Truong 2005.

Methods	Randomised, double‐blind, parallel design Randomisation: Block‐wise randomisation using a software‐generated code, stratification by centre Setting: multicentre (16 centres in USA) Duration: 4 weeks, follow‐up up to 20 weeks
Participants	80 participants enrolled (BtA group = 37; placebo group = 43) % Female: BtA: 62%; placebo: 63% Mean age, SD: BtA: 53.4 years, 11.6; placebo: 53.6 years, 12.1 Mean CD duration, SD: BtA: 7.1 years, 7.1; placebo: 5.7 years, 5.2 Mean CD severity, SD (TWSTRS total): BtA: 45.1, 8.7; placebo: 46.2, 9.4 Inclusion criteria: ≥ 18 months since cervical dystonia diagnosis TWSTRS total score of ≥ 30 de novo or previously treated with Bt ≥ 16 weeks prior to enrolment Exclusion criteria: suspected secondary non‐responsiveness prior CD surgery or phenol injections participants believed to require a Botox dose < 80 U or > 250 U pure retrocollis forms Medications such as muscle relaxants and benzodiazepines were required to be on a stable dose for ≥ 6 weeks
Interventions	BtA: Dysport (abobotulinumtoxinA); 500 U Placebo: 0.125 mg of human serum albumin and 2.5 mg of lactose Study drug preparation: BtA provided in vials by Ipsen Ltd Muscles injected: the doses and number of injection sites per muscle were determined at the discretion of the investigator. EMG guidance: left at discretion of the investigator BtA dose per participant: 500 U
Outcomes	Primary outcome: TWSTRS total and TWSTRS subscales at week 4 Secondary outcomes: TWSTRS total and TWSTRS subscales at weeks 8 and 12 Participant assessment of pain using a VAS (range, 0‐100; 0 mm: least possible pain, 100 mm: worst possible pain) Investigator assessment of change using a VAS (range, 0‐100; 0 mm: much worse, 50 mm: no change, 100 mm: symptom‐free) Participant assessment of change using a VAS (range, 0‐100; 0 mm: much worse, 50 mm: no change, 100 mm: symptom‐free) Adverse events Plasma neutralising antibodies
Notes	Participants who showed no benefit at week 4 were terminated from the study. Those who had evidence of response at week 4 continued in the study until additional injections were needed. Study discontinuations (at week 4), reasons: BtA: n = 15 (41%), reasons not described Placebo: n = 27 (63%), reasons not described
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomization was in blocks of four and was stratified by center and according to whether or not the patient had been treated previously with botulinum toxin" Quote: "All patients were randomly assigned to treatment using a randomization code generated before the study"
Allocation concealment (selection bias)	Low risk	Quote: "Dysport was provided in a clear glass vial as a freeze dried white pellet (…). Placebo was provided in identical clear glass vials (…)"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Placebo was provided in identical clear glass vials (…). Study medication was supplied in individual patient boxes, containing one vial of either Dysport or placebo. Subjects, investigators, medical staff, (…) data managers and monitors were blind to subjects' treatment group"
Blinding of outcome assessment (detection bias) Objective outcomes	Unclear risk	Quote: "Whenever possible, an investigator or research nurse other than the one performing the TWSTRS assessment who was blind to treatment condition performed the assessment for adverse events. All sites were asked to achieve as much consistency as possible with respect to assessors" Comment: insufficient information to permit judgement of low risk or high risk
Blinding of outcome assessment (detection bias) Subjective outcomes	Unclear risk	Quote: "At each post‐treatment visit, patients and investigators independently assessed the change from baseline" Comment: insufficient information to permit judgement of low risk or high risk. Although placebo was identical to intervention, the fact that most of the participants had previously been treated with Botox could have led to a degree of bias
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: post‐randomisation exclusions at week 4 were high in both intervention arms, though this difference was asymmetrical, with more dropouts happening in the placebo arm
Selective reporting (reporting bias)	Low risk	Comment: the outcomes mentioned in the study protocol matched the outcomes reported in the study
For‐profit bias	High risk	Comment: study funded by Beauford Ipsen
Enriched population – preferential enrolment of positive responders	High risk	Comment: out of the 80 participants enrolled, 21 were de novo
Enriched population – exclusion of poor responders	High risk	Quote: "Patients with pure retrocollis were not permitted to participate"