69 trigeminal neuralgia/ (5383)
 70 68 or 69 (7506)
 71 59 and 70 (1591)
 72 remove duplicates from 71 (1571)

Appendix 2. EMBASE (Ovid SP) search strategy

Database: EMBASE <1980 to 2013 Week 20>
 Search strategy:
 ‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐
 1 crossover‐procedure.sh. (36896)
 2 double‐blind procedure.sh. (114620)
 3 single‐blind procedure.sh. (17374)
 4 randomized controlled trial.sh. (342470)
 5 (random$ or crossover$ or cross over$ or placebo$ or (doubl$ adj blind$) or allocat$).tw,ot. (953986)
 6 trial.ti. (144904)
 7 or/1‐6 (1086574)
 8 (animal/ or nonhuman/ or animal experiment/) and human/ (1260257)
 9 animal/ or nonanimal/ or animal experiment/ (3386195)
 10 9 not 8 (2801568)
 11 7 not 10 (996670)
 12 limit 11 to embase (780448)
 13 trigeminal nerve/ (9260)
 14 facial nerve/ (9155)
 15 trigeminal nerve disease/ (672)
 16 (trigeminal or trigeminus).tw. (22066)
 17 (cranial or face or facial or oro?facial or tri?facial or cranio?facial or maxillo?facial).tw. (293660)
 18 (tic douloureux or tic doloureux or tic doloreux or tic douloreux).mp. (325)
 19 or/13‐18 (314032)
 20 neuralgia$.mp. (20949)
 21 19 and 20 (7928)
 22 trigeminus neuralgia/ (8062)
 23 21 or 22 (9729)
 24 12 and 23 (314)
 25 remove duplicates from 24 (314)

Appendix 3. Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library

#1 MeSH descriptor Trigeminal Neuralgia, this term only
 #2 MeSH descriptor Trigeminal Nerve explode all trees
 #3 MeSH descriptor Facial Nerve, this term only
 #4 MeSH descriptor Trigeminal Nerve Diseases explode all trees
 #5 trigeminal OR trigeminus
 #6 "tic douloureux" OR "tic doloureux" OR "tic doloreux" OR "tic douloreux"
 #7 cranial OR face OR facial OR orofacial OR trifacial OR craniofacial OR maxillofacial
 #8 (#2 OR #3 OR #4 OR #5 OR #6 OR #7)
 #9 neuralgia*
 #10 (#8 AND #9)
 #11 (#1 OR #10)

Appendix 4. LILACS search strategy

(((MH: A08.800.800.120.760$ or "facial neuralgia" or "neuralgia facial" or MH:C10.292.775$ or trigeminal or trigemino or trigemeo or trigeminus or "tic douloureux" or cranial or craneal or craniana or face or cara or facial or faciales or faciais or orofacial or orofaciales or orofaciais or craniofacial or cranio‐facial or craniofaciales or craniofaciais or cranio‐faciales or cranio‐faciais or maxillofacial or maxilofacial or maxillofaciales or maxilofaciales or maxillofaciais or maxilofaciais or maxillo‐faciales or maxilo‐facialis or maxillo‐faciais or maxilo‐faciais) and neuralgia) or "trigeminal neuralgia" or "neuralgia del trigemino" or "neuralgia do trigemeo") and ((PT:"Randomized Controlled Trial" or "Randomized Controlled trial" or "Ensayo Clínico Controlado Aleatorio" or "Ensaio Clínico Controlado Aleatório" or PT:"Controlled Clinical Trial" or "Ensayo Clínico Controlado" or "Ensaio Clínico Controlado" or "Random allocation" or "Distribución Aleatoria" or "Distribuição Aleatória" or randon$ or Randomized or randomly or "double blind" or "duplo‐cego" or "duplo‐cego" or "single blind" or "simples‐cego" or "simples cego" or placebo$ or trial or groups) AND NOT (B01.050$ AND NOT (humans or humanos or humanos)))

Appendix 5. Chinese Biomedical Retrieval System search strategy

1 suijiduizhaoshiyan

2 linchuanduizhaoshiyan

3 suijifenpei

4 mangfa

5 danmang

6 sanmang

7 *mang

8 zhutici="suijiduizhaoshiyan/quanbufuzhitici"

9 zhutici="lichuangshiyan, 1qi/quanbufuzhutici"

10 zhutici="lichuangshiyan, 2qi/quanbufuzhutici"

