Supplementary Table 4.Risk of bias in the included studies

Study / Random sequence generation (selection bias) / Allocation concealment (selection bias) / Blinding of participants and personnel (performance bias) / Blinding of outcome assessment (detection bias) / Incomplete outcome data (attrition bias) / Selective reporting (reporting bias) / Other bias
Dumais, 2012 [25] / Unclear. Randomization method not described. / Low. Opaque sealed envelopes were used. / High. Participants and personnel not blinded. “It was not possible for the participants and administrators of treatment to be masked to group allocation.” / Low. Research assistants responsible for administrating measurement instruments were masked to group allocation. / Low. Under 20% attrition in each group; attrition similar in study arms. / Low. Trial was registered on Clinicaltrials.gov (NCT01206634) and all of the pre-specified outcomes (primary and secondary) have been reported in the pre-specified way. / High. Less than 50 participants per treatment arm.
Hashemi, 2015 [26] / Unclear. Randomization method not described. / Unclear. Not described. / Unclear. Not described. / Unclear. Not described. / Unclear. Not described. / Low. Trial registration not described. The study protocol is not available; all outcomes indicated in Methods were reported in Results. / High. Less than 50 participants per treatment arm.
Hashemi, 2012 [27] / Low. Random number table was used. / Unclear. Not described. / Low. Double-blind study. / Unclear. Not described. / Unclear. Attrition not described. / Low. Trial registration not described. The study protocol is not available; all outcomes indicated in Methods were reported in Results. / Unclear. 50 patients in each group.
Jahangiri, 2014 [21] / Low. Computer-generated randomization sequence. / Low. Sequentially numbered sealed envelopes were used. / Low. Participants and personnel were blinded. / Low. The clinical assessor was blinded to the baseline evaluations and to the administered treatments. / Low. Under 20% attrition in each group; attrition similar in study arms. / Low. This trial was registered at the Iranian Registry of Clinical Trials (IRCT) website a WHO Primary Register set-up, with registration code: Irct ID: IRCT201011025088N1. All of the pre-specified outcomes (primary and secondary) have been reported in the pre-specified way. / High. Less than 50 participants per treatment arm.
Rahimzadeh, 2014 [20] / High. Robust (pseudo) random number generation software was used / Unclear. Not described. / Low. Double-blind study. / Low. Outcome assessor was blinded. / Unclear. Attrition not fully described. / Low. This trial was registered at the Iranian Registry of Clinical Trials (IRCT2013092210336N4). All of the pre-specified outcomes have been reported in the pre-specified way. / High. Less than 50 participants per treatment arm. Follow-up limited to 12 weeks.
Reeves, 2000 [23] (knee) / Low. Randomized using random number table. / Low. The group assignment was kept in a database blinded to the chief investigator and research coordinator. / Low. Participants and personnel were blinded. / Low. Outcome assessor was blinded. / Unclear. Total attrition was 11%, but not fully described per study arms. / Low. Trial registration not described. The study protocol is not available; all outcomes indicated in Methods were reported in Results. / High. Less than 50 participants per treatment arm.
Reeves, 2000 [24] (thumb and finger) / Low. Randomized using random number table. / Low. Group assignments were blinded to the chief investigator and research coordinator by using a password-protected access to the assignment database. / Low. Participants and personnel were blinded. / Low. Outcome assessor was blinded. / High. 31% attrition in dextrose group, 21% attrition in control group; imbalance between groups in attrition / Low. Trial registration not described. The study protocol is not available; all outcomes indicated in Methods were reported in Results. / High. Less than 50 participants per treatment arm.

1