Lechin 1989.

Methods	Double‐blind placebo‐controlled cross‐over trial Four‐centre study
Participants	68 patients joined the study, final number 48 outpatients, 24 men and 24 women, duration of illness 8 to 17 years Inclusion criteria: severe facial pain for at least 2 years, clinical diagnosis of trigeminal neuralgia Exclusion criteria: placebo responder (improvement of more than 20% during placebo washout period), severe physical illness, history of psychotic episodes, alcohol or drug addiction, epilepsy or any other convulsive disorder
Interventions	4 weeks of placebo washout, 8 weeks of treatment following random assignment to carbamazepine (final dose of 1200 mg with a 14‐day titration period) or pimozide (final dose of 12 mg with a 14‐day titration period), 4 weeks with abrupt withdrawal and placebo substitution, 8 weeks of cross‐over treatment
Outcomes	Pain intensity using 6‐point registration cards (0 = no pain, 6 = pain present, cannot be ignored, prompt medical advice sought), duration, frequency, basal pain, sensitivity of trigger zones, number of relief tablets Total trigeminal neuralgia score reduction of 78.1 % (pimozide group) compared to 49.7% (carbamazepine group), statistically significant (P<0.001) at week 10 Adverse effects (hand tremors, involuntary movements during sleep, slight Parkinson's symptoms) observed in 40 of 48 patients
Notes	Type of pain: trigeminal neuralgia; 9 patients were not admitted to the treatment phase, 6 patients were excluded. 11 were not included in the statistical analysis. All patients refused interruption of pimozide treatment.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Described as "identical dark capsules"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Described as "identical dark capsules"
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Imputation method not described
Size	High risk	Fewer than 50 patients/treatment arm