11 zhutici="lichuangshiyan, 3qi/quanbufuzhutici"

12 zhutici="lichuangshiyan, 4qi/quanbufuzhutici"

13 zhutici="lichuangshiyan/quanbufuzhutici/quanbushu"

14 duizhao

15 kongbaiduizhao

16 anweiji

17 1 or 2 or 3 or 4 or 5 or 6 or 7 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16

18 zhutici="sanchashenjingtong/quanbufuzhutici"

19 zhutici="sanchashenjingjibing/quanbufuzhutici/quanbushu"

20 sanchashenjing*

21 mianbutengtong

22 zhutici="mianbutengtong/quanbufuzhutici/quanbushu"

23 mian*tong

24 mian*teng

25 toumian*tong

26 toumian*teng

27 sancha*teng

28 sancha*tong

29 sancha*jibing

30 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29

31 17 and 30

Appendix 6. Clinicaltrials.gov

randomized controlled trial OR controlled clinical trial OR randomized OR randomized OR placebo OR drug therapy OR randomly OR trial AND double‐blind | Trigeminal Nerve OR TN OR Facial Neuralgia OR Trigeminal Nerve Diseases OR (trigeminal or trigeminus) OR douloureux OR trigeminal neuralgia

Appendix 7. WHO international Clinical trials Registry

Trigeminal Nerve OR TN OR Facial Neuralgia OR Trigeminal Nerve Diseases OR (trigeminal or trigeminus) OR douloureux OR trigeminal neuralgia

Data and analyses

Comparison 1. Tizanidine versus carbamazepine.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number improved (reported by participant on global efficacy scale)	1	11	Risk Ratio (M‐H, Fixed, 95% CI)	0.3 [0.05, 1.89]

Comparison 2. Proparacaine hydrochloride 0.5% versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number improved (reported by participant on verbal pain rating scale)	1	47	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.37, 2.99]

Comparison 3. Pimozide versus carbamazepine.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Improvement (reduction in total TN score)	1	48	Mean Difference (IV, Fixed, 95% CI)	28.70 [20.88, 36.52]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Kondziolka 1994.

Methods	Double‐blind, randomised, parallel‐group study Single application of drugs with 30 days of follow‐up. Baseline time of observation unclear
Participants	47 participants (25 treatment group and 22 placebo) Not reported whether participants had idiopathic TN Mean age 59 years in proparacaine group, 63 in control group 70% female. Duration of illness from 0.2 to 25 years. Baseline pain scores unreported Diagnosis clear. Inclusion criteria described. Only 12 participants (5 active, 7 placebo) were not taking any systemic medications for their TN Exclusion criteria were: surgery intended within 3 days, atypical TN, hyperthyroidism, pregnancy or with a known allergic reaction to any local anaesthetic agent
Interventions	Treatment group: 2 drops of 0.5% proparacaine hydrochloride were instilled onto the ipsilateral cornea once Control group: 2 drops of buffered saline were instilled onto the ipsilateral cornea once
Outcomes	Frequency of pain: participants kept a log of frequency of pain attacks and were telephoned at 3, 10 and 30 days. The change in severity of pain was assessed in comparison with the previous conversation Severity of pain using both a rating scale and descriptive indices: verbal pain rating (participant rating of the severity of their current pain from 0 to 10 based on their experience with TN, with 10 being the worst pain and 0 being no pain) a descriptive scale (unchanged, moderately better, markedly better or worse) to grade outcome Endpoint criteria included clinical response (defined as improved, worsened or unchanged pain in comparison to pain level before instillation) and the need for an increase in oral medication or for surgical intervention during the observation period
Notes	Funding and conflicts of interest: not reported in the paper
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly assigned to a group using a coin toss
Allocation concealment (selection bias)	Low risk	Participants could not foresee assignments
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Instilled by individual blinded to agent; participants could not distinguish the agents
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Instilled by individual blinded to agent; participants could not distinguish the agents
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts. No missing outcome data
Selective reporting (reporting bias)	Low risk	Outcomes listed in the methods section were all reported
Other bias	Unclear risk	The report does not state whether participants had idiopathic TN

Lechin 1989.

