Lopez‐D'alessandro 2011.

Methods	Single centre, placebo‐controlled parallel RCT
Participants	120 BMS patients Mean age 14.1 SD 57.5 years, median: 57 Sex: 94 F:26 M (F 78%:M 22%) Inclusion/exclusion and diagnostic criteria appropriate
Interventions	Intervention category: anticonvulsants + dietary supplements Group 1: (n = 20) 600 mg daily alpha lipoic acid ‐ for 2 months (Ciagen 600, produced by Craveri). Supplement includes excipients (sodium lauryl sulfate 30.00 mg; 329.50 mg lactose monohydrate; cornstarch 115.50 mg; croscarmellose sodium ( AC‐DI‐SOL) 72.00 mg; colloidal silicon dioxide (Aerosil 200) 5.50 mg; 35.00 mg magnesium stearate; povidone K‐30 12.50 mg; methocel E15 25.08 mg; 10.05 mg lactose monohydrate; 0.39 mg titanium dioxide; polyethylene glycol 6000 0.78 mg; yellow iron oxide 3.68 mg) Group 2: (n = 20) 300 mg daily gabapentin (GABA) Group 3: (n = 20) 600 mg daily alpha lipoic acid (ALA) + 300 mg daily gabapentin (GABA) Group 4: (n = 60) 100 mg daily starch and cellulose placebo
Outcomes	Number of sites affected ‐ "evaluated the presence of burning through a numerical scale especially created for this work, describing the burning from 0 to 4, where the 0 value corresponded to the absence of burning, the 1 value to the presence of burning in a single area of the tongue, the 2 value to two distinctive areas (tongue and gums, tongue and lips or tongue and palate), the 3 value to three areas and the 4 value corresponded to burning spread throughout the mouth. This specific designed scale, which considered the geographical distribution of burning in different areas of the mouth, allowed us to distinguish improvements or deteriorations of burning sensation in the various assessments" "Evaluation of the effects… the day before the start of treatment and thirty and sixty days, respectively. To evaluate the changes that occurred with the taking of the different drugs, it was established that the improvements (positive changes) involved the passage of a certain level or numerical category of burning to a lower one, the deteriorations (negative changes) involved an increase of a certain level of burning to a higher one and the total resolution indicated the total absence of burning, that is to say the transition from any higher value to zero value. In this way four categories were obtained for the analysis of the 5 results: Category 1: with negative changes (deterioration), Category 2: no changes; Category 3: with positive changes (improvements) and Category 4: with total recovery" Change in quality of life (QoL) was assessed at baseline only (consequently unable to assess change) by use of 2 surrogate (measuring anxiety and depression) scales: ‐ anxiety: Hamilton Anxiety Rating Scale (HARS) ‐ baseline only; and ‐ anxiety/depression: Hospital Anxiety and Depression (HAD) scale ‐ baseline only
Source of funding	Not reported
Notes	The dosage of gabapentin given is not therapeutic as an anticonvulsant Outcome evaluation was based on the numerical scale which was based on the number of sites affected. Thus only 1 (not validated) outcome measure was used in the study SES: not reported Conflict of interests: not reported Data analysis: ITT
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quotation: "The 120 patients were randomly divided into 6 groups of 20 members each" Comment: unclear as to how the randomisation was conducted
Allocation concealment (selection bias)	Low risk	Quotations: "The support staff of our service made a draw with 6 balls to link the groups with the cycles of treatment"; "allocation that was always masked to both patients and researcher" Comment: assumption that "support staff" were independent to investigators – seems unlikely that allocation concealment would be compromised
Blinding (performance bias and detection bias) Blinding of participants	High risk	Quotation: "through the use of capsules of similar size and appearance so that just the support staff was the one who recorded the information until the end of the treatment (blind)" Comment: "similar size and appearance" ‐ unlikely that formulations were identical, as weights were different and they may not look the same. Unclear how preparations were provided to participants and in what packaging. Placebo only group involved 1 tablet while group 3 had 2 tablets to take – hence unblinding group 3 that they were on the combined treatment arm
Blinding (performance bias and detection bias) Blinding of outcome assessors	Low risk	Comment: patient‐reported outcomes only
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: no dropouts were reported
Selective reporting (reporting bias)	High risk	Quotation: "adverse effects that appeared were very mild" Comment: clarification sought from contact author who provided adverse effect data. No baseline characteristics were presented for each arm of the study. Baseline anxiety and depression was presented for all participants but not for each arm. Outcome data were not presented in the prespecified way – graphs were presented which were difficult to interpret as the actual data values were not presented. The study authors combined "positive changes" with "full recovery" data for the active arms and "no change" with "worsened" data with the placebo – these outcomes should have been presented separately as combining them is very misleading. No useable data were presented for month 1, only month 2. Clarification sought from contact author who provided missing data from month 1
Other bias	Low risk	Comment: possible selection bias caused by exclusion criteria – "patients using more than 3 systemically daily drugs, those ones taking psychotropic and antihypertensives drugs as well as patients with serious psychiatric conditions previously diagnosed were excluded"