B104.

Methods	Double‐blinded, 3 arm parallel‐group randomised controlled trial 18 week treatment
Participants	Setting: Europe and Canada; 11 centres, between January 1993 and September 1993 Sample size: 114 participants Age: range years, mean age years Inclusion criteria: Age up to 90 years Diagnosis of mild to moderate dementia according to DSM III‐R, and probable AD according to NINCDS‐ADRDA criteria MMSE score between 12 and 26 points Exclusion criteria: concomitant conditions or medications that may confound assessment of dementia; current diagnosis or history of significant medical, neurological or psychiatric disorder
Interventions	1. Rivastigmine: 6 to 12 mg/day divided into 2 doses 2. Rivastigmine: 6 to 12 mg/day divided into 3 doses 3. Placebo (identical looking) twice daily Titration to a maximum dose of 12 mg/day or maximum tolerated dose during weeks 1 to 10, followed by 8 weeks maintenance of dose
Outcomes	Measured at 18 weeks 1. Cognitive function Alzheimer's Disease Assessment Scale (ADAS‐Cog) Wechsler Memory Scale (WMS) (digit span, delayed recall, word fluency) 2. Activities of daily living Nurses' Observation Scale for Geriatric Patients (NOSGER) 3.Global evaluation (physician rated) CIBIC plus 4. Behavioural symptoms Nurses' Observation Scale for Geriatric Patients (NOSGER) 5. Incidence of adverse events 6. Discontinuation withdrawal due to adverse events
Source of funding	Novartis Pharma Ltd
Declaration of interest	Sponsored by Novartis Pharma Ltd
Notes	Main hypothesis: to investigate tolerability of rivastigmine 10 week titration phase to a maximum of 12 mg daily or maximum tolerated dose, then 8 weeks maintenance phase
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Not described "patients were randomly assigned to one of three treatment groups"
Allocation concealment (selection bias)	Low risk	Not described Comment: likely to be low risk since this is a large multicentre trial
Blinding of participants and personnel (performance bias) All outcomes	High risk	Identifical placebo used, taken twice daily Comment: this effectively unblinded the group assigned to three times daily regimen
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Clinician who rated the CIBIC did not have access to baseline results and psychometric tests and also did not ask questions; however, it was unclear how effective this was
Incomplete outcome data (attrition bias) All outcomes	High risk	A total of 85/114 (75%) completed the study. Numbers of participants who completed the study were different between groups: 92% for placebo, 89% for three times daily group and 78% for twice daily group. The efficacy analysis was conducted only in "valid" patients, defined as all those patients who had completed the study according to protocol
Selective reporting (reporting bias)	Low risk	Outcomes listed in protocol reported