Freeman 1992.

Methods	Randomised double‐blind trial
Participants	Number of total participants/type of patients: 27 participants were randomised. Inpatient study Inclusion criteria: Sex: male and female Age: no information provided Diagnosis: meeting DSM‐III‐R criteria for a manic episode Other: able to give informed consent Exclusion criteria: 1. Abnormal EEGs, LFTs, TFTs, haematological findings or positive urine drug screens 2. History of drug or alcohol abuse within the previous 12 months or before the onset of major affective episode 3. Focal neurological abnormalities
Interventions	Location: USA, Houston Study duration: Not stated Treatment groups: 1. Valproate group: 1500 mg/day for first week, 2250 mg/day for second week and 3000 mg/day for third week 2. Lithium started at 0.5 mEq/kg a day, increased to maximum of 1800 mg/day or 1.5 mmol/L Concomitant medications: Rescue medication, including chloral hydrate and lorazepam, was allowed for extreme behavioural problems not responding to non‐pharmacological interventions Duration of trial: 21 days Randomisation: 14 patients randomised to valproate 13 patients randomised to lithium
Outcomes	Outcomes: 1. Changes on the YMRS scores as assessed by SADS‐C 2. GAS 3. BPRS 4. Study withdrawals were reported
Funding	Not stated
Conflict of interest	Not stated
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients were assigned randomly to treatment with lithium or valproate in a double‐blind fashion" Comment: Although participants were randomly assigned, there is no detail on the method used in the text. We were therefore unable to evaluate bias. Comment: We contacted the authors but did not receive a reply.
Allocation concealment (selection bias)	Unclear risk	Quote: "Patients were assigned randomly to treatment with lithium or valproate in a double‐blind fashion" Comment: There is no detail on any procedures present to conceal allocation. We were therefore unable to evaluate bias. Comment: We contacted the authors but did not receive a reply.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Double‐blind" Quote: "Plasma drug levels were monitored on days 5‐7 of each week, the results were sent to the research pharmacist, and the dose was adjusted by a non‐blinded, non‐treating physician." Comment: Medication was given in liquid form and always diluted to a total volume of 30 ml. Procedures in place to ensure blinding seem reasonable.
Blinding of outcome assessment (detection bias) Efficacy	Unclear risk	Quote: "Double‐blind" Quote: "the dose was adjusted by a non‐blinded, non‐treating physician. The SADS‐C, the GAS and the BPRS were administered at the time of dose adjustments." Comment: The dose was adjusted by a non‐blinded physician and the rating scales were administered at the time of dose adjustments. This implies that the outcomes may have been evaluated by an unblinded clinician. However, there is no detail in text about who administered the rating scales. On contacting the authors for clarification, we received no reply. We are therefore unable to assess bias.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: In all groups dropout rate was below 30%.
Selective reporting (reporting bias)	Unclear risk	Comment: No pre‐published protocol found. We are therefore unable to assess the risk.
Other bias	Low risk	Comment: None identified.