Methods	Randomised, double‐blind (methods not described), cross‐over study Trial duration 8 weeks with unclear follow‐up. Baseline time of observation 4 weeks
Participants	68 participants with idiopathic TN joined initially; 59 were randomly distributed into 2 groups; 48 participants completed Diagnosis of TN clear Mean age 59.3 years, 50% females. Duration of illness 8 to 17 years. Baseline pain scores unclear. Placebo responders, patients with severe physical illness, a history of psychotic episodes, drug or alcohol addiction or mental retardation, and pregnant and nursing women were excluded
Interventions	4 weeks placebo, 8 weeks carbamazepine or pimozide. 4 weeks of abrupt withdrawal by placebo substitution, and 8 weeks of cross‐over treatment Pimozide: 4 mg/day from days 1 to 4, 6 mg/day from days 5 to 9, 8 mg/day from days 10 to 14, 12 mg/day from day 14 to the end Carbamazepine: 300 mg/day from days 1 to 4, 600 mg/day from days 5 to 9, 900 mg/day from days 10 to 14, 1200 mg/day from day 14 to the end
Outcomes	Mean % changes in total TN scores from baseline at visits 5 to 12 and 17 to 24 (i.e. in each treatment phase). TN symptom score (1 to 100) was based on weekly interviews with 2 physicians. The score was the mean of the 2 assessments Adverse events
Notes	Funding: grant from the Foundation of the Institute of Experimental Medicine. Conflicts of interest: not reported in the paper All participants continued taking pimozide after the controlled trial; none of the participants showed tolerance to pimozide
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of sequence generation was not described
Allocation concealment (selection bias)	Unclear risk	Method of allocation concealment was not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "All medications were administered in identical dark capsules under double‐blind conditions"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Stated to be "double‐blind"; insufficient information to permit judgement of high or low risk of bias
Incomplete outcome data (attrition bias) All outcomes	Low risk	11 dropouts. No ITT analysis. Details of missing data and reasons were reported in each group
Selective reporting (reporting bias)	High risk	This trial reported data primarily in graphs. The first period data for reduction in TN score were reported in the text after 6 weeks of treatment, which the trial authors describe as the point of "maximal improvement" for both pimozide and carbamazepine
Other bias	Unclear risk	No baseline characteristics; carry‐over effect might exist

Lindström 1987.

Methods	Randomised, double‐blind (methods not described), cross‐over study
Participants	12 participants. Age 41 to 78 years All participants had idiopathic TN 58% females. Duration of illness 5 to 19 years All participants were on carbamazepine but none experienced complete relief. 8 were able to be without medication to record the pain before the trial. Baseline scores of pain 4.0 to 10. Only participants whose condition was active, with several pain attacks over a long period of time, were selected. Cardiovascular disease, liver and renal insufficiency were exclusion criteria
Interventions	Tocainide and carbamazepine were given for 2 weeks each. No details provided of how cross‐over was done and it is impossible to look at each arm separately Tocainide 20 mg/kg divided into 3 daily doses Carbamazepine was used at the maximum tolerated dose
Outcomes	0 to 10‐point scale summarising frequency and severity (duration and intensity) of attacks. Mean score values before treatment and for the last 10 days of each 2‐week period with carbamazepine and tocainide A telephone interview twice weekly "to supplement the evaluation procedure and to obtain information concerning the patients' activity pattern, pain precipitation factors and possible side effects"
Notes	Funders: Folksam Research Foundation and the Vivian L Smith Foundation for Restorative Neurology Conflicts of interest: not reported in the paper
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of sequence generation was not described
Allocation concealment (selection bias)	Unclear risk	Method of allocation concealment was not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Stated to be "double‐blind"; insufficient information to permit judgement of high or low risk of bias
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Stated to be "double‐blind"; insufficient information to permit judgement of high or low risk of bias
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts. No missing outcome data
Selective reporting (reporting bias)	Low risk	Outcomes listed in the methods section were all reported
Other bias	High risk	No complete baseline characteristics; possibility of carry‐over effect

Vilming 1986.

