Lusznat 1988.

Methods	Study design: parallel, double‐blind randomised controlled study
Participants	Diagnosis: presumptive diagnosis of mania or hypomania Method of diagnosis: assessed by a Research Registrar, scored 10 or more on the BRMS Age: aged 17‐65 years (no breakdown provided) Sex: women; men. Location: Department of Psychiatry, Southampton. Co‐morbidities: not described Adjunctive therapy: no Adjunctive medication: "In the setting of an ordinary acute admission unit it proved impossible to manage severely manic participants without additional medication, so nearly all of then (n = 52) received variable amounts of neuroleptics during the acute study. It had initially been hoped to give chlorpromazine as the only rescue medication of this type, but several participants had known chlorpromazine sensitivity or were known to respond better to other neuroleptics such as haloperidol, son on ethical grounds this intended restriction had to be dropped."
Interventions	Participants were randomly assigned to either: Experimental arm ‐ lithium N = 27 Duration: 6 weeks Treatment protocol: "Patients were started on study medication as soon as their initial assessment, including DST, had been completed. They were randomly assigned to be given either tablets containing 200 mg of carbamazepine plus placebo lithium carbonate tablets, or tablets containing 400 mg lithium plus placebo carbamazepine tablets. This ‘double dummy" technique was used since it had proved impossible to obtain tablets containing lithium and tablets containing carbamazepine that looked alike. Medication was started in a dose of one active tablet plus one placebo tablet daily and was increased by one tablet of each type every second day until a serum carbamazepine level 0.6‐1.2 mg per 100ml was attained, or a serum lithium concentration of 0.6‐1.4 mmol/L. Serum drug levels were estimated at each follow‐up assessment in order to check on correctness of dosage and compliance. Medication and day‐to‐day aspects of the clinical care of each patient were managed by the Research registrar, who carefully concealed serum drug estimations from those making follow‐up ratings. Therapist/face‐to‐face contact: not described Comparator arm ‐ carbamazepine N = 27 Duration: 6 weeks Treatment protocol: Patients were started on study medication as soon as their initial assessment, including DST, had been completed. They were randomly assigned to be given either tablets containing 200 mg of carbamazepine plus placebo lithium carbonate tablets, or tablets containing 400 mg lithium plus placebo carbamazepine tablets. This ‘double dummy' technique was used since it had proved impossible to obtain tablets containing lithium and tablets containing carbamazepine that looked alike. Medication was started in a dose of one active tablet plus one placebo tablet daily and was increased by one tablet of each type every second day until a serum carbamazepine level 0.6‐1.2 mg per 100 mL was attained, or a serum lithium concentration of 0.6‐1.4 mmol/L. Serum drug levels were estimated at each follow‐up assessment in order to check on correctness of dosage and compliance. Medication and day‐to‐day aspects of the clinical care of each patient were managed by the Research registrar, who carefully concealed serum drug estimations from those making follow‐up ratings." Therapist/face‐to‐face contact: not described
Outcomes	Timepoints for assessment: Primary outcome: Change in BRMS Secondary outcome: Withdrawal Hamilton depression score Side effect rating score Length of admission
Notes	Date of study: not given Funding source: RL and PM were supported by a grant from Ciba‐Geigy. Declarations of interest among the primary researchers: none given
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias) All outcomes	Low risk	This ‘double dummy" technique Follow‐up assessments were made by a rater blind to the participant's study medication. Medication and day‐to‐day aspects of the clinical care of each participant were managed by the Research registrar, who carefully concealed serum drug estimations from those making follow‐up ratings.
Incomplete outcome data (attrition bias) All outcomes	Low risk	10/54 withdrew
Selective reporting (reporting bias)	High risk	Poor reporting
Other bias	Low risk	None identified