Greene 1990.

Methods	Randomised, double‐blind, parallel design Randomisation: stratified by CD classification; method not described Setting: single‐centre (USA) Duration: 12 weeks
Participants	55 participants enrolled (BtA group = 28; placebo group = 27) % Female: BtA: 61%; placebo: 67% Mean age: BtA: 46.8 years; placebo: 52.6 years Mean CD duration: BtA: 6.6 years; placebo: 9.8 years CD severity: BtA: 7% mild, 71% moderate, 21% severe; placebo: 11% mild, 48% moderate, 41% severe Inclusion criteria: Idiopathic CD non‐responder to at least 2 drug trials including at least 1 trial of anticholinergics Exclusion criteria: Known or suspected cause for CD prior thalamotomy or peripheral surgery previous treatment with Bt
Interventions	BtA: Botox (onabotulinumtoxinA); diluted in saline solution to a concentration of 25 U per 1 mL Placebo: saline solution Study drug preparation: BtA provided in vials by Smith‐Kettlewell Eye Research Institute (USA) Muscles injected: the doses, muscles, and number of injected sites per muscle were determined by the physician based on clinical assessment and classification of CD EMG guidance: no BtA dose per participant: 150 U, rotational torticollis and torticollis plus retrocollis; 165 U, head tilt
Outcomes	Primary outcomes: Patient Subjective Assessment of Change ‐ 3 scales: Res Scale (results of injection: marked, moderate, slight improvement, no change, slight and definitely worse); Cap Scale (functional capability; 0%: completely disable, 100%: fully functional); Pain scale (0%: no difference, 100%: complete relieve) Secondary outcomes: Columbia Torticollis Rating Scale (objective video records rating) Time course of benefit Adverse events
Notes	Study discontinuations, reasons: BtA: n = 3 (11%), adverse events: n = 1, unrelated reasons: n = 2 Placebo: n = 0
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "They were divided into 3 cells (A, pure rotational torticollis; B, torticollis plus retrocollis; and C, head tilt with or without torticollis and retrocollis). In order to ensure reasonable balance of Botox and placebo injections in each cell, randomization was stratified by cell type, which was completed for blocks of 4 sequentially enrolled patients in each cell type" Comment: insufficient information about the method of randomisation to permit judgement of low or high risk
Allocation concealment (selection bias)	Unclear risk	Comment: method of concealment not specified
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The blinded physicians then injected them with Botox or saline, using syringes filled by the unblinded physicians according to the protocol"
Blinding of outcome assessment (detection bias) Objective outcomes	Low risk	Quote: "Two blinded physicians gave the injections, determined the degree of head turning and disability, and videotaped the patients; but they did not examine the strength or size of the neck muscles, so that the presence of muscle atrophy would not identify patients receiving active injection. Videotapes of each patient visit were rated by the 2 blinded observers independently"
Blinding of outcome assessment (detection bias) Subjective outcomes	Low risk	Quote: "Patients previously treated with Botox were excluded from the trial"
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Comment: the study authors stated that some data were lost, accounting for up to 13% of total data, and it is unclear whether this had an impact on the overall results.
Selective reporting (reporting bias)	Low risk	Comment: the expected outcomes that are usually evaluated in intervention trials for this condition were reported in this study.
For‐profit bias	Low risk	Study drug was provided by Dr. A. Scott, from Smith‐Kettlewell Eye Research Institute (USA).
Enriched population – preferential enrolment of positive responders	Low risk	Comment: participants who had previously received Botox injections were excluded.
Enriched population – exclusion of poor responders	Low risk	Comment: Exclusion criteria did not include forms of dystonia known to have poorer response to treatment