Methods	Randomised, double‐blind, parallel‐group study (methods not described) 3‐week trial but baseline time of observation unclear
Participants	12 participants (6 tizanidine group and 6 carbamazepine) Whether participants had idiopathic TN was not reported Mean age 60 years. Proportion of females was unclear Duration of illness and baseline frequency of pain not reported No diagnostic criteria for TN. Typical TN participants were included No exclusion criteria
Interventions	The medication was gradually increased during the first 12 days, and participants received optimal medication for the following 9 days: 18 mg tizanidine or 900 mg carbamazepine daily in 3 divided doses
Outcomes	Frequency, duration, severity of attacks Pain relief "Inability to participate in daily activities" VAS pain recorded daily Spontaneously reported side effects recorded 4 times during test period Overall efficacy (no efficacy, slight, moderate, good, very good) assessed by investigator and the participant at the end of the study
Notes	Funding and conflicts of interest not provided in report 3 participants taking tizanidine discontinued taking the optimal dose, 1 because of the unrelated disease, 2 due to intolerable pain. Side effects not specifically reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of sequence generation was not described
Allocation concealment (selection bias)	Unclear risk	Method of allocation concealment was not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Stated to be "double‐blind"; insufficient information to permit judgement of high or low risk of bias
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Stated to be "double‐blind"; insufficient information to permit judgement of high or low risk of bias
Incomplete outcome data (attrition bias) All outcomes	Low risk	Details of missing data and reasons were reported in each group. Results for 11 of the 12 participants reported. ITT analysis was not used to analyse the results
Selective reporting (reporting bias)	High risk	Side effects not specifically reported and the VAS results were not reported in a suitable format for interpretation
Other bias	Unclear risk	The study report did not describe baseline characteristics or whether participants had idiopathic or symptomatic TN

AED: antiepileptic drug
 CBZ: carbamazepine
 ITT: intention‐to‐treat
 TN: trigeminal neuralgia
 VAS: visual analogue scale

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Carasso 1979	Single‐blind only
Fromm 1984	No washout period. The first phase of this cross‐over study was too short (only 7 days)
Fromm 1987	No washout period. The first phase of this cross‐over study was too short (only 7 days)
Fromm 1993	No washout period. The first phase of this cross‐over study was too short (only 4 days)
Gazerani 2009	The participants enrolled did not fulfil the protocol criteria
Gilron 2000	Cross‐over, double‐blind, randomised trial. 19 participants with facial neuralgia and 3 TN participants. 1 TN participant dropped out, which means only 2 TN participants were treated. We do not know how the 2 TN participants were allocated, and details of their characteristics were unavailable
Jiang 2007	Repeated report
Jiang 2008	The participants enrolled did not fulfil the protocol criteria
Kanai 2006a	The participants enrolled did not fulfil the protocol criteria and the study was not of a single non‐antiepileptic drug
Kanai 2006b	Not a randomised study
Kanai 2006c	The participants enrolled did not fulfil the protocol criteria
Laurinda 2008	The participants enrolled did not fulfil the protocol criteria
Liu 1999	Participants were divided into groups by age, course and clinical situation. Allocation was not randomised
Luo 1995	Retrospective, not prospective study
Parekh 1989	The trial lasted only 10 days. No baseline; not enough time to wash out the previous drug, such as carbamazepine; no details of how drugs were escalated; no side effects reported
Shao 1999	The study was not of a single non‐antiepileptic drug
Shao 2011	Not a RCT
Silver 2007	The study was not of a single non‐antiepileptic drug
Stajcic 1990	RCT; 5 weekly peripheral streptomycin/lidocaine versus lidocaine alone injections in TN patients. However, the dropouts were more than 20% of all patients (35%, 6/17)
Wang 2007	Not double‐blind
Xu 1998	Not a RCT
Yan 2011	Not a RCT
Zhao 2011	Not a RCT

CBZ: carbamazepine
 RCT: randomised controlled trial
 TN: trigeminal neuralgia

Characteristics of ongoing studies [ordered by study ID]

NCT00004431.

Trial name or title	Randomized study of L‐baclofen in patients with refractory trigeminal neuralgia
Methods	A randomised, double‐blind, placebo‐controlled, parallel study
Participants	Ages eligible for study: 18 years to 80 years, both genders Disease characteristics: idiopathic TN, i.e. paroxysmal attacks of facial or frontal pain lasting a few seconds to less than 2 minutes. Pain has at least 4 of the following characteristics: distribution along one or more divisions of the trigeminal nerve; sudden, intense, sharp, superficial, stabbing or burning quality; pain intensity severe; precipitation from trigger areas or by certain daily activities such as eating, talking, washing the face or cleaning the teeth; between paroxysms entirely asymptomatic; no neurological deficit; attacks are stereotyped in the individual patient; exclusion of other causes of facial pain by history, physical examination and special investigations when necessary; must be uncontrolled or refractory as defined by occurrence of at least 2 paroxysms of pain per day despite conventional treatment with maximally tolerated doses of carbamazepine Prior or concurrent therapy: no concurrent medication for TN other than carbamazepine, phenytoin and Neurontin (gabapentin) Concurrent medication for other conditions allowed, on stable dose(s)
Interventions	Drug: L‐baclofen
Outcomes	None known
Starting date	June 1998
Contact information	Pittsburgh, Pennsylvania, United States, 15261. Contact: Michael J Soso 412‐648‐1239
Notes	‐

NCT01540630.

