University of Sussex
HTA-SYMBAD StudY of Mirtazapine or carBamazepine for Agitation in Dementia
Alzheimer's disease & other dementias
DESIGN Pragmatic multi-centre double-blind placebo-controlled RCT of the clinical and cost-effectiveness of mirtazapine or carbamazepine (all with usual care) at 6/12/26/52w on agitated behaviours in dementia. SETTING Community in 8 regional centres in England. TARGET POPULATION People with Alzheimer’s Disease (AD) and agitated behaviours >4 weeks duration which have not responded to non-drug treatment. INCLUSION/EXCLUSION CRITERIA Pragmatic trial, minimising exclusions to maximise generalisability. Inclusion: probable or possible NINDS/ADRDA Alzheimer’s Disease; agitated behaviours duration >4 weeks which have not responded to the DH/AS algorithm; Cohen Mansfield Agitation Inventory (CMAI) score 45+. Exclusion: current antidepressants, anticonvulsants or antipsychotics; case too critical for randomisation; no informant to give collateral information. HEALTH TECHNOLOGIES BEING ASSESSED Experimental: Either (i) mirtazapine or (ii) carbamazepine, both with usual clinical care. Control: placebo with usual clinical care. MEASUREMENT OF COSTS AND OUTCOMES Patients and carers interviewed by research worker (RW) at 0,6&12w: Primary: (i) Agitated behaviours in dementia CMAI. Secondary – person with dementia: (ii) Costs Client Service Receipt Inventory (CSRI) and nationally representative unit costs;(iii) Quality of life – disease-specific-DEMQOL-Proxy and generic- EQ-5D; (iv) cognitive impairment sMMSE; (v) Broad Behavioural and Psychological Symptoms in Dementia (BPSD)- NPI; (vi) adherence, adverse events, concomitant medication/changes, withdrawal. Secondary carer: (vii) carer mental health GHQ-12; (viii) carer quality of life EQ-5D; (ix) carer burden short form Zarit CBI; (x) carer sleep. Long term outcomes: telephone interviews at 26&52w measuring CMAI, treatment status, institutionalisation and death. Primary analyses will compare mirtazapine or carbamazepine with placebo on CMAI score change 12 weeks post randomisation. Analyses of clinical effectiveness will be pragmatic using linear mixed modelling, based on an ITT sample on all available follow-up data from all randomized patients controlling for baseline levels of agitation and centre. Secondary analyses will compare all groups and cost-effectiveness. SAMPLE SIZE An overall sample of 400 (randomised 1:1:1) provides 90% power using 2-sided 5% significance tests to detect a drug v placebo mean difference in CMAI score at 12 weeks of 6 equating to an effect size of d=0.4 (assuming a common standard deviation of 15) or a clinically significant 30% decrease in CMAI. With a realistic 15% attrition, we require a target sample of 470 (around 156 per arm). PROJECT TIMETABLES INCLUDING RECRUITMENT RATE 8 centres will each recruit 59 patients over 24 months. -6 to 0m trial approvals sought 1-6m trial systems set up 5-6m training RWs, centres set up and priming -6-6m production of trial medication and matching placebo 7-33m recruitment of patients, randomisation 8-36m follow-up interviews (6 and 12w), report preparation 30-39m final analyses and report writing, study closeout 10-45m collection of supplementary long term (24/52w) outcome data 40-48m analysis and report preparations on long term outcome data] EXPERTISE IN TEAM A broad, experienced multidisciplinary team experienced in dementia, clinical trials in dementia, statistics, health economics, implementation, policy and service user/carer perspectives.