Trial name or title	A Phase IIa withdrawal study of CNV1014802 in patients with trigeminal neuralgia
Methods	A phase IIa placebo‐controlled, double‐blind, randomised withdrawal study
Participants	Ages eligible for study: 18 years to 70 years, both genders Inclusion criteria: • Male or female aged between 18 and 70 years, with a diagnosis of TN; IHS criteria to be used • Female patients must be of non‐child bearing potential or agree to use an approved form of contraception • Male patients must agree to use an approved form of contraception • Body weight > 50 kg for men and > 45 kg for women • BMI ≤ 34.9 • Capable of giving written informed consent. Informed consent must be obtained prior to the commencement of any study‐related procedures • QTcB either/or QTcF < 450 ms in 2 of 3 ECGs conducted at screening • AST and ALT < 2 upper limit of normal; alkaline phosphatase and bilirubin < 1.5 upper limit of normal • Approved concomitant medications must have been stable for at least 3 weeks prior to day 0
Interventions	Drug: CNV1014802 Control: placebo
Outcomes	Primary outcome measures: The number of failures on CNV1014802 versus number of failures on placebo during the double‐blind treatment period (time frame: 4 weeks) Criteria for treatment failure: 50% increase in the frequency of paroxysms compared to the final 7 days of the open‐label period 50% increase in the severity of pain experienced in the paroxysms compared to the final 7 days of the open‐label period A Patient Global Improvement of Change [sic] rating of much worse/very much worse Discontinuation due to "lack of efficacy" Discontinuation due to an adverse reaction related to study medication or poor tolerability Secondary outcome measures: Secondary pain endpoints (time frame 4 weeks) Number and severity of paroxysms of pain in the 21‐day open‐label period, both evoked and spontaneous
Starting date	March 2012
Contact information	Contact: Kevin Gunn, MD, kevin.gunn@convergencepharma.com
Notes	‐

ALT: alanine aminotransferase
 AST: aspartate aminotransferase
 BMI: body mass index
 ECG: electrocardiogram
 IHS: International Headache Society
 TN: trigeminal neuralgia

Differences between protocol and review

In the 2011 update, we modified the original protocol. We edited the methods section Assessment of risk of bias in included studies according to Higgins 2008 and completed a 'Risk of bias' assessment and table for each of the included studies. In Types of participants we included all causes of trigeminal neuralgia in which the diagnostic criteria were clearly defined. We also specified that the trials needed to have lasted at least two weeks in total.

In this 2013 update, we excluded quasi‐RCTs; the protocol stated that we would include quasi‐RCTs (He 2006) (we found no quasi‐RCTs). According to the latest version of Cochrane Handbook for Systematic Reviews of Interventions, only RCTs can prevent systematic differences in general characteristics between different intervention groups. We searched the 10 journals that we previously handsearched in the electronic Chinese Biomedical Retrieval System database. We also updated the 'blinding' assessments, and split blinding into blinding of participants and personnel (performance bias) and blinding of outcome assessors (assessment bias) (Higgins 2011). In addition, we added a 'Summary of findings' table in this update.

We revised our outcomes in this update to include 50% pain reduction and quality of life measures.

Changes to the team of authors have taken place. See Other published versions of this review.

Contributions of authors

JZ: wrote the first draft of this (2013) update.

LH: wrote the first draft of the review and instructed the other authors.

MY: wrote the 2011 update of this review.

NC, MZ and JZ commented on each draft.

Sources of support

Internal sources

• Chinese Cochrane Center, China.

• Joanna Zakrzewska undertook this work at University College London/University College London Hospitals Trust, which received a proportion of funding from the Department of Health's NIHR Biomedical Research Centre funding scheme, UK.

External sources

• Chinese Medical Board of New York (CMB), USA.

Declarations of interest

None known.

New search for studies and content updated (no change to conclusions)